Norm Externalisation and the Evolution of Cooperation

1 Introduction

In a recent article, Kyle Stanford gives an evolutionary account of what he terms “externalisation”, understood as “our distinctive tendency to objectify or externalise moral demands and obligations” (Stanford 2018a, p. 1). According to Stanford, externalisation is a distinctive feature of human moral psychology, which consists in our experience of the demands of morality as externally imposed on us. That is, we see ourselves as obliged to conform to these demands no matter what our subjective preferences are and, furthermore, we regard them as imposing the same unconditional obligations on other agents as well. Externalisation is adaptive, he claims, because it enables cooperation and protects it against free riders by “establishing and maintaining a connection between the extent to which an agent is herself motivated by a given moral norm and the extent to which she uses conformity to that same norm as a criterion in evaluating candidate partners in social interaction generally” (p. 1), and this way of choosing partners makes cooperation possible.

Stanford’s proposal, although certainly innovative and intriguing, is not without its problems, which include a lack of conceptual clarity, confusion over the norm domain to which externalisation applies, and doubtful usefulness of externalisation in partner selection in early human societies, which is the evolutionary role assigned to it by Stanford. However, I argue that the main issue with Stanford’s account is that it does not pay serious attention to the effects of the social responses and structures on the individual experience of norms. Focusing predominantly on individual psychology, largely considered in isolation from the social context in which the individual is embedded, causes Stanford to miss out on alternative explanations of externalisation. While a psychological, inwards-to-outwards focus is hardly surprising for a work in moral psychology, it nevertheless prevents Stanford from fully appreciating the workings of norms and the effects that the social reality of widespread cooperation, enacted through norms and in turn sustained by them, could have on humans. In this paper, I will show how taking an alternative perspective that turns the spotlight to social practices can better explain Stanford’s concerns, resulting in a simpler and more convincing solution to Stanford’s puzzle. Seen in this light, norm externalisation becomes a side-effect instead of an adaptation deserving of a special explanation. I will not argue for this solution based on new empirical evidence that Stanford supposedly misses. Instead, I claim that this solution is a more plausible one in light of the evidence Stanford himself proposes as well as some fairly common and uncontroversial background assumptions regarding our and our ancestors’ social environment.^[1]

In the following, I will first present Stanford’s account of externalisation in more detail, after which I will discuss the main conceptual difficulties it encounters and offer an interpretation of his theory that I deem the most charitable. Then I will proceed to sketch out an alternative, simpler, explanation which pays attention to how the social reality can influence the individual, and show that this change of focus suggests norm externalisation could be a result of a more general way in which we experience the world around us. Finally, I will argue that, even if we, for the sake of the argument, accept Stanford’s view of externalisation, it could not have played a role in partner selection.

2 Externalisation as an Adaptation

Stanford sets out to explore what he believes may be the most important distinctive feature of human moral psychology: externalisation, meaning our experience of moral demands as externally imposed on us, as an obligation to act regardless of our subjective preferences, and at the same time, as imposing the same unconditional obligations on everyone else (p. 1). To further delineate the phenomenon in question and provide empirical support for his claims, Stanford cites studies (e.g. Turiel 1983; Turiel, Killen, and Helwig 1987) which allegedly show that children between 2.5 and 3 years old already begin to “reliably and systematically” differentiate between moral norms (e.g. pulling hair or stealing) and purely conventional norms (e.g. talking out of turn or drinking soup from a bowl). Violations of moral norms are considered “more serious and more deserving of punishment” than conventional norms, and their wrongness is typically justified by appealing to harm, fairness, justice, rights, or the welfare of victims (pp. 2–3). In addition to seeming more “serious”, moral norms are systematically viewed both as more generalizable, that is, as applying to people in other places and periods, as well as authority-independent, meaning that they cannot be suspended by some authority figure or institution (p. 3).

