Camel caravan

Camel caravan
Mosaic from Deir al-Adas, Syria, 8th century (photo: J.C.Meyer)
The research project Mechanisms of cross-cultural interaction: Networks in the Roman Near East (2013-2016) investigates the resilient everyday ties, such as trade, religion and power, connecting people within and across fluctuating imperial borders in the Near East in the Roman Period. The project is funded under the Research Council of Norway's SAMKUL initiative, and hosted by the Department of archaeology, history, cultural studies and religion, University of Bergen, Norway.

Project manager / blog editor:

Sunday, 17 November 2013

Proto-globalisation in the Indian Ocean World and network approaches

Last weekend I attended the concluding conference of the ERC-funded Sealinks project. The Sealinks group, headed by Nicole Boivin of Oxford University, has investigated the early maritime links that connected societies on the Indian Ocean rim, using primarily evidence produced by archaeobotany, zooarchaeology, genetics and linguistics. Many results have already come out, and these shed light on processes going further back in time and having more profound and lasting impact, than those evident from earlier studies based on artefactual, epigraphic and literary evidence.

The conference gathered researchers working with biological and environmental proxies, along with with traditional archaeologists and the odd historian. Indian Ocean trade and navigation, as well as the  early history of globalisation being among my other research interests, I found the conference immensely informative. It struck me, however, that although almost all presenters used the N-word, there was next to no discussion on what a network is, how past networks can be approached, and what insights this might give. This is not intended as criticism of the work presented, which was generally of very high standard. In fact the observation would be valid for most archaeological conferences I've attended, and I've worked with the Indian Ocean for more than ten years myself before starting consciously to think in terms of networks myself. Indian Ocean connectivity, however, is a subject that would readily lend itself to network approaches, and in this post I'd like to share a few thoughts on how different groups of material might have benefitted from being approached from a network perspective.

Networks of people
My own research on Indian Ocean trade is primarily on merchant networks: how they formed, cohered, operated and interrelated. In my paper in Oxford I argued that merchants from the Syrian city of Palmyra engaged with local and regional social networks of power, ideology/religion and commerce, in the process of expanding their own ethnically based network from the Syrian Desert into Mesopotamia, the Persian Gulf, the Red Sea and the Arabian Sea. Our data on merchant networks are of literary, epigraphic and ceramic nature – the Palmyrene network for instance being attested by circa 50 inscriptions recording the presence or activities of Palmyrene groups and individuals in settings certainly or very likely connected with Indian Ocean trade. Similar data would be possible to utilize for other groups. Ingo Strauch and collaborators for instance recently published more than 200 inscriptions in Arabic, Indic, Ethiopic, Aramaic and Greek script from a sanctuary in the cave of Hoq on the island of Socotra. Many of these should be possible to connect with places of origin, religion, language, and script used, reconstructing the network that intersected at the node of Hoq in a manner that would be possible to approach by the toolbox of Social Network Analysis.

Networks of objects
New inscriptions are occasionally still discovered, but in general, the written evidence relevant to ancient Indian Ocean trade is delimited and finite. The archaeological data available from Indian Ocean settings, however, has exploded over the last two decades, following major excavation projects  and better practices of documenting, identifying and publishing finds. In my opinion, there are cases where distribution patterns are not necessarily well suited for network analyses, as the presence of objects does not always reveal how, when and by which routes and agents they moved from point of origin to find-spot. In the case of the Indian Ocean, however, distribution patterns clearly do have potential for network analyses, as distances, topography and climate led to virtually all communication taking place by sea. Network analysis based on distribution patterns of ceramics, beads and glass – goods that can be traced to place of production on stylistic, chemical and petrographic grounds, could clearly reveal directions, bottlenecks and clustering of trade.

Networks of places
Clearly, the same would be the case with settlements. Attempts at modeling relationships between ports and their hinterlands and between ports, based on size, kinds of imports and exports, period of occupation, distance to other centers, season of navigation etc. might help us think about why some places and some regions were important, while others seem to not to have been taking part in long-distance exchange to the same extent.

Networks of genes
Many of the papers in Oxford were on the movement of crops, animals and people, studied by way of genetic material recovered from archaeological sited. Dispersal of genes was in most cases visualized by phylogenetic trees. This is of course also a network analysis, showing genetic connections between the varieties of crops etc. found in different settings. If sufficient numbers of dated examples exist from a sufficient number of sites, I think these would also be good for showing patterns of prehistoric seafaring.


  1. Hello Eivind,
    I was also at the Oxford conference in November and I very much appreciate this summary you've offered. As one of the historians who presented, I, too, would love to know how the many kinds of evidence presented at the conference could be exploited to create, perhaps as a kind of "network analysis," thick descriptions of at least some networks of trade and communication. For infectious disease transmission (the topic of my talk), specific kinds of networks need to be documented in order to build persuasive stories of epidemiological spread. The immediate problem, it seems to me, is how to combine different kinds of evidence when they having varying levels of documentary value? I'm not saying it's impossible, but it is a challenge. No network analysis can be persuasive unless we have some faith that we're comparing like with like. Future discussions of this point will be valuable!

  2. Dear Monica,

    thanks for your comment! I remember your presentation very well, and the evidence of disease transmission was one of the real eyeopeners for me at the conference.
    Of course your point on the difficulties involved in reconstructing networks from different kinds of evidence is valid. Where a letter or an inscription might (or might not) give exact information on on past connectivity, a potsherd or a pathogen could have arrived by any number of routes, carriers or mechanisms. My defense for why thinking in terms of networks might nevertheless be useful, is that network theory offers both a way of approaching connectivity in the past, and a methodology for measuring that connectivity quantitatively. Being a historian by training as well as inclination, I mostly find myself in situations were I simply don’t have sufficient quantitative basis in my data for the powerful statistical tools offered by network theory. Still, I think that plotting fragmentary networks or networks that might have existed is useful, because it forces us to think carefully about the nature of the ties that we are proposing, and the contacts that might or must have been there. So my suggestion to your problem is that we need to combine network analyses with qualitative analyses, and that if we do that properly I hope that we should also be able to combine different kinds of evidence within the same network. What is your take on this?