Liminal space and the negotiation of care work in extra-care housing

This is a qualitative case study of care work in a liminal space, specifically the case of an extra-care housing residence, which is an innovative housing alternative for elderly people in need of care in Sweden. The study is an exploration of social care workers ’ perceptions about their workplaces and their understandings of them- selves, which are shaped by their embeddedness in architectural space. The extra-care housing residence appeared as a liminal space in which two dominant spaces – home care services and residential care – under- pinned the staff ’ s perceptions of an unclear workplace and their identity work.


Introduction
After more than a decade of research, extra-care housing still defies a common definition. International researchers maintain that there is no common terminology or standard, and architectural design and service delivery both vary substantially (T. J. Atkinson et al., 2014;Howe et al., 2012;Larsson et al., 2013;Mayagoitia et al., 2015;Means et al., 2008;Orrell et al., 2013;Tinker et al., 2007). Extra-care housing may appear in a variety of designs and may be purpose-built or located in reconstructed premises (Larsson et al., 2013;Wright et al., 2008).
Disregarding the many variables in extra-care housing models, efforts have been made to pinpoint similar features. One feature that characterises extra-care housing is that it enables an independent lifestyle for the older individual. As regards similar items in architectural design, self-contained flats are indicated as an important common attribute, as are common rooms (Garwood, 2010;Moriarty and Manthorpe, 2011;Tinker et al., 2007). A full flat for each resident is also indicated as a prerequisite for independent life in old age (Hanson et al., 2006;Tinker et al., 2007). This paper discusses the heterogenous nature of extra-care housing with support of the concept liminal space. The most original and basic definition of liminal space is in-between other spacesambiguous, undecided and in transition (Thomassen, 2016). Understood as a liminal space, the unclear design and organization of extra-care housing may cause ambiguity in the roles of those in the workforce. There may be expectations among staff regarding which tasks are appropriate (Blood, 2013;Garwood, 2010;Wright et al., 2010). There may also be confusion about the very nature of extra-care housing. This paper sets out to explore how staff members resonated about this ambiguous nature in an ethnographic case study of a Swedish extra-care housing residence in between ordinary housing and residential care. How staff dealt with ambiguities in order to make sense of a liminal space was a matter of interest.
The in-betweenness of extra-care housing 1 in Sweden is a quality that appeared at its very inception when it was suggested in a government report as a solution to an identified need for a housing alternative between ordinary housing and residential care to compensate for a reduction in the number of residential care places (SOU, 2008:113). 2 Extra-care housing deviates in certain ways from these two standard housing alternatives for older people. It is more easily accessible for residents than residential care since it is not embedded in a strict legal framework of municipality mandatory care assessments, which is the case with residential care. Simply reaching the age of seventy fulfils the requirements for moving into extra-care housing (SFS, 2001:453). As a rule, extra-care housing offers home care services as though in people's private homes, in contrast to the 24-h care and services in residential care provided by permanent staff. However, a recent evaluation showed a diversity of the design and activities in Swedish extra-care housing (Larsson et al., 2013). The extra-care housing residence in this study offered around-the-clock care provided by permanent staff. The residents' care needs varied substantially. Some of the individuals had almost no support needs and would not have qualified for a place in residential care. There were also fragile residents, as a similar level of care was offered as in residential care. It was thus a home for life, also "suitable for a profile of changing care needs" (Kneale and Smith, 2013, p. 278). With this high degree of innovation, it constituted a new workplace for care staff that appeared as a liminal space in their perceptions.

