Abstract
In human geography today little is heard any more about critical realism. A quarter of a century ago, it attracted attention because it seemed to offer a method for radical geographers and was accepted by some at least as such. Yet, despite the claim that Marx could be regarded as a critical realist before his time, there were always very significant differences from the methods of historical geographical materialism. These include different approaches to abstraction, to the distinction between internal and external relations, to causation and determination, to the question of change and to the relative merits of totalizing rather than pluralizing understandings of the world. These differences were poorly understood at the time and have never been critically examined. With a critical scrutiny as background, contrasting geographical practices can be re-evaluated. Although the author’s conclusion is that historical geographical materialism provides more convincing purchase on the world, the critical realist notion of structure provides an important means of understanding the accumulation process and its contradictory character. This is because it is through the elaboration of new structures of social relations or the transformation of old ones that capital seeks to suspend its contradictions. This insight also helps shed light on the very different concepts of space underpinning critical realist and Marxist understandings of human geography.
Introduction
In human geography today one hears very little about critical realism. But 25 years ago it was quite different. Although it was always a minority interest, it was one that appealed to a number of leading figures in the field, including Simon Duncan, John Lovering, Doreen Massey, Michael Storper and Richard Walker. It had been heralded as an elucidation of method fit for a critical human geography (Sayer, 1982). Significantly, and despite claims that Marx had been one of the first critical realists, it met with a skeptical response from some Marxist geographers, most notably from David Harvey. Even so, the issues that it raised endure and, in my view, have never been resolved satisfactorily, at least not in the human geography literature: issues of modes of abstraction, the question of necessity and contingency and of totalizing as opposed to pluralizing approaches to understanding in human geography. The fact that the flame of interest still flickers was apparent in a recent paper by Noel Castree (2002), though I think that, again, things are far from settled.
The encounter should be regarded as of more than passing interest. The critical realist intervention continues to resonate. The uncritical use of the necessity/contingency couplet is one of its more dubious legacies; the easy acceptance that one can understand particular cases in terms of some relation between more general processes and the locally specific is also worrying. Both of these bear witness to a dominant pluralism which sits, or should sit, uncomfortably alongside the totalizing ambitions of historical geographical materialism. On the other hand, critical realism raised issues of just how one might do empirical research; it is something that Marxist geographers have often had trouble dealing with, often showing a remarkable lack of self-consciousness about problems of abstraction or explanation. The issues seem to have been sidelined as people ‘move on’ and exhort others to do likewise, though just how one can ‘move on’ without their critical resolution challenges credulity.
It is not just that the issues raised then have yet to be adequately clarified and settled. It is also that historical geographical materialism can plausibly gain from it at a conceptual level and not just in terms of clarifying its own procedures. Critical realism should not be entirely rejected. There are ideas, particularly its clearly developed idea of structure, which can be engaged with usefully by historical materialists. Notably, it is through the creation of structures, new forms of institutional structure that the capitalist class seeks to suspend the always reappearing barriers to accumulation; just as the working class develops its own empowering social relations in order to resist.
The paper is divided into three major sections. In the first, I provide some brief historical background to the encounter, situating both historical geographical materialism and critical realism and why it was thought, and is still thought by some, that historical materialism was critical realist anyway: in hindsight, a major misjudgment. I then proceed to the major part of the paper identifying exactly why it is a misjudgment and why critical realism does not provide adequate tools for understanding the world. In a final section, I outline their different implications for geographical understanding.
Background
Marxist geography or what would be known later as historical geographical materialism made its initial appearance at the beginning of the 1970s. Only towards the end of the decade did human geographers become aware of critical realism and how it could be turned to effective use in human geography. Both it and historical geographical materialism were in significant part critical reactions to the spatial-quantitative revolution of the 1960s, though in hindsight they were clearly quite different responses. The Marxist critique drew on an understanding of method that saw it as one moment of the social process in general; as reflecting how the world seemed to be to those engaging with it. Marxism was first and foremost a social critique and its critique of method was entailed by that. If the world was somehow deceptive in the way it presented itself to the senses then method could simply reinforce that misunderstanding. Critical realism, on the other hand, was a philosophy of science and its understanding of people in their social relations of secondary interest. In retrospect, and despite the recognition that social objects were fundamentally different from the objects studied in the physical sciences, its naturalism was to get in the way of providing methodological guidelines for the study of people in society.
To situate it further, the entry of critical realism into the jousting fields of debate in human geography is inseparable from a particular moment: this is the moment at the beginning of the 1980s of what was known as ‘society-and-space’ and which would lead to the emergence of a new journal by that name. Marxist geography is part of the background to this development but more as a critical foil. At least two things seem to have been going on. The first were worries about the totalizing ambitions of historical geographical materialism. This had been apparent a year prior to the publication of the first issue of Society and Space in the form of a highly critical paper by Duncan and Ley (1982). This excoriated the Marxist geographers for their structuralism, their economism and their single-minded focus on class at the expense of gender, race and stratum. The latter, of course, reflected in part the emergence of feminism and feminist geography in the course of the 1970s.
The second concern turned on the geographic credentials of Marxist geography. Compared with the work of the spatial-quantitative revolution, space seemed to lose the pre-eminence it had enjoyed. Marxist geography itself seemed to struggle for answers. Social relations clearly conditioned geographies in a very general sense, but how might space condition social relations and how could the two sides be put together in some sort of relationship that was non-deterministic? 1 Only with the appearance of Limits to Capital did things become clearer, but by then the interest in something rather different, ‘society and space’, was already well underway.
Society and space, therefore, came to represent some critical distancing from Marxist geography while at the same time, as we will see, assuming a position closer to the pluralistic predilections of critical realism. Significantly, the editorial board of Society and Space as it was constituted in the first year of its appearance in 1983 included among its 17 members only one defensibly Marxist geographer (Simon Duncan) though both Sayer and Massey were part of the new venture.
