The contribution of specific movement components (such as Ekman and Friesen's action units) to the interpretation of meaning in facial expressions of emotion has yet to be explored using well-controlled stimuli that systematically vary the components and observe the effect of such manipulations on judgments of emotional meaning. It is argued that by identifying salient stimulus components and relating them to semantic labels, the stability of meaning across contexts can be better understood and a covering explanation for divergent theories devised. This study will produce data that should be very useful to other researchers regardless of theoretical orientation because it explicitly focuses upon the sources of disagreement across theories. The long-term goal is to create a process model for perception of meaning in facial expressions. The proposed research will investigate the relationship between semantic labels and salient component movements by addressing the following questions: (a) are emotional expressions perceived in terms of additive combinations of component facial movements or in terms of visual patterns formed from the conjunction of such component movements; (b) how do patterns and component movements contribute to decoding of emotional meaning, identification and labeling of expressions using emotion-terms; (c) is a Stroop-like response observed in pilot studies for certain patterns related to semantic interference or is it related to emotional response; and (d) if emotional expressions are perceived in terms of patterns, how do such patterns compare to Ekman and Friesen's basic expressions? The proposed research brings new empirical methodologies to this research issue and applies new formal models for the analysis of data (the Batchelder-Bershad numerical rating system and the Batchelder-Romney consensus model). Proposed experiments employ designs that are commonly used in cognitive science including psychophysical scaling paradigms and related data analysis techniques (such as multidimensional scaling).
