METAPHILOSOPHY Vol. 15, No. 2, April 1984 0026-1068 $2.00

REVIEWS

GEORGE D. ROMANOS, @he and analytic philosophy: the language of language, MIT Press, 1983, xviii-227 pp.

I

In aoduction Harvard’s (Professor Emeritus) Willard Van Orman Quine (b. 1908) is

widely regarded as one of the most important and influential philosophers in the Western world today. Over his long career, Quine has produced some fifteen books which have been translated into a dozen languages. His impact on contemporary philosophy has been phenomenal. As a result, countless journal articles dealing with Quine’s philosophy have been generated. Recently, however, Quinian scholarship has resorted to the medium of monographs in an attempt to interpret and assess Quine’s philosophical and historical significance. Professor Romanos’s mine and Analytic Philosophy is the most recent entrant into the field of six such monographs: three in English, one in French, one in German, and one in Italian. No doubt, there will be others to come; one suspects that the author of the last word on Quine has yet to be born.

The central thesis of Romanos’s book can be put very succinctly: the linguistic turn - the twentieth-century development of analytic phlosophy as a reaction to the inadequacies of classical metaphysics - has assumed the truth of the analytic conception of language - roughly, the view that the structure, meaning, reference, and content (i.e., ontology) of the elements of languages are absolutely fixed - and Quine’s relativistic thesis has shown this conception of language to be implausible and meaningless. The con- clusion that Romanos draws from this observation “is that the philosophical method of analysis is at root no less speculative - and with that, no more plausible or coherent - than the more traditional brand of metaphysical inquiry it was originally intended to supplement or replace” (xvii). Thus, according to Romanos, one may read his book (or Quine’s philosophy) “as either an attempted refutation of the analytic conception of language or an indirect proof of the sound footing of metaphysics” (xvi). Romanos intends the ‘or’ here to be construed non-exclusively, for he regards these two readings to be consistent with each other. In this view, I think he is mistaken. I shall argue that he is correct in saying that Quine’s relativistic thesis puts the lie to the analytic conception of language, but he is mistaken to regard this development as an indirect proof in favor of metaphysics. In short, 1 shall argue that Romanos has misunderstood the import of hs own analysis primarily because he has failed to appreciate sufficiently the importance of the roles played in Quine’s philosophy by naturalism and physicalism. The

141

142 REVIEWS

proper conclusion to draw from Romanos’s book (and from Quine’s philos- ophy) is not that metaphysics is on a sound footing but that the analytic conception of language ought to be rejected along with classical meta- physics.

m e Analytic Conception of Language According to Romanos, the linguistic turn was accomplished by two

distinct analytic factions: analytic metaphysicians (e.g., Russell, Bergmann, Austin, Ryle, Hare) and linguistic Kantians (e.g., Wittgenstein, Carnap, Ayer, Strawson, Toulmin). Analytic metaphysicians share with their predecessors (i.e., classical metaphysicians) the conviction that philosophy ought to seek to discover “the first principles or truths concerning the ultimate nature of reality or the basic categories of existence” (xiv). However, “impressed with the seriousness of epistemological limitations on our ability to directly ‘grasp’ or apprehend such metaphysical truth” (xiv), the analytic meta- physicians parted company with classical metaphysicians by adopting “an indirect way of approaching these questions by looking closely at the way we do, or ought to, speak about the world. Here the basic structure or mean- ing of language is viewed as presenting a picture of the metaphysical facts, or at least providing suggestive hints or clues to the ultimate truth” (xiv). Thus, like classical metaphysicians, analytic metaphysicians “are ultimately interested in the fundamental nature of extralinguistic facts and our know- ledge of those facts. Their concern with language does not reflect a distaste for traditional philosophical problems, but only a belief that the most fruit- ful approach to solutions for such problems is through an investigation of language, because of the special role language is imagined to play in mediating between us and the world around us” (xiv-xv). The linguistic Kantians, on the other hand, represent a more radical response to classical metaphysics:

Here the traditional philosophical questions are rejected out of hand, or perhaps reinterpreted as merely linguistic, as questions concerning the grammar of a language, the reference of terms, the usefulness of a linguistic framework, and so forth. Talk about the world as it really is, apart from considerations of language and theoretical science, is viewed as unintelli- gible, incoherent, and meaningless. Historical philosophical controversies are held to arise either from confusion caused by the vagueness and ambiguity of ordinary language or, conversely, from philosophical misuse and abuse of ordinary language. (xv)

Despite their differences, both the analytic metaphysicians and the linguistic Kantians, according to Romanos, presuppose the analytic conception of language. However, Romanos chooses to focus his attention primarily on the position of the linguistic Kantians:

