Numerous recent papers on the phonological adaptation of loanwords have discussed the claim by Sharon Peperkamp that this adaptation is essentially just cross-language perception, of the same kind as is generally operative in second language acquisition. This talk will compare the adaptations evident in an extensive collection of English loanwords into Korean to a corpus of perceptual responses to English productions by Korean students of English in Korea. The analysis selects 10 obstruents situated in four prosodic contexts, initial, final, and pre- and post-stress intervocalic positions, and attempts to predict adaptation maps from the perceptual maps for these segments. It finds a strong logistic relationship between them, indicating a process of loanword adaptation as a regularization of the cross-language perception patterns. This conclusion is also supported by differences in the maps across the prosodic positions, wherein loanword differences are correlated with perceptual differences. Further, lexical variants are also more common in cases where consonants that do not have a very robust perceptual map, suggesting that the perceptual maps constrain the process of adaptation. All this indicates a pervasive and powerful role of perceptual assimilation in lexical adaptation. However, the data also exhibit strikingly exceptional cases that apparently indicate a number of aspects of adaption that make it diverge from perceptual mapping. These show especially that adaptation is a historical process with a long time-window, and that adaptations are directly subject to explicit socio-cultural effects. Thus, loanword adaptation in the Korean-English case, though largely indicative of a perceptual base, is much more than just synchronic perception.