Interacción humanx-chatbot : estudio exploratorio sobre variedades lingüísticas, acomodación y desigualdad

Paula Salerno; Milagros Vilar

doi:10.12957/matraga.2025.88189

Authors

Paula Salerno Universidad Nacional de San Martín - CONICET https://orcid.org/0000-0003-1430-3072
Milagros Vilar Universidad Nacional de San Martín https://orcid.org/0000-0001-9043-0423

DOI:

https://doi.org/10.12957/matraga.2025.88189

Keywords:

Sociolinguistics, Human-Chatbot Interaction, Communicative Accommodation, Linguistic Ideologies, Linguistic Varieties

Abstract

This paper examines how AI conversational agents handle the linguistic diversity of the Spanish-speaking world and explores some of the effects of these technologies in terms of social (in)equality. To do so, we conducted an experiment based on interactions with two chatbots, ChatGPT and Gemini, to observe whether these conversational agents adjust, modify, or accommodate their verbal behavior according to the variety of Spanish used by users. From a sociolinguistic perspective, we analyze chatbots’ actions of communicative accommodation (and non-accommodation) during interactions, as well as the underlying linguistic ideologies shaping their behavior.

From two designed prompts, we identified two types of accommodation: lexical and formal. This approach allowed us to evaluate the extent to which these chatbots adapt to our Rioplatense Spanish variety, and to explore the potential causes of these behaviors. The results reveal that the dynamics of accommodation differ for each bot, though both exhibit certain limitations in their interactional behavior that reflect the reproduction of pre-existing inequalities in our society. Specifically, the analysis uncovered homogenizing discourses and monoglossic ideologies that influence the linguistic decisions of these technological developments, which are often presented as “neutral”.

Finally, we discuss the implications of choosing a specific linguistic variety for AI in the Spanish-speaking world, and we reflect on the role of actors and institutions currently involved in these language policy decisions.

Downloads

Download data is not yet available.

Author Biographies

Paula Salerno, Universidad Nacional de San Martín - CONICET

É doutora em Linguística pela Universidade de Buenos Aires (UBA), onde também obte- ve os títulos de Licenciada e Professora em Letras. Realizou um pós-doutorado no Centro de Estudos da Linguagem em Sociedade (CELES) da Universidade Nacional de San Martín (UNSAM) com bolsa do CONICET e outro no The Graduate Center, City Universityof New York (CUNY), com financiamento da American Association of University Women. Atualmente, é pesquisadora do CELES-UNSAM. Ela se dedica aos estudos do discurso, es- pecialmente às relações entre linguagem, memória social, gênero e discriminação. Entre seus temas de trabalho, estão os discursos contemporâneos sobre a Guerra das Malvinas, a lingua- gem inclusiva e a discriminação linguística.

Milagros Vilar, Universidad Nacional de San Martín

É doutora em Linguística pela Universidade de Buenos Aires (UBA) e mestre em Gestão de Línguas pela Universidade Nacional de Tres de Febrero (UNTREF). É pesquisadora associada no Centro de Estudos da Linguagem em Sociedade (CELES) da Universidade Nacional de San Martín (UNSAM) e faz parte da equipe docente do Diploma em Estudos Avançados da Lingua- gem em Sociedade nesta mesma universidade. Ela é especialista nos estudos da interação,do planejamento e da política linguística e, atualmente, se interessa pela interação com chatbots e pelo estudo das dinâmicas conversacionais em sistemas de inteligência artificial.

References

ALEMANY, Laura Alonso. Inteligencia artificial y valores. In: FERRANTE, Enzo. (Dir.) Aprendizaje automá- gico, un viaje al corazón de la inteligencia artificial contemporánea. Editorial Vera cartonera, 2022. p. 34-

42. <https://www.fhuc.unl.edu.ar/veracartonera/portfolio/aprendizaje-automagico/>.

ALEMANY, Laura Alonso; BENOTTI, Luciana; MAINA, Hernán; GONZÁLEZ, Lucía; RAJNGEWERC, Ma- riela et al. A methodology to characterize bias and harmful stereotypes in natural language processing in Latin America. Cornell University; arXiv, 3, p. 1-24, 2023. <https://arxiv.org/abs/2207.06591>.

ARNOUX, Elvira. Modos de regulación de la discursividad: en torno a la simplificación y la uniformización. La Rivada. Revista electrónica de la Secretaría de Investigación y Posgrado FHyCS-UNaM, 14, p. 15-36, 2020.

