A Use-Therapeutic Approach to Meaning
A Use-Therapeutic Approach to Meaning


This paper makes two related points: first, that the content of Wittgenstein’s grammatical explanations, including his explanation of word-meaning, is not dissociable from the fact that it is genuinely addressed to the person that is the reader; secondly, that Wittgenstein’s notion of secondary meaning poses a problem for an attribution of a univocal use conception of meaning to him. It is argued that these points support the main contention that Wittgenstein’s approach to meaning is not use-theoretic but use-therapeutic.

Table of contents

    Wittgenstein’s reservations about philosophizing in a theoretical spirit are well known. In §§126-128 of the Philosophical Investigations, he rejects both explanatory claims and the usefulness of theses in philosophy. When the question turns to a theory of meaning, the answer to be expected is the same. However, the problem is that a good account of Wittgenstein’s non-theoretical attitude in philosophy inevitably contrasts theses and explanations with descriptions of language games and grammatical remarks. The latter, the account continues, is what Wittgenstein really offers. It is natural to ask, then, in what sense grammatical remarks are supposed to be independent of a theory of language. This is the question I want to address with particular regard to Wittgenstein’s famous dictum, “the meaning of a word is its use in the language” (§43).

    The first apparent identification of meaning and use in the Investigations occurs in §30: “[A]n ostensive definition explains the use – the meaning – of a word if the role the word is supposed to play in the language is already clear.” Is already clear to whom? Obviously, to the person to whom the explanation is given. To say that an ostensive definition explains the meaning of a word only to those who have a prior understanding of its general role involves relating the concept of word-explanation to the abilities of the learner. This is more explicit in §6, where Wittgenstein distinguishes ostensive teaching of words from ostensive definition. He restricts the latter to contexts in which the learner is able to ask for the names of things. It takes two to give an explanation. For Wittgenstein, this is true not only in a platitudinous sense. In his view, all forms of explanation are bound up with shared practices and differ according to individual practices. This he takes to be equally true of philosophy, which, accordingly, is not a matter of giving explanations that are immune to future experience, but dependent on the type of perplexities that require dissolution.

    This point can be deepened by discussing a difficulty Paul Horwich finds in reading Wittgenstein. The central question of Horwich’s Use Theory of Meaning is this: “[W]hich underlying, non-semantic property of a word provides it with the particular meaning it has?” (Horwich 1998: 103) For Horwich, the qualification “non-semantic” is crucial. If Wittgenstein were to allow semantic notions in characterizations of word-uses, “his identification of meaning with use could not be genuinely illuminating”. If, on the other hand, he is interpreted as keeping semantic notions away from his concept of use, his “thesis” that meaning is use would be controversial and thus conflict with the demands of §§126-128 (cf. Horwich 2004: 68f., fn. 5).

    In my paraphrase of Horwich’s “dilemma”, I have been slightly ambiguous about the distinction between giving an explanation of something and giving an explanation to someone. In Horwich’s account both variants are present, but he takes it for granted that the former is the only possible option in philosophy. This is what he implies in drawing a distinction between ordinary and “genuinely illuminating” explanations. My first point is that for Wittgenstein this is not the case. He is in effect addressing his reader, since, in his view, the problem is due to us. Secondly, what Horwich sees as threatening the value of Wittgenstein’s explanation of meaning as use actually provides an argument in favour of such a view. For what does meaning have to be like in order for semantic concepts to be explainable through sentences containing other semantic concepts, or even the very same words that are to be explained? What we exhibit through examples, in the explanation of word-meaning, is the way in which the word is used.1 Such explanations are genuinely illuminating for people, also in those cases where ostension plays no role.

    In this sense, Wittgenstein is not propounding a theory of meaning in the Investigations. As I read him, a philosophical problem stems from an incongruity between the way we think we understand certain words and the actual way we use them, together with a shift from what is manifest to the speakers of a language to an impersonal level of representation. The result is a “picture”, as Wittgenstein often calls it – a picture such as that of an “inner process”. Wittgenstein addresses his reader in order to enlighten him, thus changing his attitude toward the problem. In a related sense, Wittgenstein addresses his own tendencies to be misled, which he shares with the reader. Thus the continuing illuminative value of Wittgenstein’s remarks partially depends upon the extent to which he succeeds in articulating shared reactions. This is part of his mastery. If he suggests that such and such is the source of some particular philosophical puzzlement, his claim cannot be independent of the acceptance by the reader that, indeed, it is something like this that has led him astray. It is important to note that he is not only pointing out a particular tendency of our thinking but also giving an explanation as to how a philosophical problem might emerge from it. It is not this kind of explanation that he refuses to give. What he refuses to do is resort to assumption, postulation, generalization and other methods that play a vital role in scientific practice, since if he did proceed in this way he would no longer be speaking to the person that is the reader. He would no longer be giving explanations to someone but making someone an object of explanation.

