Contestation and constraint: commuting domestic workers and their employers in contemporary Kolkata

ABSTRACT The shift from live-in to live-out and part-time forms of domestic work has brought greater autonomy and bargaining power for many domestic workers in India, many of whom now live outside the cities where they work. At the same time, the advent of more impersonal and transactional labour relations has led to a series of everyday ambivalences and tensions, with workers and employers simultaneously embracing and resisting these more impersonal and transactional working relationships. Drawing on ethnographic fieldwork in Kolkata and rural West Bengal, and engaging with debates about ‘servitude’ and ‘pragmatic intimacy’, this article explores the work-life accounts of commuting domestic workers in Kolkata. It illustrates the pragmatism with which commuting workers approach and negotiate work and employers, and how they value, inconsistently, both physical/emotional distance from and closeness to employers, depending on their circumstances, experiences, and needs. Job security is, as the article shows more broadly, a key concern for commuters given intense competition and high turnover; thus, while commuters sometimes leave jobs for new ones, in some cases gaining better employment as a result, in many cases they stay and ‘adjust’.


Introduction
Domestic service in India has undergone several important changes in recent decades, including the shift from live-in to live-out and part-time forms of work and the shift from primarily men to primarily women workers. 1 These changes have in turn contributed to an increasingly transactional and impersonal relationship between workers and employers; as Ray & Qayum describe, the 'rhetoric of love', which previously bound servants to their masters/mistresses as part of a broader familial discourse, is being gradually replaced by the language of contract. 2 Despite these changes and developments, however, Ray and Qayum argue that the 'big house' remains the defining trope for understandings of domestic service in Kolkata and that the contract is 'at best partial and contradictory' -a paradox that they believe stems from the 'peculiar nature of a culture of servitude whose definition, content, and practice were scripted in a feudal/colonial past'. 3 Sen and Sengupta's more recent work supports Ray and Qayum's argument about the persistence of the paternalist family idiom but at the same time challenges the more general lens through which paid domestic work is understood in their study. 4 To view paid domestic work as 'servitude' is, they argue, to overlook the complex and contested nature of the relationship between employers and workers -particularly part-time/live-out workers who typically spend short periods in employers' homes daily and who describe their work in terms of kaaj (work/labour). 5 Sen and Sengupta prefer their own concept of 'pragmatic intimacy', which, they contend a) connotes the paradox of a relationship that is 'at once dominating and mutual, distant and intimate, exploitative and caring' and b) speaks to workers' agency, recognizing their attempts to construct reciprocal relationships with employers. 6 Drawing on ethnographic fieldwork in Kolkata and rural West Bengal, and engaging with debates about 'servitude' and 'pragmatic intimacy', 7 this article explores the work-life accounts of commuting domestic workers in Kolkata. It illustrates the pragmatism with which commuting workers approach and negotiate work and employers, and how they value, inconsistently, both physical/emotional distance from and closeness to employers, depending on their circumstances, experiences, and needs. The familial ideal, as I show, continues to be reflected in commuters' accounts of the 'good' employer, but commuters' also actively seek and maintain trusting, familiallike relationships with employers -analysis which is, rather, supportive of Sen and Sengupta's theory of 'pragmatic intimacy', as well as Ozyegin's concept of 'intimacy work'. 8 Job security is, as I show more broadly, a key concern for commuters given intense competition and high turnover; thus, while there is widespread dissatisfaction with employers and jobs, and commuters do sometimes leave jobs for new ones -in some cases gaining better employment as a result -in many cases they stay and 'adjust'. 9 The following section outlines the ethnographic context and methodology of the study, addressing the differences between 'commuting' and other domestic workers and providing detail about the approach and methods used in the study. Thereafter, the article explores workers' views on employers, discussing non-wage benefits, avoidance behaviour, 'adjustment' and job insecurity.

Ethnographic context and methodology
Commuting domestic workers make up a significant segment of Kolkata's vast informal economy and a significant subset of the city's 'live-out' or 'visiting' domestic workers. Unlike city-dwelling domestic workers, who either live with or near to their employers, 10 commuters live in rural villages and peri-urban areas outside the city, travelling in daily by train and other public transport. 11 Many find work through personal networks, doing thike kaaj (cleaning work) and/or rannar kaaj (cooking work) and typically working for short periods in multiple households daily, but others work through agencies, most often caring for children or elderly 'patients' and working in longer shifts (including at night), generally in single households at a time. While recognizing the shared experience of commuting for these different groups, and certain overlaps in experiences of work, 12 this article focuses particularly on the work-life accounts of the first group -those working through personal networks, doing thike/rannar kaaj. Living at often considerable distances from their employers, 13 these commuters are among the most autonomous of India/Kolkata's domestic workers, but they also face stiff competition in the neighbourhoods where they work (often those closest to railway stations) and are, equally, among the most insecure.
