Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 27 June 2019

The Tractatus and The Limits of Sense

Abstract and Keywords

So long as we have the idea that what we want goes beyond the limit, resignation is in place. If, instead, we were to recognise that the idea that we were trying to ask something, that there was something we wanted to find out, or wanted to say, was confused, and that there was nothing that we were asking or wanting to say, there would no longer be anything to be resigned about. Thus ‘complete satisfaction’ would be possible. This article looks at Ludwig Wittgenstein's Tractatus Logico-Philosophicus and demonstrates the connection between Wittgenstein's ideas about marking out the limits ‘from inside’ and his ideas about philosophical method, as well as the connection between those ideas about method and the contrast between the two ways of understanding the limits of language. It also discusses Wittgenstein's conception of philosophy and what he thought it could accomplish. It is a mistake to argue that, of the various forms of philosophical solipsism, Wittgenstein is concerned only with solipsism of some particular type.

Keywords: Ludwig Wittgenstein, Tractatus Logico-Philosophicus, resignation, satisfaction, limits of sense, limits of language, philosophy, solipsism

When I say: Here we are at the limits of language, that always sounds as if resignation were necessary at this point, whereas on the contrary complete satisfaction comes about, since no question remains.

(BT, 310)

In the quoted remark, Wittgenstein contrasts resignation with satisfaction. If saying ‘Here we are at the limits’ appears to suggest the need for resignation, that is because it looks as if there is something we cannot do. Why else should we think we need to be resigned? It looks as if whatever question it is that took us to this point, or whatever statement it was that we wished to make, would go beyond what is sayable, and so we are stuck. So long as we have the idea that what we want goes beyond the limit, resignation is in place. If, instead, we were to recognize that the idea that we were trying to ask something, that there was something we wanted to find out, or wanted to say, was confused, and that there was nothing that we were asking or wanting to say, there would no longer be anything to be resigned about. Thus ‘complete satisfaction’ would be possible. We need to look further at that contrast between a conception of the limits of language, or of the limits of sense, which leaves us with resignation and a conception which contrariwise goes with ‘complete satisfaction’. Is there at the ‘limits of language’ genuinely something that we cannot do? Or is that a misunderstanding? I shall examine some of the ways this issue comes up in the Tractatus, although I shall start with some slightly later remarks. I want to explore the idea of making limits clear ‘von innen’, from inside. This is an idea that comes up explicitly at TLP 4.114, and in Wittgenstein's letter to Ludwig von Ficker, explaining his aims in the Tractatus. (PTLP, 15) I shall try to show the connection between Wittgenstein's ideas about marking out the limits ‘from inside’ and his ideas about philosophical method, and the connection between those ideas about method and the contrast between the two ways of understanding the limits of (p. 241) language. The recognition of the limits of language, where such recognition is not a recognition of constraints, can be achieved only through a philosophical approach that works ‘from inside’.1 Although much in this chapter bears on Wittgenstein's conception of logic, a treatment of his understanding of the relation between logic and limits would have doubled its length, and so logic (as a topic in its own right) lies outside its limits. (For a discussion of questions about the relation between logic and limits, see McGinn 2006, especially ch. 11.)

1. A First Look Round

Consider the remarks about ethics from ‘Notes on Talks with Wittgenstein’. In December of 1929, Wittgenstein spoke of the astonishment one may feel that anything exists. He said that this astonishment ‘cannot be expressed in the form of a question and there is no answer to it’; and he added that whatever we say here must, a priori, be nonsense. Here we are thrusting against the limits of language (NTW, 12–13). But a year later he rejected that description: ‘Language’, he said, ‘is not a cage’. (NTW, 16) Speaking of thrusting against the limits suggests an ‘inside’ within which we are confined and an ‘outside’ which we can't get to. He had said that our astonishment that anything exists cannot be expressed in the form of a question. So what happens, then, if we express the astonishment by uttering the words ‘Why is there anything?’ In an important respect there is no change between 1929 and 1930 in his view about the uttering of such words, the using of the interrogative form in that way. The astonishment ‘cannot be expressed in the form of a question’: so Wittgenstein's view appears to be that there is on the one hand the astonishment and there are on the other hand the words which grammatically have the form of a ‘Why?’-question; and the words and the astonishment do not meet. No question at all has been expressed.

Why not? I shall look briefly at two ways of replying. I am not at this point arguing that either approach is in fact Wittgenstein's; my discussion is meant to use what one might think of as Wittgensteinian material in formulating two related ways of responding to the question. Both approaches begin from the same point, namely that a question anticipates ways of using words in an answer. The first approach goes on this way from that initial point: Any form of words purporting to be an answer to ‘Why is there anything?’, in which it seemed words were being used to say how things stood, would not be what we want in formulating what purports to be the question ‘Why is there anything?’ No such ‘answer’ would tell us why there were things that could be spoken of at all, since it would take for granted the availability of such things to provide explanations, and would not explain the existence of the things themselves. Such an ‘answer’ would not be (p. 242) a genuine answer. Insofar as the ‘question’ is modelled on questions which look towards the world and what is the case in it for an answer, insofar as we take ourselves to be asking such a question, we are in difficulty: we want to ask a question modelled on those that look to what is the case for answers, but we also want to ask something that would lead us to reject, as an answer, any statement of what is the case. No such answer would satisfy us; and it should become clear that the ‘question’ fails to anticipate any range of possible answers.

That first account takes for granted that the ‘question’ is modelled on questions that are answerable by stating something that is the case, that is, by the use of what the Tractatus speaks of as senseful propositions. The argument underlying the approach is essentially that we can come to see that no senseful proposition would be what we would find acceptable as an answer to the ‘question’, and that, if we are able to recognize that we had, without thinking about it very clearly, used a question-form modelled on questions that are answerable by senseful propositions, we can come to give up the ‘question’ as not asking anything, as a form of words in which the question-form doesn't get so far as genuinely to express a question.

The second approach allows that there are other sorts of question than those that can be answered by saying how things are. Take, for example, ‘Why did such-and-such star become a white dwarf?’ An answer to such a question will make use of laws of physics, and such laws, on Wittgenstein's view, provide more or less useful ways of describing how things are, but do not themselves say that anything is the case.2 Words used as they are in laws of physics will not be what we want as an answer to ‘Why is there anything?’, any more than will words used as they are when we describe how things are. When he wrote the Tractatus, and in the years when he first returned to philosophy in 1929–30, Wittgenstein recognized a variety of ways in which words might be used; he did not, as is sometimes suggested, believe that there were only three sorts of case: senseful propositions, the senseless propositions of logic, and nonsense. But if we consider these various ways in which words can be used, we can see that none of them will be what we want as an answer to ‘Why is there anything?’ It is not just propositions that state how things stand and laws of physics that won't be what we want. Words used as they are in giving logical necessities, words and symbols used as in mathematics, words used as they are in giving useful definitions, or in rules of translation from one language to another, or in presenting, through the use of variables, some or other expressions: all these various ways of using words can be recognized to be not what is wanted in reply to ‘Why is there anything?’ For us to be here at the limits of language is for there to be no way words might be used which would connect with our question-form and provide an answer. If we can see that there is here nothing to be said, it is because we can see that any way of using words would not be what we want. But if we see that that is the case, we (p. 243) should not be ‘resigned’ at the impossibility of getting what we want, but should recognize that the question-form only appeared to be asking for something that we might or might not be able to get. It's not that we can't get what we want, not that ‘what we want’ is on the far side of the limits of language and therefore unobtainable. The second approach, like the first, is not one in which we reach resignation at the limits of language. In both, we see that the recognition of limits is not the idea that something ‘beyond the limits’ is out of reach.

The two rather similar approaches that I have described are constructed from what one could think of as Wittgensteinian material. The fundamental idea in both approaches is the reverse of what one often sees in discussion of Wittgenstein on ‘what can't be said’. The idea, that is, is not that the thing one is trying to say, or such-and-such sort of thing, lies beyond the limits of what can be said, but that nothing that can be said would be what one is seeking, or takes oneself to be seeking. What I want to consider is the role, or possible role, in Wittgenstein's early thought, of coming to recognize that nothing that can be said would be what one is after.

My question, what the role is of the recognition that nothing that could be said would be what one wants, is really part of a larger group of questions concerning what kind of clarity Wittgenstein took himself to be aiming at in the Tractatus, how that sort of clarity connects with questions about the limits of language, and whether such limits are constraints of some kind. Consider here 6.5, where Wittgenstein says that, to an answer that cannot be put into words, the question (that seemed to want such an answer) also cannot be put into words. Here the reader whom Wittgenstein has in mind has arrived at the idea that the question that she is asking calls for an answer that must be inexpressible. Can that understanding be transformed? For Wittgenstein appears to want to go on to try to transform it. At 6.52, he invites us to think of a situation in which all possible scientific questions have been answered. If we are able to recognize that none of the gazillion speakable answers to questions will reach to our concern with the problems of life, that none of the speakable answers can be an answer to the questions we take ourselves to be putting, we should see ourselves as rejecting all answers, rejecting anything that would be an answer to any question. If we could have, as it were, all answers, we should still not have the answer to the questions we took ourselves to be putting. Here our understanding of ourselves can reach a further point: that the purported questions are not questions at all. (They aren't, that is, questions the answers to which couldn't be spoken; they weren't questions which fail to count as questions because of what is allowed to count as a question, demanding answers which also can't count as answers.) We no longer think of ourselves as having questions which push us beyond where language can reach. There is thus the possibility of a kind of philosophical satisfaction in reaching clarity of this sort, clarity which is not a matter of recognizing and accepting that we cannot do something. Arriving at the ‘limits of language’ from the inside, i.e. seeing what can be said and seeing that nothing sayable will be what we want, may enable us, not to resign ourselves to saying only what can be said, but to see differently ourselves and our askings, and the relation of our askings to the world.

(p. 244) At 6.53, Wittgenstein says that the correct method in philosophy would really be to say nothing except what can be said, i.e. propositions of natural science. He goes on to say what should be done when someone else wants to say something metaphysical. But let us attend for a moment to the place in this method of saying things that can be said. What would make that a ‘method of philosophy’, or an element in such a method, at all? If philosophy is an activity that aims at making thoughts clear, how can talking about the weather or hydrogen or photosynthesis contribute to philosophical clarification? Because there doesn't seem to be an obvious answer, many people have understood 6.53 as if Wittgenstein meant that we should spend our time talking sense (without suggesting that such talk was part of the ‘method of philosophy’), and that the method of philosophy comes into play only when someone wants to say something metaphysical: then we respond to the philosophical provocation. But I think we should take seriously what 6.53 actually seems to say, namely that talking sense is itself a method of philosophy. Sense-making can be directed towards philosophical clarity, can help us to achieve such clarity. If we explore the idea, we can illuminate Wittgenstein's understanding of what kind of limits the ‘limits of language’ are.

The first thing we should note is that, twice in the Tractatus and once again later in the ‘Lecture on Ethics’, Wittgenstein suggests a philosophical method which is clearly related to that of talking sense. The first such suggestion, at 5.631, I shall discuss below; the second such suggestion we have already seen, namely Wittgenstein's appeal to the imagination, at 6.52. We are to imagine that all possible scientific questions have been answered; we have arrived at the gazillion true scientific propositions. Wittgenstein is asking us, at 6.52, not to utter senseful propositions ourselves, but to imagine a vast body of sayings and writings, the sayings and writings that add up to ‘all possible questions’ having been answered. Imagining this to have been achieved is what can play a role in the transformation of our understanding of what we had taken to be the questions we wanted to put. We can come to see that nothing in all those answers could be what we were seeking. As I have noted, such a realization is not in and of itself the realization that we were not asking a question at all. At this point, I simply want to note that Wittgenstein thinks that we can be helped philosophically by seeing that nothing in the realm of sense is what we are seeking, and that talking only sense, or imagining a huge amount of sense-talking, can itself help us to reach the point at which we can say: I see that nothing that made sense would be what I want; I see that I could reject a priori anything you say in answer to my question if it made sense. Talking sense is a method of philosophy if talking sense is a way of enabling someone to see that anything like any of this (all this sense) would not be what she wants. What I am emphasizing here is the difference between thinking of the inexpressibility of something as being a matter of the something lying beyond the expressible and thinking of it as not being anything within the expressible, within what can be said, i.e. thinking of oneself as in a position to reject anything sayable as an expression of the something. How much of a difference this makes may not be obvious, but I shall come back to it. At this point, though, I am concerned to bring out only that talking sense can be a method of philosophy if one thinks of it as directed towards helping someone to recognize that no talking of sense would be a solution to the problem with which she is concerned.

