Wittgenstein, Ludwig Josef Johann (1889–1951)

9. Dismantling the Tractatus picture

Paragraphs 1–242 of the Philosophical Investigations (roughly the first third of the book) are generally agreed to provide the most focused presentation of some of the central ideas of Wittgenstein’s later outlook, in the context of which his views on philosophy of mind, mathematics and epistemology can helpfully be seen. We may divide the paragraphs into three groups: §§1–88 raise a variety of interrelated difficulties for the outlook of the Tractatus, §§89–142 discuss the nature of logic, philosophy and truth, and §§143–242 contain the so-called rule-following considerations.

The first group has two main targets: first, the idea that most words have meaning in virtue of naming something and, second, the idea that meaning requires determinacy and so exactness. Against the former Wittgenstein points out that different words function in different ways. To understand ‘five’, for example, a person needs to be able to count and behave appropriately on the result of a counting; to understand a colour term might, by contrast, involve knowing how to compare the specimen to be judged with a sheet of samples. To teach language one must train a person to produce and respond to words in the context of everyday activities such as fetching things, measuring, building, buying and selling. We can throw light on meaning by reflecting on simple ‘language games’, involving such integration of speech and action. To say that every word names something is like saying that every tool in the tool box modifies something. We can describe things this way if we insist: ‘The saw modifies the shape of the board; the ruler modifies our knowledge of a thing’s length’. But such assimilation may lead us to overlook important variety rather than representing a useful insight. To get someone to understand a word it is not enough to bring them face-to-face with the supposed referent while repeating the word. In order to profit from the confrontation, the learner must know what kind of word is being taught (for a number, shape, colour, and so on). And this in turn involves already being at home with the everyday activities into which remarks using the word are woven. ‘For a large class of cases – though not for all – in which we employ the word ‘‘meaning’’ it can be defined thus: the meaning of a word is its use in the language’ (1953: §43).

On the second topic Wittgenstein remarks that drawing a contrast between ‘simple’ and ‘complex’ depends upon context and interest. Items which might be seen as complex in one context could be taken as simple in another and vice versa. So the related notion of exactness is also context-relative. A word does not become unusable and hence meaningless because its use is not everywhere bounded by rules. That we can imagine circumstances in which a given description would seem inappropriately vague or in which we would not know whether to say that it was true or false is no criticism of its current use and hence no argument that it does not have meaning. Hence the idea that every meaningful sentence must have some underlying analysis in terms of simples is mistaken.

Each sort of word is at home in its own language game. But there are not always clear-cut relations of subordination or dependence between different language games. There are many predicates (for example, ‘intention’, ‘thought’, ‘statement’, ‘number’, ‘game’) which clearly do not name simples, because they have interesting richness and apply to complex items. But it is not the case that such predicates must have an analysis in terms of ‘simpler’ predicates. Search for such an analysis may reveal instead a ‘family resemblance’. Persons who recognizably belong to the same family may have various resemblances, of build, features, colour of eyes, gait, temperament and so on, which ‘form a complicated network of similarities, overlapping and criss-crossing’ (1953: §66), without one set of such resemblances being necessary and sufficient for having the appearance of a member of that family.

On the basis of this survey of the actual workings of language, Wittgenstein then concludes that the Tractatus picture of a detailed crystalline structure present in both world and language is an illusion.

But perhaps the considerations outlined require only minor tinkering with the Tractatus picture? We might say: ‘Certainly linguistic representations of states of affairs are put to various uses, in commands, jokes, stories etc., as well as in straight reports; also what degree of exactness we need is fixed by context; hence we should accept that for many practical purposes vague remarks are adequate. But none of this shows that we must discard the Tractatus picture of a fully determinate world, structured by simples; nor does it show that the idea of constructing a complete and exhaustive description of it need be abandoned. All it shows is that our everyday language mirrors the world less accurately than an imaginable ideal scientific language.’ That such a reading is inadequate is shown by considering the remarks on rule-following found in paragraphs 143–242.