From the ALWS archives: A selection of papers from the International Wittgenstein Symposia in Kirchberg am Wechsel

1.

If there is anything uncontroversial about the later Wittgenstein’s philosophy of language, it is his view that meaning depends on linguistic practices. Many – though by no means all – commentators also agreed that he thought these practices must be not just individual but somehow social. Most – indeed, so far as I know, all – of these commentators further agree that he thought these practices must actually be communitarian, so that having a language essentially depends on meaning by one’s words what members of one’s community mean by them. I believe that a proper understanding of what motivates Wittgenstein’s claim that meaning depends on linguistic practices does entail that these practices must be social. But I also want to suggest, more controversially, that this understanding entails that the relevant practices do not have to be communitarian but rather that they can be merely interpersonal, so that having a language essentially depends only on having had (many of) one’s words, whatever one means by them, understood by others.

2.

Why does meaning depend on linguistic practices?

If it did not, Wittgenstein argues, there would be no distinction between courses of action that accord with a rule and courses of action that conflict with it (1958, #201); there would be no distinction between correct and incorrect applications of linguistic expressions. In what sense, however, must this distinction hold for there to be meaningful expressions at all? At least in this sense: there must be a distinction between saying something true and something false, between obeying an order and disobeying it, etc. In other words, the applications of linguistic expressions must be governed by standards of correctness; and these standards, like any standards, must be objective in the sense that their being fulfilled or not is not dependent simply on language users’ opinion of the matter. Now it is true that we could also distinguish between correct and incorrect use of expressions by saying that correct use is use that is in accord with communal or expert use and that incorrect use is use that deviates from communal or expert use. But it is an open question whether speaking correctly in this sense is also essential to meaning, whether one could not succeed in saying anything true, or anything false, for that matter, unless one spoke like others. Indeed whether this is the case is precisely the issue that divides the communitarian camp from the interpersonal one. And presumably, if Wittgenstein did think that speaking correctly in this sense is also essential to meaning, it would be as the conclusion of an argument and not as an assumption made at the start of his inquiry into what kind of practices can deliver objective linguistic standards.

Wittgenstein’s claim that objective linguistic standards and hence meaning could only be based on linguistic practices is meant to replace the claim that interpretations determine meaning, a claim which, he argues, all traditional theories of meaning are committed to. This stems from their contention that linguistic standards are provided by items that stand in complete independence of people’s use of words. In the sections on rule-following, Wittgenstein focuses on theories that postulate internal items as determinants of meaning, items such as mental pictures coming before the mind and abstract entities grasped by the mind of language users (1958, ##139-55). To mean something by a word is somehow to associate it with an item of one or the other kind. The problem, however, is that none of those items wears its meaning on its sleeves, as it were, but each of them, taken on its own, could be an instance of a variety of things. And so each could be an instance of one kind rather than another, and provide the standard for the applications of the word it is connected with, only if it were interpreted in some way or other. But this entails that it is always possible to interpret one’s words in such a way that, no matter how one applies them, the application will be correct, or incorrect, as the case may be. Moreover, settling on an interpretation would be of no help, for the interpretation itself would need to be given meaning, presumably by being once again associated with some item that is on the face of it neutral.

As I said, in the sections on rule-following, Wittgenstein focuses on internal items. But the same point can be made – indeed, Wittgenstein himself makes it – about external items, that is, physical objects and events in one’s environment. As Wittgenstein makes clear in his discussion of ostensive definition, external items too need to be seen under one aspect or another before they can serve as standards for the application of the words used to refer to them.

This, I take it, is one of the main lessons of Wittgenstein’s rule-following considerations: whatever exactly it is that provides linguistic standards, nothing can do it in complete independence of people’s use of words. Something can serve as a standard, as a determinant of meaning, only if it is used as such. This is actually a lesson well learnt by commentators on Wittgenstein. Thus David Pears writes: “The meanings of our words are not guaranteed by any independent pattern already existing in the world and waiting for language to be attached to it.” (1988, 363) And Barry Stroud: Nothing could play the role of “guiding or stabilizing” words’ applications “unless it meant something to the people whose actions are to be ‘stabilized’ or ‘underpinned’ by it.” (2000, 231)

3.

So far we have seen why Wittgenstein maintains that meaning depends on linguistic practices. The next question is: on what kind of practices does it depend?

Let me start by noting that Wittgenstein seems to think that for there to be a distinction between correct and incorrect applications of one’s words, one must also be aware of the distinction. This is suggested by his remarks in section 258 of the private language argument when he says that the would-be private speaker could not succeed in endowing the sign for her sensation with meaning because she has “no criterion of correctness”, that is, no way to distinguish between what are the correct applications of her sign and what seem to her to be the correct applications. It seems to me that the suggestion here is not that she has no criterion because she has failed to establish a standard of correctness, in trying to define her sign ostensively. Rather, the suggestion seems to be that she has failed to establish such a standard because she herself cannot, in this context, make the distinction between correct and incorrect applications. There is no standard because she is not in a position to regard anything as a standard.

