Linguistic Capital and Cultural Boundaries in Artificial Intelligence: From the Digital Divide to Symbolic Sovereignty

Abstract

This paper investigates how large language models (LLMs) restructure global linguistic hierarchies by introducing the concept of symbolic sovereignty—the capacity of a language to shape meaning, discourse, and cultural visibility within AI systems. This concept is grounded in a key empirical observation: certain languages, such as Chinese in DeepSeek, are not only used more frequently but also serve as default generative frameworks, guiding semantic construction across multiple outputs. This privileged role is evident in how Chinese manages stylistic variation, initiates rhetorical structures, and frames meaning for other languages. Rather than mere technical inclusion, such dominance constitutes a form of symbolic power. To support this claim, the paper combines semiotic theory and comparative analysis of multilingual outputs from ChatGPT and DeepSeek. Drawing from Saussure's structural linguistics, Barthes's mythologies, Eco's open systems, and Bourdieu's linguistic capital, the study reveals how AI models embed linguistic hierarchies into their internal architecture. As a response to this emerging power structure, the paper introduces the ethical notion of the right to linguistic generation, advocating structural parity in model design, semantic flexibility, and cultural representation. Ultimately, the paper argues that generative AI is not merely a technical tool but a regime of meaning production, where language is not just spoken—it speaks back, shaping who gets to define the world and how.

Keywords

Semiotics / LLMs / Symbolic Sovereignty / Digital Colonization / Artificial Intelligence

References

[1] Barnard M., & Otte W. (2024). Is the machine surpassing humans? Large language models, structuralism, and liturgical ritual: A position paper. International Journal of Practical Theology, 28(2), 289-306. https://doi.org/10.1515/ijpt-2023-0078

[2] Barthes R. (1972). Mythologies (A. Lavers, Trans.). New York, NY: Hill and Wang.

[3] Baudrillard J. (1994). Simulacra and simulation (S. F. Glaser, Trans.). Ann Arbor, MI: University of Michigan Press.

[4] Beijing Academy of Artificial Intelligence. (2024). DeepSeek-V3: 开启 MoE 大模型新纪元,训练 Tokens 超过 2 万亿 [WeChat Official Account]. Retrieved from https://mp.weixin.qq.com/s/_ZIJ1YwzU17TEqyM0-c27A

[5] Bender E. M., & Friedman B. (2018). Data statements for natural language processing: Toward mitigating system bias and enabling better science. Transactions of the Association for Computational Linguistics, 6, 587-604. https://doi.org/10.1162/tacl_a_00041

[6] Bickley S. J., & Torgler B. (2023). Cognitive architectures for artificial intelligence ethics. AI & Society, 38(2), 501-519. https://doi.org/10.1007/s00146-022-01452-9

[7] Boer R. (2011). The robbery of language? On Roland Barthes and myth. Culture, Theory and Critique, 52(2-3), 213-231. https://doi.org/10.1080/14735784.2011.630890

[8] Bourdieu P. (1991). Language and symbolic power. Cambridge, MA: Harvard University Press.

[9] Calderon Gomez D. (2021). The third digital divide and Bourdieu: Bidirectional conversion of economic, cultural, and social capital to (and from) digital capital among young people in Madrid. New Media & Society, 23(9), 2534-2553. https://doi.org/10.1177/1461444820933252

[10] Campbell C. (2018). Educating openness: Umberto Eco's poetics of openness as a pedagogical value. Signs and Society, 6(2), 305-331. https://doi.org/10.1086/695567

[11] Couldry N., & Mejias U. A. (2019). The costs of connection: How data is colonizing human life and appropriating it for capitalism. Stanford, CA: Stanford University Press.

[12] Danesi M. (2015). Preface: Umberto Eco and semiotics: An enduring legacy. Semiotica, 206, 1-3. https://doi.org/10.1515/sem-2015-0025

[13] Eco U. (1986). Semiotics and the Philosophy of Language. Bloomington, IN: Indiana University Press.

[14] Eco U. (1989). The open work. Cambridge, MA: Harvard University Press.

[15] Forster C. (2022). Are large language models operationalizations of Saussurean structure? Retrieved from http://cforster.com/2022/07/on-words

[16] Foucault M. (1981). The order of discourse. In R. Young (Ed.), Untying the text: A post-structuralist reader (pp. 48-78). Boston, MA: Routledge & Kegan Paul.

[17] Joshi P., Santy S., Budhiraja A., Bali K., & Choudhury M. (2020). The state and fate of linguistic diversity and inclusion in the NLP world (preprint). arXiv:2004. 09095. Retrieved from https://arxiv.org/abs/2004.09095

[18] Liu A., Feng B., Xue B., Wang B., Wu B., Lu C., … Pan Z. (2024). DeepSeek-V3 technical report (preprint). arXiv:2412. 19437. Retrieved from https://arxiv.org/abs/2412.19437

[19] Magee L., Arora V., & Munn L. (2023). Structured like a language model: Analysing AI as an automated subject. Big Data & Society, 10(2). https://doi.org/10.1177/20539517231210273

[20] Mignolo W. D. (2009). Epistemic disobedience, independent thought and decolonial freedom. Theory, Culture & Society, 26(7-8), 159-181. https://doi.org/10.1177/0263276409349275

[21] Mignolo W. D. (2011). The darker side of western modernity: Global futures, decolonial options. Durham, NC: Duke University Press

[22] Monti R. (2021). Umberto Eco and the aesthetics of vagueness. European Journal of Pragmatism and American Philosophy, 13(1). https://doi.org/10.4000/ejpap.2306

[23] Mu G. M., Dooley K., & Luke A. (2019). Bourdieu and Chinese education. Abingdon, UK: Routledge.

[24] Peirce C. S. (1998). The essential Peirce:Selected philosophical writings (Vol. 2). Bloomington, IN: Indiana University Press.

[25] Vincze A. (2024). In the footsteps of Bourdieu towards digital capital: a case study on the application of the concept of digital capital on the relationship between digital and educational inequalities. Belvedere Meridionale, 36(1), 24-42. https://doi.org/10.14232/belv.2024.1.2

[26] Vromen E. (2024). Language models as semiotic machines: Reconceptualizing AI language systems through structuralist and post-structuralist theories of language (preprint). arXiv:2410.13065. Retrieved from https://arxiv.org/abs/2410.13065