13
Framing the Rise and Variability of Past Complex
Societies
Gary M. Feinman And Linda M. Nicholas
The most important issue confronting the social sciences is the extent to which human behavior is shaped by factors that operate cross-culturally as opposed to factors that are unique to particular cultures. (Trigger 2003:3)
Cooperation, inequality, leadership, and governance have long been recognized as fundamental characteristics of the human career. Understanding the ways in which humans forge identities, form socioeconomic networks, and establish institutions remain critical issues for the human sciences today (Pennisi 2005). And yet, as Trigger (2003:3) remarked, we are still searching for the appropriate balance between general and specific factors (or processes and history) when it comes to building our frames to examine the range and nature of human behavior. In this chapter we conceptualize the issues surrounding the development and diversity of large-scale human cooperative formations from a perspective that embraces both the critical role for comparison and the multiple pathways that human historical processes take (e.g., Little 2000).
The causes and consequences related to the formation of hierarchically organized societies have been seen for centuries as key hinge points in the human career (Childe 1950; Haas 1982). As the first episodes of these transitions to chiefdoms and states occurred before the advent of writing, a focus on the rise of hierarchically organized societies has long been a core comparative issue for archaeologists (e.g., Adams 1966; Wright 1977a). Many years ago, Henry Wright outlined the challenge of this topic (Wright 1977b:215): “The construction of research strategies for the investigation of state origins is complicated by two principal factors: the type of explanations required and the samples available to test them.” Since that time, we believe we have made more progress on building that empirical record than on finding an appropriate conceptual frame (although there remain untold lifetimes more to do on the former front as well).
In fact, the last overarching archaeological framework designed explicitly to tackle this suite of questions (Flannery 1972) has in large part stood (fostering highly productive results) for decades (Brumfiel 1992:551). Yet not surprisingly, over many years certain conceptual issues (particularly in regard to considerations of diversity, agency, and scale) have arisen even among those committed to explicitly comparative approaches (e.g., Blanton et al. 1996; Blanton and Fargher 2008; Brumfiel 1992; Chapman 2003; Earle 1997; Feinman 2012).
From the outset, we stress that we do not present definitive answers here, nor is our goal to offer an in-depth review or critique of the broad corpus of literature focused on the rise of hierarchically organized societies. Rather, our intent is to guide and redirect our framing theory or conceptualizations of this question toward comparatively based explanations that aim to understand patterns of diversity rather than uniformity. To date, most of the ways that archaeologists have looked for generality have not been adequately explanatory, widely influential in other disciplinary arenas, or able to achieve broad consensus even within our own field. The recent dearth of strong theoretical explanations for the development of cultural complexity (a point noted recently by Jeremy Sabloff [2012:xvii]) encourages those of us interested in more general and comparative approaches to consider alternative ways to frame this issue.
The critical link between frameworks and scientific advances when we are tackling big issues is illustrated through a reading of Ernst Mayr’s (1985:404–408) provocative history of biological thought. Mayr recounts that despite the general accumulation of pertinent biological data, Lamarck’s efforts to understand the complex history of life did not prompt a breakthrough because he basically asked the wrong questions. Charles Lyell, in contrast, posed the right questions, but he came up with the wrong answers. Charles Darwin, building on their works, framed the right questions and arrived at answers that we now know were more fully explanatory than those of his predecessors, thereby advancing disciplinary thought. To put it another way: “Discovery requires an aggressive and critical engagement with the status quo” (Gerring 2012:28). Or, as “a brief review of the history of archaeology makes clear . . . advances have come less by resolving major questions than by superseding them with better questions” (Drennan 1987:320).
Prior Frameworks for Archaeological Generalization on State Origins
No one would deny that every historical case of archaic state emergence is in certain respects unique (Wright 2009:122) and important to unravel in detail, both to address specific historical questions and to provide the living and future descendants of those cultural traditions with information concerning their pasts and histories. Yet for centuries, scholars also have looked for more general processes or conditions that help explain this suite of evolutionary transitions, the differences as well as the parallels (cf., Pauketat 2001).
