Crossmodal correspondences refer to the fact that certain domains of features in different sensory modalities are associated with each other. Here, we investigated the crossmodal correspondences between speech sounds and visual shapes. Specifically, we tested whether the classification dimensions of English vowels (front–central–back) and consonants (voiced–voiceless, sonorant–obstruent, and stop–continuant) correspond to visual shapes along a bipolar rounded–angular dimension. We adapted eighteen meaningless pseudowords from a previous study that corresponded to either the round or the sharp concept. On each trial, the participants heard one of the pseudowords and saw a rounded shape and an angular shape presented side-by-side on the monitor. Participants judged which shape provided a better match to the spoken pseudoword. A logistic regression was conducted in order to elucidate the effectiveness of classification dimensions of phonemes when predicting variations in the sound–shape matchings. The results demonstrated that the sound–shape matchings were predictable using front–central–back dimensions of vowels, and voiced–voiceless and stop–continuant dimensions of consonants. Hence, we verified that sound–shape matching is underpinned by contrasting dimensions in both vowels and consonants, therefore demonstrating crossmodal correspondences at the phonetic level.

中文翻译：

---

看到声音：元音和辅音在跨模态通信中的作用

交叉模式对应是指不同感官模式中的某些特征域相互关联的事实。在这里，我们研究了语音和视觉形状之间的跨模态对应关系。具体来说，我们测试了英语元音（前-中-后）和辅音（浊-清音、响音-阻塞和停止-连续）的分类维度是否对应于沿双极圆角维度的视觉形状。我们从先前的研究中改编了 18 个无意义的伪词，这些伪词对应于圆形或尖锐的概念。在每次试验中，参与者都会听到一个伪词，并在监视器上看到一个圆形和一个角形并排呈现。参与者判断哪种形状与口语伪词更匹配。进行逻辑回归是为了阐明在预测声音-形状匹配变化时音素分类维度的有效性。结果表明，使用元音的前-中-后维度，以及辅音的浊-清音和塞音-连续维度，声音-形状匹配是可预测的。因此，我们验证了声形匹配是通过元音和辅音的对比维度来支持的，因此在语音层面证明了跨模态对应。和浊 - 清音和停止 - 辅音的连续维度。因此，我们验证了声形匹配是通过元音和辅音的对比维度来支持的，因此在语音层面证明了跨模态对应。和浊 - 清音和停止 - 辅音的连续维度。因此，我们验证了声形匹配是通过元音和辅音的对比维度来支持的，因此在语音层面证明了跨模态对应。