sloonz answers What is the best way to talk about probabilities you expect to change with evidence/​experiments?

[sloonz]

I think you’re trying to point towards multimodal distributions ?

If you can decompose P(X) as P(X) = P(X|H1)P(H1) + … + P(X|Hn)P(Hn), and the P(X|Hn) are nice unimodal distributions (like a normal distribution), you end up with a multimodal distribution.