The times and the debates also favored critical realism in another way. A significant number of the early Marxist geographers of the 1970s came out of a very particular experience of method, which was that of spatial-quantitative geography. We should recall here the number of people who were to be influential who had this in their background: not just Harvey but people like Michael Webber, Eric Sheppard, John Holmes and Ray Hudson. Added to these was that broad rump of socially concerned geographers, Marxist geography’s soft underbelly, which included other one-time spatial-quantitative geographers, people like David Smith and Richard Morrill. The experience of method of all these people had been quite narrow: method as technique or kitbag of tools. The positivism of the spatial-quantitative revolution was now in question but so too was what should fill the void. Even while it was clear in David Harvey’s work, particularly in the early 1980s, it was to take quite a while before Marx’s dialectical method sank deep roots among the Marxist geographers and for them to shake off the mantra of ‘method = technique’. Critical realism seemed to offer an answer and this was certainly how Andrew Sayer saw it in one of his earlier papers (1982: 68–69.)
It also helped that a case could be made for some methodological convergence between Marx and critical realism. Abstraction was fundamental for critical realists and through his own methods Marx seemed to endorse what was believed to be the ‘correct’ way. Marx placed social relations and the causal properties they entailed at the centre of his approach to understanding; the way in which the accumulation process was necessitated by the class relation is an example. Yet as something explicitly engaged with by human geographers, by the early 1990s critical realism’s star was clearly in decline. Its last hurrah was possibly the book jointly authored by Sayer and Walker (The New Social Economy) that appeared in 1992, and by 1993, Sayer, who had been a crucial missionary influence, had moved into a Department of Sociology with no one in human geography able or willing to take on the mantle.
This obviously begs the question as to why there was indeed no one to step into Sayer’s shoes. One argument could be that human geography was by then moving in directions quite incompatible with critical realism, though interestingly ones that Marxist geography would survive. The influence of the posts may have been important in the move away from critical realism. Feminist geography is a case in point. The idea of gender as a distinct structure of social relations entailing particular interests and practices had been an important one but this was displaced by the attractions of identity construction and the politics of difference. I think this is significant because critical realism, while providing analytic penetration, had in practice little to say about struggle. There was also the view, voiced forcibly by Harvey (1987: 373), that the application of critical realism could, in practice, result in a shift in the balance between concerns for universality and particularity towards the latter. 2 This is not to downgrade the value of those studies that set out to, in effect, deconstruct some of the generalizations that had become far too accepted for comfort (England, 1993; Sayer, 1985a, 1985b; 1986; 1989); Massey (1985: 12) also helps us understand how the concern for the particular came out of the ‘society and space’ moment, particularly her concern for ‘local variation and uniqueness’. But with some exceptions, like the book by Sayer and Walker referred to above, that move was not counterbalanced by a critical realist exploration of the analytic potential of particular abstractions.
Even so, the critical realist intervention did have some lasting effects on human geography. ‘Contingency’ entered into the dominant vocabulary and so remains; consciousness was changed in a lasting manner. Furthermore, the notions of pluralism secreted by critical realist method are entirely consistent with the prevalent ‘political economy’ approach in which the state and the economy can be two separate social relations connected in a contingent rather than necessary manner, as in the way in which human geographers drew on regulation theory. But all this is anticipating rather than confronting in detail the fact that critical realism is no answer. There are ways in which critical realist assumptions can facilitate concrete research in historical geographical materialism but the way in which they need to be appropriated is entirely subordinated to historical geographical materialism as a method, as I now wish to show.
The issues
The question of abstraction: empiricist/historical
In human geography, as in other human sciences, a crucial step in investigation is that of abstraction. As Marx famously said “… in the analysis of economic forms neither microscopes nor chemical reagents are of assistance. The power of abstraction must replace both.” (1867: 90). Critical realists would surely agree with this assessment. Yet Marx’s approach to the question and that of critical realism turn out to be quite different. 3 At first glance, this might seem a puzzling claim since at times there appear to be significant convergences. The notion of a continuum between the more abstract and the more concrete, the single-sided versus the many-sided, is deeply embedded in critical realism. So the human being is gendered, then gendered and classed and so on. In Marx, the same sort of hierarchization seems to be going on and, in fact, might appear to be the predominant aspect of his approach. This is very clear in Capital Volume One, where Marx continually shuttles between the ahistoric and the historic, between the labor process in general and the specifically capitalist labor process, between class society and the class society that is capitalism and so on. Yet on closer examination, and as Gunn (1989) argued, Marx here has been quite crucially misinterpreted.
Gunn distinguishes between two forms of abstraction. There is what he calls ‘empiricist abstraction’, which is the critical realist way, and there are determinate or what have also been called substantive (Gunn, 1989) or systematic (Johnson, 1982) abstractions. The latter is what Marx favored. Empiricist abstraction is abstraction from the concrete or many-sided to the more one-sided. It corresponds to what Gunn has called a “move away from (concrete) reality” in the direction of “the less real and the more purely conceptual” (Gunn, 1987: 62). 4 Ironically, it is the characteristic of both critical realism and the positivism that the former has been at such pains to critique. Sayer in particular was keen to highlight the contrast between what he called the chaotic conceptions to which positivism was so susceptible and the rational abstractions of critical realism. Positivism, consistent with its concern for form or what could be observed, abstracted in terms of what Sayer called relations of formal similarity. The ‘correct’ approach, according to Sayer, was that of rational abstraction. This abstracts on the basis of substantive relations between the things being investigated. So in this instance, one would have to separate out people not in terms of formal similarities such as their poverty but in terms of relations to others, such as those, often diverse, agents from which they receive their money, whether it is the state, an employer or a ‘breadwinner’. Unfortunately, and given the way in which in capitalist countries state revenues presuppose a private economy and ‘breadwinners’ likewise, rational abstractions turn out to be just as much an ‘abstraction from’, just as empiricist therefore, as those that are so common in mainstream social science.
