Abstract. Blanco et al. (2014) show in a novel experiment the presence of intrinsic interactions between the preferences and the beliefs of participants in social dilemma games. They discuss the identification of three effects, and we claim that two of them are inherently of non-classical nature. Here, we discuss qualitatively how a model based on complementarity between preferences and beliefs in a Hilbert space can give an structural explanation to two of the three effects the authors observe, and the third one can be incorporated into the model as a classical correlation between the observations in two subspaces. Quantitative formalization of the model and proper fit to the experimental observation will be done in the near future, as we have been given recent access to the original dataset.