The community of the realm in Scotland, 1249–1424
History, law and charters in a recreated kingdom

Social network analysis is a set of techniques and concepts in order better to understand the social relationships between individuals. This is achieved through the creation of graphs called sociograms and the application of graph theory to elucidate the links between actors. Social network theory provides the toolkit based on decades of research by social anthropologists, sociologists, social psychologists and others. All of this allows historians to think more fruitfully about how historical actors formed parts of social networks. Finally, networks do not have to link up people; other historical networks tie together, for example, texts or places.

A sociogram is a network graph, usually constructed using specialty software such as Gephi or Netdraw. The sociogram is made up of a series of individual points called nodes, which respresent individual persons, documents, places, or other networked thing. The nodes are connected by lines which are termed edges, or sometimes ties. A line indicates some kind of connection between two or more nodes. The exact nature or meaning of the connection depends on the nature of the network. This could be a family relationship, for example, co-attendance at some kind of social event, or membership in the same organization. Sociograms will often demonstrate the weight of the tie between two nodes through the thickness of the line: this weight will usually reflect the strength of the relationship as determined by how many times it is asserted in the sources. SNA software allows users to visualize characteristics about nodes and edges through such things as size and colour. Edges can be either undirected or directed: a directed edge will have arrows demonstrating some relationship trait. The link between a king and a monastery to which he is granting a piece of land might be depicted with a directed edge pointing from the king to the monastery.

The structure of the network sheds light on the social relationships of the historical actors. In SNA jargon, two connected nodes are called a dyad and three connected nodes are called a triad. A number of nodes who are all connected to each other are termed a clique. A sociogram in which most of the nodes are connected to each other is said to have a high graph density. A graph in which all nodes are connected to all other nodes would have 100% density.

Social network analysis is built on mathematical graph theory and consequently uses algorithms to produce numerical data. The most important of these is centrality. The following centrality measurements appear in the case studies:

1. Degree Centrality is the number of nodes linked directly to any given node. In other words, it is the number of contacts an individual has in a network.
2. Betweenness Centrality considers the position of the node within the structure of a network. It is calculated by adding up how many times a node is the shortest path linking two other nodes. This is another way of saying that a node with high betweenness will occupy a strategic position within the network structure.
3. Eigenvector Centrality considers not just how many contacts a node has, but also how important those contacts are. Eigenvector centrality is often expressed as a percentage. Eigenvector offers a more nuanced way of highlighting powerful or influential actors within a network.
4. Weighted Degree is calculated by adding up all the individual links between nodes, as opposed to the number of contacts. In that sense, it is more concerned with edges (the lines between nodes) than the nodes themselves. This tells us about the strength or weakness of the connections between actors.

Use the tab on the left to explore a number of case studies which have been conducted using data from the People of Medieval Scotland database for the COTR project. The two main types of networks which have been created are relationship networks and charter witnessing networks. These have been produced for the reigns of John (1292 - 1304), Robert I (1306-29), David II (1329-71), and Edward Balliol (1331-56, in opposition). The relationship networks were constructed using the relationship factoids in the PoMS database. These are split into three categories: familial, employment, and tenurial/lordship. Networks of charter witnesses are affiliation matrices which tell us whether and how often witnesses appeared alongside one another in witness lists.

Two more complicated examples of networks have been created for some of the more specialized studies. Two-mode networks demonstrate the connections between two different types of node. In the case studies of resignations, forfeitures and escheats, two-mode networks illustrate the link between historical actors and the lands being surrendered. Colour-coding of nodes and edges allows us to visualize who lost and who gained the surrendered lands. In the sociograms of the Ragman Roll fealties made to Edward I in 1296, we have two-mode networks for those fealties made before the Berwick-upon-Tweed parliament showing the people swearing fealty and the places where they did so. For the fealties made at Berwick, there is a different kind of two-mode network showing the historical actors as one mode and each separate instrument of fealty copied in to the Ragman Roll as a numbered transaction: this allows users to visualize the people swearing fealty as clustered into numerous groups.

There are dozens of interactive Gephi sociograms available on the People of Medieval Scotland website to play around with. Enjoy!

The Social Network Analysis Researchers of the Middles Ages (SNARMA) operate a website where you can find out more about SNA and how it is being applied to medieval studies.

There is also a free e-book about SNA and the 1093-1286 portion of the PoMS website, available here.