This article draws on ethnographic fieldwork undertaken in Kolkata and rural West Bengal over 16 months between August 2014 and February 2016. Working with Parichiti (a local support and advocacy organization) and three assistants, I recruited commuting domestic workers through Parichiti's networks, at train stations, and in placement agencies and I carried out observation and interviews as I travelled with workers to and from the city, as well as in their homes and communities. Given the immense power imbalances involved in researching paid domestic work, and my 'outsider' status as a white British woman, it was important to learn Bengali and spend time with workers to build trust and rapport. It was similarly important to negotiate (verbal) consent continuously, be flexible about where and how I conducted interviews/visits, and not impose any additional burdens or expectations on participants (for instance, by staying too long in their homes). I conducted loosely-structured and unstructured interviews, allowing participants the space to raise issues and themes that were significant to them. I also worked closely with the assistants to ensure that nuances of language and meaning were reflected as much as possible in my work. When transcribing and writing, I took steps to preserve participants' anonymity, changing personal names and in some cases other details.
I interviewed 31 commuting workers in total, carrying out repeat/regular interviews with five of these and speaking to many others informally over the 16-month period. Of the 31 participants, 28 were women and 28 were Bengali Hindus, many hailing from more rural, southerly areas of South 24 Parganas; only three were Muslim, one having migrated with her parents from Bangladesh as a child. The three men in my study, who fit my broad definition of a 'commuting domestic worker', worked in different roles (driver, 'male attendant', 'errand boy') while the women either worked as ayahs (carers) or did thike/rannar kaaj. 29 of those I interviewed worked in South Kolkata neighbourhoods (Ballygunge, Jadavpur, Tollygunge), commuting daily from villages and periurban areas close to Sonarpur and in a few cases further afield in South 24 Parganas; two women worked in the North, commuting instead from Naihati in North 24 Parganas. To supplement commuter interviews, I conducted observation in placement agencies and interviews with placement agency and Parichiti staff. In a more day-to-day sense, when visiting urban middle-class homes, I spoke with many employers and observed numerous employer-worker relationships, though, for ethical reasons, I generally avoided recruiting and interviewing workers for this ethnography in these spaces/through employers.

Love, care, and familial(-like) relationships
Given that it is increasingly common for Indian households to employ several live-out/part-time workers for discrete tasks rather than a single (live-in) worker for a range of work, 14 and that the employment relationship is, as Ray and Qayum and others describe, increasingly characterized by a language of contract rather than a 'rhetoric of love', 15 it is unsurprising that commuters and other live-out/part-time workers now often see themselves more as 'workers' than 'servants'. Like the citydwelling live-out workers in Sen and Sengupta's study, the commuters I met and knew during fieldwork typically described their work in terms of kaaj (rannar kaaj, thike kaaj) rejecting servant imagery and emphasizing their autonomy through statements such as 'I do rannar kaaj'. This was also often more pronounced with agency care workers, who were typically younger and better-paid than those doing thike/rannar kaaj and who often emphasized their 'professionalism', calling themselves 'centre girls' and using the English word 'centre' (meaning 'agency') to underscore this point. 16 Commuters are, furthermore, sometimes quite positive about (certain aspects of) the more impersonal, transactional relationships they share with employers, speaking about how they appreciate being able to leave employers' homes at the end of the day, and the degree of emotional separation entailed in more transactional part-time/live-out work. As Chitra, a commuting domestic worker in her mid-forties, explained, 'When I go to work, I do all the things I'm meant tocleaning, cooking, washing-up. I can finish my work, change and then leave. I don't have to worry about who's going to come to the house after I leave, what needs to be done in the evening'. 17 At the same time, commuters resent employers who fail to interact with them, or who in other ways treat them as kaajer lok -a term meaning 'working person/person who works', which is often perceived as derogatory. Malika, another commuting domestic worker in her mid-forties, complained, 'They don't care about us. They only have a relationship with your work'. 18 Furthermore, if employers do not evidence claims to familiality and affection with tangible acts of love and care, workers tend to view such claims as hollow and/or 'a ruse to make them work harder'. 19 Ranjani, a slightly younger commuter whom I met at Dhakuria Railway Station, was particularly critical of such employers/behaviour. 'Does an outsider ever become one's own (por konodin apon hoy)? I do good work, that's why they like me. They love my work. If I were to leave this job, they wouldn't recognize me on the street. As long as I work, they'll love me'. 20 Moreover, for many if not most commuters, it is important to try and establish or hold onto at least one close, familial-like relationship, where mutual trust is built over several years/decades, and where workers receive, in return for their loyalty and labour, important non-wage benefits and a degree of dignity, respect, and stability. Because of the changing nature of domestic work and the high rate of turnover, however, these relationships are becoming increasingly rare/difficult to establish -something which, again, contributes to their value.
