This thesis is about vowel-vowel sequences across word boundaries in Québécois French (QF). QF has a number of phonological processes that seem motivated by hiatus avoidance, yet hiatus is tolerated in many instances as well. This research is about why hiatus is tolerated in QF, why it is avoided, and what grammatical models can account for the relevant processes. My work is intended as a contribution to the study of how best to account for variation and opacity within current grammatical models, and as a contribution to the study of the QF vowel system. The data are drawn from a corpus constructed from recordings of web extras for a Québécois reality television series. The data primarily come from a single speaker to ensure that any variation in the data truly represents intra-grammar variation, but data from other speakers are used as safeguard. Through the use of quantitative data as a means of investigating problems in theoretical phonology, the thesis is also meant to contribute to methodological discussions and discussions about the relationship between phonetics and phonology. I propose that the patterns of hiatus and hiatus resolution in QF are best modeled through three sets of constraints organized in a serial manner. This proposal is based on the claims that the data show evidence for an anti-hiatus constraint, for feature-based analysis, for stochastic modeling, and for multiple levels. The proposed model combines insights from Stochastic Optimality Theory (Boersma & Hayes 2001), multi-level Optimality Theory (Kiparsky 2000, 2010; Rubach 2000), and the Contrastivist Hypothesis (Dresher & Rice 2002, Dresher 2009, Hall 2007). Within the model, the first constraint set targets the smallest prosodic constituents and produces categorical outputs, the second applies to intermediate-sized constituents and can model optionality, and the third handles the largest prosodic constituents and produces complex patterns of variability.

This thesis is about vowel-vowel sequences across word boundaries in Québécois French (QF). QF has a number of phonological processes that seem motivated by hiatus avoidance, yet hiatus is tolerated in many instances as well. This research is about why hiatus is tolerated in QF, why it is avoided, and what grammatical models can account for the relevant processes. My work is intended as a contribution to the study of how best to account for variation and opacity within current grammatical models, and as a contribution to the study of the QF vowel system. The data are drawn from a corpus constructed from recordings of web extras for a Québécois reality television series. The data primarily come from a single speaker to ensure that any variation in the data truly represents intra-grammar variation, but data from other speakers are used as safeguard. Through the use of quantitative data as a means of investigating problems in theoretical phonology, the thesis is also meant to contribute to methodological discussions and discussions about the relationship between phonetics and phonology. I propose that the patterns of hiatus and hiatus resolution in QF are best modeled through three sets of constraints organized in a serial manner. This proposal is based on the claims that the data show evidence for an anti-hiatus constraint, for feature-based analysis, for stochastic modeling, and for multiple levels. The proposed model combines insights from Stochastic Optimality Theory (Boersma & Hayes 2001), multi-level Optimality Theory (Kiparsky 2000, 2010; Rubach 2000), and the Contrastivist Hypothesis (Dresher & Rice 2002, Dresher 2009, Hall 2007). Within the model, the first constraint set targets the smallest prosodic constituents and produces categorical outputs, the second applies to intermediate-sized constituents and can model optionality, and the third handles the largest prosodic constituents and produces complex patterns of variability.

Denna kandidatuppsats utgår från en egen översättning från franska till svenska av delar ur pjäsen Hôtel Pacifiqueav den franskkanadensiska författaren Fanny Britt.Syftet med examensarbetet är, förutom att skapa en spelbar text som fungerar i sin målspråkskontext, att undersöka det specifika i att översätta dramatik. I syftet ingår också att belysa och skapa förståelse för Québecfranskan och problem relaterade till att översätta språklig varietet. Det talade ordet står genomgående i fokus: replikerna i dramatexten och de karaktäristiska muntliga dragen för Québecfranskan. Newmarks kommunikativa översättningsmetod används som övergripande översättningsprincip för arbetet och visar sig vara kompatibel med texttypen trots vissa undantag. Resultatet blev en modern svensk pjäs med talspråkligt och idiomatiskt språk i replikerna. För att avgöra om texten är spelbar krävs dock att dramat uppförs på scen, vilket är ett framtida mål. / This essay is based on my own translation from French to Swedish of parts of the play Hôtel Pacifiqueby the Québécois author Fanny Britt. The purpose of the essay, apart from creating a text performable in its target language context, is to discuss and analyze the specialized nature of translating drama. The essay also aims to provide new perspectives on Québécois French and problems related to translating language varieties. The spoken word is the focal point: i.e. the lines of the drama and the characteristic oral features of Québécois French. Newmark's communicative translation method is used as the main translation principle. The method proves to be compatible with the text type despite certain exceptions. The result was a modern Swedish play with colloquial and idiomatic language. However, in order to determine if the text is performable, the drama must be presented on stage, which is a future goal.