Learning of tacit knowledge in an interorganizational context is a process fraught with risks for the collaborating partners. Two conflicting perspectives on what motivates partners to cooperatively share tacit knowledge with each other emerge from extant literature: one based on calculative considerations and the other on trust-based considerations. This paper aims to show that the two perspectives can be reconciled if the unique learning-related roles of boundary spanners at the corporate and operating levels are taken into account. Operating-level boundary spanners are the primary agents of tacit knowledge learning across organizational borders, and we argue that trust is the primary determinant of knowledge sharing at that level. In contrast, the corporate-level boundary spanners shape the structures and systems of the collaboration and thus affect the extent of sharing that can take place between operating-level boundary spanners. In this role, we propose, their learning behavior is predominantly driven by calculative considerations of potential costs and benefits of knowledge sharing.