Language models are often trained on text alone, without additional grounding. There is debate as to how much of natural language semantics can be inferred from such a procedure. We prove that entailment judgments between sentences can be extracted from an ideal language model that has perfectly learned its target distribution, assuming the training sentences are generated by Gricean agents, i.e., agents who follow fundamental principles of communication from the linguistic theory of pragmatics. We also show entailment judgments can be decoded from the predictions of a language model trained on such Gricean data. Our results reveal a pathway for understanding the semantic information encoded in unlabeled linguistic data and a potential framework for extracting semantics from language models.

翻译：语言模型通常仅基于文本进行训练，缺乏额外的真实世界参照。关于此类训练过程能在多大程度上还原自然语言语义，学界存在争议。我们证明：若训练语句由遵循语用学语言交际基本准则的格莱斯主体生成，则从完全习得目标分布的理想语言模型中可提取句子间的蕴含判断。我们同时表明，在基于格莱斯数据训练的语言模型中，其预测结果可解码出蕴含判断。该研究揭示了理解无标注语言数据中语义编码的路径，并为从语言模型中提取语义提供了潜在框架。