(p. 245) Let me turn here to 5.631, where Wittgenstein imagines writing a book called The World as I Found It, which includes a report on his body and on which bits are subject to his will, and so on. The book has no mention in it of the philosophical subject. Were he to write such a book, it might not be given any philosophical use, and something similar can be said about imagining such a book: the mere account of such a book need have no particular philosophical use. Wittgenstein does, though, say that, if there were such a book, it would provide a method of showing something philosophically significant. Here again we have the idea of the uttering of senseful propositions (the ones in the book) as a philosophical method. But at 5.631, the actual method Wittgenstein is using is that of imagining the book, just as at 6.52 the method was that of imagining the reaching of answers to all possible scientific questions. In context, in the Tractatus, the idea of such a book is taken to be something that can help us to reach philosophical clarity. If we take ourselves to be in search of the philosophical subject, we can see that nothing in the book will be what we are looking for. This of course may only suggest that the philosophical subject belongs in a realm of things not included in the book, the realm of things we can't talk about; and in that case, we are working with the conception of the limits of language, the limits of what can get into the book, as constraints. We take the idea to be that we can't speak of what lies outside, including the philosophical subject. That is how Bernard Williams, for example, takes this section of the Tractatus; he takes Wittgenstein's remark that philosophy can in a sense speak of the self in a non-psychological way (5.641) to mean that philosophy can talk about it in the way in which philosophy is able to talk about anything, i.e. not in senseful sentences. (Williams 1981, 146) On this reading, then, the sense in which philosophy is supposed to be able to talk about the self, which does not lie within the limits of senseful talk, is this: it talks about the self by talking nonsense. Insofar as we remain within the limits of senseful talk, there is something we cannot talk about. On this reading, then, those limits are constraints. How else might one see what Wittgenstein is up to? He says that the I enters philosophy through this, that ‘the world is my world.’ (5.641) What then is the connection between ‘the world is my world’ and the way in which philosophy can in some sense talk about the self?

Let me pause here. I began by quoting Wittgenstein on two ways in which the notion of the limits of language can be understood. It can be understood as a constraint, and there is supposedly another way in which it can be understood, tied to the achievement of what Wittgenstein referred to as complete satisfaction. Wittgenstein's understanding of the limits of language is tied to his ideas about philosophical method, and I have been trying to explore the connections. One of the most important places in which the idea of limits comes up in the Tractatus is the passage in which Wittgenstein speaks of ‘the world is my world’. But before considering this enormously puzzling section of the book, we need to enlarge the focus, and to look in somewhat more detail at Wittgenstein's conception of philosophy and what he thought it could accomplish.

(p. 246) 2. Enlarging the Focus: More About Philosophy as an Activity of Clarification

Philosophy, according to Wittgenstein, is supposed to be an activity that aims at the clarification of thoughts (4.112). He also thinks that, in the activity of attempting to clarify something that purports to be the expression of a thought, we may come to realize that no thought was expressed. The activity of clarification can be carried on in a variety of ways. The introduction and use of the kind of notation Wittgenstein describes at TLP 3.325 is one such technique of clarification. Analysis, thought of as the rewriting of propositions so that their logical form is more clearly open to view, is a closely related technique. I have suggested that the imagined book at 5.631 and the imagined complete set of ‘answers’ at 6.52 play a role in a different technique of clarification. (That method of clarification is used also in Wittgenstein's ‘Lecture on Ethics’, in which we are asked to imagine an omniscient being, who writes everything that he knows in a big book, which thus contains the complete description of the world.3) But there is another, more obvious technique in the Tractatus itself, namely, the provision of a variable, the values of which are all propositions. We specify a variable, according to the Tractatus, by giving its values, and one way in which this can be done is by giving a rule governing the construction of the propositions which will be the values of the variable.4 So, in the case of that variable, the values of which are supposed to be all propositions, the variable can be specified by a general rule governing the construction of propositions. (The propositions constructed in this way will be capable of having sense, but won't actually have a sense unless meanings have been given to the names in the propositions (cf. TLP 4.5).) Philosophy can provide such a general rule for the construction of propositions; and whatever can be the case can be said to be the case by some such constructed proposition.

How does the method just described fit in with familiar ideas about the kinds of proposition which the Tractatus leaves room for, or appears to leave room for? When Wittgenstein says, at TLP 6, that the operation that he has spelled out gives the general form of proposition, does he think that the specification is a piece of nonsense? There are two reasons for answering No.

  1. (a) If we read the Tractatus as holding that, besides senseful propositions, there are tautologies and contradictions, which are senseless but not nonsensical, and that, besides those two categories, there are only nonsensical pseudo-propositions, we are imposing a schematism that Wittgenstein does not himself put forward at any point. (p. 247) Among the kinds of proposition-like constructions that he plainly recognizes are definitions and other rules (3.343). A definition of one of the signs in a proposition can be added to the proposition without affecting its sense, and in that respect it resembles a tautology. (One can note the contrast with adding a nonsensical conjunct to a proposition, which makes the whole proposition nonsensical.) In The Big Typescript, Wittgenstein says that when one adds a rule to a proposition, it doesn't change the sense of the proposition (BT, 189; also Z §321). I think it is consistent with the general approach of the Tractatus to suggest that a rule that helps to clarify any expression or expressions used in a proposition can be added to a proposition without changing its sense.5

  2. (b) We need to separate the actual specification of a variable from remarks that may accompany it, and in particular from remarks that appear to say what kind of meaning the values of the variable have, e.g. that they are propositions. The specification of the variable, distinguished from the accompanying remarks, is itself an instrument of clarification, something that can help us to look at our own use of language and see underlying commonalities in the expressions we use. The philosophical method of laying out a general rule for the construction of propositions is a kind of approach ‘from inside’, related to the method of imagining ‘all possible’ answers to scientific questions, and to imagining (in the ‘Lecture on Ethics’) the big book written by the omniscient being, containing the whole description of the world. These and other methods enable us more clearly to see sense. We may thereby reach a position in which we can see that nothing that made sense would be what we want in connection with the kind of philosophical quest in which we are engaged, or the philosophical difficulty in which we are enmeshed.

I quoted earlier Wittgenstein's remark that the correct method in philosophy would be to say nothing apart from what can be said, and, when someone tries to say something metaphysical, to show the person that he has failed to give a meaning to some or other signs in his propositions. Wittgenstein also says, in the Tractatus, that, if we have a proposition like ‘Socrates is identical’, the reason it is nonsensical is that no adjectival meaning has been given to ‘identical’. The implication is that, if we gave an adjectival meaning to ‘identical’, the proposition would have sense; Wittgenstein explicitly says that we cannot give a sign the wrong sense. (In this case, no adjectival meaning assigned to ‘identical’, and no resulting sense for ‘Socrates is identical’, would be wrong.) But let us here ask what the significance is of the fact that we may want not to give a meaning to a sign or signs in some proposition, if the result would be that the proposition would then make sense. Think of a proposition (p. 248) like ‘A is an object.’ It may be perfectly consistent with the Tractatus that that proposition could be given a sense, if ‘object’ were given meaning as a word for a kind of thing; but the Tractatus can lead us to a further point about such a proposition. If we were to come to see that nothing that could be said to be the case, were we to use ‘A is an object’ to say that that was the case, would satisfy us, we might want not to give it sense. We might, that is, exclude giving it a sense, not because using ‘A is an object’ as a senseful proposition would be wrong, but because we do not want that combination of signs to have any sense.

Here I am suggesting we can see a further element in Wittgenstein's understanding of philosophical method. I shall try to make it clear by contrast with the way in which his philosophical method is often understood. Elizabeth Anscombe, for example, discusses the case of ‘Red is a colour’, and says that it does not express anything that might be false. This is why, on the Tractatus view as she is expounding it, the proposition is nonsensical. That red is a colour is something that supposedly is shown in the use of senseful propositions, but which cannot be said. (Anscombe 1963, 82) On the alternative conception that I have just been suggesting, the methods of clarification of sense on offer in the Tractatus can help us to recognize that we should not want to give a meaning to the signs in ‘Red is a colour’ so that it would be a senseful proposition. If we exclude ‘Red is a colour’ from sense, and say that it is nonsensical, we are not recognizing that, because it cannot express something that can be false, it cannot be a senseful proposition. There are good reasons for excluding it, for drawing a boundary here and saying: ‘We don't want that combination of signs within senseful language.’ But we then really and truly do treat what is ‘on the far side’ of the limits of language as being nothing but nonsense. The limits of language, construed in this way, don't force us not to do something that we might want to do; rather, we have come to see that there isn't something that we want to do. We have come to understand differently our relation to words which have an attraction for us.

I don't want to suggest that all the cases in which we might exclude giving a sense to a form of words are like the case of ‘Red is a colour.’ Anscombe describes a case in which she came to see a kind of unthinkingness in something that she herself had said, ‘It looks as if the sun goes round the earth.’ (Anscombe 1963, 151) If, seeing that her words had meant nothing, she did not want to go on to give them some sense, this would not have been like the case of ‘Red is a colour.’ But what is the difference? It looks as if the difference is that what we want to mean by ‘Red is a colour’ is something that is a necessary truth; and that that is the reason why we won't be satisfied by assigning meanings to the signs in ‘Red is a colour’ that would make the whole proposition come out as what the Tractatus would count as a senseful proposition. And here the attempt to spell out what is going on in the case of ‘Red is a colour’ calls on the idea that there is something that we cannot do: the limits of language constrain us; what is necessarily the case but not tautologously so cannot be put into words. It looks as if my attempt in the paragraph before this one to spell out a difference between Anscombe's approach and the approach that I have been sketching has collapsed.

(p. 249) 3. A Serious Objection to the Idea that the Limits of Language, as Understood in the Tractatus, are not Constraints: Two Parts of a Response

Does Wittgenstein or does he not exclude from sense propositions that express substantial necessary truths, propositions that are not false in any circumstances? The idea that the limits of language are not constraints may seem to run up against the idea that the Tractatus precisely excludes any such substantial necessary truths. For if indeed it does so, then whenever we wanted to say something that was such a necessary truth, we might be persuaded by the Tractatus to give it up, but if we did so, this would be a case of resignation, not satisfaction.

My response to that argument has four parts. I should first note that some of the propositional constructions that we might be inclined to characterize as substantial necessary truths are, when looked at with a more critical eye, anything but. Readers of the Tractatus sometimes take Wittgenstein's own propositions to express claims that are meant to be both substantial and necessary, and that then count as nonsensical because, supposedly, no such claims can count as sense. A good example to consider here is one of the propositions in 5.3: ‘Every proposition is the result of truth-operations on elementary propositions.’ If that is supposed to be necessary and yet not empty, then the word ‘proposition’ as it occurs there cannot be taken to mean the same as ‘the result of truth-operations on elementary propositions’. Further, if 5.3 is supposed to be even plausible and not obviously untenable, the word ‘proposition’ in it cannot mean just any combination of signs that looks like a proposition, or that could be used as a proposition. The idea that the word means anything at all in that context (as opposed e.g. to being a sort of blur that we read past without raising questions) should be regarded as at least questionable. (It is indeed arguable that the aims of the Tractatus preclude the giving of any clear meaning to ‘proposition’ in all the talk leading up to the specification of the general form of proposition.) So, in at least some cases, a careful examination of a proposition that we had taken to be both substantial and necessary may show that we may indeed have, instead, a proposition with a blur in it where we had taken there to be meaningful words.

The case of propositions like ‘A is an object’, ‘Red is a colour’, and so on is different. In these cases an argument that there is in the proposition a word that lacks any meaning in the context can be made through appeal to the kind of notation that Wittgenstein describes at 3.325. Such a notation will sharpen our eyes to different kinds of use of terms like ‘object’ or ‘colour’. The point of such an argument would be to call into question an idea like Anscombe's about ‘Red is a colour’, that it cannot express anything that might be false. The argument would be that it does not express anything, simpliciter, because ‘is a colour’ hasn't been given any appropriate meaning; nor has ‘red’ as a substantive in (p. 250) such contexts. But such an approach does not in fact take us very far, if our concern is with the character of the limits of language, and with the decided impression that the Tractatus gives that the limits of language rule out substantial necessary truths. For it might be asked, on the supposition that we had not given to ‘is an object’ or ‘is a colour’ any meaning in contexts like ‘A is an object’ and ‘Red is a colour’, why we should not give a meaning to the supposedly as-yet-meaningless terms in such a way as to make these propositions come out to be necessary truths.

