"Two Dogmas of Empiricism" is the title of a paper written by Willard Van Orman Quine. Here he attacks the problematic assumptions of Logical Positivism, which was at the time much in vogue. I wrote this writeup to fill a curious gap in the nodegel... we have Davidson on the Third dogma, but nothing on the first two specifically. As this writeup is mainly from memory (and a poor memory at that), I would appreciate any corrections. Not having dealt with Logical Positivists beyond the surface level, I especially welcome corrections regarding my interpretation of Quine's interpretation of their work.
What are the two dogmas of empiricism, and how are they related?
The two dogmas of empiricism (illustrated for Quine by the logical positivists of the Vienna Circle) are the analytic/synthetic distinction, and the idea that any theory/statement can be reduced to a statement about sense data. Quine sees both of these dogmas as fundamental tenets of the verification theory of meaning, which he vigorously attacks.
The analytic/synthetic distinction that he attacks can be stated as follows.
All possible statements fall into one of the following three categories:
- Analytic: necessarily true either,
- a) by virtue of their terms (e.g., All unmarried men are unmarried)
- b) by virtue of their meanings (e.g.,All unmarried men are bachelors)
- Contradictory: necessarily false either,
- a) by virtue of their terms (e.g., All unmarried men are married)
- b) by virtue of their meanings (e.g., All unmarried men are not bachelors)
- Synthetic: contingently true or false. (e.g., I like cats)
Quine’s criticism of analyticity deal entirely with the second type of statement: those that are true by virtue of their meanings. What Quine wants to show is that there is no distinction between synthetic and b-type analytic statements, that, in fact, these ‘necessary truths’ are revisable.
The Vienna Circle’s verification theory holds, roughly, that 'the meaning of a statement is its method of verification.' So, for example, the meaning of “the sun is shining” is whatever means we use to verify that ‘the sun is shining’: by looking up at the sun, by measuring the amount of photons around us, etc. etc. Thus, we can isolate the meaning of a particular sentence from the rest of language by verifying it through observation. The logical positivists believe this is true of any meaningful sentence.
So, if the meaning of a statement is its method of verification, then two statements verified in the same way should have the same meaning. Yet, if there are an infinite number of ways in which to verify any statement (continuing the sun example: we can look at the sun from an infinite number of different angles, times, places, etc.) we can neither determine what a statement means or if it means the same as another statement with this theory. The biggest problem that Quine sees with this theory is that there in the assumption that the sentence is the fundamental unit of meaning. The logical positivists assume that they can determine the meaning of a single statement in isolation from any context. Quine, on the other hand, believes that meaning can only be determined in relation to a far larger context, perhaps even the whole of science itself. You cannot simply reduce the meaning of a statement to its method of verification because there is no ‘absolute’ verifiability for any sentence; there will always be untested viewpoints from which to possibly verify a statement.
So, to review Quine’s position on analytic statements of type b:
There are no analytic statements that are necessarily true by virtue of their meanings because meaning cannot be absolutely determined. It cannot be determined, because the context in which it is determined is never fully exhausted, one can always reevaluate, or reinterpret, at any time to change the meaning of any given statement. In addition we cannot assign equivalent meanings to words like ‘unmarried’ and ‘bachelor’ in order to verify statements such as ‘All bachelors are unmarried men’ because the notion of synonymy assumes that meaning is rigorously determined.
This brings us to the second dogma that Quine talks about, that of reductionism. Reductionism is the belief that all ‘theory-statements’ (the ‘sun is shining’ example above could be considered a theory-statement) are reducible to statements about sense data (‘there are photons entering my eye’ might be seen as such a statement, though problematically, as we shall see). Quine will argue that this reduction of ‘theory-statements’ to statements about sense data as a means of verification is untenable. He believes that you simply cannot verify any one statement without relation to a large number of other statements (possibly an infinite number of statements). For instance, if we verify the statement that the sun is shining by saying that "I can see the sun", this statement relies on thousands of prior assumptions about the nature of light, the action of the photon, the nature of the eye, etc. etc. So it seems that the distinction between purely theoretical and purely sensory statements is as untenable as the analytic/synthetic distinction; any and all statements are theory laden, because in making any statement, we have to make prior decisions about what statement to make, etc. etc. We cannot simply say, “I am in pain” because this already requires a theory about recognizing pain. Quine believes that any statement, rather than being meaningful in and of itself, acquires meaning (and verifiability) only through its relation to the whole of science (or, less radically, to its particular context).
This is a kind of meaning holism.
Quine's famous paper ("Two Dogmas of Empiricism")can be found in a collection of his papers entitled, From a Logical Point of View (Harvard, 1953).