Coalitional psychology and the evolution of nationalistic cultures

Abstract

The commentaries addressed various aspects of our account of historical myths. We respond by clarifying the evolutionary theory of coalitional psychology that underlies our claims (R1). This addresses concerns about the role of fitness interdependence in large groups (R2), cultural transmission processes (R3), alternative routes to nation-building (R4) and the role of proximal mechanisms (R5). Finally, we evaluate alternative theories (R6) and discuss directions for future research (R7).

We sincerely thank all commentators for their thoughtful engagement with our article and for sharing their valuable insights. We are pleased that our work has garnered attention from researchers across a broad spectrum of disciplines. Most major traditions in cognitive and psychological sciences – and almost all of the ones we explicitly discussed in the target article – are represented in the commentaries, including social psychology (e.g., Bilewicz & Bilewicz; Kardos), evolutionary psychology (e.g., Pietraszewski & Moncrieff), cultural evolution (e.g., Zefferman & Smaldino), philosophy of mind (e.g., Wildes & Andrews; Blancke), and psychopathology (e.g., Fonagy & Campbell). We are especially grateful to commentators from political science and history (e.g., Sarkar & Sarkar; Friedman; Wimmer) for allowing us to engage in productive scientific discussions beyond traditional disciplinary divides. Overall, we rejoice that our article had the intended effect: to spark a renewed multidisciplinary interest in the topic of historical myths.

The commentaries were diverse and addressed many aspects of the target article. To respond comprehensively, we first restate the theory of coalitional choice, which forms the theoretical basis of our account of historical myths (R1). We then address concerns regarding the theoretical impossibility of fitness interdependence at the national scale and its implications for our theory of cultural transmission (R2 and R3). Next, we acknowledge the point raised by several commentators that historical myths are not the only tool for nation-building (R4). We agree: group affiliation is shaped by a flexible coalitional choice psychology and can arise from any input – symbolic or material – that signals productive social interactions. Historical myths are just one recurrent example of such inputs. We also clarify the distinction between ultimate and proximate explanations (R5). Many commentaries focused on proximal factors influencing the appeal of historical myths. We argue that our ultimate-level theory accommodates several of these proximal pathways. Lastly, we evaluate alternative ultimate explanations proposed in the commentaries (R6) and explore potential directions for future research (R7).

R1. Nations as the products of coalitional choice psychology

National categories often feel self-evident to individuals who live in nation-states. For many, it seems obvious that categories such as “France,” “Spain,” or “Ukraine” delimitate something meaningful about the world – specifically, something that ought to form the basis of political sovereignty and citizenship rights and duties (Anderson, 1991). Yet, the perceived naturalness of national categories is a typical case of instinct blindness: some social phenomena feel so natural to us that we discard the fact that they actually proceed from highly complex computations (Cosmides & Tooby, 1994). As noted by Pietraszewski and Moncrieff, the task of psychologists is precisely to reveal these hidden puzzles.

Indeed, the psychological appeal of nations is puzzling once we consider the variety of alternative coalitional arrangements that are available, in principle, to an individual at any point in time. This fact becomes apparent in social contexts where the naturalness of nations is challenged: this happens in secessionist claims or during “ethnogenesis” – where individuals contest existing coalitional boundaries and support alternative ones (Horowitz, 1985). What makes some coalitional arrangements more psychologically compelling than the myriads of cognitively invisible alternative coalitions that could emerge? Why do some boundaries (e.g., “France,” “Corsica”…) become cognitively plausible as the locus of political sovereignty and citizenship and not others (“Grenoble,” “All bald people”)?

An implicit objective of the target paper was to unravel the intuitive blindness that underlies the cultural success of national categories. The perception of a sociopolitical entity as a meaningful “nation” is not an arbitrary norm that is passively acquired from the social environment, but the result of complex computations, which we have summarized under the concept of “coalitional choice psychology” and described in sections 3 and 4. This psychology involves interpreting environmental cues to determine the most advantageous coalition – among all available options – in the individual's specific context. In turn, coalitional recruiters must strive to advertise their coalition, if they wish to secure the coalitional support of other people. This part of our theory was generally accepted by the commentators, but we provide three elements of clarification that we think can provide a preliminary response to many commentaries.

The first important clarification is the exact nature of the output of coalitional choice psychology. Importantly, the output of these computations is not directly cooperative behavior, but support for a given coalitional arrangement. What coalitional choice computations generate is the mental representation that a given coalitional arrangement is especially desirable – whether this representation applies to a group of friends, an ethnic group or a nation. Intuitively, one political preference that might emerge from this output is “I want an independent Corsica” or “I want Corsica to unite with France.” Conversely, this mechanism might also categorize some coalitional arrangements as irrelevant: no social movement has ever called for the right to self-determination of “Grenoble” or “all bald people.” Analytically, this mental representation is distinct from the motivation to actually cooperate with coalition members. At the large scale of entire nations, coalitional choice psychology merely generates the mental representation that this coalitional arrangement is especially desirable, leaving intact the typical collective actions problems that emerges in large-scale cooperation: although it would be especially productive if everyone in coalition X cooperated; individuals have a strong incentive to shirk (Olson, 1965; Ostrom, 2015). For instance, a person might in principle support the independence of Corsica without participating to Corsican collective actions.

