Recently, it has been argued that the many works on policy learning constitute a stand-alone basis for understanding policy processes. In this study, we evaluate this claim through a bibliometric analysis of 588 publications on the topic in the Web of Science database, complemented by a literature review. We find that while the study of learning is supported by an active and growing research community, it has neither definitional clarity nor a shared vocabulary. And, further, its model of agency is both incomplete and inconsistent. As such, the subject remains more a metaphor than a framework of analysis, per se, and has little potential to advance epistemologically. Given this analysis, we argue that intellectual resources are better spent organising research on learning within existing frameworks rather than attempting to create a new stand-alone one that would contribute to the further splintering of an already fragmented field of study. [ABSTRACT FROM AUTHOR]