10
Cognitive Codes and Collective Action at Mari and the Indus

Introduction

The cross-cultural and comparative study of early civilizations has a long history in studies of Old World states. The two most influential contributors are V. Gordon Childe (1934, 1951, 1964), a prehistorian, and Henri Frankfort (1956), a Near Eastern archaeologist and Egyptologist, both of whom focused on the exclusionary nature of early states. Largely based on studies of ancient Egypt and Mesopotamia, Childe considered the means by which rulers extracted food surpluses as an overriding factor in their centralized political economy. Frankfort, on the other hand, claimed that Mesopotamian and Egyptian ideologies were the principal bases for the dominance of its leaders.

While current research continues to be influenced by these early works, there has been mounting evidence that their interpretations of the Egyptian and Mesopotamian states need revision. Bruce Trigger’s (2003) comparative and cross-cultural studies of seven early states, selected for study based on strong archaeological evidence and written sources either in the form of local texts or ethnohistoric accounts, involved a close analysis of kingship among other factors. His studies demonstrated that the work of kings was often mediated by self-governing associations and councils composed of important lineages and other interest groups. Similar doubts about the centralization of power at the hands of leaders in a broad range of societies in the New and Old Worlds (Nichols and Charlton 1997) complement his research. In addition, Susan and Roderick McIntosh’s research at Jenne-jeno in the Middle Niger (AD 400–1100) has provided evidence for a self-organized community in which multiple authorities existed, none of which monopolized power (R. McIntosh 2005). In fact, a range of societies in the African examples provided by Susan McIntosh indicate that kings had relatively weak control (S. K. McIntosh 1999). In another study, Elizabeth Stone and Paul Zimansky used spatial distributions to demonstrate that southern Mesopotamian societies were “populous, entrepreneurial,” based on a dispersed pattern of settlement at Mashkan-shapir in which administrative, residential, and production areas were “both independent and connected to the public sector” (Stone and Zimansky 2004; Stone 2007:219).

The McIntoshes, Stone, and Zimansky drew on theoretical models introduced by archaeologists in their attempts to refine understanding of the complexity of leadership and power in early states. Carol Crumley introduced the concept of “heterarchy,” which she defined as “the relation of elements to one another when they are unranked or when they possess the potential for being ranked in a number of different ways” (Crumley 1995:3). Even in the most despotic states, local hierarchies and “nodes of social power” can be complementary or potentially conflictive (Brumfiel 1992; Brumfiel and Fox 1994; Janusek 2008:30). Heterarchy and hierarchy, therefore, can coexist (Crumley 1995; Feinman and Marcus 1998), as in the case of mediating bodies like the councils known in Mesopotamia (Jacobsen 1943). The self-organized landscape that the McIntoshes describe for the Middle Niger included settlements with different functions that were not hierarchically but heterarchically ordered in flexible power relations

Blanton’s dual-processual theory and subsequent publications have not been widely cited in Old World studies. In this essay, I review the ways in which the concepts of cognitive code and collective action have been employed in explaining the limitations to power in the interpretation of corporate strategies by Daniel Fleming (2004) in the case of the Mari and in my research on the Indus civilization (Wright 2014). The theory posits two forms of power, one that is exclusionary (network), in which leaders monopolize sources of power, and a corporate strategy, in which decisions are shared among groups within the society. According to Blanton, limitations to power are the result of a symbolic cognitive base in which symbols, such as ritual sanctification, limit the exercise of power (Blanton 1998a:152). The Mesopotamian citizens that held offices and acted as mediators in setting limits to exclusionary power are prime examples (Jacobsen 1943).

In their discussions of collective action, Blanton and Fargher (2008) have added new dimensions to the concept of cognitive codes that are of specific relevance to the present chapter. They suggest that cultural codes may be less “influential” in societies in which the dimensions of the world order are more secular (2008:293). This interpretation more closely follows Fleming’s research on the Mari and mine on the Indus civilization. As I demonstrate below, among the Mari and the Indus collective action in some segments of society was based on an ideology that was born out of a cognitive code embedded in longstanding social and political orders. This factor, and others to be discussed in the following, has important implications for basic questions regarding premodern and modern political processes and pathways to complexity in the Mari and Indus.

