Reconceptualising ‘integration as a two-way process’

Abstract

In this article, I advocate the reconceptualisation of ‘integration as a two-way process’. I argue that integration is, fundamentally, an issue of relational inequality, and conceptualising it as a one-way process constitutes problems of undesirability and infeasibility. I show the theoretical hiatus which characterises many dominant approaches to the two-way process, which leads scholars to build their work on internal contradictions and to implicitly (and often unintentionally) feed into a one-way integration discourse. I argue that as long as conceptualisations of integration as a two-way process reinforce a boundary between ‘people who integrate’ and ‘people who do not integrate’, they are unfit to avoid the problems of one-wayness which they intended to overcome in the first place. In the last part of the article, I put forward some initial building blocks for a new theoretical framework of ‘integration as a two-way process’ which is more attentive to the relational social processes that constitute integration.

1. Introduction

In the field of integration research, we find a (tacit) consensus that integration is a two-way process. Numerous scholars explicitly endorse this notion by arguing, roughly, that the attainment of integration does not only depend on the commitments, efforts, and achievements of immigrants and their offspring but also on the structure and openness of the receiving society (see e.g. Castles et al. 2002; Modood 2004; Lucassen 2005: 19; Ager and Strang 2008: 177; Phillimore 2012; Waters and Pineau 2015; Korteweg 2017).¹ In line with this, several scholars criticise integration policies, discourses, and fellow researchers for being too ‘one-way’: disproportionally emphasising the responsibility of ‘people with a migration background’ and underestimating the role of the receiving society (see e.g. Bommes and Thränhardt 2010; Penninx 2010; Anthias 2013; Coello et al. 2014; Schinkel 2018).

Yet the concept of two-way integration itself has been insufficiently theorised. Although the notion seems to be generally accepted by most scholars, it is not adhered to in theoretical and empirical practice (i.e. the way integration is theorised, as well as operationalised, in order to conduct empirical research), making it more of a platitude than a consistent theoretical starting point. This problem shows in many dominant approaches, leading scholars to build their work on internal contradictions, to talk past each other, and to implicitly add to or reinforce a one-way integration discourse. Drawing mainly from relational sociology, boundary theory, and relational equality theory, I aim to start filling this theoretical gap. Negatively, I argue that a relational understanding of integration is a promising approach to avoid or solve the problems of ‘one-wayness’. Positively, I address the question of what a framework of relational integration would add to our current analytical tools for researching and understanding integration.

I start this article by providing a definition of integration as an end state, which is to function as an evaluative standard for integration research: A society in which there are no social boundaries between ‘legitimate members’, or insiders, and ‘non-legitimate members’, or outsiders. I thereby argue that integration is primarily a matter of social standing, and not legal or socioeconomic status. In Section 3, I address the problems of undesirability and infeasibility that arise from one-way conceptualisations of integration. In addition to showing the relevance of conceptualising integration as a two-way process, this section also provides a first indication of which problems a theoretical framework of integration as a two-way process should avoid. In Section 4, I show that most conceptualisations of the two-way process in current integration literature do not suffice to avoid or solve the problems of one-wayness as discussed in Section 3. Based on an extensive literature review, I make a distinction between three general understandings of integration as a two-way process: (a) insiders are affected by the integration of outsiders; (b) insiders can influence the integration of outsiders; and (c) insiders and outsiders integrate with each other. Most researchers start from the first or second understanding, which, I argue, fail to avoid the pitfalls of one-wayness. A reconceptualisation of the ‘two-way process’ is therefore necessary. Finally, in Section 5, I offer a proposal for developing a theoretical framework that starts from the third understanding of integration as a two-way process: a process in which both parties take part. This, I argue, requires an ontological shift in which processes and power relations become the central unit of analysis; a methodological shift in which different types of questions take centre stage and a corresponding shift in language. Here, I lay out the foundations for the development of new measures and proxies that can be studied in integration research and used in policy design.

This article starts from the assumption that integration is desirable. The fact that a fully integrated society may be utopian does not delegitimise the goal in itself. This might be perceived as a controversial statement: through the politicisation of the concept, integration has obtained culturally imperialist, or sometimes even xenophobic, implications (Bowskill, Lyons and Coyle 2007). This has led several scholars to plea for the abolishment of the concept altogether (e.g. Rytter 2018; Schinkel 2018). Yet social boundaries between ‘legitimate members’ and ‘non-legitimate members’ are a political and social reality and not using the term would not change the relational inequality and other injustices stemming from them. Following Anderson (2010: 2), I see integration as an imperative of justice: as a matter relational equality, integration promotes more equal relations between people. It is concerned with social standing, and hence with the equal moral worth of people. To discard the concept of integration might be a firm statement against the problems of existing integration discourses and research but it does not solve the issues of relational inequality that are upheld by these social boundaries.

