A heated argument on social media may look like the beginning of a movement’s collapse. New research suggests it can be the opposite. When people from different campaigns encounter one another online, disagreement may help them identify common ground, refine their demands and coordinate collective action. In a study of the 2019 global climate strikes, researchers found that online conflict was not simply a destructive force. Instead, cooperation and contention developed together as separate environmental causes gradually became part of a larger international mobilization.
The study, led by researchers from McGill University, Northeastern University and the University of Massachusetts, examined how a “movement of movements” formed around the climate strikes associated with Swedish activist Greta Thunberg. The mobilization brought together people concerned with issues that were not identical, including opposition to pipelines, protection of forests, conservation of monarch butterflies and wolves, climate justice and fossil-fuel reduction. Although these campaigns had their own histories, priorities and political language, social media created a space where their boundaries could be crossed.
The researchers describe this process as unfolding in three broad phases: exposure, entanglement and coalescence. During exposure, activists and supporters encountered campaigns they might not otherwise have seen. A person sharing information about a school climate strike could be exposed to posts about Indigenous land rights, biodiversity loss or local resistance to oil infrastructure. This did not immediately produce agreement, but it expanded the range of issues participants considered part of the climate problem. Social media, in this phase, acted as a discovery system connecting previously separate networks.
The second phase, entanglement, began when those networks started interacting more intensely. Conversations became increasingly complex because participants brought different definitions of the problem, different political strategies and different ideas about who should be represented. Some users emphasized individual lifestyle changes, while others focused on corporate power, colonialism or government policy. Such disagreements could produce criticism, frustration and heated exchanges. Yet the researchers argue that this friction also made the movement’s internal relationships visible. By challenging one another, participants clarified which concerns could be combined and which required continued debate.
The final phase, coalescence, involved the emergence of shared goals and coordinated action. The different campaigns did not become identical, nor did every disagreement disappear. Instead, enough participants converged around common ideas—especially the urgency of climate action and the need for large-scale political change—to support joint demonstrations and shared messaging. The result was a global mobilization capable of linking local events into a recognizable international campaign. In this account, unity did not come from eliminating difference. It came from organizing difference around a sufficiently broad collective agenda.
To investigate these dynamics, the team analyzed approximately 1.9 million Twitter posts published between April and September 2019. The researchers combined natural language processing with semantic network analysis, computational methods that can reveal relationships among words, themes and accounts across very large collections of text. Natural language processing helped identify recurring concepts and patterns in the posts, while semantic network analysis made it possible to examine how issues became connected over time. In a network representation, subjects that frequently appear together are linked, allowing researchers to track when separate issue clusters begin to overlap.
This approach differs from methods based only on interviews or individual case studies. Interviews can provide detailed accounts of activists’ experiences, but they may not capture the full scale or changing structure of a rapidly expanding online conversation. By analyzing millions of messages, the researchers could observe how the vocabulary of the movement shifted and how different topics became associated with one another. The method does not mean that every post represented a committed activist or that online discussion automatically produced offline action. Rather, it provided a large-scale map of the communication patterns through which movements encountered, contested and connected their ideas.
The findings also have implications for the social media environment that has emerged since 2019. At that time, Twitter feeds were more strongly organized around chronological streams and accounts users had chosen to follow. Many platforms now rely more heavily on recommendation systems designed to predict what will keep users watching, reading or scrolling. These systems can expose people to causes they did not actively search for, potentially accelerating the discovery phase of movement building. At the same time, recommendation algorithms are often opaque, and their decisions can be shaped by engagement metrics rather than by the accuracy, representativeness or social value of the content being promoted.
That distinction makes the 2019 climate strikes a useful benchmark for studying contemporary digital activism. The basic ability of social media to connect dispersed causes remains, but the mechanisms determining who sees which messages have changed. Algorithmic recommendations may help a local protest reach people far beyond its original audience, while also creating unpredictable pathways for polarization, misinformation or harassment. The study’s central lesson is therefore not that online conflict is always beneficial. Instead, it suggests that disagreement can be part of the process by which broad movements define themselves, provided that participants eventually find enough shared purpose to act together. The same networks that generate arguments can also transform isolated concerns into coordinated collective action.
Subject of Research: Not applicable
Article Title: Entangling and Coalescing on Social Media: A Study of Issue Boundary Spanning in Movement of Movements
News Publication Date: 22 June 2026
Web References: https://www.mcgill.ca/desautels/ ; https://doi.org/10.1111/isj.70052
References: Carol Lee, Pratyush Bharati and Emmanuelle Vaast, “Entangling and Coalescing on Social Media: A Study of Issue Boundary Spanning in Movement of Movements,” Information Systems Journal, DOI: 10.1111/isj.70052
Keywords: social media, social movements, climate change, climate strikes, environmentalism, Twitter, natural language processing, semantic network analysis, collective action, online conflict
