‘You shouldn't blame religion … but the person’ – the ethnic boundary work of young second-generation migrants in Rotterdam

In the Netherlands, as in most other western European countries, the desirability and the governability of a multicultural society are topics of debate. In the last decade, this debate has increasingly centred on second-generation migrants, focusing on their high rates of crime and school drop-out. In the Dutch context, however, little scholarly research has paid attention to second-generation migrants’ own experiences. In this paper, I therefore focus on the perceptions of ethnic boundaries held by 12- to 19-year-old second-generation migrants and how they negotiate these boundaries in the low-income, multi-ethnic Feijenoord area of Rotterdam. The study shows that young people are used to living together with many different cultures and see themselves as being on both sides of the ethnic boundary between the Dutch-majority society and the culture of their parents. However, they also encounter prejudice and discrimination in their day-to-day lives, which calls into question the success of multiculturalism.


Introduction
Cities in the Netherlands, like those in most other western European countries, are becoming more hyper-diverse. Not only are the cities diverse in ethnic, demographic and socio-economic terms, but also within groups there are many differences with respect to attitudes, lifestyles and activities (Tasan-Kok et al. 2013). The desirability and the governability of a multicultural society are topics of debate. In the last decade, these debates have increasingly centred on the lives and futures of second-generation migrants. On the one hand, this is a group for which hyper-diversity is a normal part of life. On the other hand, politicians increasingly argue that the multicultural society has become a failure, referring to the high rates of crime and school drop-out among second-generation migrants (Crul, Schneider, and Lelie 2013). In the Dutch context, however, little scholarly research has been done into how second-generation migrants themselves perceive and act upon the ethnic boundaries that are commonplace in today's society. In this paper, I therefore focus on young people's perceptions of ethnic boundaries and how they negotiate these boundaries in a low-income multicultural area in Rotterdam.
Several developments suggest that this is a good moment to focus on ethnic boundaries. Firstly, in the Netherlands and other western European countries, a second and even a third generation of migrants are now emerging. It is interesting to focus on these generations for two reasons: they were born and socialized in the Netherlands and thus have the potential to cross or blur boundaries that their parents took for granted (Alba 2005); and as a consequence of globalization, children of immigrants may develop bicultural or hybrid identities (Crul and Vermeulen 2003).
Secondly, recent political developments have led to the reproduction of the boundaries between migrants and non-migrants, most notably between Muslims and non-Muslims. For example, in the campaign led by right-wing Dutch politician Geert Wilders, and picked up by mainstream media, the children of migrants, especially those from a Moroccan background, are considered to be a problem and are often excluded from the Dutch national identity (Leurs, Midden, and Ponzanesi 2012). Muslim boys are increasingly depicted as 'street terrorists' or fundamentalists, while girls are constructed as unemancipated or oppressed (Akbarzadeh and Smith 2005;Leurs, Midden, and Ponzanesi 2012;Schønemann 2013). This idea of Muslims as terrorists is further strengthened by the rise of radical Islam in North Africa and the Middle East, as well as the terrorist attack at Charlie Hebdo in Paris, which took place at the time of the interviews.
The third reason for the focus on young people's perceptions and negotiation of ethnic boundaries is that the literature increasingly calls for attention to the idea that ethnic boundaries are socially constructed (Nagel 1994;Alba 2005;Omi and Winant 2014). In the last decade, the literature (Zolberg and Long 1999;Lamont and Molnár 2002;Lamont 2009) has increasingly shown that boundaries depend on the social context people are in, and that boundary-related change is the result of a diverse set of processes (Alba 2005). This research focuses on how through everyday encounters between young people and 'others' in public space, boundaries are constructed, deconstructed and reconstructed. By doing so, I respond to the call made by Horton and Kraftl (2006) for researchers in the field of children and youth studies to explore mundane, everyday events and ongoing practices in young people's lives. This approach is embedded in the ideas on everyday multiculturalism put forward by Wise (2007). This perspective focuses on how diversity is experienced and negotiated in everyday situations such as the neighbourhood or the school, and how through these everyday encounters identities and social relations are produced and reproduced (Wise 2008). This perspective is different from the traditional approach to multiculturalism, which is mainly policy-oriented and focused on group-based rights, service provision and legislation (Wise 2008).

