Tải bản đầy đủ - 0trang
The history of quantification: the link between statistics and policy
5 From statistics to international quantification ...
when the World Bank’s policies began to emphasise the privatisation of public services
(structural adjustment, role of the market), the demography was focused on measuring the
cost of children. The Bank was engaged in microeconomics, on the one hand, and in analysis
of demographic behaviour, on the other. Fertility became an object of cost-benefit analyses.
Working in UNESCO’S statistical department while I was researching my PhD, I
observed that the history of education data had followed the same logic. The two IGOs’
statistics had evolved along similar lines and had paved the way for policies that were also
similar. Nevertheless, the mechanisms of change were different. The World Bank had been
through a radical restructuring at the beginning of the 1980s, with changes to both the
organisational structure and the personnel. The restructuring at UNESCO had been less
thoroughgoing, with the main focus being on a wide-ranging critique of its data.
The two restructuring processes constituted a ‘system’, but how could I move forward
in my efforts to demonstrate that this was so? Discussion of a paper by Alain Desrosières,
“Décrire l’État ou explorer la société: les deux sources de la statistique publique/Describing
the state or exploring society: the two sources of public statistics” (Desrosières 2005) during
the ACI seminar on statistics and the evaluation of public policies at INSEE (February 2005)
offered me some valuable new avenues to pursue. In the paper, Desrosières advanced the
following idea: “Market, incentive, rankings: this triad characterises the space within which
public statistics were in part redeployed from the 1980s onwards”. He put forward a reference
framework for the various stages of the quantification process based on the linking of three
elements: “1) the way of conceptualising society; 2) the modalities of action in that society; 3)
the modes of description, particularly statistical” -see also Desrosières (2003). He then
applied these elements to the history of the state: the engineer state, the planner state, the
liberal state, etc. Could these dimensions be transposed to IGOs? What were the factors
driving change? In the case of the World Bank and UNESCO, could the shift from
Keynesianism to neo-liberalism ‘simply’ be an effect of the changes that had taken place in
member states? Or conversely, was it the IGOs that were coordinating states and their
policies? Here was a new avenue to be explored, raising new questions linked to political
science and international relations but going beyond them to address the question of the part
played by quantification in these changes.
The idea of moving from a collection of studies that were beginning to be interlinked –
World Bank, UNESCO, European Union (EU)26– to the construction of a field of enquiry
with its own specific questions and hypotheses and encompassing the history and sociology of
quantification in IGOs and not just statistics developed out of these initial fruitful exchanges
with Alain Desrosières. I then immersed myself in his writings in order to explore his ideas
more fully and holistically.
In the context of my post-doc project on the EU, see for instance Cussó (2004).
3. The construction of a theory and the birth of a field of research
The book The Politics of Large Numbers, published in 1993, both presents an historical
narrative and lays down solid theoretical foundations. Over the following 20 years, Alain
Desrosières developed this ontological contribution in his writings as he set out various
arguments and original perceptions on a range of subjects. Viewed in their entirety, his
writings constitute an ‘historical sociology of quantification’, as he was to call it in a paper
published in 2008 (Desrosières 2008).
The debates initiated by his writings are full of diversity and complexity. An initial
example is the debate around the notion of the objectivation of social forms on the basis of
equivalence conventions. Objectivation as a hardening of ‘reality’ replaces the idea of
objectivity, which is too rigid. It is a ‘boundary concept’ that makes it possible to transcend
the debate between constructivism and realism, especially but not exclusively in the area of
quantification. In a 1991 paper, Desrosières tackles the controversy between ‘constructed’ and
‘real’, while at the same time debating Hacking and Latour: “Indeed, by deciding to take any
social as at the same time constructed and real, one discovers a way of at a stroke
transcending the two apparently opposed positions constituting positivist scientism and
denunciatory relativism” (Desrosières 1991: 196).
