1

Dave Harris says the following in "Knightian uncertainty versus Black Swan event":

In Bayesian thinking, chance doesn't really exist. What does exist is a system that is too complicated and complex to know enough to make statements that you can be sure are 100% true.

That seems to imply the universe / world (or at least the processes that are being modeled) is deterministic for a Bayesian. Is that a universally or at least widely accepted viewpoint?

For a related question about the frequentist point of view, see "Deterministic or stochastic universe in frequentist statistics?".

Richard Hardy
  • 67,272
  • 1
    This doesn't imply anything about a Bayesian conception of the universe. Indeed, the whole question appears irrelevant to statistical thinking or practice. – whuber May 07 '21 at 18:16
  • @whuber, this is how I thought: chance does not exist --> the data generating process is deterministic. Since the statement is is very general (In Bayesian thinking, ...) it applies to all Bayesian modeling. Hence, the universe (or at least the parts of it that Bayesians care to model) is deterministic. (The thought that this can be irrelevant has not occurred to me and is quite surprising; I will need more time to ponder upon it.) – Richard Hardy May 07 '21 at 18:28
  • 3
    For me, Bayesian reasoning just implies that we are modeling our reasoning through an agent which assigns plausibility to statements and updates those plausibilities according to certain logical coherence axioms. I don't see where in this framework one commits to uncertainty being attributable to a particular source (either uncertainty due to lack of knowledge in in-principle-knowable things or uncertainty due to intrinsic randomness in the universe). – guy May 07 '21 at 19:48
  • 6
    Incidentally, it annoys me when authors attempt to speak for all Bayesians (or Frequentists) about what they believe philosophically about uncertainty. In reality, the views among people who call themselves Bayesian are very diverse; there isn't some sacred text from which you can learn "what Bayesians believe." – guy May 07 '21 at 19:50
  • 2
    @guy, thank you for helpful comments. I agree with them. My difficulty is that I am only remotely familiar with Bayesian methods and ideas, and so I am trying to figure out what they are. I have no problem with them being diverse if that is the case. I am not advocating any side, just trying to learn. – Richard Hardy May 08 '21 at 09:49
  • 2
    I know from experience that some Bayesian philosophers, statisticians, computer scientists, etc. are very sceptical about any kind of chance or objective probability, but others are not. Some folks just think that Bayesian reasoning is the best way to deal with objective probability in the world. I'll post an answer if I find good illustrations, but in the mean time, here are two potentially relevant starting points from the Stanford Encyclopedia of Philosophy: https://plato.stanford.edu/entries/probability-interpret, https://plato.stanford.edu/entries/statistics . – Mars Jun 10 '21 at 23:23
  • Related: https://stats.stackexchange.com/questions/549914/are-randomness-and-probability-really-logically-dependent-notions – Galen Nov 01 '22 at 19:35

0 Answers0