Until recently, morality was largely the domain of philosophers. A typical debate might involve defining why killing another human being is wrong (taken as one of the foundations for a good society). Answers might include because it will destabilize society and lead to general unhappiness (the consequentialist position of Berkeley), or because it violates the Golden Rule of doing to others what we would want them to do to us (subscribed to by deontologists like Kant), or because people of good upbringing and hence virtuous character would not be inclined to do so (the position of virtue theorists like Aristotle). However, relatively little headway was made on such ethical issues because the standard philosophical toolkit of linguistic analysis, logical argument, and introspection provided insufficient means to settle such arguments.

Currently, a variety of disciplines have begun to engage in empirical investigations into the nature of morality. Indeed, morality is currently a hot topic in biology (Alexander, 1987), psychology (Hauser, 2006), philosophy (Joyce, 2007), law (Sunstein, 2005), primatology (de Waal, 2009), and the social sciences (Nichols, 2004). Though controversies continue to rage, there is now general agreement about some aspects of morality. It is thought to be a quintessentially human trait, although the behavioral roots of morality (such as loyalty to kin, intolerance of theft, and punishment of cheats) can be seen in related primate species (Boyd et al., 2003; Boyd and Richerson, 2006; Hauser et al., 2009).

Moral systems have psychological, social, and behavioral components: the psychological mechanisms enable individuals to recognize actions that are moralized and produce intentions to conform to normative expectations as well as to punish moral offenders; social mechanisms like ostracism or shunning help to enforce moral norms; and behavioral violations invoke the activation of the psychological and social mechanisms in the first place (i.e., immoral actions like incest or murder). All of these components need to be in place for moral systems to function properly.

Recent work in psychology has established a number of principles relevant to morality. It is broadly accepted that moral judgments have specific emotional underpinnings (e.g., shame, guilt and retribution) (Haidt, 2003; Moll et al., 2005; Nelissen and Zeelenberg, 2009) and are made rapidly without rational calculation as to harms and benefits (Haidt, 2001)—although when presented with moral dilemmas, people may derive judgments from the interaction of intuitive and rational faculties (Greene et al., 2001). Further, Mikhail’s notion of a “moral grammar” may help explain how moral intuitions are generated via a set of both rational and emotional components in the brain (Hauser, 2006; Mikhail, 2011).

Debate continues, however, as to exactly what functions morality serves—except to say that moral judgments and action facilitate social cohesion in some way (Hauser, 2006; Joyce, 2007; Bekoff and Pierce, 2009; de Waal, 2009). Indeed, there is a growing consensus that morality evolved to enable human cooperation (Rawls, 1971; Hume, 1998; Haidt and Kesebir, 2010; Ellemers and Van der Toorn, 2015; Curry, 2016). For example, Joyce (2007) argues that making actions moral (i.e., normative) increases their motivational impetus, thus increasing the likelihood of cooperative activity. But why do we think of assault, theft, eating meat, abduction, public nudity, treason, rape, counterfeiting money, and denying the Holocaust as all being immoral? Punishment would not seem to explicitly facilitate cooperation in all these cases. Here, we will argue these concerns all constitute violations of implicit obligations to obey social rules. These obligations derive from a requirement that emerged from developments in social organization during the recent evolutionary history of our species.

In particular, moral actions seek to control defections from public obligations by other members of large-scale human groups. This problem is acute because humans depend on each other—due to a significant division of labor—for everyday necessities, but cannot rely on shared genes or direct reciprocity to ensure ready supplies of what is needed to stay alive. Social organization in large groups of unrelated individuals instead depends on threats of retaliation for failures to abide by cooperative obligations, including fulfilling one’s own roles in the social group. These threats are made real by an underlying moral psychology that motivates the punishment of social infractions, even in the face of no direct benefit to the punisher. In effect, anyone in such a social group can potentially punish (or reward) anyone else in the group for “bad” (or “good”) behavior. Thus, we will defend a particular proposition that morality evolved to solve particular problems of cooperation—problems associated with ultrasociality. Essentially, morality can be hypothesized to have evolved as a system of behavioral control to aid cooperation in large groups, which cannot rely on shared genes to cause overlapping interests. Further, we will argue that to the degree that these large-scale groups cohere and function (at least partly thanks to moral threats and actions), they can be called human “superorganisms.”

The remainder of this paper will outline this new approach to elucidation of the moral domain, Human Superorganism Theory (HSoT). We conclude that identifying the adaptive function of moral action should go a long way toward clarifying the ancillary issues that dominate the contemporary study of morality.

The idea that human societies form superorganisms can be derived from a recent development in evolutionary biology called “major transition theory” (Maynard Smith and Szathmary, 1995). The basic premise of this theory is that the means by which information can be stored, translated, and transmitted from one generation to the next has itself changed a number of times during the history of evolution on Earth. Major transition theory thus suggests that the way evolution works has itself evolved, with the consequence that new levels of functional complexity are achieved with each transition (e.g., the transition from RNA to DNA as the store of genetic information, or unicellular to multicellular life, or multicellular to social life).

