Boltanski Thevenot Los Usos Sociales Del Cuerpo

Social Science Information
http://ssi.sagepub.com
Finding one's way in social space: a study based on games

Luc Boltanski and Laurent Thvenot
Social Science Information 1983; 22; 631
DOI: 10.1177/053901883022004003
The online version of this article can be found at:
http://ssi.sagepub.com
Published by:
http://www.sagepublications.com
On behalf of:
Maison des Sciences de l'Homme
Additional services and information for Social Science Information can be found at:
Email Alerts: http://ssi.sagepub.com/cgi/alerts
Subscriptions: http://ssi.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.co.uk/journalsPermissions.nav
Citations http://ssi.sagepub.com/cgi/content/refs/22/4-5/631
Downloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009
Theory and methods

Thorie et mthodes
Luc Boltanski and Laurent Thvenot
Finding one's woy in social spoce :

o study bosed on gomes
The sociology of everyday interactions presupposes that individuals

share a common ability (manifested, for example, in situations of
symbolic confrontation, challenge, bluff, etc.) to spot the social
identity of actors and to manipulate the marks of this identity . This
ability, which has rarely been empirically investigated, is, however,
often presented as an immediate consequence of the process of
socialization, and the question of differences in individual comThis is a modified and enlarged version of a paper first presented at the Colloquium
on "Social Representations and Educational Field" held at Aix-en-Provence (Nov. 30Dec . 3, 1981), and organized by the Laboratoire Europen de Psy chologie Sociale
(LEPS) of the Maison des Sciences de l'Homme (MSH) in collaboration with the
UER de Psychologie of the Universit d'Aix, Marseille l. The authors are listed in
alphabetical order .
Alain Desrosieres, Michel Gollac, Dominique Jouglet and Baudouin Seys took
part in the preliminar y discussions on the formulation of the exercises. The collection of data was done with the co-operation of Isa Aldeghi, Nathalie Heinich and
Benedicte Vallet. Yan Darr helped with the huge task of processing the material.
We would like to thank them ali . This research has been carried out under the
auspices of the Institut national de la statistique et des tudes conomiques (INSEE)
and the Centre de sociologie de l'ducation et de la culture of the EHESS. lt has provided teaching material for initiation in social classification to be used by the department "Formation et Perfectionnement" of the I NSEE .
The transla tion is by Jill Forbes and Richard Nice.
Social Science lnformation ( SAGE , London, Beverly Hills and New Delhil,
22, 4/ 5 (1983) ,pp . 631-680
Dow nloaded from http://ssi.sagepub.com at La Trobe University on A ugust 3, 2009
from the SAGE Social Science Collecti on s. Ali Ri ght s Reserved .
632
Theory and methods
Boltanski and Thvenot
petence is rarely addressed. Our current research, outlined in this

article, was originally aimed at helping to clarify the following
questions: What is the relationship between individual mental images and the social process of public or official represent ation of
groups (e.g . by trade unions and professional associations, in collective agreements, by rule-governed professional bodies such as
that representing doctors, by governmental statistical institutes,
etc.)? 1 To what extent are these images shared, and is the capacity
to use them in concrete tasks of identification equal for all? If this
capacity varies from one individual to another, are the variations
Jinked solely to unequal degrees of knowledge of the established
classification and official labels (e.g. the names of occupations,
collective-bargaining classifications and statistical taxonomies [cf.
Thvenot, 1983])?
To endeavour to answer these questions, we designed a number
of exercises and game situations which required the participants to
apply their "social sense". But, in place of the pure experimentalism of the laboratory, we sought to set up modes of objectification which would bear on ordinary social situations, particularly in
the choice of material and in devising procedures derived from the
forms used in social life (the cases exploited in the exercises corresponded to real people, the procedures were those used in parlour
games, collective bargaining, etc.).
Information has so far been gathered in the following way: the
exercises were organized in two-day training sessions, attended
each time by a group of about fifteen participants, who had
previously been informed of the aims of our research. In order to
reach a reasonably varied audience, we offered the session to institutions concerned in one way or another with adult education:
the marketing division of a major nationalized firm, a nursing college, a management training college for social workers, a group of
sales representa tives from a multi-national food compny, a group
of unemployed people on a university re-training course, a group of
retired women teachers (from primary and secondary education)
belonging to a retirement club, a group from an adult-education
course in an IUT* which specializes in communications studies, a
group of INSEE* interviewers and two groups of people employed
to encode INSEE surveys. The contract was as follows: the training
session was offered free of charge, but in exchange the observers
were allowed to keep and use the participants' scripts and the taperecordings that were made. 2
* See Olossary at end of article.
Downloaded from http://ssi.sagepub.com at La Trobe University on August 3,2009
Thorie et mthodes
633
Each session included three exercises . The first required the participants to construct nomenclatures for (social) "mileu x" and
then to negotiate how they should be combined into a single
nomenclature. In the second exercise (actually performed first), 3
the participants were asked to produce three "typical examples" of
"cadres" and "ouvriers" (manual workers). The third and final
exercise was a game in which the object was to guess the occupation
and social milieu of a real but unknown person, with the help of
clues that were supplied to the participants on request.
The practica! creation of social classifications
The aim of the exercise, which involved giving examples to the participants which they were asked to classify so as to produce a social
nomenclature, was to explore the mental categories used to think
about society. The exercise was specifically designed to allow an
analysis of the relation between the ordinary categories used in a
practica! way to place people in society, and official classifications
such as the code of socio-occupational categories that INSEE uses,
which produce a homogeneous or standardized picture of the social
world for various purposes such as censuses, government department records, social surveys , and so on. Our object, in this exercise, was to test out hypotheses put forward in the course of earlier
studies of the encoding of statements about occupational status collected by INSEE. What was interesting about such studies is that
they looked at routine, standardized situations in which classification was required, in which agents were told to make use of the
stated rules of classification that they had been given, and that were
broadly based on the use of a system of rational criteria . By defining the agents' task in such economical and almost labour-saving
terms, it was assumed that no overall understanding of what was
contained in the questionnaire would be achieved (in other words ,
no overall appreciation of the person), but that the agents would
make piecemeal use of severa! different criteria. Moreover, such a
working method, combining severa! criteria , exactly corresponds to
the way in which social nomenclatures are constructed for
academic purposes . Experience showed that in a far from insignificant number of especially complex cases, in which the rules supplied did not allow the agents to carry out their task, the agents
acted in a quite different way which might be described as follows :
Downloaded from http://ss i.sagepub.com at La Trobe University on A ugust 3, 2009
634
Theory and methods
(1) They took into consideration a great deal of information

contained in the questionnaire (instead of confining themselves to
the criteria they had been asked to use).
(2) They identified social forms on the basis of this exploration
by assimilation to social types that they were already familiar with
and which had
already been situated within the social
nomenclature.
Using these initial observations, we devised an exercise which
allowed us to record the procedures involved in classifying in a
systematic way. In addition to its academic uses, the exercise is of
particular educational interest in that it draws the attention of those
who participate to the mechanisms on which the creation of sociooccupational nomenclatures is based, usually implicitly, and the
way in which they are used in encoding operations.
The classification exercise was devised in the following way. 4
The participants were divided into teams of two -an arrangement
which has the merit of stimulating bargaining and discussion and,
therefore, of making explicit sorne of the classifying principies at
work. This is not achieved if the participants work alone (in sorne
cases we recorded ali the remarks made during the exercise). Each
team is given a package of 65 personal file cards, based on real
examples, containing the following information: age, domicile,
educational qualifications, age at which full-time education ceased,
occupation, status (self-employed, employee, etc.; number of
employees in firm, etc.) and, where relevant, level of skill or rank ,
name and address of employer/place of work.
These cards were made up from census returns chosen so as to
represent a fairly broad cross section of occupations and social
situations. We tried to mx examples which we thought could be
easily placed socially, beca use they conform to commonly held
social images (e.g. "doctor" or "primary teacher ") and examples
of a more decrystallized status (cf. Lenski, 1954) that we thought
would prove difficult for that reason (e.g. that of the "garage
hand" (laveur graisseur dans un garage) aged 26 who had a university degree (diplme d'tudes suprieures). The kind of information
used and the way it was presented recalled the questionnaires used
for statistical purposes (censuses etc.). Such similarities undoubtedly had an effect on the way the exercise was done since they
encouraged sorne rationalization of the process of classification
through the use of explicit criteria. Thus everything, in the procedures followed, which ran counter to that logic must be treated as
Thorie
et mthodes
Example
of personal 633
file
card
tl::l
o-.
i:i'
;:s
"'
1
Nom et Prnom s:
. VELL/R.. . . Mf\.t({<,,tGF .
.g
Age:
6 Profession:
f "'9lrdw'
7 Ltes-vous.
3
!'l.
. . . e tu:M(1W(;. FuraM. . . . . . . . . .
4
Catire . COIWHe.rcicl
Ad
!!l
w
"e:oo
9
5
ge !in t udcs: . .
2.2 .ans . . . . . .
Nom et ad rcssc de l 'ta bl isscrncn t :
MICH&LIN
enseignemen t tec h niq ue et forrnat ion

rirofes sion ncllc
ress
;;i
"''
"'
--::
;:s
q ua l i fica t ion , grade ou rimi t ion

hirarchiq uc
Dirilimes les rilus levs:

enseignement gnral ou su ririeu r
. H.E.C.
'
;:s
CDt,trfYlercial. .
-Mem bre d 'u ne rirofession li bralc

-Exriloi tant agricole, emri l oyeu r ou
t ravailleu r i ndriendan t : a n isa n,
com mer an t , i nd ust ricl, etc.
Nom bre de sa l aris:
-Sala ri
Dans ce cas:
2q ans .
3:;;
!::,
. Ptoce . .
. Carms .
C.W'IJ.tOllJ! l'Vran.ti. (63)
i\c1i,i1 de cct 1abli s:.erncn1 :
Mf
. de. >'1eHWls
;;i
" ''
o....
;:;:;
"'
.....
":;''.
o
q1
w'
Vi
<
636
Theory and methods
particularly noteworthy.
What happens is this. The observer tells the participants in the
exercise to "divitle the cards into piles according to the social
categories of the individuals in question". They may make as many
different piles as they wish and there is no time limit (usually they
take between three quarters of an hour and an hour to complete the
task). The definition of what constitutes a category is left
deliberately vague and observers refuse to be more specific if questioned by the participants. At the end of this first phase, participants are asked to choose the card they consider most representative, that is the card which "allows other people to understand
what the rest of the pile contains", and to place it on top of the
pile. Finally, participants are asked to give each pile a name.
The way participants usually go about the task is as follows: they
consider the cards one by one in the arbitrary order in which they
find them (each set of cards has been shuffled in advance), and
place them on the table so as to begin to divide them up in the light
of the initial differences which emerge. Next, the participants try to
add new cards to each of the piles they have started. The first cards
are usually given preferential treatment and are divided according
to a few simple distinctions. These may have to do with the usual
topography of the social world ("upper", "lower", "middle"), or
with essential or important properties attributed to certain occupations ("the land", "the manufacture of material objects",
"manual work", "art" . . . ), or with divisions that are inspired by
statistical criteria ("public sector employee", "private sector
employee", "self-employed"), or to do with differences embodied
in institutional categories ("teacher", "health worker", "skilled
worker").
Thus, two women employed in the marketing division of a large nationalized
company, of whom one was a graduate cadre and the daughter of a self-taught
(autodidacte* ) cadre, who worked in the same firm, and the other a supervisor
with a BEPC* and the daughter of a clerical worker (emp/oy) in the same firm,
began the exercise by starting three piles which they set out in a triangle. One
pile, on the left, consisted of "manual-type" (plutt manuel) jobs ("later we
might take out the students who are doing that kind of work because it's their
first job") . Another pile, on the right, consisted of "managerial-type" (plutt
cadre) jobs ("later we might separate the graduates from others like
shopkeepers"). A third pile which they started between and above the two
others, they called "in between" ("supervisor and cadre" ). This initial taxonomy was by no means clear-cut since the term cadre figured in two of the piles .
Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009
Thorie et mthodes
637
Its lack of clear definition probably had something to do with the fact that the
women themselves were borderline cases , both the group they were with and
their colleagues at work being better qualified than them and from more middle
class (bourgeois) backgrounds. The lack of assurance expressed by "in-between"
received poltica! and taxonomic expression in the final stage of their division
which resembled, more closely than taxonomies established by other participants
in the same session, the principies of classification used by the trade union (the
CGT*) to which one of the women belonged - use of the phrases "health
sector" and "education sector", references to "skill" and to the two extremes
typifying the working class, the ouvrier P2 fraiseur and the femme OS emballeuse* (the establishment of these extremes wll be further discussed below).
But the task which initially seemed easy, rapidly turned out to be
more complicated. The instruction to combine different piles of
cards means that sorne form of coherence will be required, and this
could not be met, especially as decrystallized cases had deliberately
been included in each package of cards. Difficulties arose when
participants had to deal with cards which could be placed in a
number of different piles depending on how they were considered:
Here is the example of a team (from the session with female interviewers from
INSEE) whose members had right at the beginning started the following piles:
"State employees", "cadres" , "manual workers" ( ouvriers), "self-employed",
"paramedical". With regard to the "State Registered Nurse" (64), they did not
know whether to put her in the "self-employed" pile, as a member of the "professions" or in the pile which they initially designated "paramedical". In cases
like this, the simplest soluton, and the one often adopted to begin with, is to
divide the pile, so that "State employees", for example, are separated into two
new categories consisting of "highly qualified State employees" and "State
employees who have no specific qualifications". But this form of coherence is
quickly abandoned because it simply leads to more and more sub-divisions and
considerably increases the range of the nomenclature and therefore the problem
posed by arbitrating an increasing number of borderline cases. Thus in the case
of the "professeur agrg"* (51), for example, the participants did not wish to
place this card in the "highly qualified State employee" category on the grounds
that "the teaching profession is different , it's separate", and decided to
reorganize their piles so as to create a new category consistng of "teachers". But
they encountered a new cognitive conflict with the "prmary teacher" (instituteur) (57) because of their sensitivity to differences in "qualifications",
"degrees", and "leve!".
When dividing piles becomes too difficult, the solution that is

