Why Would God Hide?

[Michael]

I wonder who it is that decides a claim is extraordinary :-)

The people who make a claim, especially one that purports to have inside knowledge of the universe that others don't, are the ones who have to bring extraordinary evidence to bear to prove it.

Or, in science, we sometimes use a different framework, a Bayesian framework. In the Bayesian framework the strength of the evidence required depends on how strongly a 'prior' position is held. So those positions initially held most strongly require the stronger evidence to overturn them. The insight Bayes had over the 'probabilists' (who viewed all scientific experiments from a neutral view) was that even in science we do not approach new data from a naive perspective, but we bring 'prior' presuppositions to the table. In Bayesian thought we don't just use data to test hypothesis, we use current paradigms to test (and accept or reject) new data. We might choose to reject the new data as being likely to be untrustworthy (or just a fluke) if it conflicts with a confidently held presupposition (we saw that in the recent example of a team who thought they may have measured faster-than-light travel). Both methods are valid, and both are used in science. It disturbs some to find that each method may reach a different conclusion based on the same experiment (it certainly made me shake my head when I first came across it, being brought up in a traditional probabilistic scientific paradigm). The discipline is to decide which method to use before doing the experiment. Bayesian stats also allows for different strengths of conviction; that is implicit in setting the 'prior'.

But I find Bayesian thought works well not only in science, but it seems to describe well what I see around me in lay discussions: those with strongly held views require greater evidence to overturn those views. What Bayesian thought also predicts is that positions can become self-sustaining: they may be held so strongly that they will almost always reject contrary evidence as untrustworthy. A sign of this is when someone is honest and says "nothing will convince me of ......x".

So I find Bayesian thought aligns with what I see around me, and it has a good scientific pedigree. Bayes predicts that you might not agree, not based on what I've just said, but based on your 'prior' position before you read what I just wrote :-)

Just another way of looking at things (but one with a good pedigree of use).