Monthly Archives: February 2018

Using the Clerical Exile Database: Benefits and Limits of Social Network Analysis

This blogpost was written by Jim Chaplain, MA student in the Department of History, University of Sheffield.

According to Andrew Prescott and Tim Hitchcock, the academic practice of history has had a difficult relationship with the ‘digital revolution’ thus far (for references see below). Historians have overall been slow in coming to terms with the changing landscape of the profession that has come about as a result of digitisation, and have almost without exception failed to use the opportunities that the modern technological landscape offers. Few digital sources, Stephen Robertson maintains, are truly digital, in so far as they effectively utilise structures and mediums made available by a computer and would cease to function once printed onto a written page. However, the database Migration of Faith: Clerical Exile in Late Antiquity goes someway in creating a digital source for historians that depends on a multi-layered digital platform, and that uses technology to examine historical problems in a pioneering way.

The database archives information on religious exiles, the location they’re exiled from and to, and by whom. Innovatively, the information is presented in two distinct ways; firstly, the data is presented on a map of the late Roman world, demonstrating each case in terms of place of residence before and after exile, which can then be altered according to chronology and according to direction of exile. These maps work to visualise the case for a Mediterranean-wide network of clerical exile, attesting to geographical concentrations in specific areas, such as the Nile delta and northern Anatolia. This focus on networks is continued in the second layer of the database, which seeks to demonstrate the social networks of individuals affected by clerical exile in the time period given. Here, data from a variety of genres of source, especially church history, hagiography and epistolography, are analysed by a computer programme in order to produce sociograms, which aim to show the personal social networks of every individual in the database.

The programme in question is designed to facilitate Social Network Analysis, which can be understood as a method of understanding social relationships, whether personal friendships, two people sharing a mutual friend or open hostility with one another. Populations are visualised through ‘nodes’, which signify individuals, and ‘edges’ which signify the ties between them. Aforementioned social relationships are understood in terms of ‘strong ties’, for example between two close friends, and ‘weak ties’ of the sort between two people who are distantly acquainted. For example, this sociogram of Martinus of Tours shows him characterised as a “node” on the sociogram with strong ties to three other individuals (including the Bishop of Milan, who has exiled him), along with weaker ties to a subgroup that has Hilarius, the Bishop of Poitiers, at the centre.

Illustration 1: Network sociogram of Martinus of Tours

Source: ‘Migration of Faith’, /sites/clericalexile/ [accessed 25 November 2017].

As Tom Brughmans shows Social Network Analysis has a long history in the social sciences, where it has been used as a means of visualising large ‘interrelated’ populations in order to analyse the flow of ‘resources’, in our case theological ideas, within them. Digital software is responsible for the creation of the sociogram according to the data in the sources, meaning that the formation of subgroups on the graph is due entirely to the data, such as how often characters interact in the narrative, and is not inputted by an historian; often these groupings are as expected but this is not always the case, which can lead to new interpretations of their roles within the network. The software also quantifies how well connected each individual is, according to their ‘centrality’ on the sociogram, which can be examined alongside the grouping of nodes in order to establish fresh conclusions. Julia Hillner, the PI of this project, has used just this connection to establish the peripheral role the emperors play in much of the literary sources concerning exile. The fact Social Network Analysis is able to render data that can be interpreted by a historian as a new understanding of the source material is integral.

The strength of this network-focussed approach to historical research, and Social Network Analysis especially, is not only that it sheds light on new information, but rather that it allows the historian to view sources that have in some cases been read for over a millennium in an entirely new way, which prompts new interpretations and new understandings. This constitutes a change in usage from the ways in which the technique has been used in the social sciences, for example by Knox, Savage and Harvey, where it has been used to analyse an existing or ‘real’ population, often to visualise structural hierarchies and interpret the behaviour of actual social groups. For the historian, as Hillner stresses, datamining historical sources and presenting the data instead offers a means not of analysing a social group in the social scientific sense, but as it is presented in the written record. It is integral that the historian views the sociogram not as an actual social network, nor even as a representation of a social network, but rather as a representation of a representation of one. Networks, when visualised on the Clerical Exile database, offers the researcher a ‘discursive gap’, as it does not look to offer quantifiable data like a graph or chart, nor does it offer qualitative data, instead, to use a term coined by Knox, becoming ‘both referent and representation’. This gap is essential in avoiding the pitfalls that often come with Big Data approaches to history, with reliance on data over literary, semantic interpretation.

