The case for “ fluid ” hierarchies in therapeutic communities

Purpose – Democratic therapeutic communities (TCs), use a “ flattened hierarchy ” model whereby staff and clients are considered to have an equal voice, sharing administrative and some therapeutic responsibility. Using the sociological framework of interaction ritual chain theory, the purpose of this paper is to explain how TC client members negotiated and enforced community expectations through an analysis of power within everyday interactions outside of structured therapy. Design/methodology/approach – The study used narrative ethnography, consisting of participant observationwithtwodemocraticcommunities, narrative interviewswith21clientmembers, and semi-structured interviews with seven staff members. Findings – The findings indicate social interactions could empower clients to recognise their personal agency and to support one another. However, these dynamics could be destructive when members were excluded or marginalised. Some clients used their interactions at times to consolidate power amongst dominant members. Practical implications – It is argued that the flattened hierarchy approach theoretically guiding TC principles does not operate as a flattened model in practice. Rather, a fluid hierarchy, whereby clients shift and change social positions, seems more suited to explaining how the power structure worked within the communities, including amongst the client group. Recognising the hierarchy as “ fluid ” may open dialogues within TCs as to whether, and how, members experience exclusion. Originality/value – Explorations of power have not specifically focused on power dynamics between clients. Moreover, this is one of the first papers to look at power dynamics outside of structured therapy.


Introduction
Democratic therapeutic communities (TCs), like some psychiatric milieus such as Star Wards (2006), advocate clients' involvement in their treatment and in the running of the unit. These TCs often use a "flattened hierarchy" model whereby staff and clients are considered to have an equal voice and share administrative and some therapeutic responsibility (Rapoport, 1960;Jones, 1976;Kennard, 1993). Whilst critics of TCs highlight power dynamics between staff and client members can be open to social control (Kesey, 1962;Sharp, 1975;Baron, 1987), explorations of power have not specifically focused on power dynamics between clients. Using the sociological framework of interaction ritual chain (IRC) theory (Collins, 2004), this paper will explain how TC client members negotiated and enforced community expectations within everyday interactions. This paper is based upon a study that used narrative ethnography within two democratic TCs in the UK for clients with a diagnosis of personality disorder (PD) to explore the role of social interactions outside of structured therapy. The findings show that social dynamics could empower clients to recognise their personal agency, particularly when all clients felt a sense of belonging within the TC. However, these dynamics could also be destructive when members were excluded or marginalised. It is argued that the flattened hierarchy approach that theoretically guides TC principles does not actually function as a flattened model in practice. Rather, a fluid hierarchy, whereby clients shift and change social positions, seems more suited to explaining how the power structure operated within the communities, including amongst the client group.

Power in TCs
This paper adopts a Foucauldian view that power is an inherent aspect of social interactions (Bloor et al., 1988). From this perspective, power can be both destructive and creative for individuals (Haigh, 2013). Moreover many contemporary TCs, particularly in the UK, adopt a "flattened hierarchy" approach (Rapoport, 1960) that views all member voices as equally valued in therapeutic and administrative decisions. Promoting group-based approaches, Winship and Hardy (2007, p. 153) argue that TCs seem "best able to facilitate exploration of inherent staff tensions occurring in the division of clinical and custodial responsibilities". Kennard (2004, p. 306) contends that TCs have a "tolerance for the expression of conflict, a desire to enable people to take responsibility for their lives", coupled with a "natural sense of democracy" that encourages all members to have a voice in the community. Through democratic structures, TCs seek to minimise the manipulative forms of power and maximise individuals' creative power. In her reanalysis of Paddington Day Hospital (PDH) and Baron's (1987) study, Spandler (2006) highlights that the legacy of the Mental Patients Union, which arose from protests against closing PDH, can be linked to current UK service movements.
