W.V. Quine and Ludwig Wittgenstein have enjoyed and suffered similar fates. Their voices were dominant in philosophy in the middle of the twentieth century and for some time afterwards. Since then, their influence has dwindled. The dwindlings have differed in nature. Wittgenstein has a dedicated group of followers who labour both to interpret his work and to apply his ideas, but these labours are often ignored by philosophers outside the group. Quine has suffered an even crueller fate: he has become an inspiration. He is routinely credited as a pioneer of philosophy’s naturalistic turn, but it is becoming rare to find philosophers engaging with the detail of his arguments.
Perhaps this is just and right; perhaps their reputations were inflated, and philosophy was right to move out of their shadows. But it is also possible that a reassessment is due. One way to reassess both thinkers is to make them communicate with each other. This paper is meant as a small contribution to that project.
There are obvious and massive differences between these two thinkers.1 Perhaps most obviously, Quine believed in integrating philosophy and science, whereas Wittgenstein’s tendency was to separate them almost entirely. Nevertheless one has a vague feeling, in reading both authors, that their thought is similar in ways that are difficult to capture. Quine himself felt this similarity, and said so, but he never expanded on the point.2
In section one I define a project in which, I claim, Quine and the later Wittgenstein3 were both engaged. In section two I discuss Quine’s approach to the project; in section three I turn to Wittgenstein. These sections together display the differences between these philosophers. In section four I defend my view that beneath these important divergences lies a deeper convergence.
In brief, the crucial difference lies in their conception of the environment which humans confront. For both, reference to this environment explains the content of our thought and talk. But Wittgenstein’s conception of the environment is much richer than Quine’s. As a slogan, I will say that for Quine the stimulus is impoverished, while for Wittgenstein it is not.
The convergence lies in a certain deflatory attitude to questions of ontology. But Wittgenstein carries out this project more thoroughly. I shall argue that it is precisely the poverty of the Quinean stimulus that prevents Quine from going as far as Wittgenstein does in the deflation of ontological questions.
§ 1. The project
Quine and Wittgenstein have an enemy in common. As Wittgenstein puts it in the Blue Book, this enemy is the illusion of explanation that arises when one attributes the meaningfulness of discourse to “something in the occult sphere” (1958: 5). What is the occult sphere, and how does this illusion of explanation arise?
Suppose that we are interested in understanding the meaningfulness of some utterance, say the command “Shut the door!” How did this utterance come to mean what it does? We notice that the command expresses someone’s wish that the door be shut. Having noticed this, it is tempting to suppose that the content of the utterance is explained by the content of the wish that it expresses. On this view, what explains the fact that “shut the door!” means what it does is the prior fact that the speaker wishes that the door be shut.
This fact about the psychology of the speaker alone cannot do the explanatory work, since the speaker could have that wish without having even the means to express the wish in that work – the speaker might not speak English, for example. So we have to postulate a further fact about the speaker’s psychology: that something about that psychology maps the content of the wish onto the words uttered.
This is the move to which Wittgenstein objects. His objection is not to the appeal to psychology as such. Nothing in this passage suggests that there is anything illegitimate about positing facts about psychology or about the speaker’s mental life. He objects, rather, to the idea that mere appeal to psychology helps in the particular task of explanation at hand.
The account supposes that an explanatorily relevant mapping between content and words takes place in the speaker’s psychology. It assumes that merely attributing the mapping to the psychological domain will provide an explanation where one was lacking before. The mapping between content and words was what was to be explained in the first place. We have attempted to do our explanatory work simply by moving the mapping into the psychological sphere, without showing why a mapping in that sphere should be any less problematic than the mapping with which we started. The explanation is illusory.
The illusion consists in supposing that mental states are, like sentences, bearers of meaning, but unlike sentences are self-interpreting. Thus it is supposed that thoughts, intentions and mental images are like self-interpreting signs. It is this idea that strikes Wittgenstein as “occult”.
Quine, too, frequently complains of uncritical appeals to a mentalistic domain of meanings, made when we think about language.4 His project can be understood as an attempt to construct an understanding of language without reference to such a domain.
