19*. Was Wittgenstein a Conventionalist?

The conventionality and subjectivity of truth (but this latter not at all in Kierkegaard’s sense) is one of the characteristically apathetic thoughts of our time. Nothing is, not only good or bad, fair or foul, but also true or false but thinking makes it so. The thinking may be a consensus within a society: that is where the conventionality comes in.

To this popular attitude the philosophy of Wittgenstein seems to make its contribution. That is, Wittgenstein of the Philosophical Investigations, not Wittgenstein of the Tractatus. That work played a part in the promotion of another attitude: truth is to be found in natural science alone. Not a new attitude, indeed, but one which in the Tractatus was backed up by sophisticated thinking about logic, language, mathematics, and which had the added cachet of a deeply serious tone about the ethical, the mystical, the higher.

If those propositions which seem to express insights into the essential natures of things—propositions whose contents are supposed to be the musts in reality—if these are to be explained as disguised propositions of grammar, or sometimes are revealed to be pure nonsense; and if their correlative contradictories in turn are not the expression of what is necessarily excluded from reality, but are merely forms of combination of words at present not admitted by the grammar of that part of language to which they essentially belong: if all this is so, doesn’t it seem that truth itself is a matter of convention? Philosophy has no deep truths to teach; its goal is not the achievement of knowledge of what must be truth; its problems are deep only because the forms of language are deeply rooted in us; and even the ‘truths’ of everyday communication are true only within the framework of a system of conventions.

‘Essence’, Wittgenstein says, ‘is expressed by grammar’. It isn’t difficult to grasp that. Think of the association of a proper name with the name of a kind of thing, such as ‘dog’, such that the proper name is used with the same reference as it was before, if and only if it is used of the same member of that kind, the same dog, say. That gives us part of the grammar of proper names. And does not this grammar express the essence of the individual named by a proper name, if anything does? Or at any rate, part of its essence.

But grammar could be different. And that suggests, not just that essence is expressed by grammar, but that essence is the creation of grammar. Couldn’t a proper name be associated with a different kind of identity? Suppose, for example, that what comes out of the mouth of one human body as personal expression (expression of personal states) were always referred to another human body so far as any body must e.g. suffer the consequences (as princes had whipping boys)—might not one personal identity be vested not in one human body but in two, be as it were distributed among two? (Or more, as we distribute the several roles of a ‘person’.)—One society, it seems, cannot criticise the beliefs and practices of another if their forms of life are different from one another. You are talking your language, and anything you say is a move in a language-game which is played in your society: the Aztec or the consultor of poison-oracles in mid-Africa is making moves in the language-games they play there.

Or again, one can describe the grammar of verbs of perception and sensation, observing that some of them take objects while others do not. E.g. ‘I see’ and ‘I itch’. One will describe the intentionality of such verbs and their double use: with an intentional or a material object, according to whether ‘I see such-and-such’ is refuted by there being no such-and-such there. One may give an account of the ‘asymmetry’ of such verbs in respect of the first person present indicative used assertorically—the senselessness of the question ‘How do you know?’ in response to such utterances (e.g. ‘I am hearing a sound’) except with reference to the material diagnosis of the object (‘I see the sky’).[1]—And now let us suppose the grammar changed. We describe a use of words which is akin to the use of the verbs of sensation and perception, but which lacks some of these features. Each person, for example, makes reports in the form ‘my retina is affected thus-and-so’ (etc.), and others may look and see whether this is so. If it is not, he is simply mistaken. So, though he says such things not by examining the way his organs are affected, but they come to him ‘immediately’—and so there is still a certain asymmetry, for the question can be ‘Did you really mean it, did you consider carefully enough?’ but not ‘Did you look carefully enough?’—still there is no ‘privacy’ and instead of intentionality there is incompleteness in what he is able to report. (But such incompleteness is usual in any report.) The question arises: should we thereby miss out things essential to perception (as it is still taking place)? Or suppose that the language we manage to describe were universal, the language of perception as we at present have it being non-existent: would that be the supposition that there was no such thing as perception—as it is at present?

At §497 Wittgenstein remarks: ‘If someone says “If our language had not this grammar, it could not express these facts”—it should be asked what “could “ means here.’ He doesn’t answer the question. Let’s try: ‘If our language (of perception) had not its present grammar but was replaced by a language of the sort we have imagined, then it would not include sentences about (e.g.) the visual experiences of people with delirium tremens.’—Well, doubtless it would not. The concept of ‘experience’, as it is used there, would make no appearance in the language we are supposing. But wouldn’t the facts appear in the alcoholic’s greater tendency to report wrongly what was on his retina? We might say: Experience, in this special sense of the word, is an ‘essence’ not just expressed but created by grammar. Now that would involve us in no general commitment.

Plato’s suggestion in the Cratylus: words for the same thing in different languages (as ‘horse’, ‘cheval’) are like tools made of different metals but all shaped to catch hold of the same object. What these words catch hold of is the same essence. This presupposes that the essence is there to catch hold of, and that the words have in some sense the same shape. Wittgenstein might concur in the last bit: the ‘shape’ is logical shape, and this is none other than that use whose similarity justifies translation of a word in one language by a sensibly very different word in another.

‘You learned the concept pain by learning language.’ I.e. not by experiencing pain. If people have a different concept do they miss something? I mean, if they have a concept which is sufficiently like ours to be called a concept of pain, but which does not apply where there is no obvious damage or physical disorder (dislocation, for example) or incapacitation—do they misrepresent, or at least miss something out? If there are people for whom red is a degenerate form of green, do they miss anything, or make a mistake?—To render this natural we need to suppose a different world: red, say, occurring only on the edges of some leaves.

On the first example, Wittgenstein imagines the objection: don’t they notice the similarity? - sc. of pain without damage to pain with damage. And he asks if we want a concept wherever we may notice a similarity.

Differences in facts of nature and differences in human practices: imagining these can render intelligible to us the formation of concepts different from ours, yet related to them. This shows us that we haven’t the right to say that certain concepts are absolutely the correct ones.[2] But Wittgenstein also seems to speak against the idea that ‘having different ones would mean not realising something that we realise’. But wouldn’t it? E.g. wouldn’t the people for whom red is a degenerate case of green fail to realise that red is a different colour?—Yet again, would they? They see a difference—and don’t treat it the same way as we do. We call the difference one of colour, they do not. That doesn’t make them wrong.

There is another question. May we not say: we don’t claim that our concepts are absolutely the right ones, but are they not nevertheless right ones? We understand indeed how other people might form different concepts. The intelligibility of this may depend on a description of some difference in very general facts of nature—for example, colour and shape vary independently in our world; in a world in which they did not, people might have only ‘simple’ concepts which correspond to our ‘round and blue’, ‘red and square’, etc. But of course we have said in our description that there are not in their world the facts for which distinct words for colour and shape give us the means of description.—However, we might suppose that there were some such facts; these might then be overlooked, or might be regarded as very unimportant. The possession of such a language as we have would presumably make it easier to notice such facts than it might otherwise be. But we haven’t reason to say that people possessed only of the colour-shape concepts would not be able to notice these facts. They can grant similarities without having terms for the respect of similarity. Here there seems no reason to say that essences are created by grammar. But: though we do thus make no claim for our concepts to be ‘absolutely’ the right ones, still we can claim that our concepts are right ones. Or at least: can’t we enquire whether our concepts are right ones, and conclude that they are, at least in some cases? And can’t we criticise others’ concepts as wrong ones, i.e. as not corresponding to anything?

