Are Hope and Possibility Achievable in Prison?

There is both hope and frustration in this article. A recent research exercise in a prison found it to be inspirational in its ethos, relationships and mission. Prisoners talked passionately about their experiences in it and its impact on their personal development. But prisoners received very little resettlement support and things sometimes went wrong as soon as they were released, not because of any &#8216;moral failings&#8217; on their part, but because they could not even navigate the journey &#8216;home&#8217;. It looked like everything we know cumulatively about &#8216;better prisons&#8217;, but its prisoners were failed as they transitioned out. More &#8216;tragic imagination&#8217; is required in penal policy.

HMP Warren Hill is a small, highly impressive prison on the Suffolk coast with a distinctive ethos and a unique regime. Most of the 244 prisoners are serving imprisonment for public protection (IPP) or life sentences and are not eligible for open conditions, since the introduction of a rule excluding those with any history of absconding from future release on temporary license or transfer to open prisons. Warren Hill offers this population a tailor-made 'progressive' regime in closed conditions. The prison hosts a Therapeutic Community (TC) and a Psychologically Informed Planned Environment (PIPE) unit. The TC and PIPE have been awarded Enabling Environment (EE) status, and the prison as a whole is currently working towards this. The culture throughout Warren Hill strongly embodies the values of an 'EE': an environment that 'creates and sustains a positive and effective social environment', meeting ten standards of: belonging, boundaries, communication, development, involvement, safety, structure, empowerment, leadership, and openness (see Benefield et al. 2018;Haigh et al. 2012;Royal College of Psychiatrists 2013).
The prisoner population is long term and prisoners have complex therapeutic and resettlement needs. Warren Hill quietly and flexibly rescues these mainly over-tariff Category C adults and recalled Category Ds, 1 imaginatively balancing risk reduction with trust-building work. This balancing act is achieved through outstanding relationships between officers and prisoners, and a pragmatic, transparent approach to risk management. The prison prides itself on a 'forgiving' but educational ethos: staff have boundaries and are not overly permissive, but in conversation repeatedly emphasised the importance of second chances, and of establishing the reasons for setbacks. Most of all, Warren Hill is a place of hope, a word that we encountered repeatedly over the course of our research. The most serious challenge faced by prisoners is resettlement. Its population does not have the opportunity to become accustomed to, or make adequate preparations for, life outside. The lack of eligibility for release on temporary license (ROTL) denies long-term prisoners the chance to demonstrate risk reduction, acclimatise to a changing world, and accrue transferable skills. Staff and prisoners are aware that current policies limited what the prison can do to address this issue.

The Research
In January 2018, a team of nine researchers from Cambridge University's Prisons Research Centre (PRC) conducted a 'Measuring the Quality of Prison Life-Plus' ('MQPL+') exercise in HMP Warren Hill, a Category C prison for men in Suffolk. This is an in-depth examination of the cultural and moral climate of a prison, which relies on a variety of methods and on the cumulative expertise of the team. It is a methodology we have developed over many years, which we call 'ethnography-led measurement' (Liebling 2015b). Painstakingly-devised MQPL surveys were administered to prisoners in focus groups, with additional surveys given to individual prisoners to complete in private. Staff Quality of Life (SQL) surveys were administered during a full staff meeting following a brief presentation about the research, and additional surveys were distributed to those unable to attend the meeting throughout the week. In addition, the team undertook observations of most areas and functions of the prison, and conducted interviews and informal conversations with prisoners, staff, and the senior management team (SMT) (see further, Liebling assisted by Arnold 2004;Liebling, Crewe and Hulley 2011a). The research team spent the evenings together, digesting and deliberating on each day's activities, and developing our understanding of the prison's climate. This is a highly-focused but interactive and emergent form of 'team ethnography' (Erickson and Stull 1998), with well-designed surveys used to check our instincts and report back in detail to the prison, but the analysis taking root while in the field. We each took detailed fieldwork notes, wrote them up fully, and circulated them among the team. The work is commissioned, in that the PRC is currently funded by HM Prison and Probation Service to carry out three such exercises a year. We have a voice in the selection of establishments and had asked to include Warren Hill in the 2018 group.
We found the highest scores we have recorded in our MQPL+ exercises to date. 2 Prisoners rated Warren Hill 7.69 (on a 1-10 scale) on overall levels of treatment and humanity. All 21 dimensions scored above the neutral threshold of 3.00 (see Table 1). 3 Figure 1 shows that Warren Hill is close to, and on some dimensions exceeds, the scores from Grendon (a Category B therapeutic community, which generally has the highest MQPL scores found within the prison system).
Warren Hill is one of very few prisons that we would situate in the 'light-present' quadrant of our 'styles of penal order' matrix developed to conceptualise the use of authority in prisons (Crewe and Liebling 2017, and see Figure 2). This reflects the highly-engaged, 'dynamic' and broadly legitimate use of authority we observed there. Staff were fully present and engaged, without displaying unnecessary power. Staff were in control of the prison, and prisoners were aware of rules and boundaries, but the regime did not feel oppressive or stifling. Authority was deployed through highquality relationships founded on mutual respect and a sense among prisoners that the establishment was fully invested in their growth. The regime assisted prisoners in finding bespoke and meaningful routes for change, through the arts, groupwork and projects. The many and varied attempts to 'make a difference' described by staff and prisoners illustrated an energetic commitment to nurturing its population. This establishment had all the ingredients of a 'desistance-focused prison' (see Maruna and LeBel 2010). Prisoners argued that that 'this model of imprisonment should be replicated' (see later).
