A challenging aspect of foreign language acquisition is learning new vocabulary. Research has demonstrated the benefits of motor action to vocabulary learning, though these findings are rarely translated into teaching and learning practices. In this work, we present a novel paradigm for vocabulary learning in a virtual reality environment, capitalizing on the prominent advantages of this technology as an 'embodied learning' tool. Hebrew-speaking adults explored a virtual room and learned the Finnish names of 40 daily objects. Every participant learned the new object names under three conditions: (1) Watch-only: watching the object without performing any movement (2) performing an irrelevant movement or (3) performing a manipulation movement that implied its habitual use. Participants completed a word-picture matching test after each training block, as well as additional word-picture matching and recall tests a week after the initial training session. The overall results demonstrate improved comprehension rates for items that were learned in the manipulation movement condition compared to the irrelevant movement condition, both in the initial learning session and 1 week following it. These initial findings suggest that VR technology may indeed hold the potential to offer a more authentic, multisensory and motor context to efficient foreign language learning.