Introduction to Wittgenstein’s Tractatus
Companion to Wittgenstein’s Tractatus
Letters from Ludwig Wittgenstein with a Memoir by Paul Engelmann
Wittgenstein: The Philosophical Investigations: A Collection of Critical Essays
Remarks on the Foundation of Mathematics
Lectures and Conversations on Aesthetics, Psychology, and Religious Belief [including the Conversations on Freud]
In the history of western philosophy, at least since the time of Descartes, it is possible to discern a definite pattern which is often repeated. First there is a period of expansion, and then follows a period of criticism and contraction. In the period of expansion a certain set of ideas will have won general acceptance, and they will be developed and applied to the whole range of human thought and experience. At such times philosophy is likely to encroach on other disciplines—on the physical sciences in the seventeenth century, more recently on psychology, and most recently on linguistics—and the central ideas will probably go too long unexamined. It is especially likely that the competence and credentials of philosophy itself will be left unquestioned. How, for example, should philosophical truths be classified? Do they belong to the realm of contingency, because they state things that might have been otherwise—such as the fact that Eisenhower was President? Or do they belong to the realm of necessity, because what they state could not have been otherwise? Eisenhower might not have been President, but it is necessarily true that he was younger than his mother, and that 7+5=12. To put the same question linguistically—are philosophical truths expressed in factual propositions or in a priori propositions?
When such questions are asked, a period of criticism begins. The vague persuasiveness of philosophy is analyzed, and there is a demand for a precise specification of the kind of statement which philosophers may legitimately make, and of the grounds on which they should be accepted. When philosophy criticizes itself in this way, the result is usually a certain contraction. If it has grown into some other discipline, such as psychology, and has become intertwined with it, those branches will be cut back, and a definite limit will be set to its competence. There are various ways in which this self-criticism may develop, but since the Renaissance the most important step has always been to inquire whether philosophy is a factual discipline, like the sciences, and, if not, how far removed from them it is. One method of arriving at an answer to this question would be to take factual discourse first and to try to determine its limits, and then to ask whether philosophy falls within those limits. This would be an indirect approach to the question what philosophy ought to be.
About the turn of the last century, philosophy in England, largely under the influence of the work of Moore and Russell in Cambridge, began to move into a phase of criticism and contraction. Looking back on it now we can see that they started the most important movement of this kind since Kant. There is, of course, a difference between their view of what philosophy ought to be and Kant’s view. They believed that the correct procedure would be the detailed analysis of the meanings of the various kinds of things that people say, including the things that philosophers say, whereas Kant believed it to be a very general analysis of the powers and limitations of human thought. But since sentences ought to be intelligible and thoughts ought to be expressible, this difference is not so great as it looks. There is, in any case, a striking similarity between their view of what was wrong with the philosophy of the period of expansion which their work was to supersede and Kant’s view of what was wrong with the philosophy which was related in a similar way to his work. In both cases the criticism was that philosophy was insecurely balanced between contingency and necessity. If philosophy was really like a science, why were its results not based on observation and experiment? If, on the other hand, it was really like mathematics, why was there so much disagreement about the validity of its arguments?
Wittgenstein’s Tractatus Logico-Philosophicus, which was first published in 1921, is in part an attempt to determine what philosophy ought to be, and particularly what its relationship to factual disciplines really is. It is also much more than this, because such an undertaking necessarily leads on to an investigation of the difference between factual and a priori propositions, and so to a general theory about the various kinds of things that people say, and the grounds on which they should be accepted. But one of Wittgenstein’s motives for seeking a general theory of meaning and truth—the one which he emphasizes in the Preface to the Tractatus—was that he wanted to put philosophy in its proper place. He wanted to show that it cannot do one of the things which it has often been expected to do: it cannot establish religion and morality on a factual basis, because it is not another factual discipline like the sciences, but more general in its scope—in short, traditional metaphysics is not possible.