From one angle it might seem that Marx also engaged in this form of abstraction. I have referred above to his use of hierarchies of abstractions: his constant counterposing of the historic to the ahistoric in particular, but also his recognition of antiquity, feudalism and capitalism as concrete forms of class society. His purpose here, though, as Gunn (1989) emphasizes, is only in part analytic. In understanding capital, it helps to know that in all social formations provision must be made for the reproduction not only of the individual who will do the producing but also of the means of production used up in the process of production. This paves the way, inter alia, for the distinction between simple and expanded reproduction and, using the latter, for an explanation of the accumulation process as a historically specific form of expanded reproduction (Marx, 1867: Chapter 23).
Marx’s primary purpose when resorting to empiricist forms of abstraction, though, is that of critique. What he wants to dispel is the reduction of capitalist practice to the ahistoric and therefore to the natural. Capitalism is a social and therefore historical creation. The contradictions to which it gives rise, the sordid, hideously exploitative practices at its core, cannot be explained away as the laws of the world; as somehow natural and therefore inevitable. 5
Rather, the dominant form of abstraction in Marx is what has been called ‘historical’, ‘systematic’ or ‘substantive’. They are abstractions created in and through practice. They are abstractions that correspond to sociohistorically specific forms of practice and therefore to the real and that are therefore a moment of a more concrete – because many-sided – whole. As Gunn emphasizes, in this way “… that which is abstract can be a mode of existence (a form) of that which is historically specific and no less real than any other aspect of the concrete totality in which it inheres.” (Gunn, 1987: 62)
The idea of relative location as we think of it and practice it today is very distinct and inseparable from the sort of capitalist society in which we live. Some conception of relative location must be very old and certainly, judging from the existence of maps both in pre-capitalist Europe and pre-Columbian societies, antedating the emergence of a form of society in which self-sufficiency gave way to a chronic dependence on commodity exchange. The idea of locations as substitutable for one another, as in the real estate mantra ‘location, location, location’, is in all likelihood much more recent; perhaps no older than the full blossoming of property capital as a distinctive part of capital’s division of labor and the emergence of the multi-locational corporation, along with the location consultants that the latter has entailed.
These different approaches to abstraction have important implications for how one views the idea of levels of abstraction. The effect of empiricist modes of abstraction is to produce levels that correspond to a hierarchy of increasingly abstract concepts as one ascends and therefore increasingly concrete ones as one goes in the other direction. Accordingly one recreates the concrete by combining the more abstract concepts one with another: a set of nested preconditions starting from the top. Work in human geography has clearly drawn on this notion of hierarchies of abstractions and presented itself as Marxist. The work of Gibson and Horvath (1983) is an early example, 6 as is that of the Cox and Mair (1989). It is an approach endorsed by Sayer in a reference to what he calls ‘Marxist research’ (1992: 140–143). It also features in his (1989) work on fordism and post-fordism, where they take on the status of more concrete forms of capitalist society.
The alternative that inheres in the processes of historical abstraction is to retain a logical ordering of concepts but not in terms of increasingly thin abstractions; rather their ordering is determined by what happened in history: what was a necessary precondition for what. The state cannot be independent of the material conditions which make it possible and so it exists on a lower level of abstraction: relations of production distinctive to a particular epoch have to be addressed before one can begin to understand that epoch’s state forms, which means that the state is never just ‘the state’ as an empiricist abstraction but a historically specific slave, absolutist, capitalist, or whatever, state. In talking about capital, one has to abstract surplus value before one abstracts the different forms in which surplus value is expressed and which depend on its appropriation at the point of production – rent, interest and profit of enterprise; accordingly they are at a lower level of abstraction. And abstract labor has to be abstracted prior to concrete labors whose specific character is determined by the need to convert it into abstract labor.
That Marx was capable of thinking in terms of the thinnest of abstractions that might appeal to advocates of empiricist abstraction is indisputable. They were essential supports for understanding more concrete modes of production; a mental check if nothing else. Yet he also recognized its dangers. In particular, it bore the problem of situatedness in a capitalist society where things did seem separated from the rest of the world. From that standpoint, empiricist abstraction comes naturally. For Marx, that problem of situatedness was the solution: abstractions, including, nota bene, concepts of abstraction, had to be situated with respect to a particular social formation, its practices and the concepts through which people made sense of them. To apply them to other social formations was to dehistoricize and to miss their specificity and therefore their fundamental logics and to be tempted into apologetics. But exactly why things seemed separate under capital and so amenable to methods of empiricist abstraction requires an examination of Marx’s ideas of internal and external relations and how they contrast with those of critical realism.
Internal and external relations
Central to the approach of critical realism to the question of abstraction is, indeed, a certain conception of internal and external relations. The objective of abstraction is to identify structures in which the parts are, by definition, internal or necessary to the modes of acting entailed by that structure. To include things that are externally or contingently related and which only affect the expression of those modes of acting and not the mode itself, is to engage in chaotic conceptions of the whole rather than rational abstractions. The point here, though, is that Marx also drew on ideas of internal and external relations. This, in turn, has further encouraged the claim that Marx was a critical realist. Unfortunately his notions of internality and externality depart in significant respects from the critical realist understanding. This needs emphasizing.
For Marx an internal relation is a relation between a subject and its objective conditions of existence: i.e. a subject–object relation, as Bologh (1979) has emphasized. The objective conditions of the subject are those through which a subject produces itself and its knowledge of the object. The subject–object relation is conceived as a property relation: the relation of the subject to its social and material conditions as its own. As Marx makes clear in the Grundrisse, this was grasped as such in pre-capitalist social formations:
… the worker relates to the objective conditions of his labor as to his property; this is the natural unity of labor with its material presuppositions. The worker thus has an objective existence independent of labor. The individual relates to himself as proprietor, as master of the conditions of his reality (1857–1858: 471).
and: The individual relates simply to the objective conditions of his labor as being his; [relates] to them as the inorganic nature of his subjectivity, in which the latter realizes itself … but this relation to land and soil, to the earth, as the property of the laboring individual – who thus appears from the outset not merely as laboring individual, in this abstraction, but who has an objective mode of existence in his ownership of the land, an existence presupposed to this activity, and not merely as a result of it, a presupposition of his activity just like his skin, his sense organs, which of course he also reproduces and develops, etc. in the life process, but which are nevertheless presuppositions of this process of his reproduction – is instantly mediated by the naturally arisen, spontaneous, more or less historically developed and modified presence of the individual as member of a commune – his naturally arisen presence as member of a tribe, etc. (1857–1858: 485).
