Does anyone know what the current expert consensus is concerning the status of the question as to whether the mapping class group of a surface has property (T)?

There is a short (21 page) paper by J. Anderson which purports to use quantum representations to prove that it does not.  See <a href="http://arxiv.org/abs/0706.2184">here</a>.  It was released in 2007, but it does not seem to have yet been accepted by a journal.  I have asked several experts (on the mapping class group and on property (T)), and none of them seem to understand the details of this paper.  One or two of them alluded to issues they had heard might exist, but they were pretty vague as to what these issues might be.