Actionable Information: Testimony Ain’t Always Knowledge

Like Wayne Riggs, I don’t think testimony always amounts to knowledge.  (I have in mind here Lackey’s “Easy Knowledge” objection to credit theories of knowledge.)  Often it does, when background evidence supports it well enough, when the testimony is part of a greater narrative which makes the hypothesis that they are telling the truth very much the best explanation of their assertion.

FREQUENTLY, testimony gives us enough evidence that, given the stakes, it is rational to act on the information.  I mean, I get off the bus in Manhattan, ask the first person I see where 30 Rock is, they give me some directions.  I might as well start walking that way, and if things don’t go well, then I can stop and ask someone else.  People are not, by nature, scientists.  To ask three people off the bat and compare their answers, generalizing over them perhaps, would be weird.  Not only that, but instinctively, it could be quite confusing, because we’re well aware that there are many roads to the same destination. So it wouldn’t be to unexpected if three people gave non-identical directions, and that is likely enough to be confusing to make it more utile to just go with the flow of the first person until need arises to do otherwise.

FURTHERMORE, it would be *rude* to ask someone and then just ask the very next person.  In normal contexts, that would convey to the first person that you didn’t trust them.  And as social animals we instinctively avoid this kind of totally un-necessary effrontery.  And to avoid this, we’d have to embark on some plan where we stand around waiting for the first person to get out of sight (and standing there might also convey lack of trust) before we accost someone else.  Messy.

SO, for multiple reasons, it seems overall best to just act on the information we are given–when there’s no obvious funny-business (and it’s evidentially relevant that we do in fact scan for this even if we’re not thinking much about it).

There’s more continuity here with cases of knowledge-conveying testimony than at first might appear.  For, typically, we are most consciously attentive to the testimony, not the background narrative in which it occurs (just as we focus on the match, not the oxygen or the powder).  But we are tacitly aware of that narrative and the testimony is “enriched” with it cognitively to yield varying degrees of positive epistemic status.  In the case of, say, grade school, we have, I think, knowledge.  That’s a rich narrative context.  But when some Manhattanite hipster gives me directions, I don’t have knowledge, I just have actionable information.

So there’s really no Problem of Easy Knowledge from testimony in Lackey’s case, for there’s no knowledge there at all.

PS – I stole the main part of the title of this post from the title of a paper by John Greco.  I haven’t seen the paper, but I hope it’s *just* close enough to what I say here.


Actionable Information: Testimony Ain’t Always Knowledge — 23 Comments

  1. LOL, fair enough: bad subtitle!

    Still, the case I mention in particular is one which received lots of attention, but I give reasons–ones that, as I understand it, go beyond what Wayne has said in reply in “Two Kinds of Easy Credit”–for thinking that that case is not a case of knowledge.

    Still, that case is so weak that it’s hard to imagine a weaker case–modulo no obvious funny-business. So it may well be that Lackey is committed to all (true) testimony being knowledge. Surely she’s written something about this.

    Myself, I can’t easily think of a case of weaker testimonial evidence in the absence of defeat. It’s hard to say sometimes where bad evidence leaves off and defeat begins, but, as I say, the friend of knowledge in her “man on the street” case, to make it perfectly obvious that they were not committed to the thesis that all (true) testimony was knowledge, would have to provide a weaker version which was not knowledge (and not just a case where the testimonial support is defeated by counter-evidence).

    Maybe Sandy’s lurking about here and can set me straight. I’d welcome that. I blog to learn.

    It could be that the idea of testimonial support–which clearly enough comes in degrees–needs a closer look, especially by non-reductionists. If my confusion lead to clarity about that, I’d be happy to be of service. Again, Sandy, you out there?

    As an evidentialist-reductionist, I’d be interested to know how testimonial non-reductionists think of the kind of epistemic support one gets from testimony. It’s always seemed like it must be a black box from the reductionist side.

    So though my point–unlike my subtitle–is really only aimed at one kind of case. Now that you mention it, it might be of more general interest after all.

  2. Trent,

    Goldberg likely won’t be online since it’s Rosh Hashana.

    I’m a little confused as to what the target of your post is supposed to be. I don’t think anyone, and certainly not Lackey, thinks all testimony results in knowledge. (Not even all (true) testimony.) As I recall from the paper, Lackey is arguing against credit being a necessary condition on knowledge. All she needs for that is a case of knowledge in which the agent is creditable. Now it may well be that we often act in the way that you describe when stepping off buses. However, are you claiming that we never can have knowledge in such cases? That seems far fetched right?

