Their experiment also led to other findings which are approximately as sociobiological theories of ethnocentrism would predict. Ethnocentrism went down the higher the cost, because cooperating with ‘same colour’ cheaters would impose a penalty and the risk of altruism not paying off would, therefore, gradually become too high. Consistent with this, it is only a tiny minority of humans — like Private Barber whom we met earlier — who are prepared to make enormous sacrifices for their ethnic group. Also, the higher the number of colours on the grid, the higher was the level of ethnocentrism. This would be because as the variety of colours on the grid increases, colour becomes an increasingly accurate indicator of relatedness, making discrimination increasingly effective in perpetuating your own colour. However, when ‘randomness’ — via increased immigration (of any colour) and increased mutation — was increased, then ethnocentrism decreased because tags became less accurate indicators of relatedness, making discrimination less effective. But, when the extent of any of these variables was either halved or doubled, around two thirds of agents still adopted an ethnocentric strategy. This would imply that reducing ethnocentrism — using colour as a proxy — could only be achieved by rendering ethnic cooperation extremely costly to one’s interests, creating a diverse population that all had the same colour tag, or by maintaining a highly ethnically homogenous population. As we will see later, these are some of the factors that reduce ethnocentrism in real life. Ethnocentrism became just as dominant even if the simulation began with a full lattice and no immigration was allowed. Altering the programme so that an agent could distinguish all four colours rather than just between his own colour and colours that were ‘other’ also had very little impact on the results. The possibility that agents would misinterpret another agent’s colour as much as 10% of the time, likewise, did not significantly impact the results. This is consistent with the argument that though there is diversity within races, as long as — on average — there are noticeable differences, races are meaningful categories and racial features are meaningful markers of ancestry.
The model allows a fascinating insight into how the ethnocentric strategy develops. In the early periods of a run, scattered immigrants create small regions of similar agents. Colonies of those willing to cooperate with those of their own colour arise relatively quickly but, over time, they face the phenomenon of free-riding egoists, who arise via mutation. Free-riding egoists cannot be suppressed by ethnocentrics of the same colour and consequently they gradually erode cooperative regions. At the same time, regions with different attributes will tend to expand until they are in contact with each other, with ethnocentric regions growing the most quickly. Once this stage is reached, ethnocentric groups will expand at the expense of less ethnocentric groups and, as such, the free-rider problem becomes, in essence, policed by this process. Egoists of one colour will fail to reproduce in the long term, because they will not receive cooperation from the other colour at their borders or from the egoists behind them. Accordingly, over time, the percentage of ethnocentrics inevitably rises even if there are periods where they decline in the history of a specific colour group.
One further result, described as ‘remarkable’ by the authors, was that the ability to distinguish between the in-group and the out-group actually promoted cooperation. As long as agents could clearly distinguish their own colour from that of others then even doubling the cost of cooperation sustained a cooperation rate of 56%. However, when agents — in the doubled-cost case — could not clearly distinguish their colour from that of others then cooperation fell to just 14%. As such, as the cost of cooperating increases, the ability to distinguish between in-group and out-group becomes increasingly essential in order to maintain that very cooperation. Indeed, the ability to distinguish between groups has been shown to be the basis for social capital (Putnam, 2000) within a group. This finding would be congruous with the importance of clear ethnic markers among relatively closely related ethnicities, such as scarring or distinctive dress. Even if we accept the veracity of Genetic Similarity Theory, there would be significant genetic variation within an ethnicity and, as such, cultural markers would sometimes be very important to discerning whether a person was part of your ethnicity and thus whether cooperation would aid the interests of the group, a desire selected for by ethnocentrism. But the authors’ overall finding would seem to make intuitive sense when there are at least two colours and the possibility to cooperate or defect on that basis. Eventually, as the lattice fills up, the colour group with the most ethnocentrics will attack the colour tag with fewer ethnocentrics. The selfish agents will not be backed up by their own side while the humanitarian agents will be betrayed by other side, leading to a selection pressure for ethnocentrism.
5. Developments of the Model
A number of researchers have developed this model. Schultz et al. (2009), running a similar simulation though including more than one ‘shot’, found that ethnocentric clumps of agents directly suppress ‘humanitarian’ (those who always cooperate) agents from different groups. At the same time, they found that ethnocentrics were more effective than humanitarians at suppressing groups of free-riders from the same group. These findings were extended by Kaznatcheev (2010) who employed the Hammond and Axelrod colour-tag model. He argued that being ethnocentric was more cognitively demanding than being either selfish or humanitarian because it involved a complex, discriminatory decision. As such, he imposed a cognitive cost upon ethnocentrism as part of the model and found that ethnocentrism was, in this context, not very robust. Imposing only a small cognitive cost on ethnocentrism allowed humanitarians to dominate ethnocentrics. As such, he demonstrated that the cost of discriminating between in-groups and out-groups needs to be extremely low in comparison to the potential pay off for so doing or people will be less inclined to do it.