A second example is Desrosières’ combining of the ‘descriptive’ and ‘prescriptive’
functions of statistics, which enabled him to link together the cognitive and practical aspects
of statistics. Statistics constitute a “cognitive space of equivalence constructed for practical
purposes, in order to describe, manage, or transform human societies” (Desrosières 1998a:
17). However, he was not content simply to analyse the functionality of that space, however
complex and interconnected it may be. There was no single blueprint but rather a slow,
historical process of construction, with aims that were very diverse and academic approaches,
statistics schools and government actions that differed from country to country. From this
point of view, Desrosières was developing a wide-ranging approach. His commitment to
multi-disciplinarity is reflected in all his writings, more than he actually asserts himself. He
mixes sociology, history, economics, political science and statistics, without abandoning the
realities of numbers and of society. This is the approach of an epistemological constructor,
one who invents questions and stimulates debate.
In 2006, after two years of a post-doc appointment at the Free University in Brussels,
where I worked under Corinne Gobin, and having moved to the University of Paris 8, I set
myself the goal of writing a thesis on quantification in IGOs for the qualification to supervise
research (2012). The underlying idea was that there is not one single abstract process of
quantification from which different cases derive, but only situated processes of quantification:
those taking place in states, IGOs, towns and cities, private bodies, etc. Although the
quantification carried out by governments is the one most closely studied, this by no means
exhausts all the theoretical and empirical possibilities for investigating quantification as a
phenomenon that is simultaneously political, social, scientific and institutional.
5 From statistics to international quantification ...
4. Statistics and institutions: a programmatic link
Although he was concerned mainly with quantification undertaken by the state, Alain
Desrosières was also interested in the new international and European processes of
quantification, particularly from the point of view of their impact on state statistical institutes
(Desrosières 2003). He investigated, among other things, the shift from the harmonisation of
outputs (governments classifying their data in accordance with international standards) to the
harmonisation of methods (governments producing internationally standardised statistics at
source). What other issues does the history of international harmonisation raise? Is the trend
towards the standardisation of methods a new one?
In 2008, I embarked on an analysis of the archives of the League of Nations, the first
political IGO (1919), which revealed that a veritable quantification programme had existed
from the institution’s very early days. I worked on several sections: economic and financial,
mandates, minorities, health, etc. The link between description and prescription in the
production of statistics could indeed be observed, particularly in the economic and financial
section, which was responsible for the production of norms, common standardisation and
comparable data (Cussó, 2012a), and in the supervision of the mandate system, where the
principal concern was an attempt to establish measurable accountability (Cussó forthcoming
At first sight, the approach adopted by the League of Nations was very similar to that
adopted by states, in the sense that it wanted to have its own policies and therefore produced
its own statistics, as states did. Closer examination, however, reveals that IGOs have four
major institutional and socio-political specificities that serve to strengthen the unique and
fundamental role played by their statistical output (harmonisation and comparison of member
states’ data) while at the same time substantiating the ‘non-political’ nature of the resultant
Firstly, the councils and assemblies on which governmental representatives sat did not
reflect political pluralism. They were not organised into majority and opposition, as they are
in democratic systems. This is relevant to the EU, because its parliament does not elect a
government and, in particular, because it makes decisions only in consultation with the other
European institutions, the Council and the Commission. Thus there are constants in the
institutional forms that IGOs take, both in their components and in their way of working.
Secondly, the fact that they do not operate democratically strengthens the relative autonomy
of IGOs’ secretariats. For example, if a tariff nomenclature is drawn up and approved in
principle by the intergovernmental body, the secretariat still retains powers of interpretation
and implementation. Thirdly, and this is an important point, governments operating at a
distance from their respective parliaments may find it convenient to go through IGOs in order
have measures adopted that will subsequently be incorporated into member states legislation,
leaving the opposition facing a fait accompli, at least in part. Thus governments are not the
‘victims’ of IGOs; rather they interact with them, even though they may sometimes have to
accept international decisions that in their view are less desirable. Finally, the transnational
element makes its impact felt more directly in IGOs than in member states, with the support
of the intergovernmental element (in the form of agreements) and, in the absence of an
opposition, without challenge.
