The Archaeology of Social Norms and Institutions: Issues and Examples The standard social evolutionary view in archaeology was based on the premise that the entities that evolved were social groups, which adapted by means of social institutions to the natural and social conditions surrounding them and attempted to maintain an equilibrium. The anthropologists and archaeologists who reacted against these ideas in the 1970s and 80s argued that such perspectives took away any concept of human autonomy and agency because people just became instantiations of groups, functions and systems. There was no possibility of them changing the conditions in which they found themselves. It was proposed that more attention should be given to individuals as the basis for understanding stability and change in societies. One of the sources for this idea was the ‘practice theory’ of the French sociologist Pierre Bourdieu, based on analysing the way individuals make and change their societies from day to day by interacting with others using the material and institutional resources available to them to achieve their ends. Others have criticised this perspective, saying that such individuals are only found in recent and modern capitalist societies. Some societies do not have any idea of an individual at all; for them only communities exist. However, very few archaeologists have made use of ideas founded in Darwinian approaches to evolution to think through these issues. It remains to be seen how archaeology may contribute to the understanding of the relationships between groups and individuals and, conversely, how evolutionary approaches to these relationships may illuminate the understanding of archaeological evidence. In fact, although most archaeologists are unaware of this literature, the evolutionary game theory framework central to this workshop is not unamenable to archaeological examination. Social institutions are structured patterns of interaction whose structure is provided by social norms that have to be adopted by individuals in a contagious process if they are to spread. Accordingly, social institutions are not something different from other aspects of culture. Potentially, archaeology offers greater possibilities than anthropology for exploring such distributions, both because of the time depth it provides and because archaeological evidence can reveal ranges of variation in behaviour, not just what is typical, whereas ethnographers tend simply to summarise average patterns of behaviour. This paper consists of a series of archaeological case-studies which have attempted to address these issues. The first case-study looks at an example from the Southwestern desert of the United States in the period between 900 and 1300 AD, from the perspective of the role of cooperation and the extent to which it did or did not conflict with individual, or at least family, self-interest. Cooperation and self-interest in the North American Southwest One of the most notable features of the archaeology of the American Southwest is the so-called ‘Chaco phenomenon’, the florescence around Chaco Canyon of elaborate architecture, large-scale ritual, structures such as ‘roads’ and indications of long-distance exchange contacts, in short the ceremonial-political centre for a large region, in the period between 900 and 1300 AD. Some have seen these features as evidence of developed social stratification and the existence of elites, while others have argued that they could have been produced by egalitarian societies of the kinds known from the recent ethnography of the region. Johnson (1989) reviewed the evidence and argued for a distinctly limited degree of inequality and an organisational pattern which had a number of levels but was essentially non-hierarchical in nature.
The amount of surplus available to these communities was small because of the marginal nature of agriculture in the region. However, if real social stratification had existed, one would expect elites to have had the benefit of what there was in terms of better access to food. At Chacoan centres such as Pueblo Alto, there appears to have been more animal bone than expected in relation to its size, which might be taken as evidence of elite consumption. On the other hand, skeletons from burials with elaborate grave goods at these centres suffered subsistence stress sufficient to generate skeletal pathologies, and the large numbers of animal bones are better seen as a reflection of consumption by large numbers of periodic visitors to the centres. As far as the concentration of large amounts of labour in the monumental architecture is concerned, Johnson argues that the construction of residential room blocks was not a particularly labour-intensive activity and that the same was true of the construction of ceremonial ‘kivas’. Furthermore, the number of kivas at sites seems to be directly proportional to community size, implying a relatively low degree of centralisation of ritual activity. Finally, although there is a considerable amount of variation between burials, Johnson notes that the most elaborate ones were almost invariably adults, and that there is little evidence of status differences ascribed at birth through kin group membership. In addition, most ‘rich’ burials only contained one type of trade good, when in fact a variety of items obtained by long-distance exchange were available. The conclusion again is that this is a pointer against social stratification. Johnson accounts for the existence of social variability, population aggregation, labour coordination and exchange in terms of the idea of ‘sequential hierarchy’ (Johnson 1982). He sees this as a solution to the problems of achieving cooperation in human groups as they increase in size. In the face of such cooperation problems, Johnson suggests, there are three possibilities: the group can split; a non-consensual hierarchy can be imposed; or a consensual ‘sequential hierarchy’ can be developed: ‘in the sequential solution, basal organisational units are aggregated into larger (and thus fewer) entities among which consensus can be obtained more easily. Lower order units are subject to minimal potential coercion by higher order organisational entities because the former retain the fission option characteristic of egalitarian [entities] that can be applied if higher order consensus is locally unsatisfactory’ (Johnson 1989: 379). On this view, the small kivas or ceremonial rooms found on small sites represent a level above the household where cooperation problems were resolved in a sanctified context. The larger special structures at large sites then represent one or more levels of sequential hierarchy above