Pages: pp. 1-10
Emotions profoundly shape behavior and lend meaning and direction to human existence. Even during the “age of reason,” when logic and dispassionate science reigned supreme, David Hume acknowledged that “reason is, and ought to be, only the slave of the passions” (Hume, 1740, p. 295). Around the same time, Enlightenment thinkers enshrined in the US Declaration of Independence the “unalienable right” to pursue happiness. And today, people passionately pursue this right through technology. The human race spends billions of dollars and hours each year, stroking their emotions in computer games, finding love in virtual worlds, “flaming” in blogs and e-mail, and cursing bad interfaces. Perhaps more ominously, technology allows the emotions of individuals to magnify into a force that can shape a society, as demonstrated when international markets rise and fall depending on the weather (Hirshleifer and Shumway, 2003).
Affective Computing is the field of study concerned with understanding, recognizing, and utilizing human emotions and other affective phenomena in the design of technological systems. Research in the area is motivated by the fact that emotion pervades human life—emotions motivate and shape our individual thoughts and social behavior, they promote social bonds between people and between people and artifacts, and emotional cues play an important role in communicating attitudes and intentions to other social actors (be they human or computer). Technology is less acceptable if it disturbs our emotions; more efficient if it engages them productively; more attractive if it appeals to them; and often it is developed with the sole purpose of enabling us to experience them (as with entertainment technology). Since the coining of the term by Picard in 1995 (and see her personal perspective on the development of the field within this issue), affective computing has emerged as a cohesive interdisciplinary field with its own international conference (the International Conference on Affective Computing and Intelligent Interaction) and professional society (the HUMAINE Association), and today, its own journal.
The IEEE Transactions on Affective Computing is intended to be a cross-disciplinary journal aimed at disseminating results of research on the design of systems that can recognize, interpret, and simulate human emotions and related affective phenomena. The journal will publish original research on the principles and theories explaining why and how affective factors condition interaction between humans and technology, on how affective sensing and simulation techniques can inform our understanding of human affective processes, and on the design, implementation, and evaluation of systems that carefully consider affect among the factors that influence their usability.
Readers will see that this first issue emphasizes an interdisciplinary perspective on the field and, for me, this is an essential characteristic of affective computing. Emotion has been studied as a science at least since Aristotle and an enormous body of theoretical and empirical work exits across a wide range of disciplines, including neuroscience, ethology, psychology, anthropology, sociology, economics, art, and literature. Computer scientists and engineers are newcomers to the game, and bring new perspectives and new tools to the challenge of recognizing, understanding, and shaping human emotions: The ability to automatically acquire enormous datas ets of human behavior will revolutionize how we understand, model and shape human behavior (see Calvo, this issue); Technology demands rigor and attempts to “computationalize” an emotion theory can highlight is vagueness and inconsistent assumptions (see Gratch, Marsella, and Petta, 2009); Further, being outsiders to the existing body of emotion research, technologists can also bring fresh new perspectives and can avoid the pitfalls of the sometimes sterile theoretical disputes that can grip any mature science (see Kuhn, 1962). That said, technologists ignore well-established findings, theories, and methodologies at their peril. For one familiar with social science research on emotion, papers in affective computing can sometimes seem naive: ignoring important distinctions, rediscovering long-documented phenomena, or employing poor experimental design. To be successful as a field, affective computing must bring new ideas and technology, but not recapitulate the common confusions and missteps that other fields spent considerable effort to overcome.
Several features of the journal reinforce my interdisciplinary perspective on the field. The editorial board draws from a wide range of scientific expertise and authors that submit to the journal can expect to be tested on their interdisciplinary knowledge. Papers will often be reviewed by reviewers with different disciplinary perspectives and the review process should be seen as a valuable tool for disseminating knowledge from other perspectives on the phenomena of emotion. Occasionally, as in this first issue, I will select an article that facilitates discussion across disciplines and invite commentary articles that highlight differences of opinion, theory, or practice across fields.