REVIEWS 143

Considered independently of metaphysics-ridden theories of an ideal language, [i.e., the theories of the analytic metaphysicians] essentially two basic versions of the analytic conception of language remain. corresponding roughly to the split between ordinary language philosophers and linguistic reformers. According to one version, usually identified with the proponents of linguistic reform and the construction of artificial languages, the basic meaning, structure, or conceptual content of a language is capable of full and explicit codification by means of a complete set of syntatical and seniantical rules. These rules are viewed entirely as the product of human artifice and convention, and they define just what basic features or con- ceptual categories the language in question applies to the world. According to the second version of the analytic conception of language, explicit codification of the meaning or content of a language is out of the question. The analyst must instead pay close attention t o the actual use or function- ing of linguistic expressions in everyday life, noting carefully the variety of different circumstances in which their use is appropriate and keeping track of the systematic relationships they exhibit with other expressions in the language. Both versions of the analytic conception of language are t o some extent distillations or idealizations of views actually held with greater or lesser admixtures of conscious or unconscious metaphysics. They are not altogether mutually exclusive, but may simply represent different areas of emphasis. (xvi-xvii)

In sum, analytic philosophy consists of both analytic metaphysicians (e.g., Russell, et al.) and linguistic Kantians, the latter group being subdivided into ordinary language philosophers (e.g., Wittgenstein) and linguistic reformers (e.g., Carnap). All of these philosophical approaches presuppose, however, the truth of the analytic conception of language - “the idea that the structure or meaning of language (individual expressions or linguistic systems as wholes) can be objectively examined and analyzed in some fundamental way in which extralinguistic reality, as such, cannot” (xv). It is this conception of language. then, which Quine’s relativistic thesis shows t o be implausible and meaning- less.

Quine’s Relativistic Thesis According t o Romanos, Quine’s relativistic thesis consists of his doctrines

of indeterminacy of translation, inscrutability of reference, and ontological relativity. The doctrine of indeterminacy of translation asserts that trans- lations of a foreign language can be set up in such ways that while each is consistent with the speech dispositions of everyone concerned they never- theless can have different sentence t o sentence correlations even t o the point where two translations of a foreign sentence can be correlated with sentences having opposite truth va1ues;and there is no answer t o the (pseudo-) question

144 REVIEWS

of which translation is the uniquely correct one - they are all correct insofar as they measure up to the speech dispositions of all concerned. The doctrine of inscrutibility of reference asserts that, consistent with the speech dis- positions of everyone concerned, the same expression (a foreigner’s, our neighbor’s, or our own) may be construed as a term or as a non-term (i.e., particle), and if as a term then as a singular term or as a general term desig- nating or denoting, respectively, any of a variety of possible objects; and there is no answer to the (pseudo-) question of which translation or inter- pretation is the uniquely correct one - they are all correct insofar as they measure up to the speech dispositions of all concerned. The doctrine of ontological relativity asserts that there is no saying absolutely what the objects of a theory are. One can specify the objects of a theory only relative to some background theory and only relative to some manual for translating (or proxy function for mapping) the objects of one theory into the other. A third aspect of relativity obtrudes in settling whether the object theory in question involves referential or substitutional quantification. Such a determin- ation is relative to both the background theory and to the interpretation (or choice of proxy function) imposed on the object theory from within the background theory. In short, Quine’s relativistic thesis claims that the struc- ture, meaning, reference and content (i.e., ontology) of a language can never be grasped in any absolute sense.

IV

Refuting the Analytic Conception of Language If Quine’s relativistic thesis is accepted, then it is “entirely meaningless to

ask in any absolute sense - that is, according to either version of the analytic conception of language - what we are really talking about, or what our words really say, mean, or refer to. This is the real thrust of Quine’s relativistic thesis” (94).

What has occurred is that a blow has been struck against philosophical analysis every bit as decisive and crippling as the positivists’ own earlier assault on traditional metaphysics. What has been shattered by Quine’s relativistic thesis is the elementary assumption of most analytic philos- ophers . . . that absolute (philosophical) talk about the structure, meaning, or content of linguistic expressions makes any objective Sense to begin with, regardless of what one supposes the precise nature and origin of these linguistic features to be. (96)

According to Romanos, this shows “that the philosophical method of analysis is at root no less speculative - and with that, no more plausible or coherent - than the more traditional brand of metaphysical inquiry it was originally intended to supplement or replace” (xvii); that “[a] bsolute semantic inquiry is, then, no better or worse off than speculative metaphysics” (105);

REVIEWS 145

that “[p]hilosophicd analysis is no worse off than metaphysics, but it is no better either” (187). But even if this has been established (which is doubtful - Carnap no less speculative than Bradley?), it still does not warrant Romanos’s conclusion that his book (or Quine’s philosophy) serves as an “indirect proof of the sound footing of metaphysics” (xvi). Two wrongs, after all, don’t make a right; “so-much-the-worse-for-analytic-philosophy” doesn’t entail “so-much- the-better-for-metaphysics”. To read Quine as endorsing metaphysics even in this backhanded way is to mis-read him, and Romanos’s tendency to do so is spawned, perhaps, from not taking seriously Quine’s alternative to both metaphysics and the analytic conception of language, namely, his naturalism ~ the very naturalism that gave birth to his relativistic thesis. Quine is not arguing for the reinstatement of first philosophy; he is arguing for its super- session by natural science. Thus, in the same Dewey Lecture that Romanos makes so much of, Quine can say: “With Dewey I hold that knowledge, mind, and meaning are part of the same world that they have to do with, and that they are to be studied in the same empirical spirit that animates natural science. There is no place for a prior philosophy.”’ This is a far cry from calling for the reinstatement of classical metaphysics to its former place of prominence. Thus, I think Romanos goes too far when he reads Quine’s attack on the analytic conception of language as an indirect proof for the soundness of metaphysics, and I believe he made this erroneous inference because he overlooked the central role that naturalism plays in Quine’s overall philosophy.

