Extreme droughts, heat waves, fires, hurricanes, floods, and landslides cause the largest losses in the United States, and globally, from natural hazards linked to weather and climate. There is evidence that the frequency of such extremes is increasing, particularly for heat waves, large fires, and intense precipitation, making better understanding of the probability and consequences of these events imperative. Further, these events are not isolated, but rather interact with each other, and with social and ecological vulnerability, to amplify impacts. Less is known about the nature and strength of these interactions. Natural and social science subfields frame extreme events with different definitions and analytical approaches, and most analyses neglect interactions and the subsequent novel extremes that can arise. Here we propose a framework for socio-environmental extremes, defined as extraordinary events that emerge from interactions among biophysical and social phenomena and have some degree of social impact. We review how different fields approach extremes as interacting phenomena and propose a synthetic framework for conceptualizing and defining extremes from both an environmental and social perspective. This approach recognizes multiple drivers and responses that yield extreme events and extreme outcomes, and reconciles the gap between understanding extremes as biophysical processes and their social underpinnings and impacts. We conclude with a future research agenda that adds clarity and direction to understanding the extreme events that matter to society. This agenda will help to identify where, when, and why communities may have high exposure and vulnerability to socio- environmental extremes—informing future mitigation and adaptation strategies.