The social sciences of quantification

Logic, Argumentation & Reasoning 13 Isabelle Bruno Florence Jany-Catrice Béatrice Touchelay Editors The Social Sciences of Quantification From Politics of Large Numbers to Target-Driven Policies Logic, Argumentation & Reasoning Interdisciplinary Perspectives from the Humanities and Social Sciences Volume 13 Series editor Shahid Rahman Logic, Argumentation & Reasoning explores the links between Humanities and the Social Sciences, with theories including, decision and action theory as well as cognitive sciences, economy, sociology, law, logic, and philosophy of sciences It’s two main ambitions are to develop a theoretical framework that will encourage and enable interaction between disciplines as well as to federate the Humanities and Social Sciences around their main contributions to public life: using informed debate, lucid decision-making and action based on reflection The series welcomes research from the analytic and continental traditions, putting emphasis on four main focus areas: • • • • Argumentation models and studies Communication, language and techniques of argumentation Reception of arguments, persuasion and the impact of power Diachronic transformations of argumentative practices The Series is developed in partnership with the Maison Européenne des Sciences de l’Homme et de la Société (MESHS) at Nord - Pas de Calais and the UMR-STL: 8163 (CNRS) Proposals should include: • • • • A short synopsis of the work or the introduction chapter The proposed Table of Contents The CV of the lead author(s) If available: one sample chapter We aim to make a first decision within month of submission In case of a positive first decision the work will be provisionally contracted: the final decision about publication will depend upon the result of the anonymous peer review of the complete manuscript We aim to have the complete work peer-reviewed within months of submission The series discourages the submission of manuscripts that contain reprints of previous published material and/or manuscripts that are below 150 pages/85,000 words For inquiries and submission of proposals authors can contact the editor-in-chief Shahid Rahman via: shahid.rahman@univ-lille3.fr or managing editor, Laurent Keiff at laurent.keiff@gmail.com More information about this series at http://www.springer.com/series/11547 Isabelle Bruno Florence Jany-Catrice Béatrice Touchelay • Editors The Social Sciences of Quantification From Politics of Large Numbers to Target-Driven Policies 123 Editors Isabelle Bruno CERAPS (UMR 8026) University Lille Lille France Béatrice Touchelay IRHIS (UMR 8529) University Lille Villeneuve-d’Ascq France Florence Jany-Catrice CLERSE (UMR 8019) University Lille Villeneuve-d’Ascq France ISSN 2214-9120 ISSN 2214-9139 (electronic) Logic, Argumentation & Reasoning Interdisciplinary Perspectives from the Humanities and Social Sciences ISBN 978-3-319-43999-0 ISBN 978-3-319-44000-2 (eBook) DOI 10.1007/978-3-319-44000-2 Library of Congress Control Number: 2016948621 © Springer International Publishing Switzerland 2016 This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland To the memory of Alain Desrosières (1940–2013) Contents Introduction The Social Sciences of Quantification in France: An Overview Isabelle Bruno, Florence Jany-Catrice and Béatrice Touchelay Part I Sociology of Quantification: Alain Desrosières’s contributions Introduction to the work of Alain Desrosières: the history and sociology of quantification Michel Armatte 17 33 Alain Desrosières’s reflexive numbers Luc Boltanski Alain Desrosières’s spectacles: one lens realist, the other constructivist Emmanuel Didier 43 From statistics to international quantification: a dialogue with Alain Desrosières Roser Cussó 55 Learning from the history of the probabilistic revolution: the French school of Alain Desrosières Fabrice Bardet 67 Part II The Statistical Argument in the Neoliberal Era Quantifying the effects of public action on the unemployed: disputes between experts and the rethinking of labour market policies in France (1980-2000) Etienne Penissat 83 vii viii Contents Counting the homeless in Europe: ‘compare before harmonising’ Cécile Brousse 97 The statistical backbone of the new European economic governance: the Macroeconomic Imbalance Procedure Scoreboard 109 Gilles Raveaud 10 Evaluating public policies or measuring the performance of public services? 129 Florence Jany-Catrice Part III Uses of Quantification : Power and Resistance 11 Private accounting, statistics and national accounting in France: a unique relationship (1920-1960s) 139 Béatrice Touchelay 12 Figures for what purposes? The issues at stake in the struggles to define and control the uses of statistics 149 Marion Gilles 13 The uses of quantification: power and resistance The example of unemployment statistics 161 Pierre Concialdi 14 Statistical argument: construction, uses and controversies Prices and purchasing power 171 Alain Gély, Bernard Sujobert and Béatrice Touchelay 15 The quantification of the social sciences: an historical comparison 183 Alain Desrosières List of contributors Michel Armatte is a statistician and economist, teacher in Paris-Dauphine University and research worker on history of statistics, probability and econometrics at Centre A Koyré He was a long date compagnon of Alain Desrosières His last book is on Economics as an Engineering (2010) Fabrice Bardet is a senior researcher at University of Lyon (ENTPE) He is specialist on sociology of quantification and urban development He holds a PhD in Political science from the University of Paris La Sorbonne (2000) and the habilitation to be PhD supervisor (HDR) from Lyon University (2013) He has recently published La contre-révolution comptable Ces chiffres qui (nous) gouvernent (Paris, Les Belles Lettres, 2014) Luc Boltanski is director of research in sociology at the French School for Advanced Studies in Social Sciences (EHESS) Regarded as a leading figure in critical thought and what has come to be known as French pragmatic sociology, his influence in the fields of sociology, political science, economic history and social economy has been wide-ranging Among his latest published works in English are, to name but a few, Mysteries and Conspiracies Detective Stories, Spy Novels and the Making of Modern Societies (Polity Press, 2014), On Critique: A Sociology of Emancipation (Polity Press, 2011) or The New Spirit of Capitalism (with Eve Chiapello, Verso, 2007) Cécile Brousse was a senior official at the French National Institute for Statistics and Economic Studies (INSEE) and a researcher at the Center for Research in Economics and Statistics (CREST) She is currently working at the Directorate for Research, Studies and Statistics (DARES) in the French Ministry of Labour, Labour Relations and Solidarity Isabelle Bruno is associate professor in political science at University of Lille (France) and member of the Lille Centre for European Research on Administration, Politics and Society (CERAPS, CNRS/Lille 2) Among other research interests, she works on benchmarking as a technology of fact-based government and on EU processes of quantification