非对称对话中的基础性误解：基于视角主义的MapTask标注方案

摘要

协作对话依赖于参与者逐步建立共同基础，但在非对称情境下，对话者可能自认为达成共识却实际指涉不同实体。我们为HCRC地图任务语料库（Anderson等人，1991）引入了一种透视主义标注方案，该方案分别捕捉说话者和受话者对每个指称表达的情境化解读，从而能追踪理解如何随时间推移形成、分化及修复。通过采用方案约束的大语言模型标注流程，我们获得了1.3万个带可靠性评估的标注指称表达，并分析了由此产生的理解状态。结果表明：在统一词汇变体后，完全误解现象较为罕见，但多重性差异会系统性地引发理解分歧，揭示出表面共识可能掩盖指称错位的机制。本框架既为研究情境化误解提供了资源与分析视角，也为评估（视觉）大语言模型在协作对话中建模视角依存性 grounding 的能力提供了方法论基础。

English

Collaborative dialogue relies on participants incrementally establishing common ground, yet in asymmetric settings they may believe they agree while referring to different entities. We introduce a perspectivist annotation scheme for the HCRC MapTask corpus (Anderson et al., 1991) that separately captures speaker and addressee grounded interpretations for each reference expression, enabling us to trace how understanding emerges, diverges, and repairs over time. Using a scheme-constrained LLM annotation pipeline, we obtain 13k annotated reference expressions with reliability estimates and analyze the resulting understanding states. The results show that full misunderstandings are rare once lexical variants are unified, but multiplicity discrepancies systematically induce divergences, revealing how apparent grounding can mask referential misalignment. Our framework provides both a resource and an analytic lens for studying grounded misunderstanding and for evaluating (V)LLMs' capacity to model perspective-dependent grounding in collaborative dialogue.

非对称对话中的基础性误解：基于视角主义的MapTask标注方案

Grounded Misunderstandings in Asymmetric Dialogue: A Perspectivist Annotation Scheme for MapTask

摘要

Support