Statistics for Lexical Affinities and Language Applications