Universalist Republic and ethnic statistics

General ambition

The RUSE ("République Universaliste et Statistiques Ethniques") project is part of a controversy surrounding ethnic statistics in France, a divide that has emerged over the last twenty years around the possibility or prohibition of producing and disseminating official figures on "ethnic" or "racial" origins and religious affiliations, i.e. going beyond the only criterion authorized for censuses in France: the legal criterion of nationality. In fact, for several years now, some people have been arguing that ethnic statistics provide a lever for political action, making it possible to gain a better understanding of the discrimination (e.g. in hiring, housing, etc.) to which certain "minorities" are liable to be subjected (Simon, 2015 a & b, Tribalat, 2016), and thus to act in favor of the fight against discrimination without the risk of letting it dissolve into integration policies (Mazouz, 2017). Conversely, others, notably the members of CARSED (Commission Alternative de Réflexion sur les Statistiques Ethniques et les Discriminations), argue that this would be a veritable "return of race" (Collectif, 2009), insofar as counting and classifying go beyond the instrument of measurement: To measure "ethnic" and/or "racial" groups is to run the risk of entrenching ethno-racial classifications in society, thus ethnicizing or racializing social relations by fossilizing differences that are destined to gradually disappear (the process of integration or, in the longer term, assimilation). Between these two points of view, there are attempts to construct more "reasoned" indicators for combating discrimination, for example, by retaining only the country of origin or nationality of ancestors. Indeed, in the medium term, this solution avoids "fossilizing" differences, since with the "droit du sol" mechanism adopted and applied in France since 1889, children born in France to parents born abroad ("immigrants") become French when they come of age; thus, since their children are born to French parents, the question of their foreign origin is destined to gradually disappear from official statistics (Weil, 2005). As we can see, the question of counting the population is a real political issue at a time when debates on immigration often oscillate between denial and misinterpretation (Héran, 2023; Savarese, 2023).

This recent controversy surrounding ethnic statistics also goes beyond the technical issue of the measuring instrument, since it implies a point of view on "living together", and therefore on citizenship. Moreover, it should be pointed out that this controversy only makes sense in countries such as France, Germany and Italy, where the production of ethnic statistics is forbidden to official bodies such as INSEE, as they are considered incompatible with the republican universalist model of citizenship. Conversely, in countries where the multicultural model has been adopted, ethnic statistics do exist, as is the case in England, the United States and Canada. More specifically, in France, the republican model of citizenship developed under the Third Republic, and constantly reaffirmed ever since, despite numerous controversies (Nicolet, 1994), has three characteristics. Firstly, it is individualistic, in the sense that rights can only be attributed to individuals, which explains why minorities are not legally recognized; secondly, it is universalist, in the sense that individuals are defined as universally equal before the law; and thirdly, it is secular, in the sense that the State is neutral in matters of religion, and is committed to protecting freedom of conscience and worship, which since 1905 have been a matter of free individual choice - religion then becoming a private matter. Under these conditions, there can be no question of counting groups of individuals according to ethnic, racial or denominational criteria, since these characteristics have no bearing on the attribution of the civil, social or political rights associated with citizenship in France.

