So, even in the case of the ecosphere, where oceans, atmosphere, rocks, biota, etc., self-produce each other under closure (according to Earth Systems Science/Gaia Theory), it would still happen via molecular synthesis (not mechanical replacement)

We need to distinguish organization and structure. Chemical processes are nomologically necessary to instantiate structurally the autopoietic organization in the physical world, because without them there's no material transformation. But the organization as such is defined without reference to them

No, they didn't have different views on this matter. Mechanical replacement is neither necessary nor sufficient for autopoiesis. They agreed completely on this.

Oh, and here's a primer in relation to the free energy principle (also not sufficient to characterize autopoiesis) philosophymindscience.org/index.php/ph...

academic.oup.com/book/5967

Chapters 5 & 6 www.hup.harvard.edu/books/978067...

For further details, see these three books (in chronological order) + mitpress.mit.edu/978026255140...

What you need is structural transformation+destruction (synthesis). The chemical domain is paradigmatic. Could it happen at a higher level? That's an open question but your system wouldn't qualify (but organs, multicellular organisms and maybe the whole ecosophere might). +

The simplest process that's minimally autopoietic would be an autocatalytic chemical reaction network that produces its own topological boundary (membrane).

That's not synthesis (chemical production/transformation + destruction). Synthesis = breaking + forming chemical bonds to create new molecular structures. (This could conceivably happen with different molecules than we know on Earth, so it is compositionally plastic.)

*or do they come from outside?*

Does the system generate its own replacement elements (as a cell does) or are they do come from outside? If the latter, then the system is allopoietic, not autopoietic. Some elements can be replaced from outside but only into a system that is already autopoietic in the case of cellular life.

No. Foreign elements can/must enter/exit (the system must be thermodynamically open) but it's definining organization (internal self-production network + self-produced topological boundary) must have closure.

See M&V, 78-9: "production" = "transformation & destruction" (instantiated in/by molecular synthesis) within the system's topological unity (instantiated in/by the membrane), which itself results from (auto-)transformation & destruction. Your example doesn't meet any of these requirements.

That's not the issue. A network/system is any definable set of processes but for them to be autopoietic they must be processes of production (molecular synthesis) - yours are not -that exhibit operational closure (every element is synthesized by another element in the system) - yours do not.

Yes.