Landau, Amer. J. Public Health 54, 85 (1964). tion on property values. See R. J. Anderson, 69. P.
, 69. P. Stocks, "British Empire Cancer Campaign,"
If one accepted this evidence as conclusive, Jr., and T. D. Crocker, "Air Pollution and supplement to "Cancer in North Wales and it would follow that the annual cost of air residential property values," paper presented Liverpool Region," part 2 (Summerfield and pollution, because of health effects, would run at a meeting of the Econometric Society, New Day, London, 1957). between $14 billion and $29 billion. York, December 1969; H. 0. Nourse, Land 70. G. Dean, Brit. Med. J. 1, 1506 (1966). 65. See J. H. Schulte, Arch. Environ. Health 7, Econ. 43, 181 (1967); R. G. Ridker, Eco- 71. A. H. Golledge and A. J. Wicken, Med. 524 (1963); A. G. Cooper, "Carbon Mon- nomic Costs of Air Pollution (Praeger, New Officer 112, 273 (1964). oxide," U.S. Public Health Serv. Publ. No. York, 1967); R. G. Ridker and J. A. Hen- 72. W. Haenszel, D. B. Loveland, M. G. Sirken, 1503 (1966); Effects of Chronic Exposure to ning, Rev. Econ. Statist. 49, 246 (1967); R. J. Nat. Cancer Inst. 28, 947 (1962). Low Levels of Carbon Monoxide on Human N. S. Harris, G. S. Tolley, C. Harrell, ibid. 73. The research discussed in this article was Health, Behavior, and Performance (National 50, 241 (1968). supported by a grant from Resources for the Academy of Sciences and National Academy 67. P. Buell, J. E. Dunn, Jr., L. Breslow, Cancer Future, Inc. We thank Morton Corn, Allen of Engineering, Washington, D.C., 1969). 20, 2139 (1967). Kneese, and John Goldsmith for helpful 66. Another way to estimate the cost of air pollu- 68. E. C. Hammond and D. Horn, J. Amer. Med. comments. Any opinions and remaining errors tion is to examine the effect of air pollu- Ass. 166, 1294 (1958). are ours.
It was not the product of "genius" or
the result of chance, but the outcome of a regular and determinate cultural process. Moreover, it was not a unique event but a recurring phenomenon: states arose independently in different places and at different times. Where A Theory of the Origin the appropriate conditions existed, the state emerged. of the State Voluntaristic Theories Traditional theories of state origins are considered Serious theories of state origins are and rejected in favor of a new ecological hypothesis. of two general types: voluntaristic and coercive. Voluntaristic theories hold that, at some point in their history, Robert L. Carneiro certain peoples spontaneously, ration- ally, and voluntarily gave up their in- dividual sovereignties and united with other communities to form a larger For the first 2 million years of his it seems desirable to discuss, if only political unit deserving to be called a existence, man lived in bands or vil- briefly, a few of the traditional theories. state. Of such theories the best known lages which, as far as we can tell, Explicit theories of the origin of the is the old Social Contract theory, which were completely autonomous. Not until state are relatively modern. Classical was associated especially with the name perhaps 5000 B.C. did villages begin writers like Aristotle, unfamiliar with of Rousseau. We now know that no to aggregate into larger political units. other forms of political organization, such compact was ever subscribed t6 But, once this process of aggregation tended to think of the state as "nat- by human groups, and the Social Con- began, it continued at a progressively ural," and therefore as not requiring tract theory is today nothing more faster pace and led, around 4000 B.C., an explanation. However, the age of than a historical curiosity. to the formation of the first state in exploration, by making Europeans The most widely accepted of modern history. (When I speak of a state I aware that many peoples throughout voluntaristic theories is the one I call mean an autonomous political unit, the world lived, not in states, but in the "automatic" theory. According to encompassing many communities with- independent villages or tribes, made this theory, the invention of agriculture in its territory and having a centralized the state seem less natural, and thus automatically brought into being a sur- government with the power to collect more in need of explanation. plus of food, enabling some individuals taxes, draft men for work or war, and Of the many modern theories of state to divorce themselves from food pro- decree and enforce laws.) origins that have been proposed, we duction and to become potters, weav- Although it was by all odds the most can consider only a few. Those with ers, smiths, masons, and so on, thus far-reaching political development in a racial basis, for example, are now creating an extensive division of labor. human history, the origin of the state so thoroughly discredited that they Out of this occupational specialization is still very imperfectly understood. In- need not be dealt with here. We can there developed a political integration deed, not one of the current theories also reject the belief that the state is which united a number of previously of the rise of the state is entirely satis- an expression of the "genius" of a independent communities into a state. factory. At one point or another, all of people (2), or that it arose through This argument was set forth most fre- them fail. There is one theory, though, a "historical accident." Such notions quently by the late British archeologist which I believe does provide a con- make the state appear to be something V. Gordon Childe (3). vincing explanation