R4.1. Beliefs grounded in experience

Because cognitive systems work smoothly in the background of our conscious mental life, their operation is in many cases entirely hidden from conscious inspection. This favors the emergence of what could be called “cognition-blindness” or spontaneous realism, the view that there is no need to assume a complex architecture of neurocognitive systems, to figure out what is simply “out there” in our experience. But cognition blindness is of course highly misleading. Even parsing a simple sentence, or seeing a scene in three dimensions, requires a vast amount of computation.

That is why there may be limits to explaining FEBs in terms of direct physical experience, as Lee & Schwarz propose. True, many abstract views about the economy have physical counterparts – for example, zero-sum thinking is analogous to balancing weights, or physical labor seems more “real” than intellectual work because it is grounded in proprioception. And abstract thoughts gain salience by being couched in terms of physical metaphors, many of which become conventional metaphors (Lakoff & Johnson Reference Lakoff and Johnson1980). But there are many distinct types of physical experience, and FEBs are associated with only some of them. Seeing the economy as zero-sum does bring up imaginative associations with intuitive physics, but the workings of the economy could also be construed in terms of other forms of experience, for example, hunger or walking or sexual desire. That we see the economy in terms of balancing weights, suggests prior intuitions about exchange that subsequently make the analogy intuitively compelling, more so than other possible metaphors.

R4.2. Domain-general properties of cognition

We emphasized the contribution of domain-specific learning systems in the acquisition of explicit, reflective beliefs. One alternative to our proposal would be that specific FEBs in fact result from domain-general properties of human cognition, as several of our commentators (Bhattacharjee & Dana, Lee & Schwarz, Leiser & Shemesh, and Ross) argue. In that view, human cognition is characterized by general operating constraints, for example, processing capacity and speed, size of working memory, and so forth, and by highly general operating principles, such as association and reinforcement, or high-level processes, such as approximations of Bayesian belief-revision (McNamara & Fischer). Naturally, we agree that such properties of human cognition exist and matter. The point of domain-specific models is not to deny the operations of general constraints, but to examine how the constraints differ from domain to domain (Hirschfeld & Gelman Reference Hirschfeld and Gelman1994a).

For example, one may consider that the complexity of economic processes, what Leiser & Shemesh describe as “relational complexity,” poses some general challenges to human cognition, and that simplified pictures of processes will be favored in the process of social transmission (Hirshleifer & Teoh). As Leiser & Shemesh point out, processes of osmosis or even simple selectionist arguments are difficult to entertain without sustained effort or special training. It is quite clear that (in the absence of systematic training), people seldom if ever think in terms of populations, preferring to think of social matters in terms of typical individuals (Leiser & Shemesh) or of generic agents, which leads to many misunderstandings of social phenomena (Boyer Reference Boyer2018, pp. 222–26). For example, we know that very small differences in the distribution of individual racial preferences in housing can lead to dramatic differences in the resulting pattern of housing segregation (Schelling Reference Schelling1971), but that kind of population thinking is unintuitive and effortful. In the same way, describing policies in terms of how they impact “the employers” versus “the employees” or “the rich” versus “the poor” ignores variance in preferences within these categories, when such variance is crucial to understanding the effects of policies – and is crucial to formal models of political choice (Munger Reference Munger2015, pp. 58–78). And, as Jern argues, other domain-general properties of imagination might influence the content of FEBs, among which is a tendency to construe close and distant reality in a qualitatively different manner (Liberman & Trope Reference Liberman and Trope2008).

While agreeing that memory and storage constraints impact the activation of representations and their inferential role, we would also contend that such constraints are insufficient to explain the particular contents of beliefs about the economy. Contrary to what an account in terms of cognitive load or constraints would predict, we do not observe a random set of uninformed or disorganized beliefs about the economy. The beliefs go in specific directions – which is why they are sometimes described as the outcome of specific “biases” (Caplan Reference Caplan2006). In our model, the most relevant biases observed in the transmission of FEBs consist of what Acerbi & Sacco describe as “content-based biases,” to adopt the terminology of Boyd and Richerson (Boyd & Richerson Reference Boyd and Richerson1985; Mesoudi Reference Mesoudi2007). That is, the diffusion of particular mental representations is a function of their content, rather than features like their frequency per se or the prestige of their sources (although such biases might also be relevant).

