Identification Paradoxes and Multiple Belongings: The Narratives of Italian Migrants in the Netherlands

In a time identified by many as one of “multicultural backlash,” we can observe a growing negative discourse on the integration of migrants with Islamic backgrounds in most European countries. Criticisms are rooted in the assumptions that cultural and religious differences are the source of social problems and that these migrants are unwilling to integrate. The aim of this article is threefold. First, it criticizes the linear and simplistic assumptions of integration informing the present negative dominant discourse in the Netherlands. Second, it shows that sources of belonging are more layered than the often-assumed exclusive identification with national identity. Third, it broadens the scope of discussion on integration (which is now mainly fixated on Islamic migrants) by showing the somewhat similar experiences of Italian migrants on their path toward integration and belonging within the Dutch context. Through this study, we argue that the process of ethnic othering in the Netherlands is broader than the often-assumed cultural difference of non-Western migrants.


Introduction
At the turn of the 21st century, we observed excessive negative attention-or as some call it a "moral panic" (Parekh, 2008;Vasta, 2007)-toward migrants with an Islamic background. In most European states there has been a so-called "multicultural backlash," with the Netherlands often mentioned as an example because of its shift from a tolerant integration policy toward a restrictive, assimilative policy (Vasta, 2007;Vertovec & Wessendorf, 2009). The Dutch discourse of migration has been dominated by the arrival of the socalled guest workers in the late 1950s (Ghorashi, 2010a), despite a longer and diverse history of immigration to the Netherlands. Postwar economic growth and the need for unskilled labor forced the Dutch government to look be-yond its borders, fostering labor contracts, first with Italy and Spain and later with Turkey and Morocco (Wilterdink, 1998). This background contributed to the persistent image of immigrants as being low educated and low skilled even as the reality of migration to the Netherlands became much more diverse. Despite this dominant image and the continuity in the policies from the 1970s until now (particularly in terms of attention for economic integration), we can see a clear shift in the Dutch discourse, particularly in how migrants' cultures have been approached in policies and politics. In the 1970s, migrants (then guest workers) were encouraged to maintain their cultural traditions and identities, with the idea that they would eventually return to their country of origin. In the 1980s, when the idea of returning was revealed to be unrealistic, there was a shift in policies to-ward promoting integration in addition to, but not replacing, the preservation of one's own culture. During that time, integration and the cultures of ethnic groups were not considered mutually exclusive. But by the turn of the century, we saw that the cultures and religions of non-Western ethnic others (including both migrants and their children) were increasingly considered to be the main source of social problems and disruption of social cohesion in society (Schinkel & van Houdt, 2010). This has been the backbone of the growing restrictive and assimilative turn in Dutch discourse. Loyalty to Dutch identity is increasingly presented as a dilemma for non-Western migrants: either they choose the Netherlands or they choose their home country (Scientific Council for Government Policy [WRR], 2007). This shift in the Dutch discourse, from integration to assimilation with an emphasis on loyalty toward Dutch national identity, is argued to be the result of the perceived failure of integration during the "multicultural era" (Schinkel & van Houdt, 2010;Vasta, 2007).
We aim to show that the complexity of the integration process goes beyond the binary division of success and failure. Our study shows that the process of integration is multilayered, ambiguous, and even paradoxical, not, as many assume, linear. We also aim to broaden the discussion on migrants' integration in the Netherlands by focusing on an under-researched group in the context of migration, namely, European migrants within Europe, particularly the experiences of Italian migrants in the Netherlands. Most studies on migrants in the Netherlands focus on Turkish or Moroccan migrants (Prins, 2011;Vedder & Virta, 2005;Vermeulen & Penninx, 2000). This study focuses on the experiences of integration and belonging of two waves of Italian migrants who came to the Netherlands after 1960. With this choice we hope to add nuance to the assumption that the "integration problem" is a problem of non-Western ethnic groups who are considered to be others with very different cultures and religions. The main assumption within the multicultural backlash is that the cultural and religious contrast and apparent incompatibility of Western self and non-Western other is the major source of the decline in social cohesion and unrest. Anthias (2013, p. 2) rightly argues that this strong "culturalization of social relations" leads to the reification of difference as dangerous and blinds us to other broader sources of exclusion. The lack of studies on migrant groups considered to be closer to the European self-reinforces this assumption. For that reason, the aim of this research is to show the broader sources of exclusion, deconstructing the culturalist foundation of the present discourses of integration.
