Language norms, linguistic diversity and sociolinguistic hierarchies in the age of AI

La conferenza fa parte del ciclo delle Conversazioni linguistiche, una serie di incontri che offrono a linguiste/i italiane/i e internazionali un'occasione per confrontarsi sui rispettivi studi e per tenersi aggiornate/i sui nuovi sviluppi della ricerca.

• Britta Schneider, University of Frankfurt

Descrizione

In this presentation, I discuss how digital language technologies affect language ideologies, that is, concepts and epistemologies of language. Language technologies based on the exploitation of large data sets – from machine translation and automatic text generation to digital voice assistants (see e.g. Bommasani, Liang et al. 2021) – have an impact on what we understand as language and on social and academic evaluations of language variation and multilingualism. I ask whether and how normative, monolingual concepts regarding language as well as the formation of linguistic prestige are influenced by digital language technologies. I first introduce theoretical concepts that understand languages as discursive constructs (e.g. Irvine and Gal 2009, Pennycook 2004) and review the role of media technology (e.g. writing and print) in shaping such constructs (e.g. Linell 2005, Ong 1982). Secondly, I present language ideological discourses found in the context of digital language technologies. I discuss how language is technologically and discursively constructed in language technologies that make use of Big Data corpora and machine learning and give insight into the language ideological discourses of programmers and users of these technologies. The analysis draws on expert discourse as found in the realm of computational linguistics (e.g. Bender et al. 2021, Blodgett and O’Connor 2017, Vanmassenhove, Shterionov, and Way 2019, Zhao et al. 2017), in commercial settings (e.g. Crawford 2017, Microsoft 2018) as well as on interviews with programmers and users of language technologies. I finally scrutinize the potential effects of commercial language technology on language hierarchies and on epistemologies that define how we understand language.

Modalità di partecipazione

• partecipazione in presenza, ogni venerdì dalle 12.00 alle 13.30. Gli utenti esterni a Unitn sono invitati a prenotarsi inviando una mail all'indirizzo romano.madaro [at] unitn.it. Al momento dell'ingresso dovranno inoltre compilare il documento cartaceo di autodichiarazione del rispetto delle norme anti-Covid (dichiarando contestualmente il possesso del Green Pass).
• partecipazione da remoto su piattaforma multimediale Zoom, ogni venerdì dalle 12.00 alle 13.30. Il link potrà essere richiesto inviando una mail all'indirizzo romano.madaro [at] unitn.it.