Introduction: exploring the International Statistical Institute, 1885–1938

ABSTRACT

This introduction does two things. On the one hand, it explains why investigating the history of the International Statistical Institute is of interest not only for students of the history of statistics, but also for those tackling more general questions like the relationship between power and knowledge; the scope and development of globalization; or the closely entangled genesis of ‘the national’ and ‘the international’. On the other hand, it introduces the individual papers of this special issue and highlights how the authors contextualize the organization in the outlined more general historiographic framework.

KEYWORDS:

• History of statistics
• international history
• transnational history
• global history
• history of knowledge
• internationalisms and nationalisms
• history of globalization

The nineteenth century witnessed many secular transformations on a global scale, which later proved to be significant turns towards the present world order. The emergence of the nation state (or nationalized empires) as a model for political organization, the development of world markets for certain commodities, the building of potentially global transport and communication infrastructures or the founding of international organizations for a multitude of tasks are examples of such interconnected and complex long-term processes.¹ Closely entangled with all of these processes was the increasingly practised technique to permanently observe social phenomena by means of numerical statistical methods. Emerging during the eighteenth century and fuelled by both the scientific curiosity of (often liberal) scholars and intellectuals, as well as the administrative needs of the evolving bureaucratic state, statistics became ‘the most important tool for the constant self-monitoring of society’ in the nineteenth century.² This was true at least in large parts of Europe, the British dominions, the United States and, from the 1870s onwards, in Japan. Mainly due to the significant political and economic influence these mostly ‘Western’ countries had on other parts of the world, the establishment of statistical infrastructures, concepts and methods to describe, observe, understand, plan and regulate social phenomena were ‘globalized’ from the early twentieth century onwards. Until today this has not changed. We are living in societies which are highly dependent on and significantly shaped by statistical observations.

As many scholars from the humanities and social sciences have highlighted, statistics is characterized by a most relevant double character.³ Simultaneously, it is descriptive and prescriptive. It observes the ‘real world’ in a specific, clearly defined way and at the same time it creates a ‘statistical reality’ which, in turn, affects the ‘real world’. ‘Migration’, ‘society’, ‘unemployment’ or ‘economic growth’ are examples of such artefacts. They are based on permanent statistical observations; they are obviously constructs that undergo changes when the concepts and methods of the observation are altered; and they are politically of the highest importance. Historically, it is therefore of significant interest to investigate who defined statistical concepts and methods and when, where and in which contexts this occurred; to explore how the actual observation was organized and conducted and how this shaped the knowledge on a certain phenomenon; and to study how and why these quantified worldviews have gained such widespread trust and influence in the nineteenth and especially the twentieth centuries.

This special issue addresses some of these questions by focusing on the International Statistical Institute (ISI) in the first five decades of its activity. The ISI still exists today and has been – since its establishment in 1885 – the most important academic association of statisticians at the international level. Founded as an institution which officially was supposed to be an apolitical gathering of an explicitly limited number of statistical experts, it has changed its character several times since then. However, there are only a few historiographical studies on the ISI and its development; its relevance for the emergence of international statistics, in particular, remains widely unknown. It is therefore one aim of this special issue to contribute to a better understanding of the ISI’s history and its role in international statistics as well as to encourage more research on it, especially because the authors reveal that the organization was of importance for the development of statistical observations in manifold ways. In particular, they ask how and to what extent the ISI has helped to shape statistical observation techniques applied on state or regional levels and what role it played in the emergence of internationally recognized social scientific concepts like ‘population’, ‘migration’, ‘national income’, ‘economic development’ and the ‘world economy’. By this the special issue sheds light on the political and institutional contexts in which these concepts were developed, investigates the protagonists who stood behind them and examines how and why these actors made use of the ISI for reaching their aims. In a more general sense, and this is the second aim of this special issue, it contributes to the growing literature on the role international non-state institutions played in the shaping of the modern world.⁴

Before the remaining part of this introduction summarizes the existing research on the ISI and outlines how the individual articles of this special issue contribute to a better understanding of the organization, it briefly sketches out why a focus on international statistics (including, but not limited to, the ISI) is a worthwhile endeavour not only for scholars interested in the history of statistics.

