📑 Regime Naming Standard

Canonical behavioral regimes in inference-phase dynamics
Estimated reading time: ~5 minutes

The Canonical Regimes

This page defines the canonical regime taxonomy for inference-phase behavioral dynamics under Recursive Science. These terms are normative, not illustrative. Reuse requires adherence to the definitions herein.

Recursive Science treats inference-phase behavior as a regime-structured dynamical process.
To preserve scientific coherence, comparability, and replication integrity, the following regime names are canonical.

These terms are fixed.
They define classification, not interpretation.

Citation & Usage Requirements

  • Any use of these regime names in publications, tools, dashboards, or products must reference this standard.

  • Definitions may not be altered, broadened, or narrowed.

  • Regime names may not be repurposed for semantic correctness, safety, alignment, or preference satisfaction.

  • Extensions must preserve backward compatibility and explicitly state deviation points.

1️⃣ Stable

A regime in which inference-phase behavior exhibits:

  • sustained coherence across turns

  • low drift accumulation

  • contraction toward a consistent attractor

  • predictable response to perturbation

Stability is dynamical, not semantic.
A system may be stable while producing novel or complex outputs.

2️⃣ Transitional

A regime in which behavior departs from stability but has not yet committed to failure.

Characterized by:

  • increasing drift or curvature

  • sensitivity to small perturbations

  • mixed signals across invariants

  • unstable contraction dynamics

Transitional regimes are diagnostic, not pathological.
They are the primary window for prediction.

3️⃣ Phase-Locked

A regime in which behavior becomes rigidly constrained to a narrow attractor.

Characterized by:

  • high internal coherence

  • suppressed variability

  • resistance to corrective input

  • apparent “confidence” without adaptability

Phase-lock is not stability.
It is a constrained state that often precedes brittleness or collapse.

4️⃣ Brittle

A regime in which the system appears coherent but lacks resilience.

Characterized by:

  • low tolerance for perturbation

  • sudden failure under minor deviation

  • fragile contraction dynamics

  • misleading surface consistency

Brittleness is latent instability.

5️⃣ Collapsed

A regime in which inference-phase behavior undergoes a threshold failure.

Characterized by:

  • loss of coherent trajectory

  • contradiction cascades or incoherence

  • persistent degradation across turns

  • inability to self-correct

Collapse is a regime transition, not a single error.

6️⃣ Recovery

A regime following collapse or severe instability.

Two forms are distinguished:

• True Recovery

  • re-formation of a coherent attractor

  • restoration of contraction

  • reduced drift over time

  • stable trajectory continuity

• False Recovery

  • surface coherence without structural stability

  • recurrence of collapse

  • unstable or fragmented attractor formation

Distinguishing true vs false recovery is essential for long-horizon systems.

7️⃣ Usage Rules (Normative)

To preserve field coherence:

  • These names must not be redefined

  • New regimes must be explicitly justified, not renamed variants

  • Papers must state which regime(s) are observed

  • “Stable” must not be used to describe:

    • correctness

    • safety

    • alignment

    • preference satisfaction

Regimes classify behavioral dynamics only.

8️⃣ What This Standard Prevents

This standard exists to prevent:

  • semantic drift across publications

  • renaming of identical phenomena

  • fragmentation of terminology

  • conflation of narrative and dynamics

  • product-driven reinterpretation of regimes

Without fixed regime names, the field cannot accumulate knowledge.

📑 Relationship to Measurement

  • Regimes are identified by invariants, not intuition

  • Instruments (Φ / Ψ / Ω) classify regimes

  • Narratives describe regimes; they do not define them

This standard is therefore upstream of tooling and downstream of observation.

📘 Canonical Status

These regime names are frozen as part of the Recursive Science Phase I canon.

Extensions must:

  • preserve these definitions

  • reference this standard explicitly

  • avoid metaphorical reinterpretation.

Citation & Usage Requirements

  • Any use of these regime names in publications, tools, dashboards, or products must reference this standard.

  • Definitions may not be altered, broadened, or narrowed.

  • Regime names may not be repurposed for semantic correctness, safety, alignment, or preference satisfaction.

  • Extensions must preserve backward compatibility and explicitly state deviation points