Assemblages and affect: migrant mothers and the varieties of absent children

Drawing from stories told by migrant women in Hong Kong, this article builds on previous studies of ‘left-behind children’ and calls for greater attention to the spectrum of sorts of absent children and to the formation of queer or less normative forms of migratory families. Taking a two-pronged approach, I present an on-theground ethnographic and affective approach through several vignettes, and consider key elements of a more mid-range and distanced ‘global assemblage’ approach to the institutions and expert knowledge that shape the experiences and practices of migrant mothers, migratory families, and the spectrum of absent children. This article posits that one’s biological children, perhaps the most familial of kin, can become familiar or even unfamiliar strangers through contemporary processes, technologies and practices of migration and separation, and that the process of migration makes and unmakes conventional and unconventional sorts of families. While affective and assemblage approaches are independently valuable, combined they offer richer understandings of the complex interplay of factors – at various levels – that shape normative and queer families and different types of children’s absences.

greater attention to the spectrum of children's absences and to queer or less normative forms of migratory families.
International labour migrations, especially those of Southeast Asian women, have increased greatly in recent decades. Many scholars have examined the well-being of migrant workers' left-behind children in Southeast Asia. Such studies often counter the prevailing and widespread 'dismal' public view of such families (Parreñas 2005: 30) and public perceptions that 'offer a dark prognosis of the impact of parental absence on the children left behind' (Asis 2006: 46) in contrast to the assumed well-being of children in idealized non-migrant families. Several recent large-scale surveys examining left-behind children in two-parent migrant families in which one or both parents migrate support the view that these children are better off than commonly assumed. Materially, educationally, physically and emotionally they often do better than children in two-parent non-migrant families (Asis 2006: 51-6;Graham and Jordan 2011;Graham et al. 2012;Hoang and Yeoh 2012;Hoang et al. 2014). Scholars also highlight the ways in which new media can facilitate a 'virtual' presence, allowing for long-distance mothering (or parenting), with somewhat mixed results, as parents and children phone, text and Skype across vast distances and time zones (Asis 2006;McKay 2016;Madianou and Miller 2012). 1 Authors of survey-based studies of left-behind children in two-parent families note the practical need to narrow the variables, and they acknowledge the limitations of doing so. As Asis (2006: 30) writes of the 2003 Philippines survey: it was limited to children whose parents were living together (as a proxy for 'stable families'), or in the case of children of migrants, their parents were not together because of migration. The children in the study, thus, belonged to twoparent families. Limiting the study to this group of children helped control for variations that could be introduced by different family structures. On the down side, the results can only hold for children belonging to two-parent families, which are only one kind of family.
Regarding the 2008 Child Health and Migrant Parents in Southeast Asia (CHAMPSEA) study, which serves as the basis for numerous publications, Graham and Jordan (2011: 767) write: The sample excluded single parent households. Qualifying households were those in which either (a) both parents had been usually resident at the same address as the target child for a period of at least 6 months prior to interview, or (b) one or both parents had been working overseas for a continuous period of at least 6 months prior to interview. 2 A focus on 'one kind of family' -two-parent ones -however useful for questioning dominant state and public views of deviant migrant families and damaged or delinquent left-behind children, can also be problematic. The practical reality of a two-parent family can easily shift through time, and its existence might also depend on who one asks. Many of the officially married migrant women I met in Hong Kong had long been estranged from their partners, some before they migrated. Others identified themselves as married but officially and legally were not. The situation, and who asked about it, largely determined the answer. Moreover, however methodologically useful a twoparent family (married or not) might be for survey purposes, it serves to highlight certain familial patterns and obscure others. Such studies of two-parent families suggest that children withstand the challenges of migration quite well, possibly reassuring policy makers that migrants whose children are doing poorly are unusual or in the minority. While not intending to examine family forms, parents' relationships, or types of absent children, studies that focus on heteronormative two-parent families can nonetheless inadvertently normalize those experiences and, in addition, would exclude many migrant women and their children (Constable 2014;Pratt 2012).
