Limitations

This paper is conceptual. It proposes a theory and proposition system but does not provide empirical validation. The construct dimensionality, mechanisms, boundary conditions, and failure modes require empirical testing. The revised proposition system is intended to guide such testing, not to claim that the relationships have already been observed.

The paper also depends on uncertain AI development trajectories. If AI agents remain narrow and low-risk in most organizational contexts, the need for full constitutional architecture may be limited. If AI agents become more autonomous, multi-agent, and embedded in organizational workflows, the theory becomes more central but also more difficult to operationalize.

The constitutional framing carries transfer risks. Concepts such as constitution, legitimacy, jurisdiction, and amendment have political and legal associations. This paper uses them as organizational-theoretical concepts, not as legal prescriptions. Future research should examine cultural and institutional variation, including whether the theory reflects corporate or Western governance assumptions. The paper also does not claim that every organization should formalize a constitution to the same degree. The appropriate level of constitutional formalization is contingent on risk, agent density, regulation, volatility, and governance competence.

Formalizing organizational meaning is difficult. Some principles, roles, and evidence requirements can be represented explicitly; others depend on human judgment, interpretation, and context. Machine-readable provisions may help govern action but cannot replace interpretation. Measurement remains immature, especially for legitimacy, meaningful human governance, representational correspondence, and resilience.

The theory is multilevel. It spans actors, roles, teams, units, enterprises, and sometimes ecosystems. This creates analytical complexity and may make empirical research challenging. The theory also assumes a minimum level of representation quality and governance competence. Organizations lacking these capabilities may need staged development before the full architecture becomes useful.

Finally, the paper preserves a productive tension between adaptation and constitutional continuity. It does not resolve that tension by privileging either flexibility or control. Future work should examine when constitutional architecture enables adaptation and when it becomes a source of inertia, capture, or symbolic compliance.