4. Two Further Parts of a Response to the Objection

Let us turn directly to the supposed ruling out of such truths. This ruling out, it might be said, is done in two ways by the Tractatus: in its explicit remarks, and in the general form of proposition itself, as given by the Tractatus. The explicit remarks that might be appealed to include ‘Whatever we can describe at all could be other than it is’ (5.634) and also the remark about logical propositions (i.e. tautologies), that they alone can be recognized to be true from the propositions alone. No propositions other than tautologies can be recognized to be true or false from the propositions themselves; and the fact that tautologies can be seen to be true from the proposition alone is inseparable from their not actually saying anything (6.1–6.113). Mathematical propositions can also be proved to be correct without any comparison with the facts, and that is why Wittgenstein speaks of mathematics as a ‘logical method’; mathematics, like logic, is not a sphere in which we express thoughts (6.2–6.2321). Philosophy does not result in ‘philosophical propositions’ (which is what propositions like ‘Red is a colour’ or ‘Propositions are truth-functions of elementary propositions’ might be thought to be). ‘The totality of true propositions is the whole of natural science (or the whole corpus of the natural sciences).’ (4.11–4.112) Every possible sense can be expressed by a symbol satisfying the description given in the general propositional form (4.5). The significance of these remarks cannot be seen, though, unless we turn to Wittgenstein's specification of the general form of proposition, at TLP 6; and my discussion of what TLP 6 does is the third part of my response to the objection. In the fourth part, I shall get back to explicit remarks like those just mentioned.

There is an important kind of problem concerning the way I treat TLP 6, but a problem which is too complex to get more than just a mention here. Peter Sullivan has argued that Wittgenstein does not succeed in specifying a variable at TLP 6, and that, even if he had succeeded, there would still be a question what the use might be of the variable. I have tried elsewhere to sketch an alternative approach to the difficulties raised by Sullivan. (Sullivan 2004; Diamond, forthcoming) There are certainly fundamental difficulties with Wittgenstein's ideas about the general form of proposition, difficulties which (p. 251) he explores in Philosophical Investigations, especially in the sections beginning at §89. In what follows, I touch on those difficulties but I don't go into them; my aim is to make clear the role that the specification of the general form of proposition has in the Tractatus.

At TLP 6, Wittgenstein provides a variable, and says about it that it is the general form of a truth-function, and that it is the general form of proposition. A variable like the one provided gives us (or is meant to give us) what is common to some group of symbols. Which ones? The specification of the variable itself gives us the answer to that question. We can say that the variable provided at TLP 6 specifies a way of using signs. The specified use could be called ‘the use as picture-proposition’, where what picture-use is is given by the variable, although we have been helped to see what this amounts to in large part through such metaphors as that of sense as ‘reversible’. What reversibility comes to in the specified kind of use of signs is the role given, in the specification, to truth-functional construction.6 What I am suggesting is that Wittgenstein, in specifying the general form of proposition, is (in intention) specifying language ‘from the inside’. This means that we cannot assume that his various remarks about symbols of the form described are to be read as his taking up a position outside the specification and telling us, from that position, what the specification does. After the specification of the variable, we should recognize that that is what we have: the specification itself, a general rule for the construction of signs. We should, when we read the Tractatus, take seriously the question what our situation is, once we have been given this specification of a variable. If we have a rule for churning out a whole bunch of signs of a certain construction, why should we take it to be anything to do with language, with saying things that we can understand? If the word ‘proposition’ is attached to this variable, and we are told that it is ‘the general form of proposition’, it is something more than a mere label, and can be seen to be something more than a mere label for a calculus of signs, only in the speaking-use of the signs. I can put the point here another way. There are three ways to look at TLP 6.

  1. (1) We can take proposition 6 to provide a method of construction, and use ‘senseful proposition’ as a label for what it tells us how to construct.

  2. (2) We can see our own talk and talk that we understand as the use of signs of the form given in proposition 6; we can see ‘into’ language as we ourselves use it, and see in it the general form that has been specified.

  3. (3) We can take the Tractatus to convey a view about senseful propositions: that all such propositions have the form specified in proposition 6.

I am suggesting that (1) and (2) are available to us, if we read the Tractatus as I think Wittgenstein meant it to be read. That is, the specification of the general form of proposition can be thought of as a rule for constructing signs labelled ‘propositions’, (p. 252) but, for it to be anything more than that, we have to see speaking, making sense, our own talk, as the use of signs of the specified form; we have to see Wittgenstein as having, through a variable, presented language, the language we understand, ‘from the inside’. I am suggesting that we not read the Tractatus in the third way; and I shall come back to this. In Philosophical Investigations §103, Wittgenstein spoke of how the ideal conception of what a proposition is had been for him like a pair of glasses on his nose. The ‘strict and clear rules’ of propositional structure appeared to him to be present in ordinary propositions, but only because the idea of such a structure was being read into them. We can take Wittgenstein's remarks in §§102–3 to indicate also how he had wanted the Tractatus to be read. If he himself could see the ‘strict and clear rules’ of propositional structure in ordinary propositions, his book (I am suggesting) invited its readers to see their own ordinary propositions as the use of signs whose essence they had been given in the general form of proposition. Wittgenstein doesn't want to assert that ordinary senseful propositions are all of the form that he has specified, but rather to lead us to see in our ordinary propositions exactly the formal structure laid out in the Tractatus. We are being led (that is) to put onto our own noses the glasses that Wittgenstein speaks of at §103, and to see our own language with those glasses on our noses. We are, in doing so, imposing an order on our use of language; see PI §105. It is not a matter, though, of our first picking out the propositions, and seeing in them the structure specified at TLP 6. Rather, the book read as it is meant to be read imposes an order in the sense of both identifying some of our talkings as the projective use of propositional signs, as senseful propositions, determinants of a place in logical space, and persuading us that ‘in’ such signs-in-use there is the common essence, glimpsable down deep.

It sounds like a Tractatus thesis to say that Wittgenstein is not putting forward as a thesis that senseful propositions are values of the variable given in proposition 6. We may think he is doing so because we don't see an alternative possible account of what he is doing. I have tried to sketch an alternative view. But we can also ask what exactly Wittgenstein would be doing were he to be asserting that what his variable presented was: all propositions. Denis McManus, for example, describes Wittgenstein as ‘in some sense’ asserting that his general form of proposition is the general form of proposition. (McManus 2009; cf. the similar claim in McManus 2006, 141.) What then is it that Wittgenstein is supposed to be ‘in some sense’ asserting? What it seems McManus means is something like this: Wittgenstein is ‘in some sense’ asserting that the values of the variable that he has produced are the propositions, i.e., that propositions have in common that they are values of Wittgenstein's variable. But any such attempt to give what Wittgenstein is ‘in some sense’ asserting requires that some term like ‘proposition’ be used, and that it not be a stand-in for ‘values of the variable presented at TLP 6’. But surely the Tractatus, if it is committed to anything in this general region, is committed to the idea that the word ‘proposition’ is the sign for a formal concept, the expression for which is a variable, the variable presented at TLP 6. The suggestion that Wittgenstein is ‘in some sense’ asserting that his general form of proposition gives what propositions have in common appears to require that he be giving the word (p. 253) ‘proposition’ a use other than that which the book as a whole can be taken to have laid out. The word can be given any use one wants to give it; but the idea that Wittgenstein is ‘in some sense’ asserting something, the assertion of which requires some distinct unspecified use of the word ‘proposition’, is an idea one should greet with scepticism. The book should teach us that we are all too likely to take ourselves to be using words with some determinate meaning when we are not doing so; and the idea that Wittgenstein is ‘in some sense’ asserting that what propositions have in common is given by his general form of proposition should be regarded as an instance of our fooling ourselves by taking a word with no determinate meaning in the particular context in which it occurs to mean something. The alternative is to say: he himself saw the general form of proposition, as specified in proposition 6, as present in language as he used it, and he wanted to lead his readers to take themselves to be seeing into what was before them in their own talk, and to see the general form that he had laid out as present in it. I am not suggesting that the trouble with McManus's suggestion is that the word ‘proposition’, as we use it (and related expressions, like ‘sentence’, or ‘what she said’, etc.) in ordinary talk, has no determinate meaning. Its use (on Wittgenstein's view) is that of a variable (TLP 4.126–4.1272), while its use in any sentence like ‘What propositions have in common is that they are values of Wittgenstein's variable’ treats the word ‘propositions’ as a term with the logical character of ‘apples’; it is as a term apparently with that logical role that the word has no determinate meaning. (A further argument about these issues is presented in Section 6.)

The significance of my argument is that, if what proposition 6 does (or is meant to do) is fundamentally to present the picture-use of signs, it is not in the business of ruling anything out. Making clear a use of signs does not itself rule out any other use. This is, indeed, evident in the structure of the Tractatus, in which three sets of remarks (the 6.1s, the 6.2s, and the 6.3s) are concerned with three different non-picturing uses of signs, none of which is meaningless.7 What we can see clearly from these three sets of remarks is that there are sign-constructions that may look as if they are picture-propositions, but that have a quite different use. The same point is indeed evident in Wittgenstein's remarks about expressions of the form ‘a = b’: that they are aids to representation. I am claiming then that there is no argument in the Tractatus, about any use of signs, that because it is not the picture-use it is excluded, and lies beyond the limits of language.

I said that we could not see the significance of the Tractatus remarks that explicitly rule out propositions that are substantial and necessarily true except in the light of the specification of the general form of proposition. If we take the general form of (p. 254) proposition to be the specification of a way of using signs, the picture-proposition way, we can think of the defender of substantial necessary propositions as having to provide some account of how the propositions that he takes to be substantial and necessarily true are being used. This would be a matter of providing an account of a mode of use of signs, different from the picturing use, or of providing an account (different from anything the Tractatus takes to be available) of how picture-use in some perhaps considerably modified sense can accommodate necessary truths. Thus, for example, suppose we think, or try to think, of there being, for any necessarily true proposition, circumstances in which it would be false, but which cannot conceivably be realized, and we think of the proposition as having truth-conditions which are necessarily fulfilled. So we have a kind of bipolarity of the supposedly necessarily true proposition, but a bipolarity which does not lie ‘within’ the existence and non-existence of states of affairs: there is no possible situation in which the proposition would be false. At the heart of such an exercise of the logical imagination is the idea of the circumstances which cannot be realized; and at 5.61, Wittgenstein says that what we are in effect trying to do here is to station ourselves, as we set out what can be the case and also the circumstances that cannot be the case, ‘outside’ the limits of the world. We are, as it were, thinking or trying to think in a language in which we can consider, even if only to rule out, things that we take to be not thinkably the case.

I suggested above that the position in which the defender of substantial necessary truths is put by the Tractatus is that of needing to explain in what way the propositions supposedly expressing such truths are being used, on his view, since they are not being used in the picture-proposition way. When Anscombe writes about Wittgenstein's account of propositions as pictures, she says that his theory is powerful and beautiful, and that ‘there is surely something right about it if one could dispense with “simples” and draw the limits of its applicability’. (Anscombe 1963, 77) One thing she wanted to exclude from its applicability was statements of what is necessarily the case which are nevertheless not tautologies in the Tractatus sense. Suppose we agree with her and say: ‘Fine, necessary truths should be excluded from the applicability of what Wittgenstein says about picturing. We shouldn't say that there can't be such things, merely because they aren't picture-propositions.’ But what the Tractatus is at least implicitly suggesting is that the cases we have in mind of such truths are cases we have not examined, and that there are more problems than we may have suspected in laying out the use of such sentences themselves, or the use of the words within them. Take one of Anscombe's own examples: ‘ “Someone” is not the name of someone.’ (Anscombe 1963, 85–6, 162) She shows that such a sentence can have a use. But what is that use? Her sentence is meant to clarify one use of ‘someone’. Nothing in the Tractatus rules out there being proposition-like constructions, the use of which is to help us to use signs in statements of what is the case; and ‘ “Someone” is not the name of someone’ is meant to be helpful in such a way. A sentence which is used as an ‘aid to representation’ (‘Behelfe der Darstellung’, see TLP 4.242) may seem itself to be a representation of something that must be the case, and therefore to fall under a supposed Tractatus interdiction, but that would be a misunderstanding. The misunderstanding arises from the propositional look that an ‘aid to (p. 255) representation’ may have, which stops us seeing its quite different sort of use. Anscombe's example is actually quite complex and I cannot go into details. (See Diamond 2004.) Here I want to argue only that the specification of the general form of proposition excludes nothing at all. It specifies a use of signs, the picture-proposition use. The various remarks in the Tractatus which do explicitly rule out substantial necessary truths need to be thought about as, in a sense, tasks for us. If Wittgenstein says ‘Whatever we describe at all could be other than it is’, can this be read as excluding nothing? It can be read (I am suggesting) as a challenge to anyone who takes herself to be describing something, something which could not be other than it is, to consider whether the kind of way in which words are supposedly being used to describe in the sentence in question can be clarified. Insofar as there really is a claim made by remarks like ‘Whatever we describe at all could be other than it is’, it is that whenever we take ourselves to be describing something which could not be other than it is, the impression that that is what we are doing reflects unclarity of mind. The unclarity may involve an implicit appeal to the picture-proposition use of words, an appeal which would break down into incoherence if made fully explicit. Wittgenstein's metaphorical description (TLP 5.61) of our wish to station ourselves ‘outside’ the world suggests just such a half-meant appeal to a kind of picture use. We might see such an attempt in, for example, Schopenhauer's talk of the thing-in-itself as ‘something’, something groundless which underlies phenomenal happenings and which withdraws itself from investigation. Speaking of this ‘that which’, this ‘what’, this ‘something’, uses the apparatus of quantifier-and-variable, but Schopenhauer's language doesn't involve our quantifying over anything about which we can talk in any way without quantifiers.8 He is as it were gesturing towards a language which we ourselves cannot use; he is (with the use of quantifiers) invoking logic, but a logic which is not the logic of the language we speak in describing the world. We may have an extremely vivid conception of ourselves as describing something when we do this sort of thing, e.g. as in the case of Schopenhauer, this talking about the thing-in-itself by using quantifiers. But we should take seriously that there is nothing here but Luftgebäude, i.e. that ‘Whatever we describe at all could be other than it is’ excludes nothing.