Addressing the conceptualisation of two-way integration is not merely of analytical importance. As several scholars have shown, there is a tight connection between the field of integration research on the one hand, and integration policy and discourse on the other (Bommes and Thränhardt 2010; Dahinden 2016). Scholten et al. (2015: 330–4) show how integration research increasingly influences both political and social discourse via symbolic knowledge utilisation: integration research is used indirectly to substantiate policy choices or to legitimise political actors. In addition, they point to the expanding role of the media as an intermediary between integration scholars and policy-making, thereby affecting the public discourse on integration. Integration research is thus involved in shaping and legitimising immigration and integration policies and discourses and can thereby, at least indirectly, influence the course of integration processes. It has, in other words, non-epistemic impact (see Douglas 2000). The reconceptualisation of integration as a two-way process, such that we really try to avoid the pitfalls of ‘one-wayness’, is also of normative and policy concern.

2. Integration as an end state

As indicated in the Introduction, I define an ideal-type integrated society as a society without any social boundaries between legitimate and non-legitimate members. Before explaining this definition, I have a short remark on the necessity for such an end-state definition. Any statement on how integration ‘works’—any claim on the conditions of successful integration or any measurement of integration—is based on such an understanding of integration as an end state, or ‘completed integration’.² If we want to be able to make such statements, we need an idea of what ‘more’ or ‘less’ integration looks like, and what a successful integration process should, ideally, be leading towards. Making this end state explicit is crucial since there is so much conceptual confusion in the field: concepts like integration, assimilation, incorporation, acculturation, and multiculturalism are used differently throughout various theoretical traditions, research contexts, and academic disciplines. Being explicit about one’s definition is therefore essential to avoid misunderstandings.

My proposed definition consists of two main elements that need explanation. First of all, it engages with literature on boundary theory. Social boundaries are social distinctions that people make which constitute their attitudes and behaviour towards others (Alba et al. 2012). They have a categorical dimension (how we categorise ‘us’ and ‘them’), and a behavioural dimension (how we behave towards ‘us’ and towards ‘them’) (Wimmer 2013). In the operation of a social boundary, we find specific relations between people on either side of the boundary; specific relations between people across the boundary; and some shared representations of the boundary itself (Tilly 2005: 134). The more consensus there is on social boundaries, the more they are revealed in (relatively) stable patterns of behaviour: They become ‘objectified forms of social differences manifested in unequal distribution of resources (material and nonmaterial) and social opportunities’ (Lamont and Molnár 2002: 168). Social boundaries can be politically institutionalised, but this is not a necessary condition (see Ozgen 2015).

Second, in the context of an integrated national society—the level of analysis for most integration research—there is no social distinction between ‘real citizens’, and people whose citizenship is questioned (see also Schinkel 2010). This distinction between legitimate and non-legitimate members pertains to the issue of social standing: legitimate citizens enjoy higher social standing, which gives them a superior power position within a society. This position provides them better access to all forms of capital (economic, political, social, cultural, and symbolic), and gives them the power to deny such access to non-legitimate citizens through mechanisms of social closure and stigmatisation (see also the distinction between the Established and the Outsiders by Elias and Scotson (1994)).

Note that ‘legitimacy’ here does not refer to a legal status. The legitimacy of some people with formal citizenship may be questioned more than that of others without this legal status. Yet, at the same time, the two do mutually influence each other. First, formal citizenship provides (mostly political) rights and duties of membership that one does not enjoy without this legal status. Thereby, legal status does play an important role in integration processes. On the other hand, processes of boundary-making between legitimate and non-legitimate (potential) members do not just start after immigration: the ways in which border regimes select people who are eligible for acquiring citizenship are influenced by ideas of what constitutes a legitimate member. The ‘legal’ and ‘legitimacy’ aspects of citizenship thereby mutually reinforce one another.

The distinction between legitimate and non-legitimate members can be based on many different characteristics, and integration as a concept is not necessarily limited to contexts of migration. In the American context, integration is mostly referred to as an issue of racial integration (see e.g. Stanley 2017; Freixas and Abott 2019), and one can also think of homeless people or felons as categories designated as non-legitimate citizens. This means that scholars when using the concept of integration must be specific about what kind of integration they are investigating, and what are the characteristics of the boundaries at play. In the context of European integration, I argue, it is the conception of ‘foreignness’ that constitutes the most important social boundary. ‘Non-legitimate members’ are, in one way or another, regarded as ‘too foreign’ to enjoy the standing of a ‘real citizen’. This notion is not just a referral to someone’s migration status or ethnic background but intersects with race, gender, class, and religion (see e.g. Bonjour and Duyvendak 2018; Rytter 2018; Hadj Abdou 2019). Some immigrants are therefore seen as more ‘foreign’, and hence less legitimate citizens than others.

