C. criticizes Grice's 'conversationalist hypothesis' about logical particles of natural language, according to which those particles (e.g. not, and, if... then) do not diverge in meaning or function from the corresponding formal-logical symbols and the possible appearance of diverging truth-functionality is due to the presumptions with which natural language utterances are understood. C.'s alternative: the 'semantical hypothesis' according to which occurrences of the particles sometimes do and sometimes do not differ in meaning from the formal-logical counterparts, but according to which both kinds of occurrences should be explained within a semantical theory of natural language, without recourse to a theory of conversational presumptions.