The later theory puts language back in its place in human life, and this may give the false impression that all the old ideas have been superseded. Certainly, the abstract elegance of the Tractatus has gone. The point of origin of factual discourse no longer stands in the center like a surveyor’s landmark. There is, instead, a babel of different uses of language, all demanding an equal hearing by philosophers. The philosophical investigation of this diverse material is itself diversified. There is no longer a single recommended philosophical procedure. The old type of analysis, which deals with a puzzling type of proposition by translating it into a more detailed and explicit form, is not entirely given up. But more often Wittgenstein uses another method, which is to compare propositions and concepts taken from different areas of discourse, in order to draw attention to features of their logical grammar.
These comparisons are based on the actual use of the propositions and concepts in everyday life, and not on any theory about the ultimate basis of their meanings. If theories are involved they will not be philosophical theories, but the half-formed theories and vague imagery on the fringe of the consciousness of anyone who uses the propositions or concepts. For example, is the privacy of my mind like some kind of physical privacy? Are my secret thoughts like entries in my locked diary? Am I the ever present witness of them, or am I related to them in some other way? These questions probe the analogy between mental and physical phenomena. They may seem to be superficial, whereas the old method was at least an attempt to explore the depths. But the point of the new method is that the depths are on the surface. The features of our concepts are evident in the ways in which we use words. If we fail to see them, we have probably been misled by some primitive analogy between two areas of discourse. So the remedy is to take up the analogy and explore it in detail, in order to determine the point up to which it is valid, and beyond which it produces nonsense.
So the later exploration of language involves the same kind of incident as the earlier exploration. It leads us again and again to the limit of language, and there we are brought up short. This may happen because we feel that it must be possible to push an analogy a little further, or that at least it must be possible to explain why another step would produce nonsense. Or it may happen in a rather simpler way, if, for example, we ask why it is nonsensical to ascribe certain pairs of colors to the same area. In all such cases the philosopher can do little more than record that the limit has been reached, and describe the logical grammar of the relevant words which determines the position of the limit. He can, of course, try to loosen the grip of philosophical perplexity by inventing a slightly different logical grammar for the words, thereby shifting the limit to a new place. But he cannot describe the things in a way that would show what it is about them that forced us to give the words their present logical grammar. For the things did not force this choice on us.
At this point it is necessary to separate Wittgenstein’s later theory about the nature of philosophy from his later philosophical theory, if any, about the nature of things. His later theory about the nature of philosophy is built around the same central ideas as his earlier theory: philosophy is not a factual discipline, but an a priori discipline with certain distinctive features. It is an a priori discipline, because it investigates the structure of language and its account of that structure is based on the meanings of words. Its distinctive features are its scope and its detachment. It stands back from other disciplines and modes of thought, and it investigates all the varieties of discourse in which they find expression. But in the later period this investigation has become less systematic, partly because factual discourse no longer stands at the center. There is still the same concern with the limit of language, but it is now a much more intricate line, because it is plotted in relation to many different uses of language. The world of meaning of the Tractatus has exploded, leaving a number of separate systems each in its own space.
There is also another reason why the later investigation is less systematic. Instead of concentrating on truth-functional analysis, it more often uses the method of comparison, and the comparison of different types of proposition is a complicated and untidy process, particularly when it is illustrated by copious examples. Any attempt to theorize across the whole field of different types of discourse is disclaimed. In fact, the goal of philosophy is no longer supposed to be the production of theories, but the achievement of understanding through the activity of philosophizing. Wittgenstein is particularly hostile to the idea that philosophers should produce theories. He even equates the achievement of philosophical understanding with the removal of misunderstanding, and compares this process with psycho-analysis.
So it might be expected that his later work would contain no theory about the nature of things. But in fact it contains a repudiation of his earlier theory about the nature of things, and this repudiation might count as another such theory in its own right. If so, it would be classified as nominalism in ontology, and as conventionalism in logical theory.
His retractation of his earlier theory of elementary propositions, and of the ontology which he deduced from it, has already been described. After the retractation he might appear to be maintaining that the logic of factual discourse reveals not only the general structure of reality, but also many specific features of its structure. But at the same time he changed the very meaning of the thesis that the structure of reality exhibits this pattern, or that pattern. Instead of treating the structure of reality as a rigid framework, fixed in advance, he treated it as a projection of human thought.
This shift toward nominalism and conventionalism can be seen at many points in his later work. It is conspicuous in the criticism of the Tractatus which he develops at the beginning of Philosophical Investigations, where he dismisses his earlier ontology as an illusion, because it was offered as an objective account of the general structure of reality, but is in fact only an account of one of the ways in which we view reality. Here he is rejecting far more than the particular theory that reality consists ultimately of simple things. He is denying the general possibility of arguing in the way in which he had argued in the Tractatus for that theory or for any similar theory. His view now is that there is no valid argument from the present structure of language to the pre-existing structure of reality. This is a general condemnation of the kind of argument on which metaphysical systems most commonly rest—not only the metaphysical system of the Tractatus, which was avowedly expressed in propositions that do not belong to factual discourse, but also traditional metaphysical systems, which were expressed in propositions that were supposed to belong to factual discourse.