However, TCs, like any therapeutic milieu, can also create negative social environments. Prominent critics of TCs, including Kesey's (1962) novel One Flew over the Cuckoo's Nest and Baron's (1987) ethnographic account of PDH, have all acknowledged power as problematic. Additionally, Rose (1999, p. 50) is critical of TCs for moving away from "direction" to "interpretation", and implies residents within TCs are more exposed to social control. Similarly, Sharp's (1975) critique of TCs illustrates communities use their interpretive position to manipulate participants and exercise conformity and control in their lives. Sharp's (1975, p. 81) main argument is that social control and therapeutic interpretivist approaches are "systematically inter-related". Though Bloor (1986) challenged Sharp's critique of power, arguing that staff members orchestrate, rather than control, dynamics within TCs, he acknowledged that through labelling and assigning blame in the service of "interpreting" social behaviour, TCs exploit the vulnerability of those with social difficulties (Bloor, 1986). Furthermore, TCs have also been criticised by Basaglia (1987) and Donnelly (1992) for ignoring critiques of wider society, merely viewing client members' behaviour as problematic in relation to societal norms, as they risk becoming the oppressors they originally set out to challenge (Spandler, 2006). Laing's experiment at Kingsley Hall and Cooper's at Villa 21 were, in part, attempts to overcome this problem. More recently, Spandler (2006) highlights linking democracy with psychoanalysis can potentially individualise and pathologise client members' perspectives, particularly when clients' views are in opposition to staff.
Efforts to reduce the problematic forms of power in TCs have changed the way communities operate. TCs now function with more explicit boundaries (Haigh, 2013) and drivers, including the service user movement that validates the client's perspective and client empowerment (Stickley, 2006). TCs undergo more accountability and responsibility through identifiable standards and clinical constitutionalism to make the process of client member safety more transparent (Winship and Haigh, 2000). Nevertheless, the criticisms of TCs regarding the use of interpretation in therapeutic work, and the tensions between empowerment of clients and problematizing clients, highlight the importance of sensitivity to power and social control within the community.
For this study, the most significant limitation of previous research related to power within TCs is that it has not been fully explored between client members and has not specifically looked at what happens outside of structured therapy, such as meal times and community meetings. Studies by Sharp, Baron and Bloor focus on the imbalance of power between staff and clients but do not fully explore the differing social hierarchies and power that may exist between clients. Furthermore, Rapoport (1960) devotes a chapter on the social roles of staff but says very little regarding the system of social roles amongst the clients. Even in relation to TC practice, Spandler (2006, p. 130) acknowledges that "psychoanalytic discourse" in TCs focuses "more readily on sibling 'rivalry' rather than on solidarity". This paper sets out to explain the role of power within TCs, between staff and clients and amongst the client member group. As power is negotiated through everyday interactions, it is vital to understand the dynamics and mechanisms of social encounters within TCs. In particular, this paper looks at power dynamics that occur outside of structured therapy.

IRC theory
The theoretical framework of the IRC theory (Collins, 2004) is particularly useful in exploring the role of power within interactions outside of structured therapy. IRC theory has its origins in the work of Durkheim (1912Durkheim ( /2001, Goffman (1967) and more recently, Collins (2004). It is characterised by individuals coming together to share their attention and emotions, which in turn generate feelings of solidarity or belonging, long-term positive emotional feelings called emotional energy (EE), symbols of group membership and standards of morality (Collins, 2004). Successful rituals leave individuals motivated to repeat similar rituals. Failed rituals, in contrast, produce feelings of alienation, and at worst, despair.
Of relevance for this paper is the fourth outcome of IRCs: moral standards, or values, that define the social rules for the group. Through repeated interactions that generate feelings of belonging and long-term positive emotion and EE, the shared values of a group become morally charged (Collins, 2004). These values can be explicit, such as written TC boundaries, and implicit, such as an expectation to engage with others. Failure to conform to the values can result in exclusion from within the group, which can lead to a lack of EE, or a lack of solidarity (Summers-Effler, 2002). Moreover, Brown and Harris (1978) suggest that individuals who are part of close-knit groups may feel less depressed but can feel anxious about having to continuously adhere to group values. Power therefore is intrinsically linked to moral standards, as abiding by community values entails a level of conformity.
Power is an inherent aspect of social interactions, and, as ritual theory identifies, it can be inclusive or exclusive. Within IRC theory, Summers-Effler (2004) distinguishes between solidarity rituals, whereby individuals maximise EE collectively through solidarity, and power rituals, whereby more dominant members maximise their EE at the expense of subordinates. Vertesi (2012) highlights that inclusive power is voluntarily shared between all participants in the group, which facilitates the process of solidarity and shared ownership. Conversely, in power rituals, positive emotions build up but are centralised with dominant members gaining more positive feelings at the expense of others (Summers-Effler, 2002). TCs conceptualise inclusive forms of power as a "flattened hierarchy", as therapeutic and administrative power is distributed between staff and clients (Rapoport, 1960). Inclusive power is encouraged and is seen as empowering (Haigh, 2013), whilst divisive forms of power and social control are considered problematic (Bloor, 1986;Spandler, 2006). To counteract problematic forms of power, TCs seek to increase solidarity rituals and creative forms of power through reducing staff hierarchal power within everyday interactions. In other words, TCs seek to minimise power as control and maximise power as empowering through shared emotional feelings. For instance, Haigh (2013, p. 13) proposes that for modern TCs, "a major part of the non-clinical work is to specify those limits and ensure that the space within them is kept free from authoritarian or managerial contamination". Managing power dynamics is therefore a significant component of times spent outside of structured therapy.
