We resonate with Tomasello's observation that much of human cooperation is driven by something akin to obligation. Pro-social behavior is not so much an act of “giving” as it is of “giving in” – to social rules, roles, and duties (see Cain et al. Reference Cain, Dana and Newman2014).

We not only find much agreement, we suggest that the scope of his analysis applies far wider than Tomasello asserts. He proposes that other-directed obligations arise when people enter explicit collaborative activities with shared goals and joint intentions.

We, however, would argue that obligation governs human behavior much sooner than that. For communities to thrive and develop, their members must honor certain rules and obligations no matter how anonymous, transient, or minimal their supposed relationships might be (Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010).

Work in our labs on interpersonal trust directly demonstrates the obligatory nature of pro-social behavior even at zero acquaintance (for reviews, see Dunning et al. Reference Dunning, Fetchenhauer and Schlösser2019; Fetchenhauer et al. Reference Fetchenhauer, Dunning, Schlösser, Van Lange, Rockenbach and Yamagishi2017). In laboratory trust games, we give participants $5 and tell them they can either keep the money or hand it over to a complete stranger whom they will never meet and who will never meet them. Importantly, if they hand over the money, we will inflate it to $20 and give the stranger their own choice. They can keep the $20 or give $10 back to the participant. Such a choice represents the core decision at the heart of many definitions of trust: People must decide whether to open themselves to potential exploitation by another person in order to produce a chance of personal benefit (Rousseau et al. Reference Rousseau, Sitkin, Burt and Camerer1998).

The choices that participants make contradict a strict account of neoclassical economics. Although participants on average think there is only a 40% to 50% chance the stranger will hand money back, clear majorities decide to trust that stranger (Fetchenhauer & Dunning Reference Fetchenhauer and Dunning2009). Pointedly, total strangers trust one another in one-off anonymous exchanges even though many expect their trust to be exploited and their welfare to suffer (see also Berg et al. Reference Berg, Dickhaut and McCabe1995). They gamble on the sociality of the stranger even though they would rarely make the same bet if gambling instead on a lottery wheel (Fetchenhauer & Dunning Reference Fetchenhauer and Dunning2012). To an economist, the source of risk should not matter. However, to our participants, it very much does.

Our research converges toward an explanation of this high rate of trust that, if not the same as Tomasello's, certainly rhymes with it. Many indicators point to social obligation as the source of this decision. Participants on average report that trusting a stranger is more something they should do rather than something they want to do (Dunning et al. Reference Dunning, Anderson, Schlösser, Ehlebracht and Fetchenhauer2014).

More important, the emotions that people attach to trust follow the logic of social rules and obligations. The prospect of breaking a social obligation produces high levels of anxiety and guilt (Higgins Reference Higgins1987). In our games, those who feel more anxious, guilty, and remorseful about distrusting the other person than they do about risking their money are the ones who trust the most – and these emotions predict trust beyond any economic factor (Dunning et al. Reference Dunning, Anderson, Schlösser, Ehlebracht and Fetchenhauer2014; Schlösser et al. Reference Schlösser, Fetchenhauer and Dunning2016).

Further work suggests the specific obligation that participants feel they must honor, again one that is compatible with Tomasello's account. Participants trust because they feel they must respect the character of the other person (see Dunning et al. Reference Dunning, Fetchenhauer, Schlösser and Elliot2016). To distrust the other person would be to call that person's integrity into question. People express intense anxiety at the prospect of doing so (Schlösser et al. Reference Schlösser, Mensching, Dunning and Fetchenhauer2015).

In addition, anecdotally, participants express that they choose to trust not out of expectation but rather out of “hope” and “faith” (Dunning et al. Reference Dunning, Fetchenhauer and Schlösser2012). When asked, our participants explain their trustfulness by noting, “I just want to live in a world where people trust each other.” Thus, they appear to take a “we” perspective, as Tomasello would put it, even at zero acquaintance. And as “we > me,” they feel obligated to trust even if from a “me” perspective this might be irrational.

In retrospect, the adaptive value of this obligation to trust at zero acquaintance is clear. First, it allows strangers, not only those already in collaborative relationships, to benefit from the behavior of others (Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010). We hand over our credit cards to complete strangers in restaurants. We purchase that trinket on eBay and assume it will show up in the mail. We procure our daily calories by going to stores run by people we never meet. Perhaps another way to put it is that we are embedded in crucial webs of implicit and fleeting collaborative relations with many of the strangers we encounter every day.

In addition, obligations to trust and respect each other from the start helps people to build the mutual goodwill needed for more intense and complex relationships involving valuable shared goals, joined intentions, and smooth cooperation (Lawler Reference Lawler2001).

Yet, what is largely missing in Tomasello's article is any reference to the delicate nature of a “we > me“ perspective. In our evolutionary past, there have always been temptations to act selfishly. It seems plausible that humans have developed all kinds of monitoring systems to protect them from exploitation by other group members (Cosmides et al. Reference Cosmides, Barrett and Tooby2010). Thus, the observation that our participants so cynically underestimate others’ trustworthiness might be one such device to avoid being naïve and gullible.

Finally, being skeptical about others’ trustworthiness, but still feeling an obligation to trust behaviorally, turns out to be a beneficial strategy, at least in our laboratories. The strangers who receive the $20 seem to respond to an obligation as well. Although they are expected to return money only about half of the time, they do so at a rate of roughly 80% (Fetchenhauer & Dunning Reference Fetchenhauer and Dunning2009). As such, trust at zero acquaintance may be more an obligation than a preference, but it is one rewarded handsomely.