The inference
CLAIM-037 holds that formal contextuality is a failure of global extension, distinct from ordinary context dependence, and needs Contextuality-by-Default “when marginals shift with context”. It grounds on the 2016 paper, LIT-777 (read in NOTE-600). The manuscript applies the criterion to the general scenario of its §3.
Critical questions
- Does the work say this, read in context? Yes, for the 2016 criterion. But the paper’s own footnotes 1, 7 and 10 point to the authors’ later “multimaximal” version (arXiv 1604.04799, 1604.08412). CLAIM-069 (2026-10-09): “That version changes which systems count as contextual when a content appears in more than two contexts. The authors say the cyclic binary case, the consistently connected case and the measure theorems are unchanged.”
- Is it said within the scope the claim needs? No. This is the question that fails. CLAIM-069: “The manuscript cites the 2016 paper; its general observational covers need the later version.” The point bites on the general scenario of §3 and on Case II’s overlapping elicitation contexts, where an observable may be asked alongside several others. CbD 2.0 (LIT-831) and its companion (LIT-836) confirm it: verdicts change only for contents measured in three or more contexts.
- Does the claim lean on the reading for more than the reading holds? Not for the manuscript’s parity example: each of S, M and H appears in exactly two contexts, a cyclic binary system, so the 2016 criterion suffices there. The canonical-systems paper (LIT-790, THEORY-168) adds a qualification: whether a fixed set of measurements counts as contextual depends on how the system is represented. “The parity example is unaffected” assumes S, M and H are binary. If any is multi-valued and its coarsenings matter, even a single observable can come out contextual.
Proposed. The owner ruled CLAIM-069 Proposed on 2026-10-27 rather than narrowing it. Its second clause, that the change matters for observables in more than two contexts, is the part THEORY-168 keeps Proposed: LIT-831 asserts it with no worked example, and THEORY-722 shows only that the change can bite in three or more contexts, not that it does. This inference goes Rejected on THEORY-168’s own promote_when: a worked system that is noncontextual under CbD 1.0 and contextual under 2.0. THEORY-168 already holds the reading, and gains nothing.
Filed 2026-10-28 by the ADR-061 migration, from CLAIM-069.
Note of 2026-10-29: rejected on its stated condition
THEORY-168’s note of this date works the case this argument named: a binary content measured in four contexts, inconsistently connected, that is noncontextual under CbD 1.0’s maximal couplings and contextual under CbD 2.0’s multimaximal ones, with a subsystem that is contextual under 1.0. So the 2016 criterion and its replacement disagree on a content in more than two contexts, which the manuscript’s general observational covers allow, and the 2016 paper does not ground what those covers need.