She recently coedited with E Didier and T Vitale a special issue on « Statactivism: State Restructuring, Financial Capitalism and Statistical Mobilizations », published in Partecipazione e Conflitto The Open Journal of Sociopolitical Studies (vol 7, n° 2, 2014) Pierre Concialdi is currently working as an economist at France’s Institute for Social and Economic Research (IRES) Earlier he worked as a deputy director at the Center for Research on Incomes and Costs (CERC) His studies have focused on income inequalities and poverty, social protection and the financing of social security, tax progressivity, low-wage employment and unemployment Roser Cussó is Professor at University of Paris-I She works on the history and sociology of international quantification (statistics produced by intergovernmental and supranational organizations), and more generally on international and transnational technical activity Her interest is specially focused on the impact of expertise on the public debate and on the ix x List of contributors functioning of democracy Her main research areas are education, population, and minorities, as well as related economic questions Alain Desrosières (1940-2013) is the founder of the social history of statistics After graduating from France’s Ecole Polytechnique, he was concerned about social issues and, therefore, chose to study at the National School of Statistics and Economic Administration (ENSAE) He became an administrator at the INSEE in 1965 His commitment to the dissemination of statistical information and his interest in history led him to play a key role in the development of the critical approach to statistics Analyzing in detail their conditions of production, their uses and their impacts on societies, Desrosières published in 1993 The politics of large numbers : a history of statistical reasoning (Harvard University Press, 1998) as a first synthesis of his thought and research Concerned with the effects of neoliberalism on democracies, he dedicated work to the role and place of statistics in new forms of governance, notably marked by the withdrawal of the Welfare State His studies were published in 2008 in the two volumes of The Statistical Argument Desrosières has played a leading role in the evolution of the social sciences of quantification both in France and in the Anglo-Saxon countries Emmanuel Didier is a former student of Alain Desrosières and became a friend even more than a colleague He is a CNRS researcher and at the same times a visiting professor at the UCLA Institute for Society and Genetics where he acts as the deputy director of a joint laboratory called EpiDaPo In addition to writing books on the socio-history of quantification which is his main topic, Emmanuel Didier edited and wrote the introduction to the last book of Alain Desrosières 2014, Prouver et gouverner, un analyse politique de la statistique publique La Découverte, Paris Alain Gély is a public statistician, former INSEE official and CGT activist involved in the National Council on Statistical Information (Conseil national de l'information statistique or Cnis) He takes part in the “comité du label”, committee which issues the certificates of general interest and statistical quality in the public statistical surveys Marion Gilles is a sociologist, research officer in French National Agency for the Improvement of Working Conditions and a Research Associate at the Max Weber Center (University of Lyon) She has defended a thesis at the EHESS on the process of quantification of health at work and its effects within companies, entitled "Counting to weigh in? The quantification of ‘health at work’ : in between institutionalization and conceptual ambiguities" Florence Jany-Catrice is full Professor in economics at the University Lille (France) She conducts her researches at the Clersé (CNRS-UMR8019) She is an honorary member of the Institut universitaire de France (IUF, 2006-2012) Her researches concern mainly the economy of quality in the field of the service economy: quality of labor and employment; quality of output, wealth and their statistical measures She published The New Indicators of Well-Being and Development (Palgrave McMillan) 2006, with J Gadrey; and La performance totale: nouvel esprit du capitalisme? (2012, Presse Universitaire du Septentrion) 190 A Desrosières are incommensurable, thereby calling into question the one-dimensional scale that had dominated the eugenicist human sciences since Francis Galton Confined at first to psychology, this idea was to spread to other sciences after the 1960s in the wake of studies by Jean-Paul Benzécri and Brigitte Cordier-Escoffier Thus two very different ways of inferring and interpreting were adopted by these two sub-branches of psychometrics One was based on the probabilistic inferential statistics used in Fisher’s tests (and which are taught to social science students and provide the key elements of scientific rhetoric), while the other developed a multifactorial form of descriptive statistics derived from Spearman’s work and presaging Benzécri’s correspondence analysis Critiquing commensuration: the role of context in history and sociology The quantification of the human and social sciences has always been subjected to critical analysis However, these critiques cannot be analysed independently of the uses to which quantified statements have been put The five disciplines differ not only in their objects and paradigms but also in their roles in society Thus the critiques may couched in the language of academic discourse and its internal norms: these are the critiques that might be made by professors, along the lines of the comments made by members of PhD examining panels Conversely, they may be linked to the disciplines’ social uses, as when quantification is associated with experts offering advice on policy to decision-makers or used in support of demands and denunciations of unjust situations123 Any investigation of the various critiques of quantification must seek to shed light on the links between these two dimensions – one inward-looking and cognitive, the other worldly, outward-looking and pragmatic Of the various disciplines, economics and psychology seem to be the two most directly concerned by the outward-looking uses to which experts put quantification One of the most frequent criticisms concerns commensuration and the equivalence conventions it implies These conventions have a price – something, namely context, is sacrificed or lost – but there is a gain in intelligibility; things previously invisible are made visible, as the macro-social regularities revealed by Quetelet on the basis of Bernoulli’s urn model showed This notion of loss with a view to a subsequent gain of a different kind is analysed by Thévenot (1986) under the name of an ‘investment in form’ These gains are both cognitive and economic, as the use of the word ‘investment’ suggests This question lies at the heart of many of the criticisms levelled at the quantification of the social and human sciences However, it takes different forms from one discipline to the next It is most evident in history, sociology and psychology