Now my head really hurts, Wildcat!

The two interpretations are on a different logical level. Syntactically, there should be no continuum between different universes of discourse. The problem is linguistic, not empirical. Classical laws have use value and apply to the data to which they have use, not to an assumed totality. You do't need to know how an aspirin works to get rid of a headache, and you wouldn't expect an aspirin to fuel your model carousel.