Ethnic boundary work
In this study, ethnicity is conceptualized as a boundary with both symbolic and social aspects (Lamont and Molnár 2002;Wimmer 2008). Lamont and Molnár (2002, 168) describe symbolic boundaries as 'conceptual distinctions made by social actors to categorize objects, people, practices, and even time and space'. Symbolic boundaries separate people into groups and produce feelings of similarity and group membership, often based on social and cultural characteristics, and shape people's actions and mental orientations towards others (Epstein 1992, 232;Alba 2005). When symbolic boundaries are widely agreed upon, they can constrain social interaction and opportunities. Symbolic boundaries can then translate into social boundaries, which are 'objectified forms of social differences manifested in unequal access to and distribution of resources (material and nonmaterial) and social opportunities' (Lamont and Molnár 2002, 168). Thus, when symbolic boundaries are agreed upon, they can translate into patterns of social exclusion and segregation (see e.g. Massey and Denton 1993).
The formation of these symbolic and social boundaries is not, however, a one-way process. Identities, and the boundaries between these identities, are always constructed in relation to the other, and should thus be seen as a simultaneous process of collective self-identification and external categorization (Jenkins 1996). Researchers increasingly emphasize the importance of ethnic 'boundary making' or 'boundary work' rather than seeing ethnic boundaries as something static. This development is in line with the general trend in social sciences to move away from structural determinism towards theories that emphasize 'agency' and social constructivism (Emirbayer and Mische 1998). In this paper, I therefore foreground the importance of individual negotiations and strategies in the production of ethnic boundaries. In the words of Lyman and Douglass (1973, 349): Just as an ethnic group must affirm and reaffirm its boundaries (or be reminded by others of what they are) in order for such boundaries to retain social relevance, so also individual ethnics must affirm and reaffirm an ethnic identity (or have it reaffirmed by outsiders) in order for it to be a feature of any social situation in which they are participants.
Useful in this context is the three-way distinction that Zolberg and Long (1999; see also Alba and Nee 2003;Alba 2005) make when discussing the change in a person's position towards the ethnic boundaries; namely between boundary crossing, boundary blurring and boundary shifting. In the context of this study, it is particularly relevant to look at boundary crossing and boundary blurring. Boundary crossing means that individuals adopt the norms, values and practices of the majority society; they move from one side of the boundary to the other side, but the boundary itself does not change. In other words, assimilation takes place (Zolberg and Long 1999;Alba and Nee 2003;Alba 2005). Boundary blurring, on the other hand, describes the way in which a group can become part of the majority society, without having to abandon most of their norms, values and identity. At the same time, the majority society also allows the participation of migrants and multiple memberships by changing the social, cultural and legal structure of society (Zolberg and Long 1999). We can speak of a blurred boundary when the position of a group of individuals with respect to the boundary is indeterminate or ambiguous (Alba, Reitz, and Simon 2012). Individuals could be members of groups on both sides of the boundaryand thus feel that they are members of both an ethnic minority and the mainstreamor switch their membership, depending on the context. The final process, boundary shifting, involves the relocation of the boundary itself. This means that groups of people who were outsiders now become insiders (Alba 2005).
In this study I focus on two ways in which ethnic boundaries can be produced, reproduced or changed, namely through the everyday encounters of individuals and through the political discourse present in society.

Everyday encounters
Several scholars have argued that in order to develop this constructivist notion of ethnic boundaries, a further exploration of how ethnicity is produced and reproduced in the everyday encounters between individuals is needed (Barth 1998;Brubaker 2002;Lamont and Molnár 2002). In this study, ethnic identity is conceptualized as the continuously changing outcome of social encounters (Hale 2004). It is generally argued that perceptions of other groups, perceptions of diversity and identification of the self (and thus ethnic boundaries) can be formed through regular encounters in shared space. Some authors argue that regular encounters in shared space can lead to the blurring of ethnic boundaries. Amin (2002), for example, discusses how through encounters with others in 'micropublics', ethnic, religious, class and other boundaries can be blurred, which leads to less stereotyping and prejudice (see also Wessendorf 2014). On the other hand, Valentine (2008, 332), for example, is more sceptical and argues that 'positive encounters with individuals from minority groups do not necessarily change people's opinions about groups as a whole for the better'. Daily courtesies in public space often coexist with privately held prejudices (Wise 2005;Noble 2011).
