Drawing on ecological and dialogical perspectives on language and cognition, this exploratory case study examines how vocabulary learning occurs during a quest-play mediated in English between a Japanese undergraduate student and a native speaker of English. Understanding embodiment as coaction between the player-avatar and player-player relations (Zheng and Newgarden 2012; Zheng et al. 2012), as situative embodiment in a perceptually and narratively rich context (Barab et al. in "Sci Educ" 91:750-782, 2007), and as a dialogical achievement (Zheng and Newgarden 2012; Zheng et al. 2012), this research provides an alternative explanation of how players embodied in their avatars appropriated semiotic resources imbued in World of Warcraft (WOW). Two hours of co-quest play provided instances of vocabulary learning unique to the WOW environment and co-play. Through iterative multimodal analysis, vocabulary learning became salient as we analyzed both chat and avatar action data and provided a thick description and dynamic process of co-play. Using the eco-dialogical model, we display how language learning as appropriation of resources and as result of eco-dialogical embodiment.