Heavy–light, absent–present: rethinking the ‘weight’ of imprisonment Crewe, B., Liebling, A. and Hulley, S. (2014), The British Journal of Sociology, 65: 387–410. doi: 10.1111/1468-4446.12084 HEAVY—LIGHT, ABSENT—PRESENT: RE-THINKING THE ‘WEIGHT’ OF IMPRISONMENT1 Word count: 10,479 including abstract, notes and references Last amended: May 30th 2014 Abstract Since King and McDermott (1995), following Downes (1988), defined the psychological oppressiveness of incarceration in terms of ‘weight’, little has been written about the ‘weight of imprisonment’. None the less, it is generally assumed that prisons that are ‘light’ are preferable to those that are ‘heavy’ – in part because of an assumption among many penologists that power, and its application, is dangerous and antagonistic. This article does not dispute that ‘heavy’ prisons are undesirable. Its argument is that there can also be dangers if prisons are excessively light. Many of these dangers are linked to the under-use of power. The tone and quality of prison life depends on the combined effects of institutional weight with the ‘absence’ or ‘presence’ of staff power. Drawing on prisoners’ descriptions of their experiences in public and private sector prisons, and their assessments of important aspects of their quality of life, the article outlines what these concepts mean in practice. The authors develop a four-quadrant framework for conceptualising penal legitimacy and the experience of penal authority. Keywords: Imprisonment; ‘ weight’; power; legitimacy; prison staff 1 In Contrasts in Tolerance (Downes 1988), a comparative analysis of English and Dutch penal policy and practice, David Downes argued that the extent to which imprisonment was ‘damaging and repressive’ depended on a range of factors: ‘relations with staff; relations with prisoners; rights and privileges; material standards and conditions; and a sense of the overall quality of life which the prison regime made possible or withheld’ (Downes 1988: 166). Summarizing the impact of these factors through the concept of the ‘depth’ of imprisonment, Downes argued that, in English prisons, imprisonment was experienced as ‘an ordeal, an assault on the self to be survived, time out of life’ (1988: 179), whereas in Dutch prisons, the rupture of confinement was ‘not so marked, the passage of time less prolonged, the sense of social distance from society less acute, and the problems of psychological survival less chronic’ (1988: 179). Drawing on Downes’ analysis, King and McDermott (1995) suggested a modification to his terminology: when prisoners used the term ‘depth’, they were generally referring to ‘the extent to which [they were] embedded into the security and control systems of imprisonment’ (1995: 90). When talking of ‘the deep end' of the system, they meant being in high-security establishments, many years from release, almost subterranean (‘buried deep’) relative to the surface of liberty. What Downes had described was ‘weight’ – the sense of the conditions of confinement ‘bearing down’ upon prisoners, and the almost palpable burden of psychological invasiveness and oppression.2 The metaphor was apt in part because it evoked the sense that the prison experience felt like a burden on one’s shoulders, or a millstone around one’s neck (McDermott and King 1995: 90). Perhaps because of an association of power with coercion, and an enduring view among critical penal scholars that power, and its application, is always dangerous and objectionable, little reflection has occurred in relation to the concept of ‘weight’. This article does not dispute that ‘heavy’ prisons are undesirable. Its argument is that there are dangers too if prisons are excessively light, and that the tone and quality of prison life depends on the combination of institutional weight with a related phenomenon: the ‘absence’ or ‘presence’ of staff power. Drawing on prisoners’ descriptions of their experience, and their assessments of important aspects of their quality of life, this article outlines 2 what these concepts mean in practice, and develops a four-quadrant framework for thinking about the experience of penal authority. The weight of imprisonment While the concept of the ‘weight of imprisonment’ has received little explicit attention in the research literature, many studies have explored the nature and consequences of different kinds of penal regimes. The early, classic sociological studies of imprisonment emphasised the core similarities of prisons (Sykes 1958; Goffman 1961), giving little consideration to the possibility that management might be a significant variable in shaping the prisoner experience. However, it was soon apparent that patterns in inmate organization and attitudes varied according to organizational goals and conditions (Grusky 1959; Berk 1966). Street (1965), for example, noted that the culture and social structure among prisoners was determined by the form and degree of deprivation that they faced and the manner in which staff exerted authority and control. A clear implication of such ‘deprivation’ perspectives was that the inmate experience would be shaped by the particular pains and deficiencies that prison managers chose to alleviate or ignore. Sykes himself was pessimistic that the pains of imprisonment could simply be ‘managed’ out, noting that there was a trade-off between different priorities, such as freedom and safety. As described by DiIulio (1987), in Governing Prisons, sociological conceptions of penal order had also promoted an assumption that prison administrators had little choice but to indulge prisoner power structures if they were to prevent breakdowns in institutional stability. In many prisons, DiIulio argued, the outcome of this form of resigned accommodation was a prisoner world characterized by idleness, chaos and pervasive violence, with staff discouraged from using their authority to regulate prisoner conduct. DiIulio used his comparison of levels of order, amenity and service in the penal systems of Texas, Michigan and California to propose an alternative vision of penal control. Key differences in prisoner outcomes were ‘rooted in differences of correctional philosophy’ (1987: 6), with the Texas model of strong, security-conscious, paramilitary bureaucracy more capable of delivering ‘a calm, peaceful and productive daily life’ for prisoners than more consensual models of prison governance. 3 The first tenet of DiIulio’s argument – that management practices are the key variable in determining the quality of prison life – is more persuasive than the second – that these practices produce the best outcomes when they comprise a ‘control style of prison administration’ (1987: 179). Research into prison riots supports DiIulio’s implicit contention that ruptures of penal order occur in prisons where there are breakdowns in administrative organization or staff-prisoner interactional assumptions (see Useem and Kimball 1989; Useem and Goldstone 2002; Boin and Rattray 2004). Historical accounts of prisons and prison systems (e.g. Jacobs 1977; Crouch and Marquart 1989) have documented how shifts in penal administration – and thus in the exercise of staff power – lead to changes in forms of prisoner safety, leadership and social structure. But while DiIulio rightly highlights the disadvantages of lax and under-supervized regimes, it is far from clear that rigid rule enforcement, tight supervision and impersonal procedural efficiency (see Crouch and Marquart 1989) represent the only solutions to penal disorder. In the UK, recent prison research has been animated by Tom Tyler’s (1988) argument that the most stable forms of order and compliance are generated by conditions that are legitimate. Seen in this way, the kind of ‘strong governance’ that DiIulio recommends is deficient because it is unconcerned with important components of procedural fairness, in particular, the degree to which authority is exercised in a manner that is respectful of those subjected to it and acknowledges their dignity and rights. Order, of a certain kind, can no doubt be established through more coercive and impersonal arrangements - as in ‘super-max’ establishments – but it is likely to come at the expense of personal wellbeing (Liebling et al. 2005), co-operation, and longer-term compliance. For current purposes, the most notable text is Prisons and the Problem of Order (Sparks, Bottoms and Hay 1996), which draws on ideas of legitimacy to compare the different means by which two high-security prisons accomplished order (see also Sparks and Bottoms 1995 and 2007). Albany ran a restricted regime, with a strong emphasis on control and supervision, based on a ‘situational’ model of control. Long Lartin allowed prisoners greater autonomy, adopting a ‘softer mode of policing’, promoting closer relationships between prisoners and staff, and relying on a more ‘social’ 4 form of control. On the face of it, Albany appeared more punitive and antagonistic – in our terms, ‘heavier’ – with more incidents of attrition between prisoners and staff, and conditions that seemed inappropriately constrained for long-term prisoners. Yet, in Long Lartin, although there was less friction between prisoners and staff, incidents and antagonisms between prisoners were more complex and serious. Long Lartin’s more relaxed (or ‘lighter’) regime enabled the development of a more elaborate informal economy, a sharper hierarchy between prisoners (particularly between ‘mainstream’ prisoners and sex offenders, who felt somewhat ‘thrown to the wolves’ [Bottoms, personal communication]), and a greater degree of backstage violence.3 Some prisoners, particularly older men who sought a quiet and predictable existence, expressed a clear preference for the more controlled regime that Albany offered. In highlighting some of the risks of the Long Lartin regime (see for example, Sparks and Bottoms 1995: 57—9), Sparks and colleagues anticipate our critique of penal lightness. They are careful to note that neither prison ‘had achieved a fully satisfactory synthesis’ of penal priorities. Yet there is an unresolved tension between their exposure of the problems that resulted from Long Lartin’s culture of relaxed supervision and their judgment that it was nonetheless a somewhat more legitimate prison than Albany (see Sparks et al 1996: 328). Part of the reason for the incomplete analysis is that Sparks et al. did not then make explicit something they have written about since: that is, the ‘self-legitimacy’ of Long Lartin’s powerholders, and the degree to which they were dialogically successful in their claims to legitimacy. Governors and