Language mixing in people with aphasia: A cross-linguistic analysis using the 4M model

Language mixing (LM) is common among individuals who speak two or more languages. LM in healthy multilingual individuals has been studied extensively, often by applying the Matrix Language Frame (MLF) and the 4M model, both developed by Myers-Scotton and colleagues. Our study examined LM in multilingual people with aphasia (PWA), to determine: (1) frequency of LM in the first-acquired language (L1) and later-acquired languages (L2/L3); (2) whether the 4M model accounts for LM patterns; and (3) how language-specific characteristics affect LM. Narratives produced by 18 multilingual PWA (English-Hebrew; English-Spanish, English-Hebrew-Spanish and Russian-Hebrew speakers) with and without agrammatism were analyzed using the 4M model. More LM was observed in L2 contexts compared to L1 in participants with agrammatism, yet in participants without agrammatism no differences were found between L1 and L2. The patterns of LM in target and non-target languages were similar across L1 and L2: content morphemes were more likely to be embedded into the matrix language, aligning with previous research on healthy multilingual controls. Differences were found between the types of LM across the four studied languages. Our findings indicate that multilingual PWA mix languages in a typical manner, similar to neurotypical multilingual individuals, aligning with the predictions of the 4M model. Furthermore, in multilingual PWA, as in healthy multilingual people, differences in LM across languages are related to the morphological richness of the matrix language and the congruency of the languages involved in LM. Clinically, based on our results, LM can be encouraged as a communication aid.