of how states began. metaphysical or adventitious, and thus It is a theory which I proposed once place it beyond scientific understanding. The author is curator of South American before (1), and which I present here In my opinion, the origin of the state ethnology in the department of anthropology at the American Museum of Natural History, more fully. Before doing so, however, was neither mysterious nor fortuitous. New York, New York. 21 AUGUST 1970 733 The principal difficulty with this Coercive Theories -cxception to this rule? Might there not theory is that agriculture does not au- be,. somewhere in the world, an ex- tomatically create a food surplus. We A close examination of history indi- ample of a state which arose without know this because many agricultural cates that only a coercive theory can the agency of war? peoples of the world produce no such account for the rise of the state. Force, Until a Iew years ago, anthropolo- surplus. Virtually all Amazonian In- and not enlightened self-interest, is the ~gist&genrally believed that the Classic dians, for example, were agricultural, mechanism by which political evolution Maya provide such an instance. The but in aboriginal times they did not has led, step by step, from autonomous archeological evidence then available produce a food surplus. That it was villages to the state. gave no hint of warfare among the technically feasible for them to pro- The view that war lies at the root early Maya and led scholars to regard duce such a surplus is shown by the of the state is by no means new. Twenty- them as a peace-loving theocratic state ifact that, under the stimulus of Euro- five hundred years ago Heraclitus wrote which had arisen entirely without war pean settlers' desire for food, a number that "war is the father of all things." (14). However, this view is no longer of tribes did raise manioc in amounts The first careful study of the role of tenable. Recent archeological discov- well above their own needs, for the warfare in the rise of the state, how- eries have placed the Classic Maya in purpose of trading (4). Thus the tech- ever, was made less than a hundred a very different light. First came the nical means for generating a food sur- years ago, by Herbert Spencer in his discovery of the Bonampak murals, plus were there; it was the social mech- Principles of Sociology (8). Perhaps showing the early Maya at war and anisms needed to actualize it that were better known than Spencer's writings reveling in the torture of war captives. lacking. on war and the state are the conquest Then, excavations around Tikal re- Another current voluntaristic theory theories of continental writers such as vealed large earthworks partly sur- of state origins is Karl Wittfogel's "hy- Ludwig Gumplowicz (9), Gustav Rat- rounding that Classic Maya city, point- draulic hypothesis." As I understand zenhofer (10), and Franz Oppenheim- ing clearly to a military rivalry with him, Wittfogel sees the state arising er (11). the neighboring city of Uaxactun (15). in the following way. In certain arid Oppenheimer, for example, argued Summarizing presmnt thinking on the and semiarid areas of the world, where that the state emerged when the pro- subject, Michael D. Coe has observed village farmers had to struggle to sup- ductive capacity of settled agriculturists that "the ancient Maya were just as port themselves by means of small- was combined with the energy of pas- warlike as the . . . bloodthirsty states scale irrigation, a time arrived when toral nomads through the conquest of of the Post-Classic" (16). they saw that it would be to the ad- the former by the latter (11, pp. 51- Yet, though warfare is surely a prime vantage of all concerned to set aside 55). This theory, however, has two mover in the origin of the state, it can- their individual autonomies and merge serious defects. First, it fails to account not be the only factor. After all, wars their villages into a single large po- for the rise of states in aboriginal have been fought in many parts of the litical unit capable of carrying out irri- America, where pastoral nomadism was world where the state never emerged. gation on a broad scale. The body of unknown. Second, it is now well estab- Thus, while warfare may be a neces- officials they created to devise and ad- lished that pastoral nomadism did not sary condition for the rise of the state, minister such extensive irrigation works arise in the Old World until after the it is not a sufficient one. Or, to put it brought the state into being (5). earliest states had emerged. another way, while we can identify This theory has recently run into Regardless of deficiencies in par- war as the mechanism of state forma- difficulties. Archeological evidence now ticular coercive theories, however, there tion, we need also to specify the con- makes it appear that in at least three is little question that, in one way or ditions under which it gave rise to the of the areas that Wittfogel cites as ex- another, war played a decisive role in state. emplifying his "hydraulic hypothesis"- the rise of the state. Historical or arche- Mesopotamia, China, and Mexico- ological evidence of war is found in full-fledged states developed well before the early stages of state formation in Environmental Circumscription large-scale irrigation (6). Thus, irriga- Mesopotamia, Egypt, India, China, tion did not play the causal role in Japan, Greece, Rome, northern Eu- How are we to determine these con- the rise of the state that Wittfogel rope, central Africa, Polynesia, Middle