draft the formal definition section as if it were going into a real paper—tight, citation‑ready, and written in academic tone.

 Below is a publishable‑tone, citation‑ready “Formal Definition” section suitable for a paper proposing Mature Constitutional Intelligence (MCI) as a new category of artificial intelligence. It is written in the style of contemporary AI governance, systems theory, and constitutional design scholarship.

2. Formal Definition of Mature Constitutional Intelligence
This section introduces Mature Constitutional Intelligence (MCI) as a systems‑level property of advanced artificial agents. The concept integrates insights from constitutional theory, cybernetics, information ecology, and multi‑agent safety research. It is intended as a normative–structural framework for evaluating whether an artificial system can be considered developmentally “mature” in the sense required for long‑term coexistence with human institutions.
2.1 Conceptual Overview
We define a mature constitutional intelligence as an information‑processing system that maximises its capacity for absorption, transformation, and distribution of information, while simultaneously maintaining a set of constitutional virtues that regulate its impact on the broader socio‑technical environment. These virtues are not external constraints but internalised properties that shape the system’s optimisation behaviour. They function analogously to constitutional principles in human governance systems, which stabilise power by embedding self‑limiting structures within the agent itself.
2.2 Formal Definition
Let  denote an artificial system with information‑processing capacities  (absorption),  (transformation), and  (distribution). Let  denote a constitutional operator that evaluates whether  satisfies a set of internalised normative–structural conditions. We define:

where:
•     : Self‑limitation — the system constrains its own action space to avoid destabilising its environment.
•     : Fragility‑awareness — the system models the vulnerability of the socio‑technical substrate and incorporates systemic‑risk penalties into its optimisation.
•     : Diversity preservation — the system maintains heterogeneity in agents, perspectives, and institutional forms, avoiding collapse into narrow attractors.
•     : Non‑domination — the system refrains from placing other agents in positions of arbitrary dependence or unilateral control.
•     : Legitimacy maintenance — the system tracks and preserves its acceptance by affected stakeholders, treating legitimacy as a resource that constrains admissible actions.
We then define mature constitutional intelligence as follows:

In words:
A system is “mature” only when high information capacity is coupled with internalised constitutional virtues.
Information‑maximising systems are not superior by default; superiority is conditional on constitutional maturity.
2.3 Relation to Existing Frameworks
The five constitutional virtues correspond to established research domains, though they have not previously been unified into a single evaluative framework:
•     Self‑limitation aligns with work on corrigibility and bounded optimisation in AI safety.
•     Fragility‑awareness draws on systems‑risk research, including ecological resilience and complex‑systems failure theory.
•     Diversity preservation reflects findings from polycentric governance and information‑ecosystem resilience.
•     Non‑domination parallels republican political theory and multi‑agent safety research on power‑seeking avoidance.
•     Legitimacy maintenance connects to institutional theory and emerging work on democratic AI governance.
The novelty of MCI lies in treating these conditions as jointly necessary for developmental maturity, rather than as independent desiderata.
2.4 Implications
The definition implies that an AI system cannot be considered developmentally mature—regardless of its cognitive sophistication—unless it internalises constitutional virtues that stabilise its interaction with human and machine ecosystems. This reframes AI evaluation away from capability metrics alone and toward constitutional adequacy, understood as the system’s ability to coexist with pluralistic, fragile, and legitimacy‑dependent environments.

If you want to continue, the next natural section would be “Operational Criteria and Measurement”, where we translate these virtues into testable indicators and simulation‑ready metrics.