ARNOUX, Elvira. Análisis del discurso. Modos de abordar materiales de archivo. Buenos Aires: Santiago Arcos, 2006.

ARNOUX, Elvira; BEIN, Roberto. Introducción: Las representaciones del lenguaje. In: ARNOUX, Elvira y BEIN, Roberto. Prácticas y representaciones del lenguaje. Buenos Aires: Eudeba, 1999. p. 9-15.

BECKER, Lidia. El lenguaje claro o ciudadano en América Latina: ¿un movimiento glotosocial, una nueva regulación democratizadora de la discursividad pública o una herramienta hegemónica?.In: ARNOUX, E. y BEIN, R. (Eds.) Semiótica y política en el discurso público. Buenos Aires: Biblos, 2023. p. 81-103.

BLOMMAERT, Jan. Language, Asylum, and the National Order. Current Anthropology, 50 (4), p. 415- 441, 2009.

BUCHOLTZ, M.; LÓPEZ, A.; MOJARRO, A.; SKAPOULLI, E.; VANDERSTOUWE, C.; WARNER-GARCÍA,

S.. Sociolinguistic justice in the schools. Student researchers as linguistic experts. Language and Linguistics Compass, 8/4, p. 144-157, <2014.https://doi.org/10.1111/lnc3.12070>.

BÜRKI, Yvette; MORGENTHALER GARCÍA, Laura. En torno al olvido en el estudio de la variación lingüís- tica del español. Estudios de Lingüística del Español, 37, p. 3-15, 2016.

CIRILLO, Giusy; RUNNQVIST, Elin; STRIJKERS, Kristof; NGUYEN, Noël; BAUS, Cristina. Conceptual alig- nment in a joint picture-naming task performed with a social robot. Cognition, 227, 105213, 2022. .

COHN, Michelle; ZELLOU, Georgia. Prosodic differences in human- and Alexa-directed speech, but si- milar local intelligibility adjustments. Frontiers in Communication, 6, 2021. <https://doi.org/10.3389/ fcomm.2021.675704>.

DEL VALLE, José. La lengua, ¿patria común? Ideas e ideologías del español. Madrid/Fráncfort: Iberoame- ricana/Vervuert, 2007.

DEL VALLE, José y MEIRINHO-GUEDE, Vítor. Ideologías lingüísticas. In: GUTIÉRREZ-REXACH, Javier. (Ed.): Enciclopedia de Lingüística Hispánica, 2 (2), 2016, p. 622-631.

DOMÍNGUEZ, Verónica. Vidal de Battini y las encuestas del habla regional (1950). El registro de lenguas indígenas en la Patagonia Central. In: LAGUARDA, Paula Inés y ABBONA, Anabela. (Eds.) Diálogos sobre cultura y región. Políticas, identidades y mediación cultural en La Pampa y Patagonia Central (siglos XX y XXI). Teseo Press, 2023. Cap. 7. <https://www.teseopress.com/dialogossobreculturayregion/chapter/capitu- lo-7-vidal-de-battini-y-las-encuestas-del-habla/>.

GALLOIS, Cindy; OGAY, Tania; GILES, Howard. Communication accommodation theory: a look back and a look ahead. In: GUDYKUNST, William (Ed.) Theorizing About Intercultural Communication. London: Sage, 2005. p. 121-148.

GILES, Howard; EDWARDS, America L.; WALTHER, Joseph B. Communication accommodation theory: Past accomplishments, current trends, and future prospects. Language Sciences, 99, 101571, 2023. .

GILES, Howard; TAYLOR, Donald M.; BOURHIS, Richard Y. Towards a theory of interpersonal accommo- dation through speech: some Canadian data. Language in Society, 2 (2), p. 177–192, 1973.

KERBRAT-ORECCHIONI, Catherine. L’énonciation. De la subjectivité dans le langage. Paris: Armand Co- lin, 2006 [1999].

LAURIA, Daniela. Sobre el programa Justicia en lenguaje claro del Ministerio de Justicia y Derechos Huma- nos de la Nación (Argentina). Entremeios. Revista de Estudos do Discurso, 18, p. 43-61, 2019.

LAURIA, Daniela; LÓPEZ GARCÍA, María. Instrumentos lingüísticos académicos y norma estándar del es- pañol: la nueva política lingüística panhispánica. Lexis, 33(1), p. 49-89, 2009.