We can surely prove this. We have a concept of superstition. Is there not such a thing? If not, then the concept is a wrong concept.[3] If so, then some concepts are wrong. So in any case some concepts are wrong. And you, Wittgenstein, yourself called some things superstitious. To be sure, what you called ‘superstitious’ were not concepts but beliefs. But if someone says ‘It is superstitious to believe that there is such a thing as magic. Certainly there are people who practise magic, but their belief about what they do is superstitious’—is that not equivalent to saying that ‘magic’ (as believed in) is a superstitious concept, and therefore a wrong one. I am not at all supposing here what Wittgenstein found so stupid, namely that belief in magic is a sort of mistaken natural science. But I take it that superstition is wrong, false. Wittgenstein too presumably means that it is false, when he diagnoses something as superstition, as for example at §110: ‘ ‘Language (or thought) is something unique’—this proves to be a superstition (not a mistake!) itself produced by grammatical illusions.’

However, he says ‘not a mistake’. We need to ask why. We shall find that Wittgenstein would say ‘It’s not a mistake’ in a number of cases. Suppose someone were to say, apparently seriously and not speaking in metaphor or by anticipation, ‘I am dead’, or to ask ‘Surely I am dead, aren’t I?’ He can’t be said straightforwardly to be making a mistake. What would it be like for him to be right in saying this? And if there is no answer to that, then what is he supposing himself to be saying? He is using language in such a strange way that we ought to hesitate to think we understand him. Of course, if he has just emerged from an accident, we may understand him in a way, as suffering from confusion of mind, and we may firmly reassure him by telling him that he is alive. This may get him out of it—but it would be comical to think of the situation as one in which we inform and he learns from us. Again, if someone is brought up, trained, to think he is a god and that the world began with and depends on him, his belief is a wrong, an appalling belief for him to have, and we might want to get him out of it, but it would be absurd to treat it as something the truth of which would be established by these observations, and refuted or rendered doubtful by those, and to think he is under the impression that these rather than those observations can be made.

The conception of ‘mistake’ which Wittgenstein has is obviously something like this: in order to say that someone thinks that p, and it is a mistake on his part, it has to be clear what it is for p to be true (or correct); he must be acquainted with this. E.g. the situation must be such that if he had noticed the relevant features, he would have agreed that p was false. Or, if it should be a calculation, he must be ready to take back a slip or false step.

Thus not every case of saying p when, by all familiar criteria, it is true that ~p is a case of making a mistake. The situation may be all wrong for this utterance of ‘p’ to be straightforwardly the assertion which is the contradictory of that apparent truth that ~p.

It may be that the assertion ‘p’ is something which we cannot understand. An example is the assertion of the deaf-mute, Mr Ballard about his childhood that

... even before he could speak he had had thoughts about God and the world.—What can he have meant?—Ballard writes: ‘It was during those delightful rides, some two or three years before my initiation into the rudiments of written language, that I began to ask myself the question: how came the world into being?’— Are you sure—one would like to ask—that this is the correct translation of your wordless thought into words? And why does this question—which elsewhere seems not to exist at all—raise its head here? Do I want to say that the writer’s memory deceives him? I don’t even know if I would say that [presumably, if confronted with the man and able to converse with him]. These recollections are a queer memory phenomenon,—and I do not know what conclusions one can draw from them about the past of the man who recounts them.[4]

That is, accepting the man’s sincerity, we don’t know what it means for the report to be true.

On the other hand, consider the Catholic doctrine of transubstantiation and the irrelevance of referring to chemical analysis or what everyone can see if he looks at the consecrated host.—What is said may be an assertion in a different line of country, where once again there is a definite way for the thing to be true. If it were a convention—say a legal convention—that someone was dead in certain circumstances, the assertion ‘He is dead’ would be an assertion in a different line of country from the usual one, and the method for showing it to be false is known; it can once more be a mistake.

About the Law of Excluded Middle, where p is e.g. ‘There are 3 consecutive sevens in the decimal development of ‘, Wittgenstein’s objection to p v p is that it can have as shaky or as solid a meaning only as its elements, p and p, have themselves.

The key thought on all this comes at §241: ‘ “So you are saying that human agreement decides what is correct and what is wrong?”—It is what human beings say that is correct and wrong; and they agree in language. That isn’t any agreement in opinions, but in form of life.’

