Challenging knowledge production on migration with statactivism: the category ‘migration background’ and some destabilizations

ABSTRACT

‘Migration background’ has been an official category of population statistics in Germany since 2006 and has become an integral part of knowledge production about Germany as an immigration country. However, ‘migration background’ is not only a way of constructing ‘migrant’ and ‘native’ populations; it is also employed as a marker of ethnic inequality and often differentiated into subgroups to stand for national links or ethnic descent. I reflect on these observations as utterances of methodological nationalism inscribed into data practices and (non)knowledge production in migration research. I link them to ideas about statactivism (Bruno et al. 2014. “Statactivism.” Partezipatione & Conflitto 7 (2): 198–220) as a means of intervention. I give some examples of how I have tried to destabilize the usual (non)knowledge production in migration research and reflect on the possibilities for individual researchers to do so. My ethnographic analysis is based on a close reading of national statistical reports, fieldwork in statistical offices, participatory observations in workshops, informal and formal interviews, and my own involvement in (non)knowledge production.

KEYWORDS:

• Migration background
• migration research
• statactivism
• knowledge production
• methodological nationalism

Introduction

Since 2006 |migration background|¹ has been an official category of population statistics in Germany and has become an integral part in knowledge production about Germany as an immigration country.² The recent definition of the Federal Statistical Office for the microcensus reads: ‘a person has a migrant background if he or she or at least one parent did not acquire German citizenship by birth’ (Statistisches Bundesamt 2021).³ The microcensus is an annual population survey called ‘small census’ with mandatory participation (Statistisches Bundesamt 2022). Several definitions for |migration background| can be found in other statistics. Nevertheless, the Federal Statistical Office’s definition of |migration background| using microcensus data functions as a benchmark; it was a blueprint for further definitions, and its results serve as target values in other surveys.

Population statistics play a vital role in narrating ‘stories or tales about collective identity’ (Yanow 2003, 22). Grommé and Scheel (2020) delineate the enacting powers of categories describing nationally coded memberships. They add an important component to studies exploring the performativity of census categories (Geschiere 2009; Gorodzeisky and Leykin 2022; Kertzer and Arel 2002; Krebbekx, Spronk, and M’charek 2017; Leibler and Breslau 2005). This article supplements these studies by exploring how |migration background| is enacted in (non)knowledge production on migration.

I address in a first step the construction of a German descent-based community also discussed as ‘nativism’ (see Bosniak 1994; Geschiere 2009) and in a second step ‘methodological nationalism’ (see Wimmer and Glick-Schiller 2002) in the enactments of the category |migration background|. As migration research is ‘enacted’ (Scheel, Ruppert, and Ustek-Spilda 2019), it might also be changed from within, following the ideas of statactivism (Bruno, Didier, and Vitale 2014). Statactivism does not refute quantitative methods and related practices of categorization entirely; instead it engages in the politics of categories and quantification to challenge the (non)knowledge production. Therefore, I try to figure out how the concept of statactivism might guide future enactments of (non)knowledge production on migration and how it can destabilize, pluralize and mind the ethics of categorization. The questions I try to answer are: How and where are the logics of descent-based nativism and methodological nationalism evident, and how can researchers destabilize and possibly overcome them by practicing ways of statactivism?

My primary data sources are materialized products of statistics like tables, graphs, questionnaires, and reports, which I analyze with a close reading approach. This is backed up by field research in statistical offices, and my personal involvement in migration research and thus the (non)knowledge production in question. I use material-semiotic approaches (Grommé and Scheel 2021; Law 2011; Mol 2002). This means that there is no reality out there. Instead, realities are enacted by the involved actors (human and non-human), who are linked through networks with each other. I suppose that statactivism and subversive semiotics can guide more plural future enactments and address criticized re-enactments of power hierarchies in nativism and methodological nationalism.

This article consists of three major parts. The first is a conceptual one, where I will explain the theoretical foundations of my work including the concept of statactivism, followed by a description of my material and some methodological notes. This lays the groundwork for the second part, an empirical section in which I sketch out how the seemingly inherent properties and functions of |migration background| are enacted firstly as a marker for alterity and ethnicity, and secondly for specific national ties and ethnicities. In the third and last part, I reflect on how practically statactivism can be used for intervening in migrantizing, racializing and ethno-nationalizing effects of statistical (non)knowledge production in migration research. Giving some personal examples and discussing their emergence and production I call for a more critical, reflexive, and active engagement of all researchers in (non)knowledge production in migration research.

Conceptual groudings

STS-approaches to (non)knowledge production in migration research

Scheel, Ruppert, and Ustek-Spilda (2019) call for more research on the enacting effects of knowledge and non-knowledge about migration. They link material-semiotic ideas from science and technology studies (STS) to the field of ‘migration’ and stress that without data, there is no ‘migration’ as a public phenomenon, only the movement of individuals, some crossing imagined lines enacted as national borders, others not. To perceive those movements as ‘migration’ in political and scientific discourses, they must become materialized in numbers, graphs, and maps. Producing these materializations presupposes a package of politically charged knowledge (for instance, the belief in nation-states, borders, and entitlements to free movement), which is already inscribed into the very notion of ‘migration’. In their view, data practices ‘are: situated in and produced by sets of relations; sociotechnical in that they involve relations between humans, materials, infrastructures and technologies; performed by actors and function as stakes in struggles over authority and power within specific professional fields of practice; and contingent in that they do not have a ‘prior and determinate form’ (Law, 2004, 38) but involve practical adjustments to address complex and changing conditions’ (italics in original)(Scheel, Ruppert, and Ustek-Spilda 2019, 583 f.).

