General domain of applicability of the quantum averages of weak values (QAWV) framework: mixed and complex apparatus priors, multi-trajectory classical correspondence, quantitative robustness bounds for eccentric weak values, and the validity domains of the framework's approximations

Aharonov and Botero's quantum averages of weak values (QAWV) framework reinterprets the outcome statistics of arbitrary-strength von Neumann measurements on pre- and postselected systems as a Bayesian re-assessment and averaging of weak values over a posterior distribution of the apparatus coupling parameter. The framework was derived under restrictive assumptions: pure pre- and postselected ensembles, a real and smooth initial apparatus wave function, a single classical trajectory in the WKB (semiclassical) limit for the quantum-classical correspondence, and approximations (first-order group-velocity expansion, Gaussian posterior distributions, large-j spin asymptotics) whose domains of validity were not characterized. The open problem is to establish the general domain of applicability of the QAWV framework: extend the weak-value-averaging interpretation to general mixed system and apparatus states and to complex, non-Gaussian apparatus priors where interference between distinct sampling regions is essential; extend the quantum-classical correspondence to regimes with multiple classical trajectories and strongly non-WKB dynamics; quantify the robustness condition on prior distributions needed to sample eccentric weak values against exponential likelihood suppression, including practical experimental bounds and protocols; and characterize the parameter regimes in which the framework's approximations accurately reproduce full quantum statistics. Later literature has developed adjacent formalisms (nonperturbative weak-measurement theory, mixed-state quasi-probability descriptions, superoscillation mathematics, meter-interference analysis) but has not completed this program within the QAWV framework itself.

Background

In the von Neumann model of quantum measurement, a system observable $\hat{A}$ is coupled impulsively to a measuring device (the 'apparatus' or 'meter') whose pointer variable $\hat{p}$ is displaced by an amount proportional to $\hat{A}$ . For a system prepared in a state $|\psi_1\rangle$ (preselection) and later found in a state $|\psi_2\rangle$ (postselection), the weak value of $\hat{A}$ is defined as $A_w = \langle\psi_2|\hat{A}|\psi_1\rangle/\langle\psi_2|\psi_1\rangle$ . Weak values arise as pointer shifts in 'weak measurements', where the system-apparatus coupling is made small so that the back-reaction on the system is minimized; the real part of $A_w$ can lie far outside the eigenvalue spectrum of $\hat{A}$ , which motivated a long debate on their physical meaning.

Aharonov and Botero (Phys. Rev. A 72, 052111 (2005)) addressed this debate in two steps. First, they gave an operational interpretation of the weak value as a definite mechanical effect on a probe designed to minimize back-reaction. Second, they introduced the quantum averages of weak values (QAWV) framework: for a pre- and postselected measurement ensemble and an arbitrary-strength von Neumann measurement, the pointer statistics can be written as an average of weak values $A_{12}(q)$ over a posterior distribution of the coupling parameter $q$ , where the initial apparatus wave function $\phi(q)$ is re-assessed by a likelihood factor $\sqrt{P_{12}(q)}$ (a quantum analogue of Bayesian updating) and then transformed by a polar-decomposition unitary generated by the weak value. In the classical limit they showed, under semiclassical (WKB, named after Wentzel, Kramers, and Brillouin) conditions and assuming a unique classical solution of the boundary-value problem defined by the pre- and postselection, that the quantum posterior expectation of a dynamical variable coincides with the posterior expectation of the corresponding classical quantity evaluated on that trajectory, so that ordinary classical properties can be viewed as weak values.

The derivations rest on simplifying assumptions. The QAWV construction is presented for pure pre- and postselected ensembles with an initial apparatus wave function taken to be real and smooth; more general ensembles are treated only by a classical (convex) decomposition into pure subensembles, which discards interference between sampling regions. In the spin- $j$ illustration (a spin whose total angular-momentum quantum number is $j$ ), the likelihood $P_{12}(q)\propto J_z(q)^{-2j}$ suppresses sampling of the most eccentric weak values — weak values lying far outside the eigenvalue range of the measured observable, arising near the boundary point $q_+$ of the allowed coupling-parameter domain — exponentially. This suppression is a Fourier superoscillation effect (band-limited functions oscillating locally faster than their fastest Fourier component), and the authors give only a qualitative 'robustness' condition on the prior (sufficient sharpness plus fast tail fall-off) for observing such values. The first-order group-velocity expansion, Gaussian posterior approximations, and large- $j$ asymptotics are used without a quantitative characterization of where they fail. The purpose of the present problem is to remove these restrictions and thereby decide how general the QAWV description of the measurement process really is.