Furthermore, one should keep in mind that this definition is not a road map for the realisation of integration, nor does it provide any specific guidelines about who should adjust to whom. As such, it does not take a stance in the normative debates between acculturationists on the one hand and multiculturalists on the other, or between difference-blind and difference-sensitive approaches. The assumption that integration is a process of boundary change is compatible with either (or a combination of) the acknowledgement of group rights or individual cultural adaptation (see also Anderson 2010). Consider, for example, Hartman and Gerteis’ (2005) taxonomy of approaches to integration. They distinguish assimilationism, cosmopolitanism, interactive pluralism, and fragmented pluralism as four different normative frameworks for an integrated, multicultural society. The four frameworks disagree on whether integration should be attained through substantive moral bonds or through procedural norms, through individuals or through groups. Still, for all four frameworks, the outcome of an integrated society would minimally be a society in which there is no social boundary between legitimate and non-legitimate legal members. This definition should therefore appeal to researchers from a broad variety of theoretical and empirical approaches. I get back to this when discussing the feasibility problems of one-way approaches.

Three additional comments need to be made here. First, whereas most integration research uses the terms ‘migrants’, ‘immigrants’, or ‘people with a migration background’ on the one hand, and ‘natives’, ‘non-(im)migrants’, and ‘receiving society’ on the other, I prefer to stick to the more general terms of ‘insiders’ and ‘outsiders’ as much as possible. This is because they are less vulnerable to the stigmatising consequences of political debates around migration, and they also provide more accurate terms when thinking about integration in general. Whereas in many cases the word ‘immigrant’, or one of its alternatives, can be used as a synonym for outsider and the ‘non-immigrant’ for insider, this is not necessarily the case (see e.g. Crul and Schneider 2010). I therefore only use ‘(non-)immigrants’ when discussing other people’s work that explicitly uses these terms.

Secondly, I do not follow a definition of integration in terms of socioeconomic equality. The underlying assumption of such approaches is that integration is attained when outsiders are, on average, equal to insiders on indicators such as employment or education levels. Insiders are thus by definition the benchmark which outsiders need to live up (cf. Carens 2005: 42). Integration hereby becomes a phenomenon which can be objectively measured by socioeconomic indictors. In contrast, I argue that distributive equality does not necessarily indicate integration as a situation in which the social boundaries between insiders and outsiders are overcome. Because of redistribution mechanisms, distributive equality can coexist with relational inequality (Anderson 1999; Schemmel 2012). One can imagine a situation of economic prosperity in which outsiders reach the same level of employment as insiders. Yet, as long as social boundaries between legitimate and non-legitimate citizens are maintained, an economic or political crisis may change this equal position: they may be the first to lose their jobs or suffer all kinds of social harms. Integration therefore cannot be measured in socioeconomic terms alone.

Thirdly, my proposed definition does not imply that integration, as an end state, is incompatible with any social boundaries. People can still identify with many different, local or transnational, mutually exclusive groups which are separated by social boundaries. The identification with these groups can be much stronger than the identification with their fellow national citizens. Yet, in an integrated society, these social boundaries do not affect the social standing of citizens as legitimate members of the national society. As long as diversity is not marked as deviancy, as lacking certain qualities, which make someone a legitimate member, it does not raise problems for attaining integration as an end state. Consequently, the definition does not infer that a fully integrated society is a homogeneous society.

3. The problems of conceptualising integration as a one-way process

In a one-way conceptualisation of integration, only outsiders take part in the process and it is primarily their responsibility to make it work. In this section, I shortly outline the problems of conceptualising integration this way to demonstrate the relevance of conceptualising integration as a two-way process in general and provide an indication of which problems a theoretical framework of integration as a two-way process should avoid. I distinguish two types of problems: problems of desirability and problems of infeasibility.

3.1 Problems of undesirability

In existing critical literature, we find three different undesirable implications of conceptualising integration as a one-way process. First, it presupposes an underlying picture of a society composed of domestic individuals and groups (‘non-immigrants'), who are already ‘integrated’ (Joppke and Morawska 2003: 3). Such a society, however, has never existed and portraying it in such a way restricts us in dealing with the increased fluidity and super-diversity of modern society (see also Grzymala-Kazlowska and Phillimore 2018). One-way conceptualisations constitute ‘an erasure, or minimisation of the significance of past social cleavages and conflicts’ (Astor 2012: 344). Hence, it makes no sense to use this image as a starting point for academic research.

Secondly, if the goal of integration is for outsiders to become part of an already integrated society, then ‘society’ takes on a prescriptive character. Rather than a description of an assemblage of social life in all its diversity, society becomes a concept which prescribes a dominant norm of what the social life should be (Schinkel 2018). Integration is then a process for those who lack certain qualities that are needed to adhere to the norm of ‘being a member of society’, a process for people who are somehow deviant (Kostakopoulou 2010). Such assumptions of what is ‘normal’ in society have very undesirable stigmatising and marginalising consequences (see also Haslanger 2014).