All these characteristics are true for |migration background|. Even though Scheel and colleagues (2019) use ‘perform’, I prefer the term ‘enactment’ in this article. In doing so I follow Annemarie Mol, who introduced the metaphor ‘enactment’ in contrast to ‘performance’, which is more common. She explains her choice: ‘It is possible to say that in practices objects are enacted. This suggests that activities take place – but leaves the actors vague. It also suggests that in the act, and only then and there, something is – being enacted’ (italics in original)(Mol 2002, 32f.). Thus, ‘enactment’ is more explicit about the accidentality or collaterality as John Law (2011) put it. The metaphor ‘performance’ might suggest a script, a stage and a background as Mol (2002, 32) also explains. However, there was no masterplan for |migration background|. Therefore ‘enactment’ fits well in this context.

|Migration background| is contingent. However, there are significant patterns, such as methodological nationalism (Wimmer and Glick-Schiller 2002), resembling and reproducing the (non)knowledge of former and other networks. Those patterns are crucial for coordinating (non)knowledge production, and the impression that scattered ‘collateral realities’ (Law 2011) form one body of knowledge that scholars may communicate as certainties. Bowker and Star (2000) also stress the contingency of categories, but add the idea of categorical ethics. Even though categories are contingent, there are better and worse ones concerning their performative influence on individuals’ lives. This is crucial in the case of ‘identity categories’ (Grommé and Scheel 2021). The plea for ‘better representations’ is therefore not a direct indicator of remaining in the ‘realist register of representation’ (Grommé and Scheel 2021, 174), but rather a political claim against hegemonic and powerful misrepresentations.

Creating categorizations includes several steps of silencing and valorizing different points of view (Bowker and Star 2000, 5 f.). For census categories, statisticians must decide what should (and should not) be counted. The same applies to migration researchers (Bartels 2024; Amelung, Scheel, and van Reekum 2024). There is a heated debate on whether ethno-racial data should be collected at all (Simon 2017). The German state, represented by different actors such as the Federal Statistical Office, the Federal Ministry of Interior, and the Federal Ministry of Justice, use |migration background| as a kind of workaround to not explicitly collect ethnic data but feed a discourse about ethnically framed differences and disparities.

If ethno-racial data is to be collected explicitly, the wording of questions and answer options again silences and valorizes – or rather, ignores or illuminates and, in doing so, evokes – certain (collateral) realities. Some people may tick a box, and others may not tick one at all, for instance, if there are no categories for hybrid or ‘other’ memberships. Sometimes categorical boxes exercise violence upon the categorized. |Migration background| in the microcensus-definition of the Federal Statistical Office does so by inventing a second-order Germanness for those not born as Germans and those whose parents were not German by birth. It denies ‘nativity’ to those constructed as ‘descendants of immigrants’.

The counting itself silences and valorizes different points of view, and of course, the last step in (micro)census-taking, the presentation of results, does this as well. Which data is published, and how are categories listed, cross-tabulated, graphed, mapped, and formulated? There are many decisions at stake, often implicit ones. Migration researchers must consider these factors of their (non)knowledge production if they want to work in a reflected (Dieterich and Nieswand 2020) or critical (Labor Migration 2014) way. Even statisticians in statistical offices have some freedom to change established representations and to create new categories. Statactivism suggests tools to take a step further and systematically destabilize taken-for-granted migratory (non)knowledge.

Challenging hegemonic representations with ‘statactivism’

Bruno, Didier, and Vitale collapsed the words ‘statistics’ and ‘activism’ together to create the concept ‘statactivism’, ‘a term to be employed in describing those experiments aimed at reappropriating statistics’ power of denunciation and emancipation’ (Bruno, Didier, and Vitale 2014, 199). They specify three possible strategies: (1) generating parallel numerical evidence, (2) inventing new statistically described categories of ‘kinds of people’ (Hacking 1999, 164), and (3) demanding other, newly arranged, or more data (Bruno, Didier, and Vitale 2014). Bruno, Didier, and Vitale do not spell out the answers to the crucial question of how to perform statactivism but give some helpful examples. The reflections of Eviatar Zerubavel about the unmarked, normalized and hegemonic positions and the marked, othered ones and his collection of counter-strategies (Zerubavel 2018) can inspire statactivist enactments.

There are essential contributions from Data Feminism (D’Ignazio and Klein 2020), QuantCrit (Castillo and Gilborn 2022.), and Counter-Mapping (Bonilla and Hantel 2016; Mason-Deese et al. 2018), to mention some praxeotheoretical fields which explore similar questions. It is not by accident that (non)knowledge production and hegemonic enactments are challenged mainly through researchers affiliated with marginalized communities. As categorization is a crucial part of ruling and re-enacting power asymmetries, researchers from marginalized communities are rare in contexts of knowledge production. They met several exclusions from prejudice to financial restraints and have to handle these additional barriers compared to researchers with a background in privileged communities. If some of them succeed in entering prestigious positions, they tend to be instrumentalized by the hegemonic knowledge production as examples of inclusiveness and have to represent their research and their communities. This puts additional pressure and expectations on them and stabilizes hegemonic enactments.

Since the word here is about categories that invent ‘kinds of people’, people who are ‘marked’ (Zerubavel 2018) by these categorizations have to participate in data collection, analysis, and dissemination. Luckily, in the case of anti-discrimination data, communities of the categorized have already formulated corresponding guidelines (Gyamerah and Wagner 2018). Such procedures have to become state-of-the-art. Critical researchers may (and should) function as a bridge between community representatives and colleagues who shape hegemonic discourses with their data and professionals from other fields.