    There are a couple of diagnoses of what’s going on in the traveler case. Here is one. If you’re a non-reductivist, then the thing that’s doing the work in the traveler case is the trust and not any ability due to the agent. So the knowledge isn’t properly creditable to the agent.

  3. 1. I saw he left some comments on Facebook, but looking closer, I now see that those were yesterday.

    2. I’m quite sure she wouldn’t *assert* that. The worry was that she was *committed* to it, and the challenge was to provide a case of testimonial support that was weaker, not knowledge, and relevantly different from the man-on-the street case.

    3. I claim that it is far fetched that we have knowledge in that case and cases relevantly similar to it.

    4. I have virtually no idea what you could mean by “trust” in this context. I’ve read the Lehrer book and the Foley book and I still don’t get trust.

    However, it at least *sounds* like something that would be a skill of an agent. Some people have a good sense for that thing and others don’t: they’re gullible, they lack skill in knowing who to trust.

  4. Trent,

    Lackey has a whole book that details her view on testimony. It’s good philosophy, you’ll hate it, and it will be good for your soul. You’ll also see why she’s not committed to thinking all (true) testimony results in knowledge. I’ll leave her view aside and just focus on the Chicago traveler case.

    I think that denying knowledge in the Chicago traveler cases is a loser route for several reasons. One is that it just strikes me as obvious that the traveler knows where she’s going after she receives the testimony. To think that she doesn’t know you have to make some pretty uncharitable assumptions about what’s going on with her as an agent, which I think Riggs does in his paper. You can give the agent all kinds of discerning and defeater sensitive conditions, and the objection still goes through. Second, if you want to deny knowledge in this case, then, contra Riggs, I do think there are skeptical worries because there seem to be a lot of cases where we come to have knowledge in this way that you’re going to have to deny. Third, you have to show that it is impossible that there could even be cases like Chicago traveler, and that seems implausible. It could be that we often don’t want to be rude, that we sometimes get less than knowledge (I actually like to think of these things in terms of justification), that we sometimes just get actionable evidence, but that’s all compatible with us often getting knowledge in a lot of cases.

    (FWIW, I do think Riggs is right to focus on the cognitive abilities, mechanisms, etc, that an agent brings to bear in cases of testimonial knowledge. I’ve raised a similar line to Lackey in the past and no longer think it works.)

    I’ll be the first to admit that my standards for knowledge are probably lower than yours, but it isn’t the case that just anything gets through. I might say to you. Man if that’s knowledge we don’t know much of anything.

    You need Goldberg trust. I think it’s in his new book coming out with OUP. Trust in this case is a social phenomena, so it’s going to be external to the agent.

  5. My reaction has already emerged in the comments: testimony ain’t ever knowledge, though knowledge may sometimes result from it.

    I also think the traveller would know 30 Rock is after being told how to get there. Assuming the testimony is correct, and has been properly understood. The proof is the pudding on arrival.

    My sense it that a lot of the issues are resolved not in the evaluation of the testimony itself but in the decision to ask for directions. After getting off the bus, the traveller decides to ask someone who “looks to be from around here”, let’s say. And the first question is, “Do you know how to get to 30 Rock from here?” What we trust is not so much the testimony (though we trust that too) but the answer to that question. We trust that someone who can’t tell me how to get won’t try. “Sorry, I’m not from around here either,” they might say.

    Error is always possible in the usual way. The directions may be flawed or misunderstood. I may not arrive at 30 Rock by the route indicated. In that case, of course, I did not know where 30 Rock was (“from here”) after being given the testimony. But that’s only because it did not lead the formation of a true belief.

  6. 1. LOL, you’re probably right on both counts!

    2a. “It’s obvious to me” Alas, not much is obvious to me. And when doing philosophy, though I ultimately have nothing else to go on than what seems obvious, I take my takings with a large grain of kosher salt.

    2b. What uncharitable assumptions do you have in mind? I can’t think of any. It’s just that it seems like too weak evidence for knowledge, that’s all.

    2c. I don’t think widespread skepticism looms for the reasons I mention: much testimony comes in the context of a much richer narrative. It’s the lack of background here that distinguishes–in my view–knowledge-giving testimony from non-knowledge-giving testimony.

    2d. Why would I have to show it impossible? That I sure don’t see.

    3. LOL, I’ll stick to my high standards! But I think I know lots of things: I have two hands, I’m typing, there’s a nice, flowering tree outside my window up and to the right of my monitor, that I had toast for breakfast, that my kids love me, that if I scoot my mouse too far it will drop off the desk, that all the lottery tickets purchased during the time it took me to hit the ‘l’ key in typing “lottery” are loosers, that the Earth only has one moon, that there is a great wall in China, that I won’t die of a heart attack before the end of this sentence, what my phone number is, that I caught a large spider last night, that my dog’s name is Caesar, that I want to write more about the problem of evil but also don’t in a way, that I like to blog and think through things. And much, much more!

