A chemist can abandon a hypothesis when an experiment goes badly. Could a logician, faced with enough trouble, abandon a law of logic? For much of early analytic philosophy, the answer seemed to be no. Experiments tested claims about the world; logic and mathematics supplied the framework within which testing took place. In 1951, W. V. O. Quine challenged that division. His most arresting sentence was brief: “No statement is immune to revision.”
The boundary philosophy wanted
Quine’s target in “Two Dogmas of Empiricism” was the distinction between analytic and synthetic statements. An analytic statement was supposed to be true by virtue of meaning: “All bachelors are unmarried,” for example. A synthetic statement, such as “The street is wet,” depended on how the world happened to be. The distinction promised to explain how logic and mathematics could be certain without reporting discoveries about the world.
The promise mattered especially to logical empiricists. They wanted philosophy to clarify scientific language, not compete with science by announcing truths beyond possible experience. If some sentences held solely because of meaning, philosophers could identify the rules of inquiry while leaving factual investigation to scientists. Quine did not deny the usefulness of logic or definitions. He asked whether “true by meaning” marked a boundary that could be explained without already assuming it.
Where definition stops helping
Consider the bachelor example. We can say that “bachelor” means “unmarried man,” making the sentence analytic. But how do we establish that the two expressions mean the same thing? A dictionary records usage; it does not create the synonymy it reports. A definition can establish synonymy if we introduce a new term by stipulation. Most words, however, were not introduced that way. They arrived with a history of use.
Perhaps we should say instead that the expressions can be substituted for one another without changing a sentence’s truth. Even that test needs care. Substitution works in “Every bachelor is a bachelor.” It does not work everywhere: someone may believe that a bachelor lives next door while not believing that an unmarried man does. We must specify which contexts count, and the most promising specifications tend to rely on notions of necessity or meaning—the very notions the test was meant to clarify.
Quine’s point was not that nobody understands “bachelor.” It was that familiar examples do not, by themselves, explain the philosophical distinction. Moving from “these words are interchangeable here” to “this sentence is true independently of any fact” requires an argument. Quine thought the available arguments moved in circles.
The second dogma: testing one sentence at a time
The other dogma was reductionism: the thought that each meaningful factual statement could, at least in principle, be tied to its own range of possible sensory experiences. On this picture, we could test a sentence against experience while keeping the rest of our beliefs fixed. Quine argued that scientific testing does not work so neatly.
Suppose an astronomer predicts where a planet will appear and the telescope shows nothing there. The failed prediction does not identify a single false sentence. The astronomer might revise a theory of motion, a calculation, an assumption about the telescope, or a report about what was observed. Each option carries a cost; none is selected by the bare observation alone. As Quine put it, “The unit of empirical significance is the whole of science.”
That formulation can sound as if every belief is equally exposed to every experiment. It does not mean that. Some commitments are close to particular observations and are readily changed; others are so central that revising them would force extensive changes elsewhere. Quine’s claim was that this difference is one of position and practical cost, not an absolute division between sentences answerable to experience and sentences forever exempt from it.
A web with no protected center
Quine described our beliefs as a “web.” Around its edges lie statements closely connected to observation. Nearer its center sit mathematics and logic, which help organize countless other statements. When experience conflicts with the web, we can make adjustments in more than one place. We normally protect central principles because changing them would disrupt so much. Protection, though, is not immunity.
This is the force of the sentence “No statement is immune to revision.” Quine did not propose discarding ordinary logic whenever a prediction fails. He denied that its standing could be secured simply by calling its truths analytic. If a deep enough scientific difficulty made an alternative logic worth adopting, its centrality would not rule out the change in advance.
There is a risk of exaggerating the result. A choice to revise logic would require an account of what the alternative explains and what it costs. Quine’s view does not make evidence optional or turn disagreement into a matter of taste. It makes justification comparative: we assess a proposed change against the larger body of inquiry it must preserve, repair, or improve.
What the argument did—and did not—settle
The challenge drew a direct reply. H. P. Grice and P. F. Strawson argued in 1956 that our ability to distinguish disputes about words from disputes about facts gave the analytic–synthetic distinction more standing than Quine allowed. Rudolf Carnap’s work on constructed languages offered another complication: within an explicitly specified system, rules can determine which sentences count as true by those rules. Quine’s objections to ordinary appeals to meaning do not simply erase that achievement.
The harder question is what such a constructed system tells us about our actual practice of revising beliefs. We may fix a rule within a language and still ask whether that language remains useful for inquiry. That question connects even our rules to purposes and experience. It helps explain why Quine’s essay remained influential despite continuing disagreement over its conclusions.
Before Quine, an analytic philosopher could hope to separate the framework of knowledge from its contents, then examine the framework from a secure position. “Two Dogmas” made that position difficult to hold. It asked philosophers to see their own methods as part of the same revisable enterprise as science. The boundary did not disappear without argument, and its defenders did not fall silent. But after Quine, calling a sentence true “by meaning alone” was no longer the end of an explanation. It was the point at which one had to begin.
Leave a Reply