an interactional interface?
How emotion is made and measured
Kirsten Boehner, Rogério DePaula, Paul Dourish, Phoebe Sengers
This is an important paper in the field of HCI and interaction design generally, and affective computing specifically. The paper puts forward “an interactional account of emotion and the role that it plays in action and practice.” It contrasts this with what the paper describes as an “informational model” of emotion, that the authors propose is the predominately used model in many systems designed to respond to and communicate human emotions.
The paper uses anthropological arguments to critique the cognitive psychology led arguments of, among others, Stuart Card (1983) and Don Norman (2003). Both Norman and Card’s models of human cognition describe processes that occurs ‘inside’ the human, hidden from others. Models for understanding emotion have been built on these cognitive foundations, treating emotion as something that is “objective, internal, private, and mechanistic”.
The authors believe this model “systematically ignores” emotion as a product of interactions between people in a cultural context. They reference anthropological research that describes emotion as a culturally grounded phenomena, and suggest that such a model of understanding affect as interaction has impact not only on how things may be designed, but also on how to evaluate these designs. They list a number of differences in an interactional approach:
“…instead of the system interpreting the emotional meaning of the input, the users interpret the emotional meaning of the output, and tune its output to support their readings.”
“The interactional approach:
recognizes affect as a social and cultural product
relies on and supports interpretive flexibility
avoids trying to formalize the unformalizable
supports an expanded range of interaction acts
focuses on people using systems to experience and understand emotions”
“The informational approach is based on a notion of emotion as transferable, communicable units. … The interactional approach instead sees emotion as constructed through interaction and expression.”
The paper addresses evaluation, and proposes that, at a theoretical level, the interactional approach to emotion questions “what evaluation is”, but that “a more practical alternative is to focus less on what evaluation is and more on what it does. Evaluation, we would argue, is one amongst a range of strategies of critique.” This, in particular, brings this approach more in line with what I understand as a design-led practice.
Perhaps most importantly, the paper uses well chosen examples to illustrate the differences between an informational and interactional approach. These examples, and the way they clearly communicate the thesis of this paper, are fitting testament to the constructionist origins of this thesis.
