Dear friends -- Pedro had suggested that the mini-session of responses to my New Year Lecture should last about two weeks. I think that the lesson of QBism for FIS is indirect. In order to make sense of quantum mechanics, which has been spectacularly successful, the interpretation of probability should be Bayesian. If we want a unified worldview to include both quantum physics and classical physics, we should adopt a Bayesian philosophy for ALL of science. From this point ion, quantum mechanics itself can be left out of the debate.
So what are the implications of Bayesianism for the definition of the term "information"? It was surprising to me that the discussion veered off to ancient concepts like Shannon information and entropy, but I suppose that's what FIS is all about. Bob Logan and Rafael helped to tighten the discussion by putting it into its into historical context. Gordana linked the philosophical outlook of personalist Bayesians to that of instrumentalists. I hope that this connection will be further explored by philosophers of science. On the more technical level, I was very pleased to read Robert Ulanowicz's statement: "So Bayesian forms of Shannon indices are far more useful about gauging information than most realize." In this way Bayesianism can help to put "meaning" back into the information concept, from which Shannon summarily, but effectively, banished it. The aspect of probability theory with which QBists have the most trouble is the treatment of certainty -- i.e. probabilities 0 and 1. QBist struggle to define those by insisting that probability 1 for an event does not mean that the event is a fact, or an Einsteinian "element of reality" -- only that an individual agent is very, very sure that the event will occur. Loet goes in the opposite direction by excizing the probabilities 0 and 1 from the normal, Kolmogorovian closed range from 0 to 1. This approach (Logic in Reality) is too new to me to comment on, but it does offer an interesting alternative. Hans
_______________________________________________ fis mailing list fis@listas.unizar.es https://webmail.unizar.es/cgi-bin/mailman/listinfo/fis
