The Illusion of Analyticity Part 2: Quine’s Collapse of Analyticity

Breaking down the Analytic-Synthetic Distinction

All too often high quality assignments from coursework ends up gathering dust on shelves or in forgotten folders. In an effort to publicize some of my work, this is part two of my final paper from an Epistemology course this past semester.

Part One

While this is primarily meant for an audience with a working understanding of epistemology, I’ll provide relevant links to explain common terms. The paper primarily deals with the distinction between analytic and synthetic truths, so skimming through that explanation will be useful.


To collapse the dogma of analyticity and highlight its illusory nature, Quine attempts to find a satisfactory explanation for what exactly comprises analyticity. Quine’s argument first defines a set of sufficient conditions for analyticity. Namely, an analytic truth must be either (a) logically true or (b) true through some form of synonymy with (a), hoping that either the former or latter provide the key to explaining analyticity satisfactorily. As an example of (a) and (b), take the following statements:

1a. All unmarried men are not married.

1b. All bachelors are not married.

Epistemologists normally characterize both (1a) and (1b) as analytic. (1a) is an analytic truth logically as unmarried seems logically equivalent to not married. Epistemologists would say that (1a) is true come what may since one statement (unmarried man) is logically equivalent to the other (not married). Then, (1b) is analytic because it is directly synonymous with (1a), an analytic truth. Bachelors seems to mean the same as unmarried men and replacing bachelors with unmarried men results in a logical truth. So, the notion of analyticity appears tied to statements which are either direct logical truths or truths that are reducible to logical truths through synonymy.

So, it’s possible that synonymy satisfactorily explains analyticity. To clarify the idea of analytic truths through synonymy, Quine proposes that two terms are synonymous either through the definition of the terms involved or the terms’ interchangeability. Consider first synonymy through definition. If two terms, bachelor and unmarried man, are synonymous through their definition, they are synonymous insofar as

def: a bachelor is defined as an unmarried man

is true. If def is true and definition satisfactorily describes synonymy, then bachelor and unmarried man are synonymous by definition. It seems reasonable to think that def entails that bachelor and unmarried man are synonymous and a statement equating them would be analytic. However, this notion is mistaken as Quine points out that “the lexicographer’s report of an observed synonymy cannot be taken as the ground of the synonymy” (p 24). In essence, the definition of a term relies on the idea of synonymy itself. A definition is nothing more than a report on the synonymy of two words or phrases and does nothing to actually describe synonymy in the first place. Since definition relies on synonymy, it further obscures analyticity rather than explaining it. This result seems entirely unsatisfactory for pinning down synonymy through definition, as definitions themselves rely on the notion of synonymy.

Taking another approach to definition, one can look to abbreviations. Scientific language or technical terms that are just abbreviations of longer terms appear straightforwardly synonymous. Quine argues otherwise, highlighting that the basic and precise definitions of scientific or abbreviated languages, still rely on synonymy between concepts to establish the terms themselves. The relation between abbreviated words and the concepts they abbreviate is nothing more than one of synonymy, simply reporting a definitional relation between one term and another. Once again, this is merely an application of synonymy, not an explication. So, it seems that one cannot explain synonymy through definition without presupposing it.

Let us then turn to another potential explanation for synonymy, interchangeability. To more formally describe interchangeability consider two terms, X and Y. According to Quine, when one says that X is interchangeable with Y, they in fact mean:

3. Necessarily, all X are Y

X and Y are interchangeable insofar as (3) is true. Put another way, if every X is a Y, then one can replace Y wherever one uses X making them interchangeable. However, Quine notes that (3) itself assumes analyticity through reliance on necessity. Necessity relies on the idea of analyticity since necessary truths are merely an extension of analytic ones. So (3), rather than providing an explanation for analyticity, simply restates that

3a. All X are Y is analytic

When looking at (3a) in relation to (3), it simply replaces analyticity with necessity, a concept that relies on analyticity in the first place. Therefore, since interchangeability restates necessity, and necessity relies on analyticity, we’re left with no actual explanation of synonymy and analyticity, just a restatement of them.

With the failure of both definition and interchangeability to explain analytic truths through synonymy with logical truths, it appears that synonymy provides no satisfactory explanation for analyticity. Quine then returns to analytic truths that are logically true. Again, Quine aims to determine a satisfactory explanation for what “logically true” means in regards to analyticity. Diving into explaining logical truths, it appears that logical truths are true through the semantic rules of language. Consider the supposed analytic truth

4. frozen water is not unfrozen water

which seems to be true through nature of English itself. The statement uses the logical operators not and un- creating a statement that seems equivalent to the logical statement:

4a. P is not ~P

(4) and (4a) seem equivalent and analytically true through the semantic rules of the language. However, Quine still hangs upon the idea of a “semantic rule” in the first place, taking issue with the fact that one must define the inherent truth of semantic rules in the same way as analyticity itself. Again, relying on semantic rules to explain analyticity does not explain analyticity itself; it only provides a concept to replace analyticity that requires just as much explanation. Even if one considers a hypothetical language with only logical operators, semantic rules become synonymous with analyticity, doing nothing to actually explain the concept. In fact, according to Quine, this hypothetical language either presupposes the concept of analyticity or simply makes analyticity an irreducible concept. Of course, neither of those do any work in describing analyticity in the first place.

So, without logical truths or truths by synonymy with logical truths, it seems, that analyticity itself is insufficiently explained and does not warrant an epistemic distinction between analytic and synthetic truths. Quine, argues that this distinction is only a “metaphysical article of faith” (p 34), and thus should be abandoned in epistemological inquiry. As a result, given the other distinctions between necessary and contingent truths as well as a priori and a posteriori, Quine denies the existence of any analytic, necessary, or a priori truths. Instead Quine proposes that beliefs about the external world can only be evaluated “not individually, but only as a corporate body” (Quine 38).

Part 3: Grice and Strawson’s Objections


Feldman, Richard. Epistemology. 2003. Print.

Grice, H.P and Strawson, P. F. “In Defense of a Dogma.” Philosophical Review, vol 65 (2), p. 141–158. 1956. Print.

Quine, Willard V. O. Two Dogmas of Empiricism.” Philosophical Review , vol 60 (1), p. 20–43. 1951.