Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Tower of Babel in Cross-Cultural Communication: A Case Study of #Give Me a Chinese Name# Dialogues During the "TikTok Refugees'' Event

Jielin Feng, Zhibo Yang, Jingyi Zhao, Yujia Li, Xinwu Ye, Xingyu Lan, Siming Chen

Abstract

The sudden influx of "TikTok refugees'' into the Chinese platform RedNote in early 2025 created an unprecedented, large-scale online cross-cultural communication event between the West and East. Although prior HCI research has studied user behavior in social media, most work remains confined to monolingual or single-cultural contexts, leaving cross-linguistic and cultural dynamics underexplored. To address this gap, we focused on a particularly challenging cross-cultural encoding-decoding task that remains stubbornly beyond the reach of machine translation, i.e., foreign newcomers asking Chinese users for Chinese names, and examined how people collectively constructed a digital "Babel Tower'' through various information encoding strategies. We collected and analyzed over 70,000 comments from RedNote with a creative human-in-the-loop approach using large language models, deriving a systematic framework summarizing cross-cultural information encoding strategies, how they are combined and layered to complicate decoding, and how they relate to engagement metrics such as the number of likes.

Tower of Babel in Cross-Cultural Communication: A Case Study of #Give Me a Chinese Name# Dialogues During the "TikTok Refugees'' Event

Abstract

The sudden influx of "TikTok refugees'' into the Chinese platform RedNote in early 2025 created an unprecedented, large-scale online cross-cultural communication event between the West and East. Although prior HCI research has studied user behavior in social media, most work remains confined to monolingual or single-cultural contexts, leaving cross-linguistic and cultural dynamics underexplored. To address this gap, we focused on a particularly challenging cross-cultural encoding-decoding task that remains stubbornly beyond the reach of machine translation, i.e., foreign newcomers asking Chinese users for Chinese names, and examined how people collectively constructed a digital "Babel Tower'' through various information encoding strategies. We collected and analyzed over 70,000 comments from RedNote with a creative human-in-the-loop approach using large language models, deriving a systematic framework summarizing cross-cultural information encoding strategies, how they are combined and layered to complicate decoding, and how they relate to engagement metrics such as the number of likes.
Paper Structure (30 sections, 2 equations, 6 figures, 6 tables)

This paper contains 30 sections, 2 equations, 6 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Social Media Research in HCI
  4. Naming Practice in Cross-Cultural Communication
  5. LLMs for Content Analysis
  6. Data
  7. Method
  8. Crowdsourcing LLMs with Consistency-Guided Correction: A Human-in-the-Loop Approach for Extracting Name–Explanation Pairs
  9. Iterative Refinement of the Extraction Approach
  10. Final Application to the Full Dataset
  11. Trade-off Analysis of Different Approaches
  12. Model Performance Comparison
  13. Channel-Specific Generation for Missing and Incomplete Explanation
  14. Name-Explanation Pair Processing
  15. Explanation Generation across Three Channels
  16. ...and 15 more sections

Figures (6)

  • Figure 1: Example posts and comments collected from the naming practice. Blue boxes indicate the posters' photos, green boxes indicate the posters' foreign names, and red boxes indicate example comments associated with each post.
  • Figure 2: The process from obtaining comment data on RedNote to constructing naming strategies framework.
  • Figure 3: Projection of semantic patterns using BERTopic embeddings and applying UMAP for dimensionality reduction.
  • Figure 4: Framework of naming strategies across semantic, phonetic, and visual channels. The semantic channel is organized hierarchically, including four overarching semantic strategies, ten Level-1 categories, twenty-seven Level-2 categories, and thirty-one subcategories at the most detailed level. The phonetic channel comprises three categories, while the visual channel includes seven categories.
  • Figure 5: Top ten combinations within (A) Layered Semantic, (B) Semantic + Phonetic, (C) Semantic + Visual, and (D) Semantic + Phonetic + Visual, each with an illustrative example.
  • ...and 1 more figures