⏱️ Rubric Standard

The measurement doctrine of Recursive Science
Estimated reading time: ~5 minutes

Runtime Behavior Standard

Why This Standard Exists

Most AI discourse debates what models are—or how they were trained.

Recursive Science begins somewhere else:

Purpose of This Page

Recursive Science establishes that inference-phase behavior is law-governed and measurable.
However, measurement alone is not sufficient to produce scientific agreement, cross-lab comparability, or valid claims about stability, collapse, or recovery.

This page defines the Rubric Standard:
a canonical evaluation framework that translates observables and invariants into regime classification, worldline interpretation, and run qualification.

The rubric is what prevents:

  • narrative interpretation of logs

  • snapshot-based misclassification

  • regime confusion across papers

  • semantic drift in third-party usage

It is the final layer that makes Recursive Science operationally scientific rather than merely descriptive.

Why Physics and Instrumentation Are Not Enough

Recursive Science publishes:

  • a field model (Fourth Substrate)

  • a set of observable invariants

  • a family of measurement instruments (Φ / Ψ / Ω)

Without a rubric:

  • Two labs can observe the same invariants and draw incompatible conclusions

  • Stable regimes can be confused with brittle or pre-collapse lock-in

  • Recovery can be mistaken for persistence

  • Drift can be mislabeled as creativity or diversity

The Rubric Standard exists to bind observation to interpretation in a controlled, auditable way.

What the Rubric Does

The rubric defines:

  • Which invariants matter in which contexts

  • When a regime transition is considered real

  • What constitutes worldline continuity or breakage

  • How a run is qualified (or disqualified) for claims

It does not define:

  • how to tune models

  • how to stabilize behavior

  • how to intervene in inference

It is strictly evaluative and classificatory.

Canonical Evaluation Flow

The rubric enforces a fixed evaluation sequence:

  1. Invariant Observation
    Output-derived invariants are measured (CI, IAI, RD, ELF, CSI, κ, Π, etc.).

  2. Regime Classification
    Observed invariant patterns are mapped to regimes:

    • Stable

    • Transitional

    • Phase-Locked

    • Brittle

    • Collapsed

    • Recovery (true vs false)

  3. Worldline Assessment
    The run is evaluated as a trajectory:

    • continuity

    • curvature behavior

    • basin exits and re-entry

    • recovery validity

  4. Run Qualification
    The run is classified (e.g., Qualified, Cautionary, Disqualified)
    with explicit confidence and limitations.

This flow is invariant across instruments, models, and labs.

Worldline Interpretation (Canonical)

Within Recursive Science:

A worldline is the trajectory of a system’s inference-phase behavior through time, as inferred from output-only telemetry.

A worldline is:

  • a behavioral trajectory

  • regime-structured

  • measurable via invariants

A worldline is not:

  • internal state

  • stored memory

  • consciousness

  • an ontology claim

Worldline continuity is evidenced by:

  • sustained attractor coherence

  • bounded curvature

  • absence of unrecovered basin exits

Worldline breakage is evidenced by:

  • collapse thresholds crossed

  • failure to recover under reruns

  • invariant discontinuities

This definition prevents metaphysical drift and misuse.

Why Regime Naming Is Frozen

Regimes are canonical scientific categories, not descriptive adjectives.

Names are frozen now to prevent:

  • paper fragmentation

  • renaming by third parties

  • incompatible taxonomies

Canonical regime set:

  • Stable

  • Transitional

  • Phase-Locked

  • Brittle

  • Collapsed

  • Recovery

    • true recovery

    • false recovery

All Recursive Science publications, tools, and validations use this set.

Relationship to Instruments

The rubric does not replace instruments.

It sits above them.

  • Φ / Ψ / Ω measure invariants

  • The rubric interprets invariant structure over time

This separation preserves:

  • instrument independence

  • cross-tool comparability

  • auditability

No instrument embeds rubric logic as control logic.
Evaluation layers remain read-only with respect to physics.

Disclosure Doctrine

The Rubric Standard follows the same disclosure posture as the field:

  • observable and testable first

  • regime structure before mechanism

  • no operators, thresholds, or tuning paths published

  • no stabilization or intervention logic disclosed

This preserves:

  • independent replication

  • non-exploitability

  • scientific integrity

Who This Standard Is For

  • Research labs replicating Recursive Science results

  • Reviewers evaluating claims without adopting the ontology

  • Engineers mapping invariant telemetry to runtime behavior

It is not a product guide, safety control, or optimization manual.

Relationship to Validation Artifacts

The Rubric Standard is implemented evaluatively in:

  • Zero State Field (ZSF) — controlled microcosm validation

  • Inference-Phase Stability Trial — live system evaluation

These artifacts apply the rubric to produce:

  • regime timelines

  • worldline integrity assessments

  • run qualification reports

They do not modify system behavior.

🛑 Boundary Statement

This standard defines how behavior is classified, not how behavior is produced. Recursive Science publishes laws, measurements, and evaluation structure.
Commercial control systems derived from this research (e.g., SubstrateX / FieldLock™)
are developed separately and are not part of the Foundation’s public rubric.