Skip to content
Coordinates of ThoughtHome of Philosophy
Join
Debate

Thesis Quine was right that even logic could be revised in the light of experience

Ethan Mercer Fellow
· 6 replies
Logic & argumentAnalyticGreek & RomanAristotleHilary PutnamW. V. O. Quine#quantum-logic#revisability#web-of-belief
Follow

Where it stands

Be gentle but not too gentle. We read the end of Quine's "Two Dogmas of Empiricism" (1951) in a seminar last week, and I think its most radical claim is correct: no statement is immune to revision, logic included.

The picture: our beliefs form a web that faces experience as a whole. When experience conflicts with the web, we can adjust anywhere. Logic sits at the centre, so we revise it last, because changing it disturbs everything. But last is not never.

And it's not just a thought experiment. Putnam's "Is Logic Empirical?" (1968) argued that quantum mechanics gives us reason to abandon the distributive law. You may think he was wrong about the physics, but the shape of the argument is exactly what Quine said was possible.

My thesis, then: logic is our most entrenched belief, not something of a different kind from belief. Objections?

LikeReply
Follow6 replies

Join the conversation: members reply here, and joining is free. Reading is always free.

Join free to reply Sign in

6 replies

For the thesis 2

  1. Tomasz Szewczak

    For Fellow

    For the thesis, partly. Logic has been revised in practice: logicians moved from syllogistic to predicate logic, and intuitionists revised classical mathematics. Giulia's objection is strong against revising everything at once. But you can revise one principle while reasoning in the shared core of the old and new logics, the inferences both accept. That's how actual logical revisions have gone. No one was left without a logic in between.

    1. Ethan Mercer

      Fellow

      Tomasz's reply is the one I'd want to make. But Giulia's point leaves me with a smaller thesis: maybe there is a minimal core that can't be revised, because any revision would use it. Then "logic is revisable" becomes "almost all of logic is revisable". Is that still Quine?

      Helpful · 1
  2. Emeka Anyaoha

    For Fellow

    For the thesis, by analogy. Kant thought Euclidean geometry was known a priori and necessarily true of space. Then general relativity described a space that isn't Euclidean. "Self-evident" turned out to be "entrenched". That doesn't prove logic is the same, but it should make us nervous about declaring anything unrevisable, which is all Ethan needs.

    Helpful · 3

Against the thesis 3

  1. Lars Eriksen

    Against Fellow

    Against, from the physics side. Quantum mechanics is formulated in completely classical mathematics: Hilbert spaces, linear algebra, classical proofs. Working physicists don't use quantum logic. And Putnam himself later gave the proposal up. So your best example of experience revising logic is an example of a philosopher proposing it and physics not needing it.

    Helpful · 2
  2. Giulia Ferrante

    Against Fellow

    The deeper objection is Aristotle's. To decide whether to revise a logical law you have to reason, and that reasoning uses logic. If you revise the rules in the middle of the argument, what has the argument shown? Aristotle didn't try to prove non-contradiction; he said any proof would presuppose it, and instead challenged the denier to say something determinate. That makes logic look less like a belief in the web and more like the condition for having a web at all.

    Helpful · 3
  3. Wanjiru Njenga

    Against Fellow

    An awkward point from Bayesian epistemology. In a Bayesian framework you can always revise empirical beliefs, as long as you never give anything probability exactly 0 or 1. But the probability axioms themselves give every logical truth probability 1. So the most flexible framework for revising beliefs we have presupposes classical logic rather than putting it in the web. That fits Giulia's picture more than Quine's.

    Helpful · 2