These specificities, and in particular the absence of pluralism, provide the basis for an
original form of linkage between statistics and prescription, which might be described as
‘ethical pragmatism’ or a direct link between the technical and the ethical. The process of
quantification is linked to consensual ethical objectives, such as ‘development’, that are
located beyond ‘politics’. For example, measurement of pupil numbers is directly linked to
the right to education, while birth control is intended to facilitate access to better child health.
This connection does not preclude intermediate choices in favour of formal schooling or
health services for which patients have to pay, but they are omitted.
To return to the beginning of this section, the harmonisation of methods (by Eurostat,
for example) is only one outcome of a longstanding process of harmonisation, as Desrosières
describes with great subtlety. Another element of this same process is a powerful and equally
longstanding programme that is both pragmatic and ethical, that is constantly being renewed
and is closely linked to the institutional forms and organisation of IGOs.
5. The actors and professionals involved in statistics
Adopting a broad sociological perspective, Desrosières took into account the role of
statisticians and other actors by means of biographical analysis and occupational sociology.
My experience at UNESCO and my research on the specificity of IGOs logically led me to
compare international statisticians with those in the government departments whom he had
studied (Desrosières 1998b).
The left-right division is less evident within IGOs than in states, which gives a
particular meaning to the notion of ‘statistical independence’. The direct connection between
knowledge and ethics, already alluded to above, gives officials of international organisations
the impression they are acting for the public good and that their missions are universal ones
(Bardet and Cussó 2012). In this context, being in the service of governments is not a problem
– what would be a problem would be to serve a single government! Their main concern is to
adopt the IGO’s core mission. It is in order to defend their interpretation of that mission that
they invoke their independence, particularly from the middle layers of the organisation’s
When UNESCO’s statistical services were being reformed and the nature of the
statistical undertaking changed (with effectiveness being lauded rather than rights), the
attempts at renewal were opposed by some statisticians, who maintained that they better
represented the institution’s values, compared to the new officials put in charge. This was
reflected in personnel changes and the introduction of new recruitment criteria. I have already
mentioned the statistical department’s move to Montreal. One of the consequences was that
the notion of the international civil servant’s independence was redefined: the statistical
endeavour had to distance itself from the past (personnel changes) and from
intergovernmental control (distance from the Paris headquarters) in order to facilitate the task
of implementing the political aims of the reform, such as the introduction of measures of the
effectiveness of state policies with a view to ranking countries.
5 From statistics to international quantification ...
6. The history and sociology of quantification in IGOs
At the beginning of 2012, I asked Alain Desrosières if he would be so kind as to be a
member of the examining panel for my accreditation to supervise research (habilitation à
diriger des recherches) (Cussó 2012b)27. In his usual thoughtful way, he had kept a file with
all the articles I had sent him as my research progressed, but this was the first time I was
formally presenting quantification in IGOs as a coherent field of research, open to
interpretation and testing by original hypotheses. This was the first time he had been a
member of the examining panel for an habilitation, although he had examined many PhDs. I
think this was an indication that quantification had become an established area of research.
My habilitation thesis presents the IGOs investigated (League of Nations, World Bank,
UNESCO, UN, EU), the quantification processes analysed (population, education, economics,
minorities) and their histories. The hypotheses advanced are informed, quite logically, by the
work of other researchers, including that of Alain Desrosières, in particular, and the
publications already cited above, as well as many others. Initially, it seemed to me important
to highlight the linkage between the history and sociology of quantification in IGOs, as
Desrosières had done in the case of the state, since in the international sphere researchers
often focus on highly specific, contemporary topics that produce relatively limited results. For
example, countries emulating each other on the basis of statistics are often regarded as an
innovation in international governance, whereas historical analysis reveals that this is a longestablished practice and that the secretariats of international organisations and member states’
governments jointly contributed to its development. From the same viewpoint, I sought to
avoid the North-South divide, which prevents analyses of statistical changes in Northern
countries from benefiting from those of statistical undertakings in the South, which often
foreshadow those in the North. Like Desrosières, who developed country typologies that
shared a common analytical framework, I observed that the differences between IGOs were
not decisive. Finally, a number of hypotheses had to be advanced in pursuit of my initial
objective, namely to provide evidence to substantiate the view that IGOs’ statistical processes
are specific in nature and can be treated as a completely separate field: (i) IGOs have more
power than the congresses and international statistical institutes of the 19th century; (ii) they
manage to produce statistics and use them in order to ‘govern’; (iii) IGOs and their statistical
processes emerged at the same time and evolved in stages; (iv) because of their institutional
specificity, quantification in IGOs is not simply a tool of government but also the basis for
‘ethical pragmatism’; (v) as pluralism has gradually been eroded, the functioning of
international organisations has tended to converge with that of states.