The project is a difficult one because of the many ways in which the notion of meaning is presupposed in our ordinary thought and talk about language. Let us take, as an example, the notion of synonymy. (This notion is of crucial importance to Quine). Take two synonymous sentences: the French sentence “mes pantalons sont vertes” and the English sentence “my trousers are green”. What does this relation of synonymy consist in?
Clearly, we cannot explain the synonymy relation by appealing to the fact that the two sentences have the same meaning. For two sentences to be synonymous just is for them to be alike in meaning. Thus their having a meaning is not explanatorily prior to their being synonymous.
Here a temptation similar to that with which Wittgenstein was concerned arises. Given that the two sentences are synonymous, we may infer that they both express the same thought: that my trousers are green. So we may suppose that by referring to this thought, we can provide an explanation of the meaning of each sentence, and thus of their synonymy.
But, once again, this line of thought can provide only an illusion of explanation. In order for reference to thoughts to be explanatory we have again to postulate mappings, on the part of speakers of both the French and the English sentence, of content onto sentence. But, again, how one content is mapped on to another was what was to be explained. If we are dissatisfied with appealing to a mapping of sentence on to sentence directly, it has not been made clear why we should be any more satisfied with a mapping of the same thought on to each sentence individually.
Both philosophers, then, oppose a style of explanation which, by looking inward to a psychological domain of meaning and thought, yields a mere illusion of the understanding of language. They both demand something better. The direction in which they look for an alternative is also similar. Both look in the direction of what might vaguely be called external, public points of reference and modes of response: things with which the discursive subject interacts. Just what this means will turn out to be a very delicate matter – indeed, the crucial factor separating these philosophers – but in both cases the drive is towards the outer.
§ 2. The drive to the outer: Quine
To understand Quine’s drive to the outer, it helps to understand a similar drive on the part of the philosophical tradition which Quine both absorbed and criticised – that of radical empiricism. In particular, it is an empiricist account of the meanings of utterances in which we are interested.
A simplified version of the empiricist account might run as follows. The theorist takes certain utterances to be appropriate responses to certain environmental conditions.5 The empiricist move is to identify the meaning of the utterance outright with the environmental condition to which it is an appropriate response.
Any empiricist approach of language must proceed on the basis of the empirical evidence available to the linguist: the sounds made by speakers and the environmental conditions under which those sounds are made. The simple empiricist account just mentioned has the advantage of making the job of empiricist linguistics – matching utterance to environmental condition – identical with the job of giving a theory of meaning for the language.
If we assume that our theory of meaning must preserve what we take pre-theoretically to be the
meanings of our utterances, then the simple empirical approach will fail. Many utterances – for example, truisms – are appropriate under any empirical conditions, but they do not all mean the same thing. Thus the environment underdetermines their meaning.
It is, further, very plausible that the environment can overdetermine the meanings of utterances. Some utterances depend for their appropriateness on more than the environment: they depend, for example, on the values of the speaker. Thus, plausibly, whether my utterance to the effect that the result of the horse race was a good one depends not only on which horse won but also on which horse I wanted to win (perhaps, which horse I had betted on).
One possible response is to attempt to develop empiricism such that it provides an account of our pre-theoretical conception of meaning. The logical positivists attempted to do so by providing independent accounts of truisms (in terms of logical truths) and value judgements (in terms of the expression of emotions).
Another possible response is to maintain the empiricist method as it is, and replace our pre-theoretical conception of meaning for another conception for which empiricism does provide a good theory. This is Quine’s route.
It is often said that Quine rejects verificationist theories of meaning as the “second dogma” of empiricism. This is true, but easily misconstrued if one supposes that Quine replaces the verificatonist theory of meaning with some other theory intended to carry out the same task. Instead of adopting some other theory meant to account for the pre-theoretical notion of meaning, Quine retains verificationism as the right methodology both for philosophy and for linguistics, and rejects the notion of meaning.