Staff were also highly positive. 4 They expressed deep loyalty to Warren Hill and described feeling highly fulfilled by their work. They believed strongly in the Governor's vision, whom they described as respected, progressive, and fully present. For all staff, all 18 dimensions scored positively. The highest rated items were: 'relationships with line management' (4.04), 'involvement and motivation' (4.02), 'relationships with peers' (4.13), and 'safety, control and security' (4.12). The lowest rated item was 'stress' (3.10). The overall quality of life scored a very positive 7.87 out of 10. There was some variation between uniformed and non-uniformed staff, which is typical. Three dimension score differences were statistically significant (three more approached statistical significance), all of which were rated more positively by non-uniformed staff. These include 'treatment by line management' (3.82 compared with 4.12), 'relationship with the organisation' (3.05 compared with 3.60), and 'commitment' (3.56 compared with 3.90). Table 2 summarises these overall very positive results.
In what follows we describe the prison in more detail, referring to the survey results where helpful, but also trying to capture the spirit of the prison and the practices we observed. Our aim is to articulate 'what is going on' in a prison that is so positively evaluated by its prisoners and its staff. The prison has established a virtuous circle of community spirit, clear boundaries, and deeply humane relationships founded on 'intelligent trust' (see Liebling 2015a;O'Neill 2002). The small size of the prison clearly contributes to its character, enabling a high level of interpersonal engagement. The site is set out in such a way that prisoners and staff encounter each other several times a day, without any sense that it is cramped or oppressive. The absence of internal fences and gates contributes to this sense of openness. The physical environment is well kept, and effort clearly goes into its upkeep, decoration, and the imaginative use of space. The staff culture is strikingly upbeat. What we saw looked like many of the best practices we have found or described in other prisons but acting together in one place.

Ethos and Management
A prison without a block [segregation unit]! That's massive; that sends a message.
Warren Hill has a deep community ethos, built on strong relationships and a sincere belief in personal transformation. One of the first signs seen when walking through the gate says, 'Hope', and on the wall behind, another sign reads, 'You made a difference'. The interior walls and wings around the prison display similar murals or quotations. These declarations are not slogans: they reinforce daily practice, and reflect the words, values and actions of the SMT. More than one member of the SMT came into the prison on their rest day during our research, in order to be part of the process. All were obviously highly committed to the establishment and its vision. Staff members were supportive and loyal to the Governor (86% of staff agreed or strongly agreed that, 'this prison has the right kind of Governor for current needs'), and expressed firm commitment to Warren Hill.  Staff often expressed the strengths of the prison in direct contrast to their experience in other establishments. They felt that working at Warren Hill was a privilege. Many said that, were they unable to work there, they would leave the prison service. Uniformed staff were highly positive about the SMT, which they felt was very visible: 97% agreed or strongly agreed with the statement, 'I often see senior managers around this prison'. Life was not perfect: 61% of uniformed staff agreed or strongly agreed that 'there are times where the Governors here fail to support staff in dealing with prisoners', but this score was higher than we typically find elsewhere. We saw or heard testimonies of governors taking the lead in initiatives with prisoners, but sometimes this was experienced as bypassing front-line staff. However, overall, relationships with senior, and especially line, management were very strong and based on a foundation of support and trust (nine out of ten staff members felt that their line manager was 'approachable when they needed to discuss an issue'). This was supportive cultural practice rather than 'peer resistance' (see Liebling 2011).
The prison's management team was deeply invested in the personal development of prisoners. There were opportunities for both prisoners and staff to suggest creative activities, and to pursue projects on behalf of individuals and the wider prison community. Eighty-eight per cent of staff agreed or strongly agreed with the statement, 'I am given opportunities to use my initiative in this prison' (mean score, 4.03). The overall orientation to prisoners was humane and person-centred: what we would call, after Martin Buber (1958), an I-Thou as opposed to an I-It approach. Prisoners were seen as 'experiencing subjects' rather than 'experienced objects' (see Liebling 2015b). Managers talked often with staff about 'developing empathy' and 'trying to understand where behaviours come from'. In discussions about specific incidents, members of the SMT knew a considerable amount about individual prisoners' backgrounds and circumstances, and took these into consideration when making decisions. Collectively, prisoners had a 'voice' in the prison, sitting on committees and attending meetings. A member of the psychology team explained: 'We're giving them a service and being held to account'. Many managers used a language of procedural justice; this was deliberate and was encouraged by the Governor. Warren Hill staff were keen to learn more about their practice. As one custodial manager put it: 'We want to know how to get better. We're not frightened of it'.
On the final day of our research we were invited to watch a play created by a talented group of prisoners. This play, called 'The Long Way Home', was written, produced and performed by prisoners. It expressed an empathetic appreciation for life stories and recognition of the complexity of human behaviour. It was a deeply moving portrayal of the lives of prisoners before prison: full of setbacks, trauma and pain, but also hope, humour, and personal transformation. We learned from the security team, on the morning that the play was due to be performed, that two of the prisoners involved in the production were thought to have brought drugs into the prison (accusations they denied). The way that this security incident was managed captured the spirit and vision of the prison. The play was allowed to go ahead, in spite of some real concerns (which were dealt with separately, afterwards). This was indicative of a management style that showed a serious commitment to: (i) the humane management of risk; (ii) the 'bigger picture' of seeking long-term change; and (iii) the avoidance of 'punitive excess'. The SMT understood that pathways to change were often nonlinear, without being sentimental, permissive, or naïve about individual prisoners. There was an expressed recognition that people are complex and multi-dimensional -that prisoners may not be 'good' or virtuous all of the time, but that they (and their futures) should not be defined by the 'lapses'. This position, which is unusual in prisons, resonated strongly with the rapidly-developing desistance literature (for example, Giordano, Cernkovich and Rudolph 2002;Laub and Sampson 2001;McNeill and Weaver 2010;Shapland and Bottoms 2017) and with the EE movement (see note 1). This faith in potential existed alongside clear statements that violence and drug taking or trafficking would not be tolerated.
