Why the terms linear/nonlinear are absent from the text and why use 'information', more useful in digital representation, than Signal/Noise more adequate to the analog world we inhabit (and concepts like negative feedback, transfer function, etc).
The low level description obeys to linear equations, where superposition rules are strict. The higher levels are nonlinear by nature or construction, f.i. hysteresis, losses, noise are non-reversible. Any transistor, diode, flip-flop, magnet, etc are examples. The ADN is digital and coded/decoded with transcription errors included (I take it as a law: a huge amount of data always has a large quantity of errors/noise). The components of all digital devices are, at low level, analog nonlinear components.
About the goal: is it real or apparent? A roly-poly toy, or the navy ships, do they want to stay up ? It appears so but in reality they are forced to stay up by construction. The same can be said about floating icebergs: no designer, no goal,...
My viewpoint is based in IT and electronic eng foundations by formation and practice.