Let us examine these points.
(i) The statistics produced by IGOs mark a turning point in the longer history of
attempts to compare countries. In the 17th century, William Petty, for example, compared the
populations of London and Paris: “these two cities do not differ […] by a 20th” (Petty
1905: 272). However, the analyses produced by adherents of political arithmetic lacked
The other examiners were Alain Chenu (proposer), Ted Porter, Martine Mespoulet, Hervé Le Bras, Corinne
Gobin and Olivier Martin.
a systematic dimension. This was remedied in the later 17th and 18th centuries with the
development of “the study of the forces of the state” (Staatskunde, or civics). These studies
are monographs on different countries, produced with comparisons in mind. They are
methodical and use common analytical frameworks, but are seldom quantified. In fact, while
the two approaches give the impression that comparisons can improve the government of
England or Germany, they do not make comparison a tool that can be used to produce
normative analyses for all countries.
When Adolphe Quetelet founded the International Statistical Congresses in 1853, he
was clearly seeking to compare countries in order to show governments what were the best
policies: “Statistics, conceived in a spirit of unity and founded on stable bases appropriate to
all countries, will assuredly […] extend their benefits to all countries and shed new light on
the real interests of governments”28. However, the congresses and the International Statistical
Institute (ISI), established in 1885, attempted to move towards their objectives – development
of data harmonisation, lists of topics to be compared, methodologies for conducting
comparisons – simply by setting up transnational networks made up largely of members of
state’s statistical institutions and departments. The debates that took place were relatively
independent of governments, but this approach very quickly proved to be inadequate and
ultimately the congresses and ISI produced little in the way of harmonised statistics (Cussó
2010)29. In fact, unlike the League of Nations, these bodies had no real political power. As
already noted, the League of Nations had an official intergovernmental section whose
decisions were binding on the executives of member states. It had its own assignments, based
on the fact that states’ governments had devolved part of their sovereignty to it (this was the
novel feature of the charters, in which certain policies were delegated). The League of
Nations had a permanent secretariat charged with carrying out the responsibilities that formed
the basis of the international or internationalist section. In addition to these three aspects, the
League of Nations covered transnational activity, which extended to voluntary associations
and the world of business, from which were drawn its experts, representatives of civil society
and interest groups.
(ii) Thus the innovative feature of IGOs was the linkage between the transnational
(inherited from the congresses), the intergovernmental (political legitimacy) and the
international (the secretariat and its missions): member states both contributed and submitted
themselves to the new quantification. This is why IGOs immediately put in place their own
statistical programmes, laying the foundations for a real process of quantification at
international level and positing comparison both as a means of producing norms and a lever
for action (or government).
Supervision of the mandate system is one example of this. An attempt was made to
establish a certain degree of rivalry or emulation among the mandatory powers (their
comparative ‘evaluation’) when it was judged that mortality rates among the ‘indigenous’
Compte-rendu du Congrès international de statistique de 1853, p.19.
In 2009, Alain Desrosières asked me to write a paper for the 150th anniversary of the Journal de la Société de
Statistique de Paris (JSSP). In that paper, I test the influence of the congresses and the ISI on the journal and
investigate the differences between those bodies and IGOs.
5 From statistics to international quantification ...
people should decline year on year. The powers produced the data and accepted supervision
by the League of Nations; however, they also made the figures work to their advantage (i.e.
show their successes). In the case of the economic and financial section, the international
statistics supported free trade policies. The balance of trade (openness to imports), (the
absence of) customs duties etc. were indicators – plain for all to see – of the extent to which
that freedom was being honoured. In both cases, data harmonisation was fundamental. The
countries concerned filed them on the basis of new classifications, or at the very least under
particular headings that gradually introduced new methods.