The Kingdom of Mari

The site of Tell Hariri (ancient Mari) is located in present-day Syria near the Middle Euphrates as the river bends west and north (see figure 10.1 for the location of Mari at the bend in the Euphrates). It was settled for the first time in the twenty-eighth century BCE (Fleming 2004, 6:n. 11). When it was a fully developed state, it extended over 100 ha and was rebuilt and destroyed on several occasions. By the mid-third millennium, it was contemporary with several polities, including Ebla, ancient Nagar (Tell Brak in the Khabur Basin) and Akkadian and Ur III states in southern Mesopotamia. The city was abandoned with the conquest of King Hammurabi at around 1761 BCE.

Figure 10.1. Archaeological sites in northern and southern Mesopotamia. (Drawn by Tom McClellan.)

The evidence for systems of rule described here are from texts that fall within a 13-year period during the reign of Zimri-Lim, a “tribal king” and “master of a major city center” at Mari (Fleming 2004:1, 2). What distinguished Zimri-Lim’s rulership was a longstanding system in which decision-making was based on negotiations among tribal groups and collective and exclusionary ruling bodies that acted interdependently (Fleming 2004:19). The sustainability of a longstanding agropastoral component in Mari’s political economy was pivotal to the state’s exclusionary and corporate structure.

At least by the mid-third millennium, the tell was an enclosed mound comprising palaces and temples that were residential and administrative. Positioned midway between southern Mesopotamia and polities to its north and east, the site’s location is considered by its current excavator, Jean-Claude Margueron (cited in Fleming’s [2004] text), to be

ideal for the control of commercial river traffic and the movement of goods among trading partners in several polities, especially the shipment of wood to southern Mesopotamia. Although its location on the Middle Euphrates offered an invaluable water source for agriculture and pastoralism, the valley was not sufficiently wide to sustain large-scale farming and major populations, and the nearby steppe was too arid for dry farming (Porter 2012). According to textual sources and archaeological evidence, the people at Mari engaged in a mixed economy that was carried out by sedentary and mobile populations that were linked by kinship ties. Mobile pastoralists were dispersed spatially, but their symbiosis with sedentary societies served an ideology in which, although physically apart, they were “conceptually together” (Porter 2012:13).

It is in this sense that Zimri-Lim’s rulership among his tribal group, the Sim’alites, differed from his predecessors and the traditional forms of rule carried out in other states in northern and southern Mesopotamian (e.g., at Ebla and by the Akkadian or Ur III kings, and of course Hammurabi) that are often cited as models for exclusionary power (though see my earlier discussion of the research of Stone and Zimansky).

The Mari State: Rethinking Pastoralism, Ethnography, and Prehistory

Before describing the details of Zimri-Lim’s reign, it is necessary to review some of the issues that have excluded pastoralism from anthropological views of state leadership. New evidence based on comparative studies has challenged evolutionary models that marginalized pastoralist societies. Evolutionary stage typologies consigned tribes to a position in between bands and chiefdoms (e.g., Fried 1967, 1975; Service 1975). In archaeology, pastoralists often were conceived as one step ahead of hunters and gatherers, thus falling out of evolutionary paradigms when plants were domesticated (but see Hole 1991, 2000; Zagarell 1989).

Agropastoralism was a latecomer in discussions of sedentary peoples since it often was associated with fully formed states. Unlike agriculturalists, pastoralists were thought to be untethered from the landscape and to be organized in egalitarian societies. These views of pastoralists are being turned around based on re-re-evaluations of ethnographic and prehistoric research, offering a perspective from many Old World contexts in which pastoralists are more closely aligned with agriculturalists and mixed economies (Frachetti 2008a, b; Hanks 2010; Linduff and Rubinson 2008; Honeychurch and Amartuvshin 2007; Kohl 2009; Porter 2000, 2012; Rogers 2012; Salzman 1972, 2002; Szuchman 2009) than previously acknowledged. In addition, Anne Porter’s evidence for social differences in burial contexts at Tell Banat challenges older views that pastoral organization precluded “differences in wealth that lead to social stratification” and separated them from class-based societies (Porter 2000, 2012). Finally, these studies demonstrate that pastoralism is not a unitary concept. Daniel Rogers suggests that anthropologists need to alter their perspectives by looking outward from centers (adopting “a decentralized vantage point”) and viewing mobility as a central concept (Rogers 2012:9).