Political discourse: Islamization, prejudice and discrimination The second factor that influences the boundaries between minority and majority populations in society is the political discourse, and in the case of the Netherlands particularly in relation to Muslims and Islam (see also Allen 2010). In the western European context, the dominant understandings of Islam strongly influence the formation of boundaries between migrants and majority society (Zolberg and Long 1999;Alba 2005). In the public debate it is not only ethnicity that influences the boundary between minority and majority cultures: being seen as 'on the other side of the boundary' is often tied to being from a Muslim background. Opponents of the multicultural society argue that Islam is irreconcilably different from the Dutch culture and identity (Leurs, Midden, and Ponzanesi 2012), and as such brighten the boundaries between migrants and majority society.
The Islamization of migrants can be connected to factors like the rise of radical Islam in Muslim countries, the growing visibility of Islam in Europe and the increasing fear of terrorist attacks such as the recent ones in Paris (Leurs, Midden, and Ponzanesi 2012). These global events are experienced locally by migrant young people in a range of ways that are important to their senses of identity and social inclusion or exclusion. Peek (2003), for example, showed how Muslim students in New York became victims of 'discrimination, harassment, racial and religious profiling, and verbal and physical assault' (271) after the events of 9/11. Similar results were found by Hopkins (2004) in Scotland, who showed that signifiers of Islamic identity such as dress choice, skin colour and having a beard became more important in influencing young Muslims' marginalization since 11 September. Noble (2005) further showed in the Australian context that many Muslims no longer felt 'at home'. He argues that for many of his respondents, it was not only the harsh and overt forms of racism, but also the mundane, even routine forms of harassment that were very relevant. The young people in his study experienced subtle, commonplace forms of racism, such as being ignored, ridiculed or treated differently.
Several researchers have found that discrimination and social exclusion are increasing and are a part of daily life for many minority ethnic youths (Anisef and Kilbride 2003;Galabuzi 2006;Herz and Johansson 2012). It can have an impact on the cultural identity of immigrant youths, and their sense of belonging to or exclusion from the majority society (Phinney and Devich-Navarro 1997;Phinney et al. 2001Phinney et al. , 2006Jasinskaja-Lahti et al. 2006).

Context and methods
Before discussing how young people experience and deal with ethnic boundaries, it is necessary to set their narratives within the specific context in which they live their daily lives. Young people's management of space, and their construction and reconstruction of boundaries, only becomes meaningful when placed against the socio-economic framework of a neighbourhood such as Feijenoord. It is therefore necessary to provide a brief overview of this neighbourhood.
Feijenoord is a sub-municipality of the city of Rotterdam. This area's population has a low socio-economic status: the average income is €10,500 (compared to the city average of €12,500), 29% of the residents receive benefits and 63% have a low level of education. Moreover, compared to the city average, the area is beset with high levels of unsafety, nuisance caused by youths and the use of drugs. Most of the inhabitants (65%) are from a non-Dutch background: the largest non-Dutch groups are Turks (20%), Surinamese (11%), Moroccans (10%) and Dutch Antilleans (4%) (all data: GGD Rotterdam-Rijnmond 2010). The neighbourhood context of this research provides an interesting context for an analysis of ethnic boundaries among the second generation, because the area is ethnically diverse and is not dominated numerically by any particular ethnic group.
This paper is based on fieldwork from two research projects. The first round of fieldwork (January-August 2013) was part of a study on how young people perceived and dealt with growing up in a deprived neighbourhood (see Visser, Bolt, and Van Kempen 2015). The participants were first asked to take photographs of the places that were important to them, and places that they liked or did not like, using their own camera phones. The use of photography allowed the young people to express their experiences with spaces and share information about their activities and social networks through a more tangible method than just an interview. The photos were used as the starting point for semi-structured in-depth interviews, which focused on the young people's socio-spatial practices, who they met in different settings and how participation in these settings influenced their lives. The interviews also included a narrative mapping exercise, in which the young people were asked to draw what to them were the most important places, as well as the places they did not like. The advantage of mental mapping was that it gave the participants the freedom to express themselves with limited intervention by the researcher. It was also a good method to gain insight into the relative importance of places in their daily lives (Young and Barrett 2001).