This is why the emergence, over the last twenty years, of a real controversy surrounding the political uses of ethnic statistics (the fight against discrimination that these statistics would enable us to better understand, and therefore better combat) and the risks associated with the official production of such figures (the ethnicization or lasting racialization of French society), needs to be explained. Not, in this case, by sociologizing the actors in this controversy or mapping the circulation of ideas, but by considering history as a "compass" to better address the issue. In fact, the aim here is to shift our gaze from present-day metropolitan France to colonial France, and more specifically to colonial Algeria, where, despite the expressed desire to mimic metropolitan institutions in a colony long considered a "second France", ethnic statistics were indeed introduced following several provisions enabling the colony's populations to be classified. To sum up, as early as 1830, and before any plans for colonization had been drawn up, a decree distinguished between "Europeans" (French by right of blood, as established by Napoleon, or foreigners for those who would gradually populate the colony, from countries around the Mediterranean) and "natives" (Jews or Muslims, distinguished by their religious affiliation). Subsequently, Napoleon III, informed by the Saints Simoniens and wishing to strengthen the means of government of the colonizing state in Algeria, produced, via the Sénatus Consulte of 1860, the distinction between non-citizen French (nationality without citizenship, and therefore without the right to vote, which concerns indigenous Jews and Muslims) and citizen French (nationality associated with citizenship, for French, then Europeans likely to be naturalized). However, once the indigenous Jews became French citizens collectively in 1870 (Crémieux decree), and the 1889 law on "droit du sol" contained provisions to naturalize "Europeans" more quickly, two categories of population coexisted in the colony for a long time to come: French citizens (those who were already citizens, the former natives of the Jewish faith, and the progressively naturalized "Europeans"), and French non-citizens (the French of the Muslim faith, who only gained full citizenship during the Algerian war in 1958). The classifications enshrined in colonial legislation had thus led to the creation of two legal statuses (national citizens and non-citizen nationals) to group together populations that were initially more numerous.

However, this binary distinction between two population categories (French citizens and French non-citizens) is not reflected in the official population-counting operations, regularly recorded in the Tableau général des communes d'Algérie. In fact, populations are distinguished not according to the criteria of nationality and citizenship, objectified by colonial legislation, but according to both legal and ethno-confessional criteria, since "French" (nationals and citizens), "indigenous Muslims" (non-citizen nationals of the Muslim faith), "naturalized Israelites" (citizen nationals of the Jewish faith), foreigners of European origin and foreigners from the Maghreb (Moroccans and Tunisians) are counted. For this reason, colonial Algeria represents a singular case study - to our knowledge, the only one of its kind - in the introduction of ethnic statistics within a colonizing state that was constantly reaffirming - but not always implementing - republican and assimilationist principles.

Hence the interest in questioning a case that is well and truly at odds with the model of citizenship, as part of a present-past-present approach (Braudel, 1969), in line with the perspective of a socio-history (Noiriel, 2008) of ethnic statistics in colonial Algeria: built in the present, starting from the controversy over ethnic statistics in contemporary France, the aim is to examine the question in the past, on the Algerian colonial terrain, which offers an original and little-studied case study, to better envisage the meaning it may take on and offer a better understanding of the controversies of the present day.

Theme(s) and question(s) addressed

Hence the research question to be addressed: how can we explain the emergence, within a universalist Republic, of an ethnic statistic that is in principle incompatible with the republican model of citizenship? By hypothesis, can it be explained by a departure from republican principles, given the characteristics of the colonial situation (code de l'indigénat, dissociation of nationality and citizenship, which makes sense when the process of nation-state construction is complete, late 19th century)? Was it the desire not to dissolve, probably under pressure from the "colonists" (Ageron, 1978), the question of indigénat, which remains the main issue facing all the colonial settlements historically (Veracini, 2010), whether they be former French colonies (Algeria, New Caledonia) or former foreign colonies (Australia, Latin America)? Was it out of a concern for better control of the territories and the various populations of the colony, while colonial power remains, "structurally", a "weak power" (Cooper, 2005), by turning statistics into a science of government or cameral science (Ihl, Kaluzinski, Pollet, 2003)?