    This contention is connected with Wittgenstein’s overall approach to philosophical problems. The notion of language-games is often taken to designate ways in which we actually operate with language in a multiplicity of contexts. However, another use of the term is more important. Primarily, language-games are descriptions of simple model contexts in which problematic features of a term can be studied. A language-game in this sense is similar to what Wittgenstein refers to as a “picture-object” in “Philosophy of Psychology – A Fragment” (PPF), formerly known as “Part II” of the Investigations. Wittgenstein introduces the term by way of a simplified drawing of a face and then explains: “In some respects, I engage with it as with a human face. I can study its expression, can react to it as to the expression of the human face.” (PFF §119) Though it is infinitely simpler than a real human face, we can stand in a similar relation to it. The same is true of the simplified descriptions that are langue-games, which is why our understanding of words can be directly studied by using them. In engaging in such procedures we can realise that the relationship between our concepts are other than we thought. Ordinary words can thus be identified, whose grammar can serve as a model for problematic terms in such a way that the philosophical puzzlement over them is dissolved. What Wittgenstein is offering is a sort of “after-education” (Nacherziehung) in the use of words. To invoke, as Horwich suggests, an underlying stratum of basic properties in terms of which meaning is to be explained would subvert this instructional or therapeutic aspect of Wittgenstein’s discourse by involving an impersonal level of representation. (cf. Horwich 2004: 68-70)

    Let us turn more directly to Wittgenstein’s explanation according to which the meaning of a word is its use in the language. In particular, I want to consider a variant formulation of §43, in which the well-known lines are followed by a third paragraph, whose deletion, according to the critical-genetic edition, is not unambiguous (my translation and emphasis): “Perhaps it would be more correct to say: a meaning of a word is one way of using it in the language.” (Wittgenstein 2001: 771f.) Accordingly, Wittgenstein is concerned with repeatable sounds or marks, and their relatively general meanings. In his view, generality of meaning is a matter of use, and meanings are relatively stable patterns of what people do with certain sounds or marks.

    To underline that meanings are relatively general is to admit that the meaning of a word need not be the same in all contexts in which the word is standardly applied. Relevant examples are words whose several uses share no common element, but display a kind of family resemblance. The general pattern is of the following type: word-application a and application b share a feature in one respect, b and c in another respect, while there is not necessarily any significant feature shared by a and c. In cases like these, a basic unitary meaning is hardly imaginable.

    Consequently, it is not easy to see how examples of this kind could be accommodated within use theories of meaning. The reason is that use theorists will not be content with a purely general explanation of word-meaning in terms of “the way a word is regularly deployed”. Use theorists want to answer the further question, which use regularities are responsible for the meaning a word actually has. According to Horwich, meaning is constituted by a basic regularity of silently accepting sentences in specific circumstances (cf. Horwich 1998: 94-96). As a consequence, he faces the difficulty of having to derive an indeterminate variety of uses of a given word from a set of unitary properties. I do not want to claim that family resemblance is an insoluble difficulty for Horwich’s account; I only want to indicate why this difficulty arises at all, rather paradoxically, in a theory that appeals to Wittgenstein as its progenitor.

    Unlike Horwich, Wittgenstein does not assume that word-meanings have a unitary basis. I wish to add: not even in their use, for this is what I want to suggest in the remainder of this paper. In Wittgenstein’s idea of language, the understanding of a word can be viewed as a capacity to transfer uses in partially similar ways from particular contexts into further contexts. It is basically nothing but this capacity, together with a number of ways of explaining words.