However, caution is still required on the part of the historian as alongside the opportunity to view a source in a different light, it requires a considered and precise methodology in order to interpret the data in a meaningful way. This is potentially problematic, as the adoption of Social Network Theory by historians is a very recent phenomenon and a correct methodology has therefore yet to be fully established, or a single method widely accepted. However, the Clerical Exile database forwards a methodology that understands clerical exile itself as a construct, and a product of a literary milieu where religious authority was contested as the early church establishes itself, which allows it to be used in order to understand the construct and not the reality. In the few comparable studies that have adopted Social Network Analysis, this methodology seems to have been the most successfully applied. Mairin MacCarron uses such an approach to contribute to the debate over the Venerable Bede’s treatment of women in his writings, contradicting prevailing scholarly opinion that he deliberately limited the role of women to show that in fact women play a more prominent role in his writings than in most Anglo-Saxon authors’ work, complete with accompanying sociogram demonstrating the centrality and connectedness of women in Bede’s work compared to others.

The Clerical Exile database can be put to similar use in the case of Saint Martinus of Tours. There has been debate in the historiography of early Christian Gaul concerning the historical Martin, who is renowned as having been responsible for the conversion of the province to Christianity. Some, like van Dam, have doubted his importance in this process which, so the argument goes, is due to his prominence in the late ancient Gallic sources and the emphasis that Sulpicius Severus and, later, Gregory of Tours places on him, both of whom having ulterior motives for doing so. In his place, the influence of the aristocratic Saint Hilarius of Poitiers has been stressed, placing him at the heart of Gallic Christianisation. When consulting the social networks of the two individuals, the difference between the two contemporaries is striking.  The network of Martinus is fairly limited, and has only three strong ties, one of which leads to Hilarius, while Hilarius’ network is far more complex, with more nodes around several distinct subgroups, and a web of edges connecting disparate people. Along with having more strong ties than Martinus does, Hilarius also has far more weak ties, which according to network analysists such as Everton are integral to the widespread diffusion of information across a network beyond an immediate group. Martinus possesses weak ties too, however all of them are ‘brokered’ by Hilarius, a position that allows an individual to control the flow of information between different groups. By applying the information provided and presented by the database, it is therefore possible to demonstrate Hilarius as being part of a far wider reaching Christian social network, and therefore in a much more important position for the dissemination of Christianity across an area the size of Gaul, than Martinus, the man traditionally acknowledged as having done so.

Illustration 2: Network Sociogram of Hilarius of Poitiers.

 

Source: https://www.dhi.ac.uk/clericalexile/network/person/interaction/40 [accessed 29 December 2017].

It must be acknowledged at this point that Martinus’ relative isolation is to some degree to be expected, insofar as his spiritual power was derived largely from his asceticism, his aversion to material wealth and his low birth, in contrast to the aristocratic standing of those who came to dominate Gallic bishoprics, Hilarius included, who are emphasised in the sources from which the data is drawn. Martinus’ relatively modest social network can therefore not be equated with a lack of spiritual power, as many of his associates would have been of too low birth to have been deemed worthy of recording, yet the sociograms show how the two people are represented in the written sources, so it is revealing that even there, where Martinus is lauded as a hero, he is presented as isolated. This is surely to be expected, given that Martinus was a man without senatorial rank, in a province that was not his own; it is no surprise that the Clerical Exile database finds him isolated, but the extent to which this is so surely suggests Hilarius was by far the more influential figure in the conversion of proto-Christian Gaul, even if Martinus gets more literary attention.