It is what human beings say that is true or false, and they agree in the language that they use. You have to know or find out what that language is, before you can draw conclusions about what they say, or call it false. When the sun appears at the horizon at the beginning of the day, that is called ‘sunrise’. That does not mean that the question ‘Had the sun risen yet?’ is not objective, i.e. that it is subjective in the sense that the correctness of the answer depends on, consists in, the agreement of those who said it had or had not.

You have to find out what language we are using, when we purport to be making assertions. What is it for that to be so, or not to be so? We often assume we know this when we don’t. E.g. an anthropologist tells us that in a certain tribe they have such and such a belief about the soul. It is to the point to ask him what he calls people having such a belief about the soul. It is of course no use his saying ‘That’s what they say’ because we shall need to know why he translates a word they use as ‘soul’.

The idea that e.g. personal identity is a matter of convention is strongly resisted. That there might be a convention by which this future object (indicated by some description) is called the same person as I am—that does not either comfort me or make me fear. The question is, will that be I? Conventions can be changed. How absurd a proposal, to relieve my apprehensions by proposing a new convention. (Chisholm)

But is there not a convention by which some future object may be identified with me? It is this: if any human body living with a human life is the same human body as this one (at which I lay my hand on my breast) then that will be me, E.A. The application of the convention is clear so long as we know what we mean by re-identifying a human being. Now isn’t this a convention? I mean: if you asked why that would be E.A., there is no answer except that that’s what is meant by being the same person as E.A. ‘E.A.’ is the name of a human being and of a person because a human being is a person; i.e. to use Boethius’ definition of a person: an individual substance of a rational nature. We could no doubt imagine a failure, an inapplicability of this convention; e.g. if it happened for one human being to split directly into two the same size: two as it were walk simultaneously out, one to the right, the other to the left, from where there was a single one. Or again we could imagine the identity of the organism preserved but this lacking some of the interest it at present has, so that it was reckoned that one organism was successively two different human beings. E.g. suppose the race worked like this: everyone is born female and continues so to the age of twenty or thirty, and then undergoes a metamorphosis so as to become male. Much, both in the facts of nature connected with this development, and in the treatment that Laura and Laurence (say) universally receive in society: the emotional relation between them, what sort of vanity is exercisable by Laurence for being the same human organism as Laura was, what sort of responsibility either has in connection with the other— all this, we will suppose, surrounds and in a way supports the reckoning of these as two human beings. No doubt it will have some role in this if Laurence quite lacks agent-memory in respect of Laura’s deeds. But the importance of this ought not to be exaggerated. The presence of such agent-memory would not prove anything, for ‘I’ is not a name. And the rule we at present have: If N directly says ‘I …’ then that is true if and only if ‘N …’ is true might not hold. Any more than it would hold if there were a peculiar relation between a human being and one of his grandfathers, such that he had agent-memory in respect of his grandfather’s actions. This, oddly enough, was recently made clear by Locke, who, however, took as a criterion for a new kind of identity one of the several reasons (memory) why we are (at present) interested in the identity of a human being (and not usually in the identity of a pin or a grain of sand). Locke saw that agent-memory could be conceived to pass from one substance to another, that is, that the spontaneous memory ‘I did such-and-such’ had by one substance might have a relation to the consciousness ‘I am doing such-and-such’ previously had by another substance, which constituted the former as the direct memory of the very deed referred to in the latter. He thought that such an occurrence would prove identity of a new sort of object, which he called a ‘person’. But a person, as he defines it is not a suitable term for identity at all. For he defines a person as ‘a thinking intelligent being that has reason and reflection, and can consider itself as itself, the same thinking thing, in different times and places.’ That is, he defines a person as a being with certain capacities. ‘Same person’ by this definition is not different from ‘same genius’. (This admirable insight is owing to my student Harold Noonan.) Find a being with this capacity and you find a being that is a person by Locke’s definition. But the identity of the being will be the identity of whatever kind of being it is, a human being for example. Locke had the profundity to see that his idea led to the question whether one thinking substance might have true agent-memory which was of deeds belonging to another, and invoked the goodness of God to avoid this. Jonathan Edwards on the other hand embraced the idea and used it to explain our guilt for Adam’s sin: we shall remember doing it when we are judged.

