A Socio-legal Approach to Language Use in Administrative Settings in Belgium’s Dutch-Language Area

Abstract This reflective article maps out the findings of a socio-legal study on language use in administrative settings in Belgium’s Dutch-language area. It explains how the author arrived at a socio-legal approach to this study object. Thereafter the article focuses on the methods used and the conceptual framework. It discusses the various methodological choices regarding data collection and the use of methods stemming from anthropology in this study. It also illustrates how anthropological literature on concepts such as discretion and multiple embeddedness provide conceptual tools for building a framework around which to structure and present the empirical data. The last part of the article sketches the findings of this socio-legal study. Drawing on the conceptual framework, it illustrates that even when practice deviates from statutory law, the relationship between the law and the practice cannot be easily captured in a dichotomic relationship. The author furthermore deduces some relevant findings in light of general human rights. The article concludes with some reflections on research that combines law and anthropology, both as separate disciplines and as a combined endeavor.


A. Introduction
Linguistic diversity has been present in many, if not most, societies throughout history. Currently, linguistic challenges in terms of linguistic gaps are not unique to historically multilingual states. Multilingualism is a reality in parts of most European countries, including states with a single official language and states dominated by one language. This linguistic diversity is often multilayered, consisting of historical "national minorities" recognized by a state, "new" minorities that settled in a given state after it had become independent, and other more recent immigrant groups.
Provisions on language use in the public sphere, however, rarely mirror the full scope of this linguistic diversity, as in most cases they were designed in times of a different sociolinguistic reality. As such, these legal norms are not always the most suitable for dealing with the current situation, as they might not provide satisfactory and structural answers on how to deal with linguistic gaps in practice. Although there are indeed diverging normative positions on how states should deal with linguistic diversity-as part of broader debates, for example, on integration and diversity in societies 1 -it is clear that a poor fit between existing legal provisions on language use and the sociolinguistic reality can have dramatic consequences. This is the case in emergency situations, at both an individual level-for example, a medical emergency-and on a larger scale-such as natural disasters or hazardous material spills-when linguistic barriers hamper rapid communication of life-or-death information between administrative authorities and the multilingual population.
The COVID-19 crisis has highlighted the challenges that administrative services face in reaching all groups of society and the linguistic issues related to official communication. A number of news outlets have reported a link between outbreaks of the virus and the failure of official announcements regarding preventive measures to reach non-majoritarian language speakers. 2 This pandemic is just another example of linguistic challenges in practice, as multilingual environments are realities that both administrations and residents struggle with on a daily basis.
The lack of empirical research on this topic motivated my subsequent socio-legal research, which I present below as a "case study" to reflect on my approach to the relationship between law, as a science, and anthropology. It focuses on the interactions between non-majoritarian language speakers and administrative authorities in Belgium's Dutch-language area, which is, needless to say, just one of many areas where these linguistic challenges arise.
After setting out the origins and design of the research project in Part A, I explain how I, as a legal scholar, engage with anthropology on two levels: In terms of research methods in Part B; and conceptually as a framework for the research at hand in Part C. In these parts, I explain how the topic, the methods, and the conceptual framework in this research fit with or deviate from common approaches in law and in anthropology. These parts allow me to reflect concretely on the relationship between law and anthropology and on the dialogue between the two disciplines, in Part D. This article does not contain a comprehensive analysis of the applicable legal framework regarding language use in administrative interactions. In light of this endeavor, I use "law" and "the legal framework" interchangeably in this article to refer to the legal norms within the framework of state law. In those instances where I use "law" to refer to the scientific discipline, I will explicitly inform the reader.

B. Research Topic and Research Design
A twofold interest sparked my attention to language use in administrative interactions in Belgium's Dutch-language area: An empirical interest in the actual course of administrative interactions with non-majoritarian language speakers and in the reasons given for certain practices, and a legal interest in the obligations of states in this regard, including attention to the implications of human rights in multilingual environments. These interests have been fueled by media coverage of numerous incidents and subsequent political responses, as well as by my own personal experiences that have given me cause to reflect on the applicable legal framework.
These interests link with themes and objects of study in both legal anthropology and doctrinal legal research, yet combining law and anthropology in the analysis of the specific issue of language use in administrative interactions represents a new direction. In reflecting on the relationship between law and anthropology, it is first relevant to discuss how the research topic relates to earlier research in the two disciplines.
The legal language of a state, its use, and the course of interactions between ordinary people and state institutions are all well-established themes in legal anthropology. Language-and-law research within anthropology has a long tradition, 3 and lay people's encounters with courts have been a classic topic in legal anthropology. 4 Yet linguistic diversity in everyday administrative interactions has not been the focus of such studies. A lack of empirically based accounts of the recent "social working of legal rules" 5 in the domain of language use in administrative settings is apparent.
