[extropy-chat] what is probability?

[Jef Allbright]

gts wrote:

> On Mon, 01 Jan 2007 20:42:46 -0500, Jef Allbright 
> <jef at jefallbright.net>
> wrote:
> 
> > Yes, I would assert that in a very deep sense all statements of 
> > probability entail a subjective viewpoint and that *ideally* the 
> > assessed probability will match the actual likelihood.
> 
> Subjective Bayesians, (at least the purists), follow after De 
> Finetti and deny the objective reality of probabilities 
> altogether.

I don't know about De Finetti, but it seems to me that all statements of
probability necessarily entail a subjective viewpoint as I said above.


> They say objective probabilities are but 
> illusions created by agreement between people about their 
> subjective assessments. Is this also your point of view?

I have made the point many times on this list that a statement can be
considered "objective" only within a specified context.  No one has the
ultimate objective god's-eye view.  This is not quite the same as saying
that such are illusions created by agreement.  Many observations are
"objective" for all practical purposes, and it would be misleading to
call them illusions. Rather, it's a matter of description within
context.


> My guess is it is not; you make a distinction in the sentence 
> above between "subjective" and "actual", and so I take your 
> word "actual" to mean something close or identical to "objective".

As a matter ultimately of faith (based on a maximum entropy approach to
explanation of empirical evidence) I assume a "real" universe
describable (in principle, at any particular moment) in coherent and
consistent terms from any particular frame of reference.  It is in this
context that I refer to "actual" likelihood, knowing that there will
always be some (possibly negligible) uncertainty in the observation.  In
my original post I used the term "ideally" to emphasize this point.

- Jef