An eye-tracking experiment tested the hypothesis that listeners use within-word fine phonetic detail that systematically reflects morphological structure, when the phonemes are identical (dis in discolour (true prefix) vs. discover (pseudo prefix)) and when they differ (re-cover vs. recover). Spoken sentence pairs, identical up to at least the critical word (e.g. I’d be surprised if the boys discolour/discover it), were cross-spliced at the prefix-stem boundary to produce stimuli in which the critical syllable’s acoustics either matched or mismatched the sentence continuation. On each trial listeners heard one sentence, and selected one of two photographs depicting the pair. Matched and mismatched stimuli were heard in separate sessions, at least a week apart. Matched stimuli led to more looks to the target photograph overall and time-course analysis suggested this was true at the earliest moments. We also observed stronger effects for earlier trials and the effects tended to weaken over the course of the experiment. These results suggest that normal speech perception involves continuously monitoring phonetic detail, and, when it is systematically associated with meaning, using it to facilitate rapid understanding.

眼睛追踪实验测试的假设听众使用内字细语音细节，系统地反映形态结构，当音素是相同的（DIS在变色（真前缀）对发现，当他们不同（（伪前缀））重新-覆盖与恢复）。口语对，至少与关键词相同（例如，如果男孩们变色/发现它，我会感到惊讶)，在前缀-词干边界处交叉拼接以产生刺激，其中关键音节的声学匹配或不匹配句子延续。在每次审判中，听众听到一个句子，并选择描绘这对夫妇的两张照片中的一张。在不同的会话中听到匹配和不匹配的刺激，至少相隔一周。匹配的刺激导致对目标照片的整体观察更多，时间进程分析表明这在最早的时刻是正确的。我们还观察到早期试验的更强效果，并且在实验过程中效果趋于减弱。这些结果表明，正常的语音感知涉及持续监控语音细节，并且当它与意义系统地关联时，使用它来促进快速理解。