SG Syllabus Lattice Directory v0.1 — SCIENCE (Primary → Secondary → JC)

TITLE: SG Syllabus Lattice Directory v0.1 — SCIENCE (Primary → Secondary → JC)
VERSION: CivOS Unified Spec v1.x
MODE: Almost-Code / LLM-first / WordPress paste-ready
OWNER: eduKateSG Fence Learning Systems
SCOPE: Singapore Science as a runnable lattice (node registry + couplings + sensors + NIT-S repair)
GOAL: Publish an executable Science directory:
      stable IDs, prerequisite graph, transfer tests, failure signatures (NIT-S),
      Fence repair routes, and coupling edges to Math (units/algebra/graphs) + English (explanation clarity).
CANONICAL CLAIM:
- Science failure is rarely “lack of memorisation”. It is mechanism instability + data interpretation errors under load.
- NIT-S is the collapse threshold: definition-only knowledge becomes irrecoverable during application questions.
- This directory makes Science a computable repair system.
========================================================
0) DIRECTORY MAP (WHAT YOU PUBLISH)
========================================================
/sg/syllabus/science/                       (Hub)
/sg/syllabus/science/node-registry/         (All nodes, stable IDs)
/sg/syllabus/science/prereq-edges/          (Graph edges)
/sg/syllabus/science/sensor-pack/           (Micro-tests + metrics)
/sg/syllabus/science/failure-atlas/          (NIT-S + data + practical signatures)
/sg/syllabus/science/repair-protocols/       (Truncation + stitching)
/sg/syllabus/science/coupling-edges/         (SCI↔MATH units/graphs, SCI↔ENG explanation)
/vl-os/                                      (Verification Literacy OS spine; plugs into STEM)
========================================================
1) COORDINATE STANDARD (LOCKED)
========================================================
ID Grammar:
Place×Lane×Zoom×Role×Type×ID
Place: SG
Lane: SCI
Zoom: Z0 learner (readable at Z1/Z2)
Role: Learner/Tutor/Teacher/Operator
Type: Node/Edge/SensorPack/Protocol/Test/Atlas/Panel
========================================================
2) SCIENCE “SPINE GRAPH” (HIGH-LEVEL)
========================================================
Spine:
TERMS (definitions) →
MECHANISM (A→B→C causal chain) →
REPRESENTATION (diagrams/models) →
DATA (tables/graphs) →
APPLICATION (novel contexts) →
PRACTICAL (planning/variables/evaluation) →
TRANSFER (timed exam stability)
High-weight couplings:
- MATH-ALG/FUNC-GRAPH → Physics motion/electricity (VH)  [COUP-SG-02]
- VL-05 Units/Dimensions → STEM correctness (H/VH)
- ENG coherence → explanation clarity (H)
- DATA-GRAPHS → Geography/other data tasks (H)
========================================================
3) NODE REGISTRY (CORE SET v0.1)
========================================================
NodeRecord schema:
NodeRecord := {
  ID, LevelBand, Domain,
  Prereq[], Outputs[],
  PhaseTargets,
  TransferTests[],
  Sensors[],
  FailureSignatures[],
  RepairLoops[],
  FenceRules
}
----------------------------------------
3.1 PRIMARY SCIENCE CORE NODES (P3–P6)
----------------------------------------
Node: SG×SCI×Z0×Learner×Node×SCI-OBS-QUESTION
LevelBand: P3–P6
Domain: Scientific Thinking (observe, question)
Prereq: []
Outputs: [SCI-VAR-FAIRTEST, SCI-EXP-CLAIM]
PhaseTargets: {P2 by P5}
TransferTests: {identify what is being tested; suggest question}
Sensors: {THINK-5}
FailureSignatures: {RandomGuessing}