The topic of language use in administrative interactions has received considerable attention in doctrinal legal research and in general normative approaches to the topic. Legal research shows that states enjoy a fair amount of liberty to regulate language use in interactions between their administrations and their inhabitants. As such, states have adopted divergent ways in law to deal with historical and contemporary forms of linguistic diversity. Some states impose the exclusive use of the official or administrative language; other states leave more liberty to use other languages; yet others have adopted provisions encompassing rights for speakers of minority languages to the use of their own languages. The legal literature has paid extensive attention to language use in administrative interactions-mostly in historically multilingual states-providing, among other approaches, overviews of language models and analyses of relevant minority rights, such as Article 10(2) of the Framework Convention for the Protection of National Minorities. 6 Nevertheless, dogmatic legal research, with its normative orientation, does not satisfy the interest in, or need for, empirical research on the topic. 7 This research gap is also present with regard to administrative interactions in Belgium, a country characterized by the historical coexistence of Dutch, French, and German speakers. The historical language debates in Belgium centered around the struggle for equal treatment of the languages, a security area for the Dutch language, and the full implementation of the relevant provisions. 8 The significance of language use in administrative interactions in Belgium's Dutch-language area is presently related to three intertwined elements: The historical Flemish language battles; the politics revolving around the demarcation of the internal language border, especially around Brussels; and the current language ideology building on integration. Dutch is seen as the common language and an important part of civic identity in the Dutchlanguage area. The political approach to language entails positive attention to Dutch, and in general a reluctance to use other languages in the public sphere, for example, in administrative contexts and official educational settings.
Language use in administrative interactions in Belgium has received extensive attention in case law and jurisprudence, 9 which have focused on the legality of the legal norms governing these interactions between civil servants and non-majoritarian language speakers. In these cases, 3 Elizabeth Mertz, Within and Beyond the Anthropology of Language and Law, in The OXFORD HANDBOOK OF LAW AND ANTHROPOLOGY (Marie-Claire Foblets, Mark Goodale, Maria Sapignoli, Olaf Zenker eds., 2021). conformity with higher, constitutional provisions and treaty obligations occupies a central position. The European Court of Human Rights, for example, has already held that the Belgian language model, based on the principle of territoriality, 10 is in conformity with the European Convention on Human Rights. 11 This principle of territoriality can be summarized as "the language of the area is the administrative language" in Belgium's four language areas-the Dutch-, French-, and German-language areas, and the bilingual Brussels-capital area-and it is in general applicable to interactions between local administrative authorities and residents in Belgium. 12 The Administrative Language Law indicates that civil servants who circumvent the relevant norms face disciplinary actions and that all administrative actions at odds with this law are null and void. 13 At first glance, it seems that the general principle, "the language of the area is the administrative language" is unimpaired in law. If one looks more closely, however, exceptions to this general principle emerge, for example, in the advisory practice of the Belgian Standing Commission for Language Supervision (SCLS). This advisory body, which monitors the implementation of the Administrative Language Law, 14 allows local administrations to use languages other than the language of the area under certain conditions. 15 With regard to the empirical interest in this topic, there is very limited case law, jurisprudence, or empirical studies that provide insight into the effectiveness of the relevant legal norms in monolingual Dutch municipalities in Belgium. 16 Case law and jurisprudence regarding administrative interactions in Belgium contain only a limited number of indications that run counter to the ideas and legal principles of Dutch-speaking administrations in the monolingual municipalities in Belgium's Dutch-language area. It might be tempting to deduce from the apparent lack of sources to the contrary, that practice is generally in accordance with these ideas and the legal norms. Yet one has to consider that, as in other fields of law and of society, certain practices remain under the legal radar. As such, the need for empirical research is clearly present. This observation regarding the effectiveness of the relevant legal norms is, from a social scientific view on law, just one of a number of aspects that would benefit from empirical analysis. Other such topics include how these practices emerge and how they are reflected in legal and non-legal justifications by the actors involved.
The "case study" presented here focuses on interactions between non-majoritarian language speakers and administrative authorities in Belgium's Dutch-language area. The research design, 10 The principle of territoriality entails that "rules of language to be applied in a given situation will depend solely on the territory in question," whereas with the principle of personality, "rules will depend on the linguistic status of the person or persons concerned." See Kenneth McRae, The Principle of Territoriality and the Principle of Personality in Multilingual States, 13 LINGUISTICS 33 (1975); Patten & Kymlicka, supra note 1.

11
Case "Relating to certain aspects of the laws on the use of languages in education in Belgium" v. Belgium (Merits), App. No. 1474/62, 1677/62, 1691/62, 1769/63, 1994/63, 2126/64, ¶ 7 (July 23, 1968, https://hudoc.echr.coe.int/eng?i=001-57525. 12 The principle of territoriality is softened in Belgium by rights for non-majoritarian language speakers in the so-called "municipalities with language facilities" (facilités linguistiques; taalfaciliteiten). There are 27 municipalities with language facilities located in monolingual language areas in Belgium. In these municipalities a special regime applies to interactions between the administration and its residents, with the local administrations being obliged to use the minority language in certain cases. guided by my legal and empirical interests, led to three central research questions: What occurs during interactions between civil servants and non-majoritarian language speakers? What reasons are invoked to explain the language use in these interactions? What needs are formulated by the actors involved? The answers are partly provided by a consideration of the legal framework. The sub-question of a dogmatic legal nature in this regard is: How do legal frameworks address interactions between non-majoritarian language speakers and public authorities?
