Adventures in Prosopography
Large-scale digitisation has opened up a wealth of resources for historians, both amateur or professional. Indeed, amateur genealogy has become one of the most popular pastimes. While many simply research their own families — often with great skill and commitment — others are contributing to larger crowd-sourcing efforts to reconstruct wider communities. A notable example is Ed Pope’s tireless search for information to complete our knowledge of the people named in William Godwin’s diaries.
I’m working on a paper on propertied reformers in nineteenth century Ireland — people forgotten largely because they maintained no personal archives, and whose descendants emigrated. There is also a wilful loss of memory through a process of social forgetting, as theorised by the historian Guy Beiner. I want to understand why constitutional nationalism failed in the nineteenth century, and how such failure related to the brutal policy disasters experienced by/in modern Ireland.
Because most of the figures I am researching have not featured in academic accounts, I’ve turned to whatever sources I can find to generate a picture of their social world and their activism. One inspiration is the magisterial biography of William Sharman Crawford and his family by Queen’s University Belfast historian Peter Gray. Gray analyses the rise of the Sharman Crawfords as a family of reformers. This tradition of Ulster radicalism was neglected too long, and Sharman Crawford nearly forgotten, until Gray’s comprehensive and ground-breaking work.
What Gray’s absorbing text shows is that social forgetting cannot be entirely complete. Digitisation and open or low-cost access makes it possible to recover what was not memorialised. Sources we can use include church records, the major land surveys, the Gentleman’s and Citizen’s Almanac, Thom’s Directory, the contemporary press, parliamentary records, legal cases, occasional memoirs and oral histories, and private archives.
My use of sources in combination in this way is therefore prosopographical: beginning with a name or event, tracking the individuals and events connected to the original node, and compiling as much information as is available for them.
The celebrated social historian Lawrence Stone claimed that ‘prosopography’ is simply a term borrowed from ancient historians — equivalent to modern historians’ ‘collective biography’ and social scientists’ ‘multiple career-line analysis’. Yet he, a ‘new’ social historian, accepted the more esoteric term, defining it as:
‘the investigation of the common background characteristics of a group of actors in history by means of a collective study of their lives. The method employed is to establish a universe to be studied, and then to ask a set of uniform questions about birth and death, marriage and family, social origins and inherited economic position, place of residence, education, amount and source of personal wealth, occupation, religion, experience of office, and so on’ (Stone 1971, 46).
If such data existed completely and accessibly, our focus would be on the analysis itself, rather than how to compile them. So, the value of prosopography lies in the fact that identifying lost groups demands extreme patience, as well as creativity and judgement. There is so much that is missing, and it is not always obvious whether there is anything there at all to be found.
We have to be wary of confirmation bias, working in reflective mode to judge whether we really have enough evidence to judge that a J. M’Donnell in 1842 is the same person as a John McDonnell in 1853. We have to question constantly whether we are simply trying to generate a good story where none exists. It may that we become excessively-beguiled by the quest, deep-diving into historical worlds of no interest oustide our own search.
Some years ago I assessed composers’ networks by using a selection of data on network connections recorded in the New Grove Dictionary of Music and Musicians, now available via the Oxford Music Online portal. I and my co-author Martin Everett argued that the styles characterising modern British classical music were a product of collective invention (McAndrew and Everett 2015). For that paper, data collection was more straightforward: the network was defined by who had an entry of their own, and the Grove provided dates of birth and death.
But where a high-quality directory has not defined the network or population for us, we have to work harder to determine who the members are, what we can know about them, and who should count. It is up to us to define the community, before reconstructing the lost detail, following principles summarised by historian Dion Smythe:
‘record accurately; record as much as possible; always reference any piece of information to the source from which it comes; be clear on separating the material that comes from primary sources and what comes from secondary analysis; be aware of the dangers of fusion (joining two individuals together) and fission (splitting a single individual into two records); favour fission over fusion’ (Smythe as paraphrased by Keats-Rohan 2007, xviii).
Sociologists have argued that we cannot do good social science using such methods — we are attempting to reconstruct social reality from ‘the relics’, which are necessarily partial and highly-selected (Goldthorpe 1991). This is fair. However, my own objectives are modest: to demonstrate that a particular political movement existed, and that important initiatives were attempted, without making grand claims regarding their wider effects.
We can mitigate risks of bad sociology by being precise regarding potential causal mechanisms, and conservative in our inference. And in any case, rather than harnessing historical data in the interests of sociology, we can harness sociological concepts and methods in the interest of better history.
Dylan Connor’s extraordinary analysis of two dimensions of cultural identification in Irish first names recorded in the 1911 Census — one relating to traditionalism, a second to Catholicity — shows what is nevertheless possible. While we are indeed dealing with ‘the relics’, the relics are becoming more plentiful, and can potentially be analysed at scale using computational methods. The economic historian Neil Cummins is exploiting digitised genealogical data to answer major questions relating to modernisation and social change. Topic modelling of the text in historical newspapers is a further example.
