What should a contrastivist say here?

I’ve been thinking about contrastivism lately, and just noticed an endnote (I think it’s number 34) in Jonathan Schaffer’s forthcoming piece in Phil Studies that I hadn’t noticed before. The topic of the note concerns enlarging our body of knowledge by deducing one claim from another. For contrastivists, the claim that Moore knows he has hands involves an implicit contrast, so that the proposition expressed is something perhaps to the effect that Moore knows he has hands rather than stumps.

Suppose Moore deduces from the claim that he has hands that he’s not a brain-in-a-vat. Schaffer says that’s OK, since his knowledge is now something like: knowing he is not a brain in a vat rather than having stumps in place of hands. He says that all it takes to know this is to be able to rule out the contrast claim: “Expand p [the principle that allows one to deduce not-BIV from having hands] does allow Moore to know that he is not a brain-in-a-vat rather than someone with stumps. But this is actually a very modest bit of knowledge, secured not by Moore’s possessing any sort of positive evidence for not being a brain-in-a-vat, but rather by Moore’s possessing conclusive negative evidence against having stumps (reminder: the stumps-possibility is restricted to worlds in which stumpedness would be apparent).”

Two points here. The first is the minimal standard given for knowing here: you know because you can rule out the contrast, even though you don’t have any evidence that you’re not a BIV. I don’t think Schaffer needed to say this, though. He could have said that Moore knows because he can rule out the contrast and deduction transmits evidence from the hands claim to the non-BIV claim.

The second is this. Shouldn’t contrastivism cite genuine contrasts in knowledge attributions? E.g., if “Joe knows that Texas is south of Canada,” involves some implicit contrast, I don’t see how a suitable theory would interpret it the missing parameter so as to yield “Joe knows that Texas is south of Canada rather than that North Dakota is north of Nebraska.” If I heard the latter claim, I’d interpret it as indicating that Joe knows the first claim and doesn’t know the second. When I hear a “rather than” knowledge attribution involving no logical contrast, it seems to me to have the form of a conjunction, not a three-place knowledge attribution of the sort contrastivists want us to focus on when we say things like, “Moore knows he has hands rather than stumps.”

You may be tempted to construe the case like this: Moore infers he’s not a BIV (understood to mean that all of his sensory modalities give him false beliefs) rather than having stumps (and being able to tell that this is so by looking). Schaffer notes parethentically in the passage quoted above that the stumps-possibility is restricted to worlds in which this is apparent, but that isn’t the same as building the restriction into the semantic content of the contrast. Instead, it only means that there are certain worlds epistemically relevant to whether you can rule out having stumps. Building it into the content gets the wrong results, since it makes the contrast harder to rule out (since it is now a conjunction).


Comments

What should a contrastivist say here? — 10 Comments

  1. What a nice opportunity for my first post! Let me take a shot at what the contrastivist might say here.

    1. In regard to Jon’s point about my minimal standard for knowledge, the contrastivist may require more than merely eliminating the alternatives. She may also require some positive evidence for p. Dretske, for instance, is ambivalent here, saying that the subject, “needs only a nominal justification, if he needs any justification at all…” for the unfocused aspect of the knowledge claim (1981, p. 373)

    2. In regard to Jon’s question of whether the contrasts need to be genuine, I think the answer is: yes. More precisely p and q need to be exclusive, in that the sets of worlds they denote needs to have an empty intersection. So one cannot know that Texas is south of Canada rather than that North Dakota is north of Nebraska. But one can know such things as that (i) Texas rather than Siberia is south of Canada, (ii) Texas is south rather than north of Canada, and (iii) Texas is south of Canada rather than Mexico. And these are different epistemic states. For instance, only (i) requires knowing where Siberia is, while only (iii) requires knowing where Mexico is.

    A more intuitive way to put all this may be with “knows whether” constructions. So one cannot know whether Texas is south of Canada or North Dakota is north of Nebraska, because the embedded wh-clause fails to denote a proper question. But one can know (i) whether Texas or Siberia is south of Canada, (ii) whether Texas is south or north of Canada, and (iii) whether Texas is south of Canada or Mexico. And the key contrastivist claim is still that (i)-(iii) are different epistemic states, differing in their contrasts.

