Monday, July 5, 2010

Aversion to computers in the social sciences: the story of ethno-telephony


For folks such as myself, who'd prefer to reduce human civilization and its economies and technological trends to a convenient, elegant set of equations, cultural lag is a bitch.

We'd like to think that when a better tool exists, or a new technology or practice offers to reduce hardship and increase happiness cost-free, the people of our good planet should see the light and change their ways overnight. This doesn't seem to happen. Tribeswomen in a remote African village refuse to take up long-handled brooms, instead using their traditional handbrooms, because forgoing back pain in their cleaning would be "lazy." Papuan warriors refuse to make their arrows fly truer by adding the fletchings they've seen the neighboring tribes use. The world's businesses haven't switched to OpenOffice. The list goes on.

Usually when we are confounded by a culture's decision not to adapt, the culprit is a value system that stands at odds with ours. Another culture sees value in tradition, or hardship, or battles with lower body counts. Other times, technological relinquishment seems to serve the purpose of preserving the hegemony of an incumbent hierarchy. This, I believe, had been the case, when I encountered my first textbook example of cultural lag in the field, while dealing with the exquisitely interesting culture of the American social scientist.

Ethno-telephony is a quantitative method in cultural geography, invented by Peveril Meigs in 1939. Meigs wanted a method to quantitatively map the French regions of Louisiana, so he got a bunch of phone books from different towns, counted all the French names, and made the excellent map depicted above. Now I know what you're thinking: "But a last name is hardly evidence of integration into the Louisiana Acadian cultural complex! Why, there may be whole towns with French last names where various social forces have completely anglicized the population, and the reverse might be true in other places!" Yes, good reader, you are right, but anything verging on quantitative is hard to come by in the study of culture, and we take what we can get.

Meigs's method kind of made it big. In fact, last time I checked, the premier journal of cultural geography was printing a couple articles per issue using this method. Geographers turned to the yellow pages and started asking even more interesting questions: Where is "dixieland"? Which people consider themselves "yankees"?

So when I was in school, way back in the year of 2006, I had this idea: "What the hell! All these guys are still counting phonebook entries by hand! Why doesn't someone make a program where you just type a query and it gives you a map?!" I looked online for a program like this, and after a while, I gave up and got my good friend, who had some minor scripting skills, to write a program. At the time, with a small sampling if phonebooks, some interpolation, and a bunch of graduate students working as slaves for many hours, people were turning out things like this map of dixie:



With our very modest, decidedly amateurish mechanization of the process, we made something that would require a few seconds of labor to produce this:


Then we got to thinking, "I wonder what the Latter Day Saints are, exactly?"


Who eats the most chicken?


Very soon, sitting in our living room, we had proven one longstanding theory and disproven another. We were conjuring religious denominational maps of hitherto unknown accuracy for the pleasure of satisfying our own curiosity. Every time we entered a few key strokes, we emulated months of work on the part of some poor troglodyte graduate student. We even pinned down Texas's infamous taco-barbecue line.

We said "Wait'll geographers get ahold of this! We're going to change the world!"

And so we approached geographers, first at our school and then at other places. We said "Use this! We'll make your maps for free!" And one after another, they all politely said "no thanks." Nobody was interested in using the new method (which was, of course, the old method, only done automatically). Everyone chose to continue doing things as they had been doing them.

Ethno-telephony didn't die that day. Grad students still count names in phonebooks, and the maps still get made. "But why!?", you ask? I think I might know.

I started to ask myself how research is valued. How is tenure established? Do we give people acclaim and money for contributing to the sum of human knowledge, or for suffering? I've come to believe the latter.

The problem with running culture maps automatically is that it doesn't give anyone a headache. What distinguished professor wants to open his paper with "So I typed this cool term into the search field"? Why, that doesn't sound academic at all! Sure the maps have 10X the resolution. Sure the search is exhaustive and doesn't rely on sampling. But no one will give you a medal for that kind of research; it's just too easy.

Imagine you're a cryptography analyst, handsomely paid, and someone invents a magic machine that makes all cryptanalysis completely automatic. Just press a button, and the thing is done for you, in a time frame which humans can't shorten by any stroke of anthropic genius. How would this bode for your field? I can tell you one thing: you wouldn't continue to be handsomely paid.

After the ethno-telephony incident (and before leaving the social sciences altogether), I believe I picked up on other deliberate inefficiencies in the social sciences, and I've come to believe that there's a great body of knowledge, ripe to be known, which is kept from us because it would ruin the research agendas of many distinguished professors, perhaps in one day.

And perhaps things are as they should be (though my undergraduate ambitions of singlehandedly revolutionizing the social sciences did not go quietly).

But I still know where the taco-barbecue line is.

(If you want to play with a really buggy, old, unsupported, ugly version of our ethno-telephony program, go to mapanything.net)



  Buy me a coffee to support posts like this.

  (Or, you know, a house. If you're just like a bored billionaire or something.)