Many remarks in the Tractatus play a role in the way Wittgenstein leads the reader to proposition 6; and many of these remarks use the word ‘proposition’. They have a role that can be compared with that of auxiliary construction lines in a geometrical proof. What we have at the end of the geometrical proof is the construction itself, the lines that form part of it. The auxiliary lines helped us to get there, but they are not part of what we are left with at the end. We can think of much of what we have in the Tractatus as auxiliary constructions, combinations of signs, using the word ‘proposition’, that lead us up to proposition 6, which itself enables us to have in clear view a particular use of the word ‘proposition’. Once we have that clear view, we can see also that those earlier remarks, those sign-combinations with the word ‘proposition’, were blurry, said nothing clear at all. This was not because they were necessary truths, but because they contain a word (p. 256) with no determinate meaning in the context,9 and indeed their function in the book requires that they not have any determinate meaning. (This point bears also on the use of the word ‘proposition’ in discussions of these remarks, e.g. my statement in section 2 that Wittgenstein takes philosophical method to include the provision of a variable which has all propositions as its values.)

My argument in sections 3 and 4 has been largely negative. I have tried to show that the Tractatus does not give us a conception of the limits of language which puts something that we might want to do, namely, express substantial necessary truths, on the far side of the limit. I have also had a positive aim, that of making clear how we can read the provision of a variable that supposedly gives the general form of proposition as giving the limits of language ‘von innen’. The positive and negative aims of these two sections hang together, and my claims about 6.1, 6.2, 6.3 and about ‘aids to representation’ bear on both aims. That no uses of language are excluded is tied to those various uses being clearly not excluded. The recognition of various non-picturing uses helps us to see how to investigate the cases we take to be excluded. (The issue of the supposed exclusion of substantial necessary truths is important and I shall return briefly to it in section 6.) The structure of the remarks in the 6s bring out an extremely important point: that there is a real variety of cases in which proposition-like constructions, although they do not have the use specified in the variable at TLP 6, may appear to us to be representations or reports or descriptions of something that is the case. Propositions of logic may seem as if they are totally general truths; equations of mathematics may appear to be reporting mathematical relations; laws of mechanics may seem to be true propositions; ‘Good is simple and unanalysable’ looks as if it describes something; the sceptic may seem to be claiming something that is true or false. Whether such sentences have a use other than the picture-proposition use or no use at all needs to be investigated in each sort of case.

Sentences of the form ‘P entails Q’ can illustrate how easy it is to miss what is at stake. Once one sees that such sentences look as if they are a kind of report but that they are, if true, necessarily true, and yet not tautologies, it is easy to go on to take such sentences to be excluded by the Tractatus as nonsensical, and indeed to read them as a kind of incoherent attempt to consider logical space from outside it. Such sentences do, however, have an important kind of use that is different from the picture-proposition use, and which is not in any way ruled out by the Tractatus.10 That they can have, as a kind of ‘aspect’, the aspect of being a report brings out that the existence of such an aspect doesn't itself settle whether they are nonsensical or not. All the various sorts of proposition-like constructions discussed in the 6s, as well as the aids to representation mentioned elsewhere in the book, can have the aspect of being a report or representation, which can mask what we are doing with words when we utter them. Even if we recognize that the sentences are not doing what they appear to be doing, the report aspect may block our (p. 257) capacity to see the quite different kind of use that such a sentence may have, and we may take the sentence in question to be, on the Tractatus view, an incoherent attempt to go beyond the limits of language.11

5. Viewing the World as a Limited Whole

Wittgenstein wrote in the Tractatus that the aim of philosophy is the clarification of thoughts, but he also said, in his letter to von Ficker, that the point of his book was an ethical one. His book, he said, had two parts, the one he was presenting to von Ficker plus everything that he had not written. The book drew the limits of the ethical ‘from the inside’. But in what sense a drawing of the limits of the ethical ‘from the inside’, since the ethical was precisely what he had said nothing at all about? The limits are hardly drawn ‘from the inside’ of ethics. And if the activity of philosophy, including the activity constituted by the Tractatus, aims at the clarification of things that we do actually say and think, how does keeping silent about ethics belong to the activity of philosophy? We can get some insight into Wittgenstein's ideas here by looking at the analogies that he saw (I believe) between what a work of art can do and what a philosophical work can do.

A work of art can transform the perspective of the reader or viewer or listener; and so can a philosophical work. A work of art can be thought of as treating of something, even when its concern with the thing is not explicit, and similarly for a philosophical work. The analogies in fact go somewhat further, as we can see by turning to what Wittgenstein says about how a work of art can transform the perspective we have on something. His remarks come from before and after the period of the Tractatus itself. In October of 1916, he wrote a number of remarks about the relation between ethics and aesthetics. (NB, 83) The work of art, he said ‘is the object seen sub specie aeternitatis’ and the good life is the world seen ‘sub specie aeternitatis’. Instead of seeing things as we usually do, from the midst of them, the work of art lets us see them ‘from outside’, ‘with the whole world as background’. He went on to consider the example of contemplating a stove, first as a thing among things, and then ‘as a world’: one can see the stove as having the dignity of being what it is, a dignity that is not in view in one's ordinary dealings with the stove.

In a fascinating essay on Wittgenstein's thought about such matters, Michael Fried suggests that we connect these early notes with remarks that Wittgenstein made in 1930, in which he spoke about how a work of art can alter the perspective we have on an ordinary thing. (Fried 2007) Without art, the particular thing is simply a fragment of nature like any other thing; but the work of art can force us to see it entirely differently, to see it in the right perspective, as something worthy of contemplation. In these notes, Wittgenstein went on to say that there was a way of capturing the world sub specie aeterni other than through the work of the artist; thought can do this: ‘it is as though it flies above the world and leaves it as it is observing it from above, in flight’. (CV, 4–5) I am (p. 258) suggesting that a philosophical work, on Wittgenstein's view, can transform our perspective on the world, and that it can do this through the presenting of senseful language ‘from the inside’. I mentioned another element of the analogy between a work of art and philosophy: that a work of art can treat of something indirectly, can treat of something without explicitly mentioning it, and that is exactly what he took philosophy to be able to do. Wittgenstein's remarks, at various times, about Tolstoy's Hadji Murad reflect his idea that the story has ethical significance despite its not dealing in any direct way with Tolstoy's conception of how we should live. Just as that story is ethically significant in part through keeping silent about ethics, so the Tractatus can be ethically significant through keeping silent about ethics. It can treat ethics through presenting nothing ethical; it can treat ethics (this is what I want to suggest) through presenting the general form of senseful proposition.

How then can philosophy be thought capable of transforming our perspective? I can begin to explain by going back to the idea in the Notebooks about how we see things prior to the transformation that can be effected by art: we see things ‘from their midst’. Prior to the transformation that can be effected by philosophy, we are similarly in the midst of things, immersed in seeings and doings and sayings in the world. We investigate whether things stand this way or that; we do things to bring about this or that effect; we say that things stand thus-and-so. And, while we may well think of philosophical questions as deeply different from ordinary questions, we don't really take the distinction deeply enough.12 And similarly for ethics: we take questions about the meaning of life, the nature of the good, and how it is to be discovered to be questions which differ in a variety of ways from scientific questions, but we take them nevertheless to be, after all, questions. We remain, as we see ourselves in our asking of such questions, ‘in the midst of things’; the implicit conception of what it would be to have a solution to the questions we ask is that it would add something to what we know to be the case. What difference, then, can be made by a philosophical approach that presents senseful language, that gets it clearly into focus for us? When we are in the midst of sense, talking sense, asking and answering questions, sense is not something that is an object of our awareness; or, rather, insofar as it does become an object of awareness at all, it is as something about which we may ask questions. So we remain within the asking and answering of questions. Wittgenstein's presenting of senseful language makes possible a standing back from the asking and answering of questions.

The standing back from the asking and answering of questions can be thought of as possibly leading to a transformation of our understanding of what it is to engage in philosophy, and can also be thought of as possibly leading to a transformation of our understanding of what ethics is. These are not two separable kinds of transformation. In the background of my discussion is a contrast that I think runs through Wittgenstein's discussions of ethics from the time of the Tractatus until at least 1930. He spoke in 1929 of (p. 259) the importance of ending the ‘chatter’ about ethics, the ‘Geschwätz’, but this desire to end the chatter co-existed with his deep respect for the tendency to come out with various forms of words expressive of a kind of responsiveness to life. (NTW, 12–16) Some ways of talking about ethics should be stopped, and others not. The contrast is implicitly present at the end of the Tractatus, when Wittgenstein says that if one overcomes the propositions of the book, one will see the world rightly. The suggestion appears to be that there are ways of speaking and thinking (tied in some cases to one's not having overcome Wittgenstein's propositions) that reflect seeing the world wrongly. When he spoke later about the importance of ending the chatter about ethics, related ideas are in play: philosophical talk about ethics reflects and encourages a kind of failure to see the world rightly. I believe he had Moore's Principia Ethica in view both in the Tractatus and in the later remarks, and I shall return to this point.13

At 6.4 Wittgenstein says that propositions are of equal value. The first sentence of 6.41 says that the sense of the world lies outside the world; the last sentence says that any genuine value must lie outside the world; and in the middle, we are told that genuine value, if there is any, lies outside all happening and being-so. So the idea of an ‘outside’ is pretty insistently rubbed in. What is inside is: everything being as it is and everything happening as it happens. Obviously, Wittgenstein's remarks about what can only be ‘outside’ cannot belong to presenting senseful language from the inside. When he makes similar remarks in his ‘Lecture on Ethics’, their role is clearer: he speaks there of the tendency he himself has, and that (some) other people have, to ‘go beyond significant language’. (LE, 11–12) That tendency has, inchoately within it, two elements which can be brought to awareness; the tendency itself can thereby be changed. The tendency to ‘go beyond significant language’ is the tendency to reject ordinary propositions, descriptive of how things do stand in the world, as not capable of satisfying one's desire to talk about what genuinely is of value, plus the desire to talk beyond ordinary descriptive language in order to speak of what is genuinely valuable. The tendency is evident in, for example, discussion of whether the Good can be defined, in theorizing about ethics, where ethics is taken as a sphere of discourse. What is noteworthy here is that, at this point, ‘In the world, everything is as it is, and everything happens as it happen's leads, or seems to lead, outwards. There is no finding satisfaction there in the world in the attempt to understand what genuinely is of value.

