Current Swedish Archaeology, Vol. 12, 2004 141 Sakbihs and Polyagency Sakbihs and Polyagency The Architectural Causes of Human Activities in the Cochuah Region, Quintana Roo, Mexico Johan Normark Most archaeological theories state that we should primarily study the human agent or social structure behind materiality. In this article it is proposed that we should focus on the effects that polyagents (materialities and im- materialities) have brought about in past social networks. By creating a biography of a polyagent it is hoped that we can reconstruct the way in which architecture has affected past environment and human behaviour and intentions. Causeways or sakbihs are one such category of polyagents found throughout the Maya Lowlands. Their relationships with agriculture, water and climate in the Cochuah region, Quintana Roo, Mexico, are discussed. Johan Normark, Department of Archaeology, University of Gothenburg, Box 200, SE-405 30 Göteborg, Sweden Key words: agriculture, causality, climate, materiality, Maya, polyagency, practice, sakbihs, time, water THE PROBLEM OF PAST SOCIAL PRACTICE Agency and practice theories of various sorts (Bourdieu 1977, 1990; Giddens 1979, 1984; Ortner 1984; Sahlins 1981) have gained popularity in archaeology during the past two decades (Barrett & Fewster 2000; Cornell & Fahlander 2002; Dobres & Robb 2000; Dornan 2002; Gillespie 2001; Shanks & Tilley 1987; Smith 2001). However, there is a common lack of a clear definition of what a practice really is on the operational level. I define a practice in an archaeological per- spective, as what a human agent did in relation to materiality in a temporal and spatial setting. Each practice cannot be the same at various locations and times, since the past agent had both experience and anticipations at every stage where we find archaeological data (Gell 1998; Husserl 1991). A practice can thus never be the same in itself. There is always a differance that is dependent on time (Cornell 2000; Hägglund 2002; Wood 1989). The essence of a practice is thus problematic since a practice can always be performed by the same agent, the same patient (receiver of agency) and in the same spatial location, but never in the same temporal setting. The critical problem in defining a practice is thus time. Although several authors have attacked the problem of time in archaeology, they tend to focus on the long term (Golden 2002; Gosden 1994; McGlade 1999;