This raises some intriguing possibilities in terms of the nature of ethnic markers. It implies that ethnocentrism will be heightened if ethnic difference is conspicuous. This would potentially mean, for example, that if a group — such as fundamentalist Muslims — emphasized their ethnic difference through clothing, then they would arouse stronger feelings of ethnocentrism than precisely the same people dressed in normal clothes. But they would also evoke higher levels of positive ethnocentrism and its accordant benefits. Kaznatcheev’s simulation also found that levels of in-group cooperation — within colour groups — quickly collapsed once humanitarians became dominant and cooperative action decreased. The implication of this is that humanitarianism cannot sustain high levels of in-group cooperation. Ethnocentrism does sustain this kind of cooperation and, as such, when the cost of cooperation is extremely high then ethnocentrism will be a necessary prerequisite of cooperation. This would imply that humanitarian societies will have trouble winning all-out wars.
In a related experiment using the Axelrod model, Schultz et al. (2006) found that areas of the lattice, later in the experiments, with substantial colour diversity often included lots of humanitarians. They theorized that this was probably because humanitarians cooperate with everybody and so boost the reproductive fitness of all adjacent agents. Accordingly, adopting a humanitarian strategy will causally foster a multi-ethnic region in a context of at least two competing ethnic groups. Moreover, agents — of any colour — who have selfish or ethnocentric strategies are likely to exploit the humanitarians thus diminishing the reproductive potential of populations which are dominated by humanitarians. A further noteworthy finding was that in the early stages of an experiment — before two colours began to come into contact — humanitarianism would dominate as much as ethnocentrism but selfishness and traitorousness never did. This is because by failing to cooperate with others at all, they reproduce less over time than if there were random mutual cooperation. So, the overall order, in terms of strategy success, is ethnocentrism, then humanitarianism, then selfishness, and finally treacherousness. Cheaters, it seems, ultimately never
prosper.
The most recent deployment of ethnocentrism computer modelling is Hartshorn et al. (2013). They attempted to understand in greater depth what it is that always ultimately gives ethnocentrism the edge over humanitarianism. As with other simulations, they found that, in a 2000 generation simulation, early generations are marked by intense competition between ethnocentrics and humanitarians. In some worlds, ethnocentrism wins outright, in some humanitarianism attains a fleeting dominance, while in others the two strategies are neck-and-neck. This changes at around generation 300, when ethnocentrism begins to pull ahead. Hartshorn et al. argue, as have others that there are two possible explanations.
6. Why Do Ethnocentrics Always Win?
1. Ethnocentrism beats humanitarianism because ethnocentrics are better at suppressing selfish free-riders. If an ethnocentric group meets a group dominated by selfish individuals, they’ll refuse to cooperate. Over time, thanks to the ethnocentrics’ mutual cooperation and the selfish group’s refusal to cooperate even with each other, ethnocentrics will reproduce faster than the selfish and thus expand at the selfish group’s expense. Meanwhile the humanitarians will waste their precious reproductive potential helping out free riders who give them nothing in return. This is known as the ‘mediation hypothesis’, and it is the mechanism favoured by Hammond and Axelrod (2006).
2. Another possibility is that ethnocentrism simply beats humanitarianism outright. If we imagine an ethnocentric group next to a humanitarian group, individuals on the group boundary benefit from the cooperation of their own group-mates behind them. But the ethnocentrics at the front doubly benefit from the cooperation of humanitarians of a different colour tag. This is known as the ‘indirect hypothesis’.
Hartshorn et al. wanted to understand why a tipping point, followed by a dramatic rise in ethnocentrism, is reached specifically around generation 300. They found that 300 generations was the time when the world started to become increasingly crowded. In the early stages, the world is sparsely populated, intergroup contact is rare, so there is little opportunity for ethnocentrism to beat humanitarianism, either through the mediated or direct mechanism. In order to establish what then tipped the scales in favour of ethnocentrism, the authors restricted the simulation. They ran worlds with just humanitarian, selfish, and traitorous individuals; or just humanitarian and selfish, or even just selfish individuals. They established that, contrary to what Hammond and Axelrod had argued, it was not free-riders that doomed humanitarians, as in the absence of ethnocentrism the dominant strategy was humanitarianism. 2000 generations of non-cooperation simply led to a lower population than 2000 generations of humanitarianism. Thus, humanitarianism is nearly as good at free-rider suppression as ethnocentrism but ethnocentrism has the edge because ethnocentrics take advantage of foreign free-riders. Interestingly, they also found that the strategy order was the same except when ethnocentrism was removed. In that simulation, treacherous agents performed better than selfish agents. The authors explain this as follows:
When traitorous agents of one cluster collide with ethnocentric agents of another, the ethnocentric agents earn outcome b, exploiting cooperating traitors by defecting against them. Just as ethnocentrism is poisonous to humanitarians, it is also poisonous to traitors, who incur a cost of c in such interactions.