Thirdly, if integration is a one-way process that only concerns outsiders, then the very concept itself creates a division between two kinds of people: people who either need to integrate or have once gone through an integration process (outsiders; immigrants; and people with a migration background), and people for whom integration has never been an issue (insiders; natives; and potentially some white immigrants). This is what Schinkel (2018) has called the ‘dispensation of integration’ for non-immigrants. To be ‘well-integrated’ (to have overcome the deficiencies of being an outsider) becomes the highest possible achievement for an outsider (Schinkel 2013). Thereby, even ‘well-integrated’ outsiders are still not truly insiders, since if they were, then they would not be ‘well-integrated’: integration would not be an issue at all (see also Horner and Weber 2011).

3.2 Problems of infeasibility

A conceptualisation of integration as a one-way process is not only undesirable but also infeasible, and therefore inept as an analytical conceptualisation. Social change is feasible if there is an action (or set of actions) available to an actor which could bring it about (Gilabert and Lawford-Smith 2012). Integration as a one-way process is infeasible because unilateral boundary change is impossible. Boundaries are operationalised through ‘extensive social organization, belief, and enforcement’ (Tilly 2005: 72). Boundary change thus requires extensive change in social relations. On a micro-level, boundary change involves a change in (1) the way people categorize each other (i.e. as fellow legitimate members of society), (2) the way people behave towards each other (i.e. as fellow legitimate members of society), and (3) a change in self-conceptualisation, which can no longer be based on opposition to the Other (although it may lead to a self-conceptualisation based on the opposition of a new Other). This change should take place at both conscious and unconscious levels as the literature on implicit bias teaches us (e.g. Greenwald and Krieger 2006).

Whether one understands integration as a process of individual acculturation, or of acquiring group rights, integration as a one-way process is infeasible because boundary change is logically unattainable without change on the side of insiders: there is no action available to outsiders that could bring about this acknowledgement unilaterally. There might be things they can do to allure the acknowledgment of insiders (ranging from adaptation to revolution) but it will never enable them to perform this acknowledgement themselves. As long as this acknowledgement is absent, outsiders can adjust as much as they like—they can change their norms and values, their names, or even their physical appearance—but they will always remain a separate category of people: those who are not legitimate members of the national society.

4. General understandings of integration as a two-way process

The concept of two-way integration, however, remains under-defined. Since many authors do not give any explicit definition at all, we often need to read between the lines to explore how the concept is understood and used. In this section, I show that endorsing a general two-way notion of integration is not enough to avoid or solve the problems of one-wayness: it matters how one conceptualises it. Based on an elaborate literature review, I map the different understandings of two-way integration into three general categories: (1) Insiders are affected by the integration of outsiders; (2) insiders can influence the integration of outsiders; and (3) insiders and outsiders integrate with each other.³ I argue that only the latter can help us move away from one-way integration discourses.

The clustering of authors within these categories may, in some cases, be surprising. Particularly the second category brings together scholars from very different approaches towards integration research. When reading it, one should keep in mind that this categorisation is based only on the conceptual and causal role(s) attributed to insiders in integration processes, and focuses primarily on the implications of the authors’ work, not their intentions.

4.1 Insiders are affected by the integration of outsiders

The first understanding of two-way integration concerns the effects of integration. Societies change as a consequence of migration and, therefore, integration affects both outsiders and insiders:

[Integration] is a two-way process. The host society does not remain unaffected. The size and composition of the population change, and new institutional arrangements come into existence to accommodate immigrants’ political, social, and cultural needs (Penninx and Garcés-Mascareñas 2016: 11).

The focus here is on change in the receiving society as a consequence of the integration of immigrants. Although this change is seen as an inevitable part of the integration process, it is comprehended as the output of, rather than input for, integration. Hence, in this understanding of the ‘two-way process’, both the locus of and responsibility for integration are exclusively situated with outsiders. Consequently, it faces the same problems of undesirability and infeasibility as depicted in Section 3.

4.2 Insiders can influence the integration of outsiders

This is by far the most dominant use of two-way integration in the scholarly literature, with some scholars being more sensitive to avoiding issues of one-wayness than others. To give some more insight into how these varying approaches feed into the problems of one-wayness in similar ways—albeit to different degrees—I discuss four frameworks in more depth at the end of this section. These examples not only show the persistence of the problem of one-wayness in current integration literature but also offer some interesting suggestions for what a new theoretical framework should adhere to.

What unites the scholars in this cluster is that integration is seen as a process of mutual exchange between immigrants and their migration context. This migration context is investigated both for its institutions and the behaviour of its majority population: research is conducted on the effects of structural characteristics of the host society (e.g. education systems, welfare systems, national philosophies of citizenship) and on the effects of discrimination and stereotyping. In contrast to the first understanding, change in the receiving society is not only seen as an output of the integration process but also as a potential input. As such, integration problems are partly explained by the attitudes or institutions of the migration context, and this context is also, in varying degrees, held responsible for the facilitation of immigrant integration.

