1. Introduction
In a 2012 novel entitled
The Most Human Human, the author tells the story of how the computer plays into the long-standing philosophical narrative of the unique place of humans in creation (
Christian 2012). The central character in the book serves as a human blind, chatting with a panel of scientists through an interface, who then have to decide whether the character is a human or a chatbot. Then, the author wonders the following: What can a human do with language that a robot could not? What are the ways of expressing ourselves that are the most distinctly human? How do we recognize our fellow humans?
The acute late concern about the complexity of integrating societal values into digital systems and ensuring that they behave in ways we intend them to, is not new and goes back to the days of computer infancy, when philosophical questions on the moral and technical consequences of automation and the fears over a mechanical agency with whose operation we cannot efficiently interfere were already asked. Today, as we experience a sudden leap forward with machine learning models put into the decision-making apparatus of society and the state, it seems we have put philosophy and values again into the center. The question is not anymore whether the machine is going to behave the way that we expected or wanted, but social scientists are pulling the fire alarm, so to speak, scrutinizing what actually goes into these datasets and to which extent ethical values are represented. Language, ethics, and representational harms are more relevant questions that come up in the present day.
The boundaries within European society and the governance of migration are not limited to the overt racism and xenophobia often associated with the far right. Instead, they encompass institutional frameworks and political and legal discourses that appear neutral on the surface but still function within a racialized logic. We may be at a pivotal moment, where significant shifts in diversity and representation are possible because of digitalization. There seems to be an emerging scrutiny of citizenship as a systemic structure, and the hope is that it will extend to the examination of societal borders and civic registration. Historically, we cannot explore or theorize the concept of citizenship without considering the state’s responsibilities and obligations, which others have referred to as the coloniality of power. These continue to shape the structure of the modern state. Therefore, we must acknowledge that any discussion of citizenship must also recognize how the framework still reflects existing hierarchies. Reparative interventions should be undertaken to dismantle these structures of injustice. Against this background, several questions emerge about the interplay between forms of contemporary citizenship, migration, and the politics of diversity: What is the position of the noncitizen in contemporary states? How can the nexus between citizenship, migration, and digital technology be best conceived? How can we resolve the tension in policy and practice between the intersection of regimes of rights and forms of social cleavages (migration)?
There has been little work thus far bringing together various trajectories of these spheres of inquiry despite the fact that migration has a continued presence as a construct that shapes social relations in modern society and continues to have analytical purchase as a concept toward contributing to more progressive ways of thinking about the relationship between the individual and the state.
In this paper, I wish to draw attention to the civil registry identity categorizations in The Netherlands and explore some of the complexities surrounding the relationship between digitization and the often absent dynamics in policy regarding the effective addressing of migration and diversity. Two lines of argument are discussed: firstly, the importance of focusing on diversity and how it may be recorded digitally and, secondly, to discuss the potential of the digital as language. Drawing on cultural and political themes in critical theory studies, I wish to ask what digital tech can tell us about the approach to migration. Next to this, I look at the digital as a language trope that provides a way to approach this issue and suggest how technology and its uses may work to enable the EU state to be open in the face of diversity without implicitly reproducing the colonial ethos and Western-centered vision and meaning of citizenship.
2. Diversity as Category in State Policy Practices
It is a really good laboratory in terms of state politics to think about how diversity is enacted in the everyday through practices, laws, policies, and discourses that continuously produce and reproduce processes of social cleavage (
Bhambra 2015). Here, I discuss how diversity is enacted in the context of the civil registries in The Netherlands, especially in relation to the rights of people who are not classified as citizens, and by doing so, I wish to intervene in current EU debates on digital tech, inequity, and diversity in relation to digital civil registration in the country.
This paper joins in the scholarly debate that takes issue with the fact that modern societies of Western EU nation-states carry on dealing with diversity in ways that continue to affirm their power and superiority as the distinctive feature of their liberal democracies. Within their legal framework about citizenship,
1 states are drawing on the idea of diversity as a model of migration through one overarching cause, which is their ability and power to absorb it and reunify it rather than understand it as a form of positive resource and a model for equal participation and social justice. What conceals the problem is that despite living in unequal, neoliberal racial societies, we are asking noncitizens to integrate into a kind of fictitiously homogenized society in which the so-called natives are integrated and the foreigners are not. Scholars have caught up with this to some extent, and their accounts confirm that diversity as a concept may often be misleading, and while it has lately taken off as a concept, it is often weakened and diluted to mean something along the lines of box-ticking diversity exercises, which is drastically missing the mark. (
Bhambra 2015;
M’charek 2005;
M’charek et al. 2014).
