Patricia Murrieta-Flores is Director of the Digital Humanities Research Centre at the University of Chester, UK; she is also European Research Council Senior Researcher on the The Past in its Place project (2014-2016) and an affiliate of the European Research Council-funded project Spatial Humanities: Texts, GIS, Places (2012–2016).
Christopher Donaldson is Lecturer in Regional History at Lancaster University. He is co-I on the Leverhulme Trust-funded project Geospatial Innovation in the Digital Humanities (2015–2018); he is also an affiliate of the European Research Council-funded project Spatial Humanities: Texts, GIS, Places (2012–2016).
Ian N. Gregory is Professor of Digital Humanities at Lancaster University. He is PI on both the Leverhulme Trust-funded project Geospatial Innovation in the Digital Humanities (2015–2018) and the European Research Council-funded project Spatial Humanities: Texts, GIS, Places (2012–2016).
This is the source
Exploratory studies have demonstrated the benefits of implementing Geographic Information Systems (GIS) technology in literary and cultural-historical research. These studies have done much to affirm the power and flexibility of GIS technology as a resource for humanities scholarship. At the same time, however, these studies share a common limitation in that they tend to rely on the analysis of point-based cartographic representations. Such representations are suitable for modelling quantitative geographical phenomena, but they are inadequate for modelling qualitative human phenomena. This inadequacy constitutes a significant problem for researchers who aspire to analyse the geographical experiences and spatial relationships represented in works of literature, including works that contain accounts of travel. The present article proposes a solution to this problem by demonstrating how advanced spatial analyses within GIS such as Cost-Surface Analysis (CSA) and Least-Cost-Path Analysis (LCP) can be used to facilitate more nuanced interpretations of historical works of travel writing and topographical literature. Specifically, the article explains how GIS, CSA and LCP can be combined to build coherent spatial models of the journeys recorded in the works of three canonical eighteenth-century British travellers, each of whom composed influential accounts of their travels through the English Lake District: the poet Thomas Gray (1716–1771), the naturalist Thomas Pennant (1726–1798) and the agriculturist Arthur Young (1741–1820).
Demonstrates how advanced spatial analyses within GIS can be used to facilitate more nuanced interpretations of historical works of travel writing and topographical literature.
Geographic Information Systems (GIS) technology has been at the forefront of
research in the applied sciences since its inception during the 1960s. In the
Humanities, however, the use of GIS has been much more sporadic. This latter
trend can principally be attributed to the fact that only some Humanities
disciplines routinely engage with the kinds of quantitative and geographical
data that GIS were designed to analyse
These disciplinary trends in Literary Studies have been shaped by pioneering
interdisciplinary scholarship which has enabled literary critics, theorists and
historians to work with Geographic Information Scientists on projects that
employ GIS in collaborative research. One innovation that has occurred as a
result of this research is the implementation of GIS in the study of both
individual literary works and literary corpora composed of several works. This
development has variously been labelled ‘Literary GIS’ (see
At the same time, however, these studies share a common limitation in that they
largely rely on the analysis of distribution maps and other kinds of point-based
cartographic representations. Point-based mapping techniques are useful for
plotting the position of fixed geographical features and for facilitating the
implementation of methodologies such as point pattern analysis; but these
techniques are also reductive in that they limit the real complexity of places
as sites of social engagement and interaction. Although we are accustomed to
thinking about places as being static
, or as being defined by fixed
boundaries, their cultural significance is in fact defined by other factors,
such as the interconnected flow of experiences that converge in them and the
roles they play as nodes in larger spatial networks
The present article proposes and models a GIS-based approach that can assist with structuring analyses that satisfy these requirements. In what follows, we shall demonstrate how advanced GIS-based spatial analysis, particularly Cost-Surface Analysis (CSA) and Least-Cost-Path Analysis (LCP), can facilitate more nuanced interpretations of historical works of travel writing and topographical literature. Our objective in the following pages shall be to combine these methodologies to perform a spatial analysis of the journeys recorded in the travelogues of three canonical, eighteenth-century British travellers: the poet Thomas Gray (1716–1771), the naturalist Thomas Pennant (1726–1798) and the agriculturist Arthur Young (1741–1820). The travelogues of Gray, Young and Pennant are diverse works, and they reflect the fact that these three writers undertook their journeys for distinctive reasons. For the sake of concision, we shall limit our analyses to the portions of Gray, Young and Pennant’s writings that concern the English Lake District, and we shall focus on employing GIS, CSA and LCP to investigate how the relative accessibility of the Lakeland landscape influenced both the routes these three travellers followed and, consequently, the experiences documented in their accounts.