The central difficulty in using major transition theory for explaining any evolutionary problem is showing how cheating can be suppressed at the lower level so a higher level of organization can evolve (Buss, 1987; Maynard Smith and Szathmary, 1995; Michod, 1999). This is because there are always incentives (in strategic, or even genetic terms) for the members of an organization to defect from the cooperative actions that sustain the new level of cohesion in favor of their individual interests. Mechanisms for regulating these “selfish” behaviors must therefore evolve to consolidate activities that ensure the new level of organization coheres and persists. These mechanisms are likely to be specific to each case of transition.

Once the first step of suppression has occurred, there is typically a second stage during which differentiation of function occurs within the new structure (Szathmáry, 2015; West et al., 2015). For example, explaining the evolution of multicellular from unicellular life forms may have to do with a mutation that causes cells not to separate after division, thus suppressing their independence, and then acquiring a reproductive/soma differentiation of function (Gavrilets, 2010).

The most recent transition recognized by major transition theory is that to ultrasociety (Maynard Smith and Szathmáry, 1998). The human species is unique in living in ultrasocial groups—that is, large groups of cooperative members who are essentially unrelated to one another (e.g., in tribal groups, or cities) (Richerson and Boyd, 1998). Recall that for a major transition to occur, fitness differences within constituent groups must be suppressed so that selection at the level of the group becomes the primary evolutionary force at work. Major transition theory argues that a reproductive bottleneck helps reduce conflicts of interest at the group level (Bourke, 2011; Szathmáry, 2015). Some insect societies exhibit extreme restrictions on genetic variability (and hence limit social conflict) by having a single female, the queen, serve as the sole reproducer (Hölldobler and Wilson, 2008).

Human societies have not (yet) achieved such a group-level adaptation. Nevertheless, there is a strong division in the human superorganism between “somatic” functions and the reproductive function. In particular, a specialized institution, the human family, has evolved as the vehicle for reproduction (Emlen, 1995; Geary and Flinn, 2001), which serves to restrict both genetic variability in human populations (which is very low compared to other primate species) (Bowden et al., 2012), and overall group population size. While human females are born with around two million eggs, a variety of mechanisms have evolved to reduce actual completed fertility to a very small number. Sequestering of newly reproductive females in puberty huts and restriction of their social movement until pair-bonded ensure that exposure to males is minimized early in life. Concealed ovulation ensures that males do not fight over access to a woman during a brief, obvious period of fertility, reducing sexual tensions in a group, and making it possible for some reproductive cycles to pass without insemination, reducing fertility (Geary and Flinn, 2001; Marlowe and Berbesque, 2012). During the majority of a human female’s reproductive career, she is (except in extreme ecological circumstances) limited to forming a single breeding pair (i.e., monogamy) (Zeitzen, 2008; Chapais, 2013). These single pair-bonds are reinforced through cultural rules associated with the institution of marriage, which make it more difficult for women to break out of this bond (and which also reduces male competition for access to females). Then, when a woman is older, menopause, or an evolved incapacity for further reproduction, ensures that her own reproductive effort ceases, so that she begins to aid the reproductive success of the next generation family to which she is most closely related (Davis and Daly, 1997; Hawkes and Coxworth, 2013). All of these mechanisms serve to tightly regulate female reproductive potential and ensure the isolation of reproduction from other functions in society, as one would expect of a superorganism.

With reproductive function isolated in these ways, other social structures could arise to serve various kinds of “somatic” functions. Here, too, fitness differences have been suppressed. Ultrasociality appears to have been achieved by transforming the ancestral form of social organization in related ape species of rank-based dominance (which creates significant differences in fitness between individuals within a social group) to one of enforced egalitarianism, as seen in extant human forager societies (Boehm, 1999). The initial trigger for the evolution of large groups in our human ancestors may have been a change in diet toward meat eating, which required group foraging (Tomasello et al., 2012; Tomasello and Vaish, 2013). In these societies, potential fitness differences are suppressed through enforced food sharing, monogamy, and alloparenting (sharing the parental care of particular children among families), all of which served to decrease fitness differences within groups. This first level of social interdependence for achieving basic needs was later reinforced by the need for coordinated defense against competing groups (Tomasello et al., 2012; Tomasello and Vaish, 2013). It appears that there was considerable inter-group conflict among our ancestors (Chagnon, 1983; Keeley, 1997; Peterson and Wrangham, 1997). Such inter-group aggression could have a considerable impact on fitness, because entire groups could be decimated, and their reproductive resources (women) lost to other groups from the frequent wars fought by such groups (Wrangham, 1999; Bowles et al., 2003). Battles and raids would have been more likely to be successful to the extent that war parties were large and well organized.