We agree with Bhattacharjee & Dana that some descriptions of such content-biases are very close to our own model. That is, for instance, the case for Caplan's description of an “anti-foreign bias” (Caplan Reference Caplan2008) on the one hand, and what we identified as an effect of coalitional psychology on the other. We prefer the latter explanation, because (a) there is already independent evidence for the workings of a coalitional psychology (Delton & Cimino Reference Delton and Cimino2010; Pietraszewski Reference Pietraszewski2016; Pietraszewski et al. Reference Pietraszewski, Curry, Petersen, Cosmides and Tooby2015; Tooby & Cosmides Reference Tooby, Cosmides and Høgh-Olesen2010), and (b) coalitional psychology explains why we would expect an anti-foreign bias, why it would focus on people and firms from other ethnic-national groups rather than other possible social categories. So, it is parsimonious to interpret the anti-foreign trade bias as one among the consequences of an already observed psychological mechanism.

We also doubt that folk-economic beliefs could be adequately explained by a tendency to project personal experience onto macro-social processes, as Ross suggests, mostly because of the problem of under-determination described above. Ross is certainly right that some macro-economic beliefs may extend aspects of private domestic economy (e.g., do not spend more than you earn) to a country's balance of trade, an analogy that many economists find deeply misleading. But this scaling up of ordinary experience would not explain other aspects of folk-economics, for example, the well-documented tendency to believe that it is better to export things and get money, rather than export money and get things (see our comments on Johnson below), or the idea that it is somehow bad for a country to export services and import manufactured goods (Wood Reference Wood2002, p. 22). Such beliefs are much more specific than a mere extension of domestic economics would warrant.

In a similar way, it may not be optimal to explain people's views of the economy in terms of a general bias towards anthropomorphism. It is probably the case, as Forstmann & Burgmer suggest, that representations of the economy as an agent are activated in the course of reflecting on economic processes. Journalists will, for instance, say that “the economy is recovering from the trauma” (i.e., production is increasing after a recession), or “the markets are still shy” (there are more bears than bulls). But anthropomorphism is a feature of the beliefs we observe; it cannot be by itself an explanation for the recurrence of these beliefs. It seems that agent-like descriptions of the economy are more attention-grabbing and potentially transmitted than agent-free representations for two reasons – because we have a rich intuitive psychology to start with, and because some aspects of economic processes may be similar to the input format for that intuitive psychology. Indeed, our rich intuitive psychology is often activated in the description of systems characterized by (a) high complexity, (b) recursive processes, and (c) consequences in people's actions (Boyer Reference Boyer2018, pp. 217–28). That construal of complex social processes in terms of “big agents” lends itself to a narrative format, as Acerbi & Sacco rightly emphasize. This may impose a particular set of filters on what become popular views of the economy. For example, narratives require that events have reasons rather than just causes, and that there should be a unity of plot in our description of an overall system.

R4.3. Is there a general “zero-sum” perspective?

Another potential candidate for a general bias that would influence folk-economics is a general view of society as a zero-sum process. Our model described this in the context of international trade, but the assumption is not in fact limited to nations (Johnson remarks that it also occurs between regions – our description was unduly narrow). In fact, the zero-sum assumption appears in many familiar views of the economy, for example, that the wealth produced in a country is a “fixed pie,” or that the prosperity of some countries depends on the impoverishment of others. It also feeds more detailed views about the economy (Bhattacharjee & Dana; Rubin).

How do we explain the recurrence of this assumption? It is quite clear that in many domains of social interaction, outside economic exchange, our ancestors, like most other organisms, evolved in a zero-sum world. Our evolved psychology certainly includes the intuition that mates, territories, foraging locales, or hunting grounds are rival goods. It is also possible that the mind entertains a general zero-sum template, perhaps grounded in physical experience (Lee & Schwarz), or in the simple experience of everyday bookkeeping (Ross), that is then projected onto unusually complex systems like the economy as a whole, so that, for example, the high incomes of the rich seem to cause lower incomes among the poor.

