Ludwig Wittgenstein wrote his Remarks on the Philosophy of Psychology about thirty-five years ago. Only now have they been published, rather late in a long sequence of posthumous books. The two volumes are successive attempts to sort out the same ideas. He was never fully satisfied by them, but they may well turn out to be his most enduring secondary work, fair companions to the only books that Wittgenstein did cast into final form: Tractatus Logico-Philosophicus, finished in 1918, and Part I of Philosophical Investigations, done by 1945.

Wittgenstein’s reflections on the human mind are central to his later philosophy. He pairs off quite nicely with Descartes, his predecessor by exactly three centuries, and the founder of philosophical psychology. Although great philosophers never come two of a kind, these men are strikingly alike. The similarity may seem a bit surprising, since Wittgenstein is often presented as the very opposite of Descartes, and even as the man who brought the Cartesian era of philosophy to an end.

Each man chose to be an émigré. The Viennese Wittgenstein made his base at Cambridge University while the French Descartes worked in Holland. Each lived in the midst of a foreign language but wrote most of his thoughts in his native tongue. Each soldiered in yet other countries. Both not only settled abroad but also set off for very strange parts. Wittgenstein traveled to the Soviet Union in 1935, possibly intending to become a doctor in Siberia. Descartes finally accepted the bidding of Queen Christina and went off to Sweden to die of the cold.

Both philosophers were obsessive about their work, holding it back for years. Each could be furious if any of it was leaked prematurely to the public. Their eccentricities were legion, but each had a personality that dominated, nay, obsessed, close friends. Both men told us never to hurry with their work. The few pages of a Cartesian Meditation are to be reread on successive days. Only when you have made them your own should you move on to the next thought. Wittgenstein: “My work must be read slowly.” He ensured that if you do read it, you do so slowly. Eclectics can dip into this or that choice fragment speedily enough, but if you read the matter systematically you have to take your time.

Wittgenstein wrote in numbered paragraphs, a few of which pursue the same topic and then may abruptly switch to something different. Even a single paragraph may be a series of quick exchanges between Wittgenstein and an interlocutor. A topic that has been dropped will reappear many paragraphs later. Strange possibilities are described and the same phenomenon will be held up again and again to be glimpsed from new perspectives. This style fits the content, for Wittgenstein’s thought keeps on illustrating related themes from successive vantage points, shooting off, recollecting, transcending, backsliding. It is not unlike a mind that talks to itself in a half dozen different conversations at once, but the successive paragraphs are the subtly organized, intensely disciplined product of unending toil. Both Descartes and Wittgenstein are remarkably graceful authors, graceful not only in the seemingly relaxed flow of words, but also in the meticulous sequencing of ideas that lies beneath the charm of the sentences.

It is a commonplace that the “philosophical psychology” of Descartes and that of Wittgenstein are totally different, the one the very negation of the other. That is half right and half wrong. What’s right is that Descartes starts from inside himself while Wittgenstein begins in the world of human communication. The Cartesian philosophy says that I best know my own mind. All my knowledge is based upon my private experiences, sensations, and thoughts. Thought is the movement of ideas in the mind. Wittgenstein holds in contrast that shared practices, actions, reactions, and interactions among people provide the foothold upon which all such self-description of our mental life must rest. Language is first of all public and firmly rooted in what we do together. He rightly resented casual readers who would dismiss him as a behaviorist:

Then is it misleading [Wittgenstein asks himself] to speak of man’s soul, or of his spirit? So little misleading, that it is quite intelligible if I say “My soul is tired, not just my mind.” But don’t you at least say that everything that can be expressed by means of the word “soul,” can also be expressed somehow by means of words for the corporeal? I do not say that. But if it were so—what would it amount to? For the words, and also what we point to in explaining them, are nothing but instruments, and everything depends on their use.

To ask whether Descartes believes in the human soul, while Wittgenstein does not, is simply to put a bad question. “Do I believe in a soul in someone else, when I look into his eyes with astonishment and delight?” Wittgenstein thought that it is not a question of belief founded on evidence at all.


