BOUNDING SPACE: PURIFICATION  AND CONTROL
To be sure a certain theoretical desanctiflcation of space (the one signalled by Galileo's work) has occurred, but we may not have reached the point of a practical desanctiflcation of space. And perhaps our life is still governed by a certain number of oppositions that remain inviolable, that our institutions and practices have not dared to breakdown. These are oppositions that we regard as simple givens: for example, between private space and : public space, between family space and social space, between cultural space and useful space, between the space of leisure and that of work. All these are still nurtured by the hidden presence of the sacred.
(Michel Foucault)'-
So far in this discussion, space has been hovering on the margins. I will now suggest that, in order to understand the problem of exclusion in modern society, we need a cultural reading of space, what we might term an 'anthropology of space' which emphasizes the rituals of spatial organization. We need to see the sacred which is embodied in spatial boundaries. In the quotation above, Foucault implies that a desanctiflcation of space is occurring in western societies. This lags behind the desanctiflcation of time, he suggests, but is an inevitable consequence of modernization, the progress of materialism and rationality. 1 doubt that this is the case. There seems to me to be a continuing need for ritual practices to maintain the sanctity of space in a secular society. These rituals, as in ancient Israel or Brobdingnag, are an expression of power relations: they are concerned with domination. Today, however, the guardians of sacred spaces are more likely to be security guards, parents or judges than priests. They are policing the spaces of commerce, public institutions and the home rather than the temple.
In Chapter 3, I indicated that there were parallels between social behaviour in small, high-density collectivities generally described as traditional societies
PURIFICATION   AND   CONTROL  /   7 3
and behaviour in present-day gemeinschaft-like groups. The liminal discourse of social anthropology developed in the context of traditional societies, concerning boundary rituals, taboos, and so on, can be used to illuminate the modern problem, although I would not advocate an exclusive disciplinary perspective. In the following account, however, I will try to identify the 'curious rituals' associated with the social use of space in developed societies, inter-leaving broad theoretical concepts and particular details of individual and group behaviour. Before looking specifically at the interconnections of spatial structures and social exclusion, however, we might consider a few general issues involved in unravelling socio-spatial relationships.
STRUCTURATION   THEORY AND SPATIAL. THEORY
While, in the history of modern geography, the nature of the relationship between people and the environment has been one of the more enduring concerns of practitioners, interest in the question faded in the 1960s when space was reduced to the primitives of distance and direction and served essentially as a neutral medium for the operation of social and economic processes. Foucault's observation about the treatment of space in the western philosophical tradition seems particularly apposite as a comment on the treatment of space in human geography: 'Space was treated as the dead, the fixed, the undialectical, the immobile. Time, on the contrary, was richness, fecundity, life, dialectic.'2 Subsequently, an interest in structure in the materialist sense has led to a revived interest in environment, particularly in the built environment as a product of capitalist development. Conceptions of the way in which the environment affects and is affected by human activity have been presented by several writers recently, including Allan Pred3 and Ed Soja,+ who draw on Anthony Giddens's structuration theory. Pred, for example, asserted that
Place . . . always involves an appropriation and transformation of space and nature that is inseparable from the reproduction and transformation of society in time and space. As such, place is characterized by the uninterrupted flux of human practice — and experience thereof — in time and space.
This sounds impressive although the writer is not saying anything particularly
GEOGRAPHIES   OF   EXCLUSION   / 74
remarkable. The problem is that geographers in their earlier grossly simplified spatial geometries had neglected the obvious.
Giddens's account of structure and agency in the constitution of social life provides one point of entry to this problem. Although his structuration theory is now treated as rather passé in human geography, some of his ideas are useful in the sense that his conception of structuration provides cues for the unravelling of socio-spatial relationships. While Giddens seems to me to have a rather naive view of space, working with a few key arguments from his general thesis, we can begin to give shape to a socio-spatial theory of exclusion.
Giddens's theory of structuration is concerned with social relationships which are both fluid and concrete, and it is an argument which can be readily spatialized.5 His first proposition is that human activities are recursive, that is, 'continually recreated by [social actors] by the very means whereby they express themselves as actors'. Second, the reproduction of social life presumes reflexivity in the sense that 'the ongoing flow of social life is continually monitored'. The monitoring of social life, however, also includes the monitoring of the physical contexts and the broader social contexts of experience. These contexts have structural properties which are 'both medium and outcome of the practices they recursively organize'. As this suggests,: structure does not just constrain activity but is also enabling, although the agency of actors, their capacity to affect the circumstances of their existence, will not be equal in relation to all the structured properties of the social system. Some of these structured properties 'stretch away in time and space, beyond the control of individual actors'.
