The formal-mathematical constructs that are the focus of the target article are not grounded in nervous system functioning. Conciliatory efforts would call into question informational uncertainty as a singular phenomenon mediated by higher-level neural processes. That is, because a form of informational uncertainty – reward uncertainty – is subcortically mediated (Anselme & Güntürkün Reference Anselme and Güntürkün2019; Hart et al. Reference Hart, Clark and Phillips2015). Furthermore, evidence suggests that reward uncertainty gives rise to complex behavioral phenomena that can form the basis for cultural constraints on behavior (Strand et al. Reference Strand, Vossen and Savage2019). These ideas are expanded upon in what follows, in the hopes that formal-mathematical models such as those presented in the target article will, in the future, enumerate the place of reward uncertainty within the domain of informational uncertainty.

To understand reward uncertainty and its importance for social behavior, let's review a recent study concerned with the formation of the attachment patterns (Beckes et al. Reference Beckes, Simons, Lewis, Le and Edwards2017). Attachment patterns describe social–behavioral propensities of individuals arising in response to relationships with caretakers or supportive others. Therefore, attachment patterns are relevant for considering how information affects social behavior. As we discuss below, they also align with the dominant cultural institutions, collectivism, and individualism.

Beckes et al. (Reference Beckes, Simons, Lewis, Le and Edwards2017) experimentally investigated how supportive social responsiveness influences attachment-related behavior using a shock threat support-seeking paradigm. Participants were adults who ostensibly sought help from another participant (“supporter”) seated in another room. They pressed a button to indicate their need for help every time a shock threat signal appeared. The supporter could stop the imminent shock in response to this support-seeking behavior. In a randomized design, one group of participants experienced a supporter who acted consistently and predictably to prevent the shock. Another group experienced a supporter who acted inconsistently and unpredictably to prevent the shock. Therefore, in response to support-seeking behavior, the former group experienced reward certainty (continuous reinforcement) and the latter group reward uncertainty (intermittent reinforcement).

Study results revealed that exposure to reward uncertainty led to behavior toward the supporter that was characterized by attitudinal ambivalence and high levels of approach motivation – a behavioral constellation consistent with an insecure-anxious attachment pattern (Ainsworth et al. Reference Ainsworth, Blehar, Waters and Wall1978). By contrast, exposure to reward certainty led to unambiguous positive attitudes and moderate approach motivation toward the supporter – the behavioral equivalent of a secure attachment pattern. Not studied by Beckes et al. (Reference Beckes, Simons, Lewis, Le and Edwards2017), insecure-avoidant is a third primary attachment pattern. It is thought to be induced by an extinction schedule for security-seeking behavior and characterized by attitudinal neutrality and low approach motivation to caretakers.

The Beckes et al. (Reference Beckes, Simons, Lewis, Le and Edwards2017) findings reveal the existence of complex socio-behavioral response patterns induced by simple forms of environmental stimulation (i.e., intermittent and continuous schedules of reinforcement). Moreover, neuroscientific evidence reveals that these behavioral responses are mediated by dopamine activity within the subcortical nucleus accumbens (Hart et al. Reference Hart, Clark and Phillips2015). Therefore, a confluence of neural and behavioral evidence reveals the existence of forms of complex social behavior that are schedule-induced and subcortically mediated (Anselme & Güntürkün Reference Anselme and Güntürkün2019). As such, they do not reflect the “first principles” of the target article, which are cortically mediated and include perspective-taking and mind reading. I return to the issue of neural mediation after considering the importance for culture of the attachment patterns.

The impact of the attachment patterns on culture is suggested by the results of a world-wide review of their distribution across collectivist and individualist cultures (Mesman et al. Reference Mesman, Van IJzendoorn, Sagi-Schartz, Cassidy and Shaver2016). Findings reveal that collectivist cultures have a higher relative percentage of insecure-anxious individuals. Such individuals tend to prioritize interactions within strong-tie social networks (reflecting high approach motivation), which is a defining characteristic of collectivist cultures (Yamagishi & Hashimoto Reference Yamagishi and Hashimoto2016). Individualist cultures, on the other hand, have a higher relative percentage of insecure-avoidant individuals. Such individuals tend to prioritize interactions in weak-tie social networks (reflecting low approach motivation), a defining characteristic of individualist cultures. Therefore, participation in strong-tie relationships is relatively advantageous in collectivist settings, and participation in weak-tie relationships is relatively advantageous in individualist settings. In this way, culture is defined not by co-constructed minds but, rather, foundationally, by group-level payoff structures that constrain the reward maximization behavior of individuals. Those constraints reflect the aggregate of the behavioral propensities of the individuals who comprise the group (Strand et al. Reference Strand, Vossen and Savage2019).

Consistent with the target article, work cited above supports an informational account of reward uncertainty. Behaviorally, a richer reward schedule (continuous reinforcement) led to less extreme approach motivation than did a leaner schedule (intermittent reinforcement; Beckes et al. Reference Beckes, Simons, Lewis, Le and Edwards2017). Neurally, dopamine release was highest at maximum reward uncertainty (probability of reward = 0.5), and lowest at both extremes of certainty (probability of reward = 0.0 and 1.0; Hart et al. Reference Hart, Clark and Phillips2015). Therefore, the behavioral response is determined by the informational value of the schedule, not the reward amount. The question remains: Does the formal-mathematical model of the target article align with what is known about reward uncertainty and the neural mediation of complex behavior?