On June 29th the Quill Project hosted their second The Oxford Research Centre in Humanities’ (TORCH) Negotiated Texts Network seminar at Pembroke College, Oxford. The event brought together interested researchers from different departments and disciplines to listen to our two guest speakers, and to discuss the possibility of future collaboration. The invited speakers, Dr Louise Thompson and Dr Luke Blaxill, took very different angles on the subject of political speech, which sparked a range of thought-provoking questions from the floor: It was clear from the afternoon’s events that digital humanities have a crucial role to play in the development of knowledge in this field.
Dr Alfie Abdul-Rahman, our visualization expert, opened the afternoon by welcoming the attendees and introducing our first speaker, Dr Louise Thompson, a lecturer in British Politics at the University of Surrey. Dr Thompson proceeded to presented a paper entitled “Using Parliamentary Material for Research and Teaching”, in which she surmised that, with thousands of legislative documents now available online, the UK Parliament has never been as transparent or as accessible as it is today. Changes to Parliament’s website, investment in public engagement strategies, and interest in big data initiatives over the last decade have brought positive changes to the way in which Hansard texts, bill documents and committee reports are published online. However, the use of parliamentary material continues to pose challenges for both academics and members of the public alike. Using her students’ experiences as an anecdotal case study, and drawing upon her experience of researching legislative scrutiny in the UK Parliament over the last decade, she demonstrated how the Parliament website is neither intuitive nor user-friendly, and therefore hinders research into the website’s content.
Dr Thompson explained that these difficulties could be attributed to a number of reasons: First, that the amount of data and information encapsulated within Parliamentary records is a great challenge to organize and display; Second, the sheer volume of data subsequently displayed on the Parliamentary website means that it is slow to adopt change, and that there is a lack of consistency in the formatting and presentation of legislative documents. Furthermore, the structure of the website, which reflects the bicameral nature of parliament, does not effectively illustrate the interaction between different legislative branches; the records often only include top-level information that does not effectively demonstrate the functions of the institutions.
The website therefore appears to be inaccessible and elitist, understandable only to those who have enough pre-existing knowledge to know how to navigate the site in order to obtain the desired information. Dr Thompson went on to suggest that this conclusion was also symptomatic of the apoliticism of the Parliament website being taken too far - no opinions or explanations are provided which might provide context or other useful insights into the parliamentary process.
As well as presenting on the shortcomings of the Parliament website, Dr Thompson also highlighted some of the key challenges for researchers posed by the information provided by Parliamentary sources themselves. Among other problems, official records of parliamentary events contain factual errors, documents are often duplicated, and the complexity of document layout, alongside an equivocal presentation of the relationships between respective documents, makes the tracking of amendments and changes to bills extremely difficult to follow. These failings hinder academic analysis. Similarly, and unfortunately for Parliament, the positive aspects of the parliamentary process, such as its productivity and other achievements that may not have altered the final text, are often lost in translation. This may conceivably have a knock-on effect when it comes to instilling confidence into the public regarding the efficacy of their democratic institutions.
Within a critical talk, Dr Thompson made sure to make mention of the improvements made by the website over time; visual displays of data were increasingly common, and while change is slow, an improvement to the site’s accessibility is being seen.
After a short break for cake and lively conversation between attending guests, Dr Abdul-Rahman introduced our second invited speaker, Dr Luke Blaxill, a Leverhulme Early Career Research Fellow at the Anglia Ruskin University. His talk, entitled “A War of Words? Text Mining Political Speeches in Britain in the 19th and 20th Centuries”, explored one possible solution to the “data deluge” faced by historians: Computational analysis of political linguistics.
Dr Blaxill explained that historians are increasingly surrounded by an ever-growing forest of machine-readable textual sources but that, despite this, the impact of text mining in History has been remarkably weak; historians tend to focus on analysis based on methodologies which rely on close readings of small corpora. Macroscopic computational approaches based on large corpora remain at the fringes of modern research, despite the traditional barrier of cost and manpower being considerably mitigated by the march of technology. His paper investigates the application of computational text mining in two areas of political language where large corpora have become available, arguing that such methods allow historians to analyse texts that are too large in size to read, to help overcome the inherent flaws in human ability to intuitively estimate frequency, and to allow greater verifiability, more precise communication of quantity, and a more empirical approach to working.
The first corpora is that of election platform speeches in the late Victorian and Edwardian era. This relatively small corpus (of approximately four million words), created by local constituency candidates holding scores of public meetings of large audiences during any given election campaign, was reported consistently by an attentive press. Dr Blaxill argued that his research using computational analysis of this corpora challenges historical consensus on the contents of general election campaigns, on the significance of issues such as imperialism and Irish Home Rule, and the respective visibility of party leaders such as Gladstone and Disraeli.
The second example is an analysis of the language of female MPs in Parliament since 1945. Drawing upon the outputs of the Digging into Linked Parliamentary Data (“Dilipad”) project – which has added gender, party, constituency, and hierarchical position coding to the digital edition of Hansard – Dr Blaxill presented an empirical analysis of the role of gender in the 677 million words of Commons debates from 1945 to 2015, using this to both support and dispute commonly-held feminist assumptions on political speech.
After clarifying some queries from the floor about the settings of such empiricism (including on how categorizations of under certain topic headings or as either “male” or “female” were made), Dr Blaxill showed that many feminist assumptions concerning gender-specific speech content are in fact corroborated by empirical analysis: Women's contributions to Commons debates are indeed substantively different to those of men. Questions of a more dynamic nature, which require analysis of changes in data trends over time, showed more thought-provoking results: For example, that there has been a marked increase in male contributions to topics which have been identified as predominantly led by female politicians, but not vice versa, a conclusion which stands opposed to traditional presuppositions of gendered discourse over time.
Closing his presentation, Dr Blaxill warned that careless use of this type of empirical analysis and similar methodologies can result in poor quality research, and that any temptations to overweigh superficially interesting or useful results should be avoided. The techniques explained solve many of the problems posed by the volume of information presently available to historians, and have great potential to reshape researchers’ work in the digital age, whether used in combination with traditional methodologies, or as a means of leading analysis.
The seminar was brought to a close by Dr Abdul-Rahman, but enthused conversations continued over more tea and cake, and the day concluded with some welcome additions to the Quill mailing list. The next TORCH Negotiated Texts Network event will take place next term, where we are sure that the potential of progressive research techniques in the digital humanities will continue to be a central theme.