In addition to location, which, as Giddens implies, embodies a set of structuring spatial and temporal relationships, we can recognize the built environment as a relatively stable element of the socially produced environment which provides the context for action. Here, the reciprocity of human activity and its context is fairly obvious. As Arthur Miller said about society, 'The fish is in the water and the water is in the fish.' This observation, banal as it is, captures a characteristic of the built environment which is still neglected in much urban geography, however, with space represented too often as an inactive context for something else, the 'where' in a Kantian tradition, dead space. Giddens himself is not very clear on this question. At one point, he implies that spatial structures serve only as containers for social interaction. Thus, in a passage selected by Nicky Gregson:
PURIFICATION   AND   CONTROL.   / 75
Locales refer to the use of space <o provide settings for interaction . . . Locales may range from a room in a house, a street corner, the shop floor of a factory, towns and cities, to the territorially demarcated areas occupied by nation-states. But locales are typically internally regionalized and the regions within them are of critical importance in constituting the contexts for interaction [my italics],6
As Gregson notes, this says nothing about 'the nature, form or content of the setting', or, I would add, about the interaction of people and the built environment. Later in the same work, however, Giddens claims that 'space is not an empty dimension along which social groupings become structured, but has to be considered in terms of its involvement in the constitution of systems of interaction'.7 It is this assertion which is echoed in geographical accounts of structuration theory, by Pred and Gregson. It is, then, important to contextualize structuration theory, to recognize how particular social and spatial outcomes are tied to particular cultures, to particular histories and to individual life experiences.
While structuration theory points to the reciprocal nature of the relationship between people, as individuals and social groups, and their environment, it still leaves a problem of explanation, which has been identified by Steve Pile.8 That is:
after the division of'the social' into structure and agency {or into context and intentionality), structure (context) is seen as external while agency <intentiona!ity) is seen as internal. The effect of externalizing structure is to make it taken-for-granted (not yet known) and impersonal (denying the personal in die social).
However, as I suggested in my earlier outline of object relations theory, structure is internalized (through introjection) and shaped by the unconscious (through projection). Again, to quote Pile:
Psychoanalytic theory, in its theories of the unconscious, describes how the social enters, constitutes and positions the individual. Similarly, by showing that desire, fantasy and meaning are a (real) part of everyday life, it shows how the social is entered, constituted and positioned by individuals.
1 do not think that this explanatory gap in structuration theory makes it necessary to abandon it, and it has particular value in defining the problem of power. Recognizing that people have a capacity to change their environment and, more generally, that individuals retain some autonomy as thinking and acting agents, leads to the question of the distribution of power within social
GEOGRAPHIES   OF   EXCLUSION   / 76
systems and of spatial structures as embodiments of power relations. As Moos and Dear observe:
Power relations are always relations of autonomy and dependence and are necessarily reciprocal. The distribution of power in a relationship may be very assymetrical but an agent always maintains some control in the relationship and may escape complete subjugation.'
Control by dominating agents may seem complete, but there is always the possibility of subversion. The prison, as possibly the most dominating control environment, demonstrates the existence of autonomy in the most adverse conditions. Michaei Ignatieff,10 for example, notes that in Pentonville prison in London, designed in the 1840s as a model of the total institution, the impossibility of total control was recognized after a few years of a very harsh regime of hard labour and solitary confinement. The resistance of the prisoners led to a moderation of the system. There was, then, more than a flicker of human agency which altered the relationship between the institutional environment and the inmates. We cannot understand the role of space in the reproduction of social relations without recognizing that the relatively powerless still have enough power to 'carve out spaces of control' in respect of their day-to-day lives.1'
We can envision the built environment as an integral element in the production of social life, conditioning activities and creating opportunities according to the distribution of power in the socio-spatial system. For some, the built environment is to be maintained and reproduced in its existing form if it embodies social values which individuals or groups have both the power and the capacity to retain. For others, the built environment constitutes a landscape of domination. It is alienating, and action on the part of the relatively powerless wilt register in the dominant vocabulary as deviance, threat or subversion. This contrast suggests that power relations are transparent, however, when they are not. In the routines of daily life, most people are not conscious of domination and the socio-spatial system is reproduced with little challenge. There are some groups for whom exclusion is a part of their daily experience, who will be highly sensitive to alien environments, but their spaces of control are too small to interrupt the reproduction of socio-spatial relations in the interest of the hegemonic power.
