Single-Variable Theories

Fri, 21 Apr 1995 13:18:05 EST

Bob Graber's example of the pendulum was precisely the sort of thing I
thought everyone was talking about, i.e. he has prove my point. Galileo's
law for pendulum's works great if you have a pendulum in 3-D Euclidean
space that is unaffected by any other factors such as resistance/inertia.
But what seems more important to me than restating my poing on ceteris
paribus clauses and ideal situations is to make a couple of remarks
about causes versus explanations versus laws.

1: The pendulum law or Boyle's gas law or any other example you might
want to pick from the physical sciences are used in explanations about
the behavior of something. Boyle's gas law tells us nothing about what
gas is, or what the cause of the law is. Karl Hempel, one of the logical
positivists who did more than anyone to formalize the hypo-thetico
deductive method loved to use Boyle's law in his discussions. The important
thing is that in explaining the behavior of gases Boyle could use the
law but that explanation had nothing to do with causality. This is the value
of a law: it can postdict and predict, it is uniform for all time (in theory).

2: Causality, as Bob admits, is a morass to deal with. I'll suspend David
Hume's unresovable (to date) critique of causality. What I'd like to
note is that when Bob jumped from Galileo's pendulum to Carniero's single
variable explanation for political evolution he jumped from laws and
explanations to empirical causal claims -- not at all the same thing.
Carneiro might claim that he has figured out while all complex political
societies to date formed but that is not a law of social formation. It
is an emprical claim about a cause that isn't even transhistorical: even
if its true in all existing cases that doesn't allow us to to say that
all future cases will be the same or, more importantly, that all future
cases MUST be the same. Galileo's pendulum and Carneiro's theory are two
completely different issues. Carneiro's claim doesn't allow for postdictionor prediction. It can be tested, it can serve as a model, it can even be made
to "LOOK" like a law, but that still won't change its empirical status.

3: I have to agree with those on the list who don't understand why a
single cause is so attractive. The idea, as one post suggested, that
science is supposed to be elegant and simple is a prejudice about what
science should look like that has nothig to do with the realities of
human complexity or, i would assert, with science either.

Brad D. Hume
History and Philosophy of Science
Indiana University