This thesis is primarily an investigation into Pronominalization in Efik. Pronominalization is treated from a basically Chomskian point of view as a general term for a number of related processes each of which is explicitly formulated as a rule. These are reflexivization, simple pronominalization, possessive pronominalization and relativization by which reflexive, anaphoric personal, possessive and relative pronouns respectively are derived. As rules they all operate on Noun Phrases (NPs) under certain conditions. One such condition is coreference. In general NPs on which a rule of pronominalization has operated may (and in fact in some cases must) be deleted under certain conditions one of which is also coreference. In this way, Pronominalization and NP Deletion are related.

The Interface Hypothesis (Sorace and Filiaci, 2006) conjectures that adult second language learners (L2 learners) who have reached near-native levels of proficiency in their second language exhibit difficulties at the interface between syntax and other cognitive domains, most notably at the syntax-discourse interface. However, research in this area was limited, in that the data were offline, and thus unable to provide evidence for the nature of the deficit shown by L2 learners. This thesis presents online data which address the question of the underlying nature of the difficulties observed in L2 learners at the syntaxdiscourse interface. This thesis has extended work on the syntax-discourse interface in L2 learners by investigating the acquisition of two phenomena at the syntax-discourse interface in German: the role of word order and pronominalization with respect to information structure (Experiments 1-3), and the antecedent preferences of anaphoric demonstrative (the der, die, das series homophonous with the definite article) and personal pronouns (the er, sie, es series) (Experiments 4- 8). Crucially, this work has used an on-line methodology, the visual-world paradigm, which allows an insight into the incremental interpretation of interface phenomena in real-time processing. The data from these experiments show that L2 learners have difficulty integrating different sources of information in real-time comprehension efficiently, supporting the Interface Hypothesis. However, the nature of the processing difficulties which L2 learners demonstrate in on-line processing was not determined by these studies, resulting in the question: are L2 learners’ difficulties a result of a limitation of processing resources, or the inability to deploy those resources effectively? A novel dualtask experiment (Experiment 9), in which native speakers of German were placed under processing load simulated the results previously obtained for L2 learners. It is concluded that syntactic dependencies were constrained by resource limitation, whereas discourse based dependencies were constrained by processing resource allocation.

La thèse porte sur le statut sémantico-pragmatique des propositions relatives appositives (PRA), étudié à travers l’examen du fonctionnement discursif des phrases complexes de la forme ‘Matrice, qui PRA’ du français. Dans beaucoup de langues, les PRA n’interagissent pas sémantiquement avec les opérateurs présents dans leurs propositions enchâssantes et tendent à s’interpréter pragmatiquement comme porteuses d’informations non-centrales ou secondaires pour la question en discussion (QUD) du discours en cours. Plusieurs analyses (Holler 2005, Arnold 2007, Koev 2012) dissocient ces deux propriétés en dérivant la projection des PRA de leur statut d’assertions indépendantes et en expliquant leur lecture pragmatique à l’aide de principes généraux de gestion du flot discursif. En effet, lorsque les PRA apparaissent en fin de phrase, elles sont susceptibles d’interagir avec la QUD tout en ayant une portée large vis-à-vis du reste de la phrase. Certains phénomènes discursifs semblent néanmoins contredire l’idée que les PRA constituent des assertions indépendantes. D’abord, les PRA peuvent interagir avec la QUD seulement si leurs matrices véhiculent également des informations pertinentes pour le sujet en discussion. Ensuite, contrairement à ce que l’on peut observer dans une séquence de deux propositions indépendantes, dans une séquence formée d’une proposition matrice et d’une PRA, quel que soit l’ordre de leur linéarisation, la matrice s’interprète toujours comme centrale pour le discours, alors que le statut pragmatique de la PRA dépend fortement du degré d’informativité du reste de la phrase vis-à-vis de la QUD. Enfin, les résultats de deux études comportementales conduites pour la thèse montrent que, suite au traitement de phrases complexes comme ‘Matrice, qui PRA’, les référents du type ‘individu’ mentionnés par la matrice restent hautement saillants pour le discours subséquent contrairement à ceux dont il est question dans la PRA, qui jouissent d’un degré d’accessibilité assez faible. Partant de ces données, nous concluons que la lecture pragmatique centrale des PRA n’est pas une conséquence de leur statut d’assertions indépendantes mais résulte de l’intégration de leur contenu dans le domaine focal de leurs matrices. Plus généralement, en nous appuyant sur les travaux d’Ander Bois & al. (2010) et Schlenker (2013, ms), nous défendons l’idée que le manque d’interaction entre les PRA et le reste de leurs phrases d’accueil ainsi que leur prédisposition à une interprétation non-centrale pour le discours proviennent du fait, qu’à la différence de leurs matrices, dont l’énonciation s’accompagne de l’introduction d’un référent propositionnel nouveau, les PRA sont des anaphores propositionnelles, dont la portée sémantique et l’interprétation pragmatique dépendent de la position discursive de l’expression important dans l’univers du discours le référent auquel s’applique le contenu qu’elles expriment. / The thesis focuses on the semantic-pragmatic status of appositive relative clauses (ARC). We address this question by examining discourse functioning of complex sentences of the form ‘Matrix, qui ARC’ in French. Crosslinguistically, ARC fail to interact semantically with scope taking operators contained in their embedding clauses and tend to be interpreted pragmatically as carrying non-central or secondary information for the question under discussion (QUD) in the ongoing discourse. Several analysis (Holler 2005, Arnold 2007, Koev 2012) dissociate these two properties, deriving the ARC projection from their status of independent assertions and explaining their pragmatic reading with independent principles of the discourse flow management. Indeed, when an ARC follows linearly its embedding clause, it can interact with the QUD while still receiving a wide scope relatively to the rest of the host sentence. Some discursive phenomena seem nevertheless contradict the idea that ARC constitute independent assertions. First, an ARC can interact with a QUD only if its matrix clause also conveys information relevant to the subject under discussion. Second, contrary to what we observe examining sequences of two independent clauses, in sequences formed of a matrix clause and an ARC, regardless of the order of their linearization, the matrix clause is always interpreted as being at-issue for the discourse, while the pragmatic status of the ARC depends to a great extent on the degree of informativeness of the rest of the sentence relatively to the QUD. And, third, the results of two psycholinguistic experiments conducted as part of this study show that after processing a sentence such as ‘Matrix, qui ARC’, the entity-type referents realized by the matrix clause are highly salient for the subsequent discourse unlike those realized by the ARC, which have a rather low accessibility degree. Based on these data, we conclude that at-issue pragmatic reading of ARC is not a consequence of their functioning as independent assertions but results from integration of their content into the focal domain of the embedding clause. More generally, building on the works of AnderBois & al. (2010) and Schlenker (2013, ms), we defend the idea that the lack of interaction between ARC and the host sentences as well as their tendency to receive a not at-issue reading in discourse arise from the fact that unlike their matrix clauses, whose utterance has the effect of introducing a new propositional referent, ARC are propositional anaphora, the semantic and pragmatic interpretation of which depends thus on the discourse position of their antecedent expression, importing into the discourse the propositional referent the ARC’s content applies to.