often adopted consists in abandoning any attempt to create a
nomenclature based on an unchanging system of criteria, together
with any attempt to remain coherent, and to build up piles based on
638
Theory and methods
similarities between cards which belong to categories whose

resemblance is vague, fluid and implicit, in such a way that the connexion between the various cards may well vary within any one pile.
Thus we observed that as the job of classification proceeded, the
identity of each of the piles could vary quite considerably without
necessarily entailing a restructuring of the whole system of pilebuilding as it had initially been conceived along the lines of a small
number of basic oppositions or relations (e.g. manual/non-manual
work, graduate/non-graduate, etc.) to be found in the first cards
that were looked at. Indeed, as each pile gradually grows larger, the
process may appear to be one in which memory plays no part and in
which the principies for the perception of similarities vary continually.
For example, using the first two cards that are looked at, a pile is created
which depends both on the opposition between the public sector and the prvate
sector and on the difference between high qualifications and average qualifications. The third and fourth cards that are looked at are "nurse" and
"psychiatric social worker" which, for the time being, are added to the pile that
temporarily contains (according to the recording of the participants ' conversation) ''middle-range public sector workers''. But as a result of adding these two
cards, the pile takes on a new dimension, that of "health workers", and this
gradually takes over the identity of the pile .
Another team initially started a pile of "artist" cards to which they added
"photographer" (7). But a few minutes later they decided that a
"photographer" is a "craftsman" (artisan) and they added severa! craftsman
cards to this pile as though they had forgotten its initial identity.
This suggests that categories are not constituted a priori according to sorne formal identity, but on the basis of chain association
by contiguity described by Bruner et al. (1966, pp. 216-230), and
that such a mode of category building is not confined to children
who do not yet understand the logic on which social class is based
or to adults who have little education. The reverse appears to be the
case, namely that this is an instance of the working of practica!
logic 5 which nevertheless does not prevent appeal to class logic of
the most academically respectable kind when the experimenter explicitly demands it.
Legitimate labels and the naming of categories

The method of classification that has just been described accounts
Thorie et mthodes
639
for the participants' difficulties when they are asked to provide a

name for each of the piles they have made. This new instruction
often leads them to revise ali their classifications so as to check the
interna! coherence of each pile and, sometimes, to modif y it by
moving cards from one pile to another. Similarly, when they inform observers of the names they have chosen, they often hedge
them round with qualifications such as "it's not exactly right but
it's more or less that", or "this name doesn't quite cover ali the
cards but we can't think of anything better", and so on.
Finally, the relation between the names given to the piles and the
content of the piles has to be understood in the light of what is
known about the way the piles were put together. In a great many
cases, perhaps the majority (though statistics are not yet available
on this point), it would appear that the names given to the piles are
those which are used by official bodies and social scientists or in the
nomenclature of official statistics (e.g. cadres moyens* , cadres
suprieurs et professions librales*, agents de maftrise* , ouvriers

qualifis* , etc.), and this tendency would appear to be particularly
marked in groups whose members have high educational qualifications. In such groups, it is rare to come across labels that do not
figure in the official systems of classification and description. And
what is noticeably absent are the names of groups which derive
from Marxist terminology relating to social classes, terms such as
"proletariat", "bourgeoisie", "petty bourgeoisie", or those which
are defined by reference to a system of ethical values, such as "the
poor", "the notables" , "the elite" etc. The euphemistic terms of
administrative jargon are almost always chosen in preference to the
socially and emotionally charged terms of ordinary speech (e.g.
"cultivateurs" (farmers) rather than "paysans" (peasants) or
"enseignants suprieurs" (lecturers in higher education) rather
than "intellectuels" etc.). This phenomenon must be related to the
emergence and ppularization of an official system for designating
social groups whose legitimating principie is derived from the
"social sciences". In origin, this is linked to the creation between
the wars (1936) and especially since 1945, of an official field of
representation of social groups and social classes and of negotiation between classes (pay bargaining, joint planning commissions,
semi-institutionalized relations between the State and the trade
unions etc.). It is almost as if social agents could make use of a
vocabulary t hat is apparently technical and neutral, which then
enables them, in t he ordinary cou rse of their daily lives, to tal k
Thorie et mthodes
640
about social differences -that is, social classes -in a euphemistic

fashion, without seeming to refer to the class struggle. The creation
of such a vocabulary for the representation of social classes can be
compared to the way in which the vocabulary of psychoanalysis (cf.
Moscovici, 1961) gave intellectuals, in particular, a means of talking about sexuality, and indeed their own sexuality, without using
"obscene words" or terms which were sexually charged, in a way
that was divorced from reality and had no immediate sexual
implications.
But of course this also meaos that any departure from the
official nomenclatures, however small it may be, in the taxonomies
created by thc participants, is extremely significant and expresses
an intention to say something or to make a point about the social
world and the divisions in it, about what these divisions are and
what they ought to be.
lt would also appear that departures from the norm are often
associated with conflicting ideas participants may have about their
own social identity.
For example, in a session with participants who were ali unemployed and on a
training course, a team made up of two former company directors, both of
whom were over fifty, one a graduate and the other with no degree qualifications, created a taxonomy in which "graduate directors" (dirigeants dip/ms)
were distinguished from "self-taught directors" (dirigeants autodidactes). In the
same session, a technician who had bec:n promoted to the rank of engineer within
the company, and had ther. been sacked and had been unemployed for two years,
after a disagreement with the "real" engineers (ingnieurs* ) who were Grande
Eco/e* graduates, created a nomenclature in which the space given to cadres and
ingnieurs in other taxonomies was practically empty. As if he did not wish to
think about or write names which brought back painful memories, he used a
system of abstract criteria to designate the various piles, a system which was based on what he saw as an opposition between the education system and skills (i.e.
"employees who have very specialized higher educational qua!ifications" as
against "employees who are graduates'', "employees with particular skills,"
"employees without particular skills" etc.)
Still in the same session, a team which was made up of two young graduates
who were unemployed created a category they called "temporary underemployed" for the card concerning the graduate garage hand (11) whicn had
been included in the package precisely because it was an extreme case of
decrystallization (the person in question was a 26-year-old holding a higher
degree). This was also the reason why most of the participants considered his
case "unrepresentative" and "marginal" .
Although participants generally used a relatively limited and fix-
640
Theory and methods
ed vocabular y to designate social groups (which, as has been seen,

is the result of the imposition of the official statistical categories), it
is still the case that one !abe! or two associated labels may be used
to designate different groups of cards .This was so in the case of the
session with social worker team leader s, a fairly homogeneous
group in many respect s (sex, age, income, qualifications etc., most
of the major differences in fact relating to social background and
the spouses' occupations). In the six taxonomies produced by this
group of twelve people (working in teams of two) there was one
they termed "cadres moyens et maftrise" , or alternatively one
category called "cadres moyens" and one called "maftrise"* (or,
in the fifth team a "classes moyennes"* category and in the sixth
team a "technicien" category) . When the cards in these different
categories were checked , it was apparent that not one was to be
found in ali six collections of cards. One card was found in five out
of six collections ("chef de service" (65)), four cards were to be
found in four out of the six collections ("attach de portefeuil/e"
(28), "contremaftre" (40), "infirmiere DE" (64), "assistante en
psychiatrie socia/e" (2)). Thirteen cards appeared in only one or
two collections . Thus it would seem that the relation between labels
and the groups they refer to is particularly ill-defined in the intermediate zones of social space . Thus ali six teams in this session
created a category that they labelled "cadres suprieurs" (or
"cadres suprieurs et professions librales" or "cadres suprieurs
et ingnieurs" ) . In such cases, three cards were to be found in ali six
collections of cards (these were "ingnieurs TP" (45), "avocat"
(48), "mdecin" (22)), and three cards were to be found in five out
of the six collections of cards ("recherche scientifique" (59),
"pharmacien" (37) and "professeur agrg" (51)). The range of
card collections in which any one card may appear would appear to
increase in proportion to the fluidity of the characteristics of the
occupation designated on the card. Thus "ingnieur conseil en
chauffage central" (5), for example, whose job description contains a reference to an advanced educational qualification ("ingnieur" ) and to a professional activity ("conseil" (consultant)),
but whose educational qualifications go no higher than the BEPC,
and who is not self-employed, was, during this same session, placed
in the following groups: "professions" (professions librales)
(twice), "cadres suprieurs" , "cadres moyens" , "ingnieurs" ,
"clerical workers" (employs).
642 Theory
and methods
Boltanski
and Thvenot
Bo/tanski
and Thvenot
Thorie
et mthodes
642
TABLE 1
Frequency with which cards were chosen to head piles

Card
No. Occupation
22 mdecin
36 PDG (President-DirecteurGnral)
33 cultivateur
42 fraiseur P2
38 ouvriere d'usine emballeuse
45 ingnieur T P
61 entrepreneur BTP
51 professeur agrg
24 bniste artisan
48 avocat
57 instituteur
49 femme de mnage (entreprise)
4 ingnieur commercial
25 hydraulicienc
12 dcorateur
40 contrema'tre
44 lectricien
56 cadre commercial
20 secrtaire comptable
21 ngociant en vin
66 fleuriste
10 vendeur en antiquit
39 ouvrier agricole
62 agent de bureau
11 laveur graisseur
18 agent SNCF
19 plongeur
34 agent PTT, facteur
41 caissiere
46 secrtaire de mairie
59 chercheur scientifique
65 chef de service installation industrielle
13 dessinateur publicitaire
17 exploitant forestier
30 antiquaire, brocanteur
2 assistante en psychiatrie sociale
5 ingnieur conseil chauffage
central
8 auxiliaire puricultrice
35 professeur de collge
37 pharmacien
53 inspecteur villed
English translation of
Freq. occupational termsb
30 doctor
29 company chairman (UK)
president (USA)
21 farmer
20 skilled milling machine operator
18 packer (female)
17 civil engineer
17 building contractor
16 lyce/university teacher
15 cabinetmaker (craftsman)
15 lawyer
15 primary teacher
12 office cleaner (female)
11 sales engineer
11 hydraulic engineer or technici an
10 interior decorator
10 foreman
10 electrician
10 sales executive
9 accounts clerk
9 wine merchant
9 florist
8 antiques salesman
8 farm labourer
8 office clerk
7 garage hand
7 railway employee
7 dishwasher
7 postman
7 cashier (female)
7 town hall secretary
7 research scientist
7
6
6
6
5
industrial superintendant
graphic designer (advertising)
forester
antique/junk shop owner
psychiatric social worker (female)
5 central heating engineer

