This study tests the hypothesis that imitators of different native languages imitate the same targets in distinct ways predicted by their native phonology, by investigating the role of a non-contrastive phonetic property in spontaneous imitation of English voiceless stops by English monolingual and Seoul Korean-English bilingual imitators. The primarily contrastive phonetic property for English voiceless stops is voice onset time (VOT), with the fundamental frequency (f0) of the post-stop vowel being non-contrastive but still informative for the voicing contrast. On the other hand, in Seoul Korean, stop VOT is a non-primary cue, but it is necessary to maintain the full three-way laryngeal contrast in the language. Post-stop f0 is the primary cue for the Seoul Korean aspirated stops. Seoul Korean speakers have been reported to imitate aspirated stops with longer VOT by raising their post-stop f0 (Kwon, 2019). In this study, English monolingual speakers and Seoul Korean-English bilingual speakers heard and shadowed model speech containing English voiceless stops manipulated by either raising post-stop f0 or lengthening VOT. Their imitation was assessed with two acoustic measurements, stop VOT and post-onset f0, of the voiceless stops, before and after the imitators heard the model speech with the two manipulations. A separate discrimination test confirmed that both manipulations were reliably perceived by both the monolingual and the bilingual imitators. English monolingual speakers' imitation data suggest that their shadowing productions reflect the phonological significance of the two phonetic properties, and only the imitative changes induced by a contrastive cue last beyond the immediate shadowing targets. In addition, Seoul Korean-English bilingual speakers, when performing the spontaneous imitation tasks in English, do not draw on their native (Seoul Korean) phonology. Implications of these findings on the role of phonology in the spontaneous imitation of bilingual and monolingual speakers are discussed.

本研究通过调查非对比语音特性在英语单语和首尔韩语-英语自发模仿英语清音塞音中的作用，检验不同母语的模仿者以他们母语音韵预测的不同方式模仿相同目标的假设双语模仿者。英语清塞音的主要对比语音特性是语音起始时间 (VOT)，塞音后元音的基频 (f0) 没有对比，但仍能提供语音对比信息。另一方面，在首尔韩语中，停止 VOT 是非主要提示，但需要在语言中保持完整的三向喉对比。停站后 f0 是首尔韩国吸气站的主要提示。据报道，首尔韩语使用者通过提高他们的停顿后 f0 来模仿具有更长 VOT 的吸气停顿（Kwon，2019）。在这项研究中，英语单语者和首尔韩英双语者听到并隐藏了包含英语清音停止的模型语音，这些语音通过提高后停止 f0 或延长 VOT 来操纵。他们的模仿是通过两个声学测量来评估的，即停顿 VOT 和发声后 f0，在模仿者通过两种操作听到模型语音之前和之后的清音停顿。单独的辨别测试证实，单语和双语模仿者都能可靠地感知这两种操作。英语单语者的模仿数据表明，他们的影子产生反映了两种语音特性的语音意义，只有由对比线索引起的模仿变化才能持续超越直接的阴影目标。此外，首尔韩英双语者在用英语进行自发模仿任务时，不会借鉴他们的母语（首尔韩语）音韵。讨论了这些发现对语音学在双语和单语者自发模仿中的作用的影响。