Learning Castle Return Schema

This schema defines how every Learning Castle route returns: first to the correct canonical owner where necessary, then toward the learner’s portable capability, and finally to school, examinations and life. It is the machine-readable companion to the Castle Machine Manifest, Castle Registry, Route Schema, Fort Schema and Edge Schema.

schema_id: /learning-castle/0105
schema_version: 0.1
system: eduKateSengkang Learning Castle
object_type: return
runtime_parent: /learning-castle/0100
keep: /learning-castle/0019
open_gate: /learning-castle/0020
return_road: /learning-castle/0048
principle: support_is_temporary_capability_should_return_to_learner

Return Contract

return:
  from: castle_object_or_live_node
  local_owner_return: castle_owner_or_null
  learner_return: /learning-castle/0019
  world_return: /learning-castle/0020
  support_state: full | partial | faded | removed
  transfer_test: required_when_learning_job_occurred
  evidence_required: true
  completion_status: unresolved | improving | transferred | independent
  dead_end_allowed: false_for_learning_routes

Return Order

1. restore canonical owner if route crossed ownership
2. confirm the intended learning or information job is complete
3. reduce scaffold or tool support when appropriate
4. test under changed or delayed conditions when appropriate
5. return capability to /learning-castle/0019
6. open /learning-castle/0020 when independent deployment is justified

Local Owner Return

if route_type = bridge
AND source_owner_remains_primary
THEN return_to_source_owner_before_keep

if route_type = archive_road
THEN return_to_owner_hub

if route_type = workshop
THEN return_to_parent_fort_or_runtime

if route_type = service_road
THEN preserve_distinction_between_service_owner_and_knowledge_owner

Capability Return Test

return_to_keep only_when one_or_more_apply:
  learner_can_reconstruct_without_model
  learner_can_retrieve_after_delay
  learner_can_select_route_without_prompt
  learner_can_handle_meaningful_variation
  learner_can_detect_and_correct_error
  learner_can_explain_or_justify
  learner_can_perform_with_reduced_support

note:
  one_successful_familiar_attempt_is_not_sufficient_evidence_of_independence

Support Fading

support_fading:
  stage_1: model_available
  stage_2: prompts_reduced
  stage_3: learner_reconstructs
  stage_4: conditions_vary
  stage_5: delay_introduced
  stage_6: independent_performance
  stage_7: constrained_or_real_world_deployment

rule:
  do_not_remove_support_before_useful_success_is_possible
  do_not_keep_support_after_it_becomes_the_bottleneck

Fort Return Rules

english_fort:
  return_when: language_capability_survives_changed_context

mathematics_fort:
  return_when: structure_method_and_verification_can_be_deployed_independently

science_fort:
  return_when: learner_can_observe_model_explain_and_update_with_evidence

examination_fort:
  return_when: learned_capability_can_be_deployed_under_realistic_constraints

Bridge Return Rules

science -> english:
  return_to: science_fort
  when: language_bottleneck_repaired

mathematics -> english:
  return_to: mathematics_fort
  when: mathematical_language_bottleneck_repaired

subject_fort -> examination_fort:
  return_to: keep_or_open_gate
  when: constrained_deployment_is_secure

rule:
  bridge_does_not_transfer_canonical_ownership_unless_explicitly_declared

Service Return Rules

if intent = tuition_service:
  service_owner_may_teach
  but_learning_goal_remains_in_relevant_fort_or_runtime

successful_service_return:
  less_external_support_needed_over_time
  more_capability_owned_by_learner
  clearer_route_back_to_subject_owner_or_keep

block_if:
  service_dependency_is_mistaken_for_learning_success

Archive Return Rules

archive_job: enumerate_and_help_discovery
archive_must_not_become_dead_end
archive_return_target: canonical_owner_hub
archive_ownership_transfer: false

World Return

open_gate_when:
  capability_is_sufficiently_portable
  AND context_requires_real_deployment

world_targets:
  school
  homework
  independent_study
  examination
  new_teacher
  unfamiliar_problem
  project
  later_stage_of_learning
  life

success_question:
  what_can_the_learner_now_do_outside_the_castle_that_was_not_reliably_possible_before

Dead-End Control

fail_route_if:
  no_local_owner_return_exists_when_required
  OR no_keep_return_exists_for_learning_route
  OR support_never_fades_without_reason
  OR archive_traps_user_without_owner_route
  OR bridge_loses_source_ownership
  OR service_route_becomes_permanent_dependency

AI Return Procedure

1. ask whether the current node is owner, support, bridge, archive or service
2. restore canonical owner if necessary
3. check whether the learner can now perform with less support
4. if not, route to the smallest remaining weak link
5. if yes, test transfer where meaningful
6. if transfer holds, return to /learning-castle/0019
7. if real-world deployment is appropriate, open /learning-castle/0020
8. do not create a dead end

The Learning Castle is complete only when movement does not end inside the website. Its final purpose is to make more capability portable outside the website, outside tuition, and outside the immediate support environment.