Saturday, August 24, 2013

Quine and the Analytic–Synthetic Distinction

I am taking a quick detour through analytic philosophy and epistemology at the moment, as a type of prolegomena to economic methodology.

In 1951, Willard Van Orman Quine (1908–2000) published the now famous paper “Two Dogmas of Empiricism” (1951; reprinted in Quine 1981), in which he argued that the conventional idea of analyticity (or the analytic nature of a proposition) cannot be defended, and that the distinction between analytic and synthetic truths is not clear cut.

“Two Dogmas of Empiricism” is often said to be one of the most important papers in analytic philosophy of the late 20th century, though Quine continued to develop his epistemological views later in life, so that he modified or shifted the arguments used in “Two Dogmas” (Creath 2004: 47).

In “Two Dogmas of Empiricism,” Quine examines the intensional definitions or meanings given to the concepts of “analyticity” and “synonymy.”

Quine complains that the process of definition necessary for understanding “analytic truth” rests on the concept of “synonymy,” and that affirming or establishing that these two linguistic forms are proper synonyms is “far from clear” (Quine 1951: 25). Moreover, Quine asserts that appeals to “synonymy” supposedly fall back on “analyticity,” so that we have a circular argument.

Quine therefore argues that all attempts to define “analyticity” fall back on intensional terms in a way that is viciously circular, and that intensional definitions do “not hold the key” to the concepts of synonymy and analyticity (Quine 1951: 27). Ultimately, intensional terms like “analyticity” must be defended in extensional terms, or, that is to say, in terms of their reference to verbal behaviour (Glock 1996: 204).

Quine therefore concluded that the standard ideas of analyticity and analytic truth were indefensible, and so the analytic and synthetic distinction unclear.

Many responses to Quine were made (Grice and Strawson 1956; Putnam 1962; Quinton 1967; Glock 1996; Nimtz 2003; Gutting 2009: 11–30).

Glock argues that, although the process by which “analyticity” is defined is circular, it is nevertheless not a vicious form of circularity (Glock 1996: 204; Glock 2003: 75). Quine’s demand that an intensional concept like “analyticity” needs to be reduced to extensional ones is unreasonable and unnecessary (Glock 2003: 75).

We can consider the following proposition:
(1) All bachelors are unmarried.
It is not possible to deny the truth of this proposition without simply redefining one of the words, and the definition of “analyticity” in terms of synonymy is not unjustified if intensional meanings can be sustained without being reduced to extensional verbal behaviour.

Quine’s complaints, then, about the circularity involved in defining “analyticity” cannot be sound, nor are they sufficient to overthrow the definition of analyticity in terms of synonymy.

Quine himself later denied that his major criticism of the concept of “analyticity” in “Two Dogmas of Empiricism” was simply that attempts to define it are circular (indeed Glock 2003: 77 contends that Quine dropped the “circularity” charge as early as 1953), and other philosophers seem to accept this, and instead claim that Quine’s actual fundamental complaint was that the term of “analyticity” must be analysed in a behavioural sense.

So with Quine’s behaviourism very much the relevant background theory, Quine’s objection appears to be that neither “analyticity” nor “synonymy” can be reduced to verbal behaviour or behavioural criteria (Gibson 1996: 99; Creath 2004: 49; Gutting 2009: 21; Hylton 2006: 183). This view was confirmed when critics charged that Quine’s standard for the definitions of these terms was impossibly high, and Quine responded by saying precisely that he wished “no more, after all, than a rough characterization in terms of dispositions to verbal behavior” (Quine 1960: 207).

So it is clear that Quine wants a behaviourist semantics (Gibson 1996: 99). Quine also redefined analytic sentences to mean those that all people in a community of speakers learn are true by learning the meanings of the words involved (Gibson 1996: 100). But, Quine argued, neither this behaviourist semantical definition nor a “popular” or ordinary language notion of analyticity (Elugardo 1997: 15) can provide a strict scientific clarification that can defend the intensional semantic one that justifies the necessary verbal truth of analytic propositions (Gibson 1996: 100).

In essence, Quine’s attack on analyticity is to be understood in the logical positivist tradition of the verification principle: how is the term of analyticity to be related to the empirical verbal behaviour of human beings as judged by a methodological behaviourism? (Creath 2004: 49). The paradox, as Creath points out, is that:
“Quine is pushing against Carnap the very demands that Carnap had pushed against the metaphysicians” (Creath 2004: 49).
But Quine’s whole attempt to reject strict analytic truth would seem to collapse once we recognise that (1) the verification principle cannot be accepted, and (2) the whole behaviourist project is unsound (Gutting 2009: 21; Burgess 2004: 51–52).

Ultimately, then, Quine’s attempt to reject the analytic versus synthetic distinction is a failure.

These conclusions are broadly in line with the arguments of Quine’s critics, who find that analytic truths do exist, but they are, as in conventional empiricist epistemology, trivial or non-informative (Putnam 1962; Nimtz 2003).

For example, Putnam argued that there is indeed an analytic versus synthetic distinction but that it is ultimately a trivial one (Putnam 1962: 361).