ditions? One promising approach is to appears to attribute to it (7). America, Peru, and Colombia, to name look for those factors common to areas This and all other voluntaristic the- only the most prominent examples. of the world in which states arose in- ories of the rise of the state founder Thus, with the Germanic kingdoms digenously-areas such as the Nile, on the same rock: the demonstrated of northern Europe especially in mind, Tigris-Euphrates, and Indus valleys in inability of autonomous political units Edward Jenks observed that, "histori- the Old World and the Valley of Mex- to relinquish their sovereignty in the cally speaking, there is not the slightest ico and the mountain and coastal val- absence of overriding external con- difficulty in proving that all political leys of Peru in the New. These areas straints. We see this inability mani- communities of the modern type [that differ from one another in many ways fested again and again by political units is, states] owe their existence to suc- -in altitude, temperature, rainfall, soil ranging from tiny villages to great em- cessful warfare" (12). And in reading type, drainage pattern, and many other pires. Indeed, one can scan the pages Jan Vansina's Kingdoms of the Sa- features. They do, however, have one of history without finding a single genu- vanna (13), a book with no theoretical thing in common: they are all areas of ine exception to this rule. Thus, in ax to grind, one finds that state after circumscribed agricultural land. Each order to account for the origin of the state in central Africa arose in the of them is set off by mountains, seas, state we must set aside voluntaristic same manner. or deserts, and these environmental fea- theories and look elsewhere. But is it really true that there is no tures sharply delimit the area that simple 734 SCIENCE, VOL. 169 farming peoples could occupy and cul- since virtually any area of forest is land was created failed to keep pace tivate. In this respect these areas are suitable for cultivation, subsistence ag- with the increasing demand for it. very different from, say, the Amazon riculture could be carried on in the new Even before the land shortage became basin or the eastern woodlands of habitat just about as well as in the old. so acute that irrigation began to be North America, where extensive and It was apparently by this process of practiced systematically, villages were unbroken forests provided almost un- fight and flight that horticultural tribes undoubtedly already fighting one an- limited agricultural land. gradually spread out until they came other over land. Prior to this time, But what is the significance of cir- to cover, thinly but extensively, almost when agricultural villages were still cumscribed agricultural land for the the entire Amazon basin. Thus, under few in number and well supplied with origin of the state? Its significance can the conditions of unlimited agricultural land, the warfare waged in the coastal best be understood by comparing po- land and low population density that valleys of Peru had probably been of litical development in two regions of prevailed in Amazonia, the effect of much the same type as that described the world having contrasting ecologies warfare was to disperse villages over a above for Amazonia. With increasing -one a region with circumscribed ag- wide area, and to keep them autono- pressure of human population on the ricultural land and the other a region mous. With only a very few exceptions, land, however, the major incentive for where there was extensive and un- noted below, there was no tendency in war changed from a desire for revenge limited land. The two areas I have Amazonia for villages to be held in to a need to acquire land. And, as the chosen to use in making this oom- place and to combine into larger po- causes of war became predominantly parison are the coastal valleys of Peru litical units. economic, the frequency, intensity, and and the Amazon basin. In marked contrast to the situation importance of war increased. Our examination begins at the stage in Amazonia were the events that tran- Once this stage was reached, a Pe- where agricultural communities were spired in the narrow valleys of the Pe- ruvian village that lost a war faced con- already present but where each was ruvian coast. The reconstruction of sequences very different from those still completely autonomous. Looking these events that I present is admit- faced by a defeated village in Ama- first at the Amazon basin, we see that tedly inferential, but I think it is con- zonia. There, as we have seen, the van- agricultural villages there were nu- sistent with the archeological evidence. quished could flee to a new locale, sub- merous, but widely dispersed. Even in Here too our account begins at the sisting there about as well as they had areas with relatively dense clustering, stage of small, dispersed, and autono- subsisted before, and retaining their like the Upper Xingu' basin, villages mous farming communities. However, independence. In Peru, however, this were at least 10 or 15 miles apart. instead of being scattered over a vast alternative was no longer open to the Thus, the typical Amazonian com- expanse of rain forest as they were in inhabitants of defeated villages. The munity, even though it practiced a Amazonia, villages here were confined mountains, the desert, and the sea-to simple form of shifting cultivation to some 78 short and narrow valleys say nothing of neighboring villages- which required extensive amounts of (18). Each of these valleys, moreover, blocked escape in every direction. A land, still had around it all the forest was backed by the mountains, fronted village defeated in war thus faced only land needed for its gardens (17). For by the sea, and flanked on either side grim prospects. If it was allowed to Amazonia as a whole, then, population by desert as dry as any in the world. remain on its own land, instead of being density was low and subsistence pres- Nowhere else, perhaps, can one find exterminated or expelled, this conces- sure on the land was slight. agricultural valleys more sharply cir- sion came only at a price. And the Warfare was certainly frequent in cumscribed than these. price was political subordination to the Amazonia, but it was waged for rea- As with neolithic communities gen- victor. This subordination generally en- sons of revenge, the taking of women, erally, villages of the Peruvian coastal tailed at least the payment of a tribute the gaining of personal prestige, and valleys -tended to grow in size. Since or tax in kind, which the defeated vil- motives of a similar sort. There being autonomous villages are likely to fis- lage could provide only by producing no shortage of land, there was, by and sion as they grow, as long as land is more food than it had produced before. large, no warfare over land. available for the settlement of splinter But subordination sometimes involved The consequences of the type of war- communities, these villages undoubtedly a further loss of autonomy on the part fare that did occur in Amazonia were split from time to time (19). Thus, vil- of the defeated village-namely, in- as follows. A defeated group was not, lages tended to increase in number corporation into the political unit domi- as a rule, driven from its land. Nor faster than they grew in size. This nated by the victor. did the victor make any real effort to increase in the number of villages oc- Through the recurrence of warfare subject the vanquished, or to exact cupying a valley probably continued, of this type, we see arising in coastal tribute from him. This would have been without giving rise to significant changes Peru integrated territorial units tran- difficult to accomplish in any case, since in subsistence practices, until all the scending the village in size and in de- there was no effective way to prevent readily arable land in the valley was gree of organization. Political evolution the losers from fleeing to a distant part being farmed. was attaining the level of the chiefdom. of the forest. Indeed, defeated villages At this point two changes in agri- As land shortages continued and be- often chose to do just this, not so much cultural techniques began to occur: the came even more acute, so did warfare. to avoid subjugation as to avoid further tilling of land already under cultiva- Now, however, the competing units attack. With settlement so sparse in tion was intensified, and new, previ- were no longer small villages but, often, Amazonia, a new area of forest could ously unusable land was brought under large chiefdoms. From this point on, be found and occupied with relative cultivation by means of terracing and through the conquest of chiefdom by ease, and without trespassing on the irrigation (20). chiefdom, the size of political units in- territory of another village. Moreover, Yet the rate at which new arable creased at a progressively faster rate. 21 AUGUST 1970 735 Naturally, as autonomous political units they have made the fullest possible Resource Concentration increased in size, they decreased in use of their own. This implies that number, with the result that an entire every autonomous village has an un- Theories are first formulated on the valley was eventually unified under the tapped margin of food productivity, basis of a limited number of facts. banner of its strongest chiefdom. The and that this margin is squeezed out Eventually, though, a theory must con- political unit thus formed was un- only when the village is subjugated and front all of the facts. And often new doubtedly sufficiently centralized and compelled to pay taxes in kind. The facts are stubborn and do not conform complex to warrant being called a state. surplus food extracted from conquered to the theory, or do not conform very The political evolution I have de- villages through taxation, which in the well. What distinguishes a successful scribed for one valley of Peru was also aggregate attained very significant pro- theory from an unsuccessful one is that taking place in other valleys, in the portions, went largely to support the it can be modified or elaborated to ac- highlands as well as on the coast (21). ruler, his warriors and retainers, offi- commodate the entire range of facts. Once valley-wide kingdoms emerged, cials, priests, and other members of the Let us see how well the "circumscrip- the next step was the formation of rising upper class, who thus became tion theory" holds up when it is brought multivalley kingdoms through the con- completely divorced from food pro- face-to-face with certain facts that ap- quest of weaker valleys by stronger duction. pear to be exceptions. ones. The culmination of this process Finally, those made landless by war For the first test let us return to was the conquest (22) of all of Peru but not enslaved tended to gravitate to Amazonia. Early voyagers down the by its most powerful state, and the for- settlements which, because of their spe- Amazon left written testimony of a mation of a single great empire. Al- cialized