5 auxiliary nursery nurse (female)
5 comprehensive school teacher
5 pharmacist
5 urban inspector
Downloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009
Bo/tanski and Thvenot
Thorie et mthodes
643
TABLE l conlinued
Card
No. Occupalion
54 patissier
63 monteur dpanneu r
64 infirmiere DE
bucheron
4 bniste vernisseur
43 em ploye PTT, perforatrice
47
60
6
15
16
26
27
50
55
7
31
58
dbitant de boissons
femme de mnage (particulier s)
magasinier automobi le
jardinier
technicienne en classement
vendeuse fruits et lgumes
lustreur
moniteur-ducateur
attach a l'informat ion mdicale
photographe
aide-cuisiniere
veilleur de n uit
English 1ransla1io11 of
Freq. occupa1ional 1erms1'
5 pastry cook (male)
maintenance man
State Registered Nurse (female)
4 woodcu tter
4 furnit ure polisher (male)
4 pu nch card operator, Post Office
(female)
4 inn keeper
4 cleaning lady
3 garage warehouseman
3 gardener
3 filing technician (female)
3 salesgirl (greengrocers)
3 polisher
3 coach, special teacher (mate)
3 medica! information officer
2 photographer
2
assistant cook (female)
2
nightwatchman
a. The job terms in this list (and on the file cards) are those found on t he census
retu rns
b. The "equivalents" offered here are necessarily arbit rary a nd debatable : t he usual
difficulties are compou nded when one is dealing wit h job-titles t hat are points
and meaos of struggle for rights and recognition . (Transl. note)
c. . One of the ambiguous terms, imprecise with respect to "leve!"
d . Very vague term from which type of occupation cannot be determined
The groups and their typical representatives
Despite the fact that a name given to a particular category, which is

usually taken from legitima te statistical and ' administrative taxonomies (e.g. "ouvrier qua/ifi" , "cadre mayen" etc.), may
well, in different teams, cover groups of occupations which do not
match and whose connexions are often very limited, nevertheless
the participants do share similar views about certain salient features
which allow them to find their way around the social world.
If one looks at the frequency with which certain cards were
chosen to figure at the head of the piles (in order, it will be recalled ,
"to !et other people k now what is in the pile" without having to use
Downloaded from http://ssi.sagepub.com at La Trabe University on August 3,2009
644
Theory and methods
a name or a label), it is clear that certain individ uals whose social

properties are committed to cards were chosen more frequently
than others to represent social groups , and sometimes even to be a
kind of paradigmatic embodiment of social groups .
One example was a group of cards which might a priori have
been thought to exemplify the relatively similar positions of various
kinds of unskilled manual workers and therefore to have been
equally likely to be selected to illustrate piles that had been constituted on that principie: "ouvriere d'usine emballeuse" (38),
"employe des PTT perforatrice" (43), "femme de mnage" (60),
"magasinier automobile" (6) , "aide cuisiniere (31), "veilleur de
nuit" (68). However, the first was chosen sixteen times 6 to head a
pile , and the remainder only t wo or three times each (cf. Table).
Such a regular occurrence suggests that, in the agents ' representations of the social world, sorne positions stand out more than
others and are incarnated in social persons who take on a
paradigmatic value. lt is as if the female packer in a factory were a
better example than the assistant cook or the unskilled worker,
without educational qualifications and with no possessions to his or
her name, who is badly paid and who was sometimes descri bed by
the participants as being "on the bottom rung" or "at the bottom
of the heap". Similar observations migh t be made about most
social grou ps. Thus the "fraiseur P2" (42) is the paradigmatic
representative of the skilled worker (chosen twenty times to head a
pile as against "ouvrier pitissier" ( 54.) five times, "monteur
dpanneur" (63) five times and "ouvrier bniste vernisseur" (4)
fou r times etc.)
Similarly, the "ingnieur tra vaux publics" (45), chosen seventeen times, is a bet ter example of a cadre than th e "cadre commercial" (56) who was chosen ten times, or the "chef de service d'installation industrie/le" (65) chosen seven times, or the "ingnieur
en chauf fage central" (5) who was chosen five times, etc.
The hypothesis might be advanced that these salient positions in
social space are filled by occupations which, over the years, have
benefited from an effort o/ social representation. This work of
representation has inevitably taken place in t he realm of social and
mental representations, in the sense that the properties which
attach to a position and to those who occupy it are the result of a
collective labour of construction, expression and simplification (or,
to use Goff man's term; "dramatic accen tuation"). But che
"representat ion" in q uestion also belongs to the domain of politics
Thorie et mthodes
645
or trade-unionism, with the creation of organizations which

mobilize groups and the appearance of "representatives" or
"spokesmen" who are mandated to speak for a particular group.
Thus it could be shown that the individuals whose cards were
chosen by our participants as embodiments of social groups either
belong to professions which are highly institutionalized and highly
organized ("doctors", "engineers", "company chairmen", etc.)
or else, as in the case of manual workers, are in occupations to
which there attach social properties that conform to the
paradigmatic representation that trade union work around
representation has given to their group. Thus the "fraiseur P2" ,
whose very name refers back to a traditional and typical kind of
industrial activity ("metalworker") and a degree of skill which has
been institutionalized by pay bargaining ("P2") is, out of all the
cases included here, the best example of a manual worker ( ouvrier)
as traditionally defined, those "skilled (male) industrial workers"
who, according to Edmond Maire (General Secretary of the CFDT*
trade union), "dominate the trade union movement" (Maire, 1980,
p. 171). As for the "ouvriere d'usine emballeuse" , her title brings
together a reference to the factory (" usine" ), to a job that is not
directly productive ("packer" = "emballeuse" ), which has no skill
and no dignity, and to both the "condition of the wor king class"
and the "condition of women". She too constitu tes what, today,
has become a paradigmatic example precisely by contrast with the
skilled male industrial worker . The case of the female packer
distills all the properties attached to so-called "marginal" wor kers,
women and unskilled workers, those who receive no attention from
the "traditional" trade union organizations, and whom the more
recently developed sections within the trade union movement, and
particularly the CFDT (whose spokespersons in fact use this as a
means to differentiate themselves from the CGT*), have undertaken to represent, i.e. to defend with the weapons of trade union
and political struggle and by the same token to constitu te as typical
examples of "the working class today".
Conversely, what one might call bad or untypical examples,
cards which were rarely selected to head a pile, consisted either of
occupations which are less well known or less frequently
represented ("jardinier" (5) was chosen three times as opposed to
"cultivateur" (33) which was ch0sen twenty-one times), or of occupations which were perceived as being on the borderline between
a n um ber of different constit uted categories (t h us the
Downloaded from http://ssi.sagepub .com at La Trabe University on August 3, 2009
646
Theory and methods
"photographe" (7), chosen twice, was not seen as being a real

craftsman [artisan] in the same way as the "bniste" (24) chosen
fifteen times, nor as much of an artiste as the "dcorateur" (12)
who was chosen ten times, nor as much of a technician as the
"hydraulicien" (25) who was chosen eleven times, but was not,
however, assimilated to the "ordinary" "clerical" or "manual"
worker). Alternatively, such cards related to individuals who did
not really possess the properties implicitly associated with the name
of the profession they presented themselves as belonging to and
whose status has, as a result, become decrystallized. One such case
was the "technicienne en classement de bureau" (16), chosen three
times, who has no degree, left school at fourteen, does an unskilled
job and who therefore, despite the fact that she calls herself a
"technician", is closer to being a manual worker ar unskilled
clerical worker.
Thus the variations in the way different individuals and different
groups represent the social world seem to be that much greater
when they are required to put individuals into categories, ar in
other words to create discrete groups (separated from each other by
clear dividing lines) of examples which are supposedly
homogeneous in relation to each other, to which a generic name is
given. The task seems to be easier, on the other hand, when what is
required is the identification of salient features within a given
cognitive space whose different segments need not necessarily be
organized ar structured in the same way. The job of categorization
in fact presupposes an ability to divide social space into discrete
categories, and this requires participants to have a homogeneous
representation of social space and of its divisions. It is, rather, as if
ordinary social sense worked as though it were directed not by
reference to a homogeneous social space that is divided up and
mapped out in a certain way (which is what academic nomenclature
presuppo ses) such that ali social positions can equally easily be
placed within it. But instead it works by reference to a social space
which is very varied, with sorne features that stand out and that individuals have a clear enough image of to be able to pick them out
in the social landscape, and to identify and classify them, but also
with grey areas and positions that are ambiguous or difficult to
classify, that is think about , except by assimilation to positions that
stand out and are recognised as being the least remate in social
space (cf. Bourdieu, 1980a and Centlivre, 1979).
Thorie et mthodes
647
The stylization of social groups

Organizing categories around typical occupations which are used to
envisage different social milieux is a device which is commonly used, as is confirmed by provisional analysis of the findings of the
first exercise carried out by the participants at the opening of a session.
This is an exercise in which participants are asked to think of
typical examples of various social categories, and it allows the
observers to investigate the individuals' ability to generate social
forros, by obliging them to combine properties in specific forros
that are given paradigmatic value . We hoped that this exercise
would contribute to our study of the structure of the mental
categories in which social groups are thought about.
The work of Eleonor Rosch (1973, 1977, and also Rosch and
Lloyd, 1978) has shown that "popular" or practica! taxonomies
(i.e. those not given objective reality by means of totalizing or
schematic representations such as diagrams, graphs; tables and so
on, 7 and not rendered legitimate by a body of specialist workers),
u nli ke academic nomenclat u res, were not made u p of
homogeneous categories or "classes", separated from each other
by clear boundaries, nor were they composed (as "componential"
analysis, for example, assumes) 8 according to a limited number of
criteria. The "semantic categories of ordinary languages" are
organised, in these hypotheses, around the best examples of each
category, the "clear-cut cases" which are easy to envisage, and
these are surrounded by a cluster of less clear-cut cases which make
up the rest of the category. It follows, first, that each category has
an "interna! structure" (it is not made up of equivalent and undif ferentiated elements) and, second, that no category has "precisely
defined limits". The tasks devised by Rosch to test these
hypotheses have up to now related to categories of shapes and colours and to biological and zoological categories. She has, for example, shown that for sorne people sorne dogs "are more doggy" than
others. Retrievers, for example, are quintessentially doggy because
of their floppy ears and feathered tails, and they occupy a central
position in the category. Pekinese, on the other hand, are "bad
examples".
When we asked participants to give three examples of cadres
and three examples of manual workers ( ou vriers) - exam ples
which they could invent if they wished - we gave ourselves the
648
Theory and methods
material with which to create an "imaginary" sample of "typical"

representatives or "good examples" 9 of these two categories , and
these were examples which we could subsequently compare with
samples that had been conceived as "statistically representative" of
these two categories (e.g. the samples provided by INSEE). In addition to the exercises devised by Rosch, which simply consisted in
placing the names on a list in rank order according to how exemplary they were felt to be, our exercise required participants to
use greater skills and, indeed, different skills, because they had to
translate into institutionalized and standardized variables (census
or statistics), individual forms that they typically associated with
the generic name of the category.
The preliminary analysis of the imaginary sample would appear
to show that the examples officially included in the cadre category
are not all "good examples" of cadres to quite the same degree. w
Participants were asked to find examples of "very cadre cadres"
and to say what their job was, their age, sex, qualifications, monthly
income, the name and size of the firm they worked for, where they
lived and what kind of car they drove. They were allowed to partially or totally invent their examples -and did so in about half the
cases -or to base them on people they knew at work or elsewhere .
By putting together these identikit pictures we were able to constitute an "exemplary" or, perhaps, imaginary sample which <loes
not reproduce any statistical sample of the kind that INSEE produces. The cadres chosen or invented as examples are, in general,
better qualified than those in the statistical samples (with a particularly heavy representation of HEC* graduates). They are also
more likely to live in Paris, closer to forty years of age (the "prime
of life") than to youth or old age. They are extremely likely to be
involved in commercial activities, especially those related to
marketing or advertising, and unlikely to be on the manufacturing
side of their companies. Public sector cadres are very rare in the
imaginary sample (although participants were not told to limit their
choice to the private sector). Small firms and organizations are
under represented, while large firms are strongly represented,
especially large private sector conglomerates or multinationals with
a high profile, such as BSN, Pchiney and, above all, IBM. (The
small businesses concerned are often advertising agencies or service
or consultancy firms.) Thus there emerges a stylized image of a
"top level" cadre whose attributes have a particularly strong
emblematic value (the HEC graduate, employed by IBM in their
Thorie et mthodes
649
marketing, advertising or data proce ssing divisions, who drives a

Mercedes or a BMW etc. recurs time and again and functions as the
central component in the representation).
As far as the case analysed here is concerned, it can be shown that
the structure of this mental category has been produced by the
history of the formation of the group . A study of the history of
cadres, who openly formed themselves into a social category in
France during the struggles of the years 1936-1940 when they
acquired their name and their organizations, shows that this "new"
group was in fact the result of the merger of several categories who
defined themselves by including sorne and excluding others and
thus delimiting the group. The category was constituted by a process of successive assimilations grouped round a central nucleus
made up of ingnieurs who were graduates of the Grandes Eco/es
(especially the Eco/e Centra/e) , who carne from bourgeois
backgrounds, and so on. This central nucleus formed a sort of
magnet to which others were attracted. The ingnieurs coming out
of the Grandes Eco/es and the top level cadres controlled the
group 's system of representation for a ver y long time; they controlled, in other words, both how the group was represented at
political and trade union level, and the ways in which it was socially
and mentally represented. Thus the representation of "authentic"
cadres was created in these people's image even though the group
itself was extremely heterogeneous in terms of its professional and
social composition. The group in fact included fractions of both
the big and the petty bourgeoisie and contained far more
autodidactes (about 65 per cent) - who were negatively defined
by the fact that they were not graduates -than it <lid graduates of
the most prestigious Grandes Eco/es (between 10 and 15 per cent).
Thus the petits cadres , who were numerically dominant but
politically and socially dominated, were represented , in every sense
of the word, by those who occupied a central position in this vaguely defined and constantly changing group (cf. Boltanski, 1975).
Thus the exemplary cases which come to mind "spontaneously"
when the category is mentioned and who, so to speak, stand out in
the social landscape (in such a way that less good examples are
defined as belonging to the category by virtue of how closely they
do or do not approximate to certain examples), are stylized and
schematized representations of agents who occupy a central position in the category . It is, in other words , as though the people
questioned, when asked to give content and substance to the term
650
Boltanski
and Thvenot
TABLE 2
v.
The imaginary and lhe slalistical sample of cadres
Theory and methods
'
lmaginary sample
S1a1is1ical samp!e
Percentages
o
1l.
3
m
,..
.
m
g
'
o
Cadres
commerciaux
GROUP D
Primary &
ALL GROUPS
secondary
1eachers
(ret ire<!)
(N = 52)
(N = 192)
Refere nce sample:
t:i
INSEE 1970 survey

"Forma1ion, qualifi-
::i...
::::
(N =49)
39
20
11
19
56
67
31
40
38
44
26
66
24
45
16
35
28
13
18
11
11
31
Age: between 35 and

