I am generally positive about Muysken's (M) approach, and the potential use of unifying various seemingly related phenomena is obvious. The approach could also serve as a tool in determining to what extent these phenomena actually are sides of the same coin (I am somewhat less convinced of this than most contact linguists).