RepairLoops: {observe→ask→predict template}
FenceRules: {if RandomGuessing repeats → reduce content load; rebuild process}
Node: SG×SCI×Z0×Learner×Node×SCI-TERMS-CORE
LevelBand: P3–Sec4
Domain: Definitions & Keywords (but not alone)
Prereq: []
Outputs: [SCI-MECH-2LINK, SCI-DATA-READ]
PhaseTargets: {P2 by P6}
TransferTests: {define + use in a correct sentence/explanation}
Sensors: {DR-5}
FailureSignatures: {DefinitionOnlyIllusion (NITS-SG-01)}
RepairLoops: {term→because-line→example}
FenceRules: {if DefinitionOnlyIllusion → forbid definition-only answers; require mechanism line}
Node: SG×SCI×Z0×Learner×Node×SCI-MECH-2LINK
LevelBand: P4–JC
Domain: Mechanism Chain (A→B→C)
Prereq: [SCI-TERMS-CORE]
Outputs: [SCI-MODEL-DIAGRAM, SCI-APP-NOVEL]
PhaseTargets: {P2 by Sec2}
TransferTests: {90-second explain with because/therefore}
Sensors: {MR-90}
FailureSignatures: {KeywordSalad, NoCausalChain}
RepairLoops: {2-link chain drills; replace vague words with causal verbs}
FenceRules: {if NIT-S rising → truncate topics; stitch 2-link chains first}
Node: SG×SCI×Z0×Learner×Node×SCI-MODEL-DIAGRAM
LevelBand: P5–Sec4
Domain: Diagrams/Models (labels + meaning)
Prereq: [SCI-MECH-2LINK]
Outputs: [SCI-APP-STRUCTQ]
PhaseTargets: {P2 by Sec1}
TransferTests: {label diagram + explain function}
Sensors: {DIAG-8}
FailureSignatures: {LabelButNoMeaning}
RepairLoops: {label→function→mechanism sentence}
FenceRules: {if LabelButNoMeaning → ban “label-only” scoring; enforce function line}
Node: SG×SCI×Z0×Learner×Node×SCI-DATA-READ
LevelBand: P4–JC
Domain: Data Reading (tables/graphs)
Prereq: [SCI-TERMS-CORE]
Outputs: [SCI-DATA-INFER, SCI-PRACT-EVAL]
PhaseTargets: {P2 by Sec2}
TransferTests: {trend + value + compare; identify variables}
Sensors: {GRF-10 (graph/table 10-min)}
FailureSignatures: {AxisConfusion, DescribeOnly}
RepairLoops: {trend→evidence→because template}
FenceRules: {if AxisConfusion repeats → force axis+unit annotation before any answer}
Node: SG×SCI×Z0×Learner×Node×SCI-DATA-INFER
LevelBand: Sec1–JC
Domain: Data Inference (claim from evidence)
Prereq: [SCI-DATA-READ, SCI-MECH-2LINK]
Outputs: [SCI-APP-NOVEL, SCI-PRACT-CONCLUDE]
PhaseTargets: {P2 by Sec3}
TransferTests: {make claim + cite data + explain}
Sensors: {DATA-CEE-10}
FailureSignatures: {ClaimNoEvidence, WrongConclusion}
RepairLoops: {CEE (claim-evidence-explain) + mechanism link}
FenceRules: {if WrongConclusion → truncate evaluation questions; stitch inference chain}
----------------------------------------
3.2 SECONDARY SCIENCE CORE NODES (Sec1–Sec4/5)
----------------------------------------
Node: SG×SCI×Z0×Learner×Node×SCI-APP-STRUCTQ
LevelBand: Sec1–Sec4
Domain: Structured Application Questions
Prereq: [SCI-MECH-2LINK, SCI-MODEL-DIAGRAM, SCI-DATA-READ]
Outputs: [SCI-APP-NOVEL, SCI-EXAM-TRANSFER]
PhaseTargets: {P2 by Sec3}
TransferTests: {multi-part questions with consistent logic}
Sensors: {STRUCT-15}
FailureSignatures: {PartAOkPartBCollapse}
RepairLoops: {markpoint map + “what does this part test?”}
FenceRules: {if PartAOkPartBCollapse → treat as P1 masquerading; rebuild mechanism + data links}
Node: SG×SCI×Z0×Learner×Node×SCI-APP-NOVEL
LevelBand: Sec2–JC
Domain: Novel Context Transfer
Prereq: [SCI-APP-STRUCTQ, SCI-DATA-INFER]
Outputs: [SCI-EXAM-TRANSFER]
PhaseTargets: {P2 by Sec4, P3 top band}
TransferTests: {apply concept in new scenario; justify}
