They have no distinct ontology; their ontology consists of the relationship of their constituents.

It seems like you don't believe that configurations are relevant ontologically, and that you discard predictability of temporal evolution and compressibility of descriptions of states as also beneath ontological consideration.

Is this correct?

So if someone says, "the interesting part isn't the simple rules that the components follow, but the consequences of their configuration," your intuition is: "No! If it's the same old stuff in some new configuration, I'm not interested, philosophically."