Burgess, John P. 2004. “Quine, Analyticity and Philosophy of Mathematics,” The Philosophical Quarterly 54.214: 38–55.

Creath, Richard. 1990. Dear Carnap, dear Van: The Quine-Carnap Correspondence and Related Work. University of California Press, Berkeley, CA and London.

Creath, Richard. 2004. “Quine on the Intelligibility and Relevance of Analyticity,” in Roger F. Gibson, (ed.). The Cambridge Companion to Quine. Cambridge University Press, Cambridge, UK and New York. 47–64.

Elugardo, R. 1997. “Analytic/Synthetic, Necessary/Contingent, and a priori/a posterori: Distinction,” in Peter V. Lamarque (ed.), Concise Encyclopedia of Philosophy of Language. Pergamon, New York. 10–19.

Gibson, Roger F. 1996. “Quine’s Behaviorism,” in William O’Donohue and Richard F. Kitchener (eds.), The Philosophy of Psychology. Sage, London. 96–107.

Glock, Hans-Johann. 1996. “Necessity and Normativity,” in Hans Sluga and David G. Stern (eds.). The Cambridge Companion to Wittgenstein. Cambridge University Press, Cambridge and New York. 198–225.

Glock, Hans-Johann. 2003. Quine and Davidson on Language, Thought and Reality. Cambridge University Press, Cambridge.

Grice, H. P. and P. F. Strawson. 1956. “In Defence of a Dogma,” Philosophical Review 65: 141–158.

Gutting, Gary. 2009. What Philosophers Know: Case Studies in Recent Analytic Philosophy. Cambridge University Press, Cambridge.

Hylton, P. 2006. “W. V. Quine (1908–2000),” in A. P. Martinich and David Sosa (eds.), A Companion to Analytic Philosophy. Blackwell, Malden, Mass. and Oxford. 181–204.

Nimtz, C. 2003. “Analytic Truths – Still Harmless After All These Years?,” in H. J. Glock, K. Gluer, and G. Keil (eds.), Fifty Years of Quine’s ‘Two Dogmas’. Rodopi, Amsterdam and New York. 91–118.

Putnam, Hilary. 1962. “The Analytic and the Synthetic,” Minnesota Studies in the Philosophy of Science 3: 358–397.

Putnam, Hilary. 1975. “The Analytic and the Synthetic,” in Hilary Putnam, Mind, Language and Reality. Philosophical Papers. Volume 2. Cambridge University Press, Cambridge. 33–69.

Quine, Willard Van Orman. 1951. “Two Dogmas of Empiricism,” Philosophical Review 60: 20–43.

Quine, Willard Van Orman. 1960. Word and Object. M.I.T. Press, Massachusetts.

Quine, Willard Van Orman. 1981. “Two Dogmas of Empiricism,” in From a Logical Point of View. Harvard University Press, Cambridge, MA. 20–46.

Quine, Willard Van Orman. 1991. “Two Dogmas in Retrospect,” Canadian Journal of Philosophy 21.3: 265–274.

Quinton, Anthony. 1967. “The a priori and the analytic,” in P. F. Strawson (ed.), Philosophical Logic. Oxford University Press, Oxford. 107–128.


  1. A good deal of the wonder of the analytic/synthetic distinction melts away when your realise that so-called analytic statements are really just tautologies.

    Then when you realise that tautologies are ultimately just statements that assert their self-contained authority you begin to see that the Enlightenment project, which aims at multiplying tautologies (analytic judgments) in human discourse, is really just a self-reinforcing ideology with little relation to the real world.

    It's a sort of deductive poison that tries to turn human discourse into a bland, homogenous puddle of goo. It fails of course, no matter what guise it takes -- from neoclassical theory to econometrics -- because human discourse doesn't function like that.

    1. Well, the logical positivists did stress that analytic propositions are ultimately just non-informative tautologies (as did early Wittgenstein), and that the necessary truth of an analytic proposition is only ever a *verbal* (or de dicto) necessity from the conventions of language, not a metaphysical one.

      I think the problem you're really identifying when you refer to the "Enlightenment project" is Rationalist apriorism: the belief that starting from allegedly self-evident or a priori axioms and using deduction (and spurning empirical investigation), you arrive at profound new, informative knowledge of the world, and those deductive truths aren't amenable to empirical refutations.

      E.g., Misesian praxeology is the ultimate and crowning stupidity of this sort of Rationalist apriorist project.

      In reality, you need highly empirical study or the world in order to get new informative knowledge of it, and your theories must always be tested against reality.

      Would you agree with this?

    2. I do of course agree with this, because philosophy aside you and I agree on every substantive issue (which, by the way, is what's really important).

      But I still think that the Enlightenment project contains within it the seeds of spreading this crap. Enlightenment hates induction/empirics. Why? Because it cannot theorise it. It cannot get a handle on it. Hamann is right when he equates Enlightenment and the mathematical mindset. They're two sides of the same coin.

      Frankly I think that empirics/induction is anti-Enlightenment. Which is why I think that Keynesian economics is, properly conceived, anti-Enlightenment.

      As for Mises, he's like the exception that proves the rule; the crank clown who shows what the goal of the circus really is.