At the very least, researchers have to ensure that they do not ignore the perspectives of the communities of the categorized but face them. To do so, representatives of these communities need a space on academic stages and funding for participation. Academics with a community background should be primarily supported in academia (by mentors and also, even if structurally more challenging, by peers alike), without regarding them as a representative of their communities per se, but valuing and supporting their standpoints, particularly if they question apparently self-evident certainties. Without them and their perspectives, no fundamental changes will take place. Statactivism will not fix the outcomes of social inequalities; in the best case, it might render knowledge production more plural and reflexive. For illustration, I will give examples of recent hegemonic knowledge production and then personal examples of attempts of statactivism in the last part of this article.

Data, methodology, and positionality

The following analysis is based on a close reading of various national statistical reports, three months of fieldwork in statistical offices, participant observation in workshops, and informal and formal interviews with statisticians and researchers in educational and health research. The material is also the result of my involvement in migration-related research and policy consultation. As I have studied the category |migration background| for more than ten years, my analysis reflects several publications and discussions. Many of these publications reached me by chance,⁴ brought to my attention by people I call colleagues because I often worked with them on projects, met them at workshops or panel discussions, or have a similar disciplinary background. They and I are part of the networks that enact the production of (non)knowledge about migration.

I use ethnographic methods, including collecting and analyzing documents, interviews, and participant observation in workshops and scientific and administrative discussions, where I regularly participate as an expert or part of the public. So I am, myself, not a neutral bystander or commentator. For roughly five years, public administration units and NGOs have asked me to explain the problematic usage of the category |migration background|. However, it seems that the (non)knowledge-networks will not abandon it soon. But why should one continue with something known as insufficient and not find better⁵ enactments? How is it possible to keep the ambivalence of categories and their enactment visible?

These questions drive my inquiry into the categorizations of humans. Unsurprisingly it is not only connected to my wage work and research but also my positionalities. The deprivileged and marked ones (social origin, ethnicity, gender, and sexual orientation) are the most productive. At the same time, I am part of a normalized and privileged whiteness. Additionally, I am temporarily able-bodied, largely without caring obligations, and socially advantaged. My interest also stems from my training as an ethnographer, whose discipline focuses on human interactions, order systems, beliefs, and tools and whose discipline's (non)knowledge production is highly woven into colonialist and nationalist power structures. All this together urges me to reflect on the above questions and consider them in my everyday work. Having laid the groundwork, I will now turn to the enactments of |migration background|.

The category |migration background|: multiplicity and several functions

|Migration background| has different definitions and functions in various fields. I concentrate (1) on the feature of marking alterity, (2) the employment of |migration background| in anti-discrimination contexts, then turning to (3) the further differentiated category by national links, and showing (4) typical hegemonic enactments of the category as a starting point for the discussion of interventions in the last part of the article.

|Migration background|: marking alterity while (re)producing an implicit ethno-national norm

Although there is no standardized definition⁶ for identifying a set of people as having a |migration background| or not, the most influential version of how it is defined and practically coded is that of the Federal Statistical Office, using data from the microcensus. The definition is based on two generations’ lack of German citizenship by birth. ‘Nativeness’ (Bosniak 1994; Geschiere 2009) is granted only to individuals whose both parents are already German by birth. This is a scientific convention based on a common-sense understanding of (‘pure’) German descent. In fact, the expansion of German citizenship law in 2000 to include conditions for ius soli citizenship for children of long-term resident foreigners contributed to the development of |migration background|. Keeping with the dominant ius sanguinis logic is a precondition for the interest in making Germans statistically visible whose parents were foreigners. To differentiate Germans with German ancestry from those without questions about one's parents’ citizenships were included in the microcensus (Deutscher Bundestag 2004).

At the same time, the intergenerational possession of German citizenship by birth evolves as a national standard and reference category. However, only the deviance from the imagined norm is marked and articulated in the definition of |migration background|, leaving the norm unmarked and implicit (Zerubavel 2018); no definition of persons without |migration background| is needed. The decisions about defining and understanding |migration background| are political, and built into the definition (Grommé and Scheel 2020). Thus, the operationalization never included German refugees from Eastern (and other) territories due to the changed German borders after WW II in the category ‘person with |migration background|’, even if they immigrated to Germany's current territory.

In 2016 the Federal Statistical Office decided to exclude foreign-born German citizens from the category of persons with |migration background| if their parents also had German citizenship by birth. These decisions imply that a particular understanding of German ethnicity rather than the act of immigrating is the core feature of |migration background|. It leads to the paradox that, on the one hand, the German state officially denies that ‘ethnic data’ using self-identifications have been collected since WW II and should be collected in the future. On the other hand, the category |migration background| is widely used as an implicit but supposedly objective proxy for ethnicity.

Haraway points out that the God-like-viewpoint called ‘objective’ is also situated (Haraway 1999). To believe that it is more objective to hand the decision regarding who is counted in or not to experts ignores the situatedness of them. As these experts construct ‘equivalence principles’ (Desrosières in Horvath 2019, 561), defining which combinations of features should be sorted to which cell of a given table, the experts are also limited and swayed by their taken-for-granted knowledge and beliefs about national belongings. Statisticians of the Federal Statistical Office followed their common-sense assumptions, some political considerations, and other scientists’ former work in educational research⁷ (Konsortium Bildungsberichterstattung 2006; Statistisches Bundesamt (2007) 2009, p. 329) while designing and revising⁸ |migration background|. By doing so, they included non-immigrants (so-called ‘second generation’) and excluded some immigrants (if they and their parents had been German citizens by birth) from a concept that – on the surface – is based on migration. The borderline between those with and without |migration background| is grounded in nationally-specific notions about who is perceived as ‘native’ and who is not. ‘Who sheds and who retains their migrancy is often bound up with nationally specific ways of encoding and remaking of race’ (B. Anderson 2019, 8).

