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

• Home
• About
• Login
• Search
• Current
• Archives

Table of contents

• References

Wittgenstein's Tractatus can, and has been plausibly interpreted as an exercise in Kantian critical philosophy. As Kant's critical philosophy is an attempt to determine the foundations of knowledge through considerations of the necessary conditions of possibility of knowledge, and to set limits to knowledge claims in this way, so the Tractatus can be understood as an attempt to determine the foundations of language through considerations of the necessary conditions of possibility of language, and to set limits to language use in this way. Some commentators have suggested that Wittgenstein's later philosophy, too, could be understood as critical philosophy in this sense. In this paper I will argue, however, that this is problematic. Although Wittgenstein's later philosophy is concerned with limits of language, it is not an attempt to give foundations and set limits to language, - these being attempts to give reasons or grounds for why language must / can only be used in certain ways. Rather Wittgenstein questions the idea of philosophy as giving foundations. Thus, he does not e.g. move from a strong absolutist view of foundations to a weaker relativistic view of foundations, or deny that foundations can be determined for language either in the strong or weak sense. Rather, he points out that the idea that philosophy gives foundations does not make sense, - and hence that there is nothing to deny. Consequently, he puts forward a different conception of philosophy: philosophy as description of language use, where description is not a new way of giving foundations.

Kantian interpretations of Wittgenstein's later philosophy are sometimes (cf. Apel 1998, 10; Hintikka & Hintikka 1986, 4) taken to get support from a certain remark by Wittgenstein - one of the very few in which he mentions Kant: "The limit of language manifests itself in the impossibility of describing the fact that corresponds to [...] a proposition without simply repeating the proposition. (We are dealing here with the Kantian solution of the problem of philosophy.)" (MS 110, 61; TS 211, 171) I will, however, argue that rather than identifying himself as a Kantian philosopher, Wittgenstein identifies here a problem in Kantian philosophy as an attempt to give foundations. It is in this sense that Wittgenstein says he is dealing with the Kantian solution to the problem of philosophy. Yet, this question of interpretation cannot be decided on the basis of the remark alone or its immediate context. It must be decided in the broader context of Wittgenstein's philosophy. To do this I will draw on other remarks to be found in MS 110, as well as on remarks in the Philosophical Investigations (PI). But let us start by taking a look at the immediate context of the remark.

The remark occurs in a context where Wittgenstein discusses a tendency to think that the meaning of a symbol were determined by something else, namely by what it talks about. Thus, there is a tendency to think that e.g. the meaning of the word "red" derives from the colour red, or the meaning of the word "wish" from a mental state, say, a feeling of dissatisfaction. Yet, this is not so, according to Wittgenstein. Singling out the colour red and the feeling of wishing presupposes the corresponding concepts. Thus one cannot point out a state of affairs, say, that a ball is red, without presupposing the appropriate concepts, without the proposition's being embodied in one's actions of pointing, as it were, and without the proposition's being repeated in this sense. Accordingly it is impossible to describe the fact without repeating the corresponding proposition. (See MS 110, 60-64; cf. TS 211 169-173)

Let us now return to the quoted remark. According to Wittgenstein, as this repetition of a proposition occurs, we are at the limits of language. But what is the issue and problem here? Why would one want to have a description of the state of affairs independent of the proposition in the first place? And why is it important to notice that one just ends up repeating the proposition while attempting this? The context of the remark suggests an answer. If one had a description of the state of affairs independently of the proposition one would have found a foundation for language in what it talks about. However, this attempt to give foundations involves the unnoticed repetition: one imagines not to be employing concepts while employing them nevertheless. In the attempt to ground the proposition one simply repeats it, and thus the attempt to give grounds does not succeed.

But how does this show that Wittgenstein is not a Kantian philosopher? Surely no Kantians wish to ground language or concepts on what is found in the world in the way described above. Is it not precisely the point of Kant's Copernican turn not to try to do this? Consequently, does not this interpretation support the idea of Wittgenstein as a Kantian philosopher, rather than show that he is not a Kantian? I agree with all of this except for the conclusion. Hence I would rather say: from Wittgenstein's point of view Kant moves to the right direction, and both Wittgenstein and Kant reject what Kant calls "transcendental realism". Yet, the very same problematic repetition occurs in Kantian attempts to give grounds and reasons for language too, and this, I believe, is why Wittgenstein mentions Kant in the remark in question. - But to see this the remark must be seen in the wider context of what Wittgenstein says about philosophy. First, however, let us consider an example of an attempt to give foundations which could more naturally be called "Kantian".

By Kantian attempts to give foundations for language use I understand attempts to give reasons or grounds for why language should be employed in certain ways through considerations of necessary conditions of possibility, - or better: why it should not be employed in certain ways, i.e. in ways conflicting with its necessary conditions. (The latter formulation, I believe, captures better the idea of a Kantian critique.) In this purpose, one might then put forward an argument of the following kind concerning the concept of rule following: "If rule following were not a practice there would be no criteria for following or going against a rule. Therefore rule following must be a practice.". Allegedly this would then show that one should call "rule following" only something which is (at least potentially) part of a practice.

