发表机构
CRIL, CNRS – Université d’Artois; CEPED, Université Paris Cité; EDA, Université Paris Cité（CRIL，法国国家科学研究中心——阿图瓦大学; CEPED，巴黎西岱大学; EDA，巴黎西岱大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出无需标注训练数据的经验互文性检测任务，采用多种方法分析法语移民叙事，发现Qwen2.5-7B零-shot及监督混合方法表现最优，且叙事位置显著影响互文性。
AI 中文摘要
穿越跨撒哈拉、巴尔干等地理迥异路线的移民，常讲述出高度相似的亲身经历：警察暴力、蛇头剥削、危险过境、家庭分离。本文提出经验互文性检测任务，即无需标注训练数据，自动识别移民叙事间共享的经验共鸣。我们从涵盖上述两条路线的108篇法语移民叙事中，自动生成句子对，并采用多种无标注方法对其打分：词汇基线、句子嵌入、基于词性（POS）的结构特征、移民专用主题词典、上下文感知叙事特征，以及采用三种提示策略的Qwen2.5-7B、Mistral-7B零-shot大语言模型（LLM）打分。我们以816条专家标注的互文性判断验证所有方法（标注者间Krippendorff's α=0.27）。结果显示，所有表面、结构及嵌入方法与专家判断的相关性均较弱（r≤0.30）；Qwen2.5-7B零-shot单方法相关性最佳（r=0.38）；少样本示例会降低Qwen的表现，但能显著提升Mistral的性能；叙事位置可显著预测互文性，出发阶段的句子对经验共鸣最强；结合全部31种特征的监督混合方法取得r=0.45，较最佳单方法提升21%。
英文摘要
Migrants traversing geographically distinct routes such as the Trans-Saharan and Balkan corridors often recount strikingly parallel lived experiences: police violence, smuggler exploitation, dangerous crossings, and family separation. We introduce the task of experiential intertextuality detection: automatically identifying shared experiential echoes across migration narratives without requiring annotated training data. From 108 French migration narratives spanning both corridors, we automatically generate sentence pairs and score them using annotation-free methods: lexical baselines, sentence embeddings, POS-based structural features, a migration-specific theme lexicon, context-aware narrative features, and zero-shot LLM scoring with Qwen2.5-7B and Mistral-7B under three prompting strategies. We validate all methods against 816 expert-annotated intertextuality judgments (inter-annotator Krippendorff's $α= 0.27$). Our results reveal that all surface, structural, and embedding methods correlate only weakly with expert judgments ($r \leq 0.30$); Qwen2.5-7B zero-shot achieves the best single-method correlation ($r = 0.38$); few-shot examples degrade Qwen but dramatically improve Mistral; narrative position significantly predicts intertextuality, with departure-phase pairs showing the highest experiential echoes; and a supervised hybrid combining all 31 features achieves $r = 0.45$, a 21% improvement over the best individual method.
Comments11 pages, 3 figures, 5 tables. Accepted at SIGDIAL 2026 (27th Annual Meeting of the Special Interest Group on Discourse and Dialogue)
Journal refProceedings of the 27th Annual Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL 2026), Association for Computational Linguistics, 2026
DOI：10.18653/v1/2026.sigdial-1.NN
