The extent to which science and metaphysics are intertwined (one could almost say bound up) with each other is, at times, astounding – and I think this intertwine-ness comes from language. The conceptual grammar which makes science possibly comes almost entirely from metaphysics – cause, purpose, etc – but are often confused with the science itself. Causality is an example of this – causality is a metaphysical category, yet it’s nearly impossible to say anything having to do with science without using some kind of causal-language. The conceptual grammar becomes inextricably part of the actual science.
Consider a similar, but negative, situation: the elimination of formal/final causes in modern science (initiated, by, say the early moderns). This shows, to me, a remarkable conceptual confusion by eliminating metaphysical categories from empirical study. Formal and final causality, being metaphysical categories, cannot be studied under a microscope – (though biology will often use the language of purpose and form, it’s a decidedly different kind of thing than formal and final causality).
I say all this not to offer a solution to this problem but to simply note the striking manner in which science and metaphysics are conceptually confused – though a solution, to paraphrase L.W., may involve the release of the fly from the fly-bottle. Matters of method (say, the elimination of formal/final causality) are fine kept strictly as a matter of method – to confuse a matter of method with reality, though, is nothing but a conceptual confusion.