Observing Disciplines: Data Practices In and Between Disciplines in the 19th and Early 20th Centuries Emma Mojet Abstract: This dissertation observes how disciplines shared data practices. Data practices enable scholars and scientists to transform observations into data that can be systematically collected and analysed. Observations and observation practices have shaped the foundations of the modern sciences and humanities, providing the basis for arguments, evidence, or inspiration to scholars throughout all disciplines. The sharing of observation practices between disciplines shows that disciplinary boundaries are permeable, it does not tell us how they are maintained. Practices of data collection and analysis from observations were shared between disciplines while at the same time disciplines also enforced certain boundaries. This tension between shared practices and creating boundaries, between the disciplinary and interdisciplinary, is the central theme of my dissertation. Relationships between disciplines and the sharing of practices are rather abstract things to study. To do this I apply a new set of historiographical tools. In this dissertation I use the historiographical framework of flow of cognitive goods. Cognitive good is an umbrella term for the shared epistemic tools of knowledge-making disciplines that can be transferred across disciplinary boundaries, such as methods, instruments, concepts, theories, or models amongst many others. In order for these cognitive goods to travel, or to flow, they need to have a certain degree of autonomy: they need to be recognisable. Nevertheless, cognitive goods are not immutable and are dependent on the context in which they are used: they are defined by and used in a community of users. The cognitive goods in this dissertation are the data practices of using the statistical approach as propagated by Adolphe Quetelet and the questionnaire research as employed by Jules Gilliéron, Antoine Meillet, and others. The flow of cognitive goods framework enables a coherent overview of how certain data practices are shared between disciplines, and thus a study of the mesolevel of historical analysis, the level on which disciplines act. These cognitive goods are subsequently embedded in the individual disciplines, which sometimes require adaptations. By considering the shared data practices as cognitive goods this dissertation can examine processes of discipline formation. The modern notion of discipline—as institutionalised social entities encompassing research and education with agreed-upon methods, topics, and practice—is relatively recent, emerging in the first half of the nineteenth century and following structural transformations of the university systems. Discipline formation is multifaceted: social, political, and institutional factors play a role in decisions of what is considered part of the discipline or outside it. The content of the research and the methods or objects of study also clearly play a role in the forming of an academic discipline. Discussions between scholars on the methods or objects of study of disciplines have their influence on and are often influenced by the work of individual scholars. In order to combine the abstract level of disciplines and the particular level of individuals, and in order to analyse not only the sharing of cognitive goods but also the emergence of disciplinary boundaries, this dissertation employs the concept of disciplinary activity. Disciplinary activity can signify the appropriation or embedding of certain practices within an individual scholar’s research in order to meet disciplinary criteria. It can also be expressed by the debates and discussions between scholars about such disciplinary criteria. I show how disciplinary activity involves interdisciplinarity: the boundaries of disciplines are negotiated after they have been crossed. Both interdisciplinary and disciplinary perspectives are necessary when studying the developments of disciplines and their practices. Deriving from the central theme of the tension between sharing data practices and creating disciplinary boundaries I pose a main research question and two subsequent questions: How did comparable observation practices become part of different nineteenth and early twentieth century academic disciplines? How did these practices develop in different disciplinary contexts? How were the different disciplines influenced by the sharing of these practices? I have concentrated on practices dealing with data from observations by nineteenth and early twentieth century scholars through two historical cases to operationalise these research questions. The cases focussed on disciplines whose boundaries were unclear or in the process of being defined, namely botany and linguistics, which uncovered multiple instances of tension between disciplinary and interdisciplinary interactions. Recent scholarship has described how a focus on data practices can connect historiographies which have been treated separately in the past: history of data is material, inclusive, and political. My research takes this approach further: centralising data practices can help to cross disciplinary boundaries since these practices are shared and borrowed between disciplines. I focus specifically on statistical data practices used to study plants and on the questionnaire research method to collect linguistic data. My dissertation consists of