Navigation – Plan du site

AccueilNuméros13-2Revue des livresComptes rendusLaurent Feller and Agnès Gramain,...

Revue des livres
Comptes rendus

Laurent Feller and Agnès Gramain, L’évident et l’invisible. Questions de méthode en économie et en histoire

Loïc Charles
p. 591-598
Référence(s) :

Laurent Feller and Agnès Gramain, L’évident et l’invisible. Questions de méthode en économie et en histoire [The Obvious and the Invisible. Questions of Method in Economics and in History.], Paris: éditions de la Sorbonne, 2020, 300 pages, 979-103510546-4

Texte intégral

Afficher l’image
Crédits : Editions de la Sorbonne

1The book is a follow-up of an extremely interesting project organized through a series of workshops that took place mostly at the Université of Paris 1 Panthéon-Sorbonne from August 2014 to June 2016. They involved PhD students at the time of the seminars and their advisers, to which a few senior researchers were added. The result is a rich volume made of a general introduction, a general conclusion and 3 parts, totaling 15 essays of different length. Its general purpose is to “reflect collectively upon the notions of sources and data and the alternative conception that economists and historians might have on those.” (7) To some extent, it is an understatement since through the lens of sources and data, the volume provides an interesting point of view on two related issues: that of the complex and uneasy relationship between history and economics as well as that of history and statistics. Indeed, one of the interest this reader has found in this volume is that it is part and parcel on an on-going discussion taking place in parallel in history and in economics on the status of the interactions between the two disciplines, and more generally on the purpose of history as a discipline and its place relative to social sciences and statistics. Without any pretentions for exhaustivity, I have in mind here the following references: Boldizzoni (2011), Thomas Rawski (1995), and Daumas (2012).

2The first part of the book “Building the Data: Actors and Processes” (“Construire les données : acteurs et processus”) is dedicated to the issue of extracting significant sets of economic and social data from documents that were constructed in a very specific context and which, therefore, bear the imprints of this context.

3In an introductory chapter, Pierre Gervais, an economic historian, argues that data are built in a four steps process, each of them encompassing a certain number of epistemological pitfalls. First, each document is created by an original author, who has in mind a specific public that may modify it (second step); a third step occurs with the archivist, who manipulates the document or its public record, which finally arrives on the desk of the social scientist, who uses and interprets it in the context of his research.

4The three following chapters illustrate this process. The first chapter by Philippe Lefeuvre discusses an archive—the Diplomatico—made of a huge set of medieval Italian parchments of notarial acts. Delving into the history of the construction of the archival fund with a use of a database he created from it, the author navigates through the various written traces registered on the parchments to unearth the social and economic meaning they carry for the historian. His argument is that in this case, it would be mistaken to separate the exegesis of the document from its exploitation as data: the two operations are inseparable.

5The second chapter by Louise Quillien takes us to ancient Mesopotamia to discuss texts written in cuneiform on blocks of clay circa the sixth-fifth centuries BC. These documents were produced in the region of Sippar, a town about 40 kilometers north of Babylonia, to record orders of textile fabrics made by craftsmen for the temple of Siddar. These goods were intended as religious gifts. The documents are not situated in one archive, but have been dispersed—a large part of them are preserved in the British Museum, but many are in privates hands and other Museums throughout the world. They are also available in various forms, some can be observed directly in museums, other are known only through databases of photographs or reproductions in academic publications. Historians put to use these documents in various contexts to produce models of the Babylonian economy, to understand the economy of Babylonian temples and the relative proportion of the private and domestic sector among their suppliers in textile. Interestingly, Quillien shows that these interpretations varied according to the nature of corpus of documents used as source for the research.