Some of the first comparative efforts advanced the notion that early civilizations arose in a specific environment (river valleys) or in response to particular prime movers (Wittfogel’s [1957] irrigation management) (e.g., Steward 1955). Such attempts neither adequately accounted for the full suite of empirical cases nor drew convincing causal chains to explain the relevant societal transitions or the specific timing of the focal historical developments (Flannery 1972). Perhaps, over the years, the most cited and discussed of such prime movers has been Carneiro’s (1970) circumscription model, which advanced population pressure, geographic boundedness or circumscription, and offensive/conquest warfare as a lineal set of triggers that led to state development. Over the years, both conceptual limitations and empirical shortcomings of this model have been repeatedly outlined (Blanton et al. 1979; Flannery 1972; Roscoe 2000). In turn, these critiques have prompted efforts by the author (Carneiro 1987, 1988) to broaden the original causal scenario to accommodate problematic cases. And yet, despite this tinkering, population pressure remains elusive, the envisioned impermeable or circumscribed region remains historically more mythic than real (Ofek 2001; Wolf 1982), and the proposed empirical link between offensive warfare and the emergence of states has not been shown to be universal.
The so-far futile search for a single, specific prime mover or set of universal conditions that invariably provoked early state development prompted a reorientation toward a series of approaches that more explicitly compared historical sequences in an effort to define more general properties. For example, refining earlier studies by Carneiro (1962, 1968), a recent cross-cultural study (Peregrine et al. 2007:77) uses Guttman scaling on 20 cases from the Collection of Archaeology in the Human Relations Area Files to outline a broad sequence of cultural change. Basically, their scaling confirms what decades of archaeological analyses have already evidenced, that sedentism, domestication, and increasing population densities preceded the formation of states (cf. Jacobs 1969; Taylor 2012). Yet even this broad-brush scaling does not conform with various historical specificities (region to region), such as precedence of sedentary life before agriculture in the Levant, the emergence of villages and inequality among sedentary foragers in several global areas, and the much later use of metal (compared to this scaled sequence) in Mesoamerica.
Another tack has been to look for general, cross-cultural parallels in sequences leading up to state development. For example, Yoffee (2005:44) has argued that when states arise they tend to be small, ensconced in networks of city-states or peer polities. Out of these networks, a large, more dominant polity can eventually emerge. In sharp contrast, Marcus (1998) has proposed that early states tend to be large polities that oversaw expansive territories, and that only later did these entities sometimes break down into smaller states. Yet neither of these sequences has turned out to be universal when we consider the suite of cases of complex society development. In fact, in the two regions where we have investigated, we see the latter pattern in the Valley of Oaxaca, where one center, first San José Mogote and later Monte Albán, seems to have dominated the region basically from the outset of sedentary villages (Blanton et al. 1999; Kowalewski et al. 1989). Whereas in coastal Shandong (China), several roughly coequal and large centers arose and were rather evenly spaced across the landscape, more in line with the former model (Feinman et al. 2010; Underhill et al. 2008).
Comparative Data: Diversity Rules
To date, archaeologists have for the most part looked for uniformities in the underlying or initial conditions, commonalities in the processes leading to, or the basic properties of, early complex societies. But these efforts have all been confounded by seemingly exceptional or nonconforming cases. In a sense, this realization should no longer be surprising. Frankly put, the more data that scholars collect relevant to the emergence of complex societies across the globe, the more variation becomes apparent, and, significantly, that variation is not simply associated with elements that might be considered local, historical, idiosyncratic, or purely culture bound (Earle 1997; Smith 2006:13; Trigger 2003:660–661).
