Narratives can be powerful tools for inspiring action on pressing societal issues such as climate change. While social science theories offer frameworks for understanding the narratives that arise within collective movements, these are rarely applied to the vast data available from social media platforms, which play a significant role in shaping public opinion and mobilizing collective action. This gap in the empirical evaluation of online narratives limits our understanding of their relationship with public response. In this study, we focus on plant-based diets as a form of pro-environmental action and employ natural language processing to operationalize a theoretical framework of moral narratives specific to the vegan movement. We apply this framework to narratives found in YouTube videos promoting environmental initiatives such as Veganuary, Meatless March, and No Meat May. Our analysis reveals that several narrative types, as defined by the theory, are empirically present in the data. To identify narratives with the potential to elicit positive public engagement, we used text processing to estimate the proportion of comments supporting collective action across narrative types. Video narratives advocating social fight, whether through protest or through efforts to convert others to the cause, are associated with a stronger sense of collective action in the respective comments. These narrative types also demonstrate increased semantic coherence and alignment between the message and public response, markers typically associated with successful collective action. Our work offers new insights into the complex factors that influence the emergence of collective action, thereby informing the development of effective communication strategies within social movements.

翻译：叙事可以成为激发人们应对气候变化等紧迫社会问题采取行动的有力工具。虽然社会科学理论提供了理解集体运动中产生的叙事的框架，但这些理论很少被应用于社交媒体平台所提供的海量数据中，而这些平台在塑造公众舆论和动员集体行动方面发挥着重要作用。在线叙事实证评估方面的这一空白限制了我们对它们与公众反应之间关系的理解。在本研究中，我们聚焦于以植物性饮食作为亲环境行动的一种形式，并运用自然语言处理技术来操作化一个专门针对纯素运动的道德叙事理论框架。我们将该框架应用于YouTube视频中推广Veganuary、Meatless March和No Meat May等环境倡议的叙事。我们的分析表明，该理论定义的几种叙事类型在数据中均有实证出现。为了识别具有引发积极公众参与潜力的叙事，我们使用文本处理来估计不同类型叙事下支持集体行动的评论比例。倡导社会斗争的叙事视频，无论是通过抗议还是通过努力说服他人加入该事业，都与相应评论中更强的集体行动感相关。这些叙事类型还显示出信息与公众反应之间更高的语义一致性和对齐度，这通常是成功集体行动的标志。我们的工作为影响集体行动产生的复杂因素提供了新见解，从而为在社会运动中制定有效的沟通策略提供参考。