PL
The dispositional account of quantum properties faces the following circularity problem: properties of a system are defined as dispositions (probabilistic or deterministic) to give rise to certain outcomes upon measurements, but measurements in turn are generally characterized with reference to the very same dispositions. I consider one way of escaping the difficulty with regard to probabilistic dispositions by applying a theorem due to Peter Mittelstaedt. The theorem enables us to give a probability-free characterization of quantum measurements, thus eliminating the need of referring back to probabilistic dispositions of the system. However, the circularity problem remains for deterministic dispositions. I give arguments why we should resist the temptation to interpret eigenstates as categorical properties, and I discuss possible alternative solutions to the problem.