To provide the necessary empirical data to fill the research gap identified above, the underlying research borrowed methods from anthropology, more specifically, ethnography. 17 This enabled me to gather and analyze empirical accounts of interactions between administrative authorities and resident non-majoritarian language speakers. The research furthermore considers local language policies, as well as the legal and linguistic needs formulated by the actors involved. The data allows for empirically based insights into the practices that lie beneath or beyond law, in a strict sense, at different levels, 18 as well as of the impact of situations on law 19 and bureaucracies. 20 The answers to the research questions are thus also partly generated by cumulative empirical accounts of interactions between administrative authorities and resident nonmajoritarian language speakers. Additionally, this research connects with social theory and concepts stemming from the social sciences, anthropology in particular, to analyze and interpret the gathered data. 21 Through its twofold response to the research questions, this article combines the "descriptiveanalytical perspective of anthropology with normative dogmatic and pragmatic perspective of legal science." 22 It is largely descriptive, including the relevant legal provisions, the practice, and the reasons indicated by the actors involved. 23 The normative undertaking is rather limited; the goal is not to take a strong normative stance on how the law ought to regulate language use in the administrative interactions. The minimal normative stance I do take here is based on two elements, namely, general principles embedded in the rule of law opposing arbitrariness, 24 and a minimum core of obligations stemming from general human rights. 25

C. Methods
The decision to conduct socio-legal research on the topic of language use in administrative interactions still leaves open several questions of methodology and methods. Very concretely, for example, the question of how to gather the empirical data remains unanswered. Some broad choices in terms of, for example, where to conduct the research, had to be made before the fieldwork started. The focus in this research project was geared toward administrative interactions with French and Turkish speakers in multiple field sites in Belgium's Dutch-language area. In this second part, I discuss the methodological choices I made for this research project and how they relate to common approaches and expectations in the disciplines of law and of anthropology. Through an extended explanation of these choices, I develop these aspects as concretely as possible, for example, with regard to "meaningful" data and fieldwork expectations.

I. Selection of Field Sites
The diversity of local norms, apparent in the media and in initial exploratory contacts with experts, led to a focus on several services in multiple field sites instead of one field site, in order to discover the source of the normative diversity. This choice was also initiated by the input from the legally dogmatic-oriented audience that found a focus on one municipality or one administration too narrow and too restricted to the particularities of a single municipality to be meaningful for law.
Regarding the selection of the field sites, I sought out places where linguistic needs might occur, where enhanced minority rights could be applicable, 26 and where language use in administrative interactions is a matter of debate. This was a two-step process. First, I developed a set of profiles for the municipalities in the Dutch-language area in Belgium, according to broader, legally significant, differences. Six profiles were chosen, namely, the periphery around Brussels-that is, the ring of Flemish municipalities around Brussels-the province of Flemish Brabant, outside the periphery around Brussels; municipalities at the language border; cities responsible for implementation of the Flemish Integration Decree; 27 a touristic center; 28 and municipalities with a considerable number of Turkish speakers. Municipalities with language facilities for French speakers 29 in the Dutch-language area were not included, both because a different legislative body-the federal level-holds competence in this regard, 30 and because of the rights French-speaking inhabitants of these municipalities already enjoy under the Administrative Language Law. 31 Second, within the broader profiles, the further selection of specific municipalities was based on criteria such as the size of the language groups, information on local language policies, and prior language-related incidents.
Primary legal sources, reports in the media on language incidents, and exploratory talks with academic and administrative experts working on this issue were used to select specific municipalities. The selection of field sites for Turkish speakers was also based on statistics on the number of Turkish nationals and persons with a Turkish migration background, while bearing in mind the potential over-and-under-inclusiveness of these indicators. 32 Several municipalities within these profiles fulfilled these criteria, which I narrowed down to eleven field sites in Belgium: Two for each profile except for the fifth profile, for which only one field site was selected. This selection covered all five Flemish provinces. 26 See VENICE COMMISSION, Report on Non-Citizens and Minority Rights, 69th Sess., Doc. No. 294/2004(2007 These criteria can be over-inclusive, as they might include persons who speak and understand Dutch perfectly. They might also be under-inclusive when compared to people's self-identification.

II. Administrative Services
The selection of administrative services was based on the different characteristics of these services, the relevant legislation in Belgium, and the scope of Article 10(2) of the Framework Convention for the Protection of National Minorities. Characteristics that were taken into account relate to several aspects of an administrative interaction: An initiative by an administrative service or by an inhabitant, the urgency of the interaction, and the presence of a certain target group. These characteristics were chosen according to-potential-variances in the legal framework regarding these criteria. The selection of services aimed at verifying whether there are differences in practice according to these characteristics.