Thus comparing these various sets of criticisms is a way of investigating, a contrario, the style of reasoning characteristic of each discipline, above and beyond their internal divisions However, balancing out the losses and gains cannot be separated from analysis of the economic, social and political uses to which quantified statements are put 123 In his study of the ‘scientific field’ (1976), Bourdieu, drawing on Marx’s analyses of capitalist production, divided scientific outputs into the ‘production of producer goods’ (sector I) and the ‘production of consumer goods’ (sector II) The distinction made here between the two types of statements and criticisms of those statements is similar in nature, although it should be noted that quantified statements generally fall within the scope of sector II 15 The quantification of the social science: an historical comparison 191 Critiquing the credibility of sources is one of the cornerstones of historians’ professional culture The long time series they construct by sifting patiently and fastidiously through archives put them in a good position to evaluate the material and cognitive cost of this investment, in contrast to many economists who automatically have recourse to ‘data banks’ developed by others Historians clearly saw what was lost by selecting one element from a far richer archive Furthermore, the sequential history thus constructed enabled them to produce grand narratives that were in tune with the social theories then in vogue The history of quantitative history since the 1950s is marked by this tension, which is encapsulated in the path taken by Franỗois Furet (having been a quantification enthusiast in the 1960s and 70s, he turned away from it in the 1980s) Speaking of studies of the libraries owned by the 18th century bourgeoisie, he remarked: ‘Before, we used to count their books, now we read them’ Thus in 1992, Olivier Marchand and Claude Thélot’s book Deux siècles de travail en France was the object of a lively debate between statisticians and historians The journal Genèses contributed to the debate by publishing various points of view that reflected the tensions arising out of these authors’ investment (Weber et al., 1992) They were hailed as ‘courageous’ by economists with an interest in history On the other hand, historians (even those with a knowledge of economics) were more perplexed They argued that the type of comparisons undertaken in the book disregard the historicity, that is the historical actuality, of the conditions under which ‘data’ from different ages and contexts are produced – and hence, according to this critique, fail to acknowledge that those data are relatively incommensurable from one period to the next This critique has since been the starting point for various research projects, carried out by statisticians in particular, on the social, political and cognitive historicity of the production of statistics124 In this sense, it has extended and broadened the tradition and expertise of professional historians, who have always maintained a critical stance towards their sources Other paradigms emerged during the 1980s The Italian microstoria school, whose principal proponents were Carlo Ginzburg, Maurizio Gribaudi and Giovanni Levi, rejected large-scale quantitative studies in favour of analyses of small units in all their complexity, for which The Cheese and the Worms (Ginzburg, 1980) was the model This decline in enthusiasm for quantification was one of the aspects of the ‘critical turning point’ mooted by the journal Annales in 1989 One way out of this recurrent debate is to ask not: ‘Is it justified to quantify’ but rather: ‘What purpose does it serve?’ In the case of history, the grand narratives to which statistics contribute have in themselves a utility that has to be examined on its own terms before such narratives can be rejected However, history seems to be the least ‘expert’ of the five disciplines The critique of quantification in sociology followed a different course Its origins can be traced back to the debates on freedom and determinism triggered by Quetelet’s work on the macrosocial regularities revealed by statistics (Lottin, 1908) Is man ‘truly free’ if it is known in advance more or less how many people will commit suicide next year? This debate went on throughout the 19th century and the same criticism was levelled at Bourdieu’s ‘objectifying’ 124 Four examples of such historicisation: Brian (1994) on censuses in 18th century France; Desrosières (2003) on ‘family budget’ surveys in the 19th and 20th centuries; Monso and Thévenot (2010) on Insee’s ‘vocational training and qualifications’ surveys, from 1964 to 2003; Amossé and de Péretti (2011) on the relative positions of individuals and households in demographic and social surveys, from the 1950s to the 2000s 192 A Desrosières sociology, which was described as ‘deterministic’ What is sacrificed in the reduction brought about by statistical conventions is no longer a macro-historical context, as in history, but the meaning of individuals’ actions, in a unique context, linked to situations125 An unimaginative form of this debate is the recurrent question of the relative merits of quantitative and qualitative research In the USA, this debate was marked by the criticisms levelled by ethnomethodologists (Cicourel, 1964) against the quantitative sociology of Paul Lazarsfeld (1901-1976) Aaron Cicourel analysed the traps and misunderstandings inherent in questionnaire-based surveys and the difficulties of coding, rendering visible everything that is lost by the mechanics of categorisation These critiques had very different sets of consequences As far as struggles within the field of university sociology were concerned, they constituted an attempt to discredit the efforts of quantitative sociologists in order to obtain recognition and credit; in this case, the quantitativists simply ignored them and did not mention them On the other hand, they also helped them to improve their procedures, their questionnaires and, above all, their ways of interpreting their results, by qualifying and recontextualising them Critiques and their uses: political science, economics and psychology In political science, the main object of criticism was opinion polls In this sense, the critique stemmed from a pioneering and very rich article by Bourdieu entitled ‘L’opinion publique n’existe pas’ (1973) In this article, Bourdieu accuses these surveys of fabricating an improbable entity, namely ‘public opinion’, instead of measuring something that pre-existed its quantification This critique differs from previous ones It opens up the way for questions about what quantification creates, about what the object thus created does126, and not just about its reductiveness, which is a critique of quite different import, heralding as it does the listing at length of a whole series of things said to be ‘socially constructed’ (Hacking, 1999) The non-relativist version of this critique leads to our proposal that the verb to measure should be replaced by the verb to quantify, defined as ‘to express in number what was previously expressed in words’ Among