Epiphenomena, emergence, holarchy (Relation is the smallest unit reprise)

I promised (on FB) to post today and so here I am: in the hospital getting freaky with it (don't ask) but here nonetheless, and in California at least it's still May 7th! This'll be a short one, taking off from a comment on a comment on Facebook posted by a new friend and interlocutor, Paul Rossi. Regarding the axiom "relation is the smallest unit", Paul's friend Juno Moon pointed to the distinction between "epiphenomena" and "emergent phenomena", suggesting that while in the first case the direction of causation heads from "part" to "whole" - and never vice versa - in the latter case the "whole" might be said to precede and cause the "parts". For some reason my dim memory (of undergrad readings in philosophy of mind?) had assigned "epiphenomenon" a different meaning, one that was more synonymous with "emergent" phenomenon, but it does seem that the concept of the "epiphenomenon" implies an effect that is, indeed, a mere "side effect", one that cannot have any impact on the "fundamental" elements in play. The most common example of this is the mind: if the mind is merely an epiphenomenon then the mind can have no impact on the "fundamental" parts that cause it - the brain, the senses, the neurons, etc. If, on the other hand, it is an "emergent phenomenon" then the mind itself exists only in the relations: in this case you have to have enough neurons, enough brain mass, enough relations - perhaps - to get to this emergent mind. If we introduce the axiom that Two Descriptions Are Better Than One into this mix, then perhaps we can see that epiphenomena and emergent phenomena are two different descriptions of the same thing. Can both be true? Undoubtedly, although they contradict each other! Now we're talking. Of course, the axiom Relation is the Smallest Unit seems to refute in principle the methodological reductionism implied by the concept of the epiphenomenon.
Urg. I have the feeling that my last few posts are needlessly overcomplicated. I suppose I can hold onto the hope that they are complex rather than complicated... In any case, I'd like to point here in the end to one of my favorite distinctions, Arthur Koestler's "holarchy". Unlike the "holism" that often comes to mind as the opposite of reductionism, "holarchy" confuses in the first case any facile distinction between parts and wholes: in a holarchy, there are no parts and no wholes, but only "holons", that is, partsy wholes and wholsy parts. A holon is part of a larger whole and contains its own parts, which are themselves both wholes and parts. And in the end? It's turtles all the way down, of course!
Until the 11th!