docs: record the correspondence corrections — grounding, faculties, wonder, consolidation #163

2026-08-16T18:32:06Z

will.anderson commented

2026-08-16 18:32:06 +00:00

Documentation only — no .el, .c, .h, .py, .sh, .plist or .yaml file is touched. git diff --name-only is 12 .md files.

What this records

foundation/el, branch design/correspondence-and-censorship, lang/spec/correspondence-and-censorship.md supersedes this repo's architecture docs on four points. Its companion, lang/spec/runtime-ownership.md, addresses a different residual in the same substrate. The corrections are transcribed, not re-derived — every earlier version of this reasoning was wrong in an instructive way and each correction was argued down hard.

Grounding is not a subsystem — it is the edge weight (the hebbian weight). One quantity, not two fields. grounded-by as a relation type should not exist: grounding is a property of a relation, not a relation between nodes, so minting the edge is the error, not merely which endpoints it chose. Never computed on demand, never a score. Traversal is already grounded inference; decision provenance is the path.
Faculties are operations, not parameters. reason changes the estimate (read), induce changes the parameters (the correspondence-beat, which exists and works), abduce changes the structure (write). A write cannot be a parameter of a read.
Wonder is the boundary, not a manifest. Any structure at all has an edge. There are about six wonders, the same for everyone, and they never close. Curiosity is wonder crystallized at a nucleation site — one thing at two phases, which is why curiosity can be satisfied and wonder cannot.
Consolidation is ambient, not scheduled. A brain has no cron job. The presence of a ticker is the diagnostic. soul.el:731's continuous awareness_run() loop is the one fragment with the correct shape; the others fold into it.

Plus: co_registration is deprecated in favour of per-edge GeoEdge.discord, and — in an immutable substrate — any mechanism that refuses a write is either redundant with immutability or an epistemic constraint misfiled as a protective one.

Approach

Stale passages are left standing, marked inline, and pointed at their replacement rather than deleted — per 03 §Immutability and 06 §4.1, supersession is residue: the trail of how the understanding matured is kept, because sometimes the truth was in the old idea even when the old idea was not itself the truth. New 06 §12 carries the transcribed corrections and the measured inventory.

Every factual claim was re-verified against source at file:line before being written. Line references that had drifted are corrected in place and labelled.

Files

file	change
`docs/architecture/06-cognitive-architecture.md`	Top banner + inline supersession markers at 12 sites; new §12 with the four corrections, the edge-vector model, the live code-residue tables, and the measured consolidation inventory
`docs/architecture/03-data-and-memory.md`	§Write-protection superseded; notes that the doc contradicts itself 35 lines later; `cultivate` line ref corrected `:781-816` → `:960`; backup cronjob marked as ticker-but-not-consolidation
`docs/architecture/02-components.md`	`awareness_run` loop marked as the correct shape; curiosity-scan superseded on three counts; stale line refs corrected; `/consolidate` and `mem_consolidate` marked as 2 of 11
`docs/architecture/00-overview.md`	Curiosity is not a peer Engine — one activation process, two seed sources
`docs/architecture/01-vbd-decomposition.md`	Axis 2 curiosity-rotation Engine superseded; its deep git-archaeology re-read as evidence of churn around a wrong model
`docs/architecture/07-storage-coherence-and-distribution.md`	`subtract` is a geometry op, not a faculty; store-bloat measurement re-read as the ticker showing up on disk
`docs/architecture/08-dharma-sovereignty-and-governance.md`	"The grounding governor" (6 occurrences) superseded as terminology — a rename, not a retraction; the projection is right, the governor is not
`docs/architecture/04-runtime-and-deployment.md`	Backup cronjob distinguished from consolidation; local ticker inventory; `crontab -l` has zero neuron entries
`AGENTS.md`	`learn(faculty=...)` flagged as a faculty-as-parameter
`council/README.md`	Architecturally superseded: a write-refusal in an immutable substrate and a resident consolidation service in Python outside el; "reject entirely" withdrawn
`neuron-dev-setup/README.md`	The exclusion list re-read as a census of the launch-agent fragmentation, with the measured table
`docs/sessions/2026-08-13-…md`	Pointer note only — session records are not amended

Measured consolidation inventory (2026-08-16)

Eleven implementations across three languages and two processes. Six are launch agents; three of those run in Python, outside el, so part of Neuron's consolidation cannot touch the geometry at all. The 23:55 / 06:00 / 08:30 trio is a sleep cycle implemented as three unrelated scheduled scripts, none aware of each other. Every name in the set is a consolidation verb — compress, cultivate, digest, integrate, review, reify, beat.