Nevertheless, this conceptualisation is limited by the fact that insiders and their institutional context are always perceived as exactly that: a context. They are, to use more causal terminology, perceived as a potential independent variable. Their attitudes and behaviour are attributed instrumental value for integration: they can facilitate the integration of others—for example by banning discrimination or by adjusting institutions towards more inclusive ones—but are never scrutinised for their integration. Being part of a context, they stand outside the process.

One of the implications of this approach is a tacit differentiation between ‘outsiders as active agents’ and ‘insiders as passive entities’, such as, for example, in Esser’s (2004) framework of ‘intergenerational integration’ which exclusively focuses on the different choices and strategies available to migrants within an integration process (1135). Apparently, the choices and strategies of non-migrants are not of interest for the framework of intergenerational integration. This distinction is discernible, in more or less tacit versions, in many different research endeavours. Migrants are portrayed as agents who integrate through ‘purposive action’ (Alba and Nee 2003: 14). They enhance their human capital (Kogan 2011) and levels of education (Kristen and Granato 2007); adapt their cultural values in order to diminish cultural distance (Koopmans 2016); develop relations with insiders to enhance their bridging social capital (Tselios et al. 2015) and an integrated identification with the host country (Verkuyten and Martinovic 2012), etc. By contrast, non-immigrants are not regarded as integrating agents in these approaches. They are addressed as part of an institutional structural context (the context which outsiders need to become an accepted part of).

Hence, integration is still portrayed as a one-sided affair: outsiders need to integrate with insiders, not the other way around. If outsiders do not identify with insiders, outsiders have an integration problem. If insiders do not identify with outsiders, it is still the outsiders with an integration problem. Again, this is not to say that these research endeavours ignore the receiving society altogether, or do not hold the receiving society responsible for successful integration. Penninx and Garcés-Mascareñas (2016), for example, explicitly state that the receiving society is far more decisive for the outcomes of integration processes than immigrants because of power differentials (17). But even if scholars plea for measuring the attitudes and behaviour of individuals towards newcomers, integration is still conceptualised as a process that only outsiders can take part in.

Consequently, the issues of one-way integration as discussed in the previous section are not solved. We still find a tacit assumption that the non-immigrant does not need integration, and non-immigrants are still used as the benchmark for measuring integration. Furthermore, the distinction between ‘people in need of integration’ and ‘people not in need of integration’ is maintained. Nor is the infeasibility problem avoided: as long as insiders are merely perceived as an independent variable, the relational character of integration is not fully appreciated.

A good example is provided by Portes and Rumbaut’s (1990, 2001) theory on modes of incorporation. Whereas they acknowledge the potential effects of government policy and discrimination on integration, it is the immigrant’s responsibility to deal with this context. Take, for example, their approach towards labour discrimination: ‘The main difference [between immigrants] lies in the ability of different types of immigrants to neutralise labour discrimination’ (Portes and Rumbaut 1990: 87). Neutralising labour discrimination, seemingly, requires action from outsiders, rather than insiders. In later work, Portes and Rumbaut (2001: 55) state that:

Once second-generation youths [of European descent] learned unaccented English, adopted American patterns of behaviour and dress, and climbed a few rungs in the social ladder, they became by and large indistinguishable from the rest of the population.

Integration is portrayed as something that just happens once immigrants have sufficiently adapted. Changing the context of incorporation into a more favourable one is primarily the immigrants’ responsibility.

A second example is provided by Crul and Schneider’s (2010) ‘Integration Context Theory’, in which they emphasise the importance of context for the success of integration. Importantly, their theory is less migrant-centred. As a consequence of emerging ‘super-diversity’, they argue that non-migrants need to integrate too: ‘The idea of assimilation or integration becomes at any rate more complex in a situation where there is no longer a clear majority group in which one is to integrate or assimilate’ (see also Crul 2016: 57). This incorporation of literature on super-diversity opens up the possibility that not only immigrants but also non-immigrants need to integrate. In a way, their claim is similar to Bommes’ (2012) statement that every individual needs to assimilate to a dominant structural order to competently participate in society (113). Nevertheless, this does not make these approaches fully ‘two-way’: integration remains a process that is done by outsiders, be they immigrants or non-immigrants. In both instances, insiders and their institutions still only provide a context for integration processes: they do not actively participate in it themselves.

Another interesting framework is that developed by Richard Alba and his colleagues, who actually do incorporate some ideas of a more relational understanding of integration. Alba (2009) endorses the idea that assimilation and integration are processes of boundary change in which ethnic distinction declines and ‘the mainstream’ expands. Alba and Nee (2003) use Shibutani and Kwan’s (1965) theory of ethnic stratification, which focuses on ethnic differences as a social construction rather than an objective reality. In doing so, they emphasise that assimilation is to be understood primarily as a change in subjective status and the reduction of social difference, rather than a change in ‘objective’ socioeconomic status and the reduction of cultural difference. In the more historical parts of his work, Alba (2009) investigates the conditions that led to ‘boundary relaxation’ in post-war America such that white ethnic immigrants could assimilate in large numbers (63). Hence, in researching the conditions for integration, Alba (2009) shifts focus from the change that migrants go through towards the change within the receiving society. In later work, he also acknowledges the importance of recognition as ‘an insider’ as a central element of the integration process, and points to the reciprocal character of bridging social capital (Alba and Foner 2015).

