1

RULES AND IMPLEMENTS ; INVESTMENT IN FORMS*

Laurent Thevenot

Thévenot, L., 1984, "Rules and implement: investment informs", Social Science Information, 23, 1, pp.1-45.

INTRODUCTIONA code is a tool well known to lawyers in the sense of a

set of texts which have the force of law. But the term"code" is also used to refer to any collection of rules ofprinciple in matters of morality or honour, and this showsthat alongside the civil or criminal code, there existcodes which are less formally laid down, and sometimes noteven written, whose effects are not as carefullycontrolled. Thus the term "code" is also applied to the setof conventions which govern "regulated" communicationsbetween people where the available equipment necessitatessuch a treatment of messages. Indeed, the spread ofinformation technology has contributed to a huge increasein coding operations. By redefining the term, linguisticand semiological research has also greatly enlarged therange of fields in which codes may apply. However, thereare such obvious differences between a code of trade, acode of honour, a telegraph code and a code of dress, thata simple demonstration that they all share a logicalfunction does not amount to proof that they are the same.

* This article was translated from an original version inFrench by Jill Forbes.

Social Science Information (1984), pp. 1-45, London,Beverly Hills and New Delhi, SAGE, 23, 1.

2

Statistical coding operations lend themselvesparticularly well to the comparison and analysis ofdifferent code forms. Any scientific activity and, moregenerally, any attempts at rationalization depend on theprocess of formal categorization involved in suchoperations. But statistical coding also brings together thelegal and administrative codes used in devising questionnaires,the linguistic codes which govern the statements of those whoare interviewed, the occupational codes which providerecognized labels that can then be stated, the cognitivecodes used by the coders who have to interpret the answersin order to place respondents under a heading in thenomenclature, and the technical codes needed so that theanswers can be fed into computers. The first part of thisessay, therefore, will review the findings of research onstatistical coding, and especially on occupational status.This offers the greatest number of links with other codeforms and is therefore a suitable object around which toattempt a consideration of the characteristics of the codeform, as well as the kinds of equivalents it produces whichmake possible actual articulation with the different codeslisted above.

However, the research presented in this article is notconcerned simply with the problems of statistical coding,and the discussion of them merely serves to introduce andillustrate a more general problematic to which theremainder of the text is devoted. In the second sectionTaylor's prescriptions are reconsidered and the code formis, as a result, compared with a whole series ofinstruments with which Taylor would wish to equip thescientific manager of labour—instruments which range fromthe tool itself, to the written instructions, the slide-rule and the "task" as he defines them. Sections three andfour put forward a definition of investment which is able tocover this range of form-giving activities as well as theimmobilization of capital usually meant by the expression,with a view to providing a better economic analysis of therespective use of capital and labour factors1. Section fivesuggests that the main practical application of this

1 This was the approach of a research project undertakenwith François Eymard-Duvernay. For the application ofsuch an analysis to a specific company see

3

theoretical framework is in the investment in forms used tomanage labour, and these are mostly readily observed, as isshown in section six when management methods change.Finally, section seven contains a discussion of some of theeffects of the production of State forms with generalvalidity.

Such a framework should allow economic analysis to takeaccount of a set of established forms which are usuallyabandoned to other disciplines because they are thought ofas "symbolic". However, grave inadequacies may result fromthe failure to integrate such forms into economic analysisas is the case with companies in which investment in formsis more decisive than investment in plant in determiningperformance.

Thus the final section provides a more generaldiscussion of the effects of the investments whichcontribute to the establishment of equivalent forms. Thefact that they are extremely costly is usually ignored ineconomic formalizations which, for the purposes ofanalysis, tend to assume that they have been establishedfrom the outset, rather like the currency. The article endswith a consideration of the effect of such investment invery general forms which constitute the State, on the powerrelations between agents who make use of more specificforms.

1. OCCUPATIONAL IDENTITY AS AN EXAMPLE OF THE CREATION OF CODEFORMS

All statistical production depends on coding, which is aprimitive form of data processing since it implies that theinformation produced and analysed will be put into standardform. Statistical coding as a scientific activity istherefore regulated by logical principles which determinethe correct forms of classification. But when a statisticianis required to classify a post or status which has notalready received technical, administrative or legal coding,these logical principles are flouted, and this is the casewith young people whose family situations and occupational

Section 6 and also Eymard-Duvernay and Thevenot (1982).Eymard-Duvernay and Thevenot (1983a) contains a re-examination, from a theoretical standpoint, of theconcepts of "entry barrier" and "specific investment".