Finally, we aim to look beyond national identity as a single source of belonging. Several studies show that migrants today increasingly maintain ties with more than one nation, facilitated by technology (Castles & Miller, 2009;Glick Schiller, Basch, & Szanton Blanc, 1995). However, knowledge remains limited about the layered identification process within a new host society over time and the sources of belonging beyond national identity (see also Zontini, 2015). Using in-depth open interviews, we show the struggles of two generations of Italian migrants in the Netherlands to become a part of that society. We use the WRR's (2007) concepts of functional and emotional identification to demonstrate the layers, dynamics, and tensions of different forms of identification over time. Moreover, the in-between position of these migrants (as both the European self and the non-Dutch other) provides us with challenging material for unraveling a number of assumptions, including linear integration. Duyvendak and Scholten (2010) argue that studies on immigration and integration are often squeezed into simplified theoretical models, whereas, in practice, the process is more complex, paradoxical, multisided, and multilayered. A model that is often used by policymakers is Esser's (2003) four dimensions of integration. These dimensions are presented as subsequent stages, where successful completion of the successive Culturation, Positioning, and Interaction dimensions will eventually lead to the final phase of Identification, where the immigrant feels she or he belongs to the new society. The importance of cultural adjustment (including language proficiency) to the society has become a benchmark of Dutch integration policies, visible in policy documents and in public statements (Entzinger, 2009). The prominence of policies as well as studies assuming integration as a linear process or a straight-line approach is being increasingly criticized, claiming that integration is a complex and multilayered process that, in practice, includes paradoxes (Alghasi, Eriksen, & Ghorashi, 2009;Brenninkmeijer, Geers, Roggeband, & Ghorashi, 2009;Eijberts, 2013).

Theoretical Framework
A good example to illustrate this critique is language as part of culturation. In a linear approach to integration, language is the key factor in the integration process because it contributes to societal inclusion by enhancing job opportunities (positioning) and facilitating contact with natives (interaction; Esser, 2003). However, other studies have shown the double-edged sword of language, meaning that it can be a source of both inclusion and exclusion. Ghorashi and van Tilburg (2006) found that refugee women from Afghanistan and Iran considered the Dutch language essential to obtaining a job in the Netherlands, but even when they were able to speak Dutch fluently, employers continued to perceive this as "not good enough" (positioning). Through their autobiographical research, Davis and Nencel (2011, p. 5) argue that, as immigrants-in their case, from the United States-they still have to deal with comments about their accent or grammar on a daily basis, despite their Dutch fluency and their living in the Netherlands for most of their lives (interaction).
Another example is the assumption that, once migrants are integrated into society economically and so-cially, thus having passed the positioning and interaction stages, they will eventually enter the stage of identification with their host society (Esser, 2003). Buijs, Demant and Hamdy (2006) oppose this line of reasoning by showing how radicalized Muslim youths are not among those who are socially isolated or economically deprived; instead they are more active and involved in society and, as a result, sensitive to feelings of unfair treatment by the host society. The researchers describe this as the "integration paradox": when migrants are actually eager to integrate into dominant society, they are most sensitive to feelings of exclusion (Buijs et al., 2006, p. 202; see also Ghorashi, 2010b).
Despite these critiques, the linear model is still prominent in the Dutch discourse on integration, and it still guides policymaking. There is growing pressure for migrants to adjust themselves to what it means to be Dutch (Ossewaarde, 2007;Schinkel & van Houdt, 2010). According to the WRR (2007), a Dutch scientific board that advises the government, this current understanding of integration and its link to belonging to the Dutch national identity should be revised to embrace the complexity of the process. First, considering belonging to be based solely on connection to a specific national identity limits the scope of belonging to a rather static notion of identity (e.g., identifying as Dutch; Eijberts, 2013). This is particularly problematic because the construction of Dutchness as such is exclusionary toward migrants and thus does not stimulate belonging (Davis & Nencel, 2011;Ghorashi, 2003;Ghorashi & Vieten, 2012;Vasta, 2007). Second, the national framing of belonging blinds us to the multiple ways and levels of creating and maintaining bonds with a society (WRR, 2007). One of the most imaginable identification routes for many is identification with the city they live in. Kasinitz, Mollenkopf and Waters (2002) have shown that immigrants in New York are more likely to identify as New Yorkers than as Americans. The authors argue that identification with the city is more open to diversity, whereas being American is less accessible because of its association with whiteness (Kasinitz et al., 2002). Another growing route of identification is the transnational sense of belonging that is becoming more imaginable for different diaspora groups (Ghorashi, 2016). Meer and Modood (2013, p. 309) argue that the transnational capacity of Islam, for example, provides youth with "emancipatory qualities" to create a sense of belonging and connection transcending the duality of choice between the territorial limitations of their past and present countries.

Methodology
We chose qualitative methods for this study, specifically, in-depth interviews with a biographical angle. This particular approach enabled us to understand participants' dynamic and multilayered perspectives. Respondents' narratives provided space for understanding the processes of negotiation in the past and the present, helping us to grasp the meanings that respondents attached to the central themes of this research (integration and belonging) and to get a picture of their experiences and challenges (Kvale, 1996). To facilitate the required openness, we avoided asking direct questions about "integration," as this could limit the research due to the variety of personal interpretations and narrow concepts people may have about the topic. Instead, respondents were asked to share their experiences and talk about their lives in the Netherlands. Through follow-up questions we tried to obtain specific descriptions that helped us better understand the meanings they attached to these experiences. The biographical angle, as expressed in the chronological structure of the interviews, enabled us to capture the diversity of stories about participants' choices and experiences in the Netherlands in different phases of their lives. The narratives gave us the opportunity to discover what was important to the narrators about their points of connection or disconnection to places and people in their lives, what and whom they identify with, and the meanings they attached to experiences (Bryman, 2008;Kohler-Riessman, 1993;Riley & Hawe, 2005).