the arguments in his article, Bourdieu raises in particular the question of non-responses and the artefact created by presenting results without taking them into account In economics, the critique of quantification (or rather the reluctance to engage with it) stemmed from the pre-eminence of a hypothetico-deductive epistemology, which was initially literary in nature and was then mathematised from the time of Walras onwards It was exemplified by general equilibrium theory, which was clearly distinct from empirical economics and its accumulation of observations These two approaches did not converge until the 1930s and the development of econometrics, which sought to bring together theoretical economics, empiricism and mathematics Michel Armatte (1995 and 2010) has analysed in detail the tension between the deductive and inductive methods in economics In particular, he notes that astronomy was long marked by the same duality until Laplace and Gauss synthesised the two tendencies in the form of probabilistic mathematical statements which, 125 Jacques Siracusa (2012) made a detailed study of the arguments advanced in these debates in the various schools of sociology in the USA 126 The ritual debate on the ‘effects of opinion polls on voting behaviour’ is a question of precisely this type 15 The quantification of the social science: an historical comparison 193 almost a century and a half later, were to be the basis for the Cowles Commission’s formulation of econometrics This explains why some famous 19th century economists, such as Jean-Baptiste Say, Augustin Cournot and Léon Walras, seemed to spurn statistics or, in the words of the French economist Claude Ménard, ‘resist’ it Their argument was that the statistics available at that time were inadequate and of poor quality: they were not on the same level as the theory and could not be used to explain the complexity of economic interactions127 However, the question was undoubtedly more profound than that It concerns the very status of the emerging discipline of economics, which lay somewhere between two models: history, on the one hand, and the natural sciences, on the other Thus this ‘critique’ of quantification was very specific to economics and to the controversies aroused by the expression of economic theories, particularly the theory of value This very directly poses the question of whether there is a general equivalent, between the labour theory of value, the marginalist theory of value and the notion of a non-substantial, purely relational and social value that cannot be aggregated and is hence not commensurable over a long period, as analysed by André Orléan (2011) The history of certain controversies shows how the role assigned to quantification in economics has evolved In each of these controversies, the inductive and deductive versions were at odds with each other, but in different ways In the Methodenstreit (‘method dispute’), statistics was on the side of the German historicist Gustav Schmoller, who took issue with the Austrian Carl Menger, for whom ‘human motives and social interaction are too complex to lend themselves to statistical analysis Consequently, the starting point for developing the laws of economics must be fundamental knowledge of the individual and of his behaviour’ (Labrousse, 2008) And then in 1949, a famous controversy brought economists from the NBER (Mitchell and Vining), who were analysing business cycles empirically and without any a priori theoretical model, into conflict with those on the Cowles Commission (Koopmans), who were advocating the then new version of econometric modelling as a means of testing theories Koopmans critiqued the NBER method, describing it disparagingly as ‘measurement without theory’ (Vining and Koopman, 1949; Mirowski, 1989) However, the arguments put forward by the two sides were then published in the same journal (Review of Economics and Statistics), which would not happen today Finally, in the year 2000, a controversy arose in France around the interpretation of measures of unemployment Neoclassical econometricians put forward a model which, according to them, could be used to quantify ‘voluntary unemployment’ and ‘inactivity traps’ In this case, the very relevance of this ‘econometric quantification’ was called into question by those whose task it was to measure unemployment on the ground by means of surveys (Coutrot and Exertier 2001; Mirau, 2002) Critiques of the quantified statements produced by economists are closely linked to the uses to which they are put by non-specialists, as is demonstrated by two recent debates One 127 This criticism of the ‘poor quality’ of the sources and their ‘failure to meet the needs of science’ is a recurrent one, as is shown by the success and the many subsequent new editions of the book by the American economist Oskar Morgenstern entitled On the Accuracy of Economic Observations (1963), although in general little consideration has been given to the history and sociology of these sources, in contrast to the situation in other social sciences 194 A Desrosières of them concerns the critique of gross domestic product, which was conceived in the 1950s as part of a much broader and complex edifice, namely that of national accounting equilibria (Fourquet, 1980; Vanoli, 2005) Fifty years later, however, GDP had become an ‘indicator of wealth’, detached from its foundations in the coherent and balanced overall economic picture presented in national accounts It was now required to reflect a country’s social and environmental health Thus the critique was linked to its new status and its new role (Gadrey and Jany-Catrice, 2005; Thiry, 2012) Another debate concerns randomised public policy experiments The question raised here is whether it is possible to ‘generalise’ from experiments carried out on small samples and often at local level, in both the scientific sense (degree of generality of the statements produced) and in terms of public action (transition from a one-off measure to a general systemic action) (Labrousse, 2010; Bardet and Cusso, 2012) The possible debates range from ‘Is it true?’ to ‘Does it work?’ The issue at stake here is the nature of the equivalence conventions underlying the indicators that are supposed to be assessing the effectiveness of the public action being evaluated, by comparing a treated sample with a control sample In the case of psychology, the often virulent critiques are aimed both at the ‘construction of…’ and at commensuration Quantitative psychology seemed to crash headlong into an idealist or spiritualist form of psychology, just as Quetelet’s sociology clashed with the notion of free will An example is to be found in the debates on the intelligence quotient (IQ) What does it measure? Can it evolve or is it an innate biological attribute and thus immutable? Is it dependent on the socio-cultural environment? The so-called ‘measurement of IQ’ is a social practice, which is very widespread in the English-speaking world, (as yet) somewhat less so in France In this case, it is preferable to use the verb ‘to quantify’ rather than ‘to measure’, since it denotes nothing other than this social practice by drawing attention to its conventions and procedures This does not involve prejudging whether there is something as tangible as the height of the Eiffel Tower to be measured However, this quantification of intelligence has had very significant social effects, particularly in the UK, where it was used for a long time in the form of the 11+ examination as the basis for allocating children to different types of school at the end of primary education The critiques are particularly intense in the case of psychiatry Attempts to quantify mental states, whether for clinical purposes, for managerial reasons or for a combination of the two as part of efforts to compile hospital record databases, have been attacked by people who regard them as a double repudiation, firstly of the incommensurability of an individual’s mind or psyche and secondly of the professionalism of healthcare workers, who are being replaced by technocrats and machines The contradiction is particularly pronounced in the debate between psycho-analysis and cognitive behaviour therapies (CBT) The quantitative evaluations carried out by Inserm, The French Institute of Health and Medical Research, are refuted by psychoanalysts on the grounds that the cases are each unique and therefore incommensurable and that the socially constructed nosographic categories in the DSM III catalogue of mental pathologies are artificial (Kirk and Kutchins, 1998) If these critiques of quantification are set alongside those observed in other disciplines, it becomes clear that the only way of going beyond a purely cognitive confrontation is to carry out a political analysis of the social uses to which they have been put 15 The quantification of the social science: an historical comparison 195 Returning the gaze: what the various social sciences have to say on the subject One way of doing this is by adopting an approach best described as ‘returning the gaze’, in which the five disciplines are examined in order to ascertain what they have to say about quantification as a social process and about its effects, not only on the disciplines themselves but also on society more generally: the history of quantification (and not just quantitative history), the sociology of quantification etc Without reviewing what is already an extensive bibliography on the subject, a number of relevant studies can be cited, with the word ‘quantification’ being broadly defined as follows: ‘to move from words to numbers for the purpose of taking action, making decisions or making demands’128, and with corporate accounting also being included (Chiapello and Desrosières, 2006) Three disciplines – history, sociology and political science – have become interested in quantification in recent decades However, the nature of this research varies, less by discipline than by whether or not the formalisms and tools of quantification are seen as falling within its scope In the history of science and technology, a distinction has long been made between socalled internalist history, practised by specialists in the various sciences and concerned with their content, and externalist history, which is concerned with their social and institutional contexts and is practised by historians, sociologists and political scientists Under the influence of the new sociology of science, one line of research has sought to go beyond this distinction Under the general title of ‘Science, technology and society’ (STS), it takes into account and closely links together all the technical and social resources deployed in scientific activities (Callon, 1989) It describes, often at a detailed level, the trajectories and mutual interactions of both scientists and the objects of their research129 Probabilities and statistics have been studied from this perspective, notably by the Bielefeld group referred to above, who researched the ‘probabilistic revolution’ Historians have approached quantification from a number of perspectives, some examples of which – among many others – are cited below On the internal history of mathematical statistics: Benzécri (in French) (1982), Stigler (1986), Hald (1998) and Salsburg (2001) are useful books On the external history: Anderson (1988) on the American census, Touchelay (1993) on Insee in the 1940 and 50s, Beaud and Prévost (1993) on Canadian statistics, Prévost (2009) on fascist Italy and Blum and Mespoulet (2003) on the USSR under Stalin From the perspective combining these two points of view (STS130): MacKenzie (1981) on English eugenicist statistics at the end of the 19th century, Brian (1994) on the relations between scientists and administrators in 18th century France, Armatte (1995) on the origins of 128 Without, however, broaching esoteric social practices such as numerology that endow numbers with an intrinsic power Hervé Le Bras (2000) gives some interesting information on the way in which in the late Middle Ages thinking of this kind on death was gradually transformed into mortality statistics A history of the progress made in quantification in the Middle Ages is provided by Crosby (2003) 129 In order to describe these three approaches, Michel Armatte uses a very telling metaphor, namely that of a football match broadcast on television The internalist point of view would be comparable to a camera following only the movements of the ball without showing the players The externalists would follow the movements of the players but the ball would be invisible Adherents of STS, finally, would show the match as a whole This metaphor has the great merit of suggesting that concepts and formalisms cannot be separated from their uses and trajectories in actual situations 130 Even though some authors not acknowledge this label 196 A Desrosières econometrics, Szreter (1996) on English birth statistics in the 19th and 20th centuries and Didier (2009) on socio-economic polls in the USA in the 1920s and 30s These historical approaches have played a role in some of the attempts to relativise if not condemn quantified statements out of hand The theme of the ‘social construction of …’, effectively analysed by Hacking (1999), is part of a more general tendency that is sometimes associated, rightly or wrongly, with the new sociology of science, or the linguistic turn movement, which emphasised the importance of language in the construction of meaning, or even Derrida’s notion of deconstruction There is a non-relativist, but not naively realist, way of presenting this The historical and sociological approaches to quantified statements show that they arise out of a specific, historic and situated way of shaping and expressing a reality that could be expressed in a number of other ways, whether quantified or not To say that unemployment statistics are socially constructed is not to deny the existence of unemployment Such considerations make it possible to examine statistics from a sociological perspective without ignoring the social uses to which they are put as a basis for constructing arguments, where realism is demanded, and quite rightly The tension between this social demand and the results of the sociology of quantification is a political issue