These criteria resulted in a selection of services and situations ranging from highly formal to almost informal, of differing duration and frequency of interactions, and exhibiting diverse backgrounds of the actors involved-for example, in terms of educational level and language competences. 33

III. Non-Majoritarian Language Speakers
Given the research focus on practices and needs identified by the actors involved rather than onpolitical-discourses, I looked for "ordinary" persons residing in municipalities where the administration was willing to participate in this research.
The large group of non-majoritarian language speakers was narrowed down to two groups. The selection aimed at including a potentially "old" minority-French speakers in the Dutch-language area in Belgium-and a "new" minority-Turkish speakers in Belgium-the members of which might enjoy different levels of national and international legal protection. 34 Further criteria were that the situation of the group needed to be a matter of debate, and that access to the group for research purposes was possible. The selected groups had in common that they have been the subject of discussions with regard to administrative language rights and/or their possible protection by the Framework Convention for the Protection of National Minorities. Belgium has signed but has not ratified the Framework Convention, and debates on the national minorities in Belgium are still not fully resolved.
French speakers in Belgium's Dutch-language area have, in general, 35 no right to expect or demand that French be used in administrative interactions; the use of Dutch is a legal obligation in their contacts with administrative services. 36 Yet French speakers in certain parts of Flanders have long-standing ties to the territory and thus some argue that they could qualify as a national minority. 37 French speakers have settled more recently in other areas in Flanders, leading some to argue that the French speakers are a rather "new" minority-in these areas. These French speakers are, furthermore, an interesting object of research, as their numbers are augmented by Frenchspeaking immigrants from, for example, African countries, giving rise to legal questions regarding the extent to which these persons are covered by national and international interpretations of the concept of "minority" and whether they can or should be able to enjoy the protection of the relevant provisions. 33 This selection included disaster/emergency communication, municipal registration offices, waste collection, local ombudsmen, and public social assistance centers. 34 For the contested distinction between "old" and "new" minorities, see Asbjorn Eide, The Rights of "Old" Versus "New" Minorities, 2 EUR. Y.B. MINORITY ISSUES 365 (2004 In addition to French speakers, Turkish speakers in Belgium were included in the study for three reasons. First, the considerable size of this group in comparison to other "new" minorities in Belgium. Behind the statistics based on nationality and migration background, however, there is a linguistically heterogeneous group, including, for example, Turkish and Kurdish speakers. Second, within Belgium, the group comprises several generations, dating back to the 1960s. Belgium only rather recently started emphasizing the-linguistic-integration of Turkish newcomers. Third, their long-standing presence has led to a situation in which people with a Turkish migration background are now in positions in public administrations, 38 which opens up the possibility of pragmatic solutions in relation to Turkish-speaking service users. Language use in administrative interactions with these two groups is still a matter of debate. As highlighted above, language use and the implementation of language regulations in administrative settings are part of long-standing discussions in Belgium. The animosity swirling around the situation of French speakers in the Dutch-language area is evidenced by several legislative initiatives at the federal level, as well as the advocacy work of the Association for the Promotion of French Culture in Flanders (l'Association pour la Promotion de la Francophonie en Flandre, APFF) at the international level. 39 There are also several empirically documented indications of the need for administrative authorities in Belgium to use Turkish with residents. Several of the SCLS's advisory opinions have been issued with regard to the use of Turkish by public authorities. 40 Moreover, during the pre-fieldwork phase, Turkish was one of the languages in greatest demanded among telephone interpretation services for administrative interactions in the Dutch-language area. 41

IV. Description of Methods
As a consequence of the choice to conduct fieldwork in multiple field sites and with diverse kinds of informants, the data was gathered through interviews, observations, and the collection of written sources between the autumn of 2015 and the summer of 2018. The collected written sources comprise local language policies and documents issued by administrations, including leaflets and municipal websites, as well as individual written interactions between civil servants and residents. Additional data includes observations on elements surrounding these administrative interactions, such as signs and the spatial organization of public buildings and meeting rooms. I further attended training sessions for frontline civil servants on the theme of communication with non-majoritarian language speakers.
