Models of spoken word recognition in monolingual, native listeners account for the dynamics of lexical activation of intended words and phonologically similar “competitors,” in terms of continuous, cascaded processing. Here we explore how the dynamics of spoken word recognition differ for second language listeners. Groups of native Korean speakers and native English speakers listened to recordings of words in three conditions: phonological overlap at the beginnings of words (cohort), at the ends of words (rhyme), or without phonological overlap (unrelated), and used a computer mouse to select the matching stimulus from an array of two pictures. The results did not reveal any interactions between language background and performance. Nor were effects of similarity related to overall performance on independent tests of speech sound categorization or vocabulary. The results suggest that the cohort and rhyme effects are robust features of proficient second language spoken word recognition.