The second round of fieldwork (January-March 2015) investigated young people's everyday lived experiences of urban diversity and their practices related to urban diversity. The interviews focused on perceptions of urban diversity; socio-spatial activities; the diversity of encounters in places like the neighbourhood, school, youth centres, etc.; and the diversity of social networks and the resources present in those networks. In both rounds of fieldwork, respondents were recruited through community organizations and schools for secondary education and vocational education (MBO: middelbaar beroepsonderwijs). Snowball sampling was used to recruit additional participants.
Twenty-one interviews with second-generation migrant youths were conducted in the first round, and 18 in the second round. The interviews took place at the youths' homes or schools, or in youth centres. Most were conducted individually, but sometimes, when the respondents preferred this, they were conducted in groups of two or three friends. The sample consisted of 22 boys and 17 girls from diverse ethnic backgrounds (Turkish, Moroccan, Surinamese, eastern European, Dutch Antillean, Afghan and Pakistani), who were receiving different (mainly low and middle) levels of education. A large share of the participants were from a Muslim background.
The interviews were transcribed in their entirety and then coded and analysed in NVivo. General patterns identified in the data in the first round were further refined during the subsequent rounds. Furthermore, text query and negative case analysis (i.e. discussing elements of the data that did not support patterns that were emerging from the data analysis) were used to determine the relative strength of themes emerging from the data. Each participant was assigned a pseudonym, which is used throughout this paper.

Young people's position in relation to ethnic boundaries
The first theme that emerged from the interviews was the 'normality' of difference in the lives of the young respondents. They were used to living in a neighbourhood with many different cultures and they indicated that most young people in the neighbourhood got along very well, hanging around with each other across ethnic boundaries. As most of the respondents had been born and raised in Feijenoord, ethnicity was not perceived as the main dividing factor. Most young people indicated that they felt positive about living in an area with many different cultures. One of the possible reasons for this is that it allows them to preserve their own culture. Moreover, many respondents felt that they could learn from other cultures and could enjoy each other's foods and festivals. These findings are in line with the literature on everyday multiculturalism. Several authors in this field (e.g. Hoerder, Hebert, and Schmitt 2005;Wise 2007;Harris 2009) found that young people normalize heterogeneity; in other words, cultural diversity is more familiar to young people than to older people. Today's young people increasingly grow up in a diverse world and have not experienced times before hyper-diversity. It should be noted, however, that this does not necessarily mean that these young people embrace this diversity, although they are far less likely to perceive it as unusual and undesirable (Back, Räthzel, and Hieronymus 2008).
Even though most of the respondents regarded diversity as normal and positive, and indicated that there were no bright boundaries between several minority groups, they were well aware of the boundary between minorities and majority society and of their own position in relation to this boundary. They said that they identified with the national Dutch culture and with the culture of the country of their parents, and as such created new, hybrid identities. Most of the young people thus saw themselves as being on both sides of the ethnic boundary. When asked whether he regarded himself as Dutch or Moroccan, Arslan (male, aged 17, Dutch-Turkish 1 ) replied: 'No, not really, I'm half/half. I was born here, but I often go to Morocco to get back to where my roots are. You have to care about where you're from'.
A negotiation and balancing was going on between the desire to root themselves in the communities of their parents while at the same time feeling that they are Dutch, since they were born in the Netherlands and had spent their entire childhoods there. The following is an excerpt from an interview with Stanley (male, aged 17, Dutch-Surinamese): Interviewer: How would you describe your own culture? Stanley: I was born in the Netherlands. It also says so on my passport. I feel Dutch as well, in some ways. Interviewer: Why? Stanley: Because I have lived in this country for my entire life. I also speak the language. That's it, basically. Interviewer: Is there a difference as well between different situations or places? That you feel more Dutch or more Surinamese? Stanley: Yes, sometimes there is. It depends who I'm with. When, for example, I'm with many people from a Surinamese background, we sometimes speak our native language, but mostly we speak Dutch.
The above illustrates that although citizenship leads to a sense of belonging to the national culture, there are certain attributes, such as language, that still bind the young people to the culture of their parents. Similarly, we also saw a number of young people speaking to their parents in Dutch while the parents replied in their mother tongue. As Bailey (2007) pointed out, language is the primary way in which people can represent and negotiate social reality. Through language, people can position themselves in the social world; in other words, language is an important part of their identity formation. The young people in our study bridged existing social worlds through their use of language (Zentella 1997;Bailey 2002).
