This dissertation explores language comprehension in native speakers (NSs) and second language (L2) speakers of French. Recent findings suggest that whereas NSs process complex sentences using both syntax and semantics, late learners of a L2 process shallowly, relying on lexical, semantic, and pragmatic cues to interpretation. Studies supporting this Shallow Structure Hypothesis (Clahsen & Felser, 2006b) rely on limited methodologies, however, and are challenged by reports demonstrating proficiency and cognitive effects on processing. In addition, recent research suggests that native language comprehension is not always complete or accurate (Ferreira & Patson, 2007) and is subject to variability (Dabrowska, 2012). This dissertation brings new evidence to bear on NS-L2 differences through the structural priming paradigm and investigates several factors thought to contribute to NS-L2 differences, including the exploratory effect of relative language dominance. Evidence from a self-paced reading task examining off-line and on-line priming of relative clause attachment height suggests that prior exposure to structural information through comprehension influences NSs' subsequent comprehension at the post-interpretive stage. Results argue for priming at the level of abstract hierarchical syntax and an implicit learning account of persistence. This study is one of few to demonstrate priming of ambiguously attached modifiers during comprehension and the first to do so within a L2. Unlike for NSs, the nature of the L2 priming effect is linked to discourse information. Age of acquisition was found to be a more important factor in L2 priming than language dominance. The results also argue that both native and L2 speakers are susceptible to shallow processing, though they use slightly different strategies. While NSs in the current study were more willing to accommodate competing syntactic and semantic analyses, ultimately accepting a less-than-complete analysis, the L2 parsing mechanism preferred to settle on one interpretation. The evidence here lends partial support to the hypothesis that L2 processing relies more on semantic/pragmatic information than NS processing but crucially does not exclude the possibility of L2 syntactic processing and highlights NS-L2 similarities in terms of the contexts that trigger shallow processing. / text

This dissertation investigates two facts about language processing. The Good Enough Approach claims that language users do not form a fully detailed representation of the input unless the task at hand requires it. On the other hand it has been shown that language users display internal preferences when they are faced with ambiguous input, as to what direction disambiguation should take. It has been proposed that these preferences are based on previous experience with similar inputs. This thesis investigates these two issues using tools from the fields of decision making and reinforcement learning. Specifically feedback and payoffs associated with sentence interpretations are manipulated to explore reading behavior, understood as a process of information seeking, and disambiguation choices. In four eye-tracking-reading experiments, the experimental stimuli are sentences containing a relative clause attachment ambiguity. Experiment 1 investigates whether the combination of the degree of ambiguity of a sentence and the possible payoffs, affect people’s reading times for the potentially ambiguous parts of a sentence, as well as their disambiguation choices. Experiment 2 investigates the role of feedback in such processes, a combination related to expected utility maximization. Experiment 3 studies how participants learn from feedback under risky or non-risky conditions. The last experiment investigates whether participants adjust their responses to evidence provided by feedback even overriding their internal initial bias towards a default response. / text

Cette thèse porte sur les différents facteurs impliqués dans le traitement des relatives, notamment le traitement des relatives sujet et objet ainsi que celui de l’attachement des relatives sujet. En prenant appui sur les approches multifactorielles, (Trueswell et Tanenhaus, 1994 ; Spivey et Tanenhaus, 1998), j’insiste sur l’insuffisance de l’approche monofactorielle souvent adoptée dans l’analyse du traitement des relatives, et je mets en avant le rôle des propriétés fines des langues. Je propose alors dans cette thèse de dépasser les notions de langue avec un avantage pour la relative sujet ou objet, ou avec une préférence avec attachement haut ou bas.Je présente d’abord les ressemblances et différences des propriétés des relatives dans les quatre langues étudiées : l’anglais, le cantonais, le français et le mandarin. Cette description permet de mieux appréhender leur rôle dans le traitement, notamment pour comprendre les différences constatées dans les expériences entre les langues.Je me concentre ensuite dans les deux Parties suivantes sur l’interaction entre des facteurs principalement syntaxiques, cognitifs et grammaticaux en jeu dans le traitement des relatives, avec des expériences de jugements d’acceptabilité et de mouvements oculaires avec le paradigme monde visuel (relatives sujet et objet dans les quatre langues étudiées, et attachement en anglais et en français). Je poursuis par une réflexion sur l’approche multifactorielle en tenant compte des facteurs liés aux domaines sémantique et pragmatique. Pour cela, j’observe le traitement des relatives sujet et objet en français et en anglais dans une étude de corpus, des expériences de jugements d’acceptabilité, de lecture par autoprésentation segmentée, et de mouvements oculaires en lecture. Enfin, pour élargir les facteurs en jeu dans le traitement des relatives au-delà du domaine linguistique, je montre l’influence d’un domaine non linguistique (amorçage mathématique) sur le domaine linguistique (attachement) en français avec des expériences de choix restreints, et de mouvements oculaires avec le paradigme monde visuel. / The present dissertation focuses on the factors implied in relative clause processing, mainly subject/object relative clauses and relative clause attachment. Based on previous multifactorial approaches (Trueswell et Tanenhaus, 1994; Spivey et Tanenhaus, 1998), I show that the usual monofactorial way of thinking about processing is inadequate and that fine-grained language properties should also be taken into account. I propose to go beyond the notion of subject/object languages or high/low attachment languages. I present the relative clause properties in four languages (English, Mandarin, Cantonese and French). This description is necessary to grasp their influence in processing, especially the differences observed across languages in experiments. Based on the linguistic description, I look at the interaction between cognitive, syntactic and grammatical factors in relative clause processing. I present acceptability judgment tasks and visual world eye-tracking experiments in order to analyse subject/object relatives in the four languages, and relative clause attachment in English and French. In line with the multifactorial approach, I analyse the influence of semantics and pragmatics in relative clause processing, especially in French and English. For that, I show a corpus study, acceptability judgment and self-paced reading tasks and eye-tracking while reading experiments. Finally, to broaden the perspective beyond the linguistic domain, I show the influence of other non-linguistic factors on relative clause processing, by presenting visual world eye-tracking and forced choice experiments about the influence of mathematical priming on relative clause attachment in French.