However, Alba (2009) does not explicitly take two-way integration as a basic assumption of his work, which makes him inconsistent at times. In Alba’s integration theory, too, there is an underlying assumption that integration is a process for immigrants, which the mainstream can influence, but cannot take part in. Thereby, at times, his (Alba and Nee 2003: 14) work also reinforces the distinction between immigrants as agents and non-immigrants as part of a structural context: ‘[O]ne key to understanding trajectories of incorporation lies in the interplay between the purposive action of immigrants and their descendants and the contexts—that is institutional structures, cultural beliefs, and social networks, that shape it’ (emphasis added). Another example is when Alba and Nee (2003: 14) state that the mainstream can be understood as ‘structures of opportunity offering powerful incentives that make assimilation rewarding for many immigrants and their descendants’. The implication here is that the mainstream itself does not assimilate.

A fourth body of literature is the research on ‘white flight’ and ‘white avoidance’. These concepts originally stem from research in the USA and refer to the inclination of white people to move away from, or avoid, neighbourhoods and schools with a high percentage of African-Americans or other ethnic minorities. White avoidance is found to be the most persistent mechanism in both the USA and European context for sustaining segregated neighbourhoods and schools (Quillian 2002; Bråmå 2006; Andersen 2017). In contrast to the most research on integration, this body of literature does actually take into account the behaviour of insiders, as ‘private choices have public consequences’ (Saporito 2003). Yet, again, it focuses on the effect of spatial segregation on the integration of migrants (see e.g. the special issue by Bolt, Özüekren and Phillips (2010))—the integration of insiders is not an issue.

In rejecting this second conceptualisation of two-way integration, my point is not that observations about the positions of migrants, such as ‘less cultural distance correlates with better labor market positions’ (Koopmans 2016), or that ‘host societies provide powerful incentives to make assimilation rewarding’ (Alba and Nee 2003), are empirically unsound. Nor do I claim that the identification of migrants with a host society is of no importance at all when it comes to investigating integration. Instead, I want to emphasise the following: as long as integration is conceptualised as a process in which only outsiders can take part, this is the only side of the social relation that will be observed. Doing so, one can only engage in the one-sided effects of boundary (un)making processes rather than investigating these processes themselves. Conceptualising integration as a process in which only outsiders take part, thus, constitutes a serious analytical problem: it prevents us from observing and understanding this relational nature.

4.3 Insiders and outsiders integrate with each other

The last option is to conceive of integration as a relational process in which insiders and outsiders integrate with each other. This entails that insiders do not just influence the process but instead are an intrinsic part of it. Ending discrimination or other practices of social closure is not perceived as instrumental to the integration of outsiders: it is integration in itself. In this approach, insiders are integrating too, which means that the ‘integration dispensation’, as described by Schinkel (2018), is lifted: the distinction between ‘people who need to integrate’ and ‘people for whom integration is not an issue’ does not hold. As such, it is the only two-way conceptualisation thus far that can avoid the one-way problems of undesirability and infeasibility. From this approach, it does not make sense to talk about outsiders integrating into a society. The approach thereby reduces the risk of both depicting ‘society’ as a harmonious whole that is already integrated, and as a prescriptive norm to which outsiders need to live up to. Consequently, there is less of a risk of portraying ‘insiders’ as the norm, and ‘outsiders’ as somehow defective.

The merits of a relational approach to integration can be illustrated by the example of investigating political integration: the majority of the literature on political integration focuses on the conditions under which immigrants participate and feel engaged with the polity (e.g. Bloemraad and Vermeulen 2014; Morales and Giugni 2016). While this research generates knowledge about the extent to which people with a migration background participate politically, it rarely takes into account how the receiving society reacts to this participation. It is often just assumed that more political participation automatically indicates more integration. However, as I show in an article with Marjukka Weide (Klarenbeek and Weide, 2019) the participation of immigrants may be not unconditionally interpreted as a positive development. Participation by immigrants may even be perceived as a threat and trigger practices of social closure and boundary maintenance (see also Verkuyten 2018). An example here would be the electoral accomplishments of the party ‘Denk’, at the Dutch general elections in 2017, and the local elections in 2018. ‘Denk’, founded by immigrants, has mostly mobilised voters with a Turkish or Moroccan background, including people who used to refrain from voting (Vermeulen 2018). The party thereby fosters the participation of people with a migration background in the Dutch polity. Yet, its electoral success has been a reason for many politicians and media to question their supporters’ ability to integrate. Thus, what matters for political integration is not just participation but the social standing as a legitimate participant. Without this social standing, participation may actually be met by practices of Othering. The concept of relational integration enables us to capture this dynamic.

