This dissertation aims to investigate the be-insertion phenomenon in L2 English. In L2 production, be-forms are often inserted before thematic verbs, creating nontargetlike forms. There have been two main hypotheses concerning the function of such be-forms: The Topic Marker Hypothesis claims that the learners from topic-prominent L1s confuse subjects and topics and analyze be-forms as morphemes marking topics, while the T/AGR Morpheme Hypothesis argues that learners acquire suppletion (e.g. is/was) before affixal inflection (e.g. –ed) and overgeneralize be-forms as exclusive tense/agreement morphemes.

Focusing on the proposed hypotheses of be-forms in previous research, this study examined their explanatory adequacy in the case of Korean and Russian learners at different proficiency levels. Their oral production and grammaticality judgments suggested that regardless of the L1, be-forms had multiple functions in the early stages of interlanguage. Due to L1 transfer, however, there were relatively stronger relationships between be-forms and topic marking in the interlanguage of the Korean group and between be-forms and encoding agreement in that of the Russian group.

As proficiency increased, the functions of be-forms were concretized and the two learner groups followed different developmental paths. The Korean group associated be-forms with both Topic and T features, but they retained the function of past tense marking relatively longer. The Russian group associated be-forms only with the AGR feature and replaced the bare verbs following be-forms with the -ing form to match L2 input. These findings provide a clear picture of how and where Topic and T/AGR features are checked in early interlanguage, how interlanguage changes as proficiency increases, and ultimately, how the complicated interplay between L1 transfer and UG shapes the process of Second Language Acquisition.