RA-016 · Research Report · 2026-05-16 · DOI 10.5281/zenodo.20234567

Holonic Systems, Boundary Architectures & Nested Autonomy

Cameisha Smith

Read online Download PDF
The Inquiry

The Inquiry: Do independent research traditions — holonic systems theory, autopoiesis, polycentric governance, complex adaptive systems, and social systems theory — converge on the same structural requirement for how complex adaptive systems maintain viability through boundary-governed nested autonomy? If so, has any tradition or combination of traditions produced a computable primitive grammar that operationalizes this convergent insight?

The question is motivated by a striking pattern: five traditions spanning seven decades, rooted in different disciplines (philosophy of biology, theoretical biology, institutional economics, complexity science, social systems theory), each independently discovering that boundaries are the critical architectural element of complex adaptive systems. Koestler (1967) named the holon. Maturana & Varela (1972/1980) formalized autopoiesis. Ostrom (1990) empirically validated polycentric governance. Holland (1995) / Holling (2001) modeled complex adaptive dynamics. Luhmann (1984) theorized autopoietic social systems. Each tradition hit the same wall: the insight remained theoretical because no computable infrastructure existed to operationalize it.

Falsifiable formulation: If any tradition had produced a computable boundary governance infrastructure satisfying all five traditions' requirements simultaneously, the convergence claim would be trivial (someone already did it). If the traditions do not converge on the same structural requirement, the synthesis is forced rather than discovered.

Executive Summary

Five traditions, one structural requirement. The most striking finding across this sprint is the convergence itself. Five traditions separated by discipline, geography, and decades independently discover that complex adaptive systems require bounded, self-coherent subsystems interacting through selective boundaries in nested architectures with genuine autonomy at each level. Koestler (philosophy of biology), Maturana/Varela (theoretical biology), Ostrom (institutional economics), Holland/Holling (complexity science), and Luhmann (social systems theory) never cite each other's work on boundaries as a central concern — yet each arrives at the same structural conclusion.

The convergence is evidence that the structural requirement is necessary, not preferred. If bounded nested autonomy were merely one viable architecture among many, five independent traditions starting from different problems would not converge on it. The convergence suggests that any complex adaptive system — biological, ecological, institutional, social, or organizational — that does not implement this architecture either finds it emergently or fails.

![Figure 1. Five independent traditions converge on bounded nested autonomy as a structural requirement — each contributing a dimension no other provides.](images/rr-016-fig-01.png)

Boundary as constitutive, not merely protective. The deepest shift in this sprint's theoretical arc is Maturana/Varela's contribution (F5, F6): the boundary is not a container wall separating inside from outside. It is the generative process that creates and maintains the distinction. Without the boundary process, the system does not exist. This upgrades Koestler's structural description (the holon has a Janus-faced boundary) to an ontological claim (the holon IS the boundary process). Luhmann extends this to social systems: the system/environment distinction is produced by the system's own operations. Cagle independently implements it in SHACL: "the SHACL shapes graph is not auxiliary documentation. It IS the boundary."

![Figure 3. Three independent traditions arrive at the same ontological upgrade: the boundary is not a container wall — it IS the system.](images/rr-016-fig-03.png)

Empirical grounding from Ostrom. Unlike the other traditions, Ostrom provides empirical evidence. Her eight design principles are not theoretical requirements but observed characteristics of institutions that have actually survived for centuries. The convergence between theoretical requirements (Koestler, Maturana) and empirical observations (Ostrom) is the strongest form of architectural validation available: requirements derived from theory match characteristics derived from observation, from independent starting points.

Medium downward causation as constraint propagation. Campbell (F14) provides the precise classification: higher levels constrain lower levels through boundary conditions (medium causation), not through commands (strong causation) or retroactive selection (weak causation). This is architecturally critical because strong causation destroys lower-level requisite variety (violating Ashby's law) and weak causation captures governance too late. The tighten-only constraint propagation described in the architectural literature is medium downward causation made computable.

![Figure 2. Campbell's three types of downward causation — medium causation (boundary conditions, not commands) is the only mode that preserves lower-level requisite variety while maintaining higher-level coordination.](images/rr-016-fig-02.png)

The rules-in-use convergence. Three independent traditions — Ostrom's rules-in-use vs. rules-in-form (F8), Luhmann's communication vs. intention (F7), Meadows' system purpose vs. stated purpose (F13) — arrive at the same conclusion: the governance reality is what the system does, not what it says it does. This three-tradition convergence establishes that any governance infrastructure must capture behavioral reality (state transformations, actual decisions, observed patterns) rather than declared intent (policy documents, organizational charts, mission statements).

