Brian Leiter recently reposted a post from last October on “The Relevance of Motives, or the Hermeneutics of Suspicion, or Ricoeur Meets Gettier”:http://webapp.utexas.edu/blogs/archives/bleiter/001520.html. I don’t really know that much about motives, hermeneutics, suspicion or Ricoeur, but I have a few opinions about Gettier so I thought I’d respond to those bits in his post. Brian claims, following Phillip Kitcher, that the lesson of the Gettier cases is that for a belief to constitute knowledge it must have the right kind of causal history. I think that if that’s the lesson of the Gettier cases, then they are much more conducive to scepticism than is usually appreciated. (I think Gettier intuitions just are sceptical intuitions, but no one believes me on this, so here I’ll just defend a conditional version of that claim.)

The following discussion owes _a lot_ to John Hawthorne. Most of the example is basically his, though I doubt he likes the way I use it. I haven’t got the book in front of me, though I think the example is covered in his book _Knowledge and Lotteries_. (Note that by blog standards that’s an adequate citation!)

Consider a standard Gettier case, in fact Russell’s version of a Gettier case. Alice sees the clock on the green and it says 4.23. So she sets her clock (in fact the clock on her cell phone) to 4.23. The clock on the green is right, as it usually is, but right now it’s stopped, as it has been for exactly 12 hours now. So Alice has a justified true belief that it’s 4.23, but according to most philosophers she doesn’t know this.

(Even if the clock was working Alice wouldn’t know it was 4.23, since the margin of error of the clock on the green is probably a minute or two even when it’s working. Strictly speaking, she would only know it’s between 4.20 and 4.25. I’ll suppress this consideration throughout in what follows, because I don’t think it’s important, and it is potentially misleading.)

Now for the Hawthorne variations.

A minute later Bob asks Alice what the time is, and she looks at her phone and (truthfully) reports that it’s 4.24. So Bob forms the justified true belief that it’s 4.24, and sets his watch accordingly. Does he know this is the time?

A minute later Carol asks Bob what the time is, and he looks at his watch and (truthfully) reports that it’s 4.25. So Carol forms the justified true belief that it’s 4.25, and sets her PalmPilot accordingly. Does she know this is the time?

[You can see how this is going]

A minute later Zinedine asks Yael what the time is, and she looks at her iPod and (truthfully) reports that it’s 4.48. So Zinedine forms the justified true belief that it’s 4.48, and sets his laptop’s clock accordingly. Does he know this is the time?

I don’t know what it is for a belief to have the right causal history, but it seems to me that if Alice’s has the wrong causal history, then so does Bob’s and Carol’s and … and Yael’s and Zinedine’s. But it seems to me (and this is just a straight intuition report though I’ll say more about it below) that it would be a very sceptical attitude to regard Zinedine as not knowing what the time was in this situation.

On the other hand, we can push the intuition that Zinedine does not know. After talking to Bob Alice took a small nap on her office couch, and then went out of her office leaving her phone behind. She runs into Zinedine a minute after he spoke to Yael and asks him what the time is. (Alice doesn’t nap often, so she has really very little idea how long she’s been asleep.) Zinedine checks his computer and says it’s 4.49. Alice forms the justified true belief that it’s 4.49. Does she know this?

If Alice can get knowledge this way, it’s hard to see why she wouldn’t know what the time was just by checking her phone. After all, Zinedine’s computer is just tracking the time according to Alice’s phone. Well, maybe Alice _could_ find out the time when she wakes up by looking at her phone.

You can probably guess where the case is going now. Change the case so that Alice naps for only a few minutes, and wakes up at 4.28. Does she know what time it is then? Or change the case so that she (a) forgets the time as soon as she talks to Bob and (b) accidentally drops her phone into a nearby gorge. She quickly asks Bob the time, and he says it’s 4.24. Does she know now what the time is? If so stopped clocks can really quickly lead to knowledge.

There are a ton of variations on these cases, but I’ll spare you them. Instead I’ll just return to Zinedine’s case and note there are three options we can take.

First, we can be sceptical and say he doesn’t know what time it is. I think this is quite a sceptical attitude because it’s plausible, highly plausible in fact, that most things we take ourselves to know are grounded _somewhere_ in the distant past on something that’s at best an accidental truth. So this could easily lead to a generalised scepticism.

Second, we can say that although Alice didn’t know what the time was when she looked at the clock what time it was, Zinedine does know the time. The error gets ‘washed out’ as it were by passing through all these hands. There’s two non-problems with this position and one real problem.

The first non-problem is that it means somewhere X gets knowledge of p by testimony from Y even though Y doesn’t know that p. Some argue that violates a constitutive property of testimonial knowledge, but I think the arguments put forward by “Jennifer Lackey”:http://www.soci.niu.edu/~phildept/faculty/Lackey.html to the effect that testimony can be a generative epistemic source are convincing, so I don’t think there’s a special problem here.

The second non-problem is that there might look like being a slippery slope from Zinedine knowing what the time is to Carol or even Bob or even Alice knowing what the time is. But it is easy to say that this is just a regular Sorites series with a vague boundary.

The real problem is that it makes it suspiciously easy to get knowledge. If Alice truly justifiably believes that p, she might not know it. But if she tells lots of her friends that p, and they believe her and tell lots of their friends and so on, and eventually word gets back to Alice that p, she then comes to know that p. This seems unfortunate. (I don’t think this kind of consideration undermines Lackey’s argument, though I don’t have her paper in front of me so I can’t check.)

Note that even if this is also a non-problem, it suggests we have to at best qualify Kitcher’s claim that knowledge requires the right kind of causal history. For on its most natural reading, that would imply Zinedine does not know that p. There might be a technical sense of ‘right causal history’ on which his belief has the right causal history, but it isn’t the most natural one.

So the third option is simply to say that Alice knows what the time is. This is counterintuitive, but I think the intuitions here are sceptical intuitions here we are better off without. As far as I can tell, this position has only two costs, and they are significantly smaller than the costs of the other options.

The first cost is that it really is very counter-intuitive to say that Alice comes to know what time it is from looking at the stopped clock. But it’s even more counter-intuitive I think to say that Zinedine doesn’t know, or that knowledge is generated through this chain of questions-and-answers.

The second cost is that we could potentially give the same kind of argument for the claim that knowledge doesn’t even require justification. It’s very tempting to try and outsmart that argument, but I don’t think I’ll try that here.