Interpreting weak value amplification with a toy realist model

Constructing an ontology for quantum theory is challenging, in part due to measurement back-action. The Aharonov-Albert-Vaidman weak measurement formalism provides a method to predict measurement results (weak values) when back-action is negligible. The weak value appears analogous to a classical conditional mean, yet can be complex and unbounded. We study weak values in the context of a recent quantum optical experiment involving two-photon interactions. The results of the experiment are reinterpreted within a realist ‘stochastic optics’ model of light. We show that the conditional means of the intensities in the model correspond to the experimentally observed weak values and study the breakdown of the model outside the experimentally probed regime in the limit where the weak value predicts ‘anomalous’ results.