UC Berkeley

This dissertation analyzes the production and perception of lexical stress in trilinguals’ first, second, and third languages (L1, L2, and L3) to evaluate how the cue-weighting transfer hypothesis applies to L3 acquisition. According to this hypothesis, acoustic cues to stress have different weights across languages, and results from both production and perception studies of bilingual speakers indicate that bilinguals transfer cue-weightings from the L1 to the L2 (Chrabaszcz et al., 2014; Ingvalson et al., 2012; Iverson et al., 2003; Tremblay et al., 2021; Zhang & Francis, 2010). Acoustic correlates of lexical stress were analyzed in English, Spanish, and Catalan, as produced and perceived by L1 English-L2 Spanish-L3 Catalan speakers. Productions of stress minimal pairs in the three languages were collected in three separate experimental sessions via a sentence elicitation task, and perceptions of stress were gathered in three separate experimental sessions via a word identification task using nonword stress minimal pairs.

This study focuses on acoustic cue-weightings to word-level prominence in English, Spanish, and Catalan, which all have lexical stress as indicated by the existence of stress minimal pairs and which are reported to belong to different rhythm classes (Prieto et al., 2012). Little, if any, research has been done to investigate how the cue-weighting transfer hypothesis may extend to the L3, which in turn allows for L2 transfer to be better understood (Flynn et al., 2004). Whereas vowel quality is the main correlate to lexical stress in English, duration is the most prominent cue in Spanish, and in Catalan, cue-weighting is vowel-dependent, with duration and vowel quality being prominent cues. The majority of models of multilingual phonetic and phonological acquisition that posit some degree of interaction between acquired language systems have been theorized to extend to L3 acquisition in a similar manner (e.g., Amengual, 2021; Chan & Chang, 2019; de Leeuw & Chang, in press; Escudero et al., 2013; Wrembel et al., 2019). Accordingly, I hypothesize that cue-weightings from previously learned language(s) will transfer into the L3 in similar ways that cue-weightings from the L1 transfer into the L2. Specifically, I hypothesize that cue-weightings in each language will be mediated by relative language dominance. I additionally predict that transfer will be bidirectional, where cue-weightings in the L2 and the L3 can influence cue-weightings in the L1.

The results of the production and perception tasks indicate that relative language dominance does affect cue-weighting to a different extent in each language in the trilinguals’ repertoires. There was additionally evidence for regressive transfer of cue-weighting in both production and perception, indicating that all languages in a trilingual’s repertoire are susceptible tocrosslinguistic influence. Principal component analysis is shown to be a viable way to extend the Bilingual Language Profile (Birdsong et al., 2012) to obtain relative dominance scores for trilinguals. Lastly, through a comparison of theoretical frameworks of L3 phonetics and phonology, the Attrition & Drift in Access, Production, and Perception Theory (ADAPPT; de Leeuw & Chang, in press) was determined to most closely align with the findings of this study.