Another Knowledge Thought Experiment

Most intuition checks in epistemology involve made-up cases rather than found cases. The lottery cases may be an exception, but I think we should have a few more. Here’s my current favourite.

Consider Hamlet as he stands at the end of Act I. Does he know, or if you like know, that his uncle murdered his father? Let’s review the case for and against.

On the pro side, he has testimonial evidence, and the provider of the testimony presumably has first-hand evidence of the murderous act. So there’s a causal chain from the act leading to Hamlet. And that should usually do.

On the other hand, and it’s a big other hand, it’s a ghost. This is not the standard kind of testimony we’re talking about.

There are a couple of textual points that seem relevant. Hamlet seems at least a little familiar with ghosts, which might matter if one thinks, for instance, that internal justification of the evidence matters. On the other hand, it isn’t really made clear in the text just how the King knows how he was murdered. As he says,

Thus was I, sleeping, by a brother’s hand
Of life, of crown, of queen, at once dispatch’d:

So maybe the causal chain here doesn’t stretch straight back to the actual killing, but to when (in the afterlife?) the King found out about his method of murder. So it’s all a bit messy.

And if one is a two-dimensionalist about justification it will turn out that Hamlet clearly does not have a justified belief, because lectures from ghosts are not actually a reliable source of evidence.

There’s a Williamsonian point coming behind all this. Possibly it’s a little hard to say whether Hamlet knows at the end of Act I whether his uncle murdered his father. Perhaps he does, perhaps he doesn’t. But the question of whether he does or not is not a distinctively philosophical question, nor even a distinctively conceptual question. We’re just asking what the state of play is at that stage of the story. It’s just the same kind of question as if we ask whether Laertes knows at the end of act IV that Hamlet killed his father. (Yes, I think, though I can sort of imagine an argument to the contrary, what with Claudius having a vested interest in having Laertes come to believe it was Hamlet.) There’s no generalised ground for scepticism about the judgement about Hamlet that doesn’t extend to scepticism about the judgement about Laertes. And that seems preposterous – very often we know exactly what characters in plays do and don’t know. Someone who wants to argue that we could not know, or even have justified beliefs about, whether Hamlet knows how his father died at the end of Act I has to appeal, in some way or other, to facts particular to the case. That’s not a lesson I’ve always followed in the past, and it undermines some of my pro-JTB arguments.

10 Replies to “Another Knowledge Thought Experiment”

  1. I should preface this comment by noting that I know nothing about epistemology; so excuse my ignorance of what’s going on here. But, it seems to me that this IS actually a conceptual question (namely, about the term “knowledge”), since there aren’t any facts of the matter to disagree about. We can, of course, think ourselves into Hamlet’s position and wonder what his thought processes are — but to ask whether his state of mind is properly called knowledge is, surely, both a philosophical and a conceptual question.

  2. Three quick points.

    1) We could say the same thing about Laertes’s case, and I think that would be a bad reason to think it was a conceptual question.

    2) There are too facts to disagree about, e.g. the (alleged) fact that Hamlet knows who killed his father.

    3) I made a point of not asking whether the word “knowledge” applied to Hamlet. That would be a verbal question. I just asked whether Hamlet knew who killed his father. To be sure, that’s (materially) equivalent to a question about the applicability of words. But so is the question of whether Ophelia drowned equivalent to the question of whether the word “drowned” applied to Ophelia, or whether Ophelia satisfied the concept DROWNED, and the question of whether Ophelia drowned isn’t a verbal or conceptual question.

    Having said all that, I don’t have an enormous stake in what we want to call philosophical or conceptual questions. What I do have a stake in is the following claim: There’s no generalised argument for scepticism about judgements like “At the end of act I, Hamlet knew who killed his father” that doesn’t extend to scepticism about judgements like “At the end of act IV, Laertes knew who killed his father”. And since scepticism about the latter is unreasonable, general arguments for judgement (or if you prefer intuition) scepticism have to fail.

  3. How are you using the word “scepticism”? Does it just mean “refusal to attribute knowledge”? ‘Cos it seems obvious that either someone who wants to argue that Hamlet doesn’t know would have to appeal to facts about the case—what else would she appeal to?—or to accept, on generalized grounds, general skepticism about judgments. And if she accepted general skepticism about judgments, then she’ll be skeptical about Laertes—she won’t accept the Moorean move “Of course Laertes knows.”

    So, in short, I’m puzzled.