Even though many young people switched between languages and thus blurred ethnic boundaries, most of them were well aware of what the hegemonic ideas were of 'good' language use. They realized that divergences from this ideal standard were considered less desirable in certain contexts, which illustrates the power differences between the majority and the minority (Bailey 2007). Here is what Timur (male, aged 14, Dutch-Turkish) said about it: Timur: They call it street language, but sometimes they use it at school as well. I think that's bad. Interviewer: Why? Timur: Sometimes you speak to the teacher using one of these words, and then he misunderstands, and he sends you out of the room. That happens most of the time. For example, when you want to say 'Sir, that isn't right', then you use palla, which means something like 'nonsense' in street language. But he thinks it's a swear word and sends you out.
Another aspect of boundary blurring that emerged from the study was that participants shifted from one identity to another depending on the space they were in at a particular time. Most of the Muslim girls we spoke to, for example, were able to decide precisely when and where they 'could be themselves'. Girls belonged to various groups, such as their family, the ethnic community, their school community and their peer group. They activated different parts of their identity repertoires according to what was appropriate in that particular group and related space. For example, they emphasized their Muslim identity when they were with family or community members, but emphasized their Dutch identity when they were with their friends. This idea of boundary blurring allows us to understand what might seem to be contradictions and inconsistencies in the practices of the young people. The same person could simultaneously be a Dutch citizen, of Moroccan origin, a Muslim, a woman, a mother and a drugs dealer (Althoff 2013).
In the same vein, many of the young people emphasized identity categories that were more complex than those based simply on ethnicity. This primarily happened as a reaction to the media and the public depicting young migrants as not being part of Dutch society and as a problem (see the following section). I often saw that boundaries between 'us' (the decent ones) and 'them' (the troublemakers) were constructed that did not necessarily coincide with ethnic boundaries and thus blurred existing ethnic divisions (see also Wimmer 2008). In other words, young people reduced the importance of ethnicity as a principle of categorization. Other, nonethnic principles were emphasized, thus undermining the importance of ethnic boundaries. Several participants tried to blur the ethnic boundary by saying that there were also differences within the group and by contesting the image of second-generation migrant youths as a homogeneous ethnic category of people causing trouble or being terrorists. Jamila (female, aged 16, Dutch-Moroccan) and Sumera (female, aged 16, Dutch-Pakistani) illustrated this as follows: Interviewer: Have you seen the news about Wilders and what happened in Paris? People are blaming the Muslims. What do you think about that? Jamila: You shouldn't blame religion for the things somebody does … Sumera: … but the person.
Besides emphasizing that not all Muslim people were terrorists, our respondents also indicated that not all second-generation migrants should be seen as being engaged in deviant behaviour. By contrasting themselves with people and places they associated with deviant behaviour, young people felt they were emphasizing important characteristics of themselves, such as behaving well, not engaging in violence, and being anti-drugs and anti-alcohol. In other words, a certain identity was constructed by constructing and maintaining boundaries other than ethnic boundaries. Through their behaviour, young people drew together the 'right' spaces and people and as such constructed or reinforced socio-spatial boundaries. Here is how Yasir (male, aged 17, Dutch-Afghan) explained it: Well, actually I know … I never hang around with people who are bad examples or something like that. Always with people I think are good people. Never somebody who smokes or drinks. I never hang around with those kinds of people.
The divisions that young people made between identities correlated with the divisions they made between different social-spatial environments. They linked going to school, the community centre and the mosque with a process of emphasizing their identity as responsible people who engage in 'good' behaviour, as opposed to the street, which was linked to undesirable behaviour.
By distancing themselves from certain places, people and identities, they were trying to become more 'in sync' (Johansson and Olofsson 2011) with the norms and values of the Dutch-majority society. They were well aware of the prejudice some people have against migrant youths, but many had found a strategy to dissociate themselves from these images. By choosing the 'right' friends and using clear demarcations between 'good' and 'bad', the youths were attempting to create a positive image of themselves and reduce the stigma attached to them. This is an ambivalent practice, one that must be understood in relation to the stigmatization of migrant youths (see also Johansson and Olofsson 2011). It is ambivalent because in order to do well in life, young people have the feeling they should adhere to majority, middle-class norms and values, thereby repudiating their own background.
