The Rural Landscape of Jordan in the Seventh-Nineteenth Centuries AD: The Kerak Plateau

Article excerpt

Introduction

The archaeological story of the rural landscape in Jordan has often been portrayed as opposition between the nomad (Bedu) and farmer (fellah), between the 'Desert' and the 'Sown'. In fact, current historical and anthropological research shows that the relationship between nomad and farmer should more correctly be seen as complementary (Johns 1994; Lancaster & Lancaster 1995: 106; Palmer 2001), as a continuum: a sliding scale from totally nomadic and mobile communities through semi-nomadic communities to sedentary communities. Concomitant with this sliding scale of mobility is the sliding scale of specialist pastoralism through extensive to intensive agricultural regimes. The regions of Jordan exhibit contrasting environmental and exploitation zones within the same small area. These zones may well be exploited at the same time by several discreet communities with totally different economic strategies. This is a particularly apt framework for considering the rural landscape of the Kerak Plateau which is often considered to be part of the transitional zone between 'Desert' and 'Sown' where the proportions of communities engaged in animal and plant husbandry have shifted both geographically and temporally (Lewis 1987). Halstead (1987) has described these various agricultural regimes for the Aegean basin, and while it is not suggested that they will be identical in detail for Jordan, the concept of differing types of agriculture certainly does apply. He distinguishes between extensive/traditional agriculture with its low-risk and less labour-intensive strategies and intensive/alternative agriculture that requires greater labour input and offers greater potential for the production of a surplus. Davis (1991: 137-8) has suggested the visible physical attributes for these various regimes. The challenge is to translate this theoretical framework into archaeological reality and thereafter historical interpretation. La Bianca's 'food-system model' used to analyse the survey material of the Hesban Project (La Bianca 1990; McQuitty 1993) is based on these concepts of different agricultural regimes producing different site and off-site artefact distributions. The shortcomings of the resulting interpretation should not be blamed on the validity of the model but rather on the inadequacy of the survey data collected for this type of analysis.

Survey in Jordan

Survey is the main archaeological source for data relating to information about rural settlement in this period in Jordan (Figure 1). The surveys conducted have largely been 'purposive' and 'extensive' i.e. related to one period of time or to sites prominent in the landscape. The emphasis is placed upon the site itself, rather than the site in its landscape and context. Inevitably the interpretation of this survey data devolves to statements regarding presence/absence of particular occupation periods rather than the more sophisticated conclusions of 'how' the landscape was used. This is by no means a unique situation to Jordan. Survey methodologies have been developed elsewhere in the Mediterranean world that are designed to counter this bias. Just one example is given here: the Boeotia Project working in the area north of Athens employs intensive survey methods that 'pay attention not only to dense clusters of surface artefacts (primarily potsherds) which would also represent towns, villages or villas, but smaller and more vestigial surface scatters that could be interpreted as family farms, animal shelters, rural cemeteries, quarries and the like' (Bintliff 2000: 38). Such techniques have been used to great effect in Jordan (e.g. Kennedy & Freeman 1995; Watson & O'Hea 1996) but the temporal and geographical coverage is by no means uniform or synthesised (Finkelstein 1998; McQuitty 2001). Added to this, for the post-ninth century AD, our knowledge of the material culture, particularly ceramics that are the main dating tool, is patchy and thus even the presence/absence conclusions can be skewed (Johns 1998; Milwright 2000, 2001). …