This relational approach is by far the least common approach towards two-way integration. As indicated above, Alba’s (2009) work does lean towards it. Further, we find some useful socio-psychological literature on the attitudes and preferences of natives when it comes to interethnic contact and integration (see e.g. Kosic, Phalet and Mannetti 2012; Martinović 2013). Yet this literature focuses mainly on the determinants of individual behaviour, rather than of societal integration (see also Elias and Scotson 1994: xx).

Anderson (2010: 115–16) is one of the few scholars to systematically use this conceptualisation of the two-way process:

Integration does not view disadvantaged communities as the only ones that need to change. Integration aims to transform the habits of dominant groups. It is a tool for breaking down stigmatization, stereotypes, and discrimination.

Anderson’s (2010) work, which is embedded in her broader theoretical framework of relational equality, does not focus on a migration context, but on the integration between the black and white population in the USA. This means that there still is a need to further develop this notion of a two-way process for the context of people with a migration background in (western) societies. I use her work in the next section, where I present some building blocks for such a framework.

The second author who does concentrate on a relational understanding of integration in a migration context is Jiménez (2017: 11), who in his work on relational assimilation states:

Relational assimilation involves back-and-forth adjustments in daily life by both newcomers and established individuals as they come into contact with one another. This volley of back-and-forth adjustments starts off with rapid intensity as new arrivals and established individuals first meet, and it gradually moderates over time (often across generations), as a working consensus around ethnic, racial, and national belonging develops. When considered side by side, the processes of established and newcomer adjustment are mirror images: both parties feel a tremendous sense of loss, often longing for the way things were; both also feel a sense of gain because of the dynamism that migration can produce. But the deliberate and incidental adjustments that established and newcomer individuals make are more than just mirror images—they implicate each other.

In Jiménez’s approach, the change amongst insiders is not a factor that could potentially affect the integration of outsiders but is an intrinsic and essential part of the integration process in and of itself.

5. Building blocks for a relational approach

Building on the preceding analysis, I argue for the development of a new theoretical framework that takes the third understanding of integration as a two-way process as its starting point. Integration should be conceptualised as a process that takes place between insiders and outsiders, which implies a relational understanding of integration. The specifics of such a framework will need to be further developed, and are outside the scope of this article. I do, however, offer some initial building blocks by depicting three types of ‘shifts’ that would be necessary for the development of such a framework: a shift in ontology, methodology, and vocabulary.

First of all, a relational framework of two-way integration requires an ontological shift. I propose to develop a framework that relies on the basic assumptions of process theory (see also Emirbayer’s (1997) plea for ‘relational sociology’). A relational understanding of integration involves a configurative ontology: We need to understand the integration process as a configuration of social interactions that together constitute the structural (integration) context (see also Cederman 2005: 871). This means moving away from taking the changes that outsiders go through (their commitments, efforts, and achievements) as the central unit of analysis. Instead, integration is to be located in the power relations between insiders and outsiders, who are both intrinsically part of the process. Consequently, a relational understanding of integration cannot be used as a descriptive term for an individual. ‘Integration’ provides information about the relationships between people, not about (categories of) individuals.

A relational perspective on integration requires that integration research primarily focuses on the relationships in which the social boundary between legitimate and non-legitimate citizenship is constructed, reiterated, and transformed. This makes it a framework of relational inequality, as it concentrates on the relationships within which (material and immaterial) goods and freedoms are distributed, rather than this distribution in itself as an outcome (Anderson 1999). In focusing on this social boundary, relational integration is not (necessarily) a process in which objective differences between people decrease, but rather a process of change in the meaning that is attributed to perceived differences. The two are not unrelated: perceived and objective differences mutually affect each other. Yet, objective differences only become an integration problem if they are used to question the legitimacy of membership.

Further, the framework allows for the notion that more integration does not necessarily equal more harmony, since changing power relations cause social friction. Insiders may feel threatened by a decrease in objective difference, and in response put extra emphasis on what distinguishes them from ‘the Other’. As Elias (2000) describes in his Civilizing Process, when the social power of outsiders increases, their self-consciousness and resistance also rise, usually leading to a phase of ‘repulsion’ or ‘differentiation’ (430). Such exclusive backlashes make the meaning of perceived difference extra important. Practices of boundary contestation, conflict, and resistance can thus be investigated as integrating practices. Conflict may indicate more, rather than less integration.

This puts power at the core of integration analysis: What constitutes this notion of legitimacy, and who is identified as legitimate and who is not? As we learn from the boundary theory, insiders have the power to maintain and reiterate the boundaries that secure their dominant position as legitimate citizens through their ‘gatekeeping power’, which refers to the power to decide who is ‘in’ and who is ‘out’ (Jiménez 2017). As such, they both have the power to determine the access to different forms of capital, and to impose their expectations concerning the right integration strategy (Horenczyk et al. 2013: 208). In investigating gatekeeping power, it is important to differentiate different power mechanisms. Whereas outright discrimination may be the most obvious, other utterances of power, such as identity power (Fricker 2007), epistemic power (Dotson 2014), and symbolic power (Hall 1997) are crucial to understand the scope and complexity of a social boundary.