One goal of this article is to shift the focus away from heteronormative two-parent families and their associated left-behind children and to focus instead on the sorts of families and absent children such a lens overlooks or obscures. The picture changes when we look beyond normative parameters. The family forms of single mothers or fathers, remarried mothers, or adoptive lesbian mothers or couples, and the forms of children's absences based on whether the child was adopted (locally or internationally), sold, died or was 'thrown away', form a spectrum. Children's absences (like those of their parents), take many forms, with many shifts and permutations. At one end of the spectrum are stories of assumed (or hoped to be) short-term or temporary parent-child separations. Mothers and others (including scholars) commonly tell such stories about children who are far away and sorely missed, perhaps cared for or fostered by family members. At the other end of the spectrum are more permanent separations. Two types of children's absences that are less common and that the scholarly literature on migration has largely overlooked are those caused by a child's death or by closed international adoption (Butt 2015;Constable 2014Constable , 2016. Stories of death and of children said to have been 'given away' for adoption or 'sold' are less commonly heard, and in conversation often expressed as fragments of stories, or told in whispers. The subject of gossip and rumour, and sometimes reported in the news, are stories of abandoned babies, 'thrown away' babies or foetuses, or of children dying under suspicious circumstances.
Below are several short vignettes that illustrate varieties of migrant mother-child separations and absent children, described mostly by Indonesian and Filipina migrant workers who were single parents, at least for a time. The children span a spectrum of sorts of absences, ranging from those assumed to be only temporarily left behind on one end, to children who are permanently absent due to events such as international adoption or death on the other. Next, shifting away from the ethnographic immediacies of the vignettes, I consider how migratory technologies, especially technological, political and ethical reflexive practices, such as government apparatuses of citizenship and governance, border-crossing social welfare regimes, and competing systems of morality and expert knowledge, shape these experiences.
Attention to assemblages marks a shift from more conventional closed, demarcated, or localized social spaces studied by anthropologists. It also shifts away from what Deleuze (1992) and Foucault (1979) described as 'disciplinary societies' in which there are enclosed spaces such as prisons, hospitals, or schools towards what they describe as the newer 'societies of control' that reach beyond conventional institutional or social spaces. Assemblages, like Appadurai's (1996) 'global imagination', direct attention away from bounded states and social spaces to the multiple, shifting, often competing, contingent forms and logics of development, progress, migration and reproductionthe networks and technological links that, according to Deleuze (1992), flow like liquid or gaseous substances. These border-crossing logics, networks, flows, or assemblages are useful for understanding migratory family formations, the borders and boundaries of which have become more fluid, existing or ceasing to exist across unbounded space.
Among the relevant global assemblages are migratory regulations, circulations of media and non-governmental or civil society reports, and other forms of 'expert knowledge', including scholarship that shapes migration, migratory families and our understandings of them. Scholarship and expert knowledge can reflect, produce and reproduce normative and hegemonic views of migratory processes and families. In calling for the queering of paradigms in migration studies, Manalansan (2008) argued that much scholarship on domestic workers presumes a limiting heteronormative lens. I argue that, since patterns of global migration are often highly gendered and heteronormative in their logic and practice, such scholarship presents an important part of the picture. Following Manalansan, focusing on normative gendered patterns clearly risks reinforcing and reproducing such hegemonic and heteronormative assumptionsregardless of their prominence -rather than elucidating their cracks, openings and opportunities. It thus inadvertently risks silencing other, less normative, possibilities.
Ultimately, as I argue below, some forms of global knowledge obscure and others reveal non-normative family formations, or 'queer families' -not in the sense that they are necessarily homosexual, but in that they do not follow dominant and normative monogamous marital or sexual expectations. 3 Single-mother migrants, and gay and lesbian migrants with children or absent children, have been largely excluded from studies of migration. Focusing on some of the exclusions reveals a bigger, messier, ever-changing, and less hegemonic and normative array of possibilities. Stories of the making and unmaking of different sorts of migratory families, among many other stories we might tell, reveal ever-fluid migratory family formations and types of children's absences -not just the dominant, normative, or ideal types.

Ethnographic vignettes: the spectrum of absent children
The following vignettes draw from ethnographic research among migrant mothers in Hong Kong in 2011 and 2012. Methods included formal and informal interviews with more than 55 migrant mothers and extensive participant observation. Some stories draw from updates (in person or online) from women in Hong Kong, Indonesia and Singapore (see Constable 2014: chapter 2).

Left-behind children, fostering, and temporary caregiving
Indah, an Indonesian domestic worker whom I met in Hong Kong in 2011, gave birth to two children after she overstayed her domestic worker visa. Her partner, a South Asian man whom she called her 'husband' (though they were not officially or religiously married) was the children's father. When Tika, the eldest, was five, Indah left her partner because of his drug addiction and inability to support the family, and surrendered to the immigration authorities. In her court hearing for overstaying her visa, the sympathetic judge granted her a suspended sentence on condition she immediately return to Java with the children.
