‘E Kuchú Ta Korta Mi’ Versus ‘I Cut Myself with a Knife’: Cultural Assumptions in Child Protection Intervention Processes

I conducted a longitudinal ethnographic PhD study between June 2009 and June 2012 to examine how intervention practices took place in 30 Dutch and Dutch-Curaçaoan single-mother families with multiple problems, because interventions were often regarded as ineffective in 'multi-problem families'. Drawing on this study, this article aims to demonstrate that cultural assumptions, along lines of, among other things, gender, ethnicity and socio-economic class, play a major role in family care and child protection intervention processes in the Netherlands. I will provide different examples of both social work text and talk to illustrate underlying categorical presumptions. It will be seen how the culturally embedded social policy category ‘multi-problem family’ plays a role in Dutch social work policy and practice. I will then demonstrate how important cultural notions are (made) in social work encounters between the mothers and the institutional representatives in question through giving an insight into how cultural beliefs are mobilized in encounters. I argue that one should be more aware of how (essentialist) cultural notions are mobilized in child welfare and protection intervention processes in order to bridge the gap between the life worlds of the mothers and the professionals, as disconnection appears to lie at the core of why family care and child protection intervention practices are often flawed.


Introduction
There is an official state policy on so-called 'multi-problem families' in the Netherlands. 2 'Multi-problem families' are often underprivileged families who are subjected to preventive and protective intervention practices to prevent their children from developing emotional and behavioural problems, delinquent behaviour, or child abuse and neglect. 3 Normative judgments about ethnic, cultural or socio-economic class status being deficient are, in this regard, motives for intervention. 4 So-called 'multi-problem families' either enter the legal child welfare and protection system directly, through referral, 5 or indirectly, if voluntary interventions are inadequate or no longer adequate. 6 More than 75 per cent of child welfare and protection measures are undertaken with single-mother families 7 while the number of these measures also increased between 2002 and 2011 by 151% to 51,326 8 and are expected to rise even further. 9 Since so many single-mother families are involved in the child welfare and protection system while intervention practices take place away from the public gaze, it is in my view crucial to examine the processes of child welfare and protection interventions in single-mother families in practice. Specifically, this article Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 is concerned with how Dutch and Dutch-Curaçaoan single-mother families with multiple problems are 'classified', or to borrow Agar's word, 10 'diagnosed' by Dutch policymakers and social work professionals and how cultural assumptions, along lines of e.g. ethnicity, socio-economic class, gender and age play a role in the child welfare and protection intervention processes of these families.

Aim and relevance of the study
The results presented in this article form part of a larger explorative ethnographic study on so-called 'multiproblem families' in the Netherlands. In 2009, a project consortium consisting of five stakeholders, the municipalities of Amsterdam, Dordrecht and Rotterdam, the Nicis Institute (now Platform 31) and Utrecht University, formulated a collaborative research proposal whereby the ins and outs of 'multi-problem families' in the Netherlands and Curaçao would be studied on account of the lack of knowledge about these families and, related to this, the supposed ineffectiveness of family care and child protection interventions in these families. A longitudinal ethnographic research design on fifteen Dutch and fifteen Dutch-Curaçaoan singlemother 'multi-problem families' was chosen in order to compare the effect of interventions undertaken with these families. Since family care and child protection interventions remain largely invisible from the public, I chose to analyse from the inside (emic) and outside (etic) 11 how intervention practices are implemented in single-mother families and the effect they might have on these families. This article addresses the research question as to how cultural categories, or notions, are mobilized in social work encounters between single-mother families and the social work professionals involved with the families. It aims to reveal how (essentialist) cultural notions are mobilized in family care and child protection intervention processes.
A critical discourse analysis of the encounters between the mothers and the professionals is crucial to understand how interventions are carried out in single-mother families, as an intervention practice is mainly communicative work. 12 Hence, the longitudinal approach that I adopted combines an ethnographic study of families who were subjected to interventions -particularly child protection interventions, including: a supervision order (ondertoezichtstelling), an extended supervision order (verlenging van ondertoezichtstelling), in-care placement (uithuisplaatsing) and the abrogation of parental authority (ontheffing of ontzetting uit de ouderlijke macht) -with a critical discourse analysis of social work encounters. This approach is relevant because it 'systemically' reveals important patterns and processes which underlie certain categorical assumptions of both family care and child protection practices (ethnography) and interactions (critical discourse analysis). This dimension is generally overlooked; an ethnographic inquiry is especially suitable for revealing intervention processes in families. 13 With regards to the aim and limitations of the study at hand, it must be noted that the study will not delve into the question whether it was, or is, the 'right' decision to subject the single-mother families to compulsory state interventions. Accordingly, the article does not aim to discuss whether the intentions of the state representatives involved were 'right' or 'wrong', nor 'effective' or 'ineffective'.
Furthermore, although generalizations cannot be drawn from the small sample of case studies, through an insight into the lives of the single-mother families we may be able to better understand how other (single-mother) families who are subjected to interventions may perceive and experience these intervention processes in the Netherlands.

