Training for Malay (eduKateSG) — Fence Learning Systems for Malay Language (Z0–Z2)

TITLE: Training for Malay (eduKateSG) — Fence Learning Systems for Malay Language (Z0–Z2)
SUBTITLE: Reach P3 reliability in Malay: comprehension + kosa kata + tatabahasa + penulisan + lisan
VERSION: CivOS Unified Spec v1.x
MODE: Almost-Code / WordPress paste-ready
SCOPE: Z0 learner + Z1 parent + Z2 tutor/school; primary → secondary foundations; applies to ML/2nd-language learners
GOAL: Malay becomes reliable meaning production under time pressure (not “memorise and forget”)
========================================================
0) DEFINITION LOCK BOX
========================================================
Malay mastery = reliable comprehension and expression with correct structure.
Malay.P3 :=
A learner can:
- read with accuracy and inference,
- use vocabulary and common phrases correctly,
- apply grammar (tatabahasa) reliably,
- write coherent paragraphs/essays under time,
- speak with enough fluency and structure,
and do all this with low volatility under exam pressure.
Fence principle:
Malay collapses when:
- vocab is passive-only,
- grammar rules are known but not applied,
- writing has no structure,
- comprehension lacks evidence use,
and TTC shrinks.
So Malay training must be instrumented, fenced, repaired, and verified.
========================================================
1) THE MALAY STACK (WHAT MALAY IS MADE OF)
========================================================
Malay = 5 sub-systems:
ML1 Vocabulary & Phrases (Kosa kata + frasa)
- meaning + usage + collocations
ML2 Grammar Control (Tatabahasa)
- imbuhan (awalan/akhiran), kata ganti, kata hubung, struktur ayat
ML3 Reading Comprehension (Kefahaman)
- main idea, details, inference, evidence
ML4 Writing Output (Penulisan)
- planning, paragraph control, connectors, format
ML5 Oral & Listening (Lisan / Mendengar)
- fluency, response structure, confidence
If any subsystem fails, performance collapses.
========================================================
2) UNIVERSAL DASHBOARD (MALAY EDITION)
========================================================
Metrics:
- TTC: days to test/exam/oral
- Buffer: sleep/routine/stress slack
- R: are weak nodes shrinking weekly?
- Θ: reaction speed after a drop (24–72h)
- Λ: enforceability of routine same day
Malay-specific sensors:
- wrong imbuhan count
- connector usage variety (kata hubung)
- “blank writing time” (minutes)
- evidence use (can quote/point to passage line)
- oral hesitation time (seconds)
========================================================
3) Z0 ROUTER: FAST DIAGNOSIS (10 MINUTES)
========================================================
Step 1: Identify failing lane:
- vocab/phrases (ML1)?
- grammar (ML2)?
- comprehension (ML3)?
- writing (ML4)?
- oral (ML5)?
Step 2: Dominant failure mode (pick ONE):
F1 Vocab weak (generic words, wrong usage)
F2 Tatabahasa drift (imbuhan/structure errors)
F3 Comprehension weak (cannot locate evidence/infer)
F4 Writing collapse (no plan, weak paragraphs)
F5 Oral confidence collapse (short hesitant answers)
F6 False competence (can copy model but cannot produce)
F7 Time collapse (cannot finish under time)
Step 3: Fence (truncate):
- shrink to smallest failing node
- reduce scope; increase correct reps
- protect buffer if low
Step 4: Stitch (repair loop):
- 7-day targeted loop
- verify with timed mini-test + short writing
========================================================
4) FAILURE MODES → SENSORS → REPAIR LOOPS
========================================================
--------------------------------------------------------
F1 Vocabulary weak (kosa kata + frasa)
--------------------------------------------------------
Sensors:
- repeated basic words, awkward phrases
Repair Loop MLR1 (Phrase Frames + Active Retrieval):
- 5 words/day → 5 sentences/day
- add 2 common frasa per word (collocations)
- daily mini-paragraph using 4 target phrases
Fence:
- fewer words, correct usage first (quality fence)
--------------------------------------------------------
F2 Tatabahasa drift (imbuhan/structure)
--------------------------------------------------------
Sensors:
- wrong awalan/akhiran, wrong kata hubung, broken sentences
Repair Loop MLR2 (One Grammar Focus Per Week):
- choose ONE: imbuhan / kata sendi / kata hubung / ayat majmuk
- 10 drills + 5 sentence constructions daily
- error log: tag the grammar type
Fence:
- truncate writing length while fixing grammar clarity
--------------------------------------------------------
F3 Comprehension weak (kefahaman)
--------------------------------------------------------
Sensors:
- wrong inference, cannot justify answers
Repair Loop MLR3 (Evidence Ladder):
- cari ayat bukti → jelaskan dengan kata sendiri → jawab
- daily 1 passage + 6 questions
Fence:
- time fence: mark-and-move; return later
--------------------------------------------------------
F4 Writing collapse (penulisan)
--------------------------------------------------------
Sensors:
- blank page time high
