(1) Introduction: the foundations of social internationalism

At the International Peace Congress of 1849, the French writer Victor Hugo affirmed a hope that animated both the organizers and attendees of this international gathering: “Yes, the period of revolutions is drawing to a close - the era of improvements is beginning.” (Peace Congress Committee, 1849, p. 13). From the middle of the 19^th century onwards, a growing number of intellectuals sought to advance ideas, initiatives and organizational forms that would optimize and elevate their societies. Their optimism about the benefits of industrialization and urbanization was beset by fears of instability and decline. The harmful side effects of modern transformations were becoming difficult to ignore. Poverty plagued the neighborhoods of laborers living and working in dire circumstances. How the young were to be educated or what position women were to have in modern society were disquieting questions yet to be settled. In 1848, a wave of revolutions had swept Europe, while the competition between nation states made war a looming threat. Various inheritors of the Enlightenment’s belief in progress turned to a different solution: social reform. For advocates of reform, the improvement of society was to be gradual, guided by science, and advanced through collaboration.

Congresses were in essence laboratories for the development and diffusion of reformist ideas, and hence represent an outstanding platform for further research into transnational exchange of reform agendas (Rasmussen, 1990; Randeraad, 2015). The TIC-Collaborative project¹ focused on this diffuse transnational field of social reform, whose actors and activities tended to merge the political categories that mark standard historiographies of the period (Topalov, 1999). By collecting and enriching a dataset of international organizations and congresses associated with social reform, TIC sought to map cooperation across national lines and across thematic categories. For social reform, it was intended as a comprehensive effort, both geographically and in terms of content. It was meant to counter a wide set of tensions and crises that became known as the ‘social question’, a catch-all term used by scholars as well as contemporaries (Randeraad & Leonards, 2010). From hygiene and criminality to education and world peace, the social question encompassed a range of issues that needed to be diagnosed, studied and addressed if ‘mankind’ was to reach its pinnacle of civilization and harmony. Whereas most European governments still held onto the non-interventionism of classical liberalism, social reformers recognized that economic and technical innovations by themselves did not inevitably benefit collective well-being. Instead, they argued that different aspects of social life could, and should, be actively steered. In this top-down vision of social change, reformers saw a role for themselves as engaged citizens working towards the piecemeal and scalable advancement of local, national and supranational communities.

Despite the various professional backgrounds of its proponents, social reform was well embedded in bourgeois circles (Van Praet 2019, D’haeninck 2018). The progressive worldview of reformers tended to range from unorthodox liberalism to moderate socialism and drew from ideologies such as pacifism, internationalism and feminism. The cause of social reform thus formed a web of linked and mutually reinforcing commitments. Especially in its early phase, reformers often had all-embracing and eclectic interests. What they shared, however, was not just a general problem - the social question - but also the ambition to solve it in a gradual, orderly, collaborative and scientifically-sound manner. The accumulation and exchange of knowledge on social issues was thus central to their self-imposed task. Before these issues could be brought under public scrutiny and eliminated, the various qualms of society had to be identified and studied. Urban centers, for example, were thus scrutinized for traces of decay or ‘backwards’ elements not yet touched by the blessings of modern civilization.

To disseminate their diagnoses and remedies, reformers organized congresses that were regularly open to participants from across the globe. Originating in progressive liberalism, but soon also embraced by socialists, these congresses were used as part of a repertoire of tools to contest existing power relations and the social and cultural status quo. They were political in the sense that they were often inspired by certain worldviews and aimed at concrete reforms in the form of social legislation. However, before the 20^th century, political parties were rarely the main organizational unit (with the exception of socialists in some cases). The international Catholic world had its own networks for the exchange of information and community building (Liedtke & Weber, 2009; Viaene, 2015), just like various other christian denominations. Freemasons, on the other hand, were ubiquitous in congresses outside their own, but generally chose to remain discreet about their affiliation (Jansen 2015; Randeraad & Leonards 2020).