the ‘household cluster’ level. The larger sites were in fact multiples of these household clusters, which represented the ‘social modules’ from which the larger sites were built up, and a constant ratio off roughly one kiva to 4-6 units was maintained. In the Chacoan settlements a three-level settlement hierarchy can be seen , reflecting different scale aggregations of such modules and their need for the sacred legitimation of cooperation at different scales. The basic household units, however, were essentially autonomous, potentially mobile and able to disperse, an option which was exercised at intervals: ‘social complexity’ of the Chacoan type was something that appeared and disappeared at intervals in the Southwest. Chaco itself seems to have had storage facilities on an exceptionally large scale, given that the resident population on the basis of the number of rooms and kivas was relatively small. Nevertheless, as we have noted already, the fact that individuals in elaborate burials seem to have suffered dietary stress suggests that the local population did not gain any special benefit from these storage facilities. Johnson suggests that participants in the Chacoan system
could have made use of these stored reserves in times of difficulty, which would justify their contribution to their maintenance. However, the system was fragile, in that the reserves were not great on a regional per capita basis, so that the emergence of continued subsistence problems, for example, as a result of climatic change, would rapidly have exhausted them and made them impossible to replenish. At this point there would have been few advantages in joining the system , and withdrawal by only a small proportion of participants would have led to the collapse that actually occurred. In other words, the Chacoan represented a fragile cooperative solution to problems posed by increasing population size and the associated increase in mutual interactions, a solution based on ritual institutions. What the analysis does not tell us though is whether the cooperation represented a Prisoner’s Dilemma, in which cooperation achieved at the expense of individual interests, or whether it was also the option which gave the best individual pay-offs. A wider explanatory framework for these phenomena has recently been developed by Kohler and van West (1996) from a similar basis in the self-interest of individual households. They begin by showing that pooling of food is most likely to develop in circumstances of high mean productivity, high variability in productivity from year to year, and great spatial differences in productivity. This contrasts with the standard models of risk and its implications which have been used in archaeology, since these presuppose great variability in space/time but low mean productivity – in other words, where there appears to be most need of sharing. However, an analysis of likely consumption functions suggests that sharing is not likely to be in the best interest of individual households when little food is available, since greater utility is obtained, on average, by not sharing. In such circumstances, on the self-interest assumption, sharing will be likely to break down if it is present, and if it is not present it is unlikely to develop. Kohler and van West obtained palaeo-productivity estimates for a 1500 km2 region in south-west Colorado, covering the period AD 900-1300, taking into account soil depth and estimates of soil moisture derived from palaeoenvironmental studies. These were used to define periods of high and low average production. Measures of spatial and temporal variability were also calculated. Aspects of the archaeological record believed to be relevant as evidence for food sharing included community growth and aggregation, the existence of great kiva ritual structures, and the presence of reservoirs. The breakdown of sharing was taken to be evident in the dissolution of aggregated sites. In general terms, the patterning in the archaeological record follows that predicted by the model, in that aggregation episodes are associated with periods when expected cooperation is high; this is the case, for example, with the appearance of the Chacoan, discussed above. The break up of this system and the final abandonment of the region both occur in periods when defection from sharing arrangements is the advantageous thing to do from the point of view of household self-interest. Evidence for cannibalism also occurs in periods when abandonment of sharing is predicted. However, the pattern is complicated by indications of high levels of cooperative behaviour at times when population is high. This is the opposite of what is anticipated by the model, since higher population would be expected to lead to lower levels of food per household. It appears that, while defection into an open landscape at times of low production was easy when population was low, this changed as population increased, because alternative resources to defect to were no longer available. When the region was abandoned towards AD 1300, climatic factors were cutting productivity – and thus the utility of sharing – at the same time that the increasingly full landscape was removing the option of defecting. From this study it emerges that a model of behaviour based on the self-interest of the participants generates patterns in the settlement of the region which correspond closely to those actually found, in terms of the aggregation and dispersal of settlement. There is no
trade-off here between individual and group interest: cooperation occurs when it is in individual interests and disappears when it is not. This appears to be in marked contrast to ethnographically known patterns of social organisation in the Southwest, which place great emphasis on the group at the expense of the individual. However, there is also a contrast in settlement processes: while the period from AD 900 to 1300 was characterised by the cyclical aggregation and dispersal patterns we have seen, since that time aggregated village settlement in areas actually occupied has been uninterrupted. The reason for this, Kohler and van West suggest, is the emergence of new sharing rules, emphasising village level activities, combined with new sanctions against defection. In the societies which then emerged egalitarianism and an emphasis on group values represented an active commitment to cooperation at the expense of maximising individual pay-offs. From cooperation to conflict among Europe’s first farmers The second example examines the history of the first farming communities in Central Europe which, after nearly 500 years, seems to have ended in conflict and collapse. The