In this sense, and this will be our second clarification, our theory does not call into question standard theories of large-scale cooperation. We agree with previous theoretical work that, ultimately, what stabilizes cooperation in large groups is a combination of cultural systems of monitoring, reward and punishment (often subsumed under the concept of “institutions”) and reputational pressures (Acemoglu & Robinson, 2012; Liénard, 2014; Lie-Panis, Fitouchi, Baumard, & André, 2023; Olson, 1965; Ostrom, 2015; Powers & Lehmann, 2013; Powers, Van Schaik, & Lehmann, 2016). While institutions can explain how cooperation can be enforced in large groups, it does not explain why people consent to them. Yet, consent is central for the success of nation-building: when institutions are perceived as enforcing a coalitional arrangement that does not optimally benefit citizens, they risk being perceived as irrelevant – at best – or as unfair, extractive or oppressive. Therefore, institutional enforcement is insufficient to explain nation-building: we also need to explain why some coalitional arrangements are more appealing than others (e.g., why can the concept of “France” move the masses more than the concept of “Grenoble” or “all bald people”) and why some large and abstract entities appealing at all (why do some people even care about something like “France”).

The answer, we suggested, is in the way standard theories of large-scale cooperation intersect with coalitional choice psychology. Successful nation-building requires solid institutions and reputational pressure – but these mechanisms are only possible if they are seated on a coalitional arrangement that is perceived as desirable in the first place – which is the product of coalitional choice computations. Institutions matter, but coalitional choice psychology explains why people support the creation of institutions around a given coalitional arrangement. Reputation matters, but coalitional choice psychology explains why people come to consider national cooperation as a special moral duty, and hence why it becomes beneficial to one's reputation to abide to national institutions. In fact, this account resolves a feature of nationalism that would otherwise seem illogical: in high times of national threats, people support (and elites execute) both strong levels of patriotism and the implementation of authoritarian institutions to enforce national cooperation (e.g., Kuzio, 2016). These are the two legs of nationalism: the shared belief that the nation is a desirable social entity, and standard enforcement mechanisms that make this social entity possible. As a result, the task of the nation-builder is twofold as well: build stable institutions and advertise the nation as desirable.

Finally, our account of coalitional choice psychology emphasizes the flexibility of its outputs. Coalitional choice mechanisms merely generate the mental representation of desirable coalitional arrangements in the form of social categories that “feel” compelling, but the specific nature of this arrangement can manifest in myriad ways: nations, ethnic groups, world religions, fandoms, political parties (as noted by several commentators; e.g., Hoffman & Moya; Moser; see also Moya, 2023). In this sense, we certainly agree with Maryanski and Turner who perceived fitness interdependence – and any other input that might activate coalitional choice psychology – might “underpin[s] all sorts of diverse social formations.” However, we do not concur with the claim that this necessarily assumes domain-general cognitive mechanisms. It is perfectly plausible that domain-specific cognitive mechanisms – mechanisms that take a narrow subset of stimuli in the world and process it in a specific way – can generate very different outputs. Moral cognition, for instance, can use the same narrow subset of stimuli (costs and benefits in social interactions) to generate a variety of justice principles (e.g., equality, equity, merit) (André, Fitouchi, Debove, & Baumard, 2022).

R2. Can fitness interdependence incentivize cooperation in large groups?

The most important implication of this underlying theory is that – contra the assumption of several commentaries (e.g., Hoffman & Moya; Zefferman & Smaldino; Brusse & Sterley) – our target article does not claim that fitness interdependence directly incentivizes cooperation in large groups. Perceived fitness interdependence between one individual and other group members is merely one of the inputs that can be fed to our coalitional choice psychology, leading to its immediate output: the mental representation of the group as a desirable coalitional arrangement. It is only indirectly, and in combination with more standard mechanisms of cooperation stabilization, that fitness interdependence can facilitate nation-wide cooperation. By making the group seem like a desirable coalitional arrangement, it can increase individuals' support for the emergence of institutions that will stabilize this arrangement and create a moral reputational pressure to act cooperatively with the nation (see esp. sect. 3.2).

Some commentators were concerned that we might have translated models of fitness interdependence in dyadic relationships to the realm of large-scale cooperation (Brusse & Sterelny), or that we were ambiguous in how we sought to do so (e.g., Friedman). We clarify this point here: when applied to non-dyadic contexts (or “groups”), we defined fitness interdependence as the correlation between the fitness of a given individual with the average fitness of other group members. This is very different from dyadic fitness interdependence, in which the correlation of fitness benefits is calculated between two individuals – as in the interdependence modelled by Zefferman & Smaldino). This definition was dismissed by Brusse and Sterelny, who argued that our focus on the “general welfare” of group members is an “uncashed metaphor”; yet we believe that it precisely captures the type of interdependence that should matter at the scale of large groups – because important acts of cooperation in large groups usually involves giving resources that are distributed widely across the group and not targeted towards one specific individual (e.g., war effort, taxpaying, nation-wide redistributive policies) (see Hechter, 2000; Tooby, Cosmides, & Price, 2006). We do agree with Brusse and Sterelny, however, that our use of the word “interdependence” is misleading in our context: as these commentators note, if an individual has a stake in the welfare of other group members, the opposite is unlikely to be true. Rather than talking about “interdependence” between an individual and other group members, a more adequate phrasing would be that individuals have a stake in the average welfare of other group members (Barclay, 2020).