Thus in my fantasy neither the absence nor the presence of agent-memory in Laurence in respect of Laura’s deeds has to determine anything. Laura and Laurence (who are but examples, these names are dummy constants and what is said using them shows what is said in all cases) can be supposed to be counted distinct human beings.—But, it may be asked, would they be distinct human beings? What can that mean? The case is after all imaginary: the facts of nature are different from ours. Among us, the single human organism is not distinguished from the single human being except in the twinning embryo and in very closely united Siamese twins. Nothing is determined by our present application of the concept ‘human being’ about what would be the continuation of that concept in circumstances so different from ours. But if anyone objects that, since various continuations are imaginable, any one would be merely conventional, we reply that our present use too [is] a convention, by which the aged man (or other animal) is the same X (‘X’ being the name of the animal kind) as the new born or new hatched. But it is an illusion to think that, because something is a convention and because we could imagine a different one, it must therefore be in human power to alter it, as it is presumably in human power to adopt a fixed date for Easter or a phonetic system of spelling for English. Altering a convention may be beyond our power.

‘If it is at present a convention that the aged man is the same human being as the baby from which he developed.’ It may be protested: but that, e.g. is what a single human being is! And that is quite true. But does it mean anything other than: that’s what is called a human being—is an example of what is called being the same human being? If[5] an Oriental sect calls some abbot the same human being as an abbot who existed before this one was born—which we do not call being the same human being—that is a convention that parts company with ours [and] this suggests that we cannot call it a mistake, as Wittgenstein understands ‘mistake’. It certainly is not like the mistake of thinking that this man I see here is the same as such-and-such a baby, when not that baby but some other grew up into him. It appears, then, that we, who are speaking within our convention, cannot contradict the belief of the sectary. But to say that is to assume that contradiction is just one thing. Negation is rejection: but there are different kinds of rejection. If any superstition is to be rejected as fake, but can be contrasted with [a] simple ‘mistake’, then so may be the convention and associated belief of our Oriental sectary. The interesting question is whether this is a matter of mere denunciation of one another by people of opposing parties—people whose several beliefs and practices excite powerful feelings of mutual opposition— but in which there can be no showing of anything as false or wrong. My supposed Oriental no doubt grasps, say has, the concept of the identity from conception or post-conceptional splitting to death, which is the only thing that we in the West are agreed on calling the identity of an individual animal, but he regards that as only one stretch of a continued identity over many such stretches which (I am supposing) he calls the identity of a human being. Now to say this is to assume that we shouldn’t after all use some other word to translate his word, rather than ‘man’ or ‘human’. That can surely be a justified assumption.

It may be quite clear that this is the word in that language for ‘human being’ but that they make these identifications, different from all otherwise normal identifying practice. This indeed we know that the Tibetan Buddhists do, identifying the present so-and-so with all former ones; but I do not know whether they regard them all as different human beings, in whom the same individual something else (spirit, say) is successively embodied, or as all the same human being who again and again becomes an embryo, is born, grows to maturity and dies. It is this latter idea that I am ascribing to my hypothetical sectary. Now how could this be shown to be wrong? We may say we don’t know what the identification means, and that no one could know: it is in fact meaningless because unintelligible. But ‘it means something to them’: it has consequences including logical consequences.[6] And we may very naturally suppose that it has a point in their lives: e.g. quasi-divine worship is paid to these identified Enlightened Men. If we say we cannot understand the identification, it may be retorted upon us: (a) what would you say to someone (of another utterly different society) who ‘could not understand’ your identification of the baby as the same human being with the grown man or woman? You may say: look, this is an identification we just do make, and it shows what our concept of a human being is. But then the objector might find that senseless, he might not be able to enter into it.—And that other one is an identification those people make, and it shows something of what their concept of a human being is. (For there is enough overlap for us to call it that.)

