Fallibilism and Multiple Paths to Knowledge

Abstract

Being a fallibilist isn’t easy. A fallibilist about empirical knowledge, in Lewis’s (1996) sense, holds that an agent can know a contingent empirical proposition P , even if she has not ruled out every last way that P could be false.1 In this sense, it seems that most contemporary epistemologists are fallibilists, at least relative to some way of understanding what it is to “rule out” an alternative. And with good reason: if knowing a contingent empirical proposition P required ruling out every last way that P could be false, then we would have little if any empirical knowledge. Radical skepticism would reign. Yet fallibilism, despite its promise for defending the possibility of knowledge, also faces problems. To borrow an analogy sometimes applied to philosophical projects, trying to fill in the details of a fallibilist theory of knowledge is like trying to install an unstretched carpet: flatten a problematic lump in one place and a new one appears elsewhere. But then again, the alternative of radical skepticism about knowledge is like having the rug pulled out from under your feet. The primary goal of this paper is to argue that what I call the standard alternatives picture, assumed by many fallibilist theories, should be replaced by a new multipath picture of knowledge. In §2, I identify the problematic lumps in the standard picture: fallibilists working with this picture cannot maintain even the most uncontroversial (single-premise, logical) epistemic closure principles without having to make extreme assumptions about the ability of humans to know empirical truths ∗For helpful feedback on this paper, I wish to thank Justin Bledin, John Campbell, Peter Hawke, Thomas Icard, Ethan Jerzak, Krista Lawlor, John Perry, Michael Rieppel, Shane Steinert-Threlkeld, Justin Vlasits, Seth Yalcin, and two anonymous referees for Oxford Studies in Epistemology. For helpful conversations or correspondence on issues discussed in the paper, I wish to thank Johan van Benthem, Keith DeRose, John MacFarlane, Sherrilyn Roush, Crispin Wright, and Stephen Yablo. 1The term ‘fallibilism’ means many different things to many different people. I explain in more detail what I mean by ‘fallibilism’ in §2.1.