Here is an argument that can be found in some “epistemology” publications: those books and papers, you know, where one is supposed to show he’s read Kant and Popper and will pretend that the stuff he deals with is so difficult that cannot possibly be rendered in plain English.
(Don’t go searching: to find the 1% publications that are beautiful and worth reading, you’d have to sort through a load of futile stuff, serving the sole purpose of adding page count for their authors and publishers).
Things that exist in Nature aren’t complex or simple per se: they merely exist. It’s only after a human observer looks at them, that some such qualifier can be attached to them.
Take, as an example, Poincaré’s 3-body problem. The fact, that is to say, that the system composed of Earth, Moon and Sun can become unstable and unpredictable. Well, the argument goes, neither Earth, Moon or Sun is unstable and unpredictable: it’s only the combination of the three, once we regard them as a system, that behaves that way.
The conceptual process is this: You consider the three bodies as a system, then you design a state space where the behavior of the system in time can be represented, and then you discover that curves in such state space are chaotic. You conclude that the system’s dynamic is chaotic.
However, none of the participants in the artificial system you have studied is chaotic. And it does not even know that it belongs to such system! For example, Earth might believe that she belongs in the Solar System or the Milky Way system: neither of which is the 3-body system you’re ascribing her to. Sun, who is more pretentious, perhaps believes to be a system in itself, with no need or wish to be mixed with inferior participants.
For that matter, if you study the Earth-Moon (or Sun-Earth, etc.) system, you are bound to discover that it never becomes chaotic!
So, what to make of the statement: there exist “systems” that are complex (a.k.a. deterministically chaotic or showing emergent behaviour)?
Do such systems really exist in Nature or do they just belong in our imagination? Or again: should we rather not talk about “systems” but “systems with their observers”?
Some argue that, because of this issue, no “simple” system can be conceived and, effectively, a system should be defined as a set of parts that, when acting as a whole, produces effects that the individual parts cannot. According to this view, there is no such thing as a simple system, and complexity is in itself a definition for the “system” concept.
The issue is tricky. We shall leave it at that for the time being.
Those interested in the idealism vs. realism issue in science, will find a good compendium of it in the first chapters of Roger Penrose’s The Road To Reality, Vintage Books, 2004.
Those who want to think more about the meaning of the systemic approach, especially in relation to complexity and emerging behavior, should read Minati, Pessa, Collective Beings, Springer, 2007.