Philosophical Connections: Quine

Willard Van Orman Quine was born in Akron, Ohio. He studied mathematics and philosophy at Oberlin College and gained his doctorate at Harvard with a thesis on 'The Logic of Sequences', one of his supervisors being Whitehead. He spent some time studying mathematical logic in Vienna, Prague (where he met Carnap), and Warsaw before returning to teach at Harvard in 1934. He was appointed full professor and Senior Fellow there in 1948. He was Visiting Professor at Oxford in 1953-4 and Wolfson Lecturer in 1973-4.

PHILOSOPHY OF LANGUAGE AND LOGIC

[1] On matters of existence and reality Quine argues that only those entities should be admitted for which strong and appropriate criteria can be given [see Ontology — sec. 2]. In the case of theoretical terms of language and logic he tends therefore to try to explain intensional ideas (sense, meaning, propositions, identity, synonymy, analyticity, and the like) in terms of extensional ones (for example, reference and truth) [a]. Indeed his tendency is to dismiss such intensional notions altogether in so far as he considers any importation of them into our theories as leading to circularity and obscurity. It is in the light of this attitude that we can understand (1) his treatment of the analytic-synthetic distinction, and (2) what he calls 'regimentation' of our language.

The analytic-synthetic distinction. Quine calls this "the first dogma of empiricism" [see 'Two dogmas of empiricism']. Traditionally (at least since Frege) a distinction has been made between logically necessary and logically contingent statements. The former are ('broadly') analytic, the latter synthetic. But this class of 'broadly' analytic statements can be subdivided into (a) logical truths and (b) 'narrowly' analytic statements — these latter being always true by virtue of substitution on the basis of definitions, which in turn depend on synonymy. Quine rejects this notion of narrow analyticity [b]. Consider these examples: (a) No unmarried man is married; (b) No bachelor is married. The first is a logical truth and remains so whatever we may substitute for 'married' — the logical particles ('no', 'is') being unchanged. The second is not a logical truth but can be made into one by the substitution of a synonymous term (say, 'unmarried man') for bachelor'. But then how are we to understand the intensional concept of synonymy? We cannot appeal to definitions, for (with the exception of stipulative and nominal definitions) definition presupposes synonymy. Nor can we appeal to the idea of inter-substitution without a change of truth-value in all contexts, because to say the truth of 'It is necessary that all bachelors are bachelors' is unchanged by the substitution of 'unmarried man' is to reason in a circle; 'necessary' here means 'analytic'. Indeed, underlying Quine's arguments here is a wider consideration: his dislike of what he calls the 'referential opacity' of modal contexts and quantification into opaque contexts (examples of modal terms are 'necessary' and 'possible'). This can be illustrated by his argument in 'Reference and Modality'. The substitution of the number of planets, (in the true statement 'The number of planets = 9') for '9' in the true statement '9 is necessarily greater than 7' results in a false statement — 'The number of planets is necessarily greater than 7'. Likewise the formal rendering of '9 is necessarily greater than 7' as '(∃x)(x is necessarily greater than 7)' makes no sense unless we suppose objects to possess properties essentially (de re necessity as against de dicto necessity) — a position the nominalist Quine finds unacceptable. Indeed, he regarded modal logic in general as being pseudo-logic [c].

He does nevertheless allow a minimal sense to synonymy and analyticity as defined in terms of observable sensory stimuli (behaviour and dispositions). [See Word and Object.] Inter-subjective stimulus-synonymy is sameness of stimulus meaning for a given speaker; while a sentence is stimulus-analytic if he would assent to it after every stimulation [d]. A socially stimulus-analytic sentence is then one which is stimulus-analytic for nearly every speaker of the language, if not all. That there are these minimal allowable senses of the terms is assumed in Quine's discussion of radical translation. By this he means the translation of the language of a previously unknown and isolated tribe. This would have to start, he says, by considering sentences (rather than individual words) which relate directly to stimulus conditions. To explain this he distinguishes between occasion sentences ('It hurts', 'His face is dirty') and standing sentences (for example, 'Each year crocuses come out'). To assert the former we require stimuli each time we observe the relevant subject. But in the case of the latter, once we have given our assent (having looked at the crocuses) we do not need to observe again. Now suppose we go to this community whose language we do not understand and observe the circumstances in which a native says Gavagai in the presence of a rabbit. To know whether he means rabbit, rabbithood, a particular segment of the rabbit, and so on, we need to frame 'analytical hypotheses'. But because there can be more than one set of such hypotheses that fit the native's speech dispositions (that is, to respond to the relevant stimuli) have no way of determining which is the right one. This is consistent with his general thesis that reference is indeterminate in the sense that the facts of our linguistic behaviour can be mapped in a variety of ways. (And this is one reason why senses cannot be admitted; for they fix references without regard to circumstances — contrary to the flexibility of our linguistic behaviour) [e]. Thus Quine posits his indeterminacy of translation thesis, from which he draws the conclusions (1) that any statement can be regarded as true — independently of empirical observation — provided appropriate adjustments are made elsewhere in the conceptual scheme to which the statement belongs; and (2) that therefore the traditional distinction between analytic and synthetic statements cannot be sustained [f]. Conversely, for Quine, it is also the case that identity statements, for example, 'Hesperus is Phosphorus' or 'Cicero is Tully', while true are not necessary truths, on the grounds that the discoveries of these identities are empirical and that therefore the statements are contingent [g].