Here is another passage that suggests the same conclusion. In Remarks on the Foundations of Mathematics, Wittgenstein argues that we would have no reason to believe that two creatures were using signs meaningfully unless we could observe, besides the regular production of signs by the creatures, actual interaction between them, e.g., “the phenomenon of a kind of instruction, of showing how and of imitation, of lucky and misfiring attempts, of reward and punishment and the like” (p. 345). What difference does the kind of interaction Wittgenstein has in mind, viz., that of a teacher instructing her pupil how to continue a series of signs, make? Surely this one: through the interaction the student is taught the distinction between doing the right thing and thinking that she is doing the right thing, and so the distinction between correct and incorrect applications of her signs. Thus Wittgenstein seems to think that we cannot say of someone that she has a language simply because we can describe her regular behaviour as rule-governed; we also need some reason to believe that the creature recognizes her rule-following behaviour as such. But why is this needed?

After all, it is easy to imagine a solitary person who regularly produces sounds or marks in response to what appear to be salient aspects of her environment. And it is easy to describe these responses as converging on particular aspects or, to put it in Wittgenstein’s terminology, as agreeing with each other, and so as being of the same kind. This may lead us to conclude that the particular aspects converged or agreed upon do provide the standards of application for the solitary person’s signs since they certainly seem to be used as such. But they might well not be so recognized by the solitary person herself.

I think, though, that these remarks belie a failure to appreciate the significance of the claim that standards are partly constituted by people’s activities, that they are the product of people’s use of words and the various kinds of items these words refer to, that is, in the first instance, objects and events in their environment. Neither side can produce standards on its own – on its own, each side is neutral among the many standards to which it may contribute. Neither people’s environment nor their words’ applications are by themselves such that these applications must latch on to one rather than another aspect of their environment. In themselves, any two or more applications may be seen as different in kind or as of the same kind in more than one respect. Similarly for any two or more items in people’s environment. It is for language users somehow to determine which applications are of the same kind, and in which respects, which is to say that it is for language users somehow to determine what are the standards that govern the applications of their words. But, if so, someone’s applications, be they solitary or social, cannot merely happen to be in agreement, linguistic standards cannot be provided by items that merely happen to be used as such. Rather, agreed upon applications must be recognized as being agreed upon, as of the same kind, in order to be of one kind or another. Thus linguistic standards must be recognized as such in order to serve as standards, which is to say that, in order to have a language, one must have a sufficiently robust concept of objectivity. For to recognize something as a standard is to recognize it as something conformity to which is independent of whether one thinks one has succeeded or not in conforming.

It is for this reason that the practices that establish linguistic standards must be social. For it is hard to see how a solitary person could be in a position to have the requisite concept of objectivity. No matter how complex and regular we may find the connections between her productions of signs and her environment and activities in it, the fact is that whatever she counts as the same kind of objects will be, for her, the same kind, and thus whatever applications she deems to be correct will be, for her, the correct ones. The solitary person has no need to draw the line between correct and incorrect applications one way rather than another and so to settle for some rather than other linguistic standards. She “communicates” with herself no matter what she says at any given time. There is no point at which, suddenly, she has to settle on one rather than another answer to the question, “what kind?”, and so no point at which she could be in a position to distinguish between what is the same and what seems to her to be the same.

On the other hand, if a person is socially situated, then she is situated in a context that makes it possible for her to make the relevant distinction and thus for items in her environment to serve as standards governing the applications of her signs. This is so not just because she is situated in a context in which there already are linguistic practices and hence standards to which she can be introduced. Rather, it is the interaction with other people that makes it possible for her to distinguish between what seems to her the same kind of applications and what is in fact the same kind. And thus it is the interaction with other people which makes it possible for there to be linguistic standards in the first place. Interpersonal interaction puts people in a position to make the distinction because it can confront each of them with perspectives different from their own, thereby allowing them to realize that matters may be different from what they seem to be. Interpersonal communication could not succeed if it were simply up to any given language user to settle on the answer to the question, “what kind?”. Thus interpersonal communication makes it possible for language users to regard aspects of the world around them as providing objective linguistic standards.

There is, however, no reason here to believe that the social practices one must be involved in in order to have a language must be shared with others, that the standards governing the applications of an individual’s words must be those of her community. Of course, given the usual way in which someone acquires a first language, viz., through training by members of her linguistic community, what she means by her words will often be similar to what her teachers mean by them. But nothing we have seen indicates that this has to be so. The crucial part of the training lies in the acquisition of the ability objectively to distinguish between correct and incorrect applications of words, not in the learning of how the distinction is to be drawn with regard to the applications of any particular word.

If the foregoing is right, then there is good reason to believe that the linguistic practices that meaning depends on do not have to be communitarian, but that they only have to be interpersonal.