In other words, this relation of property is, in its turn, presupposed by some form of social organization: the family and the extended family as clan or the commune. It is through membership of this that the individual acquires possession of the material conditions of production (p. 472).
Capitalism is then the dissolution of these relations so that the objective conditions of the individual’s self-production and reproduction, both material and social, now appear as external to her: as not presupposed to her existence but existing, thing-like, acting indifferently with respect to the human agent, and to be appropriated as necessity determines. The objective conditions of material life appear to be independent of any subjectivity, while subjects appear independent of any objective conditions of existence. The means of production, objects and instruments of labor, now appear in the (social) form of capital; as something external to the worker which employs her as objectivity-lacking labor power: a being without any objective existence in the form of land or means of tilling it, livestock and the like. Subject and object are now split apart with a reversal of subject and object. The object in the form of capital now employs the subject in valorizing itself. Capital seemingly becomes the subject and the worker the object in the form of abstract labor power.
If the objective conditions in the form of object and instrument of labor now appear as external to the individual, as thing-like in the form of capital and so lacking presuppositions, this is also the case with the social relations into which people enter in order to produce.
The wage relation appears now as what Marx called ‘external necessity’: as necessary from the standpoint of the individual but contingent as to whether or not wage work will actually be found. This is entirely different from the experience of pre-capitalist formations where the social relations appear as natural, as inherited, as something a person is born into whether as a member of a tribe or in a relation of servility. On the one hand, the individual is liberated from the connections conceived as natural and unavoidable and experiences a degree of freedom through the labor market. On the other hand, the social relation appears as thing-like, as something to which the individual must adapt. So while one can choose one’s line of wage work, one cannot choose not to so choose. 7
To repeat: the objective conditions of the individual’s being now appear as external, as acting indifferently with respect to the subject. What are being related, the relata, now seem self-sufficient, without necessary preconditions and therefore without history and in consequence, one might add, without historical geography. Object-less subjects and objects that appear to be independent of any subjectivity: just existing as capital with their own logic beyond the control of anybody. As such, the relata essential to the production of the individual can enter into contradiction with one another: sellers without buyers and buyers without sellers; capital that remains uninvested; debts that cannot be paid; factories without raw-materials. In short, limits are experienced which have to be suspended if production is to continue to develop. Institutions, the relation to nature, discourse and so on, have to be transformed so as to bring the development of the individual parts of the whole into conformity with the accumulation process: with, as Harvey puts it, the need to achieve 3% growth (2010). The totality gets reproduced, per necesita, and it is through re-totalizations that this occurs.
This is a very different view of internal and external relations from that of critical realism. For Marx, relations of internality and externality are subject–object relations and so necessarily bound up with the study of people in their relations to nature and to other people; his project is, in other words, a science of humanity. This sits uncomfortably alongside the naturalist claims of critical realism where what is taken to be a couplet, internal and external, can apply regardless of the involvement of subjects with objects. Marx’s approach is also and characteristically historical: what is an internal relation between people and their objective conditions of existence comes to appear as external only under particular historical circumstances; therefore, a relation between things without historical presuppositions. In contrast, the empiricist abstractions of critical realism are ahistoric. And finally, the use of the word ‘appearance’ is crucial. What is internal only comes to appear as external while remaining necessary: Marx’s relation of external necessity. In critical realism, on the other hand, talk of internal and external is necessarily a reference to the real and not to how things appear.
Totalizing/pluralizing
Marxism is a totalizing understanding of the world. While the world is conceived as a unity in which the constituent parts are what they are in virtue of their relations to the whole, as a totality, therefore, it is a question of relations subject to centrifugal forces and that are always in the process of being reformed: a constantly reoccurring tension between part and whole, therefore, that is resolved by the totalizing force of a whole that itself changes in its concrete form in response to changes in its constituent parts. Critical realism also has a relational view of the world but as we have seen, it draws an important distinction between what it calls internal and external relations; otherwise expressed as necessary and contingent relations. Crucially, this both confirms and expresses its view of the world as essentially pluralistic or, as it might express it, ‘differentiated’: as made up of numerous structures of relations, each generating the distinctive causal properties of the elements composing them, but contingently related to other structures of relations: in other words, relations that are possible but not necessary. Accordingly, and as in Sayer and Walker’s The New Social Economy (1992) the division of labor is regarded as having causal powers distinct from those of capital and therefore a separate structure. 8 As such, critical realism finds itself sharply opposed to historical materialism.
At this point two qualifications should be entered regarding what it means in historical materialism to talk about totalities and totalization. First, it does not mean a deterministic understanding of the world. The future remains open and cannot be predicted. The different parts can and may develop in relatively autonomous ways; but only ‘relatively’. The totalizing impulse, the necessity for coherence or what more orthodox views would call ‘social order’, originates in production; everything starts with production and returns to it. But production is always social. Under capital, the necessity to produce takes the form of the necessity to accumulate. In an important sense the different aspects of the totality are all conditions or relations of production (A Sayer, 1987; D Sayer, 1987) and are so structured. This structuring occurs in a way that allows some autonomy of development that will in turn condition capital’s development. In consequence, the totality is not something static but always changing through the process of totalization. Autonomy of development results in contradiction just as it throws up new opportunities through which the challenge of those contradictions can be met: changes, say, in the state or in gender relations that, with some reworking, can be exploited by capital as it confronts its contradictions. There is no inevitability here. What is certain is that without flows of value through those social relations, flows of value that depend on the continuation of the accumulation process, they will cease to exist: mere cul-de-sacs of history and for that matter, given its internal relation to the social process as a whole, of geography too.
