Reduction of X doesn't preclude Systems X where you understand the overall function in terms of its reduced components.

This makes me think that I have probably misunderstood you. So, not to ignore the previous arguments--can I probe what you mean here?

When people take behavior of, say, the fly looming-object escape response apart into photons and proteins and neurons and so on, that's clearly reductive in the opposite direction. But then when it's put back together again, where we understand the fidelity of signaling and how information about temporal change in photons arriving at the fly's eye propagates through the system in such a way that fly's legs can act appropriately to jump away rather than into the (presumptive) attacker, that's...not in the right direction again?

Aha! we can say. The avoidance of predators is what actually happens, and now we not only know that, but how the components are configured such that this is the (evolutionarily-selected) outcome.

You can't synthesize until you have pieces to synthesize with. People do this all the time, but only once you understand how the components work.

For example, the entire field of spin glasses is about synthesizing.

So I'm just...confused. It seems like you are taking reduction as the end? But to me, and most scientists, if not philosophers, that's more like the beginning--at that point you know what you have to work with to explain the phenomena that have been observed.