Madhumita (34) explained that in two houses where she worked as a cook, she was expected to go about her work swiftly and silently -as if she had 'sellotape' on her lips -but in the third, her employers, a kind couple in their sixties, treated her 'like family' (ekdom poribar'r moton) and provided an additional Rs.1,000 per month for her daughters' education. 21 In that house, it's completely like home for me (ekta barite, amar puro barir moto). They love them [her daughters] a lot. You'll probably laugh but this man is quite old so I call him meshomoshai. 22 Shobna (30) too shared a close, familial-like relationship with one of her four sets of employers and, for this family, likewise performed tasks and roles that she would not for others. Shobna was employed by this family to clean and grind spices but during one visit to her home she told me that she was planning to forgo her usual responsibilities to help at the younger daughter's weddingwork which she had previously undertaken and enjoyed at the elder daughter's wedding.
Lauren: What sort of work will you do? Will it be different to what you do usually?
Shobna: Many people will be coming so I'll need to look after them, feed them. Since I've been working there a long time, I'll recognize a lot of people who are coming. I'll look after them and see what they need. I'll be there to take care of the guests. And the other work I do in the house -they've arranged for someone else to come and do that. [. . .] Sohini: 25 Will you get more money than you usually do?

Shobna:
Yes. For the wedding, I have to stay day and night so I'll get more money. I'll stay there and see who is coming and going. I'll also take people to the rented houses [wedding venues]. If they [employers] need something, they'll call me and ask me to bring it. I'll also make food for them [guests] and they'll eat and go to the rented houses. I'll go along with them by car. This is what happened last time, with the elder daughter's wedding.
Lauren: Are you looking forward to it? You seem happy.

Shobna:
Yes, and I'll get 2,000 rupees for the two days. I need to work a bit harder, otherwise why would they pay me this much? But I'm happy too. It will be good to see so many people and to look after them. They've left everything to me. [. . .] They trust me a lot (ora amake khub bishshash kore). They let me wash and eat there. I can do it all there. They open the almirah [cupboard] and ask me to lock it again. That's how much they trust me. 26 Shobna did not, as this excerpt shows, resent the change in schedule, which may or may not have been proposed at short notice. Rather she seemed to be looking forward to the work -greeting and looking after guests -which she had enjoyed in the past and for which she expected to be paid well. She was, moreover, proud that, owing to her previous efforts, the family had this time 'left everything' to her, trusting her to carry out the work with minimal supervision. Shobna was keen to stress how much she is trusted by these employers, using the example of the almirah -where, presumably, the family's valuables are kept -to underscore this point. This emphasis on trust is often discernible in domestic workers' accounts, 27 and, in India, it is connected to an increasingly prevalent discourse which casts domestic workers as (potential) thieves and criminals. 28 There is a similar significance to Shobna's comments about washing and eating given the persistence of avoidance behaviour, 29 and the fact that, in other households where she worked she and her son were treated with (casteand class-based) contempt.