45
-B
71
57
69
58
59
35
V,
Resident in Paris or
Paris region
87
90
75
78
80
41
Firm with less 1han

500 employees
23
28
36
27
26
75
Bacca laurat highest

educationa l qualification
Graduate of Ecole
!'ingnieurs or
cation, profession"
:::=
s.
o
Ecole de commerce
Nature of job - sales,
mark et ing, finance
Narnre of job production, me1hods,
maintcnance
"'
GROUP C
(N = 48)
GROUP B
Unemployed
big
company
(N = 43)
.g
GROU P A
Graduate
cadres, big
company
::::
::'5:
t :i
*Percentages calcu!ated on the 101al of examples provided by each of 1he groups .
::::
::i...
'
'<!
::::
o...,.
Thorie et mthodes
651
cadre (rather than to give an abstract definition based on certain

"criteria"), spontaneously thought of paradigmatic examples and
did not think of borderline cases in a category whose limits are, in
any case, vague. Good examples of cadres are modelled on the
stereotypes produced from the work of representation that the group
has carried out. In order to explain the choice of these "good examples" and to justify the structure of the mental category, we cannot discuss here the "natural" properties which are linked to the
physiological characteristics of instruments of perception (cf.
Berlin and Kay, 1969; Kay and McDaniel, 1978). In the case of the
cadres, the focal point is provided by examples which are a
stylized expression of the properties of the people who have succeeded in monopolizing the category's means of social and political
representation. Examples that stand out in this way are the product
of the conflicts and struggles that surrounded the formation of the
group, and of its history which is thus encapsulated in mental structures in sketchy or outline form.
But what distinguishes categories of groups from so-called
"natural" categories (such as shapes and colours, or even
zoological and botanical categories) cannot simply be explained by
recalling that they are the product of history. A more fundamental
property gives them a completely original status. Because they are
used to classify the native social world, the categories refer to a
universe that also belongs to the person who makes use of them and
who, when organizing a set of exterior objects, simultaneously
determines his own social position in relation to them. Thus these
categories are always value-laden and are never received or used in
a neutral or passive way. For example, the dominant representation
may be familiar, and recognised as dominant, but need not nevertheless be accepted as "accurate", or it may become the object of
complex strategies containing a mixture of recognition and refusal,
acceptance and rejection. Thus it can be seen that agents are more
willing to provide examples of cadres that conform to the most
official representation of the category when they themselves occupy
a dominant position within the category (Group A) or when they
are completely outside the category and do not feel personally
implicated in the task they are asked to perform (Group D). On the
other hand, those who occupy a dominated or vulnerable position
in the business field (such as Group B which was made up of
unemployed people from industrial or commercial companies) are
able both to demonstrate that they are familiar with the dominant
652
Theory and methods
representation, and that they do not accept it. They do this by

combining examples that conform to the central positions within
the category with examples of petits cadres* , whose positions are
marginal, and by presenting the latter as more "typical" or more
"representative" than the former. In this way they perform the
basic political action of using statistical representativeness to
challenge political representatives and of refusing those who occupy a position of power within the group the right to speak and act
in the group's name or to behave as if they incarnated it.
To be more precise, a full analysis of the responses shows that
the examples provided by the participants did not correspond in
any mechanical or passive way to a "stereotype" that could have
been produced by the right kind of "stimulus". When responses
are not "stereotypical" it does not mean that the "stereotype" is
unfamiliar, so that this exercise can never be used to "test" the
ability to create (and therefore to recognise) basic social forms. It
can, in fact, be shown that every response points to a stand that has
been taken and that this stand expresses each participant's
representation of the category but also his own social position and
therefore the way he relates to the dominant representation of the
category. This can be seen particularly clearly through an analysis
of the structure of the three examples of cadres that the participants
were asked to provide. At the beginning of the exercise, these were
written down in the arder they carne to mind (the arder of writing),
and at the end of the exercise participants were instructed to rearrange them, if necessary, according to how exemplary they felt
them to be (the arder of exemplarity). When we asked participants
for these three examples, we gave them the opportunity both to
provide illustrations of the dominant representation of cadres and
to include less paradigmatic examples in their lists - in other
words, by suggesting that there might exist a system of differences
within the category itself , we were asking them to outline the configurations of a space. David, an unemployed, junior sales executive (petit cadre commercial) was one of the participants who
took advantage of this opportunity, and showed his familiarity
with the category by inventing three examples which were very different, though ali were taken from the prvate sector. The first was
a sales manager, aged 35, a non-graduate (autodidacte) who earned
6000 F. a month; the second was a mining engineer aged 40 who
was a graduate of an Eco/e d'ingnieurs* (though not a Grande
Eco/e) (petit diplome d'ingnieur), who earned 8000 F. a month;
Dow nloaded from http://ss i.sagepub.com at La Trobe University on August 3, 2009
Thorie et mthodes
653
the third was a woman of 30 who ran an advertising agency, a

graduate of HEC* who earned 12 000 F. a month. The order of exemplarity and the order of writing were in this case the same: the
cadre autodidacte was presented as more "typical" than the HEC
graduate. But the strategies at work in the production of these examples become even clearer when the most exemplary cases are not
written down first.
This was what happened with Marcelle, a social worker team leader, daughter of
a manual worker ( ouvrier) and married to a bank executive (cadre de banque).
(The results of this session are not included in the table.) She first wrote down the
example of an oil executive with a university degree (diplom de l'Universit* )
who worked for a big company and earned 20 000 F. a month. Then carne a
woman executive officer in the Department of Social Security whose highest
qualification was the baccalaurat* and who earned 8000 F. a month. Finally, a
metallurgist with a university degree (diplom de l'Universit) who earned
15 000 F. a month.
The second example, which was of a woman who had the properties Marcelle
herself possessed and who was least similar to the dominant representation of the
"cadre" (because she worked in the public sector, was female, had no higher
education and was not well paid), was stated to be the most exemplary, whereas
the first example, though it closely resembled the perfect "cadre" in social type,
was relegated to third place. The social workers participating in the session frequently adopted the strategy of putting public rather than private sector cadres
into first place (even though they did not immediately think of doing so), and of
relegating the most highly paid or commercially oriented to third position. No
doubt this represents a moral stand which should be linked to the
"disinterestedness" to which most social workers, teachers, and indeed public
sector workers in general attach a great deal of importance .
In this last case "exemplary" has a double meaning. The word

<loes not just refer to the paradigmatic cases, but also cases worthy
of being proposed as examples. Once again, this is an illustration of
the differences between the processes of classification applied to
objects which are socially neutral, as with the shapes and colours
studied by Rosch, and processes of classification applied to social
groups which involve the social position and the value system of the
person classifying. 11 In Gestalt language, one never asks whether
the "right" form is worthy of being "right" or whether it is
"right" to organize the world of forros around it.
654
Theory and methods
The struggles over social classifications

The social interests involved in cognitive operations applied to the
social world and its divisions (as opposed to operations on abstract
or socially neutral categories) are expressed particularly clearly in
the exercise in which a common nomenclature is negotiated on the
basis of the taxonomies produced by each team .
The exercise,is as follows: each team writes on a paperboard the
classification it has constructed in the course of the classifying
operation, indicating the occupation considered the most representative of each pile (and not the labels, which are recorded by the
observer but not used in the exercise). lt was in fact observed, after
a first attempt, that a negotiation procedure based on the names of
social groups hinders expression of the ordinary modes of
categorizing social milieux. As was seen in the pile-naming phase
which concludes the classifying exercise, attachment to the name
leads to a formal representation of the category. A debate about
names , with the formalistic arguments over the correct definition
which inevitably accompany it, makes it possible to avoid discussion of the individual cases on each card; this is more risky because
it forces each participant to indulge in specific individualized
judgements and to return to the personalized expression of everyday rivalries. In the same way , exchanges of arguments between
occupational representatives, which are occasioned by the
establishment of a new nomenclature of the occupations - of
which the exercise described here is in many ways an in vitro
reproduction - employ the jurdica! or quasi-juridical procedures
of instituting the correct criterion of definition and choosing the
true designation. They thus fail to cast light on the difficulties
which arise from actual delimitation of the categories, which the
exercise is designed to demonstrate, just as it is only in the final
clauses of laws governing certain occupations that one finds the
conditions of approved membership, i.e. l:he relation between the
name of the category, its definition and the population of individuals who constitute it.
For these reasons, we drew up a procedure for combining the
various nomenclatures constructed by the teams which is mainly
based on the cases chosen as representative of the different
headings. Starting from the shortest nomenclature, represented by
the list of typical examples, the participants proceeded by comparison with the more detailed nomenclature s, either assimilating
one case to another or adding a new heading in the form of an addiDow nloaded from http://ss i.sagepub.com at La Trobe University on August 3, 2009
Thorie et mthodes
655
tional typical example. The instruction was to arrive at a common

classification, a single nomenclature, which would win the agreement of ali the participants while conserving as much as possible of
the information which each team had wanted to put into the
classification it had constructed. Once again, the instruction is contradictory, since a common nomenclature can obviously not maintain ali the distinctions contained in each team's classification.
"Negotiation" therefore quickly leads to conflicts, which are proportionately sharper in the socially less homogeneous groups. The
participants tend in fact to want to make a larger number of subtle
distinctions in the areas of social space to which they themselves
belong, or to which they feel close, and to suggest fusions and
assimilations in areas from which they are remote.
Let us take the example of the session involving the INSEE interviewers, a mainly female, socially heterogeneous population including both women graduates of
high social origin, whose husbands have bourgeois occupations, and women with
low qualifications whose origins and whose husbands' occupations place them in
the lower fractions of the middle classes or the working class. Almost every
example of an occupation that was presented gave rise to a conflict which is , in a
euphemistic and tacit way, a class conflict, in which the participants who
originate from or belong to more bourgeois milieux clash with those of workingclass or petty bourgeois origins . What is to be done, for example, with the
"cleaning lady" (card no. 60) whom Rene, a graduate, the daughter of a
wealthy farmer, married to a psychiatrist, wants to assimilate to the "female
packer" (card no . 38)? In the classification exercise, Rene has already
manifested her tendency to lump ali working-class positions together indiscriminatel y , by making a large pile (20 cards, including "farm labourer",
"chambermaid", "storekeeper", "pastry-cook", "wood-cutter", "railway
employee'', "clerical worker" , etc.) which she defined, for the moment, as
"miscellaneous" (le tout venant) and designated at the end of the exercise under
the euphemistic heading "wage-earners, unskilled worker s".
Martine (BEPC, daughter of an office worker, married to an offset printing
machine operator): "! don't agree . . .Chambermaid and female factory
worker . . . lt's not the same background, it's not the same way of life."
Rene: "Ali right, but in the end it doesn't make much difference ."
Martine: "! dunno . . . l'm trying to follow you . . . Ali the same, they're two different life-styles, the factory girl gets dirty , she works much harder than a
chambermaid working in someone's hou se."
Rene : "A domestic servant doesn't sit around ali day. 1 think they can go
together."
Martine: "I think it's not the same sort of life at ali. Working in a factory and
working as somebody' s cleaning lady or chambermaid isn't the same sort of
thing at ali. Now, we put chambermaid with cleaning lady ."
Rene : "What they ha ve in common, is neither of them needs any qualification s,
that's an important factor, after ali."
656
Theory and methods
This is only one of rriany arguments of this sort, in which the participants clash over seemingly "technical" points (such as the
relative value of two diplomas) in the course of sometimes interminable exchanges in which what is really, but always tacitly, at
stake is nothing other than the representation of the social world
and of its division into groups and classes. In these debates, the
participants sometimes speak as if they were the representatives or
spokespersons of a social group, appointed to defend its specificity,
identity and interests in political or union bargaining. At other
times, their remarks suggest another language, that of the authorized specialists on the social world, the language of "sociologists".
This is especially the case when the struggle involves not the position of a particular occupation within the taxonomy, but the definition of social groups and milieux, i.e. the very meaning of the task
they are engaged in. One then sees the emergence of a system of
antagonistic positions in which the participants speak as a function
of the dispositions and properties of habitus (cf. Bourdieu, 1979)
which they derive from their class origin and class position.
For example, in the session involving the INSEE interviewers, there is the case of
Denise, who has followed a somewhat downward social trajectory (the daughter
of a retired naval officer who subsequently became a painter, she has a typing
certificate and is married to an autodidacte sales executive). Her discourse is based
on the opposition between, on the one hand, "occupation" or "situation" and,
on the other hand "milieu", or, to put it another way, between "class" as defined by reference to a determnate position occupied in the socio-occupational
structure and "class" in the sense in which someone is said to "have class",
defined by reference to another space, that of family relationships in which the
name is the main qualifying label. This definition indicates attachment to an
earlier state of the representation of social differences, characterized (to put it
schematically) by less State intervention in the principie of qualification. This
distinction corresponds to different degrees of standardization and anonymization of the properties qualified, ranging from the entitlement criterion which
constitutes "estate" or function in professions with regulated access, to personal
features such as manners or tastes. (Aristocratic families which had come down
in the world and lost their money could thus dissociate the properties a person
owed to his origins, objectified in the aristocratic title, and the sometimes
demeaning properties of the "bread-and-butter" activity with which he earned
his living.) But, in her efforts to illustrate the permanence of "class", Denise
encounters the scepticism of Rene, who, being a bourgeois by origin and by
marriage, is less inclined to separate "occupation" and "milieu" (and who
refers to statistics as a way of "scoring a point" off Denise).
Denise: "From the point of view of their work situation, you can't put the nurse
and the doctor on the same level. But from the social point of view, the nurse
may be from a better background (mi/ieu) even than the doctor. That doesn't
Thorie et mthodes
657
mean anything at ali. You find ancient families with fancy long names and
they're bank clerks. But they wouldn't so much as have the bank manager to
dinner, because ...You see, it's a whole question of breeding (ducation), that's
what background is."
Rene: "You may get a nurse with more breeding than a doctor . . .But in the end
(speaking to the observers), your categories are meant for making samples, that's
what it's ali about. There are nurses from very modest backgrounds. She (referring to Denise) turns everything upside down. You may find a nurse who is better
brought up than a doctor. Of course that can happen. But if you take two thousand doctors and two thousand nurses .. ."
Denise: "But you can't identify that milieu. You know what social milieu to
classify someone in once you (i.e. herself, as an interviewer) have been in their
flat and had contact with them. But you mustn't confuse breeding with situations. They are two quite different things. What is social milieu? Social milieu is
breeding . . .You see, when you meet someone, you know at once. The way they
open the door, that already tells you something, the way they talk . . ."
Rene: "Whether or not they kiss your hand" (laughter).
At this point in the discussion, Josiane, who is of working-class origin, unqualified, and married to a railway inspector, intervenes to back up Rene. But
Rene refuses this support, precisely because it comes from a woman she regards
as socially inferior; changing her strategy and her argument, she shifts alliances
and aligns herself, at least in part, with Denise.
Josiane: "lf 1 may say so, 1 think the situation is a good reflection of the individual."
Denise: "Not at ali."
Rene: "lt doesn't reflect breeding at ali. You find flats with thirtythousand
francs' worth of furniture in the room and it may be in bad taste ...A flat
reflects the money someone earns, 1 agree, but it may be in dreadful taste."
Josiane: (her voice expresses annoyance): "I wasn't talking about breeding."
As the exercise proceeds and (as in the discussions after meals)

