| Summary: | This thesis presents a theory that explains binding and control with common mechanisms. Although a unified account of the two phenomena was frequently attempted in the 1980's and early 90's, such an approach seems to have been unpopular due to certain problems since the late 90's. In this study it is shown that these problems are merely superficial and can be overcome with careful examination. In chapter 2, I start with Reinhart & Reuland's (1993) assumption that the application of Condition A of binding theory is restricted to anaphors in argument positions of syntactic predicates. I then adopt Fox's (1993) modification to Reinhart & Reuland's (1993) theory, whereby Condition A is reduced to the Chain Condition. As the discussion proceeds, further revisions are suggested. In particular, the domain for A-chain formation is defined over syntactic predicates, which is called the A-Chain Projection Domain (ACPD). With respect to anaphors that are not in argument positions of syntactic predicates, I entertain the hypothesis by the above authors that they are what are often called logophors, which are governed by a set of discourse conditions. I identify some of these conditions and explore some possibilities as to how such discourse conditions can be organized into a formalized theory. In chapter 3, it is maintained that PRO in obligatory control (OC) contexts is licensed by the Chain Condition. Discussed also are some of the objections raised by a number of researchers to the idea that the same mechanism underlies OC PRO and anaphors (in argument positions of syntactic predicates). Furthermore, it is illustrated that PRO in non-obligatory control (NOC) environments is logophoric and is licensed by discourse conditions. Lastly, it is argued that there exist control constructions that look very much like OC but are, in fact, properly classified as NOC, which I term pseudo-obligatory control (POC). In chap === Cette these presente une theorie qui explique le liage et le controle a l'aide de mecanismes communs. Bien qu'une explication unifiee des deux phenomenes ait ete frequemment tentee dans les annees 80 et 90, une telle approche semble avoir ete impopulaire a cause de certains problemes depuis la fin des annees 90. Dans cette etude il est demontre que ces problemes ne sont que superficiels et qu'ils peuvent etre surmontes par un examen attentif. Dans le Chapitre 2, je commence avec la presomption de Reinhart & Reuland (1993) selon laquelle l'application de la Condition A de la theorie du liage est restreinte aux anaphores en position d'argument de predicats syntaxiques. J'adopte ensuite la modification que Fox (1993) a apportee a la theorie de Reinhart & Reuland (1993), selon laquelle la Condition A est reduite a la Condition de Chaine. Au cours de la discussion, d'autres revisions sont suggerees. En particulier, le domaine de la formation de la chaine A est definie en termes de predicats syntaxiques, que j'appelle le Domaine de Projection de la Chaine A (DPCA). En ce qui concerne les anaphores qui ne sont pas en positions de predicats syntaxiques, je considere l'hypothese des auteurs cites plus haut selon laquelle elles sont souvent appelees des logophores, lesquelles sont gouvernees par un ensemble de conditions discursives. J'identifie certaines de ces conditions et explore certaines possibilites selon lesquelles de telles conditions discursives peuvent etre organisees en une theorie formalisee. Au chapitre 3, je maintiens que PRO en contexte de controle obligatoire (CO) est licencie par la Condition de Chaine. J'examine egalement certaines des objections a l'idee que le meme mecanisme sous-tende PRO OC et les anaphores (en positions d'arguments de predicats syntaxiques) soulevees par un nombre de chercheurs. De plus, je montre que PRO en environnements de controle non obligatoire (CNO) est logopho
|
|---|
