My aim in this chapter is to clarify, by way of looking at PI 38–64, what it is that Wittgenstein is doing where he engages in philosophical diagnosis and therapy, and to show, at the same time, that this is what he is engaging in in that sequence.
Wittgenstein is famous for having canvassed, in his later work, for the idea that the main task of the philosopher consists in freeing philosophizing people of their problems rather than just in clarifying their problems or even in helping them to find solutions. This has commonly been called his ‘therapeutic’ way of doing philosophy. The word ‘therapy’, however, can be misleading since it tends to hide the fact that Wittgensteinian therapy is a completely rational, intellectual enterprise. Let me give an example of what it is not, and contrast it with one of what it actually is.
Imagine you have just proven to your own satisfaction that non-human animals do not have non-derived rights because man, as the most highly developed creature, is the aim and purpose of nature, which means that non-derived rights apply to man only whereas rights of animals, if they exist, must be derived from human rights and from corresponding human duties (like the duty not to do cruel things to animals because this will weaken your disposition to be kind to humans). Now as night falls you are out in a place where there is a wonderfully clear sky with myriads of twinkling stars. You lie down on your back and gaze at the sky in amazement. Stars, enormously far away, enormously many. Where are we? At the centre of the universe? Or are we just lost and lonely, negligible things in a negligible place? Stars, all of them suns like our own one, perhaps with planets like our earth, and perhaps with life on some planets–intelligent life? Who are we? Anything special? Remarkable in any respect? Suddenly you find yourself sobbing.
Your sense of superiority is gone, and thinking again about non-derived rights you simply do not understand any longer how you could think such silly things about a human top position in nature. You do not trust your own reasons any longer–surely they must have been special pleading. You may even come to feel that there is no case to be argued–you have taken resort to moral reasoning because it promised to help you in avoiding a decision on what kind of person you want to be. (That may be different with different people, of course.)
You have undergone philosophical therapy–you have been cured of a moral conviction that had been established by an argument in moral philosophy. Because you permitted yourself to wonder at the stars, and at your position with respect to them, your moral make-up has been disturbed by a deep emotional shock. What distinguishes this from Wittgenstein’s way of therapy is mainly the fact that you did not need to analyse your own reasoning, to look for mistakes in it, and to find out why you made the mistakes in the first place. Part of a therapeutic attempt in Wittgenstein’s way might have gone somewhat like this: ‘Well, there is a top of the staircase, and whoever is on top is the one who has climbed highest. And in a democratic republic, there will be public offices, and you may be elected to a higher office, and if you have been elected to an office such that you cannot be elected to a higher one, then you are president (or prime minister, as the case may be). And if you march steadily north until you cannot march north any further, then you will be at the north pole. But if you march further east than anyone else, does that mean that you have approached the east pole?’ That is, Wittgenstein might have directed your attention to your faulty reasoning from being the most developed creature to being the end of development–without, however, explicitly saying that this was a mistake you committed. On the contrary, his usual practice is to illustrate your mistake without even saying what he is illustrating. You have to find out for yourself; otherwise, you won’t see through your mistake, let alone why you made it. You will not have gained real insight. In my little example, he might have pointed, in this indirect way, to other mistakes instead: that of confusing the last step of a process with its purpose; or that of mistaking the top of phylogenetic development for the top of a moral hierarchy; or that of confusing the top of a hierarchy of rights and duties with the origin of such rights and duties–or what have you. In all likelihood, he would also have directed your attention to possible motives for defending a special human position, and would have done it in the same way–by adducing parallel examples. (Of course, he would have done better than I did just now.) What matters for present purposes is that it is all being done in a perfectly cool, calm and collected style; there is nothing emotional about it. In the interest of insight, it is all reason and argument, even if implicitly so, designed to provoke his readers to think for themselves. And this procedure remains implicit to a degree that can, at times, drive his readers crazy because it can be terribly difficult to find out what he wants to show.
As therapy, in Wittgenstein, is all reason and argument, it can easily be mistaken for the philosophical activity of establishing or refuting a position. Now, both activities can be found in the Investigations, practising therapy as well as establishing certain views. Very often, it is a matter of hard exegetical work to find out what he is doing in a given passage. Our sequence (PI 38–64) is a case in point: at first sight, Wittgenstein is busy refuting a philosophical theory; in fact, the sequence consists of ever so many examples of critical reasoning. But as we have just seen, this is what we have to expect from his kind of therapeutic activity, too; and what I want to show is that it is the latter task Wittgenstein is pursuing here. I understand that such an interpretation is anything but trivial: first, because the theory Wittgenstein seems to be proving wrong is a close relative of a view defended in his Tractatus (which he famously rejected in his later work), and second, because I am obliged to offer a hypothesis on what the target of Wittgenstein’s diagnosis is.
In section 2 of this chapter, I shall briefly indicate why it is acceptable to treat our sequence as a coherent string of remarks. In section 3, I shall try to understand the metaphysical theory, stated in §39 and 46, that it is concerned with, and I shall conclude that Wittgenstein does not tell us what a metaphysically minded philosopher might think its explanatory value to consist in. In sections 4 and 5, I shall argue, by way of detailed analyses of §§40–5 and a brief summary of the remaining remarks, that the text is restricted to philosophical diagnosis and therapeutic suggestions; it does not make a case for or against anything. In section 6, I shall argue that the basic diagnosis is the following: the theory’s alleged achievement consists in a fake explanation of the invented ‘fact’ that a solitary speaker can create a language by helping himself to ostensive definitions.
Since I want to present PI 38–64 as aiming at diagnosing exactly one philosophical misconception, I have to give reasons for thinking that the sequence really forms a coherent whole. This is not seriously controversial (although there may be good reasons for letting it begin with PI 37), which permits me to be brief.