This largely "forgotten" experience deserves in-depth examination, for two reasons. Firstly, because the development of knowledge about the former colonial settlements has made it possible to decipher the different ways in which the question of indigénat was dissolved - confinement in reserves, ethnocide, physical elimination, regrouping camps as was the case during the Algerian war (Sacriste, 2022), but also access to full citizenship - while ethnic statistics record and measure differences. And secondly, in the sense that the issue of managing cultural and denominational diversity, in former metropolises and in reverberation with former colonies, makes sense today in all Western democracies, whatever models of citizenship are adopted (Joppke, 2010). Thus, multicultural regimes sometimes adopt rather assimilationist measures, such as the requirement for immigrants to take language courses in the Netherlands (Constant, 2000), while universalist regimes also seek to make adjustments by adopting more particularist measures, as shown by the many memorial laws passed over the last thirty years in France, on Jews, descendants of slaves, repatriated immigrants, Armenians and the harkis (Savarese, 2020). In this way, the experience of the introduction of ethnic statistics in colonial Algeria, which historically constitutes a truly negative case, must be interrogated in an attempt to produce a better understanding of controversies seized in the present, in particular that on ethnic statistics in France. In the former metropolises, as in the colonial terrain (De Mari, Savarese, 2019), the construction of citizenship always corresponded to drawing a line between inclusion and exclusion, by specifying the conditions of belonging to the community of citizens. Conversely, the aim here is to understand what it means to define and measure indigénat on the basis of an ethnic statistic, and how the effects of such a statistic on social relations and political choices - recurrent refusal of access to citizenship, specific expression of anti-Semitism, which is neither a property of the "colonial situation" (Balandier, 1951), nor a specificity of colonial Algeria (Savarese, 2019) - can be if not "measured", at least identified. It is therefore the sociogenesis, practical uses and effects of an instrument of measurement in the former colony that are targeted here, in order to better understand the issues at stake in the present.

Methodology and approach

1/ In this way, we need first to examine the "sociogenesis" of ethnic statistics in colonial Algeria: in what context were they introduced in 1830, and maintained under the Third Republic despite the promotion of the individualist, universalist model of citizenship; who were their potential promoters or detractors; what were their social properties and institutional positions in the former metropolis or colony, their interests and strategies, their alliances and networks of relations. The identification of the actors involved and the factors favoring (processual analysis) the introduction of ethnic statistics must be achieved by cross-referencing the decisions of the colony's general government, available at the Centre des Archives d'Outre-Mer (Aix-en-Provence), with the resources and trajectories of the actors involved, most of which are listed in the library of the former colonial school (founded in 1889 by Auguste Pavie to train the empire's future executives), and whose decisions are also commented on in the colonial press - these last two sources, especially the first, are also available at the Centre des Archives d'Outre-Mer.

2/ Next, we need to define how the census was introduced and carried out in practice, identifying not only the population categories already identified, but also the procedures used to count them. Since 1801, the French population has been regularly counted, as has colonial Algeria, where 19 population censuses were carried out between 1830 and 1962. The general government of Algeria endeavored to ask the administration to base the counting method on the instructions given in metropolitan France. However, in military territory, indigenous populations were estimated on the basis of counts of constituted tribes, and so adaptations were made at local level, at least until the transition to civilian rule (1970), when the method used in metropolitan France was adopted (Kateb, 1998). However, ethnic statistics, as used in countries that have adopted a multicultural model of citizenship, presuppose that it is the individuals enumerated who specify their own ethnic, racial or religious affiliations (Schor, 2009; Kertzer, Arel, 2009 ; Perlmann, 2018), whereas the census carried out in metropolitan France, based on a universalist model, focuses solely on "objective" criteria (age, gender, profession then socio-professional category, nationality, type of residence), which are identified on the basis of questions asked by the interviewer. Under these conditions, the archives of the Arab offices and military administrations, also available at the Centre des Archives d'Outre-Mer, seized at the time of the census periods, must be exploited to establish how the census method of metropolitan France is likely to be locally adapted: is the definition of ethnic or racial category and membership produced by the interviewer, as in metropolitan France, or by the respondent, as is the case in countries using ethnic statistics? Did the colonial government make any original or unusual adjustments - depending on the territory, whether civilian or military, and the period (sabre-rattling, transition to civilian rule) - to carry out the 19 census operations recorded between 1930 and 1962?

