We face an urgent need for better theories of rapid change in social-ecological systems. I will first use an ancient philosophical puzzle, modern ideas about networks and hierarchy theory, and an analysis of 17 well-documented historical cases of societal collapse to explore the hypothesis that human social structure influences both the likelihood of collapse and the nature of the feedbacks between ecosystems and societies. I will then extend these ideas to consider the need for structure-process theories in social-ecological research and the need to move beyond descriptive ideas (such as ‘planetary boundaries’) to better understand the mechanisms that drive environmental change. An empirical analysis of national-level use of ecosystems suggests that global sustainability requires substantive structural changes to our current economic model.