The model for the efforts to supplement existing communication flows (such as correspondence or magazines) with structural meetings was modern science, with its systematic method, claims of universality, and ideals of collaborative knowledge building and identity formation. In tandem with the emerging disciplines of the social sciences, reformers were convinced that social and economic life was also subject to certain ‘laws’ that could be discerned through the observation and comparison of social ‘facts’. Likewise, the solutions proposed by social reformers were tested in an experimental fashion. Plans for educational methods, housing projects or penitentiary reforms were ideally tried out, compared, adjusted and then propagated to be tested in other environments. However, many of the propositions that were diligently worked out, and extensively debated by reformers, were never put into practice. Before the late 19^th century, their impact on governmental policies and legal frameworks was limited or indirect at best. To increase their leverage, reform-minded intellectuals and activists organized themselves in leagues and associations. These organizations - which could be internationally oriented - sought to share their expertise in order to pressure (local) governments and raise public awareness about the topics that concerned them. If the optimization of social life was the horizon of reform, then collaboration was considered to be a crucial vehicle to achieve this goal. Often, it was even strategically necessary to link up with like-minded others across borders. Keck and Sikkink (1998) have described this as the ‘boomerang pattern.’ When it comes to social issues, the problem was very much local or national, but the solution lay partly in finding allies beyond one’s own state borders. Julia Moses, for example, has shown that the international congresses on accidents at work held between 1889 and 1914 never pushed for a binding international convention, but “provided a forum that governments could choose to ignore, manipulate or search for new ideas” (Moses, 2015, p. 62). Randeraad and Leonards (2020) have also argued that many delegates used their participation in international congresses not only to promote causes of reform across borders, but also to strengthen their authority in their own country.

The first international congresses on topics such as peace and penitentiary reform were held in the 1840s. The earliest congress in our selection is the first penitentiary congress, which took place in Frankfurt am Main in 1846. Yet their number and importance grew rapidly (Figure 1). No fewer than 1,655 congresses dealing with social change were organized before 1914. Depending on the selection criteria of what qualifies as ‘social reform’, this number could be even higher. Peaks in the organization of congresses often coincided with World Expositions, events which similarly trod a fine line between internationalist ideals and national competition. The overwhelming majority of these congresses were held in Europe, with Paris, London and Brussels as the major hubs of the congress scene. The prominence of Brussels is hardly a surprise. The new Belgian state actively developed and supported the phenomenon to gain international influence (Aubin 2020; Verbruggen, Laqua & Deneckere, 2012). At these events, experts, activists, intellectuals and officials gathered to exchange knowledge and propose guidelines and action plans around a certain topic. These topics were as varied as the social problem that the congresses were meant to pinpoint and counteract. During the 1910 World Exposition in Brussels, for example, congresses were held on food hygiene, vegetarianism, mathematics education, tuberculosis, cremation, family education, feminism, occupational diseases and many other topics and areas of social reform.

After a congress, the different contributions and the ensuing discussions were often published in the form of a congress report. As part of the TIC project, we assembled and made accessible a digital collection of over 1000 congress reports, with extensive metadata, that give an insight into the proceedings of these events. Local situations and numbers were compared, and the effectiveness and (international) suitability of different measures were discussed. This could lead to serious debates among the predominantly white male reformers who frequented the congresses (women only started attending by the late 19^th century). Nonetheless, the ideal of an international consensus - reached through open debate and rational dialogue - was the driving force behind the expanding circuit of congresses. Collaboration, across borders and in the name of science and universal welfare, was itself also a rationale that these events were meant to embody. To institutionalize these collaborative bonds, many congress series transformed into international organizations from the 1880s onwards. International congresses also became increasingly specialized and professionalized by the belle epoque. The communities of experts formed through the congresses now began to professionalize their activities, a process which concurred with and stimulated the maturation of new knowledge fields into separate disciplines (Schroeder-Gudehus 1990).

(2) Sources and data enrichment

A thorough assessment of these early sites of international sociability can not only help to reexamine the welfare arrangements before the welfare state from a transnational perspective, but can also shed light on the ascendancy of scientific internationalism and international organizations. Focussing on initiatives related to social reform, the goal of the TIC project was therefore to assemble a database containing systematic information on the phenomena of international congresses and organizations from the 1840s until the beginning of the First World War. We focussed on the attendance of reformers originating from the Low Countries as well as in-depth thematic sub-projects (see below), However, for the overall mapping of conferences related to social reform and the compilation of bibliographic metadata, the aim from the outset was to be as exhaustive as possible.