Non-wage benefits
While from the employers' perspective, developing closeness with workers may be considered an effective way of making workers more pliable, reliable, and trustworthy, 30 engendering affection and intimacy with employers can, from the workers' point of view, make employers more generous, flexible, and amiable, and the employment relationship more stable, reciprocal, and trusting. 31 Familial-like relationships, as Madhumita's account shows, often yield important non-wage benefits for workers, one of the most highly-prized benefits being support for children's education. Given the insecurity of men's work and the range of other issues that domestic workers face, workers also greatly value access to credit. 32 Taking loans from employers can push women into debt and destitution, and by taking them workers also arguably sacrifice their ability to negotiate for higher wages in the longer term; however, loans from employers are vital to workers, both in times of crisis and when buying land/property. 33 Additionally, small gifts are appreciated by workers. Madhumita proudly pointed out that the set of windows leaning against her bedroom wall was given to her by the kind couple mentioned earlier; while Madhabi (mid-forties) recalled nostalgically how a former employer had once given her a pair of dolls for her sons to play with. 'They said to me, "Here, you have young children at home, take these home with you." They [her sons] used to dress them up and play with them. Still they decorate them and keep them clean. They won't part with them'. 34 Madhabi had kept these dolls, which were placed on a high shelf in her bedroom, and she also captured one of them in a photograph for our photo-voice project ( Figure 1). 35 Although domestic workers sometimes resent gifts from employers, regarding them as reinforcing the inequality of the employment relationship, 36 the commuters I knew generally accepted such gifts. This was in part because of the power relationship and intense job insecurity which made it difficult for them to refuse, but also because commuters are highly pragmatic and will accept and even ask for items (old clothes, food) which they know have little value to their employers but which can help to blunt the edge of the hardships that they endure. 37 Furthermore, because giving and receiving gifts is closely associated with ideas of kinship in India, 38 gifts from employers help commuters to feel that their relationships are more kin-like -something which speaks to wider, sociological literature on kindness 39 and is particularly important in the context of paid domestic work given persisting hierarchies of caste and class.
Offerings of food and drink are similarly valued by commuters, in part as a material contribution, but, again, as a recognition of their humanity. 40 Commuters often travel considerable distances to get to work (in many cases on empty stomachs), and they also cannot go home during the day for breaks and meals as city-dwelling workers can. Offerings of food and drink, although often meagre and sporadic, 41 thus become for commuters an expression of kindness and compassion on the part of employers. 'There are some who understand that we come from afar and that we may be tired and hungry. They think, "how can we eat in front of them. We should offer them food"'. 42 Shobna was offered a tiffin in all four of the households where she worked, but she was particularly grateful that in the house where she had a good relationship with employers -where she was trusted to lock the almirah -her employers also encouraged her to feed her son. 'Boudi likes kids a lot and says, "Even if we're not at home, you can keep him here while you work." She says, "There's biryani in the fridge, give some to your son." They keep things just for him -this family is very kind'. 43

Outsiders in the home
Although some workers are given food and drink in employers' homes and caste-based discrimination in relation to food and drink is generally much less overt than it was in the past, many others are given nothing at all when they arrive on employers' doorsteps tired and hungry after their journeys into the city -something which is understood by workers as a lack of kindness and compassion if not also a mark of caste-and class-based contempt on the part of employers. 44 In one of the other three houses where Shobna worked, she and her son also received very different treatment to that described earlier in relation to the kind boudi. Rather than welcoming Bijoy and offering him something to eat, this employer disapproved of Shobna bringing Bijoy with her to work (which Shobna was sometimes forced to do) and expected Bijoy -a boy of seven -to sit quietly at the foot of the stairs while Shobna carried out her work in the rooms above. This employer also disapproved of Shobna giving Bijoy her own tiffin to eat, so she would do this secretly, or else give him snacks which she would bring with her. 'They should at least give him a biscuit, but they don't. I have to spend money on biscuits for him'. 45 Shobna was, furthermore, not allowed to use the bathroom in this house (something which she again interpreted as a mark of indignity and disrespect), and on occasions when she had momentarily left her work to go and relieve herself elsewhere (often in the house mentioned earlier, where she had a close relationship with employers and where there was a separate, kaajer lok bathroom), her employers had scolded her for taking too much time.
First you don't allow me to use your bathroom, so I have to go to some other place, then you say, "Why are you late?" I cannot go there and come back quickly. In the rainy season, it takes time to wash and change my clothes. [ Toilet access is, because of structures of class and caste, a problem all live-out domestic workers face, particularly if employers do not have separate kaajer lok toilets, 47 and particularly during the rainy season and menstruation; as scholars note, the middle-class language of 'cleanliness' and 'hygiene', which is often used by employers when talking to/about domestic workers (particularly in relation to toilets), has its roots in ideas of caste, ritual pollution now being 'translated into real or actual dirtiness' and 'justified by the scientific discourse of hygiene'. 48 However, commuting workers face even greater difficulty since they typically spend long periods away from home and often cannot access (free/usable) toilets en route to or in the city, as well as in many employers' homes. 49 Commuters, like other live-out workers, sometimes relieve themselves in secret at work ('I just keep the door shut and do what I need to do'), risking being dismissed if caught; but more often they do not use the toilet at all while in the city -something which is both uncomfortable and injurious to health. 50 Just as workers speak of the various indignities they suffer in employers' homes with respect to food/drink and bathroom access, they also resent being rushed, complaining about employers who hurry and scold them for being late. Chitra often grew frustrated with employers who tried to hurry her, explaining: 'There are times when I talk back and say, "It won't work if you rush me like this"'. 51 Shobna also spoke of her wish to work in a single home (rather than rushing between several homes each day), where she could spread out her work and still be on time for her train home. 'Now the houses are all in different places and I'm often late'. 52 Due to the changing nature of paid domestic work, however -namely, the shift to live-out work and increasing role specialization -few have the sort of job Shobna describes. Most workers doing thike/ rannar kaj juggle several different jobs, rushing between employer homes, as well as to and from the city each day. Moreover, because such workers spend typically short periods in employers' homes daily, employers often wish to maximize this time, expecting workers to hurry through tasks with little or no breaks for rest. 53 Failure to complete the given workload places strain on the relationship 54 ; but equally, undertaking a heavy burden of work, or indeed extra work at employers' behest, can make workers lateboth for other employers' homes and workers' own homes at the end of the working day.