These views set the stage for Fleming’s interpretation of Mari leadership under Zimri-Lim’s reign and its tribal confederation. Using texts that contain almost day-to-day accounts of decision-making, his book predates much of the recent analyses of tribe/state relations, with the exception of the works of Anne Porter. He draws on Blanton and his colleagues, whose researches are cited throughout his interpretation of the Mari texts. His description of the Mari state provides a detailed analysis of the political structure in the millennium that preceded the rise of Zimri-Lim as its king and contributed to state organization during his reign. I provide here a sketch of his findings and synthesis of a wealth of nuanced levels of organization existing during Zimri-Lim’s reign.

Table 10.1 outlines the structure of the state and some critical points regarding how it operated.

During the reign of Zimri-Lim political power was bifurcated so that the king was at the head of the core people and the groups dominated as outsiders. Both were mutually dependent sedentary and mobile populations, each of which had a separate social structure with different roles for their leaders. The Sim’alites, the core people, were tribally affiliated with Zimri-Lim. Although they consisted of more than five divisions (perhaps aligned with herding groups), these divisions had lost all political function in his relations with this tribal constituency. The king dealt with a broad-based body politic that functioned in a corporate mode in which he consulted with “elders” who were not attached to specific units. Whether they met at a hamlet, village, city, or specific center is inconsequential, as the main point is that it was a “collective expression” of the population acting as a “body of people” (Fleming 2004:170).

In distinction, the king’s rule of the three elements defined as “outsiders” was palace based. Although Yaminite populations extended far beyond the territories dominated by Zimri-Lim, he maintained supremacy over them mainly through their concentrated settlement in the district governed from Terqa. Each of the five tribes of the Yaminites had a king that ruled individual towns but followed a power structure in which there were no domains in which Zimri-Lim’s authority could be refused (Fleming 2004:166).

It is this “independence” of mind and collective “ideology” born out of the management of access to grazing as “whole communities” that was at the heart of a cognitive code practiced between the king and the Sim’alites. Although administrative correspondence regarding sheep makes references to shearing, there are few concerns for breeding and management. The texts also do not include references to grazing rights, restrictions on the use of land, or specifics of their location. The king did not own all land and pastoralists appear to have grazed and moved their flocks without restraint (Fleming 2004:167). It is this ideology of consensus building among the Sim’alites that embodies the corporate code, which Fleming believes provided the mutually beneficial counterpart to the state’s exclusionary structure in his relations with the Yaminites. These restrictions on Zimri-Lim’s exclusionary power make it a perfect example of the dual-processual model.

Ideology in this case was deeply embedded in the fabric of the social and political order that served as an obstacle to exclusionary power in Zimri-Lim’s relations with the Sim’alites. I turn now to a similar cognitive code and collective action in my discussion of the Indus civilization.

The Indus Civilization

The concepts of cognitive code and collective action are of equal interest in understanding structures of power in the Indus civilization. For the Indus, I concentrate on its community infrastructure as a framework from which to observe the nonexclusionary aspects of its political economy (see figure 10.2 for the location of the Indus civilization and sites referred to in this chapter). This example focuses almost exclusively on the civilization’s material culture. Unlike Mari, the Harappans did not have lengthy texts that recorded activities, much less the day-to-day activities of the sort in Zimri-Lim’s archive. Inscriptions are found on stamp seals and small “tablets” or etchings on pots in which only a few signs are recorded (Meadow and Kenoyer 2000). A signboard with 12 inscriptions was found at the entrance to one of its cities and may indicate fairly widespread literacy, but for the present, we can only speculate, since Indus inscriptions have not been deciphered.

Figure 10.2. Locations of Indus civilization sites.