Access to families
The aim of the study was to capture the perspectives and experiences regarding the intervention processes from the families' point of view. 14 In this respect, it was beneficial that I got access to the families and the intervention practices undertaken with them via 'gatekeepers', i.e. lawyers and doctors. These gatekeepers had a professional relationship with the families based on trust, but were, as opposed to the single-mother families and social service agencies, not the object of scrutiny.
I am aware that the way of entering the field influences the research design, the position in the field, the production of knowledge and data collection. 15 The families were the 'principals' 16 of the study (i.e. those for whom the research is likely to be consequential) and the social work professionals were 'spear carriers' (i.e. those who are tangential) because my prime responsibility was towards the single-mother families involved and not towards state officials who performed their jobs. In this respect, my aim was not to adopt a distant role in order to (re)produce so-called 'objective' knowledge. Rather, I adopted a critically engaged role and maintained multiple identities to navigate a field of power asymmetries whereby I aimed to sustain primarily 'good' relationships with the families and also with the social work professionals in question. 17 The procedure of recruiting 'multi-problem families' went as follows. Earlier, as part of my broad mapping study on 'multi-problem families', I had already interviewed different stakeholders and experts, including the doctors and lawyers who later became the gatekeepers of the study. Since I maintained a professional relationship with the doctors and lawyers, I asked them if they would be willing to approach women who might be single mothers, of either Dutch or Dutch-Curaçaoan origin, experiencing multiple problems. They were willing to help me and asked the women to whom they deliver services if they would be willing to participate in a study on multiple problems experienced by single-mother families. Whenever a woman signalled that she would be interested in participating, I phoned her to further explain the study and to make an appointment with her. The selection of the 30 women was based, besides being a single-mother family of either Dutch or Dutch-Curaçaoan origin, on the woman's self-definition of having 'multiple' problems and her willingness to talk about her life to a researcher on one or maybe more occasions. 18 When I visited the families at home, I explained the goal of the study (again) as a way of confirming the expectations of the study. I made clear that the study was based on the notion of voluntary participation and that my aim was to write a book about their lives, and the circumstances in which they, as a singlemother family experiencing multiple problems, live. I would manage their information confidentially, their identities would be concealed, and data that would disclose their identities would be either removed or changed. I also told them that whenever they (or their children) no longer wanted to participate in the study (regardless of the reason), I would stop studying the family and that this would be alright. It was not easy but it was honest to tell them that I was unable to change their lives -I did not want to make empty promises and I wanted to prevent them and myself from experiencing disappointments in the future. In this regard, I told them that I was a PhD student who was interested in their stories, experiences and lives, and not someone with formal authority or influence. I tried to equalize our asymmetric relationship by avoiding my researcher's role having connotations of 'colonialism' or a 'patronized' stance since this would, in my view, raise ethical questions about studying families with multiple problems in the first place. 19 Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 I succeeded in 'recruiting' fifteen Dutch and fifteen Dutch-Curaçaoan single-mother families with multiple problems to conduct an explorative ethnographic study on their lives by means of careful self-presentation, patience and time. 20 It was in my view beneficial that I was not affiliated with the social services in conducting the ethnographic study as I aimed to sustain enduring relationships with the single-mother families. In this respect, issues of mistrust often play a major role in relationships between families and social services. 21 'Real' access took considerable time and effort because I aimed to be 'culturally accepted' by both the gatekeepers and the single-mother families. The fact that I went to Curaçao on two occasions for a one-month period and learned some basic Papiamentu was advantageous in bridging the 'psychological distance' 22 between the Dutch-Curaçaoan mothers and myself. Being a woman of (more or less) the same age was also beneficial to developing a relationship, because I could easily relate to them. This also meant that, certainly in the beginning, I was careful not to 'look down upon' or to 'judge' their way of living. 23

Addressing so-called 'multi-problem families' in practice
I prefer to use the terms 'family with multiple problems' or 'family experiencing multiple problems' instead of the adjectival term 'multi-problem family'. The former terms are used throughout the ethnographic study in dialogue with the single-families for different reasons.
First, the term 'multi-problem family' has been criticized in the international literature by the argument that it is stigmatizes and reifies 'multi-problem families' 24 thereby doing more harm than good to the families as well as to the researcher-participants relationship and, in my view, the social work professionals-single mothers relationship as I will demonstrate below. In addition, Tausendfreund et al. and Asen 25 note that the label 'multi-problem family' may be harmful to families, as the label does not take into account that the intervention and care system are partly responsible for the families' problems, but attributes problems and risks unilaterally to families.
The second reason is that there is no uniform definition of 'multi-problem family' as will be seen below 26 while the notions of 'a family with multiple problems' or 'a family experiencing multiple problems' simply refer to a family that has or experiences multiple problems, i.e. viewing families as being in relation to problems instead of as inherently deficit. Another reason is that the term appears to be embedded in cultural and institutional customized practices, and imbued with social norms and values, that differ dramatically from, for example, the Curaçaoan context. As opposed to the Netherlands, social workers and policymakers in Curaçao (and also, for instance, in Germany) do not use the term 'multi-problem family' because they are not familiar with the term 27 while the phrase 'families experiencing multiple problems' is recognized. 28 Utrecht Law Review | Volume 12 | Issue 2 (June) 2016