These latter explanations are not altogether satisfactory, however, for reasons already mentioned. Folk-economic beliefs do not just produce a (distorted) view of the economy by emphasizing some features of exchange – they also ignore equally real features. Consider, for instance, Bhattacharjee & Dana's proposal, that people see the economy as a whole as zero-sum, because they extrapolate from an actual feature of a single trade: the division of profit between seller and buyer. As a result, the positive incentives of profit are invisible. They are certainly “unseen,” to use Bastiat's term. For example, people see surge-pricing as “gouging”; they resent the price increase, but do not spontaneously see price hikes as a signal that will motivate agents to increase the quantity supplied. Bhattacharjee & Dana's hypothesis is plausible and should be investigated experimentally. But we would point out that other, equally real aspects of trade are not picked up in the construction or adoption of FEBs. For example, profit is certainly zero-sum in a single transaction, but it is also positive on both sides – that is, both parties benefit from a trade. The former aspect may form the basis of FEBs, as our commentators point out, but the latter does not. It is indeed possible that people do not actually represent trade as mutually beneficial, as Johnson suggests (see discussion in sect. R8.2). In our view, these intuitions about trade would then combine with the intuition that a large market of mostly impersonal agents is a potential threat. Whether this is a valid interpretation is of course an empirical matter that may require more studies of the kind that Bhattacharjee & Dana and Johnson themselves pioneered.

A similar psychological enrichment may be necessary if we want to explain the salience of the additional FEBs described by Rubin. For example, although his proposed FEB 10 (“Labor market regulations … have no impact on levels of employment”) seems to be directly entailed by many policy proposals and their translation in popular opinion, it seems to us that, in psychological terms, the impact of regulation on quantities supplied and demanded is simply not computed at all. A similar point could be made about Rubin's FEB 11 (“A society can reduce or eliminate inequality with no adverse impact”). Again, this may seem logically entailed by many policy proposals. But that implication is, in our view, not represented at all. What is represented, in this case, are generic descriptions of the poor and the rich, together with a quasi-hydraulic transfer of resources from one to the other, accompanied by an emotional, ethical appraisal of the putative effect. The computation stops there.

This aspect of folk-economics is crucial. When domain-specific intuitive systems are activated (e.g., in parsing sentences, in the construction of 3D visual scenes from 2D retinal images, or in computing someone's attractiveness on the basis of their looks), they respond to a specific input format and perform principled inferences, and stop there. For instance, the parsing system does not produce a representation of the speaker's meaning – that task is handled by other, distinct systems further downstream in information processing. In a similar way, in our proposed explanation of folk-economics, when we represent the rich and the poor, this representation meets the input format of a fairness-estimation system that automatically searches for information about these generic agents' contributions to the economy seen as a collective action. But that fairness-estimation system is not in the business, so to speak, of seeking information about what caused previous supply, or about the incentives for future supply.

In our view, it is only by understanding that folk-economic beliefs are meta-representational, and that they activate domain-specific intuitive systems, that we can make sense of the apparent lapses in logic documented by Rubin. What is not handled by intuitive systems is bound to remain, barring sustained intellectual effort, in Bastiat's domain of “what is not seen.”

R5.1. Development and the acquisition of FEBs

Interaction between domain-specific expectations and relevant information from a social environment is necessary both (a) to develop an adult inference system and (b) to account for the way that inference system tends to modulate people's explicit beliefs about the economy (Liberman & Kinzler). Social essentialism, involved in some aspects of our coalitional psychology, provides an excellent illustration of these two processes. Essentialism is the general label for a variety of inferences that associate membership in a category (e.g., tigers, Germans, conservatives) with (a) an internal (and generally undefined) quality present in all exemplars, (b) the notion that this internal essence is inherited rather than externally caused, and (c) the assumption that the internal quality explains external features common to the category (Gelman Reference Gelman2004). Such essentialism is most common in representations of biological kinds (Hirschfeld & Gelman Reference Hirschfeld, Gelman, Atran and Douglas1999), but also of some social categories (Gil-White Reference Gil-White2001). However, there is an important difference here. Essentialist notions about cats and dogs develop early, do not require explicit reflection, and indeed are resistant enough to hinder the learning of scientific biology (Atran Reference Atran1998). By contrast, essentialist notions about social groups seem to consist mostly of explicit reflections on one's prior intuitions of group membership (Hirschfeld Reference Hirschfeld, Hirschfeld and Gelman1994). That is, one develops a strong intuition that other-group members must somehow be different from “us,” before hitting on their supposed essence as the explanation, or acquiring that explanation from others (Boyer Reference Boyer, Margolis and Laurence2015, pp. 38–45, 256–61). In other words, what is mostly intuitive in the biological domain seems a matter of reflective meta-representations in the social domain. That may be why, as Liberman & Kinzler emphasize, essentialist representations of social categories vary as a function of exposure and of the coalitional context – children even vary in their understanding of “race” as a fixed trait (Roberts & Gelman Reference Roberts and Gelman2016).