How, then, can philosophy affect the tendency ‘to go beyond significant language’? The activity of philosophy clarifies what belongs to ordinary senseful propositions; this can be done while remaining ‘within’ language, in the sense that there is no taking up of a perspective on language in doing so, no making of claims about language from some philosophical perspective. The lineaments of senseful language are put forcefully in front of us; our attention is drawn to them. The more clearly we see language – that language the propositions of which we understand and can use – the more clearly it will appear to us that, in ethics, no proposition at all can be what we want. This is the recognition reflected (p. 260) at TLP 6.42, and in the first clause of 6.5: the answer we want cannot be put into words. What this recognition involves, in terms of the original inchoate tendency, is the rejection of the idea of a sphere of ethical discourse, of there being a kind of proposition dealing with ethical matters. Wittgenstein is not distinguishing factual from evaluative propositions. (In fact, he has no objection to most evaluative propositions, which, on his view, would not belong to ethics as he understood it.14) Clarity of the sort reachable by presenting senseful language ‘from inside’ can enable us to see that the idea of ‘propositions of ethics’ is itself confused. Wittgenstein's view here is very different from (for example) what we have in his discussion of the laws of mechanics. Statements of such laws are not propositions but a kind of aid in the construction of descriptive propositions, whereas the very idea of ‘ethical propositions’ involves a misunderstanding of our own desires. The clarity that Wittgenstein thought could be reached about ethics through presenting senseful language ‘from the inside’ was essentially clarity about the fact that nothing that could be said would be what we wanted, that the very fact that some proposition was intelligible was enough to show that it was beside the point.

What I have said so far is only part of the story. Having our attention drawn to senseful language may do two things at once. The first is what I mentioned, the recognition that no proposition could be what we had wanted. In order to see what the second thing is, we need to return to ‘In the world, everything is as it is, and everything happens as it happens’, which, as I mentioned, at first leads our attention ‘outwards’. We should note here TLP 5.525: the possibility of a situation is expressed by an expression's being a proposition with sense. The general form of proposition gives us the rule by which all senseful propositions can be constructed. (It also gives us the construction of tautologies and contradictions.) In giving us senseful propositions, it expresses the possibility of each and every possible situation, everything that can happen, everything that can be the case. It gives us (that is) ‘the essence of the world’; cf. 5.4711: to give the essence of a proposition means to give the essence of description, i.e. to give whatever can be said to be the case and thus the essence of the world. In giving us at one and the same time senseful description and world, the general form of proposition puts before us what the will (the will as ethical subject) can attach to.

Take now again ‘In the world, everything is as it is, and everything happens as it happens.’ If we connect this sentence with what one might think of as two aspects of the general form of proposition, then the general form of proposition gives us at once both language and world: senseful language on the one hand, and all possibilities of being the case and not being the case, of happening and not happening, on the other. Think of ‘In the world, everything is as it is, and everything happens as it happens’ as emphasizing the second aspect of the general form of proposition, its giving us world. I said that initially ‘In the world, everything is as it is, and everything happens as it happens’ goes with the idea (p. 261) that the desire to find what is genuinely of value must take us beyond all happening and being the case. It seemed as if, if one could not reach value that lay beyond the being-so of this and that within the world, one could not reach value at all. Thus ‘In the world, everything is as it is, and everything happens as it happens’ went with a kind of dissatisfaction with the entire realm of the accidental. But this is an essentially ethical matter; we can re-see ‘In the world everything is as it is, and everything happens as it happens.’ That the facts of the world are whatever they are can become a changed perspective on the world.

In his ‘Lecture on Ethics’, Wittgenstein said that he was ‘tempted to say that the right expression in language for the miracle of the existence of the world, though it is not any proposition in language, is the existence of language itself’. (LE, 11) The presentation of senseful language may draw our attention to the language that is the language we understand, and this drawing attention to language can, as we might say, let us see the world in a different perspective, as being whatever it is, everything standing as it does stand. Speaking of this as a ‘perspective on the world’ is itself a kind of figurative language, since, insofar as a sense has been given to having a perspective on something, the something in question is an object or event or historical movement, or some other particular thing or things; but ‘the world’ in the context of talk of a ‘perspective on the world’ has not been given a use as an expression for such a thing. A perspective is from this or that particular spot, or reflects particular interests and concerns; and no such thing is in question here. To speak, then, about the philosophical activity of achieving clarity about language as ‘capable of changing one's perspective on the world’ is to use words in an irreducibly figurative way. Let me put the matter here another way. Having our attention drawn to senseful language may lead us to stand back from the asking and answering of questions, as we recognize that no satisfaction is to be had from any answer. We are, that is, enabled to contemplate things, but no longer ‘from their midst’, rather from a point of view that sees what is the case as whatever indeed is the case; how things are is not the centre of concern, and that they are is, in a sense, open to view. Open to view, I am suggesting, in and with the contemplation of ‘the language’. It is as if by presenting the general form of proposition, Wittgenstein had drawn a circle around the totality of language and made its ‘thereness’ for us open to view. Its ‘thereness’ stands for the ‘thereness’ of the world. The upshot of this account is not that 6.41, with all its talk about value being ‘outside’ the world, is rejected, but it is understood quite differently from the way in which we may at first try to understand it. For we may at first see Wittgenstein as suggesting that value is something that we cannot reach with words, and that resignation is therefore appropriate. A figurative use of the language of ‘outside’, a figurative use of the idea of seeing as from outside, and of the world as something ‘seeable’ is part of a perspective that we may come to occupy.15

(p. 262) There is more to be said about ‘In the world, everything is as it is, and everything happens as it happens.’ I have suggested that, when we first read the sentence in 6.41, it can be taken to express dissatisfaction with the realm of the accidental, dissatisfaction that propels us to look ‘outside’ for what is genuinely of value. There is, though, a further kind of dissatisfaction with the realm of the accidental, with the fact that things happen in the world as they happen: a dissatisfaction with the independence of the world from what we want or choose or try to bring about. This is the kind of dissatisfaction expressed in a pure form in the Grimm tale of the fisherman and his wife, when the wife, who becomes emperor and pope, is nevertheless dissatisfied because things still happen independently of what she wills; she is not God. This sort of dissatisfaction can be read into suicide, as if, in committing suicide, one were responding to the independence of the way things go by putting to an end the very world that is independent of one's will. If the way things go is not as one wants it to be, one's will is turned against the very existence of the world.

Dissatisfaction with the accidentalness of things takes a variety of forms, and here I want to include the sort of dissatisfaction which can be taken to underlie ethical ‘chatter’. I'm suggesting that, just as Wittgenstein was inclined to read a kind of profound ethical dissatisfaction in suicide, he read a kind of ethical dissatisfaction in the spirit in which ethics was discussed. Speaking in the ‘Lecture on Ethics’ of ethics as he understood it, Wittgenstein said that it doesn't ‘add to our knowledge in any sense’ (LE,12), a remark that may well be directed against Moore, who quite specifically claims that ethics is a science, and that it is directed towards knowledge. (See e.g. Moore 1922, 20.) The pursuit of ethics, understood Moore's way, can be taken to express the desire that there be a ‘being-the-case’ that is not accidental, and that is the foundation of Value. The problems of life can thus come to appear as questions, to which ethics provides a systematic approach. What might be thought to be wrong with this can be seen if we look at TLP 6.43, where Wittgenstein speaks of the ethical will as expressed in relation to the world as a whole, the world as that in which everything is as it is, and everything happens as it happens. The will can be exercised ‘happily’ in relation to the world thus understood, or ‘unhappily’. Ethical ‘chatter’, in seeking to turn ethics into a subject-matter, can be read as a kind of ‘unhappy’ responsiveness. The will to ‘go beyond’ senseful language, and to treat ethics as questions-with-answers manifests a kind of ethical spirit. How one takes ‘the limits of sense’ is itself an ethical matter. Wittgenstein isn't suggesting that what ethics demands is that we resign ourselves to remaining within the limits of sense. The point, I think, is rather that the idea that resignation is what is at stake would itself be an expression of unhappiness in relation to ‘In the world, everything is as it is, and everything happens as it happens.’

We can now see also what Wittgenstein may have meant in saying that ethics and aesthetics are one (TLP 6.421). I spoke of the way in which having our attention drawn to senseful language may enable us to contemplate the ‘thereness’ of the world, in which things are however they are, and things happen however they happen. Ethics and aesthetics are one in their relation to the world, the essence of which is given by the general form of proposition. The idea, in Wittgenstein's conception of ethics and aesthetics, is of (p. 263) a transformation in our relation to all that is accidental, all that happens and ordinarily absorbs us and matters to us, all that we usually take ourselves to know, to be aware of. Insofar as ‘In the world, everything is as it is, and everything happens as it happen's can be for us satisfaction or dissatisfaction, there is will, and thus ethics.16

The sentence ‘In the world, everything is as it is, and everything happens as it happens’ serves, I said, to emphasize the way in which the general form of proposition presents ‘the essence of the world’, in connection with the idea that the general form of proposition had two aspects. But talk of these two aspects, the presenting of all propositions and of all possible situations, the presenting of language on the one hand (that is) and world on the other, inevitably makes it seem as if we are talking about two matching systems of possibilities. This is a conception that we make use of in working our way through the Tractatus; but it is a misleading conception, for there are not here two anythings. If we say that something is the case, then what we have said may be the case. That last remark looks as if it relates language and world, saying and being the case, although it is quite empty. Here I want to make a connection with the sentence that is focal for much of section 7, ‘The world is my world.’ That sentence and ‘In the world, everything is as it is, and everything happens as it happens’ are what one might call propositional misrepresentations of the general form of proposition. They are sentence-constructions that help us to see what significance Wittgenstein took there to be in his presenting of sense ‘from the inside’. Although propositional in form, in look, they are essentially figurative in their appeal to the idea of ‘the world’ as an object of thought. Such ineliminable figures are characteristic of what Wittgenstein spoke of, in 1929, as the ‘thrust against the limits of language’ that he identified with ethics. And it is worth pointing out here too that I have relied in my discussion of Wittgenstein on a figurative use of the terms satisfaction and dissatisfaction. The meaning these terms have been given is tied to contexts in which the object of satisfaction or dissatisfaction is some particular thing or something that is the case, whereas, in my talk here about Wittgenstein on ethics, I have repeatedly used the terms as if the object of satisfaction or dissatisfaction were the limits of language (in the case of Moore) or the world in its essential nature. And insofar as my talk of will has been tied to such figurative talk of satisfaction and dissatisfaction, it too is a kind of figurative use.

It is not in general a good idea to use remarks from before or after the Tractatus to explain what is going on in the book itself. I have, however, done just that; I have taken the continuities noted by Michael Fried between the Notebooks and Wittgenstein's remarks in 1930 on seeing the world sub specie aeterni to be present also in the Tractatus; and I have made use of Wittgenstein's remarks from 1929 about our inclinations, in thinking about ethics, to use words and expressions taken over from contexts in which they have a determinate meaning, but to give them no meaning in the new context. I have been looking at the significance for ethics of Wittgenstein's presenting the limits of language ‘from inside’, but those words should be brought into question too. For no sense has been given to ‘inside’ and ‘outside’ in speaking about language.

(p. 264) 6. Philosophy and the Limits of Sense: Method and Essence

When he wrote the Tractatus, Wittgenstein accepted the philosophical idea that philosophy is concerned with what is essential, but he transforms that concern. He provides a new understanding of what it is for philosophy to treat of the essence of something through his conception of what it is for philosophy to treat of what is essential in a symbol. What is essential in a symbol is what all symbols that can serve the same purpose have in common (TLP 3.341). What they have in common is presented by the general form of the propositions which contain the symbols (3.312). Thus it is presented by a variable, the values of which are all those propositions (3.313). The stipulation of the values of the variable gives us the variable itself. The stipulation is a description of the propositions which are the value of the variable; and Wittgenstein adds that the only thing that is essential to such a stipulation is that it is merely such a description and ‘states nothing about what is signified’ (3.317). These remarks provide an outline of a philosophical method for presenting essence, in which what is essential to the method is that it does nothing but present symbols that have something in common. The method does not involve saying anything about them; it remains within language in simply laying out (in whatever way) the symbols that have something in common through which they serve the purpose that they do serve. If, for example, it is presenting names, the method does not involve saying ‘and these are names’. That the symbols presented are ‘names’ (if that is the label you attach) is nothing but what you see in the symbols themselves.

The presenting of the general form of proposition is thus an application of a general idea about philosophical method. My argument in section 4 about the presenting of the general form of proposition, that it merely presents a use and does not rule out any use, is meant to apply also to Wittgenstein's method understood more generally. The method could be described as a method through which philosophy presents what is essential, but at the same time it is meant to reshape our conception of what ‘treating of the essential’ is. It is meant to allow us to see a kind of philosophical activity entirely different from the putting forward of propositions about what is essential as philosophy treating the essential. You treat of something essential through laying out signs with the way they are used. What you have accomplished in what you have done has to be clear in what you have done. There is not going to be any matter of adding: ‘and this is what propositions’ (or names or two-termed relations or whatever) ‘have in common’.