Much of the empirical data were generated via interviews with the involved parties. Interviews were most suited for the ambitious aim of covering several administrative services in multiple field sites and examining the diversity of local policies and practices within a relatively short time period. 42 The nature of administrative encounters and the work of civil servants in interaction with non-majoritarian language speakers are not conducive to the classic ethnographic method of long-term participant observation, as they tend to be social interactions that are "spatially 38 See Diversiteit stadspersoneel naar herkomst, VLAAMSE STADSMONITOR (2015)   dislocated, time-bounded and characterized by intimacy at a distance." 43 The approach followed in this study through these interviews is comparable to what has been labeled in the literature "anthropology by appointment" 44 or "ethnography by appointment." 45 The interviews were face-to-face and semi-structured, featuring open-ended questions 46 with both civil servants and persons belonging to the selected language groups in multiple field sites. The data generated an insider's perspective-from the point of view of both civil servants and non-majoritarian language speakers. The interview topics were general and referred to situations without explicitly mentioning specific legal concepts. If an interviewee brought up legal concepts, I would explore the meaning the interviewee attributed to them. The number of interviews in a field site differed, depending on when "informational saturation" was reached, that is, the point at which additional data collection produces no new information. 47 Several informants approached me as an expert, asking me questions on how to deal with linguistic diversity, or as an ally who could improve their situation. Some civil servants, especially supervisors, showed an interest in the results, stating that they wanted to improve the functioning of their services.
Access to the administrative services was sought through the head(s) of the administration. In total, twenty-two heads of administration in eleven Flemish municipalities were contacted. Supervisors suggested names of civil servants, but for the most part I was able to speak to whomever I wanted. When supervisors recommended names of civil servants, they often included persons with allegedly different perspectives and approaches in practice. Civil servants' remarks on the local language policy, their supervisors, and the practices were, in general, frank. 48 The semi-structured interviews with civil servants were conducted in Dutch. Topics during the interview examined their experiences with certain situations, their knowledge of the legal framework, and their needs with regard to interactions with non-majoritarian language speakers. During the interviews I assumed a stance of ignorance, although I was asked on several occasions about the exact meaning of the relevant provisions in the law.
Within the selected groups, one criterion for identifying informants for the purposes of this study was having a self-declared "home language" 49 other than Dutch-that is: French or Turkish. I sought out these informants through relevant organizations such as French-or Turkish-speaking sport clubs, youth organizations, and socio-cultural associations, and then expanded the pool using the "snowball sampling" method. In my interviews with non-majoritarian language speakers, it was difficult to gather much information specifically on the topic at hand due to informants' limited number of interactions with administrative services. Moreover, the empirical findings within the administrative services indicated-or rather confirmed-at an early stage of the research a linguistic diversity far beyond these groups. In light of the multiple field sites in combination with principles of the para-site technique used by others for ethnographic research in bureaucracies, 51 the focus was, in the end, on informants who had a good overview of the situation. Non-majoritarian language speakers thus rather fulfilled a control function with regard to the data gathered with administrations. 52 Making such on-the-fly shifts in research design in light of the data collected is common practice in social scientific approaches to law. 53 The interviews with non-majoritarian language speakers were face-to-face and conducted in the language preferred by the informant-calling on the services of an interpreter when necessary. The scope of the questions was broader than those for civil servants, as it aimed at all their contacts with administrative services. The topics of the interview were similar to those for civil servants, including their experiences and needs regarding their interactions with administrative services.
At the beginning of each interview, I explained the research goals and the confidential nature of the interview, after which informants signed a consent form. The interviews were audio-recorded, and I also took additional notes during the interviews. The informants' personal data were treated as confidential, and the empirical accounts are presented so as to avoid identification of the informants in publications related to the research. Regarding local language policies, I likewise prevented identification of the civil servants by not mentioning the name of the relevant municipality.

V. A Socio-Legal, Qualitative Approach to Law
The empirical data generated by this research project led to insights on the actors' experiences, their ideological views on language use, and their perspectives on law. In general, all the methodological choices were guided by the research questions and the practical constraints on the research. Other methodological choices could have been made, and linguistic diversity is indeed studied by several disciplines, each using its own methods. Depending on the specific research object, one could approach this topic through the lens of anthropology-in particular the anthropology of bureaucracy or of the state, or linguistic anthropology-sociolinguistics, or political science.
The research and the data fit more generally in a qualitative socio-legal approach to law. Explicit methodological clarity, including on the characteristics of qualitative approaches to law, is often required for a fruitful dialogue between socio-legal and dogmatic legal research. Qualitative data attempt to capture social phenomena and their meanings. 54 The goal of this study was to inductively discover recurring themes and patterns during the fieldwork. As such, the research design can be understood in terms of "grounded theory," as the theory related to the empirical data developed as the research proceeded. 55 A different, more quantitative, approach to law and studies measuring the frequency of a social fact would require prior identification of that fact 56 rather than allowing it to emerge in the research process.
A balanced, insider's perspective, building on the experiences of civil servants and non-majoritarian language speakers in combination with repeated visits and frequent contact with different informants, leads toward an appreciation of the factual correctness of statements made during oral interactions. Some findings are also supported by written documents, other interviews, and further observation of the actual practices. The cumulative empirical accounts of administrative interactions and the needs of the actors envisaged neither statistical representativity nor an examination of the quantitative effectiveness of the legal norms involved. The selection of field sites was geared toward municipalities with a significant number of non-majoritarian language speakers and toward diversity to explore in depth the study object. The nature of the gathered data is reflected in their presentation and the derived conclusions. As such, the findings, limited by the location and timing of the data collection, should be seen as a broad and indicative insight into the practices and needs of the actors. 57 These remarks have led to the question of how this research fits within "law and anthropology" or the sociology of law. Without entering into a discussion of the differences between or the convergence of these disciplines, 58 it is clear that the methods borrowed from ethnographic research are nowadays also used within the sociology of law-and socio-legal studies-and the topic of language use in administrative settings is not restricted to one discipline. My own research for this article, however, has engaged in this regard predominantly with concepts and literature coming out of "law and anthropology" and social anthropology.