Liminality
The concept of liminality originated in anthropology through the works of Arnold van Gennep (2019Gennep ( [1909) and was later elaborated and developed by Victor Turner (1970). The concept liminality is a conceptualization of stages in the rituals that accompany transitions in human life, for instance initiation rites in which the liminal persona is in becoming between one state and another. There are three phases in these transformations: The first is the original environment from which the participant is detached when entering the second transitional phase, which is characterized by ambiguity, paradox and contradiction and where normativity is destabilized by the overturning of everyday norms and concepts (Turner, 1970). The third step is the return to society, but in the new role and status that the participant has acquired. Although both focused on rites of initiation, van Gennep and Turner ascribed a broader generality to the concept as an analytical tool for understanding any state in which there is transformation at hand. This study aims to build on this proposition by applying it to care workers' perceptions of a novel workplace. Of importance to this study is the suggestion that liminality holds the promise of potential changes of the well-known and mundane. This creative side of liminality has been an attractive approach to liminal situations in research (Thomassen, 2016). According to Turner, it is a "realm of pure possibility whence novel configuration of ideas and relations may arise" (Turner, 1970, p. 97). However, the study also takes advantage of the fact that Thomassen (2016) encourages us to include the darker and disturbing sides of liminality; the unsettling of norms and authority, leading to situations in which indifference and uncertainties or even destructiveness may reign. This can lead to highly stressful situations (Söderlund and Borg, 2018). These two approaches to liminality are found in studies of ageing, sometimes in connection to the third and fourth age (Laslett, 1991). While the third age of self-fulfilment and active ageing has been explored in playful engagement of older people together (Yarnal, 2006), the fourth age of physical and mental decline may involve a liminal experience of being in-between living and dying, a situation of prolonged, persistent liminality (Nicholson et al., 2012). The latter study indicates that while liminality is by definition transitional and limited in time, it may also appear, somewhat contradictorily, as permanent in nature when one of the three phases is frozen and has captured people involved (Szakolczai, 2003;Ybema et al., 2011). Thomassen (2016, p. 91) writes that "liminality is very essentially a spatial concept". Van Gennep (2019[1909) occupied himself with liminal spaces as territorial border zones and border lines, thresholds and doors to which an idea of transgression is connoted. Turner, who did not dwell on space in his theories, nevertheless ascertains that liminality is emplaced (Turner, 1970). Various interpretations and conceptualizations of liminal space appear in research (Roberts, 2018). The concept often refers to physical spaces "in which traditional routines, norms and activities are suspended or renegotiated" (Söderlund and Borg, 2018, p. 891). Liminal space is connected to the construction, crossing and transgression of borders and boundaries (Braude, 2012), or to situations in which borders become unclear and permeable (Wels, Van Der Waal, Spiegel, & Kamsteeg, 2011). It may harbour liminality, for instance as a prerequisite for liminal transitions to occur (S. Atkinson and Robson, 2012). Space becomes liminal by destabilizing transitional forces, or may be liminal in itself, e.g. as marginal, temporary and exterior spaces (Söderlund and Borg, 2018). These various interpretations of liminal space resonate van Gennep's (2019 [1909]) or Turner's (1970) original theories. Nevertheless, Thomassen (2016, p. 8) draws attention to the interstitial aspects of liminality by contending that "that which is interstitial is neither marginal nor on the outside; liminality refers, quite literally, to something placed in an in-between position". This proposition encourages a view of liminal space as subject to negotiation of elements from the two circumscribing spaces with which it may merge or that it may resist (Downey et al., 2016). These standpoints will constitute a theoretical framing of this study. Extra-care housing is an in-between space that evolves in carers' reflections, assembled by components that are relevant to the understanding of their care tasks, of the older residents and of themselves, negotiated, discussed and comprehended as a liminal experience. This in-between-ness is thus a third space where negotiations take place, sometimes of incommensurabilities that refuse to merge or even coexist (Bhabha, 1994). This also echoes the Foucauldian heterotopia where incompatible spaces can be juxtaposed and appear as an alternate ordering of a counter-space (Foucault and Miskowiec, 1986;Hetherington, 1997). As such, it will challenge established categories with the potential to contribute to "burgeoning social practices of new and radically different social formations" (Downey et al., 2016, p. 8). This contributes to the emergence of "an interstitial future, in-between the claims of the past and the needs of the present" (Bhabha, 1994, p. 219 italics in original). The ability to challenge established categories is very important in this study in which the extra-care housing is a novel contender in between two dominant spacesordinary housing and residential carei.e., already established forms of housing and care (Dale and Burrell, 2007). Extra-care housing is seeking its independent form. The analysis will show how staff members break up this binary and put together an interstitial future of re-claimed fragments and elements. In this work, tangible architectural spaces are a co-working feature that is a building block in processes, giving shape and meaning to inhabitants and their activities. Everyday life in the extra-care housing is formed by architectural elements such as rooms, thresholds, doors etc, linked in networks of other co-working elements. Architecture/care is a process, associating the temporality and transition in liminality to architectural space that has the capacity to mould entirely different care organizations by generating diverse care work practices in which quality of care is a product (Essén, 2008;Nord, 2018;Stanley et al., 2016;Willems, 2010). This paper builds on this conceptualization of assembled spatial and material elements in an exploration of care workers' perceptions about their workplace. Architectural space is a means for staff to give meaning to care work.