The Sample
Interviews were conducted from January through June 2013. A combination of purposive sampling and snowball sampling was used. We looked at Italian migrants who have been in the Netherlands for at least 20 years to enable analysis of what changes have taken place in their identification with the Netherlands as well as their sources of belonging. We did not include more recent migrants, assuming that a certain length of stay was necessary to study belonging, which develops over time. The respondents had to meet certain conditions to be included: they had to be Italians living in Amsterdam at the time of the interview. All respondents had to be first-generation migrants who arrived in the Netherlands during either the 1960s and 1970s (first wave) or the 1980s and 1990s (second wave). This division allowed us to consider the impacts of time (age and length of stay) and context (policies) in the positioning of these two groups of Italian migrants. We were interested in the similarities and differences between the two groups. The phases covered in the interviews were the period before their migration, the first years after their migration, the present, and the future. In total, 22 interviews were conducted, 12 from the first wave and 10 from the second wave.
Interviews lasted 1-5 hours. Some of the longer interviews were conducted at two different times. To secure respondents' anonymity, we used pseudonyms. Respondents comprised 13 males and 9 females. Males were largely overrepresented in the first wave because, at that time, Italian males were more likely to come to the Netherlands (Tinnemans, 1991). Most participants in the first group came from southern Italy and had low educational backgrounds and professional histories of un-skilled or low-skilled labor (Lindo, 2000, p. 130). In the second wave, conversely, there were more women, and most of the respondents came from northern Italy and generally had higher educational backgrounds. This second group had more diverse backgrounds as well as different motivations for coming to the Netherlands. Some came for work, but most came because they had fallen in love with a Dutch partner.

Analysis
To understand the multiplicity of positions in relation to integration and belonging, we used identification as a sensitizing concept in analyzing the data. Jenkins (2004) argued that using identification as a concept provides opportunities to follow the dynamic and ongoing process of positioning while enabling researchers to locate different levels in the process. Throughout the analysis, we worked interpretatively and inductively, guided by what the respondents shared. An essential first step was to find patterns of similarity in the interviews (Kohler-Riessman, 1993). All interviews were transcribed, and a preliminary thematic analysis took place after half the interviews were conducted.
Through further analysis and by consulting theoretical and empirical literature about integration, we distinguished the two most dominant identification patterns present in the narratives, which the WRR (2007, p. 14ff) labels as functional and emotional identification. Functional identification refers to the functional aspects of participation in a new society, like having voluntary or paid work with which you identify and are identified with by other members of society. There is thus a greater focus on the individual and not so much on ethnic group membership. Functional identification entails having common goals as an institution or a society and interdependency (e.g., being invested together in improving the well-being of a society's elderly members). There can be multiple functional identifications that can aid one's social and economic integration in society (e.g., teacher and soccer player in a club). Emotional identification refers to having a "sense of belonging" (WRR, 2007, p. 16) associated with certain places and spaces. As the WRR points out and as we mentioned earlier, public discourse and policymaking in the Netherlands often emphasizes emotional identification, which is seen as a measure of ultimate loyalty-with the underlying premise that you can only emotionally identify with and feel loyalty toward one nation, thus often demanding of migrants an impossible choice. Functional identification, according to the WRR, does not receive a lot of attention, even though it can also provide an important bond with the host society (WRR, 2007, p. 14), as our findings below confirm. When using the terms functional and emotional identification, the WRR does not discern between selfidentification and how one is identified by others, using both interchangeably for ascribed and achieved identification. However, as we shall see below, this difference is crucial in uncovering the complexities and layers of Italian migrants' experiences in the Netherlands.

Sources of Functional Identification
In contrast to the emphasis on emotional identification with the nation that is found within public and policy discourses in the Netherlands (see also Duyvendak, 2011), and partly due to the nature of the first wave of migration, work experiences (but also membership in a group or committee) were mentioned as a major source of identification with the Netherlands. Dante's story is a good illustration of this. Dante came to the Netherlands in the 1960s, intending to stay for just one year. Once back in Italy, however, his "head was still in the Netherlands." He returned to the Netherlands, married a Dutch woman, and had several jobs, which he talked about extensively throughout the interview. He emphasized his efforts to be active and "do good things," such as volunteer work with elderly people. He proudly explained that he had also received appreciation for his work, illustrating how important it was for him to contribute to society through his work, to feel needed and appreciated by society.
The importance of work was also mentioned by several second-wave migrants, even though their motivations to migrate were more varied. Eva, who arrived in the early 1990s, said she migrated, not for economic reasons, but to increase her Dutch husband's chances for a good career in the Netherlands, thereby leaving her own good job behind in Italy. At first she did not feel at home in the Netherlands. However, "the more I worked, the better I felt." Working helped her reconnect with a part of herself, but also with members of Dutch society. As a teacher, she was surrounded by students who came to learn something from her and who appreciated what she taught them, which illustrates the potential power and value of functional identification-for Dutch society as well.
