This study explored the effects of Caucasian faces and Chinese faces on native speakers of Mandarin-Chinese acquiring emotional connotations of English L2 words. Participants were presented with English pseudowords repeatedly paired with either Caucasian faces or Chinese faces showing the emotions of disgust and sadness, as well as neutrality as the control baseline. Participants’ acquisition was evaluated through both within-modality (i.e., testing participants with new sets of faces) and cross-modality (i.e., testing participants with a set of sentences expressing the emotions acquired) generalization tests. Results of the two generalization tests suggested that participants in the Caucasian-face Group acquired sad connotations better than their counterparts in the Chinese-face Group, which is reflected through their higher accuracy when matching learned L2 words with sad L2 sentences. We thus conclude that foreign faces showed a certain advantage when it comes to acquiring specific emotional connotations for L2 words with which their socio-identity is congruent.