The principle of flattened hierarchy, a model intended to create equality amongst members, appears largely an accepted, or taken for granted, component of TC life. Research exploring power dynamics in TCs (cf. Sharp, 1975;Bloor et al., 1988;Baron, 1987) has focused on power between staff and clients. The use of interaction ritual theory, with its focus on the exchange of emotions and feelings of inclusion and exclusion, is especially helpful in analysing power dynamics and social hierarchies within communities. Rather than simply describing issues, tensions and problematic examples of power in TCs, a micro-sociological analysis can explain the mechanisms involved in power dynamics within rituals.

Study design
The research was part of a larger study which used a narrative ethnography approach within two democratic TCs, consisting of 746 hours of participant observation, 21 narrative interviews with clients and seven semi-structured interviews within two democratic TCs for individuals with a diagnosis of PD. Due to space limitations, the data excerpts that appear below are illustrative exemplars, and fuller accounts of both the methods and the data appear in Clarke (2015Clarke ( , 2017. Ethical approval was granted by the NHS Research Ethics Committee East Midlands -Nottingham 1. The communities are anonymised as "Powell", a residential TC for women, and "Hawthorne", a day TC for both men and women. Treatment duration at both communities was for 8-12months. Fieldnotes were collected in the TCs and off-site, allowing for further reflection and preliminary analysis (Emerson et al., 2011), and loaded onto NVivo for further reading, coding and annotation. The majority of interviews were audio recorded and transcribed verbatim, forming the first stage of data analysis. One client member requested her interview be recorded through handwritten notes.
Data analysis was thematic (Pope et al., 2000), like much of qualitative research, and focused particularly upon client member interactions, exploring moments of inclusion and exclusion. The primary focus was on how power was negotiated in the day-to-day living outside of formal therapy. Notably, though the two communities were comparable in treatment duration, there were more hours for clients to interact at the residential community. However, this difference did not come through in the data as both TCs contained rich and complex social interactions (Clarke, 2015). Whilst the number of client member responsibilities in the communities initially made it appears dynamics in the TCs were flattened, further analysis of the data highlighted that power dynamics between client members was not so straightforward. In particular, interactions could promote empowerment through inclusion, or create feelings of depression and alienation through exclusion.

Empowerment through inclusion
During an interview with Tessa from Powell about the power dynamics in the TC, she remarked: When I first came for my first assessment, I didn't know who was staff and who was client.
Her comment illustrates one of the long-valued principles of a democratic TC: a flattened hierarchy that reduces divisions between staff and clients through sharing responsibility within the TC. Clients at Powell and Hawthorne exercised a great deal of agency through chairing community meetings, agreeing incoming new client members, challenging each other and staff members, determining meal time menus, and deciding upon aspects of the timetable.
During these times outside of therapy, clients worked continuously together, especially when a client member was struggling. The following excerpt illustrates how clients at Powell came together to support Andrea, a client who had just self-harmed, during a short Check-In ritual meeting prior to lunch: Then it is Andrea's turn. Instead of the usual "fine" she says, "I'm sorry". She explains she feels really bad because "I know I am having a negative influence". […] The clients tell Andrea they are worried for her because they care, and this is reason they are spending time to sort through it. Andrea leans over the arm of the sofa and bursts into tears. Erica passes her tissues. Andrea says she struggles to hear that because no one has ever cared about her. The room is silent, tense and heavy.
[…] Erica holds her hand and gently rubs her thumb against the back of Andrea's hand […] The group manages to talk Andrea through committing to staying safe and not using anything to self-harm (Powell, Day 38, 10 February 2013).
In addition to overt forms of support, such as with Andrea, clients' conversation and sharing stories could draw in other clients who were struggling during meal time rituals, as depicted below at Hawthorne: Jessie, who struggles with eating, does eat today but very quietly, slumped in her chair, her eyes occasionally searching around the room. After awhile, Mary shares about being a parking warden and Jessie joins in, also commenting about parking wardens. Mary then tells the group about her co-workers using handcuffs, police dogs, guns, and being at the shooting range. There are many jokes at this and everyone, including Jessie, is laughing. When we finish, Laruen remarks, "nice atmosphere at lunch" (Hawthorne, Day 21, 19 July 2013).
