CHAPTER ONE
From Empiricism to Realism
We may begin by comparing three different âstagingsâ of language. The first (in nonchronological order) is well known,1 Quineâs scene of radical translation: the linguist in the jungle studying an unknown language accompanied by a native who cries âgavagaiâ when a rabbit runs out in front of them. The second is the scene invented by Wittgenstein in §2 of the Philosophical Investigations, based on Augustineâs description in the Confessions of learning language, a description Wittgenstein cites at the beginning of the Investigations. There is a mason A and his assistant B, and their language consists entirely of the words block, pillar, slab, and beam. A pronounces, calls out these words, and B brings the object âwhich he has learnt to bring at that call.â The third scene, which will be addressed later, is from Austin, and with it begins How to Do Things with Words (there are, in fact, several scenes: the âI doâ pronounced in response to the priestâs question in a wedding; the âI name this ship the Queen Elizabethâ); this scene differs from the first two if only in that it takes place within a social and institutional context. Without having changed fundamentally, my interest has shifted from the first, Quinean scene of radical translation to the plurality of Austinian scenes by way of a specific interpretation of the first sceneâan interpretation authorized by the second, Wittgensteinâs critique. The three points of view presented here share in common a criticism, stated in different ways, of the myth of meaning. It remains to be seen what this expression means, and this issue has turned out to be more complex in Quine than one might have expected. Quineâs goal is to critique, and his thesis of indeterminacy is essentially negative. But it is still necessary to determine exactly what he criticizes and what he means by myth. This term, which occurs frequently in Quineâs work, points to what is at stake in his philosophy: the anthropological critique of the myth of meaning as the myth of a core common to different languages, as the reintroduction of a universal of thought or language that one would have supposed the very process of analysis had eliminated. This critique is thus simultaneously an epistemological critique of the myth of meaning as a myth of reference or denotation, as a myth of a shared ontology common to different physical theories or conceptual schemes.
Clearly, these two critiques and these two myths overlap; they are, Quine says, âat root identical.â I have sought to determine the relations and hierarchy between the theses of the indeterminacy of translation, the inscrutability of reference, and ontological relativity, by showing how Quine always proceeded by several pathsâproving the first thesis on the basis of the second or the third, the third on the basis of the first, and so on, thus constituting a veritable system. It then became apparent that the inscrutability of referenceâwhich became, importantly, the indeterminacy of reference after Pursuit of Truthâwas the central point. It was not that the problem of reference replaced or supplanted the problem of meaning, but that the question of the myth of meaningâand the anthropological problem it raisedâwas already the question of languageâs relation to the world, and of realism.
We know that Quine first proposed the expression âmyth of meaningâ in 1958 at the Royaumont conference,2 even though the problematic appears in several earlier texts (for example, âThe Problem of Meaning in Linguistics,â republished in From a Logical Point of View). This is because from the beginning Quine gauged the provocative aspect of the problem, which can be summarized in five points.
1. As is clear from the Royaumont conference, Quineâs thesis targets the idea of a core common to different languages and the idea of an intermediary entity that would guarantee equivalence or correspondence between them. âWhat I take issue with, in particular, is the idea of an identity or community of meaning under signs, or a theory of meaning that makes meaning into a sort of supralinguistic abstraction, of which the forms of language would be the counterpart or expression.â3 This is precisely the Fregean idea of meaning (Sinn) as a core common to languages, which constitute its âshellâ or expression.4
2. Quineâs thesis is an anthropological one. First, it raises the problem of interpretation in anthropology and in the domain of the human sciences as a whole (as Descombes has suggested). My goal in critiquing Davidsonâs, or Davidsonian, theories of interpretation has been to restore the original skepticism of Quineâs thesis. There is nothing âto be right or wrong about,â as Quine said, and thus no reason to always interpret a foreign utterance in what would be the most reasonable (for us) manner. Of course, the radicality of Davidsonâs skepticism must not be underestimated eitherâbut what I would like to question is a certain use of Davidson to trivialize the principle of charity and make it a principle of the universalization of rationality.
3. The thesis of the indeterminacy of translation bears on the indeterminacy of reference. It presents an ontological problem, although it is not based on any concern for ontological economy. âHume, not Occam, is my inspiration.â This reminder of Quineâs, made at Royaumont, should be remembered by all those who consistently approach the thesis of indeterminacy from the standpoint of âOccamâs razor, Platoâs beard.â The ontological problem of the thesis of indeterminacy is, in fact, the impossibility, or rather inanity, of asking the native, the other, but also myself âwhat one is talking about.â From this point of view, (anthropological) indeterminacy has inevitable consequences at the epistemological level, though not exactly as one would expectâit raises the question, among others, of what the objects of a/the theory of physics are. In one sense, it seems that if we follow Quine to the end, there is no answer to this question.
