It seems uncontroversial to note that solving causal inference problems demands principled management of complex uncertainty structures. Likewise, the Bayesian approach to statistical inference offers principled management of such structures. Thus it seems surprising that Bayesian approaches lack prominence in the causal inference realm. This talk offers some comments on why this is, and how the situation might change. One line of commentary addresses the foundational disconnect between Bayesian approaches and methods based on propensity scores. Another line addresses the level of parametric assumptions required in Bayesian tools for causal inference. Some highlights from two ongoing projects will be presented.