tension rises, increasing the chances of arguments and squabbles,
the participants have to be careful that what they say is not identified with their own situation or taken as an ad hominem attack or
innuendo. To calm the tension, they have to escape into generalizations and this leads them to draw on the available forms of
representation of the social world and to reinvent the principal
social philosophies, which they gravitate to in accordance with their
social trajectories and the properties of their habitus.
To take just one example, during the session involving sales representatives
from a large multinational firm, an opposition arises between on the one hand,
Pierre and Marc, whose argument is implicitly based on the idea that there exists
(to put it schematically) a "social reproduction" and, therefore, "social classes"
and, on the other hand, Robert and Grard, who deny the idea of "social rigidity" and put forward a "fluid", "meritocratic" representation of the social
Downloaded from http://ssi.sagepub.com at La Trabe University on August 3,2009
658 Theory
Theory and
and methods
methods
658
Bo/tanskiand
and Thvenot
Thvenot
Boltanski
order. This is no doubt linked to their social trajector y (the sons of manual
workers, they see their recently acquired "cadre" status as significant mobility)
and also, no doubt, to their image of their future (they are younger than Pierre
and Marc and have subjectively greater chances of continuing to rise).
Pierre: "When you state a qualification and an occupation you're also impl ying
the parents' income and with it a certain comfort and a certain class, it ali hangs
together. If, you put 'father, roadsweeper' and 'mother, cleaning lady' ,
there'll be a social milieu that , er. . . "
Robert: "That doesn't mean anything, Pierre . The parents may be very workingclass and the children very academic."
Pierre: "I agree . But, to begin with, it's clear that if the father is bringing in a lot
of money, that affects the children's upbringing, living standards, the whole way
of life."
Robert: "I don't think that's true any more. lt was still true a while ago, but now
it's much more a linear system . . . "
Grard intervenes later in the discussion to argue that people with different
education levels "can meet each other. They may take to each other or not,
that 's a question of personalities, but nothing stops them meeting each other."
Pierre: "Ali of us around this table have a family inheritance which affects our
culture, even our political reactions or whatever, and you can 't deny that a large
part of what we are can be traced back to our origin. So even if a man has reached a certain academic leve!, in his childhood he was immersed in a climate which
perhaps wasn't favourable to certain arts or certain types of music. "
A discussion follows, around the case of the "sales executive", as to the relative
chances of promotion of "autodidacts" and "graduates" .
Marc: "One of them (the autodidact) ends up more or less where the other (the
graduate) begins. He (the graduate) can manipulate knowedge and a way of
behaving which aren't necessarily part of the salesman's (i.e. the autodidact's)
repertoire ."
The negotiation is balanced unstably between the temptation to

use the task of classification as a way of settling personal seores,
and an escape into generalities. Thus, as the conflicts become more
intense and the situation more strained, there often emerges from
the group (as in meetings of specialists) a "methodologist", who
takes advantage of the collective disarray to seize authority within
the group. The other participants, if only by their silence, help to
assign him this leadership role, because he reduces an unpleasant
tension. He <loes so by pushing the logic of the collective work of
generalization to its extreme, inasmuch as he substitutes an
epistemological discourse on the method of classification for
categorizing operations which are always Iiable to get bogged down
in particular cases. He indicates the steps that must be followed in
order for the discussion to "move forward rationally", setting out
a procedure that almost invariably refers to the logic of the classes
and the definition of the "criteria" of classification.
Thorie et mthodes
659
This is how we should understand the remarks made, again during the sales
representatives' session, by Marc and then, a little later, by Grard (both of
whom have sorne experience of higher education, which gives them the necessary
authority over the group to use the language of method).
Marc: "What are we looking for? We're enumerating differences, enumerating
parameters, but we still have no objective . What are we trying to pin down?"
Later, Grard proposes a "method":
Grard: "Would it be possible to list somewhere the criteria which led each of us
to put people a priori in this or that category? We put the lawyer in such-andsuch a category for such-and-such a reason. What are the points of convergence
and what are the points of divergence?"
But the positions adopted by these amateur sociologists do not

have the same rigidity as those taken by specialists in scientific or
political representation of the social world (sociologists, tradeunionists, politicians, etc.). Unlike the professionals, they do not
have a language strong enough to pitch itself at a general leve!, so
that their discourse is constantly in danger of falling back into
individual opinion. Finally, not being set in a field of institu
tionalized positions, they are not required to maintain unchanging
statements of position. The positions stated and the alliances they
allow thus vary as a function of less objectified variables, linked to
the structure of the group itself ("likes" and "dislikes",
"alliances" and "oppositions" between individuals based on a
previous shared history, etc.), the effect of which no doubt works
in different ways in the course of the exercise, depending on the
situation , which may be exciting or boring, relaxed or tense, etc.
ldentification and calculation of indices

The final exercise, even more than the previous ones, takes the
form of a "parlour game" (jeu de socit) on "society", since the
participants compete to guess information which is hidden from
them, and there are one or more winners. The game is played three
times. Each "round" followed by a discussion in which each participant in turn explains the tactics he used (this sequence is taperecorded each time). To play the game, the participants need to
mobilize their "social sense", or "sense of the social structure'',
i.e. their tacit knowledge of the social world, its differences and
divisions, its groups and classes, the social markers, emblems,
guides to classification used implicitly in the ordinary course of
social life. 12
Downloaded from http://ss i.sagepub.com at La Trobe University on August 3, 2009
660
Theory and methods
The game consists of assembling sufficient information about a

real person, who is unknown to the participants, to be able to
discover the social milieu he or she belongs to and, more precisely,
bis or her occupation. The information that can be gleaned about
this person consists of the answers given to a questionnaire of forty
questions. Each participant is given a pack of cards, on which these
questions, but not the answers, are written. Only the umpire has
cards also containing the answers supplied by the hidden mystery
person. Each participant can thus only obtain the information he
needs by asking the umpire. To do this, he has to purchase the
answer. Each answer has a price (marked on the ,card), ranging
from 1 franc to 40 francs, and each participant has just 100 francs
to play with. The answers are sold individually and secretly to each
player by the umpire, who charges each purchase to the player's
account. When a player thinks he has found the solution, he writes
it on a piece of paper and hands it in an envelope to the umpire,
who doses bis account. On the piece of paper, is indicated the
social milieu identified in the classification agreed by the group,
and the more precise occupation. The winner is the player who
finds the exact occupation while spending as little money as possible. The questionnaire covers six "themes": (a) civil status (sex,
age, marital status . . .), (b) working life, (c) relationships (best
friends, relatives, spouse . . .), (d) standard of living and wealth,
(e) cultural activities, (f) ethical attitudes and political opinions.
Within each of these groups, the question differ considerably;
sorne call for answers of a general nature on official characteristics
of the individual, whereas others yield indices which have not been
subjected to institutional codification. Thus, legally guaranteed
properties such as educational qualifications can be distinguished
from prvate or optional properties which are precisely defined by
the fact that they do not form part of the official identification of
the person (such as "tastes and colours", which, as the French saying goes, "are not matters for argument"). The former are
amenable to statistical recording, defined by administrative codes,
or determined by law. Their use in an exercise which is similar in
form to a statistical survey therefore seems natural to the participants. By contrast, the second set of questions, which are
perceived as relating to the most prvate, individual, even physical
aspects of the person, strike them as out of place in this context, as
if they were not amenable to standardized questioning.
Sorne questions are more closed than others, sorne answers more
Thorie et mthodes
661
wordy than others. For example, a closed question can be used to

ask if the person belongs to a professional association (yes/no), or
to request a list of the associations he belongs to. In the first case,
the information is entirely contained in the formulation and
therefore depends on the mode of categorization which has been
adopted. In the second case, the information, which has not been
pre-coded, provides data about the person which are not reducible
to a homogeneous, pre-established system of categories. For example, in reply to the question "How did you spend your last summer
holidays?", which includes the item "at a hotel", one of the
unknown persons had provided the following answer, which was
duly transcribed onto the card supplied to those who bought the
question: "At home in Paris and at a hotel (with swimming-pool)
in the daytime". This answer drew the participants' attention in
a direction which was not that of the question asked. Simply by
being intriguing , the sentence was to become one of the salient
points around which the participants' effort at interpretation was
organized.
Sorne semi-open questions, while less intriguing, may also give
rise to interpretations which closed questions would not permit.
Thus, when Claude received, in response to the question on the
three best friends, the answer "first salesgirl", he commented: "It
must be someone who has had sorne promotion, otherwise they
would just have put 'salesgirl'." He thus constructed a piece of
information which had previously escaped everyone's attention,
and was the first to do so. He interpreted the answer. He gave an
objective intentionality, and hence a meaning, to a word, the
presence of a word, which "should not" have been there . This exposed him to the charge of "over-interpretation": "that word is
there by accident"; "the fact it's there has no particular
significance", etc.
The price structure, which charges heavily for the situational
variables most linked to occupation (e.g. qualifications, income,
nature of the organization, position in the hierarchy, etc.) set up a
handicap which was designed to show to what extent the participants would be able to discover the unknown person's occupation by using properties from private life (e.g. books read,
favourite records, holidays, hobbies, etc.)
One particularl y economical example will help to make clear the aptitudes used.
Jean first asks the sex (1 franc) of the mystery person. He is told that it is a
662
Theory and methods
woman. He then asks her age (1 franc) -she is 50 -and her marital status. On
the pre-ceded card given to him by the umpire the category "single" is ringed,
followed by the hand-written words "common-law wife" . He then asks if she
owns or rents her accommodation (1 franc). The answer is "she rents it" (his
comment: "She rents it? At the age of 50?"). He then buys the place of residence
(3 francs). The answer is: "Paris, 7th arrondissement". (He comments: "That's
interesting." Jean knows the area well; he has friends there and for a time he
worked ther: as a ladies' hairdresser. He knows it is a wealthy , "bourgeois"
district, is familiar with its many luxury shops, etc.). His next question is on her
three favourite TV programmes (3 francs): "Number one, serials and Les
Dossiers de l'cran* (comment: "lt's a concierge. A common-law wife, with a
rented flat, who watches programmes like that, and in the seventh! lt's got to be
a concierge.") Jean then asks the question about her car (3 francs): "Alfa
Romeo Sprint, 1979". He stops playing and formulates his answer, which he
hands to the umpire :
"She isn't a concierge, she must be a successful shopkeeper. Yes, she has to be a
shopkeeper to have tastes like that in TV programmes. A hairdresser, or a
delicatessen shop, yes, serials, Les Dossiers de l'cran and an Alfa Romeo, in
the seventh !"
J ean did not hit on the exact occupation of the unknown woman,
who is an independent jeweller running her own shop, but he got
close to it with an economy of means which makes his performance
an exemplary demonstration of successful strategy in this game.
Schematically, the qualities of his play can be described as follows:
(1) he plays very economically (spending only 13 francs), i.e. he
asks only for prvate and therefore inexpensive characteristics
(behaviour, consumption, etc.), which he uses as if they were signs
pointing to more general features of the person to be identified.
Thus, he does not buy her educational qualifications (which cost 40
francs) but clearly uses the information about favourite TV programmes as a substitute for this. Such signs cannot be sufficient to
infer qualities such as level of schooling, with which they are not
absolutely correlated, but combined with other indices, referring
indirectly to attributes of another type, they provide pointers to
situate the hidden person in a social milieu.
In the case analysed here, the player is mainly guided by twice
comparing indicators of "cultural capital" and
"economic
capital". He first compares the district, a sign of economic wealth,
with the tastes in television (indicators, in his eyes, of cultural
poverty). This first comparison initially suggests to him that the
person is a "concierge" (i.e. someone poor in a rich
neigh bourhood, "working-class" in a "bourgeois" neigh-
Thorie et mthodes
663
bourhood, etc.), because he uses the fact that she "still" rents
her flat and is a "common-law wife" as indications that she
belongs to the working class; this is congruent, in his view, with her
cultural tastes as manifested by her choice of TV programmes ,
which might have seemed incomprehensible in a "real" bourgeois
woman in such an area. The second comparison is effected when he
learns that her car is an Alfa Romeo, which for him is an
undeniable sign of wealth. He can no longer imagine this woman in
a working-class milieu. He then spontaneously moves towards
categories (shopkeepers, small employers . . .) which, according to
the statistical information on them (which is of course not available
to him, and which he may not even know exists), are precisely
characterized by a decrystallization between "economic wealth"
and "cultural wealth", between educational level and income level,
etc.
(2) He plays cumulatively: a new piece of information <loes not
lead him to abandon the knowledge he has previously built up
about the person . For example, the surprise caused by the make of
car <loes not make him start again from zero, forgetting everything
he knows about her district, age, marital status and cultural tastes.
(3) He plays without redundancies: by means of a single indicator he arrives at a set of dispositions which statistical analysis
shows to be strongly correlated. For example, having asked for her
favourite TV programmes, he <loes not ask for the number of
books she reads per year, the title of the last book she read or her
favourite records, which would considerably increase his expenditure.
(4) He plays flexibly and shows he is capable of modifying the
image he has formed of the hidden person in response to the information he received during the game (shifting from "concierge" to
"shopkeeper").
(5) He knows when to stop. He <loes not pile up information,
and expenditure, to confirm the image he has formed of the hidden
person. In other words, he acquires, at minimum cost, sufficient
certainty to stop the game and offer an answer.
(6) He is able to mobilize his practica! knowledge of the social
world, which is linked to his previous social experiences, and to
make it sufficiently explicit to be able to use it as a means of
decoding the fragments of information he has assembled .
When the indices are perceived in extended order, as is often the
case in the first phase of the game, they are mostly regarded as
664
Theory and methods
uncertain , perhaps irrelevant, an unreliable basis for predictions