There is a place in Africa where, if a head of an immediate family dies, someone belonging to that village but not to that family is identified as that man by a visiting witch-doctor, who is summoned for the purpose from a distance (so that he shall [not] have a personal stake in the decision). The person so designated has a certain position of authority in the family; he can settle quarrels, tell them to stop misbehaving, and so on. He does not acquire any material interests in connection with it. This role can be socially useful.[7] Now this is not a mysterious story. We can take it that the identification does not in the least show a different concept of human identity—any more than Christ’s saying ‘Behold your mother, behold your son’ to John and Mary. There is, so to speak, no meta-physics in it. It is the meta-physics in the idea of the Tibetans, or of the Oriental sect that I have supposed, which puzzles us. This man ‘really is’, not in a manner of speaking or for particular social purposes, that man and that one and that one, all long dead.

Of course (for I am modelling my own supposition on something we know of) the belief is connected with a role: the role of the enlightened, the carrier of the solutions [to] the problems of life, who only continue with their successions of lives to confer blessing and enlightenment on men who are still struggling with desire and sorrow. In order to reject the identification one brought up to such a role would have to deny the role itself—deny its validity, or perhaps think some mistake had been made in assigning him to it. No information on scientific facts about life and birth, or his own origin and history, could supply a refutation.

That is why it is so futile to oppose such a belief with the characteristic beliefs going with what is called the ‘scientific outlook’. To the extent that one of this outlook points to facts ascertainable by making repeatable observations, and ‘mundane’ facts ascertainable by historical enquiry, and to canons of evidence and checkability said to be connected with the adoption and maintenance of scientific theories—to that extent he cannot touch the belief that he is opposing. To the extent that he has a philosophy of the sort we may call ‘scientific’, he will want to oppose the whole conception of enlightenment which is the real key to the belief. But here what he says is likely to be weak. There is no incompatibility, even, between the pursuit of knowledge of physics and cosmology and biology according to scientific canons, and the practice of the grossest superstitions. Natural science cannot refute superstition. Take the superstition that it is unlucky to upset salt, but that the bad luck may be averted by throwing a pinch of it over your shoulder; or again to have a rowan tree growing in your garden. How could this possibly be refuted by any scientific considerations? People will speak of such a belief not measuring up to the proper canons of evidence but I don’t suppose anyone knows, and it hardly even seems relevant, whether the general run of people who have upset salt or had such trees in their gardens have led as fortunate lives as others. (For of course there will be other causes of bad luck as well.) And a believer in these superstitions may have some tales to tell in illustration, or believe that his ancestors did. It is very difficult to say what superstition is, though it is rather easy to recognise in some of its forms. Wittgenstein said in the Tractatus: Superstition is belief in the causal nexus. I am not quite sure what he meant, but I believe it was this: superstition is belief that ‘causality’ names a tie between things, that there is such a thing as causality itself which connects up a cause with its effect, so that the cause drags the effect after it by means of the tie. The cause would drag the effect into existence. The idea of the pure causal tie is obviously involved in some kinds of superstition: there is no need felt that there be a ‘how’ in this causality, different in different cases according to the natural properties of things.

The thesis that the type now called ‘scientific’ knowledge is all that enlightenment can be is not itself a truth of any of the sciences, say of physics, cosmology, or biology; nor yet of such a claimant to the name of ‘science’ as sociology. It is rather an—absurd—philosophical thesis. It is absurd because it would itself have to be enlightenment but is not part of the sciences in which it says enlightenment consists.