Literary GISthrough the spatial analysis of historical travel writing and topographical literature
In engaging with a selection of historical works about the English Lake District, this article aims to extend and refine aspects of the research undertaken by the
Significantly, these divergent patterns point up a number of the key
differences between Gray’s and Coleridge’s excursions. Gray’s tour was a
fifteen-day trip by coach and horseback from Brough to Lancaster via the
lakes of Ullswater, Derwentwater, Thirlmere and Windermere. Coleridge’s
tour, by contrast, was a nine-day odyssey taken on foot from his house in
Keswick to the coastal village of St Bees, and thence famously over Scafell
Pike to Eskdale and Coniston before returning homeward over Dunmail Raise.
These contrasting routes indicate crucial differences in touristic attitudes
and practices
In using GIS to ground these interpretations, the study completed by the
As the map displayed in Figure 1 indicates, the
connect-the-dotsstrategy does portray the sequence of places that Gray and Coleridge visited, the geometrical depiction it produces gives us, at best, a superficial understanding of how the physical geography of the Lake District shaped the experiences documented in their accounts. Integrating CSA and LCP, we decided to undertake a more nuanced spatial analysis of a selection of key historical Lake District travelogues. Once again, we decided to focus on Gray’s account, which was composed in 1769 as a series of epistles to his friend Thomas Wharton. But instead of working with Coleridge’s account (which was written more than four decades later), we elected to draw on the accounts of two other writers who toured the Lake District at roughly the same time as Gray did. The first of these writers, again, is Arthur Young, who paid a visit to the Lakes whilst completing an agricultural survey of Northern England in 1768; the second is Thomas Pennant, who passed through the region as part of his tours of Scotland in 1769 and 1772.
Our decision to focus on Gray, Young and Pennant is not, however, solely based on chronology. It is, more significantly, based on the fact that they were the three chief authorities cited in Thomas West’s highly influential
We shall proceed to explain the use of CSA and LCP momentarily. Before we do,
however, we will first summarize succinctly the steps we took to meet each
of our three objectives. The first step in this process was to create a
GIS-readable version of the tours documented in Gray, Young and Pennant’s
accounts of the Lake District, and this itself involved several stages.
First of all, we had to digitise the portions of each writer’s account that
dealt with the Lakeland region. This gave us a set of text extracts
comprising
In order to achieve our first objective, we used CSA and LCP to simulate the natural corridors, determining the most efficient route for travelling over the terrain taking into account the relief of the Lake District’s topography. We then implemented a spatial proximity test in order to assess whether the places in itineraries we had extracted from Gray, Young and Pennant’s accounts were closer to the natural corridors formed by the region’s topography than would be expected by chance alone. Performing this test allowed us to ascertain the relative accessibility of the places each writer decided to visit. In order to achieve our second objective, we simulated the routes most likely taken by the three writers according to the places they visited. Taking the journeys of the three authors into account, we then employed a Line Density Analysis (LDA) to identify the most frequented routes in order to determine whether there was a preponderance of attention given to any particular area or region. Finally, in order to achieve our third objective, we carried out another spatial proximity test. In this instance, however, we analysed the spatial relationship between the simulated pathways and those places that were mentioned in Gray, Young and Pennant’s accounts, but which were not actually sites these travellers visited.
The use of CSA and LCP is well established in the applied sciences
CSA is a spatial methodology that allows one to determine the cost
of
travelling over a surface by considering not only Euclidean space, but also
more complicated factors, such as slope and elevation. In other words,
instead of limiting us to the examination of the Cartesian spatial
properties of the data, it allows us to consider additional factors such as
the contours of the terrain, as well as other variables (such as boundaries
and borders) which influence how one moves through the landscape
The second step in performing CSA is to establish the associated cost of
traversing the terrain and any impediments (such as rivers or bodies of
water) it might contain. This is accomplished by creating a
The third step in performing the CSA is to derive from the friction surface,
a final cost surface
that depicts the cumulative cost of
travelling from a source or point of origin to any other location within the
terrain represented (Figure 4). This step is performed multiple times to
calculate the cost to travel from each location to any part of the
landscape, which result in cost surfaces derived for each place investigated
in the corpus.