Finally, our definition does not presume any higher-level selection mechanism – as suggested by Kaufman, Kashdan and McKnight or proposed by Zefferman and Smaldino. For an individual to develop an objective stake in the fate of other group members, it is enough that, for some reason, she perceives them as good recurrent cooperation partners who are difficult to replace (Barclay, 2020). If cooperation with a given set of individuals (e.g., “the French”) is especially productive and forging alternative alliances is difficult, then people objectively have some stake in preserving the welfare of these individuals – which, again, does not necessarily translate into actual cooperative behavior because of the typical social dilemma posed by cooperation in large groups. Our paper precisely provides one such reason: because group members have been involved in repeated interactions over time, which considerably increases the cost of switching coalitional arrangements.

R3. Implications for the cultural evolution of historical myths

These points clarify a few recurrent concerns raised by the authors about the cultural transmission mechanisms that we outline in our article. On the one hand, several commentators asked why people would accept historical myths that are transmitted to them if they do not provide reliable information about their coalitional interests (e.g., Wimmer; Blancke; Shao & Barlev). In the target article, we argue that we actually do not expect that historical myths will be accepted without question. Humans are equipped with psychological mechanisms for epistemic vigilance that allows them to discard information that contradicts more reliable evidence (sect. 6.1). For instance, we took the blatant example of Ukrainians receiving historical propaganda from the Russian regime while experiencing oppression from Russian troops. In this setting, it seems unlikely that people will change their mind in the face of historical myths.

This has led other commentators to ask the related question: if historical myths cannot easily sway other people's opinion, why then would there be an interest in transmitting them? (e.g., Blancke). First, our theory allows for contexts in which historical myths can alter people's coalitional preferences: for instance, in situations where this information is congruent with other reliable evidence (sect. 6.1). This might be the case, for instance, when historical myths are propagated to the masses along with other concrete material benefits – typically in the form of access to public goods (Weber, 1976). In this case, there is no strong counter-evidence that might contradict the information brought by the historical myth. The other condition is that historical myths should be believable. This is precisely the endeavor that elites strive to achieve when propagating historical myths: they mobilize existing historical events and support their claims (often dubiously) by archaeological and historical material, to increase the credibility of the historical myths (sect. 4.3). Humans are not gullible, but they are not infallible either: in the absence of incongruent cues, and provided with sufficiently solid evidence, historical myths might have a chance to impact coalitional preferences.

Moreover, individuals may have incentives to transmit historical myths even when they have a small probability to be effective (e.g., due to the prevalence of incongruent cues). Indeed, the production of historical myths is not always costly, especially for leaders of nation-states. National ceremonies, commemorations and public discourses can be produced with a negligible cost relative to the total state budget. Another explanation, suggested by Blancke and Shao and Barlev, is that publicly advocating historical narratives can increase one's reputation as a committed coalition member. This account is plausible, although it is unclear how this signal might be perceived as credible. One solution is that the transmission of historical myths can become reliable signal of group commitment when it occurs in the presence of coalitional rivals – as in “burning bridges” signaling (Mercier, 2020, p. 193; see also Williams, 2022). A complementary solution is that while the transmission and endorsement of historical myths are not sufficient conditions to establish group commitment, they might be necessary. Individuals who reject historical myths that advertise their nation might be perceived as having alternative coalitional preferences (e.g., prefer unification with France over the independence of Corsica). This mirrors a similar phenomenon in the domain of moral religions: expressing the belief in a moral God that punishes moral violations is probably insufficient to establish one's moral character but might still be a productive reputational strategy – because atheists are perceived as less trustworthy (Fitouchi & Singh, 2022; Fitouchi, Singh, André, & Baumard, 2023; Gervais, 2013, 2014; Gervais, Shariff, & Norenzayan, 2011; Gervais et al., 2017).

Most importantly, it is enough that people believe in the efficacy of a cultural technology to explain its cultural success – without having to assume that their folk-intuitions are indeed accurate. Historical myths can spread in a population as long as people perceive that they have an interest in activating other group members' perception of fitness interdependence and have reasonable reasons to believe that historical myths can activate such perceptions (see sect. 6.2 for an extensive discussion of this point; also see Fitouchi et al., 2023).

R4. Historical myths are not the sole pathway to coalitional recruitment

R4.1. Reiterating the centrality of fair public goods provision

Our account of the coalitional choice mechanisms underlying the cognitive appeal of national categories addresses the point, raised by many commentators, that historical myths are not the only cultural technology that can be used for nation-building purposes. For instance, Wimmer noted that historical narratives have been used in a wide diversity of formats across social contexts, like “imperial stories of superior civilizational origins” that do not necessarily fit our definition of historical myths. Likewise, Akers emphasized the diversity of societal “meta-narratives” that structure group identities across the world but do not match with our description of historical myths: “meta-narratives” often emphasize the nation's shared beliefs (religious or secular) and shared goals, while our target article only addressed the rhetoric of shared roots. This focus makes it difficult to account for the varieties of nationalism that are not based on the belief in a shared ancestral history – like American nationalism.

