Skip to content

Canonical Path v2: From Emergence to Survivability

The current research arc through the repository, after the Viable Corridor / TEO work.

Purpose

This document describes the present canonical path through the repository. It does not replace the older reading path in the project's README, nor the project's primary spine, The Generator Question. It clarifies the second spine that the recent work has sharpened.

Two spines run through this project, sharing one root:

  • The Generator Question (the epistemic spine): emergence is cheap to run forward (generator → trace) and structurally hard to invert (trace → generator). This organizes the project's work on recoverability, identity, and the three walls (P vs. NP, Kolmogorov, Gödel).
  • The Viability Arc (the dynamical spine, this document): what happens to an emergent optimizing system over time, and what does it take to keep it alive? This runs emergence → optimization → constraint architecture → survivability and terminates in the Viable Corridor paper.

Both begin at emergence. One asks whether a system can recover its generator; the other asks whether a system can survive its own optimization. This document maps the second.

The paper is a node, not the project. The Viable Corridor is one citable artifact produced by this repository — a stabilized synthesis at the sharp end of the arc. The repository remains the generative system: a map of related models, essays, simulations, narratives, and open problems. This path is written to keep the paper in proportion.

A note on tags. Each station below carries the status of its strongest content, using the project's vocabulary: [FOUNDATIONAL ASSUMPTION], [DEMONSTRATED], [HYPOTHESIZED], [FORMAL], [CONJECTURE], [HEURISTIC], [EMPIRICAL CONJECTURE], [MODEL ASSUMPTION], [SIMULATION ARTIFACT], [NARRATIVE STRESS TEST], [OPEN PROBLEM]. Mixed nodes are tagged per-claim, not globally.

1. The Generator Question — the shared root

Why the project begins with an inverse problem: a trace does not reveal its generator, and recovering the generator is, in general, intractable. This is the foundational asymmetry the whole repository circles.

The Viability Arc picks up downstream of this: take emergence as given, and ask what its optimizing products do.

2. Emergence and Self-Organization — [DEMONSTRATED] (forward direction)

Local rules produce global structure; the forward direction is cheap and is demonstrated many times over.

3. Optimization and its Blindness — the hinge [HYPOTHESIZED]

The arc's pivot, and (until now) the repository's thinnest-threaded station. Emergent systems contain local optimizers. Local optimization is globally blind: each step is locally rational, the aggregate consequence is invisible from inside. Unconstrained, such optimization outruns the carrying capacity of its substrate — the paperclip pattern, instrumental convergence. This is the problem to which constraint architecture (§4) and survivability (§7) are the answer; it is also where capability loading enters (capability is a single driver that pushes a system against several limits at once).

(This station consolidates existing material; it asserts no new claim. It is named here because the arc is unintelligible without it: constraints answer a question, and this is the question.)

4. Constraint Architecture — TEO [HYPOTHESIZED] / [MODEL ASSUMPTION]

The repository's constraint model: orchestration under a regulatory brake, value coupling, and a thermodynamic substrate budget. The claim is that survivable optimization requires constraint architecture, not a single fix.

5. The Viable Corridor — the stabilized artifact [FORMAL] + [CONJECTURE] + [HEURISTIC]

The formal synthesis derived from the TEO branch — a paper-style node, not the whole project. It proves a necessity result (the conjunction of three constraints), conjectures sufficiency, and advances a clearly-hedged isomorphism hypothesis. Its strongest current contribution is the constraint-architecture framing and the capability-loading result (capability growth pushes a system out of the corridor through multiple boundaries at once; single-axis fixes fail).

  • The Viable Corridor — necessity theorem [FORMAL], sufficiency [CONJECTURE], isomorphism [HEURISTIC] / [EMPIRICAL CONJECTURE], derivations in Appendix A.
  • Evidence: Appendix C (TEO ODE, Class A) and Appendix D (agent-based ecology, Class C) — see §6.

6. Agent Ecology and Alignment-Veto Simulations — [SIMULATION ARTIFACT]

The executable side: do constraints actually matter? Two structurally independent models reproduce the regime claims (hard vs. soft budgets; capability loading; single-axis interventions failing at high capability).

7. Survivability: Substrate, Veto, and the Civilization Interpretation — [HYPOTHESIZED] / [EMPIRICAL CONJECTURE]

The arc's endpoint. The theory of survivability is the veto/viability cluster; its application to civilization is an explicitly heuristic interpretive layer, not a proven isomorphism.

8. Fiction and Narrative Stress Tests — [NARRATIVE STRESS TEST]

Fiction is the project's instrument for noticing when a clean formal claim is impossible to live with. It stays linked but separate from the formal layer.

  • fiction/README.md — full annotations. Substrate-limit pieces (Impedance Crash, The Vital Floor, Dashboard of the Commons) dramatize §3–§7.

9. Open Problems

The honest frontier — what the arc has not closed.

  • Proving sufficiency / deriving \(\gamma_c\) rigorously (not just single-trajectory existence). Paper §3.4, Appendix C.5.
  • Operationalizing \(\gamma_{\text{eff}}\), \(K_{\text{eff}}\), \(D_{\max}\) for real systems; \(K_c\) for societies. Paper §5.4.
  • A finite-\(N\) / general-network version of the coherence (Kuramoto) result. Paper §3.3.
  • Validating P7/P8 on real AI-agent systems (the Appendix D evidence is synthetic). Companion paper.
  • Genuine overshoot-collapse dynamics (delays, endogenous \(D_{\max}(t)\)) vs. the current rate-threshold self-arrest. Paper §6.4.
  • Whether the Viable Corridor should remain in this repo or become a spin-off. (Governance open problem.)
  • See also theory/reference/open-problems.md for the project-wide list (Mirror Problem, Co-Instantiation, Trace-to-Generator reconstruction).

Working Rule — keep the writing modes separate

Do not route everything through the paper. Per the Information Architecture:

  • papers/ — citable, compressed artifacts (the Viable Corridor).
  • theory/ — conceptual development and argument.
  • simulation-models/, lab/ — executable claims.
  • fiction/ — narrative stress tests.
  • logs/ — speculative architecture notes.
  • meta/ — repository governance and epistemic framing (this document).

One source of truth per concept; everything else links to it.