D. Conceptual Framework
The anthropological literature on embeddedness, on legal consciousness, and in particular on discretion is highly relevant to the topic of language use in administrative interactions. This literature made it possible for me to consider a theoretical framework before fieldwork and to develop the analytic tools for this research. The literature on front-line officials or so-called street-level bureaucrats sheds light on the nature of practices not reflected in the case law and jurisprudence. This was particularly the case with the extensive literature on discretion from a range of disciplines, including the anthropology of bureaucracies.
Divergent meanings have been attributed to the term "discretion," 59 and it is possible to distinguish between legal, perceived, and de facto discretion. A common element of these three types of discretion is the absence of effective limits to action, which generates a freedom of choice to act or not to act. 60 The sources of discretion, however, differ. Legal discretion is mandated by the legal framework; perceived discretion follows from an interpretation of the law; and de facto discretion emerges out of the factual circumstances. These different forms of discretion can be falsified, respectively, by verifying whether the legal framework allows for discretion; by analyzing the legal knowledge of the actors involved; and by considering the room that the actors in a particular situation have to ignore or to avoid the applicable legal rule.
The concept of legal discretion is also referred to as "formal discretion" and as "de jure discretion." 61 Generally, this concept is understood to indicate the power to decide on the basis of a legal entitlement  or of rules that allow or generate space for different outcomes. This concept includes the power of administrations to translate law into action, which involves interpretation and choice. 62 Whereas the contours of legal discretion might be clear for legal insiders, there can be confusion among the actors involved as to whether they hold any legal discretion. Several studies on legal consciousness have indicated different understandings among citizens and civil servants of the law. 63 It is entirely possible that neither civil servants 64 nor inhabitants have a comprehensive understanding of the applicable legal norms, possibly due to a plurality of applicable procedural and substantive norms. Some civil servants might consequently interpret the same legal norms differently or invoke other norms or principles. They might perceive the legal framework as providing discretion. Practices at odds with the legal framework might subsequently not be reflected in the case law, as the involved actors-local authorities, civil servants, and inhabitants-are convinced of the legality of their approach. Regardless of whether there is legal discretion or not, the actors involved might perceive that the legal framework provides them with discretion on a certain topic. This perceived discretion is, however, not to be equated with legal discretion, as the legal framework does not necessarily allow for the former type of discretion.
Lastly, there is another form of discretion, de facto discretion, or a factual space that the involved parties have to apply or not apply the relevant legal framework. A vast body of literature from sociology, political science, 65 and more recently anthropology-under the heading of the anthropology of the state 66 or the anthropology of bureaucracies 67 -with mutual influences between the disciplines, has illustrated this de facto discretion. Several terms hint at the phenomenon of de facto discretion. De Herdt and Olivier de Sardan use the notion of "practical norms." 68 Other authors also recognize the difference between formal (de jure) and informal (de facto) discretion, 69 and Triandafyllidou distinguishes between formal discretionary practices and informal discretionary practices. 70 The literature refers to situations in which actors might be aware that the legal framework provides clear legal norms that they might even support, yet still choose not to invoke it or even to knowingly ignore it out of pragmatism or personal convictions. Furthermore, it is possible, especially in less formal legal situations, that the involved parties are unfamiliar with the relevant legal norms or that they are not consciously deliberating on their actions in legal terms. They might, however, rationalize them ex post in legal terms.
This de facto discretion stems from the circumstances of a given situation, such as the physical space in which an interaction occurs 71 or a lack of supervision, which generates the possibility of choosing from a number of potential actions, without this freedom being attributable to the legal framework. 72 This factual room to maneuver makes it possible to subvert rules, whether intentionally or unintentionally, whether out of necessity or out of ideological choice. This de facto discretion thus refers to a factual space that can be used by civil servants, the administration, or citizens to deviate from the applicable legal framework. Together, these three forms of discretion not only formed the foundation of my socio-legal theoretical framework for this research project, but also provided the analytical tools that allowed me to understand and interpret the empirical data.
An element at play within these types of discretion is the multiple embeddedness of the various actors in administrative interactions. Embeddedness is a classic concept in anthropology that highlights the webs of relationships and the several networks that actors are simultaneously part of. For example, one and the same actor might be an employee of an administration, part of a kin group, and an active member of a local community. 73 Several anthropologists have stressed "that the practices of local bureaucracies cannot be separated from the society in which they are embedded." 74 My fieldwork data can also be analyzed through this concept of embeddedness.