The philosophical error that Wittgenstein is combatting is that of thinking of the contrast between essential and accidental too superficially, an error that is reflected in the use of what we take to be substantially necessary propositions in talking about what is essential. To say (for example) ‘propositions are pictures’ is to use the word ‘proposition’ as if we were talking about a class of things that had a certain property; but ‘proposition’ hasn't been given that sort of meaning, and (if we achieve philosophical clarity) we shall not want to give it that sort of meaning. Thinking of the statement ‘propositions are (p. 265) pictures’ as if it were a necessary truth (or as if it were something that would be a necessary truth if per impossibile the Tractatus allowed for such truths) makes the difference between essence and accident too slight. Essence is, for Wittgenstein, a matter of the generality of a variable. The propositions of the Tractatus are not attempts to say something that we need to be resigned about our inability to say. Resignation is not in place. Essence, we might say, cannot be spoken about, but does show itself.

What I have argued here puts in another way the argument given by Michael Kremer, that the difference between saying and showing is the main problem of philosophy. That is, it is a problem in that we misunderstand the character of the difference, and make it too slight. (Kremer 2001, 61–5) Insofar as the specification of the values of a propositional variable is the form that the presentation of what is essential takes, on the Tractatus view, there is a sense in which what we do in presenting an essence can fail to be what we were trying to do, but such a failure would not be a matter of our getting the essence in question wrong, but rather of our giving something else than what we meant to give.17 If, when you stipulate the values of a variable, there is no going on to say e.g. ‘and this is the general form that all propositions have’, you aren't making a philosophical claim about what the values of the variable that you have presented all are. The idea that the activity of philosophy doesn't result in such claims is inseparable from Wittgenstein's view that his book was not a textbook, a book which presented a doctrine. The presentation of a kind of symbol, of a mode of use of signs, including the presentation of picture-proposition use, is not capable of truth or falsity, and this is part of what is involved in Wittgenstein's claim that the result of philosophy is not ‘philosophical propositions’ but clarification of propositions. There is, in intention anyway, a profound commitment in the Tractatus to abstention from the putting forward of theses, and I have tried to show how that commitment is connected with the idea of philosophical method as proceeding from ‘within’ language, and to the idea that the limits of sense are not constraints on what we can say.18

(p. 266) The philosophical practice of the Tractatus itself involves a kind of reflexiveness: the presentation of the general form of proposition, insofar as that succeeds as philosophical clarification, makes plain that the concern of philosophy with what is essential is not a matter of its having a subject matter, truths about which it comes up with; insofar as the method succeeds we shall not take philosophy, in its concern with what is essential, to be different from science only in the questions it asks and in the methods through which it arrives at answers, but not different from science in asking questions and arriving at answers. That (as I have suggested) would not make the distinction deep enough. But the method of presenting the general form of proposition, insofar as it is a case of presenting an essence through description of symbols, uses the method the point of which can be seen only through the achievement, through the method, of clarity about language and logic. The achieving of philosophical clarity is possible through the directing of attention to what is in some sense before us (though Wittgenstein's idea of what was ‘before us’ depended on those spectacles on his nose); and the propositions of the Tractatus, though they certainly are meant to have a role in the re-directing of attention, are not themselves failed attempts to say what we have to resign ourselves to being unable to say. I have tried here to emphasize that Wittgenstein's intention was to make clear the difference between the scientific asking and answering of questions and the philosophical task of clarification: to make that distinction go deep enough. He later thought that he had failed to do so, and that in various ways the conception of philosophy in the Tractatus was still in the grip of the idea of the method of science.

7. Philosophy and the Limits of Sense: The Philosophical I and the Challenge of Solipsism

A general principle of my discussion is that it is a mistake to argue that, of the various forms of philosophical solipsism, Wittgenstein is concerned only with solipsism of some particular type. Very different modes of treatment of the self and subjectivity, including those of Schopenhauer, Lichtenberg, Weininger, and Russell, raise issues all of which are present in Wittgenstein's very condensed treatment of solipsism.

I shall look briefly at three subtle discussions of Wittgenstein on solipsism, which I think will help us understand the issues. In discussing the Tractatus view of solipsism, Anscombe summarizes some remarks of Wittgenstein's from 1929–30 about the language we use in describing experience.19 We could have a language in which someone, A, is (p. 267) treated as a kind of centre for talk of experience. People using this language say ‘There is pain’ when A is in pain; and they say ‘X is behaving as A behaves when there is pain’ when, as we might say, X is in pain. Similarly, with ‘It is thinking’, said when A is thinking (where the ‘it’ is analogous to that of ‘It is snowing’) and ‘X is behaving as A does when it thinks’. It was G. C. Lichtenberg from whom Wittgenstein picked up the idea of this way of speaking, and I shall refer to such a language as a Lichtenberg-language, and to the centre of such a language as an L-centre.20 Anyone can be taken as L-centre of such a language; but, Wittgenstein notes, the language in which I am the L-centre ‘has a peculiar but quite inexpressible advantage’ over the language with anyone else as centre. The various languages are all capable of representing the same situations: they are translatable into each other, and the same reality ‘corresponds to them all and to the “physical language” ’. Anscombe comments that this passage is quite close to Wittgenstein's thought about solipsism in the Tractatus. Her point is that the language with me as L-centre does have an ‘absolute’ advantage, on Wittgenstein's view: ‘I am the centre, but this is inexpressible.’ On Anscombe's reading of the Tractatus remarks about solipsism, the I has a unique position in relation to language and world, but this is not expressible. Marie McGinn, in her discussion of the same group of Tractatus remarks, ascribes to Wittgenstein the idea that we need a distinction between the empirical subject and the subject who represents the world to himself. She argues that, although Wittgenstein uses the first-person pronoun in giving his view, his notion of the subject ‘is not essentially first-personal’; I can recognize others as subjects. Thus, on this reading, there is a multiplicity of non-empirical subjects allowed for by the Tractatus remarks on the self; there isn't one unique subject with an inexpressible position. (McGinn 2006, esp. 275–6) B. F. McGuinness, discussing the same passage, says that Wittgenstein holds that language has to have a centre and, when I speak or think, I am that centre; but this is equally true of whoever speaks. He goes on: ‘When I speak or think, it is the World-soul, die Weltseele, speaking,’ and this is equally true of anybody else. The World-soul is something in which we can all participate, and in which we should all participate. (McGuinness 2001, 10)

Anscombe, then, thinks that the Tractatus view is that there is, inexpressibly, a unique non-empirical self; McGinn ascribes to Wittgenstein the idea that the notion of the non-empirical subject allows for many such subjects, each of which can represent how things are, can judge what is the case. And McGuinness allows for a multiplicity that is in some sense not a genuine multiplicity but a multiple participation in the World-soul. Reading these discussions, we may take them to treat two inseparable questions about what the non-empirical subject is and whether there is one such subject or many or many-but-in-some-sense-one. It looks as if these questions call for answers that lie beyond what counts as sayable. It may seem, then, that Wittgenstein's remarks on solipsism and the (p. 268) philosophical I constitute a particularly difficult case for anyone who wants to argue that Wittgenstein's conception of the limits of language doesn't demand a kind of resignation to there being things that supposedly cannot be expressed.

Anscombe gives us a place to start: with the idea of an experiential language in which I express what I experience, not through the use of ‘I’ but through taking myself as L-centre in the way she describes. I think she is right in suggesting that the line of thought which is worked out in some detail in the 1929/30 notes and the related conversation with Waismann can help us to see what is going on in the Tractatus discussion of solipsism.21 It helps us to see how the solipsist's view crumples, and an understanding of the way it crumples is essential to a grasp of the remarks in the Tractatus about the self. But the line of thought she describes is relevant to the later stages of the crumpling of solipsism (5.64). I shall give one way of imagining how the crumpling might go, starting with the earlier stages. (I use the idea of a Lichtenberg-language in formulating the later stages, but in fact the argument could be re-formulated using a more generic account of an experiential language with myself as centre; the important points would be (a) the absence of any terms for any owners of experience and (b) a radical asymmetry between the treatment of ‘my’ experience and that of others.)

In the initial phase, the solipsist takes himself to be able to use ordinary language and the word ‘I’ to express what he wants to say about himself as the sole being who thinks and experiences. The trouble is that the I of which we speak in ordinary language is one among many persons, a being who is thought about by others just as he may think about them. So this attempted mode of expression of his solipsism has to be given up. The next stage is the solipsist's resort to an experiential language, a language which will, he hopes, be suited to bring out his unique situation. He attempts at this stage to use the word ‘I’ in two different ways: to speak about the empirical self that is part of the world and to speak about the self as centre of experience; but at this stage the self as centre is conceived as part of the world, a part that is given a priori. But, as Wittgenstein notes, there is no a priori element in experience; the experiencing self is not something in the world.

The stage which follows and which responds to that difficulty is one in which the solipsist uses a language which appears to give expression to his unique position, but which does so without using the word ‘I’ or any equivalent. This is the stage at which a Lichtenberg-language of the sort described by Anscombe would play a role. ‘There is pain’ or ‘It is thinking’ is used when I am in pain, or when I am thinking something, etc., and other people are described by formulations like ‘A is behaving as CD does when there is pain.’ A language for my experience, with a Lichtenberg-structure, at first appears to be the ideally right language to give expression to solipsism, and the ideally right language in which I can write The World as I Found It. It is indeed the availability of a form of representation distinct from ordinary ‘physical’ language, which appears to give the (p. 269) solipsist what he is looking for, that carries the argument forward at this point.22 But there are two difficulties. One is that the language does not allow the solipsist to represent anything which is not representable in ordinary language, so the advantage it at first appears to have seems questionable; the other is that there is a multiplicity of Lichtenberg-languages, which makes it unclear what the special character is of the language with me as L-centre. Whichever of the difficulties one follows up, the result will ultimately be the same. While the language with me as L-centre appears to enable me to express my unique position, it does not enable me to represent as being the case anything that cannot be represented in other Lichtenberg-languages or indeed in our ordinary ‘physical’ language.

As Anscombe notes, all the Lichtenberg languages and our ordinary way of speaking are equivalent in terms of what situations they can represent. It is what the languages have in common that enables them to represent those situations; and there corresponds to all of them the world. Think here of TLP 3.3411, where Wittgenstein said that one could say that the real name of an object was what all symbols that signified it had in common. One could go on to say that the language, or language, is simply what all languages that can represent the world have in common. What is the point then of saying that the I is a ‘limit of the world’?

The point of saying ‘The I is a limit of the world’ can be looked at from two directions. First, there is what is granted to the solipsist, who wants that everything that is the case can be put in the language with him as centre, and in which what is described is the world as he experiences it. And, secondly, there is the shrinking down of what is thereby granted to the solipsist. Whatever can be the case can be said in any language that is language in the sense explained in the preceding paragraphs; and in particular it can be said in the language which has me as L-centre, and in which I describe the world as I experience it. In that language, everything that can be the case in the world can be said. Wittgenstein writes at TLP 5.641: ‘The I gets into philosophy through this, that “the world is my world” .’ The clause ‘the world is my world’ combines, though in a form that can be misleading, the point just mentioned (about the reach of experiential language to everything that can be said to be the case in any language) with the point that all the languages, i.e. the language, have corresponding to them the world. The world is what language/all the languages speak of; it is ‘mine’ in that everything about the world, everything that is (p. 270) the case, everything that might be thought to be the case, can be said in the language with me as L-centre, in which I describe the world as I experience it. My language encompasses the world (and by ‘my language’ here one can slide between meaning the experiential language with me as L-centre and the language/all the languages, including the one with me as L-centre; one can speak of it/them as that language which I alone understand). One could think of ‘my’ Lichtenberg language, with its translatability into all other languages, as providing a shrunk-down notion of a centre of the language, an I that is an extensionless point; one could speak of this I as a limit of the world.23 But then this talk of ‘the limit of the world’ is simply a way of putting what, on the Tractatus account, is open to view in the presentation of what propositions have in common, including the translatability into each other of different forms of description, including the translatability into each other of the language with me as L-centre and non-Lichtenberg language. Rules for translating one language into another remain ‘within’ language; Wittgenstein's conception of the intertranslatability of all languages is not thought of by him as a kind of theory about language, but as something that can be seen in the shared features of different languages, through which all of them are capable of representing reality. His discussion of the self and solipsism, then, depends on the idea at 3.343 of intertranslatability of languages, which is itself meant to be a mere expansion of the idea that what is essential can be seen in the laying out of symbols which have a common feature. The general form of operation given at TLP 6.01 is meant to cover all translation rules, all transitions from one form of description to another.