The vision for Warren Hill was underpinned by realism, reflecting a concrete commitment to 'doing things differently', but safely, in order to both enable, nurture, and support prisoners in forging their own journeys of redemptive change, and manage risk. This kind of 'supportive limit-setting' (Wachtel and McCold 2001) is deeply-skilled work, and prisoners evaluate or appreciate it very highly. At their best, staff were collectively creating a new normative order, using all of the many legitimate power bases at their disposal (see Hepburn 1985;Liebling and Price 2001). All of our existing research tells us that this is precisely what is needed, both to secure decent outcomes, and for reasons of humanity and fairness (see Liebling assisted by Arnold 2004).

Humanity, Staff-Prisoner Relationships and Family Contact
We are encouraged to be people rather than controlled as numbers. (prisoner) I love my job. How many people get to come to work and change people's lives? (staff member) Prisoners consistently described their experience in Warren Hill as 'rehumanising'. The prison outscored Grendon therapeutic prison on measures of well-being (see Figure 1). Prisoners explained that they felt trusted, cared for, and recognised as persons of value. An unusually high 88% of staff said they felt proud of their job and that the work they did was meaningful. This feeling was evident throughout the prison, from the very positive treatment of prisoners on arrival, through everyday prisoner-staff interactions, to the way visits were managed. The reception experience in particular was widely praised by prisoners ('it's a welcoming reception, the best in any jail'). The score for 'entry into custody' was relatively high (3.58), considering the stressful nature of arriving into any prison. Eighty-two per cent of prisoners agreed or strongly agreed with the statement, 'when I first came into this prison I felt looked after'. The reception area was clean, bright and colourful. It felt casual -there were magazines, a TV, and plants -similar to a waiting area in any other civil building. One prisoner said: 'it wasn't a process, it was more a normal customer experience'. Officers greeted prisoners with a handshake and a cup of tea, and allowed them to dress by themselves. These small acts went a long way: 'I wasn't humiliated, I was treated with care and respect'; 'they treated me like a human'; 'staff make you feel welcome'. These were the kinds of cultural and atmospheric characteristics that private sector prisons were apparently introduced to bring about in the public sector, before 'austerity' changed the terms of trade (see Crewe and Liebling 2017;Liebling, Crewe and Hulley 2011b).
Staff-prisoner relationships at Warren Hill were very positive. Ninetyone per cent of prisoners agreed or strongly agreed with the statement, 'personally I get on well with the officers on my wing', and 81% of discipline staff agreed or strongly agreed that they had good relationships with prisoners. The dimension, 'Staff-prisoner relationships' was scored positively by both prisoners in the MQPL and staff in the SQL; 5 this 'attunement' between staff and prisoners is not always the case. The small size of the prison, and the reasonable staffing levels, increases the likelihood of regular personal interactions and enable staff to provide individualised forms of care.
While small size and decent staffing levels undoubtedly help to facilitate 'right' relationships, the main explanation for these high scores is a deep commitment to the community ethos of the prison. In general, staff were highly motivated, enthusiastic, and proud of their work. A prisoner with learning difficulties pointed to an Albert Einstein quote on the wall. He could not quite read it on his own, but explained that a staff member had taught it to him, and he had since memorised it: 'Everybody is a genius. But if you judge a fish by its ability to climb a tree, it will live its whole life believing that it is stupid'. This staff member had helped him learn to read and had pointed this quote out because it reminded her of him -and the fact that he did have value, strengths, and capacity. There were many such humane, meaningful, and sometimes transformational, interactions between staff and prisoners that left indelible marks on prisoners' memories. They were keen to share these stories with us. Even when prisoners were critical of individual staff, they were keen to explain that, relative to other prisons, Warren Hill's culture was superior to anything else they had experienced. We have come across this kind of 'buy in', or defence of a prison, only three or four times over our entire professional lives, always for the same reasons: staff are doing everything they can, with some expertise, to help prisoners out.
It was common to hear positive accounts from officers too: 'It's inspirational to work here'; and, 'I love my job'. Eighty-five per cent of discipline staff agreed or strongly agreed with the statement, 'it is important to take an interest in prisoners and their problems'. One hundred per cent agreed or strongly agreed with the item, 'supporting prisoners is part of my job'. Officers used impressive verbal skills with confidence: 'you can solve a lot of problems if you just ask questions, show interest'. Issues were explored and resolved, rather than exacerbated, and prisoners were keen to share their experiences of this kind of 'verbal judo' (see Liebling and Price 2001), with the research team: 'I know that if I have a misstep someone will talk to me, not just knock me back . . . That's given me the confidence to keep pushing myself forward'; 'it's all about conversations, not confrontations'. The calm tone and language used by staff was not dismissive. We observed little to no tension in the prison. Interactions were friendly yet professional. Officers addressed prisoners by their preferred name and were generally happy to be called by their own first names: a symbolic gesture which prisoners valued ('staff knew my name, even though I'd barely spoken to them'; 'being called J helped me to be J').