Cagle as independent implementation convergence. Cagle's four-layer SHACL architecture (F17) is the sprint's trigger and strongest contemporary validation. Building from Koestler's philosophical biology and knowledge graph engineering — not from cybernetics, audit practice, or organizational science — Cagle arrives at the same four-layer boundary architecture (interior graph / shapes graph / projection graph / context graph) with the same principles (privacy is architectural not policy-based; holons only read projections; the shapes graph IS the boundary). The convergence from yet another independent tradition confirms that the architectural pattern is determined by the problem, not by the tradition from which the problem is approached.

Abstract

Five research traditions spanning seven decades — holonic systems theory (Koestler, Bertalanffy, Simon), autopoiesis (Maturana/Varela, Luhmann), polycentric governance (Ostrom), complex adaptive systems (Holland, Holling, Arthur, Meadows), and boundary theory (Campbell, Miller, Morin, Cagle) — independently converge on the same structural requirement: complex adaptive systems maintain viability through bounded, self-coherent subsystems interacting via selective boundaries in nested architectures with genuine autonomy at each level. This convergence is evidence that bounded nested autonomy is structurally necessary, not merely preferred. The deepest finding is that boundaries are constitutive — the boundary IS the system, not its container — a conclusion reached independently by theoretical biology, social systems theory, and knowledge graph engineering. Despite this convergence, no tradition has produced a computable primitive grammar for boundary governance. Each describes what governance requires; none provides the computational machinery to enforce it.