Developmental evidence may help us specify and perhaps modify the very broad outline proposed in our target article, in terms of the cognitive systems involved. For instance, some economic beliefs seem rooted in strong intuitions about sharing and ownership. There is an obvious connection between intuitions about communal sharing, on the one hand, and widespread beliefs about the economy as a commonwealth, as something that is “ours” only, on the other (Sheehy-Skeffington & Thomsen). This latter representation activates intuitions about group-membership, about the proper extension of the category of individuals that should count as “the same.” As Nancekivell & Friedman point out, even young children seem to deploy a sophisticated understanding of group ownership. Indeed, it seems that from an early age children understand that ownership is not just in rem, connecting a person and a thing, but also in personam, modifying interaction between agents – specifically, making it possible for one agent to interfere in another one's access to resources (Boyer Reference Boyer, Margolis and Laurence2015). And the persona in question can indeed be a group, which makes evolutionary sense as nomadic foraging, characteristic of most of our evolutionary past, involves territorial claims and competition for group access to resources such as hunting grounds (Kelly Reference Kelly1995). Nancekivell & Friedman's description of early notions of group ownership would also account for the spontaneous development of commons-management institutions in many different human groups (Ostrom Reference Ostrom1990). As we and Sheehy-Skeffington & Thomsen have emphasized, group ownership representations are often informed by coalitional representations. That is precisely why ownership of common resources in modern economies can be particularly contentious. Unless there are clear ecological boundaries to groups, it may be difficult to decide, for instance, whether a particular natural resource, such as oil or a natural park, belongs to the nation, the region, the ethnic group, or some other community.

Developmental dynamics are crucial to understanding the acquisition of some “emporiophobic” beliefs. As noted by Echelbarger, Gelman, & Kalish (Echelbarger et al.), there is considerable evidence for the early development of motivations both to maximize utility and to extend unconditional cooperation. While one system allows developing minds to make sense of mutually profitable trade, the other one imposes boundaries on desirable social outcomes. Even in children, the assumption that distribution will be proportional to contributions is limited by the requirement of a minimum level of resources for all, an assumption that modulates the acquisition of beliefs about social welfare systems (see discussion of ideologies in sect. R6.1).

R5.2. The influence of moral psychology

Folk-economic beliefs are very often moralized, as discussed by Baumard, Chevallier, & André (Baumard et al.), Bhattacharjee & Dana, Chowdhury & Yu, Malhotra, and Sznycer, Ermer, & Tooby (Sznycer et al.). Given that human moral psychology is best described as the product of highly specific selective pressure (Baumard et al. Reference Baumard, André and Sperber2013a; Reference Baumard, André and Sperber2013b; DeScioli & Kurzban Reference DeScioli and Kurzban2013; Haidt & Joseph Reference Haidt and Joseph2004), it makes sense to consider how moral intuitions and economic reflections interact in the acquisition and transmission of FEBs.

First, the activation of moral psychology systems explains why some economic questions are more likely than others to be attention-grabbing (Sznycer et al.). The question of relative pay levels between CEOs and employees meets the input conditions for cognitive systems geared to fairness in exchange, as well as cheater-detection. These systems seem to bear the traces of their environment of evolutionary adaptedness. For instance, they assume that people who are in exchange will meet again (Krasnow et al. Reference Krasnow, Delton, Tooby and Cosmides2013), which makes it potentially costly to ignore the needs of the least-favored. Also, those systems evolved in ecologies with a low ceiling on productivity and therefore on differences in productivity, compared to modern market economies. Although we tend to accept differences in allocations motivated by differences in contributions, we are generally reluctant to extend this beyond relatively small differences in outcomes. As a result, considerations of fairness often dominate policy debates and obscure questions of efficiency and final outcomes, for example in the evaluation of tax systems (Rubin).

Also, moral psychology may help us clarify the set of beliefs associated with emporiophobia, which are probably more specific than we suggested in our target article. For example, as Malhotra points out, considerations of harm may modulate folk-economic beliefs about the social benefits of markets – people consider that auctions are fine for concert tickets but not for organ transplants, because they consider the harm against a single patient in a deontic manner. That is indeed why aversion to market solutions is not limited to opinion about financial markets, as Miton & Sperber suggest (see Fiske & Tetlock [Reference Fiske and Tetlock1997] and Brennan & Jaworski [Reference Brennan and Jaworski2015] for many other examples). As Chowdhury & Yu propose, the impersonality of market transactions is especially salient when it concerns items for which we assume that personal information about the partners involved will be crucial, which is why the idea of an auction of orphaned infants seems repugnant. These suggestions, if we can follow them up with appropriate empirical studies, could lead to a more refined description of what we described in broad strokes as generalized emporiophobia.