The foundational sources of our dataset were the publications of the Union of International Associations (UIA) and its ‘complete’ list of conferences from 1681 onwards (UIA, 1960; UIA 1957; UIA, 1964). In 1907, Henri La Fontaine, Paul Otlet and Cyrille Van Overbergh set up the Office Central des Associations Internationales. Besides the development of a centre for collections about internationalism in Brussels (including what La Fontaine called a ‘Documentary Repertory of Internationalism’), one of the main goals was the foundation of the Union of International Associations (UIA), an ‘association of associations’ that would coordinate the relations and interests between international organizations throughout the world. In a short period of time the UIA became a direct forerunner of the League of Nations’ Committee on Intellectual Cooperation. It issued a periodical, La Vie Internationale, and published, in collaboration with the pacifist Alfred Fried, two yearbooks (1905; 1906; 1907; 1910–1911): the Annuaire de la Vie Internationale, which gave an overview of the various international associations. The yearbook and the UIA still exist today and many studies on international non-governmental organizations rely on the UIA data and the typologies drawn up by them (see Laqua, Van Acker & Verbruggen, 2019 for an overview).²

Data collections with a global or even universal scope will always remain incomplete or questionable (Bloodgood & Schmitz, 2013). In the case of the UIA, this is partly caused by the fact that from its inception in 1907, most of the data was provided by the organizations themselves, as acknowledged by the UIA (UIA, 2015). Moreover, the categorization of ‘genuinely’ international organizations itself “has been, and remains, a political project that embodies and shapes worldviews” (Saunier, 2019). A well-known example, for instance, is the lack of religious organizations and fraternities engaged in transnational advocacy networks, health, education and many other fields related to social reform (Bush, 2007). This being said, the data offers a representative data source for the study of various long-term developments (eg. Boli & Thomas, 1999; Grandjean & Van Leeuwen, 2019), especially with regard to Europe and North America, which can be further enriched with original archival research and/or used as a starting point for further research.

Additionally, because greater emphasis was placed on the listing of organizations rather than congresses, we had no choice but to set up a new database, powered by Nodegoat (see below).³ Aware of both the potential and the caveats of UIA-related published sources and datasets, we consulted the original sources produced by the organizations (congress proceedings, reports of assemblies, charters of foundation, etc.) and other contemporary published (Army Medical Library 1938; Baldwin 1907a, b; Bishop, 1958; Coorman 1968; Doré 1923; Eijkman, 1910; 1911; Faries, 1915; Gregory, 1939; Judge 1978; Lyons 1963; Matthews, 1940 and Wallace & Singer 1970). This enabled us to correct mistakes and to go beyond the lacunas and the contradictions between the reference books (e.g. event dates, alternative names, etc.). An important added value of our list of congresses, compared to previously published datasets, is the systematic addition and verification of bibliographic data related to published congress reports. All of the publications are provided with detailed metadata, provenance and their WorldCat permalink. A second major added value is that the data has been enriched with thematic categories, allowing improved (faceted) search and other queries. Over 600 categories and themes have been attributed to the organizations and congresses (see Figure 2). For the categorisation and assignment of thematic keywords reflecting the areas of social reform (e.g. labour legislation), main causes (e.g. anti-alcoholism, pacifism or feminism) or main ideological background (e.g. Freemasonry), we used previous work done by contemporaries such as Winifred Gregory and Pieter Hendrik Eijkman. Their approach is in line with the use of as many self-definitions as possible, as reflected, for instance, in official conference titles. In Gregory’s International Congresses and Conferences, a thematic index was published (Gregory, 1938) as a guide to the events. He acknowledges that because some of the indexing was done from title only, several errors, omissions and inconsistencies are present. We have made several corrections and have added more detailed themes to the broad categories. Bearing in mind that two sources are never alike and that some of our choices are subject to discussion, our rule of thumb was to attribute a theme to a congress when a session of the congress was organized or a debate was held on the particular theme.

A final major enrichment that contributes to the understanding of social internationalism, is the addition of information regarding organizations related to the congresses, resulting in a subset that can be queried and analysed independently. The difference between an ‘international conference series’ and an ‘international organization’ is not always unambiguous. Congresses played an important role in the establishment of governmental and non-governmental international organizations. After one or more meetings around a certain topic or point of action, it was not uncommon to set up an organization with a fixed structure, a permanent office and communication outlets. Series of international congresses can thus be considered to be precursors of both the governmental and non-governmental international organizations that would come to dominate internationalism in the 20^th century. The first international sanitary conference (1851), can for instance be seen as a direct forerunner of the World Health Organization (WHO). In assembling our dataset, the qualification ‘international organization’ was used if there was a standing committee for the organization of the conference, a permanent secretary or a bureau for the exchange of information (for instance, the ‘Commission internationale des congrès d’éducation familiale’). Again, international religious orders, fraternities and secular institutes have not been included in a systematic way as they are out of the research scope.

