Abstract
Using a detailed transcription, obtained from body-camera, CCTV, and smartphone footage of the murder of George Floyd, we examine the behavior of bystanders as events unfolded. Analysis reveals 205 direct verbal bystander interventions comprised of five forms (declaratives, assessments, interrogatives, imperatives, insults). We also describe the key physical intervention strategies deployed by the bystanders. We show that bystanders prioritize interventions based on what they "know" (rather than asking questions or making demands). We suggest that this is because assessment-based strategies are less likely to be seen as a direct challenge to the power of the police and therefore have more chance of inducing constructive engagement. Although bystanders were ultimately unsuccessful in persuading the police to change course, we identify five moments in the action sequences where the assessment concerns of the bystanders were taken up by the officers-albeit fleetingly. We argue that these bystander interventions create the opportunity for officers to break the pattern of behavior that will lead to murder. It is a failure of the officers and not the bystanders that the police are unable to take those opportunities. We argue that assessment-based interventions have the potential to breach structural and situational power dynamics that usually lead to bystander interventions being overridden or ignored. We conclude by drawing some wider implications for the way bystanders and police officers can be trained to improve the safety of individuals caught up in police arrests. (PsycInfo Database Record (c) 2026 APA, all rights reserved).