An appreciation of power relations gives meaning to space. Variations in the control and manipulation of different spatial configurations reflect different forms of power relations. As Foucault maintains,
PURIFICATION   AND   CONTROL  / 77
A whole history remains to be written of spaces - which would at the same time be the history of powers (both of these terms in die plural) - from the great strategies of geopolitics to the little tactics of the habitat.12
The range of spaces which should be of interest to the human geographer interested in power relations is somewhat wider than that which has conventionally constituted the geographer's terrain, however. In particular, I will suggest that personal space defined by the self and the intimate spaces of the home are integral elements of social space. These private spaces have a relationship with the public spaces of geography - they are reciprocally conditioned, and it is the process of reciprocal conditioning which requires illumination if we are to understand problems like the rejection of difference in localities.
EXCLUSIONARY SPACE
1 will argue that 'spatial purification' is a key feature in the organization of social space. Michel de Certeau recognized this problem as the creation of 'clean space' in Utopian and urbanistic discourse. He argued that:
In this site [the city] organized by 'speculative' and classifying operations, management combines with elimination; on the one hand, we have the differentiation and redistribution of the parts and function of the city through inversions, movements, accumulations, etc., and, on the other hand, we have the rejection of whatever is not treatable and that, thus, constitutes the garbage of a functionalist djminijiratron (abnormality, deviance, sickness, death, etc.) [my italics].'i
He continues with an observation that is close to Stanley Cohen's view of social control which I discuss later in this chapter, namely, that 'progress, of course, allows for the reintroduction of an increasing proportion of the wastes into the management network and the transformation of those very flaws . . . into means for strengthening the system of order'.
This argument, which clearly resonates with the notion of abjection and pollution, needs to be given a more explicit economic dimension. We can see that the imperative of accumulation under capitalism has made developed societies centres of consumption within the global economy, and the way in which consumption is promoted, the process of 'want creation' identified by Galbraith, contributes to purified identities and feelings of abjection in relation to the 'other'. Fred Hirsch argued that the market economy in developed
GEOGRAPHIES   OF   EXCLUSION   / 78
societies encouraged 'the strengthening of self-regarding individual objectives', and consumer advertising, he suggested, comprised 'a persistent series of invitations and imperatives to the individual to look after himself [sic] and his immediate family'.14 Thus, the never-ending invitations to consume further the privatization of the family, which is closed off from the outside world. Life beyond the home enters the private sphere through stereotyped images, conveyed by videos, television commercials and similar media messages. Within the private world of the home, advertisers foster a negative view of soiled goods and a positive view of new, completely packaged domestic environments, clearly in order to maintain the levels of demand for domestic products. The imagery of this advertising is significant. It often promotes cleanliness, purity, whiteness and spatial order, images reflecting the idea of a pure inner self as, for example, in the Persil commercial described in Chapter 4, features which Freud associated with civilization and the sublimation of instinctual feelings. Unsullied whiteness is also associated with a germ-free environment so that a concern with maintaining a state of pure whiteness becomes a virtue - mother (usually) has to fight germs by keeping the house clean in order to protect her children, notwithstanding the fact that some of these 'germs' are necessary for health. Thus, the consciousness of dirt and disorder is increased and we can anticipate that a feeling of abjection will be particularly strong in those environments, domestic interiors, neighbourhoods which arc symbolically pure. It is the identification of numerous residues, to be expelled from the body, the home and the locality, which is characteristic of this purification process. In such environments, difference will register as deviance, a source of threat to be kept out through the erection of strong boundaries, or expelled.
THE   FORM   OF   PURIFIED SPACE
The anatomy of the purified environment is an expression of the values associated with strong feelings of abjection, a heightened consciousness of difference and, thus, a fear of mixing or the disintegration of boundaries. This is one of several possible maps of social organization which we can construct, drawing particularly on schemata developed by Basil Bernstein. Bernstein's project was concerned with control in educational systems, but his ideas have a particular resonance in relation to the question of exclusion. He recognized:
PURIFICATION   AND   CONTROL   / 79
an affinity with Mary Douglas, who approached what was essentially the same problem through an analysis of the rituals surrounding purity and defilement, and, recently, similarities have been noted in the writing of Mary Douglas and Julia Kristeva.15
Bernstein's interest in educational sociology has been primarily in language and the curriculum, where he has produced a classification which links academic subjects (or other social objects) with modes of control, and it is this scheme which links the social and the spatial. It provides us with a means of identifying exclusionary structures. His general thesis shows the influence of Dürkheim, particularly the latter's distinction between mechanical and organic solidarity. Thus, according to Atkinson,'6 Bernstein characterizes a social organization displaying mechanical solidarity as one which is segmented: 'members are arranged in relatively insulated, self-contained units' and 'roles are... ascribed in terms of a small number of primitive categories'. Conversely, organic solidarity is expressed through increasing individualization and a 'weakening of boundaries which formerly defined structural segments'. Bernstein does not use the mechanical-organic dichotomy in a temporal sense, signalling a change from traditional to modern, however. The terms are used instead to indicate different forms of organization within modern institutions. Mechanical solidarity, like gemeinschaß, is a characteristic social Form in developed societies.
