I am not sure if we are "arguing" at cross purposes here and are stretching the term "empirical" a tad too far. Sticking with the Pythagorean Theorem, it is first and foremost derived from axioms and geometric proofs. Being Euclidean, it was not discovered by physical observations or experimentation. All Euclidean right angled triangles have the same properties i.e. the prediction that Pythagoras Theorem makes about them holds true in all cases. It is not like the measurable physical phenomena such as those observed in say, astronomy or biology, because these phenomena are not necessarily fixed and immutable. (Think of inductive reasoning and Popper's Falsification of Science theory here). The properties of right angled triangles on the other hand are fixed and immutable...they have abstract mathematical relationships and constructs that do not change. Now, of course, we can make or draw right angled triangles and use these physical objects to demonstrate the veracity of the theorem, with physical measurements and observation but that would just be redundant anyway because we already have the theorem which works in all cases.
I should make it plain that I am not making any epistemological value judgements on empiricism v rationalism, here.