4

status are often ill-defined (Thevenot, 1979). Even when hefalls back on made-to-measure logical categories such asage, the statistician cannot avoid using practicalcategories such as "young people" or "old people" andfalling into the paradoxes they create. Such paradoxes havebeen likened to the "how many grains make a heap" model(Dummett, 1975; Weiss, 1976), and they can only be resolvedby examining the degree of precision to be found in"contradictory" codings and seeing how these relate to theobjectives of the social agents who make use of them. Inthis way the strategy of deliberate vagueness pursued bycertain social groups can be analysed (Thevenot, 1979).

The "mental burden", as ergonomic specialists woulddescribe it, of those employed in codification (Pinsky,1980), which makes this phase the most costly in theprocess of statistical production, may be thought of as thecognitive burden of trying to adapt strictly logicalrelations (instructions for encoding) to the impreciseforms in which they appear. However, neither the researchof cognitive psychology into such processes of adaptation(Pipino, Van Gigch and Tom, 1981), nor research into thetypical, non criteria-based form of ordinary mentalcategories (Rosch, 1977), are able to give an adequateaccount of the heavy labour of adaptation, since they donot look at the relations between the form of generalcategories constituted for purposes of reckoning throughlegal procedures for example, and the particular forms used byindividuals in their interpretations. The study of how socialidentities are recognized has clarified these relations(Boltanski and Thevenot, 1983).

On the other hand, a saving is achieved in statisticalcoding by having a link with other code forms. Whenoccupations are divided into homogeneous and exclusivecategories based on explicit criteria, according to theprinciples of the science of taxonomy, this must create ade facto link between such classifications and existingdivisions—those which have already been established throughstandardization processes that have made such criteria seempermanent and useful and have turned them into a sourcethat can be used in classification.

The research that has been most helpful in understandingthis operation outside the particular logic of statistics

5

is undoubtedly contained in sociological studies of thesocial uses of categorization (Boltanski, 1970; Bourdieu,Chamboredon and Passeron, 1973; Bourdieu and Boltanski,1975). In addition, studies of the historical developmentof statistical nomenclatures have brought out the linksbetween such classifications and other attempts at creatingcode forms which are controlled by corporate organizationsor by the State. Thus the classification of economicactivities in the 1861 nomenclature was based on a notionof product thanks to industrialists who banded together toprotect themselves against free trade and thus gave weightto this criterion. Similarly, it would appear that theproposed 1942 nomenclature followed the lines of theEmployers' Corporate Organization Committees (Comitesd'organisation corporatifs du patronat) which were in factset up as cartels (Guibert, Laganier and Volle, 1971;Volle, 1982). As far as the classification of occupationalactivities is concerned, it has been shown that thestatistical category of "cadres" (executives) and thecriterion of "skill" which is used to separate workers intodifferent categories were both linked to the legalconstitution of these categories, and it was as a result ofthe collective agreements and labour regulations whichbegan to come into force in the 1930s and which becamegenerally applicable with the Decrets Parodi2 in 1945 thattheir legal constitution took place (Desrosieres, 1977).Boltanki's analysis of the way "cadres" became a well-defined group has thrown light on the "social technologies"which enable the "collective persona" of social groups tobe represented and, in consequence, identified (Boltanski,1982). In addition, when the codification practices of theencoders employed by INSEE were closely examined, clearlinks were revealed between the statistical recording andthe form in which the occupational identity of theemployment notified by respondents had been established(Thevenot, 1983b).

Thus the terms "state", "commission", "office","ministry" and "corps" all describe situations in which thepost is most clearly defined and most strongly linked to

2 A set of decrees issued by Minister of Labour AlexandreParodi instituting scales of qualification-related wagerates.

6

the post-holder, with the link sometimes being confirmed bya legal act such as the purchase of a commission or theadmission into a corps. Indeed, this link is so strong thatthe name of the individual who fulfils these functions isoften preceded by the name of the function itself (as withJustice X, Lieutenant Y or Doctor Z). This then becomes adescriptive property which ensures that statistical codingis almost automatic even when it is done manually—i.e.without the encoders using automatic data-processingprocedures. Like a computer routine, this link is theautomatic result of the combination of code forms that havegeneral validity. Moreover, this combination does notrequire any personal interpretation on the part of the codeoperatives, which may be expensive and is difficult for theproducer to standardize. In this way, the criterion formwhich can be justified on logical principles also proves tobe very economical because it makes possible articulationswith other fixed forms and avoids the costs inherent in thepersonal interpretation of such articulations and in thecollective negotiations which are needed to achieveuniformity among individual interpretations.

