The authors raise spatial analysis to a new level of sophistication--and insight--in proposing a mathematical model of 'imperfect optimisation' to describe maritime networks. This model encodes, metaphorically, the notion of gravitational attraction between objects in space. The space studied here is the southern Aegean in the Middle Bronze Age, and the objects are the 34 main sites we know about. The 'gravitation' in this case is a balance of social forces, expressed by networks with settlements of particular sizes and links of particular strengths. The model can be tweaked by giving different relative importance to the cultivation of local resources or to trade, and to show what happens when a member of the network suddenly disappears.
Keywords: Greece, Aegean, Bronze Age, maritime networks, spatial analysis
If we follow Smith's line that 'spatial relationships are the sinews of archaeological research' (2003: 77), then we might expect spatial relationships to be fundamental across all scales of archaeological analysis. However, space has received a surprisingly uneven treatment within the discipline, and an assumed equivalence between physical, geometric space on the one hand, and relational, social space on the other, seems deeply entrenched. What is required is an approach that incorporates the fundamental notion that humans create space through social practices (Harvey 1973; 1996; Tilley 1994; Thrift 1996; Hetherington 1997; Smith 2003), while also acknowledging that physical parameters are not entirely redundant in this process.
In much spatial analysis physical interactions between points are seen as primary and the social interconnections come later (Batty 2005: 149). Sites are thought to emerge and gain their character on largely local grounds, and any interactions with other communities in the region follow on from that. The connections between sites are simply drawn as lines, without weight or direction. But the likelihood is that site interactions themselves contribute to the size and status of the sites in question (Sindbaek 2007). How, then, might we give the interactions and locations equal status and achieve, borrowing from Batty, an 'archaeology of relations'?
The behaviour of a large collection of entities, like island communities, is best addressed stochastically, rather than deterministically. Our approach synthesises techniques derived from statistical physics and complex network analysis for tackling social questions (Carrington et al. 2005; de Nooy et al. 2005; Evans 2005). However, before presenting our model we shall discuss its context, the southern Aegean Bronze Age.
The study of networks in the Aegean
Our case study for the development of the method starts with Broodbank's work on the Cyclades, the most systematic attempt thus far in Aegean prehistory to explain the growth of certain sites in terms of their interactions (Broodbank 2000). His approach was perhaps encouraged by the tiny resource base of some key Early Cycladic sites, making it likely that their role in a network was all important. In Broodbank's network, each individual site is represented by a node, and each connection between sites represented by a link. This simple transformation of the Cyclades into a graph of nodes and links enables a second step, the adoption of basic techniques from graph theory to analyse network characteristics. Broodbank opts for 'Proximal Point Analysis' (PPA), a technique already used effectively in archaeology and anthropology for interaction studies in other archipelagos, notably in Oceania (Terrell 1977; Irwin 1983; Hage & Harary 1991; 1996).
Broodbank's next step is to add hypothetical sites to islands on the basis of population estimates derived from site surveys. Links are then drawn from each hypothetical site to its three nearest neighbours. …