Descartes held that mind and body are distinct substances and wondered how they interact. That doctrine, called dualism, has obsessed much Western philosophy. It will be said that Wittgenstein was no dualist, thus emphasizing his fundamental difference from his predecessor. I do not agree. The contrast is, I think, wrongly understood. Wittgenstein certainly did not hold that mind and body are two “substances,” or that “mind” names a special kind of thing. But in many essentials, he is just as much a dualist as Descartes. Both hold that psychology requires forms of description and methodology quite different from those called for in natural science. Reflection on thinking is not remotely like the study of the inhuman world of spatial, mechanical objects.

Descartes took a word—cogitare, penser, “think”—and gave it an extended sense in which it captured all the disparate but roughly mental activities such as hoping and remembering and seeing and hurting. Where Descartes unifies, Wittgenstein mercilessly divides. Something different may have to be said about each mentalistic verb. Hence there is a long sequence of items that jut out in an index to the two volumes of Philosophy of Psychology: believing, calculating, expecting, experiencing, feeling, intending, and so on. Precisely such a list might be used to elucidate Descartes’s portmanteau term “thinking.” Both philosophers understood that descriptions of these items will be unrelated to anything that goes on in the material, space-occupying organs of the body such as the brain. At the same time, Wittgenstein rejects the very possibility of any doctrine about Cartesian thought-in-general.

Before looking at Wittgenstein’s descriptions of the mental, one needs to place them within the rest of his work. Three decades separate the completion of his Tractatus and the final form of Part I of Philosophical Investigations. These masterpieces are usually said to represent his early and his later philosophy. Both are worth calling philosophies of language. In the Cartesian epoch, language had been a wonderful system of signs for conveying thoughts from one mind to another, but language was always secondary to ideas in the mind. There came at last the strange reversal; language became a necessarily public institution within which human selves are formed and by which people constitute the world they live in. The switch from the primacy of private thought to that of public discourse is not the work of Wittgenstein. In 1868, C.S. Peirce, founder of pragmatism, had published in a St. Louis philosophy magazine the remarkable sentence: “My language is the sum total of myself.” That was twenty-one years before Wittgenstein was born, and the same thoughts were circulating elsewhere about that time.

Wittgenstein no more invented the idea of human beings and their world being constituted in language-than Descartes invented the mind-body problem. Both philosophers are historical personages whose writings turn the claptrap of their time into monuments. Insofar as Wittgenstein had an earlier and a later philosophy, his monuments face away from each other. The Tractatus had a vision of a single role for language, while the Investigations tells of innumerable language games each of which is embedded in its own web of activities.

The Tractatus is written as if language had but one function: representing the world. That creates a problem to which the book addresses itself. How is it possible to represent a nonlinguistic world in words? The opening sentences begin the answer. “The world is the totality of facts, not of things.” When the penknife is to the left of the snuffbox, we tend to think of two things, the knife and the box. We think of a world made up of things like penknives that can be arranged in various ways. Not so, says Wittgenstein: the world consists simply of a set of facts, like the fact that the knife is to the left of the box. This is not to deny that there are things, such as penknives. It says only that the totality of facts is all there is to the world. Once that totality is given, you add nothing more by saying, “and there are things too, such as snuffboxes.” This idea of the world begins to explain how representative language is possible. Propositions represent the world by picturing the structure of the facts. This idea has been called “the picture of theory of meaning.”

A theory of language as essentially representative excludes an enormous amount of discourse. Much of life most dear to us, including beauty, philosophy, and moral worth, has nothing to do with representation. Although values can be lived, acted out, or displayed, they cannot, on Wittgenstein’s early account, literally be stated. This is not for the simple-minded reason later expounded by positivists, that values are mere expressions of feeling, and thus neither true nor false. On the contrary, we can represent facts about feeling just as well as any other facts, and there are truths and falsehoods about the inner world just as much as the external one. Wittgenstein’s difficulty is that neither a value nor a philosophical thesis is a representation at all, and so is not something that can be “said.” It is only something that can be “shown” by saying or doing something else. The Tractatus is written as a sequence of numbered propositions. It concludes by showing its own impossibility. This very philosophy cannot be a series of propositions at all. It, too, can at best be shown. The book ends by recommending silence.


