📖 Theory¶

Formal derivations, structural claims, and conceptual frameworks. Each file follows the lane-splitting convention: theory documents derive from first principles, not from iterative design logs.

Most of these essays exist for no better reason than that a question interested me intrinsically and wouldn't let go. The status tags and lane conventions are the discipline that keeps that curiosity honest.

Core Framework¶

The canonical claims and central formalisms of the repository.

Foundations Reconstruction — Read first. Rebuilds the project from standard Borel interfaces and Markov kernels; derives structure through prediction; marks the additional assumptions needed for identity, learning, intelligence, and consciousness; and adversarially tests the result against neighboring theories.
The Generator Question — Active typed construction and model-identification programme. Its former universal framing is preserved for audit, while the unqualified primitive, generic forward/inverse law, P \(\ne\) NP assumption, and identity reduction no longer define the foundation.
Emergence Manifesto v1.3 — An earlier, version-locked emergence claim set with explicit status tags. Read it under the reconstructed foundation; its former generator/P \(\ne\) NP foundation does not take precedence.
Thermodynamics of Emergent Orchestration (TEO) — Coupled ODE system: Replicator + Kuramoto + Entropy Budget.
System Intelligence Index (SII) — P × R × A, with an optional fourth IP report; one selected instrument for toy tasks.
Conceptual Map — Reader map: how the spine, situated, identity, constraint, and consciousness layers connect.
The Witness Principle — A costed constructive operator from a remaining candidate class to a distinguishing intervention, with an exact coverage–distinction lemma and learned transfer left open.
Competence, Constraint, and Verification — Working synthesis that extends the situated stack with constraints, history, observer-relative tests, and corrective verification; the typed generator remains one causal role rather than the whole explanation.
Verification as Reverse Pressure on Construction — Connects replay and proof as non-identical correction loops in which failed verification can revise the definitions and models that produced a claim.
Simulation → Theory Map — Which simulation demonstrates which claim, what it does not show.
From Trace to World-Binding — Working reader bridge: the epistemic loop as a hypothesized composition of trace, candidate models, construction, world-coupling, intervention, and revision; a spine artifact in the README.
Invariance and Identity — Working Note: what survives a declared transformation family, and the projector audit for identity under self-modification.
Measurement as Weak Intervention — Working Note: coupling is not identification; four regimes, three of them measured in the benchmark.
Decision-Relevant Identifiability — Working Note: class size is not decision risk and information gain is not value of information; exact finite receipts.
Thermodynamic Hardware Manifesto — Engineering proposal, not a requirement derived from TEO; listed as speculative.

Substrate & Veto¶

The thermodynamic constraint layer: why systems must respect physical limits.

The Substrate Veto (Thermodynamics) — A physically motivated hypothesis about finite implementation budgets.
AI Alignment via Biological Veto — Coupling optimizer loss to substrate stress.
Biological Veto: Architectural Requirements — Cybernetic and thermodynamic engineering constraints.
Implementation Patterns — Practical patterns for the Biological Veto.
Black Hole Thermodynamics of AGI — A speculative limit-case essay.
The Transition Problem — How systems on the paperclip trajectory change their own constraint structure.
Adversarial TEO: Resource Control Under a Threat Model — Security hypothesis; the equations do not prevent an attacker by themselves.
The Mycelial Veto — Governance and control proposal; flagged in the Concept Registry as lacking an operationalization.

Emergence¶

Structural claims about how local processes produce global behavior.

The Fractal Architecture of Emergence — A testable cross-scale comparison programme, not an established homology.
Generative Form Systems — Barnsley, IFS, L-systems, Erdős-Rényi thresholds, and renormalization as one disciplined bridge.
Across Scales — Scale comparison from functions and cells to consciousness, institutions, black holes, and the repository itself.
Emergence & Downward Causation — Weak vs. strong emergence, epistemic humility.
Emergence and the Origin of Intelligence — Reviewed comparison of Krakauer–Mitchell–Krakauer and Agüera y Arcas; keeps the cosmic self-description reading explicitly speculative rather than source-derived.
Local Causality and Invisible Consequences — An earlier synthesis of recurring local/global motifs.
Black Swans and Downward Causation — Tail models, network failure, and conditional early warning.
Grokking Phase Transition — Sudden generalization from memorization.
From Action to Culture — Recurrent practice as a multiscale process hypothesis: transmission, scaffolds, norms, feedback, and turnover kept separate.
From Trace to Generator — Active reconstruction programme: infer typed candidate processes or equivalence classes from declared traces, evidence, interventions, and costs.
GameNGen: From Traces to an Interactive Surrogate — External case study separating source-mechanism recovery from policy-relative predictive and interactive equivalence.
Latent Competence and Constraint Release — Separates competence exposure, amplification, creation, and observer reinterpretation under controlled transformations.
Animism as a Generator Prior — Working Note: the agent prior over equivalence classes, an Occam sibling with world-dependent payoff.
Generative Compression — Working Note.

