Current approaches to the design and delivery of instruction in virtual reality learning environments (VRLEs) draw heavily from traditional instructional strategies and design practices. This is problematic given that these strategies and practices were developed for learning contexts lacking the dynamic nature and capabilities of technology-rich, immersive learning environments. This directly affects the instructional efficacy of VRLEs by creating a dichotomy between the learning interface, which emphasizes knowledge as object, and the learning environment, which can emphasize knowledge as action. Drawing from theory and research in the cognitive sciences on embodied and enactive cognition, we present an instructional strategy that addresses this dichotomy by incorporating techniques and design practices that are better aligned with the learning dynamics provided by VRLEs.