This note discusses the use of Petri net languages in supervisory control theory. First it is shown that the trimming of an unbounded Petri net is not always possible and a new class of Petri net languages, that may be generated by nonblocking nets, is defined. Secondly, necessary and sufficient conditions for the existence of a Petri net supervisor, under the hypothesis that the system’s behavior and the legal behavior are both Petri net languages, are derived. Finally, by means of an example, it is shown that Petri net languages am not closed under the supremal controllable sublanguage operator.