In his earlier post, Trent raised several issues — some about the credit view, some about the epistemology of testimony, and some about Jennifer Lackey’s much-discussed testimonial case against the credit view. I will try to take these up in order. (I am taking this up in a new post, rather than commenting on Trent’s post, since for some reason it seems that the “comment” feature is no longer open regarding that post.)
One: Lackey’s case against the Credit Theory of Knowledge. If my memory serves, I think two things need to be separated out here. First, the case against the credit view (which, if I understand it properly, does not crucially depend on testimony cases); and second, what the Sears’ Tower case shows about credit in (some) testimony cases. Regarding the first, I seem to remember that Jennifer (in conversation; perhaps also in print) has urged that credit cannot distinguish Gettiered JTB from knowledge – in which case it is false that knowledge is true belief for which the subject deserves (the lion’s share of the) credit. (An argument of this sort is perfectly independent of testimonial considerations.) What testimonial considerations were supposed to show – second point – was that even in some non-Gettier cases it is false that knowledge is true belief for which the subject deserves (the lion’s share of the) credit. I didn’t read Wayne’s paper, but if he writes as Trent suggested he did, I fear he didn’t have Jennifer’s views properly in his target. As Matthew noted in his comments on the thread, Jennifer explicitly acknowledges that there is some credit that belongs to the hearer in cases of testimonial knowledge: the hearer wouldn’t have believed just anyone after all, and she relied in belief-formation on various of her own competences (with English; in being alive to signs of incompetence or insincerity; etc.). Lackey’s point was that these things are not the main explanation for why she got it right; she got it right because the speaker knew and communicated this knowledge to her.
Two: I take it that Trent and Wayne (in his discussion of Jennifer’s paper) disagree that there is knowledge in the case Lackey describes. I’m with Lackey on this one. There might be cases like this where there is no knowledge, only actionable belief. But it is worth noting that in the case Jennifer describes, the belief has all of the modal properties standardly associated with knowledge. 1. The belief is safe. Not easily would the hearer have formed a false belief in the way that she did: the speaker herself would not have spoken falsely, and given the well-known whereabouts of the Sears’ Tower and the general veracity of Chicagoans (!) it is highly unlikely that the hearer would have encountered false testimony on the matter. 2. The belief is sensitive: given the general veracity of Chicagoans, false testimony would most likely have come from someone who was either drunk or otherwise clearly incompetent, in which case the hearer wouldn’t have believed them. 3. The belief is reliable: the hearer is generally very good at distinguishing reliable from unreliable testimonies, and accepts only those she regards as reliable. These are externalist virtues, of course, and I readily acknowledge that not everyone will regard these as sufficient for knowledge. Still, I think they do give some independent support to the idea that there is knowledge in this case.
Three: when it comes to thinking about the epistemology of testimony, trust (at least in the sense that it is often discussed in the testimony literature) may be a side issue. The real issue is the one Trent noted under the label “testimonial support”. The question as I see it is whether (and if so in what ways) the hearer depends on the epistemic good-making features of the observed testimony for the epistemic status of her testimony-based belief. Virtually everyone should agree that the hearer so depends at the level of knowledge. Indeed, even those who deny that testimony always merely preserves knowledge, never generates it — people like Jennifer Lackey, Peter Graham, and others — should (and do) agree that the hearer knows through testimony only if the testimony is suitably reliable. (Turns out to be hard to say what “suitably reliable” means here; I’ve tried my hand in several places.) More difficult is the question whether the DOXASTIC JUSTIFICATION of testimonial belief depends on the epistemic good-making features of the testimony observed. I assume that any “evidentialist” view Trent favors would answer in the negative; justification depends only on the goodness of the evidence that the hearer has for thinking that the testimony is reliable, not on the actual reliability of the testimony. I for one think that this is not right; though I admit that this matter is complicated. But in a nutshell (and I think this is what Matthew was getting at in his comments on trust) I think that a view like the standard evidentialist view fails to appreciate the extent of our epistemic reliance on the speaker in testimony cases. I don’t just depend on myself, and my evidence, when I accept your say-so; I also depend on you (your say-so) when I take your word for something. (Although I have many bones to pick with the so-called “assurance view” of testimony, this isn’t one of them.) In particular, I think that in taking someone’s say-so my epistemic dependence on her is substantial – and certainly more substantial than anything a traditional “evidentialist” view will be able to acknowledge. But that’s a longer story — indeed, one might even try writing a book on it…