Purpose: Monolingual autistic children show distinct patterns of linguistic mazes (disfluencies), such as fewer filled pauses (e.g., “uh,” “um”) and utterance-initial connectives (e.g., “and”), than non-autistic peers. Maze types are multifunctional, but some (e.g., filled pauses) are used primarily for pragmatic, listener-oriented purposes such as signaling an upcoming delay, while others (e.g., repetitions) reflect speaker-internal processes such as lexical retrieval. This study examined the separate and combined effects of autism and bilingualism on children's maze production, exploring whether different types are primarily listener- or speaker-oriented, thereby contributing to the ongoing debate about their function in spontaneous speech. Method: Four groups of children aged 5–9 years participated: bilingual Russian–Hebrew autistic ( n = 20), bilingual non-autistic ( n = 27), monolingual Hebrew autistic ( n = 17), and monolingual non-autistic ( n = 22). Narratives, elicited using the LITMUS Multilingual Assessment Instrument for Narratives, were analyzed for various maze types. Results: The results indicated that while autism and bilingualism alone did not predict maze rate, their joint influence systematically interacted with specific maze types. Monolingual autistic children showed higher rates of phonological fragments, inter-utterance silent pauses, and prolongations, but lower rates of the utterance-initial connectives ve “and” and filled pauses than non-autistic peers. Bilinguals in both groups produced more intra-utterance silent pauses. Among autistic children, bilinguals used connectives more often but produced fewer prolongations and inter-utterance silent pauses than monolinguals. Conclusion: Bilingualism may enhance communicative adaptability in autistic children by strengthening narrative cohesion through greater use of connectives and fewer inter-utterance silent pauses.
Beradze et al. (Thu,) studied this question.