The Emergence of L2 Phonological Contrast in Perception: The Case of Korean Sibilant Fricatives

The perception of non-native speech sounds is heavily influenced by the acoustic cues that are relevant for differentiating members of a listener’s native (L1) phonological contrasts. Many studies of both (naïve) non-native and (not naïve) second language (L2) speech perception implicitly assume continuity in a listener’s habits of perceptual attention to acoustic cues both before and after L2 instruction begins. In this dissertation we argue that while the ability to attend to familiar acoustic cues may not change as a listener begins to learn an L2, whether and how the acoustic cues will be attended to is the result of an interaction between L1-influenced bottom-up habits of perceptual attention and top-down information about the L2. This top-down information about the L2 is totally unavailable to a non-native listener but cannot be avoided by an L2 learner, as it is typically introduced at the earliest stages of L2 instruction.

In this dissertation we look specifically at the emergence of the Korean /s^h/-/s*/ perceptual contrast in L1 Mandarin and Japanese learners of Korean. Our results show that while naïve L1 Mandarin listeners (but not L1 Japanese listeners) are sensitive to differences in aspiration when perceiving Korean /s^h/ and /s*/, L1 Mandarin learners of L2 Korean fail to demonstrate such sensitivity and perform as poorly as L1 Japanese listeners. Even after one year of intensive Korean instruction for four hours per day, L1 Mandarin listeners show very little improvement, if any at all. We reconcile these differences between non-native and L2 perception by appealing to top-down information present in the orthography, romanization, and loanword vocabulary of Korean that conflicts with L1-influenced habits of perceptual attention.