Heterogeneous Capability-Action Systems Framework (HCAS)

Human productive and institutional systems have always been heterogeneous. They combine humans, animals, tools, machines, information, resources, infrastructures, rules, organizations, and institutions. Their capabilities arise not merely from the properties of their components but from how those components are composed. HCAS examines what happens when an important capability within such a system changes substantially. Its central question is: What did the old scarcity force the system to look like, what becomes possible when that scarcity changes, what remains constraining, and what new constraints arise as the system adapts? HCAS distinguishes four analytical problems: 1. Structural Recomposition Potential — how much architectural change should a capability shock plausibly enable? 2. Transition Realization — which technically possible changes are systemically feasible, and which are actually adopted? 3. Endogenous Consequences — what new constraints and vulnerabilities arise because transition succeeds? 4. Adaptive Capacity — can the institution recognize and respond to these changes before conditions change again? HCAS is intended primarily to generate pressure forecasts rather than deterministic outcome predictions. It seeks to identify likely locations of constraint migration, interface change, authority redistribution, capability loss, assurance pressure, and governance demand while recognizing that institutional outcomes remain conditioned by economics, law, politics, culture, standards, ownership, path dependence, and other environmental factors. Artificial intelligence is an unusually consequential application because the capability changing rapidly is cognition itself, understood here functionally rather than phenomenologically. Artificial systems can increasingly participate in sensing, interpretation, reasoning, planning, execution, evaluation, and system redesign. At the same time, architecture-relevant AI capability may change faster than organizations can redesign, validate, govern, and institutionalize their responses. The resulting problem may therefore be less a one-time transition toward an “AI operating model” than a condition of recurrent recomposition under overlapping capability shocks.