Return to Revisions

There is this paper by Zuk, which gives a sufficient criterion for property (T) in terms of some spectral properties of a graph depending on a group $G$ with a generating set $S$. This graph is not the Cayley graph. But maybe it is still in the spirit of the question.