June 03, 2004

Knowledge and Stability

Marc Moffett has been putting up a lot of interesting stuff. In particular, check his posts on how stable knowledge must be, with lots of cool examples--here's the middle one. The idea, partly, is that you can fail to know because your belief is too unstable--if you have a justified true belief now, but you're going to lose that belief soon, you don't know even now. At least, that's his intuition in certain cases; mine sometimes differ.

I have a quick answer to this problem: I don't care whether you know in these cases, and anyway if I invoke my intuitions on such cases Jonathan Weinberg will burst through the wall and drag me off to Hell like the Kool-Aid Man. But I think Marc is getting at stuff that's important even for a knowledge-derider like myself.

My position is that knowledge is a rough amalgamation of several independently important desiderata--truth and justification most prominently. Another desiderata is stability. As Williamson argues in the "Primeness" chapter of Knowledge and Its Limits, stability is important because it allows for more robust explanation. When asked why the burglar found the silver in the house,

(1) She knew that there was silver in the house

is a better explanation than
(2) She had a justified true belief that there was silver in the house,
because the (2) leaves open the possibility that her belief was based on a false lemma. If the belief was based on a false lemma, then she might have discovered the falsehood of that lemma and given up looking. (1) excludes that--if her belief isn't Gettierized, there's no belief that, when flipped from false to true, should cause her to stop looking. Similarly for evidence she does not possess; if (1) is false and (2) true because the burglar lacks misleading evidence, then she may discover the evidence and so stop looking. (Though I never had a strong intuition about the evidence-you-don't-possess examples.)

Think of this in terms of practical reason. Plans take time to execute. If you're going to act on a belief, you want it to last as long as the plan does--otherwise you'll disrupt your plan midway through. So it's not enough for your belief to be true; it's got to be stably true to serve as a foundation for helpful action. That's the problem with the belief Marc discusses--it's not going to last long enough to do what he needs it to.

Posted by Matt Weiner at June 3, 2004 04:53 PM