First of all, a study of international statistics contributes to a historiography focused on the relationship between power and knowledge.⁵ From the very beginning in the eighteenth century, statistical observations were closely connected to state-building and administrative logics.⁶ It was therefore prone to apply concepts and methods adapted to the respective political, economic, social and cultural contexts and needs within individual states. This was mirrored by the founding of state statistical offices, as well as by nationally oriented statistical associations and academic institutions in an increasing number of countries.⁷ At the same time, and especially in an ‘era of [statistical] enthusiasm’⁸ in the 1830s and 1840s, statisticians and statistically interested intellectuals became convinced that societies (or even mankind) develop according to certain ‘laws’ which could be made visible by means of permanent statistical observations. Such ‘laws’ were thought of as being anthropological constants with universal validity. Consequently, statisticians were at the forefront of those who, from the mid-nineteenth century onwards, strived for an international standardization of concepts and methods.⁹ Ever since, the actual practice of statistical observation was therefore situated between a scientifically based striving for universalizing concepts and methods in order to compare social phenomena from all over the world, and the relatively narrow limits within which this was actually possible. Different cultural habits, diverse (and often competing) political and administrative systems, diverging social and economic structures and processes, institutional logics and dynamics, efforts to establish and maintain national prestige as well as limits in terms of budgets and personnel made (and still make) it very difficult to agree on conceptual and methodological standards on an international level and to apply them within individual states. It is the study of these tensions among science, politics and social contexts which makes the investigation of international statistics and related institutions like the ISI most relevant for students with broader interests.

Secondly, an exploration of emergent international statistics brings into view the complex process of the closely intertwined development of ‘the national’ and ‘the international’.¹⁰ In order to conduct statistical observations, decisions have to be taken as to what should be observed in which way. When the aim is to observe social phenomena like population, migration, production or trade, the latter aspect always has to include decisions concerning periods, possibly intervals and certainly territories. The resulting figures therefore always refer to specific times and places. As such, however, they are of limited utility for decision-makers or social scientists. These actors need diachronic and/or synchronic comparisons to make sense of the numbers and to derive decisions and interpretations from them. ‘The soul of statistics is comparison’, therefore, was and is proverbial wisdom among statisticians.¹¹ Both the mandatory definitions to conduct meaningful statistical observations and the necessary comparisons to interpret the numbers constitute the above-mentioned double character of statistics. Neither the categories nor what they represent nor the modes of comparison exist outside the ‘statistical reality’. Nonetheless, all of this develops a life of its own and begins to impact the observed ‘real world’. The political, demographic, social, economic or environmental characteristics of a certain territory – including whole countries – can be made visible in such a quantitative form and will contribute – when compared to other territories – to its specific identity. Whereas the comparison between countries was not new in the nineteenth century, it is its quantified form that has become increasingly important since then.¹² The creation of modern statehood, of the nation state and also of national identities which, in several regards, had to be different from others in order to be recognizable, was therefore closely connected to statistical observations within individual countries and to comparisons of the resulting numbers with those of other states in international statistics. However, to actually compare a certain phenomenon of one country with that of another mandatorily requires the application of the same concepts and methods. International cooperation in order to agree on such common standards was necessary in order to observe the differences (and commonalities, of course) one was looking for. In short: statistics is certainly not the only, but it is a very promising field to explore the interdependence of ‘the national’ and ‘the international’.¹³ This holds all the more true if one acknowledges that the (international) comparison of (national) numbers for the purpose of generating knowledge is not the only motive for governments or other nationally oriented actors to participate in efforts to establish international statistics. Prestige, the hope to make one’s own country globally visible as well as the willingness to secure influence in international bodies and to shape international standards were similarly important.¹⁴ Because of this close entanglement of the two spheres it has been recently proposed to introduce the term ‘inter/national statistics’.¹⁵

