Epistemic contextualism, which I’ve been revising yesterday and today, is the thesis that the word ‘know’ is context-dependent. Perhaps one initial motivation for this sort of view is the common response to classic sceptical thought experiments, that the philosophy lecturer is working with some high, overly-demanding sense of ‘know’, and the rest of us can do quite alright with our ordinary sense (thank you very much), in which case the possibility that we’re in the Matrix doesn’t mean that we don’t know we have hands.

David Lewis’s contextualism[1] is much more sophisticated than this but is also, I think, better motivated. I’m not going to try to explain how so in full here (that’s what my revision is for) but I want to note what I find most attractive about it. Lewis takes knowledge of P to be a matter of the agent having eliminated the relevant possibilities that not-P; for example, I know that the cat isn’t in the room because I’ve eliminated the possibilities that it’s in the drawer, that it’s in the wardrobe etc. We properly ignore some not-P possibilities: I don’t fail to know that the cat isn’t in the room because I haven’t checked inside the photocopier. Classic sceptical possibilities are ones that we almost always ignore: indeed, everywhere but the epistemology class.

Consider what happens if someone asks, “wait, did you check the photocopier?” While we might not go and check the photocopier and continue to believe that the cat is not in the room, it is now very uncomfortable to maintain that we know that the cat isn’t in the room, and this discomfort remains until we do actually go and check the photocopier. So: raising a possibility that we were previously ignoring, that we can’t/haven’t eliminated, destroys knowledge. This is exactly what the sceptic does when he suggests that we’re really in the Matrix: he raises a possibility to salience, and thus we cease to know that we have hands.[2] So knowing that we have hands requires us to be not thinking about brains in vats.

I think that this is exactly what Hume is getting at in the famous conclusion to book 4:

Most fortunately it happens, that since reason is incapable of dispelling these clouds, nature herself suffices to that purpose, and cures me of this philosophical melancholy and delirium, either by relaxing this bent of mind, or by some avocation, and lively impression of my senses, which obliterate all these chimeras. (Hume, Treatise,; my emphasis)

This inevitable tension in our concept of knowledge—that we are happy to ascribe it to people yet epistemic vertigo is very easy to induce—is an aspect of the human condition that contextualism does a very good job of capturing.

It’s good to have something I actually want to defend (from technical objections) on my own terms as part of my metaphysics & epistemology paper, since much of the rest of this paper has bored me. I’d just managed to convince myself that I wouldn’t hate myself forever if I don’t get into grad school, but engaging with ideas like these makes it hard not to see philosophy as the only thing for me.


[1] Lewis, David. 1996. Elusive Knowledge. Australasian Journal of Philosophy 74 (4):549–67.

[2] This is not quite right, of course—it’s that ‘know’ now refers to a different relation as we’ve made a context-switch. As I say I’m glossing over how the account actually works.