It is worth observing that in the Tractatus he had already rejected almost all arguments of this kind, and allowed only one exception—arguments from the nature of logic. For example, when he wrote the Tractatus, he did not think it possible to draw any philosophical conclusions from the a priori propositions of science. He pointed out that, though certain scientific laws—for example, Newton’s laws—are a priori propositions, that is only because we choose to adopt them, and the choice is not forced on us by any objective necessity in the phenomena of mechanics. On the contrary, Newton’s system is simply a framework or grid, through which we view the phenomena, and his laws merely determine the structure of the grid. But in the Tractatus he did not treat logic in this way. In the Remarks on the Foundations of Mathematics he extends the treatment to logic.
The anthropocentric theories, nominalism and conventionalism, are most plausible when they are applied to those rules of logical grammar which can be accepted or rejected independently of one another. For instance, the adoption of the definition of the word “planet” has no far-reaching consequences. Similarly—to take a case in which we do not depend on definitions in everyday life—it would be possible to divide up the spectrum of colors in a different way without many consequential changes in our conceptual framework (the color vocabularies of ancient Greek and Latin make distinctions which are not quite the same as our distinctions). It is true that in such cases there is often something which makes it natural to develop language in one way rather than in another way. Nominalists have usually admitted that, even if language does not uniquely reflect reality, there is something about the things themselves which makes a particular linguistic development a natural one.
Wittgenstein does not deny this. His point is only that the choice is not forced on us by the nature of things. The ultimate explanation of a necessary truth is only that we find it natural to develop language in accordance with it. Of course, there is much to be said in a general way about the concepts which are involved in any explanation of this kind, and it is at this point that his philosophy of mind comes in as an auxiliary to support his conventionalism. For the main theme in his study of the concept of following a rule is that, though a person’s application of a rule shows what the rule which he was following must have been, that rule did not dictate its own application. Even the person’s intention at the time could not prefigure all its own executions unequivocally. There will always be options which he did not think of in advance.
So necessary truth is a function of what we do with language, and we create it as we go—or, rather, we find that we have deposited it. But though this conventionalism provides a plausible account of options which are independent, or at least relatively independent of one another, it is much less convincing when Wittgenstein applies it to the a priori propositions of logic and mathematics. When the meaning of a logical or mathematical symbol has been fixed, it is hard to see what options are left, because it seems to follow that certain propositions must be accepted as necessarily true, and that certain proofs must be accepted as valid, even if nobody thought of these consequences in advance. Yet Wittgenstein insists in Remarks on the Foundations of Mathematics that there will always be options. This extreme conventionalism is the most paradoxical and perhaps the most vulnerable part of his later philosophy. The logical necessity, which, according to the Tractatus, reveals the general structure of reality, does not seem to depend on us. Though we may abandon the ontology of the Tractatus, it is difficult to abandon the objective theory of logical necessity with which it was associated.
The influence of Wittgenstein’s later ideas on the work of other philosophers has been close and penetrating. The Tractatus avoided details, and kept its distance from the actual work of logical analysis which it recommended. The later philosophy has entirely lost this aloofness, and exemplifies what it recommends in rich variety. The theory of language of the Tractatus was abstract and monolithic, and philosophers might well have been hesitant about its interpretation. But the much greater divisiveness of the new theory seems to make it possible to apply it to one problem without any thought of commitment to a similar application of it elsewhere. The single line dividing what can be said in factual propositions from what cannot be said in factual propositions has been replaced by a complicated network of lines dividing language into a great variety of different kinds of discourse, and then subdividing these types in innumerable different ways. The suggestion is that philosophy can and ought to be done in a piecemeal way without any attempt to construct general theories.
Theory or not, the achievement is a set of profound ideas about language and the mind, forged into a powerful instrument, which can be applied to the main problems of philosophy. The impacts that it makes on the philosophical controversies of the present day are clear. It is directed against certain forms of realism and objectivism, and, more generally, against any kind of systematic philosophy. In the philosophy of mind, what it offers is a carefully, but perhaps rather precariously balanced rival to behaviorism and Cartesian dualism.