    3. A very minor point on Jon’s final paragraph: I did mean the apparentness of the stumps to be built into the semantic content. In general, I take the denotations of that-clauses to be sets of possible worlds, within a contextually determined domain of quantification. And I meant to stipulate that I was using “that Moore has stumps” within a domain of worlds through which apparentness holds. Jon says that “Building it into the semantics gets the wrong results, since it makes the contrast harder to rule out.” Actually, it makes it easier. In general, the fewer the worlds in the contrast, the easier their elimination is. Apparent-stumps worlds are much easier to eliminate than not-necessarily-apparent-stumps worlds are.

  2. Jonathan, this is very helpful, and I especially like the clarification in terms of knowing whether. The issue that still bothers me is the modal profile of the contrast claim. I think you’re right that we need to be charitable when interpreting such remarks, but I think this approach is too strong. On this interpretation of the modal profile, we’ll have a hard time explaining the chagrin Moore would experience at finding out that he actually had stumps.

    Since possible cases of this sort are going to be too gruesome for me this early in the morning, change the case to the legendary lecture where Moore was reported to have listed a number of things that he knows, including pointing at a skylight and saying, ‘I know that the sun is shining through that skylight,’ or something like that. I take it the relevant contrast here would be (or could be) ‘rather than some artificial source of light.’ It turns out Moore was wrong; it was an artificial skylight with strong light bulbs shining through it.

    Now suppose we take the modal profile of the contrast here to involve not only the semantic content of the phrase used to pick it out, but also restrict the artificial light hypothesis to worlds where this is apparent. We now get the wrong result. At first blush, the artificial light hypothesis is not apparent at all; that’s why the story is legendary, and why Moore would say upon learning the truth, “I guess I was wrong in assuming that it wasn’t an artificial light up there.” If the modal profile includes the apparent world qualifier, Moore had ruled out that contrast.

    One might push harder here and say that it was apparent with just a little bit of checking, so his expression of chagrin was appropriate. But practical jokes cover a continuum of cases, from ones where a brief glance gives them away to ones that would take massive efforts to uncover. In all such cases, the subject of the joke is “gotten”; and their chagrin is an indicator of what the relevant contrast claim is. If ‘apparent’ comes to include all such modes of epistemic access, then we can construct cases like the artificial light case where the joke is that anything short of an NSF grant and a supercollider won’t reveal the joke. But Moore could still say correctly, “You got me!”

    In short, I think this way of building in the kind of epistemic access to the modal profile of the contrast doesn’t yield the right explanation of such practical joke cases. I think the relevance of what is apparent and what is not should be relegated to what it takes to “rule out” the contrast, not involve it in the modal profile of the contrast itself.

    I think the reason for putting it in the contrast is this: it has to contrast with the non-BIV hypothesis, and it won’t without this additional content. But even if we plug the hole in the dam in this case in this way, the general problem here won’t go away: what if Moore had deduced something exists from his having hands? No encroachment of epistemic access on the modal profile of the contrast claim will secure a legitimate contrast here. More generally, where p and q are exclusive of each other, p will have logical consequences that aren’t incompatible with q, and hence not with anything stronger than q either.

  3. Hi Jon (and Jonathan): Here is how I understand your main objection to Jonathan’s contrastivism.

    First, let C be the context in which Moore utters ‘I know that the sun is shining through that skylight’, but, unbeknownst to him, it is an artifical skylight with strong light bulbs shining through it. According to Jonathan, Moore’s utterance of the epistemic sentence in C involves an implicit reference to a contextually delimited contrast set of possible worlds – in this case, it’s the set W of worlds in which the artificial skylight hypothesis is both true and apparent to Moore. [W is, then, part of the semantic content of Moore’s utterance relative to C.] In uttering the epistemic sentence in C, Moore asserts that he knows it is the sun that is shining through the skylight, and thereby, he rules out W. When he is later told that he was the subject of a joke, Moore responds, ‘I guess I was wrong in assuming that it wasn’t an artificial light’. According to you, Jon, Moore’s response is appropriate but, on Jonathan’s view, Moore’s response is inappropriate. And so, given Jonathan’s contrastivism, one can’t explain the appropriateness of Moore’s response when he later learns, to his chagrin, that it was in fact an artificial skylight.

