That Sognal Feeling: Emotion and Interaction Design from Smartphones to the "Anxious Seat"

Charles Luke Stark

This dissertation examines how computational interaction design has been influenced by theories of emotion drawn from the psychological sciences, and argues that the contemporary field of interaction design would be impossible without the developments in psychology that allowed human emotions to be understood as orderable and classifiable. Interaction design, or the process by which digital media are created and modified for human use, has grappled with theories of human emotion since its inception in the early 1980s. The project examines how the longer histories of psychology and psychiatry have changed conceptualizations of emotion in relation to cognition and behavior, and how shifts in these theories have shaped the development of a burgeoning array of digital tools for tracking and managing human emotions. Examining the continuum of humans and machines desired and configured by individuals throughout this history, the project explores how these subject identities, both imagined and made, have reflected broader changes in the exercise of social power and authority. The research draws on materials from several archives, including newspaper reports; the published works and archival materials of psychologists and computer science researchers; materials from the West Coast Computer Faire tradeshow in the late 1970s; interviews with designers; and psychological texts and textbooks. Alongside a design assessment of smartphone apps for mood tracking grounded in Values in Design (VID) scholarship, the project deploys historical, philosophical, and qualitative methods, including close reading and discursive and thematic analysis. The key mechanism for understanding emotion's role in digital media design is the drive to make human feelings both technically ordinal and scalable. Through these conceptual mechanisms, human feelings have become increasingly classifiable not only horizontally as different categorical types, but also hierarchically in ways that differentiate and assign value to the emotions and moods of individuals in relation to a larger mass of data. Accomplished through both natural and symbolic language, these mechanisms combine qualitative and quantitative modes of classification, enabling sociotechnical phenomena ranging from personal applications for digital mood tracking to the analysis of emotional "Big Data" by social media platforms.