What sits underneath MO§ES™.
Research foundation for AI operator evaluation. The Conservation Law of Commitment and Commitment Theory provide the theoretical basis for measuring how people operate AI systems. MO§ES™ is backed by a formal research framework with papers, experiments, datasets, patents, and falsifiability work.
See the Methodology Build a Bespoke EvalWhat stays binding when language changes form.
Commitment Theory asks a foundational question: when language changes form — paraphrased, translated, compressed, delegated, automated — what remains binding? What part of a signal still constrains what follows?
A commitment is the identity-preserving part of a signal that still constrains what follows. This includes:
- Obligations — what must be done
- Prohibitions — what must not be done
- Permissions — what is allowed
- Modal constraints — what is possible, necessary, or contingent
The theory provides a formal basis for tracking governing state across transformations of language — including AI-mediated transformations where the form of a signal changes but its binding content must be preserved.
C(Tgov(S)) = C(S)
The Conservation Law states that the commitment content of a signal is preserved under governing transformations. Language can change form without losing what it still requires.
The commitment content of signal S — what remains binding.
A governing transformation applied to S — paraphrase, translation, compression, delegation, automation.
The commitment content after transformation. Equal to C(S) under the conservation law.
Plain-language explanation: language can change form without losing what it still requires. The governing transformation of a signal preserves its commitment content. This is the formal basis for maintaining governing state across AI-mediated interactions.
How MO§ES™ uses it.
MO§ES™ operationalizes Commitment Theory through governed commitment-bearing state, provenance, validation, and separation between authoritative state and downstream intelligent systems.
Commitment-bearing state is tracked, versioned, and governed. Transformations preserve commitment content. Provenance is embedded in every record.
Every measurement, diagnosis, and intervention carries metadata declaring its source, evidence, and decision-use status. Nothing enters the system without a declared origin.
Claims are validated against evidence. Diagnoses carry alternatives. Outcome joins require separate validation. The system is designed to be wrong — and to say so.
Authoritative state is separated from downstream intelligent systems. Downstream systems can branch, interpret, and generate — but the authoritative state is preserved.
A formal research framework. Not established science.
Commitment Theory and the Conservation Law of Commitment are a formal research framework with papers, experiments, datasets, patents, and falsifiability work. They are not presented as established science. They are presented with appropriate epistemic humility — as a framework that makes testable claims and is subject to revision.
The framework carries:
Formal papers developing the theory and conservation law.
Empirical experiments testing conservation under transformation.
Datasets for evaluating commitment preservation.
Patented methods for governed commitment-bearing state.
The conservation law makes testable predictions. It can be wrong.
The framework is subject to revision based on experimental results.
Related systems.
Public and enterprise evaluation of AI operators. Operator evals, rankings, metrics, benchmark field, public methodology.
Dual-governance agentic marketplace and economy. signomy.xyz
Applications, Questions, Answers. Turns previous application work into reusable application capital. Builds a governed answer bank, maps answers to canonical commitments and supporting proof, identifies best-fit opportunities, and generates application-specific responses without losing source lineage. mos2es.xyz
Maintaining governing state across conversational and voice interactions.
Preserving authoritative state while allowing downstream intelligence to branch.
Additional systems in active development.
Concepts derived from this research.
The Conservation Law of Commitment and Commitment Theory provide the theoretical basis for the five canonical metrics and the governance framework. Each concept has a dedicated definition page.
(R + W) / I — context reuse and building relative to new input.
O / (I + O + R + W) — productive output share of total token flow.
Signal-to-noise ratio in AI operator token flow.
W / R — ratio of new context built to context reused.
AI Operator Development Index — 0–100, weighted, DEVELOPMENTAL.
DEVELOPMENTAL / HYPOTHESIS / ASSOCIATION labels, no punitive use.