Comment: Growing a Multilevel Science of Emotion

Rogers et al. (2014) make a compelling case for the use of affect control theory as an intellectual platform from which to reach across disciplines and connect various theories of emotion. Three interrelated claims underlie their argument. The first, and most explicitly stated, is that satisfactory understanding of emotion will come only with satisfactory understanding of social process. This assertion seems not to require reinforcement. It does, however, carry an important implication—that theories of emotion should at a minimum be compatible with empirically supported theories of social life. Affect control theory offers a generative theory of action at the interpersonal level, making specific and testable predictions about social behaviors and emotions during face-to-face interaction. The theory does have a substantial amount of empirical support for its predictions about interpersonal behavior and emotions. Moreover, affect control theory already links directly to theories at other levels of analysis, including two new closely related theories—the affect control theory of self and the affect control theory of institutions (MacKinnon & Heise, 2010). These linkages offer a model of how we can develop theories at different levels of analysis which are compatible with one another.

The second claim made by Rogers et al. (2014) is that our investigation of emotion takes place at multiple levels of analysis, so our metatheoretical approaches need to be multilevel as well. Admittedly, it is fairly rare for a given study, or even program of research, to be multilevel. However, since a vast body of scientific knowledge demonstrates that emotions have mechanisms and manifestations at levels of analysis both higher and lower than the level of the person, it is critical that we employ theories at these levels that can sensibly interact with each other when necessary. As an illustration of this, the authors already offer a compelling description of how we can link affect control theory to theoretical accounts of emotion at the neural level. They also note the prevalence of the three dimensions of meaning (evaluation, potency, and activity) across a wide array of research on emotion at different levels. In addition to linkages at the conceptual level, practical resources exist that can link research paradigms across different theoretical and methodological traditions. These include numerous dictionaries of words normed in evaluation, potency, and activity in various languages ¹ as well as a vast library of words, sounds, and images normed in these dimensions ² and available for research purposes. This shared infrastructure could facilitate research at more microlevels of analysis that could meaningfully connect to research at the social and interpersonal levels of analysis.

The third claim, implicit both in the proposal made by Rogers et al. (2014) as well as in my aforementioned comments, is that there is payoff when different streams of intellectual endeavors interact with one another around a common domain of interest. Rogers et al. propose affect control theory as a potential conceptual and methodological hub in a network of theories that could connect our understandings of affective processes occurring in the brain, interpersonally, in larger social structures, and within cultural institutions. I wholeheartedly agree. What I support even more strongly, however, is the need for such a hub—or perhaps for multiple such hubs. In an article describing an approach for modularizing theories of justice, Markovsky et al. (2008) note that “Theories within the fabric of science are intertwined via threads that link intellectually adjacent theories” (p. 348). When systematic cumulation of knowledge remains contained within insular, noninteracting, theoretical traditions, it hampers our ability to see this larger fabric of knowledge. Theories can only interact through points of commonality—shared terms, principles. Scholars of emotion from a broad array of disciplines and approaches should strive to identify the instances where the outcomes predicted by a theory of interpersonal behavior become the inputs to a theory of physiological response to stimuli and vice versa, where a theory of interpersonal behavior intersects with a theory of the affective underpinnings of collective action, and so on. If we can achieve that, then we will transform the various streams of systematic investigation of affective processes currently engaged in “parallel play” into a rich and deep scientific understanding of a multilayered process.