Introduction. A system that senses and models still has to decide which of its mutually incompatible actions is expressed. This is action selection. The chapter says the five-system diagram does not name it.
The vertebrate solution, as the chapter states it.
- The output nuclei of the basal ganglia are tonically inhibitory. Their resting condition is to suppress.
- The default is suppression of the available actions, not quiescence.
- Selection happens when inhibition is released from one candidate while the others stay held back.
The chapter says these terms are borrowed and must be stated accurately. Tonic inhibition, release from inhibition and the homology claims each need checking against named sources, not “carried along together”.
What changes for the VSM. The regulatory problem changes shape. It is no longer how a command center issues an action. It is how a field of possible actions is held back, compared and selectively released. The chapter says this is a different architecture from “a pair of eyes wired to a vertical command line”, and that it is not obvious where the VSM puts it.
The 560-million-year claim. A claim is in circulation that this logic has been continuous for about 560 million years. If it holds, action selection is an ancient organizing problem, not a late refinement of cognition. The chapter records it as under verification and says the number comes out if it is not supported.
Important authors. Sten Grillner and Marcus Stephenson-Jones, named by the chapter as “the starting point we have been given for the comparative lineage”. Grillner is a neurophysiologist known for work on vertebrate motor control. The chapter gives no titles.
Importance for cybernetics / VSM. It tests whether the five-system partition is complete. If action selection has to be placed inside System 4, System 4 may become “an unconstrained remainder that absorbs whatever the other four do not cover”.
Importance for the chapter. Carries “Action selection: the function the diagram omits” and item 5 of “The receipt: what falls with the substrate” (the five-system partition is complete). It also bears on the closing claim that a reconstruction could give “a more exact account of modeling, prediction, and action selection”.
Sources in the reading list.
- Grillner and Stephenson-Jones, comparative work on the basal ganglia — for tonic inhibition, release from inhibition, homology and the dating. No title given in the chapter †.
- Primary sources and reviews for each of the three claims separately — none named ✗.