Recent Presentations

The emergence and rapid growth of Digital Humanities (DH) is a crucial development in interdisciplinary research that is nonetheless impeded by the incommensurable definitions and methodologies of its intersecting domains of arts and humanities, and information and computer science. Divergent definitions include authorship, language, and text, while methodologically the hermeneutic tradition of text interpretation clashes with the method-oriented research strategies of the sciences. It is often proposed that we overcome these differences by embracing the innovation of hybrid methodologies. Instead, I maintain that meeting the impediments to interdisciplinarity requires us to strengthen rather than disavow disciplinary boundaries. I make this argument with respect to recent developments in Natural Language Processing (NLP), and the ascendency of Large Language Models (LLM). The way that these developments replace formal, rule-based approaches to language with statistical and probabilistic ones has led to their over-determined equivalence with humanities understandings of contextual meaning inclusive of materiality, social and historical context as well as the subjectivity of the reader and reading communities. Yet there is still a very substantial difference between these modes, not least of all the emphasis in accounts of language in the information and computer sciences on semantic (as opposed to syntactic) meaning as word concordance (Skelac and Jandrić, 2020, p. 41). In this paper I argue for an approach that does not remove, but rather works at disciplinary boundaries as a means of grounding and normalising concepts before and as they move between disciplines.

Digital mapping is becoming an increasingly common tool for historical research in Australia, providing historians with new ways of visualising and representing the past. How can historians use the tools, methods, and outputs of digital humanities to gain new insights into the Australian past? How can these tools be used to tell accessible stories of space, place, and Country? What kinds of sources do historians require to produce these histories? What skills are required and how can historians learn them? And are there political or ethical considerations when mapping and representing the Australian past in this way? 

In this seminar, Mike Jones (ANU), Fiannuala Morgan (ANU), Bill Pascoe (Melbourne), and Emma Thomas (UNSW) will reflect on the use of digital mapping in their work and discuss the promises and pitfalls of these methods. This is a rare opportunity to see how Australian historians are using the tools of digital humanities to investigate deep time, bushfires, 'blackbirding' in the Pacific, and frontier massacres, and to discuss the questions, research, technology, and skills that underlie these kinds of outputs. 

Recent advances in Natural Language Processing software as well as theaccessibility of digital mapping software have dramatically expanded thepossibilities for spatial analysis in the humanities. The analysis of place-names in large-scale corpora has been theorised as a ‘geographicimaginary’ and an ‘imagined geography’; cartographic models thatrespectively reflect geographies of cultural significance, and construct andmediate the readers’ understanding of space. From the outset, thisapproach attracted criticism for its reductionism and abstraction of thetext. For literary studies, it is argued that maps as a reduced, approximate,and symbolic representation of reality work only in the domain of space andcannot approximate the complex representations of place found inliterature. What remains absent from these debates, however, isconsideration of how cartographic approaches enable scholars to advancehistorical spatial claims based on a presumed equivalency between textualworlds and reality. In this presentation, I draw attention to thisphenomenon as a means of returning to some more fundamental questionsin the spatial humanities. Namely, what do we seek to achieve when wemap textual sources, and more importantly, what do we create? I draw onrecent theoretical scholarship on Topic Modeling, a machine-learningtechnique that groups thematically related documents, to conceive of acomputational cartographic approach as productive of a constellation of‘latent geographic associations’ that ultimately facilitate diverse and non-singular interpretations. In this presentation I apply this framework to mycurrent research mapping the locations of bushfires published in journalistic accounts and serialised fiction in 19th Century Australiannewspapers as a means of exploring the affordances and limitations thatdigital mapping yields for historical and literary analysis

The theoretical over-turning of the doctrine of ‘Terra Nullius’ in Mabo v Queensland (no 2) both aligned with and consolidated a tradition in Australian literary theory and cultural studies that emphasises unsettlement, anxiety and (un)belonging in settler Australian colonial literature. Based on my own research into 19th century settler Australian bushfire narratives, I consider how an ecological perspective, underpinned by an acknowledgment of Indigenous agricultural practice, has the potential to problematise this reading. What would it mean, for example, to read the ‘Bush Tradition’ or The Australian Gothic as records of ecological decline, rather than simply reflections of the settler Australian psyche? Through a critical consideration of shifting definitions of ‘disaster’ in settler Australian literature and history, this paper looks towards alternative ways of conceptualising settler representations and interactions with land, landscape and property that neither adhere to nor oppose notions of (un)belonging and anxiety. 

In this seminar I present my research process mapping the locations of bushfires in Australian colonial fiction and newspapers. This includes a critical overview of the use of Natural Language Processing Software, its limitations, and affordances, as well as consideration of the kinds of questions most amenable to the use of this tool. Further, I outline the use of Trove, an Australian online library database aggregator and service which includes full text documents and digital images, in conducting historical research. This includes discussion of harvesting data, making use of available meta-data, and data quality considerations. In so doing, I intend to bring to light some of the difficulties in conducting large scale place-based analysis as well as outlining alternative methodological approaches to place-based questions. As only a recent practitioner of Digital Humanities based research, this seminar intends to balance theory and practice with practical direction for moving towards digital research.