Theory without history (or geography)
In a recent forum on methodology in physical geography, Hirschboeck (1999) notes a tension between 'theory development' (often equated with mathematical modelling) and 'practice' (often equated with field observation) in hydrology. She argues that a dissociation between these two pursuits can lead to dilettantism, with modellers developing theory for theory's sake and remaining apart from their colleagues in field measurement. She sees future promise in combining both, but more worries that 'there is rarely any hypothesis testing of model performance to evaluate the validity of the underlying theory' (ibid.: 701).
Theory, on this view, is macroscale and expressed in mathematical terms but the crisis arises because its predictions do not accord with the empirical evidence of flow processes collected in the field. Hirschboek's is a so-called positivist view of theory, one common in hydrology and physical geography more widely. What counts as a theory is a set of universal propositions about the behaviour of drainage channels that have been 'tested' against detailed observation of particular cases. The crisis identified in hydrology, or so it seems to me, can be broadened to the whole of geography for its roots lie in a fundamental tension between the universal and the particular (see Chapter 7 in this volume by Burt). The positivist account of theory emphasizes universality, thus treating historical time and geographical space as immaterial in science. Think of Boyle's law and its constant order. The expectation encapsulated in this law is that gases will behave in the same basic way whenever and wherever they are encountered; and, we may add, whoever might be 'testing' their behaviour. Theory in science thus becomes theory without history (or geography). If we are to accept such a positivist view of theory, we must face the serious question of whether geography could ever be either theoretical or a science.
Physical geographers who reflect on these matters recognize a tension between a scientific interest in the general and a geographical interest in the particular (Spedding, 1997). The problem for theorizing in geography is that the traditional scientific view of what counts as a theory appears to exclude both time and space, at least in the sense of historical time and geographical space. There are two main responses to this (beyond denying the possibility of 'geographical' theory ). The first is to challenge the received view of scientific theory by arguing that not all theory worthy of the adjective 'scientific' ignores historical time. Theory in evolutionary biology might provide some ammunition here. Moreover, the argument might be extended to encompass the thornier issue of geographical space by appealing more generally to the time-space context. The second broad response would be an outright rejection of the traditional scientific view of theory as inappropriate to geography and the substitution of a different account of the nature of theory and its role in geographical understanding. Most human geographers have adopted this last strategy and long since abandoned any aspiration to be theoretical scientists in the positivist mode, although it is worth noting that the legacies of positivism are more pervasive than is often recognized. I will examine both these possible responses to the apparent incompatibility of scientific theory and geographical inquiry but, before I do, a word of warning: it is easy to slip into dualistic thinking that equates physical geography with 'hard' science and human geography with non-scientific approaches to theorizing. Indeed, the present discussion may even encourage such thinking. This would be unfortunate because, on the matter of theory, the division between physical and human geography is much more fluid.
Evolutionary theory Charles Darwin is one of the heroes of modern Western science. He presented his evolutionary theory in The Origin of Species in 1859 and it would be difficult to exaggerate the influence of what became known as Darwinism on the understanding of the biological world over the next century. Nor were geographers immune; W. M. Davis's cycle of erosion, Friedrich Ratzel's views on the state as an organism and Harlan Barrows' call for geography as human ecology are a few among many examples of evolutionary ideas percolating into geographers' thinking (Stoddart, 1966). In these circumstances it might seem foolhardy to challenge the scientific credentials of Darwin's theory, although it is pertinent to note that the borrowing of Darwinian ideas by geographers has not had the enduring impact that the original theory still retains in biology.
Darwin's evolutionary theory provided an account of the origins of life that famously challenged creationism. Rather than a world, in all its variety, created by God several thousand years ago, Darwin argued that all life on Earth has evolved by a process of random mutation and natural selection in favour of variants that fit within their niche. To many of us now, this evolutionary explanation seems much more convincing and scientific than the creationist view. But why is this so? On what grounds do we prefer one theory over another? We might try to argue that our preferred theory fits better with the empirical evidence. Darwin's hypotheses entail that human beings evolved from other species. If we assume that primates like the ape are closest to humans, then evolutionary theory leads us to expect (predicts) that there must have been species intermediate in the evolutionary chain between humans and apes. And since Darwin's time evidence of such hominids has literally been unearthed. Evolutionary theory, it would seem, has observational evidence on its side.
The problem that taxes philosophers of science, however, is that such evidence does not seem to justify choosing evolution over creation because the creationists can use exactly the same evidence to support their theory. What if God created the world with all the supposed evidence of evolution already there? The discovery of skulls and bones of the so-called hominids could be seen as part of God's plan and taken as empirical evidence for creationism. Although your inclination may be to think that there is something wrong with this argument, it does highlight one of the difficulties of judging between competing theories that has challenged philosophers. This is the problem of under-determination, namely, that the empirical data are often insufficient to determine which of several theories should be (provisionally) accepted as true. Further, the relationship between theory and observational evidence may be especially problematic for those theories that take context seriously.
Classical evolutionary theory embraces a conception of time as process (a process of adaptation) and space as a general container, without specific geographical location (the ecological niche). Neither time nor space is constitutive of the mechanism of change; cause is separated from context. This provides only the thinnest view of context because the mechanisms of change remain universal. Process geomorphology and geographical climatology take context more seriously insofar as they explore differential ways in which universal processes play out on the surface of the Earth. Scale and scale-switching thus become methodological issues but the universalizing imperative of normative science retains its hold: 'our goal may be to produce methodologies that allow an interpretable, comprehensive representation across all spatial and temporal scales that is somehow simpler and more compelling than a representation that includes all the separate components' (Bauer et al., 1999). Whether or not this is possible, the temptation is to conclude that it is a fit goal for 'proper' science.