Data collection on the term 'multi-problem family'
I used a triangular approach to study the so-called 'multi-problem family' phenomenon from multiple perspectives. The (inter)national literature was first studied to assess the scientific definition of the term 'multi-problem family'. 29 Second, national policy documents and those of the municipalities of Amsterdam, Dordrecht and Rotterdam were scrutinized to evaluate how a multi-problem family is defined and how multi-problem family policy is formulated. The fifteen policymakers involved were interviewed to verify whether the local policies concerning interventions in multi-problem families corresponded to the consulted documents.
Third, I examined the perspectives and experiences of 40 social workers working with multi-problem families in a voluntary framework, as they -as opposed to family supervisors working with families in a legal framework -must define their target group, namely, the so-called 'multi-problem families' for which their services are intended. I conducted semi-structured and open face-to-face interviews about how the social work professionals characterize multi-problem families and how they, subsequently, target multi-problem families for 'multi-problem family' interventions. 30 I will discuss the policymakers' and the social workers' categorizations of multi-problem families in so far as they make reference to cultural or other categorical assumptions, as this article is particularly concerned with cultural assumptions in intervention processes. 31 Fourth, a major part of the ethnographic study consisted of gathering field data through the informal interview with the mother, her narratives of her life history and the participant observation of the families (atmosphere, relationship, interaction) over a longer period of time and in different family settings (I participated in their daily lives and observed social work encounters). In addition, I relied on field notes of the observations. The interviews and narratives were either recorded with an audio-recording device (with the mothers' consent) and/or the content was written down in a notebook.
Fifth, since from the dataset of thirty families, seventeen families were subjected to compulsory child protection orders, I aimed at obtaining access to more private sources of data on these interventions. I made notes or hard copies of various documents such as their case files, diaries, bank statements, newspapers, letters and other test results in order to analyse, and cross-check, data.
Sixth, through progressive focusing, 32 it has appeared that the role and function of diagnosis 33 (I will elaborate on this in Section 4 below) plays a major role in the families' intervention processes. Since the encounters between the single mothers and the institutional representatives became an additional object of study, I tried to capture 'what was said' on tape during home visits because of methodological and analytical advantages. The use of an audio-recording device 34 enabled me to concentrate more on observing (nonverbal) communication and the setting, 35 which was analytically crucial because I could carefully re-listen to the conversations to hear what was said and thereby to assess which topics were essentially important to the further process of child protection interventions and to allow for critical discourse analysis.

Intervention practices and culture
Intervention practices are embedded in cultural and historical contexts, and since human beings are embedded in cultural systems, they both shape and are shaped by their cultural realities. For the most part, they act only within the cultural realities of their lives. The cultural norms, values, symbols and traditions an individual grows up with become 'natural' through enculturation and socialization processes. Yet while certain (dominant) cultural norms seem natural to people who are part of that cultural reality, these patterns are neither sustainable nor intrinsic manifestations of human nature, albeit reshaping the way people act and understand the world is not easy. Culture is not a distinct system functioning alongside political, societal, judicial or economic systems, but comprises beliefs, customs, values, norms, traditions, institutions and symbols. These elements form the overarching frames that shape how humans perceive reality. Cultures, as broader systems, or as a 'matrix' 36 do not only emerge from the complex interactions of various elements of social behaviours but also form people's attitudes and behaviour. 37 Within the discipline of cultural anthropology the view on 'culture' has changed from an essentialist perception -i.e. culture has its own essence and is something people can possess or own -to a social constructivist belief that people make or create culture, and produce cultural realities. Culture is, according to this view, situationally accomplished and fluid, as people stress different cultural beliefs in different contexts. This means that people may use or mobilize 'culture' instrumentally if they think they might benefit from certain (i.e. essentialist notions of) cultural beliefs if, for instance, language, land or other rights are at stake. 38 To borrow Wikan's words: 'Culture is not a thing, a material object; it is just a concept, an abstraction. But because many people clearly think of it as a [monolithic] thing, "culture" becomes amenable for use in defence of all kinds of special interests. (…) We are deeply affected by our social environment and the models of and for life that we have acquired in the course of living (...). But cultural models are not sui generis. They derive their force or power -their impetus -from their ability to give direction to, and mobilize, human energy.' 39