He adopts verificationism as a theory of another notion – stimulus meaning – which unlike our ordinary notion of meaning is, he considers, scientifically respectable, precisely because an empiricist theory of it is possible. Thus, the notion of stimulus meaning is not meant to approximate the ordinary notion of meaning. It only does so for a limited class of expressions (1960: 36-7). For many others, there is no approximation. Indeed, part of Quine’s point is that the scientifically respectable notion of stimulus meaning falls well short of the pre-theoretical notion of meaning. The gap between the two notions is the measure of the poverty of the stimulus.
The stimulus meaning of an expression is a class of stimulations: the class consisting of those stimulations upon which assent to the expression would be elicited. In order to understand the notion of stimulus meaning there are thus two further Quinean notions to be understood: that of a stimulation and that of assent.
What, precisely, is a stimulation? Tentatively, Quine identifies it (in the visual case) with a “pattern of chromatic irradiation of the eye” (1960: 31). There are two features of Quine’s choice which are of particular relevance to us. First, stimulations are to be described in physical terms. The notion of chromatic irradiation is drawn from physical theory. It is based on what we know about the physiology of vision and its physical basis. Second, the physically described features of the world chosen by Quine are located at the surface of the subject’s body. They are proximal rather than distal.
Since they are to be described in physical terms, Quinean stimulations are not private mental entities. It can be ascertained by a third party just what stimulations a subject is undergoing. At least in principle, one could ascertain this by examining the retina.
More importantly, one can without expertise in physiology or the use of specialised technology empathise with the point of view of others by the use of what one might call the geometric imagination. The difference in retinal pattern from subject to subject is a matter of angle and distance from the objects of vision. Thus I can imagine how things look to you by allowing for your location and point of view.
We might identify Quinean stimuli with perspectives on the world, provided we remember that these perspectives are not mental entities. This is why Quinean stimuli are still external. In the relevant sense of “perspective” my perspective would be the same as yours if we were to stand in the same place.
The choice of proximal over distal stimulations is also crucial. Retinal patterns vary with the disposition of the subject as well as with the subject’s physical environment. Because they are proximal, Quinean stimulations can vary independently of their distal cause. A physical object in the dark or at a distance creates a different retinal pattern from the same object in better visual conditions.
Further, distal objects can differ independently of stimulations. Different objects can create the same retinal pattern. This occurs whenever two different objects look alike from a given location under given conditions. These variations can, in principle, be predicted and taken account of by third parties.
Quine counts these features of retinal patterns as reaons for choosing them over physical objects as candidates for the role of stimulation (1960: 31). The reason is that a closer correlation can be made between utterances, on the one hand, and stimulations, on the other, if the stimulations in question are proximal. If I know not only what sort of physical environment a speaker inhabits, but also how that environment affects the speaker’s sensory organs, I will be in a better position to predict the speaker’s utterances.
What of the notion of assent? For Quine, empiricism demands that assent must be understood in purely behavioural terms, on a narrow conception of behaviour (1960: 57-8). Assent must consist in the making of some sound, audibly distinguishable by a third party, or in some bodily movement.
This notion of assent is far removed from our ordinary notion. To assent to something is, roughly, to agree to it. Such agreement can be signalled by sound or gesture, but does not in general consist in either. Once again, this gap between the ordinary concept and the Quinean concept is deliberate. Quinean assent is the form of assent that can be appealed to in a theory of language, under empiricist principles.
To see why, suppose that we were to liberalise this conception of assent, and say that a subject assents to an utterance whenever he or she expresses agreement with it. This assent can then take the form of expressing what the utterance expresses. We are now appealing to our prior knowledge of the equivalence of the contents of utterances – that is, of their synonymy. But it is just this relation of synonymy which was to be explained.
Let us suppose, optimistically, that for a given speaker a good theory of stimulus meaning can be formulated. That is, we can predict with some accuracy what utterances a speaker will assent to given certain stimulations.