Staff knew their prisoners well, were able to recognise subtle changes in their demeanour, and had the emotional confidence to broach concerns empathetically. This signalled to prisoners that their feelings mattered, and conveyed a sense of 'true care' (Tait 2011): 'I trust them, I confide in them, they treat me like a human being. I've never been treated like a human being in my whole life'; 'staff have showed me that I'm not useless. I have worth'. Prisoners gave us frequent examples of officers providing support during difficult emotional periods, or helping them navigate 'prison bureaucracy' ('officers took me over to get my meds [medication] when I missed the time. You wouldn't get that in most places').
We witnessed many interactions that exemplified a form of relational 'decency'. For example, officers put aside meals for prisoners who were engaged in activities ('I've been in some prisons where you go hungry. Here, they save your meal for you'), or demonstrated a good working knowledge of prisoners and their families ('they go out of their way to get to know you'). Many prisoners said this was the first prison in which they had engaged with staff outside of 'just asking for the basics', and several spoke highly of their keyworker (a new kind of 'personal officer') as someone they felt comfortable confiding in: 'I trust my keyworker'; 'my keyworker knows what works for me'; 'In other jails you have nothing to do with staff . . . Here those relationships are different. You have your keyworker who you talk to and that's fine and normal'.
Staff were always there, but also kept a 'respectful distance' and gave prisoners ample personal space, which they appreciated ('they are there if you need them, but they're not always on your back'; 'they're approachable, but not overbearing'). There were clear boundaries or red lines that were not to be crossed (there was no tolerance for violence, for example), but otherwise prisoners were allowed to 'just get on with it'. This promoted a sense of independence and positive self-reliance, and was reflected in the survey results (in the high score on personal autonomy for example, at 3.66).
The relatively remote location of Warren Hill, and the fact that the prison population came from many different geographical areas, meant that visits were a major challenge for many families. Eighty-eight per cent of prisoners said that they were more than an hour's drive from their home area, and over a third (37%) disagreed or strongly disagreed with the item, 'I am able to receive visits often enough in this prison'. The prison's management team had taken steps to try to overcome this, encouraging prisoners to make use of accumulated visits time by temporarily transferring to prisons closer to home (but this was not a popular method). Prisoners who had reached the highest level on the progression regime (stage 3) were able to have unsupervised family visits within the 'virtual village'. The majority of prisoners (94%) were on this regime. Special family events (Christmas dinners, birds of prey displays) where prisoners were given the opportunity to relax with their families were highly valued: I gained contact with my son after seven years. The prison put on a Christmas dinner for me and my family and took photos of us. For me it's brought my son and me closer, and I will always be grateful to the staff and Governors for their help. (prisoner) One prisoner praised the five-hour visit he was allowed with his family: 'It was amazing, I can't remember the last time I sat down for a meal and broke bread with my father. It made a hell of a difference'.
Staff observed visits but were unobtrusive and respectful. Prisoners valued the relaxed, bright and comfortable visits hall ('feels more like a café than a visits room'), and appreciated the way their visitors were treated ('our families are treated like VIPs'; 'you can sit next to your loved one. It means a lot'). Similarly, the prison was praised for 'moving mountains' to allow prisoners to attend funerals.
While prisoners were well catered for in general, some groups were overlooked. For example, foreign national prisoners felt that they were uninformed about their status, and Muslim prisoners reported inadequate provision of halal food options, described lack of access to an Imam, and described problematic culinary practices (mixing utensils or using shared microwaves for cooking pork). As in many prisons, we received mixed reports about the quality of health care. On the one hand, prisoners described wing staff as deeply invested in residents' physical and mental health. One prisoner, for example, explained that staff noticed when he stepped out of the dinner line and asked if he was being bullied or feeling unwell. Having learned that he had suffered a mild panic attack, an officer brought him food and sat with him while he ate: this action signified a personal touch that was meaningful to him. However, prisoners persistently complained about formal health care services: they described delays in getting appointments and treatment, as well as problems with medication. These kinds of complaints are routinely related in other prisons we have studied.

Safety, Security and Order
Overall, the prison felt safe, secure and well ordered. Eighty-two per cent of prisoners agreed or strongly agreed with the statement, 'Warren Hill is a well-controlled prison' (3.99), and prisoners did not have to worry about their physical welfare (80% of prisoners disagreed or strongly disagreed with the item, 'generally I fear for my physical safety'). Residents praised staff for being 'security conscious' without being overbearing, and trusting without being negligent. They talked at length about the relief of leaving chaotic and frightening local prisons. We were impressed by the detail of the 'security and safer custody' meetings (unusually, a single forum) we attended. Each security case was meticulously outlined and evidenced. These meetings were also widely attended by both discipline and non-discipline staff from across the prison, giving a sense that security was a community effort. Traditional stressors that sometimes burden staff elsewhere were not of concern in Warren Hill: residents did not need to be escorted around the establishment; nor did the population need to be separated due to gangs or offences. Security was undoubtedly made easier given the reasonable ratio of staff to prisoners in Warren Hill.
The prison was 'tranquil' without being permissive. The atmosphere was light and relaxed, and contributed to prisoners' feelings of being trusted and not overly monitored. Some of the security procedures were inconsistent, but not in major ways, and these 'omissions' (or rule-bendings) were strategic, and counterbalanced by the cultural and relational strengths of the regime. For example, we noted problems with the roll count, which often seemed to take a long time given the relatively small prisoner population. We were, in general, less supervised and observed than is typically the case in other prisons. This was, of course, welcome in the circumstances. The balancing act was fine, and it is notable that in other outstanding prisons, the same recasting of security as 'for desistance' (for example, at HMP Berwyn) or 'for progress' (for example, at HMP Frankland), rather than 'for itself ' (see Liebling and Williams 2017) can be found. There are risks that, were an incident to occur due to any underenforcement of security practices, the prison's overall ethos might be undermined. This is the sad history of many previous attempts at humanity and reform in criminal justice (see, for example, Liebling assisted by Arnold 2004; Sparks 1997).

