Professor Ellis, thank you for your reply.
From your answer, it seems to me (just as a participant in this essay contest) that there may be a specific hypothesis to test this idea about a system of systems-- that a hierarchy with many levels of structure built upon each other has a different description and vocabulary suitable for each level of the hierarchy that is related to the effective entities which occur at that level.
And, that a selection from possibilities at a higher level can, by changing context, change the possibilities for a lower level-- perhaps making some possibilities at the lower level impossible; perhaps changing the probability for a possibility to be selected at that lower level.
As an example, you have talked about "cell" as one such description. I would use hypersets to say: "cell = (cell)." It is using a computer science concept called "coalgebra" or "stream." Specifically when you use this language, say, to talk about information produced by the cell: "cell = (signal, cell)."
Today, there is some talk by physicists about "consciousness." Classically this may be self-awareness, or "consciousness of self." It does seem to me that to be conscious, one must be conscious of something. So before consciousness, there must first be "self," which again I would phrase in terms of hypersets as: "self = (self)."
Then there can be language, for example, about thinking, and the beliefs which are based on thinking: "self = (thinking, self)".
Using such language for the "higher level module" of interest, how would it emerge from lower level systems?
Again to use concepts from computer science, I think of a single transition in a Petri net, with a single output arrow to a single "place" in the Petri net, and another arrow from that single place-- an input arrow back to the same transition. The label for this Petri net place is "self."
So in terms of this picture, the question of how "self=(self)" emerges from lower level systems or lower level modules is then "How does the feedback loop in the above Petri net get going?"
The feedback loop would then comprise those "lower level modules" of which you speak-- in this case ion channels.
An experiment:
Radioactively label the ions. Study a pregnant laboratory animal. One hypothesis is that the feedback on which "self = (self)" emerges is the breath, involving specifically the proprioceptors involved in breathing and their ion channels. On the birth of the laboratory animal, when it first breathes, what can be seen with neuro-imaging?
In opposition to this hypothesis, test for the same kind of images, but for the ion channels for the interocepetors required for the heart to beat.
According to this language of Petri nets, a feedback loop is required for "self=(self)" and it may involve either the interoceptors or the proprioceptors. And-- the respective ion channels.
(It seems to me-- again just as a contestant-- that the KAVLI Institute might be interested in funding such an experiment. If you were a principal.)
Best Regards,
Lee Bloomquist