Lightning talk: Anti-Agreement Effect in the speech of Arabic-dominant heritage speakers of Tashelhiyt Berber
Investigation of agreement overmarking in novel dyad of morphologically complex languages spoken by unbalanced bilinguals
Heritage language (HL) speakers are recessive bilinguals dominant in the majority language (ML) of their society. In the majority of ML-HL dyads explored so far in the literature, it has been observed that HLs undergo morphosyntactic simplification. In some cases this is attributable to the ML being morphologically simpler than the HL; in fact, research has virtually always been done with English as the ML (Scontras & Putnam 2020). We investigate a novel language dyad: heritage Tashelhiyt Berber and dominant Moroccan Arabic. In this dyad, both languages show comparable complexity which make it harder to predict the effects or direction of simplification. We compare these languages with respect to the anti-agreement effect (AAE), a phenomenon productive in Berber but absent in Arabic. Under AAE, subject-verb agreement is suppressed whenever the subject undergoes A-bar movement, for instance, in subject relative clauses (Ouhalla 1993), which are the focus of our study. HL speakers are known to overmark structures that are not immediately analyzable (Polinsky, 2008:173-181). Our hypothesis was that the opaque morphology instantiated by the AAE forms would be overtly overmarked. Production data include narratives based on a silent cartoon and relative clauses elicited in a picture-disambiguation task (25 HLs, 11 native Tashelhiyt controls). In terms of AAE, 7 HL speakers overmarked agreement, supporting the initial hypothesis. Eighteen HL speakers were comparable to baseline Berber controls both on the fluency and AAE fronts. We suggest that in equally complex dyads, frequency and length of exposure might influence dominance effects.
This is a short talk about the findings of my 2019 research project where I involved a group of participants from the University Hassan II of Casablanca, Morocco. The ages of the HS group ranged from 19 to 55 years, whereas the control group ranged from 20 to 42 years of age. What the control group members had in common was that they had moved to Casablanca from the Berber-dominant southern region of Morocco when they were around 18 years of age. Thanks to a metadata questionnaire that the participants had filled prior to taking part in the data collection, it was possible to have an overview of the length and regularity of exposure of the participants to Berber. The language background questionnaire was adapted from a similar questionnaire found in Albirini et al, 2013, a study that provided insight into the design of the present one. The subsequent metadata made it possible to observe a correlation between the amount of exposure to the heritage language and the native-like performance of 72% of the heritage speakers in both the spontaneous narrative and in the AAE marking task.