Thirdly, focusing on international statistics helps to extend and differentiate the historiography on ‘globalization’ or, better, on ‘globalization processes’. This has several dimensions. One of them is obviously the border- and continent-crossing material infrastructures which institutions or individuals needed to obtain data from other world regions and to distribute their own figures and interpretations. Whereas such infrastructures hardly existed in a permanent form at the beginning of the nineteenth century, many of them were established and expanded in the following decades. By the mid-twentieth century, such infrastructures allowed for a steady and practically global exchange of data and statistical publications. To monitor ‘global’ migration flows on a regular basis had become possible just as this was true for a permanent observation of ‘world trade’ or ‘world production’. Historians should pay more attention to this material dimension of international statistics as it has often limited the possibilities to observe social phenomena ‘globally’. In a similar vein, they should also be more attentive for the actually covered areas in seemingly ‘global’ or ‘world’ statistics, as well as for the ways data is presented in them. Both aspects mirror not only existing (and non-existing) statistical infrastructures; they also simultaneously reflect and solidify particular notions of world orders.¹⁶

A second dimension refers to new interpretations of ‘the world’ which international statistics facilitated. As a result of the global proliferation of the Western-modelled bureaucratic state from the nineteenth century onwards, the different (but similar) forms of statistical administrations were also established in an increasing number of countries. Although in a limited form and often only from the early twentieth century onwards, this also included politically dependent territories like colonies or protectorates.¹⁷ This made it possible to include these countries in international (or even world) statistics and to compare them with others. Due to the methodological necessity that only entities of the same kind can be compared to each other (commensurability), this successively led to a discursive equalization of all these countries. For the creation of concepts of economic development for example, such an equalization was mandatory. Only then was it plausible to think of universally applicable means to ‘develop’ a certain country or society with the aim of reaching levels of production, trade or standards of living like those of other countries. Former modes of distinctions like racial, cultural or civilizational hierarchies which were often thought of as irreducible could now be replaced (or at least supplemented) by principally reducible differences like national income, literacy rates or potential work forces.¹⁸ By means of international statistics, ‘the world’ could be portrayed – and treated – as a planet seemingly composed of principally kindred countries.

Furthermore, the quantification of social phenomena in the form of international statistics made it possible to aggregate numbers into regional, continental or even global figures. A ‘global population’, ‘world production’ or numbers for ‘world-wide migration’ are apparently all-encompassing aggregates that are based on a high number of individual sets of data and numerous conceptual and methodological decisions. The wider the international statistical infrastructure became, the more countries were regularly involved in relevant statistics and the higher the frequency of data-gathering and computing was, the more convincing and ‘normal’ such ‘global numbers’ appeared.

For example, one can argue that the global economic meltdown of the early 1930s was not so much a starting point for statistically observing and analysing the ‘world economy’ as some historians have claimed.¹⁹ Rather, it seems that it was the other way around. It was only the League of Nations’ supply of monthly updated trade and production figures on a seemingly global level that made it plausible to contemporaries of the early 1930s to witness that and how the ‘world economy’ was about to collapse. And it was (also) this high-frequency coverage of an apparent ‘global’ process that evoked the harsh reactions of politicians and businessmen we all connect with those years. A decade before, such coverage would have been impossible because the infrastructure necessary for it did not yet exist. To put it differently, in the case of the Great Depression, it needed a process of accelerated ‘globalization’ – the establishment of a well-working international statistical infrastructure in the 1920s – to reveal a process of considerable ‘de-globalization’ – the diminishing of the volume of ‘world trade’ and the crashing of the ‘world economy’.²⁰

The articles in this special issue contextualize the ISI in such a history of international statistics mainly in two ways. On the one hand, they provide new insights into its institutional development, membership and political relevance. On the other hand, the articles analyse how and to what extent the ISI actually was involved in the shaping of some widely acknowledged social-scientific concepts which are essentially based on statistics (‘population’, ‘migration’, ‘national income’, ‘economic development’ and the ‘world economy’). Doing this, the authors could make use of a rich collection of very diverse sources. Particularly, this meant voluminous publications of the ISI as well as of other institutions and individuals, but also the archival documents of governments, organizations and associations.²¹ Unfortunately, the unpublished material of the ISI itself was hardly available except for its direct correspondence, circular letters and similar material sent to cooperating and interested institutions and individuals.²² This certainly makes it harder to understand the ISI’s internal evaluation processes, viewpoints and decision-making. However, the way the authors of this special issue approach the ISI’s activities and relevance and their utilization of a multitude of published and unpublished sources helps to avoid the danger of biased analyses which comes along with the lacking accessibility of the ISI’s archives.

