Substrate Mapping Across Hypotheses

Version	Symbolic Substrate	Latent Substrate	Unified / Integration Layer	Notes / Mechanistic Role
CIv13	✔ Captures discrete rules and abstract patterns from sequences; symbolic encoder extracts regime-specific motifs	✔ Latent encoder captures compressed representations of sequence dynamics	✖ Minimal integration: divergence between symbolic and latent streams signals structural breaks	Focused on detection; divergence (symbolic vs latent) is the measurable signal; no forecasting or adaptation
CIv14	✔ Symbolic encoder continues to extract minimal generative programs; outputs φ-scored sketches	✔ Latent encoder continues to capture compressed dynamics	✔ Decompression/forecaster integrates symbolic and latent streams; ΔBDM and ΔMDL combine with divergence to produce regime probability	Focus shifts from detection → dynamic modeling; system uses both substrates to forecast next sequences and explain regime changes; interpretable symbolic outputs exposed to stakeholders
CIv15	✔ Symbolic programs become self-editing, stored in versioned program library; supports causal perturbation and counterfactual testing	✔ Latent encoder continues to capture compressed trends; used for divergence + forecasting	✔ Autopoietic loop unifies symbolic + latent + action evaluation; controller edits programs based on ΔMDL, φ, and utility	Focus moves from dynamic viability → self-maintaining, planning, open-ended learning; full neurosymbolic autopoietic integration; actionable outputs for downstream systems

Perfect — let’s map CIv13 → CIv14 → CIv15 onto the substrates (symbolic, latent, unified) so you can clearly see how each hypothesis incrementally engages each representation layer. I’ll structure it rigorously, mirroring CIv13 style.

---

Key Observations

1. Progression from CIv13 → CIv15

   • CIv13: Divergence detection, substrates separate.
   • CIv14: Forecasting & dynamic adaptation, substrates partially unified via decompressor.
   • CIv15: Self-maintenance and planning, full autopoietic loop unifies symbolic, latent, and decision-making.

2. Substrate Roles

   • Symbolic: abstraction, minimal program encoding, human/interpretable outputs.
   • Latent: compressed patterns, statistical dynamics, anomaly/shift detection.
   • Unified: combination of symbolic foresight + latent representation + internal controller/planning → autopoietic viability.

3. Mechanistic Metrics Across Substrates

   • Symbolic: φ-metric, program length, edit acceptance.
   • Latent: ΔBDM, ΔMDL, forecast MSE.
   • Unified: regime probability, autopoietic viability, action utility.

---