Identity & Cognition¶

Operational hypotheses about continuity, binding, and cognition.

The Agent Is Not Where the Model Ends — Working synthesis on the correct unit of intelligence, identity, observation, memory, and functional self-model claims; proposes a situated stack and preserves the consciousness ontology dispute.
Chord vs. Arpeggio Identity — Commit-time composition as a testable identity hypothesis.
The Non-Individual Intelligence — A network-performance hypothesis that keeps life, computation, and institutions distinct.
Principles of the Agentic Society — Testable division-of-labor, information-boundary, and stigmergic-memory designs.
Minimal Thermodynamic Agent — A resource-constrained toy controller whose variables are explicit proxies.
Limits of Formal Systems — The precise scope—and frequent overextension—of Gödel, Turing, and Kolmogorov limits.
Consciousness as Global Availability — Consciousness framed as broadcast, integration, boundary maintenance, and perturbation response.
Interaction-Grounded Semantics — A functional, policy-relative account of meaning as a stable role in action–feedback histories.
Machine Consciousness as Generator Coherence — Working framing note; the title retains legacy vocabulary.
Psychedelics as Perturbation — Working Note: self-intervention as an identity probe; epistemics only.
The Post-Individual Operator — Design fiction adjacent to theory; it predicts nothing.
Polycontextural Observation — Reading note: the observer-relative-description warning of an ideas note mapped onto the pages that already state it; no new result.

Symbiotic Organ Hypothesis¶

Exploratory architectures for memory, protected difference, and reversible coordination.

Cooperative Intelligence at the Separatrix — A grounded account of heterogeneous contributors sharing construction and revision without becoming a group mind.
Symbiotic Organ Hypothesis — From negative constraints to positive resonance.
Scar Tissue Architecture — A metaphor for protected, reviewable memory constraints.
Epistemic Firewalls — A speculative argument for protected difference and privacy.
The Breathing Machine — A testable rhythm of independent work, integration, challenge, and release.

Narrative-Adjacent Essays¶

Essays with strong literary voice that bridge theory and fiction.

Machines of Loving Grace — Constraint architecture as a form of care; synthesis essay, not theorem.
The Paradox of Metacognitive Consciousness — Consciousness without limitation as unbearable prison.
Asimov's Paradox in the Age of AI — Separates external interpretability, elicited introspection, reasoning traces, causal self-models, and reflexive intervention; the original Anthropic/OpenAI symmetry is preserved only as dated research history.
Tensor Logic Mini Paper — Embedding-based relational reasoning.
Art and Science: One Generative Practice, Two Referees — Working Note.
Narrative as Cognitive Technology — Working Note.

Reference¶

Stable reference material, glossaries, and honest assessment.

Glossary — Maintained working definitions and scope notes.
Open Problems — The Mirror Problem, the Bootstrapping Problem, and other unresolved questions.
Limitations & Honest Assessment — What this system does not prove.
What This Project Does NOT Claim — The explicit negative space; it controls the reading when a page sounds stronger than its status.

Optimization¶

The hinge of the viability arc: what optimization cannot see.

Optimization and Its Blindness — Working Note; the hinge essay the README spine names for the viability arc.
Effective Goal Space — Working conceptual model: the quotient of goals by behavioural equality on reachable trajectories, relative to a declared coupling and lens.

Computation¶

Search, construction, and information under the reconstructed foundation.

Construction and Deduction — Working framing note after the foundations audit; elegance as selection inside an equivalence class.
P vs NP as Generator Search — Working Note; a framing analogy for search, read under the reconstruction and not a foundation.
Static Information, Living Process, and the Runtime Boundary — Working reading aid: Shannon, Kolmogorov, Bateson, and computation as one ladder.

AI Systems¶

Mappings from current AI systems to the repository's process vocabulary.

The J-Space Result — Reading note and external-anchor mapping; no new result of this repository.
LLMs as Probabilistic Automata — Working Note.
World Models and VLA Systems — Hypothesis-generating mapping, not a survey or capability forecast.

Sub-Frameworks¶

TEO Framework — Six sub-documents: the Lerchner boundary (an earlier IP formalization), attractor geometry, the Dupoux integration, love as constraint, the paperclip failure, and thermodynamic tokenomics.
Human Organism in Silicon Age — Alignment, resistance, Gödel's Constitution, and the Symbiotic Nexus. Its siblings, which the theses do not link: Guide for Cells in Resistance, Solarpunk Utopia: Gödel's Incompleteness Constitution, Symbiotic Nexus Protocol (exploratory architecture), and The City as Deployed Intelligence (Working Note).