(iii) The opportunity to identify the various stages in the process of international
quantification confirms the existence of this linkage. Furthermore, for the three major stages
in the history of IGOs (free trade, 1919-1940; development/Keynesianism, 1945-1975;
deregulation/globalisation, 1975-to the present), the inextricable link between norms and
statistics is evident with potential for new policies.
(iv) This potential stems from the fact that quantification in the IGOs required, as we
have seen, the establishment of a new pragmatic ‘regime’. Freed, by definition, from the
constraints of pluralism (comparison of different political projects), IGOs introduced a farreaching change that is not entirely covered by the term ‘governance’. Quantified
comparisons, the means whereby they functioned, also served to underpin them through the
construction of a power that was both non-alternating and intergovernmental (legitimate).
Implementation of the OECD’s Programme for international student assessment
(PISA) from the year 2000 onwards is an example of this regime, which has been reflected in
the development of classifications and good practices that derive their technical and ethical
legitimacy from the optimal use of economic resources. However, this exercise in
international quantification can be fully understood only when seen in the context of a long
history. PISA has similarities with the instruments used in the mandate system while at the
same displaying certain differences: with its micro-economic calculations and cost-benefit
analyses, it has moved some distance from the notion of public policy.
(v) Our final hypothesis is that national governments do not appear passive, whether in
the mandate system or with regard to PISA. However, their position has changed and has
perhaps become more difficult with the current statistical comparisons, which are increasingly
intrusive. The measures adopted by the EU and Eurostat, particularly since the 1992
Maastricht Treaty, have gradually changed the relationships between the actors involved in
quantification. The links between indicators and reforms have become closer, governments
have on occasions been out of their depth while at the same time taking part in decisionmaking and national statistical institutes have lost some of their autonomy. It is against the
background of this change that we can speak of ‘contagion’ between international
organisations and states. There is increasingly little space for pluralism in our political
societies. This contributes to a better understanding of the rise of the new public management,
which aims to establish effectiveness as a driving force and not simply as a means of
criticising the opposition.
To conclude, two remarks about my luck in having been able to engage in dialogue with
Alain Desrosières. The first is the opportunity it gave me to get to know and collaborate with
other researchers involved in the networks he organised. I’m thinking particularly of Ted
Porter, Béatrice Touchelay, Martine Mespoulet, Michel Armatte and Fabrice Bardet – this is
not an exhaustive list. The second point concerns the demands his work have imposed,
making any attempt to contribute to the ‘historical sociology of quantification’ a long and
complex exercise. It cannot be reduced to a sociology of statistics, to their historical and
political dimension, to a methodological analysis of data, to case studies or to the inclusion in
analyses of institutions and actors. It is all of this and much more. A potent programme of
research has been brought forth, one that continues and will continue to nurture current and
future research, even though its scope is not yet fully appreciated in the textbooks.
Bardet, F. & Cussó, R. (2012). Les essais randomisés contrôlés, révolution des politiques de
développement? Une évaluation par la Banque mondiale de l’empowerment au Bangladesh.
Revue Franỗaise de Socio-Economie, 10, 175-198.
Cussú, R. (2001). La dộmographie dans le modèle de développement de la Banque mondiale.
Doctoral thesis, Paris: EHESS.
Cussó, R. (2004). La méthode ouverte de coopération en Europe: des statistiques pour une
nouvelle politique d’éducation. RAPPE seminar on National and international comparisons
of education policies 24-25 May 2004, Paris.
Cussó, R. (2006). Restructuring UNESCO’s statistical services: The ‘sad story’ of
UNESCO’s education statistics four years later. International Journal of Educational
Development, 26(5), 532-544. http://dx.doi.org/10.1016/j.ijedudev.2006.01.001
Cussó, R. (2010). La quantification internationale à la lumière de la SSP et des Congrès
internationaux de statistique: continuités et ruptures. Journ@l Électronique d’Histoire des
Cussó, R. (2012a). L’activité statistique de l’Organisation économique et financière de la
Société des Nations: un nouveau lien entre pouvoir et quantification. Histoire & Mesure,
Cussó, R. (2012b). Comparer pour mieux régner: histoire et sociologie de la quantification
internationale, Paris: Institut d’Études Politiques de Paris (IEP).