When the recent reform brought about changes in theoccupational nomenclatures used for statistical purposes,this was an extremely opportune moment to examine theprocess of form-giving and how links are established withexisting forms. It also provided a good opportunity toobserve how the representatives delegated by occupationalgroups took steps to influence the statistical recording oftheir occupation in an attempt to achieve legalcodification of a more advantageous kind (Desrosieres, Goyand Thevenot, 1983). The wording of a heading, theformulation of a definition, the disappearance of certainterms (such as "trade" or "hygiene") and the introductionof others (such as "information" or "health"), thestandardization of a name and the promotion of a particularcriterion for definition are all costly operations whichare not exactly legal but which take place prior to suchterms being enshrined in law. When an occupation is legallyrecognized it acquires a title, its prerogatives are laiddown as are the way in which recruitment takes place andthe qualifications needed. In other words, competition is

7

regulated in a field of occupations having to share amarket.

What emerged was a continuum of more and less stronglyconstituted forms, ranging from the most consolidatedoccupational forms, anonymous titles with universal valuebecause they had legal validity, to individual names whichdepended on the choice of the person notifying them. Theless strongly constituted the form, the weaker its linkwith the statistical recording and the more costly it is tomaintain. Thus a "craft" or "trade" may not be as tightlyregulated as a commission or an office, but it at leastensures a permanent, quasi-universal guarantee of quality(almost an "appellation controlée"). By contrast such tradenames have broken down in the processing industries, wherejob descriptions have ceased to have general validity, andare only specified at the more particular level of thecompany "trademark". The importance of such forms in companymanagement has been brought out in the studies of the useof job classifications in individual companies (Thevenot,1983b). For example, the opening of a plastics processingfactory, designed for large-scale highly automatedproduction, involved the taking of a number of strategicmanagement decisions, one of which concerned jobclassification which is a classic tool of management. Thefirm had made heavy investments in injection presses andthis had influenced its decision to set up in a rural areaso as to keep down wage costs by employing labour withfewer educational or craft qualifications and lessexperience, which would have called for specialremuneration. The manager of this factory had thereforefollowed up these investment options by a deliberate phasingout of the employment scales recognized by the nationalagreement which he considered outdated because they werebased on craft definitions. At great cost, both in terms ofdesign and because of the lengthy negotiations requiredwith representatives of the employees, he drew up a newsystem of classifications in which craft names werereplaced by references to an entirely artificial codesystem (OS lA; OS lB; etc.) and to the particular shopconcerned (injection, extrusion, and so on). The phasingout and replacement of scales based on crafts, which areproperties recognized beyond the particular company and

8

which can be linked to qualifications controlled by theState, was in fact accompanied by much criticism of theclaims of CAP3 holders who consistently attempted toexploit their qualifications—in other words, to make aprevious connexion work within this company.

Similarly, a study of the watch industry has thrown newlight on the relation between job descriptions and theeconomic analysis of the companies which employ them(Eymard-Duvernay, 1981; Bony and Eymard-Duvernay, 1982). Inthis industry, job codification is less highly developed incompanies in rural areas which are often small and nothighly technical (since they tend to be concerned withassembling and marketing watches rather than themanufacture of components). The very personal relationsbetween employers and workers are an extension of familyand neighbourhood relations and are not mediated throughoutside institutions which might, for example, decide thecriteria used to define jobs. Any classification validbeyond the individual company is therefore looked on by theemployers as an attempt to brand their employees on thepart of outside institutions who are attempting to meddlein the direct employer-employee relations. The companieswhich have the most highly codified relations with theirlabour force are those inside Besançon which draw on atruly urban labour market in which the employer is notpersonally acquainted with the employee before he takes himon, and where labour moves easily from one company toanother. These companies are more capital-intensive and aremore likely to manufacture components, and this means thatthey are more closely connected to the metal industry whereconventional classifications are very firmly established.It is therefore these companies which have played adecisive role in breaking down the traditional crafts inwatchmaking (titres de metiers horloger) that have "losttheir value" as a result of technological change (quartz),and in substituting the more general classifications of theUIMM (Union of Metalworking and Mining Industries).

