learning

---

## Learning (study layer)

A facet of Play: a way to be *brought up to speed on* a domain and tested against it. Purely additive and post-hoc — like the Decision Matrix, it never mutates the force deltas.

**Question bank** — Per-scene **multiple-choice questions** the engine generates from a finished scene, stored on `scene.questions`. The atomic unit of the study layer.

**Scoped practice** — Running the banks across any slice of the domain — a scene, an arc, the whole domain, or a **topic** — from the **Learn** modal. The same banks back every learning surface; the scope decides which questions are drawn.

**Curriculum** — A reorganisable **Topic** tree that organises the question bank into a teachable structure (each question assigned 1:1 to a topic). The substrate read as a syllabus. The knowledge tree is **compression of understanding** — prerequisites, conceptual ordering, dependency, progressive refinement toward teachability — not merely UI hierarchy; it is how the artifact externalises "what a domain expert knows, in the order you'd learn it."

**Coverage** — Per-**member** spaced-repetition recall layered over the curriculum: who has been tested on what, and what is due. Turns *Onboarding decks* (passive review) into *active recall* — the testing counterpart to the deck.

Raw Markdown source