Therefore, the category |migration background| operates as a category of alterity (Renard 2019) and (non)belonging rather than mobility. The mechanism of ‘choosing the salient features’ which define ‘membership in the polity at that moment in time’ (Yanow 2003, 22) explains why census data is crucial in the creation of the national imagination and narratives about national identity (Yanow 2003, 22). Population statistics is a tool for performing national boundary work (B.R.O. Anderson 1991; Wimmer 2013), and census tables mirror official grand narratives, which pick out certain features and suppress others (Yanow 2003, 22). But these narratives are challenged by basic facts of German history.

The category |migration background| in its statistical and wider societal use is a means of boundary work defining who is ‘in’ and who is ‘out’. Understanding this explains why |migration background| is also employed to measure ethnic inequalities and underrepresentation.

|Migration background| as a marker for ethnic inequalities and underrepresentation

Due to the refusal to collect data on ethnic grounds, the actual state's data collection on |migration background| is framed as data on migration, not ethnicity (Bundesministerium des Innern, für Bau und Heimat 2020). However |migration background| is used for measuring ethnic inequalities. As immigration stabilizes as a person's feature (B. Anderson 2019; Renard 2019) and even becomes inheritable by descendants |migration background| stands for ethnicized populations. Furthermore, it serves to indicate the inherited nationally defined links of persons with and without |migration background|.

The entry of |migration background| in the discourse about equal opportunities and making institutions more ‘interculturally open’ was initially seen as an improvement. Shifting to the alternative |migration background| was supported by the field of anti-discrimination (Salazar 2010). Attention was drawn to data collection on discrimination because it can be used as evidence in court and for positive measures, i.e. measures which, under specified conditions, privilege particular groups where there is evidence of structural underrepresentation (Beigang et al. 2017, 71 ff.). In these contexts |migration background| serves as an approximation for unequal access and resources.

Meanwhile, actors in the anti-discrimination field do not deem the concept of |migration background| adequate for measuring discrimination (Ahyoud et al. 2018, 9). On the one hand, not all persons with a |migration background| have to deal with othering experiences due to racialization. On the other hand, not all racialized persons fit the definitions of |migration background|. For this reason, some speak of a ‘detectable’, ‘visible’, or ‘ascribed’ |migration background| to differentiate the statistical category from a socially relevant ascription in situations of othering.⁹ Given that it is still relatively uncommon for research in Germany to allow survey respondents to identify their own ‘ethnic’ denomination, the established category |migration background| is also being used in anti-discriminatory contexts.

The Federal Ministry of Interior, Construction, and Community plans to continue data collection based on |migration background| permanently. It presented this as part of its anti-discriminatory and even anti-racist efforts in the ‘Maßnahmenkatalog des Kabinettausschusses zur Bekämpfung von Rechtsextremismus und Rassismus’ [catalog of measures of the cabinet committee for combating right-wing extremism and racism, own translation] (Presse- und Informationsamt der Bundesregierung 2020, 3 (measure 7)). To use |migration background| in these contexts only makes sense if the category is perceived and understood as a marker of ethnic descent and ‘race’.

But mobility is not problematic in terms of discrimination; only ethnicized or racialized mobility is. Bridget Anderson gets to the heart of the matter when she writes: ‘once migration is no longer at the border, it becomes ‘race’’ (B. Anderson 2019, 8). The biographical fact of having crossed an international border is already a very constructed and imagination-loaded information (Amelung, Scheel, and van Reekum 2024). This information is attached to a person's description, making her an ‘immigrant’ or ‘foreign-born’ for her whole life. The ‘race’ feature is evident when this ‘immigration’ is handed over to descendants, as is the case in |migration background|. It becomes especially clear in the selectivity to mark certain immigrants and the logic of inheritance of certain immigration experiences. Here ethnicized and racialized logics are at work in defining ‘us’ and ‘them’.

Furthermore, the immigration issue is getting more complex because of ‘borders moving over people’ (Brubaker 2010, 75). The German territories changed after WW I (dismembering of Alsace-Lorraine, Danzig, East Belgium, Hultschiner Ländchen/Hlučínsko, South-Jutland, Upper Silesia, Western Prussia), from 1938 to 1945 (expansion of Nazi Germany and loss of former Eastern territories), in 1957 (integration of Saarland into the Federal Republic of Germany), and in 1990 with the joining of the German Democratic Republic. Many people were at the time of their birth ‘born abroad’ or were not at that time ‘born abroad’, but later after they immigrated to the territory of the Federal Republic of Germany. Even if respondents did not experience the changes of borders themselves, as the questions also concern their parents, that information is kept relevant.

|Migration background| as a marker for national links or ethnic descent

The Federal Statistical Office sorts ‘persons with |migration background|’ by their assumed national ties. The employed reference changed from ‘citizenship’ to ‘country of birth’ due to the harmonization of population statistics in the European Union. Since 2017 the country of birth of the surveyed and their parents have been recorded. Until then, information on (former) citizenships was collected. The new questions inquiring into the current national territory location of respondents’ birthplaces are controversial. Older respondents explicitly criticize them, as I observed during my fieldwork in a statistical office collecting the data. My colleagues who conducted interviews stressed that they regularly get into discussions about former and recent German borders with respondents born in former Eastern German territories. These respondents found the questions insulting and saw them as questioning their belonging to the German nation.