3 Certificat d'aptitude professionnelle, i.e. vocational trainingdiploma.

9

2. THE CODE AND THE TOOL: TAYLOR'S MACHINERYThis examination of the coding of occupational identity

and, more generally, the operation of statistical recordinghas made it possible to compare a whole series ofoperations which produce code forms. The attempt has beento look beyond the differences in the nature of the coding(whether statistical, administrative or legal) and thevariables that have been observed, and to relate them allto the common function of producing standard forms whichpermit the establishment of equivalents — i.e. the stableand economic articulation of these forms.

Such an analysis is helpful in giving an account of thelimits of statistical recording and the ways it might beimproved (Desrosieres, Goy and Thevenot, 1983). However, itis not epistemological considerations that are of mostinterest in such an account of form-giving operations. Whathas been said about the savings that might be expected fromthe use of such forms, and especially from the links theypermit with rules of a legal, administrative or technicalkind, suggests that a theoretical framework is needed inorder to give an account of all of these operations fromthe standpoint of the economist.

F. W. Taylor's Principles of Scientific Management (1967) hasbeen used to develop this theoretical framework. There havebeen numerous exegeses of elements of Taylor's book andmany of its contents have frequently been condemned, butthe text has not previously been used in the mannersuggested here. One of the most difficult aspects of thisstudy has been the effort of reconstruction needed torelate different kinds of coding to the same model, and inwriting his book Taylor was similarly led to link thesedifferent objects. The operation was, of course, madeeasier by the fact that he was writing a handbook, which isa work intended to produce a set of systematic and unifiedrules out of a collection of disparate habits. Handbooksare, moreover, frequently based on the notion of the toolwhich is the most rectified state of habit and custom.

Taylor's handbook, therefore, contains a particularlylarge repertoire of form-giving instruments, all of which,as his text makes clear, need to be carefully adjusted sothat they work together to produce what Taylor calls "themechanism of scientific management" (p. 122). Among the

10

instruments noted may be mentioned (in an order discussedfurther below):

— "the implement": Taylor stipulates that a standardimplement should be developed using the empirical formsgiven to the existing tools and in the light of theirrespective performances.

— "the adjustable scaffold" (for bricklayers), which is anaddition to the preceding tool (p. 80).

— "elaborate slide-rules (...) especially made for the purposeof determining the capacity of metal-cutting machines" (p.99).

— "implements and methods for properly making time study".— "books" and "records": "the practical use of scientific

data also calls for a room in which to keep the books,records, etc. and a desk for the planner to work at" (p.38).

— "experiments", leading to "the establishment of (...)rules, laws and formulae which replace the judgement of anindividual workman and which can be effectively used onlyafter having been systematically recorded, indexed, etc. (pp. 37-8).

— "the task", which lays down not only what must be done,but how and in what length of time, and which thereforeprovides the workman with a precise yardstick ofmeasurement.

— "the written instructions" on an "instruction card" which isgiven to the workman and provides detailed indications ofthe best way of doing each job. The instructions are drawnup in advance by the special "planning department".

— "the bonus and premium" which are to reward the workmanwhenever he carries out the task in the time laid down (p.122).

— "writing and talking" (with a view to) "educatingworkmen" (p. 27).

— "the scientific selection of the workman" (p. 43).— "the shoulder to shoulder contact" (p. 27) and the

"close, intimate personal cooperation between the managementand the men" (p. 26).

This is in many respects an odd list, made up ofheterogeneous objects taken from many different fields ofactivity. In it are to be found objects used in productionon the shop floor, instruments, plans, conventions,

11

scientific formulae, school precepts, ways of givinginstructions which are close to military orders, methods ofpayment to be used in companies, principles, advice andexamples to follow when deciding what action to take. It iseven difficult to categorize these objects and manydifferent terms might be used to describe them. Yet Taylorbrings all these objects together into a single mechanismso that they become articulated parts of a singlestructure, and it therefore seems essential to use a singletheoretical model to describe them all, which would explainhow they fit together and what economies are to be expectedfrom their introduction. These forms may be either machine-tools or rules of conduct, made of very differentmaterials, cast in metal or written on paper (Callon andLatour, 1981). But beyond these immediately perceptibledifferences there is a further characteristic whichdistinguishes these forms—and this explains the order ofthe list above—which might be called a rigidness orinflexibility (the ability to resist efforts to distort,adjust or negotiate them). It is clear that a machine helpsto reproduce, with the minimum of human intervention, astate of relations between standard objects (in thisinstance raw and manufactured materials). And in general, a"good form" is one which fulfils this requirement, as isclearly shown with the case of the law.4

