An ancient stateless civilization: bronze age India and the state in history.

• Author: Thompson, Thomas J.
• Position: Harappa

Few sorts of "public goods," including basic "law and order," have not somewhere, sometime been privately produced. (1) Nonetheless, the idea persists that imposed systems of "legitimate" violence have been essential to the long-term functioning of all reasonably complex societies--that is, to their avoidance of seriously "suboptimal" production of critical public goods. (2) This idea is understandable because the only well-known case of a long-lived, nonprimitive, stateless society has been medieval Iceland, a society still lacking in cities (Friedman 1979; Byock 1988).

In this article, I show that in early antiquity a whole group of interacting urban societies almost certainly lacking the state existed for approximately seven hundred years; that merchants specializing in long-distance trade organized the production of the largest-scale public goods; and that an unusually early emergence of long-distance trade probably produced these societies. My analysis (albeit of a single civilization) suggests strongly that the extreme frequency of state organization in civilized societies has been, in a perfectly straightforward sense, an accidental feature of our world's development.

Harappan Archaeology and Its Interpretation

South Asia's first civilization, labeled "Harappa" by archaeologists (after the location of one of its important sites of excavation), flourished from the mid-third millennium to the very early second millennium bc e on the plains of the Indus and Ghaggar-Hakra rivers. A rapid desertification, especially by causing the disappearance of the Ghaggar-Hakra River, brought about its eventual collapse. (3) Harappan civilization was the product in particular of farmers and herders who spread out from the western margins of the plains in the late 3000s bc e, displacing very few, if any, of the earlier inhabitants. The people of this civilization used writing, at least for limited purposes (the Harappan writing system, available only in short inscriptions, is as yet undeciphered), made extraordinarily widespread use of metal tools (Shaffer 1982, 46-47), and inhabited a number of commercial cities that achieved considerable scale (the five largest had peak populations in the tens of thousands) and remarkable levels of urban amenity (virtually every house had a bath connected to a municipal drainage system). (4) The similar layouts and similar public buildings of Harappan cities strongly suggest that no one of them served as a capital. Contrary to what was believed for decades (see, for example, Piggott 1950, 151-71), neither their similar, highly regular layouts nor their many uniformities in construction practices need indicate that a great planning entity was at work: gridlike layouts were the norm even for small settlements of the preurban era, and by far the most impressive uniformity--Harappa's common system of dimensions for bricks--has been shown to have had its origins in a preurban diffusion of technically superior practice (Kenoyer 1998, 52, 57). Entirely distinct regional material cultures are identifiable in Harappan remains (Possehl 1998, 274-75). In light of all the foregoing considerations, it seems unlikely that the civilization had any overarching political unity, although a widely patronized ritual center may have existed (one site contains evidence of what was almost certainly a large ritual bathing complex).

Harappa's urban remains, subjected to numerous excavations since the 1920s (in particular at the two earliest identified sites), are unusual in the extreme in that they offer up not a single obvious palace or imposing temple, but only simple public halls; (5) not one massive tomb (no great mounds, no pyramids); and not even any large statuary. This set of absences, which seems to indicate a complete lack of great public cults (religious or political), has played an important role in leading two of the best-known figures in Harappan studies to view the Harappans as sophisticated but probably stateless peoples (Shaffer 1982; Possehl 1998, 280,287-90). (6)

Remarkable as well and equally worth taking into account for its institutional implications (if only because it is technically conceivable that entirely unavoidable inadequacies in excavation and analysis are the reason that no large palaces and temples have been confirmed [Kenoyer 1994, 76]) is the fact that in diverse ways the Harappan remains indicate that neither war nor threats of war played an important part in intercity relations.

First, the Harappans do not appear to have constructed any memorials to military campaigns; nowhere is any battle damage reflected in physical structures; (7) and human remains reveal no evidence of violent death (Kenoyer 1998, 15; Possehl 1998, 269-71).

