**Links to**: [[Markov blanket]], [[Maturana]], [[Varela]], [[Cybernetics]], [[Limit]], [[Organism]], [[Organization]], [[Active inference]], [[Predictive processing]], [[Dialectic]], [[Tautology]], [[Self-organized criticality]], [[Simplicity]], [[Complexity]], [[Life]], [[Partially-observed Markov-decision process]], [[Notation in ActInf]]. _Structural coupling_ is what Humberto Maturana and Francisco Varela (1987) referred to as the autopoietic, mutual interaction between a defined (but open) system and its environment, or between two or more systems, where each loosely defined entity triggers _structural_ changes in the other, while maintaining their own organization. It is now widely applied in active inference and predictive processing, often simply referred to as _coupling_. Structural exchange does not transfer information in a linear cause-and-effect fashion, we can think for example of a constraint-based dynamic exchange such as the one provided here by Juarrero (2023, p. 52): >[In] the body’s vascular system ... the vesicles direct and channel blood flow but do not themselves directly transfer energy ... Like traffic roundabouts, the design and condition of arteries and veins nevertheless have consequences. But their “causal power” operates as constraint, not efficient cause. Juarrero also refers to the work of Mossio and Montévil, who present a diagrammatic version of constraint closure:^[Juarrero 2023, p. 94, our emphasis: “In this schematic illustration, the sole independent constraint, C5, is an enabling constraint that shows that biological systems are open systems. ... Constraints C1 – 4 are dependent constraints. As an individual catalyst, C4 directly depends on C2, on which C1, a by-product, also depends. Catalyst C3 depends on C4 and C5. Each of these constraints, however, is also a generative or enabling constraint: catalyst C4 generates C3, which generates C2, which in turn generates C4. Unlike [a] BZ reaction, no individual reaction [here] is autocatalytic: **no individual catalyst catalyzes a reaction that produces that catalyst.** Collective autocatalysis is realized in the overarching cycle of constraints C2 – 4 (outlined as a black oval). The illustration shows **how the cycle as a whole constitutes a coherent unit, simultaneously a dependent and an enabling constraint**. Its formation depends on a **specific sequence of individual catalysts, C2 – 4.** Some are enabling; some are governing, but it takes **the closure of the encompassing loop of constraints** (when constraint C4 directly catalyzes C3) to regenerate the sequence of reactions on which the loop’s very formation depends. By generating and regenerating the constraint interdependencies that **persist and delay dissipation**, such recursive and multidimensional organizations display qualitatively different properties than physical and chemical convection cells. Processes that realize self-constraint are self-determining. They represent a major transition in evolution (Maynard Smith and Szathmary 1995) brought about by recursively organized constraints.” Note also how the Montévil and Mossio diagram resembles the right hand side diagram on agency, by Froese (2023, p. 4), in a paper dealing with the dynamics of motivated agency, too: ![[froese autopoeisis agency diagram.png|400]]] ![[Structural coupling, montevil, juarrero, constraint closure.png|500]] <small>Constraint closure diagram, source Juarrero 2023, p. 95, original: Montévil and Mossio 2015, p. 186.</small> This closure of constraints loops can help us visualize how different elements, when structurally coupled, co-evolve through recurrent interactions that can lead to congruent structural changes. Systems can mutually adapt to and influence each other while preserving their *separate* operational closure. Structural coupling does not _determine_ changes in either system, it is rather the interstice, the processual “space” for context in which potential structural changes occur. In another, more abstract and speculative, presentation, we can understand the constraint-based dynamics of structural coupling through Hui’s observation here: >It is frustrating when cognitivists want to demonstrate that it is possible to reach the spiritual by modifying neural activity, without realizing that it is the symbols that construct the cathedral rather than the mind alone. Noetic recursivity, through which the subject-object distinction is obliterated, is fundamentally a technical recursivity, which complexifies through the course of history and the evolution of symbolic systems. > >Hui 2019, p. 196. &emsp; There remains more to write about this. ### Footnotes