It should also be noted that the distinction between governmental and non-governmental organizations is also difficult to make for events and organizations in the 19^th century. Unless explicitly stated in the names or mission statement of congresses, no systematic distinction has been made between the two. Some local and national governments (especially those of smaller countries and France) were actively involved in many initiatives. After the First World War, when internationalism further institutionalised and international relation theory developed, the distinction between organs of inter-governmental organizations (IGOs) and international non-governmental organizations (INGOs), as well as the self-definitions became clearer (Davies, 2019; about secretariats: Reinalda, 2019). In LONSEA - League of Nations Search Engine (directed by Madeleine Herren), these distinctions are made more systematically.⁴ LONSEA is complementary to our project because it is mainly based on the Handbook of International Organisations which was edited by the League of Nations from 1921 to 1938.

(3) In search of long-term patterns and dynamics: following the actors

The biographical data, memberships and congress attendance in our dataset are less comprehensive or complete as not every reform area has been studied to the same level of detail. The dataset consists of the result of several sub-studies and PhD research in which reasoned choices were made in each case. Three different thematic fields of reform were researched in the sub-projects of the PhD-students: (1) The first sub-project (Amandine Thiry) challenged the predominant narratives of the Belgian prison reform. (2) The second sub-project (Thomas D’haeninck) applied Social Network Analysis (SNA) techniques to a large sample of attendees (> 7200 people) who attended international congresses concerned with the moral improvement of society. (3) A third project on colonial hygiene also resulted in a significant amount of information that has been added to the collective dataset (a discussion of the topic in: Vandersmissen and Van Diem, 2021). In addition to this research, we integrated previous research on Belgian feminists and their ‘coalition partners’ among for instance pacifists and the temperance movement (Carlier, 2010). Most importantly, we integrated the datasets of Nico Randeraad and Chris Leonards on a wide range of reform themes, such as penitentiary reform, public and private charity, and social and moral statistics (Leonards 2015, Leonards and Randeraad 2015).

By bringing together all this data from various sub-studies, new possibilities emerge. Clusters of co-memberships at congresses give us a lot of insight into the different ‘causes’ that actors were likely to share (for this approach and a ‘genealogy of causes’, see: Verbruggen, Blomme & D’haeninck, 2020). We looked, for example, at educational internationalism at the intersection of the conferences related to moral reform, education, women’s rights and prison reform (see Figure 3, discussed in Thiry et al 2018).

TIC-Collaborative used Nodegoat as a data management system to collaborate and enrich datasets. Nodegoat (developed by Lab 1100) is a “web-based database management, analysis and visualization platform.” We experimented with harvesting meaningful information from (semi-)structured documents, but most of the data was collected via the traditional full-text search and manual labour intensive (meta)data creation. In the early data collection phase of the project, Nodegoat had no API or linked data ingestion module, so in collaboration with Lab1100, preparations were made to improve the capabilities to query SPARQL endpoints from Nodegoat. This eventually resulted in a Linked Data Module.⁵ Still, most of the disambiguation of persons and linking objects of persons to external identifiers had to be done manually. An interesting proof of concept of the added value of linking data was offered in a CLARIAH-NL pilot project (2TBI).⁶ The main objective was to link a research database of persons who were internationally active in the 19^th and early 20^th centuries, with the person entity repository of currently online available biographical resources in the Netherlands. The result was a group of 1115 Dutch social reformers, whom we can now easily trace in various datasets and link to existing biographical resources (Randeraad, 2018). In order to achieve this, we set up a ResourceSync connection between the Nodegoat software and Anansi, one of the data hubs of the Dutch CLARIAH infrastructure.