This kind of logic is part of an object of government and biopolitical intervention that encloses a body politic within a territory for the purpose of absorbing and eliminating (biological and cultural) diversity (
Foucault 2008;
Isin 2018;
De Certeau et al. 1975). From this perspective, othering ends up being biopolitical dispositifs
2 to separate people in a sort of national branding device (cf.
Agamben 2009), which uses the passport, the symbolicity of the state, to maintain which passports are more valuable than others from other countries, thought to be at the bottom of the global inequalities list.
3 This is how the ‘brand’ is maintained by sustaining inequalities
4 between people, who then have to go through tests and procedures in order to show that they deserve to be equal. We recognize the work of many prominent authors, like
Arendt and May (
1958), in this argument. Modern-day racialized neoliberalism does nothing to address some of the really pernicious deep problems about identity categorizations, classifications, and the system of global inequalities that this sustains. The classification schemes of populations are tools of discrimination and the parsing of people based on a social, institutional sort of fabrications used for nation-state building and policy assumptions ingrained in policy practices.
Consider, for example, how the datafied representation of cultural categories is being produced in state registration systems and how these, in turn, influence the way (non)citizens are addressed by the socio-political system of power. Let us take a specific example of the categories used in state registries in The Netherlands. Population statistics in The Netherlands are based on the digitized municipal population registers. A good example to look at is the technology used in the Civil Registration System or the national identification systems, which then link the national ID database with welfare and law enforcement records. Register data (System of Social-statistical Datasets (SSD), Municipal Population Register (BRP), and additional registration records from the Immigration and Naturalization Service (IND, Immigratie en Naturalisatiedienst)) cover all people who officially reside in The Netherlands and are compulsorily registered in Dutch municipalities (
Bovens et al. 2016;
CBS 2016,
2017;
Prins 2017). The Netherlands abides by the conventional EU mode of identifying migrants, which is to focus on national origin, on those who are born abroad, on those who have parents or grandparents born abroad, on those, in other words, who are deemed immigrants, and on those who, therefore, have “ethnicity” in relation to their immigrant status via their national origins. Thus, people are identified by nationality, citizenship, origins, the citizenship of their parents, their (main) motive of residence application, birthplace, and so forth. Like other EU countries, The Netherlands is careful about not appearing to reproduce categories that evoke the sorts of race classifications that took place in the past, but, of course, such national data practices used for the classification and quantification of residents continue to simultaneously address, frame, and govern people on the basis of cultural identity categories. Such identity categories are artifacts that reveal information about membership and the relationship between people, noncitizens, with the state at any given time (cf.
Yanow 2003). To explicate that relationship categorizing works powerfully in policy ensuing respective action, even if immigrants arrive already categorized by mainstream politics and public discourse (
Van Hulst and Yanow 2016), it is not until they are made to fit the established categories of the receiving state policy and administrative practices that the state decides how they are to be handled from then onward (cf.
Lipsky 1980).
The identity categories are not neutral administrative terms, and certainly, they are not neutral designators in their effects.
5 The sort of language used is, therefore, very important in policy. The meaning implicitly enables a race discourse in which birthplace stands also for a tacit understanding of behavioral traits (
Yanow and Van der Haar 2013). This is one aspect of language that performs “othering” (the racialized Other) and suggests that integration is and will not be possible for the non-Dutch into Dutch society.
6 In Stuart Hall’s writings (
Hall 1996), language sends them into symbolic exile. Another point to note is, of course, that many of these populations are effectively not immigrants, and in this sense, they do not need to be integrated.
7 This is a very obvious point from the perspective of diversity and migration studies. Long-term-settled ethnic minorities or migrant-origin populations (perhaps first-generation immigrants or second- or third-generation migrant-origin children) are noncitizens living in the society in which they are integrated, but they are not being recognized or represented as full citizens, and they are being distinguished from the so-called native population.
It is clear that identity categories are classification indexes that reproduce the national unit over time, effectively constructing its sovereign power over the society and putting people and identities in boxes. The statistical constructions of different groups in order to identify what is typically seen as immigrants in EU societies, i.e., from a non-EU origin, are, of course, problematic. They clump together different sorts of ethnic, racial, national-origin and ethnic, and cultural groups as immigrant groups, contrasting, as such, the non-Dutch origin as non-Western, and there are assumptions of race taking effect here as well. Policy administrators say they are not constructing groups with data, but they have individual-level data with variables, and they are controlling for class, gender, education, culture, religion, where people come from, ethnicity, and their behaviors. There are, of course, many different statuses in the normative model of the sovereign nation-state categorizations that are contrasted to an alleged homogenized mainstream of natives who, by definition, do not have ethnicity or, in other words, they are statistically invisible in the mainstream, which, in turn, enables us to see racially and ethnically distinct migrant-origin populations, and that is problematic. The construction of diversity is the mirror of the racialized production of identity classification categories.
