Abundant evidence across languages, structures, proficiencies, and modalities shows that properties of first languages influence performance in second languages. This paper presents an alternative perspective on the interaction between established and emerging languages within second language speakers by arguing that an L2 can influence an L1, even at relatively low proficiency levels. Analyses of the gesture viewpoint employed in English and Japanese descriptions of motion events revealed systematic between-language and within-language differences. Monolingual Japanese speakers used significantly more Character Viewpoint than monolingual English speakers, who predominantly employed Observer Viewpoint. In their L1 and their L2, however, native Japanese speakers with intermediate knowledge of English patterned more like the monolingual English speakers than their monolingual Japanese counterparts. After controlling for effects of cultural exposure, these results offer valuable insights into both the nature of cross-linguistic interactions within individuals and potential factors underlying gesture viewpoint.