To sum up, the young people in our study perceived themselves as being on both sides of the boundary between the majority and the ethnic minority, or they sometimes appeared to be members of the majority and at other times members of the minority. They felt Dutch because they were born in the Netherlands, spoke the language and had Dutch citizenship, but they also identified with the culture of their parents and the local culture of the neighbourhood. From their perspective, the ethnic boundaries between the minority and majority cultures should not be seen as fixed but rather as 'blurrable'. However, we should not overlook the importance of social power in the production of these boundaries. Whereas the young people in our study, for example, feel that they are Dutch citizens, and therefore part of the majority, the majority society produces and reproduces boundaries between second-generation migrant youths and the Dutch-majority society. The dominant groupin this case, the Dutch-majority societyplays a critical role in the production of ethnic boundaries and the hierarchy of the groups on both sides of the boundary. In the following section, I therefore focus on young people's experiences of social exclusion and discrimination.

Young people's experiences of social exclusion and discrimination
Although the young people in our study indicated that diversity was a normal part of their everyday lives and that they hung around with peers from different ethnic backgrounds, they were also aware of social distances of class and ethnicity and how 'others' thought about them. As Meliha (female, aged 19, Dutch-Turkish) pointed out: 'People have a totally different image of you. Because you're an ethnic minority. They have the image of criminality, bad upbringing, a bad childhood, a bad future, no school, no job'.
Many of the respondents said that the prejudice people hold against (second-generation) migrant youths was related to the media's dominating narratives of migrants, and talked about how often they were portrayed within the context of criminal activities. Whereas in the public opinion youth gangs and conflicts between groups of young people were seen as the main problem, the respondents challenged this idea and indicated that it was rather the behaviour of the police and the media that contribute to local conflict. Selami (male, aged 16, Dutch-Turkish), for example, said the following: When you put 30 police vans in a neighbourhood, and the police are very frustrated, they look at you with this tense look on their faces, that's what you get [boys causing even more trouble]. In [another neighbourhood]: two police vans, police are smiling at you. But here, you'll get arrested right away, get a preventive pat down, things like that. 'We saw you hanging around here today and there was a robbery, so we are going to take you to the police station'.
Ossama (male, aged 15, Dutch-Moroccan) said that the media strengthen the negative image of young Moroccan males: 'The media really gives the wrong example, because it is not all Moroccan people that are wrong. Most of them are just normal'. This is in line with the findings of Harris (2010), who argues that in particular racism and discrimination are some of the most important factors damaging social cohesion, possibly even more so than tensions and violence between groups of young people from diverse ethnic backgrounds.
Changes in the socio-political climate in the Netherlands as well as the larger global political milieu affected our respondents' sense of belonging to or exclusion from Dutch society. Global problems, like terrorist violence and the growing influence of radical Islam, have led to a negative image of all Muslims, which makes outsiders see Muslim (or migrant) youths as potential terrorists. This is in line with Said (1985Said ( , 2003Said ( , 2008, whose work shows how the western media depicts Islam and Muslims as primitive, barbaric and violent. Also the nationalist politics of right-wing Dutch politician Geert Wilders further shaped the public's negative opinion of second-generation migrant youths. The changing socio-political climate in Rotterdam was also noted by Aamina (female, aged 18, Dutch-Surinamese): A while ago I was sitting in the tram, on my way to school. It was morning and I was only half awake. And then a girl wearing a headscarf got on. And the driver said, we're not going to leave, because there's a girl wearing a headscarf ( … ) I asked him what was going on, because we'd been standing still for so long, and I had to go to school. And the driver said: yes, there's a Muslim, she is going to terrorize us.
The narratives of young people show that it is primarily the visibility of religious identity that leads to ethnic boundaries between the young people and the majority. Aamina, too, is Muslim, but she does not wear a headscarf, which leads to less prejudice: And this guy [the tram driver] talks normally to me. So I tell him: so you're talking normal to me and not to her. And he said: there's a large difference between you two. And I said: no, there's not. She's my Islamic sister; I'm a Muslim as well. And then he looks at me … You don't see me terrorizing anybody, right? I'm just too lazy for that. So neither is she.