Note that the claim that integration is a two-way process is not synonymous with the claim that integration is a symmetric process in which two parties go through a similar process of change. The power relations between insiders and outsiders bring along different resources and thus different degrees of control over the process. A relational approach to integration enables us to map the different roles within an integration process, different resources for attaining integration and, consequentially, normatively different responsibilities (see e.g. Carens 2005; Miller 2016).

Methodologically, a relational perspective on two-way integration requires a shift in both the type of research questions and the kind of data that is used to answer them. To study relational integration, one needs to move beyond comparing the average achievements of outsiders with insiders who function as the benchmark of this comparison (objective differences). Instead, one needs to ask whether and how such inequalities are constituted by social boundaries (perceived differences) between legitimate and non-legitimate citizens. At the institutional level, relational integration research would focus on the question of how such differences and correlating status hierarchies are expressed, reinforced, or transformed through existing formal and informal institutions. At the inter-subjective level, it would be concerned with these boundaries in terms of the relationships between people across and on either side of the social boundary. This brings forth the need for a different type of data that enables us to measure such boundaries: data on boundary maintenance and boundary change in formal and informal institutions (norms, policies, and symbols) and inter-subjective relationships (attitudes, roles, and behavioural schemes) (see also Yuval-Davis 2006). This data, I want to emphasise, would not serve to investigate the ‘integration context’ (as it does in one-way approaches) but provides information on the integration process itself.

When researching relational integration, ‘mapping’ the characteristics of the social boundaries at play is key. In doing so, one needs to be open to complexity: social boundaries are intersectional, and categories of insiders and outsiders are internally heterogeneous. Furthermore, one needs to take into account the different contexts of local, national, and transnational levels, and the interaction between them (see also Grzymala-Kazlowska and Phillimore 2018). As is custom in boundary theory, the framework would therefore not have a law-like, universal character but rather should serve as a tool to generate local, context-dependent explanations (see e.g. Wimmer’s (2013) framework of ethnic boundary-making).

Third and last, researching integration as a two-way process requires a language shift. Many authors contribute unwillingly to one-way discourses by structurally talking about ‘the integration of migrants’. Instead, I argue, we should be speaking of the integration between migrants and non-migrants, or, preferably, outsiders and insiders. In order to truly break with one-way integration discourses, to work with power relations as a unit of analysis, we need a more relational vocabulary.

6. Conclusion

Although many authors do embrace the notion of a two-way process, the current dominant conceptualisations are unable to avoid the problems of ‘one-wayness’ that they are supposed to overcome. I have argued that we need to avoid conceptualisations that maintain a distinction between ‘people who integrate’ and ‘people who do not integrate’, and thereby place insiders outside of the process. Instead, we need to understand two-way integration as a process in which both insiders and outsiders take part (albeit in different roles).

A shift towards a configurative ontology should allow us to observe dimensions of the integration process that, until now, have been largely left out of perspective. Insiders will no longer be perceived as a potential independent variable that can be used to explain the integration of outsiders. Instead, their attitudes and behaviour become an inherent part of the unit of analysis: The relationship between insiders and outsiders will be the dependent variable of relational integration research. Starting from the power relations between insiders and outsiders, the framework should provide a more reciprocal perspective on the various roles and resources in different integration contexts. Through the analysis of these roles and resources, we can rethink assumptions about conditions for successful integration and the way we measure integration in current empirical contexts. Furthermore, it could inform a reconsideration of normative perspectives on responsibilities for integration.

Acknowledgements

I could not have written this article without the input of Luara Ferracioli, Eric Schliesser, and Floris Vermeulen, with whom I have had many stimulating discussions and who provided very useful feedback on several versions of the manuscript. Further, I want to thank Richard Alba, Jan Willem Duyvendak, and Enzo Rossi for their critical reading of the article. I also thank the two anonymous reviewers for their time and their elaborate comments which substantially improved the quality of the article. Last, I want to thank the participants of the CPPC seminar at the University of Amsterdam for a fruitful discussion after my presentation of an early draft.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Conflict of interest statement. None declared.

Footnotes

1. Joppke (2007: 3) is an exception, calling the ‘two-way integration notion’ a ‘platitude’.

2. A surprising number of scholars measure the concept of integration while not giving any definition of integration at all, relying instead on tacitly shared assumptions of what integration is and is not (see e.g. Ager and Strang 2008; Crul and Schneider 2010; Koopmans 2010). Consequently, many scholars seem to be talking past each other while ostensibly discussing the same concept. Notable exceptions here are Alba and Nee (2003) and Penninx and Garcés-Mascareñas (2016).

3. The first two understandings do not exclude each other, and can be found conjointly in the work of several authors (see e.g. Lucassen 2005: 19; Penninx and Garcés-Mascareñas 2016).