Liminality, space and later life
The concept of liminality or liminal space/place has been used in studies of a variety of housing alternatives for later life: ordinary homes, residential care homes, assisted living facilities or long-term care facilities. Much of this research does not give support to a view of ageing in peace and tranquillity, but to a period in life during which considerable changes may challenge the older person in various ways. Liminality is ascribed to phenomena such as the individual's feeling of increasing frailty or the feeling of approaching death (Black, 2006;Nicholson et al., 2012;van Wijngaarden et al., 2016). These liminal transitional experiences give shape and understanding to both well-known environments and new living environmentsnot only for the perspective of the older individual; they can be expected to be important to staff perceptions as well. The ordinary home may become liminal because of a perceived threat of being forced to move due to lost capacities (Leibing et al., 2016), or during recovery after hospitalization (Rush et al., 2016). Older peoples' liminal experience in assisted living may have a negative impact on the residents' ability to settle down and feel at home (Black, 2006;Buse et al., 2018;Cutchin, 2017). Residents' negotiations between the independent healthy and active third age and the more care-demanding and less independent fourth age may be correlated to different spaces. While public space such as common rooms in an extra-care housing facility is associated to the third age, the private flat is less clearly defined and could be both for independent life as well as for the fourth age (Cutchin, 2017;Laslett, 1991;West et al., 2017). These suggestions associate to the suggested bond between independent living and a full flat in extra-care housing. While these studies have focused on older peoples' experiences of liminality and of liminal space, fewer studies have examined carers' experiences of how care work and older people's liminality appear in space. One exception is a study of emergent care ethics in the context of liminality of palliative care practices in which ethical rules that apply in other care situations are irrelevant (Braude, 2012). There are a few studies about semi-professional private companions in long-term care facilities; these are paid workers without formal employment who occupy liminal spaces as they are invisible in policy and legislation and their tasks and roles are unclear in practice (Daly and Armstrong, 2016;Daly et al., 2015). Liminality and liminal space seem to be two concepts in limited use in studies of professional carers' experience. This is the gap this study aims to address.

