Mapping Foreign Migration to Belgium
The Digitization of the Index Cards of the Belgian Aliens’ Police (1832-1889)
TSEG 17 (2): 83-94
Migration caused a lot of turmoil in recent years. Large numbers of people have travelled to Europe in search of a better life. Governments have fallen and political parties have risen or crashed based on their policies concerning these ‘migration crises’. But how exceptional are these migration trends when placed in a historical perspective? As the subject of migration proves to be one that is to stay on the political agenda, professionals better have the right numbers to construe the policies that will make up the future of the continent. More so, a historic understanding of migration, how migration now differs or resembles past migration, is crucial to put current movements of people into perspective.
In 2015, the Belgian State Archives entered into a partnership with Vrije Universiteit Brussel (VUB), Universiteit Antwerpen (UA) and Université Libre de Bruxelles (ULB) for the IMMIBEL-project. IMMIBEL is a research project funded by the federal research programme BELSPO BRAIN (Belgian Research Action through Interdisciplinary Networks). The project was entitled ‘Outcast or embraced. Clusters of migrants in Belgium (1840-1914)’ and aimed to digitize the large collection of nineteenth-century index cards from the Belgian Aliens’ Police held at the State Archives. The project also investigates the political, economic, social and cultural dimensions of interactions of foreign nationals with different layers of Belgian society. This part of the project was in the hands of the different university partners who used the database to study three subgroups in foreign migration: (1) undesirable migrants that got confronted with repressive public policies (VUB); (2) seafarers working in an internationalized labour market segment (UA); and (3) engineers as ‘actors of knowledge’ spreading new ideas and technologies at the start of the second industrial revolution (ULB). 1
The project resulted in a user-oriented entry tool, a database and a research guide2 which allow historians, genealogists, museum professionals and others to easily access and explore the archives of the Belgian Aliens’ Police. This data-column describes the conception, construction, structure and physiognomies of the database. It discusses the new research possibilities and knowledge that arise from the construction of the dataset and considers the practical limitations and challenges of working with a large historical dataset.
In 1832, the young Belgian State established a Public Safety Office (Sûreté publique), tasked with ensuring public order on the territory. This service, which from 1840 onwards officially resided under the Ministry of Justice, was responsible for registering newcomers in the country. The registration was done by the Aliens’ Police (Police des étrangers) who opened a file for every foreign national entering the country. With this measure, Belgium was one of the pioneers in Europe. Newcomers had to register in the municipality of arrival where the local authorities gathered information about their identity, financial means and behaviour via a standardized information bulletin. Afterwards, this bulletin was forwarded to the Aliens’ Police in Brussels where a central, national ‘database’ was created that allowed State officials to identify each registered newcomer and assess whether or not he or she could stay in the country.
Between 1832 and 1890, the Aliens’ Police opened around 500,000 individual files. In 1914 the number of registered aliens had risen to one million. In the course of the twentieth century, the growing mobility of the population and the turbulent political situation in Europe led to an increasing influx of migrants. Between 1919 and 1943, the Public Security produced another million files. In 1977 the Aliens’ Police were renamed the Immigration Office, which, until now, has been responsible for the central registration of aliens on the Belgian territory. Today, the Belgian State Archives conserve about two and a half million foreigners’ files for the period 1840-1959. The series of foreigners’ files of the Aliens’ Police is one of the oldest systematically collected and preserved datasets concerning migration and mobile groups in Europe.
Up until recently, this vast source was difficult to study and remained largely unexploited by researchers and the public due to the difficulties to examine the files in a structured, comprehensive way. The files were solely made accessible with the aid of alphabetically arranged index cards. In practice, the first series of the Aliens’ Police’s index cards covers five decades (1840-1889) and is the source of this database.3 The nineteenth-century index cards are kept in 265 cardboard chip trays and can be consulted on microfilm in the Belgian State Archives reading room.4 However, based on these microfilmed index cards, it is
not possible to get an overall feel of the profiles of the persons travelling and moving to Belgium. As such, the goal of the creation of the IMMIBEL-database was to make the nineteenth-century index cards more accessible and searchable for the public at large, thus opening up the corresponding individual files.
Collecting and cleaning the data
The digitization of the index cards was a labour-intensive process that took about eighteen months. As the handwritten cards did not permit for OCR reading, they were manually transcribed in Excel tables. To
guarantee the quality of the input, a data-collection procedure was tested by State Archives employees and PhD students working on the IMMIBEL-project. After a month of encoding, the method was fine-tuned and corrected where necessary. A detailed manual was drafted to standardize the data collection. The majority of the input was done by State Archives employees and final year history students who had the detailed manual at their disposal and worked in a group supervised by archivists of the State Archives. Problematic spellings and writings were discussed until consensus was reached. Subsequently, random samples were conducted on the complete transcribed dataset and common transcription mistakes were traced and corrected. The final product is a dataset that contains personal information on 151,857 individuals, with one entry record per person and eighteen variables collected for each one of them, comprising a total dataset of approximately 2,000,000 fields.
