How to Theorize about Statistical Evidence (and Really, about Everything Else): A comment on Allen