    So, why is Moore’s subsequent response inappropriate on Jonathan’s view? Well, by hypothesis, the actual world (relative to C) is not in W since the artificial light hypothesis is in fact true but Moore fails to believe it – I take it that the members of W are worlds that if they were actualized, Moore would have believed, correctly, the artificial light hypothesis. Thus, when Moore later says, ‘I guess I was wrong in assuming that it wasn’t an artificial skylight’, he’s wrong to say that (on Jonathan’s view, according to you) because the artificial light hypothesis wasn’t apparent to him at the time; in which case, it was right for him to assume, at the time of his epistemic utterance, that it wasn’t an artificial skylight. After all, it would have been unreasonable for Moore to investigate whether the artificial light hypothesis were true if he had no compelling evidence for thinking that it might be true.

    If that’s the argument, then it seems to me that Jonathan could reply as follows. The context changes once Moore is told that he is the subject of a joke: he now has information about C that he didn’t have before in C. So, Moore’s response may be appropriate relative to the new context C*, since its content will be determined by a different contrast set of worlds, one that is implicitly referenced in Moore’s response. Relative to the old context C, Moore’s response is inappropriate. I’m not sure how the details of this strategy might go, but I think that it’s a possible move for Jonathan to take.

  4. Hi, Ray–I’d like to hear what Jonathan says about your suggestion, but my prediction is he won’t want to go this way. He’s intent to keep contrastivism from being a form of contextualism, and this response looks like it won’t preserve that point.

  5. Hi Jon and Ray,

    1. I would say that Moore’s chagrin is simply due to p being false. The contrast doesn’t really play a role here (so no context-shifting maneuvers are needed here). Whether Moore claims to know that p: the sun is shining through the skylight, rather than q1: artificial light is shining through the skylight, or rather than q2: the moon is shining through the skylight, or rather than q3: the sun is shining through the floorboards, p remains false. And since Kspq requires that p is true, this is cause for chagrin.

    2. There is still Jon’s point that p will have logical consequences that are no longer incompatible with q. For instance, p entails (p or q), which is compatible with q. Here expand-p needs restriction. The restriction should be that what is entailed remains incompatible with q: the p2-worlds and the q-worlds are disjoint.

    So can Moore still come to know whether something exists, from knowing that he has hands rather than stumps? I would say: no. Moore’s having hands rather than stumps presupposes Moore’s existence (and ipso facto presupposes the existence of something) and so cannot be used to prove it. But Moore can come to know whether something exists, from expanding on his knowledge that his hands exist, rather than nothing.

  6. Hi Jonathan: Thanks for your reply.

    [1] Can Moore really know that his hands exist rather than that nothing at all exists? If I got you right, the contrast set will be a contextually delimited set of possible worlds in which nothing exists. You also require that contrast sets contain only worlds that are epistemically apparent to the subject (relative to the context in question). Suppose that Moore knows his hands exist rather than that nothing exists. Then, he excludes worlds that are such that, if they had obtained, it would have been apparent to him that nothing at all exists — including himself. Given the context, we hold fix the fact that Moore is rational, reflective, that he can reason properly, etc. So, my question is whether there are such worlds given those constraints. For, if there aren’t any, then there aren’t any for him to rule out; in which case, he doesn’t know that his hands exist rather than that nothing exists. Clearly, there are worlds in which nothing at all exists (or at least, worlds in which no contingent being exists – but I’m waiving that issue aside). Moore would (or could) admit that there are worlds such that, if they had obtained, nothing would have existed, and thus, there are worlds that, if they had been actual, he would not have existed. However, that isn’t the same as admitting that there are worlds such that, if they had obtained, it would have then been apparent to Moore that he and nothing else existed. My gut intuition is that there are no such worlds, but I have no argument to back it up. No doubt, a scope-ambiguity arises when the construction, ‘it is apparent to S that P’, is embedded in the consequent of a counterfactual. So maybe that’s what is tripping me up here. Perhaps you can clarify things for me if that’s the problem.

    [2] Given your constraint on your ‘expand p’ principle, Moore can’t come to know that he has at least one hand from knowing that he has two hands rather than just one. True, he could come to know the former, by way of the ‘expand p’ principle, from knowing that he has hands rather than not having any. Still, if Moore can do simple logical inferences, it should be a cinch for him to come to know that he has at least one hand on the basis of his knowing that he has two rather than just one, assuming that he knowingly deduces, from his contrastive knowledge, that he has two hands. If that’s right, then your constraint on your ‘expand p’ principle might be too stringent.