4. The thesis of indeterminacy does not counter mentalism (even if it could). It is true that in âThe Problem of Meaning in Linguisticsâ Quine states that âthe idea of an idea, the idea of a mental counterpart of linguistic form, is worse than worthless for linguistic science. I think the behaviorists are right in holding that talk of ideas is bad business even for psychologyâ5 (he was thus rejecting, one might say, psychologism even in psychology). But this is not the problem. In âOntological Relativityâ Quine specifies that âthe naturalistâs primary objection to this view is not an objection to meanings on account of their being mental entities, though that could be objection enough.â6 From there to saying that the thesis of indeterminacy is compatible with a restored or even minimal mentalism is a step I refuse to take. It seems to me even less possible to find in Quine a reintroduction of mentalism or psychologism via an examination of the radical interpreterâs âattributions of beliefs.â
5. The thesis of the indeterminacy of translation bears on the ordinary use of language. This is the point that today seems most worthy of examination. At first glance, the thesis can be expressed in terms of the problem of the synonymy of statements or terms. Why isnât the thesis of indeterminacy an ontological thesis? âReaders have supposed that my complaint is ontological; it is not. If in general I could make satisfactory sense of declaring two expressions to be synonymous, I would be more than pleased to recognize an abstract object as their common meaning. The method is familiar: I would define the meaning of the expression as the set of its synonyms. Where the trouble lies, rather, is in the two-place predicate of synonymy itself; it is too desperately wanting in clarity and perspicuity.â7 In an important article, âUse and Its Place in Meaning,â Quine takes up Wittgensteinâs question from the beginning of the Blue Book: âFor a provisional solution, consider what we often actually do when asked the meaning of a word: we define the word by equating it to some more familiar word or phrase. . . . We may persist, then, in the old routine of giving meanings by citing synonyms.â8 This takes us back to the very definition of the âmyth of meaning,â an expression, let us recall, that comes from Wittgenstein: âNegation: a âmental activity.â Negate something and observe what you are doing.âDo you perhaps inwardly shake your head? And if you doâis this process more deserving of our interest than, say, that of writing a sign of negation in a sentence? Do you now know the essence of negation?â In a note to this passage, Wittgenstein comments, âThe fact that three negatives yield a negative again must already be contained in the single negative that I am using now. (The temptation to invent a myth of âmeaning.â)â9
If, as Wittgenstein says, we are âtemptedâ to invent a myth of meaning, it is for good reason; we know how important the voice of temptation is in Wittgenstein, and it is not necessarily to be resisted, but rather simply to be heard and understood. Despite all his denials of a âfact of the matterâ of translation, Quine never says that statements âdonât mean anything.â They âsignifyâ or rather mean, in the ordinary sense. In âThe Myth of Meaning,â Quine asserted at the outset: âIt is not my intention to demonstrate that language presents no meaning. I do not disagree that the words and phrases we use have meaning, in the ordinary sense of âhave meaning.â10 For Quine, it is better to think of the verb âmeanâ as intransitive: statements mean, but this is not to say that they mean or signify some thing, a meaning. Thus, in âUse and Its Place in Meaning,â Quine proposes beginning with âmeanâ as an intransitive verb: âAn expression means; meaning is what it does, or what some expressions do. To say that two expressions are alike in meaning, then, is to say that they mean alike. Some expressions sound alike, some mean alike. It is significant that when we ask for the meaning of an expression we are content to be given another expression on a par with the first, similar to it in meaning. We do not ask for something that the two of them mean. The French idiom is more to the point: cela veut dire.â One may posit objects that are meant in common by statements that âmean alikeâ (rather than mean âthe same thingâ).
For once we understand what it is for expressions to mean alike, it is easy and convenient to invoke some special objects arbitrarily and let them be meant. . . .
. . . If we can manage this, then we can blithely say thereafter that expressions that mean alike have the same meaning.11
The problem is that we do not know what it is âto mean alikeâ; norâin particular, since the expression is more commonâdo we understand âto mean the same thing.â The empirical criteria Quine proposes in Word and Object collapse by themselves. This is the very object of his thesis of indeterminacy: there are no empirical or behavioral criteria for the synonymy of expressions or terms. Quine pursues the work of destruction begun in âTwo Dogmas of Empiricism,â claiming that there is no empirical meaning, not only (from the point of view of holism) because one cannot distinguish the empirical content of an isolated statement, but also for a supplementary reason, one that thus subordinates the epistemological to the logical in Quineâs philosophy and unifies the double criticism found in âTwo Dogmas.â This is the impossibility of correlating two statements in two different languages, or, more precisely, the fundamental indeterminacy and arbitrary nature of any kind of correlation. The thesis of indeterminacy and the assertion that meaning is a myth âcontaminatedâ Quineâs entire system, to such an extent that they then spread to all of analytic philosophyâs discussions and determined the very terms of the debate between realism and relativism.