In Collective Action in the Formation of Pre-Modern States Blanton and Fargher used macroregional clustering of 30 states to identify “shared aspects of the cultural and social histories of state formation” (2008:3). Of relevance to this discussion is a cognitive code for South Asia that they trace in a five-phase sequence that begins with the Indus civilization and ends with the Mughal Empire. Taking into account the region’s broad geographical scale and regional and cultural diversity, they focus on polities “that were based on Hindu and Buddhist political theory and culture” (Blanton and Fargher 2008:60). Using a published interpretation of the Indus civilization, they describe it as “corporately organized,” one that lacked “a ruler cult,” and that possessed “an egalitarian ethic” and “comparatively limited wealth differentiation” (Blanton and Fargher 2008:62). They rely on a widely cited paper by Daniel Miller (1985a), in which he outlined a long cycling of “egalitarian and more centralized forms” based on Vedic Hinduism and Buddhism and their “distinct theories of rulership,” as the cognitive code that was in place in early and subsequent phases in the history of South Asia that ended at around 300 BCE. They are uncertain whether the Indus represented an “initiating phase” of “collective orientation” similar to what is known from the later South Asian states and a second urbanism in South Asia (Blanton and Fargher 2008). I find this interpretation problematic in view of the limited archaeological evidence with which to establish continuity between the Indus civilization and phases one and two (Wright 2010:325, chapter 11). While I am in agreement that the Indus political economy employed a corporate strategy and collective action, the cycling model that links the Indus civilization to Hindu Buddhist ideologies proposed by D. Miller (1985a) cannot be supported by the existing evidence, and I offer a different perspective.

In chapter 12, “Collective Action and Political Evolution,” in Collective Action in the Formation of Pre-Modern States, Blanton and Fargher (2008) point to selected features (apart from the phased recycling referred to in the preceding paragraph) of the Indus that argue for collective action. They include its planned cities, population numbers that reached 50,000 and “vast communal grain storage facilities” (Wheeler 1968; Blanton and Fargher 2008:291). Although there are other features that argue for collective action, such as the extensive public amenities at Indus centers and at some rural settlements (Wright 2010), my focus here is on collective action among craft producers and merchants.

The Indus Civilization: Geography, Urbanism, and Ecologies

Working toward that end, I provide a brief background to the varied geography of the Indus in order to update important misconceptions about this civilization. The discovery of the urban centers of Harappa and Mohenjo-daro in Pakistan in the 1930s has left a lasting impression that settlements were located solely on the rivers of the Indus valley. The subsequent discovery of large numbers of Indus settlements in northwest India as well as on the Ghaggar-Hakra River in India and Pakistan, and two major centers at Rakhighari and Ganweriwala, have received less attention (figure 10.2). The Ghaggar flowed into Pakistan from India, where its name changes to the Hakra, a river that flowed into the lower Indus and as far as the Arabian Sea. These rivers along with the Indus placed the city of Mohenjo-daro in an ideal location between two hydrographic systems. A fifth center at Dholavira is in Gujarat, not on a river at all and may have been a port city.

Not surprisingly, the ecology is varied when the totality of the civilization is considered, so we should expect significant differences in organization. Studies at Harappa (Weber 2003) have identified an agropastoral, double-cropping system, while in the south in Gujarat, pastoral practices and cultivation of crops differed. There, proximity to the sea and port locations opened to a wider world and intercultural trade with contemporary societies throughout the Greater Near East (Wright 2010:215ff.).

Finally, in many early accounts the civilization was described as homogeneous, but recent research has revealed regional differences in city plans, settlement patterns, and the pace of urbanization (Wright 2010:126ff., 81ff.).

Indus Heterarchy/Hierarchy, Cognitive Models, and Collective Action

V. Gordon Childe’s interpretations on the early Indus state were based on intellectual trends of the day and were not so different from what Blanton argues against. As discussed earlier, Childe paid little regard for the economic and social infrastructure of early states, such as Stone and Zimansky’s (2004) documentation of spatial arrangements reflective of independent specialists in Mesopotamia. In distinction, Childe assigned control of production and distribution to centralized leaders and overlooked restrictions on exclusionary power in the form of independent specialists and collective action.

In the Indus, recent research on craft production and its specialized technologies and spatial distributions suggests that the variability observed by Stone and Zimansky followed similar patterns (Wright 2010:145ff.). I draw on the production of ceramics and seals, their uses and distribution. During the peak of urbanism, Indus production was intensified (new products, large numbers of producers and output); diversified (products, new skills, and uses), and specialized (divisions of labor and distribution of resources), and objects were produced for internal and external distribution. Artisans worked in large and small-scale workshops. Some were independent of the state while others may have been state controlled. They produced mundane and prestige goods from locally available materials and from distant resources.