Upon their return to Central Java, Indah and the children lived with her parents. As a single mother with visibly mixed-race children, they were subject to teasing, taunts and gossip from neighbours, which created tension with Indah's family. Within less than a year, given Indah's sense that she was a burden to her impoverished parents and ever-complaining mother, she felt compelled to leave through the only available means. She took up with a local Indonesian man long separated from his wife and especially kind to her children. Although unmarried (he had not officially divorced), she spoke of him as her husband. They had a baby and, in 2014, when the baby was under a year old, she went to work as a domestic worker in Singapore.
While Indah was in Singapore, her three left-behind children moved between caregivers in Central Java and Jakarta. At first, Indah's partner insisted that he could care of all three children with his mother's help, but that proved impossible. Tika went to stay with Indah's mother nearby; then, because she was unhappy there, she went to live with her aunt (Indah's sister) and uncle in Jakarta during the school holiday, then back to her grandmother's when school resumed. Nina, several years younger than Tika, moved between the same three caregivers and, by 2016, had settled happily with her aunt and uncle in Jakarta. The baby was in the care of Indah's husband and his mother.
Each month Indah sent remittances to the three caregivers. When she spoke to them (when her stay-at-home employer was out or during the two hours she had off once or twice a month), it was often stressful and upsetting. Her mother and partner complained that the money was insufficient to cover the children's expenses. Indah worried about Tika, who complained about her grandmother's cooking and shouting. Old enough to remember her birth father, Tika spoke of him incessantly. She begged Indah to buy her a folding bicycle, a new school uniform and other things Indah could not afford. Tika suffered taunting and bullying at school, where the other children told her that Indah was not her mother and that she was adopted.
Nina did not remember Hong Kong and seemed happy with her aunt and uncle, but she also began asking about her father. Pointing to a recent photograph of Nina and her caregivers, Indah said that her sister and brother-in-law loved Nina. Their children had grown up and they wanted, but could not have, another one. Indah's partner complained of missing Nina and rarely seeing Tika because Indah's mother scolded him at every turn, said he was 'not her father' and had no right to see her. Indah's mother's attitude was fuelled by the couple's lack of marital legitimacy, but they planned to marry when Indah returned home and his divorce was final. Indah's mother complained about the burden of having Tika, but would not give her up, perhaps because Tika ensured her some of Indah's remittances.
When we met monthly in 2015 and 2016, Indah would cry and say how much she missed her children, worried about her finances, and struggled to tolerate her demanding, controlling and ever-present employer. Ultimately, Indah decided not to renew her contract, despite the offer of higher pay and more free hours. She reasoned that she could be with her children and have no money, or be without them, lonely and depressed, yet still not have enough money to make ends meet.
Indah's separation from Tika, Nina and the baby is one type of absence. Tika, Nina and the baby are left-behind children. Yet, partly due to their ages, each child is differently positioned and each responded differently to Indah's absence. Like many migrant parents, Indah tried to talk to her children regularly, but she faced time constraints. Nina sometimes spoke hesitantly, the baby was shy, and Tika often spoke angrily, with guilt-inducing questions and demands. 4 Indah's worries about each child differed, though she missed 'all of them' and desperately 'want[ed] to be together again'.
In mid-2016, Indah anticipated going home, marrying her partner, and living together as a family. Within a month, however, she and her partner split up, and three months later she re-entered the cycle of migration and was back in Singapore in an effort to support her children. Indah's family and its transnational permutations (her time in Hong Kong; her first two children's father now living in South Asia; her status as a single mother; her unrecognized marriage; her children's ever-shifting and temporary childcare arrangements) illustrate the flexible and negotiated meanings of such concepts as family, marriage, two-parent households and left-behind children. Such variable familial and household patterns are, notably, not exclusive to regions of transnational migration. Flexible patterns of child fostering, child borrowing or lending, and adoption across a wide range of kin and non-kin caregivers have taken place in Indonesia and Malaysia long before the current era of women's international labour migration (see Beatty 2002;Butt 2008Butt , 2015Carsten 1991;Geertz 1961;Jones 2002;Newberry 2010;Schrauwers 1999). Nonetheless, a range of migratory assemblages have expanded and intensified these patterns.