Each index card contains the basic identifiers of the corresponding foreigners’ file. The eighteen collected variables in the database per foreigner are (1) a unique identifier; (2) the tray number and place of the card within the tray allowing an identification of the physical carrier; (3) the conservation status of the corresponding file as mentioned on the index cards; (4) the last name; (5) the first name(s); (6) gender; (7) birthplace; (8) region or nationality of origin; (9) birth day; (10) birth month; (11) birth year; (12) profession; (13) file number; (14) relationship status (married, widowed, divorced); (15) last name of the partner; (16) first name of the partner; (17) follow-up of the interactions with the legal system (e.g. if a foreigner was expulsed or a vagrant); and finally (18) remarks such as the nationality of the spouse, the number of children and their names or the date of death. Most cards are written in French, some in Dutch. To avoid erroneous interpretations, all fields contain a literal transcription of the information on the index cards, except for the unique identifier, the tray number, the gender and the remarks column. This way, the database can not only be studied for historical purposes but also provides possibilities for linguists and other scholars.
To improve the quality, the collected data was cleaned with the open access ‘Open Refine’ software, which allowed for semi-automatic clustering and editing of large quantities of messy data.5 The programme had to be used with care and understanding off the caveats and biases of using new technologies on historical datasets. Column by column, recurrent abbreviations were resolved, similar input was clustered, errors were corrected and ambiguities were traced. Strongly corresponding data were detected by automated clustering and mass editing was done using the GREL coding language. When possible, the methodology of Demographica Flandria Selecta, which recommends cleaning in clusters, was followed. This means cleaning clusters of names, such as ‘J.n B.te’ as a unit (‘Jean Baptiste’), rather than cleaning ‘J.n’ (‘Jean’) and then ‘B.t’ (Baptiste).6 However, the enormous variety in combinations of first names limited the usefulness of the method for this specific dataset. In the end, several abbreviations could not be resolved as it was impossible to determine their meaning with certainty.
The birthplace proved to be the most challenging variable of the dataset. A simple difference of accent could imply a different place on the map. As such, the Open Refine clustering tools could only be used in a limited capacity. When data were available, the cleaning of the birthplaces was done in correspondence with data from the origins column,
but even then, it was not always possible to determine which exact place was meant due to the great number of homonyms. To avoid interpretative mistakes, the spelling of the birthplaces was largely retained as it appeared on the index cards. This means that a lot of places are written in French even if they were not located in France or Wallonia and that places that are, for example, in present-day Poland, are mentioned by their nineteenth-century German name.
Finally, the dataset was expanded and enriched with contextual data (e.g. information gathered during the collection process about the physical state of the carrier of the data) and via links to other datasets. For example, after cleaning and clustering, a total of 5,900 unique professions and combinations of professions were identified. 1,300 of these could directly be linked to the HISCO database (History of Work Database of the Institute of Social History).7
Possibilities and limitations
By digitizing the index cards, the created dataset allows answering research questions that were hard to tackle before. The dataset lends itself to conduct quick searches of individual migrants or specific groups of migrants based on their names, professions, birthplaces, age or marital status. For example, researchers will be able to track down all shoemakers, sculptors and patisserie bakers. They can track down all registered maids born in Berlin, all mosaic workers from Seguals or prostitutes who migrated from Rotterdam. As such, the database allows to make statistical analyses on the registered migrant population, to analyse migration patterns, to derive professional and social roles of migrants and to visualise mobility in the Belgian migration field.
However, the database is first and foremost a gateway to the information in the foreigners’ files and, ideally, should be used in combination with the files. Almost all files contain information on the successive changes of residence of foreigners, making it possible to trace the mobility of persons on the Belgian territory. Files of very mobile groups, such as travelling workmen, can also contain a workmen’s booklet with a detailed overview of all the places where they were active. The foreigners’ files mainly contain documents relating to the marital status, morality, activities and movements of persons within the borders of the country. The municipal declaration of arrival or the information sheet (bulletin de renseignements), which can be found in all files, provides the basic identity details of a person using several standard questions: name, place of birth, marital status, nationality, family composition, profession, address in Belgium, last official address abroad, a certificate of morality and a residence intention. Each time a person moved within the country, a new information sheet was drawn up, which gave the Aliens’ Police insight into the mobility of aliens on the territory.
Files also contain documents and deeds on the foreign national’s interactions with the government services. The files can hold police documents concerning morality investigations, marriage and death certificates, residence advice, official reports, extracts from criminal records, exchange of data with the National Guard (Gendarmerie), documents concerning extraditions and expulsions and correspondence with Belgian state benevolent colonies. In files that extend up to the 20th century, one can find passport photographs. However, the use of photographs only became common after 1918, which is why most nineteenth-century files contained little or no illustration material. The information sheet does always contain a personal description, which makes it possible to form a picture of the physical characteristics of the foreign national in question. In addition, files of persons who represented a potential danger to public order were often supplemented with newspaper articles and engravings containing images, portraits and sketches.