Viewed from inside, all sections of this sequence concern a metaphysical theory about what a genuine name is. This question is never again taken up in the book. Also, our sequence has an unusually perspicuous organization: §§38–9 outline the interlocutor’s problem and his theory. Sections 40–5 examine the idea that having a bearer is necessary for a word to be a name (the name-bearer relation being the primary source of meaning for the metaphysical theory). Section 46 quotes Plato for some themes that are involved in the §39 theory and to be discussed in what follows: simplicity of ‘primary elements’ in §§47–9, their necessary existence in §50, and the naïve relation of designation (of ‘primary elements’ by their names) in §§51–4. The text of §§55–64 makes repeated attempts to diagnose claims, taken verbatim or almost verbatim from §§39 and 46, that are central to the opponent’s theory: §§55–6 on the connection between indestructibility and simplicity, §57 on indestructibility, §58 on the necessary existence of name-bearers, and §59–64 on simplicity and unique analysis.
Viewed from outside, the preceding text, up to §37, has discussed themes from §1, and has succeeded in driving Wittgenstein’s interlocutor into a corner (on which see more below in section 6) where resorting to a helpful theory seems natural. The subsequent text, starting with §65, is highlighted as introducing a new theme by the sentence:
Here we come up against the great question that lies behind all these considerations.
Wittgenstein then starts discussing a new question, maybe the question whether or not, in natural languages, terms are used according to explicit rules, maybe the question whether or not there is just one kind of linguistic activity–whatever you think he is discussing there, it is not the question of the existence or necessity of ‘genuine names’ and of what may follow if they do exist. Thus, our sequence looks coherent from inside, and it appears clearly demarcated from the remaining text. We may safely treat it as a coherent string of remarks.
In this sequence, Wittgenstein examines a kind of metaphysical reasoning; so much will be uncontroversial. That is, Wittgenstein deals with an opponent who invents all kinds of funny things and surprising ‘facts’–all well and good, but why should anyone be interested? Or, to ask the question in more practical terms: why should an author who has been concerned, in PI 1–37, with how things must stand for words and sentences to be meaningful, all of a sudden devote twenty-seven sections to that funny theory? Postponing a detailed answer until section 6, let us use the following heuristic device: Wittgenstein was an outspoken opponent of philosophical theories, as misotheoric as they come. For this professed aversion, just compare these famous lines from PI 109:
And we may not advance any kind of theory. There must not be anything hypothetical in our considerations. We must do away with all explanation, and description alone must take its place.
This is what his declarations come to. It is controversial whether he really followed, or might, indeed, have been able to follow, this sweeping battle-cry in practice; but the controversy need not concern us here since the theory under consideration is sure to be one of those which he really wanted to teach philosophers to get rid of. Note that the lines just quoted also refer to Wittgenstein’s reason for warning philosophers not to build theories: they tend to do it in order to explain an alleged discovery that they have made up in order to extricate themselves from some deep puzzlement. So our heuristics will consist in interpreting the theory as an attempt at explaining an alleged fact, and this alleged fact will have to be one that Wittgenstein’s interlocutor has resorted to somewhere between PI 1 and PI 38. Whether or not the heuristics is fruitful will, of course, have to be measured against its exegetical success.
The metaphysical reasoning is laid out explicitly, and it is laid out twice (§§39, 46). The text, however, interweaves its themes to such a degree that the focus of attention becomes unclear: real names are supposed to need bearers to be meaningful; bearers of real names exist necessarily; bearers of real names are simple; bearers of real names cannot be described (they can only be named); real names are the elementary parts resulting from an analysis of ordinary sentences; unless real names resulted from such an analysis, ordinary sentences would be meaningless. Some parts of the text link these themes in such a way that individual claims look as if they were explained or established by other ones; however, the resulting sketch of a metaphysical theory is incomplete and ambiguous.
I shall begin by trying to understand the ‘Excalibur’ argument from §39 and the major supplementary ideas from the Theaetetus, quoted in §46, in order to find out why a proponent of the metaphysical theory might find it useful.
Part of the idea of the ‘Excalibur’ argument is that when a whole has been destroyed in such a way that its destructible parts, too, (including the destructible parts of destructible parts, etc.) have been destroyed, we arrive at indestructible elements; and these are simple, i.e. they do not have parts. The destructibility of a whole is explained, quasi-causally, by its not being simple (its having parts); this is Wittgenstein’s reason for using ‘to break up’ (‘zerschlagen’) with respect to the destruction of Excalibur. (If Excalibur had been destroyed by melting it down, no unmeltable parts would have resulted.) The existence of simples, on the other hand, or what comes to the same, the ‘fact’ that indestructible parts are simple, is not explained; it is postulated for the purpose of getting the explanation of destructibility going. This is a metaphysical inference to the best explanation; it makes use of the additional independent postulate that the destruction of parts, of parts of parts, etc., cannot go on forever. Given the speculative boldness manifested by the whole enterprise, the additional postulate may seem to reflect the natural and insignificant mistake of supposing that for a complex to be composite, there must have been a finite process of composing it. It is, however, also an instance of a more general idea which I call the ‘this-cannot-go-on-forever’ postulate. I shall comment on its relevance to the interlocutor’s clinging to his metaphysical theory in section 6(b).