administrative, oommercial, or culture along that river higher than though this step may have occurred religious functions, were growing into the culture I have described for Ama- once or twice before in Andean history, towns and cities. Here they were able zonia generally. In the 1500's, the it was achieved most notably, and for to make a living as workers and arti- native population living on the banks the last time, by the Incas (23). sans, exchanging their labor or their of the Amazon was relatively dense, wares for part of the economic surplus villages were fairly large and close to- exacted from village farmers by the gether, and some degree of social strati- Political Evolution ruling class and spent by members of fication existed. Moreover, here and that class to raise their standard of there a paramount chief held sway over While the aggregation of villages into living. many communities. chiefdoms, and of chiefdoms into king- The process of political evolution The question immediately arises: doms, was occurring by external ac- which I have outlined for the coastal With unbroken stretches of arable land quisition, the structure of these increas- valleys of Peru was, in its essential extending back from the Amazon for ingly larger political units was being features, by no means unique to this hundreds of miles, why were there elaborated by internal evolution. These region. Areas of circumscribed agri- chiefdoms here? inner changes were, of course, closely cultural land elsewhere in the world, To answer this question we must related to outer events. The expansion such as the Valley of Mexico, Meso- look closely at the environmental con- of successful states brought within their potamia, the Nile Valley, and the Indus ditions afforded by the Amazon. Along borders conquered peoples and terri- Valley, saw the process occur in much the margins of the river itself, and on tory which had to be administered. the same way and for essentially the islands within it, there is a type of And it was the individuals who had same reasons. In these areas, too, au- land called varzea. The river floods distinguished themselves in war who tonomous neolithic villages were suc- this land every year, covering it with were generally appointed to political ceeded by chiefdoms, chiefdoms by a layer of fertile silt. Because of this office and assigned the task of carrying kingdoms, and kingdoms by empires. annual replenishment, varzea is agri- out this administration. Besides main- The last stage of this development was, cultural land of first quality which can taining law and order and collecting of course, the most impressive. The be cultivated year after year without taxes, the functions of this burgeoning scale and magnificence attained by the ever having to lie fallow. Thus, among class of administrators included mobi- early empires overshadowed everything native farmers it was highly prized and lizing labor for building irrigation that had gone before. But, in a- sense, greatly coveted. The waters of the works, roads, fortresses, palaces, and empires were merely the logical culmi- Amazon were also extraordinarily temples. Thus, their functions, helped nation of the process. The really funda- bountiful, providing fish, manatees, to weld an assorted collection of petty mental step, the one that had triggered turtles and turtle eggs, caimans, and states into a single integrated and cen- the entire train of events that led to other riverine foods in inexhaustible tralized political unit. empires, was the change from village amounts. By virtue of this concentra- These same individuals, who owed autonomy to supravillage integration. tion of resources, the Amazon, as a their improved social position to their This step was a change in kind; every- habitat, was distinctly superior to its exploits in war, became, along with the thing that followed was, in a way, only hinterlands. ruler and his kinsmen, the nucleus of a change in degree. Concentration of resources along the an upper class. A lower class in turn In addition to being pivotal, the Amazon amounted almost to a kind of emerged from the prisoners taken in step to supracommunity aggregation circumscription. While there was no war and employed as servants and was difficult, for it took 2 million years sharp cleavage between productive and slaves by their captors. In this manner to achieve. But, once it was achieved, unproductive land, as there was in did war contribute to the rise of social once village autonomy was transcended, Peru, there was at least a steep eco- classes. only two or three millennia were re- logical gradient. So much more re- I noted earlier that peoples attempt quired for the rise of great empires and warding was the Amazon River than to acquire their neighbors' land before the flourishing of complex civilizations. adjacent areas, and so desirable did it 736 SCIENCE, VOL. 169 become as a habitat, that peoples were circumscription." By this he means that point on, the consequences of warfare drawn to it from surrounding regions. a high density of population in an area for that village, and for political evo- Eventually crowding occurred along can produce effects on peoples living lution in general, would be essentially many portions of the river, leading to near the center of the area that are as I have described them for the situ- warfare over sections of river front. similar to effects produced by environ- ation of environmental circumscription. And the losers in war, in order to re- mental circumscription. This notion To return to Amazonia, it is clear tain access to the river, often had no seems to me to be an important ad- that, if social circumscription is opera- choice but to submit to the victors. By dition to our theory. Let us see how, tive among the Yanomamo today, it this subordination of villages to a para- according to Chagnon, social circum- was certainly operative among the tribes mount chief there arose along the Ama- scription has operated among the of the Amazon River 400 years ago. zon chiefdoms representing a higher Yanomamo. And its effect would undoubtedly have step in political evolution than had The Yanomamo, who number some been to give a further spur to political occurred elsewhere in the basin (24). 