Quite a few of the female participants said that wearing a headscarf fundamentally affected the extent to which they could access the opportunities afforded to the majority. We can thus see that the symbolic boundary is translated into a social boundary, since the social differences manifested themselves in unequal access to and distribution of resources (Lamont and Molnár 2002). Olcay (female, aged 16, Dutch-Turkish) reported the following: A while ago I went to the local supermarket to apply for a job, and they told me that they don't hire people wearing headscarves. And I also went to this electronics shop, and it was the same there, at different places.
The young people's experiences also show that ethnic discrimination is interwoven with gender discrimination. As noted by Frost (2010), ethnic discrimination seldom happens in isolation from other forms of discrimination, including gender. As shown above, for girls, wearing the headscarf was mainly seen as the problem, as it is regarded as signifying women's general subordination or sympathy with radical Islam. Migrant boys, on the other hand, were often characterized as criminals or gangsters, and as such were seen as potential threats and a source of anxiety. Blokland (2003) argues that when people have little personal information they are more likely to draw on available categories and ascribe stereotypical identities to others. In the case of Feijenoord, many adults in the neighbourhood applied their stereotypical knowledge of non-western boys hanging around in public space in general to the situation in their neighbourhood. How this had an impact on young people is illustrated by the story of Ali, Musa and Selami (three Dutch-Turkish male friends, aged 17, 16 and 16, respectively). They talked about the places they could or could not hang around in the neighbourhood. They mentioned a whole range of places from which many of them felt excluded. Ali said that people in the neighbourhood sometimes felt uncomfortable when the boys hung out in a group: 'People just tell us that they feel unsafe when we're standing there. Not that we're doing anything, but some people just don't like it when there are five people standing in front of their door'.
Moreover, Selami said that there were often many police officers walking or driving around the neighbourhood, and that they reprimanded the group even when the boys were doing nothing wrong. Again we can see that the symbolic boundary between the majority society and minority youth is translated into a social boundary, as it results in unequal access to resources. These youths said that false representations of their culture affected not only how others viewed them as individuals, but also how they viewed themselves. They thought it might be better to stay indoors and generally be as little a nuisance as possible. Specifically the stigmatization of Muslim boys played a role here, as pointed out by Selami: Particularly the Muslims, for example the Moroccans or Turks, then they say: it's the Muslims again. When we are outside, people are like: 'Look, another Muslim'. The more we stay inside, the better it is for the Muslims and the Dutch.
Finally, although some of the youths experienced very explicit forms of discrimination, as emerged, for example, from the story of Aamina, for most of the young people in our study their encounters with discrimination were intangible and hidden. Only few mentioned physical violence or name-calling. Prejudice mainly emerged from subtle signs of rejection, for example a certain way of looking at or avoiding a group of migrant boys on the street. Possibly as a result of this, many of our participants did not actively respond to the discriminatory practices. As Stanley (male, aged 17, Dutch-Surinamese) put it: 'I don't really care what people say about me. About me and about others. I just let it go'. Noh et al. (1999) found that forbearing responses, such as passively accepting discriminatory experiences or not reacting to them, could be seen as a form of adaptive coping, which would reduce the impact on the young people's well-being.

Conclusion and discussion
In all immigration societies, symbolic and social boundaries between migrants and second-generation migrants on the one hand and the majority society on the other hand are part of everyday life. These boundaries can be seen, in a sense, as 'fault lines along which other differences and distinctions pile up' (Alba 2005, 41). In the Dutch context, however, little scholarly research has been done into how second-generation migrants perceive and act upon the ethnic boundaries that are commonplace in today's society. In this paper I therefore focused on the perceptions of ethnic boundaries held by 12-to 19-year-old second-generation migrants, and how they negotiate these boundaries in the low-income, multicultural Feijenoord area in Rotterdam.