The only way of partially avoiding this point was offered by the Tractatus. According to it, knowledge—what can be expressed—is only knowledge of the sciences. But what can’t properly be expressed, and is far more important—is, indeed alone important, is the right way of seeing the world. For the mere facts of the world are indifferent. Enlightenment will then by no means consist in scientific knowledge, but it will not consist in anything else that can be said either. The ‘clarifications’ of philosophy turn out, paradoxically, to be improper expressions, strictly speaking nonsense, not framed according to the only strictly correct way of doing philosophy. This would consist in saying nothing except ‘what can be said’, i.e. the propositions of natural science, and, if anyone tried to say something metaphysical, in showing him that he had failed to give meaning to some part of his would-be sentences. The ‘clarifications’ were a sort of apparatus to help human weakness: Wittgenstein calls them a ladder, but that seems the wrong metaphor. A sort of apparatus for cripples, rather, by which we—being so far incapable of standing upright and walking and looking around—may somehow crawl to where we can do so.

The thing that is most striking here is the extraordinary restriction laid on language and truth: restriction to the expression of the different facts about how things are in the world—which way things are that may be one way or another, and where which way they are is indifferent: ‘how the world is, is indifferent for what is higher’. In the same vein Wittgenstein also wrote ‘God does not manifest himself in the world’. At first sight this looks like something we can set in clean opposition to that of Paul (of Tarsus): ‘The invisible things of his— his eternal power and divinity—are manifest to sight since the creation of the world, being understood by the things that are made.’† But, on further reflection, this is not so. What Paul says is paradoxical: ‘invisible’ things are ‘seen’. The invisibility is that of which Wittgenstein spoke: ‘God does not manifest himself in the world’, that is, not in the facts of the world’s being this way rather than that. This is ‘indifferent for what is higher’, i.e. it doesn’t make any difference as far as concerns the existence of ‘what is higher’ which way the world is. But [it is] the existence of the world itself—the ‘that, not the how’, as Wittgenstein called it—that makes God manifest. Only, according to the restriction which by his theory must be laid on language, this cannot be said. It is one example of ‘What can be shown cannot be said’.

The Tractatus theory of language, however, is incredible for many reasons. E.g. it requires that all of logic be a decidable calculus; it requires the existence of atomic propositions and simple objects which, however, can never be found; there are a host of propositions which it is impossible to suppose to be truth-functions of elementary propositions, and which yet are statements of mundane fact. The restriction on language is incredible. In one way there was something greatly attractive about the silence prescribed by the Tractatus: an area to cut off not only much nonsense in philosophy but also much disgusting religious drivel. But that work must be left to criticism. We must be at liberty to discuss, to reason about, all these things: to discuss whether indeed natural science is the only kind of knowledge and enlightenment that there is. And, that being so, our former argument holds. For the decision on that question is [not] in itself a scientific one.

In the Investigations Wittgenstein seems to imagine a purely negative role for philosophy. Its achievement is to clear the ground of the ‘Luftgebaüde’ which we are hoodwinked into constructing, hoodwinked by the suggestions made by the forms of language. There is no suggestion that the ground is cleared for the construction of a better edifice. The positive value of philosophy is the clarity for living in—if one could attain it.

 

* From a typescript without title or date with some written corrections to its first few pages by the author. Insertions in square brackets in quotations from Wittgenstein are by the author. Insertions in square brackets in the author’s text are by the editor; a number were required in what was a largely unrevised typescript. Title supplied.

† Epistle to the Romans, 1:20.

1 This does not mean that no case is imaginable where the question is given a sense. It would be given a sense, for example, if a use of the verb of sensation in the first person present is described, where the seeing referred to is itself an hypothesis or diagnosis on the part of the user.

2 See Philosophical Investigations II, xii.

3 I don’t mean to raise any objection to empty concepts as such; but this one has its native place in an application to certain things. If it doesn’t apply to them, or at least if it can’t be called ‘superstitious’ to do or think any of the things that have been called ‘superstitious’, then surely it is a wrong concept.

4 Philosophical Investigations §342.

5 I only say ‘if’, as I don’t know that any such do say so.

6 It is reported of the present Dalai Lama that when he was a small boy he was rebuked by those who were bringing him up; reference was made to the Great Fifth Dalai Lama—whereupon he rebuked his rebukers with the question ‘Who is the Great Fifth Dalai Lama?’ which quite confounded them.

7 I had the matter described to me by a man from that society, who seemed to me to be truthful and very sensible.