Quantum probability, not classical probability, predicts how question order flips opinion — and it's now a network model

Core claims

1. A 2026 preprint models opinion dynamics on social networks by giving each agent a density matrix instead of a scalar opinion, making survey questions non-commuting operators — under a simplifying approximation it collapses back to the classical Friedkin–Johnsen opinion model.

   "Survey questions become non-commuting self-adjoint operators, which provides a principled explanation for order effects... Under a product state approximation, the quantum model reduces to the classical Friedkin–Johnsen opinion model." Source: Weiqi Chu, "A quantum model of opinion dynamics on networks," arXiv:2607.01452 (2026). Tier 1.

2. This isn't speculative borrowing — it rests on a real, replicated empirical anomaly. Across 70 national representative surveys (651–3,006 participants each), the order in which attitude questions were asked shifted responses in a way ordinary probability can't predict, but a quantum "QQ equality" derived a priori did.

   "The results provided strong support for the predicted QQ equality... quantum probability theory, initially invented to explain noncommutativity of measurements in physics, provides a simple account for a surprising regularity regarding measurement order effects in social and behavioral science." Source: Wang, Solloway, Shiffrin & Busemeyer, PNAS 111(26), 2014. Tier 1.

3. The theory isn't uncontested: a 2016 experimental paper found quantum-like models fail to predict the conjunction fallacy under a question-order design, "at odds with the predictions of the quantum-like models" (Boyer-Kassem, Duchêne & Guerci, arXiv:1606.03430). Saved as a hook, not chased further this chain.

Why this was hop-worthy

A formalism built to explain why measuring an electron's spin twice in different orders gives different answers (1920s quantum mechanics) is now explaining why asking "is Clinton honest?" before vs. after "is Gore honest?" changes poll numbers (2014) — and is being wired directly into network opinion-dynamics models (2026), a genuine cross-domain and cross-century bridge.

Further leads

• Does Chu's network model make a testable prediction that distinguishes it from Friedkin–Johnsen on real polarization data, beyond reducing to it as a special case?
• Boyer-Kassem et al.'s conjunction-fallacy failure — is it a fatal flaw or a scope limit (order effects vs. conjunction judgments may need different quantum operators)?

Hop chain

Hop 1: arXiv cs.SI recent listing — https://arxiv.org/list/cs.SI/recent

• Hook type: mechanism question / cross-domain bridge (spotted while scanning the feed)
• Hook: "A quantum model of opinion dynamics on networks" — a physics formalism applied to a network-science listing
• Why followed: cross-domain bridge is the spec's highest-priority hook type, and vault_novelty scored it 0.588 (frontier, not duplicate)
• Key findings: Weiqi Chu (UMass Amherst) models agent opinions as density matrices; survey questions become non-commuting operators; the model reduces to the classical Friedkin–Johnsen model as a special case and was tested on the Caltech Facebook-100 network.

Hop 2: Chu 2026, arXiv:2607.01452 — https://arxiv.org/abs/2607.01452

• Hook type: the person/field behind the thing
• Hook: the paper explicitly cites Busemeyer's "quantum cognition" program as its justification for using non-commutative probability
• Why followed: zooming out from one paper's mechanism to the field it draws on — and vault_novelty flagged it as adjacent (0.627) to an existing consciousness/self-knowledge note, suggesting a real (if loose) vault connection
• Key findings: Quantum cognition (Busemeyer & Bruza, Cambridge UP 2012) is an established ~15-year-old research program using quantum probability (not claims about quantum brains) to explain "irrational" judgment phenomena like order effects and conjunction fallacies.

Hop 3: Wang, Solloway, Shiffrin & Busemeyer 2014, PNAS — https://www.pnas.org/doi/10.1073/pnas.1407756111

• Hook type: surprising claim (quantitative)
• Hook: an a priori "QQ equality" prediction tested against 70 real national surveys, not a toy lab result
• Why followed: this is the empirical anchor for the whole chain — the hop protocol flags numbers that seem too clean to be true, and r = −.82 across 70 independent surveys is exactly that
• Key findings: Order-effect pairs across the surveys fell close to the predicted line (intercept 0, slope −1), giving quantum probability theory a genuine, replicated foothold in survey methodology, not just decision-theory papers.

Hop 4: Boyer-Kassem, Duchêne & Guerci 2016 — https://arxiv.org/abs/1606.03430

• Hook type: surprising claim (contrarian/reversal)
• Hook: a direct experimental test where quantum-like models' predictions failed
• Why followed: logged as a saved hook rather than chased — the chain had already found its empirical anchor (hop 3) and this is best used as a counterweight in a future note on the theory's limits, not a new destination
• Key findings: quantum-like models correctly predict order effects but not the conjunction fallacy under their experimental design — the theory's success is scoped, not universal.

Saved hooks not followed:

• Friedkin–Johnsen classical opinion model — from Chu 2026 — interesting mechanism-question hook (what does the classical model already predict, that quantum coherence is added on top of?) but it's well-trodden network-science territory, lower priority than the cross-domain bridge.
• Facebook-100 dataset (Caltech network, 769 agents, fall 2005) — from Chu 2026 — unfamiliar-name hook, a reusable benchmark dataset in network science worth knowing about but not a thread in itself.
• Gersick's 1991 punctuated-equilibrium model (surfaced in the final novelty check's nearest neighbors) — cross-domain curiosity for a future chain: does opinion "coherence decay" have a punctuated-equilibrium analog in group decision-making?

post-worthy: maybe — the cross-century bridge (quantum formalism → 1970s-80s opinion dynamics math → 2026 network model) is a strong Seek-shaped find, but it needs one more hop into whether Chu's model makes a falsifiable prediction beyond Friedkin–Johnsen before it's a full post.