IntroductionMazes are linguistic disfluencies such as filled pauses, repetitions, or revisions of grammatical, phonological, or lexical aspects of words that do not contribute to the meaning of a sentence. Bilingual children are believed to increase the numbers of mazes in their native or heritage language, the minority language, as they become more proficient in the second language, the societal language. Mazes may increase over time in bilingual Spanish-speaking children as they become more proficient in English, the societal language in the United States. However, current studies have not been conducted longitudinally. Higher rates of mazes in the heritage language over time may be due to changes in language proficiency and differences in processing demands in the children as they use more complex language. Moreover, children with developmental language disorder (DLD) can also present higher rates of mazes than children with typical language. Heritage speakers, therefore, are at risk of being misdiagnosed with DLD due to high rates of mazes. Currently, we do not understand what the typical rates of mazes are as heritage speakers get older and become more proficient in the societal language. The current study examined the type and frequency of Spanish mazes longitudinally in a group of 22 Spanish heritage speakers with and without DLD and determined the changes over time.MethodsA total of 11 children with typical language development (TLD) and 11 with DLD participated in this 5-year longitudinal study. Using a wordless picture book, children completed a retelling task in Spanish during the spring of each academic year (PK to 3rd grade) as part of a 5-h testing battery. Narratives were transcribed and coded for types of mazes (filled pauses, repetitions, grammatical revisions, phonological revisions, and lexical revisions).Results and conclusionThe results of the study indicate that TLD children increased their overall percentage of mazed words and utterances. The opposite pattern was observed in the DLD group, which decreased their percentage of mazed words and utterances. In contrast, both groups demonstrated a decrease in repetitions in first grade and an increase in third grade. Additionally, the TLD and DLD children decreased in the percentage of fillers in first grade and then increased in the third grade. Results suggest that maze use is quite variable in heritage speakers and does not necessarily differentiate groups. Clinicians should not rely solely on mazes to determine ability status. In fact, high use of mazes can reflect typical language development.