Many respondents talked about compliments they received at work, and how this helped them feel less foreign and more like a colleague, pointing to the importance of functional identification and work as such in being able to build bridges. Although work proved to be a strong source of identification with the Netherlands, the experiences were not all positive. This is because members of the majority group saw our respondents primarily as Italian ethnic others rather than as potentially valuable employees and full-fledged members of society. Matteo, a first-wave migrant, recounted how he had encountered discrimination at his job because of his ethnicity. This is in line with Lindo's (2000) findings that Italian migrants initially encountered prejudice from colleagues and employers, which faded when Moroccan and Turkish labor migrants arrived (see also Wessendorf, 2007). Second-wave migrants mostly experienced a "subtle" form of discrimination, as this quote from Angelo, who worked in the construction industry, demonstrates: In the beginning almost nobody spoke to me. It is just…yes, a bit set aside. You are at the table with colleagues and, uh, they talk about many things, but nobody talks to me, since they knew I did not speak the language. So you are really isolated. Just all those jokes [about me], you know...in the beginning, yes, all make jokes.
Van Laer and Janssens (2011) argue that jokes are mentioned as one the manifestations of "subtle discrimination," a form of discrimination that is less easily recognized, and which has an ambiguous character because the one who commits it often does so unintentionally. Though the act may be unconscious, the consequences are real; the "victim" may experience lower emotional and physical well-being in the context of work. In Angelo's case, while language may have formed a barrier, by making jokes, his coworkers highlighted his ethnic otherness, thereby making it harder for him to feel included.
In summary, work (paid and voluntary) was a source of functional (self-)identification with the Netherlands for Italian migrants. They wanted to contribute to Dutch society. However, there were also experiences of (subtle) discrimination at work stemming from the majority group's fixation on ethnic othering, which impeded respondents' feelings of inclusion.

Language and Functional Identification
While language in Dutch policy discourses is considered an integral first step (a part of culturation) in the supposedly linear integration trajectory (Esser, 2003), it was not always a part of the integration trajectory of Italian migrants. When the first-wave migrants arrived to find work in the Netherlands, there were no expectations or facilities for them to learn the language, which might have heightened their identification as "other" by ethnic Dutch people and thereby initially impeded some migrants' functional (self-)identification and possibly even emotional identification with the Netherlands (as we shall see later). Since conditions were better for the second-wave migrants and not everyone in the second wave had to find a job immediately (because they tended to migrate for love), one might expect that, upon arrival, they would have quickly learned Dutch. Yet, 5 out of the 10 respondents preferred to speak in Italian or English and had low Dutch fluency. For them, it seemed that an English-speaking environment was an additional obstacle to practicing Dutch, which is illustrated in Barbara's story. Barbara is a second-wave migrant who came to the Netherlands "out of love." She did not need to work immediately and started taking Dutch classes. Despite the opportunity to learn Dutch at school, she encountered difficulties due to the lack of opportunities to practice Dutch: My only frustration was when I was going around and when I wanted to speak in Dutch, people were answering in English. So I made a lot of effort to speak Dutch, and I think the people wanted to be nice and help me, but it was not a great help, because I wanted to learn Dutch. And then…I found a job where I didn't have to speak Dutch. All of my jobs in the Netherlands,…were in international firms, where people didn't speak Dutch. So I speak Dutch, but I don't feel as confident as in English, because I hardly speak it. Unfortunately.
Since Dutch language proficiency has become such an important marker of successfully following the supposed linear integration path, the past and present challenges for first-and second-wave Italian migrants to learn Dutch has likely encumbered their functional identification by Dutch colleagues and employers, as indicated in the previous section and by other researchers (Ghorashi & van Tilburg, 2006). Therefore, an important source of belonging for migrants may be difficult to obtain, and this may even impede the formation of emotional (self-) identification with the Netherlands (WRR, 2007).

Emotional Sources of Identification
When it comes to emotional identification, we see that there are different spatial levels of identification (neighborhood, city, continent, world) and that belonging can be quite complex and multilayered. Nationalityand culture-wise, most respondents entirely or predominantly identified themselves as Italian, but this did not imply that they did not feel at home in the Netherlands. An illustration of this is Roberio, a first-wave migrant who married a Dutch woman. He explained that he felt at home in both Rome, his city of origin, and Amsterdam. He is fascinated by Amsterdam and loves its "allure," with all its museums, restaurants, and theaters. However, he emphasized that he felt 100% Italian at the same time. Thus, for Roberio, emotionally identifying as Italian and feeling at home in the Netherlands-that is, in Amsterdam-are not binary oppositions.
Other first-wave migrants expressed similar feelings about Amsterdam and talked at length about all the different streets and places they knew in the city. The same connection was also found among the second-wave migrants. Eva, the only woman who had lived outside the city in the eastern Netherlands before coming to Amsterdam, stated that Amsterdam was "a different world" compared to the rest of the Netherlands. "I don't feel different at all anymore. I am a foreigner, but not different." Jano's connection to Amsterdam took on a different aspect. He came to the Netherlands in the 1980s at the age of 17. He recounted feeling more at home in Amsterdam than in Italy due to his sexual orientation: I felt safe here. [More] than in Italy, I felt more free in my spirit. Because soon I understood that Amsterdam was a gay city. And I was already convinced that I was gay. And I thought: Yes, that's it…yes… In addition to emotional identification at the city level, other levels were present as well. The neighborhood was mentioned by some as a strong source of emotional identification. Pasquale, one of the first-wave migrants, indicated he feels at home in the Jordaan neighborhood because he was received there with open arms.