Although these societies lacked centralized policing authorities, they apparently could nevertheless coordinate large-scale inter-group competition via punishment mechanisms. Norms are more likely to be enforced as the degree of social interdependence increases (Horne, 2009), as is the case with large groups. For example, the Turkana, a contemporary African pastoralist society, engage in frequent cattle raids on other ethnic groups with parties of several 100 members drawn from most or all of the clans composing that ethnicity. Because these are risky endeavors, men can desert or otherwise defect, but such cheats are informally judged by the community, and beaten, fined, or socially excluded, even by those not present at a raid (Mathew and Boyd, 2011). It is also more likely that “parochial altruism” (cooperation against others) will evolve when there is significant inter-group conflict (Bowles, 2006; Choi and Bowles, 2007). Thus, prestate societies without formal institutions (only informal organizational principles such as age grades, gender roles, and endogamy) are able to organize mechanisms that protect and sustain groups of many tens of thousands of individuals.

This combination of enforced egalitarianism within the group, coupled with significant inter-group competition, was an ideal way to minimize within-group selection and maximize inter-group selection, thus facilitating the formation of organizational adaptations at the ultrasocial group level. If this combination was a general condition for prestate societies during human evolution, it would give considerable depth to the history of ultrasocial organization, consistent with the expectation of psychological adaptations for such a lifestyle (e.g., shame, guilt, and moral concern). Ethnic groups of this size probably arose tens of thousands of years ago—plenty of time for psychological adaptations to have been established. Then, more recently, the agricultural revolution brought the first economic surpluses, and therefore allowed for social specialization (Gowdy and Krall, 2016)—the sort of differentiation of roles within the larger structure expected in the second step of a major transition.

Gene-culture coevolutionary theory provides evidence that specific psychological machinery to support morality—in the form of perceptual biases, social learning abilities, and neurological mechanisms—could have evolved in human populations through the kind of intensive cultural group selection just discussed, accounting for the appearance of moral systems in our recent history (Bowles et al., 2014). According to this theory, humans should have evolved expectations to find some behaviors are sanctioned in the groups to which they belong, to have innate abilities to recognize which behaviors are moralized, to rapidly learn how to perform these behaviors from others in their vicinity through sophisticated forms of social learning, and to experience both conformity with norms and punishment of norm violators as rewarding, so that moral experiences are reinforced (Chudek et al., 2013). Further, this machinery should produce defaults that lead to automatic conformity with social norms, such that cognitive efforts are necessary to override this tendency, so that the prosocial psychology has been internalized in the form of internal sanctions like feelings of shame or guilt. These suppositions are consistent with experiments using behavioral economic games, social psychological studies, and neuro-economic evidence.

Recent evolutionary biological theory also suggests that societies which become truly large share two characteristics: a multi-individual (group-level) production system involving a division of labor and defensible outputs (e.g., nest), and a multi-individual reproductive system in which the young require significant nurturing (e.g., a colony) (Wilson, 2012). These features ensure high levels of social interdependence and form the organizational principles on which natural selection at the level of the group can act to secure cohesive ultrasocial groupings. In primate societies, raising an offspring successfully went from the job of the mother to the family, and in some cases, a troupe of alloparents (Mitani et al., 2012). Simultaneously, the basic economic production unit has gone from individuals to families to professional organizations (such as guilds and companies), as more and more resources have been dedicated to these processes, and a wider variety of interdependent roles have been required for the production of increasingly specialized outputs. Social groups with these two features thus developed both productive and reproductive stores of resources—in the form of domesticated animals and women—that could be contested by other groups, leading them to become desirable targets for takeover through inter-group conflict, as suggested above. Like eusocial insect groups, these human populations could be considered a “factory inside a fortress,” which constituted the foundation of selection for superorganisms (Wilson, 2012). This self-reinforcing process of social differentiation leading to increasing resource accumulation allowed further increases in group size, culminating in the very large groups one can find in both social insects and humans.

These large groups of humans with a low average level of relatedness require novel mechanisms for maintaining social cohesion. The suggestion made here is that morality is a control mechanism punishing social defection in human ultrasocial groups (Haidt, 2007, 2012).

The primary question, then, is what set of functions is needed to organize cooperation in ultrasocial groups. Such a set should define the domain of moral action. To answer this question, we will treat human ultrasocial groups as superorganisms. A superorganism can be defined as a “collection of single creatures that together possess the functional organization implicit in the formal definition of organism” (Wilson and Sober, 1989). This is a perspective that has been adopted in evolutionary biology to describe the social organization of a variety of species, ranging from insects like ants, termites, and bees, to mole rats (Wheeler, 1911; Seeley, 1989; Hölldobler and Wilson, 2008). The idea has also been extended to human social groups (Spencer, 1896; Campbell, 1958; Richerson and Boyd, 1999; Foster and Ratnieks, 2005; Wilson et al., 2007; Christakis and Fowler, 2009; Szathmáry, 2015). However, the extent to which human societies cohere is less than that seen in eusocial insects (which have the advantage of high levels of average kinship). Thus, we might say that the human superorganism is relatively “crude” (Richerson and Boyd, 1999; Kesebir, 2012). It might also be the case that our major transition to superorganism status is still in progress and that future developments (e.g., the world wide web) will increase the degree of social cohesion, leading to an even greater relative importance of group selection pressures (Stearns, 2007). In any case, my contention will be that not only is human society organized in a fashion worthy to be called a (crude) superorganism but also that the function of morality is to police defections from cooperative activities among individuals in their obligations as “cells” in a human superorganism.