  7. Jonathan, on the Moore example, I should have been more explicit. When moore says he knows that the room is illuminated by natural rather than artificial lights, we can remonstrate with him in either of two ways. We can claim that he’s not entitled to the knowledge claim (one way this can happen is that the claim is false, as in this example) or that he hasn’t ruled out the contrast. I was assuming that we ought to say here that he hasn’t ruled out the contrast either. Now that I think about it, maybe he has ruled it out, since he’s got perceptual evidence that it is false. But what I was thinking was that, just as he’s not entitled to the knowledge claim because it’s false, so he hasn’t ruled out the contrast claim because it’s true. If that sounds plausible, then we don’t want to restrict the modal profile of the contrast claim in the way you suggest, since if we do, we’ll have to say that Moore has ruled it out because it’s not apparent. I guess, in general, I’d rather see mode of access to the contrast characterizing what it takes to rule out a claim rather than being built into the contrast claim itself.

    The other point is that if you restrict Expand p in the way you suggest, you’re going to block knowledge by deduction in unacceptable ways. If I know your pet is a dog rather than a cat, then given your restriction, I can’t learn by deduction that your pet is alive, that your pet isn’t a horse, or an alligator, or even a rock. All of these things are things I can learn by deduction from knowing that your pet is a dog rather than a cat. These are more examples of the same kind as Ray’s.

    I’m also puzzled by your response to the ‘something exists’ example. It seems to me that you are suggesting even more restrictions on Expand p than merely that it must remain incompatible with q? I think you are suggesting more restrictions since presuppositions of p can nonetheless be incompatible with claims that p itself is incompatible with. I’m also uncertain about the notion of presupposition here. I suppose I knew I existed before I knew I had hands, but I don’t see why I couldn’t come to know that I exist by seeing my hands (though it’s not obvious to me that the order of knowing can go in this way, either).

  8. Ray:

    On (1), I think we may have had a misunderstanding. I do not require that contrast sets contain only worlds that are epistemically apparent to the subject. So being a brain-in-a-vat can be a contrast, such as in â??Moore does not know that he has hands rather than being a BIVâ?? (or: â??Moore does not know whether he has hands or is a BIVâ??) In general, the worlds in the contrast set are a function of (i) the content of the â??rather thanâ??-clause, and (ii) the domain of quantification. So non-apparent contrasts are permissible, if the domain permits. All I meant to be stipulating in my paper was that my particular discussion of the claim: â??Moore knows that he has hands rather than stumpsâ??, was to be understood in such a way that non-apparent stumps were not in the domain. Sorry if I was unclear earlier.

    Anyway, now there should be no problem with eliminating the nothing-exists possibility. If we plug in, for instance, the Lewis definition of elimination (possibility w is eliminated for s (at t) iff w is inconsistent with sâ??s total experience e (at t), weâ??ll get the result that any experience for s eliminates the nothing possibility. Which seems fairly intuitive.

    On (2), I disagree that it should be a cinch for Moore to come to know that he has at least one hand, on the basis of his knowing that he has two hands rather than one. Imagine how Moore might come to know that he has two hands rather than one. Perhaps he has already looked at his one hand (â??Here is one handâ?¦â??), and then turns and exclaims: â??and here is another!â?? Here Moore may be presupposing that he is embodied, perceiving veridically, etc. In this way of coming to know that he has two hands rather than one, Moore has done nothing to establish that he has at least one hand, rather than being a brain-in-a-vat, or an amputee dreaming, etc. Would you agree that knowledge requires discrimination? If so, I would put the point as follows: Moore might be able to discriminate between the two-hands scenario and the one-hand scenario, without necessarily being able to discriminate between the at-least-one-hand scenario and certain no-hands scenarios. Moore is presupposing that he has at least one hand, he has not proved it.