At a higher/superordinate/transnational level, some identified with being European or as an international citizen. These identifications were mentioned exclusively by the second-wave migrants, which may be related to the gradually intensifying cooperation in the European Union and/or generational differences, for instance, the usage of technology and social media and easier access to affordable travel. Jano, for example, identified both with Amsterdam and as a European. He argued that, even though each country had its own characteristics, he considered all European countries to be related, based on their common history and Christianity. A more internationally oriented identification occurred among three of the second-wave respondents, one of whom identified herself as a world citizen, while the other two identified as expats instead of migrants. They mentioned that they felt less attached to one place solely, but to multiple places throughout the world. Barbara explained that she travels a lot and feels at home in three places now: the Netherlands, Rome, and New York.
In addition to places, we also identified sources of belonging connected to people. Family members served as an important source of belonging. This created tension for several migrants because they had children who were born and/or raised in the Netherlands, while their parents and other family members still lived in Italy. The lack of family support was especially missed by the women participants. However, children seemed to be a major reason for choosing to remain in the Netherlands, as participants often stated that they probably would have returned to Italy if they had had no children (for similar findings for Italian immigrants in the UK, see Zontini, 2015). They either decided to stay because their children felt at home in the Netherlands or because they considered the Netherlands a better place to raise their children.
A good example is Dori (a second-wave migrant), who feels emotionally divided between Italy and the Netherlands. She married a Dutch man, came to the Netherlands in the 1980s, and became pregnant. However, she felt homesick and returned to her family in Italy to give birth to her child. Nevertheless, she kept trying to return to the Netherlands. After six years, she finally decided to settle in the Netherlands to be with her partner and live together as a family. Her son is an adult now, and this has renewed her desire to return to Italy. She mentioned that she wants to spend her "old days" in Italy. She still identifies with being Italian and feels she will never be able to become Dutch like her husband, who was born "here." Nonetheless, she also feels at home in the Netherlands, stating: "I have two feet, one in the one part, the other in the other part," pointing to the fact that an emotional self-identification with being Italian can go hand in hand with a sense of home in the Netherlands, expressing different kinds of belonging.
Another example is Maria, a first-wave migrant, who came to the Netherlands as the wife of an Italian labor migrant and who also struggled with her emotional identification. Her children were born in the Netherlands, and her husband wanted to stay, but she longed to go back to Italy because she missed her Italian family. While working, this went "unnoticed," as she was active. "[I] was young, 21 years, and I liked the work. Run here, run there, work, everything…it was nice, the work was nice." Hence, her functional identification with her job in the Netherlands was able to override her emotional identification with Italy for a certain time. However, time went on and Maria found herself living in the Netherlands for 50 years, despite her desire to leave: You should not stay too long, if you want to leave later. After three to five years in a different country, you should try to go back. Otherwise, you don't make it anymore. Children arrive, and when children are born here, you have roots here, too…..These roots become deeper, then you do not leave anymore. I am happy here, but I am always alone.
In sum, most respondents indicated various levels of emotional identification, but when it came to the national level, most identified with Italy rather than the Netherlands. Within the Netherlands, the highest emotional identification was at the city level for first-wave migrants and a combination of the city level and/or European level for second-wave migrants. Moreover, respondents from both groups did not exclusively identify with being Dutch themselves, but their children did. This generational difference can become a source of tension, bringing the well-being and preferences of the children into conflict with those of their parents-when the parents desire to leave, but their children desire to remain (see Zontini, 2015). The following section shows how tensions in emotional identification are related to functional identification, and the significance of age.

Functional vs. Emotional Identification
One paradox we encountered in the narratives was that an emphasis on functional identification in the early stages of migration might conflict with emotional identification with Italy at a later stage. An example of this was recounted by Jano, who was retired. He had tried to enter into Dutch social life, but upon his retirement, he did not know anyone in his neighborhood or any other Dutch people, which he partly considered his own fault: So it also…was also our fault. Because we were both busy at work, full time, both of us. On holidays, we went to England and to Italy, so we didn't stay here, we didn't meet people on the street. We didn't have occasion to have a chat, and so on, and so on… For Jano, an emphasis on work limited his time to learn Dutch, and his travels to Italy did not leave room for building contacts with Dutch people. Now that he is at home more, increased contact with his neighbors has fostered a sense of belonging at the neighborhood level: "We are no longer the foreigners, the weird people. We are part of this group of people. And that is a good feeling." In addition, he started Dutch conversation classes, which diminished his hesitancy to speak Dutch.