Sensors: {TT-V-SCI (variation transfer set)}
FailureSignatures: {MemorisedButCannotApply (NITS-SG-01)}
RepairLoops: {variation ladder: near→mid→far contexts}
FenceRules: {if cannot apply → stop new content; stitch variation + mechanism before proceeding}
----------------------------------------
3.3 PRACTICAL / EXPERIMENT CORE (Primary→JC)
----------------------------------------
Node: SG×SCI×Z0×Learner×Node×SCI-VAR-FAIRTEST
LevelBand: P4–Sec4
Domain: Variables + Fair Test
Prereq: [SCI-OBS-QUESTION]
Outputs: [SCI-PRACT-PLAN, SCI-PRACT-EVAL]
PhaseTargets: {P2 by Sec2}
TransferTests: {identify IV/DV/controls}
Sensors: {VAR-6}
FailureSignatures: {VariableMixUp}
RepairLoops: {IV/DV/control card sort + rewrite}
FenceRules: {if VariableMixUp → truncate experiment planning; stitch variable identification}
Node: SG×SCI×Z0×Learner×Node×SCI-PRACT-PLAN
LevelBand: Sec1–JC
Domain: Planning Practical
Prereq: [SCI-VAR-FAIRTEST, SCI-DATA-READ]
Outputs: [SCI-PRACT-CONCLUDE]
PhaseTargets: {P2 by Sec3}
TransferTests: {method steps + controls + safety}
Sensors: {PLAN-PRAC-10}
FailureSignatures: {MissingControl, VagueMethod}
RepairLoops: {steps→controls→measure template}
FenceRules: {if VagueMethod → enforce measurable verbs + quantities}
Node: SG×SCI×Z0×Learner×Node×SCI-PRACT-EVAL
LevelBand: Sec2–JC
Domain: Evaluation (reliability/accuracy/improvements)
Prereq: [SCI-DATA-INFER, SCI-PRACT-PLAN]
Outputs: [SCI-EXAM-TRANSFER]
PhaseTargets: {P2 by Sec4}
TransferTests: {identify errors + propose improvement}
Sensors: {EVAL-PRAC-10}
FailureSignatures: {GenericEvaluation}
RepairLoops: {error source → effect → fix template}
FenceRules: {if GenericEvaluation → require “because” line tied to data/measurement}
Node: SG×SCI×Z0×Learner×Node×SCI-PRACT-CONCLUDE
LevelBand: Sec1–JC
Domain: Conclusion from Evidence
Prereq: [SCI-DATA-INFER]
Outputs: [SCI-EXAM-TRANSFER]
PhaseTargets: {P2 by Sec3}
TransferTests: {claim + data + limitation}
Sensors: {CONC-8}
FailureSignatures: {Overclaim}
RepairLoops: {claim scope limiter + cite data}
FenceRules: {if Overclaim → enforce boundary check (VL-04) and limitation sentence}
----------------------------------------
3.4 STEM VERIFICATION SPINE (PLUG-IN)
----------------------------------------
Node: SG×SCI×Z0×Learner×Node×SCI-UNITS-DIM
LevelBand: Sec1–JC
Domain: Units/Dimensions (VL-05)
Prereq: []
Outputs: [SCI-PHY-CORRECTNESS]
PhaseTargets: {P2 by Sec2}
TransferTests: {unit tagging in steps; dimensional check}
Sensors: {UNIT-5}
FailureSignatures: {UnitBlindness (NITS-SG-02)}
RepairLoops: {unit every line rule + sanity check}
FenceRules: {if UnitBlindness → truncate formula use; rebuild units before continuing}
========================================================
4) PREREQUISITE EDGES (GRAPH) — v0.1 (CORE)
========================================================
Edge format:
Edge := {FromID, ToID, Weight(VH/H/M), SignatureLink}
Edges (selected):
- SCI-TERMS-CORE → SCI-MECH-2LINK (VH)   [NITS-SG-01 risk if skipped]
- SCI-MECH-2LINK → SCI-MODEL-DIAGRAM (H)
- SCI-TERMS-CORE → SCI-DATA-READ (H)
- SCI-DATA-READ → SCI-DATA-INFER (VH)
- SCI-DATA-INFER + SCI-MECH-2LINK → SCI-APP-NOVEL (VH)
- SCI-VAR-FAIRTEST → SCI-PRACT-PLAN (H)
- SCI-PRACT-PLAN + SCI-DATA-INFER → SCI-PRACT-EVAL (H)
- SCI-UNITS-DIM → SCI-APP-STRUCTQ (H/VH in Physics topics)
Coupling edges (published in /coupling-edges/):