prison staff at Long Lartin were mutually committed to a well-articulated strategy of liberal governance, and their confidence in this philosophy was affirmed back to them by the majority of Long Lartin prisoners. However, as Hinsch (2010) notes, and Sparks et al. clearly recognize, alongside this empirical concept of legitimacy, based primarily upon the subjective approval of prisoners, we should also consider a more normatively grounded concept of legitimacy, based on objective criteria of justice (see Bottoms and Tankebe 2012). We might also need to distinguish between the demands of prisoners and those of other relevant ‘stakeholders’ in the prison, such as the general public. As Bottoms and Tankebe (2012: 104) note, ‘Legal officials sometimes have to consider their legitimacy in relation to more than one 5 audience and […] these audiences might have significantly different priorities’. We shall return to this general point in due course. The contribution we wish to make here is to add to this dialogue more concretely. Among practitioners in England and Wales, any presumption that ‘light’ forms of imprisonment were inherently more desirable than heavier regimes lost credence in the years following Sparks et al’s fieldwork, not least due to two prisoner-on-prisoner murders in Long Lartin, alongside some high-profile escapes from other high-security prisons (see Liebling 2002; Liebling and Arnold 2004). These events indicated that power was being under-used by staff in ways that were hard to defend or legitimate either to prisoners or to the wider public, and that the Radzinowicz philosophy for dispersal prisons of a ‘liberal regime within a secure perimeter’, was potentially highly hazardous (Liebling 2002).4 In Long Lartin and similar establishments, staff had yielded too much power to prisoners, under-enforcing legitimate rules, and allowing the development of ‘no-go areas’ on the prison landings (McDermott and King 1988). As the Prison Service sought to claw back control from the mid-1990s, the (heavier) Albany model of firm but fair control became more credible in the eyes of practitioners and politicians than the (lighter) Long Lartin model of negotiated order.5 There were important counter-currents to this development, which are relevant to our concerns. A series of prison disturbances in 1990, many of which occurred in Local prisons with traditional staff cultures and generous staffing levels, testified to the dangers of excessively oppressive environments. Tasked, in part, with modelling more progressive penal cultures, the privately-managed prisons that were opened in England and Wales from 1992 appeared to foster staff- prisoner relationships that were more decent than those in the public sector. In James, Bottomley, Liebling and Clare’s (1997) evaluation of the first private prison to open in the UK in modern times, prisoners rated most aspects of their interpersonal treatment more positively than did their peers in a comparable public sector establishment. The ability to create ‘lighter’, more respectful staff cultures became part of the competitive promise of the private sector. 6 Several subsequent studies (see Liebling, assisted by Arnold 2004) and official reports (NAO 2003; HMCIP 2007) appeared to support the view that private sector prisons were more respectful and their regimes less distressing than their public sector counterparts. At the same time, however, they consistently identified weaknesses in privately-managed establishments in the domains of safety, control and security. James et al. (1997) found significant problems relating to staff inexperience, low staffing levels, and staff supervision; Taylor and Cooper (2008) described an environment in HMP Kilmarnock (Scotland) that was chaotic and unsafe; and Liebling (2004) noted that the escapes that occurred from HMP Doncaster could be linked to excessive levels of trust between staff and prisoners. In these prisons, staff-prisoner relationships were friendly and informal but not necessarily ‘right’ (Liebling and Arnold 2004; Liebling 2011) – to put this another way, the trust expressed in these relationships secured the approval of one audience (prisoners), but could not be considered legitimate in terms of the norms of society at large. Moreover, prisoners themselves expressed reservations about some aspects of the ‘lightness’ of their experience: poor regime organization, lack of boundaries, and inadequate policing by staff. Such misgivings are significant for our understanding of penal legitimacy. Sparks et al’s (1996) attribution of greater legitimacy to Long Lartin than Albany rests in part on their conclusion (Sparks, Bottoms and Hay 1996: 314—6) that the former’s ‘social’ crime prevention strategy was ‘grounded in the belief that one had to relinquish some manifest control in order to keep control in the long term’ (1996: 322). Yet prisoners appreciate safety and security as well as interpersonal decency (Liebling, assisted by Arnold 2004), so that once control is relinquished beyond a certain point, a prison’s legitimacy is eroded in their eyes as well as those of external stakeholders. Our argument, therefore, is that order and control are not just outcomes of legitimacy but are in themselves aspects of legitimacy, providing that they take certain forms