Moral psychology triggers motivation through emotional arousal. We do not think that this constitutes an alternative to a description in terms of specialized cognitive systems, as claimed by Chowdhury & Yu, who add that such an explanation is of “greater efficacy.” Our point is that emotions are part and parcel of the operation of evolved cognitive systems. For instance, the identification of free-riders depends on fine-grained representations of who contributed what and when, whether those who contributed less could have given more, and other such criteria (Delton et al. Reference Delton, Cosmides, Guemo, Robertson and Tooby2012). It is part of the design of our anti-free-riding system that it leads us not just to categorize free-riders but also to entertain emotions that lead us to avoid them, to denounce them, and to diminish their welfare (Price Reference Price2005). In more general terms, emotions are involved because access to emotional arousal is one of the ways in which an evolved computational system can direct behavior.

A radical alternative to this model is to explain most beliefs about the economy, as well as most explicit beliefs and norms about collective life, as the output of a single fairness system, as proposed by Baumard et al. In that view, the only cognitive system engaged in many folk-economic beliefs would be one that is geared to optimizing fair allocations (Barclay & Willer Reference Barclay and Willer2007; Baumard et al. Reference Baumard, André and Sperber2013a; Noë & Hammerstein Reference Noë and Hammerstein1994). Folk-economic beliefs are often compatible with intuitive fairness criteria. Indeed, this fit with intuitive fairness is an important factor in their acquisition and transmission.

But there are computational problems with the proposal that FEBs, and in fact most components of people's representations of politics, are influenced by a single intuitive system. It is of course an empirical matter, whether one or two or more distant systems are involved in a particular domain of representations or behavior. In general, it makes sense to consider a system as relatively modular, if (a) it handles only specific aspects of the available information, just like an enzyme binds only to particular kinds of molecules (Barrett Reference Barrett2005); (b) it operates on specific inferential principles; and (c) we can detect a specific developmental path. (Other criteria, such as anatomical localization or neuropsychological dissociation, are sometimes relevant; Hirschfeld & Gelman Reference Hirschfeld, Gelman, Hirschfeld and Gelman1994b). Seen in those terms, it seems unlikely that a single fairness system would handle all of the computations that lead to the intuition that something is unfair. For example, it seems intuitively unfair (a) to appropriate the cake someone else just made, and (b) to beat people up because they disagree with you. The “unfair” label results from representations of a gross imbalance between the welfare of two agents, violating principles of mutualism. But those representations of welfare are delivered by different systems, one of which links labor invested to entitlement (as described in our target article; see also commentary by Nancekivell & Friedman), whereas the other one compares the levels of harm caused by disagreements and physical violence. So it seems more sensible to accept that fairness-judgments are delivered by a system whose inputs are generated by other dedicated systems (including a representation of differential welfare and respective contributions of different agents). On the basis of these descriptions, the system for fairness-judgments then produces deontic judgments and associated emotions.

R6.1. The role of elites and the limits of propaganda

It may seem straightforward that people acquire folk-economic beliefs from elite agents, such as political entrepreneurs and political parties, and we certainly concur with Bisgaard & Slothuus that political parties in modern societies are key providers of information. But we should add that political parties share the role of information providers with many other sources such as the media, experts, social media platforms, and interpersonal discussion partners. Indeed, it seems increasingly plausible that social media in particular provide information that media organizations and political parties try to follow and accompany, rather than the other way around. More importantly, although all of these information sources are relevant for understanding FEBs, variation in elite behavior is not sufficient to empirically explain variation in FEBs. We should use a broader set of theoretical tools than the notion of top-down influence, when we try to understand the salience of particular beliefs.

With modern information technology, political parties and other elite actors emit information all of the time. However, only a fraction of that information is likely to “stick” in the minds of citizens. Indeed, we should think of and analyze the mind as part of the selection environment for information. In this perspective, it seems that information will have a higher “survival rate” – that is, it will be acquired more easily, and stored and communicated more frequently – if it meets input conditions of some of our intuitive systems. Investigating the degree of fit between elite rhetoric and evolved cognitive systems should, in other words, help explain why particular pieces of rhetoric are more effective than others. Malhotra provides an excellent illustration of this point when he refers to the widespread notion of “welfare queens.” In our view, the reason why this particular motif became so successful is a consequence of its fit with evolved cognitive systems for exchange and, in particular, systems for cheater-detection.