Quine’s thesis of the indeterminacy of translation is then as follows. Such speaker-relative theories of stimulus meaning will not be sufficient to determine a single scheme of translation between the utterances of different speakers. But the notion of stimulus meaning is that on which, on empiricist grounds, a theory of language must be based. Since the notion of stimulus meaning is not sufficient to distinguish between systematically different schemes of translation nothing empirical can distinguish between them. The famous thesis of the indeterminacy of translation is nothing more than empiricism plus a recognition of the poverty of the stimulus.
§ 3. The drive to the outer: Wittgenstein
Wittgenstein’s drive to the outer is encapsulated, above all, in his notion of use. According to the Blue Book, the “life of a sign”, what distinguishes a meaningful utterance from a mere sound, is best thought of as the use of a sign, and not as an occult accompaniment of it (1958: 4-5). The same idea is reflected in the slogan repeated throughout the Philosophical Investigations, to the effect that for many expressions and in many contexts, the meaning of an expression can usefully be identified outright with its use.
But the notion of use is itself in need of clarification. This is so not so much because of any ambiguity in the word “use”. Rather, it is because there are many ways to describe the use of a given expression. The content of the dictum that “meaning is use” will depend on what vocabulary is available to describe how an expression is used.
An economic analogy may clarify matters here. Suppose I claim that one five pound note, on the one hand, and five one pound coins, on the other, are of equivalent value. Wherein does this equivalence consist? It does not, of course, consist in the value of the materials out of which the note and coins have been made. We are within reach of a better explanation if we say that they are of equivalent value insofar as they can be used for the same purposes.
But this will only be true if we restrict the range of uses which count as relevant in this context. It is relevant to point out that they can be used in exchange for the same goods and services. But other facts about possible uses are not relevant. The note can be used as a bookmark, for example, whereas the coins cannot. Equally, the coins can be used to stop the needle of a record player from jumping, by being laid on top of it, whereas the note cannot. Such considerations should not lead us into supposing that the note and coins differ in value.
Analogously, suppose that we ask of two expressions whether they have the same meaning. Wittgenstein’s slogan suggests a criterion for answering this question. The two expressions have the same meaning if and only if they have the same use. In the Philosophical Investigations, for example, Wittgenstein claims that the sentences “Bring me a slab” and its elliptical form “Slab!” have the same meaning because a builder can use either for the purposes of acquiring a slab from his assistant, and that the latter counts as elliptical merely in virtue of being shorter (1953: § 20).
But such a criterion will only give us a general theory of synonymy if we know, in general, what it is for two expression to have the same use. And there will be a clear answer to that question only if our notion of what counts as a use is narrow enough.6
We could, for example, define the use of an expression as follows. The use of an expression is to elicit a signal of assent – a sound or bodily gesture – under certain environmental conditions. So two expressions are equivalent in meaning if they elicit assent under the same conditions. On this highly specialised notion of “use”, Wittgenstein’s criterion would then simply reduce to Quine’s.
Though it is highly specialised, this is a perfectly legitimate notion of “use”. However, it is obviously not the only one that is explicable to expressions. Wittgenstein, indeed, places great emphasis on the enormous variety of uses to which expressions of a language can be put, and gives a long list of such possible uses in the Philosophical Investigations (1953: § 23). Some of the items on this list are:
Giving orders, and obeying them.
Describing the appearance of an object, or giving its measurements. …
Reporting an event.
Speculating about an event. …
Making a joke; telling it. …
Translating from one language into another.
Asking, thanking, cursing, greeting, praying.
Each of these activities, and many others, are called by Wittgenstein “language games”. It bears emphasising that these games are meant as the explanans and not as the explanandum of the meaningfulness of utterances. When we try to understand what it is for two expressions to be synonymous we appeal to their use, where to use an expression is to be involved in activities of the sort mentioned in the list.
It is clear that, if these activities are to be understood as types of language game, we must conceive of a language game as something more than a game the rules of which enjoin the production of sounds under defined environmental conditions, certainly if those conditions are understood in the narrowly visual way which Quine proposes.
