Ripped at the seems

I don’t often have ideas about epistemology. Longtime readers may remember me talking about epistemic virtues of personal concern, and for a while I misguidedly tried to push the idea that Gettier was wrong about knowledge not being justified true belief. Recently though, I was listening to Peter Adamson’s excellent podcast series on the history of philosophy, and when he got to the ancient sceptics I started thinking about epistemology again. Here’s my idea.

In epistemology, there are (of course) lots of debates, arguments and projects about knowledge. For example, sceptical arguments say that nobody knows anything, at least in some domain of inquiry. There’s also a kind of Kantian project that starts instead from the fact that we do have knowledge, and asks how everything else might be set up to make that possible. My idea is to have analogous debates but replacing “S knows that P” or “I know that P” with “it seems to S that P” and “it seems to me that P”. This goes off in a few directions.

First, there’s a sceptical argument that nothing seems to you to be true other than that you’re having the experiences you’re now having. Just as with knowledge, you can argue for this by considering sceptical scenarios, such as that you’re a brain in a vat hooked up to a virtual reality machine (a bit like the Matrix), or that you’re dreaming, or that Descartes’ evil demon is tricking you. Now, maybe you don’t know that these things aren’t happening, but at least they seem not to be happening, right? Well, that isn’t obvious. I mean, what would these things seeming to be happening be like? It’d be just like it is now. So maybe it actually does seem exactly like you’re in the Matrix, apart from the greeny-grey colour filter which we can ignore for present purposes. At least, it seems you’re in the Matrix just as much as it seems you’re in the regular external world. The only thing these seemingly-true scenarios have in common is your experiences, so maybe all we can really say is that you seem to be having the experiences you’re having now.

For fans of modal logic, I guess this is an understanding of seeming according to which something seems to be true iff it’s true in all possible worlds where you’re having the experiences you’re actually having. Maybe you could try replacing “all possible worlds” with “close possible worlds”, producing a kind of externalist understanding of seeming which is a bit like Nozick’s understanding of knowledge. I wonder if it’d be open to similar objections to Nozick’s view, including the absolute pummelling of it that Kripke published in 2011.

We don’t have to revise our understanding of seeming that way, though. We can revise it a different way which allows some things other than our experiences to seem to be happening, while avoiding the objections to Nozick's view. We can think about what seeming, and people, and the world would have to be like for there to be non-experiential seemings. This is like the Kantian project I mentioned before. Maybe biology can help. To me it, er, seems that this might be more plausible than coming up with a biologized conception of knowledge that lets us have what we want, because knowledge is a more normatively loaded concept. Being lucky enough to have cognitive biases might make things seem to be true, but could they make you know things? I don't know. Maybe we can construct a plausible naturalized epistemology around seeming which isn't vulnerable to the same normative criticisms as one constructed around knowledge.

If you’re still on board at this point, there are a couple of ways you might want to go. One is to be like Sextus Empiricus (I think), and never make claims about knowledge: you just say how things seem to you. You might think a position like that was unstable, because when you assert the position you’re committing to knowing it, not just to it seeming to be true. I don’t think that’s necessarily right though: we could posit a speech act which is correct when its content seems to you to be true, as assertions are (let’s say) correct when you know their contents. (More boringly, we could just allow that we do know how things seem to us.) We could investigate what kind of logical norms would apply to such a speech act. Maybe you can’t know contrary propositions, but can contrary propositions both seem to be true? If P&Q seems to be true, does that mean P seems to be true as well? There’s room for productive debate about those and similar questions. I think it’s perfectly possible that an epistemology constructed around seeming might be stable in a way that full-on scepticism of a kind which doesn’t allow any assertions or put anything in assertion’s place isn’t stable. The seeming-based epistemology might not be stable either, but I’d like to know, and if it is, and it’s what the ancient sceptics had in mind, that would be very cool.

Maybe we don’t have to content ourselves with seeming, though. One way epistemologists sometimes frame the problem of scepticism is as the problem of getting knowledge from non-knowledge. Maybe a beefed-up notion of seeming can give us some traction on this. That’s because if something seems to be true, that’s plausibly a defeasible reason for believing that it is true. So from seeming, which isn't knowledge, we get reasons for belief. Can we get all the way to knowledge? Maybe we could with a bit of ingenuity. Mark Schroeder wrote a book called Slaves of the Passions in which he tried to reduce all of our normative concepts to the idea of facts being reasons for people to do actions. I don’t actually think he succeeds, but it’s impressive how far he gets. (It really is a fantastic book; I can't praise it enough.) And if you can construct a bunch of normative concepts out of ‘P is a reason for S to do X’, maybe you can construct some epistemic concepts out of ‘S has a reason to believe P’. That would also be very cool.

The homogeneous spinning disc

People like me who like Humean supervenience, four-dimensionalism and things like that probably ought to pay more attention to the homogeneous spinning disc argument. It gets attributed to Kripke and Armstrong in the 1970s and CD Broad earlier than that. (Broad had currents in a homogeneous liquid.) The argument means to show that how the world is qualitatively at each time doesn’t determine how it is qualitatively overall, because a world containing only a disc of homogeneous matter would be the same at each time whether the disc was spinning or not. Katherine Hawley seems to have fairly similar inclinations to me, but she still responds to the argument by rejecting Humean supervenience, saying that there are non-supervenient relations holding between instantaneous bits of the disc at different times, in virtue of which an instantaneous bit at one time is a future stage of an instantaneous bit at another. It seems to work. I don’t have a big problem with the very idea of immanent causation grounding temporal counterpart relations, but I don’t generally think of it as fundamental, because violations of Humean supervenience are odd. The spinning disc argument is powerful though, and I should have a response. I think Hawley’s arguments against the responses she considers look pretty good, but lately I’ve been wondering about a different one. I don’t know if it’s already out there, but I’ve not seen it before.

