Quantum-like cognition models predict question-order effects but fail to account for the conjunction fallacy

The quantum-cognition program is not uncontested. A 2016 experimental paper by Thomas Boyer-Kassem, Sébastien Duchêne, and Eric Guerci — "Quantum-like models cannot account for the conjunction fallacy" (arXiv:1606.03430; published in Theory and Decision 81:479–510) — designed a question-order-effect configuration to test quantum-like models on the conjunction fallacy (judging a conjunction "A and B" more probable than one of its conjuncts). The observed behavior came out "at odds with the predictions of the quantum-like models."

The upshot is a scoping, not a refutation of everything upstream: the same family of models that lawfully captures survey question-order effects does not, under this experimental design, capture the conjunction fallacy. That tension is the natural counterweight to hold against the network model that inherits the order-effect machinery (claim-quantum-opinion-model-reduces-to-friedkin-johnsen): a formalism's success at one anomaly is not a warrant for its success at a neighboring one.

Whether this is a fatal flaw or a scope limit — order-effect judgments and conjunction judgments may simply require different quantum operators — is an open thread (question-quantum-cognition-conjunction-fallacy-fatal-or-scope-limit), now partially addressed by reading the source paper's own hypothesis structure and conclusion in full: see claim-boyer-kassem-refutation-rejects-non-degenerate-and-degenerate-models for which model classes actually fall, and claim-boyer-kassem-names-scope-retreat-and-povm-fix-as-untested-escape-routes for what the authors themselves leave open.