Observing the Social Production of Race and Racial Justice Using Word Embeddings

• In Event: Disinformation, Extreme Rhetoric and Online Toxicity in Politics

Abstract

A long line of research suggests that language is a tool that can be leveraged by stakeholders in the world of politics to describe the causes, consequences, and populations affected by social problems and their broader political reality. In this article, I argue that the linguistic journey of the term “woke” can be used to analyze and make insights about the status and social production of race, racism, and racial justice in the United States that are not as easily observed using traditional research methods such as survey, experiments, and ethnographic observation. Specifically, I use word embeddings, a natural language processing method developed by computer scientists, I analyze an original dataset of news articles using the terms “woke” and “politics” from left-leaning, conservative, and international newspapers between 2000 and 2023 to observe how the term “woke” is constructed, reconfigured, and sometimes strategically employed in attempts to craft the “realities” at the center of political debates about racial justice and reconciliation. I draw on study findings to conclude that greater attention to triangulating data on racial language and political discourse could offer scholars a new opportunity to provide ideological context to aggregate patterns in racial classification and opinion formation.

Author

• M. Brielle Harbin, United States Naval Academy