that cannot be discussed while remaining within the academic world of knowledge, even in its most reflexive forms In his search for traces of ‘crystallised social facts’ already available for investigation by the emerging discipline of sociology, Durkheim mentioned two such forms: the law and statistics (Héran, 1984) In emphasising the similarities between the law and statistics, he was drawing attention to the instituting and constructed nature of statistics, even though his main purpose was to find data to substantiate his arguments Ted Porter developed a persuasive sociology of quantification in his book Trust in Numbers (1995) and in numerous articles, notably ‘Making things quantitative’ (1994), in which he adopts a very global perspective in order to highlight the key notions of this sociology Quantification creates a common language, provided that trust can be created and the possible uncertainties and misunderstanding arising out of their construction ignored131 The history of the requirement for objectivity (already worked on by Lorraine Daston, 1992) draws on the changing conceptions of that objectivity to reveal the routes through which the quantification of economic and social relations was developed and, on occasions, imposed from the Middle Ages onwards The alignment and standardisation of procedures turn out to be more important than the accuracy of any particular measure The unification of these procedures is closely linked to that of the institutions of nation states, what the Abbé Sieyès termed in 1789 the adunation (or union) of France, including the establishment of the départements as France’s main administrative districts and the introduction of the metric system and even of the Revolutionary or Republican Calendar (although this failed, since the Gregorian calendar was already unified and, even more importantly, the Christian tradition was too strong) The impersonality and objectivity of quantified statements are congruent with those of the law and of egalitarian democracy that govern modern societies, notably through universal suffrage: one man, one vote Mechanised judgements based on quantification can be used to 131 The controversies and critiques outlined above with regard to the use of quantification in the social sciences arise again with regard to the uses to which it is put by society, although their impact is quite different 15 The quantification of the social science: an historical comparison 197 make decisions in an impersonal way that transcends local contingencies Everything that sustained the critique of the decontextualisation that quantification requires here contributes to maintaining its strength and its social effectiveness, in the same way as universal suffrage can no longer be blamed for separating electors from their social context, as the polling booth does symbolically In the same vein, Porter also puts forward the idea that quantification regimes are the tools of the weak, which is the complete opposite of the widespread belief that statistics are the tools of the powerful, which is not wrong: they can be overturned by dominated groups within society that use quantified arguments to challenge modes of domination based on tradition, implicit attitudes or even eternal nature Demands for equality from the labour movement and, more recently, from feminists have deployed the same statistical arguments Whereas the quantified social sciences use this technique as an instrument of proof, political science employs it in a completely different way, as an instrument of government, in a way comparable to that explored by Ted Porter This strand of research is symbolised by a few pages in Michel Foucault’s book Security, Territory, Population (2007) He suggests here that statistics emerged, around the 18th century, at a time when monarchs were no longer concerned simply with maintaining power but also had to interest themselves in the wellbeing of the population The various forms of the state that have succeeded each other over the past two centuries (the engineer, liberal, welfare, Keynesian and neoliberal states) can be associated with different forms of statistics (Desrosières, 2003 and 2008) The neoliberal state is based in particular on performance evaluation and benchmarking, for which quantified indicators (Salais, 2010; Eyraud 2011) are used This has conferred on quantification a political role that is quite new; it is no longer used solely for the purposes of description or representation but also to incentivise and control, as in the planned economies of the former communist countries These questions have been examined intensively in the UK (Hood, 2002; Miller, 2004), doubtless because this country was one of the first to be confronted, from the 1980s onwards and well before France, with this mode of government by numbers Paradoxically, economics, which is the most quantified of the social sciences, does not appear to have concerned itself with quantification as an activity to the same extent as the previous three disciplines Studies of the privatisation and marketisation of the production of public statistics, or at least the outsourcing of part of that process, were carried out in the UK during the Thatcher period Government statistics, which were once free and then became available only on payment, are now free again since they have been easily accessible on the Internet, as part of the wider trend towards open data, which has made accessible an enormous quantity of unprocessed raw data, thereby marking a clear break with the previous practices of professional statisticians The development of data banks and then of billions of data items, relating to an ever increasing number of human activities, stored in data warehouses has radically changed the economics of quantification, although it is impossible to obtain an overall view What might a psychology of quantification look like? Although this question cannot be examined here, a little hypothesis can be advanced Statistics is often presented as being able to calm tensions and passions by serving as a mediating factor and protective screen between individuals and the real world The sociologist Barbara Laslett (1991) described the 198 A Desrosières interesting case of William Ogburn, a pioneer of American sociology in the 1920s and 30s Now what is surprising is that he was an active advocate not only of quantitative sociology but also of psychoanalysis132, which at that time was growing rapidly in the USA Laslett explains this curious combination of interests by pointing to Ogburn’s very disturbed childhood and his permanent state of anxiety In her view, these two techniques – statistics and exploration of the unconscious – were ways of controlling his anxiety and enabling him to lead a successful professional life This does not of course provide sufficient material to fill the last box of our table comparing the five human and social sciences133 Economics: a world apart? Where does this investigation of the five disciplines leave us? The first three – history, sociology and political science – have several points in common from the point of view of the entries