However, Stanford also refers to more recent studies (e.g. Haidt, Koller, and Dias 1993; Kelly et al. 2007) that have shown that norms which do not include considerations or justifications based on harm, fairness, justice, welfare or rights, can also be seen as moral. Similarly, norms that do include these considerations and justifications need not be seen as moral, but can also be thought of as purely conventional. For these reasons, Stanford concludes that, although moral norms are frequently or prototypically connected to harm, justice, welfare, or fairness, these are not defining features of the norms themselves (p. 3). Instead, he turns to the scale developed by Goodwin and Darley (2008, 2012, whose experiments have shown that subjects “reliably locate moral judgments at a particular point along a scale of increasing objectivity” (p. 3), ranging from judgments of taste or preference, to judgments of social convention, to moral judgments, to judgments of empirical or scientific facts. Stanford’s claim is that these “categorical differences” in subjects’ willingness to tolerate the possibility of disagreement without error provide “perhaps the clearest way to characterize the sense in which we treat moral norms and judgments as systematically more objective than judgments of taste or preference or judgments of social convention” (p. 3).

Stanford believes that this kind of externalisation presents us with an evolutionary puzzle, since merely subjective preferences for interaction with those who act prosocially would serve equally well to steer us towards interacting with those willing to cooperate, and away from those that are not. The question is, then, why did we develop this particular kind of motivation instead of a purely subjective one? Stanford’s answer is that externalisation was beneficial because it ensured the correlated interaction between agents willing to act cooperatively, thus protecting them from exploitation by free riders (p. 8). This correlated interaction is achieved by establishing and maintaining a connection between an agent’s motivation to adhere to a certain moral norm, and her use of conformity to that norm as a criterion when choosing potential interaction partners. In other words, if an agent is motivated by a norm and experiences it as somehow externally imposed on herself and others, she will have a preference for interacting with those who respect the norm, rather than those that do not. In this way, those who conform to the same norms will prefer to interact with each other, instead of interacting with those who do not respect the norm. This is a beneficial outcome for them, as interacting with those who do not respect the norm could make them lose out by cooperating when the other does not.

3 Externalisation as Seen from a Social Perspective

The main question that Stanford poses is why do we, as individuals, externalise some norms (but not others), and his answer is that our psychological experience of norms as externalised enables and sustains cooperation by influencing our partner selection. On this view, our experience of norms as externalised is an evolved adaptation the function of which is to make cooperation possible and sustainable. However, once we turn our attention to our (and our ancestors’) social worlds, rather than their psychological experience, a simpler and more parsimonious answer suggests itself to us.

To see how, let us begin by asking a counter-question: why do we “externalise”, that is, experience as objective or as externally imposed on us, various empirical facts, even those that we only know through testimony, as well as empirical regularities and so-called natural laws (as Goodwin & Darley scale also seems to show)? For example, why do we in this way “externalise” our belief that a pen pushed off a table falls onto the ground?

A simple answer is that, every time we saw a pen rolling off a table, it fell to the ground, and that it is a part of our mental constitution somehow that we experience such regularities a part of objective reality, or even something of a natural law. Analogously, we regularly, from the earliest age, observe people acting in accordance to the norms of our society, for example, acting in ways which avoid causing harm to others, speaking against causing harm, negatively judging and punishing those who cause harm, hiding, apologising or justifying oneself when causing harm, in ways that are mandated by the local social norms. Hence, analogously to the pen case, we come to experience these regularities as a part of reality. More precisely, we do not only experience norms as a part of reality, they indeed are a part of reality, namely, of our social reality. Our experienced reality does not consist only of pens falling down instead of shooting up, sun coming out every morning instead of every evening, sugar melting in water instead of hardening. It also consists of people joining in to help push cars stuck in the snow rather than digging them in even further, waiting for their turn rather than talking when they please, and paying in a restaurant instead of wrecking the place and skipping away merrily.

It could be argued that norms are not perceived or experienced as objective nearly as much as empirical regularities or natural laws are, and this would likely be correct. However, this could be explained by the fact that, while we have never witnessed or received reliable testimony about a pen shooting upwards upon release, we sometimes do in fact observe, hear about, and commit norm breaches. Sometimes we do witness or hear gossip about people taking out of turn, stealing, pulling other people’s hair and picking their noses in public. This could likely result in a weakening of our experience of the respective norm’s “objectivity”, or rather, validity, just as the validity of the pens-fall-down-when-released regularity would have been shaken after experiencing a pen shooting up in the air upon release.