The second point is that, as implied above, this is a centred totality. This means not just a totality in relation to production but more specifically to the actions of the producing agents. It is the producing agents who give the totality its coherence and reshape it, and who are reshaped by it, albeit in and through those others that through their actions constitute the whole. As Marx and Engels expressed it in the German Ideology:
… life involves before everything else eating and drinking, a habitation, clothing and many other things. The first historical act is thus the production of the means to satisfy these needs, the production of material life itself. And indeed this is an historical act, a fundamental condition of all history, which today, as thousands of years ago, must daily and hourly be fulfilled merely in order to sustain human life (1978: 48).
and: This mode of production must not be considered simply as being the production of the physical existence of the individuals. Rather it is a definite form of activity of these individuals, a define form of expressing their life, a definite mode of life on their part. As individuals express their life, so they are. What they are, therefore, coincides with their production, both with what they produce and with how they produce. The nature of individuals thus depends on the material conditions determining their production (1978: 42).
People produce their world: not just their means of subsistence and the means of producing those means of subsistence but also their social relations, their culture, their politics and so on.
To consider the world as a totality is to consider it from the standpoint of production and its objective conditions, physical and social, political, discursive, spatial and so on, all conceived as moments of production, as Marx makes clear in his discussion of production in its relation to distribution, consumption and exchange. 9 Under capitalism, these conditions are viewed as separate, as thing-like, pursuing their own logics, developing in their own way, providing both challenges and opportunities. The important point though is that given their internal relations to production – and they are internally related only through production – their uneven development results in contradiction. As a result the totality is pregnant with change.
Contradiction and change
Critical realism recognizes the fact of change in the world. Without it, its concept of the openness of systems, so crucial for its critique of the humean concept of cause, would be meaningless. It has had difficulty, though, in providing some understanding of why change comes about. Historical materialism is quite different. The crucial concepts here, as we may have deduced from the discussion of totality and totalization, are those of contradiction, its suspension and its eventual abolition. Contradiction looms large in the work of some historical geographical materialists; that of Jamie Gough (1996a, 1996b, 2002; Gough and Eisenschnitz, 1996) is a case in point.
Central to Marx’s concept of contradiction is the unity of opposites: a unity which opposes the producing individual, conceived socially, as an irreducibly social individual, therefore, to the conditions of production. The latter serve to simultaneously facilitate productive activity and to undermine or inhibit it. It is then the suspension of the contradiction through some change in the conditions of production (institutional, technological, etc.), which allows production to develop further, though raising new barriers to the expansion of production. Marx’s treatment of the division of labor in German Ideology provides an example. The development of the social division of labor and the assignment of particular moments of the social process to particular institutionalized roles, as in the separation of mental from manual labor and the emergence of state forms, is both predicated on the production of a surplus and facilitates the expansion of that surplus. But the different parts of the social division of labor then become centres of independent action that can inhibit and even undermine production. The totality is threatened and totalization through the suspension of contradictions is therefore a necessary counterpart.
Under capitalism, the immediate producers encounter capital as their necessary condition of production: something external to them but as per the earlier discussion ‘an external necessity’. As with all such externalizations, there is a history here as I discussed in the same section. What Marx called the ‘natural unity’ of pre-capitalist formations where the worker relates to his or her objective conditions as belonging to him or her, as his or her ‘inorganic body’, the dissolution of these relations gives way to the unity of opposites that is capital and labor: labor power and means of production are then brought together by those with money wealth.
Capital now becomes the necessary condition for the reproduction of the immediate producer. Labor in effect reproduces itself by working for a wage. In this sense, capital is a support for that reproduction. But as a result of capital’s own drive to reproduce itself, subsequent downward pressure on wages and the replacement of workers with machinery, the ability of workers to reproduce themselves is undermined. Historically this was the condition for the development of the labor movement, resistance to wage cutting and the drive for a welfare state: in other words, attempts to suspend the contradiction. Meanwhile, of course, an essential condition for the reproduction of capital is labor power. But through the expulsion of labor power from the labor process it removes that one condition of production capable of producing more value than it costs and so able to produce the capitalist’s profit. This in turn, and highly schematically, results in attempts on the part of capital to suspend the contradiction to its advantage, including expansion elsewhere. 10
Unlike historical geographical materialism, critical realism has no concept of contradiction. On the other hand, the unity of opposites has been of interest even while the preferred term has been that of dualisms – individual and society, culture and economy, economy and polity and so on. How these have been treated by critical realism is significant. The overall goal seems to be to soften the contrasts and explore/define/elaborate the mutual dependence of the two terms comprising the dualism; how they are, in critical realism’s terms, internally related and to show how, far from having independent causal powers, they can be subsumed by a single causal structure. Sayer’s (1997) examination of what he called the ‘The Dialectic of Culture and Economy’ is a good example of this approach. Lacking in it, though, is any sense of the creative tensions inherent in these relations; and any investigation of where the dualism came from: its history, its preconditions, what the world had to be like for people to start thinking in terms of these dualisms. 11
Towards the end of the 1970s, a dualism that particularly captured the imagination of human geographers was that of agency and structure or, alternatively expressed, individual and society. This in fact was part of the background to the emergence of the interest in ‘society-and-space’ discussed earlier. The view was that during that decade, human geographers had either opted for the structuralism of Marx in the form of Marxist geography 12 or the unstructured agency of humanistic geography. Salvation came in the form of Gidden’s structuration theory and the closely related 13 ‘transformational model of social activity’ (TMSA) advanced by the dominant force in critical realism Roy Bhaskar. Both recognized the problems inherent in the dualism and proposed a solution emphasizing the relational character of the terms: that people could only be agents in virtue of their social relations; but without that activity, those social relations would not be reproduced and could not be transformed: in short, a relation of mutual dependence. 14
As it happens, Marx himself addressed this dualism at various points in his work and very differently. For a start, he was acutely aware of how the dualism had developed through history in response to changing production relations 15 and how understandings of the relationship had changed with the advent of capitalist social relations. In this regard, the opening passages of the Introduction section to the Grundrisse are particularly helpful. The notion of development is also crucial. Marx saw human beings as developing themselves through their labor; the empirical record could mean nothing else. It was not, however, enough to appeal to a distinctly human nature. 16 Rather development occurred through the successive emergence of contradictions, their suspension, their abolition and replacement by new production relations, only for the process to be repeated.