3. A BROADER DEFINITION OF INVESTMENT: INVESTMENT IN FORMThis section will expand on the outline above by looking

at the various instruments listed under the model proposed,as forms which arise from investments. These needspecification in such a way as to take into account a range

4 The prescribed forms must be respected because they arewhat enable the distinction between a "formal" and"informal" act, between an acte solemnel (an act such asmarriage which is only valid if accomplished withincertain forms before a clergyman or registrar) and anacte consensuel (which may be a private agreement orcontract needing no special forms). Forms guarantee aresult, i.e. a relation with other objects, whether suchforms are "legal", "probative", "capacitative" or"executory" (forma dat esse rei),

12

of forms which includes tools, wage-slips, trademarks,instructions, training and habits.

Economists have usually conceived of investment asdepending directly on a definition of the capital whosegrowth is measured. The conventional imagery peculiar tothe teaching of economics consists of formulations whichare half way between theoretical formulae from which theytake their formal rigour and ad hoc applications to exampleswhich attempt to provide realistic instances, and suchimagery makes a clear distinction between capital andconsumer goods. Thus illustrations based on hypotheticaldesert islands show that in subsistence economies there isa clear division between natural resources and capitalgoods, between the water and the pail. However, thisdistinction becomes less clear cut when the initial stocksof raw materials or cash advances for wages are taken intoaccount as they are by the classical economists. Suchassets are not included in the modern definition ofinvestment which excludes working capital and only takesaccount of the growth of fixed capital, that is durableproduction assets. Furthermore, the example of consumerdurables such as motor cars suggests that although they donot count as investments, unlike real estate, references toproduction—that is, to the type of agent taking theinvestment decisions—do not provide the most usefuldistinction on which to base the concept of investment.There have been many proposals to stretch the definition ofcapital to include knowledge that is acquired, as in thenotion of human capital, and the fact that patents arecounted as assets shows that a more satisfactory definitionof investment is not supplied by the distinction betweenmaterial and non-material goods.

However, lifespan does appear to be a fundamentalcharacteristic since it is this which determines whethercapital is fixed or working. It would therefore appear thatthe productivity which should derive from the "sacrifice ofconsumption" and the "roundabout method of production", andwhich is more often illustrated than explained in theliterature, depends on this very engagement of temporaryliquidity in a durable asset. What is suggested here isthat it is such an immobilization which permits the fixingof a relation that can be reproduced between, for example,

13

certain forms of input and output, and leads to theeconomies of labour looked for from the investment plan.Such an investment cannot be thought of simply as amaterial form, like a machine, since it also requiresstandardization, the definition of norms and thecodification of input and output. Because there is a needfor articulation between the exploitation of the tools andthe form-giving operations which make them function, thereis every reason to put forward a definition of investmentwhich can take account of all of these operations, whateverthe material nature of the forms produced. From this pointof view, it would appear that the most relevant way toconceive of investment, which takes account both of theclassical use of the term and the extension of its meaning,is as a costly operation to establish a stable relationwith a certain lifespan. One illustration of this definitionof investment might be the purchase of a patent, since thisis an expense which gives the right to reproduce aparticular form of a more or less material kind (dependingon the kind of patent) over a certain period of time. Themonetary equivalent, which is rendered objective by thetransaction of purchasing the patent, makes this example aparticularly clear illustration of this definition ofinvestment since it means that it is counted as a fixedasset or a capital immobilization. Classical references to"sacrifice" in theories of investment and interest ratesmay be reinterpreted within this framework, even if theoutlay is not considered an establishment cost or afinancial commitment, but as a sacrifice of some other moreimmediate expenditure or of unproductive savings as inSenior's "abstinence theory of interest".

