Friday, October 10, 2014

Introspective and Reflective Distinguishability

Mooreans, including neo-Mooreans, think that we know lots of ordinary stuff, and that we also—maybe on this basis—know the denials of extraordinary skeptical scenarios. Duncan Pritchard defends a particular disjunctivist brand of neo-Mooreanism, according to which, in cases of successful perception, one has reflective access to factive reasons of the form I see that p, and perceptual knowledge based on such reasons. So for instance, when one looks at red wall under ordinary circumstances:

  • One sees that the wall is red.
  • One has reflective access to the fact that one sees that the wall is red.
  • One knows that the wall is red on the basis of the fact that one sees that the wall is red.
Since Duncan also accepts a closure principle on knowledge, he accepts:
  • One knows that the wall isn't a white wall illuminated by red light.
Like all forms of Mooreanism, Duncan's view is in tension with certain skeptical intuitions. For example, it is in tension with this intuition:
(S) One can't tell by introspection that one is faced with a red wall rather than a white wall with red light.
As Duncan puts it,
If, in the non-deceived case, one has reflective access to the relevant factive reason as epistemological disjunctivism maintains, then why doesn't it follow that one can introspectively distinguish between the non-deceived and deceived cases after all, contrary to intuition? ... In short, the problem is that it is difficult to see how epistemological disjunctivism can square its claim that the reflectively accessible reasons in support of one's perceptual knowledge can nonetheless be factive with the undeniable truth that there can be pairs of cases like that just described [ordinary perceptual cases and corresponding deceptions] which are introspectively indistinguishable. (21)
(Duncan defines 'introspective indistinguishability' as the inability to know by introspection alone that the cases are distinct. (p. 53))

If I wanted to be a neo-Moorean of broadly Duncan's style (something I might well want to do), I'd just deny S, along with the many other skeptical intuitions that come out false on this view. But Duncan doesn't want to go that way; as this passage indicates, he considers S and claims like it to be 'undeniable truths'. (On p. 92 he even says that disjunctivists in particular are "unavoidably committed to denying that agents can introspectively distinguish" between the relevant cases.) I confess I don't see why it's so important to hold on to this particular skeptical intuition while happily rejecting others, such as the intuition that an ordinary person at the zoo doesn't know that she isn't looking at a cleverly disguised mule.

How does Duncan go about resolving the tension between his disjunctivism and S? By leaning on the 'by introspection' qualifier. He does think that, if one in the good case, one can reason thus, resulting in knowledge of the conclusion: "I have factive reason R. Only in the good case would I have factive reason R. Therefore, I'm in the good case." But, he says, this is consistent with intuitions like S, which are about introspective abilities. And while one may be able to tell by introspection what reasons one has, one cannot tell by introspection that factive reasons obtain only in the good cases. This is something one can come to know by a priori reflection, but not by introspection. (And maybe the same goes for the epistemic standing of the inference from the two premises to the conclusion.)

This is ultimately a much milder concession to skeptical intuitions than at first it appeared. Although he preserves the letter of his interpretation of the claim that we can't introspectively distinguish the good cases from the bad cases, he does so by pointing out that "introspectively" is a stronger qualifier than one might have realised. He does think (p. 95) that one can reflectively distinguish between good and bad cases, where reflective distinguishability is the ability to know distinct base on a combination of introspection and a priori reasoning.

So two thoughts. First the smaller one: is it really right to exclude a priori reasoning from the considerations that establish 'introspective distinguishability'? It's very hard for me to even make sense of just what that constraint is. (In The Rules of Thought, Ben and I argue that we can't divorce any kind of thought from a priori reasoning.) Consider these two cases: (1) I am presented in ordinary circumstances with a blue ball. (2) I am presented in ordinary circumstances with a black ball. Given the way my perceptual faculties work, we should consider these cases to be distinguishable in the relevant sense if any are. But is it clear that I can know them to be distinct without using a priori reasoning? It's not like the proposition that they're distinct is made available to me directly via introspection. Instead, I have introspective access to how one case looks, and to how another case looks, and I observe that they're different. From this I infer, using something like Leibniz's law, that they're distinct.

