Space as a text: James Duncan and the concept of landscape

In this note I make some remarks on the concept of landscape as I have seen it used in geography through the work of one of its most notable English-speaking advocates, James Duncan. Contrarily to previous reviews, I do not focus on a single article, but rather on a few of them. The main paper I will nonetheless use is “The Power of Place in Kandy, Sri-Lanka: 1780 – 1980”, which can be found in The Power of Place, edited by John Agnew. All unreferred quotes are from this document.

The Rio São Francisco and Fort Maurits, by Frans Post, as presented in Philippe Descola's lecture "Les formes du paysage"

The Rio São Francisco and Fort Maurits, by Frans Post, as presented in Philippe Descola’s lecture “Les formes du paysage”

The concept of landscape

It is regularly pointed in readers of geography that the first approach of modern human geography organised its particular interest based on the concept of region, taken as a given whole, closed upon itself, which worked as a basic enviromental unit. The goal of geography as a science, from that perspective, would be to explore the various regions and uncover their specificities, not in order to built a sort of global agenda of the different regions, but for comparative or interpretive purposes. The synthesis of this tradition made by Richard Hartshorne in The Nature of Geography and Perspectives on the Nature of Geography places this position well: what matters for the author is not as much the separation of a spatial realm which to study, but instead “the study of aerial differentiation”, concerning men, nature, climates, and every other aspects as parts of the region, following the grand tradition of possibilism. This trend of research placed the discipline under the banner or uniqueness, and in the positivist turn marking the 1960s, got considerably marginalised as a non-conceptualised and non-generalisable accumulation of monographs.

As often, a later consensus would recognise that if the answer was questionable, the question still stood of how a specific area is given a particular unicity which transpires, among other things, in the human experience of it. This is where the concept of landscape comes in handy, as presented by Philippe Descola in his lecture “Les formes du paysage”: one can start with the contrast between Descola’s own experience of a view of a river in South America after a long time in the forest, marked by his European socialisation which “taught” him through paintings and readings to see the magnificient wild nature as a wonder, and the experience of his Achuar companion, who sees in the same set of material stuff (the river, the small islands, animals, woods, and beaches) a completely different thing, such as memories of encounters, a river well maintained, and so forth. What we see here indicates the two basic elements for apprehending the notion of landscape: a landscape implies at the same time a spatial, material arrangement of stuff, and the experience of an individual looking at it with their acquired capacity to look at it.

Before this conceptualisation, though, and underlying the debate that will surround it, comes the question of: do all people in the world have “landscapes”? Indeed, the concept is closely attached to specific social groups of cultures (the rich Western European and Eastern Asian ones, for example), and has often been used in colonial practice. In his review on the topic, James Duncan, one of the leading cultural geographers and prolific user of the concept, specifies: “landscape is a painterly way of seeing the world that creates a picturesque view. Such a painterly way of seeing, as they point out, is an élite way of seeing, not only because it was the wealthy classes of Europe who commissioned paintings but also because there developed a dialectical relationship between the rural landscape and painting” (1995). Similarly, on the painting of the Rio São Francisco chosen to illustrate his lecture, Descola shows how the painter has chosen, when representing a Brazilian landscape, to do it as if it were a Dutch one. A first remark to be made is therefore that an aesthetical definition of landscape is necessarily anchored in a particular tradition or history and cannot be detached from it. This nonetheless does not render the concept unoperational for analysis once this basic precaution has been set.

As James Duncan points out, once the aesthetic content of the concept has been left aside, a landscape “is a culturally produced model of how the environment should look. It is, therefore, not only an environment but a type of arrangement of hills and trees, or towns and houses”. This definition has the advantage of extracting the concept of his original aesthetic understanding, while maintaining a considerable use for it, that of transcribing the interaction between power, culture, and the taste for certain shapes, sounds, feelings, smells, and their relations as perceived by people (once again, consider how what appears to Descola as a grandiose interaction between a river and beaches appears to his companion as the promise of safe passage, for example). As such, the concept of landscape, as Duncan insists, is undetachable from the concept of hegemony, as in both case ideology goes beyond what is explicitely stated in an ideological fashion: saying a specific landscape is “beautiful”, “ugly”, “harmonious”, or “disorderly” is already making an ideological statement.