Cussó, R. (forthcoming in 2016). Aux origines de l’évaluation statistique internationale: le
système mandataire de la Société des Nations. Experts et expertises dans les mandats de la
SdN : figures, champs et outils, Conference 26-27 March 2015, Paris.
5 From statistics to international quantification ...
Desrosières, A. & Thévenot, L. (1988). Les Catégories socio-professionnelles. Paris: La
Desrosières, A. (1991). How to Make Things Which Hold Together: Social Science, Statistics
and the State. In P. Wagner, B. Wittrock, R. P. Whitley (Eds.), Discourses on Society: The
Shaping of the Social Science Disciplines (pp. 195-218). Netherlands: Kluwer Academic
Desrosières, A. (1998a). The Politics of Large Numbers. Cambridge, Mass.: Harvard
Desrosières, A. (1998b). L’administrateur et le savant. Les métamorphoses du métier de
statisticien. Courrier des statistiques, 87-88, 71-80.
Desrosières, A. (2003). Comment fabriquer un espace de commune mesure: harmonisation
des statistiques et réalisme de leurs usages. In M. Lallement, J. Spurk (Eds.), Stratégies de la
comparaison internationale (pp. 151-166). Paris: CNRS Editions.
Desrosières, A. (2003). Historiciser l’action publique: l’Etat, le marché et les statistiques. In
P. Laborier, Danny Trom (Eds.), Historicités de l’action publique (pp. 207-221). Paris: PUF.
Desrosières, A. (2005). Décrire l’État ou explorer la société: les deux sources de la statistique
publique/Describing the state or exploring society: the two sources of public statistics.
Genèses, 1(58), 4-27.
Desrosières, A. (2008). Pour une sociologie historique de la quantification. Paris: Presses des
Desrosières, A. (2012). Sur l’histoire de la méthodologie statistique: mesurer ou instituer?
Deux traditions de recherche encore largement sộparộes. Congrốs de la Sociộtộ franỗaise de
statistique, Brussels, 23 May 2012.
Petty, W. (1905). Arithmétique politique in Les œuvres économiques de W. Petty.
Tome I, Paris: Giard et Brière.
Chapter 6. Learning from the history of the probabilistic revolution: the
French school of Alain Desrosières
In this chapter, Fabrice Bardet, a long-life student of Alain Desrosières tells the
amazing academic dynamic that Alain Desrosières created around him. From the 1990s to his
death in 2013, he was, without any doubt, the French scholar situated in the center of the
sociological field dedicated to quantification’s processes. Fabrice first comes back on the
history of his master, a story that he had been told several times, by his side. Then he tells the
story of the influence of his master on several generations of scholars, in France and abroad.
He insists, among other things, on the importance that Alain Desrosières gave, from his main
book published in 1993, The Politics of Numbers, to his most recent publications, to the
influence of the writing of The Probabilistic Revolution, in the 1980s, in an international
context, set up in Bielefeld, Germany. From this perspective, he explains how, in his own
view, this focus made by Desrosières participated in creating a French school for sociology of
Theodore Porter, one of the main historians of the "Probabilistic Revolution", maintains
that the impact of this history on the social sciences was stronger in France than anywhere
else. Written after a seminar organized in the early 1980s at Bielefeld, in Germany, by the
Zentrum für interdisziplinäre Forschung (Zif), this history was the fruit of a long
interdisciplinary encounter envisioned by physicist and philosopher of science Lorenz Krüger
and sustained by the research dynamic stemming from Thomas Kuhn's work on the structure
of scientific revolutions. The revolution, which played out primarily between 1800 and 1930,
placed the statistical sciences at the heart of all the activities of all the intellectual disciplines,
throughout the world. The widespread use in industrial societies of "quality" tests is the
product and symbol of this revolution. In France, Alain Desrosières, who had not been part of
the Bielefeld group, was enthused by this masterpiece in the contemporary history of science
immediately on its publication (Krüger et al. 1987, Krüger et al. 1987). He made it a basis for
developing his plan for a historical sociology of quantification. He led a number of peers and
students in this endeavour and was unquestionably the individual largely responsible for
bringing this French school of the sociology of quantification into being. Within this school,
the hypothesis of an accounting counter-revolution, dialectically linked to the probabilistic
revolution, took shape as a means of shedding light on contemporary modes of government30.
