Mapping qualitative geographies: Tourist, traveller and inhabitant places in the literary Lake District

Significant progress has been made on analysing the geographies in textual sources using techniques such as Geographical Text Analysis (GTA). These approaches identify place names (toponyms) in the text and allocating these to coordinate-based locations to allow mapping and further analysis. From here a combination of techniques from geographical information science and corpus linguistics is used to analyse and visualise the geographies within the texts under study (Paterson & Gregory, 2018; Taylor & Gregory, 2022). This enables us to describe, visualise and analyse the geographies in large textual corpora, potentially billions of words, to ask questions such as ‘what is at this place?’ or ‘what places are associated with this theme?’

While this approach represents a valuable way of analysing text, GTA as currently configured is limited in three key ways: first, GTA does not use any theoretical concept of space or place, the concepts of place used are driven by the opportunities and limitations of conventional geospatial technologies such as geographical information systems (GIS). Second, geography in a GTA is solely represented using toponyms and their associated coordinates identified by a process known as geoparsing (Gregory et al. 2015). Any geographical references that are not locatable toponyms are not included, thus much that is geographical is excluded from any analysis. Third, the concept of geography in subsequent visualisations and analysis remains coordinate based with most spatial analytic approaches using straight-line distance on a Euclidean plane. While this representation is useful and recognisable, it is not necessarily the way in which writers experience, describe and aggregate place. In summary, to date, GTA has been restricted to a subset of geographical references – namely place names – onto which it imposes a quantitative, coordinate based concept of geography. In this paper, we describe a new approach to GTA that is based on a qualitative, network based representation of geography that, we argue, provides a more holistic and theoretically robust approach to representing geography in ways in which writers themselves describe the places that they experience.

This paper starts from the theoretical framework for place described by Agnew (2011) in which place can have three key characteristics: the location of an object or event; the locale which refers to the natural or built environment that makes up a place; and the sense of place, the accumulated events, actions, and memories attributed to a location (Agnew 2011). Location can be sensibly represented using coordinates for named places in the ways described above. Locales can be represented using what we term Geographical Feature Nouns (GFNs), features such as ‘lake’, ‘hill’ and ‘village.’ Geographical feature nouns can be identified within the text using approaches from corpus linguistics, however, they cannot be located geographically using coordinates as their locations are usually ambiguous. Sense of place, is a somewhat different concept. Locations and locales identify that there is a place, while sense of place describes the writer’s conception of that place or what happens at it. Sense of place can be derived from adjectives, nouns and verbs, in particular, that co-occur with toponyms (locations) or GFNs (locales).

The above framework provides a theoretically-based approach to summarising the geographies in large corpora however it presents us with a challenge of how to locate places whose locations are ambiguous, and how to analyse the relationship between them without using Euclidean distance. To do this, we use textual space rather than Euclidean. Where a writer refers to two or more locations or locales within the same sentence they are assumed to be related. The locations and locales then form nodes on a network with edges being formed when locations or locales co-occur. This then allows us to identify clusters of locations and locales within textual space – ie geographical references that the writers see a being closely associated, rather than those that are necessarily found close together on the map. These clusters can then be allocated to a sense of place based on other words that are also found in the sentence.

The paper uses this analytic framework to explore the contrasting geographies of three different types of people – tourists, travellers and inhabitants – as described in a large corpus of pre-twentieth century writing about the English Lake District. These three groups were identified as having contrasting geographies using conventional GTA (Taylor & Gregory, 2022: chap. 3). In this analysis we take this further using the approach described above. The locations and locales the writers assign to the three groups are identified along with the senses of place assigned to the different groups. These are mapped in textual space and the different patterns for the different groups are compared and contrasted to provide novel views on the different geographies associated with these three groups by the writers themselves. The different senses of place for each textual-geographical cluster are also defined. The wider implications of this paper is that it provides an analytic framework for the geographical analysis of texts within which is theoretically grounded, incorporates all types of place described by the writers, and clusters them in ways that the writers described.

References:

Agnew J. (2011) “Space and place” in Agnew J. and Livingstone D. (eds.) Handbook of Geographical Knowledge. London: Sage

Gregory I.N., Donaldson C., Murrieta-Flores P. and Rayson P. (2015) “Geoparsing, GIS and textual analysis: Current developments in Spatial Humanities research” International Journal of Humanities and Arts Computing, 9, pp. 1-14

Paterson L.L. and Gregory I.N. (2018) Representations of Poverty and Place: Using geographical text analysis to understand discourse. London: Palgrave.

Taylor J.E. and Gregory I.N (in press, June 2022). Deep Mapping the Literary Lake District. Lewisburg: Bucknell University Press