These two concepts-discretion and embeddedness-are essential to my presentation of the data to audiences in law and in anthropology, as they allow me to establish a shared framework. These concepts also enabled me to dissect the gathered data with sufficient precision and to consider all aspects in practice, while still respecting disciplinary views on the general concept of discretion. An understanding of the different facets of discretion makes it possible, for example, to show a legal audience the origins and impacts of deviations from the legal framework in a concrete manner. Within the presentation of complex data, the concept of embeddedness offers the possibility of explaining to legal interlocutors the different roles that persons working in administrative services have and how their networks can impact their work.

E. Research Results and the Contribution to the Dialogue Between Law and Anthropology
The publications that have come out of this study describe at length the complexity of situations, the internal and external dimensions and pressures surrounding these administrative interactions, and the often quite apparent balancing exercise that people go through in situations of linguistic diversity. The qualitative approach to data collection provides insights into the perspectives of the involved actors, including civil servants who are faced with legal and linguistic obstacles and dilemmas on a daily basis. The result is a more refined and contextualized view of language use in administrative interactions.
Two extracts from interviews that I conducted are illustrative of several themes and patterns that emerge in the empirical data. A female social worker, active in an administration in the province Flemish-Brabant, explained: The agreement is in fact : : : that we may not speak French : : : . [D]efinitely at the reception, we have to be sure that everything occurs in Dutch. The guideline within our department is to start and to end the interactions in Dutch. And yes, if it isn't possible, you switch. But yes, I notice that one person switches very quickly, another really tries to continue [in Dutch] : : : . I think that every social worker does : : : more or less what he feels fine with, and that differs quite a bit from one person to another. I know of myself that I switch rather quickly, maybe too quickly, because I find that people who come here [to the public social assistance center], that it isn't easy [for them], especially the initial interactions. It is really a hurdle and if they then have difficulties speaking Dutch and expressing themselves, then I feel that I want to stop their suffering. So I speak French. I do tell them that it isn't my native language and that it might happen there are things that might be unclear or that I might have to ask for clarification, but that isn't really a problem.
Another civil servant with a Moroccan background working in a municipal registry office in a large Flemish city explained that in her interactions with the service users, she refuses to speak any language besides Dutch, although she speaks Dutch, French, English, and Arabic fluently. She indicated that she wants to avoid becoming the preferred contact person for every Arabicspeaking client. Moreover, she explained that she made the effort to learn Dutch and sees the benefits of doing so, and feels that others should learn the language, too.
These extracts hint at several recurring themes in the fieldwork data, such as the norms mentioned by the actors that go beyond the legal framework, the different approaches among colleagues-for example, in terms of languages used-and the norm-making power of local situations. Civil servants referred to specific aspects such as a migration background to explain their language use. Some of these civil servants indicated that they were willing to use French and English, but not Arabic, while other persons with a migration background stated that they "protected" themselves by only using Dutch. Some then made a distinction within this attitude, being more willing to use their native tongue with elderly people, but not with younger newcomers speaking the same language.
The empirical data reveal several elements that are particularly prominent, such as the importance of the local context and the de facto discretion of civil servants. The data shows that actual practice, even under a strict legal framework at the domestic level, is not so rigorous or unequivocal. During the actual course of administrative interactions, local language policies soften or even replace statutory provisions with local approaches to language use in administrative interactions. Moreover, there is often both a perceived discretion and a de facto discretion that generates the possibility of deviating from the legal framework.
As is illustrated by the data, legal and linguistic challenges constitute obstacles to the enjoyment of other rights in some situations, as non-majoritarian language speakers do not receive easy or unencumbered access to various basic services, do not understand the requirements for social benefits, or cannot readily articulate their needs.
The fieldwork data indicates that civil servants would generally not outright refuse to use a particular language or to otherwise find a solution to bridge the linguistic gap, but they did not always feel completely free to use a non-majoritarian language, even if they could. The data also shows how the relevant actors relate to previously established legal norms, adopted in a society that has since evolved in terms of its linguistic diversity. Civil servants invoked a range of arguments to justify their approach, such as their individual identity, a collective linguistic identity, pragmatics, and the legal framework. Their subsequent approach has an influence on language use in administrative settings, the functioning of administrative services, and people's access to administrative services. One could argue from a social scientific approach to law that these practices and approaches generate new normativities in terms of policies and shared norms in local administrations. They might also inspire the legislature to shape potential modifications to the legal framework. Yet these normativities do not fit into a strict definition of law as state law, as they are at odds with the constitutional and statutory framework.