    Jon:

    On the Moore example, you say we can scold Moore on two counts here: claiming to know what is false, and failing to rule out the contrast. I guess I just donâ??t have this intuition. Perhaps you can say more to motivate it? In any case, if the contrast is true then (i) it cannot be ruled out by definition, and (ii) p must then be false. So it seems to me that when you say â??he hasnâ??t ruled out the contrast claim because itâ??s trueâ??, that this need not be a separate failure from claiming to know what is false. Which is a failure that the contrastivist can explain.

    On the restrictions to expand-p, I have the same sort of response here as to Rayâ??s second point above. I can know that Fido is a dog rather than a cat by some very simple means, such as seeing Fidoâ??s snout (given the presupposition that cats do not have protruding snouts), hearing Fidoâ??s bark (given the presupposition that cats do not bark), etc. Here I am presupposing that Fido is either a cat or a dog (and ipso facto not a horse, or an alligator, or a mineral). But I have done nothing to establish that Fido is an animal rather than a mineral. For instance, I have done nothing to rule out the possibility that Fido is a cleverly designed robo-dog, and hence mineral rather than animal. I am presupposing that Fido is either a dog or a cat, I have not proved it.

  9. Jonathan, very good, I think we are on exactly the same page here now. So here’s how I see your response: you want to restrict Expand to cases where the contrast is preserved, and you want to use the distinction between presupposition and proof to explain the restriction.

    I think the distinction is useful in pointing out how to formulate the closure principle, but I also think that it will allow more latitude in the closure principle than you’re suggesting (though I also think that the greater latitude can still secure the antiskeptical virtues that you wanted to preserve with the original closure principle).

    The details are a little messy for me at this point, since the notion of presupposition is not a precise one. For example, what I presuppose in a given context might be somewhat a psychological fact about what I attend to and what I don’t. In this sense, I might presuppose that my wife is female in a given context of inquiry even though I also know this to be the case.

    It is in this partially psychological sense that your response is perfectly fair about the dog/cat case. If we limit presuppositions to what we don’t know to be true, however, we get a different answer here, for I might not attend to elements of my background knowledge in a particular context of inquiry. In such a context of inquiry, I might conclude by observation that your pet is a dog rather than a cat, I might know this to be true and not be presupposing that your pet is either a dog or a cat.

    That’s the messy part about presuppositions, but here’s the easier part, I think. Let’s take presuppositions to be partly psychological, so that you can presuppose in a given context information that is nonetheless part of your background knowledge. Then we should be able to appeal to your background knowledge to justify weakening the contrast clause as well. If I know that your pet is either a dog or a cat rather than any other of the pet-ideas I’ve ever heard of, then even though my context of inquiry makes salient only the dog/cat contrast, when I infer from the dog claim that you don’t have a pet rock, I have available in my background knowledge a suitably weakened contrast to your not having a pet rock–you don’t have a pet rock as opposed to having as a pet some other pet-idea I’ve already heard of, one of which is a pet rock.

    So, a first start on allowing weakening of the contrast claim is by appeal to background information in a particular context of inquiry. We might want further restrictions on when background knowledge can be used (maybe in the process of the deduction, the background info has to become salient in some sense–this is what I’m working on to talk about in Aarhus in February), but the general point is that contrastivism can still be antiskeptical without requiring that the contrast never be weakened.

    This move will allow one to be less restrictive as well about expanding the known claim. Whatever combination of strengthening/weakening of the known/contrast is used, contrastivity will still be preserved and without sacrificing the antiskeptical idea central to the view.

  10. Jon, very good, I am glad we have come to a common understanding here.

    Let me just say a bit more about the notion of presupposition. What I have in mind is the Stalnaker picture on which a context is a set of alternatives, the ‘live options’ under discussion. What is presupposed is that actuality is one of the ‘live options’.

    This would count for you as a psychological sense of presupposition, I think. It is certainly not a sense in which what is presupposed = what is known true. In the one direction, we might presuppose that p without knowing p to be true, by stipulating p, or even by ignoring the possibility of ~p. We might presuppose that p even when p is known to be false, in the scope of a counterfactual antecedent. In the other direction, we might know p to be true without presupposing it, as happens when we have just proved that p.

    I agree with the rest of what you say then, about how the contrast can be expanded or replaced in certain cases, given this sense of presupposition.

    I’m looking forward to hearing your talk in Aarhus.

Leave a Reply

Your email address will not be published. Required fields are marked *