Relatedly, it could be further argued that, since “conventional” norms, in Stanford’s sense, are also regularly observed, this explanation could not account for the “pattern of externalisation” that Stanford seeks to account for, where people deem some norms to be “more objective” than others. If norms in general are externalised to a weaker extent than empirical facts because we sometimes learn about breaches of the former as opposed to the latter, why is it so that some norms are more internalised than others? A part of the answer could be that some norms are breached more often than others – we witness people picking their noses in public more often than we witness people murdering each other. Furthermore, when we do witness someone picking their nose, the social response to the fact (which is a further social norm) is different, to put it mildly, than the social response to murder.

This brings us to the second part of the answer, namely, that all norms are not alike with respect to the type of reasons, justifications and the like people appeal to when acting out of them or breaching them. For example, one can appeal to the word of God or Gods in order to justify a certain practice, or to the fact that this is what others in the group have previously been doing (compare the difference between wearing a hijab and between wearing a similar headscarf because that is the current fashion). Alternatively, one can appeal to ancestors, prophets, leaders (“this is in accordance to our ancestors’/leader’s/guru’s word”), or to more abstract authorities such as “nature” or morality (“this is good/right/natural”). Here, the parallel is with empirical facts that we come to learn through testimony. Just as our group members can tell us that there are raspberries in that forest, or that the Earth is flat or round, they can tell us that sharing berries is objectively morally good or natural, or that our Gods want us to dress in this way or that.

It is important to note that these “reasons” and “justifications” are not meant to be taken as being in agents’ minds, but rather as being socially accepted as valid standards in a particular group (see e.g. Chwaszcza 2017), which is itself constituted by further social normative practices. Thus, reasons and justifications here are not meant to be agents’ mental states which they report to other group members, but rather standards which are established through social practices, taken to be valid by the group in question, and to which the agent can make an appeal. As already stated above, the reasons can vary. They can include, for instance, practice-based reasons that refer to or derive from background rules which constitute social practices, as well as moral reasons which refer to the moral rightness of the act (Chwaszcza 2017). Socially accepted reasons of our ancestors likely also included those that refer to various gods and religious principles or teachings, ancestor and authority-figure based reasons, and a variety of others, each backed with its own system of norms. Brennan et al. (2013) seem to pick on a similar point when they argue that different kinds of norms create different kinds of accountability or obligations, for example accountability to each other as individuals, accountability to each other as members of a certain social group, and similar.

These different kinds of reasons, justifications, accountability and the like are usually presented as objective features of the world, even though they ultimately depend on the group acceptance of them as such. That is, if the group norms start to shift in such a way that a particular justification (e.g. “this is in accordance to God’s word”) is not anymore taken as acceptable, meaning it generates a different group response, this would reflect on the perception of their “objectivity”, or validity. This might be expressed by the subjects in Goodwin & Darley-like experiments as some norms being “more objective” than others. Here we return to the previous considerations on what the subjects might mean by “more objective” in these contexts. It is possible that people in these experiments and outside of them conflate the validity of different reasons or justifications for norm obedience, or different kinds of responses to norm breaches, with different levels of their objectivity. This, however, does not mean that norms that are experienced as, e.g. moral will be ranked higher on the scale of objectivity. Rather, it may be that some norms are taken to be, by the social group at a specific point in time, more “objective”, as in, backed by valid reasons, justifications, accountabilities, authority etc., and consequently eliciting a more reliable and strong response.

Norms to which the social response is more regular, widespread, intense, organised, and the like, are likely to be experienced as more “regular” or “valid”, which could translate to them being ranked higher on the scale of objectivity by the subjects. For example, some norms are deemed as more significant than others in a society, in the sense that greater importance is given to obeying them, norm breaches are very rarely tolerated and those who breach them are consistently taken to account, their obligation to follow the norm is evoked, asked to justify the breach, sanctioned and so on. This sort of social response may result in subjects deeming these norms “more objective”. Alternatively, certain norms apply only to a particular subgroup in the society, while others apply to every member, which may reflect on the perception of their “objectivity”, as well. To return to the comparison with empirical cases, the statement such as “January is the warmest month of the year” might be ranked as “less objective” than the statement “Pens fall to the ground when released”, for example because, while January might be the warmest month in some parts of the world, such as in the Southern Hemisphere, it is usually the coldest month in other parts of the world. In contrast, the workings of gravity are ubiquitous.

