Wei wrote:
> If this is the right way to think about probability theory, then we should
> make clear in our discussions whether we're talking about it as a positive
> theory or a normative theory.
Yes. (Though if the descriptive and the normative theory came too far
apart, so that what people purportedly think is hardly ever related
to what they purportedly ought to think, then that would be an
indication that at least one of the theories had got it wrong.)
>But it is still not clear on what basis we
> should choose a normative probability (or decision) theory.
The best normative decision theory is the one it would be most
rational to follow. So in order to determine which decision theory
has this property, we must make sure we know what "rational" means.
"Rational" is a word in English, so in order to determine what it
means, we need a grasp of the English language, a good theory of
meaning, and possibly some other facts as well. Once we know the
meaning of "rational", all we have to do is to compare various
proposed normative decision theories with this meaning, to see
which one is most "rational". What meaning "rational" has is partly
an empirical question -- it depends on how that word happens to be
used by English speakers --, but in practice what is needed is not so
much an empirical investigation of linguistic habits but rather a
careful analysis of all the implications of what we already know
about the usage of the word "rational". We need to check proposed
explications of "rational" to see if they agree with most of our
intuitions about how that word is used; we need to see which
explication best fits in with all the things we tend to say and
associate with "rational", and yet is as simple and powerful an
explication as possible.
_____________________________________________________
Nick Bostrom
Department of Philosophy, Logic and Scientific Method
London School of Economics
n.bostrom.domain.name.hidden
http://www.hedweb.com/nickb
Received on Mon Mar 16 1998 - 12:14:46 PST