This project explores the efficacy of the argument that social structural factors generate a public discourse conducive to the formation of vocabularies of punitive motives. The argument has implications for a model of punishment since several theorists posit that how and when a society punishes its criminals depends upon its social environment. To date, this argument has remained at the level of conjecture. This project is an initial step in determining if structural variation in public discourse exists. County-level data on discourse were derived from a content analysis of newspapers over a 30-year period. I found limited support for a structural explanation. Specifically, time-series analysis (ARIMA) for two counties in postbellum Georgia showed that under selected conditions, certain claims of threat were predicted. These conditions required acute and/or extended economic crises and contested political events. Claims most vulnerable to structural variation were those differentially targeting the activities of black offenders for condemnation. This last point has implications for the social production of a deviant population that is black. Smaller incremental social change, such as diminishing racial inequality, registered intermittent effects as well. Although no one factor was decisive across contexts and measures, the project has implications for a model of punishment. Specifically, an adequate model of punishment necessitates not simply the examination of penal events and statistics, but an identification and analysis of those contextual processes underlying the generation of these events.