Prior Assumption = Incomplete Presumption @jonmbutterworth @guardian

Another interesting and typically honest down-to-earth Grauniad piece by Jon Butterworth, following on from a couple of weeks ago, he’s obviously had plenty of correspondence from two quarters. Scientistic types who find Bayes Theorem the thin edge of a statistical wedge, admitting subjectivity into their hallowed ground, and philosophical types (aka nut-jobs) using the chink to insert suggestions of alternative physics into Jon’s “standard model” domain. Against the scientistic types Jon is happy to point out the value of honesty when it comes to admitting Bayes; to the nut-jobs he says:

For example, as a writer and head of a physics department, I get quite a few unsolicited communications about new theories of physics, often involving Einstein having been wrong, or the Higgs boson actually being a macaroon or something. I have a prior bias here, based on the enormous amount of existing evidence. Einstein might have been confused about the cosmological constant on occasion, but given prior evidence it is highly unlikely that the whole thrust of relativistic mechanics is up the spout. Likewise, I personally have quite a lot of evidence that the Higgs boson is consistent so far with being the fundamental Higgs of the Standard Model, and inconsistent with the macaroon theory.

Well I’ve not been sending Jon any pet theories, but I do highlight two of Jon’s points:

(1) it is highly unlikely that the whole thrust of relativistic mechanics is up the spout.

(2) a lot of evidence that the Higgs boson is consistent so far with the Standard Model.

Firstly, that prior assumption, his bias,  has a massive impact on interpretation of new results. Perversely, Einstein was right, and there is a great deal of “non-inflation” evidence the standard model is way off the mark. Once that is more generally recognised, that prior assumption (1) is gone, totally.

Secondly, as I reported when I heard Jon speak on the latest LHC Higgs evidence, it is quite explicit that the increasingly significant (5-sigma plus) evidence is pointing to internal self-consistency of the incomplete standard model. (2) does doing nothing to prove the fit between the model and the real world, other than to reinforce the subjective impression in (1).

As well as Bayes, we need a little Godel here. Jon already highlighted last time, the need to look elsewhere. It is a wonderfully healthy situation to have an honest scientist thinking out loud in the mainstream press. Could save science form the scientistic extremists. Be even better if we could find Jon a philosophical type for similar dialogue, with mutual respect.

[Post Note : More on Bayes in science from the NYT

“Statistics sounds like this dry, technical subject, but it draws on deep philosophical debates about the nature of reality,” said the Princeton University astrophysicist Edwin Turner, who has witnessed a widespread conversion to Bayesian thinking in his field over the last 15 years.

Countering Pure Objectivity

Woohooo – maybe more scientists will gradually see the need to respect and/or get to grips with philosophy?]


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.