Several recent studies, for example, have compared multiple archaeological sequences of demographic change, village formation, and population nucleation prior to the emergence of states. In one, Bandy (2008) examined 36 cases (in two cases, large villages as he defined them never emerged), finding that there was significant variation in the time lag between the first sedentary communities in a region and the advent of large villages in that area (> 300 people and 3 ha in size) (figure 13.1a). In some cases, there was little or no lag, while in others as many as 4,000 years elapsed. In certain cases one dominant village emerged, while in others, multiple large communities arose (figure 13.1b). Bandy also isolated eight of his cases where states ultimately developed. For each of these eight cases, the initial time lag from the earliest sedentary villages to large villages was 1,000 years or less (figure 13.2a) (a point that we return to), although the lag from the presence of sedentary villages until the rise of states was more variable (figure 13.2b), as was the timing from large village formation to state development (figure 13.2c).
Figure 13.1. Time from the first settled agricultural villages in a region to the presence of large villages: (a) 34 cases where large villages formed; (b) cases where multiple large villages formed versus those where a single village was dominant. (Adapted from Bandy 2008:figures 2 and 3.)
Figure 13.2. Village formation and state development in eight world areas where states ultimately developed: (a) time from the first settled agricultural villages to the advent of large villages; (b) time from the first settled agricultural villages to primary state formation; and (c) time from the advent of large villages to primary state formation. (Adapted from Bandy 2008:336–337, figures 5 and 6.)
In an amplification of this analysis, Peterson and Drennan (2012) compared 11 archaeological sequences, some overlapping with Bandy’s and others not. They examined the process from the first sedentary agricultural villages in each region (figure 13.3) through the advent of more nucleated centers, again finding considerable variation in both the timing of change and the size and number of centers that were founded. They also made rough estimates of the tax rates per citizen in each context and were surprised when the highest tax rates were seemingly being paid in three of those cases in which the central settlements were comparatively small and where more hierarchical formations did not directly develop (more on this as well).
Figure 13.3. Sequence from the presence of the first sedentary agricultural villages (the beginning of the Neolithic) to the advent of more nucleated centers for 11 archaeological cases. (Adapted from Peterson and Drennan 2012:figure 6.14.)
Fostering this point, Fletcher (2012) has reminded us that the cities associated with early states were not always compact; sometimes they were dispersed. They also vary markedly in size (figure 13.4) and seemingly in their basic systems of governance (e.g., Fargher, Heredia Espinoza, and Blanton 2011; Feinman 2001), with some stressing the focal primacy of individual rulers and others exhibiting less personal glorification and greater indications of power-sharing. Michael Smith (2009:28) has noted this diversity in regard to early cities, but the same is at least as valid for archaic states. What we need are comparative approaches that can accommodate and explain elements of diversity in the evolution of human institutions.
Figure 13.4. Size of the largest cities in early states. (Data drawn from Blanton et al. 1993; Chandler 1987; Culbert et al. 1990; Kowalewski et al. 1989; MacSweeney 2004; Modelski 1997, 1999; Morris 2010; Stanish 2010; Underhill et al. 2008; Yoffee 2005.)
Building Theory to Account for Human Organizational Diversity
Clearly, our focus on the diversity of early states and the processes leading up to them is not meant to discourage comparison or the building of theory on a more general level. Rather, it serves to underpin alternative ways to conceptualize this complex issue. Since Childe (1950), it has not proven productive to search for the unilineal pathway, the uniform proximate cause, or the universal set of unvarying properties that can account for early complex societies and their diversity. But fortunately, there are conceptual roadmaps for building and using cross-culturally applicable constructs to account for human behavioral variation (Ellen 2010:399; see also Goldstone 1998; Kiser and Hechter 1991).
Here, we outline four behavioral tenets (table 13.1) that, based on a broad reading of literature in the social and behavioral sciences, we see as at the heart of future framework construction.
Table 13.1. Human behavior: basic principles
1. | The evolutionary legacy of our species has tendencies toward both dominance hierarchies and high degrees of sociality with great potentials to cooperate (e.g., Boehm 1993; Dubreuil 2010). |
2. | Agency is universal but also constrained by structure and resources (e.g., Schelling 2006; Sewell 2005). |
3. | Human groupings may be open and permeable to varying degrees, but they rarely are entirely closed for lengthy periods (e.g., M. L. Smith 2005; Wolf 1982). |
4. | Multiscalar perspectives are essential for understanding human groups, as humans generally participate simultaneously in networks of varying scales (e.g., Parkinson and Galaty 2009; Turchin 2003). |
Oddly, none of the five principal paradigms (table 13.2) that have been most influential in anthropological archaeology over the last 50 to 60 years comfortably conforms to all of these tenets, nor, in our opinion, have these extant paradigms found just the right balance between accounting for generalities and specifics.