In technical terms, most of the ethnicity assumptions about noncitizens fall apart. We do not have the models and the language to do it in the way of bringing rights without discriminating against noncitizens. The existing models are long-standing national practices, such as naming patterns and categories that are fed into state policy practices. We need to try and shift the way that the system is designed to work regardless of how difficult it is to get policies to operate outside of the models that work for them. This is an agenda that needs re-working and refining.
Critical scholars are doing this kind of work, addressing the big moral questions at the heart of the matter related to inequity and hierarchy of cultures. We may start focusing on the different ways in which polity can behave in relation to diversity. There is a need not to fudge with the issue of diversity and inequality. My point is not to get society organized on the basis of diversity but the simple point that the goal of our society should be to dispose of the structures of exclusion that continue to underpin the ways in which societies are organized. We have to deal with this in order to be able to think further about where we want to be with this in the future as we try to decolonize European societies. There will be models that can pull the ideas that I am advancing forward, so the discussion in the next section is on what digital tech can do for citizenship and migration in today’s digital age.
3. Digital as a Language Trope
In the early days of the internet, scholars of digital technology were putting forward their ideas on the fundamental challenges to all things social
8, including possibilities to transcend inequalities, by the emergence of virtual culture (cf.
Castells 1997;
Turkle 1997). They expected it to be so exceptional, in fact, that people would have to be provided with new rights and legal frameworks as a result of the new power of tech. Yet, already back in the early 1990s, there were others, like the work of Oscar
Gandy (
1993), who stated that power relations between individuals and sociotechnical systems are not exceptional, but they are shaped by political, legal, and social choices.
9 Since then and despite the ubiquitous usages of digital technologies, we are still not conclusive as to how digital tech is fed into the realm of policy and whether it can lend itself to producing an understanding of the social world on the basis of equity and inclusion. Inclusion and equity, being key values for democratic polity, must be addressed, in particular when contributing to a digital government. With a strand of literature highlighting that we need to use better analytical tools in order to advance our analysis of technology, the attention of scholars has moved into theorizing the extent to which liberal democracies may construct a coherent political digital vision that advances democratic values. They are looking to uncover how tech shapes the structures that underlie social inequality and discrimination and vice versa. Their work on shaping ethical technology, society, and, in some ways, politics concentrates on the intersection of values and technology, in general, and examines the deep roots of inequality in existing social structures that shape technocultures in particular. (
Arora 2019;
Zuboff 2019;
Oomen 2016).
While automation technologies are often accompanied by narratives of progress and freedom, they are also associated with increased surveillance and a shift in surveillance practices from discipline to prediction (
Andrejevic 2019;
Zuboff 2019). Automated technologies, particularly machine learning algorithms, function by sorting and categorizing data, which are then used to identify patterns and make predictions based on those patterns. The goal of the Panopticon in Western states’ surveillance, founded on mathematical calculations of human behavior and predictive algorithms, is not only to predict our behavior but also to influence and modify it to maximize social efficiency. It is, therefore, unsurprising that such technologies contribute to the replication of discriminatory patterns from the past with potentially disastrous consequences for democracy and freedom. In this context, the application of AI and related automation technologies is clearly at odds with human rights and the protection of vulnerable communities. We are witnessing the negative effects of technological power in areas such as national identification systems, fraud detection by tax authorities, migration control, and so forth.
Societies and individuals face crucial junctures. Researchers in the social sciences and humanities are examining the challenges to rights and freedoms posed by technology, as well as the structures underpinning the rise of social inequalities and discrimination. They argue that technology is symptomatic of the way we have constructed our current political and social systems, and its outcomes are shaped by our political, legal, and social choices (
Ferrari et al. 2023). The challenge lies in understanding how we can reconcile these differing perspectives on the appropriate use of technology if we wish to govern it effectively (
Taylor 2023).
A re-examination of the legacy of modernity and a questioning of its structuring principles as a political project that emerged under specific conditions within the context of the European nation-state are necessary in the current context of racism against individuals often categorized along ethnic, racial, and religious lines, whether they are citizens, migrants, asylum seekers, refugees, or stateless persons (
Lentin 2020).