The authors found that chance is the factor behind early humanitarian dominance is some simulations. If a group of humanitarians of the same colour simply happen to end up together on the lattice by chance, isolated from others, then humanitarianism will spread through the population via a founder effect, a phenomenon we have already discussed. In general, however, ethnocentrism will always dominate humanitarianism.
7. Limitations to the Computer Modelling of Ethnocentrism
Clearly, there are a number of limitations to Axelrod and Hammond’s model. Real-life inter-ethnic interaction is more complex because there are many markers of ethnicity rather than just one and reproduction occurs sexually rather than asexually and can happen across ethnic boundaries (although there is a degree to which the model controls for this by testing for the ability to distinguish colours). Also, all of the cooperation is ‘one shot’ whereas real life involves continuing interactions and thus, for example, a greater likelihood of cooperation between those of different ‘colours’ when they happen to be in close proximity. This is perhaps why, in real life, ethnocentrism is not as dominant as this model would predict even in conditions of natural selection, a point we will explore below. In real life, one’s chance of dying is not random and set at 10%. It will depend on factors such as one’s mutational load, one’s lifestyle (and thus one’s personality and intelligence), and one’s environment. Moreover, the simulation assumes a high level of viscosity; that people do not move around much. This was probably quite true in the early evolution of humans but it is clearly less true now and, therefore, it would be interesting to see what impact this factor has on the different strategies. Indeed, in a critical examination of the Axelrod model, Jansson (2013) attempted to render the model more realistic by allowing people to discriminate not just in terms of ethnic markers, but in terms of kin; as in real life they would be likely to make this distinction. Janssen found that kin-discriminators soon came to dominate. Where ethnic markers coincided with kin-markers then the two strategies simply coalesced. So, this research would imply that a successful evolutionary strategy is to clearly display kin or ethnic markers, as can be seen in religious groups, for example (see Dutton, 2008).
These models are highly abstract and heavily simplify the human experience. However, they are useful because they are likely to have captured some fundamental principles of evolution that occur in all species and environments. These models would seem to imply that it is inevitable that some form of ethnocentrism is always likely to develop whether two ethnic groups are placed together or whether, due to mutation over time, one original group gradually splits into to two decreasingly similar groups. In addition, some of the models give us reason to believe that ethnocentrism requires strong rewards in order to be sustained. This allows us to cautiously make predictions regarding why ethnocentric behaviour might rise or fall in certain environmental circumstances. And if there was no longer a strong selection pressure for it then it would be especially likely to decline.
8. Bruce Charlton’s Model
The computer modelling analysed above would appear to imply that ethnocentrism is a product of intergroup competition. However, English psychiatrist Bruce Charlton (15th December 2015) has argued that, though this is indeed the case, ethnocentric behaviour can also be selected for even if a group is isolated from other groups of the same species for a long period of time. In other words, a group can become ‘group selected’ even in the absence other groups. In many ways, the model Charlton proposes makes intuitive sense. If it is in the interests of the survival of the species to become pack animals, then certain kinds of pack will be more likely to survive the pressures of natural selection than others, even if there is no alternative pack of the same species against which to compete. Consequently, a certain kind of pack will be selected for meaning that, in a sense, the pack, in itself, will select for certain kinds of individual. As we have seen from the computer modelling, these will tend to be ethnocentric individuals who are prepared to make sacrifices for the good of the pack as a whole, and thus for other pack members. This being the case, it would be possible for high levels of positive and negative ethnocentrism to develop even if a pack was relatively isolated from other packs of the same species. This would happen if the group in question needed to be highly group selected; especially it was evolved to an extremely harsh environment in which the benefits of a cohesive group were very strong. We will see below that this would potentially make sense of the high levels of ethnocentrism found among Northeast Asians.
9. Conclusion
In this chapter, we first examined the Big 5 model of personality in an attempt to see whether or not ethnocentrism can be reduced to a certain kin
d of personality. We found that although there are dimensions of personality which are associated with elements of ethnocentrism it is not really possible to talk of an ‘ethnocentric personality’. We have, however, noted evidence that ethnocentrism, like religion, may be regarded as an instinct. We then looked at the research attempting to discern the reasons for the spread ethnocentrism and we found that, drawing upon computer modelling, the more ethnocentric group will always outcompete the less ethnocentric group when other factors are held constant. Finally, we have looked at Charlton’s theory that environmental conditions alone could select for ethnocentrism. We will now turn to ethnocentrism’s specific genetic and environmental causes and how this explains differences between races and between ethnic groups.
Race Differences in Ethnocentrism Page 12