By forging a monetary equivalent between goods thatbecome disposable at different dates, the financial marketreduces the temporal dimension of investment using interestrates as an operator. As is well known, Fischer, by puttingforward the concept of temporal preference which resultsfrom the relation between the amount of future revenue asan "equivalent" of an immediately consumable unit,formalizes for the consumer the counterpart of theproducer's "rate of return" on investment. He reconstructsthe classic conditions of equilibrium on the basis of amarket composed of goods disposable today and goods

14

disposable at a later date. The projection of the temporaldimension onto the realm of goods, which tends to disguiseits specificity, is nevertheless linked to the "investmentopportunity principle" of converting present revenue intofuture return. By making this "capacity" derive from theoperation of establishing economic regularity at any givenmoment, it becomes possible to recognize it both in theinvestor and in the consumer who has committed himself fora certain period by sacrificing immediate consumption. Inhis discussion of "binding oneself" or "precommitment" inUlysses and the Sirens, Jon Elster emphasizes the human"capacity to relate to the future" and to calculate the"global maximization that requires by-passing a localmaximum" (Elster, 1979, p. 10).5 As Leibniz observed:"Appetitions are like the path of the stone which fliesstraightest toward the centre of the earth, but does notalways take the best road, being unable to foresee that itwill meet with rocks that will break it; whereas it wouldhave come closer to its end had it had the wit or the meansto turn aside (...) which makes us know that it is reasonand the will which lead us toward happiness, but sentimentand appetite lead us only toward pleasure" (Leibniz, 1875-1890, vol. 5, pp. 175, 182, cited in Elster, 1979, p. 10).Elster discusses the reasons for "using rules rather thandiscretion in formulating economic policy": cost ofinformation, cost of adjustment, cost of uncertainty. Heprovides examples of institutions which, in his view, maybe considered as "devices for precommitment": the centralbanks, in the British and American systems, who decide onmonetary policy at a distance from the political sphere),periodical elections which are not plebiscites or aconstituent assembly "that lays down the ground rules to befollowed by future generations".6

5 Like the definitions referred to above which were basedon terms such as "abstinence" and "sacrifice", Leibniz'scontrast between "reason" and "sentiment" or between"will" and "pleasure" is taken from ethics. And it isethics which provide the rules for living, that isformulations made stable by written or spoken repetition,some of which are almost as firmly laid down as legalprinciples.

15

4. THE AREA OF VALIDITY OF THE CODE FORMIf Taylor's list is given further consideration, it

becomes clear that lifespan is not sufficient to describethe effects of investment. Thus the "written instructions"drawn up for the workman may have the same lifespan as the"wage rates" but they differ insofar as the writteninstructions are specific to the company, whereas wagerates may be applied over a whole industry through wageagreements, or may even be nationally applicable whenagreements such as the SMIG7 are introduced. This meansthat a further important characteristic which must be takeninto account is the area over which a form is valid as well

6 Elster is most concerned to explore the limits ofrationality and irrationality in human behaviour and doesnot, therefore, examine the form-giving activities whichmake up commitment. Nor does he consider, as has beenattempted here a comparison between investment and otheroperations designed to set up temporal regularity. Thus,in his axiomatic of commitment ("binding oneself"),Elster puts forward a condition which he believes mustexclude investment from his conception of commitment: "Ifthe act at an earlier time has the effect of inducing achange in the set of options that will be available atthe later time, then this does not count as bindingoneself if the new feasible set includes the old one" (p.42). It would, however, appear that this condition onlyexcludes investment if investment is defined (as it is byElster) as "any sacrifice of present goods in order tomake more goods available later on" with no considerationof the form-giving operations needed to establish ahighly equipped relation of production as so manyoperations to reduce the set of what is feasible. For itis surely such a reduction, along with the increase inpredictability that accompanies it, which is the sourceof any yield from investment.

7 Salaire minimum interprofessionnel garanti (i.e. nationalguaranteed minimum wage) was created in 1950 and wasreplaced in 1970 by the SMIC (Salaire minimum interprofessionnelde croissance), (i.e. national guaranteed increasing wage)which is fixed in relation to the rise in the cost ofliving and also raised every year on the 1st of July inrelation to general economic conditions.

