# Foundation / Prep Science · ten 25-minute scripts, Days 11–20

Read the [safe-source protocol and complete original models](MATERIALS.md) before the week. Each lesson has **3 + 4 + 5 + 7 + 4 + 2 = 25 minutes** and a matching [learner card](LEARNER-CARDS.md). Choose an access route afresh; routes meet the same target, not fixed learner types. Keep source/date, first response, method, actual result, prompt and access mode. If no preflighted movement trial is safe, write **REAL OBSERVATION NOT AVAILABLE** and teach with the labelled paper model; never credit a fictional arrow as local movement. Teacher-held [checks](ASSESSMENT.md) stay separate until first independent response.

## Week 3 · Day 11 — Which movement question could we check?

**Goal:** distinguish an investigable question about a specific approved object's movement from a preference and an unsupported statement. **Proposed codes:** AC9SFH01, AC9SFI01. **Prepare:** Model P, blank [motion-question mat](motion-question.svg), safe preflight option, Card 11.

1. **Boundary · 3 min.** Name the clear flat lane and no-touch/pass choice; no object is released yet.
2. **Source · 4 min.** Read Model P verbatim as four *paper cards*. If a locally approved actual object is present, log its exact identity/date separately; do not assume its movement.
3. **Model · 5 min.** Teacher turns “Which colour do you like?” into “What happens when this approved object is gently released on this lane?” Explain that a question can be checked; a prediction is a later guess, not an answer.
4. **Child try · 7 min.** Child uses one Card 11 route to sort all four statements, then asks a new question about a safe specific object or model-only possible try. Adult records the child's exact wording and whether an actual source is available.
5. **Reason · 4 min.** Ask why “the drawn arrow proves rolling” is unsupported; child names what actual source/action would be needed. Do not run a surprise test.
6. **Exit · 2 min.** Save one investigable question and a still-unknown answer for Day 12.

## Week 3 · Day 12 — Make a safe, answerable plan

**Goal:** predict and select a locally safe method that could answer a specific motion question, identifying why two alternatives do not. **Proposed codes:** AC9SFI01, AC9SFI02. **Prepare:** Model Q, [safe-motion routine](safe-motion.svg), blank question mat, Card 12.

1. **Check · 3 min.** Confirm local approval, a broad slow-moving object and clear flat START/CATCH lane; if unavailable, mark plan-only model.
2. **Question · 4 min.** Reopen the child's Day 11 question or use “Where might this specific object stop after one gentle release?” Record prediction **before** any real action.
3. **Model · 5 min.** Read Model Q's three exact choices. Think aloud: one is a candidate only after approval; small hard bead is unsafe; drawing alone is not an actual test.
4. **Child try · 7 min.** Child uses Card 12 to set/indicate START, CATCH, who will release, and a no-touch access route. Ask what should stay the same if the try repeats. Adult signs off before any release.
5. **Check · 4 min.** Child explains why pushing harder would change the condition; record NOT TESTED if no actual try occurs today.
6. **Exit · 2 min.** Keep the prediction in its own box, with the result still blank or NOT TESTED.

## Week 3 · Day 13 — What did this object do here?

**Goal:** safely observe or attribute one actual gentle movement and record action, path/stop, and source without pretending a prediction was seen. **Proposed codes:** AC9SFU02, AC9SFI02, AC9SFI03. **Prepare:** one approved broad light object, flat marked lane, catch, [motion-log](motion-log.svg), Card 13.

1. **Preflight · 3 min.** Adult checks object, surface, distance, catch, bystanders and accessibility immediately before the trial; stop if unsuitable.
2. **Predict · 4 min.** State exact object/source/date, let child predict “stay / slide / roll / another movement / unsure” without treating these as results.
3. **Model · 5 min.** On a **different fictional record**, point out START, release action and result boxes. Explain “I saw it stop…” can only be said after an actual sight route; an adult report must be named.
4. **Child try · 7 min.** Adult performs one approved slow release or, with consent/site approval, child directs it. Child observes directly if accessible or receives an exact attributed adult report; adult writes action and what actually happened, including “no movement” if so. No safe trial means model reasoning and REAL OBSERVATION NOT AVAILABLE.
5. **Compare · 4 min.** Child compares prediction with actual result only if a result exists; SAME / DIFFERENT / NOT TESTED are all valid. Do not infer why it stopped from one try.
6. **Exit · 2 min.** Say what one arrow would leave out (e.g. surface or release), then save the dated log.