The ‘Excalibur’ argument assumes that if an object a consists of the same parts as an object b, and if both objects have been composed from these parts in different ways,1 then if α is the name of a, it is not the name of b. Furthermore, the argument assumes that if the composition of the parts of an object a has been undone, then a does not exist any longer, and if α used to be the name of a, it is no longer a (meaningful) name. The second case may even be considered a special case of the first one; for the heap of pieces that remains after Excalibur has been decomposed is just another object consisting of the same parts differently composed. In Wittgenstein’s view, it is important for the metaphysical theory that putting together the same parts in different ways results in different complexes; for he spends several sections in examining its converse, namely, the claim that there is just one way of dismantling a complex into parts. In §39, the idea is presupposed: ‘So the word “Excalibur” must disappear in the course of an analysis of the sense’*;2 I take ‘an analysis’ to mean ‘the only correct analysis’, which implies that for a given sentence, there is just one translation that contains only ‘real names’. The ontological counterpart is, of course, that there is one and only one way in which a complex is composed of simples; putting together the same elements in a different way would not result in the same complex. Sections 48–9 treat this idea directly with reference to the Theaetetus passage, and §§59–64 treat its semantic version that sentences have just one analysis by which to arrive at their real meanings. (In §60, there is the same transition from the idea that a broom is actually composed of broomstick and brush to the idea that a sentence about a broom is really about broomstick and brush, as there is, in §39, the transition from the idea that if Excalibur has been destroyed–which means recomposed in the form of a heap–then the sentence ‘Excalibur has a sharp blade’ is no longer about anything.)
Now the difficulty the argument is designed to treat is not anxiety about the stability of the universe; it has to do with the fear that names might loose their meanings if all things whatsoever were constantly recomposed:
But if ‘Excalibur’ is the name of an object, this object no longer exists when Excalibur is broken in pieces; and as no object would then correspond to the name it would have no meaning.
(PI 39)
I suggest the connection between that fear and constant recomposition to be the following: recomposition means change, change means alteration, alteration means loss of identity, and what has lost its identity cannot keep its name. Thus name-bearers must have no parts because they must be unalterable; or to put it differently, they must be simple because they must have no properties. Only that which has properties can change. Elements are, in fact, treated as properties rather than as something which has properties when it comes to the opponent’s last line of retreat in §§56–8.
This aspect of simplicity is emphasized by the quotation from the Theaetetus. Wittgenstein quotes the passage3 as an answer to the following question: ‘Now what is the point of saying that names really signify simples?’* (§46). We may assume that he has chosen the most informative answer he has found in the text. And what this passage–in his heavily edited version, see section 6(d)–contributes to the meaning of ‘simplicity’ is just this:
everything that exists in its own right can only be named, no other determination is possible,
and
what exists in its own right has to be … named without any other determination (Wittgenstein’s omission).
The result of attributing a property F to an object a is complex in the sense that if you point to a which is F, you may be pointing either to the object a or to its property F, so that your pointing gesture will be ambiguous. So I take ‘simple’ to mean ‘without properties’, and this fits very well with ‘indestructible’ meaning ‘unchangeable’.
In §50, the metaphysical picture is enriched by the idea that ‘primary elements’ exist necessarily. Necessary existence is more than indestructibility, which only means that if a primary element exists, it will do so forever. Indestructibility is not a property worthless in a guarantor of meaning; for it ensures that a name-bearer will last and that christening it will fix the given name’s meaning forever. But it must first be there in order to be there forever. The general idea underlying the proof of their necessary existence is that otherwise ‘primary elements’ could not even be named, and a related idea is used in §55: since the truth of a description is independent of whether or not it is actually given, the state of affairs that would exist if it were true does exist if it is true (even if nobody has actually given the description). Therefore, the primary elements of which this state of affairs would be composed if it did exist, do exist even if it does not exist itself, because otherwise the possible description could not become true.4 I do not see a connection between this kind of necessary existence and indestructibility and take it to be a relatively independent element; in fact, Wittgenstein’s way of extracting it from Plato’s report that ‘no other determination is possible, neither that it is nor that it is not’ (which, according to Plato, was just one among several forbidden determinations) testifies to his particular interest in adding this element to his target theory.
Even if we have a moderately clear idea of how the elements of the theory are thought to be connected with each other, we still do not know what, in Wittgenstein’s view, it is intended to explain. (Remember that, according to our heuristics, he denounced philosophical theories as pretending to have genuine, while possessing merely bogus, explanatory value.) He remains silent on this question. Section 39, containing the ‘Excalibur’ version of the theory, begins as follows:
But why does it occur to one to want to make precisely this word [‘this’] into a name, when it evidently is not a name?–That is just the reason. For one is tempted to make an objection against what is ordinarily called a name. It can be put like this: a name ought really to signify a simple.
Then the first statement of the theory follows. It is intended to ‘show’ that ‘real names’ of ‘simple objects’ are the only guarantors of linguistic meaning; but there is eloquent silence, on Wittgenstein’s part, as to what the theory could possibly explain about the fact, presupposed implicitly, that ‘this’ could be used, in the role of a ‘real name’, to signify simples. Diagnosing a philosopher’s motive for holding a theory requires finding out about its alleged explanatory value; holding the theory is just a syndrome to be diagnosed.
Wittgenstein’s silence on this matter is worth noting because he describes the metaphysical theory–a syndrome characteristic of an, as yet unspecified, illness–in so much detail. More often than not, he leaves to his readers even the task of working out what the symptoms are he is busy diagnosing. Take, as a concise example of his more usual style, this beautifully phrased5 exchange from PI 435:
If it is asked: ‘How do sentences manage to represent?’–the answer might be: ‘Don’t you know? You certainly see it, when you use them.’ For nothing is concealed.
How do sentences manage it*?–Don’t you know? For nothing is hidden.
Here, the reader is not even told what the symptoms of the underlying philosophical illness are, how it manifests itself; fortunately, however, he has been told before, at PI 93, that what Wittgenstein has in mind is the idea of sentences achieving it to represent a state of affairs (Anscombe’s word for the German ‘Satz’, translated as ‘sentence’ in PI 435, is here ‘proposition’):
[The enormous importance of propositions,] together with a misunderstanding of the logic of language, seduces us into thinking that something extraordinary, something unique, must be achieved by propositions.–A misunderstanding makes it look to us as if a proposition did something queer.