The hurried pace of commuters' daily lives explains why they tend to speak positively about employers who allow them to take their time doing their work. These employers are, not coincidentally, those with whom workers share familial-like relationships; indeed, workers often emphasize that they work in these homes as if they are working in their 'own homes' -statements which may be read as an attempt by workers to distance themselves from an understanding of domestic work as dirty and menial, 55 and thus in conflict with dominant ideas about femininity and domesticity, 56 but which also highlight the importance of autonomy and flexibility for workers.
Madhumita explained that in the kind couple's home, where she was treated 'like family', she spent two to three hours daily, even though she had less work to do in this home than in the other two homes where she worked for considerably shorter periods. In other words, she could space out her work more, and indeed often became distracted from her work, chatting with the man she called meshomoshai; in the other two homes, she was, by contrast, expected to focus on the task at hand, cramming her work into much shorter periods -an hour and a half in one home and just forty minutes in the other. This arrangement had some advantages: by doing only what was required in two of the three homes, Madhumita had more time to spend in the other, the third home, where she was respected and where she received in return for her loyalty and labour important non-wage benefits.
Another key theme emerging from workers' accounts, and linked to the issue of autonomy, is suspicion and monitoring. The Bengali 'mistress' or boudi is often at home during the day to supervise workers, and there is also a prevalent discourse about workers as (potential) thieves and criminals, which, although not new, has intensified with the move to apartment-living and parttime/live-out work. 57 Employers are, in many cases, fearful that their domestic workers (often complete strangers) will steal from them or harm them -ideas which are reinforced through sensationalist newspaper reports and advertisements, 58 and which have contributed to a cycle of suspicion, dismissal, and turnover. 59 Domestic workers are often the first to be accused when something goes missing from the house and employers are increasingly opting for 'verified' workers (those supplied through agencies and/or registered with the police). 60 The discourse of suspicion and criminality plays out most starkly in the case of gated communities where workers must negotiate identity checks, body/bag checks, and CCTV monitoring, 61 but it also plays out in other contexts -including in more ordinary middle-class homes and apartments.
The phrase khit khit kore, repeated time and again by workers, encapsulates the experience of being followed about, picked at, and scolded by employers who are in turn perceived as being constantly irritated and dissatisfied. 'Some are always behind my back, khit khit kore. They ask me to wash my hands constantly and tell me not to spill things'. 62 Even if workers are not suspected of trying to steal from their employers, such treatment is clearly infantilising and reflects a lack of autonomy at work. There is, moreover, often a distinct class/caste flavour to employers' admonishments, with workers being instructed to clean their hands or avoid touching things: 'They say to me, "don't open this, don't touch this." [. . .] They hate people [like us] (ora lok ke khub ghrina kore)'. 63 Unknown women workers are thought to be less threatening than unknown men workers, but there is still, from the employers' perspective, a need for supervision given concerns about theft/ criminality, pollution/'hygiene' and immorality/sexuality. On the latter, there is an undeniable link between domestic work and sex work, with studies showing how women combine and/or move in and out of these forms of work, 64 but there is also a more insidious narrative, discernible in many employer accounts, which constructs domestic workers, and poor, working-class, low-caste women generally, as 'promiscuous' and 'immoral'. 65 This narrative, which is not dissimilar from that which positions men servants as potential sexual predators, 66 and which is linked to widely-held notions about women's sexuality as 'dangerous, threatening, and uncontrollable', 67 explains why many employers fear that domestic workers will engage in illicit sexual activity in their homes. 68 It also renders domestic workers, and poor, working-class, low-caste women generally, (particularly) vulnerable to abuse, 69 though because of the strength of ideas about women's 'honour' and the associated feelings of shame, workers tend to only speak about such abuse indirectly -as something that happens to others. 70 In various ways, then, whether it be through avoidance behaviour in relation to food/drink or toilet access, or through mistrust and monitoring, workers are made to feel different and 'other' in employers' homes and it is this treatment which they resent the most. 71 This experience of 'otherness' applies to all kinds of domestic work/er in India since this labour has been, and continues to be, shaped by structures of caste/class and colonialism 72 ; but it is arguably more pronounced in the case of commuters, who come from outside the city and thus are not only figurative but also literal 'outsiders'. Indeed, as interactions with middle-class employers demonstrate, commuters are often viewed as a dirty, dangerous rabble to be approached with caution and care. 73 They are, additionally, often stereotyped by city-dwelling domestic workers -blamed for creating competition and pushing down wages. The latter of course relates to the wider context of job insecurity and the lack of alternative, 'stable' employment for poor women and men in West Bengal 74 -an issue which will now be discussed in greater detail.

