[mannykannot]

> I suppose if we take the philosophical position of Bayesian statistics, where even one sample is sufficient basis for estimating results from some sane priori model...

Some care should be taken to see that this principle is not being used to excuse just making things up. Bayesian statistics is rational but not magical, and it cannot create information out of nothing.

In practice, this comes down to the question of whether the model is plausible, accurate, complete and constrained enough to deliver an answer that is informative about the external world, as opposed to the choices made in modeling. Something more than reasons to believe various probabilities are non-zero is needed.

I feel that, in your final paragraph, you are grappling with this issue.

[fsloth]

I totally agree with you. My intent was not to promote pseudoscience but just to try sketch out why I don't intuitively feel the question would be totally beyond the scope of rational discourse.