The present study proposes an L1 grammar-driven loanword-adaptation model with three intermediate steps — L1 perception, L1 lexical representations and L1 phonology — between L2 acoustic output and L1 output by examining how the distinctive features, syllable structure constraints and structural restrictions of one’s native language steer speakers in their search to replace foreign sounds with native sounds. The main source of data in support of this model comes from differences between the Korean adaptations of English and French voicing contrasts on the basis of a recent survey of English and French loans in the year 2011. In word-initial position, for example, English voiceless plosives are borrowed as aspirated plosives, while French voiceless plosives are borrowed as either aspirated or fortis plosives in free variation. Considering the data examined here, it is suggested that the different Korean adaptations of English and French voicing contrasts in plosives are based on Korean speakers’ perception of redundant phonetic variants in the donor languages (L2) and that this perception is conditioned by the acoustic cues to the laryngeal features [±spread glottis] and [±tense] of Korean, the host language (L1). In contrast to some current models, it shows that the distinctive feature composition of L1 segments plays an important role in loanword adaptations. It is also suggested that not only L1 laryngeal features but also L1 syllable structure constraints and lexical restrictions influence L1 perception of the L2 voicing contrasts in word-final postvocalic plosives and that variation in vowel insertion after the plosives in our 2011 data collection is motivated by L1 phonology in both English and French loans. Variation in vowel insertion after English and French word-internal preconsonantal coda plosives is also affected by the native phonology in the 2011 data, no matter whether the plosives are released, as in French, or unreleased, as in English.