Thus at PI 435, unless the reader remembers the earlier passage, he can only try to make sense of Wittgenstein’s repetition of the seductive phrases ‘How do sentences manage?’ as well as of the reassuring remarks ‘Nothing is concealed’, ‘Nothing is hidden’, and if he proceeds this way, he may be lucky enough to hit at what Wittgenstein wants to expose, namely, the idea of a sleight of hand by which the sentence manages to do the representing in the role of an active subject. (By stressing, in §93, ‘misunderstanding’, Wittgenstein also hints at the diagnosis: the expression ‘a sentence’ can occur as a syntactical subject in sentences whose predicates denote ‘activities of representation’–a sentence expresses, reports, explains, refers, etc.)
As far as I can see, this is Wittgenstein’s usual way of prompting his readers to work out for themselves that a given phrase really displays symptoms of philosophical confusion–they really have to take the trouble to find out what his remarks are diagnoses of, rather than look for theses he might be engaged in refuting. In order to establish that our sequence, as a whole, exhibits this very character of an attempt at diagnosing philosophical confusion, I shall show this in detail in the next two sections, dealing chiefly with §§40–5 (postponing some remarks on §38 to section 6, and giving only a brief summary of the diagnostic points in §§47–64).
At first sight, the beginning of §40 looks as if it introduced a critical discussion of the ‘Excalibur’ argument that precedes it:
Let us first discuss this point of the argument: that a word has no meaning if nothing corresponds to it.
If construed as arguing against this premise, the remainder of the section would be weak for three reasons. First: if
the word ‘meaning’ is being used illicitly if it is used to signify the thing that ‘corresponds’ to the word
then there simply are no cases where ‘the meaning of “Mr N.N.”’ is (in fact, i.e. successfully) being used to signify Mr N.N.; on the other hand, if that is supposed to mean that the sentence ‘Mr N.N. is the meaning of “Mr N.N.”’ is linguistically deviant, then Wittgenstein is begging the question. Second:
When Mr. N.N. dies one says that the bearer of the name dies, not that the meaning dies. And it would be nonsensical to say that, for if the name ceased to have meaning it would make no sense to say ‘Mr. N.N. is dead’.
Wittgenstein here presupposes that a person who dies ceases to exist; but a defender of the argument is not forced to accept that, and it is far from clear that when we say, ‘Pegasus does not exist’, we are using ‘Pegasus’ as a name. Third, nobody ever tries to refer to Mr N.N. by using the expression ‘the bearer of the name “Mr N.N.”’. Therefore, ‘one says that the bearer of the name dies’ etc. has to be read de re: ‘one says of the bearer of the name that he dies; one does not say it of the meaning of the name’. In this context, that cannot be used as a premise because it is the very point at issue. On the whole, then, the section would be completely inconclusive if it were construed as an argument against the point ‘that a word has no meaning if nothing corresponds to it’. However, it is quite convincing if construed as a diagnosis of how someone might come to hold this opinion. For if he, in fact, confused the meaning of a name with its bearer then he would be forced to say that without a bearer nothing can be a name.
In §41, Wittgenstein refers to the language-game of §15, where each tool was marked with its own sign; the builder could show his assistant one of the signs (as inscribed on a label), and the assistant would have to fetch the corresponding tool. The present section describes three situations which agree in that the sign the builder shows his assistant is no longer attached to any intact tool and that this failure, on the part of the sign, causes the mark to lose or change its use and, thereby, its meaning. Therefore, being attached to a tool–‘having a bearer’–is relevant to a mark’s being a meaningful name to the extent that the language-games the mark is used in causally require that people can actually have dealings with the bearer. This may be the case in some language-games,6 and if we think only of these we will be prone to think that there are no names without a bearer. This, then, is the diagnostic point of the section.
There are two ways to read the first sentence of §42. (We are still dealing with variations on the language-game of §15.) Anscombe translates the German ‘etwa auch’ as ‘for instance also’:
But has for instance a name which has never been used for a tool also got a meaning in that game?
This is lexically possible; however, it fits better with the exchange between Wittgenstein and his interlocutor to read the sentence as the latter’s rhetorical question (translating ‘etwa auch’ as ‘even’):
But have even names that have never been used for a tool got meaning in that game?
That is, the interlocutor expresses his bewilderment in view of the possibility of someone’s going so far in neglecting the necessity of name-bearers. Once again, Wittgenstein’s answer would be a weak reason for the conclusion that there may, in fact, be such eternally void names; for he describes just one possible use, namely, that the assistant has to shake his head when shown such a sign, which might be ‘a sort of joke between them’. Certainly, this use does not characterize the sign as a name. On a charitable interpretation, Wittgenstein does not discuss the interlocutor’s question but asks him to imagine situations, perhaps even fantastical ones, that will help him to see that having dealings with a bearer may be just one among many features of using a sign that turn it into a name.
Postponing my interpretation of §43 to section 5, I now turn to §44, where Wittgenstein asks us to imagine a use of names that ‘are used only in the presence of the bearer’; would these be the metaphysical theory’s ‘real names’, as distinct from ‘Excalibur’ and other ordinary names? This might be a diagnostic step: has the opponent rashly generalized from such examples? Actually, there are such names, e.g. certain ceremonial forms of address such as ‘Your Majesty’. However, they are simply too rare to lure anyone into a general theory; there may be more to the idea:
But we can imagine a language-game with names (that is, with signs which we should certainly include among names) in which they are used only in the presence of the bearer; and so could always be replaced by a demonstrative pronoun accompanied by* the gesture of pointing.