6The third chapter, written by a collective of three authors, focusses on the set of public or official data produced in the United States to measure racial discrimination either from surveys or in an experimental setting from the 1970s onward. There are two surveys: the Current Population Survey (CPS) and the Panel Study of Income Dynamics (PSID). The CPS is a database produced from information collected by the Census Bureau managed directly by the US Administration and the PSID a byproduct of Johnson’s administration war on poverty managed by the University of Michigan. These public data were used by social scientists to measure statistically the economic weight of racial discrimination in the society. In addition to this, experimental economists have imagined new ways to explore the issue of discrimination. The notion of race and the various categories used to account for it and its role in the social fabric are not given, but are on the contrary one of the, if not the, outcome of the research and the interaction between the individuals who are taking part in the experiment(s). The authors insist, and rightly so, that the rise of experimental economics implies that the social scientist is now the author and curator of the data used to measure discrimination as compared to the earlier situation when he was only using data produced and managed by the administration.

7The second part of the book “Contextualizing Data” (“Contextualiser les données”) is introduced by the historian and sociologist of science Michel Armatte who discusses contexts and regimes of production (and use) of data. This part focusses “on the process that linked the production of data to its uses and to the contexts be they social or cognitive” that frame this issue. The author makes four suggestions to accompany the reading of the three case studies that follow the introduction. First, the data are not given, but produced by a sophisticated apparatus; second, this production can be related to a specific regime of knowledge that makes it relevant for a given society; third, when dealing with data at some point in time, one should account both for the historical context in which data were produced (the reconstruction of producers’ goals and intentions) and the way these data, in turn, have been used and manipulated by historians and social scientists of the past ; finally the use of series and aggregated values over long periods of time creates issues of homogeneity of the data and of their contexts.

8In the first chapter, Jean-Christophe Balois-Proyart analyses statistical compendiums on respectively criminal, civil and commercial justice published between 1827 and 1864. Drawing on various published and manuscript sources, the author studies both the rise of the demand for statistics in the French ministry of internal affairs and the practices that allowed the administrative actors to produce them. He shows that these practices differed from one city to the other, which resulted in a significant variety in the processing of the legal decisions of the prud’hommes courts (industrial courts) and their administrative accounting, as well as the information linked to it.

9The second chapter discusses how numerical data were used in a very specific context, that of the German occupation of France during World War II, to gain some traction against political authorities, the German occupant and the Vichy regime. The producer of data was the Compagnie française d’éclairage et de chauffage par le gaz, a mid-size company that produced and distributed coal, electricity, gas and water with plants all over France. These statistics were made by each factory by the director and either an engineer or the chief-accountant. They were sent to the head office, which produced two monthly reports, one on gas and one on electricity. In addition to these technical reports, each central administrative division (human resources, commercial supply, etc.) edited a monthly report. Djouhra Kemache shows that the data produced and communicated to the political authorities were altered by the company to obtain some degree of liberty, for example the coal supply was systematically underestimated. Also, the company did not hesitate to negotiate directly with the Germans to improve its situation.

10The third and last essay of this part by Tristan Jacques and Erich Pinzón-Fuchs expounds the construction of a statistical system of French internal trade and the implementation of a bureau in charge of it from 1945 to the end of the 1960s. This construction parallels the transformation of the French distribution system, from one dominated by very small retail shops to one dominated by mass retailers, large stores and supermarkets. Hence, 1959 was the year of both a major legislative change in the retail industry as well as of the beginning of an active policy to rationalize the inconsistent and incomplete array of existing commercial statistics. By 1970, French commercial statistics provided detailed and standardized information and formed a complete sub-system of national accounts. The authors point out that big retailing firms as well as the organizations of retailers played an important role in collecting the data.

11The last part of the book “Serializing Data” (“Sérier les données”) is introduced by Laurent Feller’s essay on the legitimacy of serial analysis in history. He recalls to us the traditional suspicion of historians against quantitative history, based at least for a significant part on their lack of training in advanced statistical methods. In the first chapter of this part, he and Dominique Margairaz, throughout a handful of case studies belonging to the medieval and early-modern history, ponder upon the necessity for the historian to understand and disentangle the process(es) of (re)classification through which each document went through. This analysis is, they argue, an essential step that the historian has to go through before he is able to extract from the document its pertinent historical meaning and to insert it in a statistical series.