Adoption and 'selling babies'
Mia, like many other migrant mothers, arranged a birthday party at a migrant charity organization she frequented for her son when he turned two. She brought Indonesian food, a birthday cake with candles and party favours for the children. Many migrant mothers and their children attended. The striking difference between this party and others I attended in 2011 and 2012 was that Mia's son was not there. She did not know where he was. Someone had adopted him 18 months earlier. Mia imagined, and was quite sure, that this was a wealthy expatriate family and that he was either living in Hong Kong's prestigious mid-levels or that he had gone to England, where he would get a better education and a better life than she could give him or the children she had left in the care of her mother in Central Java. Like other women I knew, Mia said she 'gave him for adoption'. She did not use the term 'give up' but rather 'give' (kasih), as in giving a gift of a child to someone who could not have one. She left a note for the social worker to put in her son's file in case he sought her out when he grew up. She was hopeful that one day he would. Absent children, like Mia's, may be present despite their absence or absent despite their presence. When Mia (whose husband had left her before she went to work abroad) finally returned to Central Java in 2012, with a new baby girl in tow, her sons, who had been living with their maternal grandmother, were very angry with her. It was, she explained, as though the absence continued in their presence. The eldest ignored her or lashed out at her. He was angry because she could not buy him a motorcycle, or fancy clothes. He asked why she had returned without money. Her younger son was happier to see her, but jealous of his new sister.
Elle, another woman who gave her baby for adoption in Hong Kong, gave birth, then caught but a glimpse of her baby. In her groggy, sedated, exhausted and emotional postpartum state, she did not see him for long enough to ascertain the child's paternity or guess at his racial identity (perhaps I was more curious about that than she was). After labour, relieved that the ordeal was over, she was eager to put the episode behind her so that she could return to work and be 'back to normal' as quickly as possible.
Elle's story of conception involved what she described as a drunken blackout and gang rape. I had no reason to doubt her version of events, but the social worker who introduced us stated point blank that such stories of rape are just a cover for women's immorality. The other women with whom Elle lived during her pregnancy thought that she was a 'tomboy' -a butch Indonesian lesbian -who did not like men. She made no attempt to convince them otherwise.
Elle expressed sadness to me about the ordeal of pregnancy and childbirth, but she did not second-guess her decision to give the child for adoption. Adoption was a solution to the larger upset that included pregnancy, lack of clarity about the father, and a necessary absence from work. A few days after she gave birth, Elle hinted that she regretted not having had a better look at her baby. But when she finally saw him, very briefly, when she formally relinquished her parental rights, it pained her. 'He just looked Indonesian like me', she said.
Unlike several other women I knew, or knew of, who frequently changed their minds about giving the child for adoption, Elle never wavered. She accepted the NGO's assurance that she was 'doing the right thing' and that 'he will have a better future' with Hong Kong expatriate parents (those most likely to adopt non-Chinese-looking babies) than she could provide for him in Indonesia. After the birth, she briefly visited her family in East Java; they were unaware that she had given birth. She then returned to Hong Kong and resumed work. She intentionally lost touch with the women who knew her during the pregnancy.
Ree, in her late teens when she got pregnant in Hong Kong, repeatedly changed her mind about adoption. Her employer terminated her contract on learning of her pregnancy, so she overstayed her visa and worked illegally. As her pregnancy progressed, she learned about an NGO that could help her and went to stay in its shelter (as had Elle a year earlier). The residents were migrant workers (Indonesians and Filipinas), but all intended to keep their babies, so, like Elle, she kept quiet about her adoption plans.
Within a day of her baby's birth, Ree changed her mind and, to the surprise of the Indonesian caseworker who had been her main support and confidante, she left the hospital with the baby without being discharged. For a few weeks, Ree and the baby were off the grid. Ree did not respond to messages or calls from her caseworker.
Eventually, the caseworker heard that Ree had been seen in the New Territories makeshift village where she had lived before going to the shelter. I accompanied the caseworker to look for her.
The baby was staying in an illegal structure with a lesbian couple (both Indonesian overstayers in their early twenties) who said Ree had given them the baby and that they planned to raise her as their child. The caseworker explained gently but directly that the Hong Kong authorities would not allow them to keep the baby, that adoption in Hong Kong -unlike fostering in Indonesia -does not simply depend on the mother giving them the child. At worst, they (and Ree) could face criminal charges for kidnapping or child trafficking if they tried to keep the baby. At that point, the women seemed to lose confidence and explained that Ree had asked them to take care of the child and had said she would return but never did. Their impassioned plan to keep the baby was short lived. They asked us to look at the baby, who was listless, visibly ill, and needed medical care. The caseworker concluded that the couple lacked the means, the knowledge, and the legal right to care for the baby.
Within a few days, the caseworker arranged to 'coincidentally' appear at a meeting the couple had set up with Ree. The caseworker appealed to Ree, as a fellow Indonesian and herself a mother, explaining that the NGO for which she worked had not reported Ree for child neglect or trafficking (for abandoning or selling her child), and said that had they done so she could have ended up in prison. Instead, they wanted to 'help her do what was best for the child'. Ree, who had abandoned the sick baby in a panic, agreed to relinquish her maternal rights.