The difference between the spinning disc and the stationary disc is supposed to ground counterfactuals about things like what would have happened if some paint or a volleyball had landed on the disc. I’m wondering whether a homogeneous disc really would do different things to some paint or a volleyball when they landed on it. If a homogeneous disc would interact with its surroundings in exactly the same way whether or not it was spinning, it’s easier to deny that it makes sense to say that a homogeneous thing is spinning or not. One way of thinking about it is that when something is internally moving like when a disc spins or there are currents in a liquid what’s really moving are the inhomogeneities. Now, in the actual particulate world a particle can be looked at as an inhomogeneity, especially if you’re a supersubstantivalist. So on this way of thinking, particles still move. But to say that the contents of a homogeneous region of space was moving internally wouldn’t make sense, for the banal reason that inhomogeneities are what moves, and it doesn’t contain any of those.

So why will the disc interact in the same way with the volleyball or the paint whether it’s moving or not? We don’t have a true theory of physics for homogeneous matter, so we have to think about it in an intuitive way, or at best in a Newtonian way. Essentially the idea is that the disc is perfectly smooth, and that means it won’t have any friction and won’t exert an angular force on the volleyball. It’ll just spin smoothly underneath it. The paint is a bit harder to picture, but I’m not sure the paint will properly stick to the disc. If the disc is smooth it’ll exert no angular force on the paint and just spin underneath it without affecting the circular puddle that forms. At least, I don’t know that that’s wrong.

I’m not sure how to make the case that the disc would behave the same way whether or not it was spinning if you drilled a hole in it, but I can believe there’s a case to be made. Perhaps a spinning disc would just flow around the drillbit. I don’t have a brilliant grip on the difference between a homogeneous liquid and a homogeneous solid. There’s an intuitive way of thinking about it, where the liquid disc would flow around the drillbit and the solid disc would get an arc-shaped hole in it, but that doesn’t fit well with the Newtonian picture in which under the microscope all matter is in the same state of being composed of tiny indestructible billiard balls. Would a piece of homogeneous matter behave like a giant impenetrable indestructible billiard ball, or more like a gas? I don’t know how to settle a question like that. I’m suggesting that to get homogeneous matter to behave anything like normal matter behaves we may already need the non-supervenient properties or relations, like non-supervenient frictional properties of smooth surfaces and cohesion properties of homogeneous substances. If that’s the case, it isn’t that the counterfactuals are there and the Humean can’t explain them; the counterfactuals aren’t there in the first place unless the non-supervenient properties and relations are introduced. This is because inhomogeneities aren't just the things the move; they're also the things that act. The Humean can say that non-particulate matter would have to be metaphysically weird to behave anything like normal matter does, so pieces of homogeneous matter would have to behave like big particles, and particles don't internally move. I suspect that showing whether or not they'd be right would take some hard work.

Counterpart theory

The other day I was trying to explain to another philosopher what counterpart theory was in terms which didn’t make me sound like a maniac for sympathising with it, and it wasn’t that easy.

I said that according to a counterpart theorist, she could have been a chef because the world could have contained somebody who was very like her, except she was a chef. (I glossed over technicalities about multiple counterparts and multiple good candidates.) This contrasts with a transworld identity theorist, who says that she could have been a chef because the world could have had her in it being a chef. Similarity is vague and context dependent and identity isn’t, so counterpart theory makes de re modal predication vague and context dependent, and transworld identity theory doesn’t.

I think this is more or less right, but there’s a problem with it. In my explanation of counterpart theory I talked about how the world could have been, and taken at face value that’s a de re modal predication of the world. Now, some people don’t think the world exists, like compositional nihilists, and for them I suppose there’s a difference between a de dicto modal statement and a de re modal predication of the world. Either way you’re having to talk about ways things could have been. If we give a counterpart-theoretic gloss of how I could have been, why don’t we give the same sort of gloss of the way things in general could have been? And if we take ways the things in general could have been as primitive, why not do the same with ways I could have been?

Well, if you’re a modal reductionist you’ve got an answer to this. David Lewis had a bunch of spatiotemporally connected mereological sums representing ways things could have been. Yagisawa had the same things as Lewis but let you cut them up however you like, pretty much making the accessibility relation a counterpart relation. Ersatzists had abstract possible worlds. Once you’ve got ways things could have been you’re up and running and can give your counterpart-theoretic account of who could have been a chef. This keeps the reductive isthmus small, just like Mark Schroeder tells us to.

But what about me? I don’t want to sound like a maniac and I don’t see how ersatz worlds could bear much explanatory weight, so I’ve got nothing to reduce ways things could have been to. I need to take them as primitive, and that means I can’t be a counterpart theorist all the way down.

Perhaps this isn’t so bad. It’s pretty much what Ted Sider does in ‘the ersatz pluriverse’, taking overall possibilities as primitive and doing the rest with counterparts. What intrigues me is whether we’ve any reason to think, after letting in primitive possibilities and necessities for the world, that there aren’t primitive possibilities and necessities for me as well. I guess what I’m pushing is a kind of essentialism: there are some context-independent facts about possibility given by the essences of things, and the rest is counterparts. It’s inspired by a footnote to Naming and Necessity, though I suppose Kripke would be upset if he knew.