The particular form of the mechanical and the organic are presented in an educational context. Here, Bernstein represents the school curriculum as a number of subject areas insulated from each other in different degrees, according to the prevailing ideology. First, in 'Open schools, open society?' he distinguishes two polar types of curriculum organization which have the characteristics shown in Table 5.1.17
Table S.I Characteristics of open and closed curriculum organization
Open
Closed
1	Ritual order celebrates participation and	Ritual order celebrates hierarchy and
	cooperation	dominance
2	Boundary relationships with outside	Boundary relationships with outside
	blurred	sharply drawn
3	Opportunities For self-government	Very limited opportunities for self-
		government
4	Mixing of categories	Purity of categories
GEOGRAPHIES   OF   EXCLUSSON   / SO
In a later paper,'8 he rephrases the problem, using the terms classification and framing to describe the characteristics of mixing or purification in curricula. With strong classification, the contents of subject areas are strongly* bounded and kept separate, while strong framing suggests a clear distinction between what may and may not be transmitted within subjects. Decisions; on what is permissible come from above and inter-subject communication is minimized. Conversely, with weak classification, subject boundaries are weakly defined and there is less concern with the singular and distinct identities of subjects, and weak framing allows the transmission of a wide range of ideas within a subject. Strong classification and strong framing tend to go together, as do weak classification and weak framing, although alternative combinations are possible. When the curriculum is strongly classified, new ideas on pedagogy or academic content are seen to be; threatening because they challenge the hierarchical control structure. A; weakly classified system, by contrast, is a tolerant one in which new ideas-are absorbed. They do not threaten non-hierarchical power relationships precisely because power is diffuse. A hierarchical power structure does not like ambiguity. Fuzzy boundaries between subjects in the school curriculum;;: for example, suggest communication between subjects which could represent a challenge to those in power. Therefore, some knowledge within a strongly classified system would be seen as 'dangerous knowledge', to be suppressed, ignored or rejected, jf it did not fit the classification.19 This is characteristic of polluting objects and ideas in Mary Douglas's thesis. They do not fit a society' s classificatory system.
Bernstein's educational model provides a clear analogue for the structuring; of social space. Thus, we can speak of strongly classified space, where there is; internal homogeneity and clear, strong boundaries separate that space from others. Alternatively, we could identify a strongly classified spatial systems-consisting of a collection of clearly bounded and homogeneous units, like land-uses in a city or the rooms in a house. The contents and arrangement of the; contents of strongly classified space, like the furniture in a room, would be; strongly framed if there were inflexible rules determining those internal; arrangements. Difference in a strongly classified and strongly framed assem--blage would be seen as deviance and a threat to the power structure. In order to minimize or to counter threat, the threat of pollution, spatial boundaries; would be strong and there would be a consciousness of boundaries and spatial; order. In other words, the strongly classified environment is one where abjection is most likely to be experienced. Strong classification will reinforce
PURIFICATION   AND   CONTROL   / 61
feelings of abjection and the two may be recursively related. Weak classification and framing as forms of spatial structure would be associated with social mixing, a tolerance of difference and little interest in boundary maintenance. It is also possible, but less likely, that strong classification will be combined with weak framing. Alaszewski,20 for example, drawing on Douglas rather than Bernstein, describes fluid and relaxed regimes, incorporating a wide range of therapies, in some of the wards of a mental hospital. The hospital as a structure and in its institutional organization is strongly classified but, within it, there are instances of weak framing. Bernstein is concerned with polar types and in practice we might expect some problems in classifying environments and forms of social organization which do not match his model. He docs provide us, however, with a basis for connecting social structures and spatial structures and, at the same time, we can make his model relate to psychoanalytical, anthropological and economic theory at the point where Freud, Kristeva, Sennett, Douglas and Hirsch converge.