Among the many themes in that strange and powerful book, I would here emphasize only its unified conception of the role of language: representation of facts by propositions. That vision is abandoned in the later philosophy. Wittgenstein came to see that language is not one monolithic system of representations for picturing reality. Instead it is composed of myriad fragments that loosely overlap and intersect. Most of these are not used to represent anything. We are told to look at little bits of real or invented discourse to see what nonlinguistic activity—what social context or use—must accompany each one in order for it to make sense. A case in point of this approach is the way in which Remarks on the Philosophy of Psychology replaces the Cartesian concept of “thinking” by detailed study of lots of different mentalistic verbs—introspecting, calculating, remembering, intending—each of which demands its own social setting.

Philosophical Investigations begins by imagining a game in which builders call out a few words such as “block,” “pillar,” “slab,” or “beam,” thereby indicating to their helpers what materials they want where. Actions and words are formed in a language game from which the words take their meaning. Later, we learn that what we say about knowing and feeling and pain are likewise a mélange of small language games, each with its own family of preconditions and applications. Collectively, these display how wrong it is to seek a single model for language. They show us the disunity of language.

The game of the builders began an attack on the idea that words work chiefly as names. We are led through many related themes, including the famous maxim, “Don’t ask for the meaning, ask for the use.” One point of that slogan is: don’t ask for objects that might serve as meanings of our words. Meanings are not objects that are saved by words. Consider what we do with words, not what they represent. That comes out even in my long Philosophy of Psychology quotation about the soul, above. It is part of the reason that Wittgenstein could doubt that we use the words “soul” or “mind” to name a thing.

Philosophical Investigations is also famous for its “private language argument,” that there cannot be a language that is in principle inaccessible to anyone else. There can, for example, be no language with names for just my sensations. A word like “pain” does not get its use by first naming something that we feel, and then telling others about it. Instead, it is necessarily embedded in various kinds of things we do in connection with being hurt. That is not to say that there is no pain without behavior, or that pain is a kind of behavior. But the idea of stoically concealed pain is nested in and parasitic upon more public ways of talking about pain—evincing it, wincing, trying to comfort the victim or relieve the suffering. We must particularly resist the idea of a private object (pain) that is named by the word “pain.” Remarks on the Philosophy of Psychology imaginatively applies this line of thought to all kinds of mental goings-on. For example, knowing is not in general a state named by the “know”-words, nor does “I know” have the same family of uses as “he knows.” Again, Wittgenstein asks, “Why can a dog feel fear but not remorse?” Not because there is something in the dog’s mind correctly named by the word “fear” while there is nothing in the dog’s mind to which the word “remorse” applies.

Many philosophers nowadays make eclectic use of these and many more ideas. Wittgenstein may, however, have hoped that his thoughts would one day be understood less for his theses than for his method and attitudes. He had a gloomy view of scientific culture and a deep pessimism about the possibility of his work’s being understood in the darkness of our times. Maybe he was right. After his death, his thought was briefly quite fashionable among leading philosophy teachers, but now it is not. It continues to attract young people, and the name “Wittgenstein” is often enough invoked in high culture. Yet most serious philosophers seem to have put him aside, tending to find his concentration on diverse cases unproductive for the systematic analyses that philosophers have traditionally preferred. The other day, I heard a distinguished older philosopher, no “Wittgensteinian” herself, say crossly: “Really, it is astonishing. It is just as if Wittgenstein had never lived!”

Wittgenstein carefully preserved his work in tin boxes. His literary executors have edited it piece by piece in what sometimes seems haphazard order. They will be made fun of (or insulted) when a later generation prepares an Academy Edition of the Great Man’s Work, but the somewhat personal style of editing may well coincide with the author’s own intentions. At any rate, chunks of Wittgenstein’s writings have appeared every year or so since his death in 1951. They have the effect of time-release capsules. This is salutary if you think, as I do, that many of us skim his words and forget.