It would, I think, be a mistake to try to explain the Tractatus view of solipsism (and the sense in which what it ‘means’ is correct) without looking at the significance of a language which appears to give special expression to my experience. The equivalence between that language and other languages, including ‘physical’ language, all of which have corresponding to them the same reality, is what enables us to see why the self of solipsism ‘shrinks to a point without extension’.24 But although the self of solipsism shrinks down to an extensionless point, the solipsist is nevertheless granted something: something more, indeed, than that the world can be represented in language that he understands. At the heart of what is granted by Wittgenstein's argument to the solipsist is the translatability of ‘physical’ language into the experiential language with me as centre; the possibility of translation in that direction is every bit as important to the argument as is the translatability of my experiential language into ‘physical’ language. The solipsist is meant to be genuinely satisfied: he is to be genuinely offered the reach of his experiential language to everything that is the case. To follow through the implications of solipsism strictly is to begin by trying to express the uniqueness of the I, and, from (p. 271) that starting place, to reach the stage at which one might say ‘My language encompasses the world’ (where ‘my language’ simply is language, the language, but is also at the same time language-with-me-as-L-centre).25 The forms of words we call on here, like ‘The world is my world’, are empty; they nevertheless provide useful images, and represent stages in the collapse of the solipsist's attempt at saying his solipsism. In particular, the language used in putting the different stages of the collapse draws on ordinary ways of using ‘the’ in contrast with ‘my’/‘me’/‘mine’ in identifying something, but the ordinary ways of making such contrasts are lacking, and no other mode of making them has been provided. So the final stage of the collapse of the solipsist's attempt to say his solipsism involves a recognition that no saying of anything would be what he wants, but that what he wants is in a sense open to view, just as much as the translatability of English into German and vice versa would have been taken by Wittgenstein at the time of the Tractatus to be open to view, and to be expressible in a rule. In suggesting that the translatability of experiential language into physical language, and of physical language into experiential language, gives the solipsist what he wants, I do want to insist that this is not nothing, though it is nothing that would be the content of any senseful proposition, nor is it some ineffable quasi-content beyond the reach of the sayable, any more than is the translatability of English into German.26

To see what the solipsist is granted, consider ‘The Earth has existed for more than four billion years.’ The picture that may accompany such a sentence is of an existence wholly independent of any thinking or experience, an existence that could perfectly well have been what it was, even if no being with experiences had ever existed, and that is causally related to experiences of things that count as evidence of the age of the Earth. What the solipsist is granted by translatability into his experiential language is that, insofar as the sentence about the Earth says that something is the case, what it says is so is sayable also in sentences in the book The World as I Found It, i.e. in experiential terms. The sentence about the Earth has a complex use, and the way in which it is descriptive isn't straightforward: it doesn't state any facts over and above experiential facts.27 The use of the sentence, that is, is not what the picture of the Earth out there, spinning along in its orbit independently of anybody's experience, suggests. And the solipsist is further granted (p. 272) that the relation between the Earth's existence and experiences of things that may be counted as evidence of its age is not a causal relation but a kind of logical relation.28 The issue here is that of the ‘limits of the world’. Before his line of thinking collapses, the solipsist takes there to be a would-be understanding of the sentence about the age of the Earth that puts it beyond the limits; he says ‘There is only my experience.’ What lies within the limits is only what can be said in experiential language. Realism about the world appears to require going beyond the limits recognized by the solipsist. What is offered by Wittgenstein puts the world, including the four-billion-year-old Earth, as we speak about it in ordinary ‘physical’ language, within the limits of sense, but its lying within the limits is inseparable from translatability of all sentences about the Earth into experiential language. (In figurative language: ‘The subject is a limit of the world.’) Thus to get something that looks as if it lies beyond the limits, we have to get something that appears to be outside the reach of such translation. Here the Tractatus can be thought of as offering us a distinction between realism and a kind of fantasy of realism. A fantasy of realism would involve the idea of a language in which things could be said that could not be said in my language, an idea which Russell (for example) did hold.29 The picture here is of the subject as not a limit of the world; it is a picture of the world as extending beyond the reach of the subject. How the contrast should be made between realism and a fantasy of realism is a question which, apparently settled in the Tractatus, gets unsettled and rethought later.

The self or subject shouldn't be thought of as a special sort of thing, a non-empirical thing, or an active thing, or an impersonal representing thing. Any such way of thinking makes the difference between it and the self of our ordinary thought and talk too slight. ‘It’ isn't a thing at all; and the thing-y language we may use for it in philosophy is deeply misleading. Anything spoken of isn't it. What we are trying to do in our thing-y language is put into words what can be seen in the translatability between ‘my’ centred-experiential language and our ordinary ways of representing the world. What then does Wittgenstein mean by saying that there really is a sense in which philosophy can talk about the self in a non-psychological way?

The sketch that I have given of the collapse of the solipsist's attempt to say his solipsism suggests, I think, that if we follow through what is involved in the collapse, we shall take the provision of a variable that presents what all propositions have in common to be the way philosophy does, in a sense, talk about the self in a non-psychological way. The earlier stage of the collapse, in which one might say ‘The world is my world’, even though it is a way of speaking which is ‘overcome’ (in the sense in which Wittgenstein's propositions are, as he says, to be overcome), nevertheless enables one to take the variable given at TLP 6 to satisfy the desire that philosophy treat of the self. That one moves through and beyond30 saying such things as ‘The world is my world’ or (p. 273) ‘The subject is a limit of the world’ is what enables the collapse of solipsism as a distinct view to be a satisfaction of the desire to speak solipsism. The figurative sentences through which one progresses mark the progress itself as one in which thus-and-such specific philosophical problem has been resolved. What I am here taking for granted is that a theme of all of Wittgenstein's philosophy is the transformation of philosophical desire, i.e. the idea that such desire can be understood to be satisfied, at the end of the activity of philosophy, by something different from what one had originally taken oneself to want. For the solipsist, then, who thinks through the implications of his view, there isn't something on the far side of the limits of language that he would like to be able to say, if only it were possible to do so.

There is a parallel between the structure of the account that I have given and that given by Thomas Ricketts of the collapse of a far more extensive structure of reasoning about sentences as ‘logically interconnected representations of reality’. (Ricketts 1996; see especially 88–94.) The crumpling of what appeared to be a stretch of thought, as we attempt to think it through, is an important element in the method of the Tractatus. After the crumpling, we may take ourselves to be left back ‘within’ senseful language. But that picture of our being ‘within’ language is itself the after-effect of the crumpling of talk of such things as ‘The world is my world’, talk that might appear to go beyond what can be said. In his letter to von Ficker, Wittgenstein described the book as drawing limits ‘gleichsam von innen’, ‘as it were from inside’ and it's time now to emphasize that ‘gleichsam’. To say, without that ‘as it were’, that Wittgenstein draws the limits from inside carries the strong suggestion of an outside where he did not go, and hence of limits which are constraints. The picture there of inside and outside is misleading. (cf. Minar 1995, 443.) Philosophy carried out ‘as it were from inside’ is simply philosophy, making clear what we say and think, showing us what we do. That was the intention; and measured against that intention, the book could later be recognized to have failed.

There is much in the Tractatus discussion of solipsism and the self that I have not touched on at all, including the connections with ethics. I have tried only to show how Wittgenstein's remarks about solipsism do not require us to think of the solipsist as having to be resigned to the impossibility of saying something that he wants to say, and to bring out the shifting character of a remark like ‘The world is my world’, which shifts from its appearance as a kind of conclusion at one stage in the thinking through of the implications of solipsism to its ultimate appearance as something we move through and beyond in the collapse of solipsism as a distinct view. There is also much in the Tractatus use of the notion of a limit or limits that I have not been able to discuss. The most important thing that I have not discussed except tangentially is the Tractatus understanding of logic and its relation to giving the limits of language ‘from within’.31


Anscombe, G. E. M. (1963). An Introduction to Wittgenstein's Tractatus. London: Hutchinson.Find this resource:

    Conant, James (2002). ‘The Method of the Tractatus’, in Erich H. Reck ed., From Frege to Wittgenstein. New York: Oxford University Press.Find this resource:

      Conant, James and Diamond, Cora (2004). ‘On Reading the Tractatus Resolutely’, in Max Kölbel and Bernhard Weiss eds., Wittgenstein's Lasting Significance. London: Routledge.Find this resource:

        Diamond, Cora (2000). ‘Does Bismarck Have a Beetle in His Box?’, in Alice Crary and Rupert Read eds., The New Wittgenstein. London: Routledge.Find this resource:

          —— (2004). ‘Saying and Showing, an Example from Anscombe’, in Barry Stocker ed., Post-Analytic Tractatus. Aldershot: Ashgate.Find this resource:

            —— (2010). ‘Inheriting from Frege: the Work of Reception, as Wittgenstein Did It’, in Thomas Ricketts and Michael Potter eds., The Cambridge Companion to Frege. Cambridge: Cambridge University Press.Find this resource:

              —— (forthcoming). ‘The Hardness of the Soft’, in James Conant and Andrea Kern eds., Scepticism, Meaning and Justification. London: Routledge.Find this resource:

                —— (forthcoming a). ‘What Can You Do with the General Form of Proposition?’, in José Zala- bardo ed., Wittgenstein's Early Philosophy. Oxford: Oxford University Press.Find this resource:

                  Donatelli, Piergiorgio (2005). ‘The Problem of “the Higher” in Wittgenstein's Tractatus’, in D. Z. Phillips and Mario von der Ruhr eds., Religion and Wittgenstein's Legacy. Aldershot: Ashgate.Find this resource:

                    Floyd, Juliet (1998). ‘The Uncaptive Eye’, in Leroy S. Rouner ed., Loneliness. Notre Dame, IN: University of Notre Dame Press.Find this resource:

                      Fried, Michael (2007). ‘Jeff Wall, Wittgenstein, and the Everyday’, Critical Inquiry. 33.3: 495–526.Find this resource:

                        Friedlander, Eli (2001). Signs of Sense. Cambridge, MA: Harvard University Press.Find this resource:

                          Griffin, James (1964). Wittgenstein's Logical Atomism. Oxford: Clarendon Press.Find this resource:

                            Kremer, Michael (2001). ‘The Purpose of Tractarian Nonsense’, Noûs, 35: 39–73.Find this resource:

                              —— (2002). ‘Mathematics and Meaning in the Tractatus’, Philosophical Investigations, 25.3: 272–303.Find this resource:

                                Lee, Desmond ed. (1980). Wittgenstein's Lectures 1930–1932: From the notes of John King and Desmond Lee. Oxford: Blackwell.Find this resource:

                                  McGinn, Marie (2006). Elucidating the Tractatus. Oxford: Oxford University Press.Find this resource:

                                    McGuinness, Brian (2001). ‘ “Solipsism” in the Tractatus’, in David Charles and William Child eds., Wittgensteinian Themes: Essays in Honour of David Pears. Oxford: Oxford University Press.Find this resource:

                                      McManus, Denis (2006). The Enchantment of Words. Oxford: Oxford University Press.Find this resource:

                                        —— (2009). ‘The General Form of the Proposition: The Unity of Language and the Generality of Logic in the Early Wittgenstein’, Philosophical Investigations, 32.4: 295–318.Find this resource:

                                          Minar, Edward (1995). ‘Feeling at Home in Language (What Makes Reading Philosophical Investigations Possible?)’, Synthese, 102: 413–52.Find this resource:

                                            Moore, G. E. (1922). Principia Ethica. Cambridge: Cambridge University Press.Find this resource:

                                              Mulhall, Stephen (2007). ‘Words, Waxing and Waning: Ethics in/and/of the Tractatus Logico-Philosophicus’, in Guy Kahane, Edward Kanterian, and Oskari Kuusela eds., Wittgenstein and His Interpreters. Oxford: Blackwell.Find this resource:

                                                Ricketts, Thomas (1996). ‘Pictures, Logic, and the Limits of Sense in Wittgenstein's Tractatus’, in Hans Sluga and David G. Stern eds., The Cambridge Companion to Wittgenstein. Cambridge: Cambridge University Press.Find this resource:

                                                  (p. 275) Russell, Bertrand (1932). ‘Knowledge by Acquaintance and Knowledge by Description’, in Mysticism and Logic. London: George Allen & Unwin.Find this resource:

