Authoritarian stability is a major concern of contemporary comparative politics. Scholars have long debated what makes regimes stable, arguing for the importance of state power (Slater 2010), legislatures (Gandhi 2008), political parties (Magaloni 2008, Svolik 2014), international support (Levitsky and Way 2010), new technologies of rule (Heydemann 2007) and rentierism. Tenacious authoritarian systems in the Middle East loom large in these discussions, whether as outliers or as paradigmatic cases. Yet the concept at the center of all this scholarly attention, "stability," remains worryingly underconceptualized. Indeed, the primary empirical measure of regime stability in political science is still the number of years a regime lasts. Stability-as-duration fails us in at least three ways: (1) it cannot distinguish between regimes whose tenures are uneventful and those that barely hang on to power; (2) it elides shifts and policy changes that take place within a single regime (i.e., it assumes that stability is stasis); and (3) it leaves us unable to define what makes stability appealing—surely political actors who speak longingly of "istiqrar" do not only mean that they want the current regime to last as long as possible. This paper offers a more meaningful and illuminating concept of stability that can better anchor empirical research. Thinking through academic and colloquial diagnoses of stability over the nine years of the long Arab Spring, I suggest that stability inheres in state-society units rather than in regimes. A stable system is one in which the kinds of challenges most likely to emanate from society are challenges with which the regime has historical experience and to which it has adapted. This is a coevolutionary model of stability: patterns of social life, economic shifts, and other stressors (often originating from state policy) produce certain kinds of threats to regimes; regimes, for their part, evolve over time to respond to some threats rather than others. I demonstrate the analytical leverage of my approach through a case study of Morocco in 2011. Despite dramatic, sustained mobilization, Morocco's "Arab Spring" protests were (and still are) regularly dismissed as not posing any real danger to regime stability. Indeed, observers often remark that Morocco "didn't have an Arab Spring," though it witnessed protests both larger and more long-lasting than many breathlessly-covered demonstrations elsewhere. Drawing on fieldwork and archival research, I explore this seeming disconnect, locating its justification in patterns of state-society interaction established long before February 20.