As we develop below in section R6.2, within our evolutionary-cognitive framework, political elites provide information in the sense of (a) coalitional signals that elicit cognitive systems for coalitional coordination among partisans; (b) cues that, depending on their specific content, can elicit the entire suit of intuition-producing human cognitive systems, including systems for sharing or cheater-detection; and (c) factual information that people can build particular FEBs with on the basis of their intuitions. To sum up, it is crucial to describe the cognitive systems and intuitions that these informational pieces tap into, in order to understand the power and limits of elite information.

R6.2. Why partisan politics? The role of coalitional psychology

Describing the way individuals filter and modify top-down information naturally leads to the question: Why do people follow a particular political party? That is, why do they not just receive information from parties, but also favor information from one particular source, and consider their interests as closely bound to those of a particular political organization? As rightly emphasized by Bisgaard & Slothuus, people pay enormous attention to the rhetoric of their favored political party, creating – as observed by Ruisch, Anderson, & Pizarro (Ruisch et al.) and Buturovic – systematic partisan and ideological variation in FEBs.

As discussed by McDermott, humans are by nature a coalitional species and have massively benefited in terms of fitness from being part of coalitions that amplify their bargaining strength. As a consequence, a core ability for humans is to coordinate agendas within coalitions – that is, to agree implicitly (or, sometimes, explicitly) on a prioritized list of problems to tackle by means of collective action. When people come to identify with a particular coalition, the stance of the other members will, in other words, matter for the beliefs that people adopt, including their FEBs.

In our view, the underlying set of psychological mechanisms – that is, evolutionary coalitional psychology – provides the best available ultimate explanation for partisan variation in FEBs, and for the proximate mechanisms often described as “partisan” or “motivated directional” reasoning within political science (see Leeper & Slothuus Reference Leeper and Slothuus2014). As we noted in the target article (sect. 6.4), empirical evidence from carefully tailored experimental designs shows that modern political parties tap into the slots of human coalitional psychology and that political parties are psychologically represented as an instance of the mental category of “coalitional alliance,” along with other instances such as nation or ethnicity. That is why we agree with Karabegović, Rotella, & Barclay (Karabegović et al.) as well as McDermott that cognitive systems for coalitional coordination are crucial for understanding FEBs, and that the effects of such cognitive systems reach beyond the domains mentioned in our article, immigration and trade in particular.

Coalitional psychology explains how the rhetoric of political parties can significantly shape political attitudes and perceptions (including FEBs), and how partisan and ideological considerations often overpower other types of considerations, as demonstrated by a great deal of empirical research in political psychology, mentioned by Bisgaard & Slothuus, Buturovic, Ruisch et al., and Tappin et al. In our view, coalitional intuitions and inferences play a crucial role in generating these effects, including their affective nature (Petersen et al. Reference Petersen, Skov, Serritzlew and Ramsøy2013; Reference Petersen, Giessing and Nielsen2015) and the fact that that conflict increases their strength (Druckman et al. Reference Druckman, Peterson and Slothuus2013).

It should be noted, however, that there are limits to the power of coalitional psychology and, hence, to the role of partisan and ideological considerations. One obvious limitation is that coalitional psychology has specific input conditions, for example, cues of group-based conflict. When available information does not meet these input conditions, we should not expect coalitional psychology to have strong effects. Also, other cognitive systems may compete with coalitional considerations to determine attitudes and behaviors. For example, as discussed in the target article (sect. 5.2), when people are confronted with clear cues that a welfare recipient is a cheater or a reciprocator, these cues completely override partisan considerations (Petersen et al. Reference Petersen, Sznycer, Cosmides and Tooby2012).

