Concept

Analytic-Synthetic Distinction

Intro

Some sentences are true just because of what their words mean. "All bachelors are unmarried" is the classic example: once you know what "bachelor" means, you know the sentence is true without checking a single bachelor. Philosophers call that kind of truth analytic. Other sentences are true because of how the world happens to be. "All swans are white" cannot be settled by thinking about the word "swan"; you have to go and look (and when people looked in Australia, they found black swans). That kind of truth is synthetic.

The distinction matters in apologetics more often than people expect. It shows up when someone argues that God can be "defined into existence," when a skeptic says a moral truth is "true by definition," when a debate slides between "all swans in my example" and "all swans," and when a critic cites the philosopher W. V. O. Quine to say the whole distinction is a myth. The short version: the distinction is real and useful, Quine raised a genuine problem about how to explain it, and neither fact does the work people often want it to do.

◎ See this page's graph and discover linked topics →

In full

A proposition is analytic if its truth is fixed by the meanings of its terms and logic alone, so that its denial is a contradiction ("a bachelor is married" contradicts what "bachelor" means). It is synthetic if its truth depends on how things are beyond those meanings. Immanuel Kant crossed this distinction with a second one, between knowledge a priori (independent of experience) and a posteriori (dependent on it), and argued in the Critique of Pure Reason that the most important category was the synthetic a priori: truths that extend our knowledge yet are known independently of experience, such as, on his view, the truths of arithmetic and geometry.

In "Two Dogmas of Empiricism" (1951), W.V.O. Quine attacked the distinction itself. Every attempt to explain analyticity, he argued, appeals to synonymy, definition, interchangeability or meaning, and each of those, examined closely, is explained in turn by appeal to analyticity or something equally in need of explanation. The circle suggested to him that no principled boundary had been drawn, and he replaced the picture with confirmational holism: statements face experience as a whole "web of belief," and in principle any statement, even of logic, could be revised to accommodate recalcitrant experience. H. P. Grice and P. F. Strawson replied in "In Defense of a Dogma" (1956) that a distinction can be real and teachable even if it cannot be defined without circularity, since the same is true of many basic concepts.

For apologetics the distinction settles four recurring confusions: (1) a stipulation inside a thought-experiment is not an analytic truth; (2) failing to give a non-circular criterion for a distinction does not prove the distinction does not exist; (3) analytic truths tell you what follows if something exists, never that it exists; and (4) calling a moral claim analytic ("unjust killing is wrong") secures a conceptual connection, not the moral reality the connection is about.

Three categories that get blurred

The third row is where debates go wrong. Inside a stipulated hypothetical you cannot produce a black swan without changing the example, but that does not make whiteness part of the meaning of "swan." The assumption governs the model, not the word. An economics exercise that says "assume no taxes" has not made "economies have no taxes" true by definition.

Kant's two distinctions

The table matters because critics often run the two distinctions together. "Known without experience" and "true by meaning" are not the same property. Much of the debate about whether moral truths, mathematical truths or metaphysical principles like the Causal Principle can be known a priori turns on the bottom-left cell.

Quine's challenge, stated fairly

1. Analyticity is explained as truth in virtue of meaning, or as reducibility to a logical truth by substituting synonyms.
2. Synonymy, in turn, is explained as sameness of meaning, or as interchangeability in all contexts without change of truth-value.
3. Interchangeability "in all contexts" only works for the needed cases if the language already contains a notion of necessity, which is the analytic notion again.
4. So the explanations run in a circle, and no non-circular criterion has been given.
5. Therefore (Quine concludes) there is no principled analytic-synthetic boundary.

Steps 1 to 4 describe a real difficulty. The argument is also valid in form. The pressure point is the move to step 5, which requires an extra premise: that a genuine distinction must admit a non-circular reductive definition.

What the challenge does not show

1. A failed explanation is not a nonexistent thing. People could not give an airtight definition separating living from non-living things in every edge case; it did not follow that a horse and a rock are indistinguishable. Difficulty analysing a distinction is evidence about our theory, not about the distinction.
2. Not every circle is vicious. A proof that assumes its conclusion is viciously circular. A family of basic concepts (meaning, synonymy, necessity, definition) that illuminate one another may simply be basic. The critic must show the circle is explanatorily vicious, not merely that it exists. This was the core of the Grice and Strawson reply.
3. The holist still relies on logic he cannot revise. Quine's web of belief is itself revised by reasoning that obeys non-contradiction and valid inference. A proposal to revise the law of non-contradiction would have to be argued for using it. See Law of Non-Contradiction and Self-refutation.
4. "Indistinguishable" is a far stronger claim than "hard to define." Someone who says analytic and synthetic truths are simply indistinguishable owes an argument that goes beyond the circularity worry, and the bachelor and swan examples stand as obvious prima facie counterexamples.

Where it matters in apologetics

Defining God into existence. An analytic statement tells you what follows from a concept, never that the concept is instantiated. "God is necessarily good" states what would be true if God exists; existence needs a separate argument. That is why the serious theistic arguments (Modal Ontological Argument, Contingency Argument, the causal arguments) are metaphysical arguments, not definitions. The ontological argument's defenders argue from possibility and modal principles, not from a dictionary.

"Analytic" moral truths. "Unjust killing is wrong" is plausibly analytic, because "unjust" already carries the negative verdict. For the same reason it answers nothing. The real question is what makes a particular killing unjust: what the victim is due and why. An answer that says "because it is wrong" has gone round in a circle. This is the gap exploited by the Atheist Moral Realism Defeater: conceptual equivalence is not a truth-maker. (See also Is-Ought Fallacy.)

Modal arguments. Quine's hostility to analyticity went with hostility to quantified modal logic. The theistic modal arguments do not depend on Quine's targets, and the codex answers the modal side at Quinean Rejection of Modal Logic Objection Defeater and Metaphysical vs Logical Modality.

Stipulation games in live debate. A debater who sets up "all black things are swans" as an example and then claims every non-black non-swan "proves" it has confused a stipulation with an analytic truth and confirmation with proof. One black non-swan falsifies the universal. See Transcendental Argument for God (master objections) for how this misfire gets aimed at presuppositional apologetics.

See also

• W.V.O. Quine, the author of the challenge
• Immanuel Kant, the analytic, synthetic and a priori categories
• Empiricism and Rationalism, the schools the distinction divides
• Coherentism, the epistemology Quine's holism favours
• Atheist Moral Realism Defeater, why an analytic moral truth is not moral ontology
• Quinean Rejection of Modal Logic Objection Defeater, Quine's related attack on necessity
• Stroud Objection to Transcendental Arguments Defeater, the neighbouring debate about conditions of thought and of being
• Begging the Question, vicious circularity as distinct from basic concepts