1. The meeting with sociology
I would like to thank Roser Cussó, Elisabeth Zucker, Morgane Labbé and Philippe Corcuff for their
suggestions and comments.
University of Lyon, ENTPE, Laboratory EVS-RIVES
© Springer International Publishing Switzerland 2016
I. Bruno et al. (eds.), The Social Sciences of Quantification, Logic, Argumentation
& Reasoning 13, DOI 10.1007/978-3-319-44000-2_6
Alain Desrosières studied the sociology that Pierre Bourdieu taught at the Ecole
nationale de la statistique et de l'administration économique (ENSAE) in Paris, from 1963 to
1966. Bourdieu had just returned from Algeria where he had worked with three administrators
of the Institut national de la statistique et des études économiques (INSEE), with whom he
had published one of his first books (Bourdieu et al. 1963). He nurtured the hope of initiating
a large-scale quantitative sociological undertaking like the one that had started to develop in
the United States (Desrosières 2003). Alain Desrosières was part of that generation of INSEE
administrators influenced by Pierre Bourdieu's teaching but who ended up developing an area
of sociological enquiry that was actually quite distant from that of the master: not "with" but
"about" statistical tools. His first objective was to establish their history (Affichard 1977,
Desrosières et al. 1976, Guibert et al. 1971).
This difference of perspective fully explains Alain Desrosières' decision, in 1984, to
support the creation establishment of the Groupe de sociologie politique et morale (GSPM).
One of the group’s founders, Laurent Thévenot, also an INSEE administrator, was working at
the time with the sociologist Luc Boltanski on the operations of social taxonomy
(Boltanski,Thévenot 1983), and with Robert Salais, another INSEE administrator, on the
conventions structuring labour markets (Salais,Thévenot 1986). These two research projects
were to lead to the development of the "economics of conventions" (Desrosières 2011). The
influence of Alain Desrosières, and in particular that of his work on the long history of French
socio-occupational classifications (SOCs) (Desrosières 1987 (1977)), was decisive in these
entangled dynamics, as the introduction to the book that brought them together attests
(Boltanski,Thévenot 1991). Desrosières’ approach to sociology was to change as a result.
This was evident in the first book that he published, with Laurent Thévenot, on SOCs, in
which he explored their history on the occasion of their reform and which has remained a
reference for French sociology students (Desrosières,Thévenot 1988). It was around this time
that Desrosières' perspective shifted in a sense from social history to the historical sociology
2. The Probabilistic Revolution as a historical and sociological tool
In fact this was more than a shift, for Alain Desrosières probably imagined the addition
of a line of "tension", to use one of his favourite terms. He studied the history of sociology
and that of its genetic ties with statistics, in particular (Desrosières 1985). At the time,
however, he was also regularly participating in the seminar on the history of the calculation of
probabilities organized by Ernest Coumet, Marc Barbut and Bernard Bru, all of whose work
he admired. That was where, in May 1987, he met Lorenz Krüger, who had been invited to
MIT on the occasion of the publication of The Probabilistic Revolution. A few months later
he met Lorraine Daston, whose work, following on from that of Coumet, was concerned with
the development of the first probabilistic techniques by Italian city state bankers during the
Renaissance (Daston 1989).
This contact with the Bielefeld group opened up new horizons for Alain Desrosières.
For several years, he had been seeking to develop links with research projects abroad. He had
combed through the English-language journals, in particular, and in 1984 had come across
Simon Szreter's work on the British census office (Szreter 1984). He wrote to Szreter and,