Is Molinism falsifiable?

One of the challenges in disproving Molinism is whether Molinism is even falsifable. There are certain theories or thought experiments which are resistant to or even impervious to falsification, viz. string theory, Berkeleyan idealism, the multiverse, Last Thursdayism, Monadology, the brain-in-vat. 

In general, there are two ways to disprove a claim. One is empirical falsification. But suppose a claim doesn't even impinge on empirical reality? 

The alternative is logical falsification. Is the claim coherent? 

Take Monadology. From what I can tell, that's impervious to empirical disproof. But suppose, as Leibniz originally formulated his claim, that Monadology suffers from subtle internal inconsistencies. Would that disprove it? It may not be that simple.

If Monadology is an imaginary construct to begin with, then it may be easy for an ingenious defender to make it coherent by adding some imaginary refinement. Because an imaginary construct has so much play, it's open to considerable modification. 

A critic raises an internal objection to Monadology. A defender fixes it. The critics raises a new objection. The defender introduces a new fix. For every objection, the defender can add another refinement.

Molinism may be too flexible to disprove. But if so, that poses a different kind of problem for Molinism. It's unfalsifiable at the expense of being unverifiable. If it's ultimately unfalsifiable, that's because it's an imaginary construct, which is subject to endless imaginary refinements. 

But that makes it unreal to begin with. And every subsequent refinement is ever further removed from reality. Like a literary tradition which is fictional to begin with, then subsequent authors keep rewriting the story. Each fictional variant gives rise to another fictional variant, like the many permutations of the Star Trek canon.