We certainly agree that our target article is nowhere near explaining every aspect of nationalistic rhetoric. There are probably dozens of cultural technologies involved in the formation of national categories. In fact, this idea is consistent with our account of coalitional choice psychology: group commitment is the product of a flexible psychological mechanism which assesses the fitness benefits associated with membership in different coalitional arrangements. As such, any input can fuel national affiliation to the extent that it provides information about its associated fitness benefits. For this reason, we actually expect nationalists to invest in a wide range of cultural technologies beyond historical myths to attract coalitional support. In particular, we certainly expect that they will mobilize information about the present and future of the coalition. For instance, commentators have proposed that coalitional recruiters might advertise the nation's “cooperative prowess” in achieving coordination in the present (Wolf); shared goals and beliefs (Akers); the moral virtue of national heroes (Wildes & Andrews); or the magnitude of external threats (Pietraszwski; Snijder & De Dreu). In our target article, we also noted the existence of political uses of history beyond the scope of historical myths as we defined them and which probably play a substantial role in collective actions – for instance, raising awareness about historical grievances of marginalized groups (see sect. 2).

We also recognize that some of these cultural inputs should be significantly more psychologically compelling than historical myths in their capacity to alter coalitional preferences. In particular, a consistent finding in political science is that one of the most important drivers of national affiliation is the fair provision of public goods – or at least the perception thereof (Wimmer, 2018). In light of this finding, we certainly agree with Friedman that nationalists aiming to recruit coalitional support should primarily emphasize the quality and fairness of resource distributions across social groups – both in the present and in the past. And they certainly do: as historians have shown, one of the most consistent ideological tenets of nationalism is the belief in a horizontal comradeship between all group members – a rhetoric that is often produced with an aim to conceal domination and inequality (Anderson, 1991).

There is, however, one caveat to the claim that nationalists should advertise egalitarian access to public goods in their coalition – at least if “public goods” are understood in a purely economic sense. Indeed, as shown by Sarkar and Sarkar based on the Indian example, low-status individuals sometimes endorse nationalism even when the nation is highly unequal – presumably to their detriment (see Shayo, 2009). From an evolutionary perspective, this is not necessarily a paradox: individuals might still favor a coalitional arrangement in which they receive less strictly economic benefits if this loss is compensated with non-economic benefits to their fitness. In their commentary, Sarkar and Sarkar emphasize status and protection from (perceived) out-group threats as such non-economic benefits, and argue that this is what motivates low-status Hindus to tolerate high degrees of within-coalition inequality. This explanation is certainly plausible: low-status individuals might endorse an unequal coalitional arrangement if they perceive that they receive status gains and protection in exchange (Klor & Shayo, 2010). Interestingly, historical myths themselves might serve a similar purpose: not by providing status and protection but information about coalitional productivity. In advertising their coalition as especially productive and difficult to replace, recruiters might aim to convince low-status members that their subaltern social position is compensated by the added efficiency.

Overall, we easily recognize that historical myths activating perceived fitness interdependence through cues of repeated interaction over time are not the sole element with the potential to generate affiliation with large groups. Cultural evolutionary processes can generate a variety of technologies to achieve the same objectives while targeting the same human minds. For instance, fictions with imaginary worlds target specific cognitive mechanisms which explain why they have recurrent design features – yet they exist in an impressive variety of formats (Dubourg & Baumard, 2022). Likewise, the cultural technologies for nation-building that become prevalent in a given context might vary to some extent with ecological and cultural constraints: for instance, as suggested by Wimmer and Akers, historical myths might be more prevalent in countries that have an ancient history that they can exploit, but less so in countries like the United States, where European-Americans could not rely on ancient historical material.

What we do argue however is that historical myths are cross-culturally recurrent, and most importantly, that they are especially puzzling. The combination of these two factors is how we justify the focus of our main article. It is unsurprising that nationalists emphasize the present and future benefits that can be derived from national commitment; or that social movements use history to highlight historical grievances; but it is puzzling that nationalists care so much about their nation being ancestral and continuous (see sect. 2 for a discussion of this point).

R5. Proximate accounts

R5.1. Proximate mechanisms add interesting descriptive layers, but do not necessarily improve explanatory power

The coalitional choice psychology framework can also address concerns regarding the relationship between the ultimate logic of group affiliation and its proximal manifestations. Indeed, several commentaries proposed alternative of complementary accounts to our theory of historical myths by emphasizing proximal mechanisms. We recognize that our target article was focused on the ultimate logic and cognitive mechanisms driving the cultural evolution of historical myths while leaving out important proximal mechanisms at play.

Generally, we agreed with the authors who proposed proximal complements to our theory. Some of them discussed possible phylogenetic pathways that might have contributed to the human ability to produce historical myths. Benitez-Burraco stressed the decisive role of the human ability for language and storytelling in the evolution of large-scale cooperation and proposed two evolutionary pathways to account for it. Similarly, Maryanski and Turner suggested that “Machiavellian” social manipulation – similar to coalitional recruitment in humans – is found in chimpanzees, suggesting that this trait might date back to early hominoid ancestors. Both accounts are plausible, although more research is needed to establish these hypotheses. Other commentators focused on psychological proximal accounts: Kaufman, Kashdan and McKnight and Kardos suggest that the appeal of historical myths stems from psychological needs (e.g., for meaning, belonging, continuity, or entitativity). Drawing on Social Identity Theory, Kardos links these needs to social identification, which ultimately fosters within-group cooperation. Similarly, Pelletier and Fay, Wildes and Andrews and Oatley and Wu insist that historical myths do not merely provide information about the fitness interdependence of group members, but also elicit emotions or “affective states” that also play an important role in nationalist rhetoric. In particular, Pyszczynski, Solomon and Grenberg stress the role of death anxiety in motivating social identity. We certainly acknowledge the fundamental role of psychological needs and emotions on decision-making, especially when it comes to group affiliation.