Institutional discourse
In this respect, Agar's institutional discourse theory 40 is relevant as a 'cultural model' because institutions play a central role in the production and maintenance of the social or cultural order through, for instance, the distribution of services to citizens. Institutional discourse has, according to Agar, three functions. The first function is diagnosis: an institutional representative must diagnose the client (i.e. single-mother families) in order to be able to fulfil institutional tasks or agendas. Diagnosis refers to the 'problem-definition aspect' of institutional discourse because an institutional representative must define who the client is. 41 It thus constitutes the institutional reality for further action.
A diagnostic label is a form of categorization that is constructed on the identification of specific 'signs and symptoms' 42 and expresses and relies on indexicality. 43 Although the single-mother families may resist the diagnosis -as diagnoses are often deficit oriented, along lines of culturally dominant 'middle class ideals' 44institutional representatives (social work professionals, family supervisors, the juvenile court judge, police officers and other state representatives) have the authority or power to generate consequences. They make the identification matter regardless of whether the family agrees or disagrees. 45 Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 The second function of institutional discourse, and one of the goals of diagnosis, is to provide directives. The directives can be seen as the 'problem-solution aspect' of institutional discourse. 46 Institutional representatives direct the single mother and her children to do or not do certain things and/or direct other colleagues or institutions to do certain things to, or for, the single mother and/or her children. 47 This means that services are distributed according to top-down directives through the dominant, socially constructed ideas of diagnosis. For instance, through categories of what the culturally hegemonic middle class considers to be normal or abnormal/deviant. In a similar vein, social work and child protection settings are embedded with various normative socio-cultural categorizations, which are related to e.g. ethnicity, class, gender, age, neighbourhoods, (mental) health, educational status and parenthood. 48 Mäkitalo and Säljö argue that: 'Invoking a particular category is a mode of reproducing a specific type of interactional pattern and moral order.' 49 (emphasis added) The labels 'child neglect', 'domestic violence', or 'multi-problem family' lead to different actions from the label 'promising family'. 50 Hence, the directives determine the actions taken as well as how the life of the single-mother family is shaped by the endeavour. The role of language plays a vital part in this process, as it is a powerful instrument to construct certain 'realities'. It may therefore, in the institutional process, depict the 'reality' of the single mother or other state representatives as being invisible and/or illegitimate. 51 Institutional language 52 is not a neutral means but political and furthermore embedded in an asymmetric setting. 53 Hence, the most powerful and authoritative interactant, the institutional representative, primarily produces the most powerful language, which, through certain questions or topics (or forms of questioning) controlled by the institutional representative, comes together in the report. 54 The third function of institutional discourse is the report in which the institutional representative establishes the 'reality'. A report is another sense-making device that may hold the institutional representative accountable for decisions made or not made, advice given or not given, and actions taken or not taken. 55 As a matter of fact 'a report is the summary of the institutional discourse that the institutional representative [in dialogue with the single-mother family] produces' 56 . Agar states that the single mother (i.e. 'client') may not be present when the institutional representative generates a report. A report may be in written or oral form. Additionally, the report may be directed only towards other institutional representatives and not towards the single-mother family thereby remaining hidden from the single-mother family. 'The institutional frames prescribe how a report should look and what it should contain'. 57 This means that the professional discretion of institutional representatives is constrained, whilst the report is crucial because the progress of the diagnosis is measured along these frames. This indicates that a report is an example of institutional representative control of the outcome of institutional discourse. 58  Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 In sum, culturally embedded institutional discourse is very powerful and intertwined in intervention practices. These practices are mainly transmitted by, and utilized and produced in, the face-to-face interaction between the institutional representatives and the single mothers and the written reports that the institutional representatives generate. Of importance here is that cultural notions, along lines of certain beliefs of gender, age, class, religion and ethnicity with regards to, for instance, what is considered 'good' and 'bad' behaviour do not only shape institutional discourse (and the three functions), but are also shaped in dialogue and furthermore produced by the encounters between families and the social work professionals. I aim to demonstrate that a 'battle' of moral judgments 60 and perspectives appears to lie at the core of the production of intervention practices.

