Emotions in archetypal media content

Abstract

Emotion is an intriguing and mysterious psychological phenomenon. While everyone seems to know what it is, researchers have not yet come to consensus on its definition, and many questions still remain unanswered. While the nature of emotion is yet to discover, the design community has noticed is importance, and poses the challenge of how emotion could inform design. We see the necessity to follow the state of the art in psychology and initiate the undertaking by exploring the emotional qualities in various types of media content. The first part of this thesis aims at constructing a theoretical framework. Recent years have seen empirical studies suggest that emotion could be unconscious. While this is to be further justified, scientists are motivated to reconsider current theories of emotion to account for this phenomenon. In light of this, we integrate these studies about unconscious emotion into our literature review. An overview from theory to practice is illustrated to provide a reference for viewing the current states in application domains, such as affective computing and emotional design. This review offers a holistic understanding about emotion from various perspectives, which allow us to look for new directions in future studies. Based on our review, we see a promising direction by applying psychoanalysis methods to analyze the media content as affective stimuli, and these stimuli can be evaluated by using quantitative measures to investigate the connection between the content and the corresponding emotions. The analysis on the media content is based on a psychoanalysis theorythe theory of archetypes proposed by Carl Jung. He argues that there exists a universal pattern in humans' unconscious thoughts, which can be manifested as symbolic content in various forms of narratives, such as myth and fairy tales. Today, this archetypal symbolic content can be seen in modern media, particularly in movies. By applying the Jungian approach, we analyzed the symbolic meaning in movie scenes and edit these feature scenes into a collection of archetypal media content, which serve as the experimental materials for later explorations. In the second part of this thesis, we present three experimental studies that aim at determining if archetypal media content can be differentiated based on emotional responses. We adopted the psychoanalytical approach described earlier to collect feature scenes in movies as archetypal media content. Meanwhile, affective stimuli of explicit emotions are also included as benchmarks for comparison, such as sadness and joy. Self-reports and physiological signals are both adopted for measuring emotional responses. These three studies follow similar experimental design: presenting stimuli and measuring emotion concurrently. The results of these studies confirm that emotions induced by archetypal content are different from explicit emotions, and the statistical analysis further indicate that the predictive model obtained from physiological signals outperforms the model generated from self-reports while viewing archetypal media content. These results, however, are opposite to the results gained from affective stimuli of explicit emotions, leading us to the conclusion that archetypal media content might induce unconscious emotions, and physiological signals are more effective than self-reportsfor recognizing emotions induced by archetypal media content.