Extensive studies of craft producers have been undertaken at Harappa, one of the main Indus centers (Wright 2010:chapters 6 and 7). The production of ceramic vessels and stoneware bangles required related skills in which clay was processed, shaped, and fired but in unique ways. Locally available clays were used for both, but they were refined and standardized differently, according to the desired end product. The production of vessels required substantial trimming, mixing of pigments for paints and slips, and firing in two-chambered kilns (Wright 1989, 2010:152–166), while stoneware bangle production involved several steps in reshaping by grooving or cutting forms shaped on the potter’s wheel. In a final step, they were fired in airtight containers to provide optimal control of atmospheric conditions and to attain higher temperatures than in vessel production (Vidale 2000:92). Many bangles bore Indus signs as if produced for individuals, or were possibly inscribed with a name.

The independent, noncentralized organization of vessel production is based on contextual data from the excavations at Harappa. A ceramic workshop in the city was associated with a mudbrick wall believed to be part of a residential building. During the urban period, potters built a large two-chambered kiln replacing a small pit kiln that had been built in a previous period. We interpreted the presence of the two kilns as evidence for an early pre-urban era of small-scale production and the later intensification of production in the urban period (Wright 1991, 2010:187). This time capsule of social change demonstrated continuity among specialist producers in which a craft was handed down over several generations. Although there were changes in the production process over several hundred years of producing ceramic objects, their decisions about timing, technologies, and social arrangements were organized without any apparent control from groups outside of their production units.

In early ethnographies, ceramic production was often described as “a labor intensive and time-consuming” craft in which production was limited (Foster 1995:100) and economic yields were low. Because ceramics were thought to be produced by a single individual, the “potter,” their organization had little to offer in shaping society, an assumption partially based on misplaced census reports in which a single entry designated the name of the “potter” (Miller 1985b). In fact, as subsequent research has demonstrated, pottery production more often requires a cooperative network of craft workers who perform various tasks that cannot be accomplished by a single producer and that are essential to the final product. They work together in tightly organized groups in which decision-making requires consensus building and collective action to succeed (Coburn 2011; Kramer 1997; Wright 1991).

The organization of the stoneware bangle industry differed. Situated outside of a residential area within a walled enclosure at the city of Mohenjo-daro, it fits the exclusionary model often associated with specialized forms of ceramic production that are highly standardized and controlled. Perhaps because stoneware bangle production was an innovation associated with the urban period, it has been interpreted as an “industrial craft under administrative control” (Vidale 1989:178). According to a neutron activation analysis of a sample of bangles from several sites, their production appears to have been limited to Mohenjo-daro and Harappa and controlled by a restricted group (Blackman and Vidale 1992).

My second example, the production and use of seals, has a long history in the greater Near East and South Asia. Conceptually, seals and sealings are related to what Dennis Frenez and Maurizio Tosi refer to as a “Transcultural Administrative Sealing System” in which objects were sealed for storage purposes and safe-keeping, a practice known from the sixth to fifth millennia BCE (Frenez and Tosi 2005). By the third millennium, the shapes of seals signified their culture of origin; for example circular seals were from groups in Arabia, while Indus seals were square. The latter were produced from steatite and engraved with large mammals, small stands, and Indus script (Franke-Vogt 1989). At Indus sites, seals are present in great numbers and widely distributed at small and large settlements, in households, public buildings, and workshops.

Based on contextual data—stylistic and technological studies—Indus seals are associated with trade and merchants (Frenez and Tosi 2005). At the small site of Chanhu-daro, a seal workshop was separated from other craft production. Among the production debris associated with the seals, there were net weights, a standard of measurement associated with exchange systems and merchant activities (Mackay 1943). In another study, an analysis of the stylistic elements of seal iconography, Paul Rissman (1988) identified features that were sufficiently distinctive that he was able to identify the work of regional artisans. He referred to them as “schools” of producers who were organized in guild-like structures, possibly within the context of a family group (Rissman 1988), a finding that complements identifiable stylistic variations found in other studies (Vidale 2005). Close analyses of technological features have identified Indus production techniques that vary regionally (Green 2016) and interregionally (Pittman 2013).

