Vocabulary OS — Mega Index Hub (Z0–Z3 Directory) (Almost-Code, Paste-Ready)

Vocabulary := retrievable symbols (words + phrases + structures)
a person can activate under time + context + cognitive load
to convert thought ↔ meaning ↔ output.
P3 Vocabulary := <2s retrieval + correct usage frame + load survival + spaced stability.

Input ≠ Output.
Recognition (P1) ≠ Retrieval (P3).
Load reveals truth.
Stability across days beats single-day performance.

VOCABULARY_OS_STACK {
  Z0: Word-level retrieval (single units)
  Z1: Phrase/collocation engine (chunks)
  Z2: Sentence mechanics (structures + connectors)
  Z3: Output under load (timed speaking/writing)
  Control Layer:
    Sensors → Router → Repair loops → Re-test
}

Entry := {
  Z: Z0|Z1|Z2|Z3
  Skill
  FailureModes[Top6]
  Sensors[Key]
  Repairs[Minimal]
  OutputTest(P3)
}

Z0.Skill: Word_Retrieval_P3(word)
Z0.FailureModes:
  - P1 recognition-only
  - slow recall (P2)
  - wrong meaning slice
  - wrong collocation
  - wrong register/tone
  - spelling/phonology mismatch
Z0.Sensors:
  RT(word)          // target ≤ 2s
  UC(word)          // meaning+collocation+tone correct
  SD(word)          // T+0/+24h/+7d stability
Z0.Repairs:
  R0a 2-sentence anchor
  R0b 3-recall drill
  R0c speak-then-write transfer
Z0.OutputTest:
  RT≤2s + UC=true across 3 sessions

Z1.Skill: Chunk_Retrieval_P3(phrase/collocation)
Z1.FailureModes:
  - word-by-word assembly
  - wrong preposition/particle
  - synonym substitution breaks meaning
  - generic verb overuse
  - low variation (repetition)
  - idiom misuse
Z1.Sensors:
  CER (collocation error rate)
  GVR (generic verb rate)
  VI (variation index)
Z1.Repairs:
  R1a collocation pair bank (3 pairs per word)
  R1b swap-frame drill (same meaning, stronger)
  R1c sentence pattern templates
Z1.OutputTest:
  5 sentences using chunk family with natural variation + low CER

Z2.Skill: Sentence_Build_P3(structure)
Z2.FailureModes:
  - run-ons / fragments
  - tense drift
  - agreement errors
  - weak connectors
  - clarity loss (who did what)
  - over-complex collapse
Z2.Sensors:
  ErrorPer100Words
  ConnectorDiversity
  ClarityTest (actor/action clear)
Z2.Repairs:
  R2a 1-structure-per-week
  R2b connector ladder upgrade
  R2c rewrite weakest sentence 3 ways
Z2.OutputTest:
  timed paragraph: low errors + ≥4 connector types + clarity pass

Z3.Skill: Output_P3_UnderLoad
Z3.FailureModes:
  - blank page freeze
  - vocab collapses under timer
  - repetition increases
  - flat verbs/adjectives
  - imagery/emotion missing
  - coherence breaks
Z3.Sensors:
  FreezeEvents count
  LS (load survival) of targets
  TimedFinishRate
  WeakestParagraphIndex
Z3.Repairs:
  R3a 5-box plan + vocab targets
  R3b timed micro drills (8–12 min)
  R3c weakest paragraph rewrite
  R3d deliberate vocab injection (8W/4C/2S)
Z3.OutputTest:
  finishes on time + targets appear correctly + coherence stable

MV_Sensors {
  RT ≤ 2s
  UC true in sentence
  LS true under timer
}

LeakTag := Speed | Usage | Load | Drift | Selection

If RT high → Speed
Else if UC fails → Usage
Else if timed output collapses → Load
Else if day-7 decay → Drift
Else → Selection

Primary uses:
  - P1→P3 vocabulary conversion
  - composition vocabulary injection
  - timed paragraph + rewrite loops

Secondary adds:
  - register control
  - paraphrase families
  - summary compression vocabulary

O-levels adds:
  - variance control under topic variation
  - full timed sets + post-mortems

VOCABULARY_OS_SPEC v1.0 {
  Definition: Vocabulary = retrievable symbols enabling output
  Phases: P0 absent | P1 passive | P2 active-weak | P3 active-stable
  ZLevels: Z0 words | Z1 chunks | Z2 structures | Z3 load output
  Laws: Input≠Output; Recognition≠Retrieval; Load reveals truth; Stability matters
  Sensors: RT, UC, LS, SD, OAR (+CER,GVR,VI...)
  Router: LeakTag Speed|Usage|Load|Drift|Selection
  Repairs: R0/R1/R2/R3 mapped to tags
  P3 Gate: RT≤2s + UC=true + LS=true + SD pass
}