Do philosophers still believe in the analytic-synthetic distinction? Did Quine in his attack on the analytic-synthetic distinction go too far or did he get it about right?
Answer by Massimo Pigliucci
Here is how Willard O. Quine put the challenge, in his famous paper, Two dogmas of empiricism, published in 1953:
"It is obvious that truth in general depends on both language and extralinguistic fact. … Thus one is tempted to suppose in general that the truth of a statement is somehow analyzable into a linguistic component and a factual component. Given this supposition, it next seems reasonable that in some statements the factual component should be null; and these are the analytic statements. But, for all its a priori reasonableness, a boundary between analytic and synthetic statements simply has not been drawn. That there is such a distinction to be drawn at all is an unempirical dogma of empiricists, a metaphysical article of faith."
We first need some context. Quine was reacting to the then prevalent tradition in philosophy of science, logical positivism. The logical positivists in turn were interested in furthering David Hume’s famous dismissal of metaphysics in favor of math and empirical science (Hume was one of the most influential British empiricists of the 18th century). Here is Hume, in An Enquiry Concerning Human Understanding, first published in 1748:
"If we take in our hand any volume; of divinity or school metaphysics, for instance; let us ask, Does it contain any abstract reasoning concerning quantity or number? No. Does it contain any experimental reasoning concerning matter of fact and existence? No. Commit it then to the flames: for it can contain nothing but sophistry and illusion."
Hume, in other words, recognized only two types of meaningful statements: he called them ‘relations of ideas’ and ‘matters of fact.’ Today we would say that the first pertain to logic and mathematics, the latter to the natural and social sciences. Everything else is ‘but sophistry and illusion.’ (With all due respect to Hume – of which I have a lot! – it needs to be highlighted that the entire Enquiry contains no math and very few matters of fact, and yet it would really be a pity to commit that masterpiece to the flames.)
The logical positivists formalized Hume’s fork and emphasized the distinction between analytic statements (true by virtue of their semantic meaning) and synthetic ones (true in virtue of observation or experiment). A classic example of the first type of statement is: ‘All bachelors are unmarried,’ since ‘unmarried [man]’ is synonymous with ‘bachelor.’ More importantly, though, all of logic and mathematics are also into this same business of producing analytic truths. Examples of synthetic truths, by contrast, are statements such as ‘Saturn has rings,’ or ‘my house has two bedrooms.’
The reason the contrast between analytic and synthetic statements matters is this: all synthetic truths are a posteriori, i.e., arrived at by empirical means. But analytic statements fall into two categories: those that are true by definition (the bachelor case) and those that can yield non trivial a priori truths, such as the Pythagorean theorem, for instance. Quine, who was a pretty radical empiricist, was bothered by the possibility of a priori truths, and consequently also did not much like the idea of analytic ones.
His famous critique of the ‘dogma’ of the analytic-synthetic distinction, however, hinges on a very technical matter, and one that leaves a number of philosophers not entirely convinced. Basically, Quine argued that in order to claim that an analytic statement is truly such one has to provide an account of synonymy, since it is the latter concept that does the actual philosophical work: for instance, when we say that ‘All bachelors are unmarried’ we understand this as an analytic truth precisely because, as stated above, we mentally equate the terms ‘bachelor’ and ‘unmarried [man],’ i.e., the two terms are synonymous. But whence synonymity? According to Quine, at some point, even the notion of synonymity itself needs to be anchored by some sort of empirical fact, for instance about marriage, or men. If that’s the case, then the apparent solidly impenetrable barrier separating analytic and synthetic statements is no such thing and all truths, at bottom, are synthetic. This is Hume on steroids, in a sense.
As usual in philosophy, there have been thorough and numerous responses to Quine’s claim. For instance, Paul Grice and Peter Strawson pointed out that Quine’s skepticism about synonymy quickly leads to skepticism about meaning itself, which in turn leads to the problematic conclusion that one cannot actually determine whether a translation of a given sentence is or is not correct. This was a bullet that Quine later was apparently happy to bite rather than dodge, in his Work and Object (1960), where he presented the idea that translations are, in fact, indeterminate. Hilary Putnam argued that Quine’s critique actually confuses two different targets: analytic statements and a priori truths, which are not co-extensive. John Searle pointed out that even if Quine’s attack is granted some force, it doesn’t follow that the notion of analyticity is in fact useless.
I’m with Searle here: Quine’s analysis was an example of how philosophy makes progress by questioning previous assumptions (in this case the idea that there is a sharp distinction between analytic and synthetic truths) and providing reasons to think that things are different (and usually more complicated) then previously thought. But even if we admit that ‘All bachelors are unmarried’ eventually does connect to some empirical fact of the matter necessary to anchor the meaning of the phrase, it is somewhat daft to claim that there are therefore no interesting distinctions between that sort of sentence and more obviously synthetic ones like ‘Saturn has rings.’ Moreover, it seems that mathematics and formal logical truths still stand very much in the realm of analyticity, Quine’s stamping of his feet notwithstanding.