Finally, we agree with Elster and Glowacki that historical myths activate cues of self-concept and thus help group members represent their group as a “quasi-coherent self with distinct characteristics, goals and desires.” As suggested by these authors, representing social groups in this way is likely beneficial and might serve as a basis for more complex computations (e.g., representing stereotypes, group “goals” or “desires,” see Tooby et al., 2006). It is less clear, however, how representing the group as a coherent entity leads specifically to coalitional preferences. The authors propose that this is because the perception of a coherent group-self might increase individuals' interest in group-level benefits but this causal mechanism is underspecified. One possible answer is that the perception of a coherent group-self actually reflects perceived fitness interdependence between group members – which is why it motivates coalitional affiliation. More research is needed to test this hypothesis.

Overall, while we agree that these accounts add interesting layers to our theory, it is unclear whether they alter the main predictions of our model. Proximate mechanisms are certainly indispensable for evolutionary theories, but only to the extent that they yield novel predictions that would not be expected by sole reliance on the theoretical tools of evolutionary biology – the “phenotypic gambit” of human behavioral ecologists (Nettle, Gibson, Lawson, & Sear, 2013). For instance, evolutionary psychologists are interested in the evolved cognition of humans because it can explain why humans sometimes engage in maladaptive behaviors in industrial environments (Li, Van Vugt, & Colarelli, 2018). Similarly, considering phylogenetic or neurobiological constraints can explain why apparently sub-optimal traits evolve (Nettle et al., 2013). In our case, the contribution of proximal accounts beyond the cognitive processes we describe in the main article is less clear. More research is needed to investigate whether the study of storytelling capacities (Benitez-Burraco), emotions (Pelletier & Fay, Wildes & Andrews, Oatley & Wu) or psychological needs (e.g., Kaufman, Kashdan & McKnight, Kardos) can make new predictions about the cultural success of historical myths.

R5.2. Proximate accounts are not alternatives to ultimate accounts

Conversely, we disagree with commentaries that presented proximal accounts as alternatives to our ultimate model (e.g., Kardos). Most of the proximal mechanisms cited in the commentaries are not alternatives to our account but are likely the proximal manifestation of the ultimate logic we described. We have no doubt that people experience visceral and measurable emotions and psychological needs related to their social identity – for example, need for belonging, need for continuity, etc. What we propose is that most of these visceral affective states might be design features of cognitive mechanisms that motivate individuals to engage in behaviors that are beneficial to their reproductive success – just like the visceral desire for an attractive mate is a feature of our mating cognition. For instance, the “need for belonging” is likely the proximate psychological mechanism that motivates humans to secure membership into a reliable group of social partners and “need for continuity” might be the proximal motivation that pushes people to attend to cues of repeated interactions over time and commit to more interdependent groups. Likewise, successful historical myths might indeed reduce anxiety, not by providing an abstract sense of permanence, but by showing the group as a cohesive entity that is able to provide valuable fitness benefits. This account is compatible with the evidence provided by Pyszczynski, Solomon and Grenberg – much of which we cited ourselves and analyzed in light of our account in the target paper (see sects. 4 & 5).

The same line of reasoning applies for emotions. From an evolutionary perspective, emotions are a coordinated adaptive response to adaptive challenges in our environment – for instance, fear coordinates coherent physiological and perceptual responses to the presence of imminent threats (Al-Shawaf, Conroy-Beam, Asao, & Buss, 2016). Thus, some of the emotions elicited by historical myths, like “security” or “intimacy” (Wildes and Andrews) might represent a psychological response to the recipients' perception of fitness interdependence with other group members. However, in light of our theory, we believe that the emotions that might play the most important role in national affiliation were not mentioned in the commentaries. First, previous research suggests that empathy is the emotional marker of perceived interdependence (Fitouchi, André, & Baumard, 2024). For this reason, we expect that successful historical myths – and more generally, successful nation-building – should elicit empathy toward other group members. Similarly, because coalitional choice psychology can lead to moral representations about what constitutes a mutually beneficial coalitional arrangement, national affiliation might elicit a sense of moral duty – or, in cases where individuals are imposed a coalitional arrangement that is not perceived as mutually beneficial, anger (André et al., 2022; Fitouchi, André, & Baumard, 2023). Overall, this discussion shows the limitations of traditional philosophical divides between rationality and emotion (“Logos” and “Pathos,” following Pelletier & Fay). Emotions and other psychological needs are instincts that are felt viscerally, but they are also “rational” – in the sense that have evolved to solve adaptive problems, and result from complex computations of fitness costs and benefits (Al-Shawaf et al., 2016).

