ABSTRACT

Existing methods to determine the potential of the Milky Way from a Gaia-like snapshot of the kinematics all depend on the potential being integrable and the orbiting stars being phase-mixed. Neither of these things can conceivably be true; integrability and phase-mixing are both strongly ruled out, observationally and theoretically. I describe some ideas we are exploring to perform the dynamical inference in our "worst of all possible worlds". The key concept is that locally, phase-space structure evolves predictably. The audience is warned that I will talk much more about data than about theory, that only toy problems will be explicitly solved, and that hands will be waved.