Causation and determination
External relations such as cause and effect contrast with internal relations conceived as determinations (Bologh 1979: 52).
…while causation is certainly a particular kind of determinative relation, it is not the only such determinative relation, nor the most important (Roberts, 1999: 39).
At one level the conceptions of causation embraced by critical realism and historical materialism are very similar. Both reject the atomistic conception of cause embodied by the humean thesis; a conception in which effects are registered quantitatively without changing the qualities of the objects doing the causing or of those affected. Instead both adhere to an action theory of causation. Change occurs because people act. They act on and with other things or conditions, but these too should be seen in their relation to human agents: as products of a human action that also transforms them and gives them new properties. At the same time, the agent herself is changed through the action. This view is apparent both in Bhaskar’s TMSA and in Marx’s writings, particularly those on the labor process in Chapter 7 of Capital Volume 1 and in his discussion of capital as labor’s product. 17
Seen closer up, the differences become more evident. In the critical realist notion of causal structure, a separation is made between an underlying reality and the empirical world. The explanatory problem is to explicate the empirical in terms of a structure of relations and the contingent circumstances in which it operates. This, it should be noted, is entirely in accord with the distinction between what critical realism defines as internal and external relations, and as was discussed above.
Historical materialism also has views of the world as layered but it is a very different one. Rather, the distinction is not between the empirical and the real but between form and content or what Marx described variably as appearance/reality, form/essence, appearance/essence and form of manifestation/hidden connection. While appearance/reality might appear redolent of the distinction made by critical realism between the empirical and the real, what Marx is saying is actually quite otherwise. The appearances to which he is referring are, rather, ideological forms, and unlike critical realism’s ‘empirical’ forms, they are ones rooted firmly in the dominant, historically defined, social relations of production. They are forms of reality, even if illusory ones. 18
This again takes us back to the earlier discussion of internal and external relations but this time to the distinctive way in which Marx defined them. Externality is an appearance: the historical result of the dissolution of relations to nature and to others that were once apprehended as internal – as relations that were necessary presuppositions for the individual. It is an appearance apprehended as the thing-like nature of the social conditions of production; of all production relations. Things having their own historical trajectory, therefore, and entering into contradiction with one another but only, as Marx was quick to point out, provoking transformations that would reassert the centrality of production or, in the case of capitalism, accumulation. These include, inter alia, the state, the family, space, culture and the division of labor. These are appearances that get reproduced by what I referred to earlier as critical realism’s empiricist abstractions. Critical realism conceptualizes them, along with other appearances at what it would regard as lower levels of abstraction, as structures of relations but they break up the underlying unity of the social totality and in that way serve to reproduce its fragmenting tendencies.
Critical realism’s explanatory task follows on from this: To uncover in thought a reality in the form of a structure of social relations that entails particular causal properties and that is therefore the necessary condition for the empirical world; and then to show how it combines with contingent conditions to produce the observed form with which one started. An example in human geography would be the attempts of Massey and Sayer to unravel the relation between assisted area status in Great Britain and the geography of employment growth (Massey, 1984: 12–13; Sayer, 1986b). In other words, one works back from appearance to the real, drawing on structures of relations abstracted in the empiricist way outlined earlier.
The explanatory task for historical materialism is quite different. One might indeed start out with some empirical form, as in critical realism; something like geographically uneven development or the changing form of the city. But this is then situated with respect to the contradictory development of capital and its production relations. What in a critical realist analysis are taken to be contingent conditions now appear as conditions constitutive of a contradictory whole to be mobilized, re-harnessed, and transformed so as to facilitate the lifting of barriers to the accumulation process. What are seemingly contingent conditions get reworked and reassembled into new structures of relations that can empower in ways that allow individual capitals or capital as a whole to meet the challenges confronting them.
The analytical move is therefore also a critical one since it shows how what seemed autonomous and external to capital is an aspect of its contradictory development: a production relation that is the focus of the ongoing totalization process. The corresponding critical move in critical realism is the critique of the empiricist: of empirical regularities or the classifications typically engaged in in mainstream analysis.
Historical geographical materialism would recognize the pertinence of the sorts of causal structures that critical realist research has been so effective at unearthing: the causal structures underpinning just-in-time manufacturing or Massey’s (1984) new spatial divisions of labor, or structures like the one mediating inward investment which Cox and Wood (1997) referred to as the local economic development network. It would, however, want to situate them with respect to the contradictory social totality. At the heart of this is the accumulation process, constantly suspending contradictions through the formation of new structures of relations and just as constantly re-encountering them later in a new form or provoking the creation of new structures on the part of the working class – the class structures of the labor movement like labor unions, cooperative retailing, workers’ friendly societies, political parties, neighborhood organizations, though also structures that can be turned to capital’s advantage, like those conditioning Hagerstrand’s chain migration. These then have to be responded to with still newer structures of social relations, empowering capital, if it is to complete its circulation through its various moments. This is an approach, therefore, which is essentially dynamic, which seeks to elucidate the contradictions behind the formation of causal structures and their supersession and which therefore embraces as well as explaining change. So while the procedures of critical realism can be usefully brought to bear on the identification of structures, it is their relation to capitalist development, indeed their capitalist nature, which needs to be foregrounded. Once created structures can perform useful functions for capitalists or indeed for the working class, they facilitate particular sequences of cause and effect. This should not, however, result in a marginalization of their determination by the accumulation process and its inherently contradictory dynamic.
Practicing human geography
Every sentence in Spatial Divisions of Labor is so laden down with a rhetoric of contingency, place, and the specificity of history, that the whole guiding thread of Marxian argument is reduced to a set of echoes and reverberations of inert Marxian categories. The categories are seemingly treated as a dead weight into which no new life can be breathed through exploration of the vast new field of historical and geographical experience that Massey opens up for consideration (Harvey, 1987: 373–374).