I later heard that the baby went to hospital and, when healthy, was placed in an institutional home pending adoption. Within the next few months, the NGO also opened a separate shelter for women who did not plan to keep their children. That way, women like Ree and Elle would avoid social pressure from their compatriots to keep the baby. 5 These are but a few of the stories of adoption I heard over two decades. I heard several stories of -and from -Filipina self-identified lesbian domestic workers, tomboys and other unmarried single women (including three sisters) who had adopted the unwanted babies of other domestic workers, had birth certificates made in Hong Kong (before DNA tests were required), and took the children home to the Philippines as their own. Such babies were sometimes raised by the adoptive mother, but most were raised mainly by the women's mothers or sisters or by hired local helpers while the adoptive mother went back to work overseas. One lesbian domestic worker proudly showed me pictures of her adopted children (who were by then in high school) in the Philippines. One Filipina domestic worker in Italy showed me photographs of her now adult child, who was supported by her and another sister and raised by a third. One Indonesian mother was suspected of selling her baby to a former domestic worker in Java and then retrieved him when her brother and his wife offered to raise him as their own (Constable 2014: 146-8). Such children, whether adopted domestically or internationally, may be absent from the perspective of the migrant worker who has adopted him or from that of the birth mother. Such familial patterns clearly depart from the (hetero)normative two-parent family with left-behind children.

Death, suicide, and 'thrown away' babies
I heard stories, from pregnant woman themselves (usually months or years afterwards) about attempts to abort or 'throw away' pregnancies. In several such cases, the attempt to terminate the pregnancy was unsuccessful, and the mother would interpret the failed abortion as a sign that she should keep the child (or that the child wanted to be with her). I also heard first-hand stories and rumours about successful abortions, where to obtain them illegally, how much they cost, and what sorts of drugs one could obtain to induce miscarriage. Gossip churned when the mainstream media reported on domestic workers giving birth to and abandoning newborn babies or disposing of foetuses. Stories of panicked workers who hid their pregnancies or responded to them with confusion and shock and hid or disposed of a baby are not unheard of. Recent newspaper reports tell of domestic workers allegedly storing or disposing of a newborn baby or a miscarried foetus, echoing familiar themes (South China Morning Post 2015a, 2016). Such stories circulate as precautionary tales among migrant workers (and employers) in Hong Kong. It is legal for a domestic worker to be pregnant, but many workers and employers do not know that, and if they do they still consider it wrong or immoral for a domestic worker to get pregnant, because she is there to work.
An unusual but revealing story discussed and retold among migrant workers involved the suicide of a former domestic worker's daughter in Hong Kong in 2015. Newspapers reported that the 15-year-old girl -one of two daughters born to a Filipina who had come to Hong Kong as a domestic worker and a British insurance executive who worked for an elite Hong Kong firm -had jumped to her death from an upper floor of her family's luxurious Hong Kong apartment in Repulse Bay. In the aftermath of her suicide, it became clear that the two girls were 'paperless' and had no official status in Hong Kong. They were not registered at birth. They had not attended school but received private tutoring and presumably private health care as well. Their mother had overstayed her domestic worker visa by about 20 years, and the couple had not married (South China Morning Post 2015b-e).
One can productively compare the Repulse Bay family with Indah's Hong Kong family (in the first vignette above). The father in the Repulse Bay case was a welleducated and well-employed British permanent resident of Hong Kong; Tika and Nina's father was a low-paid, poorly educated South Asian irregular migrant worker. Both mothers had come to Hong Kong as domestic workers and overstayed their visas. Both passed as wives, despite not being legally married. Their living conditions -a luxury apartment in Repulse Bay versus an illegal New Territories village structurecould hardly be more different.
Indah was always watching out for immigration officers who patrolled their area of the New Territories, which was home to many migrants and minorities. Like many overstayers, Indah felt vulnerable at the market, on public transport and near her home. When I first met her, shortly before she gave herself up, she had become too frightened to go to the market or to allow her children to play outdoors.
According to news reports, the Repulse Bay family attended church; the children had private tutors and appeared physically healthy. Indah's daughter Tika was not in school, and the children had not received medical care since birth. Indah and her partner could not afford healthcare or education for their children, which contributed to Indah's eventual decision to give herself up to the immigration authorities and leave Hong Kong.