    4. Well, I’ll ask Sandy to explain it to me next month at the MEW. I’m actually all for it, I just don’t get it. Linda Zagzebski’s Locke Lectures (I think it is) are on trust too, and we’ve discussed it a bit. Still don’t get it though.

    PS – N.B. I’m still looking for a case of testimonial support NOT strong enough for knowledge (NOT one that is, but is just defeated).

  7. “The proof is the pudding on arrival.”

    Oh surely not. That’s getting pretty close to a true belief model of knowledge. You admitted you had low standards, but c’mon! 🙂

    When you say “we trust answers, not people” that–to my ear–just means we *rely* on the information, i.e. that it is actionable, which is just what I’m saying.

    When you use it as a propositional attitude “We trust that…” how is this any different from we *believe* that? Does trust entail belief? Ok, so you trust that: is the trust rational, is it based on adequate evidence. I mean, *somebody* (apparently about 30 souls) trust that nut-case clown of a Koran burner down in Fla. So trust can be more or less rational depending on whether the evidence supports it (whether as invested in information, persons, or propositions).

    Why do we need trust when we have belief? What does it add?

  8. Actually (I admit it) I hold to a “true belief” model of knowledge. I’m not sure what’s it called (probably “pragmatism”, though), but I normally let the standard objections to justified, true belief adjust (if need be) my definition of truth and belief (rather than knowledge).

    I don’t think I said “we trust answers, not people”. I meant we trust the answer to the question “Do you know?” before we trust the testimony itself. We assume that someone who doesn’t know won’t tell us how to get there.

    (But we know that error is possible. Someone might think they know, but have false beliefs about where 30 Rock is and pass those on to us … because we believe them.)

    Imagine the following sequence:

    1. Jones does not know where 30 Rock is.
    2. Jones asks Smith where 30 Rock is.
    3. Smith tells Jones it’s over there.
    4. Jones believes that 30 Rock is over there.
    5. Jones goes over there, and find 30 Rock, i.e., it’s where 30 Rock is.

    Now, I would say that Jones knows that 30 Rock is over there at 4 (because he believes it and it’s true), just as certainly as he knows that 30 Rock is right here at 5.

  9. I don’t think I’m making trust do any epistemological work here. Jones finds Smith trustworthy probably already before asking him for all kinds of prejudicial, irrational reasons. Nothing that happens during the presentation of testimony undermines that trust. (Though many things might: flickering eyes, lack of confidence in his own description, lots of going back and correcting himself, brief bursts of incandescent lunacy … )

    In the end, Jones comes to believe Smith’s testimony, and it’s the belief, not the trust that helped to establish it that contributes to his knowledge.

  10. 2a. In as much as we count peoples institutions, I think a lot of people are going to share the intuition that you can come to know how to get somewhere just by asking. I know your the kind of guy who likes to take into account the judgments of others, so I think this carries some weight for you.

    2b. You have to build in that the agent picks someone at random, with little to no discrimination. That she is completely unreflective. If I recall correctly, Riggs makes some moves like this in denying knowledge in the Chicago traveler case. Of course, Chicago traveler isn’t anything like that. She asks directions of a competent adult, not a child or someone obviously intoxicated. She’s sensitive to defeater, etc.

    2c. I’d have thought we come to know a lot of things by simply asking people who are strangers to us, and we do so in lots of contexts that aren’t classroom settings. I know how to get around town, how to fix my rear derailleur, that there are rapids ahead on the river. In any case, maybe you can say more about this narrative/contextual stuff.

    2d. Maybe not you, but if someone wants to defend the credit thesis, which is a necessary condition on knowledge, then they need to show that there can’t be cases like Chicago traveler.

    4. Goldberg is extending the kind of self-trust of the authors you mentioned before to other agents. It’s the way we rely on others to be reliable or dependable testifiers. We’re extending our cognitive abilities through those individuals. We can talk more about this at the MEW.

    5. Ahh now I see what you’re after. I guess I wouldn’t think such a beast exists.

  11. “I don’t think I’m making trust do any epistemological work here.”

    My point exactly! 😉

    In all seriousness, though, I remain utterly confused by talk of trust in epistemology. And I have great fondness for all the people working on it.