introduction of agriculture into Central Europe seems to have resulted from colonisation by incoming farmers. It occurred very quickly. The earliest settlements in Austria and Hungary date to c.57/5600 BC while the earliest ones in Belgium and the Netherlands, 1500 km to the west, are already there by 5400 BC. A great deal is known about these early farmers, because their settlements of long houses are archaeologically very obvious and have been the subject of extensive work. In one case in western Germany several square kilometres of land were excavated along the valley of a small stream, the Merzbach, in advance of strip mining for brown coal, so that a whole microregion of early farming settlement could be excavated more or less in its entirety. Detailed analysis of the finds and the stratigraphy has made it possible to characterise the history of settlement in this area, which lasted from c.5300 to 4800 BC. Initially there was a single colonising settlement. This initial settlement continued to be occupied for the full 500 years and continued to be the largest. However, other settlements were subsequently founded in the vicinity, the majority of which had a more intermittent occupation. It seems likely that most if not all the settlements were descended, directly or indirectly, from the founding settlement. We can trace the number of houses in use at a particular time to obtain an idea of the population history of the area. The number of houses gradually increases from the initial 3 to 16 after an estimated 200 years of occupation; it then falls back slightly to 10 before increasing again to 16 after another c. 120 years. Over the following 50-100 years the number of houses rapidly decreases until the area is abandoned altogether; in the final phases before abandonment ditched enclosures existed. The fact that the area was actually abandoned is further confirmed by the results of pollen analysis, which shows that agricultural indicators disappeared and the area was reafforested. In the latest phases there are two ditched enclosures in the area, the latest of which outlasts the latest evidence for houses. Although the history of settlement is exceptionally well documented here, the pattern it reveals, including the existence of a main founding settlement and a final phase with ditched enclosures seems to characterise many other settlement areas of these first farmers. There has been considerable discussion of the function of the late enclosures, but the idea that they were often defensive has been supported in recent years by the finding of two massacre sites dating to this late phase. At the site of Talheim in Germany a large pit contained the remains of 34 individuals, from small children to old people. At least 18 people had
unhealed wounds to the skull; in most cases the shape of these wounds matched the shape of the stone axes used for everyday wood-working. More recently, at the site of Asparn-Schletz in Austria, a similar find was made. The site is an oval enclosure made up of two concentric ditches up to 4 m wide and 2 m deep. On the bottom of the outer ditch were human skeletons and parts of skeletons, making up at least 67 individuals. All the skull remains showed traces of violence and again in many cases the holes in the skulls showed the shape of various forms of axes. Numerous indirect fractures on the base of the skull suggested that wounded individuals on the ground were struck on the side of the head. Many of the skeletons were incomplete and traces of carnivore tooth marks on some of the bones suggest that the bodies lay exposed for some time. The site seems to have been abandoned after the massacre. In summary, in the late phase of early farming occupation we have evidence of ditched enclosures and population decline as well as the two massacres. It has been suggested that conflict arose widely in C. Europe in this period, as a result of population pressure and/or soil exhaustion arising from 400-500 years of constant use of the same areas. It has also been suggested that climate change led to a drier period at this time, which would have led to lowering of groundwater levels; many later settlements show evidence of the digging of wells. A further source of information on aspects of inter-group cooperation and conflict during this period comes from the study of exchange patterns. Almost throughout the period supplies of high-quality flint for tools were obtained from special sources and exchanged very widely. However, in the latest phase exchanged flint declined in frequency at settlement sites and increasing proportions of the lithic assemblages were made up of material from local sources of poorer quality. The same is true of the hard stone resources used to make stone axes. It appears then that at the end of the early farming period, for whatever reason, the exchange routes which gave access to key sources of raw material ceased to operate, after a period of perhaps 400 years, or 16 generations, when relations between settlements seem to have been essentially peaceful and cooperative. One way of looking more closely at what was going on during this period is to examine the changes in the pottery in the Merzbach area mentioned above, where we have very detailed information, especially on the chronology. Stephen Shennan and Richard Wilkinson (submitted) carried out an analysis to test whether the diversity of the changing decorative patterns on the pottery of the two continuously occupied sites in the area could be modelled as a neutral drift process in which change is essentially random (see chapter three) or whether more directional forces were involved, which might provide information about patterns of social relations. Shennan and Wilkinson concluded that the transmission process in the case of the early farming pottery decoration does not correspond to the neutral model; or rather, while drift was occurring, other forces were also at work; specifically, in the later phases there was a bias in favour of using novel decorative motifs. Christiane Frirdich (1994:355), the author of the study whose data were the subject of this analysis, argued that the long-lasting uniformity in the use of a small number of traditional band types which characterized the early phases of occupation was maintained by the imposition of strict social norms on band type choice and that the rapid change that followed represented a relaxation of such norms. Our results do not provide evidence for the imposition of strict norms in the early phases, but do suggest that the change which occurred during the later phases was an active preference for novelty, rather than a relaxation of previous constraints.
