S. Muck, LARGE DEVIATIONS WRT QUASI-EVERY STARTING POINT FOR SYMMETRICAL RIGHTPROCESSES ON GENERAL STATE-SPACES, Probability theory and related fields, 99(4), 1994, pp. 527-548
In the work of Donsker and Varadhan, Fukushima and Takeda and that of
Deuschel and Stroock it has been shown, that the lower bound for the l
arge deviations of the empirical distribution of an ergodic symmetric
Markov process is given in terms of its Dirichlet form. We give a shor
t proof generalizing this principle to general state spaces that inclu
de, in particular, infinite dimensional and non-metrizable examples. O
ur result holds w.r.t. quasi-every starting point of the Markov proces
s. Moreover we show the corresponding weak upper bound w.r.t. quasi-ev
ery starting point.