Despite deserving a place amongst the historic milestones of the philosophy of disciplines, the system of the sciences put forward by Whewell has so far received little interest. Yet his ideas had a significant impact on the researches subsequently carried out on the topic, exerting in particular a decisive influence on Peirce and Spencer. The present paper aims to display the innovatory nature of the philosophical foundations of the Whewellian classification of the sciences. In this respect, we will argue that the most striking feature of his disciplinary system lies in a heuristic categorization of the sciences according to their "methods of discovery". This represents a double departure from both the Aristotelian and the Baconian disciplinary paradigms, which are instead underpinned by ontological and epistemological criteria, respectively. Next, we will explore the pivotal role of Whewell's classification of the sciences for his overall project of a philosophy of scientific discovery.