This chapter analyses the fundamental debate on agent rationality within an evolutionary vision of law, contrasting rational choice theory (RCT) with bounded rationality. It explores models like ‘homo economicus’, ‘homo sociologicus’, and ‘homo heuristicus’. RCT’s assumptions (e.g., perfect information) are challenged by limitations like indeterminacy and hyperrationality, and empirical counterexamples such as the sunk cost fallacy and Centipede Game. The chapter introduces Herbert Simon’s bounded rationality, where decisions are constrained by limited information, cognition, and time, leading to “satisficing”. Behavioural economics (Kahneman, Thaler, Sunstein) further reveals cognitive biases, advocating for ‘nudges’. In contrast, Gigerenzer’s ‘homo heuristicus’ posits ‘fast and frugal’ heuristics as “better than rational” solutions for efficient decision-making. Ultimately, the chapter proposes an evolutionary account of law integrating three levels: biological, cultural, and individual learning processes. This framework views legal rules and institutions as sophisticated survival strategies and emergent orders. It argues that the ‘legal order’ itself, an emergent product of millions of individual interactions across these evolutionary levels, constitutes the primary unit of study in law.