The work of 'adjustment'
Because domestic workers' bargaining power generally decreases over time, with workers often failing to receive pay increments and in many cases taking on more work for the same pay, it follows that there is a considerable degree of turnover in the sector, with workers frequently leaving jobs for new ones. 75 Santi (26) explained that she had once walked out an employer after finding out that she would not be getting a Puja bonus, 76 while Mithu (45) recounted how she had left a job after her employer's daughter had started to cause trouble: 'She used to complain (khit khit korto). She'd say, "this is not right", "that's not been done properly" [. . .] I thought, "I won't go to work [here] anymore, I'll go to some other place"'. 77 A middle-class employer also once reflected that it is in such moments -when a 'maid' fails to show up for work -that one's dependence on them is most keenly felt: 'As an employer, you will always feel that you have no rights and they are calling the shots. They can just not turn up one day and you're left waiting'. 78 In many cases, however, workers decide to stay and 'adjust' to jobs they are dissatisfied with. Namita (mid-thirties) resented the stringent security measures in the complex where she worked and was unhappy with both her work and her pay; however, because she had only ever worked in this complex and lacked the networks other women had, she felt she could not leave. Shobna, too, continued working in the house where she was not allowed to use the bathroom and where she and her son were treated with contempt, in large part because she had little time to look for alternative work.
For Ruksana, a Muslim woman in her late-thirties, other factors influenced her decision to continue working for (Hindu) employers who had repeatedly let her down.
At Eid I asked them to give me Rs.500 so that I could get clothes for my kids. She said, "We can't give you so much money." Another time, I said, "I've been working for you for fourteen years. Won't you help with my daughter's wedding?" I cried in front of her but she didn't give me anything. I felt hurt. I enter that house at eight in the morning and I work continuously until two-thirty. I know everything that's in their house but I don't think about it. They give me Rs.4,500 at the end of the month. I just take it and leave. 79 Ruksana clearly felt cheated by these employers, emphasizing how hard she worked for them, cooking and cleaning, often from the early morning to the mid-afternoon without a break. It is also likely, given the earlier discussion about workers' bargaining power -and indeed Ruksana's description of her heavy workload -that this workload had also increased over time, without any commensurate rise in pay. Her point about knowing 'everything that's in their house' could be read as implying something about her employers' relative wealth (and thus their ability to provide additional financial support), but it could also be read as implying something about the temptation to steal, and in turn her own restraint. The fact that she does not even 'think about it' underscores the broader point she appears to be making about loyalty and trust.
Although Ruksana had often considered leaving this job, periodically asking others to help her find new work, she had nevertheless stayed -a decision that was likely connected to the increasing intolerance towards Muslims in India. 80 Muslim workers often find it difficult to secure domestic work jobs, particularly cooking jobs 81 ; and, even in cases where they do secure jobs, problems can still arise, with workers being branded 'illegal' and so on. 82 Ruksana may have been worried about finding new work, or work where she could be similarly open about her Muslim identity; she explained how she had previously worked with a Hindu name and was reluctant to do this again. Ruksana's wage was also relatively substantial (compared to the wages of other domestic workers); and, importantly, she had not been dismissed for asking for a bonus and additional financial helpsomething which suggests that there was at least some understanding between herself and her employer.