Table 13.2. Theoretical frames on the preindustrial past
Theoretical frame | Agency | Scalar focus | Boundedness |
---|---|---|---|
Culture history | Elite? | Culture | Closed |
Cultural evolutionary systems | Elite | Society | Closed |
Marxism/Marx-influenced | Elite | Society, Class (rarely) | Potentially open |
Sociology (narrow Darwinian) | All | Individual, Kin | Not adequately considered |
Postprocessual | Elite (situational for commoners) | Society | Mostly closed |
The first three frameworks tend to give inadequate consideration to the “whys” and “hows” of human groups, often simply presuming their existence, continuity, and closure (e.g., Schortman and Urban 1992:12). Also, these three conceptual frameworks tend to afford little, if any, agency to the nonelite, or most people, in preindustrial contexts (see Brumfiel 1992; Wright 2009:122).
In contrast, recent postprocessual perspectives do speak of commoner resistance, yet it is largely conceptualized as an ad hoc phenomenon—for example, at times of already known societal breakdown or collapse (e.g., Joyce et al. 2001). Through this generally circumstantial or situational consideration of agency, the opportunity to understand the shift or diversity in the rules and practices of cooperation is largely lost. Strict sociobiological approaches do ascribe agency more broadly, particularly for small groups, but they are reliant on narrow definitions of self-interest and have not been able to account convincingly for large social formations and their diversity (Sterelny 2013). On their own, these approaches also fail to account adequately for the role of history and its contingencies or path-dependent aspects (Goldstone 1998:836). Clearly, we need analytical frames (Blanton et al. 1993; Little 2008:242; Parkinson and Galaty 2009) that not only take account of the multiple, scalable, and interlinked networks in which humans (past and present) participate but that also recognize that human groups and networks rarely have been entirely isolated or closed (e.g., Adams and Kasakoff 1975; M. L. Smith 2005).
Although it would be premature to try to outline an encompassing new frame here, there are, in our view, appropriate constructs in the broad transdisciplinary literature on collective action and cooperation that can serve as a foundation. Although much of this theorizing has roots outside anthropology and archaeology, such frames are compatible with the four tenets outlined earlier. These constructs are intended as an intermediate theoretical frame, in the sense of the sociologists Hedström and Swedberg (1996:281): “attention is called to an intermediary level of analysis in-between pure description and story-telling on the one hand, and universal social laws on the other.” But also as defined by Blanton and Fargher (2009:135): “a productive middle path between a ‘homo economicus’ perspective of methodological individualism on the one hand, and normative determinism, on the other.”
Recent philosophy of the social sciences also has pointed in this direction. For example, Mjøset (2009) opines that the social sciences require theories that can account for variation and allow for the contingency of history (see also Kiser and Pfaff 2010:573), calling for the decoupling of unilineality and uniformity from explanatory power. Similarly, Little (2000:89) advocates “explanations that . . . highlight both the structural factors that govern change and the multiple pathways that change can take.” Such approaches offer a means to account for “outcomes on the basis of both social context and individual action” so that they are inherently multiscalar (Hechter and Kanazawa 1997:208).
Decades ago in an effort to compare diachronically three ancient Mesoamerican civilizations, the senior author and colleagues (Blanton et al. 1993) advanced a conceptual frame that bears certain parallels to these approaches through a focus on parameters such as societal scale, integration, complexity, and boundedness and the variable relations between them. Frankly, this approach gained little traction because we did not fully spell out the expected theoretical links between these parameters (but see Kowalewski et al. 1983), nor did we give sufficient consideration to human agency, instead conceptualizing more at the societal or systemic scale. Fortunately, in the interim, advances associated with human collective action, cooperation, and social networks have provided a basis for constructing new frames that are specifically designed to account for the diversity of complex societies while also being underpinned by the four tenets outlined earlier.
