MAYA Scoring

Preamble

MAYA Scoring treats every number as a claim under uncertainty. The score has to say which audience it imagines, what evidence it carries, and what would make it wrong.

---

Acceptable To Whom

Most Advanced Yet Acceptable sounds clean until the hidden audience appears. Acceptable to a founder, buyer, reader, subculture, hiring manager, critic, child, engineer, collector, or tired person on a phone all produce different edges.

The score needs an assumed audience, context, prior exposure, reference set, and confidence level. Without those, the number is a mask for the model’s training distribution.

Score Claims Carry Evidence

The score surface covers novelty, acceptability, usefulness, carrier strength, payload clarity, testability, divergence, rejection risk, and confidence. Each score needs evidence, counter-evidence, rationale, uncertainty, and status: hypothesis, tested, rejected, or strengthened.

Low-evidence scores remain hypotheses. They can guide a next test. Validation waits for evidence.

Conformity Pressure Stays Experimental

Conformity Pressure may track rewrite pressure, payload preservation, hedging pressure, refusal pressure, and cross-model disagreement. It can suggest where a model tries to normalize an idea.

The signal stays experimental because model pressure differs from world feedback. The branch lands when score claims carry evidence, confidence, and visible uncertainty, so low-certainty judgments cannot borrow authority from numbers.