The World According to Derp

Yesterday Josh Barro wrote a post about his disagreements with Erick Erickson, concluding with this:

Basically, Erickson is derpy. And Erickson has big appeal to conservatives because lots of them are derpy. But the country is getting less derpy, and in time the Republican party will have to get less derpy, too. That’s my project, and I don’t expect Erickson to like it.

 Paul Krugman linked to Barro's piece in a post entitled "Moral Derpitude."  On Twitter and elsewhere, people have been pushing back against the use of the word "derp."  For instance, Ross Douthat just tweeted:

The problem with "derp" isn't the definition. It's the word itself.
— Ross Douthat (@DouthatNYT) June 5, 2013

I'm inclined to agree.  But what is the definition?  Noah Smith has written a post defining it as holding strong Bayesian priors and repeating them in the face of contrary evidence:

English has no word for "the constant, repetitive reiteration of strong priors". Yet it is a well-known phenomenon in the world of punditry, debate, and public affairs. On Twitter, we call it "derp".

Barro seems to like Smith's definition, and so does Matt Yglesias:

This is correct RT @noahpinion: Defining "derp" in terms of Bayesian probability theory: noahpinionblog.blogspot.com/2013/06/what-i…
— Matt Yglesias (@mattyglesias) June 5, 2013

But to me it seems problematic.

Imagine two possible examples of "derp" or "derpiness":

1.  In the first scenario, a liberal has a strong Bayesian prior belief in racial equality.  That is, he doesn't think whites are genetically superior to blacks (or any other race).  Now a study is published supporting the thesis that whites are genetically superior to blacks (these studies are piling up, Will Saletan tells us).  The liberal nevertheless argues repeatedly against the superiority of whites as a race.

2.  Fox runs a piece with the headline, "Obama's Hip-Hop BBQ Didn't Create Jobs."

Now in Smith's framework, the liberal is being "derpy" and Fox is not.  I doubt Smith himself would embrace this conclusion, but it's a fairly straightforward application of his definition.  So the definition itself does not seem to work.

I would use a different economic metaphor.  There are irreconcilable differences between different ideologies.  But a lot of disagreements are not irreconcilable, they simply reflect misinterpretations or factual mistakes.  Sometimes those errors are willful, other times they are not, but in any case they would be eliminated in an ideal deliberation between ideologues.

So think of these arguments as taking place somewhere within a "production possibilities frontier."  That is, if you are operating at the frontier, then there is a strict trade-off between the ideologies—any misunderstandings or misinterpretations have been eliminated, and only pure disagreement remains.  But if you are operating deep in the interior, you are basically making mistakes that would not exist in a perfect world.  You are misconstruing the other side's case or ignoring relevant facts.

It's not a perfect analogy, since some ideologies probably benefit from muddying the water and ignoring relevant facts (whereas in the traditional concept, moves toward the frontier benefit everyone).  But the point still holds, I think:  some people are moving the debate toward the frontier, or operating more-or-less at the frontier (making the kinds of arguments that would be made in an ideal deliberation), and others are degrading the debate or making foolish arguments that would have no place in a proper discussion.

The latter is derpiness, it seems to me.  And I think this is how Yglesias uses the term:

But instead of saying that, we’re treated to endless rounds of derp about death panels and rate shock and blah blah blah.
— Matt Yglesias (@mattyglesias) June 5, 2013

"Death panels" don't reflect a strong Bayesian prior, they are an erroneous argument against Obamacare based on incorrect factual premises.  "Derp" may not be generalized stupidity, but it is also not so narrow as Smith's definition would suggest.