As promised in my review of Water: the Epic Struggle by Steven Solomon, here is a more scholarly look at the hydraulic hypothesis.

As I wrote before, "The hydraulic hypothesis says that large-scale irrigation works that supported the land-based empires of the ancient world required a certain type of government—an absolute monarch/emperor supported by an elite caste of technocrats or priests—to implement, and that this exploitation of water resources also provided legitimacy to such governments." The academic definition is more rigorous, considering hydraulic works as a significant cause of the origin of states.

The paper that I found to read is: Michael J. Harrower. 2009. Is the hydraulic hypothesis dead yet? Irrigation and social change in ancient Yemen. World Archaeology 41(1):58-72. (It comes from a special issue on the archaeology of water).

The paper is summarized (in the abstract) as follows:

Irrigation played an important role throughout ancient Southwest Arabian histories. Irrigation structures provide some of the earliest evidence of crop agriculture and large-scale flash floodwater irrigation systems sustained ancient states; the region thus offers important potential for reconsidering links between irrigation and social change. This paper examines millennia-long connections between social relations and the increasing technological and organizational complexity of irrigation in ancient Yemen. While the hydraulic hypothesis in its original deterministic formulation does not adequately account for the complexity and diversity of regional histories, large centrally managed irrigation systems played an indisputably significant role in Southwest Arabian state formation. Irrigation not only generated the food to sustain burgeoning populations but, just as importantly, afforded ancient kings the ideological prestige of commanding transformation of hyperarid areas into lush, bountiful oases.

The author describes the purpose of the study as follows:

[This study] challenges the view that centrally coordinated irrigation was necessary to provision states in Southwest Arabia and argues that large-scale irrigation was not agriculturally required but played a vital social role as a source of ideological magnetism and religiosity that attracted hinterland peoples towards state polities.

Dr. Harrower discusses the following points:

  • Social logistics of irrigation: that is, how much centralization is required and how activities are coordinated
  • Taking a task-based view of social logistics by examining job titles and the perks that went with them
  • Interestingly, some operation and maintenance positions for traditional irrigation were elected and represented a power base distinct from tribal hierarchies
  • Using GIS (i.e. digital elevation models, DEMs) to model water flow for archaeological applications; this allowed the watershed size (which impacts the available flow) to be determined
  • Sayl ( سيل ) (torrent/flood flow) irrigation systems
  • There was a 16 m high dam at Sabaean capital of Ma'rib that redirected flash floods to sluice gates. There are records of it being repaired more than once (including by an Ethiopian viceroy) but eventually it fell into ruin shortly before the rise of Islam.
  • Size of watershed (which corresponds with water availability in arid environments—except different regions of Yemen get quite different levels of precipitation, so it's not an exact proxy) was not the only factor in the location of capitals (e.g. also want access/control over trade routes for incense)
  • Ideological significance of irrigation: this society had gods of thunderstorms and irrigation works
  • An Ethiopian viceroy (i.e. a foreign ruler) repaired the Ma'rib dam as mentioned above to boost his legitimacy

These points are discussed within the context of the two main aspects to this paper:

  1. social logistics (task-based perspective -> job titles), and
  2. spatial dimensions (using GIS to get size of watersheds)

In the end, the author has these conclusions:

Intriguingly, however, states in Yemen first developed not in Yemen’s moist western highlands, where dry-farming was possible, but instead along the hyper-arid margins of the desert, where irrigation was necessary. By harnessing vast water flows, and transforming hyper-arid deserts into lush, bountiful oases with the blessing of revered gods, states’ leaders not only benefited economically but claimed enormous ideological credit. But these large-scale systems were not agriculturally necessary to provision Southwest Arabian states; dry-farming, terrace agriculture, and smaller-scale community-managed irrigation could have provided the food necessary to sustain large populations. Indeed, these alternatives likely provided much of the food for the later highland state of Himyar and they continue to provide most of the agricultural production in Yemen today. For the earliest states, however, massive irrigation works indentured farmers to complex polities through largely state-contrived dependence on irrigation water from restricted access sources and such impressive hydraulic installations were arguably as important ideologically as they were economically.

Understanding irrigation’s role in the societal trajectories of ancient Southwest Arabia requires attention to a variety of influencing factors, including hydrology, economy and ideology. While emphasis is often placed on the chronological, technological and economic dimensions of irrigation, it was not only that large watersheds offered productive capacity, but that access could be restricted at these locations allowing elites to dominate and position themselves as cosmological authorities of food production.

So, in the point of view propounded in this paper, large-scale irrigation works seem to be about prestige and control more than being driven as a food-supply necessity.

Overall, I found this paper to be very well-written and well-organized, making it easy to follow and understand even though archaeology/anthropology/history is not my speciality. As someone with an engineering background, I missed having more quantitative data and statistical analysis, but I guess that probably reflects the amount of effort that has to go into finding each clue about the past. Harrower includes a lengthy bibliography, which includes a couple of items that interest me to read further.