The methodology has a name because it has a literature. Governance Physics is not a brand applied to a scoring spreadsheet; it is the operational form of a body of published work on what governance is, what it conserves, where its limits lie and what it is worth. This chapter names that work by its public titles and identifiers, says what each contributes to the category, and then describes the research lines still in progress by what they will let a customer do, with direction marked as direction.
#The published work.
#The Physics of Governance: Thermodynamic Limits of Autonomous Agents (10.5281/zenodo.22151454) is the volume behind the methodology's name. Its argument is that the governance of autonomous agents has limits that can be reasoned about the way a physical system's limits are reasoned about. There are quantities that are conserved, costs that cannot be avoided, and states a system cannot reach without paying for them. That is why the engine is computed rather than surveyed, and why its constants are public: a law that depends on who is asking is not a law.
#Ethical AI Series — Conservation of Intent, Volume I (Doctrine) (10.5281/zenodo.18930091), Volume II (Ontology) (10.5281/zenodo.18930283) and Volume III (Limits + Closure) (10.5281/zenodo.18930366) set out the doctrine the platform is built on: that an organization's intent is a thing to be conserved through every transformation of the systems that carry it, the vocabulary needed to say what intent is and where it lives, and the limits of what any system can conserve. Ethical AI Series — Conservation of Intent, Volume IV (Companion): Alignment & Safety Ontology + Intent Preservation (Non-Operational) (10.5281/zenodo.19304640) extends the series to alignment and safety and to the preservation of intent as a property that can be stated without describing any mechanism, which is the discipline this document follows. Intent preservation is among the earliest foundations of this work, and the one the goals of chapter eleven rest on most directly. A conservation law names what may change by naming what may not. An organization that can state the intent it approved, keep it and check it through every change to the systems that carry it is free to change everything else. Posture queried at any moment, a board question answered from evidence, a credential renewed rather than re-surveyed: each follows from that one fixed thing. What is conserved can be checked; what is only remembered must be asked again.
#The Geometry of Governance: Volume III — Topology, Swarms, and the Physics of the GPU-G (10.5281/zenodo.18665381) treats governance as a matter of structure: the shape of a system determines what can be governed in it, and systems of many agents have a geometry of their own. It is the structural view behind the category's concern with integration across domains and with systems of many agents.
#Intent-Aware Continuity in Evolving AI Systems: A Typed, Provenance-Aware Framework for Continuity Judgments in AI Governance (10.5281/zenodo.21539741) answers a question every certification has to answer: when a system has changed, is it still the system that was approved? Its framework for continuity judgments, typed and carrying provenance, is the published basis for renewing a credential on evidence rather than re-surveying from scratch, a direction chapter eleven records.
Three volumes on valuation, Ontology of Valuation (10.5281/zenodo.18663952), Structure of Valuation (10.5281/zenodo.18664069) and Harmonic Law of Valuation (10.5281/zenodo.18664118), connect governance to what an enterprise is worth: what valuation is made of, how it is structured, and the regularities that hold across it. They are cited for what they contribute to the category: a published account of how governance bears on what an enterprise is worth.
The published work is the visible part of a larger body. Behind it stand further volume series and theory that we have not released: on valuation and how governance enters what an enterprise is worth, on the discipline of decisions and the records they leave, on the identity a system keeps or loses as it changes, and on the theory that joins the three. Nothing of their contents is described here, and nothing is cited that cannot be checked. They are named as a class because they are already at work. The offerings of Part II, the research lines below and the bands of the roadmap draw on them, whether a surface is live, in progress or planned, and the depth a customer meets in the product is the depth of that work.
#The research lines, by what they let a customer do. The research lines in progress are described below, each by its benefit; none is described by its method. Where a line is live it says so, and where it is not, a Direction callout says so.
#Projection. A customer can ask what a proposed change would do to the governance posture before making it, and receive the rung it would reach, the dimensions that would move and the gaps it would open or close. The process behind the answer is a computation over the organization's governed knowledge under the engine's public constants; the internals of that computation are not part of the public description and are not needed to act on its result. This line is live as the Projection Engine.
#Governed multi-agent execution. A customer will be able to put a bounded group of agents to work on a pre-defined governance scenario without giving up control. Before anything starts, the customer knows what the run will cost and sets its limit. Each run returns a decision packet and a governance attestation, with its cost accounted for in a record a reviewer can trust. A person, not the agents, makes the decision, and a run that meets a safeguard stops for human review rather than deciding on its own.
#Intent preservation. A customer will be able to check a system that has changed against the intent that was approved for it, and see where the change has drifted from that intent before the drift compounds into a violation. The doctrine is ours, published in the four Conservation of Intent volumes and among the earliest foundations of this work; the capability that applies it to a customer's systems is in progress.
#Executable-opportunity derivation. A customer will be able to ask the operating system what it could do next with it and has not yet done. The answer is a prioritized list of concrete next actions, each ready for an agent to run or a person to fill in, and each showing the knowledge it rests on. It is advisory: a reading of what could be done next, not proof of what has or has not been done.
#The governance agent. A customer will be able to keep its governance picture continuously current from the systems it already runs. A governance agent maintains the graph, checks each change as it happens, and warns early of a violation or of risk to a certified rung. At the higher rungs it holds automated processes and other agents to governance boundaries before they act. The agent acts within limits the customer approves, and what it did and did not do is reviewable. The customer approves what is observed; enforcement begins as advice; people keep the final say.
#Knowledge-graph quality, entity resolution and retrieval. A customer will be able to ask the Knowledge Graph a question and get a complete, evidence-backed answer rather than one with silent gaps because the same concept was held in several fragments. That means duplicate entities resolved into one, each concept held once with all of its evidence together, and provenance kept through every repair and through every refresh of a source document. The benefit is trust that an answer today and an answer tomorrow refer to the same thing. It is also trust that each answer can be traced to its evidence across all seven governance domains, and that nothing becomes more visible by being cleaned up.
#Governance in the deal. An acquirer or a portfolio reviewer asks three questions of a target that today are answered by hand, late and once. The first is the target's governance posture, computed from its evidence rather than surveyed from its data room. The second is the combined entity's posture before close. The third is what a change in posture does to what the enterprise is worth, so that governance enters the price as an input rather than as a judgment applied at the end. The benefit is a diligence answer that the deal team, the board and the regulator can read from the same evidence.
What these lines share is the shape of Part I's definition: each extends the operating system without changing what it is. The kernel stays governed knowledge; the engine stays public physics; the doors stay open to agents and people alike; the evidence record stays the memory; and the credential stays the output. The next chapter is about the discipline that keeps it so.