Precision note: it is not cron. crontab -l has zero neuron entries; every neuron schedule is launchd StartInterval / StartCalendarInterval.

Discrepancies recorded, not resolved

The count. The authority doc's §7 heading says "seven implementations" while its own table lists ten rows. Measured independently here it is eleven. Recorded in 06 §12.4; the authority doc is not edited from this branch.
The values count. The spec says the values reference is thirteen regions; this repo's allowlist enumerates eight value nodes plus a values hub (neuron-api.el:20-37). Recorded in 06 §12.1 with an instruction not to cite a count without measuring it.
Backup interval. 03:215-217 says 15 min, which matches local snapshot-backup (900 s) rather than local engram-backup (3600 s). The cluster manifest was not read on this branch; the figure is marked unverified.

Live code residue flagged (not edited)

engram_cognition.h:155,158 (COG_GROUNDED_BY_RELATION) · engram_cognition.c:249 (cog_ground_edge) · mcp-wrapper/src/main.el:422,516-519,1394-1397 (the wonder-manifest, served in the live tool list) · neuron-api.el:1436,1447-1456 · mcp-wrapper/src/main.el:409 (faculty as a parameter) · engram_geometry.h:79,426 (co_registration, deprecated — removal is a GEO1 format migration and must not ride along) · el_runtime.c:14698 (keystone_write_blocked).