Second, supposing Duncan is right about introspective distinguishability: maybe this just shows that the worry wasn't properly articulated in the first place. I submit that someone motivated by the kinds of skeptical pressures that would drive someone to say that you can't tell good cases and bad cases apart by introspection, isn't going to feel better if you allow a priori reasoning along with introspection. The key skeptical intuition in the first place was just that it shouldn't be that easy to tell the good cases and the bad cases apart. And there's no getting around it: that's just an intuition that disjunctivists need to deny. Once we come to appreciate this fact, I'm not sure how important it is to conform to the letter of certain idiosyncratic statements of the intuition.


  1. Duncan Pritchard10/12/2014 02:41:00 PM

    Many thanks for the comments Jonathan (and apologies for the delay in responding--I haven't been on-line much the last few days). I think there's a key element of my view which you're missing out here, which is the distinction I draw between favouring and discriminating epistemic support (a distinction which I argue we should all accept--i.e., it's meant to be independent of ED). The idea is that with this distinction in play we can come to see how a further distinction can be drawn between a narrow sense of 'reflection' in which we clearly cannot reflectively tell the difference between good and bad cases (lack of discriminating epistemic support of the right kind), and a broad sense of 'reflection' by which we can come to reflectively know the difference (via favouring epistemic support). The point is that with this distinction in play we can resist the sceptical inference from the fact that there is nothing in my current experiences which distinguishes them from the sceptical case to the (I claim stronger) claim that there is no reflective basis available to me (even in the good case) which epistemically favours my being in the good case rather than the bad case. If we deny the latter, then conceding the former is (I claim) harmless. That's the general idea anyway. (And note that all I need is for the distinction to be genuine. I'm sure you're quite right that there are certain particular cases where the distinction becomes blurred and perhaps even artificial for various reasons. But I just need the general shape of the distinction and its application to the sceptical reasoning).

  2. Thanks Duncan.

    I am still struggling to understand the difference between what you call favouring epistemic support and discriminating epistemic support. I agree that there is difference between having some support for a hypothesis and having an ability to discriminate it perceptually from other hypotheses, but I haven't quite convinced myself that this motivates a dramatic difference in kinds of epistemic support.

    I wasn't aware you were thinking of the kinds of epistemic support as corresponding to different senses of 'reflection'. Is that discussed in the book?

  3. Hi Jonathan,

    Yes, I talk about this towards the end of part two, though the terminology is slightly different. I contrast a narrow introspective version of what you're calling 'reflection', with a broader category which includes both introspection and a priori reasoning (which I call 'reflective'). With the different terminologies in play, it might be best to characterise this distinction as 'narrow reflection' versus 'broad reflection', or something like that. The thought is that while the sceptic is quite right to say that we can't reflectively distinguish between the good and bad cases in the narrow sense (because there's a lack of discriminating epistemic support), they are wrong to conclude on this basis that one has no reflective grounds available to one in the good case which indicate that one is in the good case (because according to ED one has the favouring factive epistemic support, which entails that one is in the good case). In sloganising form: I'm claiming that there's a way of appealing to one's reflectively accessing grounds so as to know that one is in the good case (when one is in the good case) which is not thereby a way of introspectively telling the good and bad cases apart.

    1. So I'm happy to grant the narrow/broad reflection distinction. (At least for the purpose of argument; as I said in the body of the post, I'm suspicious of the idea of divorcing any capacity from a priori reasoning.) What I am not seeing is how this distinction, which concerns which capacities are made use of, is related to the distinction between favouring and discriminating support.

      Like I said in the post, I just have a hard time seeing how the kinds of internalist, skeptical intuitions that would lead someone to suppose we can't tell from the inside whether we're in good or bad cases, are plausibly read as ones specific to narrow reflection, defined as you do.

      This isn't a complaint about the approach to epistemology -- I have considerable sympathy with a lot of it. I'm not just convinced that it manages to be the kind of 'holy grail' of epistemology that does justice to deep-seated internalist intuitions.