10,000, live in an extensive region of evolution in that region. The notion of resource concentra- noncircumscribed rain forest, away We see then that, even in the ab- tion also helps to explain the surpris- from any large river. One might ex- sence of sharp environmental circum- ing degree of political development ap- pect that Yanomamo villages would scription, the factors of resource con- parently attained by peoples of the thus be more or less evenly spaced. centration and social circumscription Peruvian coast while they were still However, Chagnon notes that, at the may, by intensifying war and redirect- depending primarily on fishing for sub- center of Yanomamo territory, villages ing it toward the taking of land, sistence, and only secondarily on agri- are closer together than they are at the give a strong impetus to political de- culture (18). Of this seeming anomaly periphery. Because of this, they tend to velopment. Lanning has written: "To the best of impinge on one another more, with the With these auxiliary hypotheses in- my knowledge, this is the only case in result that warfare is more frequent and corporated into it, the circumscription which so many of the characteristics intense in the center than in peripheral theory is now better able to confront of civilization have been found without areas. Moreover, it is more difficult for the entire range of test cases that can a basically agricultural economic foun- villages in the nuclear area to escape be brought before it. For example, it dation" (25). attack by moving away, since, unlike can now account for the rise of the Armed with the concept of resource villages on the periphery, their ability state in the Hwang Valley of northern concentration, however, we can show to move is somewhat restricted. China, and even in the Pet6n region of that this development was not so anom- The net result is that villages in the the Maya lowlands, areas not charac- alous after all. The explanation, it central area of Yanomamo territory are terized by strictly circumscribed agri- seems to me, runs as follows. Along larger than villages in the other areas, cultural land. In the case of the Hwang the coast of Peru wild food sources since large village size is an advantage Valley, there is no question that re- occurred in considerable number and for both attack and defense. A further source concentration and social cir- variety. However, they were restricted effect of more intense warfare in the cumscription were present and active to a very narrow margin of land (26). nuclear area is that village headmen forces. In the lowland Maya area, re- Accordingly, while the abundance of are stronger in that area. Yanomamo source concentration seems not to have food in this zone led to a sharp rise headmen are also the war leaders, and been a major factor, but social circum- in population, the restrictedness of this their influence increases in proportion scription may well have been. food soon resulted in the almost com- to their village's participation in war. Some archeologists may object that plete occupation of exploitable areas. In addition, offensive' and defensive al- population density in the Peten during And when pressure on the available liances between villages are more com- Formative times was too low to give resources reached a critical level, com- mon in the center of Yanomamo terri- rise to social circumscription. But, in petition over land ensued. The result tory than in outlying areas. Thus, while assessing what constitutes a population of this competition was to set in mo- still at the autonomous village level of dense enough to produce this effect, tion the sequence of events of political political organization, those Yanomamo we must consider not so much the total evolution that I have described. subject to social circumscription have land area occupied as the amount of Thus, it seems that we can safely clearly moved a step or two in the di- land needed to support the existing add resource concentration to environ- rection of higher political development. population. And the size of this sup- mental circumscription as a factor lead- Although the Yanomamo manifest porting area depends not only on the ing to warfare over land, and thus to social circumscription only to a modest size of the population but also on the political integration beyond the village degree, this amount of it has been mode of subsistence. The shifting cul- level. enough to make a difference in their tivation presumably practiced by the level of political organization. What the ancient Maya (28) required consider- effects of social circumscription would ably more land, per capita, than did Social Circumscription be in areas where it was more fully the permanent field cultivation of say, expressed should, therefore, be clear. the Valley of Mexico or the