  12. Matthew,

    2a. True, but I think there’s a lot of loose talk going on. I think we’re pretty casual with how we speak generally, and certainly with “knowledge.” Unger notes this and Conee makes the point well in the Contemporary Debates in Epistemology volume.

    2b. I’m just taking the case as she gives it: “the first adult passerby that he sees.” You’ve added “competent” and that’s part of what’s in question. Now if you want to go reductionist on me and start talking frequencies in reference classes, then we can go there, but my understanding is that most epistemologists of testimony are anti-reductionists.

    2c. Note that you focus on knowledge *how* which, pace Jason, is distinct from knowledge that. Or even if it’s a sub-genre, it differs substantively from testimony.

    2d. And the other side has to show there can be, and we disagree on that.

    4. Foley has a chapter on the “interpenetration” of trust, and Linda’s current work certainly does.

    5. But if not, then almost anything can count as knowledge. If there is no such case, then it seems to follow that any degree of testimonial support AT ALL is sufficient for knowledge in the absence of defeat. And that sounds crazy. At least it makes knowledge crazy-eazy to have, and it’s hard to explain the allure of skepticism.

    See you at the MEW.

  13. 2b. Now I’m wondering if you’ve actually read Lackey’s paper. Here is a relevant passage in which she’s willing to grant competencies to the traveler.

    For instance, Morris clearly chose a human passer-by rather than a cat or dog or telephone pole; he chose an adult testifier rather than a child or infant; he chose a conscious person rather than one who is unconscious or minimally conscious; and so on. Such discriminatory behaviour, one might argue, is clearly an important part of why Morris obtained accurate directions to the Sears Tower since its absence would have most likely resulted in his acquiring either no relevant directions or inaccurate ones.

    Of course, she goes on to argue that granting all of this isn’t going to help the credit theorist.

    2c. Knowing how to how to fix my rear derailleur involves a lot of knowledge that, but I didn’t claim just know how cases. There are still the cases like knowing that there are rapids ahead on the river.

    4. Here is a pass at trust for epistemologists that I’ll paraphrase from Goldberg. I don’t think there is anything hard to grasp here

    S trusts an information source for presenting the truth on an occasion iff S regards that information source as dependable for the truth, and S’s so regarding the source of information is the basis of S’s acceptance of the information presented.

    5. I don’t like to think of these things in terms of evidential support, but I’m failing to see why there is a problem here once you’ve removed all the defeaters. Sometimes knowledge is easy to get, why should that be a problem.

  14. 2b. Dude! There’s nothing in that quote that goes against what I said! And all I can say is that I find Wayne’s reply more compelling. Well, I can add that the *reason* Morris so acts, if he’s like us, is that his background beliefs–his evidence–make it much more likely that he’ll get the truth out of those sources.

    2c. I’ve lost the dialectic on 2c.

    4. “regards”? Does that mean “treats” or “believes” or what?

    And, surely, regardings can be rational or irrational. 30-some-odd people regard that book-burning ass-clown in Fla as an authority. When I crossed Fred Phelps’s picket lines in Topeka (ah, remember Topeka and the biker bar?), those people regarded him as a prophet.

    But their regardings were irrational. Their trust was irrational. It was not founded upon any evidence of trustworthiness. Shame on them. (Well, some of them were in a bad place and are probably not culpable.)

    5. Not *that* easy. How do I know I’m not a BIV? Matt told me.

  15. Concerning knowledge as mere true belief:

    There is some interesting literature on cases in which it seems that little–or perhaps nothing–beyond true belief seems needed for knowledge. The easiest way for me point you to (some of) this literature is to send you to the short section of my book where I discuss it, with references:

    Case for Contextualism–proofs, Ch. 1, sect. 6: Which Claims to Take Seriously and the ‘Floor’ for ‘Know(s)’

    Kripke had an interesting case of this sort in his “Nozick-bashing” lectures in the 80s; since those lectures seem still unpublished [they’re supposed to be coming out as “Nozick on Knowledge,” item #9 here, but I don’t think that’s out yet, and, well, I’ll believe it when I see it], I characterized the case as best I could from memory (middle of p. 14). Other examples, I just give references to, except for one of John Hawthorne’s, which I quote on p. 17 before discussing. Since my list of References is at the back of the book, here are the pieces I refer to in that section:

    Hawthorne, John 2000. ‘Implicit Belief and A Priori Knowledge’, Southern Journal of Philosophy 38, Spindel Conference Supplement: 191–210.

    Hazlett, Allan forthcoming. ‘The Myth of Factive Verbs’, Philosophy and Phenomenological Research.

