A groundbreaking new study, “Examining Linguistic Strategies for Digital Mobilization in Social Movements on Instagram and TikTok” in the Communications of the Association for Information Systems journal (vol. 58), reveals how subtle shifts in social media language serve as the primary engine driving political mobilization, civil unrest, and viral malicious campaigns on major digital platforms. By combining core principles from linguistics and communication theory, researchers from the COSMOS team mapped how specific types of functional vocabulary shape the trajectory of activist movements across Instagram and TikTok. By studying over 130,000 social media posts and comments, the authors show how words can influence people’s reactions, decisions, and actions, leading them to gather and stay involved over time.

To organize their analysis, the researchers grouped language into four main styles based on what the words try to achieve. Conative (or, direct action) language uses strong verbs to tell people what to do, like joining a rally or sharing a post. Informational language focuses on facts, dates, names, and locations to give context. Emotional language shares feelings like anger, hope, or fear to bring people closer together. Finally, expressive language shares opinions and judgments to help supporters agree on a common cause. The authors looked at major political events in Brazil and Taiwan to see these language styles in action. In Brazil, researchers analyzed posts surrounding the post-election protests and riots. They found that anti-government groups heavily used action-focused words early on to get people into the streets. On the other hand, pro-government groups used factual and emotional language to identify protesters and build long-term support. In Taiwan, during the 2024 presidential election, TikTok users involved in misleading information campaigns used urgent commands to spread claims of voter fraud, while fact-checkers used opinionated and factual words to coordinate their responses and rebuild trust.

By mapping these post histories, the researchers discovered that successful movements naturally shift their language as they grow. Movements often start with call-to-action words to gain initial attention, shift to factual language to build context, and finish with emotional language to keep supporters connected over time. These findings can help social media platforms and researchers spot early signs of offline protests or dangerous information campaigns before they spread too far.