Cooperation and Collective Action
The empirical relationship between group size and organization long has been recognized. As Dubreuil (2008:203) has observed: “many things are debated about the evolution of human societies, but no bands of 50 individuals have ever created a bureaucracy and no society comprising millions of individuals has remained perfectly egalitarian.” While we explicitly do not wish to imply that human cooperation in small groups is “natural,” ever-present, or easy, there is an empirical basis to suggest that consistent face-to-face interaction does seem to facilitate it (e.g., Apicella et al. 2012; Fowler and Christakis 2010). Yet in repeated synchronic, cross-cultural studies, intensely interactive groups or communities over several hundred people (Dunbar 2011:table 1) generally have some kind of suprahousehold integrative institutions, while sedentary populations numbering more than 2,000–3,000 are almost always hierarchically organized (Feinman 1995, 1998, 2011) (table 13.3).
Table 13.3. Organizational thresholds of human groups
Source | 150–200 | Source | 2,000–3,000 |
---|---|---|---|
Hill and Dunbar (2003) | 150 | Carneiro (1967) | 2,000 |
Forge (1972) | 150 | Forge (1972) | 2,000 |
Adler and Wilshusen (1990) | ~200 | Sinha (1978) | 2,000–3,000 |
Kosse (1990) | 2,000–3,000 | ||
Johnson (1982) | 2,400 | ||
Bernard and Killworth (1973) | 2,460 | ||
Brown and Podolefsky (1976) | 2,500 | ||
Lekson (1985) | 2,500 |
Various streams of research tie these new forms of cooperation either to mechanisms associated with the sanctioning of free riders (Baldassarri and Grossman 2011; Dubreuil 2010:166–170; O’Gorman et al. 2009; West et al. 2011) and/or human cognitive constraints (Dunbar 1993, 1998; Johnson 1982; Kosse 1994); the two are not mutually exclusive (Hooper et al. 2010).
And yet, despite the strength of the correlation between group size and hierarchical complexity (Johnson 1982:figure 21.1; see also Bodley 2003; Feinman 2011), there is considerable variation when one zooms down to more constrained group size ranges. No magic thresholds have been identified; most importantly, the kinds of social adjustments made by groups with increasing size are by no means uniform (Adler 1989; Adler and Wilshusen 1990; Feinman and Neitzel 1984).
For example, in Bandy’s comparative data, the eight regions where states ultimately emerged saw the rise of large villages within a millennium (see figure 13.2a), but the subsequent time lag to the rise of states was much more variable (see figure 13.2b). Also, there were 12 other regions in which large villages arose just as quickly but states did not ultimately develop in the examined time horizon. So one set of factors (including available resources and geographic conditions) may have pushed or pulled humans into large communities, but once there, the specific cooperative adjustments that the occupants of these large communities made played a role in whether and how rapidly (or not) those groups established new forms of more hierarchical arrangements or states. So path dependence and the organizational/institutional choices taken also matter.
In seven of the eight cases where states ultimately arose, multiple large communities were established within 1,000 years (the Valley of Oaxaca is the sole exception in that sample with a primate center). And yet, there also were 11 cases or regions where multiple large communities or peer polities were founded within a millennium, but states did not develop. Competition matters and conceivably even may be necessary, but it alone is not sufficient.
Cooperative arrangements entail social contracts, but the natures of human social contracts are variable. This observation is not new, and for years social scientists working from different approaches and disciplines have recognized that in human groups, even at comparable levels of vertical complexity, power is worn and wielded in distinct manners and that these differences show a strong relationship to the ways that the funds or bases of power are procured (table 13.4). How rulers amass their resources affects the nature of the social contract and leadership. In other words, by definition, leadership is relational, albeit in different ways (Ahlquist and Levi 2011:5).