that might have a reasonable chance of success. This is the case, for
example, with the make and model of car. Frequently, at the beginning of the exercise, when the participants read the cards showing
questions they can choose from, they are heard to comment: "The
car doesn't tell you anything" (i.e., anybody can own any car or, in
the language of statistical regularities, the distribution of cars by
make and model is statistically independent of the distribution of
properties in the socio-occupational structure). Again, sometimes
the comment is: "Nowadays, that doesn't mean anything." This
statement of :mcertanty expresses a practica! experience of the
struggles between groups and classes to appropriate the signs of
distinction (cf. Bourdieu, 1979). Another remark is frequently
made about the question of holidays : "Winter holidays means
nothing nowadays. Everyone does it - what you need to know is
how and where." Or again, about a mystery person whose
favourite sport is sailing: "Anybody can go sailing, but is it a little
yacht or a big one, on a lake or the sea?"
As they play, the participants learn (sorne much better than
others, although in the present state of research we cannot relate
these differences to social properties) to increase the value of the indices by considering them not separately, as happens when the
series of questions is read for the first time, but with mental crossreferences (e.g. "rented accommodation" yes, but "at age 50"), as
if they were managing to mobilize their tacit knowledge of the
social world in the form of a system of conditional probabilities (cf.
Smedslund, 1955; Bruner et al., 1962, esp. Ch. 7).
Attributing signs and recognizing forms

Opening gambits
When the game begins, the players know nothing of the social identity of the mystery person, who may, with equal probability, belong
to any milieu. lf the questions did not have a price-tag, the players
would spontaneously ask for information which has a clear, institutional relationship with the occupation (as in the case of occupations which can only be entered with specific educational qualifications) or a quasi-substantial relatonship (e.g. the question on the
"official name of the establishment" or on the "first job you
Thorie et mthodes
665
had"). The differential price-structure is, as we have seen, precisely

intended to encourage the players to use other sorts of indices. The
early choices therefore depend to a large extent on the players' willingness to shift the preferences from the most expensive to the
cheapest questions, which is itself a function of their attention to
these indices .
Midd /e game
There is sorne evidence that once the player has acquired sorne
information, of whatever sort, his subsequent operations are not
reducible to a process of analysis and induction (like the procedure
of the perfect sleuth in detective stories) 13 which would take account, in a rational way, of the costs and benefits of the information. It would seem that, in a large number of cases at least, the
first items of information produce in the player's mind, implicitly
and almost unconsciously (though sorne players mention it explicitly when describing their procedure), a social form and often, no
doubt, the substantial, quasi-physical form, of someone they have
known. It also seems that these forms, which may be evoked by a
minimal index, are emotionally coloured and, to be more precise,
negatively so.
The subsequent operations are then chiefly designed to confirm
this form, the intensity of which has a considerable effect on the
success or failure of the endeavour. When the ,form is too vivid, the
player cannot break free of it, with the result that the new information he acquires is no longer used, or is reinterpreted to fit the
previous hypotheses. The player is then incapable of modifying his
conception of the hidden person . 14
Endgame
However, the constitution of a vivid social form and a precise image of the hidden person (of whom the participants are occasionally
even able to give a physical description which is sometimes fairly
accurate) does not necessarily lead the players to abandon their
investigation, stop the game and present their solution. They often
continue to buy information (which ought to increase their chances
of getting the "right answer", but reduces their chances of winning
666
Theory and methods
the game, because they are spending more), but they only use this
information to confirm their choice.
This final phase resembles a process of stabilization in which
what has to be consolidated is not so much the image itself as the
player's belief in the validity of the image which he is about to communicate to the umpire . To confirm its validity, but also to reduce
the anxiety involved in risking a personal interpretation which
might expose him to ridicule, the player is tempted to back up his
interpretation with a reasoned explanation based on official
variables (qualifications, income, etc). The price-structure of the
game penalizes such hesitations and makes it possible to evaluate
the price that is set on the explanations legitimated by law or even
statistical law as against intuitive interpretation based on hunches.
An example will make clearer what is meant by "vivid social
form". The information provided on one of the mystery persons,
the jeweller, contained sorne very discreet indications of gender.
Although not asked for in the questionnaire, these gender-relevant
indices appeared in the form of additions to two questions. In
answer to the question "marital status", the "single" box was ticked, but next to it the words "en concubinage"* were hand-written.
The second addition occurred in answer to the multiple-choice
question on holidays: "at home in Pars and at a hotel (with
swimming-pool) in the daytime". In addition, secondary indices
pointed to "femininity" defined as "frivolity". For example, the
"jeweller" reads no daily paper and only one weekly magazine,
Jours de France, which has a "feminine" image (fashion, etc.) and
which the players frequently associated with "waiting in a hairdressing salon".
These indices explain why the mystery person was almost always
identified, after the first few questions, as a "woman" (when the
question of her sex had not been asked), and as a "young woman"
in a "female" occupation, often as a "salesgirl" ( vendeuse), or,
even more precisely, as a female vendor of futile or luxury goods,
as if the players were guided in their search by the social image of
the ''midinette'' .*
These very general tendencies in interpretation are particularly
clear when the players are women, even more so in the case of
middle-aged and older women, in occupations which entail commitment to a value-system emphasizing "seriousness" (as opposed
to "frivolity") and which are socially defined as involving tasks of
ethical control, social control and moralization . In other words,
they are most common among the senior social workers and the
Thorie et mthodes
667
retired teachers, although they also manifested themselves in other

sessions in which older women from "controlling" professions
took part.
This can be briefly illustrated by two examples in which the hidden person was identified as a "model" or "cover-girl". In each
case, the oral explanation the players gave included a physical portrait of a "frivolous", "daring" young woman, excessively focused on her "body" and engaged in the business of "seduction". In
both cases, the form is so vividly present that the question of age
(though it costs only 1 franc) was not even asked, although it was
essential in order to confirm the validity of the portrait. The players
were therefore surprised to learn, at the end of the game, that the
person they were looking for was a 50-year-old lady shopkeeper, a
fact that obviously destroyed the form which they had constructed
or which, to be more precise, had taken possession of them.
Mireille is a retired primary school teacher, aged 63, the daughter of an officer.
She spent 42 francs and asked seven questions: the newspapers read, the occupation of the three best friends, the last summer holidays, the last book read, is the
home rented or owned, <loes the person speak English fluently, the make and
model of his/her car. Her answer for the occupation was "model, cover-girl".
This is her account, after the game, of the procedure she had followed:
"! could have spent even less and asked only three questions -the three questions I asked to begin with . That's to say, what newspapers or magazines do you
read. Answer: Jours de France: l thought it must be a woman . What are the jobs
of your three best friends: jewellery saleswoman, a nurse who runs an agency,
jewellery traveller. The third question was: Where do you spend your holidays.
The answer was: in Paris at home and at a hotel -in brackets -with a swimming pool in the daytime . With those three questions I formed a mental image of
that girl. I saw her as young, very average in level (she presumably means intellectually very average), very pre-occupied with the luxury goods business, with
a nurse among her friends who could help her with her beauty care, be her
beautician, not afraid to show herself off. After al!, I didn't imagine she went to
the swimming pool to admire the view or relax. I assumed she went there to
swim, perhaps even . . . (one of the listeners: to pick up men?) Yes, to pick up
men. So there you are. I built up a picture of a girl like that. lt was quite clear to
me, so I thought, why carry on. Ali the same I hesitated. I asked what book she
was reading: Kramer versus Kramer. That fitted, too. No problem. She rented
her flat, she spoke English only with difficulty, and she drove an Alfa Romeo, so
the car was right, too. So I said to myself, there you are, she's a cover-girl. A girl
with a good figure and lots of energy. A bit daring, brazen even, early twenties. I
had a very clear picture of her like that."
Like Mireille, Josette, a woman of 55, saw the mystery person she had to identify not as a shopkeeper but as a "model" ( mannequin). She is an officer's
668
Theory and methods
daughter (like Mireille) and a graduate of HEC; for a long time she ran a temporary employment agency. She is now unemployed. She spent 31 francs and asked five questions: about the person's views on strikes ("Are there too many
strikes in the public sector?"), the newspapers or magazines read, the last summer holidays, the make and model of car, the occupation of the three best
friends (four of these questions -newspapers, friends, holidays and car -were
also asked by Mireille). This is the commentary she gave after the game: "I took
the question on strikes. l started out with a pre-conceived idea. l said to myself,
since we've had a senior executive, now they're going to go to the other end of
the scale (i.e. present us with a manual worker). When l saw the person was
against strikes, l switched to the cultural leve) and asked about newspapers . That
really had me confused . . . ! l immediately decided it was a woman and one with
a rather low intellectual leve!. l can't imagine a man with a mentality that would
be interested in a rag like Jours de France, not even a manual worker with no
culture. When that is ali you read, really, it's very limited. Next l asked about her
holidays, thinking that would enlighten me. And once again 1 couldn't make
head or tail of it. l said to myself, she's a . . . well . . . spending the
daytime . . . spending her holiday in a hotel, with swimming pool! There are other
swimming pools, after ali! You'd really have to want to move in those circles.
It's got to be someone with no money trying to meet certain sorts of people. Trying to show off. That's how l saw her. Then 1 got an answer that confirmed me
in that view, the fact that she has a 1979 Alfa Romeo Sprint. l said to myself,
she's a kept woman. After ali, an Alfa Romeo is a really flashy sort of car, the
car of someone who moves in luxury circles perhaps without having the means of
living up to it. Spending her holidays . . . ! lf you have the means to do it, then
you take your holidays, l don't know, in Greece! Then l asked questions to try
and pin down her elusive personality. l asked about her friends. At that point she
rather went up in my esteem, because ali the same, l said to myself, she has
friends who do sorne work. lt's not entirely what 1 thought. When l looked at the
friends, l didn't pay so much attention to the commercial side, if you see what 1
mean, the business angle, l saw more the luxury side of jewellery, and l deduced,
it was a sort of intuition, she must be a model. Someone who moves in . . . Who
needs to keep up an appearance, not just well-groomed but luxurious, without
necessarily having the means to live luxuriously. l saw her as fairly young, thirtyish perhaps."
Here we see at work sorne of the processes which guide the work
of identification. 15 Once the form is fixed ("midinette" , "little
woman", "model", etc.), the information is re-interpreted in
terms of that form by means of a transformation or an - often
very free - "translation". (Thus, in the first example, the friend
who is a nurse becomes a "beautician"; in the second, the friends'
occupations suggest not "business", "trade", but "luxury".)
Again, in the first example we see very clearly how, once the form is
vividly established, information is used not to rectify the sketch but
to confirm the player's belief in the accuracy of the portrait she has
Thorie et mthodes
669
drawn : Mireille enumerates properties (the last book read, rented

accommodation, etc.), merely commenting "that was fine" , "that
fitted", without feeling the need to make explicit the relationship
between this new information and the social form she has constructed by means of the previous information . She is unable to say
how, in what respect, this information confirms the image she has
built up. For her, it is self-evident.
The penchant for interpretation