The item which comes closest to the purely visual case is “describing the appearance of an object”. But even in this case, there is no indication that a description of the appearance of an object is to be construed as a response to a visual stimulus in Quine’s sense. To describe an appearance can be to respond to an object – a three-dimensional material thing – and not to respond to a chromatic irradiation, or any other sort of physiological or mental event in the perceiving subject. That Wittgenstein immediately afterwards mentions the giving of measurements suggests that what he has in mind is indeed a response to an object and not to anything like a sensory stimulation. Even in this case, then, the Wittgensteinian stimulus is more enriched than the Quinean.
The other examples show an even richer conception of the environment in terms of which utterances are to be understood. To understand an utterance as an instance of thanking, for example, one must already understand the speaker’s environment as one in which there are other agents who are candidates for gratitude.
Again, if praying is a type of language game, then to know what someone is doing when they utter words one must, sometimes, know what it is to pray, and in order to know what it is to pray one must know what God is.
The point is one which operates in two distinct directions. Wittgenstein’s conception of the speaker’s environment is richer than Quine’s. And so is his conception of what speakers are doing when they use expressions. Thus both Wittgensteinian stimuli and Wittgensteinian behaviour are richer – require greater resources to describe – than their Quinean counterparts.
It is unexceptionable to call greeting, cursing, praying and so on forms of behaviour. But if we include these forms of behaviour in our descriptions of the uses to which speakers put expressions, we have flouted the empiricist restrictions observed by Quine. Wittgenstein differs from Quine, then, in the much richer conception of behaviour which he appeals to in the evaluation of the meanings of expressions.
Similar remarks apply to the description of the environmental conditions of utterances. For Wittgenstein, no restriction is placed on the way in which the environment is to be described. In particular, there is no attempt to describe the environment in exclusively visual terms, let alone in the rigorously perspectival terms of Quinean retinal patterns.
§ 4. Convergence
The discussion so far has emphasised a sharp divergence between Quine and Wittgenstein in their approaches to their common problem. They both advocate a drive to the outer in order to answer questions about meaning. But they have contrasting conceptions of what the outer consists in.
Despite this contrast, there is I believe a deeper convergence than we have seen so far. The convergence I have in mind manifests itself in a similar attitude to ontological questions. By ontological questions, however, I have in mind nothing abstruse. I mean, simply, questions about what things there are. By this criterion, if one were to ask how many chairs there are in the world, one would be asking an ontological question. With a little looseness, one could even admit into this category questions about how many chairs there are in a given room.
In order to understand the convergence, I wish to turn first to some remarks Wittgenstein makes at Philosophical Investigations § 80. I will elaborate the scenario that is imagined in that section in a way that suits our purposes. Wittgenstein imagines finding himself in a room furnished with what is apparently a perfectly ordinary chair. It looks ordinary, and feels ordinary to the touch; sitting in it is like sitting in any other chair.
But the chair disappears before his eyes. He pinches himself to make sure he is awake. Perhaps the chair has been spirited away by some trick, but no way in which this could have been done can be found. Perhaps he was hallucinating the chair, but sustained, vivid multi-modal hallucinations of this sort do not occur. Perhaps he is hallucinating its absence now, but no explanation or evidence of such can be found. What is more, he is assured by his friends that in this country it is the way of chairs to appear and disappear.
Was there a chair in the room? It may be supposed that there was, for after all Wittgenstein sat in one. It may be supposed that there was not, for chairs are not the right sort of thing to disappear into thin air (as rainbows are, say) so it must not have been a chair that Wittgenstein sat in. It may be that there are further things we can find out about the situation which will provide an answer to the question. But this may not be the case: even if the situation is fully described, it may not determine either a “yes” or “no”.
Wittgenstein’s thought is that, in the latter circumstance, neither English nor German has sufficient resources to determine an answer to the question “was there a chair in the room?” No doubt this is because those languages are spoken in a world in which such things, as a matter of fact, do not occur. Were our world different enough, then no doubt we would speak a language in which we were not puzzled as to what to say, given the circumstances.
Now suppose that there is such a language, L. It is enough like English that speakers are prepared to say the words “there is a chair in the room”, but also enough unlike English that they are non-plussed by the chair’s disappearance. It does not tempt them to withdraw their claim.