four chapters. Chapter 1 is the Introduction where I lay-out the theoretical framework for my research. I introduce the historiographical framework of flow of cognitive goods and disciplinary activity and my research cases. Chapters 2 and 3 discuss these research cases–Chapter 2 on statistics and botany and Chapter 3 on the questionnaire and linguistics. In Chapter 4 I combine and compare the results from my two research cases and draw conclusions on how data practices were shared between disciplines and on how this sharing influenced both the data practices and the disciplines in question. The separate yet connected research cases are both situated in the nineteenth and early twentieth centuries, and mostly in France and Belgium. While both cases focus on different data practices and different disciplines, together they narrate the nineteenth and early twentieth century development of emerging social science disciplines–from a mix of moral and political sciences to institutionalised social science disciplines such as sociology–and, especially, how the social sciences related to the natural sciences and the humanities. This dissertation offers a multidisciplinary and praxiological approach to the history of the sciences, the humanities, and the social sciences. I show how, through the development of statistical methods in various disciplines, it became possible to measure and observe abstract concepts such as a society or a population. By looking at the use of statistical methods in botany and in social research, I demonstrate a connection between the natural sciences and the social sciences. In addition, I examine the relationship between the humanities and the social sciences through the use of the questionnaire method in linguistics and various social sciences such as sociology (or ethnology) and psychology. The development of the questionnaire to collect data on regional dialects involved an increasing focus on the social situations of the speakers, providing a direct link between these disciplines. Chapter 2 presents how statistical and quantitative methods were developed in varying research contexts and introduces the pivotal work of Adolphe Quetelet in social statistics. Quetelet’s application of these statistical methods, which he had learnt as an astronomer, meant they could be used on many different projects involving observations, spread out over different disciplines. To Quetelet these general observations did not belong to a particular discipline but to a larger, overarching project of observation sciences. Quetelet formed and instructed a network of observers, involving correspondents across Europe and from the colonies, who sent him letters and tables containing data from their observations. These observers were not necessarily scientifically trained, yet Quetelet did recognise them as playing an important role in his new science. This led to Quetelet’s work being criticised for not being able to uphold a standard of rigour deemed necessary for the sciences. In Chapter 2 I focus on how statistical methods were increasingly employed in the discipline of botany to study the relationship between temperature and plant development. To see how Quetelet’s wide-ranging methods were adopted and adapted by specific disciplines, I examine the case of one of Quetelet’s observers, Charles Morren, and their botanical observations and discussions. Quetelet and Morren statistically studied the foliage, flowering, fruiting, and falling leaves for multiple plants over the course of several years. By doing so they hoped to gain understanding of the Belgian climate and its effect on plants, as well as compare their findings to those of other observers in different countries. This resulted in discussions on how to standardise their observations in order to make a comparative study of the topic. Morren and Quetelet disagreed on what should be observed; Morren took the side of the botanists who believed Quetelet’s approach was too general, whereas Quetelet aimed at creating an international network of observers making many general observations. Quetelet placed himself directly in the tradition of the German naturalist Alexander von Humboldt in wanting to observe a large amount of phenomena, ranging from meteorite showers to flowering of plants or from suicide rates to chest spans of soldiers. Morren and other botanists such as Jules Planchon and Julius Sachs, however, required more specific observations to investigate the topics of their discipline. Through the lens of disciplinary activity I argue that the botanists can be seen determining what should be part of their discipline and what should be left out: more specific observations on plants were required. Nevertheless, Quetelet had a broad influence on many disciplines: scholars adopted and adapted his statistical methods in their own research. This had as a consequence that the statistical methods themselves became more sophisticated and advanced, enabling the study of intangible objects such as a society or a population through the use of statistical data practices. Chapter 3 also centres on the observation of an intangible object: spoken language and dialects. To collect data on language, scholars worked with questionnaires. The questionnaire is seen as a data collection method, which, based on questions, enables the systematic study of a certain object that is otherwise difficult to grasp. Questionnaires