coast of But there is still another factor to First would come a reduction in the Peru (29). Consequently, insofar as be considered in accounting for the size of the territory of each village. its effects are concerned, a relatively rise of the state. Then, as population pressure became low population density in the Peten In dealing with the theory of en- more severe, warfare over land would may have been equivalent to a much vironmental circumscription while dis- ensue. But because adjacent land for higher one in Mexico or Peru. cussing the Yanomamo Indians of miles around was already the property We have already learned from the Venezuela, Napoleon A. Chagnon (27) of other villages, a defeated village Yanomamo example that social cir- has introduced the concept of "social would have nowhere to flee. From this cumscription may begin to operate 21 AUGUST 1970 737 while population is still relatively sparse. and R. M. Adams, Eds. (Univ. of Chicago villages could not fission so readily, and thus Press, Chicago, 1960), p. 281]. For China, grew to population levels which, according And we can be sure that the Peten was the prototypical area for Wittfogel's hydraul- to Lanning [Peru Before the Incas (Prentice- far more densely peopled in Formative ic theories, the French Sinologist Jacques Hall, Englewood Cliffs, N.J., 1967), p. 64], Gernet has recently written: "although the may have averaged over 300. times than Yanomamo territory is today. establishment of a system of regulation of 20. See R. L. Carneiro, Ethnograph.-archaol. For- Thus, population density among the water courses and irrigation, and the control schungen 4, 22 (1958). of this system, may have affected the political 21. Naturally, this evolution took place in the lowland Maya, while giving a super- constitution of the military states and impe- various Peruvian valleys at different rates ficial appearance of sparseness, may rial China, the fact remains that, historically, and to different degrees. In fact it is possible it was tie pre-existing state structures and the that at the same time that some valleys were actually have been high enough to pro- large, well-trained labour force provided by already unified politically, others still had not voke fighting over land, and thus pro- the armies that made the great irrigation evolved beyond the stage of autonomous projects possible" [Ancient China, from the villages. vide the initial impetus for the forma- Beginnings to the Empire, R. Rudorff, Transl. 22. Not every step in empire building was (Faber and Faber, London, 1968), p. 92]. For necessarily taken through actual physical tion of a state. Mexico, large-scale irrigation systems do not conquest, however. The threat of force some- appear to antedate the Classic period, whereas times had the same effect as its exercise. In it is clear that the first states arose in the this way many smaller chiefdoms and states preceding Formative or Pre-Classic period. were probably coerced into giving up their Conclusion 7. This is not to say, of course, that large-scale sovereignty without having to be defeated on irrigation, where it occurred, did not contrib- the field of battle. Indeed, it was an explicit ute significantly to increasing the power and policy of the Incas, in expanding their em- In summary, then, the circumscrip- scope of the state. It unquestionably did. To pire, to try persuasion before resorting to the extent that Wittfogel limits himself to force of arms. See Garcilaso de la Vega, tion theory in its elaborated form goes this contention, I have no quarrel with him Royal Commentaries of the Incas and General far toward accounting for the origin whatever. However, the point at issue is not History of Peru, Part 1, H. V. Livermore, how the state increased its power but how Transl. (Univ. of Texas Press, Austin, 1966), of the state. It explains why states arose it arose in the first place. And to this issue pp. 108, 111, 140, 143, 146, 264. the hydraulic hypothesis does not appear to 23. The evolution of empire in Peru was thus where they did, and why they failed hold the key. by no means rectilinear or irreversible. Ad- to arise elsewhere. It shows the state to 8. See The Evolution of Society; Selections from vance alternated with decline. Integration Herbert Spencer's Principles of Sociology, was sometimes followed by disintegration, be a predictable response to certain R. L. Carneiro, Ed. (Univ. of Chicago Press, with states fragmenting back to chiefdoms, specific cultural, demographic, and eco- Chicago, 1967), pp. 32-47, 63-96, 153-165. and perhaps even to autonomous villages. 9. L. Gumplowicz, Der Rassenkampi (Wagner, But the forces underlying political develop- logical conditions. Thus, it helps to Innsbruck, 1883). ment were strong and, in the end, prevailed. elucidate what was undoubtedly the 10. G. Ratzenhofer, Wesen und Zweck der Politik Thus, despite fluctuations and reversions, the (Brockhaus, Leipsig, 1893). course of evolution in Peru was unmistakable: most important single step ever taken 11. F. Oppenheimer, The State, J. M. Gitterman, it began with many small, simple, scattered, Transl. (Vanguard. New York, 1926). and autonomous communities and ended in the political evolution of mankind. 12. E. Jenks, A History of Politics (Macmillan, with a single, vast, complex, and centralized New York, 1900), p. 73. empire. References and Notes 13. J. Vansina, Kingdoms of the Savanna (Univ. of Wisconsin Press, Madison, 1966). 