📄 中文摘要
在动员社会、扭曲沟通和重塑集体行为方面，很少有力量能与恐惧匹敌。在计算语言学领域，恐惧主要被视为一种情感进行研究，而非一种独特的言论形式。恐惧言论内容普遍存在且日益增长，由于其表面上“更文明”且能规避审查，在传播范围和互动性上往往超越仇恨言论。然而，对恐惧言论的计算研究仍然支离破碎。建立一个能够捕捉恐惧言论多方面特征的理论框架至关重要，该框架需涵盖其语言策略、心理效应及其在社会中的传播机制。需要开发专门的数据集和标注指南，以识别和分类不同类型的恐惧言论，例如煽动性言论、恐吓性言论和制造恐慌的言论。深度学习模型和自然语言处理技术可以用于检测恐惧言论的存在、强度和潜在影响，并区分其与合法的担忧或警告。研究恐惧言论在不同文化和语言背景下的表现形式及其对特定群体或社群的影响，将揭示其复杂的社会动态。此外，还需要探索有效的干预策略，既能应对恐惧言论的负面影响，又能保护言论自由，例如通过开发智能审核工具、教育公众识别和批判性评估恐惧言论，以及促进基于事实的沟通。最终目标是构建一套全面的理论、方法和工具，以更好地理解、识别和应对数字环境中日益增长的恐惧言论现象，从而维护健康的公共对话和信息环境。
📄 English Summary
From Emotion to Expression: Theoretical Foundations and Resources for Fear Speech
Few forces rival fear in their ability to mobilize societies, distort communication, and reshape collective behavior. In computational linguistics, fear is primarily studied as an emotion, not as a distinct form of speech. Fear speech content is widespread and growing, often outperforming hate speech content in reach and engagement because it appears 'civiler' and evades moderation. Yet, the computational study of fear speech remains fragmented. Establishing a theoretical framework capable of capturing the multifaceted characteristics of fear speech is crucial, encompassing its linguistic strategies, psychological effects, and mechanisms of propagation within society. Developing specialized datasets and annotation guidelines is necessary to identify and categorize various types of fear speech, such as incendiary rhetoric, intimidating discourse, and panic-inducing narratives. Deep learning models and natural language processing techniques can be employed to detect the presence, intensity, and potential impact of fear speech, while also differentiating it from legitimate concerns or warnings. Investigating how fear speech manifests across different cultural and linguistic contexts, and its effects on specific groups or communities, will reveal its complex social dynamics. Furthermore, exploring effective intervention strategies is vital, both to counter the negative impacts of fear speech and to protect freedom of expression. This includes developing intelligent moderation tools, educating the public to identify and critically evaluate fear speech, and promoting fact-based communication. The ultimate goal is to construct a comprehensive set of theories, methodologies, and tools to better understand, identify, and respond to the growing phenomenon of fear speech in digital environments, thereby safeguarding healthy public discourse and information ecosystems.