If “less externalised” norms include those that are more often or more easily (that is, with less severe consequences or less pushback, meaning less calling to account and the like) broken in a certain community, this could also explain why some kinds of norms could be typically “more externalised” than others: aesthetic norms, for example, are arguably in most communities breached more often, and their breach leads to less response, than various norms of cooperation or rationality. For obvious reasons, having solid norms when it comes to situations likely to matter to the group interests is of greater importance than in other domains. Typically, whether you like ice-cream or not is of no importance to me or anyone else, but whether you like Nazis or not matters substantially, as it says something about the ways in which you are likely to treat others. The group, then, has an interest in holding you to account regarding your fondness for Nazis, while we have no comparable interest when it comes to your taste for ice-cream. Likewise, reaching and maintaining common ground on the acceptability of theft is arguably more important than on the acceptability of listening to Britney Spears’ music, as theft could be detrimental to the group interests or even existence, while Britney’s music could not, in most cases. “Strongly externalised” social norms, then, do not have to be only experienced as externally imposed or objective, they can indeed be externally imposed and objective, in the sense that they are an important part of the social world of the group.

The second part of the answer, regarding norm justification, would seem to at least partly reduce to the first part, i.e. norm regularity. At least for some norms, it appears vital that people make regular demands on each other with respect to them, that these particular norms are obeyed with great regularity, and that there is a significant social response when these norms are breached. This can, in turn, result in further regular practice of the norm, making the norm a stable part of the group’s social reality, which could lead to the norm being “externalised”. Furthermore, appealing to the nature, “the way it has always been done”, ancestors and the like is often based on previously existing social regularities and practices.

Therefore, norm externalisation in this account becomes a consequence of more general mechanisms such as our tendency to perceive regularities in reality, including social reality, as objective and external. This account is more parsimonious than Stanford’s, and it treats norm externalisation as a side-effect, rather than an adaptive innovation. This means there is nothing to account for – there is no role that externalisation should play to justify its existence. Making demands on each other, creating different kinds of social accountability, justification and reasons could all be adaptive in the context of the evolution of cooperation. However, this does not mean that makes externalisation, in Stanford sense, adaptive – externalisation is still the effect, rather than an adaptation itself. The claim is that Stanford has the order wrong: rather than externalisation leading to cooperation, cooperation is enabled and sustained, out of necessity, with available cooperative partners and through the existing norms, and as a consequence of this social reality these norms are experienced as somehow objective and externalised.

Similar points have been raised by Isern-Mas and Gomila (2018). They emphasise the connection between norms and values and motivations, put in terms of Darwall’s second-person view of morality. According to this account, the objective character of norms aligns with subjective motivations because the norms themselves are grounded in claims and demands that emerge out of interaction between the community members, which the community eventually begins to sanction. The process through which this happens goes as follows. Individuals frequently interact out of evolved prosocial preferences, and when doing so they explicitly or implicitly make demands on each other. Over time, they form expectations about how others will or should act. At the same time, they are themselves motivated to comply with others’ expectations, and know they can expect sanctions otherwise. Eventually, this results in group enforcement of these norms and in group members experiencing them as objective, while simultaneously being motivated by them. In this way, the experience of norms as externalised emerges as a side-effect of the process of social interaction.

These remarks are echoed in Elder-Vass’ (2014) discussion of social structures and their causal significance. He writes that the individual “has a sense, however vague and minimal, that she is acting on behalf of something wider than herself when she acts in support of a norm, and that sense increases the likelihood that she will act in its support, by comparison with the isolated individual with a purely personal attachment to the standard of behaviour concerned” (p. 50). According to him, this sense is a consequence of a process at the core of which is “repeated exposure of individuals to acts of endorsement and enforcement of the norm concerned” (p. 50). He illustrates this on an example of norm of queuing, where repeated exposure to people criticising those jumping the line results in the individual’s coming to understand the norm of queuing and to believe that she will be negatively sanctioned for failing to observe it in that environment. According to Elder-Vass, this will ultimately result in a tendency to conform to the norm as a result of the group members’ actions.

Isern-Mas & Gomila’s and Elder-Vass’ proposals are plausible and welcome alternatives to Stanford which take into account how individuals are influenced by their social interactions with others in their community. However, it seems unnecessary to remain focused on a speculative account of individual moral motivation and its connection to sanctions when explaining externalisation. As seen above, the phenomenon can be explained without referencing such particular psychological mechanisms, apart from a general feature of our constitution to experience external world as real.

