Allostery describes the process whereby ligand binding at one site on a protein transmits an effect to another distal site. Ever since its discovery in 1961, allostery has remained an important topic within structural biology because of its role in cell regulation. However, the concept has changed drastically since Jacques Monod and his colleagues first characterized it. This paper aims to recount the conceptual evolution of allostery over the past 50 years. I argue that the concept has evolved in three stages. First, from the late 1950s to the mid-1960s, allostery was a population-level phenomenon, closely tied to biochemistry and heterodox enzyme kinetics. Then, starting in the mid-1960s and extending to the mid-2000s, allostery became a feature of individual protein molecules with special structural and conformational properties. The final stage in this conceptual evolution, which introduced the ‘ensemble nature of allostery’, began in the mid-2000s (Motlagh et al. 2014). Creager and Gaudilliere (1996) have offered an explanation for the first historical transition based on the complex interplay between theory and experiment, and it is the primary aim of this paper to characterize the second transition and offer an explanation for it. I aim to show that the driving forces behind the conceptual shift from the structural-mechanistic view to the ensemble view were twofold: (1) the mounting body of anomalies that could not be explained by any of the structural-mechanistic models of allostery and (2) the increasing recognition that protein dynamics play an important role in protein structure and function.