Is what they say true? No doubt it is, provided that the description given by Wittgenstein at § 80 holds of their environment. The description, then, is in a sense all that need be said about chair-ontology, from the point of view of speakers of L. Provided that L-speakers can read English, they will be able to read from Wittgenstein’s description whether or not they would be prepared to assent to the sentence of L.
But, someone might insist, an ontological question still remains. Is there a chair there? As English-speakers, confronted with this question, how are we to proceed? It seems that are two courses of action which are open to us. We can recount the story of § 80, and allow ourselves to be satisfied with it. We could, alternatively, investigate the conditions under which chairs appear and disappear, as we never previously knew they could, thus learning something new about the world. But neither course of action, in itself, determines a “yes” or a “no” answer to the question whether there is a chair in the room.
The difficulty is that (i) our English sentence ‘there is a chair’ is not translatable into any sentence of L; and (ii) is not translatable into any sentence of the theory we develop to explain the appearance and disappearance of the chair. No doubt, were the scenario envisaged in § 80 to become common, we would begin talking in ways that accommodated the facts. The language we speak would change so that our question did after all admit of a definitive answer.
But Wittgenstein’s point is that even in such a reformed language, “the application of a word is not everywhere bounded by rules” (1953: § 84). The reform would handle the situation we are discussing, but not every possible situation.
Wittgenstein’s thought is this: we have not uncovered a flaw in English when we discover that our knowledge of the language determines no answer to ontological questions under the conditions described in § 80. In this sense, ontology is language-relative. A determinate answer is available to the question how many chairs are there in this room, but it need not be. If there were not, there would be no ontological task left to us except to understand why there were not.
My contention is that Quine’s attitude to ontological questions mirrors Wittgenstein’s. Quine advocates approaching ontological questions through a notion of ontological commitment.7 What is in question is the commitments of a particular discourse: what a given system of sentences held true says there is in the world.
According to Quine, we can deduce the ontological commitments of a discourse or theory by expressing it in a way that conforms to the requirements of first order quantificational logic. The discourse is commited to the existence of those things that the variables range over. The appeal to the logic of quantifiers is thus meant as a criterion for ontological commitment, to be applied to the utterances of speakers.
Ontological commitment is, on this understanding, language-relative. What a speaker is commited to will depend on what language is being spoken. Quine has not shown us how to go about the business of arriving at an ontology – an account of what there is in the world – until he has told us what language we ought to carry out such an investigation in. Is there, for Quine, such a privileged language?
It may seem that Quine believes there is, so that he is unable to follow Wittgenstein into a deflation of ontological questions. What disbars him from doing so, according to the objection I have in mind, is his physicalism. According to this doctrine, what exists is given by a physical description of the world, or at least will be given by a description of the world in the terms of a completed physics.
If this were so, then for any ontological claim, in any language, we could say that the claim is true if and only if it corresponds, when translated into a purely physical language, into part of a purely physical description of the world.
The essential appeal this doctine makes to the notion of translation should be enough to alert us to the fact that this cannot be Quine’s view. According to Quine, any translation of a sentence of one language into a sentence of another is at best correct relative to some translation manual. Thus physicalism could only provide a criterion for settling ontological questions if it determined a translation manual: the mere provision of a physical description of the world would not be sufficient. But Quine’s verificationism precisely rules out the possibility of a single translation manual.
Indeed, Quine’s reasons for accepting physicalism are pragmatic in nature, in the sense that he does not recommend physical descriptions of the world on the grounds that they are more correct than other sorts of description. Physics is preferred, among the other ways of talking about the world, for its high degree of predictive power and sensitivity to subtle changes in the way the world is (see Quine 1978). The reason for these advantages is, ultimately, that sentences about physical objects are “fairly directly associated with sensory stimulation” (Quine 1960: 237).
Quine’s contention then, is that physics is the best means available to us of describing the world. For the sake of argument, let us accept Quine’s contention. It does not follow that claims made within other discourses should only be accepted insofar as they are translatable into a true claim of physical discourse. The indeterminacy of translation rules out this move. In particular, then, it is not the case that the question “is there a chair in the room?” can be settled by translation into the terms of physics.
