Revising a system of beliefs › The second dogma: isolated confirmation

Reductionism initially appears as a proposal to translate meaningful statements into a language of immediate experience and logic. Section 5 considers Carnap's attempt in the Aufbau and distinguishes that ambitious project from a weaker assumption. Even if literal translation is abandoned, one might still suppose each statement has its own independently specifiable pattern of confirming and disconfirming experiences.

Quine challenges this weaker assumption too. A prediction about the temperature reading depends on claims about the instrument, its environment, and how the displayed number should be interpreted. The experience tests an arrangement of assumptions. Treating one sentence as possessing a complete evidential account independently of its companions leaves out those dependencies.

This is confirmation holism: the significance of evidence for a claim depends on a wider body of beliefs. It should be distinguished from the loose observation that ideas influence one another. The stronger philosophical claim concerns what encounters empirical testing. In §§5–6, Quine presents that unit in sweeping terms, extending it to science collectively. Quine, §§5–6.

The analytic/synthetic distinction enters again because an analytic truth could be treated as a limiting case: a sentence confirmed whatever experience occurs. But that characterization assumes we can assign an individual sentence its own confirmation conditions in the first place. Quine's attack on isolated confirmation therefore bears on one proposed way of identifying analyticity. This explains why the two criticisms belong in the same paper.