Situations in which two or more individuals make choices that affect each other's future can be captured in games of strategy, the simplest ones having two players each with a choice among two options. Well-known examples include the prisoner's dilemma, stag-hunt, and hawk-dove games, and contain a social dilemma – whereas choosing C(ooperate) maximizes social welfare and is “collectively rational,” choosing D(efect) maximizes personal welfare and is “individually rational” (Kollock, Reference Kollock1998; Van Dijk & De Dreu, Reference Van Dijk and De Dreu2021). Whether Cooperate or Defect is labeled as rational or as irrational depends on the perspective taken – by the outsider or the players themselves. Whether the perspective is taken deliberately or intuitively is irrelevant.

Collective and individual rationality can be about minimizing collective or personal loss, or about maximizing collective or personal gain. In social dilemmas, a psychological focus on either gains or losses can be due to reference-dependent framing (Kahneman & Tversky, Reference Kahneman and Tversky1979), framing public good provision as give-some or take-some (Gaechter, Kolle, & Quercia, Reference Gaechter, Kolle and Quercia2017; Van Lange, Joireman, Parks, & Van Dijk, Reference Van Lange, Joireman, Parks and Van Dijk2013), or by mindfully taking one rather than the other perspective (Bermúdez, target article). Gain–loss framing does not alter the rank-ordering of preferences and leaves intact the logic and equilibrium properties of the game.

Gain–loss framing can, however, change the strength of preferences. Humans pursue individual rationality when in a “me” frame, and because of loss aversion more rigorously so in loss-frames. Experiments corroborate that individuals more likely Defect in loss-framed than gain-framed social dilemmas (Brewer & Kramer, Reference Brewer and Kramer1986; Sun et al., Reference Sun, Guo, Wang, Zhang, Jiang and Liu2022). Individuals pursue collectively rationality when in a “we” frame – in the behavioral sciences captured and extensively studied under headers such as pro-social motivation (De Dreu, Weingart, & Kwon, Reference De Dreu, Weingart and Kwon2000; Van Lange et al., Reference Van Lange, Joireman, Parks and Van Dijk2013) or cooperative orientation (Deutsch, Reference Deutsch1960). And indeed, experiments robustly revealed that individuals in a “we” frame more likely Cooperate in loss- rather than gain-framed social dilemmas (Carnevale, Reference Carnevale2008; De Dreu & McCusker, Reference De Dreu and McCusker1997; Ispano & Schwardmann, Reference Ispano and Schwardmann2017). From a normative perspective, as in the target article by Bermúdez, frames can be pondered in isolation. From an analytic and empirical point of view, however, frames need to be considered in combination – in social decision making, the impact of gain–loss frames depends on the “we–me” frame.

In any game of strategy, individual outcomes are a function of the choices made by oneself and at least one other player. Bermúdez remains silent on the possibility that, therefore, individuals are influenced by their own frames, and by those of the other player(s). Already in two-player games of strategy, there are no less than 2 (players) × 2 (gain or loss frame) × 2 (we or me frame) = eight possible scenarios. This complicates the analysis in particular for games with repeated play where players learn about and adapt to their partner's behavioral strategies and, possibly, their partner's frame(s). Someone in a gain frame may interact with someone in a loss frame, and someone in a “we” frame may interact with someone in a “me” frame. Put differently, someone trying to minimize personal loss as rationally as possible may interact with someone trying to maximize collective gain as rationally as possible.

Although Bermúdez offers little to understand how such interactions unfold, extant work in the behavioral sciences does. For example, knowing that the other player is under a loss rather than gain frame raises empathy and motivates cooperation (De Dreu, Emans, & Van de Vliert, Reference De Dreu, Emans and Van de Vliert1992; Fiedler & Hillenbrand, Reference Fiedler and Hillenbrand2020; Van Beest, Van Dijk, De Dreu, & Wilke, Reference Van Beest, Van Dijk, De Dreu and Wilke2005; Weber, Kopelman, & Messick, Reference Weber, Kopelman and Messick2004). This incentivizes individually rational players, such as those in a “me frame,” to adopt and communicate a loss frame themselves, as it helps to extract cooperation from the counterpart. Experiments with two-person bargaining games have revealed such asymmetric convergence on loss rather than gain frames (De Dreu, Carnevale, Emans, & Van de Vliert, Reference De Dreu, Carnevale, Emans and Van de Vliert1994). With players in a “me” frame, loss framing dominates gain framing.

When Cooperate is the dominated strategy, cooperators and those in a “we” frame will be exploited by counterparts in a “me” frame. In mixed populations, “we” players only survive if they adopt their counterpart's “me” frame. Conversely, non-cooperators paired to cooperators have little incentive to switch from their “me” into a “we” frame. Accordingly, we should see asymmetric convergence on “me” frames. This prediction resonates with Pruitt and Kimmel's (Reference Pruitt and Kimmel1977) goal-expectation hypothesis, with experiments on conditional cooperation in N-person social dilemmas (Fehr & Gächter, Reference Fehr and Gächter2000; Van Dijk & De Dreu, Reference Van Dijk and De Dreu2021), and with evolutionary agent-based modeling of cooperation in mixed populations (Axelrod & Hamilton, Reference Axelrod and Hamilton1981; Gross & De Dreu, Reference Gross and De Dreu2019a, Reference Gross and De Dreu2019b; Nowak, Reference Nowak2006). Except when all players are in a “we” frame, social decisions in repeated play will be dominated by a “me” frame.

It is of note that the domination of loss over gain, and “me” over “we” frames cannot be solved with rational (re)framing as suggested by Bermúdez. When Cooperate is the dominated strategy, as in the prisoner's dilemma, considering Cooperate as creating social welfare or as safeguarding one's personal outcomes does not solve the dilemma as long as it is not (1) common knowledge that (2) the other player(s) also take a social welfare perspective. Rational framing is also of no help when neither Cooperate nor Defect is the dominated strategy, as in games with their equilibrium in mixed strategies such as the Stag-Hunt game discussed by Bermúdez (also see De Dreu & Gross, Reference De Dreu and Gross2019). Consider a simple Hide-and-Seek game in which player H can hide in location 1 or 2 and player S can seek in location 1 or 2. H wants not to be found by S, and S wants to find H. If S is expected to choose 1, H should choose 2, upon which S is incentivized to choose 2, leading H to opt for location 1, and so on ad infinitum. It is unclear how choosing a particular frame, or considering decision-options from different perspectives, can help players to achieve whatever goal they have – what to choose remains conditional on what others choose (Arad & Rubinstein, Reference Arad and Rubinstein2012; Camerer, Ho, & Chong, Reference Camerer, Ho and Chong2004). Quasi-cyclical preferences may be deliberated, rational, and defendable. They do not, however, solve the dilemma.

Experiments in the psychological and economic sciences converge on the possibility that some frames discussed by Bermúdez can exist in theory, yet rarely survive in practice – human psychology gravitates toward minimizing my loss, and this explains the break-down of cooperation. In the end, loss framing explains how much effort humans invest, and the “me–we” frame on what the effort is invested in. From a social welfare perspective, and to prevent the tragedy – or funeral – of the commons (Gross & De Dreu, Reference Gross and De Dreu2019a, Reference Gross and De Dreu2019b; Hardin, Reference Hardin1968), it is collectively rational to collectively adopt and rationally stick to a “minimize our loss” frame.