Reference Sets

Preamble

Reference Sets give the council ballast. Novelty, acceptability, carrier strength, and generic failure become challengeable only when the score has examples to answer to.

---

Examples Smuggle Judgment

An example set is never neutral. It carries a history of what the user noticed, what the domain rewarded, what the model can retrieve, and what the brief treats as relevant. Bad references make weak ideas look brave. Narrow references make familiar ideas look new.

The comparison surface separates familiar carriers, generic defaults, successful MAYA-style examples, rejected too-novel attempts, adjacent-domain patterns, and anti-patterns. Each example needs source, provenance, relevance, and the reason it belongs beside the idea.

The Score Needs Something To Push Against

A council lens can say an idea is advanced. The reference set asks advanced compared with what.

A lens can say an idea is acceptable. The reference set asks who has accepted something nearby, under which conditions, and where similar attempts failed. The examples do not prove the answer. They make the answer harder to fake.

Bias Stays In View

The reference set also exposes its own weakness: missing audiences, thin samples, overfamiliar domains, stale examples, and source clusters that tilt the council toward one taste world.

The branch lands when a score claim can point to the examples that shaped it and leave enough trail for later review.