Lanscape as text: hegemony and agency

This is where the textual analogy comes: presented as such, a landscape can be associated to a text, because it is subjected to processes of writing, reading, but also intertextuality and the mechanisms of diversion which Michel de Certeau associated with the practice of reading: in the same time as landscapes are attached to hegemony, it still mostly functions through perception and, therefore, is never entirely without ambiguity. More importantly, landscapes do still work as palimpsests: the example given by Duncan, the city of Kandy, shows this ambiguous dimesion.

To illustrate his concept, Duncan presents the transformation of the Sri Lankan city of Kandy. In his exposé (which is developed much more deeply into his book The City as Text), the author links the production of space and obedience to power to the landscape of the city. This landscape is as much a product of the existing hegemony and discourses on power as it is a resource which in the people’s everyday life reasserts the legitimacy of the current monarch. To clarify what the author means by that, we can compare two extracts, between the pre-colonisation and colonised period of the city:

The city was composed of two rectangles, the rectangle being the sacred shape of the cosmic cities of the gods. The western rectangle constituted the main city and contained both the residence of the nobles and the houses and shops of the common people. This rectangle was in turn divided in 21 squares representing the 21 administrative units of the kingdom. The kingdom, therefore, was symbolically reproduced within the city; the macrocosmos was reduced to the microcosmos.

The British usurped and transformed the Kandyan symbols in a number of ways. There existed a belief in Kandy that no foreign power could rule the kingdom unless three conditions were first met. The foreigners had to take position of the relic that waw jealously guarded by the monks of the Tooth Temple and hidden away in the mountain when foreign armies invaded. They also had to build a road from the coast up through the mountains of Kandy, and finally they had to ride a horse through the heart of a mountain. (…) Upon their arrival in 1815 they negotiated for the return of the [locally important] Tooth Relic to the temple where their soldiers could keep an eye on it. They also started to work on a road connecting Kandy to the coast. (…) Upon their arrival in 1815 the British immediately set about replacing the Kandyan power structure and symbols of authority with their own.

Considered as such, there is little difference between the notion of landscape and that of the production of space, as proposed by Lefebvre: in both case we have a relation between space and the power relations. What is specific to the notion proposed by Duncan is the importance, not only of the “writing”, but also of the “reading”: it is through their experience of landscape as familiar, beautiful, ugly, and so forth, that the people are taken in power relations. Conversely, as we see with the example of the British invaders, landscape does not work as a unilateral discourse: it is a perceived physical translation of a symbolic order of things.

This barely makes for an introduction of the implications of the concept of landscape. For example, the coexistence of a multiplicity of landscapes has not been evoked, neither have plenty of other questions raised by the concept. Although his work on Kandy had been criticised on several manners – including the question of a tendency to assert instead of proving and some factual mistakes (Meyer, 1994), Duncan’s argument on considering landscape as a text remains seducing on more than one account. By linking material forms and shapes, power, and culture, it allows us to make sense not only of the fact that the impression given by space when it is observed by researchers or people has an importance, but also serves as a key to enter the debates on the importance of the transposition of certain types of representing space, using it, and of certain ways to build or organise in it across contexts. Thinking of the concept, one cannot avoid evoking the multiplication of “normalised” urban shapes across countries, including malls or suburban neighbourhoods, boulevards, but also towers, urban divisions and planning, and so forth.


Descola Philippe, Les formes du paysage, anthropologie de la nature, Lecture at the Collège de France, 2012 – 2014

Duncan James, “Landscape geography, 1993-94”, Progress in human geography, Vol 19, No 3, 1995, pp 414 – 422

.-, “The power of place in Kandy, Sri Lanka, 1780 – 1980”, in Agnew John A and Duncan James (eds.), The power of place, bringing together geographical and sociological imaginations, Boston : Unwin Hyman, 1989

Hatshorne Richard, The nature of geography; a critical survey of current thought in the light of the past, Lancaster: The Association, 1939

.-, Perspective on the nature of geography, Chicago: Association of American Geographers, 1959

Meyer Eric, “Reviewed Work: The City as Text: The Politics of Landscape Interpretation in the Kandyan Kingdom. (« Cambridge Human Geography ») by James S. Duncan”, L’Homme, Vol 34, No 132, 1994

You may also like...

4 Responses

  1. Sam Kinsley says:

    Hi, you may find more recent work on “landscape” in geography of interest too. For example, John Wylie’s book “Landscape”:

    • Alex Mahoudeau says:

      Thank you for the advice. I may not have stressed enough that Duncan’s work is dated and has been subjected to a lot of discussion indeed.

Leave a Reply

Your email address will not be published. Required fields are marked *