A less-obvious way in which other cognitive systems constrain the power of coalitional psychology relates to persuasion. In the target article, we focused on how FEBs affect political attitudes. A coalitional psychological perspective, however, implies that the causal arrow between belief and attitude sometimes will be reversed. At times, processes of coalitional coordination will imply that we agree to a position of a fellow coalitional member and then, in a post hoc manner, construct a set of FEBs around this position. As Karabegović et al. note, we can think of coalitional considerations as (broadly defined) self-interested considerations. Past research shows that self-interested considerations operate within the limits of persuasion; that is, self-interest only shapes moral positions to the extent that the position can be justified with reference to principles other than self-interest (DeScioli et al. Reference DeScioli, Massenkoff, Shaw, Petersen and Kurzban2014; Kunda Reference Kunda1990). When people adopt certain positions on economic policies for coalitional reasons, they will, in other words, feel a need to justify this positions on the basis of particular beliefs about the operations of the economy.

R8.1. Is it possible to counter misleading FEBs?

Although the research program outlined in the target article does not include an inherent normative agenda, we agree with an important point made by Kam: FEBs might have normatively undesirable consequences. Some FEBs hinder the pursuit of egalitarian economic policies, others the pursuit of free-market-oriented policies. Depending on one's political values, either set of FEBs might appear undesirable. So it is relevant to ask whether there are ways to counter the intuitive emergence of particular FEBs. Arceneaux emphasizes the role of reflection to counter intuitively produced inferences, while Acerbi & Sacco emphasize the role of counter-narratives. Both perspectives have merit and, in fact, they can be fruitfully integrated.

In our model, this would imply prompting individuals to direct their attention to relevant experiences that resonate with normatively desirable goals and encourage them (through increased motivation or capacity) to process the relationship between these experiences and the problematic FEBs. For example, as we stress in the target article, people might have repeated experiences of smooth market transactions, yet maintain the FEB that markets in general are exploitive. This is made possible because the FEB is cognitively situated in relative isolation from the experiences – that is, the FEB is produced not from personal experience but from a range of general cues about marks, for example, impersonal transactions. In essence, by activating memories of personal experiences, one set of intuitions (related to plus-sum exchange) operates as counterweights to another set of intuitions (related to the risk of exploitation in anonymous situations).

As we noted elsewhere (Boyer & Petersen Reference Boyer and Petersen2012), our evolved intuitions do not construct an iron cage, but a rubber one. When sustained effort is exerted, our intuitions can be bent out of their natural shape. Therefore, an important research agenda not just for the next generation of studies on FEBs but for the next generation of evolutionary cognitive psychologists more generally, is to understand the ultimate causes and proximate outline of the mechanisms that in this way allow the human organism to prioritize one set of evolved intuitions at the expense of another. This should pave the way for concrete studies on how normatively undesirable FEBs (as defined by one's particular normative agenda) can be countered.

R8.2. The connection to economic psychology

In our original article, we chose not to elaborate on the connections between folk-economic beliefs, on the one hand, and our actual economic psychology, the set of cognitive mechanisms activated when people actually engage in transactions, on the other. This seemed a prudent move, mostly to avoid the temptation to see the “economic mind' as a unitary part of our cognitive makeup, a set of coherent assumptions that would inform all of our representations about economic matters. We also suggested that the connections between FEBs and economic psychology may be tenuous at best, both because reflective beliefs are only weakly constrained by intuitions, and also because most FEBs are about macro-economics. Several commentators, however, suggest that it may make sense to consider the interaction between the two domains of representations because the intuitive systems involved in supporting economic behavior may influence the contents of FEBs directly enough that it would be strange to ignore that effect.

Our economic psychology is a barely explored set of cognitive systems. A persistent problem is that it is mostly studied in terms of its divergence from normative economic theory. But that is not always adequate. For one thing, as Jern remarks, people's intuitions about other agents' behavior are often in broad agreement with normative economic theory, although people's own behavior is not. More generally, a number of economists have argued that considering human evolution, and particularly our evolved psychology, is indispensable if we want to understand apparently prosocial behaviors, and other deviations from normative economic theory (Capra & Rubin Reference Capra, Rubin and Saad2011; Koppl Reference Koppl2005; Robson Reference Robson2001).

There is unfortunately very little experimental research on the actual cognitive processes involved in valuation and exchange, as Johnson points out. For instance, every microeconomic textbook points out that both parties in a voluntary exchange benefit from a trade. The very fact that this notion is unintuitive, and needs to be spelled out, is revealing. Indeed, empirical studies by Johnson et al. show that people's intuitions are at variance with the theory, as they consider that many transactions are ”win or lose,” that in general sellers are better off after a transaction but buyers are not (Johnson et al. Reference Johnson, Zhang and Keil2018b). If people are indeed intuitive mercantilists, this suggests that cheater-detection systems may be set off in contexts in which we act as buyers, because the amount of the benefit (the good purchased) will always appear slightly less than the amount of the cost (the money paid).