3/ Finally, we need to ask to what extent these statistics have contributed to the ethnicization or racialization of colonial society. Such ethnicization/racialization is hardly in doubt, given that the actors are regularly referred to as "French", "European", "Jewish" or "Arab" (Lorcin, 1999), rather than by their socio-professional status. However, other elements also contribute to this dynamic, in particular colonial ideology or imagination, through which the differences between the components of colonial society reflect above all supposed differences in civilization (for example, between the "Kabyle" and the "Arab", the former being considered closer to the French, even though both are considered inferior to French citizens). Interrogating the effects of colonial statistics on Algerian society therefore presupposes identifying spaces for the expression of this difference, such as electoral consultations, already studied in d'Oran (Savarese, 2016) due to the presence of this department's electoral archives at the Centre des Archives d'Outre-Mer. Other areas where the racialization of society is expressed, such as the filing of complaints and court rulings (also belonging to the "archives of sovereignty" and thus fully available in Aix-en-Provence), could be addressed to enrich these analyses.

These three sets of questions should be the subject - by estimation, and taking into account our relative knowledge of the funds available - of some two months' work devoted to the examination of the archives and sources listed above, organized, taking into account the academic calendar of a teacher-researcher, into several periods of one to two weeks' work (shorter periods being, by experience, often unproductive). As the month of July offers the possibility of two weeks' work in the archives, and the weeks when classes are interrupted are dedicated to completing the work, we propose the following schedule of research and discussion operations.

Impact and potential spin-offs

1/ With regard to stays at the Centre des Archives d'Outre-Mer, we are planning two weeks in July 2024, one week when teaching is interrupted in November 2024, one week at the end of December 2024, one week when teaching is interrupted in April 2025, and the remainder in July 2025. The pace at which the project progresses may condition any adjustments and additional visits to the archives, particularly in the second half of 2025, given that most of our teaching takes place in the first semester.

2/ In the meantime, we will be organizing an exploratory research seminar in Montpellier between January and June 2025, to welcome an Algerian colleague from the CRASC (Centre de Recherches en Anthropologie Sociale et Culturelle, Oran), as well as several French and foreign specialists in population counting and census operations. These sessions will provide an opportunity to discuss the initial results of our work, to compare it with the presentation of surveys on Algeria and other European countries, and to consider the feasibility of a comparative research project on the controversies surrounding ethnic statistics in a variety of contexts. We also intend to present the results of our research in Oran, to colleagues from CRASC, a laboratory with which our own laboratory has recently finalized a multi-year scientific and educational partnership agreement.

3/ The proposed archival survey should lead to the publication of a research paper on the sociohistory of ethnic statistics which, starting from its applications in countries that have adopted a multicultural system of citizenship, and the controversies on the subject in countries that have chosen a universalized model, can account for the contributions of the Algerian colonial case to the sociology of these controversies. It could take the form of a research book in its own right, or the coordination (alone or in association with a colleague) of an issue of a peer-reviewed journal based on the results presented at the seminar we plan to organize - and in which we will propose one or more students enrolled in the Master's research program to participate.

4/ As we are responsible for teaching epistemology and research methods in the social sciences in the Master's program in Comparative Politics and Public Action, we propose to supervise a research dissertation by a student.e, as part of a laboratory internship that has become compulsory since this Master's degree became part of the IDIL program, on the sociology of the controversy surrounding ethnic statistics in France, which could be linked to the survey work we hope to carry out through a historical detour towards its application in colonial Algeria.

Project summary in French

Based on the recent controversy surrounding the possibility of taking a census of the French population by introducing ethnic statistics, this article takes a detour via colonial Algeria to draw on an original experience: the introduction of ethnic statistics to count the populations of the former colony. This is a singular example insofar as the republican model of citizenship, which recognizes only the universality of individuals, is a priori incompatible with ethnic statistics, which in principle concern only countries that have adopted a multicultural model of citizenship. The sociogenesis of ethnic statistics in colonial Algeria can thus be questioned in three ways: by attempting to identify the conditions of its introduction, the practical conditions of its application (how to count) and the effects of the measurement technique on the dynamics of ethnicization/racialization of the colony's populations, in order to better grasp the challenges of classifying the French population in the present. This work is based on the archives available at the Centre des Archives d'Outre-Mer.