Modelling dynamic cultural variance in social face signals

人类是一种复杂的社会物种，经常参与复杂的动态社会互动。这种互动的一个强大工具是脸，它可以产生无数复杂的动态模式-即，面部表情（Jack & Schyns，2015）-以及许多细微的社交信息，包括情感（例如，杰克等人，2016）、认知状态（例如，Chen等人，2015），和人格特质（例如，Gill等人，2014）。鉴于它们在社交互动中的中心地位，该领域的一个长期目标是理解面部表情的“语言”-即，面部运动的哪些特定句法组合将有意义的社会信息传递给谁以及这是否以及如何在不同文化中变化（例如，达尔文，1872/1999;杰克，2013）。然而，到目前为止，人们对面部如何产生和传递基本而复杂的社会概念知之甚少，这些概念构成了社会互动并维护了更广泛的社会功能。这种知识限制反过来又阻碍了该领域内长期争论的解决，例如这些复杂的面部信号是否被普遍理解或特定于文化，以及它们如何映射到心理过程（例如，分类与维度感知）。这在很大程度上是由于对社会概念（精神状态、个性、情感）、文化的零散研究，以及传统上以西方为中心、理论驱动的方法和静态刺激（例如，Ekman &弗里森，1982），因此，忽略了一个可能的潜在代数，句法结构的社会面孔信号跨文化。杰克教授实验室最近的工作暗示了这样一种结构，表明面部表情随着时间的推移遵循一种潜在的、从粗到细的句法结构，早期发生的、文化上常见的面部运动传达了广泛的维度信息，晚期发生的、文化上特定的口音将信息细化为更精确的社会类别（杰克等人，2014; Jack等人，2016）。目标。我的项目借鉴了这些证据，旨在利用社会/文化心理学、3D动态计算机图形学、视觉科学心理物理学方法和数学心理学相结合的方法，推导出社会面部信号的第一个生成、代数和句法模型。具体来说，我将首先寻求使用Jack教授实验室开发的最先进的数据驱动方法，将面部表情的知识扩展到超越文化内部和跨文化的基本情感的社交信息。然后，我将正式检查这些面部表情模型，以提取它们的时间结构，识别特定的早期和晚期面部运动，并使用新颖的信息理论分析（例如，因斯等人，2017年）。最后，我将对这些面部表情模型进行跨文化比较，以识别跨文化的特定面部信号，即文化特定口音，并测试它们对跨文化交际的影响。因此，该项目将产生有史以来第一个对文化敏感的面部表情句法模型，该模型将精确地表征潜在的面部运动模式，其文化特定的口音和社会信息（例如，维度，类别），它们随着时间的推移传达，具有弥合该领域长期争论的潜力（例如，本质主义与建构主义;分类与维度），并形成了一个新的社会面孔感知理论框架的基础，统一了社会信息，面孔信号和文化的知识。

Humans are a sophisticated social species, frequently engaging in complex dynamic social interactions. A powerful tool for such interactions is the face, which can generate myriad complex dynamic patterns - i.e., facial expressions (Jack & Schyns, 2015) - and thus many nuanced social messages including emotions (e.g., Jack et al., 2016), cognitive states (e.g., Chen et al., 2015), and personality traits (e.g., Gill et al., 2014). Given their centrality in social interactions, a longstanding goal in the field has been to understand the "language" of facial expressions - i.e., which specific syntactical combinations of face movements transmit meaningful social messages to whom - and whether and how this varies across cultures (e.g., Darwin, 1872/1999; Jack, 2013). Yet, to date, little is known about how the face generates and transmits the fundamental and complex social concepts that structure social interactions and uphold broader societal functioning. Such knowledge limitations in turn impede resolution of longstanding debates within the field, such as whether these complex face signals are universally understood or culture-specific and how they map onto psychological processes (e.g., categorical versus dimensional perception). This is largely due to fragmented research on social concepts (mental states, personality, emotions), culture, and the traditional use of Western-centric, theory-driven approaches and static stimuli (e.g., Ekman & Friesen, 1982), which, consequently, has overlooked a possible latent algebraic, syntactical structure to social face signals across cultures. Recent work from Prof. Jack's laboratory hints at such a structure, suggesting that facial expressions follow a latent, coarse-to-fine syntactical structure over time with early-onset, culturally common face movements conveying broad dimensional information and later-onset, culture-specific accents refining the message into more precise social categories (Jack et al., 2014; Jack et al., 2016). Aims. My project draws from such evidence and aims to derive the first generative, algebraic, and syntactical model of social face signals using methods combining social/cultural psychology, 3D dynamic computer graphics, vision science psychophysical methods, and mathematical psychology. Specifically, I will seek to first broaden knowledge of facial expressions to social messages beyond basic emotions both within and across cultures using state-of-the-art, data-driven methodologies developed in Prof. Jack's laboratory. I will then formally examine these facial expression models to extract their temporal structuring, identify specific early- and late-onset face movements, and characterize the dimensional and categorical information they transmit over time using novel information-theoretic analyses (e.g., Ince et al., 2017). Lastly, I will conduct cross-cultural comparisons of these facial expression models to identify the specific face signals that are cross-cultural, those that are culture-specific accents, and test their impact on cross-cultural communication. This project will therefore produce the first ever culturally sensitive syntactical model of facial expressions that will precisely characterize latent facial movement patterns, their culture-specific accents, and the social information (e.g., dimensions, categories) they convey over time, with the potential to bridge longstanding debates in the field (e.g., essentialist versus constructivist; categorical versus dimensional) and form the basis of a new theoretical framework of social face perception that unifies knowledge on social messages, face signalling, and culture.