Such non-normative responses suggest that people activate notions of value that diverge from microeconomic models of utility. Nancekivell & Friedman suggest that there may be a form of intuitive realism about value, an assumption that goods or services have an inherent numerical property, that prices reflect only imperfectly. If that is indeed an intuitive assumption in our economic psychology, it might explain why political economists for centuries tried to figure out the mysterious process that “creates” value, or why they found mercantilism so compelling. Indeed, some form of mercantilism would explain Johnson's finding of a difference between money transactions (people see the seller as better off) and barter (people see neither party as better off) (Johnson et al. Reference Johnson, Zhang and Keil2018b).

R8.3. The prospect of a naturalized economics

Finally, some commentaries lead us to a much broader and more difficult question, to do with the prospect of making economics itself more realistic, by considering that the agents of microeconomic processes are evolved organisms, from a particular species with highly specific motivations and capacities. That is the change of perspective recommended by DeScioli, who laments the psychological poverty of classical Homo economicus, characterized by exceedingly general, and biologically implausible, motivations and capacities for information processing, preference-ordering, time-discounting, and so forth, which is why both economists and psychologists have argued for more realistic models (see, e.g., Gigerenzer & Selten Reference Gigerenzer and Selten2001; Simon Reference Simon1982).

Like all empirical sciences, economics only predicts phenomena under some degree of idealization. But there is a fine demarcation between sensibly idealizing away some aspect of reality (e.g., friction when dealing with dynamics) on the one hand, and just ignoring those aspects of reality that the model does not explain, on the other. Critics of standard models would argue that the latter is often characteristic of economic modeling (see, e.g. Alexandrova & Northcott Reference Alexandrova and Northcott2013; Reiss Reference Reiss2012). Indeed, evidence from behavioral economics, and to some extent from neuro-economics, suggests in many domains behavior predictably goes against standard predictions.

A common way of accommodating these results has been to consider that people's minds are a composite of (a) some utility maximizing neural approximation of rational choice theory, and (b) various “hot” passions, animal instincts, emotional urges, and so forth, that would explain deviations from normative predictions (Loewenstein & O'Donoghue Reference Loewenstein and O'Donoghue2004). But such accommodations create further difficulties. First, in cognitive terms, they raise the question of how rational choice is actually implemented in cognitive systems, whether rational choice is a characteristic of the person, the entire assembly of cognitive systems, or of some of those systems, or of neural function inside systems (Ross Reference Ross2005, pp. 76–100, 114-119). Second, from an evolutionary standpoint, simply adding a measure of emotionality or error to rational choice cannot be a satisfactory account. As Lightner & Hagen remark, the results of current behavioral economics studies are often difficult to interpret, because of the wrong idealization problem mentioned above. That is, experimental protocols are generally supposed to eliminate “environmental noise” – for example, reputation, norms of fairness, and so on, from economic behavior. But Lightner & Hagen sensibly point out that such factors are not noise – they are the cues our evolved cognitive systems expect from environments, in order to deliver motivations for behavior. On this basis, we agree with DeScioli as regards the limitations of Homo economicus – and his cousins Homo sociologicus, Homo politicus, and so forth. Rather than using the wrong kind of idealization, one should pursue social sciences on the basis of Homo sapiens.

This leads to another question raised by commentaries from Lightner & Hagen as well as DeScioli. Once we see human behavior from an evolutionary perspective, is there an actual domain for economic models? Economics is usually defined as the study of all problems that involve alternative choices given scarce resources. But, in the course of evolution, humans, just like other organisms, were never confronted with a unified class of such problems. Rather, they had to solve scarcity and opportunity cost problems in a variety of domains, for example, in foraging, mating, acquisition of social status, recruitment of social support, territorial competition, and more. In actual ecologies, each of these different domains comes with highly specific invariances, therefore with a unique space of potential adaptive solutions, resulting in selective pressure for domain-specific optimization systems. So, it may be that, in terms of cognitive processes, it makes sense to study each of these specialized systems – for example, our intuitive economics of mating, intuitive economics of foraging, and so on – because that is the level of abstraction at which we can maximize explanatory power. By contrast, the “intuitive economics of everything that is economic” just never evolved as a cognitive system. That is why most of our beliefs about the economy are neither intuitive nor scientific; they are folk-economic beliefs.