Compared to the wide range of sources, there is only a small body of related research literature. It is remarkable that the organization has hitherto not received scholarly attention like the preceding International Statistical Congresses (1853–76).²³ What is known about the ISI’s history is mainly based on two book-length accounts authored by one of its former presidents and a long-time ISI member, respectively.²⁴ In addition, a few articles inform about the founding and development of the ISI.²⁵ These texts are certainly informative regarding the institutional development, important actors and activities, and they give, especially in the case of the two books, quite biased interpretations of decisions and developments. Only recently have historians put the ISI into the focus of more analytical studies. Whereas Marc-André Gagnon queried the structural position, entanglement and function of the ISI within the expanding internationalist world of the late nineteenth and early twentieth centuries, Jan-Philipp Horstmann analysed the organization and the preceding Congresses as case studies for exploring the tensions between scientific universalism and state-centred interests of statistical authorities.²⁶

Besides basic facts like the founding dates, the main topics of the ISI’s biennial congresses or the principal officers of the organization, the authors of this literature widely agree on a number of interpretative aspects.

Firstly, the history of the ISI can be divided into several periods. After its foundation in 1885, a first era of fruitful debates on statistical concepts of all sorts lasted until 1913-14. A first turning point was 1913, when, after several years of discussion, a Permanent Office was established at The Hague. During the First World War, this office could uphold a certain level of trans-border communication from the neutral Netherlands, and it could also initiate the publication of some international demographic statistics. Nonetheless, the war itself pushed the ISI into an existential crisis out of which it could escape only in the early 1920s when it began a close cooperation with the newly established League of Nations and the International Labour Office. A further crisis occurred in the late 1920s and early 1930s when the ISI’s Permanent Office was increasingly criticized by some members. The major caesura was the Second World War. After the break-up of its regular congress in Prague in September 1938, the ISI could uphold almost no meaningful form of cooperation during the conflict. Only in 1947 could it re-convene in the framework of a combined gathering of several organizations in Washington, DC.²⁷ The articles of this special issue generally confirm this periodization. However, Léa Renard and Yann Stricker, as well as Jan-Philipp Horstmann, suggest in their contributions that the founding of the ISI was not so much a new start for statistical internationalism after the failure of the International Statistical Congresses, but rather an adjustment of existing expert networks. Martin Bemmann, in turn, interprets the 1920s more distinctively than previous authors as a crucial decade for the ISI, during which it had to reinvent itself twice and after which it became the international learned academy it has been ever since. From this perspective, it was not so much the Second World War that represents the turn to the ISI’s institutional present, but the years around 1930.

Secondly, a major reason for the collapse of the International Statistical Congresses was seen by contemporaries – and later authors – in the ambition of many of the involved experts to sustainably change national statistics along internationally agreed rules. As a consequence, the founders of the ISI decided to shape the new organization as a strictly non-political institution aiming at academic debates and submitting recommendations for good statistical practices. This claimed non-political character of the ISI is a central topic for almost all the following articles. Depending on their perspective, they challenge this assertion to a greater or lesser degree. Not only is it generally problematic to see science as a sphere which could be ‘apolitical’: the very essence of statistics of being prescriptive and descriptive at the same time, as well as its crucial relevance for state administrations in the nineteenth and twentieth centuries, makes the assumption of a non-political ISI implausible. Quite concretely, Renard and Stricker, as well as Horstmann, argue that because of this close entanglement of statistics and the state, there was no escape from the ‘political sphere’ for any of the people involved in the debates of the ISI. Also, as these experts were discussing the statistical dimensions of concepts which served as crucial bases for social and economic policies (like ‘population’, ‘migration’ or ‘national income’), the debates had huge political relevance. This is also true for the concept of ‘development’. Martin Rempe sees it emerging in the ISI’s debates on labour statistics prior to the First World War. This was long before the actual coining of the term and outside the usually studied realm of colonial development. The contribution of Aykiz Dogan introduces another layer of political importance of international statistics in general and the ISI in particular. On the one hand, by focusing on Turkey’s first population survey in the mid-1920s, she reveals how important it was for the young Turkish republic to obtain an image of being a ‘modern state’ by establishing a statistical administration shaped according to European models and to become involved in the respective ‘Western’ expert circles. It was the ISI’s recommendations which served as the most important orientation for the modelling of this survey. On the other hand, Dogan reveals that it was important for the Turkish government to find a foreign statistician who, by means of an ISI membership, could confirm sufficient expertise and who came from a politically ‘fitting’ country. Finally, Bemmann highlights the political instrumentalization of the ISI by governments (or, better, government officials) in the aftermath of the First World War.