    Sartwell, Crispin 1991. ‘Knowledge is Merely True Belief ’, American Philosophical Quarterly 28: 157–65.

    _____1992. ‘Why Knowledge Is Merely True Belief ’, Journal of Philosophy 89: 167–80.

    Holton, Richard 1997. ‘Some Telling Examples: A Reply to Tsohatzidis’, Journal of Pragmatics 28: 625–8.

    As a contextualist, of course, I can, and do, hold that the “floor” for “know(s)” can get very low, without having to hold that it is always down there. That seems to be the contextualist position Hawthorne is tempted to in his 2000 paper. (You’ll see that neither John nor I are committed to the floor ever getting all the way down to mere true belief — on John, see esp. my footnote #10.)

    I also discuss cases in which it can seem (at least to some) that truth isn’t needed for knowledge (Hazlett). These are cases I don’t try to accommodate (adjust the theory to make the claims come out true). But I thought it important not to simply write such cases off as mere “loose talk,” but to give some more principled reason (the main element of which comes from Holton, but I also have some discussion of intonation) for not accommodating such claims.

  16. Thanks Keith, that’s most helpful. I’m pretty brazen about writing such speech off as loose talk–because that’s what it patently seems to me to be–but, like you, I think it’s good to have a closer look as well.

    Of course, the “loose talk” objection has been marshaled by Conee against contextualism (Contemporary debates in Epistemology), so perhaps you don’t want to lean on that too heavily.

    Conee also says, however–and I can’t think where right now–that he’s only interested in analyzing “factual knowledge” which is what epistemologists have traditionally been interested in.

    It is only one of the properties expressed by “knows” (and I think if you go this route you should think of these properties as related by no more than family resemblance so that no Socratic definition is possible. That’s me talking now, not Earl).

    Some evidence for this consists in the token cases that traditional epistemologists have given as paradigm examples of knowledge. I think it’s only since the linguistic turn that people have been worried about the kinds of expressions that seem to denote properties which are non-factive or non-normative.

    I think those expressions are interesting from a linguistic point of view, but not necessarily from an epistemological point of view because they simply stand outside the scope of what the conversation has previously been about, as I read that conversation.

  17. I think a lot of the conversation within epistemology has been in certain ways extremely confused — esp. concerning the relation of its own claims, questions, & investigations to ordinary (non-philosophical) uses of “know(s).” (An account of this relation is vital to epistemology, I believe, and, of course, can’t be worth much w/o a reasonable grasp of the meaning of the ordinary claims.) But that’s a long story…

  18. Keith, I’m somewhat sympathetic to that, but I’m not sold on it.

    I don’t see why the neo-Chisholmian couldn’t say: “Here’s what I’m interested in: what the following kinds of “factual knowledge” have in common: Sense perception, memory knowledge, introspective knowledge, rational insight, perhaps numinous perception.”

    Or she may specify by specifics: “What the following cases have in common: My knowing by seeing that I have hands, my knowing by remembering that I had toast for breakfast, by knowing by introspecting that I believe I’m cold, my knowing by rational insight that numerical identity is transitive, my knowing by perceiving that there is an Infinite Other.” And the list could be longer or shorter.

    The kinds of cases where we bear some relation to p we express by “knows” but it seems like that relation isn’t factive or normative are too dis-similar to be relevant to that particular neo-Chisholmian’s concerns.

    So as long as one is clear about what one is concerned about, I don’t see that any general inquiry into or understanding of “the relation of its own claims, questions & investigations to ordinary uses of ‘knows'” is “vital”.

    I personally still find it interesting though, just not necessary for what I do as an epistemologist. Others may have different interests. There’s no accounting for that.

  19. I don’t believe in non-factual knowledge. “Knowledge,” ordinarily so called, requires truth, and more.

    I suspect most epistemologists, in constructing and evaluating theories in epistemology, are in part being led by what it sounds right to call “knowledge.” Especially if “intuitions about cases” are in part driving them. If so, a big part of discerning when this leading is really misleading will likely be understanding how the term operates.

    Your imagined epistemologist is interested in what a bunch of cases have in common, & the items on the list all have “knowledge” in their specifications. If those instances of “knowledge” are in regular English, then I’m interested in what those cases have in common, too, since it’s likely to be that they’re all what’s designated in normal English by “knowledge”! (If this imagined epistemologist isn’t speaking in English, then I need a translation.)

  20. You seem to have more faith in the univocity of language than I do. I actually have a post planned on this, so hopefully we can continue the dialog then, as you always have enlightening things to say, even though I almost NEVER agree! (Except when we both agree against Jason.)