This dissertation presents results from a study on the expression of emotions in a second language in order to address two overarching research questions: 1) What does the acquisition of L2 emotion lexicon and discourse features tell us about the pragmatic and communicative competence of late learners and the internalization of L2-specific concepts, and 2) Knowing that expressing emotions in L2 is one of the most challenging tasks for L2 learners (Dewaele, 2008), what can late L2 learners do at end-state, with regards to ultimate attainment and the possibility of nativelikeness? Narratives of positive and negative emotional experiences were elicited from late L2 learners of English and French at end-state, both in their L1 and L2. First, the acquisition of L2 emotion words was analyzed through the productivity and lexical richness of the emotion vocabulary of the bilinguals. Analysis of L2 emotion concepts was also conducted through the distribution of emotion lemmas across morphosyntactic categories. Lexical choice of emotion words was also investigated. Results showed that although L2 English and L2 French bilinguals' narratives were shorter than the monolinguals' and the proportion of emotion word tokens were fewer than that of monolinguals', bilinguals showed greater lexical richness than the monolinguals. In terms of morphosyntactic categories, bilinguals behaved in a nativelike pattern such that L2 English bilinguals favored adjectives and L2 French bilinguals preferred nouns/verbs. This pattern was held constant across the first languages of the bilinguals. With respect to lexical choice, bilinguals used the same emotion lemmas used the most by monolinguals. On occasion, non-nativelike patterns also emerged, suggesting both L1 transfer on L2 (L2 English bilinguals favoring nouns/verbs) and L2 transfer on L1 (L1 English bilinguals favoring nouns/verbs). However, these rare instances could be explained by individual and typological variability. The findings suggest that late L2 learners can achieve nativelike levels of attainment in L2, providing evidence against the existence of a critical period for the acquisition of L2 pragmatics and culture-specific L2 lexicon. In a separate analysis, the L2 discourse of emotion was investigated under a corpus linguistic framework, in order to shed some light into the ways late L2 learners of English and French talk about emotions in narratives of personal stories. The use of stance lemmas and tokens, and the distribution of these stance markers across categories of certainty and doubt evidentials, emphatics, hedges, and modals, as well as lexical choice of stance were analyzed. This was followed by an analysis of discourse features, such as figurative language, reported speech, epithets, depersonalization, and amount of detail. Results showed that although bilinguals produced significantly less stance lemmas and tokens than monolinguals, in terms of the distribution of stance categories, the French group (L2 French and L1 French bilinguals) behaved in a nativelike pattern, favoring emphatics, certainty evidentials, doubt evidentials, hedges, and modals. The English group's results, on the other hand, were somewhat inconsistent, in that neither L2 English bilinguals, nor L1 English bilinguals followed the distribution pattern of English monolinguals. In terms of nativelike performance, we conclude that the L2 French bilinguals did perform nativelike with regards to stance marking, and that L2 English bilinguals also performed nativelike, but only for certain categories of stance. Also, L2 English transfer on L1 French was evidenced for L1 French bilinguals. Analysis of discourse features revealed between 1 up to 10 bilinguals (L2 English or French) out of 31 who used those features which were only evidenced in native speech in previous research. The findings here, once again suggest that late L2 learners can acquire aspects of L2 discourse to a nativelike degree. / text

Föreliggande studie undersökte relationen mellan startålder och slutlig behärskning i fonologi och icke-fonologi (ett samlingsbegrepp för morfosyntaktisk och idiomatisk brytning). Deltagarna var 200 spansktalande andraspråkstalare av svenska, 20 modersmålstalare av svenska (kontrollgrupp) och 28 lyssnare med svenska som modersmål. Slutlig förmåga i fonologi och icke-fonologi mättes genom två brytningsbedömningar av lyssnargruppen, en där talarnas faktiska röster användes för att mäta fonologisk förmåga och en där en syntetisk röst användes för att mäta icke-fonologisk förmåga. Studiens resultat visade att det fanns en stark negativ korrelation mellan startålder och slutlig behärskning för både fonologisk och icke-fonologisk förmåga. Korrelationen var starkare för fonologi än för icke-fonologi. Starålder visades även vara den starkaste prediktorn för inföddlik slutlig behärskning, även efter kontroller av andra prediktorer som ålder, vistelsetid i Sverige och grad av andraspråksanvändning (% i förhållande till L1). Sannolikheten att tillägna sig en inföddlik slutlig behärskning var högre ju tidigare startålder talarna hade och endast talare med startålder mellan 1 och 16 uppvisade inföddlikhet i både fonologi och icke-fonologi. Fler talare uppvisade inföddlikhet i icke-fonologi än i fonologi, vilket kan bero på den unika motoriska begränsning tillägnande av fonologi har.