"The SHACL shapes graph is not auxiliary documentation. It IS the boundary." — Kurt Cagle (2026), The Ontologist
Findings20
F-RA-016-01 · theoretical-grounding · established
Bertalanffy (1968) established that structural isomorphism across systems is not analogy but formal relationship — the same patterns recur across domains because the same formal problems recur. Three contributions: open systems and steady states (boundary as selectively permeable membrane, not container wall), equifinality (same final state from different initial conditions/pathways), and isomorphism as methodology (cross-domain structural correspondence is method, not decoration).
F-RA-016-02 · theoretical-grounding · established
Koestler (1967) introduced the holon — an entity simultaneously a self-contained whole and a dependent part — with Janus duality (self-assertive vs. integrative tendencies), the cancerous holon pathology (self-assertion overwhelming integration), fixed rules vs. flexible strategies, holarchy as organizationally distinct from hierarchy (nested autonomy, not command-and-control), and arborization (growth by differentiation).
F-RA-016-03 · theoretical-grounding · established
Simon (1962) established near-decomposability — intra-subsystem interactions significantly stronger than (but not zero) inter-subsystem interactions — making holarchic organization computationally tractable, plus the watchmaker argument (Hora's stable sub-assemblies categorically outperform Tempus's flat assembly under non-zero interruption probability) and aggregation (aggregate properties carry nearly all inter-subsystem information; exposing detail beyond the aggregate is architecturally erroneous).
F-RA-016-04 · convergent-validation · lab-originated
The Koestler-Beer relationship is convergent with mutual awareness — both independently identified holarchic organization as structurally necessary, from different disciplinary traditions (philosophy of biology vs. neurophysiology/cybernetics).
F-RA-016-05 · theoretical-grounding · established
Maturana & Varela (1972/1980) established that an autopoietic system is a network of processes that produces the components which constitute the boundary that enables the processes to continue — making the boundary constitutive, not merely protective. Key concepts: organizational closure (self-producing including its own boundary; materially/informationally open but organizationally closed), structural coupling, and autopoiesis vs. allopoiesis.
F-RA-016-06 · theoretical-grounding · established
Varela (1979) specified six criteria for autopoietic organization — distinguishable boundary; boundary produced by internal processes; components produced by internal processes; processes enabled by their products; boundary enables processes; system operates in space defined by its boundary.
F-RA-016-07 · theoretical-grounding · established
Luhmann (1984/1995) extended autopoiesis to social systems — social systems are composed of communications (not people), and the system/environment distinction is constitutive (produced by the system's own operations, not external observation). Three insights: communication (not actors) as fundamental operation; operational closure / cognitive openness; functional differentiation.
F-RA-016-08 · empirical-demonstration · established
Ostrom (1990/2005/2010) demonstrated empirically that polycentric, boundary-governed institutions successfully manage common-pool resources across centuries, and identified eight design principles characterizing enduring institutions; plus polycentricity (multiple decision centers, mutual adjustment not hierarchical command) and rules-in-use vs. rules-in-form.
F-RA-016-09 · convergent-validation · lab-originated
Vincent Ostrom (1973) demonstrated that monocentric governance — a single center of authority — is structurally inadequate for complex systems, independent of Beer's cybernetic argument against command-and-control.
F-RA-016-10 · theoretical-grounding · established
Holland (1995) identified building blocks, internal models, and tagging as the three mechanisms through which complex adaptive systems produce emergent complexity — with building-block systems tending toward minimal sets (26 letters, ~100 elements, handful of amino acids).
F-RA-016-11 · theoretical-grounding · established
Holling (2001) / Gunderson & Holling (2002) identified the four-stage adaptive cycle (exploitation → conservation → release → reorganization) and the panarchy model where adaptive cycles nest across scales with cross-scale "revolt" and "remember" interactions.
F-RA-016-12 · theoretical-grounding · established
Arthur (1994) established that systems with increasing returns exhibit path dependence — small early events lock the system into trajectories neither predictable from initial conditions nor necessarily optimal.
F-RA-016-13 · theoretical-grounding · established
Meadows (1999/2008) ranked twelve leverage points from least effective (changing parameters) to most effective (changing paradigms) — a structural analysis of intervention effectiveness; and: "The best way to deduce the system's purpose is to watch for a while to see how the system behaves."
F-RA-016-14 · architectural-framing · lab-originated
Campbell (1974) introduced downward causation; Emmeche et al. (2000) refined it into three types — strong (commands), medium (boundary conditions), weak (retroactive selection). This report identifies medium downward causation — constraints without commands — as the architecturally correct mode for holarchic governance.
F-RA-016-15 · theoretical-grounding · established
Miller (1978) identified boundary as a first-class subsystem — one of twenty critical subsystems appearing at every level of living organization — not an emergent property or architectural afterthought; the cross-level hypothesis holds that the same subsystem functions recur at every scale.
F-RA-016-16 · theoretical-grounding · established
Morin (1977–2008) established the dialogical principle — complex systems are constituted by permanent productive tensions between complementary-antagonistic logics, not by resolution of opposites — and the hologrammatic principle that parts contain structural information about the whole (each part holds the whole's organizational logic at reduced resolution).
F-RA-016-17 · convergent-validation · lab-originated
Cagle (2026) independently arrived at a four-layer holon architecture using SHACL — interior graph (structurally private), shapes graph ("the SHACL shapes graph IS the boundary"), projection graph ("holons only read each other's projections"), context graph (shared immutable audit trail of boundary crossings) — structurally isomorphic to the boundary architecture the other four traditions require; "interior privacy is architectural, not policy-based."
F-RA-016-18 · convergent-validation · lab-originated
Five independent traditions converge on the same structural requirement — bounded, self-coherent subsystems interacting through selective validated boundaries in nested hierarchies with genuine autonomy constrained by higher-level boundary conditions — and each tradition contributes a distinct dimension no other provides (holonic = what entities ARE; autopoiesis = why boundaries are CONSTITUTIVE; polycentric = how governance WORKS EMPIRICALLY; CAS = how governance CHANGES; boundary theory = how constraints PROPAGATE).
F-RA-016-19 · gap-identification · lab-originated
No single tradition has produced a computable primitive grammar for boundary governance — each describes what governance requires but none provides the computational machinery.
F-RA-016-20 · convergent-validation · lab-originated
The three-fold independent formulation of the rules-in-use principle — Ostrom (rules-in-use vs. rules-in-form), Luhmann (communication vs. intention), Meadows (system purpose vs. stated purpose) — establishes that governance reality is determined by system behavior, not by system declarations.
Concepts19
Holon (self-contained whole + dependent part)Structural isomorphism (cross-domain formal correspondence)Near-decomposabilityAutopoiesisSix criteria for autopoietic organizationSocial systems as communications (not people)Polycentric governance / polycentricityEight design principles (enduring commons institutions)Monocentric governance inadequacyBuilding blocks (CAS)+9 more
Open Questions4
OQ-060Does panarchy cross-scale interaction require lifecycle spec amendment?
OQ-061How does Miller's 20 vs. 9 subsystem count inform primitive sufficiency?
OQ-062Does biological autopoiesis post-2000 extend or challenge the six criteria?
OQ-063Are there additional independent boundary architecture implementations beyond Cagle?
Bibliography23
von Bertalanffy, Ludwig (1968) · General System Theory: Foundations, Development, Applications
Koestler, Arthur (1967) · The Ghost in the Machine
Koestler, Arthur (1978) · Janus: A Summing Up
Simon, Herbert A. (1962) · The Architecture of Complexity
Beer, Stafford (1979) · The Heart of Enterprise
Maturana, Humberto R. and Varela, Francisco J. (1980) · Autopoiesis and Cognition: The Realization of the Living
Maturana, Humberto R. and Varela, Francisco J. (1972) · De Máquinas y Seres Vivos: Autopoiesis, la Organización de lo Vivo
Varela, Francisco J. (1979) · Principles of Biological Autonomy
Luhmann, Niklas (1995) · Social Systems
Ostrom, Elinor (1990) · Governing the Commons: The Evolution of Institutions for Collective Action
Ostrom, Elinor (2005) · Understanding Institutional Diversity
Ostrom, Elinor (2010) · Beyond Markets and States: Polycentric Governance of Complex Economic Systems
+11 more citations