Exercising agency and experiencing inclusion could also occur through arguments, such as food shopping rituals, an activity that clients at both TCs organised and carried out. The excerpt below depicts a shopping experience with three Hawthorne clients: We get in to the shop. […] They pull out the list and all three immediately start arguing about where to begin. Whilst they had carefully planned the meals and all the ingredients they need before setting out, Heather insists they change around the days they do meals and also what they need. Her voice is very high pitched, loud and breathless. Daniel suddenly stands directly in front of her and tells her to calm down, take deep breaths whilst he demonstrates by inhaling and exhaling slowly and steadily. She nods swiftly and then sails off to look at something. Daniel rolls his eyes.
Then an argument starts with the vegetables and gets more pronounced in the deli-meat section. The problem is that Carl wants to buy the cheapest of everything. Daniel refuses saying that "they won't eat it!!" […] The last big argument is over sausages. Heather wants to get veggie sausages to have on hand […] But it's not on the list, as both Carl and Daniel tell her.
[…] This goes on for about five minutes, all three of them speaking in raised voices.
[…] When we are finished paying, we head out the doors and there is yet another argument about whether to hand carry the bags to the car or return the cart in the car park. Daniel who is pushing the cart opts for the latter (Hawthorne, Day 2, 21 May 2013).
Despite the arguing, the clients did continue to work together. Once we returned to the TC, the shopping trip was discussed during the afternoon meeting and members were able to express what they found stressful or difficult.
The majority of clients in their interviews reported that working together and being involved in their respective TCs had resulted in feelings of confidence and empowerment, such as Heather and Amy: Other clients spoke about how they felt able to do new things, had a greater sense of self, were able to validate themselves and felt more in control in their daily lives.

Power through exclusion
Not all interactions within the community resulted in the clients working to support one another and at times, support between client members could be inconsistently offered. For instance, Amanda from Powell noted: Some people would […] cut their wrists and we could go, there, there, there. And the next day someone else would cut their wrists and we'd go, you b***h! Get the f**k out of here! How community values, such as refraining from self-harm, were enforced was thus often dependent upon the context of the situation. That challenging power dynamics were sometimes present in both TCs amongst the client group is hardly surprising, as they exist in other mental health environments (Bloor, 1986). The question is how the community responded to these dynamics. The excerpt below from Powell depicts how during a community meeting the clients take issue with Julie's perceived lack of engagement with client members: Then the conversation goes on to talk about how Julie does not engage in conversation, in the group, at Meal Times, in the lounge, etc. Tessa, states in a soft, reassuring voice, "I would love more chances to talk with you". She goes on to say about how Julie "just sits in the lounge doing nothing. I really don't feel that is helpful".
Then there's discussion about how Julie doesn't make eye contact during meetings. Julie says, "I am trying to not disconnect, so I look down so I can concentrate on what they are saying".
Alison responds, "I feel like you are disinterested when you don't look at me". Martha adds, "There's normal behaviour and then-". Julie interrupts, "I'm trying!" She looks up with very wide eyes, and continues, "I do feel I am genuinely trying".
Speaking more gently, Erica replies, "We don't expect it to change overnight, it takes time, it's a process". Tessa says, "As long as you are trying" in a bit of a rougher, insistent voice with her legs crossed and arms folded across chest.
Julie moved frequently between insider to outsider within the group and support from clients was sometimes inconsistent. There was often tension between Julie and the other clients over her struggles adhering to the meal plan, her self-harming behaviours and engaging in groups, all of which increased others' anxieties. Often staff allowed these dynamics to surface and play out with little intervention. Margaret, a nurse from Powell, explained in her interview it was difficult for staff to intervene when a client was being marginalised as it could make the situation worse. It may seem staff favoured the excluded client member, thus prompting further exclusion.
At Hawthorne, support was also inconsistently offered at times, and clients were sometimes marginalised. The excerpt below is from a therapy break ritual where client member Mary criticises client member Robert for his extra-marital affairs: Whilst Brian, Evan and Jessie pop outside to smoke, Robert, Mary and Abby stay inside and chat. Talk moves on to men who cheat (Robert has admitted to being a serial adulterer) and Robert explains he never said he was not a b*****d for his behaviour. However, Mary will not let it go, asking what he felt when he stood up on his wedding day and made "those vows to her", meaning his wife. "It was a nice day out", responds Robert casually (tone of voice even, not emotional, face relaxed, hands behind his back. He is standing up; Mary is sitting down).
