102 Anna C. F. Collar
This potential for exploring relational space has been investigated by scholars working
with topological data in ancient texts. Isaksen (2008) used information about transport
routes in the Antonine Itineraries and the Ravenna Cosmography to explore the posi-
tioning of towns in southern Spain as nodes in a network. Using these documents, which
are essentially linear descriptions of how to move through the Roman world, allows the
archaeologist to “get inside the heads” of how the Romans viewed their world. Isaksen
demonstrates the potential of a relational database, alongside GIS and network analysis
software, “to make a spatial argument about the relative importance of key towns within
a transport network and expose the constituent elements of that argument in a visual
manner” (Isaksen 2008: 2).
More recently, Barker et al. have been using networks to explore spatial data con-
tained within Herodotus’Histories. Their project involved the digital mark-up of the
text being fed into a database, allowing them to reconstruct the world of Herodotus as
a system of interconnections and relationships. “[...] every step of the way through the
narrative points of contact are made between different places in a variety of ways [...]
By seeking to lift these connections out of the text, we hope to counter the conven-
tional emphasis on topography, which appears all too evident from our contemporary
viewpoint of the world afforded by satellite imaging, and refocus attention instead on
thetopologicalrelationships between places—the links that depend on human agency
and the associative clusters that certain places form over the course of the narrative”
(Barker et al. 2010: 5).
Material Networks and Relational Space
The network framework has also been used to understand relational space as encoded
in material culture. Analyzing the distribution patterns of ceramics, artifacts, or settle-
ments as a relational network enables archaeologists to build a picture of site interactions
and reconceptualize such phenomena as the spread of technological skills, adoption of
decorative styles, or the growth and decline of settlements as a result of network interac-
tions. Proximal Point Analysis (PPA), a technique developed in anthropological studies
in Oceania (Terrell 1977), was used by Broodbank (2000) to analyze interactions of
islands and the growth of settlements, centrality, and interaction in the Early Bronze
Age Cyclades.
PPA is a simple gravity model that predicts interactions between distributed nodes.
It begins from the assumption that each node interacts most intensely with its three
closest neighbors. Because of their geographical position, some nodes collect more than
three links—so becoming centers in the resulting network. PPA shows relative degrees
of connection, rather than absolute links, and has been used to reevaluate centrality
and isolation—couched as the degree of influence that a node exerts on regional
interactions—by highlighting the most well-connected nodes and the most active
routes. By factoring in distance, inter-island visibility, and means of travel, Broodbank
revealed that population size and network connectivity were linked: as populations on
larger islands grew, so did their self-sufficiency and the introspection of sub-clusters,