By Edmond Smith
Early modern England contained a wide range of institutions that formed vital parts of peoples’ daily lives. These ranged from parishes and livery companies to trading companies and the developing state apparatus. The ways in which peoples’ lives can be traced through their institutional activities is vital for understanding different organisations within the complex institutional environment of early modern England.
Attempts to explore networks, webs, communities and social groups have in recent years led to an increased number of historical studies utilising networks as a key concept for analysis. Networks, in this way, ‘link individual behaviour to larger macro-level social and organisational outcomes’ which ‘help us to make sense of aggregate patterns of behaviour at the individual level and to thereby see the link between micro and macro levels’ (Emily Erikson). In turn, by understanding the relationships between individuals in the network we can start to draw conclusions about how the individual might have thought, made decisions and built relationships when only limited records exist regarding their lives. In some respects, network analysis enables the analysis of institutions ‘from below’, giving voice to the vast majority of people who left no personal records.
The application of network methodology, particularly quantitative network analysis, requires considerable reliance on some complex analytical tools. It is economists and sociologists who have developed the considerable methodological tool-kit that we can draw upon digital humanities. Here, I will briefly demonstrate the means in which evidence makes its way from the manuscript, through a database and into a network analysis.
For example, to examine the foundation of the East India Company in 1599, when 134 merchants came together to promise funds and petition the crown for a charter, we can analyse the networks of its founding members. A list of all these men is recorded in the first item in the India Office Records at the British Library – an excerpt of which is above. A cursory glance reveals much about some of these individuals, with livery company involvement and commercial partnerships clear, but collection of the data allows the compilation of a list of EIC members that can then be identified in other types of sources.
A simplified version of this information, limited to the EIC and membership in four other organisations, for ten such individuals (who were all members of the EIC in 1613) is presented in chart 1 below. In the chart, a ‘1’ indicates each investment of these EIC members in another organisation. As you can see, EIC member Maurice Abbott has been identified as a member of the Virginia Company, Levant Company and North-West Passage Company. Henry Archer, on the other hand, was not an investor in any overseas trading organisation other than the EIC. Elsewhere, through collecting data from manuscript records left by other trading organisations I have similar data for all members of the EIC between 1599 and 1630, but for brevity’s sake I will limit it to 10 here!
Although there are gaps and no other organisation has as complete records as the EIC, the use of petitions, charters, letters and remaining court books for these bodies has revealed a huge amount of data of this sort. In particular, due to the importance of keeping records of members for organisations, many materials recording this sort of information have survived while more practical materials, such as account books, have not. Indeed, many early charters for trading companies were reproduced in printed material from the period. This means that for many of the identified EIC members, the only material recoverable about their lives are their membership and participation in other organisations.
In order to visualise the network shown by this membership data, the next step is to convert, through matrix multiplication, a matrix that details the relationships between individuals. This is displayed in figure 2 below. Here, each individual has a value aligned with each other individual depending on the number of shared investments. For example, Maurice Abbott has shared two investments with Edward Allen, none with Henry Archer, and three with Hugh Hamersley. These values are the strength of the relationship between two individuals in the consequent network analysis. Of course, there are limitations to this approach, as repeated shared investments do not necessarily represent a close personal or professional relationship between two individuals. However, they do suggest possible shared experiences, shared interests, shared expertise and shared ideologies about investment and overseas enterprise. The network graph below is, therefore, a representation of the connections between individuals through shared participation in similar activities.
To create the visualisation of this network, and to run data analysis on the underlying data, a program called Gephi has been used. Gephi utilises software drawing on algorithms created by a number of sociologists to analyse a matrix and represent it as a network. To do so, it assesses shared values between different individuals (or nodes) within the network and the strength of the relationships (or edges) between them. Nodes with stronger edge strength are drawn together, creating clusters of nodes that all have similarly strong relationships with similar nodes for similar reasons. In the case of this example, clusters will develop around nodes with shared experiences. Furthermore, the Gephi analysis prioritises more influential nodes (nodes with the highest value edges) towards the centre of the network, and sends less influential nodes towards the periphery. This is particularly useful for identifying particularly influential figures within the network, who in turn might be examined as brokers between different groups. The following network, figure 3, has been developed in this method from the smaller matrix of the ten EIC members detailed above.
In addition to simply creating a visualisation of the network, Gephi also enables the analysis of the data through a number of different statistical models. In this network, two of these have been used. First, the relative strength of each node within the network has been measured by its ‘degree’, the total value of its relationships within the network, and this value has been used to adjust the size of each node. As such, Hewett Staper and Hugh Hamersley are the largest two nodes and the most central, as they share the largest number of shared experiences with the others. Henry Archer and Richard Persons are the smallest nodes and exiled to the periphery of the network, as they lack any shared investments with other nodes. This can be a particularly useful tool in larger graphs to help pick out and identify the most influential nodes.
A second tool that has been applied here, and which is vitally important for effectively analysing networks, is modularity analysis. Essentially, this is a community detection algorithm that identifies communities within a network depending on the shared relationships between different nodes. In this graph four communities were identified, one for each of the peripheral nodes, and two more each containing four of the nodes in the main network. Shown here in red or blue, these two clusters are identified through the shares investment practices of their members, and the strength of these relationships as opposed to others. For example, Edward Allen, George Tucker, Maurice Abbott and Humphrey Robinson are all drawn into a single community due to their shared experience investing in the Virginia Company and North-West Passage Company. Although simple here, the same modularity analysis has been used later in the thesis to delineate similar communities in much larger and more complex analyses (see below). This helps to demonstrate how the EIC included many members with interests and experiences that overlapped, and how different groups within the Company could form depending on the shared experiences of individuals involved.
This graph, visualising all the data for members of the EIC in 1613 (which presents 40,449 connections between 392 merchants) clearly demonstrates how investors in the company came from a diverse range of different backgrounds and had experiences across the full range of England’s global overseas activities. I won’t say more about it here, so please keep an eye out for the article!
Although limited here to shared investments, the methodology examined in this post can be used to analyse how people’s lives were affected through relationships revealed in their shared geographies and spaces, familial relationships, other institutional memberships, and could easily be expanded to include personal correspondence and other sources where they exist. Quantitative network analysis, therefore, presents an opportunity for historians to ‘peek under the hood’ of seemingly monolithic institutions, gaining an insight into the lives of people in the early modern world.