Linguistically Differentiating Acts and Recalls of Racial Microaggressions on Social Media

TR Number



Journal Title

Journal ISSN

Volume Title


Virginia Tech


Experiences of interpersonal racism persist as a prevalent reality for BIPOC (Black, Indigenous, People of Color) in the United States. One form of racism that often goes unnoticed is racial microaggressions. These are subtle acts of racism that leave victims questioning the intent of the aggressor. The line of offense is often unclear, as these acts are disguised through humor or seemingly harmless intentions. In this study, we analyze the language used in online racial microaggressions ("Acts") and compare it to personal narratives recounting experiences of such aggressions ("Recalls") by Black social media users. We curated a corpus of acts and recalls from social media discussions on platforms like Reddit and Tumblr. Additionally, we collaborated with Black participants in a workshop to hand-annotate and verify the corpus. Using natural language processing techniques and qualitative analysis, we examine the language underlying acts and recalls of racial microaggressions. Our goal is to understand the lexical patterns that differentiate the two in the context of racism in the U.S. Our findings indicate that neural language models can accurately classify acts and recalls, revealing contextual words that associate Blacks with objects that perpetuate negative stereotypes. We also observe overlapping linguistic signatures between acts and recalls, serving different purposes, which have implications for current challenges in social media content moderation systems.



Natural Language Processing, Human Centered Computing, Race and Ethnicity