L3
The Didactic Model
The score for L1's performance
L3 defines the principles by which learning happens; L1 executes them in the moment. Seven principles: intent-based navigation, the Socratic method, dynamic difficulty, functional embedding, content polymorphism (the musician learns fractions as rhythm intervals), progressive de-adaptation that systematically reduces support as competence grows, and collaborative learning that connects learners across hubs and time zones.
Score and performance
L3 is the score; L1 is the performance. L3 defines the principles by which learning happens. L1 executes them in the moment. The separation matters: when pedagogical decisions are baked into the generative model, you cannot inspect them, contest them, or swap them out per jurisdiction. When they live in L3, you can.
Seven principles
Seven principles govern the model.
Intent-based navigation. The process begins with an intention, not a curriculum. "I want to understand why my bridge collapsed."
The Socratic method. The Mentor asks back rather than answering.
Dynamic difficulty. The learner is held in the zone of proximal development — challenged but not crushed.
Functional embedding. Fundamentals are woven into the learning path at the moment they become necessary, not as a prerequisite gate.
Content polymorphism. The musician learns fractions as rhythm intervals; the architect as geometric ratios. The logical understanding is non-negotiable, but the form adapts.
Progressive de-adaptation. With growing competence, the system deliberately reduces individualisation, building the learner's capacity to function without it.
Collaborative learning. The system connects learners — co-located at a hub, synchronously across distance with real-time translation, or asynchronously across time — matched by complementary strengths, shared interests and compatible levels.
Progressive de-adaptation
This principle deserves elaboration because it addresses the strongest pedagogical objection to the architecture: a system this responsive creates dependency. The objection is valid. A learner who has never navigated a poorly structured resource, endured ambiguity, or decided what to study next has not developed the self-regulation the real world demands.
De-adaptation is the architectural answer. It begins at intermediate competency and escalates through defined stages. First, content polymorphism is reduced (the musician must see fractions as geometry, not just rhythm). Then scaffolding is reduced (hints become less specific; wait times before intervention grow longer). Then deliberate friction is introduced (ambiguous problems, resources with gaps). Finally, the adaptive profile is withdrawn entirely for defined periods — the learner works with generic materials and self-directs.
The system measures de-adaptation readiness through one signal: can the learner achieve comparable outcomes with reduced support? If performance collapses, the system re-engages. The goal is not a learner who needs the Mentor forever. It is a learner who outgrows it.
Eleven doors. No floor.
Why this is L3, not L1
A platform that bundles pedagogy into its generative model has no seam at which a school district, a national education ministry, or a learner cooperative can disagree. They can ban the platform, or accept it; they cannot rewrite its didactics. L3 is the seam. The same Mentor (L1), constrained by a different L3, produces a different learning experience — and the disagreement happens in the open, in a layer designed to be inspected and contested.
Reference
Architecture paper, Section 5, L3. DOI: 10.5281/zenodo.18759134. CC BY 4.0.