Directions in Geography

Richard J. Chorley

• Publisher: Routledge
• Year: 2019
• eBook ISBN: 9781000007046
• Topic: Physical Sciences
• Subtopic: Geography
• Index: Geography

Originally published in 1973. This collection of essays looks at the 'quantitative revolution' and the 'new geography' by some of the geographers who had a significant part in those innovations and looks ahead to further developments. The views in the chapters are diverse and offer a fascinating glimpse of the discipline of geography as the subject was undergoing such change and becoming more socially committed. They cover theory, spatial-systems theory, forecasting, human ecology and climatology alongside the teaching of the subject. The concerns of the contemporary geographer come across and are of interest today as these areas have developed still more.

PART I

Theoretical

1· A paradigm for modern geography

BRIAN J.L. BERRY

Time is like a river made up of the events which happen, and its current is strong; no sooner does anything appear than it is swept away, and another comes in its place, and will be swept away too.

Marcus Aurelius Antoninus, Meditations. IV, 43

Several converging threads of dissatisfaction with certain elements in contemporary geography led me to write this essay. In thinking about the implications of spatial field theories I had become increasingly frustrated with what had all too rapidly become “traditional” statistical geography – with the mindless use of conventional inference statistics and measures of association in geographic research without regard for the validity of their assumptions. Many statistical manipulators were ignoring what Dacey was showing clearly enough in the point pattern case: that static pattern analysis is incapable of indicating which of a variety of equally-plausible but fundamentally different causal processes had given rise to the patterns he was studying. This finding seemed generalizable to many other circumstances. At the same time, the “new” environmentalists – today’s ecoactivists – were pointing out how irrelevant it is to theorize about the uniform plain so dear to location theorists. Behavioral geographers were calling for, but not producing, new types of theory. New insights were showing the importance of perceptual filters in decision-making. Phenomenology was indicating the limitations of “the” scientific method in the human sciences, a point all the more telling as large-scale planned intervention has appeared in the attempt to guide social change and modify the systems of concern.

What this paper represents is a personal effort to come to terms with the many sources of confusion and doubt about the continuing viability of earlier research orientations that arise from these sources. It owes much to insights derived from John Platt’s 1970 paper on “Hierarchical Growth”, as well as to the lessons of several years of extra-university involvement in public affairs, confronted by immediate questions of locational and environmental decision-making on the part of city and national governments – i.e. by the real challenges of social relevance. Geography has been characterized as a “mosaic within a mosaic”, essentially pluralistic because of the persistence of a variety of strands arising from diverse origins and changing philosophies (Mikesell 1969). It is tempting, therefore, to suggest that any personal statement simply adds another pluralistic element. I have come to feel that.there is a unity that transcends the apparent disunity, however, and that the essence of this unity, when distilled, will be the basis for more general geographic theory that will call forth reevaluation and substantial rethinking of the partial theories within each of the separate strands. I view the personal statement as a first step in deriving the more general theory.

The plan of the essay is first to spell out the nature of my dissatisfaction with the state of inference in statistical geography. The conclusions of this critique lead naturally to philosophic preference for “process metageography”, and this in turn to a paradigm of locational and environmental decision-making in complex systems that I favor as a guiding orientation for the next generation of geographic research.

Galton’s problem: a ghost that haunts statistical geography

The basic problem, from which all other sections of this essay derive was and is the realization that the determinants of spatial variation may be such as systematically to violate one of the most basic assumptions of the conventional inferential procedures that most geographers rely upon. To clarify, implicit in most uses of correlation, regression, factor and like modes of analysis is the assumption that the observations used in the analysis are independent entities for which certain functionally necessary causal relationships between variables occurring within them are equally and generally true. An equally plausible rival hypothesis at the same aggregate level of generalization is, however, that the observations are elements set within larger systems from which they acquire common characteristics by borrowing or migration, or more generally, through the operation of some spatial diffusion mechanism. To say this is to say that much statistical analysis in geography may be confounded by what has now come to be called “Galton’s Problem” by American anthropologists (Hildreth and Naroll 1972).

Sir Francis Galton raised his problem at the meeting of the Royal Anthropological Institute in 1889 when Tylor read his pioneering paper introducing the cross-cultural survey method (Tylor 1889, 272). Tylor showed correlations (“adhesions”, he called them) between certain traits; in the discussion which followed Galton pointed out that traits often spread by diffusion – by borrowing or migration. Since this is often so, how many independent trials of his correlation did Tylor have?

Galton’s problem, then, is to distinguish the effect of functional associations (“adhesions” in Tylor’s graphic term) from the effect of more common historical association through diffusion. By a functional association is meant a relationship between one or more variables such that the presence of any one of these tends to facilitate the occurrence of any and all of the others within any given area. By diffusion is meant a process involving the acceptance over time of some specific idea or practice (or a set of them, either simultaneously or in sequence) by individuals, groups or other adopting units linked to specific channels of communication, to a social structure, and to a given system of values or culture (Katz, Levin and Hamilton 1963), producing growth that does not appear universally at any one time but manifests itself at points or poles of growth and diffuses in definite channels between areas (Perroux 1955).

Franz Boas, for decades the immensely influential dean of American anthropologists, once told his student Lowie (Lowie 1946, 227–30) that when he first read Tylor’s paper, he became greatly enthusiastic. The cross-cultural survey method seemed to him an ideal research technique. On reflecting further, however, Galton’s objection seemed to him a devastating one; unless there was a solution to Galton’s problem, Boas considered the cross-cultural survey method valueless.

The same may be said about much statistical geography. Consider, for example, a study of the votes cast for a 1968 open housing referendum in Flint, Michigan, published in the Annals of the Association of American Geographers in 1970 (Brunn and Hoffman 1970). As part of their analysis, the authors used the voting precincts of the city as observational units, and they regressed percent favorable vote separately for the black and white residential areas on median incomes, median housing values, median school years completed and distance from the ghetto core. Ignoring the problem of multicollinearity of income, education, housing value and race, they built into the problem an implicit assumption of functional necessity. First, to be able to predict the vote, their acceptance of a regression framework implied that all one needs to do for any precinct is to measure the variables said to operationalize the causes of voting preferences and to solve the resulting regression equation. Such usage of a regression equation implies that it replicates a uniform causal sequence that arises equally and independently within each of the units of observation. The model thus suggests that voting preference is the result of an evolutionary sequence of events – a unilinear combination of causes – that unfolds homogeneously within each unit of observation. Each precinct, for the purposes of such a model, therefore might well be a similarly-structured but a self-contained spaceship – a closed system – floating in a vacuum totally unrelated to other precincts.

But what if the variables used to construct the model of voting behavior are not related in functionally necessary causal sequences, but merely correlated because they are common outcomes of an institutionally structured dual housing market in which housing choices involve comparison of the relative advantages of a wide range of neighborhoods, subject to externally imposed racial constraints, and if voting in certain precincts is affected by their location relative to other precincts of the opposite kind in the dual system? The causal assumptions of the regression model then fail, as does the assumption that the precincts are independent closed systems, and the particular statistical model is clearly inappropriate.

If the variables are symptoms of phenomena that have diffused from one neighborhood to another in historical sequence following paths of functional interdependence, and subject to the barriers of racial constraint, neighborhoods are open rather than closed systems. A different spatial systems framework is then necessary to specify the web of relationships within such observational units, as open systems, are set. Spatial autocorrelation is said to exist. The statistical analysis used must then be appropriate to that situation, and an alternative theoretical stance must be sought appropriate to that alternative situation.

This involves recognizing explicitly in the model that spatial systems display territoriality (areas of organization), formed with respect to points of focus that are structured hierarchically by dominance and subordination characteristics, through operation of distance attenuation mechanisms and boundary effects affecting interaction patterns over regular lines and channels of movement and communication. It also involves assuming some theoretical stance about the elements holding individuals together in systems – and in social philosophy there have been many variants of at least three differing stances (Bell 1971

1. functional interdependence theories, including (a) theories of exchange relations and market mechanisms (Smith) and (b) ideas of hierarchies of stratification based on technical competence (St Simon);

2. value-integrative theories, of which there are at least four – (a) rule by myth (Plato), (b) society as sacred (Burke), (c) society as a moral center (Durkheim), and (d) society defined by ends, traditional or concensual (Locke);

3. domination theories, (a) by traditional or irrational forces (Weber), (b) by the sovereign or the state (Hobbes), and (c) by class (Marx).

And what then of the individual observations? It follows from the above that they can only be understood in a relativistic sense, with respect to the entire structure of the system of which they are parts. If the system changes, so does the relative position of the individuals, which thus have no absolute and independent existence of their own yet which, in combination with all other individuals, define the system of which they are part. How, then, can one proceed at this juncture?

Towards process metageography

It was Emerson who said that “a foolish consistency is the hobgoblin of little men”, and clearly what is called for in the above is not consistent application of any past methodology, but some new paradigm – a new conceptual statement that by its nature admittedly cannot be disproved, yet is open-ended enough to generate new rounds of research activity, serving for this activity as a basic statement of belief.

David Harvey (1969), among others, has argued that one step in producing such a metamorphosis of geography would seem to be that of examining interactions between temporal process and spatial form. To move beyond this suggestion, however (which is what I seek to do in what follows) a conceptual chasm must be bridged, for form can never be absolute. Not only is the “reality” of any element within a system relative to the entire system of elements; it is also time-relative. To seek any fixed thing is to deal in false imagination, therefore, for all phenomenal existence is immediately also seen to be transitory when the dimension of time is added. No particular thing is “real” in any absolute sense; it is passing into something else at every moment. Every individual, for example, is a progressively ageing, temporarily-organized “bundle” of energy flows faced with ultimate disintegration.

To be sure, the search for absolutes of form in some geometric sense is understandable. We perceive the world through screens composed of ideas, and the idea-systems are limited by a language oriented to classifying objects, naming things, and hence codifying their “reality”. Yet what is needed to advance our science is conditional thought that recognizes the relativity of existence and the relative truth of perceptions. Indeed, what is needed is the initiation of a more continuous intellectual process in geography that recognizes that every system and every interpretation needs reassessment in the light of a more complete system.

As a contribution to such a process, I propose that geographic explanation be viewed as dealing with the antecedents and consequences of environmental and locational decision-making in which man, as the prime actor, is viewed as “an information-processing, decision-making, cybernetic machine whose value systems are built up by feedback processes from his environment. These feedback processes are built into the most primitive forms of life, and they form a continuous spectrum all the way back through prehistory and to times when no life existed. Throughout this whole development of man’s history, coming up through biological evolution and extending into cultural evolution, the essential message is one in which disorder, or randomness, is used to generate novelty, and natural selection then generates order” (Potter 1971, 36).

This view implies looking at the world as a complex living system in which individuals, social groups and institutions are dynamically interrelated actors involved in continuing processes of decision-making. The nature, purpose and meaning of any actor and action can only be understood in relation to a field of forces involving other actors and actions. Many actions appear random, but disorder is ordered through responses to their consequences, which may reinforce or change nature, purpose and meaning by changing relations in the system. The behavior of the actors in such an eventuality contributes to equilibrium processes. Another way that the structure is maintained is by making homeostatic adjustments to disturbances; random trends away from system integrity are suppressed by negative feedback. Yet other actions can produce evolutionary shifts in structure by engendering or supporting morphogenetic processes involving gradual growth and change and increasing levels of organization. Finally, and more radical, are those actions that result in revolutionary transformations of structure. In each case, of course, decisions are made in the relational context of perceived organization and structure, and processes set in motion by actions therefore reaffirm...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Title Page
• Original Copyright Page
• Contents
• Preface
• Part I Theoretical
• Part II Spatial
• Part III Environmental
• Part IV Temporal
• Part V Educational
• Part VI Ethical