Dynamic Probability Logics: Axiomatization & Definability

We first study probabilistic dynamical systems from logical perspective. To this purpose, we introduce the finitary dynamic probability logic} ($\mathsf{DPL}$), as well as its infinitary extension $\mathsf{DPL}_{\omega_1}\!$. Both these logics extend the (modal) probability logic ($\mathsf{PL}$) by adding a temporal-like operator $\bigcirc$ (denoted as dynamic operator) which describes the dynamic part of the system. We subsequently provide Hilbert-style axiomatizations for both $\mathsf{DPL}$ and $\mathsf{DPL}_{\omega_1}\!$. We show that while the proposed axiomatization for $\mathsf{DPL}$ is strongly complete, the axiomatization for the infinitary counterpart supplies strong completeness for each countable fragment $\mathbb{A}$ of $\mathsf{DPL}_{\omega_1}\!$. Secondly, our research focuses on the (frame) definability of important properties of probabilistic dynamical systems such as measure-preserving, ergodicity and mixing within $\mathsf{DPL}$ and $\mathsf{DPL}_{\omega_1}$. Furthermore, we consider the infinitary probability logic $\mathsf{InPL}_{\omega_1}$ (probability logic with initial probability distribution) by disregarding the dynamic operator. This logic studies {\em Markov processes with initial distribution}, i.e. mathematical structures of the form $\langle \Omega, \mathcal{A}, T, \pi\rangle$ where $\langle \Omega, \mathcal{A}\rangle$ is a measurable space, $T: \Omega\times \mathcal{A}\to [0, 1]$ is a Markov kernel and $\pi: \mathcal{A}\to [0, 1]$ is a $\sigma$-additive probability measure. We prove that many natural stochastic properties of Markov processes such as stationary, invariance, irreducibility and recurrence are $\mathsf{InPL}_{\omega_1}$-definable.