Let us accept for the sake of the argument that there being criteria for following or going against a rule depends on rule following's being a practice. Granted this, one can say that rule following is necessarily a practice, and that only something which is part of a practice is correctly called "rule following". Yet, the alleged argument above gives no reason or ground for this being so. For a reason or ground should be independent of what it is a reason or ground for. Here, however, this is not the case. Thus, were one to adopt another kind of way of using the word "rule following", the supposed reason and ground for using the word in the first way would thereby vanish too. That is, if one were to use this word in such a way that there would be no distinction between following and going against a rule, it is not that the distinction would now be missing, as it were, and the concept incomplete. It is not that there would still be worries about whether someone follows or goes against a rule, and yet no way of deciding this. Rather, there simply would no longer be questions about following or going against. To worry about deciding this here is a misunderstanding. Likewise the alleged reason "otherwise there would be no criteria for following or going against" has no existence independent of the usual use of the word, and this shows that it cannot function as a reason or ground for adopting this use.

As to the relation of this example to the red and wish examples in MS 110, the idea that the clause "otherwise there would be no criteria for following or going against" expresses a ground for why only a practice can be called "rule following" involves precisely the same problematic repetition which takes place in the earlier examples. What strikes one as a ground for one's having to use the word in this way, is not given independently of the use, and by referring to the alleged ground one just repeats what was supposed be given grounds. That is, one gives a paraphrase for the use, but no grounds for it. Thus the second, more clearly Kantian attempt to give foundations fares no better than the two first ones. There are no grounds or reasons to be found for why the word should be used in this way. Consequently, insofar as possible ways of giving foundations are now exhausted, philosophy cannot give language foundations. Its task is merely descriptive. This is what Wittgenstein seems to think. He writes in PI, in a remark originally drafted in MS 110: "Philosophy may in no way interfere with the actual use of language; it can only in the end describe it. For it cannot give it any foundation either." (PI, 124; MS 110, 188)

Accordingly, no limits are set through considerations of necessary conditions of possibility. The limit of language comes to view here but no reasons or grounds can given for why it should be respected. Grammar is arbitrary, as Wittgenstein says (cf. MS 110, 133), which in the end, means, I believe, that no philosophical reasons or grounds can be given for why words must be used in certain ways. In a remark on the notion of the arbitrariness of grammar Wittgenstein writes: 'The rules of grammar may be called "arbitrary", if that is to mean that the aim of grammar is nothing but that of the language. If someone says "If our language had not this grammar, it could not express these facts" - it should be asked what "could" means here.' (PI, 497)

Thus, if one is inclined to say something like: "if rule following were not a practice, one could not distinguish between following and going against", one should, ask what is the sense of "could" here. And the answer is: to use the word "could" here is quite misleading. There is no could here, i.e. something one might try to do but only to find it impossible. Insofar as there were no criteria for following or going against, it is not that one could not distinguish between these two. Rather one simply would not distinguish between following and going against a rule. (Cf. MS 136, 121b)

To look at the issue from point of view of yet another remark of Wittgenstein's, he writes in MS 110 about the idea that grammar forbids something:

The grammatical rule is supposed to, for example, forbid something, say, that the word 'A' can be substituted for the word 'B'. But how can it forbid that the 'is' in "2x2=4" is substituted for 'is' in "the rose is red"? This is nonsense.
    This arises from the pernicious idea that behind the word there would be as if an invisible train of rules so that it would make sense to talk about two words which look the same. The question is about one word which can be substituted for by two words that cannot substitute for each other. (MS 110, 207)

So, there may be a word x which, when used like this, is to be substituted for by the word a, but at another occasion is used like that and to be substituted by the word b. But it is not that the word x is, as it were, two words at once because of these two different possibilities of substitution. Rather the sign x is a symbol when it is used as a symbol, and what symbol it is depends on the particular use given to it. Thus it does not happen that it symbolises in two incompatible ways at once at any particular occasion of its use, as if its identity as a symbol were independent of the particular use which is made of it, after all.

Similarly for the word "rule following". If it is used in such a way that there is no way of deciding what counts as following the rule and what as going against it, this is merely a use different from the normal one. There is, however, no inconsistent or impossible use for the word to be excluded from language, where one would like to talk about following and going against but there would be no criteria for deciding this. To be sure the word can be used in inconsistent ways on different occasions, but at none of these particular occasions is its use inconsistent as such. The idea that the integrity of the word use somehow need to be protected by setting limits, so that the inconsistent case would not arise is a confusion. It can be pointed out that one is using the word in incompatible ways on different occasions, but there is no impossible use to be excluded from language, and thus no limits to be set for language. All one can do is describe the different uses and point out their mutual incompatibility.

And finally in regard to the attempts to give foundations: it is not that one is trying to do something impossible here either. Rather, one is just being inconsistent. To determine necessary conditions of possibility is simply not what one usually calls "giving grounds". Yet, as this has been pointed out it has not been shown that the expression "give grounds" must be used in any particular way. To point out that on one occasion one is contradicting something one adheres to or presupposes on another occasion is not to lay down foundations, but just to point out inconsistencies.

Refbacks

• There are currently no refbacks.