Once the cost surfaces have been established, one can then proceed with completing LCP. LCP is, essentially, a subsequent set of calculations that determines the path across the cost surface that will cost the traveller the least. The GIS is instructed to create a path across the cost surface following the adjacent cell with the lowest value assigned.
As with the cost surface, the paths that result from performing LCP analysis
are dependent on the cost variables that were used to create the model. For
the purposes of this case study, we created two anisotropic models using
ArcGIS. Anisotropic models account for the direction of movement in the
calculation of the cost, whereas isotropic models do not. It should be noted
that for each of our experiments we made use of LCP, and that we created two
different models in order to meet our objectives. In order to avoid
confusion, when we discuss the results related to our first objective, we
shall use the expression natural corridors
to indicate those paths
for which a model was designed to identify the least cost paths for the
whole region. When we discuss the results related to our second and third
objectives, we shall refer to either simulated pathways
or
simulated itineraries
in order to describe the model that was
designed to calculate the least cost paths for each of the authors’ tours.
In order to investigate the degree to which the Lake District’s topography may have conditioned the itineraries of Gray, Young and Pennant’s tours, we calculated the least cost paths between the places each author visited. We accounted for three variables in these calculations: 1. the slope of the region’s terrain; 2. the presence of bodies of water; 3. the direction in which Gray, Young and Pennant were each travelling (Figure 5). We then carried out a statistically based spatial proximity test in order to assess the spatial relationship between the places Gray, Young and Pennant visited and the Lake District’s topography. In order to do this, we created a distance index for the accounts of each writer by calculating buffers of 500m from the natural corridors and then counting the number of visited places within each buffer (Figure 6). We then used this data to perform a statistical significance test (Kolmogorov-Smirnov). Our goal was to test whether the places Gray, Young and Pennant visited were closer to the natural corridors than one would expect by chance alone. The significance level set to test the null hypothesis was 0.01, and in all cases it was rejected. (As noted above, our goal here was to ascertain the degree to which Gray, Young and Pennant were influenced by a logic of optimality in their journeys: in other words, how closely they followed the natural corridors of the terrain.) This finding suggests that Gray, Young and Pennant did generally follow the natural corridors of the Lake District in their journeys and, accordingly, either that they were able to read the contours of the landscape successfully or, at least, that they were well informed about the most efficient ways to move between the locations through which they travelled.
Having thus satisfied the first of our three objectives, we then proceeded to determine the extent to which Gray, Young and Pennant emphasised specific portions of the Lake District in their tours. In order to do this, it was necessary to construct simulations of the journeys documented in their accounts. We built these simulations by using each of the places in the itineraries we created for each writer as nodes, and then by using cost surface analysis to calculate the least cost path between these places, following the order in which they were visited (Figure 7). After we calculated least cost paths for each account, we then performed a LDA, which allowed us to identify those parts of the landscape through which each of the writers were likely to pass most frequently (Figure 8). Performing these analyses indicates that, collectively, Gray, Young and Pennant spent a significant amount of their time in the Lake District travelling around the lakes of Windermere, Ullswater and Derwentwater. We will offer additional commentary about these findings in the following section. But before concluding, it is worth noting that in addition to examining the locations Gray, Young and Pennant visited, we also performed an analysis on the places they mentioned but did not visit.
Our intention, in this case, was to determine the role played by these unvisited places in the travel narratives presented in Gray, Young and Pennant’s accounts. Accordingly, after we identified all the places in the four accounts that were mentioned but not visited, we then carried out another spatial proximity test. This time, the distance index for each writer was calculated with buffers of 1 km created from the itineraries simulated and the statistical test was also set to a significance level of 0.01. Our intention was to assess the degree to which the representation of the Lake District in each account was influenced by the places Gray, Young and Pennant visited and the degree to which it was influenced by places that they mentioned but did not visit. The results of this test indicate that the places mentioned but not visited in Gray and Young’s accounts, and in Pennant’s account of his tour of 1772, are closer to the routes they these writers followed than one would expect by chance alone. When it came to Pennant’s account of his 1769 tour, however, the null hypothesis could not be rejected. As we explain below, this result clues us into an important aspect of how Pennant’s 1769 tour stands apart from the other texts assess in this experiment.
