From platforms to people: rethinking population estimates for the abandoned agricultural settlement at Engaruka, northern Tanzania

Abstract

The Late Iron Age abandoned agricultural settlement of Engaruka in northern Tanzania has long drawn both popular and scholarly attention due to the size and scale of its agricultural works and its apparent size and density of settlement. Estimates of the total number of inhabitant have, nevertheless, varied dramatically. While a recent reassessment by Laulumaa (2006) has greatly enhanced the accuracy of future estimates, there are a number of flaws exist in his calculations. By making a series of observations about the structure and lifecycle of Eastern African family units and drawing on comparative ethnographic data from the Marakwet/Pokot region of Kenya, this paper revises Laulumaa's estimate. It also discusses the comparative potential for understanding population densities at Engaruka and comparable extant agricultural systems.

La colonie agricole abandonnée d'Engaruka, au nord de la Tanzanie, qui date de l'âge du fer tardif, a pendant longtemps attiré l'attention populaire et savante en raison de la taille et de l'échelle du travail agricole ainsi que de la taille et de la densité apparente de la colonie. Néanmoins, les estimations relatives au nombre total d'habitants varient énormément et bien que la réévaluation récente de Laulumaa (2006) ait sensiblement amélioré la précision des estimations actuelles et futures, il existe encore un certain nombre de failles. En se basant sur une série d'observations à propos de la structure et du cycle de vie des unités familiales d'Afrique de l'est, et en s'appuyant sur des données comparatives ethnographiques sur la région Marakwet/Pokot au Kenya, cet article propose une révision de l'estimation de Laulumaa. Il discute par la suite du potentiel comparatif afin de comprendre les densités de population à Engaruka puis, de manière comparable, les systèmes agricoles d'aujourd'hui.

Keywords:

• Engaruka
• Marakwet/Pokot
• agriculture
• population density
• Kenya
• Tanzania

Acknowledgements

Many thanks are due to Vesa Laulumaa for comments and suggestions on an earlier version of this manuscript. This work was principally written during preparation for my doctoral thesis, which was funded by the Arts and Humanities Research Council of the United Kingdom, with support from the Tweedie Exploration Fund (University of Edinburgh), the Meyerstein Fund (University of Oxford) and the British Institute in Eastern Africa.

Notes

1. Pokot and Marakwet houses vary from 4–5 m in diameter, are normally spaced some 4 m apart, and are surrounded by a sizeable living compound of at least a couple of metres on all sides (Moore 1986; Davies 2009a). Modern Sonjo compounds seem similar in scale to this (personal observation).

2. This male/female house distinction is clearly over-simplified and may, for example, also be expressed as ‘sleeping/cooking’ houses. However, for the purpose of clarity within the current discussion this simplistic male/female categorisation will be utilised.

3. The agricultural Pokot referred to here live in the Cherangani foothills and are very similar in economy and culture to the Marakwet. They can be contrasted with the Pastoral Pokot of the plains to the north who lead a quite different and much more nomadic lifestyle.

4. Interestingly, figures from Bertelsen (1995) show that the modern Maasai population of the Engaruka ward have some 5.4 people per household (likely to equate to something like three people per house).

5. In my Pokot survey, a maximum of only around 70% of all house platforms are occupied at any one time, even in well populated areas that appear to be thriving (Davies 2009a). Laulumaa's (2006) two-thirds figure may therefore not be too wrong if one assumes that the whole of Engaruka was occupied concurrently. If, as seems likely, this was not the case then the total number of houses occupied at any one time would be significantly less than two-thirds.

6. Even if the figure of three people per household is an underestimate, the likelihood that many of the house platforms would not have been occupied concurrently suggests that these figures are quite reasonable.

7. Elsewhere such densities would be considered ‘urban’ in scale, yet there is little evidence for social differentiation or for complex institutions of management and conflict resolution. See Davies (2009b) for discussion of complex management in non-hierarchical irrigating societies.