[…] Mary goes on to say, "If a man I were married to ever let me down […] I'd KILL HIM! Because he promised to look after me" […].
Evan, Brian and Jessie come back into the therapy room at this point. Abby joins in, saying that "all men are b******s". Jessie adds, "yeah!" quite loudly. None of the men say anything (Hawthorne, Day 22, 23 July 2013) At both communities, subgroups would be pairings or small clusters of clients that became friendlier with one another than with others. Patterns of power inequalities emerged and EE did not always appear to be shared amongst all members. During the interviews, these inequalities became more evident: It's quite often the strong members who, not bully, but coerce or, or they just set a, a means of behaving and rather than endure a conflict, or rather than go on a limb, the weaker members will follow those stronger members (Martha, Powell Interview).
And I felt like certain members were more important than others. Because you've got like the whole senior member hierarchy ( Jessie, Hawthorne Interview).
As the excerpts illustrate, clients referred to an invisible social hierarchy that appeared to leave some clients with less positive feeling than others, suggesting that equality between client members was not always achieved.

Discussion: a fluid hierarchy model
To understand about power and social control in TCs, it is necessary to recall the spirit in which they were formed. As previously mentioned, TCs were established in part as a move away from traditional psychiatric approaches that denied clients power and agency in their treatment. In contrast, TCs seek to empower and enable clients to take a greater role in their therapy. A flattened hierarchy model recognises that everyone in the community has individual power and agency and that all members have a responsibility for their behaviours and actions (Rapoport, 1960;Winship, 2004;Haigh, 2013). Moreover, Haigh (2013, p. 34) writes that modern TCs acknowledge that each person has their own agency, and that this personal power can at times be "destructive, envious and hateful". However, the concept of flattened hierarchy does carry two problems: one, it can underestimate or ignore the role that power actually plays in interactions, as outlined above, and two, it emphasises the direction of power between staff-to-clients and assumes that client members themselves are equal. Additionally, Spandler (2006) identifies the contradictions between TC models and TC practices are rarely addressed. Using interaction ritual theory highlights the social mechanisms involved in the power dynamics of everyday encounters.
In practice, the notion of client member equality was complex and inconsistent. It was also a sensitive issue, with clients at Powell frequently stating that as a flattened hierarchy, there were no cliques or subgroups in the client membership. Hawthorne client members were more vocal in community meeting rituals and social time rituals about client member hierarchies. Whilst this in turn led to some discussions about power, most clients stated during meetings that they felt everyone was fairly equal and that no one was excluded. Yet, the clients' perspective of power was contrasted with what was observed in practice and in their interviews.
Differing hierarchies amongst the clients appeared particularly between senior and newer members, and between genders (Hawthorne only). That senior members had more status within the TCs is understandable as they had that much longer to learn the community values, get to know staff and clients and gain experience working with address difficult issues. This difference was not necessarily problematic, though senior members did report feeling under pressure to be a "good example" and newer members reported feeling "intimidated" by more senior clients. Power differences between genders at Hawthorne sometimes occurred with the women making negative comments about the men, as seen above with Mary referring to Robert as a "b*****d".
During community meetings and during interviews, the men often reported feeling conscientious to not remind the women of male abusers, yet none of the women commented that they tried not to remind the men of abusive or unhelpful women. Male clients also reported feeling "emasculated" within the community. Studies on gender within TCs, particularly focusing on the client group, are limited and gender was also not the focus of this study. As such, whilst there appears to be a gender split that at times resulted in negative feelings, it is not fully known whether the men experienced this consistently or in what ways women, including women at the residential community, may also have experienced negative feelings related to issues of gender. Moreover, it is not known how clients may have benefited from gender differences.
However, not all power differences were due to length of time in the TC or gender. Some clients moved frequently from insider to outsider status, sometimes by isolating themselves, and at other times based upon whether they were adhering to the moral standards of the community. This in itself did not necessarily appear to be problematic or indicative of a rigid hierarchy, especially as clients often moved between groups. Moreover, in ritual theory, there can still be solidarity if emotion is shared (Collins, 2004). When Andrea used a self-harming behaviour, the clients worked together to extend support and a sense of belonging to her, drawing her back into the community and adhering to the TC value of client member safety. Shared emotions also do not necessarily have to be positive. In the example at Hawthorne with the shopping trip, the emotional feeling was one of frustration, anxiety and anger. Yet, all three shared these feelings together, no one was isolated in the encounter, and when they returned to the TC, they in turn shared these difficulties with the whole group, receiving support and feedback on what happened.