As the inaugural issue of the journal, this volume contains two appetizers before we get to the meat of the field. We begin with a personal reflection by the founding figure on the field. Rosalind Picard of the MIT Media Lab. Roz first coined the term “affective computing” 15 years ago at a time when the topic of emotion had a dim reputation within the computational sciences. Her commentary, “From Laughter to IEEE,“ recounts these early challenges and the fellow visionaries who have led us to this point. We next introduce our impressive list of associate editors. Collectively, they bring a broad range of expertise to the field spanning neuroscience, cognitive science, psychology, robotics, linguistics, human-computer interaction, the learning sciences, health, multimedia, design, and engineering.
The scientific articles begin with a conversation on affect detection. Rafael Calvo and Sidney D’Mello begin with a review article of models, methods, and applications of affect detection. This is followed by two commentary articles from prominent emotion psychologists. Arvid Kappas emphasizes some of the challenges and pitfalls that can confront techniques for automatically recognizing affect. Rainer Reisenzein offers a broader perspective on the topic, arguing that affect detection is best seen as a special case of, and should be performed within the context of a larger effort on “mental state detection.”
Rounding out the issue, Ptaszynski, Maciejewski, Dybala, Rzepka, and Araki present work on inferring the emotional content of computer-mediated conversations, which often emphasize obscure symbols such as OMG, ;-), and :-*. Finally, Bickmore, Fernando, Ring, and Schulman illustrate the strong effect that human touch has on our emotional state and subsequent decisions, and present a robot that yields the same beneficial outcomes.
I close this introduction with a few personal words. My own introduction to this field began 13 years ago when, somewhat by accident, Paul Rosenbloom gave me the opportunity to explore the role of emotion with Soar, a cognitive architecture he had co-developed with John Laird and Allen Newell. Since then, I’ve come to appreciate the strange and wonderful ways that emotion colors our thinking and our interactions with technology. And as I’ve grown into more of a leadership role within the community, I also appreciate the importance of sharing this knowledge and sense of wonderment with the next generation of affective computing researchers. The launch of an IEEE journal on this topic is an important milestone. It definitively signals that the topic of emotion has established itself as a serious domain of discourse within the engineering sciences. It is an achievement only made possible through considerable personal investment from many individuals in the field. But a journal is just a vehicle. Now it is up to you to drive this vehicle in whatever way your wonderment for emotion and technology may take you.
This journal would not have been possible without tireless work of several individuals, only a few of whom are recognized in the steering committee and editorial board of the journal. We owe Roz Picard a considerable debt for giving a name to the field and tirelessly advancing the community through research and service. Efforts to launch a journal were started at the request of the membership of the HUMAINE Association. I became involved as an officer of this society and through the prodding of Roddy Cowie, and I leaned heavily on the support of Roddy, Maja Pantic, Björn Schuller, and Jianhua Tao. The initial proposal to the IEEE leveraged on an earlier effort to create a journal, spearheaded by Fiorella de Rosis and Roz. Within the IEEE I received considerable support and advice from Sorel Reisman and Alicia Stickley. And ongoing operations are only made possible through the efforts of the IEEE and ScholarOne staff, including Joyce Arnold, Kathy Santa Maria, and Kristen Anderson.
J. Gratch, S. Marsella, and P. Petta, ”Modeling the Antecedents and Consequences of Emotion,” J. Cognitive Systems Research, vol. 10, no. 1, pp 1-5, 2009.
D. Hirshleifer and T. Shumway, “Good Day Aunshine: Stock Returns and the Weather,” J. Finance, vol. 58, pp. 1009-1032, 2003.
D. Hume, A Treatise of Human Nature. Oxford Univ. Press, 1740 (1967 ed.).
T.S. Kuhn, The Structure of Scientific Revolutions. Univ. of Chicago Press, 1962.