Source
Gao (2016; Cambridge University Press 2017), LIT-121, chs. 1 and 4, read in full in NOTE-140 (arXiv draft).
What was actually shown
Protective measurement (§1.3). With Zeno protection over [0, τ] at N → ∞ times, |t=τ⟩ = |ψ⟩|φ(x₀ + ⟨A⟩)⟩ (eq. 1.10): the pointer shifts by ⟨A⟩ and the system is not entangled with it. By Ehrenfest, any non-disturbing measurement shifts the pointer by ⟨ψ|A|ψ⟩ (eqs. 1.12–1.13). Taking A as normalised projectors on regions, and B as a current operator, gives ρ and j and hence ψ up to a global phase (eqs. 1.14–1.16). This is the standard Aharonov–Anandan–Vaidman result (NOTE-140, C1, strong).
From there to ψ-ontology (ch. 4). Three routes:
- inside an extended ontological-models framework, with a preparation non-contextuality assumption: |0⟩ and |+⟩, each protected along its own spin direction, give 1 and ½ for a protective measurement of |0⟩⟨0|, so no λ can be compatible with both (pp. 33–35; C2, moderate);
- two versions said to need “no auxiliary assumptions” (pp. 35–37; C3, weak);
- a “weaker criterion of reality” outside the framework: a definite pointer value whose shift rate is also fixed by that value marks a real property (§4.4; C4, weak).
Gao presents the argument as free of PBR’s preparation independence and of Hardy’s ontic indifference (ch. 4).
The load-bearing premise. All three routes assume the protection procedure contains “no information about the measured wave function” (pp. 36–37). Gao’s fn. 10 (p. 37) undercuts it: Zeno protection needs an observable of which ψ is a nondegenerate eigenstate, and adiabatic protection needs ψ to be a nondegenerate eigenstate of H, so the procedure selects ψ from a discrete set. This is Combes, Ferrie, Leifer and Pusey’s objection, answered in one footnote (fn. 13, p. 40) by assertion.
What could have come out otherwise. The pointer could have shifted by a random eigenvalue, as in a projective measurement, or the shift could have needed an ensemble. In the ideal limit it does neither. Whether that bears on ψ’s reality is where the dispute sits.
What this does not say
- That ψ is ontic without assumptions. The claim to need none rests on the contested premise above (NOTE-140, C3, weak).
- That the charge distribution |ψ|²Q is a real field. Gao’s own next step denies it, on a premise about self-interaction that he calls “a fundamental assumption in physics” (p. 168) and does not argue.
- That particles move randomly and discontinuously. The random discontinuous motion ontology rests on that self-interaction premise and, for randomness, on an assumption Gao leaves for future work (p. 67; C7–C9).
- That a collapse model is required or confirmed. Collapse follows only from a further assumption that the Born probabilities originate in that motion (§8.3). The model’s predictions, a ¹⁸⁰Ta isomer superposition collapsing in about 20 minutes and a seasonal variation of collapse times, are untested (C16–C17).
Connections
- Rivals THEORY-621. Rovelli gives ψ no ontological weight and lists Rovelli (1994) among the critics of protective measurement; Gao makes ψ ontic and adds a preferred frame. Both cannot be right about ψ.
- PBR’s route to ψ-ontology, with preparation independence, is THEORY-605. Whether ψ-ontic and ψ-epistemic exhaust the options bears on Gao’s parity argument (THEORY-586).
- The readings Gao rejects in ch. 3 and §7.1, nomological and configuration-space, are mapped in THEORY-669 and THEORY-485.