This observation clearly illustrates the sociotechnical aspects of (non)knowledge production in migration research: some respondents question the technical performance of data collection according to EU-European requirements in the German context with changed borders since WW II. They find the procedures intrusive as they feel socially excluded from their felt German belonging by naming their birth country not as Germany but using the denomination of current states. Not including these people in the category of persons with |migration background| although they immigrated to the more recent territory of the Federal Republic of Germany, is a move for not excluding them from an imagined German core. Because statisticians at the Federal Statistical Office were aware of them, they based |migration background| actually not on immigration but on citizenship by birth.

While |migration background| continues to be determined using the logic of (former) citizenships, the newly collected data on countries of birth are employed to continue the distinction by national ties. The Federal Statistical Office displays for persons with |migration background| a ‘country of birth, respectively country of parents’ birth’ (Statistisches Bundesamt 2020, 4). This leads to a new sociotechnical side effect or collateral reality (Law 2011): All persons with a |migration background| born in Germany are listed under the country of birth of their parents if both parents were born in the same country (Statistisches Bundesamt 2020, 68). If one parent is born in Germany and the other abroad, the descendants are listed under the foreign country of birth of the parent born abroad. If both parents are born in different foreign countries, the country of birth is ‘undetermined’ instead of Germany. This procedure hides the connections to Germany of persons often referred to as ‘second generation’. Their country of birth (i.e. Germany) is substituted by the foreign countries of birth of their parents or even replaced with ‘undetermined’. Implicitly this constructs respondents’ links to their ancestors’ national descent and weighs these stronger than those to Germany, thus migrantizing them.

At the same time, there are people categorized as not having a |migration background|, although they were born in a foreign country. Since these foreign links of ‘persons without |migration background|’ are not presented, their existent mobility is silenced. Categorizing in one direction or another based on German citizenship by birth for two generations (re)produces the image of a sedentary German core alongside the implicit, descent-based logics of the category |migration background|.¹⁰ This logic turns out from the construction of the tables in the publication of results from the microcensus (Statistisches Bundesamt 2020, 68).

On the one hand, it seemed logical to the colleagues in the Federal Statistical Office to categorize people born in Germany under foreign countries of birth if they have a |migration background| (7.6 million people in 2019 (Statistisches Bundesamt 2020, 68)), prioritizing their parents’ information over their own personal association with Germany. On the other hand, it also seemed straightforward not to name foreign countries of birth for foreign-born persons if they and their ancestors were German citizens by birth (around 120 thousand in 2019 (Petschel and Will 2020, 85)). As evident from the numbers, the second case is rare, but this is also due to the fact that German refugees are excluded by definition (Petschel and Will 2020, 85). Furthermore, the collected data does not allow identifying everyone who might fit into the second category.

In the spirit of statactivism, such issues are put centerstage and not left to a footnote. The mentioned article makes the subgroups visible. The 120 thousand would not be enumerable without the effort the colleague from the Federal Statistical Office invested in it. The aim was to illustrate the complexity of |migration background| and give a numerical basis for the discussions of an expert commission. These additional numbers destabilize a notion of |migration background| as straightforward, making the construction work visible and thus debatable.

The article is, therefore, a means of statactivism. Given that it was successfully published at the time of the commission's work (which is and was not sure for several reasons), it could be and was considered by the latter (Fachkommission Integrationsfähigkeit 2020, 223). The commission recommended abandoning |migration background| and employing the category ‘immigrants and their descendants’ in the future. This category should be based only on immigration since 1950, and ‘descendants’ must have two parents who immigrated since 1950. However, the commission skipped a decision on one crucial point, and in doing so, it shifted it to the responsibility of the Federal Statistical Office: how to name the residual category of ‘natives’ which results from the differentiation of ‘immigrants and their descendants’. The latter adopted the recommendations and started a new data publication for ‘persons with an immigration history’ naming the residual category ‘persons without immigration history’.

Up to now (November 2023), there has been no published critique. However, the new categories produce collateral contradictions. One example are immigrants who immigrated before 1950. They are ‘persons without immigration history’ despite their immigrant biography. The new categorization represents immigration better because it does not rely on German citizenship-by-birth anymore but on the cost of ignoring older immigrations and the historical continuity of human mobility in Germany.

Putting naturalized ways of data production and presentation into question and searching for new forms of enactment takes effort. Additionally, it needs some theoretical groundings as developed in STS, tools like statactivism, and serendipity to head for other collateral realities. So I did notice the described procedure regarding the ‘country of birth, respectively country of parents’ birth’ incidentally and not while working with these tables as I often do. It came to my attention while finishing a report where a colleague insisted on a chart showing the ‘countries of origin’ of persons with |migration background|. I refused to create such a chart because it enacts and stabilizes methodological nationalism.

However, a junior colleague was then ordered to produce it, turned to the homepage of the Federal Statistical Office, downloaded the data, and plotted a pie chart. I wondered about the graph as I know the data and share of the so-called ‘second generation’ (roughly a third of persons with |migration background|). The information in parenthesis is part of the linking and coordination in producing and enacting taken-for-granted (non)knowledge. Researchers can agree on any ‘fact’ by ‘looking at the data’ if they recognize expected results based on already accepted knowledge. Knowing the ‘fact’ that ‘one-third is second generation’, I wondered why ‘Germany’ did not show up as a part of the pie(chart). It made me look at the tables and discover the above-described procedures. The veiling of respondents’ birth in Germany to highlight their foreign links as people with a |migration background|, often framed as ‘roots’ or ‘descent’, is a common practice. It is problematic because it distorts the data and ignores people's links to Germany. In the next section I will give some examples to illustrate the broader dynamic where one can observe the consequences and the stabilization of this kind of (non)knowledge-producing enactments.