He filled notebook after notebook. Often he clipped out the paragraphs and rearranged them. The same paragraph can appear in different settings. When he was confident of an arrangement, he would write it all down again and then dictate it to a typist. Volume I of these Remarks was dictated in the fall of 1947, and Volume II a year later. A third, later, and shorter survey was used by the editors to form a sixty-page “Part II” of Philosophical Investigations, but aside from a remarkable discussion of “seeing as,” it has neither the tightly knit craftsmanship of Part I nor the wealth of examples discussed in the two previous but newly published typescripts under review.

What is Wittgenstein’s philosophical psychology? We can see at once a number of things that it is not. It is not experimental psychology. A famous remark that concludes Philosophical Investigations runs thus: “The confusion and barrenness of psychology is not to be explained by calling it a young science—the existence of experimental methods makes us think we have the means of solving the problems that trouble us; though problems and methods pass one another by.”

Philosophical psychology is not introspection, whose noblest practitioner was William James. James is the only psychologist (besides some of the Gestalt people) to whom Wittgenstein regularly alludes. The vigor of James’s writing is used to make plain the bizarre paths into which we are led by the very idea of a faculty of introspective knowledge. The danger here lies in postulating that there is an exclusively subjective means of gaining self-knowledge.

Philosophical psychology is certainly no cousin of psychoanalysis. “Freud’s fanciful pseudo-explanations (precisely because they are so brilliant) perform a disservice. Now any ass has these pictures available to use in ‘explaining’ symptoms of an illness.”

Philosophical psychology is not cognitive psychology, which seeks models of what goes on in the brain when we think, know, talk, perceive. Cognitive psychology nowadays most often means the study of how mental representations are connected with cognitive functions in the brain. Wittgenstein would have been quite hostile to this. “I don’t care whether his brain goes green or red when he thinks of that.”

Even when cognitive psychology does not postulate representations in the brain, it still seeks explanations of what is going on in the head when we act or think or talk. That is just what Wittgenstein resists. “People who are constantly asking ‘why’ are like tourists who stand in front of a building reading Baedekker and are so busy reading the history of its construction that they are prevented from seeing the building.” “The tendency to explain instead of merely describing” gives only bad philosophy. Describing: that is what he would like to be doing. “Not to explain, but to accept the psychological phenomenon—that is what is so difficult.”

Mere description is so difficult because one believes that one needs to fill out the facts in order to understand them. It is as if one saw a screen with scattered colour-patches and said: the way they are here, they are unintelligible; they only make sense when one completes them into a shape.—Whereas I want to say: Here is the whole. (If you complete it, you falsify it.)

I find this conception of philosophical psychology oddly congenial to Descartes. That’s odd because Descartes was a great explainer. He did make explanatory models of how the body works, of the movements of the blood, even in the brain. He was fabled to have made a human robot, thrown overboard during a storm on the last voyage to Sweden because the sailors thought it was a Jonah. He was, then, intrigued by speculations about how the body works. The mind, however, is something else. He thought it is not in the same ballpark, or in another ballpark, either: mind and body are about as different as the Oakland A’s and the letter A. That is, they are not open to remotely the same styles of description, and whereas the Oakland A’s often need explanation, it is obscure what it would mean to try to explain the letter A.

Descartes was very cagey about the relationship of mind to body. He did not like the ancient formulation that “I am in my body as a pilot in his ship.” He wrote that instead “I am most tightly bound to my body, and as it were intermingled with it, so that it and I form a unit.” An interviewer asked what he meant by that. Descartes tartly replied: “It is very difficult to explain; but our experience is enough, since it is so clear on this point that it cannot be gainsaid. This is evident in the case of the feelings and so on.” It is possible to put some of Descartes’s descriptions of love, yearning, and desire alongside those of Wittgenstein and not quite know whose is which. “I find my arms reach out as if to embrace something; my soul is thereby moved to join itself willingly to this object.” “One really always thinks of the stance of the body towards an object. The stance of the soul to the image is just what one might represent in a picture: the man’s soul, as it leans with gestures of longing towards the picture (an actual picture) of an object.” The first remark was written in 1647, the second in 1947.

