The grammatical status of Romance pronominal clitics has long been the object of intense debate. Are they syntactically-independent arguments or are they affixal agreement markers? This dissertation addresses this question with respect to Québec Colloquial French (QCF). It treats the morphophonological and morphosyntactic dimensions as two independent dimensions, thus allowing either for affixes to have argument status and prohibiting them from co-occurring with an overt, lexical argument, or for non-affixal elements to behave like agreement markers and not count as syntactic arguments. The analysis reveals that all the clitics of QCF are affixes at the morphological level, since they demonstrate numerous patterns which are too idiosyncratic to be handled by syntactic rules. Only subject clitics, however, function as agreement markers, since they occur in all the environments where we would expect agreement markers; object markers, conversely, are excluded from those environments. I shall present a morphological analysis following the approach taken by Roberge and his colleagues and work within the Minimalist framework of Chomsky (1993). I assume that fully-inflected lexical items are inserted in the syntax, and I adopt Cummins & Roberge’s (1994a,b) suggestion that an additional interface, the Lexicon-Syntax Interface, handles all inflectional morphology. Chapter 1 introduces the main problem and summarizes a number of recent studies on Romance and French clitics. Chapter 2 presents the facts in QCF and applies morphophonological and morphosyntactic criteria in order to determine the grammatical status of these argument markers. Chapter 3 develops a morphological approach that allows fully inflected verbs to be inserted in the syntax. Finally, chapter 4 discusses one element which has traditionally been excluded from the paradigm of subject markers: ça. The chapter is divided into two parts; the first one discusses the semantic aspects concerning the generic use of ça, while the second one develops an approach to grammatical agreement which allows agreement targets to introduce their own features, as well as certain types of feature conflicts.