- MATH-ALG-LINEAR + MATH-FUNC-GRAPH → SCI Physics application (VH) [COUP-SG-02]
- ENG-PARA-CORE (coherence) → SCI explanation clarity (H)
- VL-OS nodes (VL-03, VL-05, VL-06, VL-08) → SCI correctness & markpoint alignment (H/VH)
========================================================
5) SENSOR PACK (MICRO-TESTS) — v0.1
========================================================
SensorPack: SG×SCI×Z0×Tutor×SensorPack×SENS-SG-SCI-v0.1
Weekly minimum (20–35 min total):
- DR-5: definitions (5 min) BUT must add 1 mechanism sentence
- MR-90: 90-second mechanism explain (1 item)
- GRF-10: graph/table read (10 min)
- STRUCT-10: short structured application (10 min) alternate weeks
Metrics:
- DR (definition recall)
- MR (mechanism reliability)
- GRF (graph/data reliability)
- PR (practical reasoning)
- NITS_w (NIT-S triggers per week)
- ERC (repeat failure signature count)
Fence triggers:
- NITS_w ≥ 2 → Emergency Mechanism Mode (truncate topics; stitch 2-link chains)
- DR high but MR low → Definition-only illusion (treat as P1 masquerading)
- GRF errors repeat → force axis/unit annotation protocol
- UnitBlindness → enforce units-every-line rule (VL-05)
========================================================
6) FAILURE ATLAS (SIGNATURES) — v0.1 (SCIENCE)
========================================================
SCI-SIG-01: Definition-Only Illusion (NITS-SG-01)
SCI-SIG-02: Keyword Salad (no causal chain)
SCI-SIG-03: Describe-Only Data Answers (no because line)
SCI-SIG-04: Axis/Unit Confusion (data/graphs)
SCI-SIG-05: Part-A OK, Part-B Collapse (structured question breakdown)
SCI-SIG-06: Generic Evaluation (practical)
SCI-SIG-07: Unit Blindness (NITS-SG-02)
SCI-SIG-08: Overclaim (conclusion too strong for evidence)
Each signature links to:
- predicted next failure
- truncation action
- stitching repair loop
- re-test
========================================================
7) REPAIR PROTOCOLS (TRUNCATION + STITCHING) — v0.1
========================================================
Protocol: SG×SCI×Z0×Operator×Protocol×SCI-FENCE-v0.1
Truncation Rules:
- If NIT-S rising → pause new topics; reduce question novelty
- If definition-only illusion → stop memorising more; require mechanism lines
- If data errors repeat → pause evaluation questions; rebuild data reading
- If units blind → pause formula plugging; rebuild unit discipline
Stitching Order (canonical):
1) Mechanism chain (A→B→C) stability
2) Data reading (trend + evidence)
3) Inference (claim + cite data + explain)
4) Practical planning/evaluation templates
5) Transfer (variation ladder) + timed structured sets
Re-test Rule:
- same micro-test category next week; must show signature extinction trend
========================================================
8) REQUIRED FAILURE TRACE (SCIENCE DIRECTORY)
========================================================
Trace:
Definitions memorised →
mechanism unstable →
data misread →
structured questions collapse →
evaluation becomes generic →
NIT-S crosses threshold →
exam shock (P2→P0) in application topics.
Fence Repair:
Sensors detect MR/GRF drift early →
truncate new content →
stitch 2-link mechanisms + data template →
units/verification discipline →
variation ladder →
transfer stabilises.
========================================================
END: SG Syllabus Lattice Directory v0.1 — SCIENCE
========================================================