in the four columns Although they are separate disciplines, methods and ideas circulate between them Moreover, each of them is partially split into two tendencies, one quantified, the other not, that often pay no heed to each other For their part, economics and psychology constitute very different worlds One specific characteristic of economics is the recurrent tension between deductive and inductive approaches, of which the latter involves the use of statistics However, less attention has been paid in economics than in the other disciplines to the theoretical problems associated with data construction On the other hand, much energy has been expended in economics on problems of inference, particularly through the development of econometrics This technique then spread to the other disciplines, particularly in Anglo-American sociology As early as the 1930s, moreover, considerable effort in psychology was put into inferential statistics and hypothesis testing University courses in these disciplines show the effects of all this Apprentice historians and sociologists frequently have a good humanist and philosophical education They are trained to critique sources and to deal with categorisation and coding problems, whereas economists are not On the other hand, economists receive advanced training in mathematical economics and econometrics and have a higher level of expertise in this area French economists in the Association franỗaise dộconomie politique (AFEP), whose chair is Andrộ Orlộan, are striving to promote a pluralism of approaches and to open up economics to other disciplines It seems to us that the approach put forward here, although it remains very incomplete, may provide some food for thought for those involved in drawing up AFEP’s plans 132 Sufficiently so to visit Freud in Vienna in 1936 and to suggest to him that psychoanalysis could be quantified, a proposition that Freud vigorously rejected (information conveyed orally by Barbara Laslett) 133 It is true that the row in our table labelled Psychology is very inadequately documented 15 The quantification of the social science: an historical comparison 199 Table 15.1: Quantification viewed by five social sciences QUANTIFYIN G From words to DISCIPLINES numbers Forms of equivalence INFERRING and INTERPRETIN G From the particular to the general Probability and variability Archive work Example: market price lists Intertemporal equivalence Abstract homogeneous time, as opposed to ‘real-life time’ Sequential economic and social history Simiand Labrousse, Braudel Notion of the ‘long run’ Comparison of long time series Tradition of Historicisation critiquing sources of categories, Incommensurabilit data recording y of data from and different periods processing Microhistory procedures, (Ginzburg) uses of data Critical turning and statistical point (1989) institutions SOCIOLOGY Equivalence based on predetermined criteria (quasilegal) or typicality (resemblance) Closed questions in surveys Law of large numbers: Bernoulli’s urn, constant causes Quetelet and the average man Macro regularities, Durkheim Statistics as a tool of revelation Analyses of correspondences Logit regressions Determinism and free will Coding Critique of equivalences Ethnomethodolog y Interaction between interviewee and interviewer Quantitative vs qualitative Linguistic Turn Opposition between explaining and understanding POLITICAL SCIENCE Election results Responses to surveys on party Electoral mapping (Siegfried) Public opinion does not exist Polls HISTORY CRITIQUING Forms of reflexivity and controversies RETURNIN G THE GAZE What the disciplines have to say Sociology of statistics as a coordinating tool Tool of the weak (Porter) Impersonality and objectivity Quantification shapes the world, creates a common language, things that hold, comparability spaces Neoliberalism Government 200 A Desrosières preferences or voting intentions ECONOMICS Money as general equivalent Corporate accounting National accounting Ecological footprint CO2 emissions PSYCHOLOG Y Test results Measurement of intelligence Subjective probabilities: Bayes How the brain functions (Kahneman and Tversky) Cognitive psychology Straw polls Gallup polls Electoral sociology: geographic or sociodemographic analytical criteria Prices indexes Graphics Cycles and barometers Econometrics, without (Lenoir, Tinbergen) or with (Haavelmo, Cowles Commission) probabilities VAR models 9Sims), Randomised experiments Two forms of psychology (Cronbach, 1957) General psychology: psyche of human being in general Inference tests (Gigerenzer) Differential psychology: factor analysis by psychologists (Spearman, Thurstone) decontextualise The problem of non-responses Critique of survey democracy (Bourdieu) by numbers Performance evaluation by quantified indicators League tables and benchmarking Marx, accounting in labour time Earlier critiques: Say, Cournot, Walras Methodenstreit Vining-Koopmans controversy Economic statistics vs econometrics Critique of GDP (Gadrey) Privatisation and marketisation of the production of statistics and of market research Activities of data warehouses charged for or free Critique of IQ Critique of the quantification of mental illnesses Evaluation of treatments by meta-analyses: psychoanalysis vs CBT Psychology of statistics, reassuring and calming Ogburn advocated both quantitative sociology and psychoanalysi s 201 15 The quantification of the social science: an historical comparison References Anderson, M.J (1988) The American Census A Social History, New Haven/London : Yale University Press Armatte, M., (1995) Histoire du modèle linéaire Formes et usages en statistique et économétrie jusqu’en 1945 Thèse de Doctorat, EHESS, Paris Armatte, M (2010) La science économique comme ingénierie Quantification et modélisation Paris : Presse des Mines Bardet, F & Cusso, R (2012) Les essais randomisés contrôlés, révolution des politiques du développement ? Une évaluation par la Banque Mondiale de lempowerment au Bangladesh Revue franỗaise de socio-ộconomie, 10, 175-198 Beaud, J.P & Prévost J.G (1993) La structuration de l’appareil statistique canadien, 19121921 The Canadian Historical Review, 75(3), 395-413 Benzecri, J.P (1982), Histoire et préhistoire de l’analyse des données Paris : Dunod Blum, A & Mespoulet, M (2003) L’anarchie bureaucratique Statistique et pouvoir sous Staline Paris : La Découverte Boumans, M (1992) A case of limited physic transfer Jan Tinbergen’s resources for reshaping economics Amsterdam : Tinbergen Institute Research Series Bourdieu, P., (1976) Le champ scientifique Actes de la recherche en sciences sociales, 2(23), 88-104 Bourdieu, P (1979) La Distinction Critique sociale du jugement Paris : Minuit Brian, E (1994) La mesure de l’Etat Administrateurs et géomètres au Albin Michel XVIII e siècle Paris : Callon, M (Ed.) (1989) La science et ses réseaux Genèse et circulation des faits scientifiques Paris : La Découverte Chiapello, E & Desrosières, A (2006) La quantification de l’économie et la recherche en sciences sociales Paradoxes, contradictions et omissions Le cas exemplaire de la positive accounting theory In Eymard-Duvernay, F (Ed.) L’Economie