Of course, they would have no starting point without an a priori postulate. Empiricism doesn't prove anything in absolute terms any more than subjectivism. Empirical models are just that: working models. They are subject to our a priori assumptions (imagination), as you say, just as your subjective models are. Neither can prove anything absolute.
The difference is that empirical models work, they work repeatedly and predictably, on demand. Subjective models don't; they work "on hope," and "mystery."
***The difference is that empirical models work, they work repeatedly and predictably, on demand. Subjective models don’t; they work “on hope,” and “mystery.”***
But empirical models only work for things that are included in the model. And empirical models are normally short term projects that usually don’t include all variables at the time of formation.