In the Repulse Bay case, money, and the lifestyle it afforded, along with the white professional identity of the father, reinforced an impression of citizenship or belonging. Those who knew the Repulse Bay family from work, church or the community, assumed that they were a 'normal' family -a married couple with children and legal residence. According to news articles, they were regularly seen together, which suggests they were not worried (as Indah and her family were) about being detained and asked for identity cards by immigration officers when they went out. The mother did not attend her partner's work-related social events, but no one considered that particularly odd or surprising.
The Repulse Bay family was the subject of much gossip and speculation after the newspaper reported the daughter's suicide. Members of the migrant worker community found it surprising that a family with significant financial resources had apparently failed to seek legal help to rectify the mother and children's legal status. Less-privileged migrants (some of whom had overstayed with their children) found it shocking that, despite the father's wealth, connections and legal residency (to which his children were entitled), his partner and children should remain undocumented. Some speculated that the couple were trying to keep the family together and were probably afraid that registering the children's births would reveal that the mother had overstayed, thus risking her imprisonment or deportation from Hong Kong. Others speculated that the couple could not marry because she had already married in the Philippines. Yet, such an obstacle is not insurmountable. Although the Philippines neither permits nor recognizes divorce, I knew domestic workers who had obtained divorces in Hong Kong, thus allowing them to remarry overseas (Constable 2003). One woman said that the man should at least have hired her as a domestic worker to maintain her legal standing in Hong Kong. Two people speculated that the woman was 'too proud', or that the man had 'control issues' and wanted to deny legal status to his partner and children so that they could not easily leave him.

Global assemblages
A global assemblage approach to the institutions and expert knowledge that shape the experiences and practices of migrant mothers, migratory families and the spectrum of absent children, requires shifting the immediate focus away from the details of the ethnographic vignettes above (which might be conceived of in terms of cultural and society-bound categories such as 'family' or 'migrant workers in Hong Kong') and focusing instead on the global patterns, networks and knowledge systems that flow over these spaces and shape migration -especially women's labour migration -and the varieties of family formations, including possibilities for the presence or absence of parents, the spectrum of absent children and the making and unmaking of different sorts of families across time and space.
Global assemblages are the systems of governance and regimes of ethics or values that structure practice and that, according to Ong and Collier (2005), offer new ways of framing spaces of enquiry. Some aspects of assemblages, alluded to above, include knowledge systems that promote or oppose temporary migration; policies on citizenship and adoption in sending and receiving states; the role of NGOs and social workers whose expert knowledge allows them to counsel pregnant migrant workers authoritatively about the 'benefits' of keeping or giving their children for adoption; and the legal parameters that allow migrant workers to obtain divorces in Hong Kong that are not recognized in the Philippines but might be in Indonesia. Collier (2010: 400) defines global assemblages as the configurations 'through which global forms of techno-science, economic rationalism, and other expert systems gain significance'. Global assemblage is 'a tool for the production of global knowledge … that strives to replace space, culture, and society-bound categories that have dominated social sciences' (Collier 2010: 400). Drawing on the work of Cohen (2005) and Scheper-Hughes (2005), Collier explains that human organs, for example, have become an 'increasingly global form' as they are extracted, tested, moved, decontextualized and recontextualized. These assemblages involve 'organs, networks of brokers and dealers, donors and recipients, sellers and buyers who interact in various moral and money economies, and through various forms of technical and political regulation' (Collier 2010: 400). The 'relationship among the elements in an assemblage is not stable' or 'reducible to a single logic' (Collier 2010: 400). Instead, it is characterized by debate, instability and conflict. The point of the global-assemblages approach is not to 'relativize cultural analyses' but to 'clarify moral and ethical positions, resistances, and possibilities that emerge around such processes' (Collier 2010: 400).
If the global assemblage of organ donation involves, as Collier (2010: 400) argues, movement of 'organs, networks of brokers and dealers, donors and recipients, [and] sellers and buyers who interact in various moral and money economies, and through various forms of technical and political regulation', then, in the vignettes above, we might ask what elements of the assemblage regulate who circulates to and/or from where, and how? I begin first with the elements of assemblages that facilitate the circulation of migrant domestic workers, since migratory circulations of workers create conditions for the circulation or absence of children. Systems of recruitment and brokerage in Indonesia (as in the Philippines and elsewhere) and partner agencies in Hong Kong and other receiving countries (like Hong Kong, Singapore, Taiwan and Malaysia) in large part shaped the movements of Indah, Mia and others.