that are neither too heavy nor too light. In other words, as we illustrate below, while prisons that are situationally oppressive or highly coercive are for all kinds of reasons disagreeable, it is a mistake to equate ‘lightness’ with quality if the environment is laissez-faire or dangerously under-policed. ‘Light’ can mean un-burdensome and easier to bear, but it can also mean ‘insubstantial’ or deficient – characteristics that are undesirable in prisons. The ideal 7 might be an establishment whose staff exert authority without relying on the kinds of situational control measures that ‘chafe and vex and arouse frustration and annoyance’ (Sparks, Bottoms and Hay 1996: 323). Rather, if control is achieved interpersonally or relationally i.e. through staff-prisoner interactions and engagement, as much as through forms of restriction and surveillance, it can be ‘present’ yet relatively un-oppressive, with protective and supportive functions that prisoners appreciate. The failure among penologists to appraise the concept of weight, to think through its different forms, or to examine its converse, ‘lightness’, exposes an assumption that the use of power in prison is inherently bad, that more power is worse than less, and that a prison that is ‘light’ is therefore preferable to one that is ‘heavy’.6 McMahon (1992) argues that the reflex repudiation of power among critical scholars in the 1970s and 80s derived from Foucault’s argument that disciplinary currents pervaded modern society. All developments in penal practice, however benign they appeared, were to be scrutinized as sinister re-formations of power, ‘involving only more social control, repression, domination, and subjection’ (1992: 218). All forms of power, all power-holders and all institutions of power were to be opposed and viewed with suspicion. This nihilistic position, McMahon argues, makes it impossible to identify ‘what forms of the exercise of penal power, and by whom, might be preferable to others’ (1992: 215).7 While much can be gained from a stance of scepticism towards apparently progressive penal developments, a position of blanket hostility towards all forms of power is unwise on both theoretical and empirical grounds (and is inconsistent with Foucault’s (1982) own position). First, it is often forgotten that there are ‘good’ and ‘bad’ forms of power, and that while coercion is an aspect of power, it is not ‘the essence of the phenomenon itself’ (Wrong 1997: 261n). Power is ‘the capacity of some persons to produce intended and foreseen effects on others’ (Wrong 1997: 2), so this means that good power – power combined with care, for example (Sennett 1993), as compared with power used with indifference – can achieve positive outcomes. As Mulgan (2007) argues – and as Hobbes suggested – freedom depends upon order: we need power to flow to liberate us from disorder. Second, 8 as Downes (1988) notes, and as we seek to demonstrate, what is required is serious investigation of the views of those who are subject to different forms of penal power. The public—private sector study Most of the data on which this article draws derives from a comparative study of public and private sector prisons, carried out over a two-year period from 2008—2010.8 Seven prisons were involved in the research, including five out of the eleven private sector prisons in England and Wales and two public sector establishments. In each of the four main prisons in the study (consisting of two ‘matched’ pairs of public and private sector establishments), the research team spent between six and eight weeks undertaking interviews with staff and prisoners alongside periods of observation and the distribution of quality of life surveys to prisoners and staff. The three additional private prisons were researched for around one week each, using the survey tools and a small number of interviews with prisoners, uniformed staff and managers. In total, in addition to surveys conducted with 1145 prisoners and 500 staff, semi-structured interviews were conducted with 114 prisoners and 133 staff. Both groups were sampled purposively, in order to represent variables such as (for prisoners) wing, privilege level, prison experience, sentence length, ethnicity and age, and (for staff) wing, seniority, job function and sex. One-to-one, semi-structured interviews were undertaken in private rooms on prison wings, and were digitally recorded with the permission of interviewees. Most lasted between 40 and 90 minutes, and were subsequently transcribed and coded using NVivo software, with coding themes derived both deductively, from the research literature, and inductively, from emerging patterns in the data – a form of adaptive theory (see Layder 1988). Alongside fieldwork notes, the prisoner interviews, which discussed issues including staff- prisoner relationships, and perceptions of treatment and safety, provide the primary material for this article. A key aim of the study was to provide qualitative insight into cultural differences, staff practices, and the experience of confinement and authority in the two sectors. Interviewees who had served sentences within both sectors were asked to reflect directly on their experiences in each. At any one time, around 13 per cent of the prisoner population in England and Wales is held in private 9