- weak connectors, no structure
Repair Loop MLR4 (Plan + Paragraph Frames):
- 4-minute plan:
  idea 1 / idea 2 / example / conclusion
- paragraph frame:
  idea → example → explanation → link
- connector bank:
  kerana, oleh itu, walaupun, selain itu, namun, akhirnya
Fence:
- plan is mandatory before writing starts
--------------------------------------------------------
F5 Oral confidence collapse (lisan)
--------------------------------------------------------
Sensors:
- long hesitation, very short answers
Repair Loop MLR5 (Oral Response Template):
- structure:
  pendapat → sebab → contoh → kesimpulan
- daily 5 prompts, record 60–90 seconds
Fence:
- low-stakes exposure; build up gradually
--------------------------------------------------------
F6 False competence (model copying)
--------------------------------------------------------
Sensors:
- can memorise karangan but cannot write fresh
Repair Loop MLR6 (Inversion Practice):
- same tema, different soalan
- rewrite weakest paragraph after marking
Fence:
- no new topics until one weakness shrinks (R<1)
--------------------------------------------------------
F7 Time collapse (cannot finish)
--------------------------------------------------------
Sensors:
- rushed endings, incomplete answers
Repair Loop MLR7 (Timed Micro-Writes):
- intro 6 min
- body paragraph 8–10
- conclusion 5
Fence:
- “finish fence”: last 8 minutes for conclusion + grammar check
========================================================
5) OUTPUT TYPES (EXAM RELIABILITY)
========================================================
OutputType O1: Kefahaman (comprehension)
- evidence-based answers
OutputType O2: Tatabahasa (grammar)
- correct application under speed
OutputType O3: Penulisan / Karangan (writing)
- structured + connectors + vocabulary
OutputType O4: Lisan (oral)
- structured response + confidence
========================================================
6) Z1 PARENT TOOLKIT (MALAY)
========================================================
Parent role = enforce routine + protect buffers, not teach advanced rules.
Home fences:
- short daily slot (20–30 min)
- require:
  - 5 sentences with 5 words (MLR1)
  - 1 grammar focus drill (MLR2)
  - plan before writing (MLR4)
- sleep fence before tests/orals
Low-effort prompts:
- “Apa bukti dalam teks?”
- “Guna ‘oleh itu/selain itu/namun’ dalam ayat.”
- “Bagi satu contoh.”
========================================================
7) Z2 TUTOR/SCHOOL ENGINE (TRANSFER UNDER LOAD)
========================================================
Weekly cycle:
- diagnose dominant failure mode
- 4 days repair loop (MLR1–MLR7)
- 1 timed kefahaman set
- 1 timed writing task
- redo weakest paragraph + grammar error log review
Non-negotiables:
- active vocab use
- one-grammar-focus-per-week
- evidence ladder
- timed writing
========================================================
8) MINIMUM VIABLE 14-DAY PLAN (TTC-RED)
========================================================
Day 1–3:
- MLR3 evidence ladder + MLR4 planning (fast stability)
Day 4–7:
- MLR2 grammar focus + error log (clarity)
Day 8–10:
- MLR1 phrase frames + paragraph drills (expression power)
Day 11–14:
- MLR7 timed micro-writes + 2 timed karangan
- MLR5 oral drills if oral upcoming
Rule:
Shrink scope until R<1 (weak nodes shrink weekly).
========================================================
9) FAILURE MODE TRACE (REQUIRED)
========================================================
Trace:
vocab weak + tatabahasa drift + TTC short
→ comprehension errors + writing collapses under time
→ R>1 repeats → late reaction (Θ>1)
→ P2→P0 near exam
Fence Repair:
detect TTC-RED → truncate to MLR3+MLR4 first
→ enforce routine same day (Λ<1)
→ fence within 24h (Θ<1)
→ stitch MLR2/MLR1 next
→ timed verification (MLR7) → R<1 → restore P3 reliability
========================================================
10) WORDPRESS PASTE-READY STRUCTURE
========================================================
H1: Malay Training: How to Reach P3 Reliability in Exams (Fence Learning Systems)
H2: What Malay Mastery Really Is (meaning under load)
H2: The Malay Stack (vocab, grammar, comprehension, writing, oral)
H2: Dashboard (TTC, Buffer, R, Θ, Λ) + Malay sensors
H2: Fast Diagnosis (pick failing lane)
H2: Failure Modes → Repair Loops (MLR1–MLR7)
H2: Output Types (kefahaman/tatabahasa/penulisan/lisan)
H2: Parent Toolkit (Z1)
H2: Tutor/School Engine (Z2)
H2: 14-Day TTC-Red Plan
H2: Failure Mode Trace (Short)
H2: Copy/Paste Weekly Plan
========================================================
11) COPY/PASTE WEEKLY PLAN (MALAY)
========================================================
Daily (25–45 min):
- 10 min kefahaman evidence ladder (MLR3) OR vocab sentences (MLR1)
- 8–10 min tatabahasa drills (MLR2)
- 10–15 min paragraph writing with connectors (MLR4)
- log 1 recurring grammar mistake
Weekly:
- 1 timed kefahaman set
- 1 timed karangan task
- redo weakest paragraph + review error log
END SPEC
========================================================