Data from: Infants prefer the faces of strangers or mothers to morphed faces: an uncanny valley between social novelty and familiarity

The “uncanny valley” response is a phenomenon involving the elicitation of a negative feeling and subsequent avoidant behaviour in human adults and infants as a result of viewing very realistic humanlike robots or computer avatars. It is hypothesised that this uncanny feeling occurs because the realistic synthetic characters elicit the concept of “human” but fail to satisfy it. Such violations of our normal expectations regarding social signals generate a feeling of unease. This conflict-induced uncanny valley between mutually exclusive categories (human and synthetic agent) raises a new question: could an uncanny feeling be elicited by other mutually exclusive categories, such as familiarity and novelty? Given that infants prefer both familiarity and novelty in social objects, we address this question as well as the associated developmental profile. Using the morphing technique and a preferential-looking paradigm, we demonstrated uncanny valley responses of infants to faces of mothers (i.e., familiarity) and strangers (i.e., novelty). Furthermore, this effect strengthened with the infant’s age. We excluded the possibility that infants detect and avoid traces of morphing. This conclusion follows from our finding that the infants equally preferred strangers’ faces and the morphed faces of two strangers. These results indicate that an uncanny valley between familiarity and novelty may accentuate the categorical perception of familiar and novel objects.

恐怖谷（uncanny valley）效应指人类成人与婴儿在观察高度拟真的人形机器人或计算机虚拟形象时，会产生负面情绪并随之出现回避行为的现象。现有假说认为，此类恐怖谷感受的产生源于拟真合成角色触发了“人类”的认知概念，却又未能完全契合该概念。这类违背社会信号常规预期的情况会引发不安感。这种源于人类与合成智能体（synthetic agent）互斥范畴间的冲突所引发的恐怖谷效应，引出了一个新问题：其他互斥范畴（如熟悉性与新颖性）是否也能引发恐怖谷感受？鉴于婴儿在社交客体中同时偏好熟悉性与新颖性，本研究不仅探讨了这一问题，还分析了相关的发展特征。本研究采用变形技术（morphing technique）与优先注视范式（preferential-looking paradigm），证实了婴儿对母亲面孔（即熟悉性刺激）与陌生人面孔（即新颖性刺激）存在恐怖谷反应。此外，该效应随婴儿年龄增长而增强。我们排除了婴儿能够检测并回避变形痕迹这一可能性，该结论基于我们的实验发现：婴儿对陌生人面孔与两名陌生人的变形面孔的偏好程度无显著差异。本研究结果表明，熟悉性与新颖性之间的恐怖谷效应，可能会强化个体对熟悉与陌生客体的范畴化知觉。