Beyond misperception: contextual effects in L2 segmental adaptation of interdental fricatives

Abstract This study examines how syllable position and voicing influence the adaptation of English interdental fricatives /θ/ and /ð/ by Taiwan Mandarin speakers learning English. Using a within-subjects design, it compares responses from an imitation-based production task and an AX similarity rating task. Production results show a positional asymmetry: the voicing contrast between /θ/ and /ð/ is robust in onsets but reduced in codas, where the center of gravity values of /ð/ shift toward the range observed for /θ/. Frication duration was longer in codas and for voiceless fricatives, although these effects were additive rather than interactive. Perceptual ratings also varied by syllable position: similarity to voiceless comparison consonants was higher in codas than in onsets, whereas similarity to voiced comparison consonants showed a smaller, marginal decrease. Overall, the interdental contrast was more robust in onsets, whereas coda /ð/ showed reduced perceptual differentiation and a more voiceless acoustic pattern. The current results suggest that L2 segmental adaptation may involve gradient, context-dependent phonetic differentiation rather than categorical replacement by a single L1 category.