or, stuff that I dragged out of my head

Location: Moncton, New Brunswick, Canada

Tuesday, February 22, 2011

Middle of the Road

I saw him face-to-face once while shopping. He's even more spectacular in person.

This is going to blow your mind. Select two or five or twelve pictures from a group of people who range in appearance from not-bad to let's be kind and say unfortunate, a little software wizardry averages out the faces, and what happens? Astoundingly, the result is almost invariably more attractive than any of the faces is individually. The averaged face isn't average at all: it's very, very nice to look at.

The reason is counterintuitive at first but obvious once you think about it. Broadly speaking, an attractive face has two qualities: it's symmetrical overall, and the proportions between its parts adhere to the Golden Ratio. (There is much information and many photos of extremely attractive people on this blog page, if you can ignore the severe hideousness of the page design.) When you average out multiple faces, you are losing the various asymmetries, blemishes, and individualities that may make a specific face less than absolutely beautiful, and retaining the overall symmetry that the faces have in common.


Where might the word "average" have come from? To look at it you can guess it's French: that "-age" suffix that turns things into nouns (haulage, passage, steerage) is very French. But what about the rest of it?

Not what you'd think. The OED suggests that the word originally came from French "aveir", "goods", from Latin "habere", "to have", which is all well and good, but does that even make any sense given the word's modern meaning? It does if you think in terms of shipping: if a ship or the goods within were damaged, the French had a word for that, "avarie", and the original meaning of "average" in English was the financial loss that such damages entailed. Eventually, the word broadened to mean the share of a financial loss that investors in the ship's voyage would take, and from there it was a very short step to the sense of something equally apportioned, evenly divided--averaged out.

Surprisingly, "have" is not descended from French "avoir" or its forebear Latin "habere", though you'd think, wouldn't you? Instead, it's a sibling of German "haben", with the same meaning. Yes, "have" really does look like the Latin, and no, it really isn't related, unless you go all the way back to Indo-European, when all the words have the same source, "kap-", "to grasp": but they split apart and went their separate ways before "have", from Old English "habban", showed up. ("Habere", "to have", is related to "capere", "to seize", which is obviously the source of English "capture", and less obviously "capable", which probably deserves its own writeup.)


Post a Comment

<< Home