Hope and Progression
If you can't make it here, you won't make it anywhere. (prisoner) You can catch your breath here. That's everything when the next step is home. (prisoner) Prisoners spoke a language of 'hope' and 'progress', terms which came up in almost every conversation: 77% of prisoners agreed or strongly agreed that 'the regime in this prison is constructive'. Prisoners commonly said that they were able to 'see the light', or suddenly felt 'a glimmer of hope' in Warren Hill. They consistently explained that this was a place to come 'unstuck', and as a result, they were positive, hopeful, and engaged. Much of this optimism was related to parole, which was a very 'present' concern throughout the establishment and a notable feature of the prisoner experience. Warren Hill excelled in preparing prisoners for, and supporting them through, their parole. Prisoners knew, and frequently referenced, the 88% parole hearing success rate. They saw other men moving forward (leaving the prison or being downgraded) with regularity. For many, this was the first time in years that they had witnessed their peers progressing ('I had forgotten that there could be a light at the end of the tunnel; it's been dark a long time').
The prison's Offender Management Unit (OMU) was one of the best and most effectively functioning that we have witnessed, with a hardworking and dedicated staffing group (84% of prisoners felt that they were 'being encouraged to work towards tangible goals and targets'; mean score, 4.01). Many prisoners described their offender supervisors in highlypositive terms. Conversations with the OMU team offered further evidence of the benefits of a small community ethos: staff related specific details about their cases with ease, and the OMU team was clearly well integrated into the prison. Although case file reviews were explicitly about risk, staff worked very hard to identify and explain to prisoners what they needed to do in order to reduce it. While there was some confusion about the unique 'stage' system at Warren Hill, and some sentiments that movement through the stages could feel somewhat arbitrary, prisoners generally perceived it positively. The prison was clearly attempting to combat 'negative entry culture': trying to incentivise positive behaviour rather than sanction negative behaviour. The allocation of 'enhanced' Incentives and Earned Privileges (IEP) status to all prisoners upon entry signalled a kind of administrative 'fresh start'. The keyworker system was a particularly effective way of individualising relationships and communicating with prisoners: officers were empowered to intervene and challenge men in their caseload through talk. Each keyworker had roughly 45 minutes' contact with each client per week, but this was flexible and was adapted on a needs basis. This relational approach to offender management meant that it was typically unnecessary to resort to formal measures like negative entries.
The prison looked to us like a successful EE. Prisoners highlighted its positive features: emphasising the opportunities to transform themselves (83% of prisoners agreed or strongly agreed that spending time in Warren Hill offers a 'chance to change'), and stand on their own feet: 89% felt that, 'the regime in this prison allows opportunities for me to think for myself ', with mean scores of 4.01 and 4.03 respectively. Many prisoners found the regime liberating and flexible: 'There's a lot of freedom to build a routine in the way that you want it to be'. The limited institutional constraint gave prisoners the space to find their own way into transformative projects or experiences. Staff clearly expected prisoners to be proactive and responsible for this process: 'You've got every opportunity to progress, but it relies on you to identify your needs and develop self-sufficiency' (prisoner). This could go too far: there was some resistance to administrative 'responsibilisation'. There was a strong expectation of prisoner engagement, and it was not easy to tune out or 'lay low'. At times, this individualised conceptualisation or operationalisation of 'personal responsibility' was unwelcome: some staff put too much onus onto prisoners without providing adequate support. Not all staff in the prison were fully engaged: 58% of all staff members agreed or strongly agreed that, 'some staff get away with coasting in this prison'. Some prisoners felt that 'privileges' (such as use of the virtual village and spending in the shop) were veiled tests of competency, and indirect ways of monitoring financial behaviour. There was a sentiment that under these conditions, 'you're relaxed but you're still policed' (prisoner). 'Enabling' sometimes slipped into permissiveness. On Maple and Elm units (the PIPE and the TC respectively), prisoners felt that there was 'a line where staff should step in'. While prisoners appreciated a sense of freedom, they felt that staff should be more active in dealing with incidents or poorer behaviour (for example, loud stereos, bad language, or threats) on these units. This was a balancing act. On the whole, prisoners benefitted from a largely 'hands-off' or trust-based approach: this promoted a sense of independence, positive self-reliance, and autonomy. Prisoners talked about humanity being 'not just the way you are treated' but having the opportunity to be meaningful agents: to suggest things, make decisions, give back, and flourish in their own way. This combination of humanity, help, order, and agency comes up frequently in our ongoing analysis of prison regimes that facilitate both personal development and well-being (see, for example, Auty and Liebling in press).
Warren Hill offers hope and opportunities for personal development to prisoners without relying solely on a therapeutic discourse in order to do so. The prison operates according to an authentic, and theoretically recognisable, strengths-based model (Ward and Maruna 2007), without staff using this term explicitly. The language we heard among senior managers and staff was all about helping prisoners get 'unstuck', both in terms of their progression and their personal/psychological problems.