Our task then is to find a way to rigorously define these obligations. This is a step that has not previously been taken in discussions of morality, partly because selection at a particular level of organization will produce novel adaptations to fulfill functions operating only at that level. Thus, one cannot assume that a superorganism (a level of coherent organization above the level of the individual organism) will exhibit all of the systems of a multicellular organism—that is, simply transferring “organs” from one level to the other by analogy is unlikely to work (Turner, 2002). Rather, selection in groups could result in the evolution of novel features to support the development and maintenance of membership in groups (Brewer and Caporael, 2006). What we require then is a foundation for making inferences about the “shape” of a superorganism from a more general theory (Mitchell, 2003).

We can find that theory by recognizing that organisms at any level of organization are complex adaptive systems—that is, they can be described in terms of dynamic processes involving flows of material, energy, and information (Bowles, 2006; Miller and Page, 2007; Sumpter, 2010). As living systems, they are complex adaptive systems that are open (i.e., exchange material, energy, and information with their environments) and self-organizing (by means of interactions among their elements), with emergent properties. Relevant kinds of organizational functions at the superorganism level can therefore be extracted from “living systems” theory (Miller, 1978), “minimal life” theory (Gánti, 2003; Rasmussen et al., 2007; Bedau, 2011), collective animal behavior (Sumpter, 2010), and eusocial insect ecology (where these societies are treated as superorganisms) (Wheeler, 1911; Seeley, 1989; Hölldobler and Wilson, 2008; Gadau and Fewell, 2009) as well by comparison to the set of organs in multicellular organisms like mammals, and the Indian caste system (taken as an example of a human superorganism) (Table 1).¹

From these sources, we can extract a set of subsystems that all living systems must depend upon to survive and reproduce, regardless of their complexity.² Any function nominated by two or more of these sources will be taken to qualify as a subsystem of a human superorganism (i.e., single nominations will be considered to be eccentric). There is a high level of agreement among these sources, with only a few nominations not qualifying by this rule (at the bottom of Table 1).

However, there are reasons to include two components, despite their not meeting this criterion, and to exclude another, even though it does. (Remember, one has to take into consideration the special features of superorganisms, rather than simply transfer all potential functions.) Enforcement can be included because it is the primary function of moral action and occurs at the organism level in humans—a close evolutionary analog to the human superorganism—in the form of an immune system. Signaling should also be included because it happens in eusocial insects (although not mentioned explicitly by our source on eusociality)—another close analog of the human superorganism. Including signaling also makes for a complete set of information-based functions to mirror those working on material and energetic aspects.

Another decision has been taken to exclude one function even though it has been nominated by several sources: the motor/movement function, which does not apply to human superorganisms, as they do not move territory or migrate as large units of unrelated people (although eusocial insect colonies can move location when local resources are depleted). (Human superorganisms are more akin to plants than animals in this respect, being concentrated on growth rather than movement.) Movements of superorganism “body” parts (e.g., displays of strength in the form of military parades, or signals of trustworthiness in the shape of free market institutions) can be considered expressive or communicative; hence, the most relevant kinds of superorganism “movement” are encapsulated in the notion of signaling in any case.

Thirteen functions of a human superorganism can be identified in this way (Table 2). For example, living systems need boundaries to hold their elements together and protect them from threats in the external environment. These boundaries can manifest as specific structures at different levels of organization, such as a specialized membrane at cell level, skin at organism level, or an army (to patrol territorial encapsulation) at superorganism level.

This list of components can be adequately summarized by a phrase used to describe a eusocial superorganism: “a factory inside a fortress” (Oster and Wilson, 1978). This phrase highlights the sometimes competing goals of defense and production (the basic political and economic functions of any social group). For example, the biological fitness of a eusocial insect colony depends on the ratio of soldiers to workers: if there are more soldiers, the colony is more secure, but this means fewer workers, and hence reduced food production for the colony (i.e., the workers’ “factory” job), and hence a smaller colony size. Such colonies have mechanisms to regulate the soldier/worker ratio such that it optimizes colony survival (Hölldobler and Wilson, 2008). Similarly, the number of plumbers, doctors, and lawyers in a human group is loosely regulated by market forces (i.e., how much these professions can earn).