Formulation of 'multi-problem family' policy
Based upon the Youth Care Act (Jeugdwet), 61 Dutch policymakers formulate social welfare policy at the national level, which is for instance established as 'multi-problem family' policy in the Netherlands. 62 Policymakers determine 'multi-problem family' policy at the local level of policy as well. Policymakers must distinguish between 'normal' and 'deviant' families in order to categorize certain families as 'multi-problem families'. In this way they outline the directives for social work professionals to implement 'multi-problem family' policy.
My analysis of policy documents on 'multi-problem families' in the Netherlands, including national and local policy documents, indicates that the policy-oriented definitions of a 'multi-problem family' are ambiguous and unaligned. The national policy document of the Ministry of the Interior and Kingdom Relations on multi-problem families, which is called Achter de voordeur aanpak/Multi-probleem gezinnen [Behind the front door approach/multi-problem families], does not define the term 'multi-problem families'. Instead, the term refers to a variety of family types such as 'vulnerable families', 'households in disadvantaged neighbourhoods', 'families at risk', 'complex and persistent nuisance families', 'Roma families with multiple problems', 'families under supervision', 'families with chronic multiple problems and/or a high risk of relapse', 'low IQ families or with psychiatric problems', etcetera. 63 Based upon this we may argue that different cultural notions, along lines of e.g. ethnicity (Roma), socio-economic class (disadvantaged neighbourhoods), intelligence (IQ), underlie the characterization of a 'multi-problem family'. 64 Furthermore, my analysis of local policy documents shows that different municipalities place 'multiproblem family' policy under different departments. Consequently, a municipality shapes not only 'multiproblem family' policy, but also the view of so-called multi-problem families. It seems that each municipality uses different definitions and includes different variables along the lines of risks, family composition and/ or ethnicity. This explains, among other things, why the number of registered multi-problem families varies and the selection processes differ (the estimated number of multi-problem families in Amsterdam is 2,000 and in Rotterdam 30,000, while the total number of residents in Amsterdam is higher). 65 Additionally, there seems to be a gap between the way of defining and selecting multi-problem families. In Rotterdam, for instance, families are not selected on the basis of 'a specific signal', but are actively approached to participate. Via the 'basic administration' of the municipality the families in 'disadvantaged neighbourhoods' with children under the age of four are selected and visited by social workers on basis of 60 It must be noted that many formulations of all people draw on both moral and factual constructions (discriminating between good and bad and correct and incorrect), as they are both rhetorical devices. 61 The Youth Care Act is the legal framework in the Netherlands that prescribes that high quality care must be available to children and their parents, and to strengthen the (at risk or problematic) position of children and their parents in the Netherlands. A family supervisor from the Youth Care Agency (Bureau Jeugdzorg) (or another mandated agency in the private sector) will execute the order mandated by the juvenile court judge (on the advice of the Child Protection Board (Raad voor de Kinderbescherming)). 62 Ministry of the Interior and Kingdom Relations 2012, supra note 2. 63 Ibid., pp 1-24. 64 It must be noted that deficit characterizations are not necessarily attributed a definition from the category 'multi-problem family' as they can stand on their own. 65 See Verhallen 2015, supra note 7. the definition of 'multi-problem family'. 66 This means that in Rotterdam families are not only targeted on the basis of demographic criteria but also actively approached by social workers. Furthermore, there was also a specific policy on intensive guidance for Dutch-Curaçaoan so-called 'multi-problem families' (this policy was called IBAG) which included mandatory participation by the families in question. 67 A consequence of the fact that various types of families are earmarked, targeted and approached as multi-problem families is that family interventions do not obviously correspond with every type of (socalled multi-problem) family. Whereas Amsterdam has never explicitly focused on Dutch-Curaçaoan multiproblem families, Rotterdam and Dordrecht have done so. More specifically, Dordrecht focuses on Dutch-Curaçaoan high-risk youth and not on families.
We may thus conclude that ethnicity and even vulnerability or risks play a role in the recruitment of 'multi-problem families'. This means that the label 'multi-problem family' has a preventive family diagnostic function; families with (as yet) no problems are diagnosed as multi-problem families and are targeted for multi-problem interventions. The fact that certain marginalized groups -such as single-mother families whose children are considered to be at a higher risk of developing delinquent behaviour 68 and who are particularly vulnerable to preventive state interventions 69 -are selected on the basis of demographic criteria corresponds with other studies conducted in e.g. Norway, Sweden and Ireland. 70 Such a preventive policy approach appears to discriminate against certain disadvantaged groups in the Netherlands, revealing racism at the system level. 71 In addition to this, Asen describes the 'racist reactions' of social work practitioners towards certain families. 72

