March 24, 2007

Epistemic Standards in the Wild

Josh Marshall provides an interesting example of explicit standards for "know," of the sort that Peter Ludlow discusses in "Contextualism and the New Linguistic Turn in Epistemology" [pdf]:

Now we know with crystal clear proof what we really already knew a week ago: that Alberto Gonzales was lying about his role in the US Attorney Purge.

The "crystal clear proof" is basically a smoking gun. An e-mail documents that Gonzales personally discussed firing the attorneys, contradicting his previous statements that he had not seen any memos or been involved in any discussions. So Marshall is essentially saying that before the e-mail was released our evidence was good enough for knowledge, but now it meets a higher standard, knowledge "with crystal clear proof."

A couple of possibly interesting things about this: "Really" isn't being used as a degree modifier here; it's applied to the weaker degree of knowledge. "Actually" could have been used instead. I'm not sure exactly how to explain this use of "really"/"actually," but it seems as though it expresses an implicit contrast: "You might have thought that without this crystal clear proof we didn't know that Gonzales lied, but he did."

Another question is what the significance of the e-mail is to our epistemic state, if we knew beforehand that Gonzales was lying. The least Marshall-friendly interpretation is that before the e-mail came out we couldn't quite be certain, but now we can.

I think Marshall would say something different, though; that before the e-mail came out we could still be certain that Gonzales was involved in the attorney firings. It was just common sense that the Attorney General would've been involved somehow in such a major and unprecedented operation within the DoJ.)

But e-mail improves our epistemic standing in two ways. First of all, previously reporters couldn't have just said that Gonzales had to have been involved in the firings based on common sense. (Though I wish they would more often write things like, "DoJ spokesbot did not say anything about why Gonzales would have had no participation in an unprecedented mass firing of US Attorneys that had been appointed by the sitting administration.") Now, reporters can mention the e-mail that contradicts Gonzales's previous statement.

Second, there's a potential dialectical issue here. Even if people on the left were perfectly certain that Gonzales knew even before the e-mail came out, those who disagreed with us might well have been unsatisfied with the claim. You can imagine arguing with someone who would demand some specific reason to believe that Gonzales was involved in the purges. Now those who think that have something specific to say: We know that Gonzales was involved in the purges because this e-mail says he was. So even if we knew before, the new evidence places us in a better position to support the claim dialectically. [Now, trained professionals may still be capable of denying that the evidence indicates that Gonzales was lying, but it's a bit of a stretch.]

It may be an interesting question whether new evidence can strengthen our dialectical position with respect to p without giving us a stronger reason to believe p, and if so what that means about degrees of knowledge or such. Those (like my colleague Allan Hazlett) who think that knowledge is closely tied to dialectics might be particularly interested in such questions.

Posted by Matt Weiner at March 24, 2007 10:12 AM

You could look at it in terms of prior and posterior probabilities. We had a very strong belief that DoJ representatives were lying, because their lips were moving and past experience suggests that administration representatives have become accustomed to misleading simply because they can (let alone when the truth is inconvenient).

So we "really knew" it then, but we couldn't quite say that we knew it. Because we also knew that it wasn't quite a respectable argument in what is left of polite society: the standards of our political discourse have a presumption that we can't go around accusing people of near-criminal dissembling just based on our opinions of them. (Which is to say that the standards of our discourse have not evolved to take account of, and can be taken advantage of by, people who lie reflexively.)

Posted by: Ben at March 26, 2007 01:41 AM