Of particular interest for us is the fact that group cohesion requires agents at sub-levels in the organization to act for the good of the group rather than in their own genetic interest. As a consequence, punishment is needed to keep individual elements in line. In cells, punishment is undertaken by lysosomes, cellular organelles that contain acidic enzymes to break down waste materials and cellular debris. They digest excess or worn-out organelles, food particles, and pathogens. In multicellular organisms, punishment is meted out by the immune system, which identifies pathogens and kills them. In superorganisms, we argue that this policing function is performed by individuals motivated by moral sentiments. These motivations inspire people to punish infractions by other members of their group. Opportunities to serve a moral function arise when individuals fail to perform their functions as components of a superorganism. (Note that we suggest all tasks are subject to policing, including failures to serve informational or other control functions—even policing itself.) In such situations, other individuals are required to serve as enforcing agents, bringing the offenders back into line, so that the human superorganism continues to cohere. It is this domain of behavior we wish to investigate here.

We have already provided (in the introduction) an argument for how human society evolved into a (rough) superorganism; however, now that we have identified the particular functions required for superorganism status, we need to elaborate what developments in moral psychology accompanied this social innovation. As with any complex adaptation, it is likely that moral capacities emerged gradually, through a number of evolutionary innovations, so not all of these functions are crucial to the evolution of ultrasociality and hence human morality, as they were present prior to the arrival of large-scale human groups. Here, we are interested in identifying which aspects of superorganism function were crucial to making this major transition, and which therefore constitute the essential elements of human morality.

Comparative biology can provide clues as to which of the functions appeared in the human lineage prior to ultrasocial living. For example, reproduction through families is characteristic of many mammals, so reproduction could not have been crucial. Similarly, xenophobia and preference for insiders (the psychological underpinnings of boundary) are evident in many mammalian species (Hammond and Axelrod, 2006; Efferson et al., 2008; Fu et al., 2012) so boundary is not key either.

Basic storage, in the form of housing, is present even in insects, but also in reptiles and mammals. Humans expanded the types of structures and devised new means of construction to allow larger structures (e.g., weight-bearing arches) (Aunger, 2010) but largely after the transition to large social groups, so this function does not appear crucial either.

Excretion was largely a personal, private matter until the evolution of cities, which required novel innovations for dealing with human waste (e.g., sewerage systems); these are of very recent origin (i.e., the last thousand years), so this function could not have held up the transition to ultrasociality.

The basic function of memory (long-term memory/recall in brains) is evolutionarily old and may have sufficed to coordinate ultrasocial life (e.g., memory among elders of earlier experiences), but ultrasociality may also have required more reliable intergenerational recall of important information—which was made possible first through writing, possibly about 5,000 years ago [although the first markings on bone date to 30,000 years ago (Davis, 1974)]. The earliest writings tend to document economic exchanges, suggesting a link with markets, and hence large-scale societies. The contemporaneity of writing and large-scale social organization suggests that this “augmented” memory may have been important in helping superorganisms to cohere. Memory continues to evolve new modalities today, as significantly new forms of non-genetic intergenerational inheritance evolved after the transition to ultrasociality (e.g., digital communications).

Perception is about monitoring external threats and opportunities—that is, keeping an accurate, up-to-date account of what is going on in the external environment. For example, chimps have been known to patrol their territorial limits as groups to investigate what is going on in other groups (Williams et al., 2004) and to identify ripe fruit sources (Janmaat et al., 2013). Human hunter-gatherers do the same (Cashdan, 1983). Specialists at performing this function are media organizations that keep the population up with any relevant news, but these obviously arose late in history, suggesting that innovations in perception did not constrain ultrasociality.

While signaling outsiders can be done with simple iconic messages (e.g., threats of aggressive intent), Communication within large unrelated groups likely required syntactic language to maintain personal reputations, and thus avoid punishment for being antisocial. Maintaining reputations in large groups requires complex language abilities—the ability to gossip about others (Dunbar, 1998). In particular, recent modeling efforts suggest that the ability to maintain reputations (and thus avoid punishment) is crucial to the formation of large unrelated groups, as indirect fitness benefits, relatedness, and memory for past dyadic interactions quickly diminish in larger groups (Powers and Lehmann, 2016). When this innovation in communication occurred remains highly contentious, however, although most would argue it happened sometime during hominid or hominoid evolution (Hauser et al., 2002; Corballis, 2012; Morgan et al., 2015). Communication is thus probably uniquely human and may have been required to coordinate ultrasocial cooperation.

Enforcement, in a superorganism context, requires third-party punishment, or the willingness and ability to punish others in the group for antisocial activity, even when the individual themselves has not been involved or injured (Gintis et al., 2008). Further, in major transition theory, it is the enforcement function, which is always crucial to the achievement of higher levels of organization (Corning and Szathmáry, 2015; Szathmáry, 2015; West et al., 2015). The evolution of social complexity thus depends strongly on the existence of effective within-group suppression mechanisms—that is, enforcement that minimizes antisocial activity among group members (Frank, 1995; Bowles et al., 2003; Joyce, 2007). In particular, sanctioning institutions fare better than non-sanctioning ones (Gürerk et al., 2006), and fair institutions (i.e., which reduce free riding) are also more likely to evolve (Kosfeld et al., 2009; Kube et al., 2015) probably because they suppress within-group variability in selective outcomes. It is evident that individuals will actually perform third-party norm enforcement occurs in real-world situations in a variety of contexts (e.g., Balafoutas and Nikiforakis, 2012; Yoeli et al., 2013; Tsusaka et al., 2015; Przepiorka and Berger, 2016).