  4. I mean scepticism about whether our judgements in the story amount to knowledge. So I’m interested in whether we know that Laertes knows who killed his father, or that this death caused Ophelia’s drowning, or that it was an immoral act on Hamlet’s part. I don’t particularly care in this instance whether the claim in the story is about knowledge.

    One might deny that we can know all these things, but that strikes me as rather implausible. That Polonius’s death caused Ophelia’s suicide seems as stable a premise as we’re likely to find in debating causation. Scepticism about those kinds of judgments strikes me as very unmotivated.

  5. OK, I wasn’t thinking about the metaknowledge claim. Two more questions:

    (1) If I say “We don’t know whether Hamlet knows because we can’t know how to apply the concept ‘knowledge’ [to Hamlet] in borderline cases,” does that count as citing facts about Hamlet’s case?

    (2) How is it that you think you’ve failed to take the lessons of these examples into account.

  6. (1) Sure – but we should have reasons for saying it is a borderline case.

    (2) I thought there could be such a thing as a general argument against the use of intutions about possible cases in argument. I never thought such arguments worked, but because I thought there could be such an argument, I thought that how we responded to such arguments could teach us quite a lot about when we should use intuitions in argument. If we can just take as a starting point the facts that Polonius’s death caused Ophelia’s suicide, and we know this, and we come to know this simply by judging it to be true when paying suitably careful attention to the story/case, at the very least there isn’t a need for the kind of theoretical response to challenges to intuitions I was considering. Hence we can’t put constraints on intuitions like “They should only be intuitions for which that kind of theoretical response makes sense.” So we can’t, for instance, write off Gettier intuitions because they don’t satisfy the constraints of the best response to challenges to the general legitimacy of intuitions, because all we need to do is point to Ophelia’s case to show we don’t need a systematic response to those challenges.

  7. Brian, I’m just not seeing how the argument is supposed to go — maybe at least in part because I don’t see what’s so special about the Ophelia case, as compared with any of a zillion other cases where it’s hard to see what could be wrong with the intuition in question. I’m taking your line here to be somewhat Moorean in flavor, i.e., “Here’s one good intuition. Therefore, at least some good intuitions exist.” Is there more to your argument than this?

    (Btw, that question isn’t meant to have the implication that, if there’s not more to it than that, it’s a bad argument. It’s probably a fine argument, and indeed I certainly wouldn’t want to attack either the premise or the conclusion. But it felt like there was something special about the Ophelia case here, and I still don’t see what it is.)

  8. Hmmm… does this leave intact the argument that knowledge is JTB because JTB is the natural concept that best fits our intuitions about knowledge? I always liked that one.

    The corresponding move wrt Polonius would be, “It certainly is intuitive that Laertes knows that Hamlet killed his father, but when we invoke [Theory X] which is the theory of knowledge that best explains our intuitions, [Theory X] says he doesn’t know—so intuitions are deceptive in this case.”

    Incidentally, a friend recently floated the idea that Polonius is really Hamlet’s father. Don’t think Hamlet was supposed to know.

  9. Jonathan,

    There’s nothing special about the Ophelia case, but that’s not quite the argument I have in mind. Or at least, that’s not determinately the argument I want to make. I want to leave open the question of whether facts about Ophelia could simply be part of our philosophical foundations. If this is true it wouldn’t so much be a case of a good intuition as a denial that we need to appeal to intuitions to justify claims about particular cases, because we don’t need further justification for the claim about Ophelia. This is the Williamson line that we shouldn’t think our epistemic justification ultimately rests on facts about our mental states, sometimes it rests on facts about actual or possible cases. I don’t have a clean definition of intuition though, so maybe saying all this just is to say that there are good intuitions.


    That argument needs at least one more premise. If there are Gettier cases that are just like the Ophelia case in that we simply know they are not knowledge, and we don’t have to appeal to intuitions (as opposed to facts of the case) to ground a premise that these are not knowledge, then the argument fails. I think there aren’t any such cases, but Williamson thinks there are, so there is some reason to think there are.

  10. I think any version of acceptable version of that argument would already have proved the extra premise. That is, that argument would show that the best account of knowledge is that Gettier cases count as knowledge—and a fortiori that there aren’t any Gettier cases such that we simply know they don’t count as knowledge. So appeal to facts of the case would never be enough to establish them as knowlege.

    But I suspect that I’m now just saying “I don’t believe Williamson’s argument,” and that I’d better go read Williamson’s paper to see what that actually is.

Comments are closed.