Source
Grangier & Auffèves (2018), LIT-077, pp. 3–5 (NOTE-044). Spekkens (2014), LIT-007, Table II. Abramsky & Brandenburger, LIT-016. Schmid et al., LIT-003.
What was actually shown
LIT-077 postulates that the real is relative to a context: modalities belong jointly to system and context, and are mutually exclusive within one (Postulates 0–2, p. 3). Its one argued result is informal: with N exclusive modalities per context, certainty on returning to a first context after an incompatible second one would give N² > N joint modalities (p. 4). It defines no measurement cover, empirical model or operational equivalence, derives no noncontextuality inequality, and uses the Kochen–Specker and Bell results only as motivation; its Born-rule claim is cited, not shown (NOTE-044).
The phenomenon the argument uses — losing certainty about one question after asking an incompatible one — is reproduced by Spekkens’s epistemically restricted classical bit, and LIT-007 files it as only weakly nonclassical (Table II, asserted). And a single system probed context after context presents no overlapping contexts, so no global-section obstruction in the sense of LIT-016 can arise. Both formal notions ask whether a context-independent model of the statistics exists; contextual objectivity is a claim about what exists.
What this does not say
- That contextual objectivity is wrong, or that its authors’ separate derivation of unistochastic transition matrices fails; that derivation is not held and has not been read.
- It says only that the work held here does not establish contextuality in the Kochen–Specker or Spekkens sense.