Mutual visual attention between the current and next speaker were analyzed from the viewpoints of grounding in communication and linguistic proficiency. Each utterance was categorized according to the grounding acts in the dialogue [21], and the gazing activities of the current and next speakers were compared between native (L1) and second language (L2) conversations. A correlation analysis showed that mutual visual attention was less prominent in utterances that ground the previous utterances without adding new information in L2, but not necessarily so in L1 conversations.