This study explores interlanguage representation of English VP-ellipsis. Previous studies have reported that L2 learners of English (L1 Arabic and Chinese) rejected both the sentences where an object is elided, such as *Tom sent Jill a letter, and Mary sent __ too , and the sentences involving the elided uninterpretable perfect feature on -en, such as Tom sent Jill a letter, and Mary has __ too, even though the latter is acceptable in native speakers' grammar (e.g., Hawkins, 2012). The results are compatible with the Interpretability Hypothesis (Tsimpli & Dimitrakopoulou, 2007), which claims that L2 learners have difficulty acquiring uninterpretable features. This study focuses on whether Japanese learners of English (JLEs) behave similarly to L2 learners in previous studies. Results suggest that their knowledge of VP-ellipsis relies crucially on the presence or absence of the parallelism of features, rather than on the interpretable or uninterpretable feature distinction.