The absence of an adequate concern for Quine’s naturalism surfaces to obscure Quine’s position again when Romanos admonishes us to refrain from concluding that Quine’s relativistic thesis merely amounts to the “banal observation that we must assume a prior understanding of some terms in order to construct explicit definitions of others by means of them, and thus that our understanding of these basic terms cannot itself be ultimately based on simple verbal definition, on pain of infinite regress, or vicious circularity . . . ” (91). Quine’s relativistic thesis, we are warned, is more profound: “According to this thesis, interpretation of a theory in terms of a background theory never really succeeds in laying our ontological/conceptual cards on the table, simply because there are no such cards to show” (92). Unfortu- nately, however, Romanos never entirely succeeds in making clear why there are no cards to show - why there is no fact of the matter to absolute questions of meaning, reference, and ontology. Nor is the issue further clarified in saying that “ [n lo matter how earnestly and tirelessly we attempt to tie one theory translationally to another, we never reach a background theory whose ontology or conceptual apparatus is any more fixed and determinate than that of the theory translated into it, simply because any sense we are able to make of the notions of linguistic meaning or conceptual content derives from

W.V. Quine, “Ontological Relativity” in Ontological Relativity and Other Essays (New York: Columbia University Press, 1969), p. 26.

146 REVIEWS

the possibility of such translation itself rather than constituting the basis upon which such translation proceeds” (92). Banality, after all, is not diminished by obscurity. The point of the claim that there are no cards to be shown - no fact of the matter to meaning, reference, and ontology - evades Romanos again and again on pages 94, 95, and 96. The elusive point that needs making, in order to understand the import of Quine’s relativistic thesis, is that Quine’s argument against semantic absolutism develops from within an already accepted naturalistic, physicalistic position:

The intended notion of matter of fact is not transcendental or yet epis- temological, not even a question of evidence; it is ontological, a question of reality, and to be taken naturalistically within our scientific theory of the world. Thus suppose, to make things vivid, that we are settling still for a physics of elementary particles and recognizing a dozen or so basic states and relations in which they may stand. Then when I say there is no fact of the matter, as regards, say, the two rival manuals of translation, what I mean is that both manuals are compatible with all the same distri- butions of states and relations over elementary particles. In a word, they are physically equivalent.2

It is just this naturalistic-physicalistic basis of the relativistic thesis that is never clearly brought out in Romanos’s analysis of Quine’s philosophy.

For a book that is widely heralded by its publisher as the “first major exposition and defense of Quine’s systematic philosophy” one would reason- ably expect to find more exposition and defense of Quine’s views than is contained in this book. The first forty pages of the 195-page text never mention Quine at all. And the remainder of the book devotes only eight and one-third pages (at the beginning of Chapter 2) t o the exposition of Quine’s systematic philosophy. But even this “exposition” amounts to no more than an enigmatic synopsis of Quine’s 1968 Dewey Lecture, “Ontological Rela- tivity”. Suffice it to say, in this regard, that Quine has not been idle since 1968. Given the fact that Romanos devotes whole chapters of his book to explaining the views of Carnap and Tarski, I suspect that, ironically, the newcomer to analytic philosophy will come away from reading this book understanding the views of Carnap and Tarski far better than he does the views of Quine.

With respect to defending Quine, Romanos’s book fares a little better. He tries to dispel1 the apparent contradiction in Quine’s views regarding the theory of reference versus the theory of meaning; he attempts to answer Chomsky’s charge that indeterminacy of translation is reducible to the uncontroversiaJ thesis of underdetermination of physical theory; and he has some suggestive things to say in the final pages of the book that may be

W.V. Quine, “Things and Their Place in Theories” in Theories and Things (Cambridge, MA: The Belknap Press of Harvard University Press, 19811, p. 23.

REVIEWS 147

useful in countering Rorty’s interpretation of Quine’s contribution to the philosophical scene. Laudible as these attempts at defending Quine are, Romanos is remiss when it comes to the larger picture: Zf Quine’s relativistic thesis is true, then the analytic conception of language is false, and this surely would have dire consequences for those analytic philosophies that presuppose the analytic conception of language. But why should anyone accept Quine’s relativistic thesis? Unfortunately, Romanos gives no more than a hint of an argument (See pp. 44, 93) designed to convince the reader that Quine’s relativistic thesis is true. In short, we are never told why anyone ought to accept Quine’s doctrines of indeterminacy of translation, inscrutability of reference, and ontolgical relativity.

Perhaps the moral of the story is that authors ought not be held account- able for what their publishers say about their authors’ works.

ROGER GIBSON WESTERN I L LI NO1 S UNIVERSITY MACOMB, IL 61455 USA