There is however a tension in Quine’s thought. Certain aspects of his approach militate against ontological deflation. As we have seen, Quinean linguistics is committed to a privileged vocabulary in which to describe environmental stimuli. Such stimuli are to be described in physical terms: in the visual case, in terms of the “chromatic irradiation” of the speaker’s eyes. Similarly, the speakers’ responses to that environment, whether spoken or gestured, are to be described in physiological terms, as bodily movement and produced sound.
A form of physicalism is thus presupposed by Quine’s way of conceiving the empirical constraints on our understanding of language. Since no such constraint is presupposed by Wittgenstein’s drive to the outer, his deflation of ontology is more thoroughgoing than Quine’s.
Arrington, Robert and Glock, Hans-Johann (eds.) (1996), Wittgenstein and Quine. London: Routledge.
Goldfarb, Warren (1983), “I want you to bring me a slab” Synthese 56: 265-282.
Heal, Jane (1989), Fact and Meaning: Quine and Wittgenstein on Philosophy of Language. Oxford: Blackwell.
Hintikka, Jaako (1990), “Quine as a member of the tradition of the universality of language”, in Robert B. Barrett and Roger F. Gibson (eds.) Perspectives on Quine. Oxford: Blackwell.
Kripke, Saul (1982), Wittgenstein on Rules and Private Language. Oxford: Blackwell.
Quine, W.V. (1953a), “On what there is”, in From a Logical Point of View. Cambridge, Mass.: Harvard University Press.
(1953b), “Two dogmas of empiricism”, in From a Logical Point of View. Cambridge, Mass.: Harvard University Press.
(1953c), “The Problem of meaning in linguistics”, in From a Logical Point of View. Cambridge, Mass.: Harvard University Press.
(1960), Word and Object, Cambridge, Mass.: MIT Press.
(1970), Philosophy of Logic, Englewood Cliffs, NJ: Prentice-Hall.
(1978), “Otherworldly”, New York Review of Books. November 23rd, p. 25.
Wittgenstein, Ludwig (1958), The Blue and Brown Books. Oxford: Blackwell.
(1953), Philosophical Investigations. Trans. G.E.M. Anscombe. Oxford:
1For scepticism about the possibility of a comparison such as I draw in this paper, see Kripke (1982: 5-7) and Hintikka (1990: 167-9).
2Quine remarks (1960: 76-7) on the connection between Wittgenstein’s dictum that “understanding a sentence means understanding a language” and his own theory of the indeterminacy of translation, making the intriguing but unspecified suggestion that the latter “will have little air of paradox” for those familiar with Wittgenstein’s later philosophy of language. He also connects (p. 260) his own and Wittgenstein’s metaphilosophical views. Wittgenstein never mentions Quine in print; neither, so far as I have been able to discover from published sources, did he ever mention him in conversation. For comparisons of the two thinkers from a largely Wittgensteinian perspective see the papers in Arrington and Glock 1996. The feeling alluded to in the text is well captured by the title of Peter Hacker’s contribution to the volume: “Proximity at great distance”. See also Heal 1989.
3Our focus here will be on the Wittgenstein of the Blue Book, composed in 1933-4 and part one of the Philosophical Investigations, complete by 1945.
4The criticism of such appeals is a theme of Quine’s “Two Dogmas of Empiricism”. But the felt need to carry on both philosophy and linguistics without unexplained appeal to meanings motivates much of Quine’s work: see for example Quine 1953c: 48, where the danger of illusions of explanation is mentioned explicitly, and 1970: 1 and throughout.
5Carnap, Quine’s major influence in the empiricist tradition, took appropriateness to consist in truth. But in putting things in this general way, I abstract from different possible ways of understanding appropriateness. It is the general point in which we are particularly interested.
6On this point I am indebted to Goldfarb 1983.
7See Quine 1953a; Quine 1960: 238ff and – especially – 243n.