Ultimate and proximal explanations are therefore complementary: ultimate theories provide the evolutionary logic for the existence of a trait, while proximal theories uncover how this trait manifests. However, proximate theories that rely on psychological needs and emotions without considering ultimate-level mechanisms run the risk of producing almost tautological explanations – equivalent to claims like: people eat because they have a need for food and have sex because they have a need for pleasure. These accounts also tend to raise more questions than they solve: why do people have such needs and emotions in the first place? In the absence of first principles from which needs can be inferred, psychological needs might become ad-hoc explanations of human behavior.

One commentary explicitly addressed this critique: Pyszczynski, Solomon and Grenberg propose that historical myths are appealing because they alleviate death anxiety; and they justify this psychological need with anxiety's presumed deleterious effect on “diverse biological and psychological systems that facilitate individual survival.” We raise several concerns about this interpretation. First, the proposed evolutionary mechanism is underspecified. The authors mention that historical myths “enable people to view themselves as valuable contributors to something great that stretches far into the past and that will endure indefinitely,” but it remains unclear why people should perceive this information as reducing the “inevitability of death.” Second, the authors wrongly assume that anxiety is systematically detrimental to survival and reproduction. In contrast, evidence shows that anxiety is an adaptive mood that is triggered by threatening environments and generates adapted physiological and cognitive responses (e.g., higher threat sensitivity) – which does not mean that there cannot be pathological forms of anxiety (Nettle & Bateson, 2012). Relatedly, the authors claim that people might feel anxious because of the inevitability of death. But from an evolutionary perspective, it would make no sense to react to a threat that is so general and, by definition, inevitable (Kirkpatrick & Navarrete, 2006). Rather, evidence shows that threat-detection mechanisms react to specific and avoidable threat with adaptive responses, like pathogens, predators or coalitional rivals (Neuberg & Schaller, 2016). Cognitively, humans do not fear “death” but fear specific features of their environment that may compromise their fitness. Thus, if it is indeed true that historical myths alleviate death anxiety, a compelling theory must provide an evolutionarily plausible reason for why information about the ancient history of one's group can lead to fitness gains.

R6. Alternative ultimate accounts

R6.1. Uncertainty reduction

Several commentators suggested alternative ultimate accounts for the cultural success of historical myths. One recurrent proposal was that historical myths might serve an epistemic function, providing a shared framework for interpreting social interactions. For instance, Moser suggested that they serve as “meta-heuristics for ensuring that members of a common group follow the same norms, envision the world with a shared ontology, and respond to new problems in the same way.” Similarly, Grüning and Krueger proposed that myths “provide[s] a behavioral framework for the world that reduces the cognitive load of decision making.” This account is convincing but might not explain the same types of historical narratives as the ones we emphasized in the target article. Uncertainty-reduction theories are well-suited to explain myths that provide information about the distinct customs and values of group members – like the myths described by Oatley and Wu – but are less suited to explain historical myths that emphasize cooperative events and long histories. Future research might disentangle the two accounts by testing how relevant psychological constructs (e.g., uncertainty aversion and national identification) predict the endorsement of historical narratives.

R6.2. Cultural group selection and arbitrary group norms

Other commentators rejected our claim that the cultural evolution of historical myths could be explained in the absence of group-level processes and proposed alternative accounts based on cultural group selection theory. First, Smaldino and Zefferman proposed that group boundaries reflect the scale of cultural variation – because this is the scale at which parochial behavior is favored by group-level cultural selection. As a result, the function of historical myths might be to “clarify the scale of existing cultural variation.” However, this idea does not account for the immense malleability of group boundaries – far beyond the constraints of cultural similarity – that we detailed in section 3. As political scientists have repeatedly shown, ethnic and national group boundaries are malleable and often reflect contextual incentives more than sincere attachment to cultural homogeneity (Habyarimana, Humphreys, Posner, & Weinstein, 2007). This is well illustrated by a seminal paper showing that the Chewa and Tumbuka ethnic groups are allies in Zambia but rivals in Malawi – despite being culturally similar across borders, and reflecting different incentives for alliances in both countries (Posner, 2004). Cultural (especially linguistic) similarity can certainly play an important role in coalitional preferences, to the extent that sharing culture reduces transaction and coordination costs, but other incentives (e.g., shared goals, common threats, efficient institutions) can substantially reduce and sometimes cancel preferences based on cultural similarity (e.g., Habyarimana et al., 2007). Finally, this account does not explain the design features of historical myths. Per Smaldino and Zefferman, nationalists should simply advertise the cultural homogeneity of their nation – which they often do – but not advertise the nation's long history of cooperation.

Second, Hoffman and Moya propose a game-theoretical model to claim that large social groups may not “be based on anything “real”” – as long as all group members have a shared understanding of group boundaries. This claim is similar to other commentaries, who stressed the coordination role of historical myths. Wolf points to evidence that shared mental representations stimulate cooperation; and proposes that it is the observation that group members have achieved a shared representation of the past that make historical myths compelling for coalitional choice. Fonagy and Campbell make a similar point when they attribute the appeal of historical myths to their capacity to elicit a “shared cognition” (i.e., shared intentionality, shared goals and mutual recognition).

