Human L0–L12 Reference Architecture (DESIGN-ONLY)

This is the document. Rendered from the repository at the commit above, with nothing rewritten for the web. A gate re-renders it on every deploy and fails the build if a single byte differs.

This document is design-only reference architecture. This repo contains no organ, language, self-model, reasoning, dreaming, or awareness dataset. Nothing in this file is runnable evidence. Nothing in this file raises any P-level.

Human L0–L12 Reference Architecture (DESIGN-ONLY)

Status: NOT-BUILT · NOT-SCORED · NOT-CLAIMED · DESIGN-ONLY

1. Why this file exists, and why it stays inert

The stacked active-inference architecture is described across two stacks. Only one of them has data in this repository.

stack	data here?	status
Human/reference `L0..L12`	none	design context only — this file
Flagellum motor `Lmotor-5..0`	yes (Wadhwa 2022)	runnable and scoreable

Implementing L0..L12 would produce scaffolding that no gate in this repository can ever test. Impressive-looking unfalsifiable code is precisely what the truth contract exists to prevent, so the human stack is recorded as architecture and not built.

2. The reference stack

Each level would carry: external states η, sensory states s, active states a, internal states μ, blanket b = {s, a} with p(μ, η | b) = p(μ | b) · p(η | b); a variational free energy F used only for belief update; and, where the level can act, an expected free energy G used only for policy selection.

level	domain	would need data of type
L0	molecular / channel / reaction	single-molecule kinetics
L1	cellular homeostasis	intracellular state time series
L2	tissue / metabolic fields	spatial metabolite fields
L3	organ systems	physiological control signals
L4	affect / precision / allostasis	interoceptive + autonomic recordings
L5	sensorimotor control	proprioception, kinematics
L6	perception / world model	multimodal sensory scenes
L7	language / symbolic action	text, speech, discourse state
L8	self-model / metacognition	confidence and error-monitoring data
L9	reasoning / model selection	hypothesis-test traces
L10	social / normative / long horizon	social and institutional records
L11	offline replay / dreaming	internally generated sample traces
L12	awareness boundary	open question — no closed model accepted

None of these datasets exists in this repository.

3. What the architecture legitimately contributes

Nature and this architecture supply a candidate structure: nested inference, typed blankets, per-level clocks, bottom-up evidence and top-down priors/preferences/precision. That motivated the motor-stack design — hierarchy over motors, hazard/survival at the observation boundary, state conditioning.

Nature supplies the architecture candidate. The gate supplies the status. A citation to biology never moves a UNI gate.

4. The scale-ladder rule

A quantity fitted at one rung may not be silently moved to another. A dwell-time shape parameter estimated from 19 holdout motors says nothing about organ control, language, or awareness. The motor stack is an instantiation at one rung of this architecture; it is not evidence for the rest of the ladder.

5. L12 stays open

The awareness boundary is not a deliverable. No closed model is accepted, and the honest program position is a developmental active-inference simulation — a toy world, not a person. Any future work here must first define a falsifiable criterion and then risk it.

6. Forbidden readings of this file

that the human stack is built, validated, or partially demonstrated
that motor-stack results transfer upward to any human level
that architectural correspondence with biology constitutes parity evidence
that G at any level is biologically tested by this repository

See MOTOR-STACK-AIF-SCOPE-RULING.md for what is buildable and scoreable.

sha256 02d1f2f545f21ffa — of the original file, so what was ingested stays checkable.

Human L0–L12 Reference Architecture (DESIGN-ONLY)

How to read this page

Precise — the source document