Practice
Through processes of 'bureau interpretation' the social work professionals translate 'multi-problem family' policy into social work practice, and by so doing they decide who is eligible for 'multi-problem family' support. 73 I examined the perspectives of 40 social workers working with 'multi-problem families' in a voluntary framework, as this group of social workers determines which families are eligible for multi-problem family interventions. 74 With them I discussed the ways of defining, selecting and recruiting multi-problem families, which enables them to exercise multi-problem family interventions in these families. However, their ways of ascribing meaning to the analytical concept 'multi-problem family' is to a large extent based on their own professional experiences and social work realities. This seems to indicate that there is no need to strictly follow the 'multi-problem family' definitions as outlined by policymakers.
When I asked the social workers to define a 'multi-problem family' almost all of them informed me that they find it either 'awkward' or 'difficult' to define. 75 They said that they used the term 'multi-problem Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 family' (only) to indicate the families they assist. Only one social worker responded immediately by defining a Dutch multi-problem family: 'A Dutch multi-problem family is a family with combined socio-economic, psycho-social and child-rearing problems. These problems often come together and are often chronic as well. Sometimes they resist, they refuse to work with you, they do not keep appointments, and they deny or downplay the problem. Consequently, they affect the resilience of the social worker.' 76 39 social workers admitted in an interview finding it difficult to determine what a multi-problem family essentially is. Most social workers indicated that they think about a 'good', 'adequate' and 'workable' definition. They found it difficult, however, to decide in practice which families fit their target group, and therefore, to determine who to recruit for 'multi-problem family' interventions. 77 This is partly due to ambiguous multiproblem family policy, their substantial space for professional discretion and the profound implications their actions might have. Based upon the way the social work professionals categorize families, for example as 'normal' or as 'multi-problem', they diagnose certain (multi-problem) families for interventions. 78 Deficit-based diagnosis is important in institutional child welfare and protection argumentation because it is a way to manage risks and serves to secure the legitimacy of intervention in the private sphere of the single-mother families. 79 Furthermore, most social workers found the concept 'multi-problem family' unworkable. In this respect, 8 out of the 40 social workers interviewed asked how I would define a 'multi-problem family'; 'maybe you can help us to formulate a workable definition'. None of the social workers used a definition of a multiproblem family, as formulated by policymakers in multi-problem family policy documents, as a guideline to select families.
Instead of defining 'multi-problem families' analytically, all 40 social workers refer to multi-problem families through the variety of problems that families have. In this regard, frequently mentioned problems are: pedagogic problems, financial problems (debts), poverty, addictions, criminality and delinquency, domestic violence and child neglect. The 'Tokkie' family is seen as a typical Dutch multi-problem family; a socio-economic low class family which engages in anti-social behaviour and dysfunctions on several dimensions. Apart from these problems, social workers seem to perceive certain other problems as being characteristic of Dutch-Curaçaoan multi-problem families which mainly emerge from a lack of knowledge of the Dutch language and the Dutch cultural and institutional context.
The following explicit and implicit normative statements were made by different social workers about Dutch and Dutch-Curaçaoan 'multi-problem families', whereby we can see that a causal relation is made between 'problems' and (presumed) categorical characteristics, such as age, ethnicity, gender, class, motherhood:  Other problems that, according to social workers, characterize Dutch-Curaçaoan multi-problem families refer to cultural differences and are mainly related to pedagogic or child-rearing issues and the incapability of formulating and communicating aid questions: Furthermore, the expression in Papiamentu: 'E kuchú ta korta mi' ('the knife has cut me') is often mentioned to illustrate that Dutch-Curaçaoans passively communicate their problems, which is seen as an undesirable cultural difference, as Dutch-Curaçaoans would not take responsibility for their own mistakes or problems. In contrast, Dutch multi-problem families would be too assertive and verbally aggressive in their communication. This is also seen as an undesirable cultural difference as they would not take responsibility for their own mistakes and problems but would either deny them or ascribe them to other people (such as state representatives). Another way of typifying multi-problem families is by risk factors such as a low educational background, unemployment, trauma, poverty, isolation, a negative view of oneself, belonging to an ethnic or sub-culturally different minority (i.e. lower socio-economic class), a lack of facilities in the neighbourhood, divorce and disabilities. This 'risk-based' method of typifying certain families corresponds to a large extent with the arguments of Dutch policymakers about who is considered 'at risk' in Dutch society. Apart from that, most social workers associate certain risk factors with 'multi-problem families' such as single-mother parenting, poverty, a low IQ, a low self-esteem, and teenage pregnancies -these factors are in particular associated with Dutch-Curaçaoan families. Consistent with Dutch policymakers, ethnicity, poverty, and risks seem to be factors in characterizing multi-problem families.
Summarizing, based on the interviews it seems that a specific policy focus on disadvantaged minority groups, such as on Dutch-Curaçaoan families or Dutch 'Tokkie' families, results in discrimination against them, although it may be unintentional. 80 In a similar vein, Križ and Skivenes argue that social workers intervene too extensively, however not deliberately, in (ethnic) minorities families' lives because of normative cultural and other categorical assumptions. 81 6. The results: the ethnographic study 6.1. Problems or circumstances of the families I recognize that the families' problems (which are sometimes regarded by the mothers as circumstances) play a major role in the possibilities for and the limitations of interventions. Since the complexity of the problems of the single-mother families are discussed elsewhere, 82 I will only briefly describe them here. A result of the ethnographic study is that I identified 43 different 'problems' or 'circumstances' in the 30 families under scrutiny. The most important are: poor financial situation, mental and physical health problems, no steady job or education and with income coming from informal or criminal activities, pedagogic questions or problems, problematic or difficult behaviour of the children (many children are diagnosed as ADHD, ADD or something similar, sexual abuse of the mothers in the past, problems or difficult circumstances with an ex-partner, difficult physical life circumstances and/or living in an insecure neighbourhood, an insufficient understanding of the Dutch language, the low IQ of the mothers (i.e. IQ according to the test in Dutch. The lack of Dutch language skills of particularly the Dutch-Curaçaoan single mothers might well explain the IQ results) 83 and a limited understanding of legal correspondence.
It is remarkable that there are hardly any differences between the 15 Dutch and 15 Dutch-Curaçaoan single-mother families with multiple problems with regards to their circumstances or problems. Differences do occur in the reasons for the absence of the husband within the family, the prevalence of domestic violence and language problems. The 15 Dutch-Curaçaoan mothers who participated in the study were almost all single mothers because the father of their children had left them for another woman. The Dutch mothers were leading the life of a single mother often as a result of domestic violence taking place against them and/or their children. These single mothers were much more often threatened by their ex-partners than the Dutch-Curaçaoan single mothers. The children of the Dutch-Curaçaoan mothers are descended from different fathers more often than the children of the Dutch mothers. The fact that the ex-partners disappeared from the Dutch-Curaçaoan mothers' lives plays a role here. Furthermore, the Dutch-Curaçaoan mothers frequently experienced Dutch language problems.
It is relevant to note that, although the single mothers experience many different problems, it was clear from the interviews that none of them considered themselves to be 'multi-problematic' or to be 'a multiproblem family', instead, they emphasized encountering problems or difficult circumstances, resembling 'survival needs', which made their lives difficult. When the mothers discovered that they were being seen or addressed as a 'multi-problem family' by, for instance, the social workers involved in the family, they found this very painful. Some mothers found out on the Internet that they were designated a multi-problem family because multi-problem family interventions took place in their family. The mothers felt stigmatized because of this label and this led to distrust, which, as a result, harmed the social work relationship and the intervention process.

