“The “No Sabo” effect”: Insights from AI into child Spanish heritage language and identity development
University of Michigan Undergraduate Research Journal · 04.06.2026 · 18
DOI: 10.3998/umurj.9837 ↗Kurz gefasst
Fachartikel in der Zeitschrift University of Michigan Undergraduate Research Journal, online erschienen am 04.06.2026. Verfasst von Jacqueline Aguiar und 2 weiteren Autor:innen. Die Arbeit nennt 6 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Studies document distinctive phonological, syntactic, and morphological features across ethnolinguistic varieties (e.g., Labov, 1964, 2006; Purnell, Idsardi, & Baugh, 1999). Research on language and identity in bilingual speech has focused on adolescents, with only recent attention to Spanish (or its absence) as a racialized/ethnic marker (Zentella, 1997; Bucholtz, 2011; Rosa, 2019). We present a speech perception experiment exploring how Spanish heritage language (HL)-speaking children (ages 7 - 9) across HL proficiencies in Michigan perceive their own and other children’s Spanish use and ethnic-racial identities, with an eye to how varying Spanish features are associated with race/ethnicity phenotypes and social attributes. We introduce a two-part experiment using AI-generated audio-visual stimuli of different dialectal Spanish speakers (Mexican, Caribbean, Southern Cone, Andean) and a low proficiency Spanish, “No Sabo”. Participants provide Yes/No responses regarding the AI child’s friendliness, intelligence, and race/ethnicity. Phase 1 examines dialectal categorization based solely on Spanish audio, while Phase 2 introduces visual stimuli to analyze interactions between language and perceived race/ethnicity. Measures include LexTale proficiency scores, response correlations, and focus group analysis. Preliminary outcomes from pilot data demonstrate that regardless of proficiency level, HL children are attuned to the Spanish dialect of their household, family, and social circle. Pilot participants express judgments about “other” Spanish dialects as being “better” or “worse.” Interestingly, 80% of pilot participants said the “No sabo” dialect was “not intelligent” and were averse to the way that child spoke Spanish. This reaction may indicate that Spanish HL children quickly internalize the complex stigma associated with non-mastery of Spanish. Pilot data suggest child participants with higher Spanish HL proficiency perceive differing Spanish varieties better, but also indexicalize Spanish language characteristics to phenotypic (racial/ethnic) markers, attributing the stereotypical racialized AI images to the corresponding audio of each child.
Zitieren
APA
Aguiar, J., Sims, S., & Satterfield, T. (2026). “The “No Sabo” effect”: Insights from AI into child Spanish heritage language and identity development. University of Michigan Undergraduate Research Journal, 18(0). https://doi.org/10.3998/umurj.9837
Vancouver
Aguiar J, Sims S, Satterfield T. “The “No Sabo” effect”: Insights from AI into child Spanish heritage language and identity development. University of Michigan Undergraduate Research Journal. 2026;18(0). doi:10.3998/umurj.9837
BibTeX
@article{Aguiar2026,
author = {Aguiar, Jacqueline and Sims, Sophia and Satterfield, Teresa},
title = {“The “No Sabo” effect”: Insights from AI into child Spanish heritage language and identity development},
journal = {University of Michigan Undergraduate Research Journal},
year = {2026},
volume = {18},
number = {0},
doi = {10.3998/umurj.9837}
}
EUVIMED