Some examples of hegemonic (non)knowledge production

My first example stems from public health research and measures subjective social status. The questionnaire has been adapted to account for ‘persons with |migration background|’ by asking for a rating on a visualized social ladder regarding one's social status in the ‘country of residence’ and in the ‘country of origin’ (Schumann et al. 2019, 57) (The questions are displayed in Figure 1.). But to ask for the subjective social status in one's ‘country of origin’ makes no sense for persons with a |migration background| if they were born and probably grew up in Germany.¹¹ As evident from the article's abstract and keywords, it deals with immigrants and their descendants, collapsed in the category |migration background|. From the title of the figure, it follows that the question is to be answered by all survey participants with a |migration background|, ignoring or at least not considering that surveyed descendants of immigrants have not moved from any ‘country of origin’ to Germany. The same is true for most participants lacking a |migration background|.

Figure 1. Adapted version of the MacArthur Scale […] for people with a |migration background|.

Source: Schumann et al. 2019, 58.

Because the category |migration background| reproduces the logic of foreign ties beyond one's own immigration, it also functions as a marker for ethnicity. The ethnicizing effects of this mechanism of non-belonging become all the more exacerbated the more a |migration background| is differentiated into subgroups. This becomes evident in two further examples where people with |migration background| apparently ‘are from’ or ‘originate from’ other countries (Olczyk, Will, and Kristen 2014, 20; Pfündel, Stichs, and Halle 2020, 14 f.), which does not apply to persons constructed as ‘descendants of immigrants’ born in Germany. The consequences can also be traced in an annual report of the Expert Council for Integration and Migration (SVR 2021). In the report a ‘duration of residence’ is visualized for the 35.9 percent of persons with |migration background| who were born in Germany (Figure 2). However, it is missing for persons without |migration background| who, if polled in the microcensus, mostly but not exclusively resided in Germany since birth, too.

Figure 2. Persons with |migration background| 2019 by time of residence (in percent).

Source: SVR 2021, 23.

In my view, these three examples show an unmindful enactment of (non)knowledge production, which echoes many everyday experiences of exclusion by people who grew up in Germany, as shared under #vonhier. This exclusion in everyday life is not surprising if even experts in migration research do not correctly display people's country of birth but rather (re)produce ethno-national exclusions.

Notwithstanding, data on a nationally differentiated |migration background| is also utilized by anti-discrimination activists in order to present evidence of disadvantages for certain groups (Diakonie 2015, 12). Doing so, they find themselves in a dilemma. On the one hand, they lobby for voluntary, self-descriptive categories which abandon the logic of |migration background| and are sensitive to discriminatory dynamics. On the other hand, they know that any ethnic data – even the explicitly collected ethnic data they lobby for – will reify ethnic categories (Supik 2014). So, how can we deal with this kind of dilemma?

Some personally tried destabilizations

Questioning particular hegemonic stances is not that complicated if taken-for-granted certainties are systematically questioned. As you see here, one can even produce text about it. Approaches from STS offer helpful tools for a systematic inquiry in individual knowledge production and involvement in thought collectives. Statactivism supports going further in challenging the production of (non)knowledge. Bruno, Didier, and Vitale suggest establishing new ‘kinds of people’ or challenging established ways of depicting society numerically. Creating new, alternative, or other representations is part of an enactment that strives for less damaging, and therefore, more ethical categories and their use. The goal is not a better representation of any reality. Because reality is enacted and produced, it is already changed by counter-representations. However there are more ethical categories and enactments which consider or serve the needs of the categorized and challenge power hierarchies. Of course, this is a never-ending process, and the intervention only produces new collateral realities. Therefore, there is no control over the consequences of the invention of new ‘kinds of people’. But if the goal is to destabilize hegemonic (non)knowledge production, to pluralize the data on migration and subsequently to de-migrantize migration research then every critical intervention will challenge the hegemony. It is, in my view, the only way for responsible, reflexive, and non-positivist (non)knowledge production.

Here is one example where I tried the latter, i.e. to de-migrantize a certain ‘kind of people’. I mentioned above that the concrete use of the category ‘persons with |migration background|’ hides the connections to Germany of certain descendants of immigrants. To reverse it and deliver the required pie chart, I reorganized the numerical data graphically to make the numerically most important country of birth of persons with |migration background| – which is Germany – visible (see Figure 3).

Figure 3. Countries of birth of persons with |migration background| 2018.

Source: Baumann et al. 2019, 10.

The graph itself did not change much, but the whole episode gave me more data and evidence of the underlying patterns of nativism and methodological nationalism. It is an illustration of the power of hegemonic ways of data presentation. In the end, there was the required graph on ‘countries of births’, even if it was not the standard way, because it did not use the common ‘copy-and-paste’ procedure. Chripa Schneller developed a scheme to classify resistant actions as persisting resistance or reframing (Schneller 2022). In that classification, the graph itself is a way of reframing, but to write about it, making the power structures behind the graph public is a means of persisting resistance and challenges hegemonic enactments.