Progressive focusing: the role of diagnosis in intervention practices
Although I did not know beforehand to which aspects of the family setting I could get access and if I could expect to play an important role in the families' lives, it soon appeared that 5 of the single-mother families received social service assistance on a voluntary basis, i.e. they received child rearing or administrative assistance, and that 17 of the single-mother families (respectively 11 Dutch and 6 Dutch-Curaçaoan families) were subjected to compulsory (legal) child protection interventions. Table 1 below illustrates the type of assistance provided to the thirty single-mothers families in the study.
Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 Table 1 Single-mother families with regard to types of assistance and the depth of the analysis (in italics)  Table 2 illustrates the judicial diagnosis of the families, respectively in 2010 (beginning of the study) and in 2012. In order to be able to understand the families' child protection intervention processes, I first aimed to reconstruct how the single mothers got involved in the child protection system, by scrutinizing their case files. In this respect, it is striking to notice that the six Dutch-Curaçaoan single mothers who voluntarily asked for child-rearing assistance all ended up in a compulsory framework, most often with the diagnosis 'neglect'. The majority of the Dutch single mothers got involved in the child protection system as a result of the diagnosis 'domestic violence', as shown in Table 3 below. I will demonstrate that two catch-22 situations of the child protection system are particularly striking: first, voluntary involvement often turns Utrecht Law Review | Volume 12 | Issue 2 (June) 2016

Single mother No assistance Voluntary assistance
into compulsory intervention, and secondly, mothers are persuaded to leave their violent partner but are still seen as uncooperative. In many Dutch-Curaçaoan cases the single mothers' children were placed under the supervision of a family supervisor because 'support in a voluntary framework turned out to be no longer sufficient'. 85 In this regard, the following argumentation was given to a Dutch-Curaçaoan mother in 2010. The mother was defined as 'not being sufficiently capable of handling her complex situation' and furthermore diagnosed as being 'at risk of possibly neglecting her children' 'by virtue of her limited cognitive abilities'. 86 An analysis of the interview data, furthermore, reveals that the mothers do not consider themselves as being in accordance with the formal categories of the child protection system; on the contrary, these categories evoked fierce resistance from the mothers. 87 A Dutch mother, for instance, was very angry with the family supervisor when she was depicted as an 'emotionally unstable and verbally aggressive mother'. 88

Continuation of intervention practices
In official documents, as arguments for the continuation of child protection orders, professionals often used certain diagnoses: in the case of the mother diagnoses such as 'incapable', 'uncooperative', 'verbally aggressive', and in the case of the child diagnoses such as a 'pervasive developmental disorder' or 'ADHD'. 89 The seven Dutch mothers who were involved in the child protection system as a result of domestic violence were either angry and/or disappointed because they had followed the advice of the family supervisor. This advice was to leave their violent partner, but they were nevertheless being 'punished twice' as a mother indicated. 90 The mother explained that she was blamed for the actions of the father, which were outside her control. In her view, she was a victim of domestic violence (her ex-partner was convicted of attempted homicide) and a victim of the system (her son and daughter had been placed in foster care and there were no signs of any change since the care order had been extended by another year due to the fact that the mother's expartner was still 'potentially dangerous' as she had a restraining order against him).
In this regard, the mother indicated that the child protection agency finds, or invents, yet another reason to continue the care order.
Utrecht Law Review | Volume 12 | Issue 2 (June) 2016 By deconstructing the diagnosis of 'being incapable', it appears that 'being capable' in cases of domestic violence, as in the Dutch mother's case for instance, means that a single mother is able to handle difficult situations on her own and takes care of the children as well as being able to protect them from their violent father. 92 The tendency to focus on the source of protection (the mother) instead of on the source of the problem (the violent father) has been highly criticized in scholarly literature. 93 This dominant 'motherhood discourse', as the focus is on being a 'good' mother and is rarely on the father 94 makes it difficult for the mother to reverse the intervention process to her advantage.
If the institutional representatives invoked a particular diagnosis, such as being an 'incapable', 'uncooperative', 'unwilling' or 'verbally aggressive' mother, this led to even more resistance and a more negative stance towards the representative involved due to the fact that these diagnoses refer to an inherent deficit quality of the mothers rather than a result of an interaction between the mother and professional. 95 Summarizing, it appears that notions of socio-economic class, gender, ethnicity, age, parenthood, and IQ play a role in the risk assessment of the single-mother families. 96