Problem Statement

Establish the general domain of applicability of the quantum averages of weak values (QAWV) framework of Aharonov and Botero (2005), in which the outcome statistics of an arbitrary-strength von Neumann measurement on a pre- and postselected quantum system are described as a Bayesian re-assessment and averaging of weak values over a posterior distribution of the apparatus coupling parameter. The problem is to extend the framework beyond the simplifying assumptions under which it was derived, namely: (i) determine whether the QAWV description can be extended to the general class of mixed system states together with general mixed, complex-valued, and non-Gaussian initial apparatus states, including the regime in which the re-assessed apparatus state develops several well-separated posterior peaks and interference between the corresponding sampling regions is essential to the pointer statistics, constructing that description where it exists and establishing where it does not; (ii) determine whether the quantum-classical correspondence between averages of weak values and posterior classical expectation values can be extended to boundary-value problems admitting multiple classical trajectories and to strongly non-WKB (non-semiclassical) dynamics, where interference between classical paths may contribute, extending that correspondence where it holds and establishing where it does not; (iii) turn the qualitative robustness condition for observing eccentric weak values (weak values lying far outside the eigenvalue range of the measured observable; the condition combines prior sharpness with sufficiently fast tail fall-off against the exponentially rising likelihood, e.g. $P_{12}(q)\propto J_z(q)^{-2j}$ in the spin- $j$ case) into quantitative bounds on the required prior and apparatus properties, together with realistic experimental protocols or an argument that none exist; and (iv) characterize the parameter regimes in which the framework's approximations — the first-order group-velocity expansion, the Gaussian posterior approximation, and the large- $j$ asymptotics — accurately reproduce the full quantum statistics. These four components together serve one objective: to determine whether the weak-value-averaging interpretation is a general description of von Neumann measurement statistics, of which the pure-state, real-smooth-prior, single-trajectory, weak-coupling cases are special limits.

References

Y. Aharonov and A. Botero, 'Quantum averages of weak values', Phys. Rev. A 72, 052111 (2005). DOI: 10.1103/PhysRevA.72.052111; https://doi.org/10.1103/PhysRevA.72.052111 (arXiv: quant-ph/0503225, https://arxiv.org/abs/quant-ph/0503225)
A. G. Kofman, S. Ashhab, and F. Nori, 'Nonperturbative theory of weak pre- and post-selected measurements', Phys. Rep. 520, 43-133 (2012). DOI: 10.1016/j.physrep.2012.07.001; https://doi.org/10.1016/j.physrep.2012.07.001 (arXiv: 1109.6315, https://arxiv.org/abs/1109.6315)
J. Dressel, 'Weak values as interference phenomena', Phys. Rev. A 91, 032116 (2015). DOI: 10.1103/PhysRevA.91.032116; https://doi.org/10.1103/PhysRevA.91.032116
J. Dressel, M. Malik, F. M. Miatto, A. N. Jordan, and R. W. Boyd, "Colloquium: Understanding quantum weak values: Basics and applications", Rev. Mod. Phys. 86, 307 (2014). DOI: 10.1103/RevModPhys.86.307; https://doi.org/10.1103/RevModPhys.86.307
Y. Turek, H. Kobayashi, T. Akutsu, C.-P. Sun, and Y. Shikano, 'Post-selected von Neumann measurement with Hermite-Gaussian and Laguerre-Gaussian pointer states', New J. Phys. 17, 083029 (2015). DOI: 10.1088/1367-2630/17/8/083029; https://doi.org/10.1088/1367-2630/17/8/083029
A. Karanjai, E. G. Cavalcanti, S. D. Bartlett, and T. Rudolph, 'Weak values in a classical theory with an epistemic restriction', New J. Phys. 17, 073015 (2015). DOI: 10.1088/1367-2630/17/7/073015; https://doi.org/10.1088/1367-2630/17/7/073015
A. Hosoya and Y. Shikano, 'Strange weak values', J. Phys. A: Math. Theor. 43, 385307 (2010). DOI: 10.1088/1751-8113/43/38/385307; https://doi.org/10.1088/1751-8113/43/38/385307
Y. Aharonov, F. Colombo, S. Nussinov, I. Sabadini, D. Struppa, and J. Tollaksen, 'Superoscillation phenomena in SO(3)', Proc. R. Soc. A 468, 3587-3600 (2012). DOI: 10.1098/rspa.2012.0131; https://doi.org/10.1098/rspa.2012.0131
Y. Aharonov, F. Colombo, I. Sabadini, D. C. Struppa, and J. Tollaksen, 'The mathematics of superoscillations', Mem. Amer. Math. Soc. 247, no. 1174 (2017). DOI: 10.1090/memo/1174; https://doi.org/10.1090/memo/1174
A. N. Jordan, Y. Aharonov, D. C. Struppa, F. Colombo, I. Sabadini, T. Shushi, J. Tollaksen, J. C. Howell, and A. N. Vamivakas, 'Superphenomena for arbitrary quantum observables', Phys. Rev. A 110, 012206 (2024). DOI: 10.1103/PhysRevA.110.012206; https://doi.org/10.1103/PhysRevA.110.012206
A. N. Jordan, J. C. Howell, N. Vamivakas, and E. Karimi, 'Superoscillations and Physical Applications', in: Operator Theory, Springer Basel (2025), pp. 1-33. DOI: 10.1007/978-3-0348-0692-3_101-1; https://doi.org/10.1007/978-3-0348-0692-3_101-1 (arXiv: 2505.22925, https://arxiv.org/abs/2505.22925)
H. F. Hofmann, T. Matsushita, M. Isobe, and J. R. Hance, 'What meter interference can tell us about the statistics of weak measurements' (2026). arXiv: 2608.05494; https://arxiv.org/abs/2608.05494
A. C. Ipsen, 'Anomalous Weak Values are Caused by Disturbance', Found. Phys. 52, 20 (2022). DOI: 10.1007/s10701-021-00534-w; https://doi.org/10.1007/s10701-021-00534-w
C. C. Payne and E. Cohen, 'Anomalous Weak Pointer Shifts as Postselected Interference among Coarse-Grained Histories' (2026). arXiv: 2608.15209; https://arxiv.org/abs/2608.15209