Intonation Patterns in Argentinean- and Venezuelan-Canadian Heritage Speakers of Spanish

Abstract

An encompassing study projected by Rao (2016) identified various factors such as transfer from English, parental input, and pragmatic context as contributing to the source of variation in the production of nuclear (i.e., final) intonation in the Spanish of heritage speakers (HSs) of Mexican Spanish who lived in the United States. The current study expands on Rao (2016) by quantifying the degree of similarity to (a) English and (b) parental intonation patterns in nuclear yes-no question intonation in Porteño (i.e., Buenos Aires, Argentina) Spanish HSs and Caraqueño (i.e., Caracas, Venezuela) Spanish HSs living in Ontario, Canada. Whereas the intonation contour in absolute interrogative questions in Canadian English is primarily ascending (Séguinot, 1976), it is often descending in Porteño (Gabriel et al., 2010) and Caraqueño Spanish (Sosa, 1999). Therefore, these groups of bilinguals lend themselves well to a study examining whether heritage intonational phonology is more influenced by parental input or whether it exhibits drift towards English. We tested eight adult English-dominant HSs – four of Argentinean heritage and four of Venezuelan heritage  – and their parents. All participants filled out the Bilingual Language Profile (BLP; Birdsong, Gertken & Amengual, 2012). The participants were also asked to complete a semi-spontaneous task, where they were presented with fictional scenarios (identical in both languages) and asked to respond (e.g., Fictional Scenario: Entrás a un negocio donde nunca estuviste antes y preguntás si tienen mandarinas (You walk into a store that you’ve never been to before and ask if they have any mandarins) Anticipated utterance: ¿Tienen mandarinas? (Do you have any mandarins?)). Pragmatic context was controlled and taken into account with the elicitation of 6 different kinds of yes-no questions: surprise questions, requests, information-seeking questions, confirmation, offers and orders. The tasks were completed by both HSs and their parents; they first completed the task in Spanish, followed by English. English-speaking controls completed the same tasks only in English. This study only focuses on the Spanish stimuli of the HSs and their parents as well as the English of the control group. Results were caluclated using the total percent of frequencies of pitch accents across all participants. To analyze the impact of group on these pitch accents, a 3 x 4 Chi-squared was run across group and pitch accent, reavealing a significant impact of group on the frequency of pitch accent (p < 0.001, φ = 0.39). While L+H* showed a numerical declining trend across the three groups, this effect did not achieve significance (max absolute SR = 1.1). L* showed an increase across groups that achieved marginal significance (max absolute SR = -1.9). L+¡H* showed a significant decrease across the three groups (max absolute SR = -2.5). H+L* showed a significant difference across groups, however, this change was driven by a greater frequency in the monolingual English group (SR = 2.8), with no parametric trend across the three groups. An analysis was then run to analyze the impact of group on these boundary tones within the Argentinean group (p < 0.001, φ = 0.54)  and then the Venezuelan group (p < 0.001, φ = 0.65), a 3 x 3 Chi-squared was run across generation and boundary, revealing a significant impact of generation on the frequency of boundary tone. Parental influence on pitch accent, nuclear pitch configuration, and boundary tones was also investigated. The proportion for each response characteristic for each child was paired with their parent’s proportion of responses (e.g. Child 1’s proportion of responses with a L* pitch accent was paired with their parent’s proportion of responses with an L* pitch accent), and bivariate Pearson correlations were performed for each type of characteristic. Parents’ pitch accents were significantly related to their child’s proportions of pitch accents (r(34) = 0.46, p = 0.005, CI = [0.15 0.68]). Neither boundary tone (r(16) = 0.18, p = 0.49, CI = [-0.32 0.59]) nor nuclear pitch configuration (r(16) = 0.21, p = 0.40, CI = [-0.28 0.62]) showed any significant relationship between parent and child. In sum, the results reveal that, while we do note parental influence on pitch accents, we do not observe this effect much on heritage speakers' boundary tones and nuclear pitch configurations, where we ascertain shifts toward English in heritage Spanish question intonation.