The coordination of minds and behaviors is certainly an important feature of group formation. As evolutionary psychologists have claimed: “groups do not objectively exist: they only exist to the extent that they are represented in mutually consistent ways in the minds of assorted individuals” (Tooby et al., 2006; p. 111). Accordingly, people are highly sensitive to cues that indicate common knowledge of relevant information among a group of participants – and use these cues to motivate cooperation (De Freitas, Thomas, DeScioli, & Pinker, 2019; Deutchman, Amir, Jordan, & McAuliffe, 2022). This probably explains why historical myths are often publicly displayed in ostensible rituals – typically in the forms of national commemorations (e.g., Ben-Amos, 2000). Contra Hoffman and Moya, however, we note that this claim does not need to presume any group-level mechanism: a plausible alternative is that elites have an intuitive understanding of human social cognition and have a direct incentive to publicly display historical myths to recruit coalitional support.

More generally, we disagree with the claim that coordination is the sole driver of coalitional choice, and therefore that what individuals decide to coordinate on is arbitrary. Consistent evidence from across the social sciences shows that coalitional choice in humans (including in large natural groups) depends on a variety of variables – all of which are indicators of the potential fitness costs and benefits associated with group membership: shared goals (Noyes & Dunham, 2017), common threats (Barclay & Benard, 2020), expectation of reciprocity (Yamagishi & Kiyonari, 2000), institutional quality (Bartoš & Levely, 2021), reputation circulation (Habyarimana et al., 2007), and, we suggest, perceived fitness interdependence (e.g., Swann et al., 2014). In psychology, the notion that mere coordination is sufficient to establish group boundaries is reminiscent of the minimal group paradigm – in which researchers found that mere allocation to an arbitrary group was sufficient to elicit in-group favoritism (Tajfel, 1982). But precisely, subsequent studies have found that in-group preferences in minimal groups readily disappear when other sources of incentives are introduced (e.g., reputational incentives and reciprocal cooperation) (Balliet, Wu, & De Dreu, 2014). Lastly, there are abundant examples of failed nation-building – cases where people refuse a given coalitional arrangement despite strong coordination cues from state-level public propaganda (see sect. 2).

These elements demonstrate that pure coordination is insufficient to establish group boundaries in individual minds. Our suggestion is that while coordination is important, individuals want to coordinate on coalitional arrangements that are otherwise desirable. This explains why cultural technologies are not arbitrary, but tend to exhibit recurrent cross-cultural (not necessarily universal) patterns – as those of historical myths.

R6.3. Andreas Wimmer's theory of nation-building

Wimmer wrote an especially detailed critique of our target article. His commentary explores a promising avenue for testing our theory using large cross-country datasets. It also gives us an opportunity to discuss the relevant measures to do so. Indeed, the “foundation year” variable that was used as an outcome variable by Wimmer in his regression model is not suited to test our main hypotheses because it was designed to measure the objective date of birth of an ethnic group. To collect this data, Kaufman asked experts when they thought that an ethnic identity first emerged and explicitly noted that he was not “concerned with the group's own claims, which often stretch back much further than what is warranted by the historical record” (Kaufmann, 2015). Yet, it is precisely the “group's own claims” on the antiquity and continuity of their nation that matters for our theory. The target article is very clear that historical myths are mental representations of the past that are likely to diverge from expert opinions. Future studies using large datasets should select an outcome variable that measures group member's subjective rather than objective representations of their history. Collecting subjective data will also shed light on intra-group variability in the endorsement of historical myths – as we expect that not all group members hold the same representation of their past (see sect. 6.1).

We actually provide such data in the main article – despite the commentary claiming otherwise. We cite numerous psychological experiments showing that people readily endorse the belief their nation has an ancient and continuous history; and that this endorsement is significantly higher in people who identify with their nation more, who are exposed to information about the disappearance of their group, and who are more exposed to threat cues (see sects. 4 & 5). However, we agree with Wimmer that this data is imperfect, relies too heavily on Western samples, and needs to be tested outside the lab.

Finally, Wimmer referred us to his own “modernist” account of nationalism (Wimmer, 2018). We believe that our respective accounts are quite compatible in that they both assume that successful nation-building is premised on the quality of intra-group cooperative exchanges. One key difference in our approaches is that Wimmer (2018) emphasizes the relationship between citizens and the state, which we also recognize as an important driver of coalitional choice (see R1 & R2) – whereas our target article also emphasizes the (perceived) horizontal bonds between individuals. Another difference lies in our relative disciplinary focus. As a political scientist, Wimmer (2018) is interested in the objective determinants of successful nation-building (i.e., public goods provision), whereas psychologists are interested in individuals' subjectivity (e.g., perceived public goods provision). Integrating the two disciplines in the study of nationalism is crucial as what ultimately drives behavior is what people perceive, and how this information is processed in our minds.

R7. Promising avenues for future research

R7.1. Positive and negative historical narratives

Finally, several commentators suggested promising avenues for extending and broadening our research work. A first line of investigation for future studies is to differentiate historical myths and their psychological impact based on their valence. Indeed, some authors questioned our decision to focus on historical myths that describe positive past interactions, when so many national narratives are based on national traumas. For instance, Wimmer cites the historical “tales of victimhood, defeat and injustice” that prevail in Polish nationalism and Zionism. Bilewicz and Bilewicz make a similar point, although they note that myths of shared suffering may in fact reduce national cohesion. These observations are consistent with historical research showing the prevalence of negative themes (e.g., crushing military defeats) in nationalist rhetoric; and with psychological evidence that shared negative experiences are more likely to generate identity fusion than shared positive experiences (Whitehouse et al., 2017).