Harvey counterposes to realism and other approaches, ‘the tough rigor of dialectical theorizing and historical analysis’. I know it is a long time since he has worked on questions of philosophy and method, but what we need is not reiteration of such tendentious phrases in the hope that we will eventually believe them but a thorough analysis of the epistemological, ontological and methodological arguments behind such claims (A Sayer, 1987; D Sayer, 1987: 398).
I want to turn now and consider the implications of the arguments developed above for the practice of human geography. I do this through a brief characterization of representative claims of major figures: on the one hand Doreen Massey and Andrew Sayer and on the other David Harvey. This is followed by a recuperation of the major differences in approach and how they follow on from the contrasts identified above.
Massey and Sayer
Central to both Massey and Sayer was the idea that we can understand human geographies through the conjoining of general processes or causal properties with contingent circumstance. A major concern was to reject the positivist notion that empirical regularities have explanatory import. This is very evident in the work that Massey did in the 1980s on spatial divisions of labor. The general process could be the formation of new firm spatial divisions of labor or some aspect of it like the female-employing and low skill branch plant. Alternatively, the analysis might be in terms of some need common to a set of firms, like the particular locational needs imposed by their technologies or by their social relations, as in relations of competition. In Spatial Divisions of Labor (1984), she explored the relation between the formation of new spatial divisions of labor for firms and their particular geographic expression. Correlatively, what she wanted to do is show how instead of understanding geographic change in terms of geographies – as in geographies of government location incentives – we have to examine the various socially entailed powers and needs of firms and how these, in combination with various contingent conditions, result in new spatial divisions of labor for them.
Some of Sayer’s work was along similar lines. The paper he contributed to the Storper and Scott collection (1986) included analyses that were similar to those of Massey. We learned, for example, of the difference that contingent conditions can make to the concrete effects of the exercise of causal properties. So while the US electronics industry has successfully decentralized semi-conductor manufacture to the Far East, this has not been possible in Western Europe even while the same conditions of deskilled labor processes and reduced transport costs applied in both instances. The contingent condition in question: high duties on their import in the Western European case compared with ones that are virtually non-existent for American importers.
Harvey
The accumulation process has always been central to Harvey’s Marxist work. Capitalism strives to mould geographies to its own needs of accumulation: ‘to abolish space with time so as to speed up the rotation of capital through the construction of highways, railroads, electricity grids, pipelines and the like; and to exploit the productive advantages of scale through the construction of concentrated physical infrastructures – cities, factories, airports, dock facilities’ (1975, 1985b). At the same time, through working, if to a variable degree, with what already exists, it transforms the geography of social infrastructures: the governance systems through which risk can be reduced, exchange lubricated and labor relations stabilized. As production concentrates geographically, so the division of labor can be deepened. Vertical disintegration occurs and constellations of interrelated firms come into being to constitute territorial production complexes. In short, in order to valorize itself, some fraction of capital has to be invested in fixed facilities of long life. Social infrastructures also have to be formed and these have their own forms of fixity owing to the substantial opportunity costs of relocation that they impose, both for firms and for workers.
Capital seeks to subordinate geography to its own logic of expanded reproduction or accumulation, therefore (much as it seeks to subordinate technology, the state, the family and not least the working class), and it certainly achieves a degree of success in this regard. But at the same time, it must, of necessity, revolutionize geography. It must develop new transportation and communication technologies, new products that may shift the balance of advantage between one place and another or result in the creation of entirely new industrial spaces elsewhere. In consequence, it invariably creates tensions between fixity and mobility and therefore between the old and the new and between here and there. Capital differentiates itself, not least geographically, in order to solve its contradictions, but this results in re-posing them in a new concrete form.
Discussion
In review, I think there are three crucial contrasts to be drawn here. The first is the greater sense of dynamism, of human geographies as always in a process of becoming, of dissolving and being recreated afresh, which is conveyed by Harvey’s work. This might seem an odd claim to make given Massey’s more recent work (1999, 2005) in which she emphasizes the openness of human geographies and the role that random juxtapositions of events play in laying down the conditions for that openness. But while she is good on the conditions for change, it is less clear why some conditions are mobilized and others seemingly filtered out of the social process and why some mobilizations are successful and others less so.
Second, for Harvey the social process always has geographical aspects that are necessary to it rather than contingent; social relations can never be anything other than spatial. This means that capitalism as an ongoing, self-transforming mode of social life cannot be understood outside of geography. Capitalism is not just a social process; it is a socio-spatial process and cannot be theorized apart from space. This is in some contrast to aspects of the critical realist approach. In his paper ‘The difference that space makes’, Sayer famously concluded that “Abstract theory need only consider space insofar as necessary properties of objects are involved, and this does not amount to very much” (1985a: 54). In order to understand the difference that space makes, concrete research into specific geographic configurations of objects, forces and practices was necessary. This, however, assumes a pluralistic understanding of the world from the start; the idea, in this instance, that the space relations are external to capital.
Third and finally, for Harvey the human geographies under study are specifically capitalist geographies. Capitalism is the driving, form-creating force. It is the contradictions of capitalism that engender the dynamism of capitalist geographies, their ceaseless inconstancy as Storper and Walker (1989) put it, as capital seeks to suspend its contradictions only to see them reappear in a new form. There is in short a strong sense of what drives social change, what brings the various social relations together into the new juxtapositions or happenstance – or perhaps not so happenstance – arrangements that Massey has highlighted and which in turn generate new mutations in capitalist geography. This is surely entailed by Harvey’s concern for the totality and the ongoing totalization that is a response to contradiction; something that is in sharp contrast to the highly circumscribed nature of the clusters of relations examined by Massey in her book on Spatial Divisions of Labor.
Even so, this does not mean that historical geographical materialism as a method has nothing to gain from an awareness of at least some of the claims and research strategies of critical realism. There is the possibility of a limited – a very limited – rapprochement and it is with that I want to conclude this paper. The critical realist concept that I believe to be relevant to historical geographical materialist research is that of structure. Structures entail causal properties. People are enabled or limited in various ways because of their substantive connections with others. Massey’s spatial divisions of labor are there for a purpose and it is a capitalist purpose whose realization they facilitate in varying ways.
