1: \begin{abstract}

2:   We study the relationship between functional inequalities for a

3:   Markov kernel on a metric space $X$ and inequalities of

4:   transportation distances on the space of probability measures

5:   $\mathcal{P}(X)$.  Extending results of Luise and Savar\'e on

6:   Hellinger--Kantorovich contraction inequalities for the particular

7:   case of the heat semigroup on an $RCD(K,\infty)$ metric space, we

8:   show that more generally, such contraction inequalities are

9:   equivalent to reverse Poincar\'e inequalities.  We also adapt the

10:   ``dynamic dual'' formulation of the Hellinger--Kantorovich distance

11:   to define a new family of divergences on $\mathcal{P}(X)$ which

12:   generalize the R\'enyi divergence, and we show that contraction

13:   inequalities for these divergences are equivalent to the reverse

14:   logarithmic Sobolev and Wang Harnack inequalities.  We discuss applications

15:   including results on the convergence of Markov processes to

16:   equilibrium, and on quasi-invariance of heat kernel measures in

17:   finite and infinite-dimensional groups.

18: \end{abstract}

19: