Doesn't this imply that, while you can invent all the axioms you like, you must discover which ones are consistent with each other and with experimental results.