                                                    Schopenhauer, Arthur (1988). Die Welt als Wille und Vorstellung. Zürich: Haffmans Verlag.Find this resource:

                                                      Stern, David (1995). Wittgenstein on Mind and Language. New York: Oxford University Press.Find this resource:

                                                        —— (2007). ‘Wittgenstein, the Vienna Circle, and Physicalism: A Reassessment’, in Alan Richardson and Thomas Uebel eds., The Cambridge Companion to Logical Empiricism. Cambridge: Cambridge University Press.Find this resource:

                                                          Sullivan, Peter (2004). ‘ “The General Propositional Form is a Variable” (Tractatus 4.53)’, Mind, 113. 449: 43–56.Find this resource:

                                                            —— (2004a). ‘What Is the Tractatus About’?, in Max Kölbel and Bernhard Weiss eds., Wittgenstein's Lasting Significance. London: Routledge.Find this resource:

                                                              —— (2011). ‘Synthesizing without Concepts’, in Rupert Read and Matthew A. Lavery eds., Beyond the Tractatus Wars: The New Wittgenstein Debate. Abingdon. Routledge.Find this resource:

                                                                Waissmann, Friedrich (1965). ‘Notes on Talks with Wittgenstein’, Philosophical Review, 74.1: 12–16.Find this resource:

                                                                  Waismann, Friedrich (1979). Wittgenstein and the Vienna Circle: Conversations recorded by Friedrich Waismann, ed. B. McGuinness, trans. B. McGuinness and J. Schulte. Oxford: Blackwell.Find this resource:

                                                                    Williams, Bernard (1981). ‘Wittgenstein and Idealism’, in Moral Luck. Cambridge: Cambridge University Press.Find this resource:

                                                                      Wittgenstein, Ludwig (1951). Tractatus Logico-Philosophicus, trans. C. K. Ogden. London: Routledge & Kegan Paul.Find this resource:

                                                                        —— (1953). Philosophical Investigations, ed. G. E. M. Anscombe and R. Rhees, trans. G. E. M. Anscombe. Oxford: Blackwell.Find this resource:

                                                                          —— (1961). Notebooks 1914–1916. Oxford: Blackwell.Find this resource:

                                                                            —— (1965).  ‘A Lecture on Ethics’, Philosophical Review, 74: 3–12.Find this resource:

                                                                              —— (1967). Zettel, ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe. Oxford: Basil Blackwell.Find this resource:

                                                                                —— (1971). Prototractatus: An Early Version of Tractatus Logico-Philosophicus. London: Routledge & Kegan Paul.Find this resource:

                                                                                  —— (1975). Philosophical Remarks, ed. R. Rhees, trans. R. Hargreaves and R. White. Oxford: Basil Blackwell.Find this resource:

                                                                                    —— (1980). Culture and Value, ed. G. H. von Wright in collaboration with H. Nyman. Oxford: Blackwell.Find this resource:


                                                                                      (1) There is a contrast between ‘limits’ and ‘limitations’ used in discussing Wittgenstein's views, a contrast related to but not identical with the contrast I am using between two ways of understanding limits. See Sullivan (forthcoming), 1–2, for a lucid explanation of the ‘limits’/‘limitations’ contrast.

                                                                                      (2) Cf. Griffin 1964, ch. 8, §5, esp. pp. 102–3. Griffin says that Wittgenstein held that the function of scientific laws ‘is not to make reports, not even very general ones, but to supply representational techniques by which reports can be made’. I discuss Wittgenstein's account of scientific laws in Diamond, forthcoming.

                                                                                      (3) For a discussion of Wittgenstein's use of imagined ‘books’, see Friedlander 2001. Friedlander's account of the role of these ‘books’ in Wittgenstein's method, and in particular of their relation to the aim and character of the Tractatus, is very different from mine.

                                                                                      (4) This claim draws on 3.314–3.317 and 5.501.

                                                                                      (5) For a general discussion of the notion of senselessness and of the ways in which a variety of senseless propositions may be used, see Kremer 2002. Philosophy can make extensive use of such senseless propositions in its activity of clarification, distinct from any use it may make of nonsense. Thus e.g. the Russellian analysis of definite descriptions can be treated as a rule for rewriting propositions, and can be added to a proposition containing a definite description without changing its sense.

                                                                                      (6) For the connection between picture-use and reversibility, and between reversibility and a truth-functional understanding of logic, see Ricketts 1996.

                                                                                      (7) I have seen it argued that Wittgenstein takes mathematical equations to be nonsensical pseudo-propositions. He certainly speaks of the equations of mathematics as pseudo-propositions, Scheinsätze (6.2). He does not identify them with tautologies, but it hardly follows that he regards them as nonsensical. Calling them Scheinsätze or pseudo-propositions certainly implies that they misleadingly appear like propositions, sign-constructions used to say that something or other is the case, picture-propositions, but it does not imply that they have no use or that they are meaningless. See Kremer 2002.

                                                                                      (8) Schopenhauer 1988, Buch II, §24.

                                                                                      (9) The point about the meaningfulness of ‘proposition’ does not apply to ordinary uses of the word, like ‘Every proposition he utters is false’, but only to those uses in which it is treated as a ‘proper concept-word’; see TLP 4.1271–4.1272.

                                                                                      (10) James Conant and I discuss the case of sentences of the form ‘P entails Q’ in Conant and Diamond 2004. We pick up the example from Peter Sullivan's discussion of it in his 2004a in the same volume.

                                                                                      (11) See Mulhall 2007, 246 on the ‘dual aspect’ of the Tractatus remarks in the 6.4s and 6.5s.

                                                                                      (12) I discuss Wittgenstein's ideas on the importance of distinguishing sharply enough, and the relation of his ideas to those of Frege, in Diamond 2010.

                                                                                      (13) See Donatelli 2005 for a discussion of the relation between Wittgenstein's remarks about value and Moore's ideas in Principia Ethica.

                                                                                      (14) Compare his treatment of evaluations in LE, 5. At the time of that lecture, at any rate, he saw no special problems in including within a general account of factual language a treatment of much evaluation, including, for example, the evaluation of chairs and pianists. A similar account of such evaluation would be compatible with the Tractatus. What the Tractatus excludes is that the ‘value’ spoken of in evaluating chairs and so on is relevant to the search for genuine value.

                                                                                      (15) Contrast Kremer 2001, 60. Kremer argues that Wittgenstein does not invite us to think in terms of a ‘perspective on the world’. He refers to James Conant's criticism of the idea of a philosophical perspective on the logical structure of the world. According to Conant, this is an illusion of a perspective, and not a point of view which we are invited to take up by the Tractatus. (See Conant 2002, 422–3.) My reading of Wittgenstein is certainly different from Kremer's, in that I think the figurative use of the idea of a perspective on the world does play a role in Wittgenstein's conception of ethics, but such a figurative use of the idea of a perspective does not involve the conception criticized by Conant.

                                                                                      (16) Compare Floyd 1998, 103, on attachment to life and the refusal of attachment to life.

                                                                                      (17) Compare Z §320: ‘If we follow grammatical rules other than such-and-such ones, that does not mean we speak falsely but rather that we are speaking of something else.’ I am arguing that the Tractatus conception of the method of philosophy implies that if we give a variable other than such-and-such one, we have not given a false account of propositionhood (say), but have given something else instead.

                                                                                      (18) My argument in section 6 draws on TLP 4.126 in a way that I haven't made explicit. It is often claimed that the Tractatus does not take propositions to have a Bedeutung; but Kremer (in 2002) has made clear the broad sense of the word as Wittgenstein uses it, and its tie to an expression's having a use. At TLP 4.126, Wittgenstein says that the sign for the marks of a formal concept is a distinctive feature of the symbols whose meanings (Bedeutungen) fall under the concept. This applies to the formal concept being a proposition as much as to any other formal concept, i.e. it applies to symbols the meaning of which is propositional meaning. When Wittgenstein says in 3.317 that the stipulation of the values of a propositional variable is not concerned with the Bedeutung of the symbols in question, this applies to propositions as well as to other symbols. The presenting of a variable like the one at TLP 6 has to be separated from the philosophical ‘gas’ that accompanies it. The essential thing about the giving of the variable is: ‘it states nothing about what is signified’. So far as Wittgenstein does actually say, at TLP 6, that what he is giving is the general form of propositions, the words giving what the general form is of drop out; they are no part of the philosophical activity of presenting the variable. It is essential to that activity that it doesn't include saying that the values of the variable have such-and-such sort of meaning.

                                                                                      (19) Anscombe 1963, 166–7. Anscombe wrote before the publication of Philosophical Remarks, but the remarks which she quotes can be found in PR, 88–9; cf. also WVC, 49–50. The material in the latter dates from late 1929.

                                                                                      (20) ‘It thinks’ or ‘It is thinking’ is a translation of ‘Es denkt’. ‘Es denkt, sollte man sagen, so wie man sagt: es blitzt. Zu sagen cogito, ist schon zu viel, sobald man es durch Ich denke übersetzt.’ The remark is on p. 74 of the Reclam edition of Lichtenberg's selected writings that Wittgenstein gave Russell in 1913; the remark is also quoted by Weininger in Geschlecht und Charakter.

                                                                                      (21) Anscombe says that the passage she quotes ‘appears very close to [the Tractatus] in thought’. Her impression is certainly supported by Wittgenstein's claim that Carnap had plagiarized the Tractatus in his claims about the translatability into each other of physical language and experiential language. For discussion of this accusation, see Diamond 2000, 263, 278–9, 287. See also Stern 2007.

                                                                                      (22) Many discussions of Wittgenstein's remarks on solipsism leave out the idea that there is any further stage beyond the criticism of the idea of the self as an a priori element in all experience. The idea that Anscombe emphasizes, that Wittgenstein's thought about solipsism takes seriously the possibility of a language which appears to give an ideal kind of expression to solipsism without mentioning the self at all is simply not considered. But I believe that this idea is an essential part of what Wittgenstein thinks of as the following through of the implications of solipsism, and that it is at the heart of what is involved in ‘the world is my world’. It is at any rate clear that there is no obvious step from the collapse of a version of solipsism that takes the self to be an a priori element within experience to the collapse of solipsism simpliciter. Readers of Wittgenstein may disagree about how far the Tractatus remains in the grip of some kind of solipsism, but discussion of the issue will be distorted if it is thought that he took the point about there not being any a priori element in experience to finish off solipsism.

                                                                                      (23) The notion of a centre introduced here is different from the notion of an L-centre. There are many L-centres, one for each Lichtenberg language. So far as all the Lichtenberg languages, including the one with me as L-centre, are intertranslatable and translatable into ordinary language, they have one centre in the shrunk-down sense, the ‘centre’ of the language.

                                                                                      (24) The importance of the ideas here for Wittgenstein is reflected also in his accusation of plagiarism against Carnap, an accusation which involved Wittgenstein's claim that what Carnap meant by physicalism was already in the Tractatus. See Stern 2007.

                                                                                      (25) There is more that could be said about the ‘peculiar but inexpressible advantage’ that the language with me as centre appears to have, but I cannot get into this question here. See WVC, 50. For a discussion of how the Tractatus can accommodate the asymmetry between ‘I have a pain’ and ‘He has a pain’, see Diamond 2000.

                                                                                      (26) I have not tried to lay out what exactly is meant by ‘experiential language’ in the context of discussion of the Tractatus. I don't think we can identify what is meant with Wittgenstein's later conception of a ‘phenomenological language’. But the questions with which he was concerned later, when he wrote about a ‘primary’ language, were close to the questions involved in the discussion of solipsism in the Tractatus. See also Stern 2007, and Stern 1995, ch. 3, esp. p. 78, where Stern quotes F. P. Ramsey's notes from the 1920s on Wittgenstein's treatment of the significance of experiential language.

                                                                                      (27) Ramsey's notes from the 1920s (see previous note) explicitly ascribe to Wittgenstein a treatment of material objects of the sort described. Talk that purports to be about such things enables us to use simplified general laws in descriptions. A corresponding account is suggested of the ordinary language ascription of sense-data to owners of sense-data.

                                                                                      (28) Compare WL, 80–83.

                                                                                      (29) Russell 1932, 216–18; the essay was originally published in 1911. I discuss Russell's view and Wittgenstein's response in Diamond 2000.

                                                                                      (30) ‘durch sie–auf ihnen–über sie’; see TLP 6.54.

                                                                                      (31) I have been greatly helped by comments and suggestions from Marie McGinn, James Conant, and Alice Crary.