What matters for ‘Excalibur’, as distinguished from such names, turns out to be that if Excalibur has been destroyed it is no longer present rather than that it does not exist any longer. The therapeutic intention has shifted from the need for a name-bearer to the question whether names are really (or whether ‘real names’ are) similar to demonstratives. Since the following §45 abandons this question and prefers to ask whether demonstratives are names, we are expected to answer the present question for ourselves. A simple answer would be: even if using the names in question could be replaced by demonstrative uses of ‘this’, that would not turn them into demonstratives, because they would still lack most of the characteristic uses of the latter ones.
Perhaps Wittgenstein is here hinting at a deeper point. Speaking about what is present in one’s surroundings may, indeed, be a use of language which is basic in some way or other. And in such cases, you can direct your own and other people’s attention to what you are speaking about. This fact, however, must not be confused with the idea that you can speak about things because you can direct your own and other people’s attention to them.
The first half of §45 presents a fine example of a peculiarly Wittgensteinian technique:
The demonstrative ‘this’ can never be without a bearer. It might be said: ‘so long as there is a this, the word “this” has a meaning too, whether this is simple or complex’.
This way of expressing the matter lends itself equally well to reporting metaphysical discoveries as to describing the trivial facts that the discoveries boil down to. Reading the sentences metaphysically, we get:
‘This’, the demonstrative, has a bearer of necessity. Or even: Let there be a possible object of attention; this will suffice for ‘this’ to be meaningful, and this can be achieved with simple as well as with complex objects.
Reading the sentences in a down-to-earth way, we get something very trivial:
Unless something is pointed to, ‘this’ is not being used in demonstration. To put it differently: Whenever you point to something while uttering ‘this’, the word ‘this’ is being used with its usual meaning; it does not matter whether what is pointed to is simple or complex.
I suggest that Wittgenstein uses this technique to remind his interlocutor that he may overstate ordinary facts as much as he pleases–it won’t help him, for
that does not make the word into a name. On the contrary: for a name is not used with, but only explained by means of, the gesture of pointing.
Before turning to my last detailed analysis, that of §43, let me summarize without argument what I take to be the main diagnostic and therapeutic points of the remainder of our sequence: the idea that there is simplicity independently of any context, i.e. prior to all language-games (§47)–an idea which is grounded in one context’s being so suggestive that one tends to be blind to the possibility of moving to another one (§48); confusing an activity like that of marking an object with a sign with establishing minimal linguistic contact through pure naming (§49); sliding down the following chute: for the name to be used in a given language-game, it is necessary that a bearer of the name exists–it is necessary that a bearer of the name, as used in this language-game, exists–it is necessary that a bearer of the name, which is used in this language-game, exists–it is necessary that a bearer of the name exists–necessarily, the name has an existing bearer–the name has a necessarily existing bearer–there is a necessarily existing object that is the bearer of the name (§50); failing to think about what ‘being the name of’ may really amount to and therefore failing to realize that this relation can obtain in virtue of the most diverse sorts of situations (§§51–4); making the very same mistake of oversimplification with respect to the roles of standards in the use of words (§§55–6); confusing the senselessness of a sentence with its expressing a metaphysical impossibility (§§57, 58); generalizing from the uniquely determined construction of artefacts out of parts to a unique composition of the world (§59); misconstruing statements about things that are composed of parts as statements about the parts and their way of composition (§60); generalizing from the observation that ‘having the same sense’ is an intelligible relation in some contexts to the prejudice that it is intelligible in any context (§61–2); confusing analysis with enlightenment and knowledge of parts with basic knowledge (§§63–4).
From these examples, we can learn a very important lesson about Wittgenstein’s diagnostic and therapeutic procedure: he introduces philosophical observations to use them in arguments; but these arguments are designed to lead his interlocutor to understand where his philosophical thinking has gone astray. In such cases, Wittgenstein’s philosophical observations and statements are not used to proclaim, or to insist upon, or to establish philosophical points of view. There is, in our sequence, an often-quoted half of a section that exhibits this characteristic in the extreme. It is perhaps the most frequently quoted paragraph in the whole Investigations, but it has practically always been quoted illegitimately, as if it expressed a statement of a philosophical point of view: §43a.
Section 43a has been treated, again and again, as Wittgenstein’s statement of his ‘use theory of meaning’–notwithstanding the blatant fact that if it were such a statement, it would lack any recognizable point and purpose in its context:
For a large class of cases–though not for all*–in which we employ the word ‘meaning’ it can be explained* thus: the meaning of a word is its use in the language.
Since the explanation of the word ‘meaning’ is restricted to ‘the meaning of a word’, the paragraph says that in cases which belong to the ‘large class’ and where the word ‘meaning’ is used to refer to word meaning, the word ‘meaning’ (nothing else, and in particular, not word meaning–the reference is unambiguous in the original) can be explained in a certain way. Now if someone says, ‘The word “A” can be explained thus: A is B’, he is employing the so-called material mode of speech–‘A is B’ looks like a statement about the objects A and B, but the utterance is intended as a linguistic explanation to the effect that the expression ‘A’ can be replaced by the expression ‘B’. That is, when Wittgenstein writes, ‘the meaning of a word is its use in the language’, we are supposed to understand this in the sense of ‘the expression “the meaning of a word” can be replaced by the expression “the use of this word in its language”’. ‘Cases where we employ the word’ are contexts where we use it. So what §43a says is: In many, though not in all contexts which contain the word ‘meaning’ as referring to word meaning, the word ‘meaning’ can be replaced by the expression ‘use in the language’.
What is Wittgenstein up to? It might be advisable to consider at least the immediate context, §43b:
And the meaning of a name is sometimes explained by pointing to its bearer.