Table 13.4. Variation in modes of finance and leadership
Autocratic | Collective | Reference |
---|---|---|
Finance-based big-man | Production-based big-man | Strathern (1969) |
Individualizing chiefdom | Group-oriented chiefdom | Renfrew (1974) |
Wealth finance | Staple finance | D’Altroy and Earle (1985) |
Predatory rule | Quasi-voluntary compliance | Levi (1988) |
Exclusionary/network | Corporate | Blanton et al. (1996) |
Extractive | Inclusive | Acemoglu and Robinson (2012) |
Each of the studies listed in table 13.4 contrasts different modes of leadership that parallel in similar ways the distinct manners in which funds of power were amassed. Building on rational choice approaches, the political scientist Margaret Levi (1988) outlined a provocative model that endeavors to account for these patterns theoretically, while also offering a means for addressing the micro-macro problem (e.g., Schelling 2006) that has plagued most archaeological approaches to state formation. Levi’s model was devised specifically to understand the development of states, albeit in more contemporary times. Nevertheless, recently, this model has been extended, expanded, and tested through application to a large comparative sample of preindustrial states (Blanton and Fargher 2008; see also Fargher, Heredia Espinoza, and Blanton 2011; Kiser and Cai 2003; Kiser and Linton 2002).
The architects of this model propose that more representative or collective forms of leadership will be found where those with power depend more directly on the local populace for their economic underpinnings, whereas exclusionary/autocratic rule is more apt to occur where leaders rely less on their immediate populace and acquire their funds of power from external sources, such as the monitoring of exchange routes, war booty, or the control of spot resources (figure 13.5). In the latter cases, leaders exact less from their immediate populace and so are freer to afford diminished representation and fewer public goods. In large human cooperative arrangements and institutions, the more rulers depend directly on their immediate sustaining population for their resource support, the more agency and voice that populace is likely to be able to assert and the more public goods are apt to be distributed (Blanton and Fargher 2008:figure 10.2) (figure 13.6).
Figure 13.5. Internal and external sources of funds. (Adapted from Blanton and Fargher 2008:254–255.)
Figure 13.6. Model of collective action. (Adapted from Blanton and Fargher 2008:figure 10.2.)
Revisiting the Peterson-Drennan sample (see figure 13.3), we see that the cases with the highest tax rates were calculated for relatively small polities in which the leaders were neither particularly powerful nor ostentatious/self-aggrandizing. Yet they also note that the higher tax costs were linked to the building of public works and common goods, such as defensive features and irrigation systems (Peterson and Drennan 2012:123–124). Consequently, in accord with the collective action framework, the high-tax case (Pueblo Grande, Hohokam) has both an astonishingly large irrigation system and rather collective forms of leadership/cooperation (see also Ross 2004).
Obviously, this model cannot provide anything close to a complete roadmap to understand this aspect of the diversity of complex societies or early states, since human institutions, such as states, are complicated, and history does matter. Nevertheless, it does present a conceptual and testable basis to understand certain key aspects of the variability in human cooperative institutions. And it links that relationship to the quantities and the means through which resources are acquired by governing authorities as well as the works and goods that are delivered. Here we stress that while such variables as primary productivity, available technologies, and, particularly, population size factor into the kinds and amount of resources these authorities could potentially accumulate, critical too are the specific ways in which (and from where) these acquisitions occurred. In other words, the nature of cooperative arrangements, institutions, and their configurations do matter, and in ways that are potentially explicable.
Addressing Empirical Questions and Conundrums
One measure of a conceptual frame is whether it can help untangle puzzles and conundrums that are left unexplained by extant paradigms. For example, if we briefly compare two prehispanic Mesoamerican state societies during the Classic period (ca. AD 200–900), Teotihuacan and the polities of the Classic Maya, we face a number of seeming paradoxes (see Feinman 2001; Feinman and Nicholas 2011 for fuller treatments of this topic). The urban center of Teotihuacan was much larger, with more monumental buildings, and was laid out impressively along a grid system. In contrast, even the largest Classic Maya cities were smaller, with less monumental structures, lower densities, and seemingly less structured urban layouts. And yet, Maya rulers wore their power more ostentatiously, they were buried with highly valued, elaborately crafted goods in special contexts, and they often lived in palaces that most archaeologists can agree as to their function. They also had a highly developed writing system that in large part relates the life histories and real and imagined exploits of named rulers.