Workers, like organs, circulate through both legal and illicit channels. Brokers, including agents and subagents (Killias 2010;Xiang and Lindquist 2014), might follow or circumvent legal regulations pertaining to recruitment. Strictly speaking, workers are not sold like organs, but the marketing and sale of their labour power is not entirely different from that of the sale of organs. Domestic workers are commodified and objectified, and, in some cases, traded and trafficked like organs (Constable 2007).
Legal and human rights discourses and debates increasingly surround the topic of domestic workers (Constable 2015). In the past decade, the rhetoric of trafficking and human rights has increasingly replaced non-governmental labour rights discourses.
NGOs and various activist and advocate organizations have proliferated in Hong Kong, and have established international networks in many sending and receiving countries. NGOs and activists promote global discourses about domestic worker rights, resulting in, for example, Convention No. 189 on Domestic Workers Rights (International Labour Office 2011).
Technological practices and political regulations shift as immigration technology becomes increasingly accurate at regulating the exit, entry and re-entry of workers. Biometrics make it more difficult than in earlier decades for workers -including migrant mothers -to re-enter territories with new names and passports, or on tourist visas, after overstaying, so recruiters promote access to new locations with new names and identities.
Political regulations also shift in response to local sentiments and media reports of death, violence, or deleterious working conditions abroad. Sending countries impose new regulations on receiving countries regarding minimum wages or mandatory days off for their citizens working abroad. The market and conditions of labour in Hong Kong (often considered the best in Asia for domestic workers) can influence a market elsewhere, such as Singapore. 6 Patterns of migration are shaped, moreover, by sending country bans on migration -and rumoured bans, like the Indonesian ban on all Indonesian domestic worker jobs by 2017. Accurate or not, it has encouraged agencies to seek new markets of workers.
Multiple intersecting border-crossing structures facilitate the mobility of some and the immobility of others. Numbers of workers -and gendered patterns -have shifted over recent decades. Construction worker, nurse, domestic worker, and factory worker are some of the occupations filled by migrant workers from Southeast Asia in countries spanning from Canada to Korea, Saudi Arabia to Singapore. Economic demands or downturns shape each pattern of training and recruitment, movement and immobility. The massive growth of medical training colleges for nurses in the Philippines in the early 2000s, for example, followed by the collapse of the US market for immigrant nurses and the subsequent closing of many such colleges, is but one illustration of how market fluctuations shape mobility (Ortiga 2014).
Scholars point to the intersections of migrations. They focus on the sending and receiving locales and increasingly follow migrant workers across a variety of local and international locations, not just back and forth, but 'on and on' (Liebelt 2008), sometimes 'stepwise' to better locations (Paul 2011). A global assemblage approach highlights, as Xiang and Lindquist (2014) and Palmer (2012) do, the heterogeneity of Indonesian state and local positions and processes and the fractures in interpretations, viewpoints and opinions at the level of policy makers and bureaucrats in relation to global and international discourses about gendered migration and trafficking.
Local and international laws, regulations and practices also shape the mobility or immobility of children. Hong Kong (and most other migrant worker destinations) prohibit temporary 'low skilled' migrant workers from bringing children or other family members with them. Adoption may take place locally or may be part of an increasingly global market. International adoption reflects border-crossing patterns that follow or echo political relations and postcolonial relations (China, Russia, South Korea and El Salvador are source countries of adoptees in the USA that have shifted in popularity). Adoption also echoes other flows of people and commodities across borders. These patterns (which involve complex elements of assemblages, including expertise, evaluations, information technologies, networks, brokers, and social workers) raise questions of ethics and fuel debates about profits, human rights and trafficking, especially in relation to poor mothers and purported expert knowledge about children's health and well-being (Constable 2016).
Echoing conflicting views on the right to sell an organ across national borders to wealthier people in wealthier nations, source countries' governments likewise may object to the adoption of children born to their citizens locally or abroad. The Indonesian government and its consular officials in Hong Kong do not endorse the adoption of Indonesian migrant workers' children in Hong Kong, but it happens anyway (as in the cases of Mia, Ree and Elle) because Hong Kong law allows it. According to Hong Kong policies, the biological mother's legal husband, even if he is not the father, must give permission for the child's adoption unless the mother proves he is not the father. Social workers, lawyers, anthropologists and others also contribute -intentionally or not -to the shape of interventions and approaches to these issues.
