yuji-tt Structural Transfer
Before naming donor fields, derive the target signature: entities, observable information, hidden state, actions, feedback, objective, constraints, timescale, dominant failure mode, invariants or monotone quantities, and information or resource bottleneck. Mark unknown structure explicitly.
Search for the same constraint-and-feedback pattern in distant fields, not matching terminology. Seek diverse donors without forcing a fixed domain list. Use transfer protocol for search, comparison, and rejection rules. Distinguish verified source mechanisms from speculative constructions; support specialized source claims with primary references when retrieval is available.
For each surviving proposal provide:
source mechanism → abstract structural principle → mapping to target → concrete intervention → predicted behavioral signature → fastest falsification test → reasons the analogy may break.
Map causal relationships and operating conditions as well as entities. Reject a proposal when a required resource, information channel, invariant, or timescale has no target counterpart. Similar labels are not evidence of transfer. A useful transfer must change an implementable target mechanism and predict behavior distinguishable from the baseline.
Select the minimal intervention with the best information value relative to cost. Do not claim novelty without a literature check or success before testing. User instructions take precedence; an explicit transfer request may be exploratory even without a demonstrated plateau, but label that uncertainty.