Last week, David Brooks wrote a column about the American Midwest. A couple people had a problem with it--particularly with the first sentence:

If Balzac were alive today, he would plant himself in that region of America that starts in central New York and Pennsylvania and then stretches out through Ohio and Indiana before spreading out to include Wisconsin and Arkansas.
There are a number of things one could say about this sentence. One might, for example, start with the "Balzac" part (is there any member of the Western canon Brooks hasn't worked into a column? Next up: how Sartre illuminates Glenn Beck).

William Easterly at Aid Watch, however, chose to focus on the geography. Since when, he asked, is New York part of the Midwest? In addition, he continued, "there has never been a single Midwesterner in two centuries who ever thought they were in the same region as Arkansas." Jesse Walker at Reason and Kathy Kattenburg at The Moderate Voice snickered in agreement.

But hang on just a second. Razib Khan from his new post at Discover is willing to defend Brooks on this one--with a "real social metric," he says. Have a look: it's the map of where carbonated beverages are referred to as "soda" and where they are referred to as "pop" (or "coke" or "other"). And you know what? The shift does actually start in "western New York and Pennsylvania." But he does have some lingering questions. For example:
I haven't heard of a good explanation why coastal Wisconsin and a penumbra around St. Louis stands out anomalously in the Midwest. Though it has been pointed out to me that the original elite of St. Louis had pretty strong New England connections.


It's worth noting, too, that this map still doesn't get Brooks off the hook regarding Arkansas--or Balzac.