des conventions Méthodes et résultats, tome 1, Débats Paris : La Découverte, 297-310 Cicourel, A (1964) Methods and Measurement in Sociology New York : The Free Press of Glencoe Coutrot, T & Exertier, G., (2001) La loi des grands noms Quand le « non-emploi » efface le chômage L’Année de la Régulation, 5, 256-279 Cronbach, L.J (1957) The two Disciplines of Scientific Psychology American Psychologist, 12, 671-684 Daston, L., (1992) Objectivity and the escape from Perspective Social Studies of Science, 22(4), 597-618 202 A Desrosières Desrosières, A (1998) The Politics of Large Numbers A History of Statistical Reasoning Cambridge : Harvard University Press Desrosières, A (2003) Managing the Economy, the state, the market and statistics In Porter T & Ross D (Eds) The Cambridge History of Science, vol.7, Modern Social and Behavioral Sciences, Cambridge : Cambridge University Press Desrosières, A (2008) Gouverner par les nombres Paris : Presses des Mines Desrosières, A & Thévenot, L (1998) Les catégories socioprofessionnelles Paris : La Découverte Didier, E (2009) En quoi consiste l’Amérique ? Les statistiques, Le New Deal et la démocratie Paris : La Découverte Eyraud, C (2011) Une comptabilité d’entreprise pour les Etats ? Un dispositif de quantification économique soumis l’analyse sociologique Le cas franỗais Sociologie et Sociộtộs, 34(2), 91-116 Foucault, M (2007) Security, Territory, Population Lectures at the Collège de France (1977 1978) London : Palgrave McMillan Fourquet, F (1980) Les comptes de la puissance Histoire de la comptabilité nationale et du Plan Paris : Recherches Gadrey, J & Jany-Catrice, F (2005) Les nouveaux indicateurs de richesse Paris : La Découverte Gigerenzer, G & Murray, D.J (1987) Cognition and Intuitive Statistics Hillsdale : Erlbaum Gigerenzer, G et al (1989) The Empire of Chance How Probability changed Science and Everyday Life Cambridge : Cambridge University Press Ginzburg, (1980) Le fromage et les vers L’univers d’un meunier frioulan au Paris : Aubier XVI e siècle, Hacking, I (1990) The Taming of Chance, Cambridge : Cambridge University Press Hacking, I (1999) The Social Construction of What ? Cambridge : Harvard University Press Hald, A (1998) A History of Mathematics Statistics, from 1750 to 1930, New York : Wiley Halbwachs, M., (2013) La Classe ouvrière et les niveaux de vie, Alcan, Paris Hautcoeur, P.-C (2002) Entre micro et macro, quelle place pour le quantitatif en histoire économique ? Historiens et géographes, 380, 139-146 Héran, F (1984) L’assise statistique de la sociologie Economie et Statistique, 168, 23-35 Hood, C (2002) Controls, bargains and cheating, the politics of public-service reform Journal of Public Administration Research and Theory, 12(3), 309-332 Kirk, S & Kutchins, H (1998) Aimez-vous le DSM ? Le triomphe de la psychiatrie américaine, Le Plessis-Robinson : Les Empêcheurs de penser en rond/ Institut Synthelabo 15 The quantification of the social science: an historical comparison 203 Le Bras, H., Todd, E (2012) L’Invention de la France Atlas anthropologique et politique, Gallimard, Paris Labrousse, A (2008) Le Méthodenstreit, une affaire close ? In Alcouffe A., Diebold C (Eds) La Pensée économique de langue allemande Paris : Economica Labrousse, A., (2010) Nouvelle économie du développement et essais clinique randomisés, une mise en perspective d’un outil de preuve et de gouvernement Revue de la régulation, Laslett, B (1991) Biography as historical sociology The case of William Fielding Ogburn Theory and Society, 20, 511-538 Lemercier, C & Zalc, C (2008) Méthodes quantitatives pour l’historien, Paris : La Découverte Lenoir, M (1913) Etudes sur la formation et le mouvement des prix, Paris : Giard et Brière Lottin, J (1908) La statistique morale et le déterminisme Journal de la Société statisique de Paris, 49(10), 317-342 MacKenzie, D (1981) Statistics in Britain, 1865-1930 In The Social Construction of Scientific Knowledge, Edimburg : Edinburgh University Press Martin, O (1997) La Mesure de l’esprit Origines et développement de la psychométrie 1900-1950 Paris : L’Harmattan Miller, P (2004) Gouverning by numbers, why calculative practices matter In Amin A & Thrift N (Eds), The Blackwell Cultural Economy Reader (pp 179-190) Malden : Blackwell Mirau, C (2002) L’affaire ‘Laroque Salanié’, une controverse avortée en matière d’expertise éconmique et sociale Genèses, 49, 110-122 Mirowski, P (1989) The measurement without theory controversy Economies et Sociétés, série Oeconomia, 23, 109-131 Morgan, M.S (1990) The History of Econometric Ideas, Cambridge : University Press, Cambridge Morgenstern, O (1944) On the Accuracy of Economic Observations, Princeton : Princeton University Press Orléan, A., (2011) L’empire de la valeur Refonder l’économie Paris : Seuil Porter, T., (1986) The Rise of Statistical Thinking, 1820-1900 Princeton : Princeton University Press, Prévost, J.G (2009) A total Science Statistics in Liberal and Facist Italy Montreal/Kingston : McGill Queen’s University Press Siegfried, A (1913) Tableau politique de la France de l’Ouest sous la IIIe République, Paris : Armand Colin Salais, R (2010) Usages et mésusages de l’argument statistique, le pilotage des politiques publiques par la performance Revue franỗaise des affaires sociales, 1-2, 129-147 204 A Desrosières Salsburg, D (2001) The Lady Tasting Tea How Statistics Revolutionized Science in the Twentieth Century, New-York : Freeman Spearman, C (1904) General Intelligence objectively determined and measured American Journal of Psychology, 15, 201-293 Stigler, S (1986) The History of Statistics The Measurement of Uncertainty before 1900 Cambridge : Harvard University Press Szreter, S (1996) Fertility, Class and Gender in Britain 1860-1940, Cambridge : Cambridge University Press Thévenot, L (1986) Rules and Implements, investments and forms Social Science Information, 23(1), 1-45 Thiry, G (2012) Au-delà du PIB, un tournant historique Enjeux méthodologiques, théoriqus et épistémologiques de la quantification Thèse de doctorat en sciences économiques et gestion, Louvain Touchelay, B (1993) L’Insee des origines 1961, évolution et relation avec la réalité économique, politique et sociale Thèse de doctorat, Université Paris 12, Paris Vanoli, A ( 2005) A History of National Accounting Amsterdam : IOS Press Vining, R & Koopman, T.C (1949) Methodological issues in quantitative economics Review of Economics and Statistics, 31, 77-94 Weber, F et al (1992) Histoire et statistique Questions sur l’anachronisme des séries longues Genèses, 9, 90-119 ... The Social Sciences of Quantification, Logic, Argumentation & Reasoning 13, DOI 10.1007/978-3-319-44000-2_1 I Bruno et al The development of the social sciences of quantification The social sciences. .. dissolution of the evaluation of public policies in the measurement of the performance of public services; and a dissolution of the collective performance, and of the social progress, in the sum of individual... Introduction The Social Sciences of Quantification in France: An Overview Outline of the book The book opens with a summary of Alain Desrosières' contributions to the sociology of quantification The chapters