Additional opportunity is provided by social network analysis. Goldthorpe argues that sociologists differ from historians because they can generate evidence: ‘[it is] such generated evidence, rather than evidence in the form of relics — in other words, evidence that is ‘invented’ rather than evidence that is discovered — that constitutes the main empirical foundations of modern sociology’ (Goldthorpe 1991, 214). This makes a great deal of sense; but we can also generate evidence from the relics by identifying connections between individuals and events, both in the ways that historians always have, and via formal methods. A new paper by Evan Bourke and Deirdre Nic Chárthaigh has identified patronage networks supporting bardic poets in early modern Ireland, extracting new insights from Gaelic sources often perceived as fragmentary — ‘bring[ing] to light those people and connections that are not immediately apparent in the larger network, and who may not have been the subject of previous scholarship’ (Bourke and Nic Chárthaigh 2023, 940).
Obvious problems regarding the ‘relics’ include source and data loss, inconsistencies in spelling, source illegibility, and low diversity of first names generating ambiguity as to who the individuals really are. Cursive handwriting is a challenge to read, particularly where letters are crossed. For digitised newspapers, the quality of the image as well as the quality of the original print may be far from perfect. Even where disambiguating the names within a family and townland is not the issue, it can still be difficult to identify people: I spent weeks assuming that Robert Owen, a director of the Farmers’ Estate Society, was the Robert Owen.
For historians of Ireland tracking individuals over time, spellings for individuals and townlands also vary considerably. Names entered into the church records can also differ from those in the civil records. The townland where my family lived and worked for over a century was spelt twelve different ways in church and civil records for births, deaths and marriages between 1864 and 1881.
It is more straightforward to trace what happened to people if they stayed in Ireland and died after the advent of civil registration in 1864. It is much harder if they emigrated. I attempted to track a woman born with a relatively unusual name for the 19th century — Penelope Mullarky.
Penelope Mullarky was godmother to my great-great-great uncle Alick (baptised John), who died in infancy. Penelope married Pat McAndrew in 1865, who was perhaps my great-great-great grandfather’s brother. Their daughter Mary was baptised in 1866. But after Alick’s christening in 1874, there is no trace of them in the church or civil records. Neither can Penelope, Pat or Mary be found in the 1901 Census, and there are no death certificates. Perhaps they emigrated — I can’t tell. The time I spent searching for her was not well-used, though I’m now expert in the surprising prevalence of ‘Penelope’ as a first name in the West of Ireland.
With other families, particularly the propertied, we can do better. One great scholarly investment is the Landed Estates database developed and hosted by the University of Galway, with details on nearly 4,000 landed families in Connacht, Munster and Donegal, Monaghan and Cavan, and their associated estates. The propertied are more likely to be found named in directories and almanacs, in the contemporary press, the Calendars of Wills and Administrations, Registry of Deeds and so on.
Of course, where the historical evidence base has been built up in this way, it is very likely to be partial: highly-selected, prone to missingness and sparsity, and with low generalisability. We will only know a sliver of the real social network, and have to take care over what we can claim.
But without such methods, it’s very likely that we would just know nothing about swathes of nineteenth-century Irish society before the advent of large-scale official data and social surveys, more systematic archiving by organisations and individuals, and projects such as the Irish Folklore Commission. The nineteenth century was a period of catastrophic failures with a long reach into the present. ‘Finding out’ brings great and immediate gratification. And more importantly, understanding how reform movements and their leaders failed, as well as where progress was hard-won, helps us make better progress now.
References
Guy Beiner (2018), Forgetful Remembrance: Social Forgetting and Vernacular Historiography of a Rebellion in Ulster (Oxford University Press).
Evan Bourke and Deirdre Nic Chárthaigh (2023), ‘Patronage Networks in Gaelic Ireland ca. 1541–ca. 1660’, Renaissance Quarterly, 76/3: 938–979. doi:10.1017/rqx.2022.436
Dylan Shane Connor (2021), ‘In the Name of the Father? Fertility, Religion, and Child Naming in the Demographic Transition’, Demography, 58/5: 1793–1815. doi: https://doi.org/10.1215/00703370-9427093
John Goldthorpe (1991), ‘The Uses of History in Sociology: Reflections on Some Recent Tendencies’, British Journal of Sociology 42(2), 211–230. https://www.jstor.org/stable/590368?seq=20
Peter Gray (2023), William Sharman Crawford and Ulster Radicalism (University College Dublin Press).
Katharine Keats-Rohan (ed., 2007), Prosopography Approaches and Applications: A Handbook (Unit for Prosopographical Research, Linacre College, University of Oxford).
Siobhan McAndrew and Martin Everett (2015), ‘Music as Collective Invention: A Social Network Analysis of Composers’, Cultural Sociology, 9(1): 56–80. https://doi.org/10.1177/1749975514542486
Lawrence Stone (1971), ‘Prosopography’, Daedalus 100/1: 46–79. http://www.jstor.org/stable/20023990.