Not everyone is able to make these sorts of investments at a later stage of the migration process. In the first stages of her "integration," Maria and her Italian husband focused on work. Now that her husband has died and she has retired, Maria admitted she is ashamed that she cannot speak Dutch fluently, despite having been in the Netherlands for 50 years. She spoke of wanting to engage in small talk with the baker, for instance, but does not do so as she feels she will not be able to express herself understandably. She really regrets not having had time to study Dutch when she was younger because of her focus on work and on raising her children. Nowadays, however, due to a lack of schooling and old age, her brain is no longer "open," as she phrases it.
Hence, the proper positioning (Esser, 2003) in the labor market did not automatically lead to more interaction, that is, social integration, as is often assumed, even though, especially in their old age, the Italian migrants realize that is what they desire. Passing the successive stages of integration is no guarantee for developing an emotional identification with the Netherlands or even the often lesser-contested sense of belonging.
The second paradox we observed was that even those who did invest in contact with Dutch people (interaction) and learned to speak Dutch fluently (culturation) experienced problems with their emotional identification at a later stage. Chiara and Marta, two second-wave women who could be considered the most successfully integrated (they have a high Dutch fluency, good jobs and housing situations, and Dutch partners), were the only ones who indicated they did not really feel at home in the Netherlands. In the initial phases of their migration, they were ambitious to become part of the Dutch society, but after approximately 20 years, they have started to withdraw from Dutch social life, speaking about their emotional struggles with living in the Netherlands while still considering Italy their "home." Chiara mentioned that, upon her arrival in the Netherlands, the first thing she did was learn Dutch. She regularly receives compliments for her level of Dutch, and she explained that she tried hard to integrate into Dutch society. However, she consciously decided to stop "integrating" at a certain point because she wanted to maintain her Italian identity: You have to choose at a certain moment-do you want to be Italian? Because you cannot be schizophrenic, it drives you crazy. And then of course, around my 35th, I chose to bring my Italian identity forward. And that, because that is what I feel more comfortable with….Now, I am always busy with something Italian, with books, with people who speak Italian. And that gives me peace, that is nice….I think, I have really always tried my best to learn Dutch, in-tegrate….I liked it, I had the energy for it…but at a certain moment then you recognize that it is exhausting, this urge to belong.
This illustrates that knowing the language, working, and seeking contact does not automatically lead to emotional identification with a new country. The paradox in the case of Chiara and Marta is that successful integration has led to higher emotional struggles with the Netherlands-which is also due to being seen as the ethnic other in daily interactions. Jano, for example, mentioned: I have always found my way in the Netherlands…the only thing which still irritates me…not so much irritating but rather annoying me, is that, when you tell people that you have been here for 32 years, you are still Italian to them. They always ask about Italy and then you think, God, it is as if I just arrived here yesterday.
Or as Marta puts is: "There is a feeling of Us and Them, you know. You are from outside, so you do not belong, these kind of things….There is always a feeling of reticence toward the other." This finding is in line with Davis and Nencel's (2011) conclusion that "meeting the implicit and explicit criteria of integration does not ensure that newcomers can be taken up into the national imaginary available for thinking about 'Dutch-ness' in the Netherlands" (p. 482). It also relates to those of Eijberts (2013) and Ghorashi (2010b), who both showed that a high level of integration could actually become a reason for a lower level of emotional identification if the expectation of being accepted as a full citizen is not met. Our data also underline other studies' findings of this socalled integration paradox (Buijs et al., 2006).

Embracing In-Betweenness
A strong line in the narratives was the simultaneous presence of Italy and the Netherlands and selective use of these two contexts to describe connections and loss, possibilities and impossibilities. Edward Said (1993, p. xxxi) refers to the condition of being both an insider and an outsider simultaneously as the condition of in-betweenness. Daniela, a second-wave migrant recounts this point: At a certain moment in time, that breaking loose, leaving everything behind you, was of course nice, and probably I needed to develop my individuality, develop myself as an individual. I did not fit anywhere, and I did not want to fit anywhere. However, at a certain point, I decided to live here, and then I thought, I recognized, that I indeed did not fit here. And that this was no longer a good feeling.
According to Daniela, the key to living as a migrant is accepting that you will never be entirely at peace, you will always have "flying roots": I accepted that my emotional house is not in the Netherlands….What is inside of me, is not here, it is there….You have to be tough enough to say: there are my emotions, here is my future and my work; you are a bit schizophrenic. But it is a nice way to be schizophrenic.
While one is active, mobile, and ambitious to build a future through hard work, this in-between position is associated with positive feelings, as several migrants regarded being a migrant as an enriching experience. However, this feeling might become less positive at an older age when the level of mobility is reduced and the resources to connect are less accessible (see Zontini, 2015). For the older generation who no longer work and whose children have left the house, the position of being inbetween has made them feel more of an outsider, accompanied by a sense of social isolation. The solution for them, as long as their health and resources let them, is to take long vacations in Italy and live the lives of pensionatos. Even then, some have discovered that, in Italy, they feel like tourists, as living abroad for so long had changed their connection to Italy as well (see Zontini, 2015).
