Sunday, August 14, 2011

Intellectual Peers and Epistemic Peers

Paul M. has a nice post on a particular application of thoughts about epistemic peer disagreement. I'm in broad agreement with the basic thrust of the post; I've argued against the position he calls ST myself a number of times. (I am inclined to think that NDT is ultimately not coherent given standard characterizations of what is meant by 'epistemic peer', but it's much less common to meet it in the wild, anyway -- most people have more moderate views.) I do think, however, that the post suffers from not looking more closely at what 'epistemic peer' means in this context, and this does have some ramifications for the argument given. To say that someone is an epistemic peer in a given context is to say that this person has no significant disadvantages or advantages with regard to background, ability, or external assistance for that context. Practically every serious discussion of the subject requires this: no argument can be made about epistemic peers and what follows from their disagreement if, for instance, we merely mean that two people arguing over a calculation are more or less generally equal in background and overall ability if it also turns out that one person has devoted hours more to the calculation, or has a greater proficiency in this particular kind of calculation, or has a calculator.

It also has the effect of complicating any discussion of practical applications. If you're just positing an abstract situation for the purpose of better understanding knowledge, or disagreement, or some such, things are easy: you just posit that there are two or more people, none of which have any relevant epistemic advantages over the others. In real life, however, there is good reason to think that one can never know with sufficient precision whether anyone is an epistemic peer. In real life, of course, when we think of peers, we think of what we might call intellectual peers -- people who have, in general, roughly the same ability as averaged out over a number of things and very broadly speaking the same kind of educational background. But this is not precise enough to say much about whether these intellectual peers are epistemic peers in any particular case. And in practice it doesn't matter -- we don't have to bother with epistemic peers. Indeed, even whether someone is an intellectual peer is not hugely important. Sally may be the world's most brilliant mathematician, and John merely a competent one, but there will be plenty of circumstances where John will be competent enough to check Sally's work, disagree with her on a mathematical issue and be right, or catch Sally out in a mistake. It reminds me of a philosophy professor I know who once told a rather condescending story about a discussion with his teenage daughter on critical thinking, in which his daughter put forward some reasonably clever arguments and the philosophy professor gave some very mediocre answers to them, answers that did not really do justice to the issues raised by the daughter's arguments. When this was pointed out, he got into something of a huff and pointed out that he was a professor of philosophy who had studied analogous issues for some time and his daughter was merely a teenager. Now, it is true that in general teenagers are not the intellectual equals of tenured professors. But the moral to be taken from the story, besides the point that any parent, regardless of intelligence and education, who thinks they will always have the better argument in a discussion with their teenager is making a truly stupid assumption, is that good arguments are good arguments regardless of who makes them. Setting aside the fact that the stupidest teenager can have a good day in an argument, and the cleverest professor a bad day (anyone can stumble onto something good, and anyone can just plain stumble), in some contexts even the stupidest teenager can be in a better position to know what they are talking about than even the cleverest professor. Set your finest college professors in the same room with a very irrational pre-teen and ask them which of the actors you will show them is the more accurate Justin Beiber impersonator, and you will see that epistemic peers for a given kind of problem need not be intellectual peers in general. In My Cousin Vinny, Mona Lisa Vito is not an intellectual peer of George Wilbur; at the very least, he is certainly better-educated than she is, and there's no implication in the movie that his intellectual abilities are inferior to hers. But in the area of general automotive knowledge, she may well be his epistemic superior, because her father was a mechanic, and his father was a mechanic, and her three brothers are mechanics, and four uncles on her father's side are mechanics, and she's worked in her father's garage, and she can tell the make and model of a car by looking at what tire tracks imply about positraction.

Intellectual peerage in other words is not epistemic peerage; it states something about one's general-purpose intellectual competence, not one's competence for drawing particular kinds of conclusion with particular kinds of evidence, and thus is mostly useful just for identifying which people are equally likely to be competent over a wide range of issues. It's an important distinction, although easy to overlook; much of the early discussion of epistemic peer skepticism was vitiated by a failure to distinguish intellectual peerage (which is broad and has no implications for particular topics because it, so to speak, is a form of assessment-by-average -- this person is my peer because in general or on balance we share backgrounds and intellectual dispositions) from epistemic peerage (which has to be defined in terms the ability and background people have for drawing conclusions in any particular context -- this person is my epistemic peer because we share the same background and ability for drawing a general kind of conclusion from the same evidence). Two people with equally good eyes and equally good observational skills looking from equally good vantage points without any advantages over the other with respect to relevant tools are epistemic peers on the question of who crossed the finish line first, even if one is for most things clearly smarter, better informed, and more thoughtful. Two intellectual peers, with the same basic education and intelligence, might not be epistemic peers in judging the significance of a bit of evidence in their field if one is highly biased with a bias that distorts precisely this sort of judgment, or if one has put considerably more effort into this particular problem, or if one has an educational gap at just the right point.

Thus (1) in practice we use intellectual peerage rather than epistemic peerage, and the two are not the same; (2) we can practically identify intellectual peers, but this has very little in the way of epistemic consequences for any particular context; and (3) identifying real epistemic peers requires a degree of precision we rarely if ever have in practice. The upshot of this is that any discussion of epistemic peers is at an abstract and hypothetical level; it can be useful in certain circumstances at precisely this level (e.g., in better understanding the nature of inquiry or how knowledge works under conditions of disagreement). It follows from this, however, that we can't use the disagreement over ST as an argument against it. Certainly intellectual peers disagree over it. But the claim is not about intellectual peers but about epistemic peers: and among intellectual peers some can be in a better position to know something. Some will have a better acquaintance with the arguments; some will better understand the arguments made; some will have a more extensive relevant background; some will be making perfectly ordinary mistakes that anyone in their position could make but that not all of them are guaranteed to make; some will have biases sufficiently strong to distort evaluations, creating unnecessary doubt or imprudent confidence, and so forth. We can posit as a hypothetical that if epistemic peers disagree over ST, the conjunction of this with ST is defeating for ST. But hypotheticals of this sort don't refute or defeat anything. Similar problems arise both with the position Paul designates as CAST, as formulated, and with his argument against it.