Neither Wilkins nor B ö hme, representative though they were of different manifestations of the search for the universal in language, contributed directly to what was to become the mainstream version of universal grammar. The very origin of the Port- Royal Grammaire generate et raisonn é e (1660) mirrors the conflicting elements at work: the meeting of particular grammar and philosophy. In the course of writing textbooks on Latin, Greek, Spanish and Italian, Claude Lancelot observed the existence of features common to these and (he surmised) all other languages; a philosopher colleague, the embattled Antoine Arnauld, brought inductive confirmation of the cognitive basis of language. Mental operations were made the basis of grammatical distinctions: the three primary operations—forming a concept such as ‘round’, making a judgement such as ‘the earth is round’, and reasoning—provided a framework for distinguishing the various parts of speech and for the study of syntax. Because these operations and their linguistic consequences are universal, they can be exemplified through any language, and French and Latin provide most of the examples. In this way the celebrated analysis of the proposition ‘Dieu invisible a cr éé le monde visible’ simply shows how three distinct mental propositions—that God is invisible, that He created the world, and that the world is visible—are included in this one verbal proposition. A distinction between mental language and the verbal language which is the province of grammarians had been part of the theological and philosophical tradition for centuries. That one might seek to derive grammatically analysable sentences from mental propositions was not an enterprise that would have struck someone trained in this tradition as worthwhile. Instead, the analysis of the justification for the parts of speech was of more immediate importance. Having defined the verb as a word whose principal use is to signify affirmation,