SPACE   AND   SOCIAL CONTROL
Bernstein provides a link between exclusionary processes which are rooted in family and group relationships and exclusion which has its source in institutional practices. Classification and boundary maintenance are characteristic of both, and families, communities and institutions are all implicated in the construction of deviance and the exclusion of deviant individuals and groups. For the moment, however, I want to examine socio-spatial exclusion as a part of the more general question of social control, with particular emphasis on controls exercised by agencies of the state. Social control is a term which has varied usage, but what I will be concerned with here is the attempted regulation of the behaviour of individuals and groups by other individuals or groups in dominant positions. Specifically, I am concerned with constraints on social interactions and the use of space which result from the actions of social control agencies.
Davis and Anderson suggest a scheme for classifying social control systems, containing several elements which are relevant to the social control problem (Table 5.2).21
In this scheme, they distinguish between those controls which are external in origin and which are transmitted hierarchically and those which are
GEOGRAPHIES   OF   EXCLUSION   / 82
Table 5.2 A classification of social control systems
Mode of control
High perrasivencs;
Low pervasiveness
External
Internalized norms and values
External and internalized norms and values
Asylums, prisons Transformative groups, for
example, Alcoholics
Anonymous Traditional families and kin
groups
Bureaucracies, firms Self-help groups, for
example, Weight
Watchers Professions (law, medicine,
etc.)
internalized, in that members of a group make a commitment to norms and values. The dimension of pervasiveness separates the total institution with a control or corrective function, like a prison, and the conformist community, on one hand, and organizations where control is an unstated objective, on the other. The externally controlled/highly pervasive category is the one which has most immediate relevance to the problem of exclusion. In Bernstein's terms, we are concerned here with hierarchy, strong classification and, by implication, a high level of visibility for those identified as deviant. However, while this scheme has heuristic value, control regimes should not be thought of as fitting into discrete categories. Thus, the asylum and the prison, rather than being considered exceptional, should be thought of as models which have a wider application in society even though they may assume a more muted form. In particular, pervasiveness should be thought of as a continuum rather than a dichotomous variable. This is the essence of Foucault's argument in Discipline and Punish, a text that has generated considerable discussion in the social sciences, including geography.22
Foucault's thesis is that the discipline of a highly controlled institution like a prison or a nineteenth-century asylum 'represents a continuation and intensification of what goes on in more ordinary places',23 and that the controls which are embedded in ordinary life legitimate the kind of regime practised in a prison, for example. 'AH micro-forms of discipline are functional to a larger system', as Michael Walzer puts it. Foucault's particular vision of a controlled society originates in Bentham's panopticon, which was a model for: a totally controlled institution, designed on the principles of discipline, surveillance and hierarchical classification. The panopticon was a prison/ factory, so designed that the controller could remain invisible and at a distance from the inmates yet control their lives in detail. The panoptic principle,
PURIFICATION   AMD   CONTROL   / 83
however, extends much more widely than this. As a metaphor for control, the panopticon 'inserted the power to punish much more deeply into the social body'.24 It 'colonizes' social life and erects boundaries between the normal and the deviant at all levels, irrespective of legal codes which define criminal behaviour. Thus, control, discipline and carceral forms of punishment are diffused through society and social control on the panopticon principle becomes much more than confinement under a particular regime: 'The prison is only one small part of a highly articulated, mutually reinforcing carceral continuum extending across society in which all of us are implicated, and not only as captives and victims'.35
This is a bold claim and, as an account of the geography of social control, it warrants critical examination. The first problem concerns the generality of the panoptic principle. Prisons, asylums and workhouses, associated particularly but not uniquely with the disciplining of the proletariat in the nineteenth century, could be seen as useful instruments for the spatial exclusion of deviant and unproductive groups at a time when the Benthamite principle of getting the maximum return from labour encompassed the factory and the institution. Thus, institutions like the prisons or the magdalens, hostels for the confinement of 'prostitutes', were places for closely supervised work as well as for the correction of deviance. The geography or spatial design of these institutions varied. They were not necessarily planned on the panoptic principle, although they did generally exhibit strong classification. Thus, in his detailed account of buildings designed for 'formation', like schools, and for 're-formation', like workhouses, in capitalist societies during the eighteenth and nineteenth centuries, Thomas Markus argues that:
In all these places, order is based on stable categories of people, objects and activities, together with a set of rules — much stronger and more explicit than in other buildings — which govern their interactions. They establish diurnal, weekly, and seasonal timetables and shifts, and they specify the duration and repetition of events. The rules are, equally strongly, built into space and its management. They define the location of persons and things, they control the paths of movement and the degree of choice as well as the visual paths, they define programmed encounters and place limits on those occurring by chance. Time and space are joined in rules which govern the opening times of specific spaces. In short, the building and its management determine who does what, where, with whom, when and observed by whom."6