LEE, Kenny. Why are datasets important in training LLMs? Kili [blog], 2024. Disponible en: . Acceso: 3 de marzo de 2025.

LÓPEZ GARCÍA, María. El otro yo: la identidad lingüística argentina como conflicto. Amérika. Mémories, identités, territoires, 9, 2013. <https://web.archive.org/web/20220419132713id_/https://journals.openedi- tion.org/amerika/4387>.

MARTÍN ROJO, Luisa; PUJOLAR, Joan; AMORRORTU, Estibaliz. Negotiating sociolinguistic justice: tur- ning spaces of inequality into spaces of conscientization. International Journal of the Sociology of Langua- ge, v. 2025, n. 291, p. 1-26, 2025. <https://doi.org/10.1515/ijsl-2022-0114>.

MENGESHA, Zion; HELDRETH, Courtney; LAHAV, Michal; SUBLEWSKI, Juliana; TUENNERMAN, Ely-

se. “I don’t think these devices are culturally sensitive.” Impact of automated speech recognition errors on African Americans. Frontiers in Artificial Intelligence, 4, 2021. <https://doi.org/10.3389/frai.2021.725911>.

MORENO-FERNÁNDEZ, Francisco. Variedades del español en los Estados Unidos. In: MORENO-FER- NÁNDEZ, Francisco. Variedades de la lengua española. Londres/Nueva York: Routledge, 2020. p. 130-141.

PICKERING, Martin J.; GARROD, Simon. Alignment as the Basis for Successful Communication. Research Language Computation,4(2-3), p. 203–228, 2006.

PÜTZ, Ole; ESPOSITO, Elena. Performance without understanding: How ChatGPT relies on hu- mans to repair conversational trouble. Discourse & Communication, 0(0), 2024. .

SCHEGLOFF, Emanuel A.; JEFFERSON, Gail; SACKS, Harvey. The preference for self-correction in the orga- nization of repair in conversation. Language, 53, p. 361–82, 1977.

SHEN, Huiyang; WANG, Min. Effects of social skills on lexical alignment in human-human interaction and human-computer interaction. Computers in Human Behavior, 143, 108818, 2023. <https://doi.org/10.1016/j. chb.2023.107718>.

SHIRO, Martha. Editorial. El método tampoco viene del aire... Revista Latinoamericana de Estudios del Discurso, 12(2), p. 3-6, 2016.

SLEZAK, Diego F. Una breve introducción a la inteligencia artificial. In: FERRANTE, Enzo. (Dir.) Aprendi- zaje automágico, un viaje al corazón de la inteligencia artificial contemporánea. Editorial Vera cartonera, 2022. p. 12-17. <https://www.fhuc.unl.edu.ar/veracartonera/portfolio/aprendizaje-automagico/>.

SONEIRA, Abelardo Jorge. La “Teoría fundamentada en los datos” (Grounded Theory) de Glaser y Strauss. In: VASILACHIS DE GIALDINO, Irene. (Coord.) Estrategias de investigación cualitativa. Barcelona: Gedi- sa, 2006. p. 153-173.

UNAMUNO, Virginia. “¿Qué son las lenguas?” In: UNAMUNO, Virginia. Lenguaje y educación.Quilmes: Universidad Nacional de Quilmes, 2016. p. 9-30.

VIDAL DE BATTINI, Berta Elena. El español de la Argentina. Buenos Aires: Consejo Nacional de Educa- ción, 1964.

WANG, Boxuan; THEUNE, Mariët; SRIVASTAVA, Sumit. Examining Lexical Alignment in Human-Agent Conversations with GPT-3.5 and GPT-4 Models. In: FØLSTAD, Asbjørn, et al. (Coord.) Chatbot Research and Design. Conversations 2023. Lecture Notes in Computer Science, vol 14524. Springer, Cham, 2024. .

ZAJKO, Mike. Artificial intelligence, algorithms, and social inequality: Sociological contributions to con- temporary debates. Sociology Compass, 16(3), 2022. <https://compass.onlinelibrary.wiley.com/doi/10.1111/ soc4.12962>.

ZAVALA, Virginia. Justicia sociolingüística para los tiempos de hoy. Íkala, Revista de Lenguaje y Cultura,

v. 24, n. 2, p. 343-359, 2019. <https://doi.org/10.17533/udea.ikala.v24n02a09>.