This study is partly an attempt at confronting the institutional structures fed by the facts of tradition and archives from a decolonial perspective. The project here is to deliberately “read” them from a different place, from a place other than a politics of representation with ties to colonialism. The intellectual exercise attempted in this study is to put these structures in relation to other current contingent concerns, namely digitalization, and to do this in a manner that allows it to speak out of past lineages that use exclusion and racism to develop discourses of otherness and discrimination.
One main question to ask is whether digital technology, rather than offering an escape route out of notions of racializing diversity, is still chained to a politics of representation with ties to colonialism, complicating diversity in new ways (
Nakamura and Chow-White 2012). We acknowledge that technology, from everyday uses to complex algorithms, while appearing neutral, has many entanglements, including the potential to discriminate. It is a challenge to try to distinguish between those two levels of tech system in operation: one that potentially reproduces and amplifies inequalities and another that acts as a virtual environment of choice, freedom, and emancipation.
The problem most readily identified next here is that cultural biases seep into the data systems being used. This hinges on the kinds of datasets that are used in order to train these systems, which adds an extra layer of complexity into this. We hear a lot about inequalities, through network-based bias in operating systems, and we question to what extent these models of training datasets are really fixable “because it is a question of data”. But, is it the case? What happens, for example, when we find identity bias in fraud-detection tech tools, as we have seen recently happening in the case of the tax authority in The Netherlands and the
toeslag affair?
10 We understand from examples like this that racialized categories are implicated in the structure of tech interfaces. The way that socio-political systems are optimized to serve the interests of state power may not be something new, nor is how new digital technologies may have internalized this logic, but it is a really interesting case to approach this debate as it is landing in The Netherlands regarding technology, diversity, and the civic registration system.
The Dutch government has the ambition to advance by forming a digital government and to use the potential of digitalization in practice, including how civil registries are accessed and used through digital platforms. National agencies and municipalities, the main authorities on residency and citizenship status, have mandates to shape their digital services and increase knowledge and efficiency about how digital technology can also be used in this domain. In the context of advancing digitalization, my concern here is with the kind of taxonomy of categorization that is placed at noncitizen registration. The focus is on the constellation of technoculture developed so that it does not evoke the language of hierarchies of difference. The solidification of cultural hierarchies has been a key function of registration procedures in practice and develops into how people are considered distinct and discriminated against on that basis. There is no denying that there are problems of discrimination in the registration process in the move to a digital registration process as well. Assumptions about culture and identity of migrants, developed within national borders, are used in interfaces (i.e., datasets are lumping different people under the same category in menus and clickable boxes) to define cultural categories and, in this way, sustain them online and trace them back as key classification categories. Such categories, embedded in infrastructures and government policy tech design, become indicators of the state policy, facilitated by clickable boxes (
Nakamura and Chow-White 2012). These strategies are clear examples of how the political techniques of the state and the technologies intersect (
Agamben 2009).
There are other intersecting components that are playing out for us to consider, as well. Probing a little bit further, datasets are, to a large extent, generated on the basis of convenience, from anywhere they are available to be harvested. Once we start approaching the issue of inequality, diversity, and digital tech from these aspects, the question is turned onto its head, and it becomes less about biasing or de-biasing datasets and more about what types of datasets we want to curate for certain purpose and hence, the models we would like to have to work toward our aim. This should include the kinds of social, ethical, and political issues that are implicated in datasets as a way of curating them better, so that the results will not be as unrepresentative of certain populations, including race, ethnicity, and so forth. Obviously, we are dealing with profiling people and communities, so it is a sensitive subject that cannot be dealt with passively and unconsciously. This is a matter of research to be fed into social policy oriented toward theory building. How can we deal with this conundrum of what are tech systems (i.e., information governance structures) built for and what are the outcomes that are desired? In a way, a key insight into this becomes a choice of how we curate them. I would suggest not to link these tech systems to the rubrics of inclusivity or discrimination but to put the digital aspect at the center and examine it as a language trope.
Latour (
1999,
2005) claimed that there are moments when technology can speak for itself and something may be learned about the social in the process of interpreting the technological. Admittedly, digital technology was born and raised in fully artificial laboratory conditions. However, digital tech is contingent, shaped and interpreted by language. Capturing and making sense of the interaction between the ‘technological’ and the ‘social’ is where language comes in. This helps us address issues in a more systematic way. Digital tech seen as a language is quickly marked by discourse. The premise of language here borrows from both the Foucauldian and Latourian senses.
