Intuition Machine
The Loops

Every recurring process is literally a word

A string of moves on the matrix that starts at a cell and walks back to it. Ten of the twelve are intra-matrix words, machine-verified as closed walks.

Notation

Four letters, one for each thing you can do

U = Φ   unfold, +1 level
D = Φ̄   fold,   −1 level
I = ι   instance ↔ system
G = γ   content  ↔ vehicle

Words: strings over {U, D, I, G}, read left to
right from a start cell.

Because processes are strings, structural facts about them are visible in the letters before any interpretation is added — and two processes can be compared by diffing their words.

The inventory

Twelve loops

LOOP  FUNCTION              INTENDED CV             START  WORD
──────────────────────────────────────────────────────────────────────
⦿O    skilled action        task completion         2.ie   DGUUDG
⦿C    coordination          coordination state      2.ie   DIGUUDIG
⦿M    management            operational conformity  3.ce   DIIU
⦿M*   audit                 truth correspondence    3.ce   DI▼UI
⦿I    intelligence          model accuracy          1.ie   IUUIIIGIIG
⦿P    policy / identity     identity coherence      3.ci   DDUUII
⦿K    knowledge creation    knowledge validity      3.ii   GIGI
⦿G    grounding             empirical grounding     3.ce   DIDGUGUI
⦿L    legitimation          normative legitimacy    3.ci   GDGU
⦿R    reflexive quality     formal quality          3.ce   DDUUIGIG
⦿H    harness (mediation)   mediation effectiveness  — coupling layer —
⦿T    transparency          perceptual integrity     — coupling layer —

The specification flags one anomaly rather than silently fixing it: the intelligence row does not close under the stated move rules. The likely reading is that intelligence is an entering walk which then settles onto a closed convention-level cycle — but the claim as printed is "ten machine-verified closed walks", so the row is marked for re-checking against the verifier.

What the words reveal

Structure visible before interpretation

EACH WORD IS MADE OF THE OTHER’S MISSING LETTERS ⦿G grounding νκω uses D and U climbs and descends; never changes perspective ⦿K knowledge νκω uses G and I rotates perspective; never touches ground
Fig. 05 — Grounding descends; knowledge rotates
⦿G — the V-word

Grounding descends, then returns

Descents, a turn at felt quality, then ascents. Its phase trajectory is its syntax. To ground a claim you go down to what the thing is actually like, and you come back up carrying receipts.

⦿K — the Klein word

Knowledge rotates without descending

GIGI: flip, flip, flip, flip. A complete tour of the four quadrants conducted entirely at the convention level, containing zero up or down moves.

The complement

Each is made of the other's missing letters

The knowledge word uses only flips; grounding's core is climbs and descents. So the old claim that knowledge-making and grounding need each other is not consultant's advice — it is forced by letter inventories.

⦿M

Management never touches quality

Its word moves only between the causation and convention levels. The inability to feel is a typing fact, not a character flaw — and it is the structural reason the reflexive quality loop has to exist separately rather than as a management agenda item.

⦿R

The no-outsourcing rule

The reflexive loop opens with a double descent within the formal column, and that level change is reachable only by folding. Sensing the quality of your own formal structures is a descent you must perform yourself.

Recovery

Always two letters away

From any cell in a formal column, a double descent reaches the quality level. Whatever else is true of quality-bypass pathologies, the route back is structurally short.

The ceiling conjecture  ·  new in 5.3

How far anything can ever be proceduralised

Which processes can become a standard operating procedure — same steps, every time, by anyone — and which will forever require judgment? Organisations answer this today by proceduralising everything and discovering the hard way what dies.

depth(⦿X) = lowest spine level its word visits

contract ∈ { path ≻ outcome ≻ semantic ≻ statistical }

depth = κ  (⦿M)      → path-approachable
depth = ω  (⦿G, ⦿R)  → semantic at best

Management's word never descends below causation — and management is precisely what organisations proceduralise most successfully. Grounding descends to quality, and quality-level content cannot be transmitted on demand, so grounding can promise the same meaning but never the same steps.

HOW DEEP A PROCESS GOES CAPS HOW FAR IT CAN BE STANDARDISED STRICTEST LOOSEST path same steps outcome same result semantic same meaning statistical stable pattern ⦿M management depth = κ  · stays at encounter ceiling: proceduralise nearly to identity ⦿G ⦿R grounding, reflexive quality depth = ω  · descends to felt quality ceiling: same meaning, never the same steps Proceduralising past a process’s ceiling is how organisations kill the work they meant to scale.
Fig. 06 — Depth caps the strictest promise a process can keep

Build it, or find out which layer is missing.

Design engagements build the architecture for a specific organisation. Diagnostic engagements establish which of its loops actually close, which of its claims reach evidence, and which of its standing rules no longer have an owner.