This is the story version. For the formal academic version with full theoretical apparatus, read the technical version.

A model that hides its weaknesses isn’t honest — it’s performing. Every Tensions and Limits section in this document was written to earn trust through transparency, not to apologise. These ten open questions are the ones I carry with me. They are not flaws in the architecture. They are the frontier — the places where the map ends and the unexplored territory begins.

The unresolved questions this architecture carries honestly:

  1. Falsifiability. Friston’s FEP is unfalsifiable at the principle level. Only process theories derived from it (predictive coding, active inference) are testable. Which level do this architecture’s claims operate at?

  2. r = 0.72 as correlation, not causation. Wilson’s plasticity finding demonstrates correlation between social support and prosociality. The causal direction — do supportive environments create prosocial people, or do prosocial people create supportive environments? — is not established. Both are likely true (bidirectional niche construction), but the formal causal claim is unproven.

  3. Bach’s consciousness claims are untested. MicroPsi was demonstrated in limited AI agent environments, not validated against human cognitive data at scale. The modulator model is computationally grounded but empirically unconfirmed.

  4. Bidirectionality. Does the architecture shape behaviour, or does behaviour shape the architecture? The niche construction framework (Chapter 4, Section 4.7) argues both, but the formal bidirectional dynamics have not been modelled.

  5. The Punishment Paradox. Nowak says winners don’t punish. Wilson’s CDP 5 says graduated sanctions. The proposed distinction — retaliatory punishment (anti-values) vs proportional correction (values-driven) — is proposed but not formally established mathematically.

  6. Can the Inversion be quantified? Is there a measurable index of cooperation collapse that tracks the inversions described in Chapter 9?

  7. Does the synthesis cover all manipulation techniques? Is there a manipulation technique in active use that these six frameworks do NOT explain?

  8. Neurodivergent development pathways. Friston’s aberrant precision account of autism has significant implications for the ESM. Does the architecture require stage-specific modifications for neurodivergent populations?

  9. Stage 6 in adversity. Can the Transcendent configuration sustain under hostile conditions, or does it require the Conditions (Chapter 4) to be supportive? Wilson’s CDPs suggest the latter — but contemplative traditions claim the former.

  10. The recursive problem. Using this architecture to judge others is itself an Inversion. The architecture describes the trap of ranking people by configuration stage. The architecture also assigns people to configuration stages. This recursion is acknowledged, not resolved. Common Humanity (Chapter 8) is the compass point that addresses it — but philosophical address is not logical resolution.

Question 8 is the one that lives closest to me. I built this architecture from inside a neurodivergent brain, using neurodivergent pattern-recognition, through a developmental pathway that looked nothing like the neurotypical version. The model needs to account for that — not as a footnote, but as a structural consideration. The fact that I haven’t resolved it yet doesn’t make the model weaker. It makes the model honest. The map is drawn. The territory keeps revealing itself.

Ethan Seow. Verixiom Pte. Ltd. First conceptualised 2017 (Emotional State Model). Integrated 2026.

以诚守正, 以恕同尘, 以拙成大, 以悟归空

With sincerity, hold the correct path. With forgiveness, walk among the ordinary. With foolish boldness, achieve greatness. With awakening, return to emptiness.