For Anjali, a single mother in her early forties, it was a combination of ill-health and financial worries that constrained her ability to change jobs. Recovering from a recent surgery for breast cancer, Anjali had various medicines and treatments to pay for, as well as rent, food, and fuel for herself and her youngest daughter. She was, in short, under severe financial pressure, often struggling to pay her various outgoings with her meagre monthly earnings (Rs.3,200); and indeed, it was this severe financial pressure that led her to both keep commuting for domestic work and periodically ask her employers for additional financial help. One couple, whom, she explained, did not usually talk harshly to her, typically refused such requests, explaining that they already paid her a fair wage and sometimes then critiquing her work: 'They say to me "We pay you Rs.2,000. Why don't you do this properly?"'. 83 Anjali also frequently clashed with this couple over leave, explaining: 'When I tried to take leave on the eighteenth of this month, they said, "Why are you taking leave?" So, I said, "What are you saying? Are you human? I need leave. I have to go to the doctor and I still have three holidays [to take]"'. 84 Anjali's mentioning of her illness, which was likely regarded as a form of emotional manipulation by her employers, was arguably intended to induce feelings of guilt in her employers, and thus to make them more flexibile regarding leave. 85 The emphasis on 'human need', articulated in the phrase 'are you human?' and discernible in similar ways in other workers' accounts, was likely for the same effect, as a means of appealing to and challenging her employers on the issue of leave. While, then, workers sometimes reinforce ideas about themselves as machine-like and indefatigable, 86 they at the same time resist and challenge this narrative by asserting their right to time off and rest. 87 Indeed, whether in the context of work or commuting, workers often make a similar point: 'You have to take a break, you can't always go on like this'. 88 Anjali similarly explained how she challenged her employers over pay. Indeed, when the aforementioned couple refused to give Anjali money and told her that she was free to leave if she was not happy with her pay, she snapped, saying, 'You pay me Rs.2,000, but what value does this have these days?' Anjali was, however, careful not to go too far, indeed explaining how, after this altercation, she had bit her tongue. She was aware that too much back-chat could result in her wages being cut or worse: dismissal. Anjali spoke directly about this issue of vulnerability later in our conversation, explaining how she cannot easily find new work and must be pragmatic and 'adjust': 'Even if I feel upset, what can I do? I have to 'adjust". 89 With so many workers competing for jobs, Anjali was arguably concerned about a gap in income and how she, as a single mother with serious health issues, would manage. Because of her ill-health, moreover, her options were already limited. These various factors, then, constrained what her employers saw as her 'freedom' to leave.

Job insecurity
Anjali's account, like many of the others discussed here, highlights a broader problem affecting commuting, as well as other live-out domestic workers: job insecurity. The degree of competition and turnover in contemporary domestic work is what leads workers in many cases to settle for unsatisfactory but (relatively) stable jobs. Labour is cheap and plentiful, and so the rate of dismissal and turnover is correspondingly high. As we have seen, workers also sometimes walk out on employers, which again contributes to turnover. The discourse of suspicion has a similar effect since workers and employers in many cases do not know and trust one another, and workers are frequently dismissed on the grounds of (suspected) theft. Workers who fail to turn up to work or who are repeatedly late thus run the risk of being fired, as do those who refuse requests for more work or ask for more pay/money. Additionally, employers sometimes perceive an advantage in hiring new workers. 90 The extreme heterogeneity in the sector means that new workers are not always more expensive: employers may get new workers at lower wages while workers who leave jobs may get new jobs at higher wages. 91 New workers are also less likely to require and ask for non-wage benefits (including money for medical treatments and operations) 92 ; and new workers are sometimes perceived to be more hardworking, and thus less likely to skimp on their work and take leave without notice. 93 While all live-out workers struggle with job competition and turnover, commuters are in a particularly tricky position since they frequently experience problems en route to the city and arrive late to employers' homes. Their numbers are also often highly concentrated in the neighbourhoods where they work (typically those close to railway stations), which in turn leads to stiff competition and undercutting. Sen and Sengupta note how city-dwelling domestic workers accuse commuters of stealing jobs and pushing down wages, and there is reason to believe that this is happening since commuting workers, like new migrant workers, have fewer/smaller networks and greater financial compulsions (which force them to undertake long, arduous journeys) 94 ; and, as mentioned earlier, commuters sometimes receive food in employers' homes, which may be considered payment in kind, thus reducing the cash component of the wage. 95 City-dwelling workers must, as Sen and Sengupta describe, accept very low wages or risk being replaced by commutersthe latter being a particular source of anxiety during periods of leave/absence. 96 Commuters not only replace workers who live in the city, but also one another. Madhabi was one of several women to speak about being dismissed by an employer after taking leave, in her case to visit Haridwar 97 -a trip she had been planning for some time and which she had discussed with her employer. Madhabi had been in this job for a year and believed that she had an understanding with her employer; although they had not liked the idea of her taking leave, they had, in the end, agreed to the trip, providing that Madhabi find a replacement (badli) to fill in for her while she was away.