Additionally, as part of my research, I scrutinize how statistics on people without and with |migration background| are represented in the official tables and share my findings with academic and non-academic publics alike. In Petschel and Will (2020), we deconstruct the category |migration background| into its ‘building blocks’ and show how results would vary for different definitions of |migration background|. In doing so, we made the contingency and constructiveness of these results more visible and could thus enter into informed debates on them.

Focusing on the ‘building blocks’ allows looking into the black box of derived variables like |migration background|. However, it is restricted to experts and far from everyday life. But it is not limited to academic researchers alone, as my colleague’s interest, collaboration and invested effort in the statistical office shows. Without her the article would not exist. Statisticians often see themselves as service providers. Academic and public number crunchers will get several parallel statistical representations if they ask for them. An example is the recently started statistical report on ‘Population with immigration history’ which will be published additionally to the classification of the ‘Population with migration background’ (Statistisches Bundesamt 2023). There has to be only a demand for it. The effort to design new plausible categories is the effort statactivists must invest.

In another text, I included a poster series from 2012, in which the civil society organization DeutschPlus had reinterpreted |migration background| in a creative manner (Figure 4). They produced posters advertising for a wallpaper with a huge family photograph on it, displayed in a living room set-up. The text reads: ‘Do you sometimes feel excluded because of your monotonous family tree? Have you also always wished for an exciting migration background? Then order the new wallpaper from DeutschPlus today. Bring atmosphere (of departure) into your four walls. Choose from different backgrounds: Whether Anatolia, Vietnam, Iran, Lebanon, or Nigeria. We are sure to have the right migration background for you. We promote young talents with exciting backgrounds with the proceeds, so they come to the foreground even more in Germany’.¹² This artistic intervention points to different issues at once: family trees as means of inclusion and exclusion, exoticization of people with immigrant ancestors, and lacking representation of migrantized persons described as having a |migration background|.

Figure 4. Poster series ‘Migration background for all!’.

Source: Will 2020.

After getting permission to use the poster images, I had to struggle with the editors,¹³ to keep Figure 4 in that article. They did not explain why they would not like to include it. But as the aim was also to use the text in political education classes in upper schools, they might have been positive about the permitted use of posters in that format. For me, however, it was important to include this humorous reversal of the stigma/mark (Goffman 1963; Zerubavel 2018) of having a |migration background|, implying that it would be a deficiency if one does not have it. This aligns with the statactivist demands to use artistic tools and Zerubavel’s ‘subversive semiotics’ against taken-for-granted certainties and epistemic stabilizations (Zerubavel 2018, 60 ff.).

The advertising rhetoric of the figure foregrounds |migration background| as an important asset. Simultaneously, the importance of a |migration background| is questioned through its equations with wallpaper. Nevertheless, the advertisement addresses the underrepresentation of immigrants and their descendants in the short explanation text at the bottom. Even if I had to convince the editors, it was clear to me that I wanted to include the posters because they provoke questions of belonging in another way than the statistical definitions and representations, but both are linked with each other. As the target group is a politically interested public, I was clear about the fact that the text would not get published without the posters.¹⁴

‘Marking the unmarked’ is one crucial strategy of subversive semiotics. The invention of the ‘cis’-gender label is a brilliant example since it explicitly names the implicit norm of gender congruence. In my field, the absence of definitions for persons without |migration background| is striking, and I find it crucial to elaborate on this side of the category (which is applied to a much larger group) with at least the same intensity as the marked side of the scientifically constructed dichotomy of |migration background|. I try so in my writings about |migration background| and my critique of the category as misleading and excluding (non)knowledge production. Additionally, I urge to counter-read, re-arrange or disaggregate data or sometimes just ask questions to make controversial inscriptions also visible to research colleagues.

The demand for categories other than |migration background| to collect anti-discrimination data is also a form of statactivism. If future statistics fulfill this demand and serve its intended aims, they might be comparable to ‘femostats’, a term coined for feminist statistics (de Rosa 2014, 314). Also, in the case of femostats, ‘NGOs intervene[d] in gender statistics and gender equality policy’, and in doing so, they ‘strategically produce and use categorizations and statistics’ (de Rosa 2014, 318). In my vision, researchers must align with NGOs to support their claims and make them heard in academia, especially in hegemonic circles, population statistics, and further relevant fields. Of course, this does not precisely create career advantages, but does a researcher’s responsibility not extend beyond pure (non)knowledge production for its own sake? Especially in migration research, the topic is people. Becoming aware of the impacts, possible dangers, and side-effects of the concepts one employs is crucial to use them ethically and carefully or even abandon them for slightly better ones.

In the end, changes have to come to a significant share from within academia employing the same means of data production, i.e. categories, numbers, and statistics. There is no other similar powerful site for producing (non)knowledge and enacting realities. The production of knowledge has to be given significantly more notice. Additionally, reflecting of its interactions with the general public is crucial for not isolating and losing the sense of its impacts on people's lives.

Conclusion

|Migration background| was designed by the Federal Statistical Office and employed by the German government to measure ethnically framed inequalities. Mainly if further differentiated by nationalities, |migration background| is also used by anti-discrimination activists, who lobby for more specific data on ethnicity and racialization. As shown, |migration background| is a means of nativism and contributes to methodological nationalism. It is employed to describe a social structure as ethnically stratified according to residents’ ancestral nationalities. As long as scientists and administrations persist in using the category and continue to enact this kind of (non)knowledge production, they undermine the demands for anti-discrimination data. These are: to employ more suitable ethno-racial variables, engage categorized persons in a participatory manner, and value self-identification. However, even if the latter dispute the term's validity, the ongoing use of |migration background| increases its factuality.