We believe that the methods outlined above can facilitate more detailed analyses of how the records of historical travellers were shaped by the contours of the landscapes through which they travelled. What is more, to the degree that we accept those travellers’ accounts as accurate descriptions of their journeys, these methods can furnish useful insights into how their accounts were conditioned by factors including the physical geography of the terrain, the existence of roads and tracks, and the availability of specific kinds of transportation. Our first spatial proximity analysis of the relationship between the tours and the Lake District’s natural corridors demonstrates that all the places Gray, Young and Pennant record visiting are nearer to these pathways than would be expected by chance alone (Figure 5). In other words, although these three writers certainly visited a few places that could be considered remote — even adventurous — their itineraries still closely followed the natural corridors of the Lake District. (Indeed, as indicated in Figure 6, most of the places they visited fall within 500 metres of the region’s natural corridors.)
These initial observations assist us in making some basic distinctions about the tours documented by Gray, Young and Pennant. For instance, they help us to discern that of the four accounts, the itinerary Pennant followed during his 1769 tour was the one that most closely adhered to the natural corridors of the region’s topography. Comparatively, Young and Gray strayed farther from the Lake District’s natural corridors. This finding makes sense, since in 1769 Pennant merely travelled along the outskirts of the Lakeland region whilst on his return from Scotland to his Downing estate. In his tour of 1772, Pennant visited other parts of the Lake District, particularly the Cumberland coast. Here again, however, the analysis shows us that, for the most part, he still followed the natural contours of the region’s terrain.
Beyond enabling us to draw such basic distinctions, the spatial proximity
analysis we conducted also helps us to perceive the correspondence between the
Lake District’s natural corridors and the routes that were present at the time
Gray, Young and Pennant paid their visits to the region. Inasmuch, this analysis
enables us to discern that, without exception, all three writers primarily
relied on local carriage roads and turnpikes in their travels. At first glance
this may seem a minor point, but it is of significant importance. Scholars
Accordingly, turning to Figure 8, it makes sense that the parts of the landscape
through which Gray, Young and Pennant passed most frequently are located in the
eastern half of the region (generally speaking, around the lakes of Ullswater,
Derwentwater and Windermere), since these were the sites around which the
majority of road improvements occurred. Incidentally, this also helps to explain
why neither Gray nor Young nor even Pennant — who travelled through much more of
Cumberland — set foot in the western valleys of Wasdale and Ennerdale. Although
turnpike acts were passed to improve access to this area in 1750 and 1762, the
roads in this remote part of the country remained in disrepair well into the
nineteenth century much improved since Mr. Gray made
his tour in 1765 [sic], and Mr. Pennant his in 1772
, as late as 1821
West’s guidebook (then in its eleventh edition) still cautioned tourists that
the roads leading to Ennerdale and Wasdale were, as yet, unimproved
So, in the first instance, Figure 7 shows us that Gray, Young and Pennant
chiefly went where the local carriage routes and turnpikes enabled them to go.
But how, we might ask, did those roadways shape their impression of the places
through which they passed? Cultural historians have often remarked, with due
amusement, that tourists in the late 1700s moved from scenic vista to scenic
vista, collecting sensations in much the same way that modern tourists collect
snapshots and souvenirs. Significantly, however, cultural histories of the Lake
District tend to quote passages from eighteenth-century travelogues and
guidebooks without commenting on the material conditions that determined the
writer’s point of view. Thus, whereas Gray’s panoramic description of Grasmere’s peace, rusticity, and happy
poverty
Twelve of the fifteen miles from
Shapp to Kendal are a continued chain of
mountainous moors, totally uncultivated; one dreary
prospect, that makes one melancholy to behold. Pass over
Shap
Fells, more black, dreary, and melancholy, than any of the Highland hills,
being not only very barren but destitute of every picturesque beauty. This
barren scene continued till within a small distance of Kendal[.]