Interestingly, the move towards a pro-novelty bias in the transmission of decoration types, which occurred in phase VIII, did not take place straightforwardly in step with the developing colonization of the Merzbach micro-region. New sites were already being established in phase V and this process continued through phases VI and VII; between phases IV and VII the number of contemporary houses also went up, from seven to sixteen, a number subsequently equalled but never exceeded. Phases VIII–IX, when the switch to pro-novelty or anti-conformist bias takes place, are actually a time when the number of houses and occupied sites decreases, to four sites and ten houses in phase VIII, with four sites and twelve houses in phase IX, although in the following phases it gradually increases to a maximum of seven sites and sixteen houses in phase XII. Nevertheless, the pro-novelty bias may reflect a concern to establish distinct local identities once the area had more or less filled up. It may be that the rise in the number of houses in the valley to its maximum of 16 led to some sort of local crisis, reflected in the drop to 10 houses in the following phase. It was precisely at this time that potters started actively selecting for novelty in their pottery decoration. The next time, a few generations later, that the number of houses in the area reached 16 it was followed by decline and final abandonment, after a phase in which the evidence for the building of enclosures suggests the probability of violence. The close initial links of common descent between communities generated by the speedy colonisation process and the continuing close proximity of the settlements within particular micro-regions would have provided ideal conditions for the emergence and maintenance of pro-social norms of interaction between people living in different settlements. However, for the first couple of hundred years of this colonisation process there need have been no conflict of interest between the individual and group levels: sufficient resources were available for everybody, hence the five-fold growth in population. The time when the growth in house numbers stops and the number actually declines must be the point where density-dependent checks on population began to be effective and this is where we might expect conflicts to start arising between group and individual interests, producing Prisoner’s Dilemma situations. No effect is seen at the larger regional scale, since the free movement of lithic materials seems to continue. Nevertheless, at a local level something seems to have happened, even if it is unclear how it related to possible density-dependent checks and conflicts of interest. Individual settlements show a pro-novelty bias in their pottery production. A cladistic analysis of the pottery decoration from all the settlements in the Merzbach valley carried out by Mark Collard and Stephen Shennan showed that individual settlements mostly have their closest links with a single ancestral settlement phase, rather than with several, suggesting that interaction between nearby settlements was not being reflected in ceramic style choices. In these new circumstances though the number of houses in the Merzbach region gradually climbed once again to the previous maximum. The subsequent decline in population and eventual regional abandonment could perhaps be seen as a ‘tragedy of the commons’ situation, in which individual households continue to try and secure their own subsistence interests at the expense of otheres, with agricultural failure for everybody as a consequence. However, it is more likely that conflict was at a higher level. Precisely what level is unclear. As we have already seen, the decline in the use of high-quality lithic resources obtained from distant sources and their replacement by local resources of lower quality can be taken as an indication of the breakdown of exchange relations, but this could result from conflict at any scale. The number of enclosures in a given micro-region may be a better guide. In the case of the Merzbach valley, the evidence is such that it is hard to say whether there were two enclosures, or only one, during the final phases. In other words, if they are indeed a reflection of conflict, as the massacre evidence elsewhere suggests, it is unclear whether there was conflict within the Merzbach
micro-region or between the Merzbach community as a whole and others. If, for the moment, we assume the latter, then we can raise the question of whether cultural group selection could have been operating, with the different micro-regions as the groups competing with one another. The fact that the Merzbach valley was abandoned rather than re-occupied by a competing group from elsewhere with different cultural traditions suggests that individual/household level natural selection is more likely to have been the relevant process, even if the abandonment resulted from emigration rather than the demise of the people concerned. However, in some areas of Central Europe successor communities with new cultural traditions did emerge and spread, probably by the sort of population mechanisms discussed in chapter five. The archaeology of hierarchies and group selection The neo-evolutionary anthropological schemes of authors such as Elman Service, with his sequence from bands to tribes to chiefdoms and states, have been extensively discussed and criticised, not least because of their implicit naïve group selectionism. In principle anyway we can now see such schemes in a rather