To rely on the enactment agenda allows researchers for critical reflections on personal (non)knowledge production. Social scientists must resist the positivist assumption that a category represents anything in a given reality and replicate any category with no explanation than ‘that is the way it is’. In contrast, they should act with caution and responsibility in their field of societal descriptions and explanations, minding the productivity of identity categories and keeping the contingency visible. This aligns with the reflection of one's positionality and a decided denial of stereotyping answers and mainstream assumptions. All scientific enactment remains ambiguous and contested as long as it does not melt into common sense (non)knowledge.

If migration research considers STS approaches in combination with forms of statactivism it can challenge the excluding ethno-national, culturally charged conceptualization of |migration background| and its usage in the future. Furthermore, it can make a worthwhile contribution toward establishing new, more suitable categories for measuring and monitoring discrimination on ethnic and racial grounds and other categories alike. These categories should be kept emancipatory and resilient to (scientific or other) taken-for-granted standards or aims and destabilize ethno-racial (non)knowledge production in migration research. Categories should be used to cure social inequalities while being highly alert to their poisoning abilities.

Acknowledgment

This work was supported by the Deutsche Forschungsgemeinschaft [DFG-Projektnummer 410871136] and the International Centre for Advanced Studies: Metamorphoses of the Political ICAS:MP funded by the Federal Ministry of Education and Research.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Bundesministerium für Bildung und Forschung through the International Centre for Advanced Studies: Metamorphoses of the Political ICAS:MP; Deutsche Forschungsgemeinschaft [grant number 410871136].

Notes

1 I write the term |migration background| within vertical slashes to highlight its constructed and computed nature.

2 I thank the anonymous reviewer and the editors for their valuable comments and suggestions. The final version has profited greatly from their knowledge and supportive way of making me think further. Once again, I am thankful to my copyreader Melisa Salazar. The writing of this article was enabled by funding from the German Research Foundation (Deutsche Forschungsgemeinschaft) with funding number 410871136 and a fellowship at the M.S. Merian – R. Tagore International Centre of Advanced Studies ‘Metamorphoses of the Political: Comparative Perspectives on the Long Twentieth Century’ (ICAS:MP), an Indo-German research cooperation funded by the German Federal Ministry of Education and Research (BMBF). All views expressed here are solely those of the author.

3 The Federal Statistical Office translates ‘Migrationshintergrund’ as migrant background. I prefer the more literal translation |migration background|.

4 However, I also did systematic analyses, for instance, on the wording in PISA reports (international, German, Swiss, Austrian, and Liechtenstein) and the reports on children and youth in Germany, census questionnaires, microcensus questionnaires, and the publication ‘Population with migration background’ of the Federal Statistical Office.

5 There will never be purely good ones since categories are always Janus-faced, valorizing, and silencing at once (Bowker and Star 2000).

6 The only binding definitions exist in area-specific laws and regulations like children in daycare and unemployed persons, and in laws of three federal states on the participation of persons with |migration background|. All definitions differ from each other, which is part of the multiplicity of the category (Mol 2002).

7 Horvath shows how ‘migration background’ is an ethnicized category in educational research (Horvath 2019). I would add that the national imagination, including perceptions of who belongs to the nation or not, is a shared one and not specific to educational research.

8 Until July 2022, there were two revisions of the wording of the definition and one major revision of the operationalization of |migration background|. The Federal Statistical Office did not officially communicate, discuss, or document these revisions but just implemented them. Knowledge about it is thereby limited to insiders.

9 This echoes findings in the US about the difference between ‘street race’ (López et al. 2018) or ‘ascribed race’ and the official ethno-racial classification or individual self-identification (see Roth 2016 for an overview of different aspects of ‘race’).

10 It mirrors the primary ius sanguinis principle of German Citizenship Law, which holds that only those who have parents with German ancestors are considered German by birth. Since 2000, the German Citizenship Law created a ius soli exception for children born to foreign citizens who have lived in Germany for at least eight years and fulfill some more conditions. Since these children are Germans by birth, only their descendants will get categorized as persons without |migration backgrounds | thus weakening the intense orientation to descent.

11 It is also senseless for foreign-born people who were just born but did not live in their ‘country of origin’ or did not live there very long, i.e., immigrated to Germany at a young age or to another country and then at a later time to Germany (where the survey questions should be administered). These exceptions unveil the dichotomous and simplified assumptions at stake in these kinds of categorizations and questions.

12 ‘Fühlen Sie sich manchmal aufgrund Ihres eintönigen Stammbaums ausgegrenzt? Haben Sie sich auch schon immer einen spannenden Migrationshintergrund gewünscht? Dann bestellen Sie noch heute die neuen Tapeten von DeutschPlus. Bringen Sie (Aufbruch)stimmung in Ihre vier Wände. Wählen Sie aus verschiedenen Hintergründen aus: Ob Anatolien, Vietnam, Iran, Libanon oder Nigeria. Wir haben bestimmt den passenden Migrationshintergrund für Sie. Mit dem Erlös fördern wir junge Talente mit spannenden Hintergründen, damit sie in Deutschland noch mehr in den Vordergrund rücken’.

13 Strikingly this figure of the poster series is not included in the pdf download of that article, which means that if teachers use a printed pdf version in class – which was one intention of the text – the illustrative artistic feature of the poster is not available.

14 Chripa Schneller (2022) conceptualizes it as ‘the power of No’. There is always an exit option. However, it might get difficult with a writing contract and often needs some negotiations, but reasonable arguments cannot be easily ignored. Nevertheless, a risk remains of not being heard and having invested time in vain.