A note on stationarity of spherical measures (Q2382335)

Let \(G\) be a connected semi-simple Lie group with finite center and let \(K\) be a maximal compact subgroup of \(G\). A compact homogeneous space \(B(G)\) of \(G\) is called boundary if each probability measure on \(B(G)\) is contractible; that is, for each probability measure \(\rho\) on \(B(G)\) there exists a sequence \((g_n)\) in \(G\) so that \(g_n \rho\) converges to a point measure on \(B(G)\). \textit{H. Furstenberg} [Ann. Math. (2) 77, 335--386 (1963; Zbl 0192.12704)] introduced the notion of boundary and proved the existence of the maximal boundary \(B_m(G)\), that is, every other boundary is a quotient of \(B_m(G)\). It is also shown that any maximal compact subgroup of \(G\) acts transitively on \(B_m(G)\). This shows that maximal compact subgroups act transitively on any boundary \(B(G)\) of \(G\). In this note the author shows that if \(B(G)\) is a boundary of \(G\) and \(\mu_1\) and \(\mu_2\) are probability measures on \(B(G)\) equivalent to the unique \(K\)-invariant probability measure on \(B(G)\), then there exists a probability measure \(\mu\) on a given lattice \(\Gamma\) in \(G\) such that \(\mu*\mu_1=\mu_2\) (in Theorem 1.2 \(B(G)\) should be taken as a boundary since the contraction condition stated above is needed in the proof). This result in particular implies that any smooth probability measure \(\nu\) on \(B(G)\) is stationary for some probability measure \(\mu\) on a given lattice (that is, \(\mu*\nu= \nu\)) which generalizes a result of \textit{H. Furstenberg} [Actes Congr. internat. Math. 1970, 2, 301--306 (1971; Zbl 0244.22010)].