12The second essay by Pierrick Dechaux and Robin Hege analyses the issue of the construction and interpretation of indicators. Indeed, a lot of economic data are not raw data, but indicators. The underlying reason is that most of economic and social quantifiable concepts are multidimensional. They use two case studies to illustrate their argument: first, an indicator built to measure the complexity of an application file to qualify for unemployment benefits in the French social system; second, the development of surveys in the US to build social psychological indicators in the 1950s. These two cases illustrate two aspects of the interplay of theory and data. In the first instance, the necessity to measure resulted in the simplification of the underlying theory to ‘fit’ the data constraints. In the second case, it is on the contrary the theory that drives both the indicator chosen and the method to measure it.

13In the third chapter, Cédric Quertier draws on a diversified set of sources to reconstruct a monthly series of Florence’s trade with Pisa from 1374 to 1382. The author wants to measure the impact of political events on the trade between Florence and Pisa and more generally, whether the turbulent political history of Pisa in the 1380s can be linked to their commercial relations. Confronted with a dearth of data in the Pisan archives, Quertier must rely on incomplete series he found in Florence. However, by themselves these data provide only a very indirect and partial view on Florence’s trade with Pisa and the author has to manipulate and complement them with other sources to create a monthly series of bilateral trade. The construction of the latter allows the author to show both the impact of the war between Florence and the Pope on its trade, the commercial boom created by its ending and the political tension that this boom created in Pisa that resolved into a riot that ousted the local government favorable to Florence’s economic interests.

14The fourth and last chapter by Camille Chaserant, Julie Claustre and Agnès Gramain presents three case studies where juridical decisions are quantified in order to understand what norms, or set of norms, drive the practice of the judges of each of these cases. The authors want to describe and compare the methodological processes that led the researchers who worked on each of these cases from the raw data to the series (step 1: standardization of the data), through the aggregation of the data which have been put into series (step 2: homogenization), to the final interpretation by inference (step 3: generalization). They draw several lessons from their investigation: first, some differences attributed to disciplinary practices are in fact linked to the informational context of the study; second, the periodization does not seem to have a significative bearing on the method chosen by the researcher; the main difference is in the type of statistical inference selected. While econometric methods are based on an a priori model which make sense of the data and frame the interpretation, more historical methods, which often do not use statistical inference, are more open and leave room for thicker (and qualitative) interpretations.

15In a concluding chapter, the editors insist on the convergence of economists and historians regarding the necessity to study the context in which the data were created. Although, they underline that there are several differences between the work of the economists and historians that figured in this collection of essays, they argue that the origin of these differences is to be found in the nature of the sources used rather than in the disciplines to which they pertain.

16The reading of the chapters as well as that of the complementary essays written by senior researchers carries a sense of purpose quite rare in collective volumes. As an economist, historian of economics and economic historian who was involved with the construction of a database and interpreting its data, I have a very peculiar interest in the aim of this book and found it most welcome. The interested reader is invited to visit the webpage of the TOFLIT18 project, dedicated to French trade statistics from 1716 to 1821 (https://toflit18.medialab.sciences-po.fr). Indeed, the issue of data is nowadays—due to the swift technological transformation that impacted both disciplines and more generally science in general—I believe crucial in the disciplines of history as well as in economics (Guldi and Armitage, 2014; Backhouse and Cherrier, 2017). The point of view defended by the editors of the collection is that the economists and historians are faced with the same kind of problems regarding the data and, therefore, the two disciplines might converge on this issue. I wonder however if that optimistic conclusion can be held as soon as one puts the problem in a sociological or a historical rather than in a methodological perspective. Indeed, this proposal may be true at the more abstract level: the data contained in each set or series of data used by social scientists need to be validated through careful contextualization before being used by social scientists and it does not matter if it is done by a historian, an economist or another kind of social scientist. However, the work historians and economists do and the uses they make of these data seem very different to me.