There is one aspect of Wittgenstein’s philosophy of mind which deserves especial emphasis. C. K. Ogden, writing to Russell about the Tractatus on November 5, 1921, says, “I should very much like to know why all this account of signs and symbols cannot best be understood in relation to a thoroughgoing casual theory.” But this particular supplement cannot be added to the theory of the Tractatus. For Wittgenstein makes it very clear in the Bemerkungen that his view was that, whereas cause and effect are connected contingently, in an external way, the connection between a picture or a proposition and the fact which would verify it is an inner logical connection which depends on the intention of the user. According to him, this logical connection cannot be explained by bringing in some third, contingently connected thing, such as the experience of recognition. For a person might use a certain proposition to signify a particular fact, and later, when he was confronted with another, slightly different fact, he might have the experience of recognition, and so he might come to think mistakenly that this is what he had originally meant. He would, as it were, settle for a meaning which his original proposition did not have. This kind of mistake tends to occur when a person is asked whether this is the taste that he meant, or whether this is the color that he meant. It also affects aesthetic judgments, and in such cases it is sometimes more difficult to guard against it. But Wittgenstein’s point is that, even if this kind of mistake never occurred, it could occur, because the connection between the proposition, with its original meaning, and the later experience of recognition is only a contingent connection, and not a logical connection—not, as he often puts it, an inner connection. Similarly, the expression of a desire is internally connected with its object, and this connection cannot be explained by bringing in the feeling of satisfaction. For this feeling might be produced by something which was not the object of the original desire.
Wittgenstein makes the same point in the Conversations on Freud where he is criticizing the interpretation of the symbolism of dreams by free association. The person who had the dream may trace a path through the complicated maze of his associations, and this path may lead him to an interpretation of his dream, but according to Wittgenstein that does not prove that this is what his dream must have meant, or even that it had any meaning at all. In the Lectures on Aesthetics he makes what is essentially the same point, but its application to aesthetic judgments is more complicated. If someone complains that a door is too low (on aesthetic grounds, not on practical grounds), he does not mean that he has a feeling of discomfort which will be removed if the door were made higher. It may be true that his feeling of discomfort could be cured in this way, but it is one thing to describe the feeling prognostically by its cure, and quite another thing to describe it diagnostically by saying that at the time it was directed at a certain object, the lowness of the door. The success of the cure may suggest that the feeling was directed at that object, but it does not prove it. The connection between the feeling and its object is an inner logical connection, but it is only a contingent inference that, if the feeling vanishes when the door is made higher, it must have been directed at the lowness of the door. It might still have had a different object. This example, which is Wittgenstein’s, is characteristic of him. The case is a simple one, and nobody could suppose it likely that in such a case the cure would indicate the wrong object. But the example illustrates a feature of aesthetic judgments which according to him is essential to them, and which according to him cannot be given a causal analysis—they are directed at objects, or, to be more specific, at aspects of objects.
If, instead of considering a simple example like this one, we took a case in which the person who was making the aesthetic judgment found it difficult to formulate it precisely, the situation would be more problematical, but essentially the same. On the one hand, if Wittgenstein is right, no experiment which showed what changes in a work of art would lead to a retractation of the original adverse judgment could establish beyond all possible doubt what the aspect of the work against which it was directed must have been. But on the other hand, if we asked the person himself, he would begin by giving a vague and hesitant answer. How then could he arrive at a firmer and more precise answer? Wittgenstein suggests that the way for him to do this would be to make certain comparisons which would bring out the relevant aspect of this work of art. The point of the comparisons would be to reveal connections which run from the inner nature of the work, just as the comparisons which a philosopher makes in his study of the sense of a proposition reveal logical connections which run from the inner nature of the proposition.
Unfortunately, Wittgenstein does not develop this idea very fully in the Lectures on Aesthetics, and he does not make it clear how the comparisons which he has in mind would work. At one point he takes as an example a question about a piece of music: “Why do these bars give us such a peculiar impression?” He suggests that the answer to this question can be reached “only by peculiar kinds of comparisons, e.g., by an arrangement of certain musical figures, comparing their effect on us. ‘If we put in this chord it does not have that effect on us, if we put in this chord it does.’ ” Of course, Wittgenstein is not suggesting that this method will provide an absolutely certain indication of the aspect of the music which was the object of the original reaction. For here too the situation is essentially the same as it was in the simpler case of the door. The reaction to the slightly altered version might suggest that the original reaction had been directed at one aspect when in fact it had been directed at another aspect. His point is only that, in a case in which the person who is making the judgment is at a loss, the method which relies on alteration and comparison is the best available one.
These scattered examples illustrate a strong and constant tendency in Wittgenstein’s thinking, the tendency to treat the explanation of certain mental phenomena as a type of explanation which cannot be reduced to ordinary causal statements. Hume had hoped to devise a sort of mechanics of the mind, which would rival Newton’s achievement. Rightly or wrongly, Wittgenstein stands opposed to any attempt by modern empiricists to carry out anything like this program in any field which involves the mental phenomenon of intentionality or “directedness.” When I know what I mean, or what I intend to do, or what the object of my feeling is, Wittgenstein would deny that it is possible to give a causal analysis of what I know, or how I know it. It is always rash to predict the history of ideas, but at the moment it does look as if it is this element in his philosophy which will be most influential in the immediate future.