However, this important mechanism, needed to drive cooperation in large-scale groups, appears lacking in non-human primates. Although they do reject lower-valued rewards than fellow experimental subject animals (Brosnan and de Waal, 2014) and thus care about relative outcomes, apes have thus far not been demonstrated to reject unfair offers in ultimatum games (Proctor et al., 2013), suggesting that they do not have a general principle of fairness. Third-party punishment in non-human primates has thus yet to be conclusively demonstrated (Hall and Brosnan, 2016; Talbot et al., 2016; Schmidt and Rakoczy, in press), making enforcement a candidate for the crucial advance. A recent review of experimental studies on responses to unequal rewards in primates suggests that the factor which best correlates (across 10 species) with aggressive responses to unfair offers is the frequency with which members of that species cooperate with non-kin; that is, species which routinely engage in cooperative hunting, coalitions, alliances, and/or food sharing (Hall and Brosnan, 2016). Of course, the primary constraint on the evolution of ultrasociality is cooperation with non-kin, so this correlation is an important confirmation of the centrality of this aspect of moral functionality.

The date at which this innovation occurred is difficult to pinpoint, but it must have been in the last six million years, since that is when our most recent common ancestor diverged from other primates. Presumably, large-scale social groups simply would not form and function without an enforcement mechanism being in place, so we can look to the archeological record for evidence of such groups. (Any sufficiently large human group will have low average relatedness, and hence require third-party enforcement.) Such groups appear to have been a fairly recent innovation, as the earliest appeared no more than about 10,000 years ago (Bellwood, 2004).

Pool punishment (the game theoretic equivalent of paying a tax toward a police force) can emerge in situations where second-order free riders are punished—that is, where those who fail to punish are themselves punished (Zhang et al., 2014). People prefer to use policing to maintain common-pool resources (Andreoni and Gee, 2012; Traulsen et al., 2012). The evolution of specialized social institutions for internal sanctioning and punishment came surprisingly late. The ancient world (China, Rome, and Greece) did not have specialized domestic police forces, but rather used slaves, local magistrates, or government administrators to enforce laws and regulations, but were notoriously lax in doing so (Hunter, 1994; Jones, 2014). Modern policing began only in the 1600s (Critchley, 1978).

Control of large social groups probably first took place through implicit rule making and reference to social norms. Examples include rules of food sharing in hunter-gatherers, rules for the using irrigation systems in agriculturalists, informal property rights, and systems for sharing reputation between traders (Johnson and Earle, 2000). Use of norms was probably made stronger through sacralization—investing normative options with extra meaning and social force through reference to absolute values—to help ensure prosocial behavior when individual- and group-level incentives conflict (Marshall, 2010).

Later, control became a specialized activity, eventually manifesting itself in the development of a professional class, bureaucrats and/or priests, and new forms of organization such as corporations (Wirth, 1938). A crucial innovation was therefore institutions, or self-created rules of social organization where cooperation can be individually advantageous even in large groups of unrelated individuals. Forming institutions requires shared intentionality, language, and other cognitive abilities largely absent in other primates (Powers et al., 2016). So organizations are quintessentially human, and coevolved with superorganism status.

In their ultrasocial form, all the economic functions—production, distribution, and structures—are related to a division of labor. However, they appear to be consequences, rather than instigators, of ultrasociality—that is, they are only necessary once ultrasociality has arisen. In particular, per capita production increased significantly in human society as social roles differentiated with economic specialization, but this only occurred once cities had developed.

Distribution is not necessary until there are significant social inter-dependencies, due to economic specialization, a recent feature of human society. Distribution can occur “naturally” through the unsupervised diffusion of goods and services via multifarious interactions. However, some distributions could lead to conflict over ownership or payment, and therefore require enforcement functions. Distribution is therefore likely to be associated with a concern with fairness, which only arises after cooperative ventures, or when exchanges take place, but again probably only post-facto, not as an initiator, of enforcement. Superorganismal distribution essentially takes advantage of an adaptation that had already arisen for other reasons. In this case, an elaboration of fairness took place: chimps sometimes engage in group hunting and exhibit a concern for fair allocation or sharing (Boesch, 2002). However, in humans, fairness was extended to cover concern for fair exchange of goods produced by separate, non-collaborative activities, once human societies had developed economic specialization.³

Similarly, structures such as road and electricity systems are of very recent origin on an evolutionary timescale, and related to city life, but again are derivative of the essential transition point, and probably without needing specific psychological innovations for their construction.

