Famine, as John Walter writes, was “…a recurring reality and ever-present fear” in the early modern period (Walter, 2015). The AHRC-funded project Famine and Dearth in India and Britain, 1550-1800: Connected Cultural Histories of Food Security ( http://humanities.exeter.ac.uk/english/research/projects/famine/) examines the practices, discourses and literary modes through which societies in early modern India and Britain articulated their concerns about the availability and distribution of food (Mukherjee, 2014). A collaboration between the University of Exeter and Jadavpur University, Calcutta, the project draws upon a large body of texts written in languages including English, Latin, Persian, Bengali and Hindi for evidence about cultural responses to landscapes of famine and dearth. Its aim is to produce a digital resource that includes encoded extracts from the source materials and maps that reflect the variety and scope of identified responses to the landscapes encountered in the sources.

The application of digital methods to large corpora of thematic texts presents both opportunities and challenges, which will be examined in this work-in-progress paper. The dataset is highly diverse, not only in terms of the languages in which these texts are written but also in the types of documents that form our body of evidence. The source materials include chronicle histories, gazetteers, official correspondence, legislation, pamphlets, periodicals, plays, poetry, surveys and prose (fiction and non-fiction), and the project will publish excerpts from each of these categories. Our markup priorities (beyond the basic structure of the document) lie in how to encode the many themes surrounding famine and dearth that are present in our texts, which we need to extract in order to address the research questions of the project. We can use natural language processing tools to help us identify names and places, for instance (at least for some of the languages), but the range of terms used in these texts to describe the features of the landscape, the people, and the food situation are extremely varied. Inevitably there are also subtleties in the ways in which particular concepts are represented in the various languages of our source materials, as well as added complications such as variant spellings. The project uses the open source software GATE (General Architecture for Text Engineering, http://gate.ac.uk/) to identify names and places in the texts written in English, and we would like to apply a similar process to the texts written in other languages. It is likely that we will create some of the gazetteers from scratch, in which case we would aim to make these available as part of the project’s outputs.

One of the most interesting challenges is in how to map the resulting data meaningfully. Exploring responses to landscapes of scarcity is a key research question of the project, and descriptions of such responses feature very heavily in some of our texts, particularly in the travel writings (see McRae, 2009 for a discussion of some of themes in travel writing of this period). The works of Peter Mundy, for instance, are full of rich descriptions of the places he visited, including very personal observations of the circumstances in which he found himself (Carnac Temple, 1914; Pritchard, 2011). We are particularly keen to place our work in the context of some of the projects that have taken place during the last decade on mapping emotional responses to landscapes at other periods and in different geographical areas. While many of the recent projects on emotional cartography use wearable technologies to measure and record responses to the landscape, such as Christian Nold’s 2006 Greenwich Emotion Map (an art project combining annotations with measurements of skin responses at different stages of a walk through the Greenwich area of London: http://www.emotionmap.net/background.htm), we see potential in learning from, and building upon, their approaches to visualising the resulting data (Nold, 2009). Projects such as Kurt Jensen’s representation of Sterne’s A Sentimental Journey, built using Neatline ( http://neatline.org/), have also suggested possible directions for representing some of the travel writings ( http://enec3120.neatline-uva.org/neatline/show/a-sentimental-journey), and our recent workshop on food security has helped to clarify the relevant user requirements ( http://foodsecurity.exeter.ac.uk/). However the question of how to integrate such a wide variety of sources and languages into useful and meaningful maps remains one of the most interesting and challenging aspects of the project. As such, it will be a key focus of our paper, and we anticipate that presenting the results of our experiments with this data could be helpful for other projects that are grappling with similar issues, potentially in very different subject areas.