Textual evidence for contemporary merchants who traveled abroad and are known from Mesopotamian texts complement these interpretations. In Mesopotamia, merchant and artisan groups were organized into professional communities, often linked through family ties (Garfinkle 2002), many of which were independent of the state. Merchants are known to have traveled to Mesopotamia, as attested by the words of an Akkadian king (ca. 2350 BCE) who boasted of having seafaring traders from Meluhha (interpreted as the Indus civilization) in his harbor. In a subsequent period (Ur III, ca. 2000 BCE), other texts record a “village of traveling merchants” from Meluhha in the village of Guabba in Mesopotamia (Vermaak 2008). These Meluhhans had become acculturated, as evidenced by their Sumerian names (Parpola et al. 1977:145), suggesting that generations of Meluhhans had traveled to Mesopotamia and became acculturated over a period of several hundred years.

Our best examples of regional variation in Indus seal iconography and technology and traveling merchants unattached to a central authority come from a recent study by Steffen Laursen (2010). His research is based on the discovery of round seals on the Arabian Peninsula in Bahrain. Although round seals are common on the Arabian Peninsula, Laursen believes that the technology of seal production was spread by entrepreneurs from the Indus who transmitted a technological package comprising sealing, writing, and weight technologies. Sometime around 2100 BCE these “breakaway” entrepreneurs established a hybrid form of the technology in Bahrain (Laursen 2010) and by 2050 BCE they were acculturated, much like the traveling merchants in Guabba. Based on the analyses of the sequence order of Indus signs (Parpola 1994), Laursen notes that among the 28 seals he researched, only three have sequences that parallel those found on Indus inscriptions. Others included signs with the double image of an inscription, which is not true to the normal order of inscriptions produced in the Indus. Further, in another group of seals from Failaka, Mesopotamia, and Iran, inscribed twin images and improperly ordered script have also been discovered. Laursen describes the signs as “pseudo-inscriptions” designed to convey a strong message of autonomy (Laursen 2010:36).

These “errors” are typical of altered forms of writing when contact occurs over long periods of time and a language becomes pigeonized. They are clear signs of the development of new forms that are not true to the original languages but that are understood by persons engaged in transactions from different language groups. Possibly the seals belonged to a merchant of Indus origin who lived in Mesopotamia or the Gulf and who either adopted a local language or refashioned his own into a version of the original, signaling a difference from the Indus script that is consistent with the examples of collective action carried out by the merchants who produced and used the seals.

Conclusions

In a recent work, Blanton looks at the marketplace and argues against earlier antimarket thinking in substantivist debates and the idea that the commercialized West was “unique in human experience” (Blanton 2013:23). His views on marketplaces, which he discusses in great detail, are closely aligned with the interpretations offered here with respect to collective action among pastoralists, craft producers, and merchants at Mari and the Indus. Like marketplaces, pastoralists, craft producers, and merchants carved out social spaces that shaped institutions, social relationships, and domains, establishing ties of sufficient political and economic import that they limited exclusionary rule. The evidence from the Indus and Mari provides a compelling example of the efficacy of the corporate model and offers a structure for understanding collective action. Both in Mari and the Indus they were especially useful tools with which to model strategies among pastoralists, where a strong, materially based ideology, drawn from textual sources, excluded centralization among the Sima’lites. In the Indus, the examples of craft production and merchant activities, when complemented by textual and modern examples of producers and merchants, demonstrated the presence of whole communities (guilds, occupational specializations) and collective action that predates modern states.

Acknowledgments. My thanks to Lane Fargher and Verenice Heredia for inviting me to participate in this volume and in the earlier session at the SAA. I especially wish to thank Tom McClellan for providing me with the Mari map (figure 10.1) and to Dan Fleming, my colleague at NYU, for countless talks about the Mari evidence. I met Richard Blanton for the first time too many years ago to remember when he appeared at the site of Harappa in the Punjab, Pakistan, during our field season. Rich was on a major trek through the Old World, nourishing his interests in comparative studies, stopping at important sites along the way. It was a treat to have him there and to share his observations. Later, I asked him and his colleagues to contribute to a series I edit on Case Studies in Early Societies (Cambridge University Press). The result was Ancient Oaxaca (Blanton et al. 1999). Rich’s productivity has been an inspiration and his advancement of theory influential in my own efforts to provide a comparative perspective on the Near East and South Asia. This chapter was revised in response to peer reviewers, to whom I am grateful. All errors are mine alone.