17The average economist uses historical datasets and series to produce analysis and papers. The increased availability of historical series, which parallels the applied turn in economics may have created the illusion that more economists are nowadays more interested in what is sometime called ‘historical economics’ or more often ‘cliometrics’. In my opinion, there is however very little which is historical in most of these works; ‘applied econometrics’ would be a better term to describe them. Average economists have no specific interest for the document and its context. Although part of their job necessitates that they check and clean the data, the techniques they use to do so are statistical in nature (looking for outliers to exclude them for example) and they rarely venture into the archives. For the economists, context is at best a burden to overcome to get the data straight: their work does not necessitate that they engage deeply with the historical context to have value for their peers, quite the contrary in fact. Here and there, some economists do engage with the context when this operation is needed to produce a usable dataset; my point is that it is outside the normal practice of what economists do to be published and to advance in their careers. For historians, the context of production and circulation of each document is their primary object of research. In contradistinction with economists’ practices, historians are reluctant to standardize documents: their limits as data is actually what entices them to delve deeper into their contexts to extract historical meaning from it. We can think of two general strategies mobilized by historians to work on data and historical series to give value to their research. Firstly, some historians concentrate less on the data than on the actors and the institutions who produced them. Several chapters (2, 4, 6) from the volume illustrate this. In this case, the data is just a point of entry to produce a rich historical narrative. Secondly, historians often use datasets and series to produce descriptive statistics—the famous ‘histoire sérielle’ favoured by the Annales school. Labrousse (1933) is a good illustration of what is still considered in the history profession as good work. A long methodological introduction with an incredibly rich discussion of the sources used which is followed by a wealth of tables of numbers and graphics presenting the data and the analysis.

18We have therefore what may be called an extensive use of the documents and an intensive use of them. In the first instance, the document is transformed into data by a process of standardization and homogenization (see the chapter by Chaserant et al.) after which it can be used in econometric models. In the second case, the document is investigated and placed in multiple contexts and provides historical meaning to each of this context. My point is not that one approach of the data is better than the other—there are in fact complementary—but I argue that the disciplines of economics and history are built on very different cultural models of science, which in turn implies a very different attitude towards the data. Hence, I do not query the methodological argument made by the editors, but I think that the sociology of both disciplines makes it utopian. This is not to say that the volume is a failure, it is an important achievement. Researchers—young and not so young—will find here a sophisticated and very useful discussion of a wide-ranging issues on the use of documents and data to produce interesting work in economic history and history of economic thought and knowledge. In this regard, it has no equivalent and, thus, should be carefully read by those confronted to this issue in their own work.

Haut de page

Bibliographie

Backhouse, Roger and Béatrice Cherrier. 2017. “It’s Computers, Stupid!” The Spread of Computers and the Changing Roles of Theoretical and Applied Economics. In Roger Backhouse and Béatrice Cherrier (eds), The Age of the Applied Economists: The Transformation of Economics since the 1970s. Durham and London: Duke University Press.

Boldizzoni, Francesco. 2011. The Poverty of Clio: Resurrecting Economic History. Princeton: Princeton University Press.

Daumas, Jean-Claude (ed.). 2012. L’Histoire économique en mouvement entre héritages et renouvellement. Villeneuve d’Ascq: Presses du Septentrion.

Guldi, Jo and David Armitage. 2014. The History Manifesto. Cambridge: Cambridge University Press.

Labrousse, Ernest. 1933. Esquisse du mouvement des prix et des revenus en France au XVIIe siècle. Paris: Librairie Dalloz.

Rawski, Thomas (ed.). 1995. Economics and the Historian. Berkeley and Los Angeles: University of California Press.

Haut de page

Pour citer cet article

Référence papier

Loïc Charles, « Laurent Feller and Agnès Gramain, L’évident et l’invisible. Questions de méthode en économie et en histoire »Œconomia, 13-2 | 2023, 591-598.

Référence électronique

Loïc Charles, « Laurent Feller and Agnès Gramain, L’évident et l’invisible. Questions de méthode en économie et en histoire »Œconomia [En ligne], 13-2 | 2023, mis en ligne le 01 juin 2023, consulté le 09 juin 2026. URL : http://journals.openedition.org/oeconomia/14701 ; DOI : https://doi.org/10.4000/oeconomia.14701

Haut de page

Auteur

Loïc Charles

Université Paris 8 Vincennes – Saint-Denis. loic.charles02@univ-paris8.fr

Articles du même auteur

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont susceptibles d’être soumis à des autorisations d’usage spécifiques.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search