If the competence required by the game were only of the order of
knowledge, i.e. either native knowledge of the social world and
previous social experience or theoretical knowledge, then variables
such as age, membership of the milieu to which the hidden person
also belongs, or educational level, might be expected to have
positive effects on performance in the game. In fact we find that
the executives of the marketing department of a large nationalized
firm did not achieve high seores despite being, for the most part ,
Grande Eco/e graduates and belonging to the same milieu as the
mystery person (a senior executive in industry). Similarly, in a
group of female retired teachers, whose educational level is high
and homogeneous, the players had to spend a lot to discover this
individual's hidden occupation .
A first analysis of the relationship between the amount spent , the
number of questions asked and the accuracy of the solutions offered suggests that the ability to play well depends not only on the
knowledge accumulated about the social world but also on more
general aptitudes for interpretation which no doubt depend at least
in part on the agents' capacity to mobilize and use particular signs
of social identity.
There are several ways of playing badly (which will be analysed
in a subsequent article). One of the most frequently observed ways
of refusing the type of interpretative work demanded by the game
leads to the asking of a small number of expensive questions ("official" variables). It is particularly common among players who in
their professional life occupy strong positions of responsibility or
authority (even subordinate authority). The professional role of
these institutionally-located men (or women) in fact requires them
to base their working relations, particularl y with their subordinates, on the qualification s acknowledged by the institutions
Thorie et mthodes
671
employing them, which are written into the rules of management

and have a prescriptive character. They are thus inclined to confer a
high predictive power on official definitions and not to take account - in the game as in life - of the particular indices used by
the best players.
Genevieve, one of the group of retired teachers, is aged 66, is the daughter of a
sales manager, and has the agrgation* in English. She spent 100 francs on only
'four questions and suggested, for the occupation, "cultivateur" (peasant
farmer). Instead of trying to use personal indices, she started with two questions,
at 40 francs each, which she hoped would give direct information as to the occupation: educational qualifications (answer: "none") and whether the person
concerned has any subordinates in his work (answer: "none"). She then asked
what newspapers or magazines were read (answer: no daily paper, one weekly,
Jours de France, no monthly magazines) and whether the person had ever been
unemployed (answer: "no"). The absence of qualifications and subordinates
directed Genevieve towards "the working classes" (she was clearly thinking of
an industrial manual worker). Learning the person had never been unemployed,
she slightly modified the image she had formed and, having used up ali her
credit, fixed on the idea of the peasant, who, without being a "wage-earner" or
an "industrial worker", possesses the properties which characterize the "working classes" - at least when defined negatively, as the absence of bourgeois or
petty bourgeois attributes.
A second example: Josphine, aged 41, the daughter of an engineer, a
graduate of the Paris lnstitut des Sciences Politiques* , married to a judge. She
identified the first hidden person as a "managing director" (rather than as a
"senior executive"), and the second as a "traveller in jewellery" (rather than as
the proprietor of a jeweller's shop). These errors arise from focusing on certain
indices isolated from the whole and associated with a social type, very
schematically characterized by a few emblems. (In the first case, the hidden person is a "managing director" because one of the publications he reads is TimeMagazine and one of his hobbies is sailing; the second person is a female "commercial traveller" (reprsentante) because she has a powerful car - an Alfa
Romeo - and because "one of her best friends is a reprsentant" ).
The interesting feature about such mistakes, which occur frequently, is that they attribute to the hidden person an identity
which figures in his/her field of lateral possibles (the "jeweller",
who started her working life as a "sales-woman-demonstrator" no
doubt had objective chances of becoming a "travelling representative") or a vocational identity inscribed in his/her social tnljectory. This is clear in the case of the "senior executive", who is
often identified either as a "research scientist" or as a "managing
director''. He is in fact a chemical engineer, a graduate of the Eco/e
Centra/e*, has an American Ph.D., and did indeed begin his career
670
Theory and methods
at the State research institute, CNRS. He now runs the research

laboratory of a large firm, is increasingly concerned with management and sales and hopes, at the peak of his career, to manage one
of the subsidiaries of the group he works for.
In the third "round", in which the mystery person was a 56-year-old semiskilled worker, Josphine was much wider of the mark since (at a cost of 48
francs) she thought she recognized a "(female) medica! social worker for the
mentally handicapped". She first asked the question on newspapers and
magazines and decided the unknown person was a woman because he/she read
no daily paper but listed: "La Vie catholique, Femme d'aujourd'hui, Tl 7
jours, Historia" (in fact it is the man's wife who buys them). Josphine was so
sure of the person's sex that she did not ask the question to find out (although it
would only have cost her 1 franc). She then asked if the person belonged to any
associations (a question which cost 25 francs and was rarely asked). The answer
was: "residents' association and a society aiding the handicapped". "At that
point," she says, "I immediately pitched into the medica! or paramedical area."
Convinced that she was on the trail of a woman, and a woman with a "social" or
"para-medica!" job, she was incapable of using the other information she
acquired in order to modify that first impression. She asked for the hobby:
"D.I.Y ." -"That threw me completely." Then, the three best friends ("dental
mechanic, tailor and salesman"): "lt was getting more and more confusing.
Eventually l decided it was a woman working with the handicapped, a medica!
social worker for the mentally handicapped."
Here the mistake is based on over-interpretation of certain indices, on which Josphine concentrates, and which precisely correspond to the properties whereby the hidden person deviates from the
stereotype of the "manual worker". An ex-peasant farmer,
devoutly Catholic, married to a clerical worker whose cultural leve!
is higher than his own (thus, asked how many books he reads a
year, he replies "none, but my wife <loes"), this semi-skilled
worker presents, in his social properties, "feminine" traits and indices of cultural and associative practice which do not correspond
to the social image of the "manual worker", defined by a set of
"virile" attributes such as physical strength, trade-union or
poltica! radicalism, team (and/or "rough") sports, etc.
The best results are scored by the groups whose members have
had, for defensive purposes in the course of an unstable working
life, to make the ironic use of interpretation which the game
demands. Just as the "official" variable, the variable of the strong,
serves for prescription, so interpretation with the aid of personal
indices can be used as a weapon to break out of the system of
offidal constraints, since it enables one to give a meaning to the ar-
672
Theory and methods
bitrary behaviour of the "bosses" by revealing its hidden intention,

i.e. by constructing a model which can be used to help control the
relationship. These individual exposure operations are linked to the
adoption of a deterministic, reductive view of the social world. This
world-view, which tends to explain great effects in terms of small
causes, the overt in terms of the hidden, the coherence of
behaviours in terms of the cohesion of a secret alliance, is opposed
both to the official representation of a world ordered in accordance
with transparent, stable principies and to the atomized image of a
social world without regularities, composed of free individuals and
based on the frec-will of the person, which is often associated with
occupation of a position of irresponsibility within a dominant
group (young people or non-working women in the bourgeoisie,
etc.). It is linked to a particular experience of the social world that of dominated private life, of participation in the world of
those whose power and controls one is subjected to. It is often
experienced by women (wives, secretaries, etc.) and always has
something feminine about it even when experienced by men (for example, no such dispositions were found in a group of manual
workers, although they had suffered the consequences of
unemployment and dclassement) It is the type of experience
familiar to the female interviewers (working on short-term contracts, at piece-rates, and often having a second job) and, even
more, to the unemployed men being re-trained (often self-taught
petits cadres), whose common feature is that their professional trajectory has been a "bumpy" one, in which rapid promotion has
alternated with sudden loss of status. Finally, within each of the
other groups, the individuals who achieve the best seores are
generally set apart from the other participants by their greater
social fragility. This may be related to a stigma, to career setbacks,
or (as in the case of one member of the salesman group), to a personal difficulty in carrying off the virile self-image appropriate to
the values of the profession. In all these cases, the best player has
no doubt been led by his relative weakness to develop the cunning
(as opposed to strength (cf. Detienne and Vernant, 1974)) which
makes him a "crafty devil".
Official classifications
The homogenization of the system of professional titles, of the
Downloaded from http://ssi.sagepub .com at La Trabe University on August 3, 2009
Thorie et mthodes
673
names of occupations and, more generally, of social classifications,

and above ali the creation of official spaces for the representation
of occupational groups (whether real ones, like the corporate
bodies (Conseils), or symbolic ones, like nomenclatures), belong to
the series of unifying processes (linguistic, educational, legal, etc.)
linked to the formation of the State. Just as the law is "presumed
to be known by ali", so nowadays, in France, everyone is sufficiently acquainted with the official system of occupational-group
representation to be able to use it, whether to reconstruct it from
memory, to use it to perform classifications, to argue with other
people about its validity, or, when the occasion arises, to situate
himself within it. This is remarkable in itself, especially since
neither the system, nor the sets of terms in which it is expressed, has
so far been the object of explicit transmission in schools. It can be
shown 16 that this relatively close correspondence between the most
individualized images of social groups and the most general
representations of the social structure results from a process of diffuse inculcation which has accompanied the progressive formation
in France, since the l 930s, of a mode of political management
which, de facto if not officially, takes account of the demands of
corporatist representation. Since the War (to put it schematically)
this mode of government has been characterized by the establishment of new institutions which bring together (e.g. in joint consultative committees, or state planning commissions) senior civil
servants, experts, employers' organization representatives, tradeunion spokesmen, personalities from the "voluntary sector" (le
mouvement associatif), etc. This new mode of political representation (whose origins could no doubt be traced back both to the
socialist tradition of the labour movement and to the tradition of
"social Catholicism" and the corporatism which inspired middleclass organization in the later inter-war period) substitutes for the
unqualified citizen of "republican reason", as Claude Nicolet
(1982, p. 365) puts it, a quaUfied man defined not only by the job
he <loes but also by his position on a scale of skills (qua/ifications)
and, increasingly, by his place in the distribution of educational
qualifications ( litres).
The prescriptive power of State classifications should not,
however, be overestimated. They are not as omnipotent as is
sometimes assumed in accounts of "l:mreaucracy" which see no
other alternative, in the face of "Leviathan", than complete
passivity or pure rebellion . Individuals know and use official
674
Theory and methods
classifications, but always, at the same time, so as to play with

them, to play on them. As we shall try to show in a later article with
an analysis of the grammar of mistakes, apparent misclassifications never stem simply from inadequate knowledge of the
nomenclatures. They are always simultaneously the expression of
the individual's relationship to the political representation of social
groups. Thus the versions of general systems of classifications
which indivduals provide, also manifest, precisely in their
singularities, the particular position of the individual and the position he takes vis-a-vis the generic instrument. In short, there is no
reason to think that State classifications are the operational instruments of social identification. For example, the ability to
reconstruct the generic nomenclatures seems to be independent of
the ability to discover a hidden person's occupation; not only is the
latter not directly related to knowledge of the official maps of
society, but it draws on a penchant for interpretation that is, as we
have seen, formed in opposition to the official representations.
Equally, however, there is no reason to suppose, in the present state
of research, that the penchant for interpretation necessarily
generates immediately social forms or, to put it another way, that it
is directly linked to the practical equivalent of a sociological competence. While it is true that economy of interpretation depends on
the ability to work with a repertoire of coherent (and, no doubt,
emotionally charged) forms, and not (or not solely) by combining a
system of discrete criteria, the hidden person under investigation
may, as the participants' accounts often suggest, be recognized ("I
see her . . . ", "I feel that she . . . ", "I can quite imagine him . . . ",
etc.) without being identified or, at least, without this identification
being directly expressible in social terms. It is the constraint of the
game that forces the partici pants to project the forms they have
constructed onto a social space (and this operation often seems to
require a specific effort subsequent to the interpretation proper),
but there is no reason to suppose that in another game, with other
types of constraints, they would not be able to use other projection spaces equally well or better - the space of personality types,
defined in psychological terms, for example.
The fact remains that the participants' frequently-attested ability
to project their interpretations onto the official maps of the social
world (as they do when they give the mystery person a position in
the nomenclature and attach to him the name of an occupation)
and to produce a discourse (comparable in sorne respects, to

"secondary elaboration"
attribution in social and
that state classifications
Schelling's (1960) sense.
places.
Thorie et mthodes
675
in psychoanalysis) which justifies the

sometimes quasi-sociological terms, shows
now play the role of salient points, in
They have acquired the power of common-
Luc Boltanski (born 1940) is Directeur d'Etudes a l'Ecole des Hautes Etudes en
Sciences Sociales and works at the School's Centre de Sociologie de l'Education
et de la Culture. He has published several books and numerous articles on a
variety of topics ranging from social class and education to comic strips and the
social uses of the automobile . One of the central concerns in his studies is to
reveal the mechanisms whereb y social structure is interiorised, a subject dealt
with in the article "Les usages sociaux du corps" (1971).
Author's address: CSEC, Maison des Sciences de l'Homme , 54 Bd. Raspai l ,
75270 Paris Cedex 06.
Laurent Thvenot (born 1949) is Administrateur at the Institut National de

Statistique et des Eludes Economiques. After work on social categorization and
on the establishment of a new official nomenclature of professions, he is now
studying "coding" behaviour (production of standards, rules and norms) in
society as a form of investment to be taken into account in socio-economic
analysis.
Author's address: !NSEE, 18 Bd. Adolphe Pinard, 75675 Pars Cedex 14.
Notes
l. Cf., for example , in the case of the formation , in France, of the "cadre"
category, the role of the symbolic work of definition and delimitation (in the 1930s),
to which academic sociology has made a major contribution, especially since the
1950s. (See Boltanski, 1982)
2. The material from the first ten sessions is now being processed. The taperecordings and game-sequences have already been transcribed. Systematic statistical
analysis will be done at a later stage, when the data have been coded.
3. Since this exercise was designed to generate typical forms, it was thought
. referable to put it in first place, before any work was done on making explicit the
social categories and their representations (cf. Desrosieres, Goy and Thevenot , 1983.
On the discover y of professional identities according to the social milieu of respondants, see Coxon and Jones, 1978).
4. The procedure that was followed drew on techniques used by anthropologists
676
Theory and methods
of the ethnomethodological school (see, for example, S. Tyler (ed .) 1969).

