M. shows (i) that it is very difficult to define the notion 'subject' in a syntactic theory, and (ii) that the pragmatic distinction between theme and rheme is a good candidate for a universal category, signalling the presuppositions connected with an utterance.