Key Texts in Human Geography

Phil Hubbard, Rob Kitchin, Gill Valentine, Phil Hubbard, Rob Kitchin, Gill Valentine

• Publisher: SAGE Publications Ltd
• Year: 2008
• Print ISBN: 9781412922616, 9781412922609
• eBook ISBN: 9781446243657
• Edition: 1
• Topic: Physical Sciences
• Subtopic: Geography
• Index: Geography

A book that will delight students… Key Texts in Human Geography is a primer of 26 interpretive essays designed to open up the subject?s landmark monographs of the past 50 years to critical interpretation... The essays are uniformly excellent and the enthusiasm of the authors for the project shines through… It will find itself at the top of a thousand module handouts.
- THE Textbook Guide

"Will surely become a 'key text' itself. Read any chapter and you will want to compare it with another. Before you realize, an afternoon is gone and then you are tracking down the originals."
- Professor James Sidaway, University of Plymouth

?An essential synopsis of essential readings that every human geographer must read. It is highly recommended for those just embarking on their careers as well as those who need a reminder of how and why geography moved from the margins of social thought to its very core." - Barney Warf, Florida State University

Undergraduate geography students are often directed to ?key? texts in the literature but find them difficult to read because of their language and argument. As a result, they fail to get to grips with the subject matter and gravitate towards course textbooks instead.

Key Texts in Human Geography serves as a primer and companion to the key texts in human geography published over the past 40 years. It is not a reader, but a volume of 26 interpretive essays highlighting:

• the significance of the text
• how the book should be read
• reactions and controversies surrounding the book
• the book?s long-term legacy.

It is an essential reference guide for all students of human geography and provides an invaluable interpretive tool in answering questions about human geography and what constitutes geographical knowledge.

---

1	INNOVATION DIFFUSION AS SPATIAL PROCESS (1953): TÖRSTEN HÄGERSTRAND
	Bo Lenntrop

---

The diffusion of innovations – the origin and dissemination of cultural novelties – is an area of study which concerns all sciences dealing with human activity, including, not least of all, cultural and economic geography. (Hägerstrand, 1953: 1)

Introduction

It is difficult to grasp the importance of Törsten Hägerstrand’s key work on innovation diffusion – his doctoral thesis from 1953 – without an appreciation of the historical context in which the work was conceived and prepared and the fact that it was first translated into English by Allan Pred some 14 years after submission. Most notable for setting out theories of spatial diffusion and adoption, Hägerstrand’s early research also contains the key to the later development of his ideas. One example is time geography, which became formalized in the 1960s, but whose conceptual roots were to be found in Hägerstrand’s writing in the 1940s and 1950s.

Hägerstrand arrived at Lund University in the late 1930s. His interest was directed, more or less by chance, toward migration and he started to work on a project intended to chart the entire demographic development of a considerable geographical area of Sweden from 1840 to 1940. Building an impressive collection of data, this enormous undertaking left Hägerstrand with a profound empirical understanding of demographic development. At the same time he also developed a deeper theoretical proficiency, setting in motion his particular geographical worldview. More particularly, the systematic collection and analysis of data on the life courses of a population over a century contributed to the germination and growth of a foundational idea; namely the importance of analysing spatial processes.

Over time Hägerstrand absorbed important theoretic ideas and trends from beyond Swedish geography. These did not emanate from the regional perspective then dominant in university teaching, with Hägerstrand stating ‘lectures in regional geography were abominably boring … Geography appeared not as a realm of ideas or a perspective on the world but as an endless array of encyclopaedic data’ (Hägerstrand, 1983: 244). Rather they came to his attention through a chance acquaintance. His future wife, Britt, was then working for the ethnologist Sigrid Svensson who was conducting research and publishing books on diffusion processes, and one of his school colleagues had a burning interest in numerical analyses and in developing computers and was an early visitor to the US.

Having gained a sound knowledge of demography and ideas on diffusion courses and simulation models, the foundations of his doctoral thesis had fallen into place. In many respects this work was to mark a decisive break with the then dominant tradition of regional studies. Hägerstrand’s principal aim in his doctoral thesis was not to present a broad regional description of an area, but instead to investigate and illuminate a problem. That the material concerned a specific area, argued Hägerstrand, was a regrettable necessity and not a methodological finesse which in itself marked a stand against regional geography.

Innovation diffusion: the Swedish version

For obvious reasons a doctoral thesis written in Swedish does not reach a large international readership. Nevertheless, it is informative to comment on its immediate reception, 14 years before it was first published in English. The first academic review of Hägerstrand’s doctoral thesis was published in 1953 – in the same year he was awarded his doctorate. The reviewer was Edgar Kant, previously professor in Tartu in Estonia, but who had by then spent many years in Lund at the Department of Geography. Kant was widely read in international scholarly literature and he was also the first opponent at Hägerstrand’s disputation.

Kant opens his review with a discussion of the respective pros and cons of research specialization. Kant was of the opinion that the disadvantages associated with specialization ‘begin to become apparent when large lacunae arise leaving poorly-lit areas in border zones, as monadnocks of the total ignorance’ (Kant, 1953: 221). It is useful to recite Kant’s concluding comments in order to relate exactly why he saw Hägerstrand’s thesis as pioneering.

The author has, to a noteworthy extent, utilized new methods and established new links to neighbouring disciplines. This must present itself as innovative to those who perceive geography as indissolubly bound to traditional methods of investigation and research subjects, such as landscape analysis, which have as their only or primary task ascertaining interactions between man and nature. … It may transpire that the author’s longest expeditions into the unknown border- and twilight-zone have been but excursions leaving many areas as yet largely unexplored. Those who follow in his path can, however, draw benefit from his pioneering work and fashion new riches. (Kant, 1953: 225)

In his doctoral thesis Hägerstrand investigated the changing extent of propagation of cultural artefacts. He did this by selecting six specific indicators; three for agriculture (state subsidies for improving pastures, control of bovine tuberculosis, soil mapping) and three more general indicators (postal money transfer, automobiles, telephones). The choice of indicators was guided by the need that these should be localizable to coordinates and that their development over time could be followed with a very high degree of precision. Furthermore, it was necessary that the indicators had been adopted by a sizeable proportion of the population.

The next stage in the investigation was to establish reduction bases for the indicators. This was necessary because it would have been meaningless to work with absolute numbers of acceptors. Hägerstrand carefully analysed the population (or demographic) development in the region and the location and size of each of the respective homesteads and residential apartments in order to construct the ‘reduction bases’ against which the number of acceptors should be matched.

An important section of the thesis, encompassing nearly 100 pages, then deals with the actual diffusion process. Hägerstrand’s primary objective here is not to identify specific details in this diffusion process, but rather to generalize about the characteristics that could serve as the basis for subsequent operationalization in models. However, it was important to observe certain characteristics as these generated variations in the diffusion process. A significant degree of the discrepancy between the indicators is based on the level of state intervention, for example in promoting the introduction of controls on bovine tuberculosis. Car purchase was principally a matter of private decision making, even if state legislation played a certain role. The introduction and diffusion of the telephone in the region of investigation was affected by the manner and the speed with which the electricity network was developed. Consequently, the diffusion of the six indicators reveals different courses and show how they are influenced to lesser or greater extents by planning and policy at national and regional levels. The larger part of this discussion is of more general interest, and is not confined to its historical or regional context.

With the help of this detailed empirical knowledge, Hägerstrand then formulated a series of experimental, stochastic models to show how innovations spread within a population. The first model was very simple and is steered entirely by chance and provides a picture that most closely approximates the manner in which a rumour spreads through a population. Hägerstrand therefore focuses on how differences in acceptance and how unevenly spread information could be modelled. Concerning the indicators for agriculture, he examines how, for example, farm size can influence the propensity to accept an innovation. In his studies, Hägerstrand identifies the importance of the proximity and thus comes to deal with private information diffusion and the question of how this should be modelled. He examines the chorological characteristics of information and how migration and telephone data can be utilized to describe the extent of the range of private information. In this particular case study he finds that migration data provides the best approximation as the telephone network was incomplete at the time of the study. Furthermore, there were zonal boundaries that acted to deform the contact field.