Purposeful, Bespoke, Creative Activities
This [working with birds] will make me give up crime. It is just more interesting. (prisoner) It's a good prison . . . Most of the activities are run and organised by prisoners, which is good, but more formal programming would be good too. (prisoner) Prisoners found new 'ground projects' at Warren Hill: that is, meaningful, bespoke, creative activities that helped them in the 'formation of personal moral experience' (Mattingley 2013;Williams 1981). Ground projects are 'the kinds of commitments that people find so integral to who they are that they might not care to go on with their lives without them, or would not know themselves if they no longer had them. They include deeply cherished and self-defining ideals, activities, and personal associations' (Mattingley 2014, p.307). They spoke with passion about these opportunities. In general, Warren Hill excelled in offering a suite of unique programmes and activities to prisoners. Staff were committed to identifying prisoners' strengths and talents, and were skilled at tapping into personal interests. Many prisoners described a process whereby they had suggested ideas for projects, and received enthusiastic institutional support to help implement them. Some managers were personally involved and invested in these different initiatives. This created an environment that promoted imagination and individuality, and encouraged residents to forge their own paths. This was reflected in the survey results: 83% of prisoners agreed or strongly agreed that their time in Warren Hill 'seems like a chance to change', with a mean score of 4.01. 6 From cooking meals for those with no cooking skills, bee keeping, taking care of an aquarium, or working with birds of prey, to becoming a barista or a gardener -prisoners were able to find their niche, and along with it, purpose and some gratification (see Toch 1992). Prisoners were enthusiastic and passionate about these roles, and mindfully described the benefits. Those who worked with animals or in the gardens spoke of the impact of having something to nurture and care for, and were moved by the levels of responsibility and trust placed in them (for example, a prisoner working in the gardens recalled: 'when I was on that tractor I felt an overwhelming sense of trust').
Welding, multi-skills, woodwork and motor mechanics workshops offered a meaningful and enjoyable work environment, particularly compared with the often menial, uninteresting work on offer in similar establishments. The workshops were calm yet productive, and the instructors were professional and engaged. Creative arts were recognised as serious routes for self-development, and there was a variety of such activities on offer ('Artistic expression is taken seriously here' (staff)). The prison newspaper ('The Know') was informative, interesting, and professionally produced. While it was formally edited by the writer-in-residence, it clearly had the imprint of its resident editorial team on it. There was a great deal of energy, participation, and goodwill on the part of staff and prisoners, and meaningful skills (collaboration, care, creativity, dedication, and patience) were being taught and developed through these activities. This kind of 'virtue development' was transformational and inspiring.

Towards Release
No prison is without its flaws. There was one area where the tragedy of the prison as an institution loomed large. The prison was not adequately resourced to resettle prisoners, despite this being an (unofficial) expectation. Many prisoners were well past their tariffs, had been recalled, or had complex legal cases in progress. They were entangled in national level policies and sentencing decisions which were beyond the establishment's control. Prisoners' ineligibility for ROTL, and lack of resources in the communities to which prisoners belonged, meant that prisoners were 'on their own', despite staff doing more than they were officially allowed to do (like transporting prisoners to a place of residence) once out of the door. 7 Two prisoners, found on recall in another prison shortly after release from Warren Hill, described a process of 'crash and burn' (what sounded like anxiety attacks) as they attempted to catch trains, deal with road traffic, or otherwise navigate their way in the world outside after more than 16 years inside. This was devastating, to them, their families, and to many others. Their accounts were particularly tragic given the progress they had made while inside.
Not every aspect of preparation for release lay in action outside the gates, but the 'bridge' between the prison and the community was no bridge. The dimension designed to reflect prisoners' feelings about the transparency and fairness of processes aimed at progression and risk reduction, 'bureaucratic legitimacy', was the lowest scoring dimension across the survey at 3.16 (this was substantially higher than in most prisons; see Figure 1). Forty-eight per cent of prisoners agreed or strongly agreed with the item, 'I feel stuck in the system', reflecting frustration with national level issues relating to risk management and release procedures. The relatively low mean score of 2.67 for the item 'I have to be careful about everything I do in this prison, or it can be used against me' (compared with the rest of the report) indicated that broader politics and sentencing strategies had tangible, 'on the ground' results that were felt acutely by prisoners however much effort was being made by staff to be supportive. There was frustration that some of the best work or education positions were not always attached to viable qualifications or meaningful accreditation certificates. Given the long-term nature of the population, and the realistic possibility of release, this caused significant anxiety. Prisoners noted a need for internet and technology orientation courses: navigating the world outside depended on developing new understandings and using relevant tools. In all of these areas of work -where the prison meets the community -the prison was constrained in what it could do to help overcome transition collapse. For these reasons, in our view, neither the prison's moral quality (or 'success'), nor the prisoner's 'moral journey', can be fairly judged by resettlement outcomes, although these are clearly relevant to the moral reckonings we should collectively make about penal systems, and constitute an area of criminal justice practice worthy of serious strategic rethinking and (re)investment.

Summary and Conclusion
I'll describe this prison for you in three words: replicate this model. (prisoner) 'You need sensitivity and resilience, enthusiasm and stoicism, honest, courage and hope, patience, empathy and interest'. ('Prison officers at their best', in Liebling and Price 1999) We were professionally relieved to conduct this research exercise in Warren Hill, after a series of more difficult and demoralising exercises carried out during the two years before (see Crewe and Liebling 2017). The prison's climate is an example of what can be achieved given the right form of leadership, creativity and attitudes among staff. The nature of the prisoner population -with a high proportion of experienced lifers and IPPs coming towards the ends of their sentences -certainly contributed to 'peer buy in' and to the prison's community-led culture. But other factors explained its high results: Warren Hill excelled in its relationships and mission. A 'strengths-based' and personal development model was being implemented sincerely (if implicitly): staff were committed to finding prisoners' qualities, tapping into specific interests, identifying talents, and not leaving prisoners to their own devices (generally). This was being achieved without forcing men to engage with institutional mandates, and allowed them some autonomy to engage with the prison in the ways they wanted to. In more penological language, the prison was far from 'new penological' (Feeley and Simon 1992) but supported individual prisoner trajectories. 8 The use of authority in Warren Hill was 'light-present' (see Figure 2). Staff were readily available, and used their authority confidently but not oppressively. They engaged with prisoners proactively by using 'soft' communicative skills. This approach was enhanced by the bespoke and meaningful activities on offer that provided creative routes for self-transformation. The prison excelled at developing 'softer' relational strengths and skill, rather than prioritising purposeful activity targets alone. The size, staffing levels, and experience base allowed staff to really know their prisoners, and gave them adequate direct personal contact: prisoners felt understood as 'whole' people, and were more trusting of staff and the institution, and of the sense of safety they experienced there, than we have seen elsewhere for a long time.