## Week 3 · Day 14 — Can an arrow show the whole movement?

**Goal:** represent a dated actual observed route in a provided template and identify what an invented or actual arrow omits. **Proposed codes:** AC9SFI03, AC9SFI05. **Prepare:** Day 13 dated log if available, invented Model M, [motion-model](motion-model.svg), Card 14.

1. **Source gate · 3 min.** Retrieve an actual dated log or mark REAL OBSERVATION NOT AVAILABLE. Show Model M is invented, not yesterday's object.
2. **Read model · 4 min.** Read START, circle, arrow, CATCH in exact order. Child names only marks shown; do not say it rolled or moved fast.
3. **Model · 5 min.** Teacher places an actual Day 13 record beside M if one exists. Demonstrate a separate arrow/word for that actual path only, labelled with date/object/action; list surface and release as details an arrow alone omits.
4. **Child try · 7 min.** Card 14 offers drawn, tactile or dictated routes. Child makes one representation from a real log **or** interprets M with MODEL label. Adult captures what each mark stands for and the source.
5. **Check · 4 min.** Ask “Can this arrow tell how fast it went?” Child says no speed/time is supplied. If their real log lacks a path, leave the arrow unmade rather than inventing one.
6. **Exit · 2 min.** Share with a consenting listener or teacher privately: one shown feature and one unknown.

## Week 3 · Day 15 — Fresh movement source check

**Goal:** independently ask, predict, distinguish actual movement from a new invented path, and state a source limit. **Proposed codes:** AC9SFU02, AC9SFI01, AC9SFI02, AC9SFI03. **Prepare:** unseen teacher-held [Day 15 check](ASSESSMENT.md), blank log, Card 15 *after* the check.

1. **Welcome · 3 min.** Explain the response chooses next teaching. Offer direct access, attributed adult report, paper-only or pass.
2. **Retrieve · 4 min.** Revisit only known Model M's arrow limits; do not show the fresh held path or object.
3. **Fresh source · 5 min.** Introduce the new preflighted broad soft object and flat lane, or the check's labelled invented two-arrow board. Record source/date/mode; no answer cue.
4. **First independent response · 7 min.** Child asks one motion question, predicts before any actual release, reports or reads what happened from the source, and names which claim remains unknown. Save first words/point/sign/AAC and support before teaching.
5. **Provenance · 4 min.** Child labels DIRECT, ADULT REPORT, or MODEL and compares prediction with actual only if a real result was obtained; otherwise NOT TESTED.
6. **Close · 2 min.** Use [teacher key](TEACHER-KEY.md) for next step. Release practice Card 15 only after scoring the first response.

## Week 4 · Day 16 — Does shape help explain a different movement?

**Goal:** describe movement of two safe objects and judge whether a shape comparison held other relevant features sufficiently alike. **Proposed codes:** AC9SFU02, AC9SFI01, AC9SFI02, AC9SFI04. **Prepare:** adult-verified matched pair if available, same flat lane/release, [compare-factors](compare-factors.svg), Model N, Card 16.

1. **Match gate · 3 min.** Check two broad approved objects for verified material, approximate size, shape difference and safe slow release. If another factor differs or no pair is safe, write this limitation or use Model N only.
2. **Predict · 4 min.** Name exact pair/source/date and shared surface/release; child predicts separately before any try.
3. **Model · 5 min.** Read Model N: invented endpoints differ but its size/material/release are unverified. Model “I cannot tell that shape caused the difference.”
4. **Child try · 7 min.** Adult conducts two safe tries from the same marked start using the same gentle action, if possible. Child uses Card 16 to record actual A/B path or stop; adult notes differences, access and any altered condition. Never supply the result in advance.
5. **Compare · 4 min.** Child contrasts predictions and observations, then names what stayed same and what differed. If controls were incomplete, say “these tries differed” and leave cause unknown.
6. **Exit · 2 min.** Choose one safer next repeat or missing fact; no universal “round things roll” rule.

## Week 4 · Day 17 — What changes when size changes?

**Goal:** plan and, where safe, compare movement of same-shape/same-material objects of two verified sizes, with a cautious claim. **Proposed codes:** AC9SFU02, AC9SFI01, AC9SFI03. **Prepare:** adult-verified broad/light size pair or MODEL ONLY, comparison mat, Card 17.

