Robert Rosen talked a lot about the distinction between simple and complex systems. Essentially, simple systems can be built from parts, like a clock. Complex objects, on the other hand, are things with many models. Like time, or music, or a plant. No single description is enough.
He went further, saying that when studying a complex system, one can "throw away the matter and study the organization" to learn those things that are essential.
He was talking about Category Theory and its use in studying the organization of organisms.
A complex system, you might say, is not decomposable into parts. In this manner the complex :: simple distinction is akin to the difference in monoidal :: Cartesian tensor products in a monoidal closed category. Cartesian products have projections, you can take them apart. A monoidal product does not. You can only make inferences about their structure (make models)