![[2602.16957] When Semantic Overlap Is Not Enough: Cross-Lingual Euphemism Transfer Between Turkish and English](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.16957] When Semantic Overlap Is Not Enough: Cross-Lingual Euphemism Transfer Between Turkish and English
Summary
This study explores the challenges of cross-lingual euphemism transfer between Turkish and English, highlighting the limitations of semantic overlap in multilingual euphemism detection.
Why It Matters
Understanding euphemism transfer is crucial for improving natural language processing (NLP) models, especially in low-resource languages like Turkish. This research provides insights into how cultural and pragmatic contexts affect language processing, which can enhance multilingual applications and AI systems.
Key Takeaways
- Semantic overlap does not guarantee effective euphemism transfer between languages.
- The study categorizes euphemistic terms into overlapping and non-overlapping subsets.
- Performance can degrade in low-resource language pairs, particularly Turkish to English.
- Counterintuitive results suggest that training on non-overlapping euphemisms may improve outcomes.
- Domain-specific alignment plays a role in euphemism transfer effectiveness.
Computer Science > Computation and Language arXiv:2602.16957 (cs) [Submitted on 18 Feb 2026] Title:When Semantic Overlap Is Not Enough: Cross-Lingual Euphemism Transfer Between Turkish and English Authors:Hasan Can Biyik, Libby Barak, Jing Peng, Anna Feldman View a PDF of the paper titled When Semantic Overlap Is Not Enough: Cross-Lingual Euphemism Transfer Between Turkish and English, by Hasan Can Biyik and Libby Barak and Jing Peng and Anna Feldman View PDF HTML (experimental) Abstract:Euphemisms substitute socially sensitive expressions, often softening or reframing meaning, and their reliance on cultural and pragmatic context complicates modeling across languages. In this study, we investigate how cross-lingual equivalence influences transfer in multilingual euphemism detection. We categorize Potentially Euphemistic Terms (PETs) in Turkish and English into Overlapping (OPETs) and Non-Overlapping (NOPETs) subsets based on their functional, pragmatic, and semantic alignment. Our findings reveal a transfer asymmetry: semantic overlap is insufficient to guarantee positive transfer, particularly in low-resource Turkish-to-English direction, where performance can degrade even for overlapping euphemisms, and in some cases, improve under NOPET-based training. Differences in label distribution help explain these counterintuitive results. Category-level analysis suggests that transfer may be influenced by domain-specific alignment, though evidence is limited by sparsity. Subject...