1. **Match gate · 3 min.** Verify shape/material and two sizes from source information; check that both are safe on the same flat lane. “Looks bigger” is a visible observation, material needs verification.
2. **Question · 4 min.** Ask “In our two gentle releases, will the larger and smaller approved objects stop in the same place?” Let child predict without claiming a physical rule.
3. **Model · 5 min.** Teacher shows blank SAME / DIFFERENT / OTHER CHANGED boxes. Explain that a larger object may also weigh more, and one trial does not isolate every influence.
4. **Child try · 7 min.** Card 17 guides two actual adult-led releases if both safe, with start/surface/action fixed as far as possible. Record dated results exactly or NOT TESTED. If pair is unavailable, use an invented empty template to plan only.
5. **Check · 4 min.** Child says which details matched, which did not, and whether the result matched each prediction. Do not infer “bigger always travels farther”.
6. **Exit · 2 min.** Ask what would need to be repeated or held alike before a stronger size claim.

## Week 4 · Day 18 — Is material the only difference?

**Goal:** identify the evidence needed to compare material's influence and report actual movement only when a safe, verified matched pair is available. **Proposed codes:** AC9SFU02, AC9SFI02, AC9SFI03. **Prepare:** possible broad/light same-shape-and-size pair of verified different materials, exact source labels, flat lane, Card 18.

1. **Verify · 3 min.** Adult checks material labels from preparation/product sources. If UNKNOWN, do not call it a material comparison. Keep the lane and release safe and identical as far as possible.
2. **Observe before movement · 4 min.** Child notices pair shapes/sizes through an accessible route or hears attributed report. Ask whether finish/weight also differs and log that as a limitation.
3. **Model · 5 min.** Use invented Model N to show that even a different endpoint does not reveal *which* factor mattered when more than one changed.
4. **Child try · 7 min.** If approved pair exists, adult performs each slow release, stopping at catch; Card 18 records source/date, actual result and access. Otherwise child marks NOT TESTED and identifies conditions a later teacher would need.
5. **Check · 4 min.** Offer “The shiny one went farther because it is plastic.” Child corrects to verified material, actual result and other possible differences; a surface look alone is insufficient.
6. **Exit · 2 min.** Leave cause unknown where controlled comparison was not possible; save question for Day 19.

## Week 4 · Day 19 — Tell what the evidence can and cannot say

**Goal:** share an accurate dated motion report with prediction, actual observation if available, relevant controls and one limit. **Proposed codes:** AC9SFH01, AC9SFI03, AC9SFI04, AC9SFI05. **Prepare:** one real dated Week 3/4 log or Model N labelled fiction, [evidence-report](evidence-report.svg), Card 19.

1. **Choose source · 3 min.** Select real log with date/conditions or write MODEL ONLY. Do not add a movement result to a blank log.
2. **Review · 4 min.** Read question, prediction, action, actual result and what stayed the same; distinguish child's direct access from adult report.
3. **Model · 5 min.** Teacher reports on Model N as fiction: “The invented A and B stop labels differ; I cannot know if shape caused it.” Identify missing controls.
4. **Child try · 7 min.** Child chooses private speech, AAC/pointing, tactile sequence or drawing on Card 19; presents source/date, one result or NOT TESTED, one cautious conclusion and one next check. Adult records only actual sharing/response.
5. **Listener · 4 min.** If a consenting listener genuinely asks something, capture their words and the child's answer. If not, write NO ACTUAL LISTENER RESPONSE; no imagined feedback claim.
6. **Exit · 2 min.** Choose a clean question for a future safe investigation, not a presumed outcome.

## Week 4 · Day 20 — Fresh two-source movement check

**Goal:** independently read two unseen movement sources, make a bounded comparison and state what could support a stronger factor claim. **Proposed codes:** AC9SFU02, AC9SFI02, AC9SFI03, AC9SFI04. **Prepare:** fresh [Day 20 check](ASSESSMENT.md), blank comparison mat, Card 20 only after first response.

1. **Welcome · 3 min.** Explain this small formative sample and no-touch/pass choices; do not preview the held pair or log.
2. **Retrieve · 4 min.** Revisit the known Model N limit only, without revealing the held fresh endpoints or objects.
3. **Fresh sources · 5 min.** Introduce an unseen safe matched pair with local dates and conditions, if approved; otherwise use the distinct invented two-record fallback. State which is actual and which is model.
4. **First independent response · 7 min.** Child identifies source, predicts before any actual release, records actual A/B result or reads model records, and explains whether one factor can be credited. Preserve exact response and prompts before feedback.
5. **Bound · 4 min.** Child marks same/different/NOT TESTED for prediction against actual trials only. They name another control or observation needed; paper values never become local data.
6. **Close · 2 min.** Use separate key to choose follow-up. One check cannot certify the whole Foundation Science year or a complete understanding description.

Original scripts © 2026 NeuroForgeIO Pty Ltd, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).