Frequentist validity of Bayesian limits (Q2656593)

This article belongs to a series of author's papers on Bayesian statistics, actually a book on ``The frequentist theory of Bayesian statistics'' is reported to be in preparation for Ch. 8. Springer, Berlin. The investigation here is mainly done around the generalization of Schwartz's Kullback-Leiber (KL) and Ghosal-Ghosh-van der Vaart (GGV) conditions to reach a wider range of frequentist applications of posterior convergence. This is done in the context of Bayesian statistics. All along the paper one continuously refers to similar results in the literature, one compares, generalizes some of the ideas, uses them in proofs or adopts changes. The second section is devoted to Bayesian test sequences, the third to remote contiguity, the fourth to posterior concentration of frequentists and the fifth to frequentist model selection with posteriors. In the sixth section one formulates a result on the relation between a credible set and its associated confidence set. Some conclusions are presented in the seventh section. Additional definitions and conventions are reported to be described in the Supplementary Material, \url{doi:10.1214/20-AOS1952SUPP;.pdf}.