The philosopher Karl Popper's project in the philosophy of science began with a search for a demarcation between science and pseudoscience. At one time, he was critical of the theory of evolution because he thought the hypothesis that the fittest species survive was tautological, or true by definition (Ladyman, 2002). In Popper's terms, this meant that the theory was not falsifiable and thus not science. However, Popper reserved his main attack for the social sciences, especially the theories of psychoanalysis and Marxism, which he deemed pseudo-science. His critique of Marxism (Popper, 1945; 1960) challenged the claim, made by Frederick Engels at Marx's funeral, that Marx had discovered the scientific principles underlying the development of societies and raises questions about the nature of social theory.
Marx's ideas about the structure of society (usually referred to as historical materialism) have greatly influenced theoretical thinking throughout the social sciences and continue to be developed by theorists in human geography. In particular, geographers such as David Harvey, Neil Smith, Richard Peet and Dick Walker have used the theoretical framework of Marxism to expand understandings of the structuring of space under capitalism (see Peet, 1991). Marx's historical materialism is only one example of social theory and has many variants but its importance within geography, its claims to scientific status and the parallels (and contrasts) that can be drawn with evolutionary theory in biology make this set of theoretical ideas specially apposite to the current discussion.
Any social theory is directed towards the understanding/explanation of the nature of society. The task assumed by social theorists is to identify, or articulate, the basic components of the social world and the mechanisms that drive social change. In this context, we can see how one set of theoretical ideas can spawn another, leading to new ways of seeing the world. For Marx, society has its roots in the material conditions of life; human beings satisfy their material needs through productive social labour and, therefore, of necessity enter into relations of production. The real foundation of society is thus its economic structure, on which is built a legal and political superstructure. Many philosophers of social science have criticized Marxist social theory as being deterministic, denying the significance of intentional human action in shaping society, and a similar criticism has been voiced by geographers (Duncan and Ley, 1982).
This directs us towards a further question of philosophical interest. Is there a fundamental difference between the social world and the natural world such that the former cannot be explained or understood in the same way as the latter? One argument is that human behaviour cannot be conceived in terms of constant order (i.e. there are no laws of human behaviour) because, unlike volcanoes or rivers, human actions involve intentionality and free will. In short, humans choose to act, volcanoes and rivers do not. If this is so, then it suggests that theories in social science might be rather different from theories in the natural sciences. In the absence of laws, predictions are harder to define and the whole edifice of scientific method begins to seem inappropriate. Somers (1998: 756) comments, 'It is obvious in the social sciences . . . that the test of falsification -in which a single counter-observation can falsify a theory - is virtually never practiced.' She ascribes this absence to the fact that more than one theoretical construction can almost always be placed on a body of evidence. Further, social theories often make claims about unobservable entities, such as social structures, classes or market forces, for which there could only be indirect empirical evidence. Although theories in natural science also make reference to unobservables (think of gravity or electrons), their predictions are, arguably, more amenable to testing. The reason for this, according to Chouinard et al. (1984), is that while research in natural science deals with closed systems, human societies are essentially open systems.
In the absence of a rigorous methodology for testing theoretical claims, or so the argument goes, theories become self-referential; theory guides observation which is, in turn, (mistakenly) taken to validate the theory. Another way of putting this is that observation is theory-laden. The possibility that no interesting observational descriptions are theory-neutral is, of course, as much a problem for the physical sciences as the social sciences but I raise it here to demonstrate the need for careful consideration of this issue in relation to social theory. If empirical evidence cannot decide on the validity of a given social theory in an unambiguous way, then what other grounds might there be for preferring one over the other?
The answer provided by Sayer (1984) proved an attractive option to some human geographers. Sayer argues for a 'critical realist' philosophy of science which contrasts sharply with positivism, most notably in its view of causation. Realism also provides us with an alternative way of warranting social theory. Rather than look to observational evidence for validation (the singular strategy of the positivists), realism emphasizes conceptual coherence and, above all, practical adequacy. Thus a theory can be accepted if it proves an adequate theory to live by. Whether practical adequacy provides a more secure grounding for theory choice than correspondence between theoretical predictions and observational evidence is debatable, but Sayer's realism introduces the possibility that there may be ways of justifying theories beyond the empirical 'testing' at the heart of positivism. Moreover, realism encourages us to theorize about underlying realities beyond the confines of the observable.
There are many theories across the social sciences that offer conceptualizations of society. Marxists theories view social relations as the product of material conditions and thus emphasize class. Feminist theorists focus on the fundamentally patriarchal nature of societies and have criticized Marxists for not giving equal weight to gender (Massey, 1991). Despite substantive differences in content, theorists in this social theory tradition share a common goal in the development of a general theory of society. The impetus for generalization harks back to the claimed scientific status of classical Marxism, and the search for such 'grand theory' has been identified with the modernist project of the Enlightenment (Barnes and Gregory, 1997). It is also a goal that has been challenged by those who doubt the capacity of any single theoretical framework to represent the (social) world.
Was this article helpful?