We can also see the finely meshed network of control, represented physically in the design of the prison or the school, extending to other locales and to other social groups who might interfere with the efficient performance of the
GEOGRAPHIES   OF   EXCLUSION   / 84
capitalist economy. This is particularly evident in specialized spaces, those which, like institutions, are based on an explicit ideology. Thus, the Utopian: creations of nineteenth-century capital, like Robert Owen's New Lanark or Titus Salt's Saltaire, extended the discipline of the workplace to the residential sphere in that tenants were selected for their respectability and conformity to! the ideals of the community. In the highly ordered space of the Utopian! settlement, deviance would have been conspicuous. A more recent example is the settlements provided by the state for minorities whose presence interferes^ with the exploitation of resources by capital and whose values are in conflict with the materialistic, progressive values of capitalism. Planned settlements for! Australian Aborigines, native Canadians and some European travelling people,; for example, express the state's interest in separation and the correction of deviance." Locations are selected which remove the minority from areas!! valued by the dominant society and, in isolation, the design and regulation of! space are supposed to induce conformity. The regularity of the design, the high!: visibility of interna] boundaries which interrupt traditional patterns of social organization, make what is culturally different appear disruptive and deviant.!; As in prison, power and domination are expressed in arbitrary rules and transgression warrants the imposition of sanctions, including eviction in the case of many English Gypsy sites. These schemes fail because there is no awareness of the capacity of the minority to resist and to maintain its own cultural values, but they do demonstrate the need of the state to secure theV interests of capital through socio-spatial control of 'deviance' and cultural difference.
The mentally ill and mentally disabled, the criminal and the racially different are all in varying degrees 'other' and beyond the bounds of normal society according to some narrower definitions of normality. Do we, however, create spaces for the disciplining of groups within mainstream society, extending the finely meshed network of domination into areas of social life occupied by the majority? Rather than thinking of the problem as one of inserting panoptical controls into the social body, I think we should recognize the reciprocal conditioning of individuals and families, on the one hand, and social institutions on the other. Object relations theory, as I indicated above, suggests that the tendency to reject difference and to value order is characteristic of the pathological personality but that this tendency is also evident in the development of the balanced, well-integrated personality. I would argue that institutional controls, manifest in schools, bureaucracies and, physically, in organizational systems like land-use planning, reinforce this tendency. If this is
PURIFICATION   AND   CONTROL   / 85
the case, it is not surprising that small features of the urban landscape, like children's playgrounds and parks, and many domestic interiors, are also supervised, controlled spaces which signal exclusion. Because many of these controls are taken for granted or register negatively only in the world-views of others, like children, who have little power to influence the design of the spaces which they have to negotiate, we, that is, the dominant majority, are implicated in the perpetuation of the carceral control system.
As a metaphor, the carceral city, in which all of us are trapped, either as agents of domination or as victims, or both, has considerable value. The recognition that socio-spatial control is not restricted to particular and obvious sites of exclusion gives geographies of the asylum movement or the geography of prisons particular potency. Characteristic geometries and patterns of domination appear widely. Although there is not much room for agency in Foucault's thesis, since we are all apparently trapped in the carceral net, there is a connection between his structuralism and Freudian theories of the development of the self. Thus, Foucault's view that 'Men and women are always social creations, the products of codes and disciplines',2" can be reconciled with the view of Erikson, Klein, Sennett and others on the production of the social self, where the other assumes both material and social forms which are articulated in rules of exclusion. This gives the thesis a more general significance, at least within developed, capitalist societies, although we should be wary of ignoring cultural difference and generalizing too far.
Foucault's analysis of social control is depressing. We are left feeling helpless. A similar conclusion might be reached from reading Stanley Cohen's Visions of Social Control.29 Cohen challenges the view that exclusion, separation and isolation are necessary features of social control. He suggests, rather, that programmes designed to bring the 'deviant' back into the community result either in the reconstruction of group conflict at a different scale or more insidious modes of inclusionary control, which are less likely to be challenged because they are relatively benign and liberal. He maintains that
when .. . boundary blurring, integration and community control take place, the result is that more people get involved in the 'control problem' . . . more rather than less attention has to be given to the deviance question. In order to include rather than exclude, a set of judgements has to be made which 'normalizes' intervention in a greater range of human life.'0
Thus, more 'humane' penalties, like electronic tagging or community service, involve more people in the corrective and caring professions, they may involve the vetting of the families of recidivists, and they reduce awareness of control
GEOGRAPHIES   OF   EXCLUSION   / 86
and tht: criminalizing of behaviour. 'Modern inclusionary .social control becomes a system of "Weepers, screens and trackers", part of the "invisibly controlling city'".1' At least, the strong boundary between the prison as a site of exclusion and 'normal' space may serve to keep carceral punishment at high level of consciousness, although diverting attention from exclusionary-practices elsewhere in society.
