Every physical theory has to confront two issues: 1) What are the meaningful questions ("observables")? 2) What do they mean?i.e., what part of the theory purports to correspond to facts in the universe, and how? In quantum gravity (QG) these issuessometimes become particularly entangled. This talk will reviewthe causal set approach to QG and then describe recent advances that go some way toward answering these questions, showing how they find a clear formulation and solution in the case of a toy stochastic model of causal set cosmology, and giving brief prospects for the extension to a quantum theory. The toy stochastic model defines an evolution for posets that is analogous to the dynamical models for the growth of massive random graphs which are of current interest.