Regimentation of language ['On what there is']. If identity conditions stated in intensional terms are unclear or inadequate, as Quine claims, then how are we to deal with abstract entities? His solution is to eliminate them from our discourse. He not only replaces all proper names by definite descriptions, but also goes on to eliminate all other singular terms (pronouns, demonstratives, and the like), replacing them by logical devices such as quantifiers, variables, and identity [h]. Thus 'Socrates' in 'Socrates was Greek' becomes 'The x which socratizes'; and 'Socrates was Greek' turns into 'There is just one x which socratizes, and whatever socratizes was Greek'. In logical symbolism: G(ιx)Fx) = df. (∃x)(Fx &(y)(Fy → y = x) &Gx) (where 'F' stands for 'socratizes') and 'G' stands for 'Greek'. This regimentation and reconstruction of natural language from within (essentially to 'improve' it), through the application of the techniques of formal logic [i], has important implications for Quine's ontology and his theory of knowledge.

METAPHYSICS/ONTOLOGY

[2] According to Quine, "to be is to be the value of a variable" ['On what there is']. By this he meant that as a consequence of regimentation and the eliminability of abstract entities the question of what actually exists in the world becomes the question what satisfies the quantified variables. It is these, he says, which carry 'ontic commitment'. Furthermore we can choose how far to go with elimination according to the things that need to be done in philosophy. In his early work his philosophy tended to be characterized by a preference for concrete, physical objects over 'abstract' entities as 'unactualized possibilities'. We can do without abstract objects if their jobs can be performed without obscurity by, say, sense-data, physical objects, and classes. (Likewise to explain human action we can, if we wish, eliminate references to 'purposes' or 'intentions'; while the resolution of conditional statements into statements about dispositions also enables us to do without modal operators. Dispositional statements are then to be understood in terms of statements about the physical objects to which the dispositions are attributable.) Nevertheless he finds room for mathematical 'entities'; and [from 'Two dogmas of empiricism' onwards] he shows increasing tolerance of abstract entities in general — if they can be seen to have a useful role to play [a]. The ideas required by science, for example, are regarded as real as any physical entity. And he eventually came to reject his earlier acceptance of phenomenalistic conceptual schemes.

Regimentation, together with Quine's views on the indeterminacy of translation, leads further to his concept of ontological relativity. If there are different sets of analytical hypotheses to account for the matching of any one language with another, it must follow that there can be different theories and conceptual schemes consistent with the empirical evidence in a given context [b]. We cannot therefore say that the 'real' world will be represented correctly by one particular conceptual scheme. Indeed, to compare different schemes presupposes a theoretical background which itself may be but one of many, alternatives.

Quine's ontology is thus hierarchical and allows what he calls 'semantic ascent'. By this means the move away from talk about putative objects or ideas to talk about the very language we use for the purpose. As philosophers we may well start out front everyday or 'ordinary' language, but we can revise and restructure it, dispensing with certain concepts to suit our explanatory needs; and we can thereby move 'up' through the hierarchy of languages [c] from the 'everyday' way of looking at things, through natural sciences, to mathematics and logic, and finally ontology — depending on how much generality and abstraction we are prepared to allow.