The other strand in Cohen's argument is that when there is dincarceration, the community replicates the territorial divisions that occur when there is a-clear policy of separation for the mentally ill, mentally disabled or criminal,; Thus, while asylums removed the mentally ill from the rest of the urban population, deinstitutionalization isolates them also, particularly within inner-city areas. We have the creation of new ghettos, described in the North American context by Wolch and Dear.32 Rather than being the inevitable:: geographical expression of de-institutionalization, however, it could be argued1: that this pattern reflects the inadequacy of community care. A properly funded : programme of half-way houses, therapeutic treatment, employment provision^; and so on could counter the tendency towards isolation and enclosure.
1 think that we have to accept Cohen's argument that exclusion is not a necessary feature of social control. Exclusion is symbolically rich and it has provided an attractive theme for literature, both fictional and academic. The : oppositions of inside/outside, pure/defiled, and strong spatial divisions are: appealing and they do apply to some cases of socio-spatial control, but we have; to recognize that social control can assume diffuse forms and may not be expressed in such stark terms geographically.
CONCLUSION
In this chapter, I have suggested that both space and society are implicated in the construction of the boundaries of the self but that the self is also projected onto society and onto space. Self and other, and the spaces they create and are alienated from, are defined through projection and introjection. Thus, the built environment assumes symbolic importance, reinforcing a desire for order and conformity if the environment itself is ordered and purified; in this way, space is implicated in the construction of deviancy. Pure spaces expose difference and facilitate the policing of boundaries. The problem is not solely one of control from above whereby agents of an oppressive state set up socio-spatial control
PURIFICATION   AND   CONTROL   / 87
systems in order to remove those perceived to be deviant and to induce conformity. A reading of Klein, Kristeva and Scnnett suggests that exclusionary tendencies develop in the individual and that the exclusionary practices of the institutions of the capitalist state are supported by individual preferences for purity and order. Feelings of abjection are reflected in consumer advertising, for example. A rejection of difference is embedded in the social system.
One difficulty with this argument is that, despite the apparently universal nature of these processes, some people and some localities arc more tolerant than others. In typologies of personality, the 'authoritarian' or 'foreclosed' is recognized as an exceptional category, and similarly, within the city, as Wolch and Dear indicate, there are contours of tolerance. Although there is no simple contrast between heterogeneous, accepting inner cities and homogeneous, rejecting suburbs, it could be the case that the experience of difference and mixing in social and spatial terms contributes to variations in the response to difference. Individuals are socialized into a variety of environments, both in the home and in the neighbourhood, and the forces of purification are not going to be equally effective in moulding all individuals, groups and localities. Furthermore, if we accept that people are active agents who think reflexively, there is always the possibility of springing the trap. Even the suburban couple in Sidcup whose home life is described in graphic detail by Cohen and Taylor may mock the bourgeois pretensions of their neighbours and create an enclave in which they are able to live a non-conforming life:
The uniformity and predictability of it all might seem to induce an unshakeable sense of routine, a soul-destroying impression of the unmallcability of paramount reality, But when the door is shut against the night, and the two children arc safely in bed, husband and wife turn to each other and laugh. They are subscribers lo the new self-consciousness, apostles of awareness. Cynically, they deride those who share bourgeois arrangements with them, but who do not see the joke. Looking around the room they declare their awareness of their apparent suburbanity, and then with a delicious sense of their own distinctive identities, record their distance from such artifacts."
Admittedly, the chances of this happening are not great, given the residential selection process and the fact that social and environmental homogeneity are mutually reinforcing, but the temptation to construct yet more social and spatial stereotypes should be resisted.
GEOGRAPHIES   OF   EXCLUSION   / 88
NOTES
!, Michel Foucault, 'Of other spaces', Diacritics, 56 (!), 1986, 22-27.
2. Edward Soja, Past-modern Geographies: The reinsertion of space in critical theory, Verso London,1989,p.119.
3. Allan Pred, 'The social becomes the spatial, the spatial becomes the social: enclosure, social change and the becoming of places in Skane', in Derek Gregory and John Urry (eds), Social. : Relation! and Spatial Structures, Macmillan, London, 1985, pp. 337—365.