Thirdly, the existing literature has underlined that due to its purportedly apolitical character the ISI could and did not want to prescribe any concepts and methods for statistical observations in individual countries. It did not understand itself as an agency for international standardization like the preceding Congresses had intended to become, and like the United Nations or the European Commission are today. However, previous authors left no doubt about the influences the ISI and its recommendations could have on national practices by means of the organization’s professional standing and by means of its members, which were often senior officials in their home countries. From the late nineteenth to at least the mid-twentieth century, the ISI helped to establish a certain kind of framework within which statistical concepts, methods and practices could develop in quite different ways, but outside of which statistics seemed not to be acceptable. Or, as Gagnon has put this, the ISI ‘joue plutôt le rôle d’idéologue, de producteur de normes d’une […] internationalisation [statistique]’.²⁸ This special issue strengthens the claim not to judge the ISI in terms of success or failure of standardization. Rather, the articles argue for understanding the ISI as a multidimensional ‘thing’ that functioned on different levels and in different ways. Most obviously, as especially Renard and Stricker as well as Rempe show, the ISI served as a forum for scholars, experts and government officials to discuss certain statistical concepts and methods which significantly shaped what was understood as ‘population’, ‘migration’, ‘economy’ or ‘development’. In this context, it is especially important to note that the ISI was not a gathering of ‘statisticians’ in a narrow sense. As the articles make clear, social scientists of all kinds attended the organization’s congresses, read and engaged with the published discussions and results, and integrated the ISI in a much broader landscape of international associations and organizations with direct links to governments. Neither the history and political relevance of this landscape nor the underlying, often informal institutional and social networks are well known. The articles of Rempe, Dogan and Bemmann in particular show that the exploration of how and where the ISI gained relevance brings into view hitherto underrated actors and contexts.

A fourth aspect which can be learned from the existing literature, although in an implicit manner, is that the ISI was dominated by European actors through the whole period of interest. British, French, German, Belgian and Italian experts provided the highest number of ISI members and congress participants; consequently their interests and opinions also shaped the debates and recommendations. But it can easily be seen from the well-documented lists of members and participants in the ISI’s Bulletins (often containing even short professional biographies of the included persons) that the organization had a much broader basis (and outreach) than this. Russia and increasingly also the United States were countries which made up a huge number of ISI members and congress participants from 1887 onwards. Experts from Latin American countries (especially from Argentina) were often present as well, and, from 1899, Japanese statisticians attended all of the following congresses. After the First World War, the number of represented non-European countries increased and even Soviet officials and scholars attended the congresses, at least before the beginning Stalinization of the state’s administration in the last third of the 1920s. Similarly remarkable is the fact that, after Chicago in 1893, extra-European places served as venues for the ISI’s congresses several times (Cairo 1927-1928, Tokyo 1930 and Mexico City 1933). As especially Dogan’s contribution to this special issue reveals, a closer analysis of the direct and indirect consequences of the increasing global outreach of the ISI would be a very fruitful endeavour, promising stories of the possibilities and limits of a Europeanization or ‘Westernization’ of the world.

Of course, this special issue cannot offer the or even a history of the ISI. Instead, its authors aim at providing some glimpses on a number of interesting aspects which demonstrate that it is a fruitful endeavour to integrate the organization into the broader historiography of (international) statistics. Even more they hope that the special issue will encourage scholars exploring topics from trans- and international as well as from global history perspectives to take into account the relevance statistics – and international statistics in particular – had for many of the transformation processes that have changed the world since the nineteenth century.