What we found at Warren Hill illustrated the general model of prisons and prison officers 'at their best' that we have described in ongoing research, a model that tends to lead to better outcomes. Officers, and other staff, operated within a finely-tuned range, balancing authority, compassion and respect for prisoners' moral agency. Staff had a clear 'mission', strongly encouraged and passionately articulated by the Governor: changing lives for the better, and generating hope. Having a critical mass of staff with this kind of vocational orientation and investment -and who themselves felt rewarded by their work -made a significant difference to the experience of prisoners. This mission underpinned often outstanding uses of discretion by prison officers. Their time horizons were long -their sight lines were set on the future. They were not, like other officers we have observed in more chaotic and underfunded prisons recently, 'just surviving till the end of the shift'. Officers were able to 'use the space that control on the wings gives to feel more comfortable in their decision-making', as we described in our first study of prison officers at their best in Whitemoor in 1999 (Liebling and Price 1999, p.65). Instead of 'this is how it's gonna be', we saw, 'this is what we are after, let's try and work towards it'. As prison officers described in the Whitemoor study (in a different context), 'you drop the seed and they start thinking about it' (p.67). Warren Hill was visibly hope-based. The staff were energised, mutually supportive, mature, engaged, and confident. The prison was exceptionally well-led. Everyone knew what the prison was about (progression, based on 'a clear belief that people can change') and understood their potential contribution to this goal. Prisoners were also able to use the space created by safety and legitimate order to 'work on themselves'.
While far too little is known about how outstanding prisons become, or remain, good, significant factors in this case include size, location, highquality leadership, 9 staff experience, and the impact of participating in the EE quality improvement programme, in which person-centred and relational values are prominent, or reinforced (Kordowicz 2018). That we have seen many of the key ingredients at Warren Hill (clear, sustained, and energetic leadership and vision, committed and experienced staff who enjoy their work, sufficient numbers, a well-operating key worker scheme, an outstanding OMU department, an overriding emphasis on 'the whole person' and on hope) elsewhere (for example, most recently at Berwyn, Frankland, Wakefield, Hull, and Durham, and clearly present at Grendon), and in the past, and in some other jurisdictions, suggests that this model could be replicated. What is unique at Warren Hill is the combined impact of all of these 'good ingredients' working together effectively so that, taken overall, the whole prison 'challenges the cultural imagination' (Mattingley 2014, p.157). What worries us is the lack of a co-ordinated policy and political effort to make this possible elsewhere, the tragic 'dropping off' of this level of support on release, and the long-term, and yet widely known, nature of these problems (see Liebling 1989). What can possibly explain this phenomenon?
David Garland (1990) argued, in Punishment and Modern Society, that punishment is a complex, distinct, and tragic social institution, 'which, in its routine practices, somehow contrives to condense a whole web of social relations and cultural meanings' (p.287). It is marked by 'conflict, tension and compromise', embodying a 'depth of stored up meaning' that narrow administrative analyses ignore. We cannot understand punishment, or the prison, if we consider their instrumental purposes alone: Punishment is, on the face of things, an apparatus for dealing with criminals -a circumscribed, discrete, legal-administrative entity. But it is also . . . an expression of state power, a statement of collective morality, a vehicle for emotional expression, an economically conditioned social policy, an embodiment of current sensibilities, and a set of symbols which display a cultural ethos and help create a social identity. At once an element of social organisation, an aspect of social relations, and an ingredient of individual psychology, penality runs like a connecting thread through all the layers of social structure, connecting the general with the particular, the centre with its boundaries. What appears on the surface to be a means of dealing with offenders so that the rest of us can lead our lives untroubled by them, is in fact a social institution, which helps define the nature of our society, the kinds of relationships which compose it, and the kinds of lives that it is possible and desirable to lead there. (p.287) The explanation for the outstanding practices we observed in Warren Hill (and elsewhere) is tied to the efforts, values, and imaginations of outstanding individuals, historical contingency, some good luck, and to finely-tuned experience. The explanation for a failure to link these practices to the future lives of released offenders lies in other areas, including 'economically conditioned social policy', 'current sensibilities', internal conflicts, and poor organisation. As Garland (1990) powerfully suggests, most of the answers to crime and deviance lie outside the 'puzzle' that is the prison. As he argues: However well it is organised, and however humanely administered, punishment is inescapably marked by moral contradiction and unwanted irony -as when it seeks to uphold freedom by means of its deprivation, or condemns private violence using a violence which is publicly authorised. (p.292) We should not be surprised, he suggests, when punishment 'fails'. It is 'a civil war in miniature . . . a society engaged in a struggle with itself' (p.292). Paradoxically, grasping this key argument may be a way to both conceptualise and organise punishment practices in better ways. Many of the failings of penal practice have their origins in the difficulties of life before and after prison, as the prisoner starkly reflects, above ('I've never been treated like a human being in my whole life'). Reform projects fail when the struggles and dramas that underlie them are denied. Garland did not put it like this, but others have: acknowledging the tragedy of 'a world like this' (Williams 2016) makes it more likely that we notice the conflicts present in policy and practice, that we notice what matters (Eagleton 2015) and that we treat 'the offender' as 'one of us', who 'could have been me' (see Muir 1977). A tragic understanding of the human condition, of the potential for civility and rebellion in all of us, of the complex causal patterns at work in lives that go wrong, leads to better judgment, and considered solutions, in a 'moral landscape' (Muir 1977, pp.178-97). Presently, offenders are not treated as 'requalified juridical subjects' after prison. We should ask why 'a liberal-democratic state might knowingly tolerate the existence of social conditions in which it is very hard for an offender to turn his or her life around' (Bottoms 2019).