In stark contrast, we know the names of no rulers at more monumental Teotihuacan. Their writing system was much more basic, leaving few evidences. Murals at the site depict important figures, but they usually are in groups, wearing masks, and present no details concerning personal history or genealogy (Cowgill 1997:152; Manzanilla 1999:111). Tellingly, if a possible Teotihuacano leader is named, it is through Maya epigraphic studies (Martin and Grube 2008), although there is debate whether this mention represents an individual, office, or military order. Archaeologists cannot agree if the rulers of the site lived in a palace (three completely different locales have been proposed) (Cowgill 1983; Flannery 1998; Sanders and Evans 2006), and researchers keep tunneling into the largest structures to see if they can find a rich ruler’s tomb but so far have come up empty.
From a unilinear scale-complexity perspective, one might think that the larger (greater urban scale and density), more monumental polity should have more aggrandizing rulers and more highly developed communication technologies, but that clearly was not the case. Alternatively, some archaeologists have traditionally ascribed these differences to culture or ethnicity; the Maya were this way and Teotihuacanos were another. Of course, that point is valid to a degree, but the later, Postclassic Maya had a different system of rule and cooperation (e.g., Chase and Chase 2006), they used writing differently, and we have not found the degree of aggrandizement evident for the earlier Classic Maya elite.
Clearly, agrarian tribute made up part of the fund of power in both regions, but Classic Maya rulers seem to have relied more than their central Mexican counterparts on external revenues from the exchange of crafted prestige goods and interpersonal networks/elite alliances, and possibly the direct control of spot resources (reservoirs) that they centrally constructed to store water at several major centers (Lucero et al. 2011). In contrast, reliant on internal revenues, the nonelite population at Teotihuacan may have had greater voice and been the recipients of more public works (wide thoroughfares, grid plan layout, large public spaces). Power was more broadly shared, and differentials of access and wealth were flatter. Classic Maya rulers were more flamboyant, and the disparities of wealth appear to have been more fully expressed.
As a working hypothesis, it seems plausible that most first-generation states, at least immediately following their foundations, were underpinned heavily by internal resources, namely the agrarian production of the local population, and hence, more similar in this respect to Teotihuacan. Once established, early state rulers had incentives to provision public goods, such as the construction of defensive features, the building of roads, and the coordination of disaster relief. Flight and various forms of resistance were consistent options for commoners (Van Vugt et al. 2003), and in-migration likely led to potential growth in extractable revenues. Governing authorities may have had few incentives to rule flamboyantly, despotically, or to pursue offensive wars (Kiser and Linton 2002), although defense and the sanction of free riders were often central to their duties as leaders (e.g., Feinman and Neitzel 1984:table 2.6). Of course, as the revenue streams sustaining these polities and the wider worlds in which they were situated always shifted, so could extant social contracts.
Alternatively, as argued for the lowland Maya and also the Mycenaean polities of ancient Greece (Parkinson and Galaty 2007), there are other contexts in which emergent states relied more heavily on external resources. Frequently, these states arose in peer polity networks, in which governing authorities monitored and derived a sizable portion of their wealth from trade and foreign networks as well as the spoils of war. In such instances, power was wielded in comparatively more authoritarian and dynastic ways. Early Egypt, where trade networks going north and south yielded key resources, also may conform to this pattern (Bard 2000; Köhler 2010:40). Fuller assessment of this hypothesis requires better understanding of the economic underpinnings and the revenue-generation practices of early states.
Another conundrum is the vexing variance in the previously discussed relationship between population size and hierarchical complexity. At smaller scales, relatively more collectively organized and tightly interdependent (Curry and Dunbar 2011; Johnson 1982) social groups seem to be able to forestall the formation of new tiers of hierarchical organization, as degrees of cooperation are high. But at larger scales, more collectively organized groups often require greater investments in infrastructure and bureaucracy to maintain and facilitate revenue generation and connectivity over space. Hence these more representative polities often are associated with larger populations, as noted at Teotihuacan and in a sample of historical cases studied by Blanton and Fargher (2008:appendix 3) (figure 13.7). Thus some of the variation in the size-complexity relationship may reflect variance in cooperative social contracts or modes of intergroup connectivity (figure 13.8).