better light in terms of multi-level selection. For example, if a particular community or group of communities makes the innovation of having a chief and that innovation results in the group concerned competing more successfully with other groups, killing or absorbing their members and fissioning into other similar successful communities, then we have an example of cultural group selection at work, since the traits characteristic of chiefdoms are the key to success and will proliferate as a result. There will not necessarily be a conflict here between the group and the individual level – individuals may gain greater reproductive success as members of the group then they would do otherwise, despite having to pay costs such as tribute. Nevertheless, they stand or fall together in success terms. If the group is defeated and ceases to exist, the conditions of defeat are likely to be such that individual reproductive success is affected for the worse, at least for males. The replicative success of their cultural traits is also likely to be adversely affected. To use philosopher David Hull’s terms, the level of the key evolutionary ‘interactors’ has shifted: it is no longer individuals or households, but chiefdoms, while the ‘replicators’ whose propagation is affected will be genetic and cultural traits at the individual level, as well as group level traits. This sort of process may be what we see in the Central European Bronze Age. At the beginning of the Bronze Age, c.2300 BC, there was a widespread social pattern of small-scale societies in dispersed groups practising varying combinations of plough agriculture and stock-rearing. Since plough agriculture is historically and ethnographically a male-dominated activity, there are clearly good grounds for believing that it would have led to a dependence of women on male-controlled resources which had not previously existed, or at least to nothing like the same extent. The process is likely to have been compounded for other reasons. Plough agriculture involved the use of fixed fields rather than temporary clearances in the forest and we know that fairly permanent clearance was becoming more widespread at the end of the fourth millennium, when animal traction had been introduced. We also know that the appearance of permanent pasture at the same time, by cutting the need for the provision of leaf fodder began to lift the constraint on the number of animals that could be kept, especially cattle. Although it is difficult to imagine that archaeologists will ever succeed in proving it, it seems likely that investment in such long-term resources would have been associated with the rise of the institution of inheritance. Since the lifetime reproductive success of sons is correlated with that of their fathers, but not that of mothers with their daughters, and inherited
resources generally give greater reproductive returns to males than to females, when they exist it is most likely that they will be transferred from father to son. Furthermore, an analysis by Holden and Mace of the evolution of patrilineality and pastoralism led to the conclusion that matrilineal societies become pastoral and pastoral societies patrilineal as inherited resources become more important. As a consequence women become more dependent on men. It seems likely that, whether or not earlier societies were patrilineal, plough agriculture and permanent clearance had the same sorts of consequences in terms of the emergence of new patterns of male control of women that had not previously existed. At the same time, there would have been competition between women themselves for the best male resource holders for long-term parental investment, involving strategies that emphasised fidelity and therefore paternity certainty. Although male-female symbolic distinctions had always been important, they certainly seem to have received a new emphasis at this time. In the western Alps about 3100 BC we find a new emphasis in the archaeological assemblages on bows and arrows, stone hammer axes and flint daggers. At the same time there is an increase in the frequency of items of personal ornament, including for the first time pendants in the form of male genitals, while the bows and arrows, axes and daggers become symbolic male attributes. Pierre and Anne-Marie Petrequin (1988) suggested that there was an asymmetry between men, who had access to exotic materials obtained by exchange for such things as hammer-axes, daggers and arrowheads, and women, whose tools were made of local materials. An analysis of north Italian petroglyphs dating to the same period by John Robb (ref), led to similar conclusions. In hunting scenes weapons were used by males while in others, halberds (a kind of dagger blade hafted like an axe), axes and especially daggers were consistently used to distinguish males from females. Furthermore, although we know from bone assemblages that a variety of animals was hunted, the only one commonly represented in art is the deer, and in particular the stag with its antlers; if the gender of the hunters is represented it is male, distinguished by a phallus. Images of ploughing show a similar pattern – oxen are represented in particular by their horns and those using the ploughs frequently identified as male. Robb suggested that these representations would have contributed to an ideology of male power, in which the male-female distinction was a hierarchical one, although relations between males were not yet stratified; this came later. By the later early Bronze