THEORY OF KNOWLEDGE

[3] A second 'dogma' of empiricism rejected by Quine is the reductionist view espoused by some logical positivists that meaningful statements can be translated into logical constructs out of terms which refer to immediate sensory experience, and which are taken to constitute verification of synthetic statements. Indeed the second dogma, he thinks, is basically identical to the first. The idea that any given true statement in isolation consists of a linguistic component and an extra-linguistic factual component is nonsense. (Analytic statements are limiting cases which have no factual component and are confirmed vacuously.) Our statements about the world can be confirmed or invalidated only 'holistically, as a "corporate body" [a]. As he says,

For Quine epistemology is to be naturalized, as a consequence of which many traditional philosophical issues may appeal to scientific facts and empirical explanations. There is thus no clear distinction between knowledge, science, and philosophy. At all levels of the semantic ascent what we can be said to know must depend on what entities we have chosen to admit into our conceptual schemes. Philosophy is simply more general [b]. Nevertheless, although he thinks of physical objects not as defined in terms of experience but as "irreducible posits" comparable epistemologically to the gods of Homer, he remains essentially an empirical physicalist. He regards the conceptual scheme of 'science' (in its widest sense) as a tool for making predictions on the basis of past experience, and the postulation of physical objects rather sense-data as a basis for successfully achieving a coherent and systematic structure out of our experience. There is thus a strongly pragmatic flavour about his epistemology.

CRITICAL SUMMARY

Quine was probably one of the most important American philosophers of the twentieth century. Initially influenced by Carnap's logical positivism he worked out over many years a comprehensive physicalist and pragmatic philosophy which combines vision with great analytical power. Its most notable features include a sustained attempt to eliminate intensional terms in favour of extensional terms. This can be seen in particular in his rejection of the analytic-synthetic distinction. His dislike of 'referential opacity' led him also to attack efforts to formalize modal terms, which he saw as requiring acceptance of de re essentialism. Quine made radical use of Russell's logical techniques to achieve 'regimentation' of ordinary language; and he regarded quantified variables as carrying ontological commitment'. In his later work he advocated an appeal to socially observable stimuli as the basis of his physicalism and account of meaning. And in his holistic ontology, implicit in a modifiable hierarchical 'semantic ascent', he made no clear distinction between science and philosophy.

As might be expected, criticism of Quine's work comes primarily from 'linguistic' philosophers (whether of the 'informal' or 'systematic' varieties). Formal objections have been made of the viability of claims to eliminate linguistic terms and of his alleged obscuring of the distinction between formal and informal logic. Many critics would say that Quine's attack on the analytic-synthetic distinction, powerful though it may be, has failed; and that intensional terms have a legitimate function in everyday discourse; no circularity is involved if they are accepted as interdependent. It is said also that he exaggerates the difficulties of intertranslatability; sufficient clarity can be achieved in normal discourse and ambiguities avoided. Indeed it is questionable whether Quine's concerns about this issue are consistent with his appeal to, in effect, a common pool of socially observable stimuli and with his attribution to language of an informational role. Another serious objection is perhaps that it is arbitrary and presumptious to suppose that ontology is governable by what one chooses to eliminate by logical techniques. Moreover, Quine himself seems to equivocate about his belief in the existence of abstract entities. Do they 'really' exist? Or is it just a matter of ontic commitment — determined by what is 'useful'?

READING

Quine: 'On what there is' (1948); 'Two dogmas of empiricism' (1950); 'Reference and modality' (1953 — originating from earlier essays) [These, and other, essays are reprinted in From a Logical Point of View (1953)]; Word and Object (1960); The Roots of Reference (1973).

Studies

G. Kemp, Quine: A Guide for the Perplexed.

P. Gochet, Ascent to Truth: A Critical Examination of Quine's Philosophy.

C. Hookway, Quine.

A. Orenstein, W. Quine

Collections of Essays

D. Davidson and J. Hintikka (eds), Words and Objections: Essays on the Work of W. V. Quine.

R. Gibson (ed.), The Cambridge Companion to Quine.

P. A. Schilpp (ed.), The Philosophy of W. V. Quine.

CONNECTIONS

Quine

Note: the general influence on Rorty of Quine's attack on traditional empiricism and his scepticism about the fact-language distinction.

Philosophical Connections

Compiled by Anthony Harrison-Barbet

QUINE

PHILOSOPHY OF LANGUAGE AND LOGIC

METAPHYSICS/ONTOLOGY

THEORY OF KNOWLEDGE

CRITICAL SUMMARY

READING

CONNECTIONS

Quine