Here, we propose a tentative explanation for this apparent paradox, although more research is needed to support our claims. We suspect that, generally, these disastrous narratives are not displayed for their own sake. Rather, they are typically used as a narrative device to highlight the heroic self-sacrifice of ancestors (e.g., Serbian nationalism: Bieber, 2002; Lomonosov, 2021; in Israeli nationalism: Ben-Amos, 2003; Gal, 2009). A striking example is how the siege of Masada – during which a battalion of Jewish soldiers self-sacrificed to fight Roman oppressors in 73-74 CE – became a central symbol of Israeli nationhood despite being a military defeat. The story of Masada was explicitly used to “inspire many Israelis and Jews to greater heroism and self-sacrifice” (Smith, 1999, pp. 179–180). The phrase “Masada shall not fall again” has become a rallying sign for Israeli nationalists, and Masada remains one of the most visited places in Israël, showing the prominence of this narrative (Gal, 2009; see also Ben-Yehuda, 1996).

Two evolutionary mechanisms may explain the appeal of tales of heroic self-sacrifice for nationalist coalitional recruiters. First, the public celebration of group members who risk their lives for the nation sends a powerful signal that such behavior is socially valued and will be rewarded. In certain conditions (e.g., when death probability is compensated with even greater potential benefits) these incentives can motivate extreme pro-group behavior (Dessalles, 2018). This idea was confirmed in a recent model, showing that a propensity to pay tribute to self-sacrificial heroes might coevolve with the presence of heroic acts in a social group (Dessalles, 2024). A second possibility is that martyrological narratives serve as a hyper-stimulus of costly helping (Barclay, Bliege Bird, Roberts, & Számadó, 2021). By observing costly self-sacrifice from other coalition members, recipients can infer that they are highly valued by other members and therefore, that the fitness interdependence that binds them is high (Barclay et al., 2021). In our target article (sect. 4), we suggested that cues of repeated interactions over time might be perceived as having a greater impact on perceived fitness interdependence if these interactions were especially costly to participants. Thus, by emphasizing heroic acts of costly helping in the past, coalitional recruiters may hope to reinforce the commitment of group members.

R7.2. Are historical myths detrimental to social cohesion?

These hypotheses address Wimmer's observation that nationalism is often founded on tales of victimhood, but do not explain Bilewicz and Bilewicz's remark that endorsing such tales has deleterious effects on social cohesion. In particular, these authors cite evidence that endorsement (or exposure) of traumatic collective memories are associated with more paranoia, conspiratorial beliefs and obsession with treason – which might ultimately damage social cohesion. These observations are consistent with the view of Kaufman, Kashdan and McKnight who also stress the negative psychological effect of historical myths – although their focus is on positive narratives. In strengthening the cohesion of the majority group, these authors argue, positive historical myths might contribute to the stigmatization of marginalized groups and dissenting in-group members – and stifle their creativity.

In fact, these psychological responses to historical myths – positive or negative – are not incompatible with their potential positive effect on national commitment. Recall that the output of coalitional choice psychology is not directly cooperation, but the mental representation that a given coalitional arrangement is desirable – leaving the possibility that people might shirk their national obligation. When the incentives and consequences of cheating are especially high – as it is usually in times of national trauma – we might actually expect that commitment to the nation will translate into a greater motivation to monitor and sanction other group members. This probably manifests psychologically in the form of paranoid thinking, conspiratorial beliefs or exclusionary attitudes toward dissenters (Greenburgh & Raihani, 2022). In line with this idea, attachment to the homeland and nationalism are positively correlated with Right-Wing Authoritarianism – showing that people who support a given coalitional arrangement also support more stringent social control to maintain its boundaries (Osborne, Milojev, & Sibley, 2017). Future research might further investigate the negative traits associated with the endorsement of historical myths.

R7.3. Historical myths for attack and defense

Third, Snijder and De Dreu raise an interesting debate on the contextual variability of historical myth prevalence. Our target article predicted that historical myths would spread more in societies engaged in warfare because group members need more costly investment from each other (sect. 5.1), but it remains unclear whether this prevalence should vary between attack and defense situations. Snijder and De Dreu convincingly show that group members are more interdependent in defense than in attack and conclude that historical myths will be more prevalent in attack situations. We generally agree with this logic but introduce a possible caveat. Defense situations are indeed more prone to fitness interdependence between group members, but they also make cooperation failures more costly – and possibly dramatic. The severe cost of losing against invaders might incentivize group members to transmit historical myths. More research is needed to determine which of these two opposite effects trump the other and how this affects the circulation of historical myths.

R7.4. Toward a psychological science of the historical discipline?

Finally, our research might open broader research question related to the psychological appeal of history in general. In their commentary, Pietraszewski and Moncrieff convincingly demonstrated that the cross-cultural interest of humans for history represents a puzzle for social science beyond the study of historical myths. Interest for history, these authors argue, recycles important cognitive templates like the ones involved in retaliatory behavior. We certainly concur with the author's call for an adaptationist cognitive science of history and hope to see future studies addressing these questions.