This paper investigated the "be"-insertion phenomenon in L2 English. L2 learners often insert "be"-forms before thematic verbs, creating nontargetlike forms (e.g. "She is love ice cream"). Based on L2 data from learners of topic-prominent L1s, a group of researchers have claimed that such "be"-forms are topic markers transferred from the L1s. As L1 transfer cannot be supported without comparing different L2 groups, however, this study examined the explanatory adequacy of the Topic Marker Hypothesis by comparing the Korean and Russian EFL learners at different proficiency levels. Their oral production and grammaticality judgment suggests that regardless of the L1, "be"-forms could mark topics in the early stages of interlanguage, supporting full access to UG. Due to L1 transfer, however, "be"-insertion by the Korean group was more relevant to topic marking while that of the Russian group was more relevant to encoding agreement. These findings show complicated interplay between L1 transfer and UG.