Nous voudrions examiner, dans son ensemble, le processus de grammémisation (Pottier 1974:68) ou passage du lexème au morphème grammatical en québécois. Il y a là un vide à combler, car le sujet n’a guère fait l’objet d’une étude large. Il convient cependant de signaler que, dans un champ lexical particulier, Thibault et Vincent (1981) ont déjà constaté le passage des lexèmes ayant un dénoté de caractère religieux et sacré au statut de sacres (jurons) puis de quantificateurs ou de modalisateurs (modalisateurs assertifs ou qualificatifs).

This paper is concerned with the formulation of a constraint on the set of NPs which can serve as “antecedent” for PRO and floated quantifiers. I will begin by showing that a principle formulated in terms of c-command encounters problems distinguishing between two classes of NPs occurring within PPs. I will then propose that a more comprehensive and successful constraint can be formulated in terms of thematic dependency, a relation holding between an NP and the lexical category which determines its thematic role.

In this paper, the consonantal coefficient is defined as the ratio of consonants to vowels occurring in continuous spoken texts in any one language. The vowel coefficient is the reciprocal, i.e. the ratio of vowels to consonants in the same type of texts. For example, if a text consisted of all vowels and no consonants, the value of the consonantal coefficient would be zero. The vocalic coefficient would theoretically be infinity, but no comparison could actually be made. Similarly, in a text that was all consonants and no vowels, the values would be reversed. In a text in which the number of vowels and consonants was equal, the value of both the consonantal coefficient and the vocalic coefficient would be 1, or unity. The use of the coefficient gives us a method for the classification of languages that can be of use in typological studies, just as the coefficient of stem to inflection or vice-versa can give us another index for typological classification.