Cultural Anthropology for Social Emotion Modeling: Principles of Application toward Diversified Social Signal Processing

Abstract

The practice of modeling social emotions has benefited from interdisciplinary engagements with other fields in the hard and human sciences; however, perspectives from cultural and social anthropology have been limited. This has at times resulted in the integration of emotion theories into emotion modeling that emphasize the universal communicability of social signals of emotion at the expense of accounting for cultural diversity evidenced in the ethnographic record. This paper outlines methods and findings of a collaborative effort between cultural anthropologists and engineers to create platforms for interdisciplinary communication and emotion modeling practices more sensitive to cultural diversity and better protected from risks of ethnic, racial, and ethnocentric bias. The paper presents five principles for applying anthropological perspectives to emotion modeling and ultimately argues for a consideration of design strategies for social signal processing based on recent ethnographic evidence of evolving human-robot relationships in Japan.