However, at this stage any conclusion drawn from a network database such as Clerical Exile needs to be heavily prefaced by the fact that methodological issues still abound, to the extent the very usefulness of this exercise, and this way of thinking about the past and the literary record, is questioned even by the few people who are pursuing it (including, as Hillner’s work shows, the project team of The Migration of Faith). The fact it involves borrowing methodology from sociology and even physics, which is a worthwhile exercise as an interdisciplinary practise, also risks losing sight of the historicity of the network that is being analysed, and involves entering an already muddled field where Social Network Analysis is developed in different academic fields potentially, as Brughmans argues, in ignorance of each other. The above analysis of Martinus of Tours and Hilarius of Poitiers demonstrates at once the potential of the database, and Social Network Analysis in general, alongside the pitfalls. Historians who have argued in favour of an interpretation of Hilarius as an influential Christian leader have done so on the basis of his connectedness and wide circle of friends, while Martinus of Tours has traditionally been viewed as an exceptional character unlike the other bishops and church leaders of his day. Both of these interpretations can be supported by analysis of the sociogram, which demonstrates the ability of the Clerical Exile database to substantiate the work of historians independently. However, as Hillner warns, it runs the risk of focussing too strongly on networks that may only have existed in literary sources, at the expense of all other possible motivations and causal factors for an event. As stated above, just because Hilarius was potentially better connected and more influential according to his Social Network centrality, this does not mean that in reality he was empirically more influential in the spread of Christianity than Martinus, only that he may have been, had he wanted to be.

A further issue with the use of Social Network Analysis is that it relies on expertise that is markedly different from that which the academy prepares historians with; a highly successful example to-date of the application of Social Network Analysis on a medieval source was carried out by two sociophysicists, Kenna and MacCarron. Indeed, further reading would be required in order for the average historian to make the Clerical Exile database useful due to the complicated nature of the visualisations that it produces. As such, many of the issues of a ‘big data’ approach identified by Tim Hitchcock are visible in the Clerical Exile database, as it creates a kind of history that requires advanced technological capability to understand, along with dehumanising historical experience in the desire to evidence findings through datamining software and hard data.

The use of digital technology alongside a historical methodology that might fairly be described as still experimental has created a source that will be of great interest to historians interested of the period of late antiquity and to anyone researching the dissemination of knowledge and the ways in which historical networks are presented in written sources. Clerical Exile is justified in its use of a ‘semantic approach’, focussing on networks as imagined by literary sources over the ‘hermeneutical’, focussing on ‘real’ networks. It that way it facilitates fresh interpretation of hagiography and epistolography. There are still issues to be resolved, as we have seen, as the problems of big data digital history persist in this project, but the way Social Network Analysis is used and thought of in this project goes a long way in pointing the way forward for future uses of the technique.

References:

Brughmans, T., ‘Thinking Through Networks: A Review of Formal Network Methods in Archaeology’, Journal of Archaeological Method and Theory, 20.4, (2013) pp.623-655

Everton, S.F., Disrupting Dark Networks, (Cambridge, 2012)

Gramsch, R., MacCarron, M., MacCarron, P., and Yose, J., ‘Medieval Historical, Hagiographical and Biographical Networks’, in Kenna, R., MacCarron, M., and MacCarron P., (eds), Maths Meets Myths, (New York, 2016), pp.46-68

Hillner, J. D., ‘Approaches to Clerical Exile in Late Antiquity: Strategies, Experiences, Memories and Social Networks’, in Hillner, J. D., Ulrich, J., and Engberb, J., (eds), Clerical Exile in Late Antiquity (Frankfurt am Main, 2016), pp.11-46

Hitchcock, T., ‘Confronting the Digital or how academic history lost the plot’, Cultural and Social History, 10.1, (2013), pp.9-23

Hitchcock, T., ‘Academic History Writing and the Headache of Big Data

Kenna, R., and MacCarron, P., ‘Maths Meets Myths’, Physics World, Vol.29(6), (June 2016), pp.22-27.

Knox, H., Savage M., and Harvey, P., ‘Social Networks and the Study of Relations’, Economy and Society, 35.1, pp.113-140

Prescott, A. ‘Consumers, Creators or Commentators? Problems of audience and mission in the digital humanities’, Arts and Humanities in Higher Education, 11, 1-2, (2012), pp.61-75.

Robertson, S., ‘What’s Wrong with Online Readings?’ The History Teacher, 39, (2006), pp.441-454

Van Dam, R., Leadership and Community in Late Antique Gaul, (Los Angeles and Oxford, 1985)