Methodology
The study was carried out with an ethnographic approach in one case with a strong focus on the understanding of the meaning of human actions. The fieldwork was conducted over nine months with several weekly visits to the residency during which the author closely socialized with both carers and residents (P. Atkinson and Hammersley, 1994). Data collection methods included interviews and observations. Individual, 1-h interviews were conducted with all staff; i.e., 15 individuals including 12 carers, two managers, and one nurse. Care work, practices, values and residents' everyday life were among the main issues discussed. All interviews were recorded and transcribed verbatim. Observations were gathered in two ways: direct observations of everyday life in the residence, and shadowing (Czarniawska, 2007). The author accompanied care workers while they carried out individual care work in the residents' flats and other spaces. About 120 h of shadowing and observations were accumulated over the full nine-month period, throughout daytime shifts (7 a.m.-9 p.m.). Field notes were used to document observations, and the physical environment was documented with photography and the collection of drawings.
Actor-Network Theory (ANT) is a major theoretical framework in which the relational and processual perspective of architecture and care sketched above can be accommodated (Cutchin, 2005;Nord, 2015;Yaneva and Guy, 2008). This links well to the becoming of liminality and liminal spaces (Roberts, 2018). Extra-care housing is considered as a network of heterogenous material, spatial and other non-human elements that are interlinked with humans in a network where care emerges (Latour, 2005). This is a dynamic network in which an exchange of elements reverberates through the whole network and afford various changes to its outcomes, in this case care. In accordance with this approach, the analysis identified liminal space as this type of network of various interacting elements. Extra-care housing in the study is assembled in staff's reflections by different elements from the two dominant spaces. This included different combinations of chosen elements which had diverse consequences for the notions of care. Thus, care appeared in various versions too. Turner suggests that liminality offers an opportunity for reflection. Interviews with carers made use of his argument that "[l]iminality may be partly described as a stage of reflection. In it those ideas, sentiments, and facts that had been hitherto for the [liminal personae] bound up in configurations and accepted unthinkingly are, as it were, resolved into their constituents. These constituents are isolated and made into objects of reflection" (Turner, 1970, p. 105). The carers in the study were identified as the liminal personae and the analysis aimed to identify how they isolated these constituentselements of the dominant spacesand how they were perceived, deconstructed and recomposed in interviews "in ways that ma [d]e sense" (Turner, 1970, p. 106). New knowledge appeared in these reflections, aiming ultimately to establish extra-care housing as a category of its own (Enosh and Ben-Ari, 2016). Of particular importance are the 'sacra, the heart of the liminal matter' and the 'sacerrima, "the most sacred things"' (Turner, 1970, pp. 102, 107). The sacra are the most important issue, subject to reflections, while the sacerrima are axiomatic principles and presumptions that do not easily lend themselves to negotiation or alteration. The sacra and sacerrima were thus sought in the analysis, as was how they contributed to staff perceptions about themselves and their care work.
The analysis was carried out with the support of situational analysis building on ANT (Clarke, 2003). This is a two-step analytic process: first, various elements of relevance to the research issue were identified from a qualitative analysis of interviews and field notes. These elements were of very different types, e.g., people, spaces, care tasks, artefacts, etc. In the second step, relationships between these elements were identified as they were assembled in staff reflections and translated into new meanings and understandings of the extra-care housing and their care work (Latour, 2005). The last step included an exploration to uncover the sacra and the sacerrima.

Results
The residence in the study was a reconstructed student home in which the former student flats were now inhabited by older residents. It consisted of two three-floor buildings at 90-degree angles to one another (Fig. 1); each was a multi-household housing unit with long empty corridors with closed doors (Fig. 2). Each resident had a small onebedroom flat with a kitchen and bathroom. There was also a common facility with a resident dining room and living room, which is one of the few requirements of an extra-care housing in Sweden. As mentioned above, the care needs of the 24 residents were varied. Aged between 70 and 99 years, some residents needed help with almost everything, while others required very limited assistance. All of them had a rental contract for their flats and a guarantee that they could remain there and receive all necessary assistance until the end of their livesa home for life, as is suggested by Kneale and Smith (2013). There were a total of 15 staff members taking care of these residents; 12 of them were care workers.
Although it had not been purpose-built, the residence in the study nevertheless adhered to most of the features included in a preliminary definition of an extra-care housing unit discussed above. It provided for an independent life in old age by the provision of individual full flats while giving the opportunity for residents to socialize in common rooms (Garwood, 2010;Hanson et al., 2006;Howe et al., 2012;Moriarty and Manthorpe, 2011;Tinker et al., 2007). Despite this, the extra-care housing residence already appeared as an ambiguous space during observations and in informal conversations with staff members in which the nature of the extra-care housing was questioned. For this reason, the theme was revisited and explored further in interviews, which revealed an almost unanimous sentiment among staff about the unclear nature of their workplace, albeit with no consensus regarding how and why. The perceived ambiguities about the nature of the extra-care housing residence appeared in various versions in the interviews. Many of the care workers interviewed had experience working in the municipality's residential care, home care services, or both, and these were important sources on which to draw when they attempted to define the extra-care housing residence. Many of these reflections placed the extra-care housing residence in-between these two dominant spaces (Dale and Burrell, 2007) and functioned as a depot for elements that were possible to explore and combine in reflections on the nature of the extra-care housing residence.

A diversity of combinations of elements
The in-betweenness of the extra-care housing residence had initially represented something new and interesting for Tora, prompting her to apply for a job when the residence had started up about ten years earlier. She said in an interview that.