24. Actually, a similar political development did 1. R. L. Carneiro, in The Evolution of Horti- 14. For example, Julian H. Steward wrote: "It is take place in another part of Amazonia- cultural Systems in Native South America: possible, therefore, that the Maya were able the basin of the Mamore River in the Mojos Causes and Consequences; A Symposium, J. to develop a high civilization only because plain of Bolivia. Here, too, resource concen- Wilbert, Ed., Antropol6gica (Venezuela), they enjoyed an unusually long period of tration appears to have played a key role. See Suppl. 2 (1961), pp. 47-67, see especially pp. peace; for their settlement pattern would seem W. Denevan, "The Aboriginal Cultural Ge- 59-64. to have been too vulnerable to warfare" ography of the Llanos de Mojos of Bolivia," 2. For example, the early American sociologist [Amer. Anthropol. 51, 1 (1949), see p. 17]. Ibero-americana No. 48 (1966), pp. 43-50, Lester F. Ward saw the state as "the result 15. D. E. Puleston and D. W. Callender, Expe- 104-105, 108-110. In native North America of an extraordinary exercise of the rational dition 9 No. 3, 40 (1967), see pp. 45, 47. north of Mexico the highest cultural de- . . . faculty" which seemed to him so ex- 16. M. D. Coe, The Maya (Praeger, New York, velopment attained, Middle-Mississippi, also ceptional that "it must have been the ema- 1966), p. 147. occurred along a major river (the Mississippi), nation of a single brain or a few concerting which, by providing especially fertile soil minds...." [Dynamic Sociology (Appleton, 17. See R. L. Carneiro, in Men and Cultures, and riverine food resources, comprised a New York, 1883), vol. 2, p. 224]. Selected Papers of the Fifth International zone of resource concentration. See J. B. 3. See, for example, V. G. Childe, Man Makes Congress of Anthropological and Ethnological Griffin, Science 156, 175 (1967), p. 189. Himself (Watts, London, 1936), pp. 82-83; Sciences, A. F. C. Wallace, Ed. (Univ. of 25. E. P. Lanning, Peru Before the Incas (Pren- Town Planning Rev. 21, 3 (1950), p. 6. Pennsylvania Press, Philadelphia, 1960), pp. tice-Hall, Englewood Cliffs, N.J., 1967), p. 59. 4. have in my files recorded instances of sur- 229-234. I 18. In early agricultural times (Preceramic Period 26. Resource concentration, then, was here com- plus food production by such Amazonian bined with environmental circumscription. tribes as the Tupinamba, Jevero, Munduruct, VI, beginning about 2500 B.C.) human settle- And, indeed, the same thing can be said of Tucano, Desana, Cubeo, and Canela. An ex- ment seems to have been denser along the the great desert river valleys, such as the haustive search of the ethnographic literature coast than in the river valleys, and sub- Nile, Tigris-Euphrates, and Indus. for this region would undoubtedly reveal many sistence appears to have been based more more examples. on fishing than on farming. Furthermore, 27. N. A. Chagnon, Proceedings, VilIth Intema- 5. Wittfogel states: "These patterns [of organi- some significant first steps in political evolu- tional Congress of Anthropological and Ethno- zation and social control-that is, the state] tion beyond autonomous villages may have logical Sciences (Tokyo and Kyoto, 1968), come into being when an experimenting com- been taken at this stage. However, once sub- vol. 3 (Ethnology and Archaeology), p. 249 munity of farmers or protofarmers finds large sistence began to be based predominantly on (especially p. 251). See also N. Fock, Folk 6, sources of moisture in a dry but potentially agriculture, the settlement pattern changed, 47 (1964), p. 52. fertile area. . a number of farmers eager . and communities were thenceforth concen- 28. S. G. Morley and G. W. Brainerd, The to conquer [agriculturally, not militarily] trated more in the river valleys, where the Anctent Maya (Stanford Univ. Press,. Stan- arid lowlands and plains are forced to invoke only land of any size suitable for cultivation ford, Calif., ed. 3, 1956), pp. 128-129. the organizational devices which-on the was located. See E. P. Lanning, Peru Before 29. One can assume, I think, that any substan- basis of premachine technology-offer the the Incas (Prentice-Hall, Englewood Cliffs, tial increase in population density among the one chance of success; they must work in N.J., 1967), pp. 57-59. Maya was accompanied by a certain intensi- cooperation with their fellows and subordi- 19. In my files I find reported instances of vil- fication of agriculture. As the population in- nate themselves to a directing authority" [Ori- lage splitting among the following Amazonian creased fields were probably weeded more ental Despotism (Yale Univ. Press, New tribes: Kuikuru, Amarakaeri, Cubeo, Urubfi, thoroughly, and they may well have been Haven, Conn., 1957), p. 18]. Tupari, Yanomamo, Tucano, Tenetehara, Ca- cultivated a year or two longer and fallowed 6. For Mesopotamia, Robert M. Adams has nela, and Northern Cayap6. Under the con- a few years less. Yet, given the nature of concluded: "In short, there is nothing to sug- ditions of easy resettlement found in Ama- soils in the humid tropics, the absence of gest that the rise of dynastic authority in zonia, splitting often takes place at a village any evidence of fertilization, and the mod- southern Mesopotamia was linked to the population level of less than 100, and village erate population densities, it seems likely that administrative requirements of a major canal size seldom exceeds 200. In coastal Peru, Maya farming remained extensive rather system" [in City Invincible, C. H. Kraeling however, where land was severely restricted, than becoming intensive.