When read in isolation, these two passages might seem simply to reflect the
late-eighteenth-century vogue for sensational landscape description. When we
turn to the map, however, we begin to discern that they are as much shaped by
the aesthetic conventions of their era as they are by the situation of the
roadway that Young and Pennant followed. With a peak elevation of some 445m
above sea level, the route of the old turnpike between Shap and Kendal is
notorious for its isolation, exposure and poor weather
This is one of the benefits of using spatial models to study historical works of
travel writing and topographical literature: it obliges us to examine local
descriptions within the broader geographic context of the traveller’s
experience. Rather than thinking in terms of isolated places, the models we have
developed emphasize that these writers were describing journeys from place to
place. In doing so, these models help remind us that the experience of
travelling in the Lake District in the late eighteenth century was not always
pleasurable, but that it imposed upon the body in other ways, through the jolts
of deep wheel ruts, the hazards of unpaved roads, the exposure to the wind and
the rain and the mud. The fact that Gray, Young and Pennant take the time to
record these details alongside the various scenes they encounter indicates the
importance of the
I know not, in the whole range of language, terms sufficiently expressive to describe this infernal road… let me most seriously caution all travellers, who may accidentally purpose to travel this terrible country, to avoid it as they would the devil; for thousand to one but they break their limbs by overthrows or breakings down. They will here meet with rutts [sic] which I actually measured four feet deep, and floating with mud only from a wet summer; what therefore must it be [like] after a winter? … These are not merely opinions, but facts, for I actually passed three carts broken down in these eighteen miles of execrable memory.
Carrying out a spatial analysis to obtain a more precise visualisation of the routes writers like Young might have followed helps put these sorts of commentaries into context.
In addition, the other GIS-based methods we employed allow us to analyse the role that particular places play in the geographies represented in Gray, Young and Pennant’s accounts. For instance, by examining whether the places that were mentioned but not visited in these accounts have a significant spatial relationship with the paths simulating the itineraries of Gray, Young and Pennant, we can assess how the routes they travelled might have influenced the perception of the places and geographies recorded in their tours. The results of the statistical significance test indicate that places mentioned but not visited in Gray and Young’s account, and in Pennant’s 1769 tour, are closer to the routes than one would expect as a matter of chance alone (Table 1). In the case of Pennant’s 1772 tour, however, the null hypothesis could not be rejected, indicating that there appears not to be a significant relationship between the places mentioned but not visited in his account and the simulated pathway we created from it.
This result is significant because it confirms how working with quantitative
research techniques can clue us into the qualitative differences between
historical works of travel writing. Specifically, it clues us into the fact that
whereas Gray and Young’s account, and Pennant’s second account, are all accounts
of journeys those parts of Lancashire,
Westmoreland, and Cumberland, which I had not before seen
Conducting this statistical test, then, enables us to draw a meaningful
distinction between the four works in our case study. More generally, it prompts
us to think critically about the role that places that were mentioned but not
visited plays within each tour. That each of the four works contains references
to such places indicates that the accounts of Gray, Young and Pennant are more
than mere itineraries; rather, they are descriptive records that attempt to
broaden the reader’s awareness of the geography of the Lake District, which was
still a relative terra incognita during the
late eighteenth century. The fact that each account contains references to
places that the writers did not actually visit suggests Gray, Young and Pennant
all endeavoured to give their readers a sense of the surrounding countryside. In
fact, when one turns back to the works themselves, one begins to see that the
place-names that are mentioned but not visited are often invoked in panoramic
descriptions. Thus, for example, Pennant’s account of the mountains visible from
the ramparts of Carlisle castle: The
castle is ancient, but makes a good appearance at a distance: the view from
it is fine, of rich meadows, at this time covered with thousands of cattle,
it being fair-day. The
Eden here forms two branches,
and insulates the ground; over one is a bridge of four, over the other one
of nine arches. There is besides a prospect of a rich country, and a distant
view of Cold-fells, Cross-fells, Skiddaw, and other mountains.
