For more than fifty years Henry Parkes (1815-1896) was an iconic figure in the political life of colonial Australia. A stalwart of liberalism, he was five times Premier of New South Wales and was regarded at the time as the “Father of the Federation'' of the Australian Colonies. Parkes emerged as a political force in the 1850s as a member extensive socio- political network in Sydney. Some of the nodes of that network have been examined - especially in biographical studies of individuals (see Australian Dictionary of Biography, https://adb.anu.edu.au/) – but the intricate structure of the relationships and characteristics that connected them have received little detailed attention. In this paper, we report on a preliminary data science examination of a prosopography of this network, as the first tranche in a capacious and multifaceted project entitled “Liberal Sydney”.
Our aim is to mine extensive digital datasets of historical documents using a combination of close reading and computational methods to answer questions about the political landscape in colonial Australia and the wider Asia-Pacific region from the 1840s. Named entity recognition (NER) tools and the Linked Data publication paradigm (following workflows from Nurmikko-Fuller, 2023) enable us to gain insight into the origins and development of political, social and ethical ideas under the rubric of ‘liberalism’. These core ideas have profoundly influenced the trajectory of Australian politics up to and including the formation of the Liberal Party (the Australian Conservatives) in 1944, which has dominated the national government of Australia for the past eighty years and played an important role in shaping the Asia-Pacific region.
The project draws upon a vast archive of physical primary sources – newspapers, letters, diaries, pamphlets, books, memoirs, public speeches, parliamentary debates, court transcripts and other official records. There is also a plethora of cultural material: posters, banners, and flags as well as plays, poems, songs, and novels.
At the time of writing, a significant percentage of the material has been digitised: the online British Newspaper Archive, for example, comprises 72,940,133 pages of text ( https://www.britishnewspaperarchive.co.uk/help/about). The National Library of Australia’s Trove site offers 14 billion digital items ( https://trove.nla.gov.au/about/what-trove/history). OCR scans of books and pamphlets no longer in print are increasingly available in Google Books or through non-profit sites such as Internet Archive (which currently contains 41 million books and text). Resources, such as Gale’s massive Nineteenth-Century Documents Online series, are also invaluable ( https://www.gale.com/intl/primary-sources/nineteenth- century-collections-online).
Yet, an equally important corpus of research material remains undigitized. Part of it has been recorded in an analogue hyper-text system (Nurmikko-Fuller & Pickering, 2021) designed and implemented for several pervious projects (Pickering 1995, 2001, 2008, 2013; Pickering and Tyrrell, 2000). The system contains hundreds of thousands of granular data points accessible via interlocking indexes and underpinned by the assignment of unique identifiers. It is much more than a simple index to resources, constituting information curation, and the ‘extraction of relevant content in response to specific research questions, themselves directed by intellectual desire and curiosity’.
Our initial proof-of-concept, both for the project’s potential contribution to knowledge and the robustness and sagacity of its research infrastructure, was a study of the guests (including Parkes), invited by the Lord Mayor of Sydney to attend an annual costume ball between 1857 and 1879 (Gatti, et al, 2022). This involved the translation of digitised historical ephemera though several data formats and methods (e.g. PDF, OCR, and RDF), as well as the development of a custom-built user interface to query of the underlying knowledge graph without the explicit need for query languages such as SPARQL. Costume balls could be dismissed as meaningless diversions from the serious day-to-day business of politics with little or no historical significance. In fact, they provide an important lens onto contemporary society. Indeed, as the editor of the Sydney Morning Herald put it, the events ‘make perceptible the true spirit of the time’ (9 July 1853). Moreover, the annual guest lists were carefully drawn-up under the supervision of the incumbent Mayor and, as such, in many ways the Costume Balls reflect a political network. In 1857 and 1879, they provide snapshots of liberal Sydney. The successful conclusion of the proof-of-concept gave us confidence in the importance and potential of the project per se in terms of knowledge creation and that our methodological framework appears to be fit-for-purpose.
More recently, we have begun to establish a comprehensive taxonomy of terms to represent the ideas captured in the discourse between committee members. However, an early examination and the use of tools, such as the Gale Digital Scholar Lab’s NER tool highlighted several issues. First was the discrepancies between the ideas conveyed and the words used in primary sources. Moreover, unsurprisingly, many terms, which are relevant today did not form part of the common vernacular in Sydney 150 years ago, and thus cannot readily be found using character-matching searches. Similarly, it goes without saying that terms familiar today have undergone multiple accretions of meaning between then and now. In methodological terms, our project provides considerable scope to explore these issues. Indeed, one of the project aims is to map structured ways to integrate domain expertise and results generated via character matching.
Subsequent investigations into the economic and administrative development of the colonies have enabled us to better understand the socio-political networks around and involving Parkes. For example, our initial examination of the census data for New South Wales spanning 1833 to 1901 demonstrated the rapidly increasing complexity of the economy, land-use and employment profile of colonial Australia, as well as the more granular and detailed approaches to record-keeping that emerged over that time span. Completely anonymised at the point of collection however, historical census data cannot provide details for individuals that formed part of the prosopography of Parkes’ committees, but it gives us an appreciation of the social landscape and urban development of the area in which it flourished. Moreover, there were successive trade directories for Sydney that were published as commercial items covering these years. In additional to the fact that they are ipso facto not anonymised they allow us to closely match patterns of trade, business and location with the relevant patterns and categories in the census data. They are available as searchable PDFs from several institutional repositories and digital libraries. Our next step is to examine newspapers not only for the articles and extensive lists of committees and boards, public and private interest groups, etc., but for the advertisements – these hold a promise of copious amounts of implicit information.