I became, you know, curious; What do you do in an extra-care housing facility? What does that mean? And that was the mix between home care and residential care, something in-between.
The nature of care work is an important feature in this reflection, as is the mixed characteristic of the extra-care housing residence. Tora's comment foregrounds the creativity that is sometimes ascribed to a liminal state, found already in Turner's (1970) work. Care work is emplaced and the in-betweenness of extra-care housing as a workplace became tangible in staff reflections by combinations of various architectural spaces where care was carried out. The residents' flats surfaced as an important element in the extra-care housing network in contradictory arguments as to whether extra-care housing was home care services or not. A few staff members raised the idea that extra-care housing was equivalent to or combined with home care services. Berit was of that opinion, and her arguments included a spatial distribution of care tasks excluding those located outside the residents' flats from home care. She said:

I feel that I go to their [the residents'] homes … Everything inside the main door is home care services … It is actually the dining/living room, the nurse and the transport to and from the dining/living room that is not home care services. That falls into the extra-care housing concept.
Berit's observation implies a perception of a conceptual combination of care and space in extra-care housing. That fact that she juxtaposes extra-care housing and home care services but keeps them clearly separated associates to the heterotopic possibility of juxtaposition of outside spaces (Foucault and Miskowiec, 1986). In this quotation, the flat is equivalent to the ordinary home where home care services are offered. However, the resident flat was also a decisive component in the opposite arguments that care in extra-care housing was not equivalent to home care services. The new flat in extra-care housing could not be compared to a dwelling in which the resident had lived many years. This statement challenged the dominant space of ordinary housing and therefore made care in extra-care housing different to home care services. Extra-care housing thus became a third space presenting an altered ordering (Bhabha, 1994;Hetherington, 1997). Susanne explained: I think there is a difference between home care services and extra-care housing. In home care services, you go to people's homes. You do that here too, but home care is in their own flats where they might have lived ten years or more.

Transitional liminality
A frequent argument among interviewees was that things had changed over time in the extra-care housing residence, so it was now more similar to a residential care facility than it had been when it had started. This brings the unsettling side of liminality to the fore and liminal space emerged as in the process of a transition into something stressful and unwanted (Söderlund and Borg, 2018;Thomassen, 2016). The residents' flats appeared here in a complex argumentation in which residents' deteriorating health, changed care work and a possible move to residential care were included and negotiated in a third space of contention (Hetherington 1997). The general opinion was that residents' health had deteriorated since moving in. "They are in too poor condition to live in their own flats", Lennart simply noted. Thus, some of the staff found it questionable that they were living in extra-care housing at all. It was possible to trace disappointment about the ways the work had changed; residents were growing increasingly weaker, and activities with the residents had gradually changed to bodily and medical care instead. Tora expressed how the work that she had expected had changed into something else: Because from the beginning, healthy, retired people were supposed to come here, because there would be activities, a lot of things were going to  C. Nord be done in the beginning. And we did them, because they were healthy and in fairly good shape.
Observations did not entirely confirm these statements. As was suggested, there were few common activities and few participants. Only a handful of the 24 residents regularly came to the weekly reading session in the common living room, where a carer read aloud from novels or other books. Entertainment, such as live music, which was offered occasionally, was more popular and attracted most residents to the coffee tables. Meals were attended by most of the residents. However, an observation that contradicted the staff's proposal was that those who could organise their food themselves, i.e., the healthier residents, did not take advantage of the common lunch and coffee offers as frequently as weaker residents did.
Nor were the two dominant spaces entirely helpful as ordering devices of a third space, as they did not offer a convincing difference as regards care work (Hetherington, 1997). Some staff argued that the care work tasks in the extra-care housing residence were similar to those in both residential care and home care services. Lennart commented on the care work: Well, you have the same tasks … yes, it is cleaning, showers, medication, assisting with meals, breakfast, fixing up … yes, the tasks are basically the same.