Researchers who aim to use the database to study migration and mobility on a bigger scale or as a social construct, should have to account for the fact that the current database is an incomplete record of the actual migrant population and movement on the Belgian territory in the nineteenth century. Numerous migrants were never registered or registered only after they had stayed in the country for a long time. Furthermore, in 1900 the Aliens’ Police eliminated several hundred thousand files and the corresponding index cards of persons that were deemed to be no longer present on the territory. Certain categories of files and cards were exempt from elimination, such as files of ‘important’ people and ‘important’ political refugees, files concerning general issues and matters of principle, files of foreigners that were extradited, expulsed or sent away for reasons other than vagrancy, files that were still actively used and files of people who were condemned for facts of a certain seriousness (theft, fraud, abuse of confidence).8
Users also need to bear in mind that the database is a translation of the index cards, which in their turn, are a momentary registration of the contents of each foreigners’ file. Changes in the migrant’s profession, marital status, etc. usually don’t show in the database. Furthermore, researchers will notice that determining an individual’s nationality was no easy task. There often was confusion between nationals de jure and nationals de facto, a consequence of the limited importance that the abstract relationship between State and citizen had in the social reality of the nineteenth century.9 Consequently, the ‘origin’ was only noted on 36 per cent of the index cards and even then, it did not necessarily contain a nationality. The index cards show a variety of ‘origins’ ranging from the nationality of a person’s parents, birth countries, marriages to foreign nationals or references to regions and old administrative territories who still had a lot of importance (such as Hanover, Grand Domaine de Luxembourg or Saxe) rather than specific countries. Because of this, the database does not provide a clear-cut way to search for historic nationalities.
Finally, one needs to keep in mind that the dataset only allows for one-sided and static analysis of the migration towards Belgium. Migrants may have passed several other countries and cities before coming to Belgium. Possibly, they only stayed in the country for a short period and travelled on to other parts of the world. Since the dataset is based on the index cards, it is not possible to search on the place of settlement in Belgium. To truly grasp the mobility of each migrant on the Belgian territory one has to consult the corresponding foreigners’ file.
The created dataset offers a wide range of new research possibilities and teaches us a great deal about the migration to Belgium in the nineteenth century. The 151,000 index cards that have been digitized, open up the access to over 80,000 individual foreigners’ files. Before digitization, searching these files was only possible by last name (via the alphabetical index) or by the period of registration (via the file number). Now, the foreigners’ files are searchable on last name, period, first name, profession, birthplace, birth date, origins, name of the partner, names of children, conservation of the corresponding file and combinations of these variables.
A basic version of the dataset was integrated into the Search Persons search engine of the Belgian State Archives in April 2018 and allows for quick searches of individual migrants or specific groups of migrants. To access the dataset, users should type ‘IMMIBEL’ in the search bar of the search engine.10 A detailed manual elaborates how the search engine can be used for searches within the dataset. The manual further contains three aid-tools: (1) a list with file numbers of nomads11; (2) a legend of the different follow-up codes and their meaning 12 and (3) a list with the opening years of foreigners’ files based on their respective file number.13 The Search Persons search engine of the Belgian State Archives contains a thesaurus with different spellings of Belgian localities which means users can type in the Dutch version of a place name and still retrace relevant sources and files in French and vice versa. Although this multilingual thesaurus is systematically expanded to include different spellings and new place names, there is currently no good coverage of foreign municipalities. The list is therefore particularly useful for tracing foreign nationals born in Belgium. When searching for foreign municipalities and professions, one should keep in mind that the index cards were written in French and therefore French should be the primary search language.
The complete enriched IMMIBEL-dataset will eventually be made available via the BISHOPS-platform (Belgian Infrastructure for Social Sciences and Humanities Open Science), currently being developed by the Belgian State Archives.14 This dataset will allow for additional analyses such as identifying contingents of migrants, the proportion of male and female migrants and the weight of certain occupational groups travelling to Belgium. All researchers and other interested parties who want to work with specific groups of migrants or large amounts of files can contact the Belgian State Archives at firstname.lastname@example.org with their research question.
About the author
Sarah Heynssens is wetenschappelijk medewerker aan de Vakgroep Geschiedenis van de Universiteit Gent. Ze werkt er voor het Quetelet Center for Quantitative Historical Research, een dienstencentrum dat advies en expertise aanbiedt over het gebruik van historische gegevens. Voordien was zij actief als onderzoeker bij het Centrum voor Historisch Onderzoek en Documentatie over Oorlog en Hedendaagse Maatschappij (Cegesoma), Universiteit Antwerpen, het Mode Museum Antwerpen en het Rijksarchief. Haar onderzoek richtte zich voornamelijk op metissage in koloniale context, migratie, gedwongen verplaatsingen van kinderen en interlandelijke adoptie. Ze creëerde de IMMIBEL-databank in het kader van het BRAIN-Belspo project ‘Outcast of Embraced? Clusters van Buitenlandse Immigranten in België, ca. 1840-1890’.