It seems to me that Descartes wants to say not only that all this is “very difficult to explain,” but also that one ought not to try to explain the way in which events in the brain are associated with feelings. There is a whole domain of descriptions about how one feels thirsty, sees trees, grieves, and so forth, where one would be making some sort of conceptual error to ask for explanations of a materialistic sort. This is the important sense in which Wittgenstein and Descartes are equally dualistic.

I would never urge that a person cannot learn from both philosophical and cognitive psychology. I say only that they are different enterprises, of which only the latter could ever be explanatory. Human interests are commonly so narrow that you won’t find many people who, like Descartes, could take pleasure and profit from both kinds of endeavor. But we must not let differences in taste or in life-projects make us think that one of these enterprises is right-minded while the other is wrong-headed.

We must not imagine that Wittgenstein, in rejecting explanation, randomly describes a “screen” of scattered mental events, refusing to “complete” it for fear that he might “falsify it.” His paragraphs are never unmotivated. They turn out always to be directed at “the problems that trouble us.” He means conceptual problems and confusions. I shall take only one example, a sustained discussion of “imaging” in Volume II, pages 13-28. “Imaging” is the best the English translator can do for the activity of forming images. The word “image” translates Vorstellung, as central a term of art as exists in the whole of German philosophy. It is a word variously translated as “representation” or “idea,” but Wittgenstein uses it more like the English translation as simply “image.” Indeed the opening shot is fired not at a German thinker but at a Scotch one. “Auditory images, visual images—how are they to be distinguished from sensations? Not by vivacity.”

The target is David Hume, who, like Descartes, would have liked to build up all knowledge from our immediate mental experiences. He badly wanted a way to tell mere ideas (which would include visual images) from sensations and sense impressions. “They differ,” Hume wrote, “only in their different degrees of force and vivacity.” To call sensations more lively than images is to suppose that they are the same sort of thing, all bits of that global Cartesian “thought.” That is so wrong that we should not even assert that seeing and imaging are different phenomena. Better to say that neither is a phenomenon, and that “image” and “sense-impression” do not name kinds of entities.

Instead consider what may be essential to our usage of words such as “image” and “see.” I call Wittgenstein a philosopher of language, but he is not a “linguistic philosopher” of the sort once dominant in Oxford, and who studied, often with great panache, the actual uses and nuances of English words and phrases. Wittgenstein directs us often to almost imaginary language games to get at what is essential to this or that concept. So imagine two games. In one of them people say, “Look at that figure!” perhaps pointing to a cube in a book of geometry. In the other they say, “Imagine that figure!” One game goes along with other instructions such as “Look over there!”—said holding up the book. The other might have, “Shut your eyes.” The verb “see” will have a role in one, but need not in the other. (“A language game comprises the use of several words.” But only several.)

We are not to think of seeing and imaging as being different phenomena in themselves, but as verbs distinguished by the ways in which they “relate to a host of important kinds of human behavior, to the phenomena of life.” The phenomena are not the seeing and the imaging but the practices in which they are embedded. Wittgenstein’s examples include: “closing one’s eyes to form an image, straining to see something, following a moving object with the eyes.”

In the mere fifteen pages devoted to this topic, numerous considerations weave in and out. Imaging, for example, seems subject to the will while seeing is not. We can call up, form, or banish images, but seeing is not like that. Is that the real difference between the two kinds of thing? No. Wittgenstein tells a story where we might say that sense-impressions are as subject to the will as are images, but that does not reduce them to the same kind of thing. We are led back to the more general point: “With the sentence ‘images are voluntary, sensations are not,’ one differentiates not between sensations and images, but rather between the language games in which we deal with these concepts.”