Documentation only — no `.el`, `.c`, `.h`, `.py`, `.sh`, `.plist` or `.yaml` file is touched. `git diff --name-only` is 12 `.md` files. ## What this records `foundation/el`, branch `design/correspondence-and-censorship`, `lang/spec/correspondence-and-censorship.md` supersedes this repo's architecture docs on four points. Its companion, `lang/spec/runtime-ownership.md`, addresses a different residual in the same substrate. The corrections are **transcribed, not re-derived** — every earlier version of this reasoning was wrong in an instructive way and each correction was argued down hard. 1. **Grounding is not a subsystem — it *is* the edge weight (the hebbian weight).** One quantity, not two fields. `grounded-by` as a relation type should not exist: grounding is a property *of* a relation, not a relation *between* nodes, so minting the edge is the error, not merely which endpoints it chose. Never computed on demand, never a score. Traversal is already grounded inference; decision provenance is the path. 2. **Faculties are operations, not parameters.** `reason` changes the estimate (read), `induce` changes the parameters (the correspondence-beat, which exists and works), `abduce` changes the structure (write). A write cannot be a parameter of a read. 3. **Wonder is the boundary, not a manifest.** Any structure at all has an edge. There are about six wonders, the same for everyone, and they never close. **Curiosity is wonder crystallized** at a nucleation site — one thing at two phases, which is why curiosity can be satisfied and wonder cannot. 4. **Consolidation is ambient, not scheduled. A brain has no cron job.** The presence of a ticker is the diagnostic. `soul.el:731`'s continuous `awareness_run()` loop is the one fragment with the correct shape; the others fold into it. Plus: `co_registration` is deprecated in favour of per-edge `GeoEdge.discord`, and — in an immutable substrate — any mechanism that refuses a write is either redundant with immutability or an epistemic constraint misfiled as a protective one. ## Approach Stale passages are **left standing, marked inline, and pointed at their replacement** rather than deleted — per `03` §Immutability and `06` §4.1, supersession is residue: the trail of how the understanding matured is kept, because sometimes the truth was in the old idea even when the old idea was not itself the truth. New `06 §12` carries the transcribed corrections and the measured inventory. Every factual claim was re-verified against source at `file:line` before being written. Line references that had drifted are corrected in place and labelled. ## Files | file | change | |---|---| | `docs/architecture/06-cognitive-architecture.md` | Top banner + inline supersession markers at 12 sites; new **§12** with the four corrections, the edge-vector model, the live code-residue tables, and the measured consolidation inventory | | `docs/architecture/03-data-and-memory.md` | §Write-protection superseded; notes that the doc contradicts itself 35 lines later; `cultivate` line ref corrected `:781-816` → `:960`; backup cronjob marked as ticker-but-not-consolidation | | `docs/architecture/02-components.md` | `awareness_run` loop marked as the *correct* shape; curiosity-scan superseded on three counts; stale line refs corrected; `/consolidate` and `mem_consolidate` marked as 2 of 11 | | `docs/architecture/00-overview.md` | Curiosity is not a peer Engine — one activation process, two seed sources | | `docs/architecture/01-vbd-decomposition.md` | Axis 2 curiosity-rotation Engine superseded; its deep git-archaeology re-read as evidence of churn around a wrong model | | `docs/architecture/07-storage-coherence-and-distribution.md` | `subtract` is a geometry op, not a faculty; store-bloat measurement re-read as the ticker showing up on disk | | `docs/architecture/08-dharma-sovereignty-and-governance.md` | "The grounding governor" (6 occurrences) superseded as terminology — a rename, not a retraction; the projection is right, the governor is not | | `docs/architecture/04-runtime-and-deployment.md` | Backup cronjob distinguished from consolidation; local ticker inventory; `crontab -l` has zero neuron entries | | `AGENTS.md` | `learn(faculty=...)` flagged as a faculty-as-parameter | | `council/README.md` | Architecturally superseded: a write-refusal in an immutable substrate **and** a resident consolidation service in Python outside el; "reject entirely" withdrawn | | `neuron-dev-setup/README.md` | The exclusion list re-read as a census of the launch-agent fragmentation, with the measured table | | `docs/sessions/2026-08-13-…md` | Pointer note only — session records are not amended | ## Measured consolidation inventory (2026-08-16) Eleven implementations across three languages and two processes. Six are launch agents; three of those run in **Python, outside el**, so part of Neuron's consolidation cannot touch the geometry at all. The 23:55 / 06:00 / 08:30 trio is a sleep cycle implemented as three unrelated scheduled scripts, none aware of each other. Every name in the set is a consolidation verb — compress, cultivate, digest, integrate, review, reify, beat. Precision note: it is **not cron**. `crontab -l` has zero neuron entries; every neuron schedule is launchd `StartInterval` / `StartCalendarInterval`. ## Discrepancies recorded, not resolved - **The count.** The authority doc's §7 heading says "seven implementations" while its own table lists ten rows. Measured independently here it is **eleven**. Recorded in `06` §12.4; the authority doc is not edited from this branch. - **The values count.** The spec says the values reference is **thirteen** regions; this repo's allowlist enumerates **eight** value nodes plus a values hub (`neuron-api.el:20-37`). Recorded in `06` §12.1 with an instruction not to cite a count without measuring it. - **Backup interval.** `03:215-217` says 15 min, which matches local `snapshot-backup` (900 s) rather than local `engram-backup` (3600 s). The cluster manifest was not read on this branch; the figure is marked unverified. ## Live code residue flagged (not edited) `engram_cognition.h:155,158` (`COG_GROUNDED_BY_RELATION`) · `engram_cognition.c:249` (`cog_ground_edge`) · `mcp-wrapper/src/main.el:422,516-519,1394-1397` (the wonder-manifest, served in the live tool list) · `neuron-api.el:1436,1447-1456` · `mcp-wrapper/src/main.el:409` (`faculty` as a parameter) · `engram_geometry.h:79,426` (`co_registration`, deprecated — removal is a GEO1 format migration and must not ride along) · `el_runtime.c:14698` (`keystone_write_blocked`).

will.anderson added 1 commit 2026-08-16 18:32:07 +00:00

docs: record the correspondence corrections — grounding, faculties, wonder, consolidation

Neuron Soul CI / build (pull_request) Failing after 4m9s

Details

Neuron Soul CI / deploy (pull_request) Has been skipped

Details

65dd2cf097

The architecture docs describe four things the design spec has since ruled out,
and each one is a supervisor invented for something that should be a property of
the substrate: grounding modelled as a subsystem rather than as the edge weight
it already is; faculties modelled as parameters of a read when abduce is a write;
wonder materialized as a maintained manifest when it is the boundary of the
structure; and consolidation implemented eleven times behind tickers when a brain
has no cron job.

Left standing rather than deleted, per the repo's own supersession discipline —
the trail of how the understanding matured is the point. Each stale passage is
marked inline and points at a new 06 §12 that transcribes the corrections and
records the measured consolidation inventory.

Authority: foundation/el, branch design/correspondence-and-censorship,
lang/spec/correspondence-and-censorship.md.

will.anderson merged commit 9a3a248675 into main

2026-08-16 20:49:29 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: neuron-technologies/neuron#163