Yet as we have seen, critical realism imposes a particular view of the world: one which precludes situating structures with respect to a developing, contradictory totality. Evidently that does not have to be. Rather, they can be regarded as coming into being to serve some purpose in the struggles around accumulation, whether those of the capitalist and working class; or those between factions that are spun off by the inevitable tensions generated by the accumulation process. There is, in other words, a history – more accurately a geographical history – of structures awaiting investigation. 19 This is one which would certainly focus on the geohistorical conditions necessary to their coming into being, as in Sayer’s discussion of the electronics industry briefly referenced above. But it would also focus on their necessary conditions in the shifting dynamics of the accumulation process, conditions possibly entailed by previous ‘solutions’ to its contradictions, but always to be understood in terms of their own contradictory nature and eventual supersession or transformation.
It is also important to be clear on the status of those particular geohistorical conditions that get mobilized and reworked in attempts to suspend emergent contradictions. In this regard, a recent attempt of Castree (2002) to arbitrate between the contrasting views of space held by Sayer and Harvey ultimately runs aground on a failure to completely understand the totalizing claims of historical geographical materialism. In his own words, “… both thinkers are saying something right about space: that is, the difference that space makes can be necessary and/or conditional since there is nothing contradictory about holding to either or both positions simultaneously” (p. 189). Rather, he argues, a proper appreciation of the difference that space makes requires work at a variety of levels of abstraction (p. 209). According to this view, Harvey works at a higher level of abstraction, demonstrating how the dynamics of the accumulation process work themselves out over space in the abstract rather than in particular places with all their contextual variety. Sayer, on the other hand, believes that concrete research into particular cases is necessary since the sort of abstract claims of people like Harvey are so general as to shed little light on how things work out in practice. Castree is not about to accept Sayer’s limited view of how historical geographical materialism has incorporated space into its claims; but nor is he willing to accept Harvey’s refusal to work at lower levels of abstraction. As a result, for him “It is, contra Sayer possible to make theoretical claims about space. But contra Harvey, we have to recognize that, in practice, the necessary properties of space are realized in and through contingent conditions” (p. 209).
Yet it seems to me that this approach is too critically realist for comfort. The particular notion of levels of abstraction drawn on is the critical realist one of empirical abstraction that entails a pluralist understanding of the world as I pointed out earlier. Accordingly, Castree refers to how “… only concrete research will be able to account for how capital, patriarchy, racism, the division of labor, etc. articulate in and through particular geographical landscapes” (p. 209; Castree’s emphasis) and to “the contingent question of how real spaces are necessarily co-determined by and co-determining of a plethora of relatively autonomous relations and structures.”
Likewise, the notion of contingency drawn on in Castree’s account is, as it has to be, given this pluralist approach, the flat, un-dynamic one of critical realism, not something that is acted on, transformed and incorporated into the ongoing social process as a necessary aspect of it as it is manifest in particular places. Sayer has claimed that capitalism has a high degree of context independence. This is because it seems to be able to exist in a vast range of situations. As a result, the settings cannot be considered necessary aspects of capital. Yet crucially this is to ignore the way in which capital works on those contingent conditions to, as Harvey puts it, ‘bring the unique qualities … of given places … into a framework of universal generality’. (1985b: 45) It does not take things as it finds them but transforms them, bringing in the missing and complementary parts, perhaps, so that they can be effective in the valorization of capital.
In short, space or spatial arrangement is always socially produced. It is never, qua Sayer, just a spatial arrangement of objects which then trigger off, via their proximity, causal powers and liabilities. As an example, we might consider the case of locating a new shopping centre. On the one hand, there is a profit-seeking corporation that needs appropriate sites; on the other hand, a set of locational possibilities determined by properties of accessibility and the cheapness of the site: in other words, contingent spatial arrangements over which the corporation seemingly has no control. But this contingency, of course, is quite erroneous, not least because capitals, including property development companies, seek to change the map of accessibility. They lobby for new freeway interchanges, sometimes paying for them, themselves. Or they build the housing that will then provide the market for the shopping centre. As for the price of the land, they can and do exercise some control over it by land banking: calculating the geographic trends of development so as to know where to buy ahead of development. They then lower their holding costs until the land is ‘ripe’ by taking advantage of tax rules originally introduced to benefit farmers and keep land in agriculture. In other words, it is not just spatial arrangement that is at issue but a capitalist spatial arrangement: far more than a contingent condition vis-à-vis property capital since it was, in various ways, internalized by them. As Harvey and Scott emphasized: “We need to show … how particular contingencies that on first sight appear as external and arbitrary phenomena are transformed into structured internal elements of the encompassing social logic of capitalism” (1989: 19; my emphasis).
Accordingly, concrete research of a historical geographical materialist sort should focus on how various configurations of agents and conditions are selected in, how they are articulated one with another, become the object of struggle and transformed into a structure of relations that can function as a necessary aspect of the circulation of capital as it seeks to suspend the barriers inherent in its own nature. The locally specific is not accepted by capital as such, and nor could it be. Capital, pace Sayer, is not indifferent to context. It has to work to transform it in accord with its totalizing impulse so that it becomes something quite different and internal to capital. Structures get constructed and draw on the conditions encountered in particular localities or regions and they get constructed to respond to the geohistorically specific forms in which capital’s contradictions are being experienced at a particular time. To research a history of structures is therefore to situate them with respect to those specific forms and how they came into being. At the same time, it is to explore the ways in which the institutional fixes arrived at will themselves get challenged, defended, reworked, as capital’s contradictions assume new forms: new configurations of the fixed and the mobile, for example, as Harvey (1985a) outlined them in discussing the geopolitics of capitalism.
Footnotes
Acknowledgements
The author would like to acknowledge the extremely helpful comments of the reviewers of an earlier version of this paper. Any remaining weaknesses in the argument are entirely the author’s responsibility.
Declaration of Conflicting Interests
None declared.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.