Age around 1600 BC local site and social hierarchies, evidenced in settlement distributions, rich burials and prestige metalwork, had developed that were competing with one another. Although prestige competition between males no doubt continued within the local polities, it was now the between-group competition that defined the conditions of social and economic life. Those groups that had managed to attain greater power within a particular local area needed to maintain it through the accumulation of wealth and followers. As John Robb has proposed (), strategies for different men may have gradually diverged, in that well-situated groups could have used their success in exchange and/or subsistence production to attract followers, which would have led to further success. As such groups became more powerful, independent, peripherally-situated men would have found increasing difficulties in gaining wealth and prestige and, as a result, in reproducing themselves, physically and socially: ‘They may have accepted second-class status in a powerful unit as a route to personal power more effective than life without powerful friends and relations. The result would have been the simultaneous enlargement of [some] units and differentiation within [and between] them’ (Robb). In addition, the emergent competition between these higher level units would have
made followers more vulnerable to the depredations of their leader’s competitors and thus yet more dependent on their own leaders. On this model, the growth of a differentiated warrior ideology by the end of the Early Bronze Age, out of the earlier more generalised male one described above, seen, for example, in the wide distribution of similar prestige weapons, would have been the outcome of a group selection process, since groups which did not adopt such innovations would have been broken up and absorbed. An intra-group corollary of this would have been reproductive stratification associated with the rise of male inequality. Thegn Ladefoged and Michael Graves have demonstrated a similar shift in the scale of selection in a very different context, that of subsistence strategies in Polynesian Hawai’i. They analysed aspects of the size and distribution of fields, arguing that these evolved through time as a response to optimality and risk factors. During the first period of settlement in the region of Hawai’i that they studied, coastally-based communities used dispersed fields focussed on areas of higher rainfall and supplemented agriculture with marine resources. In the following phase less dispersed field systems were constructed in upland areas which had higher and less variable rainfall. This entailed increased energy costs, because of the increasing distance of the agricultural areas from the coast. Costs were further increased by the construction of smaller fields than in the previous period. The field walls served as erosion inhibitors and as windbreaks, slowing evapo-transpiration from the sweet potato crops (L and G 2000, 427). In the final period additional fields were created still further away from the coast and located in areas of lower rainfall. L and G propose that this new development represented a shift in the scale of selection operating on agricultural production, because it arose as a result of the activities of the chiefly elite that had developed by this time. In other words, individual food-producing households and their decisions were no longer affected solely by the direct costs and benefits of their own agricultural investment but depended also on the political activities of chiefs. These actively instigated agricultural production in more marginal areas of the district , probably using some of the produce for ceremonies. Nevertheless, the larger organisational scale of the chiefly economies made this a viable possibility because they had the potential to make up for poor harvests in these areas in drier years by the use of distributional mechanisms. These buffered both chiefs and commoners from the occasional crop failures resulting from dry-farming in areas with marginal rainfall, but of course at the expense of yet further increased costs to the producers since some of the crop was reserved for chiefly use. As L and G are aware, their analysis takes us right back to the neo-evolutionary arguments of the 1960s and 70s about the adaptive and exploitative aspects of chiefdoms, and in many respects supports Service’s arguments about the importance of chiefly redistribution, although their their theoretical foundations are different from his group functionalism. In fact, they do not provide the basis for evaluating whether cultural group selection was taking place, in the sense that groups organised in this way were expanding and absorbing those that were not, or that other groups were adopting the new organisational pattern to be able to compete. This would need a further study, but on the basis of everything we have seen in this chapter it seems a reasonable hypothesis. In any event, it seems reasonable to suppose that the reproductive success of chiefdom members and the replicative success of their cultural traits would be affected by the fortunes of the chiefdom as an entity, just as we suggested for the previous example. Conversely, when the overhead costs of such systems become greater than the benefits they provide to the participants, they become vulnerable to collapse and fragmentation, as Joseph Tainter proposed some years ago for the best known of all collapses, the Fall of the Roman Empire.