Residents: healthy or not?
Others agreed with Tora and raised the idea that extra-care housing offered or should offer work tasks that were somewhere between residential care and home care services. This argumentation revealed the healthy residents as the sacerrima expressed as the original core of the extra-care housing. The axiomatic and non-negotiable character of "the healthy residents" seemed to be a significant stronghold in the staff's perceptions, in light of the residents' now-deteriorating health. The much weaker residents were thus put at the heart of the liminal matter, the sacra, which was the most important issue upon which to reflect since it had put the staff in a perceived situation of contradiction (Turner 1970). They had caused a change in the initial concept of extra-care housing and added a transitional liminality to their work. For this reason, many interviewees had reservations about the category of extra-care housing to which they ascribed healthier residents, seemingly not considering the fundamental purpose of the extra-care housing residence as a lifelong care service to elders. A departure from that would perhaps create a liminal and stressful situation to the residents with a potential threat to move (cf. Leibing et al., 2016). Staff seemed to propose that the residence would be easier to classify as extra-care housing without the weaker residents.
That notwithstanding, the staff who tried to define extra-care housing without a comparison to other care services did so by describing the residence from the perspective of the residents. Magnus attempted to circumvent the difficulty of categorizing the liminal extra-care housing when he argued that it is the resident that should be in focus. Even if it is a collective that lives in a house in 24 flats, it is nonetheless the resident that should be in focus. As long as he or she feels safe, then it is a form of extra-care housing (cf. "safety housing" in Swedish).
The resident flat appeared once again as an important architectural element, this time for independent living in old age, where it was possible to make individual choices. In an interview, Tora spoke about the advantages of the independent lifestyle the residence offered: They [the residents] stay in their own homes. They can decide whether they want to cook at home or what they want for breakfast and dinner. They can decide when they want to be accompanied outdoors … it's like living in an ordinary flat in town.
Here, another aspect of the sacerrimathe healthy residentswas noted: the residents who were capable of autonomy. The flat was indicated as a resource for an independent resident and not a space for a weak resident with deteriorating health. This argument did not solve the problem of the liminal extra-care housing however, but seemed rather to increase the inconsistencies in the argumentation. This liminal transition of the extra-care housing included a pathway in which the residents and staff moved from public common rooms associated with the third age to the more private resident flat unambiguously perceived as for the fourth age, overlooking that this is a space in which both the third or fourth age may be accommodated (West et al., 2017). This move is linked to residents' independent choice against what was expected. The many interpretations of the flat in interviews may indicate ambivalence and difficulties in seeing that the move was a consequence of choices made by either healthy or weaker residents. It seems reasonable to assume that in this extra-care housing residence, self-contained flats promoted independent residents, as has been suggested (eg. Garwood, 2010;Hanson et al., 2006;Moriarty and Manthorpe, 2011). As a matter of fact, extra-care housing demands an independent resident (Nord, 2018). Care workers explicitly acknowledged this as something positive in the study. However, the carers seemed to have difficulty reconciling that the choice to stay in the flat was perhaps not a consequence of deteriorating health, but the conscious free choice of a healthy resident. It is extremely important to be the agent in later life, also in cases in which an independent choice is impossible (Nord, 2016). Thus, the older individual's choice to not participate in activities organized in the residence is an unexpected reason for staff dissatisfaction, especially when staff members acknowledged the resident flat as a space for independent living in old age.

