|
|
Re: Anthropology and Prediction
Bradley David Hume (bdhume@INDIANA.EDU)
Sat, 9 Sep 1995 11:40:36 -0500
The discussion about prediction in anthropology seems very familiar to
another discussion held some months back on laws in anthropology. Both
discussions appear dogged by that old bogey "physics-envy." But in
both the "physics-envy" is of the deterministic, Newtonian sort (or
so it seems to me). No doubt, as populations grow & etc., generalizations
can be made about changes in social complexity. Just how detailed
those "predictions" can be makes me wonder if they are indeed predictions
(i.e., are they so general that they don't really constitute predictions?).
I also agree that these sorts of predictions beg the question about
causality versus correlation. But that leads me to two points:
1: We no longer live in a deterministic universe. Quantum physics
is irreducibly probablistic. The social sciences have been probabilistic
almost from the very beginning. That doesn't solve any problem about
prediction, but it is worth remembering when considering that
anthropology and evolution are traditionally regarded as "historical"
sciences where prediction is not only less "important" in some ways
but also far more difficult. Predicting how a gas will behave
according to Boyle's Gas Law is a very narrow problem compared to
making predictions about what will happen in systems with so many
variables (like changes in human society). Perhaps that's simply
the difference between controlled laboratory conditions and the
complexities of field research.
2: Having laws which allow us to make predictions has nothing to do
with causality in a strict philosophical sense. The Logical Positivists
(determinists if there were determinists) talked a lot about prediction,
laws, and covering models for the all the sciences. They avoided the
problem of causality like the plague because they didn't feel that
anyone could get around Hume's arguments about causality. Hempel
finally tried to grapple with it and Popper (not really a Lo-Po)
believed he had solved it. In any case, being able to make predictions
is often done on the basis of good statistical correlation or a
covering law. In neither case do you have to understand the cause.
Given that the universe, from particles to societies, is probabilistic,
I hate to think that anthropologists are still concerned that they
aren't making sufficient advances in science. Look at it this way:
The social sciences have been probabilistic since the 1830s. It took
physicists another century to catch up.
Brad D. Hume
History and Philosophy of Science
Indiana University
|