The possibility of 'hope' requires both faith in the potential of the individual (the opposite of cynicism) and a link to reality, to the realistic possibility of positive future outcomes (Van Hooft 2011). Without this, it becomes 'mere wishful thinking'. Hope must 'have regard for how the world is [and . . . ] what is likely to occur in it' (p.40). Decent criminal justice may require not pessimism, nor utopianism, but a more 'tragic imagination' (Williams 2016 . Ninety per cent of respondents were over 30 years old. The majority were sentenced prisoners (86.7%); 12.2% were on license recall. The population was stable: 56.7% reported that they had been in HMP Warren Hill for more than two years. The majority (86.4%) had spent more than ten years in prison throughout the course of their lives. At the time of the research, 86.4 % of respondents had enhanced status, 7.8% were on standard, and one prisoner was on basic. The ethnic background of the sample was mostly white British (76.2%), and the remaining prisoners self-identified as one of ten other ethnic minority groups. A third of the sample (33.3%) identified as Christian, a third (33.3%) reported no religious affiliation, and 13.8% identified as Muslim. Just over half (53.3%) reported receiving visits from family and friends. Only 12.2% said they were less than one hour's drive from their home area. The majority of prisoners (80%) managed to stay in contact with family and friends via phone, mail or visits. 3 Variations between units were relatively minor. 4 The Staff Quality of Life (SQL) survey measures staff feelings about their working lives and treatment, and their attitudes and orientations towards prisoners (see Liebling, Price and Shefer 2010). Staff members are asked to rate 114 statements on a five-point scale. The items form 18 conceptual dimensions. The responses for each dimension are summarised and a mean score is produced, ranging from 1.00 to 5.00. A mean score greater than 3.00 is taken as an overall positive view on that dimension, with higher scores always reflecting a more positive answer. Eighty-two staff members participated in the SQL survey. Just over half were male (52.4%), and 47.6% were female. The majority of staff were aged between 31 and 60 years (82.9%) and identified as white British (97.6%). Half of the sample (50%) comprised nondiscipline staff, while 42.3% worked in discipline roles and 7.7% were managers. Most staff reported contact with prisoners 'most of the time' (62.2%). This was a long-serving staff group, as more than half (63.7%) reported working for the prison service for ten years or more, and 43.9% had worked in Warren Hill for ten years or more. 5 3.80 and 3.84, respectively. 6 Ninety-seven per cent of prisoners reported spending less than six hours a day, between 9am and 6pm, locked in a cell, and 79% spent their days either at work, in the education department, or both. Only 3% were unemployed, and an additional 3% were retired. 7 Release on temporary license (ROTL) is managed under Prison Service Instruction 6300. The conditions were restricted in the light of 'the significant public concerns about absconds from open prison, and the impact of those absconds on public safety'. A new 'absconder policy' was introduced in May 2014 with immediate effect (by internal memorandum) so that any prisoner in closed conditions who had a history of absconding on their current sentence would in the future be ineligible for a transfer to open conditions 'save in exceptional circumstances'. These changes followed a few high-profile cases in which prisoners on temporary license from open conditions carried out serious offences. This increased weight on absconder history, or 'very strong presumption against' ROTL for those with a record, allowed the Secretary of State to transfer indeterminate prisoners with a history of absconding from open to closed conditions without notice, as well as making it impossible for prisoners who had a history of failing to return from temporary release, even if this was many years ago, or while young offenders, to receive it. The progressive regime at Warren Hill was designed for those serving indeterminate sentences with a history of absconding who could not 'satisfy the test for exceptional circumstances'. It would 'provide an alternative means to open conditions whereby an indeterminate sentence prisoner (ISP) might provide evidence of his suitability for release'. The regime is designed to encourage prisoners to take more personal responsibility to produce evidence which they need to secure release from custody on completion of tariff with which they will be supported by relevant interventions and by appropriately trained staff. There will be a secure perimeter fence, in accordance with Category C conditions. There will be no entitlement to ROTL other than in exceptional circumstances' (see McAtee v. The Secretary of State for Justice ([2016] EWHC 1019 (Admin))). The tensions here are between the advantages of phased release versus risk. 8 The key elements of 'the new penology', which coincided with mass incarceration, are statistical prediction, and the aggregate management of groups rather than intervention in the lives of individuals. 9 Weber (1919) described good leadership (in politics) as characterised by passion: the serving of a cause (which promotes inner strength), responsibility, proportion, 'knowledge of tragedy' ('trained relentlessness in viewing the realities of life, and the ability to face such realities and to measure up to them inwardly'), and understanding of the meaning of human conduct (pp.20-5). This is an apt collection of qualities which resonates with our data in this and other related studies.