Abstract
The arrival of the short video era has opened up new opportunities for informal language teaching. In this era, algorithms play a crucial role in content distribution, enabling language teaching influencers’ videos to reach a broad audience. However, the issue of algorithmic opacity poses challenges for influencers, who have to explore the effective strategies to enhance their visibility within algorithmic constraints. The exploration process is often lengthy, yet our understanding of this process remains limited.
To address this research gap, this study aims to investigate the diachronic changes in the negotiation between language teaching influencers and algorithms. Using the social semiotic approach to multimodality, this study analyses 107 short videos produced by four language teaching influencers on Douyin (the Chinese version of TikTok). In addition, the analysis is triangulated with data collected from interviews with 10 language teaching influencers.
The findings show that the evolution of content production mirrors the process of negotiation with algorithms. On the one hand, the evolving changes in content production underline the necessity of complying with algorithms. Influencers tend to imagine algorithms as intended viewers and adjust their content to align with algorithmic preferences. Three evolutionary trajectories are identified: 1) progressive refinement of multimodal ensembles, 2) enhanced interaction, and 3) stylistic stabilisation. On the other hand, influencers can assert their agency to reduce reliance on algorithms. One common strategy is to convert public platform traffic to private domains, such as establishing their own fan groups. Nevertheless, the power asymmetry between algorithms and individual agency persists, with algorithms maintaining dominance.
This study not only reveals the co-evolutionary interplay between content production and algorithms, but also offers suggestions for content creators in the digital age.
To address this research gap, this study aims to investigate the diachronic changes in the negotiation between language teaching influencers and algorithms. Using the social semiotic approach to multimodality, this study analyses 107 short videos produced by four language teaching influencers on Douyin (the Chinese version of TikTok). In addition, the analysis is triangulated with data collected from interviews with 10 language teaching influencers.
The findings show that the evolution of content production mirrors the process of negotiation with algorithms. On the one hand, the evolving changes in content production underline the necessity of complying with algorithms. Influencers tend to imagine algorithms as intended viewers and adjust their content to align with algorithmic preferences. Three evolutionary trajectories are identified: 1) progressive refinement of multimodal ensembles, 2) enhanced interaction, and 3) stylistic stabilisation. On the other hand, influencers can assert their agency to reduce reliance on algorithms. One common strategy is to convert public platform traffic to private domains, such as establishing their own fan groups. Nevertheless, the power asymmetry between algorithms and individual agency persists, with algorithms maintaining dominance.
This study not only reveals the co-evolutionary interplay between content production and algorithms, but also offers suggestions for content creators in the digital age.
| Original language | English |
|---|---|
| Publication status | Published - Dec 2025 |
| Event | 8th International Conference of the Asia-Pacific LSP & Professional Communication Association - City University of Hong Kong, Hong Kong, China Duration: 11 Dec 2025 → 13 Dec 2025 https://www.en.cityu.edu.hk/en/conference |
Conference
| Conference | 8th International Conference of the Asia-Pacific LSP & Professional Communication Association |
|---|---|
| Abbreviated title | LSPPC8 |
| Place | China |
| City | Hong Kong |
| Period | 11/12/25 → 13/12/25 |
| Internet address |
Bibliographical note
Research Unit(s) information for this publication is provided by the author(s) concerned.Fingerprint
Dive into the research topics of 'Guessing what’s in the “black box”: A diachronic analysis of the negotiation between language teaching influencers and algorithms'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver