Embodied emotion: the role of bodily feedback in emotion comprehension as seen in language

The current project explores the nature of embodiment, specifically how affective associations of physical posture interact with emotion in language. Previous research has established that perceptual and motor information is encoded alongside semantic representation, and physical feedback is used to support cognition. Evidence suggests that emotion is also embodied. Muscle feedback from the face is used to support recognition of valence in language. Observational studies also show that open and slumped body posture are associated with positive and negative emotion, respectively. However, little research investigates experimentally induced posture. Additionally, research typically employs self-referent and autobiographical language with little consideration for experimental control over language stimuli. The current projects makes important steps toward developing and validating controlled linguistic materials for emotion research. Further, the project aims to extend previously observed facilitation from facial feedback to the full body. Three repeated measures experiments compared the effects of open and slumped posture on three cognitive measures: explicit recognition of valence from sentences, memory of affective details from sentences, and memory from text. Results indicate that positive, but not negative or neutral sentences are susceptible to the influence of posture. Interestingly, explicit processing of emotion, i.e., recognising valence, was inhibited by open posture, while implicit access during the memory task revealed the classic facilitation effects proposed by embodiment research. Body feedback did not affect comprehension and memory from full text. These results are discussed in relation to three main claims. First, embodiment of emotion is observable in sentence processing, although potentially limited to specific valence types. Second, explicit and implicit processes require different levels of activation, and implicit, automatic access that does not involve conscious awareness of the emotive content is more reflective of the neural priming effects proposed by embodiment models. Third, the linguistic and conceptual complexity of longer text formats preclude embodied effects due to increased confounds from visual imagery, metaphor and other processes that do involve perceptual and sensory information, but not internal bodily feedback, and are thus not susceptible to the posture manipulations employed here. Finally, implications for future research involve continued validation and refinement of the language materials used for emotion research, with additional consideration for concreteness and imagery ratings, as well as embodied properties on the word level. Research should also consider exploring the bidirectional interplay between language and posture further