Figure 13.7. Relationship between population size, hierarchical complexity, and social capital/collective action. (Drawn from data in Blanton and Fargher 2008:appendix 3.)
Figure 13.8. Relationship between population size and increasing complexity for collective and autocratic organization.
Concluding Thoughts
Roughly 12,000 years ago, humans embarked on new social arrangements, but these groupings took variable forms and the historical processes that followed were hardly uniform. Understanding the diversity of these social formations and the alternative historical paths that people took in different regions are fundamental elements to understanding why, when, and where complex societies emerged in certain settings but not others.
Through its explicit focus on diversity and its ability to account for change, the approach outlined here is suited to make greater sense of the complexities of early complex society development. That it also provides a “micro-foundation for macro-historical and comparative phenomena” (Levi 1988:8) adds to its utility, as the micro-macro problem (e.g., Schelling 2006) rarely has even been addressed by comparative approaches in archaeology.
Working from this frame serves to facilitate the building of theoretical bridges and networks of communication to other disciplines, including other theoretical approaches focused on cooperation (e.g., Kohler et al. 2012) and social networks (e.g., Curry and Dunbar 2011). If the theoretical frames employed to understand the emergence of complex societies were made more broadly compatible with approaches employed to explain the diversity of states over time and space, we suspect that analytical benefits would be realized. At present, our analytical sample and theoretical frames are often constrained by longstanding but poorly justified disciplinary divisions, which utilize distinct theoretical constructs (Gerring 2012:3–4).
Until recently, surprisingly few systematic analyses or conceptualizations have crossed the chasms between classical and anthropological archaeology, prehistory and history, or the recent and the deeper past. If there are solid reasons to isolate certain cases in time and space from broader analytical samples, and at times indeed there may be, then such focused investigations should be based on the precise questions being asked rather than on disciplinary inertia or ossified theoretical tenets and constructs.
As an example, since all larger human cooperative groups, across time and space, were embedded in even wider social networks, we are not convinced that it makes analytical sense to isolate at the outset primary or pristine states from all others (A. T. Smith 2003:83). The corpus of primary states is internally variable itself, and perhaps these examples ought be examined in broader comparative contexts to help define and understand that diversity (Parkinson and Galaty 2007:113).
A recent commentary in the New York Times (Gutting 2012) questioned the utility of the social sciences in regard to contemporary political practice and policy because, in the view of the author, research in these fields generally is not constructed to yield predictive results. Certainly, such perspectives are not our overriding concern, nor should we measure our advances simply by this yardstick. And yet, it is clearly unfortunate that our societal institutions and relevant publics often find it difficult to derive practical insights from our discussions drawn from the archives of human histories. In part, the obstacles reflect the manner in which we have presented our findings. If we may hazard an opinion, there may be benefits and insights if the collective “we” made more concerted efforts to frame our concepts and investigations in ways that could potentially facilitate and encourage reasoned practical considerations of the lessons derived from comparative examinations of the past. The frameworks outlined should better position us for just that, while neither ignoring path dependence nor the potential diversity of human relations and institutions.
Acknowledgments. An earlier version of this paper was presented at the Santa Fe Institute. We are grateful to Institute Director Jeremy Sabloff for that rewarding opportunity. Richard Blanton, David Carballo, Mark Golitko, and Steadman Upham offered astute and helpful comments on a draft of the paper. We also thank Verenice Heredia and Lane Fargher for organizing this volume and for inviting us to participate. Finally, we wish to express our great appreciation to Richard Blanton. Over the years, Rich has been an invaluable teacher, mentor, colleague, and inspiration for us, and we are thrilled to have this opportunity to celebrate both the impressive creativity and wide-ranging impact of his research.