are best defined, however, by the research project in which they are used. This is done in Chapter 3, where several examples of questionnaires to collect data on languages are discussed. Two influential research projects—the Atlas linguistique de la France and the Sprachatlas, led by Jules Gilliéron and Georg Wenker respectively—illustrate how the questionnaire was adapted to the purpose of the specific research. The questionnaires enabled the scholars to not only collect data on spoken language, but also on various social factors that were believed to influence differences in language. Indeed, the questionnaire method was not only part of language studies, as I show, but also employed in social science research such as sociology, ethnology, and psychology. A notable development was the introduction of the fieldworker questionnaire instead of postal questionnaires. The postal questionnaire was employed by Wenker in his project to map the language varieties in several German speaking regions. While Wenker obtained a large amount of detailed data, his questionnaires were filled in by many different people who were not necessarily trained in linguistics or phonetic notations and hence their responses contained many uncertainties. To deal with this issue, Gilliéron decided to send one fieldworker to collect data on languages spoken in many different villages in France. Gilliéron argued that this method was more rigorous and could meet the standards of linguistic research. Chapter 3 also shows how the field of language studies developed towards a discipline of general linguistics. Different approaches towards a single discipline of linguistics were discussed, including different attempts to generalise the study of language by scholars like Georg von der Gabelentz, Ferdinand de Saussure, and Antoine Meillet in the final decades of the nineteenth century. These developments are linked to developments in the emerging discipline of sociology through the use of the questionnaire and especially through the work of linguist Antoine Meillet. Meillet was among the early followers of Émile Durkheim's sociology in Paris and actively collaborated on Durkheim’s journal. He used Durkheim’s notion of a ‘social fact’ to understand language change and deemed it important to collect data on many different languages to examine this. Meillet proposed to use questionnaires to collect these data and to map the world’s languages. These languages could only be collected with the help of international linguists and required agreements on the methods and practices employed, and thus Meillet initiated the organisation of a Congress of Linguists. At the first International Congress of Linguists, held in 1928 in The Hague, the role of the questionnaire in linguistic research was discussed. This resulted in a standardised international questionnaire developed by Marcel Cohen, a student of Meillet. Cohen was affiliated to the Institute of Ethnology in Paris and used ethnological questionnaires as a template for his linguistic version. This shows how closely the social sciences were involved with other disciplines, in this case linguistics. A direct connection can be made between Cohen and Meillet’s linguistic endeavours and the disciplinary specialisation of sociolinguistics. In both central research chapters of the dissertation I have found that international disciplinary congresses are important sites of disciplinary activity: questions on the preferred methodology of a discipline are discussed explicitly here. Such discussions lead to the consolidation of disciplinary boundaries. Scholars discussed in both parts of my dissertation were influential in organising these interactions: Quetelet took up a leading role in the organisation of the International Statistical Congresses and Meillet was the driving force behind the first International Congress of Linguists. I show that congresses can be seen as events that on the one hand played an important role in the creation and consolidation of disciplinary boundaries, and on the other hand where interdisciplinary activity, and consequently flows, took place. At congresses, questions about the discipline’s boundaries were discussed and research methods were explicitly debated; they functioned as sites of interdisciplinary flow as well as disciplinary activity. In my research I have used the flow of cognitive goods together with the concept of disciplinary activity to illustrate how scholars made agreements about the data practices they intended to use in their disciplines. These agreements shaped the boundaries of their disciplines, as they determined the research methods and research questions which belonged to the disciplines. This perspective on discipline formation is practice-based and discusses processes of specialisation, hybridisation, and professionalisation in an active way. All in all, there is much to see when we observe disciplines. This dissertation has not only given various indications towards interesting cases, but also provided a historiographical framework to research them. By constructing a shared foundation from which to examine multiple disciplines, new questions about the activity within and between these disciplines can be asked. In a time when knowledge is increasingly considered a multi, inter, or even extra-disciplinary product, historical studies into the dynamics of disciplines can be both informative and reflexive.