Elsewhere, one finds that place-names that are mentioned but not visited are also
invoked in descriptions of trade, manufacturing and shipping. In these
instances, the place-names are typically used to situate the industrial centres
of the Lake District in relation to the economic geography of the nation. Hence,
Pennant’s observation Ulverston, a town of about three thousand souls,
seated near the water side, and is approachable at high water by vessels of
a hundred and fifty tuns [sic]; has a good trade in iron ore, pig and bar
iron, bark, lime-stone, oats and barley, and much beans, which last are sent
to Leverpool [sic], for the food of the poor enslaved
negroes in the Guinea
trade. Numbers of cattle are also sold out of the neighborhood, but the
commerce in general declines; at present there are not above sixty vessels
belonging to the place; formerly about a hundred and fifty mostly let out to
freight; but both master and sailors go now to Leverpool for employ.
Significantly, one finds similar passages in Arthur Young’s account, which was in
large part a survey of agricultural productivity. Consider, for example, his
account of the wool market in Kendal, where he uses place-names to refer to
specific market centres as well as to specific breeds of sheep: The wool they use is chiefly
Leicestershire, Warwickshire, and Durham: They generally mix Leicestershire
and Durham together. The price 8d. 9d. and 10d. per
lb. They send all the manufacture to London
by land carriage, which is said to be the longest, for broad wheel wagons,
of any stage in England.
These observations, when viewed alongside the results from the second spatial proximity analysis, suggest that places that are mentioned but not visited played a number of important roles within the broader associative geographies of each text. They are invoked not only to describe the surrounding countryside, but also to document how places throughout the greater Lakeland region connect geographically, socially and economically with the wider world.
Although the use of spatial analysis in Humanities disciplines such as Literary
Studies is still at an early stage, the application of geospatial technologies,
such as GIS, has already begun to have a profound influence on the way literary
scholars and historians approach the study of space, place and landscape. By
implementing the geospatial methods employed in this study, for example,
researchers can achieve new insights not only into the cognitive construction of
geographies in literary works, but also into fuzzier
phenomena, such as
geographies of absence: places that are not specifically mentioned in a literary
work, but nonetheless form part of its implied geography. In addition, the
introduction of approaches such as Geographical Text Analysis
The application of GIS technology in literary studies represents an important development. Until recently, however, little has been done to implement more creative approaches to mapping a writer’s movement through space. Advancing from point-based representations to the analysis of raster-based pathways is, therefore, a major step forward in the application of spatial technologies for studying works of travel writing and topographical literature. Such works are, after all, not merely descriptive inventories of places, monuments and buildings; they are narratives that offer first-hand accounts of the journeys of specific individuals. In order to visualise such works adequately we must employ techniques that are capable of representing the places they mention not as discrete locations, but as a series of interconnected points along the line of transit that constitutes the narrator’s tour. As noted at the beginning of this article, is only by assessing works of travel writing as journeys that we can understand how the writers of those works encountered the places and landmarks they described. Thus, with respect to Gray, Young and Pennant, it is only by tracing their journeys through the Lake District that we can hope to obtain insights into how the topography of the region — with its valleys, peaks and mountain passes — shaped the local experiences recounted in their travelogues.
The implementation of these methods, however, is as yet imperfect. One
significant complication is that the data used to create the Digital Elevation
Models (DEMs) used to perform CSA and LCP comes from modern sources (Lidar
images and/or satellite or aerial photographs). This means that they contain the
elevation data of modern roads, motorways and towns. Although the routes
followed by eighteenth-century Lake District tourists often correspond to the
routes of the modern road networks in Cumbria, in many instances there are
significant differences between the two
Based on the study on which we have reported in this article, and the examples cited herein, we can confidently conclude that spatial analysis can facilitate more nuanced interpretations of works of travel writing and topographical literature. The process of creating a spatial dataset from the tours of Gray, Young and Pennant alerted us to details that most previous studies have overlooked. The simulations we created using this dataset have, moreover, provided the foundation for our on-going research into experimental forms of digital literary geography, such as using Google Earth to create interactive virtual tours. We believe that studies, such as the one presented here, can be productively combined with other emergent forms of geospatial technology to foster innovative collaboration between researchers in the Humanities and the Geographic Information Sciences.
This research was supported by the European Research Council (ERC) under the European Union’s Seventh Framework Programme (FP7/2007-2013) / ERC grant