(4) From sources to data and data archiving

In addition to the sources that were already available via digital libraries and archives, the researchers have identified complementary sources in Belgian archives and libraries, notably proceedings preserved in KBR, Royal Library of Belgium, UC Louvain Library, Ghent University Library and the Mundaneum (Mons). Many of these sources were digitized by the project and its partners to produce high-quality OCR scans of the documents. Mass digitization as such was not the project’s central aim, rather we wanted to create a virtual infrastructure which offers access to digitized versions of previously dispersed material via a central observatory. With this objective in mind, a detailed analysis was undertaken of the then available ‘text platforms’. A range of existing platforms and technologies (including Islandora, Veridian, TextGrid, Text-Image Linking Environment (TILE), Transcribe Bentham, Shared Canvas, Annotated Books Online and Trove) were evaluated against a range of criteria, including organizational aspects, data aspects (e.g. import, export formats and metadata capabilities), user management capabilities and annotation features (including OCR correction). As a result of this evaluation, Islandora, an open-source software framework for the collaborative management and discovery of digital assets, was selected for the TIC platform. Islandora is an open source digital repository system based on Fedora Commons, Drupal and a host of additional applications. However, during the development and implementation process, a number of issues began to arise, which became more pronounced when using the platform: shortcomings and bugs (e.g. import), the performance and scalability of the ingestion process (e.g. the OCR process and the uploading of text often took many hours, even days) and the management and customisation of the platform.

From-scratch development of a new platform would be unnecessary, as there are many existing solutions that already partially fulfilled the requirements of the corpus management platform. We therefore looked for promising existing solutions and reached out to them with the TIC requirements as a guideline. In collaboration with the Ghent Centre for Digital Humanities and in the framework of DARIAH-BE and CLARIAH-VL, a survey was conducted at Ghent University to gain insight into the current data collection and digitisation processes of researchers involved in text analysis.

After additional discussions with the involved stakeholders (the Ghent Centre for Digital Humanities and the Ghent University Library), it was decided to continue with the implementation of the International Image Interoperability Framework (IIIF), an open framework and standardised method to publish image-based resources, to be viewed, cited and annotated by any IIIF-compatible image-viewing application (overall architecture in Figure 4).⁷ Notwithstanding some limitations (for example its - by definition - focus on images), it is the most advanced and widely supported framework that allows researchers to build corpora across institutional borders and collections. This valuable corpus is made accessible via IIIF and is preserved by the Ghent University Library. As a collection it is accessible on request, downloadable and full text searchable via Shared Canvas.⁸ Shared Canvas is a IIIF-enabled service provided by Ghent University Library for heritage institutions, museums and libraries that want to make their images widely available in a controlled and standardized manner. It is powered by the institutional repository system Librecat. In addition to this, the TIC-collection has been used to test the new IIIF corpus Management Platform (CMP) that has been co-developed by CLARIAH-VL.⁹ The platform, now named Madoc, incorporates a range of open source services, technologies, and recommendations including IIIF, W3C Web Annotation Data model and related protocols, and linked open data.¹⁰

Notwithstanding the fact that part of the biographical data had already been successfully published in Anansi and linked to other datasets, we opted for a different LOD platform with more possibilities.¹¹ Druid is a scalable triplestore by Triply based on the HDT (header dictionary triples) technology (Meroño-Peñuela et al, 2020). The platform was also developed in the context of CLARIAH-NL and is currently maintained by the International Institute for Social History (IISH) and Triply, a CLARIAH spin off company.¹² In terms of content, the IISH is an obvious partner for making the TIC-dataset available for further use or integration with other datasets. To allow additional verification and processing of the data, the data was exported from Nodegoat as CSV files and converted to a Turtle file to be uploaded in Druid using a custom Python script using rdflib.¹³ Congresses, congress series, organisations, publications and places were mapped to the classes Event, EventSeries, Organization, CreativeWork and City of the schema.org vocabulary. Tags were mapped to SKOS Concept¹⁴ and linked to entities using Dublin Core Metadata Initiative (DCMI) Metadata Terms Subject property.¹⁵

Last but not least and as an illustration of the potential of the data, we staged an online exhibition on social reformers powered by Omeka-S. The interactive exhibition on Couples for progress has been built around the maps, graphs and other visualisations made by the researchers connected to the project in collaboration with our Belgian partners Amsab Institute for Social History and Liberas. The exhibition makes use of the entire IIIF and Linked Open Data ecosystem, finally tying up all ends. Via iFrames, the exhibition also refers to visualizations in Tableau and a public interface including sample graphs in the project’s Nodegoat installation that will remain open to new research in the coming years.¹⁶ However, even without this software, the data will remain findable, accessible, interoperable and reusable (FAIR) through the IISH Dataverse Repository.¹⁷