Foucault (
1978,
1981) clearly recognized the power that language holds over the existence of a thing, as evidenced by his claim that without something to define it or to place it within specific parameters of meaning, a thing is left formless and cannot exist within the human mind. Any thing has to be “subjugated at the level of language”, as Foucault appears to understand it, and through the subjugation of language, our very conceptions of it may be drastically altered. In correlation, Latour argued that science as practice is a discourse based on language (cf.
Latour 1999). Language, in Latour’s understanding, is the code between discourse and representation. Language is the symbology used in discourse and representation, not the mere labels of things but words and actions indistinguishably woven together. Language, thus, becomes a key metaphor for thinking about mediation by means of coordinated actions and transmitted meanings. Similarly, in a quasi-Latourian analysis, digital tech may be the language behind the meaning of actions, a system of mediators that encodes a string of signifiers dependent on the context of use for its particular meaning. Hence, the digital, as a language, is not something imposed on the social world, but it may be something that defines the social world, already composed of actions, negotiations, and transformations. The premise here is to look at the digital as a language and a catalyst for change when approaching diversity in polity, in this case. If digital tech is language, discourse and representation are the grammar, which highlights the structures within which value is anointing arbitrarily to diversity. This value has essentially no meaning until meaning is ascribed by representations based on a system of exclusion.
Hence, the mission is to separate language from representations of exclusion. In effect, bringing the digital to the foreground of the analysis, alongside language as the symbology used to conduct the tech, gives rise to two challenges. The first challenge for researchers is to avoid the problem of looking at technological systems in a way that frames them as innately positive or negative (i.e., to explicitly focus on whether there are particular kinds of racial identities that are being included or excluded in the digital registration systems and whether this is good or bad). It connotes a very restrictive way of appreciating them, which means that we are looking at them as a finished product. The challenge is to go a step further and try to demoralize them, to appreciate them as constantly emergent forms within context. This, in turn, highlights the complicity of socio-political systems within which they are embedded and allows the social-cultural elements to make a play. By doing so, we accept that systems can actually be neutral, but they are in constant motion as they are being interacted upon, and hence, they have the potential to become more or less discriminatory. Digital identity categorization schemes in government registries are digitally created to impose some sense of order on diversity. Viewing them as language tropes is challenging the marking of people through lines of difference. When we accept that language categories are not univocally connected to an outside world, our research is bound to proceed differently. In the same way, researchers may challenge power hierarchies in the digital production and circulation of such representations, suggesting that more carefully contextualized work is required to trace the political subjectivities of diverse communities in digital national categorization schemes.
11 The premise of looking at the digital as language has, in this sense, the potential to create changes in the way tech is regulated and governed.
The second challenge for researchers is to think about the way that language and digital technology shape one another, especially within a scholarly environment, which is more and more concerned with the ethical and social impact of technology (
Warschauer 2000;
Kolko et al. 2020;
Benjamin 2019;
Nakamura and Chow-White 2012). Technology has historically had an air and allure of objectivity that made us assume it rises above human patterns of behavior and language. If we want to construct a different social reality, one grounded in social justice and equity, one of the things we have to come to grips with is that technology is not an asocial or apolitical form of authority and production of knowledge. It is not simply the outputs of technology but the social inputs that inform the tech applications. However, defining the inputs is not enough to define the outputs in order to disrupt knowledge and facilitate anti-discrimination action in state policy, for example. In today’s tech parlance, the language’s imperceptible system of signifiers is digital. Language is a resource, which means that language is productive and forward-looking and can be innovative, as much as it is systemic, based on how it is instrumentalized. This, in essence, means that language can broaden the parameters of the underlying systems of established structures, which are not fixed nor complete because they never are. In fact, if we go by this logic, we can use the digital as an input and output of social order. This begs the question of how can the digital, as a language trope, develop an innovative and critical conceptualization of citizenship rights and critically assess the extent to which digital infrastructures may or may not expand spaces of recognition and representation or even contribute to the development of a citizenship regime in a meaningful way based on social justice and equality. Our quest for social justice and equity entails the concern for the role of technology and technological prowess in that quest. Too often, our investment in building up our technological capacity has not gone hand in hand with our ability to build up our moral language capacity and our social imagination. These two things have been very asymmetrical. Hence, in kicking off on the right foot, the digital language trope is a way of building toward an end goal of a revived citizenship kind of project with creative and critical capacities as we move forward about the plight of humanity for equity and the role of technology in that.