His method of placing philosophy is indirect. He takes factual discourse first, and tries to determine its limits, in order to see whether philosophy falls within those limits. So after announcing in the Preface that the Tractatus deals with the problems of philosophy, he says that the aim of the book is to plot the limit of language, within which everything that can be asserted truly or falsely will have to find a place. The two projects are connected because, according to him, the traditional problems of philosophy arose from linguistic misunderstandings—not, of course, from particular failures to communicate with other people, but from general misconceptions, and especially from the inability to see all the differences between different kinds of discourse. It is, for example, obvious that morality is not a science, but the full extent of the differences between moral and scientific discourse is far from obvious.
When Wittgenstein says that the aim of the Tractatus is to plot the limit of language, he means that its aim is to plot the limit of factual language. Although it is clear why he believed this task to be important, it is not so clear why he describes it in the way that he does. Whatever the exact scope of factual discourse—and this too is uncertain—it is obvious that not all discourse is factual. So does non-factual discourse not count as language? Is there something wrong with all non-factual discourse? Does his equation of language with factual discourse express a kind of positivistic intolerance? These are crucial questions in the interpretation of his philosophy. For beyond the limit of language there is only nonsense, and so, given the equation of language with factual discourse, all non-factual discourse will be nonsense—a conclusion that will be as harsh and intolerant as it sounds, unless either the term “factual discourse” is being used to include more than is commonly allowed, or there is some subtle distinction between good nonsense and bad nonsense.
Perhaps, however, it is not quite accurate to say that nonsense lies beyond the limit of language. For the task of plotting the limit is rather like calculating the curvature of space, because there is nothing outside it. Nonsense is nowhere, and the line between sense and nonsense can be drawn only from some point of vantage within language. Now if “language” means “factual discourse,” the point of origin of this geometrical construction must lie somewhere within the factual assertions that people actually make, or, at least, somewhere within some of them. If we did not live in a world of factual meaning, we could not imagine what such a life would be like, because to imagine something is to put together certain words, or perhaps certain ideas, in such a way that the result—a sentence or a thought—has a meaning, even if it is not true.
But the general area, within which the construction certainly had to start, contained a variety of possible points of origin, among which a choice needed to be made. It would be possible to start from empirically based factual propositions, on the ground that at least they have meanings. Or it might be better to include under the heading “factual discourse” more than is always allowed. If religious discourse were included, and other similar relaxations were allowed, it would be possible to take several points of origin and to give the space of factual meaning a more sinuous curve, which would allow for the gravitational pull of whatever it is that is happening in less scientific fields of thought.
There are also other questions which need to be asked about Wittgenstein’s attempt to plot the limit of language. For in order to understand his choice of a point of origin, we need to know something about his theory of logical analysis. Now he began to develop the ideas which went into the Tractatus soon after he first arrived in Cambridge in 1912, and like Russell, with whom he worked, he did not take ordinary factual propositions in the form in which they are current in everyday life and in scientific discourse. He believed that language disguises thought, and that the real forms of our thoughts would be apparent only when the language in which they are expressed had been analyzed and broken down into its basic components. His idea was that the assertion of an ordinary factual proposition is a gross move, which contains within itself a number—perhaps a very large number—of minute moves.
For instance, merely to assert that the watch is lying on the table is to assert by implication quite a large number of other propositions. These implied propositions would be propositions about the mechanism inside the case of the watch—or, at least, this is the natural view to take of the implications of the original assertion. (There is also another possible view about them—that they would be propositions about people’s sensory experiences.) Of course, Wittgenstein was not recommending that the assertion of each of these implied propositions should be a separate move in everyday life. The grossness of ordinary factual propositions is a blessing. But he did think that an exact account of what they mean could be given only if they were analyzed into their basic components, which he called “elementary propositions.”
This raises further questions about the point of origin which Wittgenstein used in the Tractatus. If we are analyzing ordinary factual propositions, how far shall we have to go before we reach the elementary propositions which will reveal the real forms of our thoughts? This is not a question about the scope of factual discourse, but a question about its depth. How are we to know when analysis has gone deep enough? Do we even know the general direction which it ought to take? Should the statement about the watch be analyzed into statements about its mechanical parts, as has just been suggested? Or should it be analyzed into statements about the sensory experiences (sense-data) which a person would get if he examined it—e.g., into statements about the visual impressions which he would get if he looked at it?