Citizenship regimes are of vital importance. Children born in Hong Kong do not have the right of abode ('citizenship' rights) unless there is proof that one of the parents is a Hong Kong permanent resident. None of the fathers of the children in the earlier vignettes were Hong Kong legal residents. In addition, is it impossible for a temporary migrant worker to gain permanent residence except through marriage to a local. The possibility of marriage, in turn, fuels cross-cutting legal assemblages, for migrant workers who seek divorce must locate their estranged spouses abroad. A husband abroad has the right to claim his wife's child, even if he is not the biological father (Constable 2014: 149). There are also questions about the state to which the child belongs. If migrant mothers submit asylum claims through the UN refugee or torture conventions with a view to remaining in Hong Kong with their children, as many do to avoid the stigma of single parenthood at home, the assemblages stretch even further. Once a parent files such a claim, she must relinquish her passport, and she and her child lose the rights or protections of citizens accorded by her home country. Moreover, she has no right to work in Hong Kong and no right to settle there in the future on the rare (or non-existent) chance that the state upholds her claim, since Hong Kong only supports third-country resettlement of refugees.
These elements of global migratory assemblages begin to reveal some of the contemporary global configurations that shape the current scene but that are also in flux. They involve (and shape) legal and illicit movements of people, debates about rights and values pertaining to labour and adoption, and contested meanings of what constitutes family, what the obligations of the mother (or biological father or husband) are, what is 'best' for the child, and who has the say over whether a child can or should be adopted. Migratory and adoptive assemblages avoid a unified or static picture of global homogeneity or of simple local-global binaries.
An assemblage approach, in aiming to avoid conventional bounded cultural and social categories, also, I would argue, allows us to re-examine, recapture and criticize conventional social concepts such as the family. Deleuze (1992), perhaps too hastily, claimed that social institutions like the family are in crisis and on their way out because they are no longer conceived of as enclosed social spaces. Instead, global assemblages of 'family' -by which I include circulating discourses about them in their varied and multiple forms, which include the mobility and movement of parents and children and their social, lived experiences -are far from disappearing. Rather, the fluid forms and transformations of migrants' families across time and space and across borders (absent and present children, through images and disembodied voices and text, through estrangement, and through legal and financial claims and anticipated reunions) are part and parcel of the networks and flows of global assemblages. The boundedness of family formations and their normativity are clearly contested. Ideas about families -which were never, in fact, fixed or static, even before the current era of migration -are enriched by sensitivity to their various forms and the patterns of assemblages through which they are shaped and assigned different values and meanings.

Conclusion: queering migrant families and the spectrum of absent children
Asked in an interview about the mid-range approach of global assemblages, Aihwa Ong (quoted in Sinha 2010: 93-4) said: My own approach is to look at practices, i.e. the focus is not on 'people' but social and institutional practices that are largely observable in the public realm. … You 'stay close to practice' i.e., abstract your claims from observable practices that seem constitutive of emerging situations. By following practices (individual and collective, informal and formal, scripted and spontaneous) we avoid intruding too much into people's lives (or respect the limits they imposed on our observations).
Missing from such assemblages are the face-to-face interactions and stories that reveal people's raw emotions and visceral pain. A focus on people, paired with a concern for assemblages (both parts of this article together instead of one or the other part alone), may combine the best of both approaches for anthropologists while revealing some of the worst practices of our world. The global assemblage, with many fluid elements of the 'society of control' (Deleuze 1992) is valuable, but one can productively shift the focus back and forth, between the wider assemblage (the middle view) and the close-up view of lived realities in social spaces. We cannot fully understand leaving children behind, having them adopted, or abandoning them overseas only in terms of assemblages or affect. Instead, we must understand them in terms of the practices and choices that assemblages shape. Assemblages of migration and citizenship have an impact on fluid family formations and the spectrum of absent children. A recent communication from Indah painfully reminded me of this. She told me that she was considering selling a kidney in Singapore, something a fellow migrant worker had told her about, and that she thought it might allow her to return home sooner to reunite with her children.
As I have argued, the academic and scholarly stories we tell, or the data we report, are not outside the global assemblage that circulates as expert knowledge. They can reshape, reinforce, or contribute to new hegemonic assumptions of normativity and to ideas about the well-being of children belonging to certain sorts of family formations, such as two-parent, heteronormative, migrant, or non-migrant families. There is a wide range of migratory family forms -some of which reflect more distant patterns of the past (see Geertz 1961). Yet, as the vignettes in this article illustrate, these forms change through time, sometimes appearing normative and at other times offering new shapes and opportunities for formations of queer families of various sorts -single mothers, gay or lesbian parents, and newly formed families based on adoptive children, some of whom have no birth certificate and no citizenship, or have died or disappeared. Assemblages and affect thus make and unmake migratory families and absent children.