In our modern world, being NOT connected is actually more difficult than being connected. We all use social media, messengers, email, phones and everything else that keeps us in touch with our loved ones, our colleagues, our customers and basically everyone. But how were people connected 5000 years ago? Can we even get an idea of how their social networks functioned? The answer is simple: Yes – and no. Of course, we will never now for sure how the social networks of ancient societies looked like but there is a tool that can help us to at least make some assumptions: Social network analysis.
Social network analysis (SNA) aims to understand the structure and functioning of social networks and their emergent behaviours. Such networks are developed from (inter)relational data and consist of social entities (e.g., individuals, groups, organisations) and the interactions between them. Those networks are modelled by using graph theory with nodes (actors) and ties (interactions). Common tasks of SNA include the identification of flows (like information or behaviour) and in this context the identification of important or central nodes and emergent behaviour and ties. The emergent ties may be directed or undirected. Furthermore, a tie may have an associated weight which can give information about for example the strength of a connection.
There are several configurations of social networks. If there is just one set of actors, it is called a one-mode network, a network with two different sets of actors is thus a two-mode network. A social network may be a combination of one-mode and two-mode where the actors of one set are not only connected to the actors in the other set, but also interconnected. If these connections are of different kinds (e.g., proximity, kinship, trade), they are called multi-relational. If the actors in the second set are also interconnected, the network becomes a multi-relational, multilevel one.
An example for such a network is a university department. The actors in the first set are the people related to the department like students and staff. They all share a general interest in one broad subject (like Geography), and therefore belong to this department. However, in this department, there are several sub-departments or clusters that specify their research interest (human geography, physical geography). These are the second set of actors. Each actor of the first set is connected to one or more actors of the second set. But the people in the department are also interconnected. Some may have a professional relationship, some may be friends, and some may be working together and be friends as well, i.e., there are multiple kinds of relations between these actors. If the clusters are also interconnected in some ways, the network becomes multi-layered.
There are several possibilities to analyse social networks, such as centrality (degree, betweenness, closeness or eigenvector centrality) or statistical measures (diameter, distance, degree, reciprocity, density, clustering). Static networks can be examined as well as evolving ones, ego networks as well as sociocentric ones. This broad range of tools and possible applications leads to a wide variety of SNA models, concepts, and approaches.
In my project, I want to use SNA to identify ”key player” settlements, i.e., settlements that cause a significant disruption of the system when removed, and see if settlements form clusters or groups. Understanding how those settlement systems where connected, how the systems functioned and evolved will help me to evaluate if the networks influenced their resilience and to what extent. I assume that if two settlements were already connected through social ties like migration or trade it would be more likely that they would support each other in times of crisis, for example with food if one settlement lost part of its harvest due to a drought.
So, although SNA is usually used with a vast amount of input data to examine modern day social networks, it also can help us to learn something about connections between long gone people or – in my case – settlements.
I’m really looking forward to the results of my analyses and if you are too, spread the word on your social network of choice!
Borgatti, S. P., Everett, M. G. and Johnson, J. C. (2013) Analyzing social networks. Los Angeles: SAGE.
Scott, J. (2013) Social network analysis. Third edition. Los Angeles: SAGE.
Tabassum, S. et al. (2018) ‘Social network analysis: An overview’, WIREs Data Mining and Knowledge Discovery, 8(5). doi: 10.1002/widm.1256.
Wasserman, S. and Faust, K. (1994) Social network analysis: methods and applications. Cambridge; New York: Cambridge University Press (Structural analysis in the social sciences, 8).