<p>A growing literature indicates that local violence impairs children’s cognitive performance, but it has not systematically theorized and examined the range of potential cognitive and emotional responses and adaptations to violence that could have performance implications. I draw on the psychological study of risk/adversity and competence/adaptation to broaden yet formalize and clarify the multiple theoretical pathways that connect violence to cognitive performance. I illustrate the utility of this approach by combining test scores from Mexican children with homicide data. The results suggest that repeated recent homicide events in places with previous homicidal violence have a U-shaped association with children’s cognitive performance. A similar pattern of results can be found when child attentiveness is examined as the outcome. I argue that children’s cognitive performance decreases and then increases with homicides–along with impulse control and attentiveness – as a result of processes of sensitization, desensitization, and challenge.</p>