On returning to Kolkata, then, Madhabi was maddened to find that she had been permanently replaced, and that her employer had given the other woman -the replacement -money Madhabi believed she was owed. 'I was meant to get Rs.600 but they gave it to her. [. . .] I said, "Where will I go?" but they wouldn't listen'. 98 Madhabi's employers had told her that they had simply preferred this other woman's work, but it is likely, given the prevalence of undercutting, that the replacement had (also) offered to work for a lower wage. Indeed, the fact that Madhabi later reflected on this incident differently, commenting calmly that it had perhaps been 'God's will' for her to find new work, suggests that undercutting occurs frequently among commuters and is also something she may herself have done in the past. Madhabi had spoken about the significant number of women travelling to work in the city, noting that the crowd was 'much bigger than before'; she was, then, arguably aware of her own dispensability and that by taking leave -pre-arranged or not -she was taking a serious risk, rendering herself vulnerable to being replaced by the badli or any of the other women who were ready at any given moment to step in and take her place. 99 For those who lack networks in the city, or those who want and are able to work longer hours and earn more money, 100 placement agencies or 'centres' can play an important role in helping workers to find and connect with employers. Several of my participants worked through agencies when I knew them, and many others had worked through agencies in the past, often working night shifts as ayahs (carers) and earning more money as a result. However, while agency work has its benefits (for those who are able to do it), it can nevertheless involve a similar degree of turnover and instability with jobs lasting anywhere between one day and several years -in the case of elderly care, for example, until a 'patient' recovers or dies. It guarantees neither uninterrupted income nor familiality/reciprocity with employers -both of which are important to workers.

Conclusion
Commuters are among the most autonomous, as well as the most insecure of the various categories of domestic workers in Kolkata/India -a point which not only helps to explain the contradictions in their accounts, but is also in keeping with findings from other parts of the world where increased flexibility has meant increased insecurity for workers. 101 While commuters reject servant imagery, describing their work in terms of kaaj and embracing elements of more transactional, impersonal relationships with employers, they also reject the label of kaajer lok, resenting employers who fail to interact with them, or who in other ways treat them with a lack of respect. Moreover, like other live-out workers, commuters value familial-like, reciprocal relationships with employers and will seek to establish and hold onto at least one or two of these closer relationships -in large part because of the important non-wage benefits and level of protection, stability, and dignity these relationships can bring, but also because of the wider context of instability and indignity within which paid domestic work in India continues to take place. Furthermore, flexibility is particularly important to commuters, and it these trusting, familial-like relationships which offer workers the greatest flexibility to negotiate leave and other benefits.
Because of the high degree of dismissal and turnover, however, opportunities for forming close, familial-like relationships are increasingly hard to come by. New workers are particularly vulnerable to being replaced, but long-standing workers can also face this threat, particularly if they ask for leave or an increase in pay. Workers do sometimes leave jobs they are dissatisfied with (longstanding relationships are valued by workers, but only where there is reciprocity with employers); however, in many cases, workers decide, for various reasons, to stay and 'adjust'. Older workers and Muslim workers face a particularly difficult set of constraints, as do single mothers and those with health issues: but, for commuters generally, the decision to leave a job is risky given the intense competition which renders them and other live-out workers vulnerable to undercutting and replacement -particularly during periods of leave/absence. Furthermore, while agency work provides commuters and others with a route into what is an incredibly overcrowded and cutthroat market, it can entail a similarly high level of turnover and instability.
In line with Sen and Sengupta's work, then, the more general picture that emerges from commuters' accounts is one of intense pragmatism and negotiation. The high degree of turnover is, as noted, in part caused by workers walking out on employers. Like city-dwelling live-out workers, commuters reject the pressure from employers to work continuously, demanding time off on moral and humanitarian grounds 102 ; and their accounts contain numerous other examples of everday agency and resistance 103 -including talking back, humour, and demanding time off. Furthermore, even in cases where workers decide to stay and 'adjust', the decision to stay should not be read only as an indication of their dependence on employers or their powerlessness in what is an intensely crowded and cutthroat market, but also as evidence of their pragmatism and intense 'intimacy work'. Given the absence of formal rights and protections for workers and the degree of turnover that is entailed in live-out/part-time work, it is these steady and stable relationships which appear to offer workers the best hope of gaining at least some level of protection and dignity.