The language of political violence and impunity: News framing of Israel and Palestine

Within the field of political communication, it is widely recognized that text reflects complex, enduring interests that point to deep-rooted aspects of ideology and political influence. A conduit of such influence is the news media, which constructs narratives and makes deliberate rhetorical choices in an effort to reinforce existing power structures, framing events to legitimize and advance specific aims, worldviews, and doctrines. When examining news media bias, specific indicators (e.g., word choice and labeling) and sources (e.g., interest groups and corporations) become evident. Historically, research has shown that media representation tends to privilege Israeli perspectives, while coverage of Palestinian suffering and experience remains comparatively limited and less sympathetic. While the existing literature on the pervasiveness of this imbalance in U.S. news outlets is exhaustive, there is less research on comparisons made between Western media and those based in the Middle East. To pursue a data-driven understanding of media bias, we leverage natural language processing (NLP). Our NLP approach relies on quantitative analyses of article headlines from four news sources: Al Jazeera, The Guardian, The New York Times, and The Washington Post. Notably, Al Jazeera’s headlines were, on average, more negative in sentiment, employed the passive voice least frequently, and featured more critical language of political state actors and their actions. Our findings contribute to discussions on media accountability in conflict reporting. The current study draws attention to the challenges facing international human rights and the role of news media in producing impunity.在政治传播学领域，学界已达成广泛共识：文本承载着复杂且持久的利益诉求，其背后指向意识形态与政治影响力的深层维度。承载此类影响力的核心渠道之一便是新闻媒体：新闻媒体会构建叙事框架，并通过刻意的修辞选择强化既有权力结构，通过对事件的框架化报道，为特定目标、世界观与信条赋予合法性并推动其传播。在探讨媒体偏见时，诸如措辞选择与标签化这类具体指标，以及利益集团、企业这类消息来源，其偏向性会清晰显现。过往研究表明，媒体报道往往偏袒以色列视角，而对巴勒斯坦民众的苦难与经历的报道则相对匮乏，且缺乏共情性。尽管现有研究已对美国新闻媒体中这种失衡现象的普遍性进行了充分探讨，但针对西方媒体与中东本土媒体的对比研究仍较为匮乏。为以数据驱动的方式理解媒体偏见，本研究采用了自然语言处理（Natural Language Processing）技术。本研究的自然语言处理方法依托对四家新闻媒体的文章标题进行的量化分析，这四家媒体分别是半岛电视台（Al Jazeera）、《卫报》（The Guardian）、《纽约时报》（The New York Times）与《华盛顿邮报》（The Washington Post）。值得注意的是，半岛电视台的标题平均情感倾向更为负面，被动语态使用频率最低，且更多使用针对政治国家行为体及其行动的批判性表述。本研究的发现可为冲突报道中的媒体问责相关讨论提供参考。本研究还提请学界关注国际人权事业面临的诸多挑战，以及新闻媒体在助长有罪不罚现象中所扮演的角色。