Rationality in Joint Action: Maximizing Coefficiency in Coordination

Abstract

When people perform simple actions, they often behave efficiently, minimizing the costs of movement for the expected benefit. The present study addressed the question of whether this efficiency scales up to dyads working together to achieve a shared goal: Do people act efficiently as a group (i.e., coefficiently), or do they minimize their own or their partner’s individual costs even if this increases the overall cost for the group? We devised a novel, touch-screen-based, sequential object-transfer task to measure how people choose between different paths to coordinate with a partner. Across multiple experiments, we found that participants did not simply minimize their own or their partner’s movement costs but made coefficient decisions about paths, which ensured that the aggregate costs of movement for the dyad were minimized. These results suggest that people are able and motivated to make coefficient, collectively rational decisions when acting together.

The Naïve Utility Calculus as a unified, quantitative framework for action understanding

“…When reasoning about social interactions, the Naïve Utility Calculus can reveal when an agent is pursuing non-social, cooperative, and competitive goals, based on whether they are maximizing egocentric, group-level, or antagonistic utilities, respectively. Furthermore, social interactions often require combining our own costs and rewards with those of others (Kleiman-Weiner, et al, 2016;Török, Pomiechowska, Csibra, & Sebanz, 2019). To do this effectively, we must consider how other agents' costs differ from our own, and the Naïve Utility Calculus may also be instrumental in providing these estimates.…”

Progress in Joint-Action Research

“…A further important goal for future research is to achieve a better understanding of how individuals decide whether to engage in joint action given the implicit and explicit commitments it entails (Michael et al, 2016) and the individual and joint costs that are involved (Török et al, 2019). Answering this question will require going beyond collective decisions (Roberts & Goldstone, 2011) and judgments about perceptual features (Bahrami et al, 2010) to understand how jointaction partners choose a course of joint action in the face of asymmetries in perception, knowledge, and abilities (Voinov et al, 2019).…”

Mechanisms for mutual support in motor interactions

“…Both joint actions and collective decision making 68,69 are characterized by the agent's investment in behavioral adaptations that imply costs in terms of cognitive and/or motor resources (e.g., deviation from maximal efficiency of movement trajectory [70][71][72][73][74][75] ), which are nevertheless repaid by smoother coordination and higher co-efficiency (see Ref. 76,77 ). Admittedly, we did not test in our task whether a joint payoff is what guides the agents' joint motor planning, but our results suggest it might be a relevant avenue for future research in the field.…”