5. On the functioning of "practica) logic" in system s of representations of the
social world, see Bourdieu, 1980b.
6. The analysis was applied to nin-: groups , i.e. 120 individuals teamed in pairs .
7. See Goody (1979) for the effects of schematic (or cartographic) substantiation.
8. See, for example, Goodenough (1956) and Romney, D'Andrade (1964) .
Criticism of the componentiali st view was the starting-point for Rosch's analyses:
"The present trend in anthropology to study 'folk' classification has tended to emphasize methods, such as 'componential analysis', which seek to find the minimum
basic criterial l!ttributes by which folk use of the terms of a domain can be formally
ordered" (p. 142-143) and "psychological and linguistic research has tended to treat
categories (whether perceptual or semantic) as though they were internally unstructured - that is, as though they were composed of undifferentiated, equivalen! instances - and as though category boundaries were always 'well defined' (Neisser,
1967)." . . . whereas "It is the contention of the present paper that most 'real '
categories are highly structured internally and do not have well defined
boundaries;" (pp. 112, 113).
9. The task of giving "good examples" of members of various social categories
(in particular, examples of "cadres") has so far been performed by 135 people (in
groups of 12 to 15) which gives a sample of 400 "typical cadres" . Since more
material is yet to be collected, comparison will be limited to the answers provided by
four groups: a group of highly qualified executives working for a large nationalized
firm; a group of unemployed people undergoing re-training; a group of sales
executives , mostly without higher qualifications , working for a multinational firm;
and a group of retired women teachers (primary and secondary school) .
10. Since coding of the whole set of material is not yet completed, sorne prop erties of the examples of "cadres" provided by the participants in four sessions have
been processed manually.
11. Even in more established statistical nomenclatures, which flaunt the neutrality of the natural (Tolosan's activity categories: animal, vegetable, mineral) or the
measurable ("levels" of education, measured in years), one can point to similar instances of the taxonomist who enhances the representation of the positions which
surround him (senior civil servants) and diminishes that of private-sector executives
(cf. Desro sieres and Thvenot, 1979).
12. The material provided by this last exercise has not been statistically analysed
in its entirety. On the other hand ali the "moves " by the players and the recorded
commentaries on them have been processed, giving a corpus of sorne 500 cases. A
statistical analysis comparing the diff erent strategies and the properties of the
players (qualification, age, sex, occupation, social origin, etc .) will obviously be
required in order to test the hypotheses put forward here .
13. For an historical analysis of the mode of reasoning applied since the late I 9th
centur y in a new literary genre, the "detective story", see Messac (1929).
14. As George Kelly pointed out, when someone is asked to classify his friends
and acquaintances, he tends to group these people in pairs on the basis of their
overall similarity or dissimilarity with a third person . If the subject is asked to say in
what respect these persons are similar, he may reply, for example, that both remind
him of a striking third person in his circle (e.g. "Mary"). The attribute in relation to
which the categorization is performed is thus, Kelly says, "degree of Mary-ness"
Thorie et mthodes
677
(Kelly, 1955, p . 139).

15. The work of identification the participants have to perform is close in sorne
respects to ordinary perception of other people (see Heider, 1958) and, in other
respects, to "attribution" (see Jones et al., 1971). As Doise (1976, pp. 134-39) has
shown, these two dimensions are closely related to the process of categorization.
16. We shall endeavour to demonstrate this in a subsequent article .
References
Berlin, B.; Kay, P .

1969
Basic color terms: their universa/ity and evolution. Berkeley, Calif.,
University of California Press.
Boltanski, L.
1982
Les cadres . La formation d'un groupe sociale. Paris, Minuit.
Bourdieu, P.
1979
La distinction. Paris, Minuit.
1980a
"L'identit et la reprsentation: Elments pour une rflexion critique
sur l'ide de la rgion", Actes de la recherche en sciences sociales 35:
63-72.
J980b
Le sens pratique. Pars, Minuit.
Bruner, J.S.; Goodnow, J.J.; Austin, O.A.; Brown, R.
1962
A study of thinking. New York, Science Editions.
Bruner, J.S.; Olver, R.R.; Greenfield, P.M.
1966
Studies in cognitive growth. New York, Wiley.
Centlivre, P.
1979
"Groupes ethniques : de l'htrognit d'un concept aux ambiguits
de la reprsentation . L'exemple afghan", pp . 25-37, in: E. Ehlers
(ed .), Baarage zur Kulturgeographie des islamischen Orienta,
Marburg, Lahn.
Coxon, A .P.; Jones, C.L.
1978
The images of occupational prestige. New York, St Martin's Press .
Desrosieres, A.; Goy, A.; Thvenot, L.
1983
"L'identit sociale dans le travail statistique. La
nouvelle
nomenclature des professions et catgories socio-professionnelles",
Economie et statistique 152: 55-81.
Desrosieres, A.; Thvenot, L.
1979
"Les mots et les chiffres", Economie et Statistique 110: 149-65.
678
Theory and methods
Dtienn e, M.; V ern a nt, J.P .

1974
L es ruses de l'intlligence: La Melis chez les Grecs. Pars, Flammarion.
Doise, W.
1 976
L 'articulation
psychosociologique
et
les
relations
entre
groupes.
Brussels, A. de Boeck.
Goodenough, W.H.
1956
"Compon ential analysis and the scud y of meani ng", Language 32:
195-216.
Goody, J .
1978
La raison g raphique. Par s, Min uit.
Heider, F.
1 958
The psychology
of inlerpersonal relations. New Yor k , Wil ey .
Jon es, E.E .; Kanouse, H.H.; Kelley, R .; Nisbet t, E.

1971
A11ribu1ion: Perceivin g !he causes of behavior. Morristown, NJ,
General Learning Press.
Ka y, P.; McDaniel, C.
1978
"The lingui st ic signi ficance of th e m ea ni n gs of ba sic color rerms",
Language 54 (3): 610-643 .
Kelley, G.A.
1955
The psychology of personal construcls. New York , Norton.
Lenski, G.E.
1954
"Statu s cry stallization: A non vertical dimen son of social s1at us" ,
American Sociological R eview 19 (3): 405-413 .
Ma re, E.
1 980
Reconstruir e l'espoir. Par s, Seul.
Messac , R.
1929
Le "de1ec1ive novel " el l'inji'uence de la pense scienlij ique. Pars ,

Librairie ancienne Honor Champion.
Moscovici, S.
1961
La psychanalyse, son image el son public. Pars, PUF.
Nicolet, C.
1 982
L 'ide rpublicaine en Franc e. Pars, Gall mard.
Romme1v it, R.
1 960
Selec1ivi1y, iniuition and halo ejfec1s in social percep1ion. Oslo ,
Universite1s-Forlaget .
Thorie et mthodes
679
Romney, A.; d'Andrade, R.

1964
"Cognitive aspects of English kin terms", American Anthropologist
66: 146-170.
Rosch, E.
1973
"On the interna! structure of perceptual and semantic categories",

pp . 111-144, i.n: T.E. Moore, (ed.) Cognitive development and the
acquisition of language , New York, Academic Press.
"Classification of real-world objects: Origins and representations in
cognition", pp. 212-222, in: P.N. Johnson-Laird, P.C. Wason (eds),
Thinking, readings in cogn itive science, Cambridge, Cambridge
University Press.
1977
Rosch, E.; Lloyd, B.B.

1978
Cognition and categorizat ion. New York, Er lbaum.
Schelling, T .C.
1960
The stra1egy of conflict . Cambridge, Mass ., Harvard Universit y
Press.
Smedslund, J.
1955
M ultiple-probabi/i1y /earning. Oslo, Akademisk For lag.
Thvenot, L.
1983
"L'conomie du codage social", Critique de l'conomie poli tique 23:
186-216.
Tyler, S. (ed.)
1969
Cognitive anthropo/ogy . New York, Holt, Ri nehart and Wi nston.
Glossary *
agrge:
autodidacte:
baccalaurat:
BEPC:
cadre :
CFDT:
holder of highest university qualification (agrgation ),

member of top category of secondary teachers
"self-taught", but in the sense of holding a post as
cadre or ingnieur by virtue of experience and
promotion rather than a prestigious diploma
18-plus school-leaving examination
Brevet d'tudes du premier cycle, secondary schooling
diploma for those who do not proceed to baccalaurat
"executive/manager", but cf. this text, passim; cadre
suprieur, "top executive"; cadre mayen or "petit
cadre", "junior" or "minor" execut ive
Confdration jranraise dmocratique du travail, tradeunion organization, non-Communist left
For the other names of occupations, see Table 1
680
Theory and methods
Confdration gnrale du travail , trade-union

organization, close to Communist Party, strongest in
traditional industries
classes moyennes:
"middle classes", but tends to be used to designare
"lower-middle classes"
diplom de l'Universit: holder of university (rather then grande cole) degree
engineering grande cole
Ecole Centrale:
"living as man and wife", a status now officially
en concubinage:
recognized
femme OS emballeuse: female packer (OS = unskilled factory worker)
Ecole des Hautes Etudes Commerciales, a grande cole,
HEC:
the top French business school
holder of an engineering qualification
ingnieur:
minor engineering school (as opposed to grandes coles)
cole d'ingnieurs:
the State institute for socio-economic statistics
INSEE :
Paris poli tical science institute (on a par with the
Institut des Sciences
grandes coles)
politiques:
lnstitut universitaire de technology. a higher-education
IUT:
college offering vocational courses
supervisors, foremen
Maitrise (agents de):
19th century term for shop-girl, connoting "flightiness"
midinette:
skilled metal-work er (P2 indicares a recognized leve! of
ouvrier P2 fraiseur:
skill and remuneration)
skilled workers (generic)
ouvriers qualifis:
the independent professions, such as doctors, lawyers,
professions librales:
etc.
CGT:

Boltanski Thevenot Los Usos Sociales Del Cuerpo

Uploaded by

Copyright:

Available Formats

Boltanski Thevenot Los Usos Sociales Del Cuerpo

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Boltanski Thevenot Los Usos Sociales Del Cuerpo

Uploaded by

Copyright:

Available Formats

Social Science Information

Finding one's way in social space: a study based on games

Downloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Theory and methods

Luc Boltanski and Laurent Thvenot

Finding one's woy in social spoce :

The sociology of everyday interactions presupposes that individuals

Dow nloaded from http://ssi.sagepub.com at La Trobe University on A ugust 3, 2009

from the SAGE Social Science Collecti on s. Ali Ri ght s Reserved .

Theory and methods

Boltanski and Thvenot

petence is rarely addressed. Our current research, outlined in this

Boltanski and Thvenot

Downloaded from http://ss i.sagepub.com at La Trobe University on A ugust 3, 2009

Theory and methods

Boltanski and Thvenot

(1) They took into consideration a great deal of information

Downloaded from http://ssi.sagepub.com at La Trobe University on August 3,2009

Boltanski and Thvenot

Nom et ad rcssc de l 'ta bl isscrncn t :

enseignemen t tec h niq ue et forrnat ion

q ua l i fica t ion , grade ou rimi t ion

Dirilimes les rilus levs:

-Mem bre d 'u ne rirofession li bralc

Theory and methods

Boltanski and Thvenot

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Boltanski and Thvenot

When dividing piles becomes too difficult, the solution that is

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Theory and methods

Boltanski and Thvenot

similarities between cards which belong to categories whose

Legitimate labels and the naming of categories

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Boltanski and Thvenot

for the participants' difficulties when they are asked to provide a

suprieurs et professions librales*, agents de maftrise* , ouvriers

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Boltanski and Thvenot

about social differences -that is, social classes -in a euphemistic

Although participants generally used a relatively limited and fix-

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Theory and methods

Boltanski and Thvenot

ed vocabular y to designate social groups (which, as has been seen,

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Frequency with which cards were chosen to head piles

5 central heating engineer

Downloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Bo/tanski and Thvenot

The groups and their typical representatives

Despite the fact that a name given to a particular category, which is

Theory and methods

Boltanski and Thvenot

a name or a label), it is clear that certain individ uals whose social

Dow nloaded from http://ssi.sagepub.com at La Trobe University on August 3, 2009

Boltanski and Thvenot

or trade-unionism, with the creation of organizations which

suprieurs et professions librales, agents de maftrise , ouvriers