In this paper I shall sketch a novel conception of the syntax and semantics of natural languages. According to this view, the semantics of a natural language cannot be presented solely in terms of the technical notions of truth and satisfaction, and yet formalizations of certain parts of natural languages within that framework can lead to theoretical insights. Thus my view will not please either the relentless formalizer, nor the unregenerate philosopher of ordinary language. The view also upholds the pretheoretic and empirical significance of syntactic evidence and syntactic mode of argumentation. Finally, my construal of natural language as a biological phenomenon will lead to a defense of the partial independence of the grammars of natural languages from semantic considerations, and it will attribute to such grammars mental reality of a sort, under idealizations. This part of the view will please neither the practitioners of Montague-grammars, nor the most influential schools of thought concerning the issue of "psychological reality' for rules of language; i.e., it will please neither those who regard such questions beneath their (mathematical) dignity, nor those who aim at a "real time process' model of linguistic performance.