This study found that these youths were used to living in a neighbourhood with many different cultures, and most of them said that most young people in the neighbourhood got along very well, hanging around with each other across ethnic boundaries. They were often in favour of multiculturalism, which is part of their everyday lives. The interviews also show that there are differences between the institutional and formal dimension of belonging and citizenship on the one hand, and the meaning attributed to it in daily interactions on the other hand. The participants saw themselves as being on both sides of the ethnic boundary between the Dutch-majority society and migrants. They described themselves as, for example, Dutch-Moroccan, Dutch-Turkish or Dutch-Indian. Their position in relation to the ethnic boundaries was dependent on the social and spatial context they were in. Citizenship, being born in the Netherlands and speaking the language led to a sense of belonging to the national culture. At the same time, they addressed their ethnic and religious backgrounds in the family and community context through festivals, food and language. They thus saw themselves as simultaneously members of groups on both sides of the ethnic boundary, or they sometimes appeared to be members of the majority society, and at other times members of the ethnic minority. Depending on who they were with, or the situation they found themselves in, they could emphasize certain aspects of their identity, code-switching between the Dutch aspects of their identity and the identity markers of the culture of their parents. The findings of this study show that the participative and geographical dimension namely where they were, what they were doing and who else was presentwas more important than the dimension of belonging determined by blood (see also Colombo 2010).
It is important in this context to realize that ethnic boundaries are not static. The study showed that boundaries were contested by the young people. Through their everyday behaviour, they were able to challenge the ideas of what it means to 'be Dutch' and to 'be a migrant'. The results of this study confirm Harris' statement that young people should not be seen as 'passive vessels for identity or attitude problems ' (2009, 202), but that they develop strategies to manage encounters with diversity, such as challenging stereotypes and celebrating difference (see also Butcher and Harris 2010). Our participants, for example, often constructed boundaries between 'us' (the decent ones) and 'them' (the troublemakers) that did not necessarily coincide with ethnic boundaries and thus blurred existing ethnic divisions. By distancing themselves from certain places, people and identities, they tried to become more 'in sync' with the norms and values of the Dutch-majority society. We must understand that this is a rather ambivalent practice, however, since it is clearly based on ideas of the superiority or inferiority of certain languages or conventions.
The results of this study thus confirm the importance of young people's agency, but also illustrate the obstacles and the power relations that influence youths' behaviour and identity formation. There were still bright boundaries in Dutch society that influenced the youths' behaviour, life chances and sense of self. The restrictions on the ways in which ethnic boundaries can be negotiatedin other words, on young people's agencymust be recognized. Racism and prejudice and a lack of equal opportunities can exclude young people from particular spaces and resources (Butcher and Harris 2010). The young people in this study encountered prejudice and discrimination on a daily basis. For girls, wearing the headscarf was mainly seen as the problem, as it is regarded as signifying women's general subordination or sympathy with radical Islam, whereas boys were generally seen as terrorists and criminals. This affected the processes by which individuals gained access to the opportunities afforded to the majority. We saw that the symbolic boundary between migrants and majority society was translated into a social boundary, as the social differences manifest themselves in unequal access to and distribution of resources (Lamont and Molnár 2002). Our findings suggest that discrimination and prejudice are damaging to cohesion as well as to young people's sense of self.
This calls for further research into how ethnic boundaries function as obstacles to the social mobility of young, second-and third-generation migrants. At least some of the respondents felt, to some extent, socially excluded from Dutch society, primarily that outside the Feijenoord area. In order to 'fit in', young people can adopt practices of self-censorship, such as trying to 'erase' themselves and changing their appearance, behaviour and language (De Finney 2010). As Johansson and Olofsson (2011) showed, there is a strong connection between conceptions of being the 'other' and decisive turning points in young people's life plans. The life plans and the educational and occupation careers of the young people in their study were strongly linked to structured inequality in Swedish society. An interesting avenue for future research would be to focus on young migrants' adaptive coping, and how the conception of being 'the other' relates to their social outcomes or subjective well-being.
Finally, the findings indicate that ethnic identity may be directly affected by the global valuation of different cultures, or the stigmatization of migrant identities within the Dutch context (see also Khanlou, Koh, and Mill 2008). Young people's ethnic identity and experiences in the Netherlands are shaped by events occurring in their countries and cultures of origin, such as the rise of radical Islam, and by how these events are portrayed or responded to in the Netherlands. Further research in the field of children and youth would benefit from paying more attention to the global as well as the national context in shaping youth's identities, for example by means of comparative research. This means we should go beyond the 'local' that is often the focus in the field of children and youth studies. The socio-political climate of the host country as well as the larger global political milieu can affect how youths relate to their identity, and their sense of connection and belonging to or exclusion from the host country.