Now, since names are words, the meaning of a name is the meaning of a word; therefore, §43b is a context which contains the word ‘meaning’as referring to word meaning. Section 43a reveals itself as a recommendation to reformulate §43b–provided Wittgenstein intends the case of §43b as one which belongs to the ‘large class’ mentioned in §43a, where ‘meaning’ can be replaced by ‘use in the language’. Does he intend that? Compare: ‘On many evenings–though not on all–I go to the movies. And tonight…’ in contrast to: ‘On many evenings–though not on all–I go to the movies. Tonight, however,…’. Unless §43b is in the ‘large class’, the text should read ‘The meaning of a name, however,…’; however, it reads ‘And the meaning of a name…’. Accordingly, §43b is in the ‘large class’ where the word ‘meaning’ can be replaced by the expression ‘use in the language’. If we perform this substitution, we get:
And the use, in the language, of a name is sometimes explained by pointing to its bearer.
This is so true that it is close to being trivial. We explain the name’s use in the language by explaining, with the help of a pointing gesture, who is to be called by that name. We do this only sometimes, namely, when the bearer is present and when the conditions discussed in §§30–1 are satisfied (as they usually are): The addressee has to know that it is the name of a person which is being explained, and he also has to know how to use a person’s name. What is the point of telling us all that?
Remember the diagnostic points made in the context of §43. Section 40 has diagnosed the mistake of confusing the bearer of a name with its meaning. Sections 41–2 try to understand why one might nevertheless be misled: we are so used to having dealings with a name-bearer when using a name that we mistake the name-bearer for the only thing that is relevant for the name’s meaning. Section 44 envisages language-games where names are used only in the presence of their bearers; §45 notes that the demonstrative ‘this’ cannot be used without something being pointed to (something similar to the bearer of a name, only that ‘this’ is no name).
Placed among these remarks, §43 can be seen to be a further diagnostic suggestion (which loses its cleverness when paraphrased): ‘You think that the name’s bearer is the name’s meaning because you can point out its meaning by pointing to its bearer. Now look: isn’t it possible very often–sure, maybe not always, but don’t bother about other cases now–to speak of “the use in the language” instead of “the meaning” of a word; and isn’t it certainly so when we speak of explaining the meaning of a name by pointing to its bearer (instead of explaining the name’s etymology or describing the bearer)? What we explain in such cases is whom to call by that name, i.e. how to use the name in the language. But you would not, for this reason, mistake the name’s bearer for the name’s use; then why should you, for that very reason, mistake it for the name’s meaning?’
The job of §43a is quite obvious, then. It helps us to remember what it is that is being explained when a name’s meaning is explained by pointing to its bearer. The diagnosis is helpful only for an interlocutor who agrees that in the class of cases under consideration the word ‘meaning’ can, in fact, be replaced by the expression ‘use in the language’; so this is supposed to be common ground, and Wittgenstein has no business at all to state it as a ‘use theory of meaning’.7 This is not to deny, of course, that he did subscribe to a view that comes close to such a theory; it just means that he cannot be taken to have proclaimed this view, carefully qualified by ‘though not for all [cases]’, in this very passage–the only one where it looks as though it had been formulated explicitly.
I think that PI 43a presents a particularly convincing example in support of my suggestion that, whenever Wittgenstein is busy clearing up puzzles and confusions in a completely perspicuous, rational, well-argued way and thus cannot help using premises and establishing conclusions, his way of presenting things must not be misunderstood–establishing certain conclusions is not, then, the point of his discussion. For diagnosis and therapy to be successful, any philosophical background that is necessary for the imagined rational dialogue has to be common ground among the partners.8
At the end of section 3, I noted that even if content and structure of the metaphysical theory can be clarified, it still remains unclear which alleged fact it has been designed to explain. That is, even if readers agree that what the theory purports to establish is that there is a guarantee for linguistic meaningfulness thanks to ‘real names’ of ‘simple objects’ and that the latter exist in some preferred way, they may still disagree about the problems a philosopher might intend to solve by proposing such a theory. For the puzzlement the theory is meant to help him get out of has not been stated explicitly.
I take the alleged fact the metaphysical theory is designed to explain to be the following: a previously a-lingual subject that is restricted to his own mental resources can, by attending to some object–by mentally pointing to it–christen this object in such a way that its name is now a word with an unambiguous meaning. There are four main reasons for this interpretation: first, read in this way, the sequence fits in well with the preceding text of the Investigations; second, it makes good sense of the philosopher’s using the ‘this-cannot-go-on-forever’ principle in his theory; third, §38 is meant to attribute precisely this idea of an isolated christening to Wittgenstein’s opponent, and to ridicule it; fourth, the interpretation makes sense of the fact that Wittgenstein treats demonstratives as instances of ‘real names’.
(a) From the very first section of the Investigations, Wittgenstein has been dealing with an opponent who has tried to explain the meaningfulness of expressions by attributing to language learners or speakers competences they are claimed to possess independently of their mastery of a language but which are, in fact, disguised linguistic competences. This begins in §1. It is uncon-troversial that among its targets is the idea that words have meanings because they designate objects and that sentences have senses because they are combinations of such designations. But Wittgenstein’s own comment, in §32, on his quotation from Augustine’s Confessions in §1 shows that there is more to the idea, something that makes it interesting and important:
Augustine describes the learning of human language as if the child came into a strange country and did not understand the language of the country; that is, as if it already had a language, only not this one. Or again: as if the child could already think, only not yet speak. And ‘think’ would here mean something like ‘talk to itself’.
This comment is correct because what the quotation in §1 claims little Augustine did in learning his first language presupposes competences which might even be considered as metalinguistic: he ‘saw this’, namely, that his elders ‘named some object’; he ‘grasped that the thing was called by the sound they uttered as they meant to point it* out’; what they wanted to point out he ‘found out from their gestures’*.