Staff: responsible or not?
It is not surprising that the residents' flats appeared as important elements of architectural space in staff reflections, considering that this was where they spent most of their working time, according to observations. Care work was largely independent. This solo work dispersed in residents' flats had consequences for the ways in which the staff perceived themselves and their work. Independence was also apparent for the staff.
The liminality of the extra-care housing contributed to a professional role taking form that also benefitted from components borrowed from the two dominant spaces. A staff subjectivity of responsibility emerged, although help was not far away. In interviews, several staff members said that they maintained a strong feeling of independence and responsibility through working alone in the residents' flats. Each resident flat offered a space for private meetings with residents. This is a working situation that can be compared to the constantly changing situation of home care service staff. Most of the staff stated in interviews that oneon-one time with residents in their flats was a highly rewarding part of their work. They appreciated this intersubjective space of care (Braude, 2012). Maggan situated this experience explicitly in time and space: Yes, I enjoy it, because then I'm with the care receiver. I've been cleaning in Henry's flat for an hour or so now. It's really pleasant to make jokes to him and chat. Yes, it is us, then and there.
Magnus also agreed: "that is when you come close to the residents". He continued with a comment on the responsibility that solo work invoked: "And you need to lift yourself too. You have to take on more responsibility". Most of the caregivers enjoyed that the solo work implied more responsibility. Berit commented on her natural affinity for working alone: I have no problem with that. I am very independent, I have to say that … I have always seen myself as someone who can manage things, kind of … and then you don't have to be worried … You know that you can take care of things.
The appreciation of being alone with the resident expressed, which appeared in contrast to the lament of the lost activities with healthier residents, is a somewhat surprising argument, however. Although these work tasks are not mutually exclusive, the deteriorating health of residents must have increased the possibilities for the much-appreciated one-on-one moments between staff and residents in the flats, compared to activities which are often in public spaces in a group, as was observed. The fact that the former grew at the expense of the latter is a logical outcome in a place where people grow older, and their increasing weakness is highly likely. In this case, the increasing weakness among the residents, which was associated with residential care, was not used as an argument for an understanding and appreciation of the extra-care housing, even though that had been possible.
In contrast, the staff appreciated another aspect of the liminality of the workplace that belonged to residential care: the availability of colleagues nearby. The presence of workmates contributed to the negotiation of a third space (Bhabha 1994). Not all of the staff members expressed the strong self-confidence that the above quotes on solo work suggest. Lennart commented on the lonely decisions a staff member had to take: "I think it feels chancy. You get worried. Damn it; what if I do something wrong?" He drew on his experience from working in residential care when he indicated the advantages of presence of workmates: In home care services and in this place, there is more solo work. You are never alone in residential care … It's always nice to have someone beside you in case there is something you don't know, if you need to ask for advice or something.
Birgitta agreed with this statement by distancing herself from home care services: I would never want to work in home care services because I want to have my workmates close by, so this is optimal for me.
The staff members' understandings of themselves appeared from the interviews as assemblages of elements from the dominant spaces. The fact that they saw themselves as responsible care workers on their own whilst appreciating the proximity of their colleagues is a possible combination and not a contradictory outcome. It appeared as though the liminality of the extra-care housing allowed an accommodation of a new combination of solo work and the support of workmatestwo organizational elements that were picked up from the dominant spaces, home care services and residential care, respectively. This was an outcome of novelty of extra-care housing. However, the fact that permanent staff was there because some residents needed care at the same level as residential care was overlooked in this argumentation.

Conclusions
The study revealed a situation of permanent liminality explored in staff reflections in which comprehensions emerged that sometimes fused, but sometimes refused to become a working whole. In some cases, their argumentations remained suspended, without logic or coherent categorisation. The extra-care housing as liminal space can in this respect be understood as a third space of negotiations, but also a heterotopia where the dominant spaces were challenged but also incoherently juxtaposed in staff reflections (Foucault and Miskowiec, 1986;Hetherington, 1997). The tensions between the sacerrimathe non-negotiable presumption of healthy residentsand the sacrathe consequently problematic presence of residents with deteriorating healthwere the main issue contributing to a perpetual liminal space that was impossible to reconcile (Szakolczai, 2003;Turner, 1970). The sacerrima, the healthy residents, seemed to be amalgamated with the concept of extra-care housing, making it inert and immobile. The diversity of the residents' care needs was considerable in the residence studied, and integrating this in the idea of an extra-care housing appeared to present the greatest challenge for the staff, thus creating frustration and disappointment. The staff members experienced a growing departure from their original view of the extra-care housing and their interpretation of its associated aims and goals. The care work had changed in certain respects that were not appreciated. Working in a continually changing facility seemed difficult to accept for staff, although this is the case in any housing type for older people. These circumstances give rise to liminal experiences with which the older people themselves have to deal (Black, 2006;Nicholson et al., 2012;van Wijngaarden et al., 2016). Contradictorily enough, the solution to the situation of permanent liminality would be to accept this transitional state in which the residents constantly change and new care constellations are continuously emerging (Szakolczai, 2003;Ybema et al., 2011). The argument that fragile elders should be relocated to residential care removes the "all-inclusive" character promised to the residents of the extra-care housing residence in the study, constituting a questioning or a misunderstanding of the basic idea of the project in its entirety. The study seems to have revealed a missed opportunity to argue for a home for life that harbours liminality and transitions as a natural part of everyday life in extra-care housing.

Funding source
This work was supported by the Swedish Research Council for Health, Working Life and Welfare, Forte; under Grant number 2014-00427and Kungsleden AB.