Second, of greater significance (because direct evidence of ancient warfare is usually difficult to find and because portrayals of battle might have been made on perishable materials), the Harappans produced extraordinarily few kinds of specialized fighting weapons (Mate 1985, 80-81) and no defensive armor, although they possessed sophisticated metalworking techniques even in the early phase of their flourishing (Kenoyer 1998, 135). One of the major combat weapons they employed was the stone mace, a sort of weapon that everywhere in history quickly became militarily obsolete--it had already become so in contemporary Mesopotamia--with the production of helmets (O'Connell 1995, 118). (8)

Third, the walls that surrounded each sizeable section of a city on the Harappan floodplains were unaccompanied by moats, and in each case they were singular--that is, they established no system of concentric barriers. Charging fees for access seems to have been the sole consideration in the construction of their gates. The entranceways had no turns to make for ease of ambush, but the gates were just wide enough to accommodate an oxcart and were apparently accompanied by stations for the weighing of goods (Kenoyer 1998, 55-56). The only known case of defensive wall design at an urban location--that at the great coastal settlement at Dholavira, defended by a system of concentric walls--probably reflected not any concern for intercity military activity but rather the city's need to fend off pirates making sea-to-land raids. Inland settlements could not conceivably have required defense against highly mobile bandit groups, first because during Harappan times neither horse nor camel riding was as yet remotely common in the world, and second because there were likely never any horses in the area at all (see Witzel and Farmer 2000, debunking contrary claims) or any camels, at least until era's end (Kenoyer 1998, 40, 89, 167).

Of course, it is conceivable that Harappan military science, including logistics and planning, simply did not evolve over a period of seven hundred years to the point that setting a large city to siege was a practical option, but, if so, that fact in itself would be significant.

Thus, the evidence suggests that Harappans experienced few if any pressures toward the elaboration of improved fighting technology and that Harappa's cities had no need of military infrastructural investments beyond the walls defending one particular site against pirates. Because we can scarcely imagine a centuries-long interstate system seldom disturbed by serious warfare or even by its prospect (and so experiencing much less military development than its technology would have allowed), we have solid grounds for concluding that the Harappans did not know the state, precisely as the absence of any physically impressive "signatures" of power also leads us to infer.

Perhaps naturally (especially if one assumes that military activity was episodic and small scale), one of the Harappa scholars most associated with an emphasis on the civilization's comparatively pacific character (and responsible for the preceding analysis of inland city walls and gates), though fully prepared to envision some level of elite military competition across societies, has been inclined to the view that diverse "means of control" not including armed force were at work in the management of Harappan cities (1994, 77; 1997, 263).

Although, as I have indicated in the preceding discussion, some of the leading figures in Harappan studies have suggested that Harappan societies may have been organized on a nonstate basis, detailed speculation about this possibility has been lacking.

A Model of Harappan Public-Goods Production

By the very unusual apparent simplicity of public style and by what seems to have been the comparative insignificance of military affairs, the archaeological record suggests that Harappa did not know the state. Any intergroup violence was evidently so infrequent and of such low intensity that no one's fighting skills and resources ever rose within reach of the ability to commit full-scale urban plunder, and therefore intercity defenses were unneeded. In light of these facts and the inferences that I (along with others) have drawn from them, a critical challenge--especially given that proponents of the state-civilization model are prone to make a great deal of Harappa's urbanism (see, for example, Ratnagar 1991, 16-18, 23-49)--is to explain how the planning, building, and repair of Harappa's major city walls, streets, and larger urban drainage systems were accomplished (to which might be added the question of how the defense of the walls at Dholavira was organized). With regard to other public goods, it is reasonable to assume the following: (1) neighborhoods and artisans' sections, which had distinct boundaries and corporate identities (Kenoyer 1998, 44, 55, 81), had little difficulty in creating and maintaining local walls and streets; (2) marketplace and community actors managed public order; and (3) law was a decentralized thing. Regarding the third point, commercial disputes might have been settled by guilds, as they normally have been in many times and places (Benson...