Thus the way Augustine, as quoted in §1, purports to have learnt to speak presupposes that he could talk to himself all along–that he could talk all along, for short. The same point has been made in §25: the explanation that animals ‘do not think, and that is why they do not talk’ is refused bluntly, being an explanation of something by itself. Augustine assumes that some process of acquiring linguistic competence is necessary, but he implies that in the course of this process, the learner can avail himself of precisely the competence he is supposed to be acquiring. In this picture, the ability to speak needs a foundation; however, this foundation is supposed to be provided by the speaker himself. The child Augustine that appears in §1 is described, in §32, as mastering language single-handedly, without any external backing. Now in §32, Wittgenstein explicitly refers to §1, and there he comments on Augustine’s learning story that it gives us a ‘picture of the human language’ according to which ‘the individual words in language name objects–sentences are combinations of such names’. The minimal explanation of milking this idea out of the learning story, as commented on in §32, seems to be: in Wittgenstein’s view, the picture of words being names is fostered by the idea that man can be a speaker all by himself.
Sections 32 and 24 are no incidental remarks; both carry particular weight because they mark clear caesuras: with §24, Wittgenstein has finished the direct discussion of the §1 theory of the meaningfulness of expressions (words from §2 to §19a, sentences from §19b to §24), and §32a is the end of the discussion of ostensive definition. Ostensive definition, in turn, has no clear connection with the §1 theory unless it is the means by which the ‘meaning is correlated with the word’, i.e. unless pointing is involved in the origin of meaning; this is why Wittgenstein turns to the idea of pointing in §§33–6. He tries to show his opponent that even the capacity to point to something unambiguously presupposes linguistic competences, and having done that, he brings the discussion to an end with the exclamation mark of §37. Isn’t it a kind of last resort if the opponent now falls back on a theory that would explain how an a-lingual subject can, in fact, point to something unambiguously, and can thereby christen it, and can thereby create linguistic meaning? There may be other reasons for inserting the sequence §§38–64 right here; but this one seems to be quite a good reason anyway.9
(b) We noted in section 3 that the metaphysical theory commits the mistake of supposing that breaking a complex down into parts cannot go on forever, that this process must eventually conclude with parts that cannot be broken into yet further parts. This amounts to the same as saying that for a complex to be composed of parts there must have been a finite process of putting it together. I called this an instance of the ‘this-cannot-go-on-forever’ principle. This is a common kind of philosophical fallacy; think of the ancient atomists, or of Aquinas’ idea, implicit in his First Cause argument for the existence of God, that for a chain of causes reaching to the present it must have begun a finite number of steps ago. It has a double function for the alleged explanatory achievement of the metaphysical theory. First, remember from (a) that the holder of that theory presupposes that linguistic meaning is brought about by linguistic means (which was the point of §§32 and 25). Now if you explain words verbally, ‘this cannot go on forever’, and the normal way out is resorting to ostensive definition: you explain what ‘red’ means by pointing to something red and saying, ‘This is red’. So a holder of the principle needs to be able to appeal to an opportunity of ostensive definition at the start of explanation. Our example of the ‘this-cannot-go-on-forever’ principle provides such an opportunity–partless objects which you can point to in order to define your words unambiguously. In fact, from the days of the British empiricists, philosophers have tried to explain how words become meaningful (an enterprise that has not always been distinguished from an explanation of how concepts are formed) by postulating what they called simple ideas to start with.
(c) I think that §38 confirms my view of our sequence’s point and purpose to a reasonable degree. For if meaning originates in pointing, then it is tempting to jump to the conclusion that original names should be very much like demonstratives (§38a), a temptation which Wittgenstein explains twice, in b and d. His explanation in d is easy to understand:
Naming appears as a queer connexion of a word with an object.–And you really get such a queer connexion when the philosopher tries to bring out the relation between name and thing by staring at an object in front of him and repeating a name or even the word ‘this’ innumerable times. For philosophical problems arise when language idles*. And then* we may indeed fancy naming to be some remarkable act of mind, as it were a baptism of an object. And we can also say the word ‘this’ to the object, as it were address the object as ‘this’–a queer use of this word, which doubtless only occurs in doing philosophy.
The philosopher ignores the actual use of the word, trying instead to use it all by himself in addressing an object by means of it, and hoping in vain to manage in this way to call the object by its name–a deplorable sight, signalled by Wittgenstein with a pun that turns on the double meaning of the German ‘Verbindung’, translated as ‘connexion’. It can also mean a chemical compound, and if we take up this suggestion and read the second occurrence of the word accordingly, then what is accomplished by the attempt to connect word and object is nothing but a kind of tableau consisting of an object and a stammering philosopher.10
Let us now turn to Wittgenstein’s comment, in §38b, on §38a:
[…] Yet, strange to say, the word ‘this’ has once* been called the only genuine name; so that anything else we call a name was one only in an inexact, approximate sense.
This queer conception springs from a tendency to sublime the logic of our language–as one might put it.
In order to understand this, we have to make up our minds what the ‘tendency to sublime the logic of our language’ might be. The text of this section contains no hint to anything elevated (and is in this respect arguably different from the use of the word ‘sublime’ in §89); on the other hand, something has saliently been removed from the tableau consisting of the philosopher and the object he tries to address, namely, uses, in language-games, of the word ‘this’–the word the philosopher utters in his vain attempt at addressing the object. For this reason, I take the ‘tendency to sublime the logic of our language’ to be a tendency to separate or purify the rules of our language from what they are rules for, namely, from language use. When language is not being used, it is, figuratively speaking, inactive, ‘is idling’, and the philosopher gets into trouble by trying to find out anything about it while actively disregarding its embedding in human activities.