The late development of these “economic” functions suggests that the major transition to superorganism status constitutes what is called a “fraternal” transition (Queller, 2000). In fraternal transitions, similar units join together, reaping the benefits first from new economies of scale, and then evolving a division of labor through differentiation (similar to the development of complex multicellular organisms with multiple organ systems from simple multicellular organisms without internal differentiation). On the other hand, in “egalitarian” transitions, dissimilar units come together, so that their functions are complemented through the formation of a higher-level unit—e.g., the eukaryotic cell, which combines elements from multiple, more primitive cell types. Obviously, in fraternal transitions, the differentiation takes place as the final step of the transition and can continue to increase once the transition has successfully established a higher-order unit (Szathmáry, 2015).

The crucial innovations required to support the transition to ultrasocial living therefore seem to be associated with control, enforcement, and communication—essentially, social institutions, exchange-based punishment, and syntactic language. These developments are in fact interrelated, in that social reputations are maintained by language (gossip) (Joyce, 2007; Tse, 2008), and stories about miscreant activity can lead to third-party punishment. Further, punishment has been institutionalized into police forces in modern societies.

The account produced via this logic is similar in many respects to that devised by Krebs (2008) from a somewhat different evolutionary foundation. He suggested that the first development in human morality was social life, which introduced differences of rank (requiring new social emotions such as deference and awe), as well as conflicts of interest. This in turn required the evolution of resolution mechanisms such as social sanctions, followed by the internalization of moral norms via fear of punishment, and anticipated regret and remorse, all of which were facilitated in turn by the ability (through so-called “theory of mind”) to adopt the perspective of others. Such internalized motivations for moral action have lower biological and social costs, as actual conflict can often be avoided when individuals prejudge the consequences of their actions and curb their antisocial proclivities. A fourth development was then the development of language and symbolic thinking, which facilitated the formation of moral judgments (the foundation of a sense of fairness). These judgments were finally themselves internalized into moral intuitions (i.e., subconscious predilections to find particular acts immoral or wrong). Such a sequence is consistent with the overall picture painted above, identifying several of the same mechanisms (e.g., complex language and third-party sanctioning) while making some nice theoretical additions in terms of explaining the origin of particular psychological adaptations such as moral intuitions.

Another prominent theorist of the evolution of morality, Franz De Waal, also admits that “objectified” concern for fairness, expressed positively as a generalized empathy, and facilitated by language, is a uniquely human aspect of morality (de Waal, 2009). He says, “We have no evidence that other animals judge the appropriateness of actions that do not affect themselves. The great pioneer of morality research, the Finn Edward Westermarck, explained what makes the moral emotions special: “Moral emotions are disconnected from one’s immediate situation: they deal with good and bad at a more abstract, disinterested level.” This is what sets human morality apart: a move toward universal standards combined with an elaborate system of justification, monitoring and punishment.” (De Waal, 2010)

So punishment of antisocial behavior by anyone, facilitated by the development of instinctive moral tendencies, the maintenance of reputations through language, and the eventual institutionalization of punishment and norms in military and police were the human elaborations of morality that were central to the transition of human society to a superorganism.

Having laid out the basic claims of HSoT, and provided an evolutionary story for its development, we can now compare it to alternative explanations of the moral domain.

Human Superorganism Theory suggests that a group’s morality reflects how that society is organized. Morality is designed to produce “good” behavior, but what is “good” is a function of the ways in which a given superorganism has been organized (including via cultural rules). “Good” people are therefore those who fulfill their roles in their superorganism; those who do not are “bad.” In this way, specific actions may be required by one superorganism, but not another.

Moral action feeds off failed demonstrations of particular practices, and the sanctioning thereof. It achieves this by moralizing specific cultural practices, to which individuals must conform or risk punishment from their fellows. How these behaviors are selected for moralization is an interesting question. HSoT suggests this process is related to the action’s ability to demonstrate prosocial tendencies, or to the action’s relationship to some superorganism function. While there may be some arbitrariness in the initial choice of practices used to demonstrate prosocial intentions and trustworthiness, there is likely to be some selection of those which serve a social function. This is certainly the case for manners, for example, where European rules of etiquette that had hygienic functions (e.g., urinating in private) were more likely to persist over several centuries than those which were simply local conventions (such as where to place one’s hands during a meal) (Nichols, 2002). Inter-group variation in moral codes can thus be arbitrary and non-functional in their origin, but probably become functional once they become sanctionable moral practices, as they reinforce social cohesion and limit cheating.