16

as the length of time for which it is valid. The larger thearea of validity, the more likely it is that establishedforms will be interconnected and the greater the savingsthat may be expected from the investment. As a generalrule, the two characteristics of investment that have beenproposed here (lifespan and area of validity) go together.Thus an individual's personal action is fleeting and oflimited impact; it has only momentary significance withinthe context of personal interactions and affects only theindividuals involved. However, an individual may give thisaction a less peculiar and more standard form if, forexample, he does something remarkable which is intended to bereported and which will therefore acquire greater validityand lifespan (as a result of being recalled).8 Conversely,international time is a standard form of exceptionalstability and universality which may seem obviously usefuland necessary, but which was in fact only set up after along period of hard work (Zerubavel, 1982). Thanks to thisexample, some idea may be gained of the cost of investmentrequired to achieve standard time, without which otherinvestments which depend on regularity could not have takenplace. The development of heavy industry at the end of thenineteenth century is one investment which obviouslydepended on such regularity (Thompson, 1967), and an evenbetter example is Taylor's machinery which is clearly builtaround the measurement of time. The creation of a time formwhich is valid over a large area depends on thearticulation of a number of different forms such as thetechnical instruments and scientific formalizations9 used

8 Both this observation and the remarks on general andindividual forms owe a great deal to Luc Boltanski's workon the technologies of representation (Boltanski, 1983),and to his current research into the way individuals canbe businesses, that is, may construct individual causeswith general value.

9 The treatment of scientific methods as some, among theseveral, methods used in the "social construction ofreality" is a major topic in the work ofethnomethodological sociologists, see. e.g. Cicourel(1964), Garfinkel (1967), Berger and Luckmann (1967). Fora very stimulating analysis of the techniques forconsolidating scientific proof see Latour and Woolgar

17

to create Greenwich Mean Time, relations of equivalenceequipped with communications networks (mail coach,telegraph, railway, etc....) in order to extend thevalidity of this time, a legal definition of States inorder to create standard time zones, and national andinternational institutions to agree about time.10

Between the most universal of forms such as measurementsor international law, and the most individual of forms suchas personal interactions, there exist intermediate formswith lesser lifespan and validity, such as company wagescales drawn up in the way described above, or companytraining which carries no State validation, or the networkof customers for a particular trademark, or a set of "houserules", or, indeed, the "standardized labour implements"Taylor recommends should be introduced throughout a company(Taylor, 1967, p. 66). However, these two majorcharacteristics are not found together in the case of someforms, and that is why it is necessary to distinguish them.Thus the craftsman's own personal tools may be aninvestment with the same lifespan as the standard tools,but they will not have the degree of standardization whichwould allow the equivalents to be valid over a large area.

When standard time was given a form, the importance ofwhat was spent on investment was clear even if the

(1979).10 The work of uniting the world offers many examples ofhow code forms articulate with forms that already exist.At the 1884 conference to choose a prime meridian to beused worldwide—that is, in all the countries which haddiplomatic relations with the United States—the Britishand American delegates "repeatedly insisted on the factthat choosing Greenwich as the prime world meridian wouldentail the least modification of the system already inuse, whereas choosing a different meridian would only addpointless confusion to the existing expense and otherinconveniences" (Zerubavel, 1982). The French delegateunsuccessfully opposed this realist argument based on thesavings to be made by linking up with investments alreadyin place, by an appeal to a wider cause which was thatthe meridian must of necessity be neutral.

18

economies were achieved by recouping earlier investments ofa scientific, technological, administrative or legal kind.Some of this expenditure can be expressed in monetary termsand can therefore be easily quantified. This would be trueof the installation of plant and machinery (measuringequipment, the construction and connexion of networks,timetables and hours of work, etc.). Other kinds ofexpenditure, such as the personal time spent in negotiationsto reach agreement, are less easy to objectify. It isinteresting that the more individual the investment, thegreater the amount of personal time that has to be spentinitially. With the most individual of investments, time isspent gaining the habits and experience which make up theskills or the network of relations belonging to anindividual or a company. This points to another possiblemode of investment which is not an immediate and costlyacquisition but training that takes place over a period oftime, in order to get into a habit, a relation forged as a pathis traced by repetitions which fix its form. This mode ofinvestment can affect the established forms which will notbe strongly objectified if there is no complementarymaterial investment. Bourdieu's discussion of thedifference between "embodied" properties, known as thehabitus, which function, for example, in transactions inthe local markets of traditional societies, and thequalifications which set up a unified and objective market(Bourdieu, 1980) demonstrates that this feature of theestablished form is extremely important. There is thereforea third characteristic of investment in form which must betaken into account, linked to the way it is constituted andthe extent to which the form is objectified or materially"equipped", the extent, in other words, to which it existsin the form of anonymous implements by which it isrestricted, whether these implements are of a conventionaltechnological kind, or are of a legal, scientific or othernature. It is clear that the equipment derives from andcontributes to the fixing and diffusion of a form, and thatit is therefore generally linked to its lifespan and areaof validity. But distinguishing this characteristic doeshelp to illustrate a number of cases. Thus the etiquettewhich "civilizes manners " (Elias, 1973) is, like a code ofpractice, a set of relations practised within a given