Problems occur when power builds with a particular group or some members feel excluded. Summers-Effler (2006) argues that for those who have a negative history of power interactions, such as some TC clients, power becomes less about consolidating positive emotional feeling than it is about minimising emotional loss. From this perspective, marginalising or excluding other members is not necessarily to create a power base, but to prevent potential increases in negative feelings that may result from other clients' destructive behaviours. This is one interpretation of the interaction above with Julie and the other Powell clients regarding her isolation and lack of eye contact. Through drawing her more into the community, the clients may have been attempting to reduce her self-harming behaviours, and thereby reducing their own anxieties. Nonetheless, problematic forms of power occurred if clients were repeatedly excluded and unable to consistently experience positive feelings of belonging.
Changes in the power dynamics between client members in everyday interactions could be minute and very fluid. One moment could be temporarily exclusive, such as violating a boundary before becoming inclusive as clients worked to re-establish solidarity. Given these changes, the traditional flattened hierarchy model does not appear to fully explain the complexity and flow of power within TCs. Instead, a fluid hierarchy framework acknowledges clients at any one point might hold differing levels of power and social status. Whilst Haigh (2013, p. 13) does recognise that the recognition of personal agency in modern TCs "goes much further than the original 'flattened hierarchy' ", he stops just short of departing from it. A fluid hierarchy model allows for hierarchal components and power imbalances within the client cohort. However, what is crucial is whether these power roles become flexible or rigid resulting in imbalances of positive emotions. This follows Haigh (2013, p. 13) who argues, "[a]uthority is fluid and questionablenot fixed but negotiated". Flexible social roles, whereby positive feelings are shared, suggest there is room for client members to change their social position and hierarchy status without power being consolidated in any one individual(s). It allows for the implicit social values and norms of community living to be continuously negotiated rather than merely enforced.
Recent studies of power have not been widely represented within the literature. Though this paper considers the power dynamics between clients outside of therapy, an analysis of power during everyday interactions could be expanded by considering staff to client interactions and exploring hierarchies within the staff team that may impact the overall organisation and fluidity of power within communities. As noted above, issues of gender remain limited within TC research. More studies that explore gender and power would significantly aid our understanding of power dynamics within communities. Furthermore, this study is limited by its focus of UK-based communities for those with a diagnosis of PD. Future research could explore power dynamics between clients within other types of TCs, and with TCs and other therapeutic milieus globally to understand how power permeates amongst client members within different cultural contexts. Lastly, though IRC theory has been used to understand everyday forms of power, it is important to acknowledge that social psychology, and in particular the group psychotherapy literature (cf. Nitsun, 1996) may also provide relevant insight into in-group and out-group dynamics.
In the main, most of the clients at Powell and Hawthorne could challenge one another and shift between power roles whilst maintaining positive emotions. At times of course, this did not always happen and rituals became power based (Summers-Effler, 2004). Whilst some occupied insider status more frequently than others, there was at least some movement. For TCs, there are some implications to consider. First, moving the discourse away from "flattened" to "fluid" hierarchies may open more discussions in the community about how power, and power shifts, are experienced in everyday life. This is especially important as everyday forms of power are not static but are constantly in motion. Open reflection about these changes may help identify where clients feel excluded. Second and related to first point, involving clients for both identifying and actively responding to problematic issues of power may facilitate more positive forms of power in everyday interactions. Lastly, TCs could consider how power could be used in interactions more creatively, particularly in light of Spandler's (2006), Basaglia's (1987) and Donnelly's (1992) critiques noted above.

Conclusion
This paper has explored how TC members negotiated and enforced the moral standards of the community through an analysis of power within interaction rituals. Crucially, explorations of power in TCs have not specifically focused on power dynamics between client members. Data from Powell and Hawthorne highlight that power dynamics within everyday interactions contained complex social process. These dynamics could be helpful, as in empowering patients by recognising their personal agency, yet these dynamics could also be destructive when members were excluded. The flattened hierarchy approach that theoretically guides TC principles does not actually function as a flattened model in practice. Rather, a fluid hierarchy, whereby clients shift and change social positions, seems a more accurate and nuanced explanation as to how the power structure operated within the communities, particularly amongst the client group.