v0.4.2 — Zenodo DOI archival

Adds .zenodo.json metadata for Zenodo integration. No code changes from v0.4.1.

v0.4.1 — Replication data and recursive audit

Includes replication data for all three analysed sessions (Grunch, Echo, ZoryaGPT), RAA-SC-001 recursive adversarial audit (Phase A + B + convergence), and lock-in language qualification. No changes to analysis code from v0.4.0.

v0.4.0 — Public Release

First public release of the Semantic Climate Phase Space library.

What's New (since v0.3.0)

Features

• SafetyGate — Observation-only coupling feedback system with 4-tier gate output (LOW/MODERATE/HIGH/CRITICAL). Evaluates after each turn, logs to session history, exposes via /api/gate/status. Does not inject into LLM context yet (reversible architecture).
• Hybrid affective substrate — GoEmotions integration for richer emotion detection beyond VADER sentiment
• Configurable LLM settings — Temperature and system prompt now configurable per session
• ECP session context — Accept Ecological Coherence Protocol session context for experiment correlation
• Frontend trajectory visualisation — Real-time trajectory viz in the web app with performance improvements

Fixes

• Remove causal language from cross-correlation lag interpretation — notes instrumentation asymmetry (1Hz cardiac vs turn-boundary semantic) rather than claiming temporal precedence
• Map EBS entrainment field to coherence for biosignal display

Documentation

• Comprehensive system docs updated for modular architecture
• .env configuration instructions added to README
• Basin count references corrected (7 → 10)

Infrastructure

• 272 tests passing
• Root requirements.txt for core library
• ESL-A v0.1 license, CITATION.cff, ATTRIBUTION.md all present

License

Earthian Stewardship License (ESL-A) v0.1 — non-commercial use permitted for research, education, and community projects. See LICENSE for full terms.

v0.3.0 — Pre-release: Movement-Preserving Semantic Phase Space

v0.3.0 — Pre-release: Movement-Preserving Semantic Phase Space

Status: Private research pre-release
Intended audience: collaborators, reviewers, future maintainers
Relation to publications: Companion implementation to forthcoming Springer chapter
and preprints on Entangled Cognition, Semantic Climate, and Trajectory Integrity.

This release marks the first internally coherent instantiation of the Semantic Climate
Phase Space as a movement-first semantic analysis framework. The system treats meaning
not as a sequence of static states, but as trajectory through a constrained phase space,
with explicit attention to path-dependence, hysteresis, and temporal integrity.

This is not a public API release. Interfaces, naming, and internal structures remain
subject to change as theoretical and empirical work continues.

---

Conceptual Contributions (Primary)

The following elements constitute the core intellectual and architectural contributions
of this project:

• Movement-preserving semantic framework

  • Meaning modeled as trajectory rather than pointwise state
  • Classification preserves how a configuration was reached, not only where it is

• Trajectory Integrity

  • Coherence defined as the relationship between a path and its own history
  • Path-dependence, memory without lock-in, and gestalt structure treated as first-class
    analytic properties

• Hysteresis-aware semantic dynamics

  • Explicit separation of entry and exit thresholds
  • Resistance to spurious state-flipping under noise or short-term perturbation

• Phase-space architecture

  • Semantic dynamics represented via velocity, acceleration, curvature, and entropy shift
  • Soft membership across attractor basins rather than hard classification

• Interpretive stance

  • Instrumentation designed to constrain interpretation without reifying metrics
  • Intended for use alongside first-person accounts and contextual analysis

These contributions are authored and integrated by Mathew Mytka (2025) and form the conceptual
basis for related academic publications.

---

Core Metrics (With Repairs and Reinterpretation)

This implementation incorporates semantic metrics originally proposed by
Daria Morgoulis (2025), with mathematical corrections, reinterpretation, and embedding
within a trajectory-based system:

• Semantic Curvature (Δκ)

  • Frenet–Serret local curvature applied to semantic trajectories

• Fractal Similarity (α)

  • Detrended Fluctuation Analysis (DFA) over semantic velocity
  • Corrected scaling behaviour and validity constraints

• Entropy Shift (ΔH)

  • Shared clustering with Jensen–Shannon divergence
  • Interpreted as change in semantic dispersion, not static uncertainty

Metric repair, validation, and integration into the phase-space model were performed as
part of this project.

Architecture Overview

• TrajectoryBuffer

  • Maintains temporal continuity for semantic velocity, acceleration, and curvature

• Movement-preserving classification stack

  • TrajectoryStateVector
  • SoftStateInference
  • HysteresisConfig
  • Classification outputs retain temporal context and path structure

• Attractor basin system

  • 10 canonical semantic configurations
  • Soft assignment via centroid proximity
  • Hysteresis-aware transitions

Web Application (Experimental)

• Multi-provider LLM support
  • Together AI, Ollama, Anthropic, OpenAI
• Real-time semantic coupling visualisation
• Session export and replay
• Optional integration with EarthianBioSense for biosignal coupling experiments

UI components are exploratory and not yet stabilised.

Test Coverage

• 244 tests covering core metrics, trajectory logic, and classification behavior
• Emphasis on regression detection for movement and hysteresis properties

License

Earthian Stewardship License (ESL-A)

This license reflects the project’s orientation toward responsible use, research
stewardship, and regenerative application. Licensing terms may evolve alongside
governance work.

Attribution

• Original semantic metrics:
  Daria Morgoulis (2025), MIT License

• Phase-space framing, trajectory integrity model, hysteresis architecture,
  movement-preserving classification, system integration, and interpretive framework:
  Mathew Mytka (2025)

Attribution reflects conceptual authorship, not only code provenance.