You can do historic comparison. 500 hundred years ago people in Europe acted very differently than they do today. On the other hand their genes didn't change that much.

Or cultural variation is mostly determined by genetic variation. It's hard to empirically distinguish the two.

It is even theoretically possible? If there are causal influences in both directions between X and Y, is there a meaningful way to assign relative sizes to the two directions? Especially if, as here, X and Y are each complex things consisting of many parts, and the real causal diagram consists of two large clouds and many arrows going both ways between them.