This often arises in the discussion of concepts like "what is somebody's Erdos number", or when one discusses what field a contribution belongs to. When "does it count" or when it is or isn't a contribution in a certain area. It's something that I have thought about when these conversations come up.
Personally, what do you think a mathematical paper is if you were to classify one as "one" or "isn't one"?
For example, I have seen people describe papers as mathematical for simply having experimental data in it that happened to just test against a model. Sure Mathematics is involved, but would you call that a mathematical paper? I mean any reasonable scientific paper could be called mathematical if that were the case.
I'm more of the stance that a mathematical paper contains original mathematical results. For example, it presents a theorem (or more than one) and proves it (pretty traditional). In the empirical sense, maybe the paper presents a new model that could be used empirically, and proves theorems for the model?
Where do you stand on this? What in your opinion is what classifies a paper as a "mathematical paper"? Is it the difference between what a mathematics paper is (in the vaguest sense), or is it even more general than this to encompass non-formal results (e.g., just invokes the use of numbers)?