Acknowledgements

I would like to thank Martin Rempe as well as the two anonymous reviewers for their valuable comments and suggestions on earlier drafts of this article.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

Martin Bemmann

Martin Bemmann is a lecturer (Privatdozent) at the History Department of the University of Freiburg (Germany). He is interested in European history of the nineteenth and twentieth centuries in global contexts, as well as in economic and environmental history from a global perspective. He has recently finished a second-book project on the emergence and development of international economic statistics from about 1850 to 1950.

Notes

1. Bayly, The Birth of the Modern World; Osterhammel, The Transformation of the World; Rosenberg, A World Connecting.

2. Osterhammel, The Transformation of the World, 26; see more generally Desrosières, The Politics; Porter, The Rise; Randeraad, States and Statistics; Prévost and Beaud, Statistics; Frängsmyr, Helbron and Rider, The Quantifying Spirit; Beaud and Prévost, L’ère du chiffre; Bilo, Haas and Schneider, Die Zählung der Welt.

3. Desrosières, The Politics; Porter, The Rise; Heintz, “Numerische Differenz.”

4. Davies, NGOs; Davies, “A ‘Great Experiment’”; Davies, “The Roles”; Reinalda, “NGOs.”

5. Fundamentally on this tension in regard to statistics see Desrosières, The Politics.

6. Behrisch, Die Berechnung.

7. For the emergence of statistical systems within individual states see, out of a huge body of literature, Bréard, “Reform”; Brückweh, Menschen zählen; Göderle, Zensus; Goldman, Victorians; Mespoulet, Statistique et revolution; Okubo, The Quest for Civilization; Patriarca, Numbers and Nationhood; Prévost, A Total Science; Randeraad, States and Statistics; Schneider, Wissensproduktion im Staat; Schweber, Disciplining Statistics; Stapleford, The Cost of Living; Tooze, Statistics.

8. Westergaard, Contributions, 136.

9. Desrosières, The Politics; Porter, The Rise; Randeraad, States and Statistics; Goldman, Victorians; Mazower, Governing the World, 100–1.

10. Sluga, Internationalism.

11. E.g. in Zahn, “Statistik,” 937; or Wyler, “Das schweizerische Volkseinkommen,” 174.

12. Steinmetz, Force of Comparison; Steinmetz, “Macht – Leistung – Kultur.”

13. Hansen, Mapping the Germans; Göderle, Zensus; Speich Chassé, “Nation”; Tooze, “Imagining”; most instructive for this reasoning is Heintz and Werron, “Wie ist Globalisierung möglich?”

14. Bemmann, “Comparing”; Bemmann, Weltwirtschaftsstatistik.

15. Wobbe, “Die Differenz.”

16. Heintz, “Welterzeugung.”

17. Bayly, Empire and Information; Speich Chassé, “The Roots”; Bemmann, Weltwirtschaftsstatistik.

18. Speich Chassé, “Fortschritt”; Speich Chassé, Erfindung; Bemmann, Weltwirtschaftsstatistik.

19. Ekbladh, “American Asylum,” 633; Slobodian, Globalists, 57.

20. Bemmann, Weltwirtschaftsstatistik.

21. This includes mainly the Bulletin de l’Institut International de Statistique (1886 ff.) and the Revue de l’Institut International de Statistique (1933 ff.).

22. Fragments of the ISI’s archival material are part of the collection of the Dutch Centraal Bureau voor de Statistiek in the Dutch National Archives (2.06.118/746). Most of the documentation, however, is currently not accessible (emails of Ada van Krimpen, director of the ISI’s Permanent Office, The Hague, to the present author, 25/26 February 2018).

23. For the history of the Congresses see especially Randeraad, States and Statistics; but also Brian, “Y a-t-il”; Brian, “Observations”; Schneider, “Internationalisierung”; and Goldman, Victorians, 241–56.

24. Zahn, 50 années; Nixon, A History.

25. Especially Campion, “International Statistics,” 108–11; Menges, “Versuch,” 44–6; Bott, “Die Statistik,” 69–71; Horvath, “Le concept,” 290–2.

26. Gagnon, “Les réseaux”; Horstmann, Halbamtliche Wissenschaft.

27. Speich Chassé, “In Search.”

28. Gagnon, “Les réseaux,” 210.