Social work encounter
In the case of ten single-mother families (five Dutch and five Dutch-Curaçaoan) out of the dataset of the 22 families, I gained access to observe the mothers' successive child protection encounters with the institutional representatives involved; these mainly concerned encounters with the family supervisor, often accompanied by social workers in a voluntary framework. Since the family supervisor, the social worker(s) and the mother use, produce and negotiate their identities in a social work encounter, I particularly aimed to assess how the mothers and the institutional representatives (re)negotiate the diagnosis of the singlemother families in real time. In the negotiation of diagnosis, they may orientate to the institutional setting, professional or backstage knowledge, cultural categorizations and the wider child protection policy context. 97 Certain categorizations could be made significant in the institutional argumentation of the institutional representatives, whilst the single mothers also make use of certain categorizations in their argumentation.
Hence, I observed more than 70 encounters between the 10 single mothers and the social work professionals during home visits, court hearings and supervised meetings, so that I could examine how the mothers were addressed in social work encounters and whether certain 'categories' such as, for example, Curaçaoan ethnicity, single-mother parenthood, motherhood gender, low IQ, or low socio-economic class were invoked in institutional argumentation and functioned as legitimation for decisions or indications. 98 In this respect, it must be noted that the interaction between the single mother and the social work professionals is embedded in an institutional setting. This setting is by definition asymmetrical. Neither the family supervisor, nor social workers in a voluntary framework and the mother can escape the asymmetrical setting. This setting underlies intervention practices and underwrites the knowledge of the interactants.
This excerpt from an interaction demonstrates that the family supervisor reinforces the diagnosis that the daughter 'can't live with you' by the argumentation that the daughter is 'hyperactive' and has 'a developmental disorder'. Furthermore, the family supervisor claims that powerful authority bodies -'the judge', 'the supervisor's Agency' and 'psychologists' -have also come to this conclusion. Moreover, the excerpt discloses a competing moral judgment of abnormality/normality: in the view of the family supervisor 'children ought to cry when they fall' whilst the mother's opinion is that 'not all children cry when they fall'. It seems that the family supervisor claims truth by reaffirming her moral judgment. 100 Later in the conversation we see a moral judgment of the mother with regards to what she considers as culturally normal/abnormal: Not only cultural, but also class categories were frequently invoked in child protection encounters. Illustrative is an encounter between a Dutch mother and the family supervisor, whereby the mother had raised her voice and told the family supervisor in question that she lied. In her response, the family supervisor said to the mother that 'shouting might be common to you people'. 103 Based upon the conflict that followed, the family supervisor decided to continue the child protection order.
In this regard, it is striking that the family supervisors often refer to certain 'normalizing' standards in their judgments, as for instance, 'a standard income' (i.e. as opposed to unemployment benefit), 'a standard family' (i.e. two parents), 'Dutch cultural norms' (i.e. as opposed to Dutch-Curaçaoan norms), 'normal housing' (i.e. as opposed to subsidized housing), and furthermore, 'proper food', 'normal IQ', 'good education' and 'good parenting'. Although the single mothers also use certain (similar) categories to make their normative claims, they often find themselves in a disadvantageous position, as they cannot defeat the family supervisors' assumptions of what is considered a 'good' mother or 'in the child's best interest'.

Inconsequent diagnosis
Normative judgments tend to be situational, as these have appeared frequently inconsequent dependent on the circumstances. For example, a family supervisor, following the argumentation of a previous social worker, said to the mother that 'the reason for placing her daughter in care was that there was no proper food in the fridge' 104 whilst taking the mother and her daughter during supervised meetings to McDonalds to eat some French fries, hamburgers and milkshakes. 105 This example shows that the family supervisor in question applied double standards with regards to what s/he considered to be 'proper food'.

Misunderstanding
It is not only explicit examples of cultural categories invoked through language in social work encounters that play a role in the intervention processes. As a result of the 'immersed role' of the ethnographer and an understanding of the life worlds of the Dutch and Dutch-Curaçaoan single-mothers, it has become apparent that certain culturally embedded linguistic features generated major consequences for the Dutch-Curaçaoan family in question. The use of language also plays a major role. An example is the word 'e' that in Papiamentu refers to both 'he' and 'she'. Hence, due to a lack of knowledge of how grammatical and lexical categories are embedded in language, it has become apparent that misinterpretation and misunderstanding often lie beneath the surface of social work encounters. The following excerpt is an example. A Dutch-Curaçaoan mother frequently used the wording 'he is thinking' ('hij is denken', whilst she means to say 'she thinks'), 'he is saying' ('hij is zeggen', but means 'she says') and 'he is finding' ('hij is vinden', whilst meaning 'she has the opinion that'). Later, in an official report, the family supervisor had written that the mother did not seem to know that she had a daughter since the mother always used the word 'he', and used the plural to make her statement. 106 In this regard, it must be noted that after a noun the plural form of the verb follows in Papiamentu.
In sum, based on a critical discourse analysis of more than 70 social work encounters between the single mothers and the social work professionals, it seemed that both cultural categories embedded in language and essentialist notions of 'culture' are frequently invoked by both social work professionals and the single mothers to reinforce arguments in discussions about what is 'good' behaviour and what is not. 107 These notions appear to lie behind the construction of the mothers by the social work professionals. For instance, 'being not capable of raising her children' is such a construction.