    Cases of secondary meaning are uses which are transferred in a way very dissimilar from the usual one. The first example I want to discuss is saying that Wednesday is fat and Tuesday lean instead of the other way around. Wittgenstein declares to be strongly inclined towards the first option and continues to ask: “Now have ‘fat’ and ‘lean’ some different meaning here from their usual one? – They have a different use. – So ought I really to have used different words? Certainly not. – I want to use these words (with their familiar meanings) here.” (PPF §274)

    It is important to see that Wittgenstein’s example is not metaphorical. Moreover, meaning and use seem to come apart. Wittgenstein wants to use “these words (with their familiar meanings) here”, while he obviously does not use them in the familiar way. Nevertheless, these familiar meanings do inform the expression “Wednesday is fat and Tuesday lean”. This is clear from the fact that Wittgenstein “could only explain the meanings in the usual way”, as he notes in the next section. This is what distinguishes secondary meaning from metaphor. To explain a metaphor, such as “the chairman ploughed through the discussion”, one has to use terms that are connected with the context of the metaphorically deployed word. Metaphorical meaning, as Max Black explains, is the result of an “interaction” between a word and a context in which it normally does not occur;2 it is not the meaning of a word. In instances of secondary meaning, by contrast, the element of interaction is absent. In explaining them, one has to fall back on the word’s regular pattern of use, its familiar meaning. Consequently, there is nothing secondary meanings could be meanings of, except for words. This is why they can, whereas metaphors cannot, be in conflict with Wittgenstein’s explanation of word-meaning in §43.

    Wittgenstein’s second example, “For me the vowel e is yellow”, is quite obviously a kind of aesthetic judgement. This is perhaps due to the kinship of this type of example with metaphors. This kinship is based on our practice of contrasting “dark” sounds with “bright” or “light” ones and the admissibility of saying that a vowel has its place “between” two others. As a consequence, many people might agree that English vowels could be classified from bright to dark in the following manner: eaiuo, and similarly with regard to colours: yellow – orange – red – green – blue. Given such a practice, they could see the point of devising a metaphor like “The vowel e is yellow”, at least in certain contexts. However, there is no reason to suppose that all people would agree on the above orderings. Somebody may want to include violet instead of green; another person insist on a different ordering of the vowels in the middle range. Consequently, some of the corresponding expressions, say, “For me the vowel i is red”, may indeed be like “Wednesday is fat”, while others, such as “The vowel o is blue” may be metaphorical. In a uniform series of items, some might be instances of metaphor and others instances of secondary meaning.

    Wittgenstein’s second example indicates that it may appear plausible to include secondary meaning in the broad category of aesthetic judgments. Accordingly, matching weekdays with “fat” or “lean” would be an expression of an aesthetic reaction. Although it presupposes an understanding of ordinary concepts, the corresponding rules do not determine which correlations are correct and which ones are not. Instead, it is up to each participant to make a decision. That this may be a rule of this peculiar language-game of “matching” does not alter the fact that the meanings of the words remain the same, while their uses differ. In this particular language game of “matching”, the very idea of attributing a colour to an object is transformed.

    To conclude, interpreting Wittgenstein, as Horwich tends to do, as propounding a use theory of meaning non-trivially conflicts with the discussion of secondary meaning. The alternative I suggested in the first part of the paper is to see Wittgenstein as adopting not a use-theoretic but a use-therapeutic approach to meaning.


    I would like to thank Joachim Schulte and Britt Harrison for assistance and discussion.


    1. Black, Max 1962: “Metaphor”, in: id.: Models and Metaphors: Studies in Language and Philosophy, Ithaca: Cornell University Press, 25-47.
    2. Horwich, Paul 1998: Meaning, Oxford: Clarendon Press.
    3. ––– 2004: “Meaning, Use, and Truth”, in: id.: From a Deflationary Point of View, Oxford: Clarendon Press, 67-85.
    4. ––– 2005: Reflections on Meaning, Oxford: Clarendon Press.
    5. Wittgenstein, Ludwig 2001: Philosophische Untersuchungen, critical-genetic edition, Schulte, Joachim (ed.), Frankfurt a/M: Suhrkamp.
    6. ––– 42009: Philosophical Investigations, rev. ed. by P. M. S. Hacker and Schulte, Joachim, Malden MA: Wiley-Blackwell.
    7. ––– : „Philosophy of Psychology – A Fragment“, in: Wittgenstein 42009, 182-265.
    Horwich’s “Implicit Definition Argument” comprises an analogous element of conferring meaning to a term by accepting postulates in which the very term is contained (Horwich 1998: 50, 138).
    Black 1962: 38f.; for a discussion of the ‘chairman’-example cf. 26-28, 30f.
    Bernhard Ritter. Date: XML TEI markup by WAB (Rune J. Falch, Heinz W. Krüger, Alois Pichler, Deirdre C.P. Smith) 2011-13. Last change 18.12.2013.
    This page is made available under the Creative Commons General Public License "Attribution, Non-Commercial, Share-Alike", version 3.0 (CCPL BY-NC-SA)


    • There are currently no refbacks.