Published online by Cambridge University Press: 24 October 2022
Among archaeologists concerned with the study of cultural regularities in the evolution of the early civilizations, few general hypotheses have stirred such controversy as that which postulates a causal link between the phenomenon of irrigation agriculture and the origins of the state. Adumbrated initially by Karl Wittfogel in the 1920s and stated by him most completely in Oriental Despotism (1957), the hydraulic theory has been subject to much discussion and to varying fates in recent literature and research history. Although Wittfogel's original formulation was based chiefly on Old World data, the theory has had major impact on research and interpretation in the New. The first large-scale use of the approach in American archaeology followed World War II, when the Institute of Andean Research sponsored the Viru Valley Project in North Coastal Peru (Bennett, 1948; Willey, 1953). In recent years American archaeology has become concerned with questions of process beyond the limits of simple historical reconstruction (Binford and Binford, 1968). Concepts derived from systems theory (Flannery, 1968a; Hole and Heizer, 1969) are increasingly invoked to explain the cause-and-effect feedback mechanisms involved in the evolution of culture. Concomitantly, a virtual explosion of data has occurred concerning the chronology, size, and sociological and demographic matrix of New World irrigation systems. Thus, both investigative techniques and theoretical frameworks have undergone considerable recent modification and the body of relevant data is large and growing. While some might consider the hydraulic agriculture hypothesis a dead issue, such is not the case. Changes in the total conceptual context of any theory, and new evidence both pro and con, necessitate reevaluation of the theory.
Given the recency of much of the field work cited in this paper, many substantive data remain unpublished. I wish to express my thanks to those who have graciously permitted me to quote often preliminary results of their recent work and recent thinking sent to me as personal communications. Thus, many investigators have contributed significantly to the preparation of this paper, although they are in no way responsible for its conclusions, with which many of them would probably more or less disagree. Special thanks go to Pedro Armillas for showing me selected areas of his chinampa survey in the southern Valley of Mexico; to Kent V. Flannery for taking me to Hierve el Agua, Oaxaca; to M. Edward Moseley, Carol J. Mackey, and Kent C. Day for my extremely productive stay with the Harvard Chanchan Project, Trujillo, Peru; to Richard S. MacNeish and Gary Vescelius for an instructive visit to Ayacucho; and to Thomas C. Patterson for reporting data from his recent research on the Central Coast of Peru and for copies of several unpublished manuscripts. This paper has benefited further from comments and suggestions from Edward Calnek, Marvin Harris, Jeremy A. Sabloff, and William T. Sanders.