4. Soja, op. cit., chapter 6.
5. Anthony Giddens, The Constitution of Society, Polity Press, Cambridge, 5 984, chapter 1. :.
6. Nicky Gregson, 'Structuration theory: some thoughts on the possibilities for empirical.! research', Environment and Planning D: Society and Space, 5, 1987, 73—91.
7. Giddens, op. cit., p. 368.
8. Steve Pile, 'Human agency and human geography re-visited: a critique of new models of the self, Transactions, Institute oj'British Geographers, NS, 18, 1993, 122-139.
9. A. Moos and M. Dear, 'Structuration theory in urban analysis, 1: theoretical exegesis', ii Environment and Planning A, 18, 1986, 231—252.
10. Michael IgnatieiT, A Just Measure of Pain: The penitentiary in the Industrial Revolution, / 750- / S 50, Macmillan, Basingstoke, ! 978.
11. Anthony Giddens, Profiles and Critiques in Social Theory, University of California Press,': Berkeley, 1982, pp. 197-198.
12. Foucault, 1980, op. cit.
13. Michel de Certeau, 'Practices of space', in M. Blonsky (ed,), On Sign*, Basil Blackweil;! Oxford, 1985, pp. 122-145.
14. Fred Hirsch, Social Limits to Growth, Routledge, London, 1978, p, 82,
15. Julia Kristeva, in Powers of Horror, pp. 65-67, describes Mary Douglas's work on pollution/exclusion as fundamental. Mary Douglas, in Natural Symbols, devotes a chapter to Basil Bernstein. Bernstein, in Class, Codes and Control, vol, 1, acknowledges Douglas's influence. Both Douglas and Bernstein were clearly influenced by Emile Durkheim's thinking on the sacred and;', profane.
16. P. Atkinson, Language, Structure and Reproduction: An introduction to the sociology of Basil. ■ Bernstein, Methucn, Andover, 1985, p. 27.
17. Basil Bernstein, 'Open schools, open society?', New Society, 14 September 1967, 351-353.
18. Basil Bernstein, 'On the classification and framing of educational knowledge1, British Sociological Association Annual Conference on Sociology of Education (reprinted in Class, Codes -and Control,™], t, Paladin, St Albans, 1971, pp. 202-230).
!9. I return to Bernstein and the idea of 'dangerous knowledge' in my account of excluded geographies in Part II of the book.
20. Andy Alaszewski, Institutional Care and the Mentally Handicapped: The mental handicap,..: hospital, Croom Helm, London, 1986. Alaszewski trained as a geographer and then as a social anthropologist. His application of Mary Douglas's ideas in an analysis of the organization of ■ interior spaces has been largely unacknowledged in geography,
PURIFICATION   AND   CONTROL   / 39
21. Nanette Davis and Bo Anderson, Social Control: The production of deviance in the modern state, Irvington, N.Y., 1983,
22. The value of Foucault's arguments in Discipline and Punish for the development of a more nuanced socio-spatial theory has been recognized by Chris Philo, initially in The Same and Other: On geographies, madness and outsiders, Loughborough University, Department of Geography, Occasional Paper I 1, 1987; by John Lowman, in 'The geography of social control: clarifying some themes', in David Evans and David Herbert (eds), The Geography of Crime, Routledge, London, 1989, pp, 228-259; and by Soja, op. cit., 1989.
23. Michael Walzer, 'The politics of Michel Foucault', in David Hoy (ed.), Foucault: A critical reader, Basil Blackweil, Oxford, 1986, 51-86.
24. Lowman, op. cit., p. 237.
25. Walzer, op. cit., p. 60.
26. Thomas Markus, Buildings and Power: Freedom and control in the origin of modern building types, Routledge, London, 1993, p. 97.
27. I discussed the dual role of planned settlements for indigenous minorities and Gypsies in Outsiders in Urban Societies, Basil Blackweil, Oxford, 1981, particularly chapters 9 and 11.
28. Walzer, op. cit., p. 61.
29. Stanley Cohen, I'iilom of Social Control, Polity Press, Cambridge, 1985.
30. ibid., pp. 230-231.
31. ibid., p. 230.
32. Jennifer Wolch and Michael Dear, Landscapes of Despair, Polity Press, Cambridge, 1987.
33. Stanley Cohen and Laurie Taylor, Escape Attempts, Allan Lane, London, 1976, pp. 32-33.