Similarly, there can be justifiable variation in moral codes within societies. This is seen quite effectively in the degree of concern about the rule of primogeniture—the need for a male heir to continue the family’s ability to hold onto wealth. The strength of this need differs between heads of state, and poor people. In the line of English kings, not only wealth but also continued political power depends on a male offspring. The consequences are that any woman who marries into such a family must bear a son, and therefore is treated largely as chattel. She can be beaten, raped, and humiliated by her husband without compunction, and must not have any lovers (because then the fruit of her womb becomes suspect), while the husband can philander to his heart’s content—all while entirely different morals hold in lower classes of the same society around the institution of marriage (e.g., men cannot rape wives without compunction). Only a theory that allows morality to be a function of social organization can explain such variation in what is considered to be “right” behavior within societies.

Further, HSoT suggests that changes in moral rules over time should be associated with changes in the ways societies are organized. Thus, particular actions can be held to be “right” in a given situation at a given time in a particular superorganism, but not in another because some aspect of social organization has changed. For example, women can only be treated as chattel when they are economically irrelevant. This can be seen in societies that adopted agriculture, where women no longer brought in significant calories (they had been primary provisioners in forager societies). Instead, men were needed to manage beasts of burden to plow the fields, and still did the hunting, so women no longer contributed to primary economic production, allowing them to become just another kind of male property (Sanday, 1973). Thus, HSoT suggests diversity in social rules—whether between societies, within a society, or in the same society over time—can be explained using a single function that is implemented using a universal psychological mechanism. It is simply that substantively variable rules for organizing social life apply in different cultural groups.

All of this depends on the legitimacy of the suggestion that human societies can profitably be considered (crude) superorganisms. As indicated earlier, the truth of this assertion depends on human society having undergone a major transition toward strong group-level regulation of social life. An example of regulation at the social level can be seen in eusocial insect colonies, where the lifespan, growth rate, and rate of reproduction of these colonies are nearly indistinguishable from those of individual organisms (scaled to their difference in sizes), suggesting that superorganisms obey the same kinds of metabolic rules as units at the lower level of organization (Gillooly et al., 2010). Similarly, human social networks regulate metabolic features of the individuals within them. In particular, a person is 10% more likely to be obese if the friends of their friends’ friends are obese, even if they do not know those people directly (Christakis and Fowler, 2009). People are not aware of the structure of their own networks and do not control them. Yet how people are connected determines what kinds of functions the network can perform, as well as its qualities (e.g., people tend to be happier in denser networks). Human social networks thus have emergent properties—the primary quality of a complex adaptive system. For these reasons, the collective of all social networks in a population can be called a “human superorganism” (Christakis and Fowler, 2009).⁹

Of course, the mechanisms through which social control occurs remain important. Human superorganisms differ from eusocial insect colonies in not being able to depend on the genes in such groups sharing a common fate—genetic diversity is much higher in human societies than among haplodiploid social insects. Nor do individuals acquire all their goods directly from those who produce them. Hence, kin selection and reciprocal altruism cannot explain the level of cooperation observed in modern human societies. Here, we have argued that the function of moral concerns is to induce behaviors that help such ultrasocial groups cohere, both through fear of punishment for antisocial behavior and the promise of rewards for prosocial behavior (Janoff-Bulman and Carnes, 2013). This need for regulation arose because of the incredible degree of interdependence between people living in societies characterized by a sophisticated division of labor. We have also sought to delineate the dimensions of the moral domain by outlining the kinds of obligations individuals can owe to a human superorganism.

A major advantage of this approach is that it brings a society’s legal systems much closer to its moral systems—they can be shown to overlap in content to a much greater degree than when moral justification is based on traditional approaches limited to fairness and harm. In particular, the categories of moral concern identified in this way also map closely onto the types of offenses in a consensus-based criminal code derived from legal codes from around the world, suggesting that HSoT is consistent with the kinds of problems that social groups actually need to solve through sanctioning systems. HSoT can thus provide a parsimonious yet powerful way of delineating the domain of real-world moral concerns. This could present a significant boon to legal scholars and others seeking to legitimate the criminalization of particular kinds of antisocial behavior.

Hume (1998) thought that the need for justice is a function of a society’s size: in small societies, members can rely on kinship to align interests and thus minimize conflict; he thought that only when society extends beyond the narrow circle of kin does the need arise for rules to regulate human affairs (Hume, 1998). This is essentially equivalent to the claim (without an evolutionary context) of HSoT that morality is necessary when human societies become composed of large groups of unrelated individuals. HSoT claims that moral concerns arose from the need to police social actions in human “superorganisms” (i.e., large groups without shared kinship). It is grounded in major transitions theory (Maynard Smith and Szathmary, 1995), which explains how human groups organized into crude “superorganisms.” HSoT identifies a broader range of moral concerns than competing approaches such as MFT (Haidt, 2007), the MMM approach (Janoff-Bulman and Carnes, 2013), or Dyadic Morality theory (Gray et al., 2012, 2014). It also allows one to develop a sophisticated story about how human morality differentiated itself from simpler forms in our primate relatives. For these reasons, HSoT appears to be an advance in understanding the nature of human morality.

RA: conceived of the theory and wrote the paper.

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.