When Language Shapes Thought: Cross-Lingual Transfer of Factual Knowledge in Question Answering
This addresses cross-lingual information access for users of non-English languages, offering a novel approach to improve factual knowledge transfer.
The paper tackles the problem of multilingual large language models' sensitivity to input language in factual knowledge use, showing that Language-to-Thought prompting outperforms English-based reasoning by aligning internal thinking with the knowledge source, achieving consistent gains across three languages and four models.
Multilingual large language models (LLMs) offer promising opportunities for cross-lingual information access, yet their use of factual knowledge remains highly sensitive to the input language. Prior work has addressed this through English prompting and evaluation, assuming that English-based reasoning is universally beneficial. In this work, we challenge that assumption by exploring factual knowledge transfer from non-English to English through the lens of Language and Thought Theory. We introduce Language-to-Thought (L2T) prompting, which aligns the model's internal ''thinking'' language with the source of knowledge. Across three languages and four models, L2T consistently outperforms English-based reasoning, reversing the expected advantage of English prompts. Our code is available at https://github.com/GeomeunByeol/Language2Thought.