This sense of social isolation was illustrated by Rosa, a first-wave migrant who came to the Netherlands as the wife of Bernardo, an Italian labor migrant. First, she remarked that living in the Netherlands for 50 years had made her Dutch as well. Nevertheless, later she stated: "There is always something. Since I am Italian. Do you understand what I mean? I cannot really become Dutch." Feeling different is strengthened by the lack of social contacts-Rosa indicates that lately she has "almost nobody anymore"-as well as an increasing number of unfriendly encounters. Although Rosa and Bernardo began their interview by emphasizing they never experienced problems with Dutch people, toward the end of the interview they recounted several instances of having been treated in an unfriendly manner due to their background. During a conflict with a neighbor, they were told to go back to their own country. This was an especially sore spot for Rosa, who repeatedly stated how much the comment hurt her. Bernardo said that Dutch people used to be nice, but that more recently they had become less friendly, which has been echoed by other migrant groups in previous research, as the public and political discourse on integration has become harsher in tone (e.g., Eijberts, 2013).
Unfriendly and unwelcoming remarks were often mentioned by the first-wave migrants. This was less true of the second-wave migrants, possibly due to their greater focus on functional identification or their stronger feeling of identification as European. In sum, although both waves of migrants found themselves in an in-between position, this had become a stronger source of confusion and isolation for the first wave. Rosa, for example, stated: "Sometimes I really think, what am I actually? Not Dutch and not Italian. I am a strange person." This outsider position, which some call a sense of social isolation, may increase as people age. Getting old had various consequences for this group. On different occasions, they indicated their frustration at being afraid or unable to deal with new technology that could help them to stay in touch with family and friends in Italy (see also Zontini, 2015); thus, for them it is more difficult to engage in transnational spaces, which could actually enhance integration in both countries. They do not feel as part of Dutch society and have lost connections with Italy both emotionally (because of the duration of their stay in the Netherlands) and physically (because of the lack of mobility they enjoyed when younger). Physical distance, combined with a lack of knowledge about new technology that could help bridge this distance virtually, has added to their sense of isolation. In this way, first-wave migrants feel both out of place and behind the times because of their age (Zontini, 2015).

Discussion
As demonstrated above, the narratives show paradoxical processes at work when Italian migrants talk about their lives in the Netherlands. Many had a very strongly developed functional (self-)identification in the Netherlands and wanted to contribute to society (especially in the earlier stages of their life and stay in the Netherlands). They wanted to be "needed," as one respondent said, and be a valuable contributor to the welfare of their organization and society. In this sense, they can clearly be seen as an asset to Dutch society. However, their colleagues did not always recognize their efforts and approached them primarily as ethnic others, which led to their feeling excluded and may have called their self-identification into question. According to the WRR (2007), functional identification can be one of the most powerful means to fostering social cohesion and integration. However, overemphasis on migrants as ethnic others may hamper its effects.
Furthermore, at the level of functional identification, we heard narratives that challenged the assumption that language, as part of culturation, serves as a critical bridge to integration, a dominant assumption in the literature and policy discourse of integration in the Netherlands (Entzinger & Biezeveld, 2003;Esser, 2006). We refer to this contrast as the paradox of language. In the past, policymakers saw no need to focus on language because the migration approach of that time focused on the temporary stay of guest workers (Entzinger, 2009). However, migrants may realize later in life that their lack of language knowledge means they were unable to invest in a social life in the Netherlands outside of work. Thus, while their lack of culturation in terms of language did not result in a lack of positioning, positioning in turn did not automatically lead to more interaction, challenging the assumed linear process of one following the other. However, in old age, social interaction with their neighbors was exactly what migrants desired, and the shift in importance from functional to emotional identification caused them to feel a longing for their country of origin.
However, with the second-wave migrants especially, it becomes clear that culturation in terms of language does not open all doors either. This group of Italian migrants enjoyed higher education, and most of them worked in international environments. Some identified as expats, which could be considered a way of distancing themselves from the dominant image of the guest worker who is low educated and traditional. Though the label "guest worker" does not fit the second group of participants (from the 1980s and 1990s), they are often associated with it. Because of the "negative" image of this label, there is a clear pattern of de-identification by the second group, emphasizing their difference based on education (being more educated than the average Italian who came in that period), salary (gaining more salary than the average Italian in that period), and choice (choosing to migrate rather than needing to). In this way, they are maintaining the image of mobile migrants or Eurostars (Favell, 2008), who are from higher classes compared to the earlier group of Italian migrants. Even though all of the second-wave migrants were motivated to learn Dutch, half of them had not succeeded in learning enough to feel comfortable speaking it. The main causes for this were having a schedule that prevented the combination of working and learning Dutch, and being in a Dutch-speaking environment with Dutch people who switched to English.
However, even those who received compliments for their Dutch (culturation) and who had successfully socioeconomically integrated (positioning and interaction) did not develop the emotional identification with the Netherlands predicted by Esser (2003) and the Dutch government. This might be partly due to the fact that these kinds of mobile migrants often combine their sense of mobility with a functional integration, as Favell (2008) has argued. Another argument could be that Amsterdam has been less permissive and open than it presents itself. In an ethnographic comparison between different European cities, Favell (2008) showed that Amsterdam is a deeply regulated and controlled city in which some processes of social and ethnic closure are ingrained in daily interactions. Thus, the linear assumption underlying Dutch integration policies does not even apply to European migrants in the Netherlands, who are considered culturally close. More importantly, the "integration paradox" applies to those who are most integrated, who have successfully passed all stages, in contrast to those who are less integrated and do not even feel at home.