4 Externalisation and Partner Choice

Nevertheless, let us accept, for a moment and for the sake of the argument, Stanford’s view of externalisation as an adaptation in need of explanation. Even so, externalisation is unlikely to do the job Stanford assigns to it, as it has already been argued by a number of commentators (e.g. van Prooijen 2018; Voorhees, Read, and Gabora 2018). Since others have previously presented similar ideas, the arguments will not be developed here in much detail, but rather it will just be briefly pointed out that, in the circumstances in which human cooperation is thought to have first evolved, choosing partners on the basis of the norms they adhere to would probably have not been possible.

Even though Stanford clarifies that he takes externalisation to influence one’s preference ordering of partners, that is, one’s judgment of their desirability relative to other potential cooperation partners, rather than leading directly to shunning of agents with different norms, it is unclear whether even such choice was possible in early human groups. That is, it is doubtful that individuals could normally just avoid contact with or reliance on less desirable partners. While in large communities it might be true that individuals are, to a degree, free to choose their cooperation partners, it is unlikely that this was the case in the early hunter-gatherer societies in which cooperation and prosociality evolved, which were small, and emigration was not a likely option. People in these societies likely did not had the luxury to interact or rely less on individual members of their tribe, even if they disagreed with them (Voorhees, Read, and Gabora 2018, p. 32).

Furthermore, even if choosing between partners were possible, it is far from clear that there were multiple norms on the basis of which one could choose. If norms are established and transmitted via enculturation, which seems the most plausible explanation by far, all or almost all members of a group would share the same norms. This homogeneity would be the most probable result not only of their modest size, which makes it less likely that significant variation arises (Mesoudi 2011), but also of their relative isolation and lack of communication across different groups, which shields them from coming into contact with people who behave differently. Another factor could be the lack of a written record which could expose them to their own history and offer an opportunity to compare the way they live at the moment with the way their ancestors lived. The effects of conformity are also not to be discounted (Boyd and Richerson 1985). In small and tight-knit communities, where everyone is raised in similar ways and by similar people, and where people lead similar lifestyles that highly depend on others in the group, it is likely, if not inevitable, that everybody will adhere to the same norms, which would make it impossible to choose between individuals on this basis. It might be the case that modern societies are comprised of different, more or less integrated social groups whose norms differ to a greater or lesser extent. But in this respect, too, modern societies differ from ancient small-scale ones. In short, in these circumstances, it would be difficult to imagine a different way of life, or even that a different way of life is possible. That is, when it comes to partner selection, there would be nothing to choose from, if the basis of choice would be different norms.

This fact also bolsters the case for the account of “externalisation” presented in the previous section: if you never come into direct or indirect (e.g. through testimony or even just fictional stories) contact with any people who practice different norms, the norms of your own society may seem much more real or robust, and a different social reality would be very difficult to imagine. In fact, we can observe similar situation even in our seemingly very heterogenous, well-connected and informed world: the more widespread and long-lasting some social norm or structure is, the more difficult it is for people to imagine that it could be otherwise. A case in point is the seeming irreplaceability or naturalness of capitalism as a form of socioeconomic organisation, that is experienced by large numbers of people who fail to even conceive of possible alternatives to this system. Another example are certain gender roles and racial stereotypes that stubbornly resist change.

5 Conclusion

In summary, exploring the situation from a perspective which fully takes into account social relations, norms, structures, and their effects on individuals, can help us make sense of what Stanford terms externalisation and provide an explanation of it which is simpler and more fitting in respect to both empirical evidence and common assumptions about human social environment and psychology. According to Stanford, externalisation influences partner selection, which enables and sustains cooperation. I suggested that Stanford has the order wrong. Instead of externalisation first appearing as a psychological adaptation which leads people to cooperate better, it could be argued that our ancestors must have already been obligatory co-operators whose social interactions were taking place in a rich social landscape. Their shared practices constituted their social worlds whose regularities are perceived in much the same way as regularities of the empirical world. In this way, focusing on our ancestors’ social circumstances instead of on their psychological experiences leads us to the conclusion that there is no special adaptation to explain. Regarding partner selection itself, even of the weaker “partner ranking” type, no such luxury was likely available in circumstances in which cooperation first evolved, that is, in small and relatively isolated groups. In such circumstances, the co-operators are “stuck with what they got” in their groups, which we can safely assume are largely homogenous.