Modeling the auxiliary phrase asymmetry in code-switched Spanish-English

Chara Tsoukala, Stefan L. Frank, Antal van den Bosch, Jorge Valdés Kroff, & Mirjam Broersma


Spanish-English bilinguals are moderately likely to code-switch between the Spanish progressive auxiliary estar ("to be") and the participle (e.g., "Ella está voting"; "She's voting"). However, a switch in the perfect structure between haber ("to have") and the participle ("Ella ha voted"; "She has voted") is rarely produced. This phenomenon is known as the "auxiliary phrase asymmetry". One hypothesis as to why this occurs is that estar has more semantic weight as it also functions as an independent verb, whereas haber is almost exclusively used as an auxiliary. To test this hypothesis, we employed a connectionist model that produces spontaneous code-switches. We performed three simulation experiments, showing that i) the asymmetry emerges in the model; ii) the asymmetry disappears when using haber also as a main verb; iii) and this is not due to the increased frequency. Therefore, the lack of semantic weight of haber may indeed cause the asymmetry.

Keywords: auxiliary phrase asymmetry, code-switching, computational cognitive modeling, sentence production, Bilingual Dual-path model