Wittgenstein is very thorough. He gently probes the idea that still lures many a person: a sense-impression and an image might have the same “experiential content.” He does not sarcastically urge (as did the late Oxford philosopher J.L. Austin) that an “image” and a “sense-impression” of the Taj Mahal would never look the same. On the contrary. Could I not form a detailed image of a face, and later see exactly that face in real life? Would not my earlier image have the same “experiential content” as my later seeing? (Wittgenstein often hands his interlocutor a better example than we read in his real-life opponents.) He holds that “one can not say the two are not the same on the grounds that an image and an impression never look alike.” We might even draw just one picture to illustrate what first one imaged and later saw. If you like, you can call that very picture—the one you drew on the paper—the “experiential content” of both events. “Only one mustn’t allow oneself to be deceived by the myth of the inner picture.” It was of course just that myth that forced Hume to distinguish images and sense-impressions in terms of their vivacity.

“In philosophizing,” he wrote, “we may not terminate a disease of thought. It must run its natural course, and slow cure is all important.” My speedy snippets from a few pages are no substitute for slow reading, especially since they cannot adequately illustrate Wittgenstein’s thesis of disunity—that we engage in endless loosely overlapping language games. Once we have described imaging, other stories must be told for other mental activities such as hoping, wishing, describing, hearing, etc. His remarks about “cures,” alas, are too often overemphasized. We are sometimes told of a Wittgenstein as therapist who would gladly bring philosophy to an end. He certainly did think that we are prone to certain kinds of conceptual mistake, and he detested philosophers who feel good being quick, clever, and flashy, three arguments on the blackboard and a hundred more in the pockets. From none of that does it follow that Wittgenstein either ends philosophy or stands outside it. My little example about imaging reminds us how much he stands in our own tradition. He himself did not quit, but kept on filling up notebooks on new topics until he was on his deathbed. His last written thoughts were about concepts of color. They are published in a short book, Remarks on Colour, and are very much an instance of work in progress.

Emotions, intentions, and ever so many more aspects of being alive are illumined in these two volumes of Remarks. Why, he asks, may sorrow and grief be so naturally described as gray, or as a heavy cloud descending from the sky? Such an array of particulars does not, however, stay put as a row of isolated insights. There are general implications. For example, many of our leading philosophers at present debate the merits of this or that proposed systematic and general theory of meaning for a natural language. This takes many forms. There is no current work by an American philosopher that I more admire than that of Donald Davidson, at Berkeley.* His account of meaning, truth, understanding, and translation, combined with his theory of human action, has an enormous range of consequences. It starts with a precise version of the idea that language, action, and belief are to be approached as a whole. Our evidence may be piecemeal, but our interpretation of another person’s speech is a theory about one unified thing.

Michael Dummett at Oxford has for some years been mounting an “atomistic” attack against Davidson, holding that we confer meanings upon our words almost one by one, situation by situation. Even he, although once strongly influenced by Wittgenstein, has fallen prey to the idea that there could be such a thing as a “theory of meaning” in general, and that there could be a general theory of the conditions under which what we say is true. I think that Wittgenstein’s theme of the disunity of life, reason, and language runs counter even to Dummett, let alone Davidson. Neither of these men, nor their students, will be much moved by the disunity aspect of Remarks on the Philosophy of Psychology, but in the end Wittgenstein’s little guerrilla army of unlike examples may begin to tell against the big guns.

More immediately, the disunity thesis helps vindicate the claim that there is a legitimate project to be called philosophical as opposed to cognitive psychology. It even connects with a phenomenon familiar to ethnographers, namely that once you have ceased being flummoxed by an inability to recognize phonemes and “words” in a really foreign language, you can make quick headway with translation. But as soon as you get to interesting concepts, things go poorly. You may find that hoping or expressions of anger or joy don’t have a place in that culture, thanks to a lack of the same array of practices that we have in ours. Likewise for their important concepts. Moreover, having grasped “hope,” the other people needn’t by analogy grasp our “joy” or “anger,” for each is embedded in its own web. This may even be true for speech acts, like promising or even stating, that are sometimes held out as neutral between cultures. Moreover, the disunity thesis is germane not just to psychology, but also to the “motley of mathematics” which Wittgenstein did discuss at length, and to the manifold styles of scientific reasoning about which he was silent.

This Issue

April 1, 1982