It is clear that a part of this debate stems from Quineâs rejection of meaning. However, Quine does not reject meaning, or rather, the âto mean.â Instead, his goal, given the absence of a âfact of the matterâ of translation is to render meaning intransitiveâthe cela veut dire. There are no entities that are meant by language; there is only language. For Quine this does not mean that, as Rorty would say, we have lost the world, for it is only in language that it is to be found. âTruth is immanent.â
It is now time to explore the sense of these claims Quine makes. My orientation is radically opposed to any theory of interpretation inspired by Davidson and to any relativism Ă la Rorty. (The two are not as different as one might think, and, paradoxically, they join each other at a certain level). One might think that in the absence of a âfact of the matterâ of translation, in the absence of meaning or an intermediary Sinn, all that we are left with is the theory of interpretation, the play of conceptual schemes and the communality of language. But it seems to me that Quine means just the opposite. It is now no longer a problem of interpreting, adequately or not, Quineâs work and theses. A new problem has been raised by his successors and their endless discussions of realism, which are only now beginning to exhaust themselves. Quine poses the question of realism, but he resolves it immediatelyâhe thinksâby affirming an âimmanent truth.â12 I would like to show that he does in fact resolve the matter, but in a way that has not been understood (this lack of understanding was inevitable, and it led to the debate over realism) and could not be understood within the framework of his work alone, which wants to affirm empiricism and realism at the same time. This misunderstanding has been shared by Quineâs successors and critics Davidson, Putnam, and Rorty, each in his own way. It is my belief that in order to understand what realism consists in, or what the ârealist spiritâ is, we must stop putting the question of realism into the framework of empiricism. This would amount to undoing the connection between empiricism and realism that seems fundamental in the work of Quine and others. But, does realism in Quineâs work depend on empiricism? I am always brought back to Wittgensteinâs affirmation that the most difficult thing is to have ânot empiricism, and yet realism.â13 To determine the sense and stakes of this, it is useful to return to the empiricism-realism connection in Quine. As an examination of Quineâs relation to Carnapâs work would show, there is, despite everything (and in spite of the great coherence of his work), a tension in Quine between two tendencies. There is the naturalist or (although it is not the same thing) pragmatist tendency proclaimed in the famous last lines of âTwo Dogmas,â and there is the conventionalist or âlinguisticâ tendency in the style of Carnap. It is true that, after âTruth by Conventionâ (1935), and even as early as his 1934 lectures on Carnap, all the way down to his most recent works, Quine never stopped criticizing the doctrine of âtruth by convention.â But he criticized it less in order to give up on this doctrine than to give content to conventionsâin a sense, to justify them. This was Quineâs initial project, formulated in a text from 1937, âIs Logic a Matter of Words?â: to empirically reconstruct logic. âSuch a reconstruction would lend substance to the linguistic doctrine of logicâand would seem, incidentally, to merge with the empiricistâs program of delineating the connections of all statements with direct observation.â14 Such a project, as hopeless as it may seem, contains the seed of Quineâs entire work and all its contradictions, along with todayâs tensions within analytic philosophy. It is a matter of preserving the rigor of the laws of logic while connecting them to experience. âA principal virtue of the doctrine [that logic is wholly a matter of linguistic decision] is the clarity with which it explains the a priori character of logic,â Quine writes in this piece.15 But if logic is a matter of languageâand Quine has always said that it is, in the trivial senseâit is not therefore a matter of convention (except in the trivial sense in which anything that has to do with language is conventional), not mere convention. This is because every convention has its reasonsânatural reasons, as Cavell says in a different context16âand also because logic, inscribed within language, within the conceptual scheme, is no more or less connected to experience than the scheme taken as a whole. As Skorupski has noted, convention is not a matter of choice: âIt is not that we âagreeâ on a âconvention.â Rather, we discover that we share in common a certain constraint or limit in our thinking. The âlimits of empiricismâ are not conventions. They are ways in which we naturally go on.â17
The influence of Carnap may explain the tension between empiricism and conventionalism in Quineâs work. Quine says he is an unswerving empiricist, and he proclaims this empiricism anew in response to Davidsonâs lucid criticism of âthe very idea of a conceptual scheme.â18 For Quine it all beginsâto use...