Even more, in this study, we observed a rather different pattern-a clash between functional (e.g., economic) and emotional sources of identification. In the case of first-wave migrants, we showed that the exclusive emphasis on work in the first stages of migration could become a hindrance to emotional identification at a later stage. Their in-between position became a source of social isolation later in life. For the second-wave migrants, we found that those who had succeeded in learning fluent Dutch, who worked, and could be considered successfully integrated, as defined by Esser's (2003) first three dimensions of integration, indicated that they felt neither Dutch nor at home in the Netherlands. Knowledge of the Dutch language and active participation in the workforce led to exposure to the exclusionary aspects of not being Dutch, which had a negative impact on their sense of belonging (see also Eijberts & Ghorashi, 2016).
In addition to the paradoxes of language and integration, we also identified layered sources of emotional identifications. Identification as "being Dutch" has become an important measure of integration in the Netherlands (WRR, 2007). As we stated earlier, several studies have shown that the construction of Dutchness is exclusive of difference, which does not provide space for a sense of emotional belonging for migrants (Davis & Nencel, 2011;Ghorashi & Vieten, 2012;Vasta, 2007). This is confirmed by our research: migrants in our study did not (or could not) fully identify with a Dutch identity but instead predominantly identified as being Italian. However, their sense of belonging was certainly much broader than identification with a national identity. The most prominent source of emotional identification was at the city level. The multicultural character and openness of Amsterdam proved an important basis for identification. This matches Kasinitz et al.'s (2002) results of immigrants identifying as New Yorkers rather than as Americans. In addition to the city level, we also found emotional identifications with particular neighborhoods and at the European level. Notably, European identification was only present in the narratives of the second-wave migrants. Their identification as expats along with intensified European cooperation and attempts to construct a European identity may have contributed to this level of identification. Identifying as European gave them an opportunity to connect their worlds (the Netherlands and Italy) and choose the best of both worlds when relating to the condition of in-betweenness. In contrast, we found that the first-wave migrants actually faced the downside of what Said (1993) refers to as the in-between position. Their focus on work in the earlier years of their migration, their limited contacts with Dutch people combined with social contacts mainly in Italy and with Italians in the Netherlands seem to have become a source of isolation in their old age (feeling out of place and behind the times).
In addition to these different (either overlapping or conflicting) layers of belonging to places, we identified family as an important core of belonging in both Italy and the Netherlands. For both migrant groups, we saw some contradictions at play, such as between the place one lives in (the Netherlands) and the people with whom one identifies (Italian family members in Italy) or the place where one's children live (the Netherlands) and the place one wants to be (Italy).

Conclusions
This article aimed to demonstrate the complexity of the integration process and to broaden the scope of existing studies on migrants living in the Netherlands by including the experiences of Italian migrants. To understand Italian migrants' routes to identification in the Netherlands, we distinguished between functional and emotional identification (WRR, 2007). The narratives showed that the experiences of the participants were complex, nonlinear, multilayered and did not fit a linear definition of "integration." The narratives also showed that different forms of identification overlapped and clashed over time.
All respondents in our study had to deal with a new language, with finding their way in the Netherlands, and with re-figuring the question of who they are. A similarity in all the narratives, despite differences in age and migration background, was the in-between position of these migrants in Dutch society. The first-wave migrants in our study have a less positive narrative about this inbetween position due to their age and isolation in Dutch society. They miss social contacts with their families in Italy, but when they are in Italy, they feel less at home, as over time they have come to feel like an outsider there as well. Yet, we saw that the second-wave migrants were better able to cope with this in-between position and profit from its positive aspects. As Eva put it: "…you have the ability…to look at both parts, from a distance. You can be more critical. That is fun. That is fun, yes." The secondwave migrants are part of many places at the same time, are connected virtually to the world (but also to their friends and families in Italy), and have the resources to be mobile. Identifying as European, for many of them, is a way to make sense of this mobility and the multiple sources of connection. As Antonsich (2012, p. 7) argues: "Europe answers the needs of the modern individual to travel, to communicate, to exchange information and experiences. This clearly resonates with the image of the mobile European citizen put forward by Verstraete (2010)." Nevertheless, we agree with Favell (2008) that this notion of mobility should not be overemphasized, because the interactions of these mobile migrants are situated within specific national settings that also include exclusionary practices, as we discussed above.
This study has also shown that the struggle for integration and belonging in the Netherlands is not limited to migrants with Islamic backgrounds. There are many similarities in the paradoxes (of language and socioeconomic integration) present in the narratives of this research with those found in other studies on various generations of migrants and refugees with Islamic backgrounds (e.g., Eijberts & Ghorashi, 2016). We hope, with the results of this study, to weaken the "culturalist" foundation of the present discourse on integration. Moreover, we have provided a more comprehensive insight into the various ways identification takes place, beyond the singular source of national identification.