The starting assumption of this thesis is that to fully understand legal practices – including legal reasoning – we need to get a grasp of the complex body of knowledge into which they are immersed. Legal studies have often assumed that legal knowledge can be reduced to the knowledge of legal rules. This research departs from this perspective and argues for an understanding of legal knowledge that includes the complex set of conceptual, procedural and affective considerations which shape legal practices in general, and legal reasoning in particular. Herein we argue that not only legal knowledge is wider than the knowledge of rules, but that there are also some aspects of legal practice that cannot be properly addressed by explicitly drafted legal rules. We purport to build such an account upon epistemologically-informed comparative legal perspectives and insights of the cognitive sciences, by way of discussing a particular factual problem. The case to be studied in this thesis is the apparent loss of certainty in Mexican legal practice, when legal professionals engage in precedent-based reasoning. The situation, which was first reported in 2006, has remained broadly unexplored, and by default has been reputed as a problem concerning the set of explicit rules that regulate the system of legal precedents in that national context. We argue that the situation cannot be fully comprehended and remedied if we exclusively focus on the dimension of legal rules, but that it would be better understood if we direct our attention to the deeper knowledge structures in which that practice is immersed. This thesis builds a case for a broadened approach to legal knowledge by unveiling the historically built knowledge structures in which the Mexican understanding of precedents is embedded. As we shall see, this particular framework has acted as a deterrent to precedent-based reasoning, as accounted by a set of theories of law and legal reasoning. By focusing on the several processes of legal change and the collateral epistemic revisions that Mexican legal professionals seem to be experiencing for the past decades, this thesis argues that changing deeply embedded knowledge structures is a difficult task that needs to be supported by revising the processes of knowledge construction, and most importantly legal education.