The concepts and analytical tools provided in the literature on discretion allowed me to structure the empirical data that I gathered in the course of my research. 75 The broader themes and patterns are more revealing than the expected gap between "law in the books" and "law in action," and they make it possible to connect in a very productive way to anthropological research on bureaucracies. Whereas there might still be a tendency in empirically informed legal research to contrast "law in the books" and "law in action," social scientific studies on law have moved beyond this dichotomy and have started to investigate the relationships and interactions between 75 For the relevant literature, see supra Part D.
the two dimensions. 76 The data I generated can indeed not be easily captured in the dichotomy represented by "law in the books" and "law in action." They rather paint a picture of a recursive relationship between law and practice that goes beyond a strict distinction between the "law in the books" and the "law in action." This research project shows how actors, including administrative bodies, in practice create for themselves a gray zone between "law in the books" and "law in action," making it difficult for those involved to know and to articulate how they should and can act in accordance with the law in situations of linguistic diversity.
Compared to a dogmatic legal study, this study, with its empirical data, reveals a broader and more nuanced picture of linguistic diversity, one that is currently, only to a very limited extent, represented in the case law or jurisprudence on this issue. The empirical data fills a gap in the available knowledge on language use in administrative settings. This topic has long been of interest in the legal discipline, and the data I collected are recognized by legal scholars as thoughtprovoking and insightful. But for my study to have any impact within the legal realm, it had to be complemented with a dogmatic legal approach. As such, I used the empirical data to inform the doctrinal legal thinking. In doing so, I hope to have demonstrated the value of combining law and anthropology.
In this exercise, the data should not be considered revealing solely in light of differences between "law in the books" and "law in action." The finding that the use of other languages is against the black letter of the applicable provisions in certain situations is merely a very initial observation, not unique to this area of the law, and the approach of certain actors in the legal framework already distorts this deceivingly simple observation. It is, rather, the diffuse picture in terms of practices, influencing factors, explanations, and considerations in practices that is illuminating and relevant for legal thinking. The pertinent question is whether the diverging practices, factors, and motives can be captured and, if so, whether they should be reflected within the legal framework.
As states have considerable room to determine their language policies in administrative interactions, and in light of the diverging normative views on this topic, I restricted my legal thinking within this study to a minimal normative stance formed by general rule-of-law principles and obligations stemming from human rights. The empirical data enabled me to identify a minimum core of obligations and a framework for dealing with the language diversity in administrative interactions. The situations represented in the data I collected fall within the scope of several human rights and demonstrate the importance of the linguistic component of these rights, thereby illustrating the relevance of human rights for language use in a number of contexts. As some informants explained rather bluntly, information in a language one does not understand is pointless in a life-or-death situation. The data indeed points toward a minimum core of linguistic obligations based on the right of the public to information if the right to life or the right to physical integrity are at stake (Articles 2 and 8 ECHR). 77 The public's right to information is well established in the ECtHR's case law, 78 but the Court has not linked it with a linguistic component. In light of the requirement of adequate information, the use of additional languages besides the official language(s) might, in my understanding, be required in certain emergency situations that fall within the scope of Articles 2 and 8 ECHR. An approach to the public's right to information that acknowledges the linguistic component would push states to consider a regulatory framework and preventive operational measures regarding how to deal with linguistic diversity in society. This could take the form of proactively drafted multilingual messages with essential information on major emergencies, just as authorities prepare these messages in the official language(s). One should not forget, however, that the choice of means and practical measures are in principle a matter for states to decide upon, 79 and that an impossible or disproportionate burden should not be imposed on state authorities. 80 A combined reading of the legal and empirical findings also shows that the discrepancy between the domestic legal framework and actual practice leads to legal uncertainty and differential treatment. In light of the requirements of the rule of law and the obligations under general human rights, the empirical data allows me to argue through concrete examples that legislatures should create a framework to enhance legal clarity within the current multilingual realities. The need for such framework is not limited to Belgium's Dutch-language area; it is equally relevant for other regions and states faced with multilayered linguistic diversity.
The data on the divergence in practice in terms of approaches and languages used, including in emergency situations, do stimulate legal doctrine and legislatures to rethink legal frameworks in this regard. In light of the empirical data presented in my study, it is difficult to ignore the effects of the current implementation of legal provisions on language use in administrative interactions. This study thus invites lawmakers to consider the implications of the empirical data and the legal minimal core presented here, which is relevant for just a segment of all administrative interactions. Clearly, one could imagine more far-reaching recommendations, but this invitation is to be understood in light of the considerable room that states have to decide on this matter.
The empirical data leads to evolving insights that slowly penetrate legal doctrine and, ultimately, the activities of the executive. As such, the empirical data serves as a stepping-stone to contributing to dogmatic legal approaches to law, and further a meaningful exchange between myself, as a socio-legal scholar, and doctrinal legal scholars and practitioners.
As a conclusion to my "case study" and my subsequent reflection on this dialogue, I argue that research that aspires to combine the disciplines of law and anthropology does not necessarily have to fall between two stools, yet it has to be aware of the basic requirements of the two disciplines to engage meaningfully with them. In this regard, an understanding of the legal framework and awareness of the anthropological literature on relevant key theories and central concepts are crucial from the perspective of both doctrinal legal research and anthropology.