Quasar pairs are either physically distinct binary quasars or the result of gravitational lensing. The majority of known pairs are in fact lenses, with a few confirmed as binaries, leaving a population of objects that have not yet been successfully classified. Building on the arguments of Kochanek, Falco & Muñoz, it is shown that there are no objective reasons to reject the binary interpretation for most of these. In particular, the similarity of the spectra of the quasar pairs appears to be an artefact of the generic nature of quasar spectra. The two ambiguous pairs discovered as part of the Large Bright Quasar Survey (Q 1429-053 and Q 2153-0256) are analysed using principal components analysis, which shows that their spectral similarities are not greater than expected for a randomly chosen pair of quasars from the survey. The assumption of the binary hypothesis allows the dynamics, time-scales and separation distribution of binary quasars to be investigated and constrained. The most plausible model is that the activity of the quasar is triggered by tidal interactions in a galatic merger, but that the (re-)activation of the galactic nuclei occurs quite late in the interaction, when the nuclei are within 80±30 kpc of each other. A simple dynamical friction model for the decaying orbits reproduces the observed distribution of projected separations, but the decay time inferred is comparable to a Hubble time. Hence it is predicted that binary quasars are only observable as such in the early stages of galactic collisions, after which the quiescent supermassive black holes orbit in the merger remnant for some time.