The resulting models operated with a real, coordinate based population. The diffusion of an innovation in a population is determined by constructing a so-called mean information field (a concept still outlined in many standard textbooks on human geography), which illustrates how the probability of making contact with another individual decreases with increased distance. The empirical basis for the MIFs comprising 5 × 5 cells (each cell being 1 square kilometre) is grounded on migration data. The matrix illustrates the probability of a contact being made from the central cell to one or more of the surrounding cells. The probabilities are cumulated (from 1 to 10,000) and each cell is attributed an interval proportional to the probability. The matrix is centred on an individual having knowledge of the innovation in question. By drawing a random number one decides which cell (interval) is met. The matrix is used in the manner of a floating grid – i.e. it moves over the fixed population and is centred over those individuals who have knowledge of the innovation in each generation and who are prepared to spread this knowledge. In this manner the innovation is continually diffused over time to new generations and gives rise to spatial patterns of acceptors, which are randomly determined but always within the given probability intervals. Even when the rules of the game and the probabilities remain unchanged, the results of different simulations are often very divergent as a result of the stochastic factor.

In these models Hägerstrand experimented with different forms of physical and social barriers. The propensity to accept an innovation was modelled (for example, a person must be ‘hit’ two or three times before accepting) and the degree of correspondence with the empirically ascertained patterns became increasingly close. As a result of the degree of correspondence between the actual and the modelled courses, Hägerstrand was able to conclude that the key elements deciding the course of innovation diffusion had been captured. It is worth noting that all of these comprehensive modelling experiments were carried out by hand as computers had not as yet been developed. Each of the thousand upon thousand of random numbers was collated from tables in a strict order.

The first international commentary on Hägerstrand’s doctoral thesis was written by John Leighly and published in the Geographical Review in 1954. Leighly, then Professor at Berkeley, was well acquainted with Swedish geography, which he had considered for several years (Leighly, 1952). Leighly emphasized his admiration of the scope and precision of the empirical work, not least concerning the map representation: ‘His “relative” mapping uses refinements (isarithms at numerical intervals given by geometric progressions, interpolation of isarithms by logarithmic intervals) that make it exemplary’ (Leighly, 1954: 440). However, and in a similar manner to many other commentators, Leighly viewed Hägerstrand’s interpretation of innovation diffusion and the following operational modelling as the ‘culmination’ of his work. Leighly (1954: 441) concludes by commenting that anyone doing research in this area ‘cannot afford to ignore Hägerstrand’s methods and conclusions’.

While the attention surrounding Hägerstrand’s doctoral thesis waned in the period following this appraisal, the same cannot be said regarding interest in his theoretical ideas and methodological approach. Aspects of his principal research were gradually disseminated via lectures, conferences and minor publications. Shortly prior to the completion of his doctoral thesis, Hägerstrand laid out his research field in an article entitled ‘The propagation of innovation waves’ (Hägerstrand, 1952). By 1967 he had published three articles on migration, two on diffusion, and one on simulation. A very widely read and often cited article is ‘A Monte Carlo approach to diffusion’ (1965), which directly deals with models and is published in four journals and has even been translated into Japanese.

Innovation diffusion: the English version

Although Hägerstrand’s research had become relatively well known on the international scene, the broader understanding of his research remained fragmentary. Accordingly, both Gilbert White in Chicago and Allan Pred at Berkeley argued that his doctoral thesis should be translated into English. Pred took on this task as he was both fluent in Swedish and well versed in the specific area of research; two attributes which Hägerstrand argued were prerequisites for the successful translation of his work. Pred also wrote a Postscript in which he introduced Torsten Hägerstrand providing a detailed background to the research, and how the field of research had developed.

The book was well received. However, the fact that 14 years had passed since the publication of the original Swedish edition is at least partly evidenced by the content of a number of more critical commentaries. Quantitative geography had further developed and was quickly establishing an important set of new methodologies and theories (see Chapters 2, 3 and 4). One particular line of development concerned the statistical and objective comparison of patterns in time and space, which made it possible to elucidate more precisely differences and similarities in these patterns. In his review, Gunnar Olsson pointed out that ‘the evaluation of the model was based on more intuition and visual inspection than objective statistical testing’ (Olsson, 1969: 310). None of the reviews published during the 1950s had raised this issue.

Simulation had become a popular modus operandi within the discipline, and against this background Olsson had somewhat critically stated: ‘In conclusion, most of those who now experiment with Monte Carlo models seem to have missed the point. Thus it is rare that the relation between theory and model is as explicit as in Hägerstrand’s own work’ (Olsson, 1969: 311). This observation further supports the proficiency with which Hägerstrand had succeeded in fusing thorough empirical material with a well-developed command of theory. This is what gives his work its credibility and strength.

Richard L. Morrill, himself a former visiting fellow at Lund who also had experience of simulations, argued that although Hägerstrand’s earlier work had been disseminated relatively widely, the opportunity now existed for a wider audience to discover that this was by no means a curious approach, but instead constituted a theoretical framework founded on an unusually rigorous empirical grounding. Morrill (1969) went to particular lengths to emphasize the linkages between spatial patterns and individual behaviour, and argued this represented a revolutionizing insight and a powerful break with tradition within the discipline. In a more anecdotal vein, we might also note that he concluded his review in Economic Geography by protesting against the ‘outrageous price’ of the book (then $16).

L.J. Evenden (1969) forwarded a comprehensive and positive review in the journal Social Forces. He argued that: ‘… Hägerstrand allows a glimpse of a fertile imagination and, simultaneously, teaches the useful lessons that patience is a research virtue, that theory is fundamentally about something, and that reliable theory in social science stays close to this something’ (Evenden, 1969: 9). Evenden further argued that evidence that Hägerstrand’s research had left a lasting legacy was to be found in the wide range of fields in which Monte Carlo simulation had subsequently been applied. Simulation was well known and widely used at the time of the publication of the English edition, so there was no question of its subject matter being received as breaking news. Evenden emphasized that the work represented a combination of a careful factual analysis and the best of traditional and contemporary geographical scholarship, and contended that Hägerstrand’s doctoral thesis would long be considered a classic work in the development of geographical theory.

The economist Harvey Leibenstein, then based at Harvard University, reviewed the book in the Journal of Economic Literature. He maintained that it was of interest for economists who were interested in diffusion processes and location theory, as well as modulated processing of innovation diffusion. Leibenstein further contended that it was difficult to assess the value of the models’ predictions, in spite of all the empirical information. He also found the approach altogether mechanical – a critique which one can at least partly agree upon – if one did not take into consideration the fact that the translated thesis was 14 years old. A further point made by Leibenstein was th...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contents
• List of Contributors
• Acknowledgements
• List of Figures and Tables
• Editors’ Introduction
• 1 Innovation Diffusion as Spatial Process (1953): Törsten Hägerstrand
• 2 Theoretical Geography (1962): William Bunge
• 3 Locational Analysis in Human Geography (1965): Peter Haggett
• 4 Explanation in Geography (1969): David Harvey
• 5 Conflict, Power and Politics in the City (1973): Kevin Cox
• 6 Place and Placelessness (1976): Edward Relph
• 7 Space and Place (1977): Yi-Fu Tuan
• 8 The Limits to Capital (1982): David Harvey
• 9 Uneven Development (1984): Neil Smith
• 10 Spatial Divisions of Labour (1984): Doreen Massey
• 11 Geography and Gender (1984): Women and Geography Study Group
• 12 Social Formation and Symbolic Landscape (1984): Denis Cosgrove
• 13 Capitalist World Development (1986): Stuart Corbridge
• 14 Global Shift (1986): Peter Dicken
• 15 The Condition of Postmodernity (1989): David Harvey
• 16 Postmodern Geographies (1989): Edward Soja
• 17 The Capitalist Imperative (1989): Michael Storper and Richard Walker
• 18 The Geographical Tradition (1992): David Livingstone
• 19 Feminism and Geography (1993): Gillian Rose
• 20 Geographical Imaginations (1994): Derek Gregory
• 21 Geographies of Exclusion (1995): David Sibley
• 22 Critical Geopolitics (1996): Gearóid Ó’Tuathail
• 23 Logics of Dislocation (1996): Trevor J. Barnes
• 24 Hybrid Geographies (2002): Sarah Whatmore
• 25 Cities (2002): Ash Amin and Nigel Thrift
• 26 For Space (2005): Doreen Massey
• Index