Wittgenstein is here using ‘to sublime’ in the sense of the chemical method of purifying substances by evaporation and subsequent condensation–a cunning comparison, because in order to get at the real thing, you have first to convert it into something ghostly. The word ‘to sublime’ is explicitly glossed as meaning ‘to purify’ in §94, and allusions to the idea of an essence’s being something to be got at in its pure form only by wrenching it from seclusion pervade the first part of the sequence §§89–133, which treats the mistake of regarding rules of language as use-independent, in §§89–108. (See §§91–2, 97, 98, 100–2, 105, 107, 108.) The points of the explanations in b and d are one and the same, then: it is a mistake to fancy one could make linguistic contact with an object all by oneself. It is clear from the example–demonstratives–that Wittgenstein takes his opponent to think that this solitary kind of contact is to be achieved by pointing or attending; I turn to this issue now.
(d) Supposing a self-sufficient subject’s alleged capacity to point unambiguously to be ‘explained’ by the metaphysical theory makes sense of the close connection Wittgenstein sees his opponent as attempting to establish between ‘real names’ and demonstratives. (This was the objective entertained by Russell, whom I take to be the target of the unspecified reference in §38a;11 cf. 1918: 524–5, and 1914: 445–6.) Letting alone the fact that the sequence commences, in §38, with this very example, it is also discussed in §§44–5 (see above, section 4), which the reader can only construe as discussing the §39 Excalibur argument; however, in the Excalibur argument there is no mention of demonstratives at all. Wittgenstein is apparently eager to present the quest for ‘real names’ as something very close to the quest for expressions that are as close as possible to demonstrative pronouns. He even goes so far as to touch up his Theaetetus quotation in a way that would have earned him some red ink from a tutor. Remember that the passage is quoted as a classical example of the metaphysical theory; there are two omissions (which Wittgenstein has marked by dots), and I supply the text of the first one (translating Preisendanz’s translation):
[With any attribution over and above pure naming of a primary element] one would attribute to it being or non-being; but one ought not to attribute anything to it if one only wanted to name it in and of itself. For it ought to be loaded neither with an ‘itself’ nor with a ‘that’ nor with an ‘each’ nor with an ‘alone’ nor with a ‘this’, nor with any other attributions of a similar kind. For these are the most general expressions since they are joined to all words; however, they are not the same as the words they are connected with.
The italics are mine; Wittgenstein suppresses Plato’s report that his informants had forbidden to load a ‘primary element’ with a ‘this’; he conceals a detail because it does not fit his presentation.
Demonstrative pronouns are, of course, precisely those linguistic expressions whose use depends on successful unambiguous pointing. (Sections 44–5 are particularly explicit in this respect.) That is, for pointing to be successful, there must be something that is pointed to, and this something must be simple for the pointing to be unambiguous.
Let me add, as a concluding remark, that my interpretation of what the alleged explanatory achievement is intended to be does not explain why the theory postulates the breaking down or ‘analysis’ to be unique. Therefore, the interpretation is certainly not the best one could think of. However, I hope it is an interpretation that gives one something to think about.12
1. ‘If they [the parts] are combined differently Excalibur does not exist’ (§39).
2. The asterisk marks modifications of the Anscombe translation. I read the German text as implying that there is just one analysis.
3. From Karl Preisendanz, Platon, Protagoras und Theaitetos, Jena 1925 (201 E–202 B).
4. This is not new, of course (see Baker and Hacker 1980: 165), but I like the idea of finding it in the text without help from the Tractatus.
5. This is a comment on the German text.
6. Wittgenstein returns to this idea in §50, when he presents his diagnosis of the idea that ‘primary elements’ exist necessarily.
7. Specht has seen that paragraph b is what matters without, however, recognizing the diagnostic aim (1963: 108–9); Baker and Hacker have noticed the diagnostic aim without, however, recognizing the connection between the two paragraphs (1980: 243).
8. To the extent that Wittgenstein uses statements, suppositions, or observations for purposes of rational diagnosis, he is, of course, committed to sticking to them in other contexts, too. By making a large number of such moves he would come to commit himself to certain comprehensive views on some questions. This, however, is not the point of the present chapter.
9. Baker and Hacker have clearly shown that, as regards ostensive definition, Wittgenstein disagrees with his opponent precisely on the point whether or not it needs linguistic background (1980: 171–2, 179–80), and they have made the same point in passing about pointing gestures (ibid. 184). My reading differs from their interpretation in making this matter Wittgenstein’s starting point for our sequence.
10. I would like to add an observation which is irrelevant to the interpretation because probably unknown to Wittgenstein, and not used by him anyway, but, to my mind, very amusing. The German for ‘address’ is ‘ansprechen’; ‘entsprechen’, German for ‘correspond’–a word often used for referring to the relation between an object and its name–meant ‘to answer’ in early Modern High German but became obsolete. (Interestingly, ‘to answer’ seems to be proper English for ‘to correspond’ in the sense of ‘to fit’.) An object that ‘entspricht’ (corresponds) to a name would (in that obsolete state of German) be an object that answers if it is being addressed with its name.
11. Not an original idea, of course; cf. Baker and Hacker 1980: 224–5.
12. I would like to thank Joachim Schulte for turning my draft text into intelligible English.
Baker, Gordon P. and Peter M.S. Hacker (1980) An Analytical Commentary on the ‘Philosophical Investigations’, vol. I, Wittgenstein, Understanding and Meaning, Oxford, Blackwell.
Russell, Bertrand (1914) ‘On the Nature of Acquaintance’, The Monist 24: 1–16, 161–87, 435–53.
Russell, Bertrand (1918, 1919) ‘The Philosophy of Logical Atomism’, The Monist 28, 1918, 495–527, and 29, 1919, 32–63, 190–222, 345–80.
Specht, Ernst K. (1963) Die sprachphilosophischen und ontologischen Grundlagen im Spätwerk Ludwig Wittgensteins, Kant-Studien Ergänzungsheft 84, Köln.