The held cases sample paper-source interpretation and one-property reasoning. There is no fixed local material answer: an approved real classroom sign's composition and visible features depend on its exact preparation, provenance and view. Record first response/access/neutral reread before feedback. A child can respond by speech, sign, AAC, pointing, drawing or directing an adult. A prompt that supplies the material rule changes the evidence. One paper check cannot certify year mastery, growth, actual AC9SFI02 investigation or a real local material property.
Day 135 · Case AD
- Literal picture evidence: The invented triangle has FOUR LARGE DOTS at LEFT and THREE ZIGZAG LINES at RIGHT, divided by a dark border. Those are depicted marks. No maker/provenance, verified fabric/card, real object, local date, back view or touch is provided.
- Claim boundary: The caption cannot convert drawn dots or zigzags into physical fabric/card. A two-region drawing does not establish two material components. An acceptable answer says what is pictured and keeps materials UNKNOWN. A child can ask a willing maker what they used, how they know, and inspect a safe approved actual object with provenance/date; no outcome is promised.
- Follow-up: If the child names fabric from dots, revisit K's picture and L's separately labelled maker source after scoring. If left/right marks are confused, offer two fixed broad regions and read the exact held words neutrally after the independent record. If the child says “not enough information”, ask for a smaller testable question next.
- Do not award: actual AC9SFU03 material property/composition, AC9SFI02 safe investigation, a verified two-material sign, a real maker answer or observed touch from AD.
Day 140 · Case AE
- One-property sort: By DRAWN DOTS, P and Q go together; R goes in NO DOTS. The imaginary maker reports P as paper-based card and Q/R as woven cloth. Material identity is in an attributed fictional note, not in the dot pattern itself.
- Counterexamples: P has dots but is maker-labelled card, so “every dotted patch is cloth” fails inside this model. R is maker-labelled cloth without dots, so “every cloth patch has dots” also fails. The child need not use formal universal-logic language; “P is dotted card” and “R is cloth with no dots” suffice.
- Honest share: “In this invented note, P and Q have drawn dots, but the maker labels them as different materials. R is labelled cloth with no dots. The rule fails here. We don't know how real class patches would look or what they are made of.” A next actual source needs local approval, date, broad safe known samples, verified provenance, direct feature view or correctly attributed ADULT REPORT and raw unknowns.
- Follow-up: If a learner sorts by maker label instead of dots, name the requested one drawn-feature rule and model a new example after scoring. If a learner calls dots texture, distinguish paper raised-dot adaptation from an actual material surface. If a maker label is treated as the child's direct sense, separate source and access fields. Preserve mixed/unknown answers rather than forcing a pattern.
- Do not award: a real fabric/card property, real sample composition, actual safe investigation, prior-prediction comparison, child improvement or listener feedback from AE. The imaginary maker's words are not the child's own direct sensory observation.
Interpretation: AC9SFU03 includes both multi-material objects and observable properties. This pair of paper checks only probes whether a child can keep a model, maker attribution, visible mark, counterexample and genuine-source need separate. If an actual approved inspection also occurred, assess its dated log separately. If not, record REAL OBSERVATION NOT AVAILABLE and plan a safe accessible local opportunity when possible. Original key © 2026 NeuroForgeIO Pty Ltd, SubjectNest, CC BY 4.0.