19

social group, which has great stability (since generationconflicts are often what break these relations). However, acode of practice is much more objectified than a code ofbehaviour and it is therefore much more restrictive.However, the difference between these codes may also begauged in relation to their areas of validity. Thus the lawrelating to codes of practice is not linked to common lawwhich has a national rather than a specificallyprofessional or occupational validity. Similarly, habits orskills peculiar to an individual or an occupational groupare more or less flexible to the extent that they are moreor less fully equipped with tools which, even if they arespecific, ensure that such skills and habits can be handedon from one person to the next. The significance of suchequipment became very apparent through study of howoccupations in the health service were constituted. Theless equipment there is, the more difficult it is for theform to be anonymous, so that the form remains highlyindividualized and is sometimes even invested in aparticular individual, which implies high maintenance costsin terms of personal time spent. The same may be said ofthe time taken in the repetitions which create fixed habitssuch as that of "soldiering" (i.e. skiving) (Taylor, 1967,p. 24)11, or the interpersonal relations which only functionwhen individuals see one another frequently and which breakdown more quickly than material implements. Because nocomparable equipment is available, such relations cannoteasily cease to be individual. Thus a business finds iteasier to sell its "custom" to another if its custom ismaintained through a network of dealers and an establishedtrademark. Similarly, personal relations can only be handedon if a general institutional objectification takes place

11 In order to indicate how relatively widespread thishabit was and to give some idea of the extent of thisform's area of validity, Taylor mentions how it had beenlabelled in certain standard ways in the differentcountries: "Underworking, that is deliberately workingslowly so as to avoid doing a full day's work,'soldiering' as it is called in this country, 'hanging itout' as it is called in England, 'ca canae' as it iscalled in Scotland, is almost universal in industrialestablishments" (p. 13).

20

and an anonymous equivalent, such as an educationalqualification, an association, or a marriage alliance whichcreates an equivalent at the more limited level of thefamily unit, can be produced.

5. INVESTMENTS IN FORM IN THE USE OF THE WORKFORCE

An analysis of Taylor's principles along the linessuggested here is of particular interest because theseprinciples establish a radical model of the company inwhich it is presupposed that the company controls everyinvestment (in the sense the term has been given here)relevant to the use of the workforce. Thus the employee'sindividual activity is changed into the time-equivalentform which makes possible economic connexions, and thesemay occur through technical relations (by means ofimplements), regulations (through hours of work) and socialand contractual relations (through wages). The key form inTaylor's system is the task which makes possible theelementary coding of the individual employee's personalactions. It is a stable form that is equipped by means ofwritten instructions valid throughout the company whichhave nothing to do with the qualities which make up theindividual employee's identity.

"Perhaps the most prominent single element in modernscientific management is the task idea" (Taylor, p. 39).And if Taylor lays down ways to get the workman to fulfil his task,it is because in his view a workman's individual activityis something which falls outside the general forrn of thecompany time-equivalent, and must therefore amount only tounprofitable "soldiering". Thus in some sense Taylorattempts to create the material conditions for a productionfunction which corresponds to the formulations in economictheory, with a work factor that is assimilable to aquantity of time which can be consumed without anyresulting workforce immobilization. As has been seen, thismodel implies significant material immobilizations in theform of machines which the production function couldaccount for, but also a series of other investments whichdo not count as investments but which are nevertheless

21

sustained by the company. Theoretically, however, thecompany is not responsible for the costs which arise fromoutside investments concerning the people employed, whetherthese are nationally valid (as in the case of educationalqualifications) or whether they are constituted in a moreindividual way inside the family or in a company in whichthe employee previously gained work experience.

Experience with this model has shown that it is verywell suited to a workforce which is not highly skilled andis not inclined to invest in work. In the examples Taylorprovides of how recruitment should be conducted, heemphasizes that the ideal workman for the task is "of amentally sluggish type" (ibid., p. 46) and has no specialabilities. Similarly, his reward should be as "liquid" ashis service; it should be immediate and should involve noform of participation since this would suggest commitmentover time. If lack of skill (in the formal sense of theterm—"skill" meaning a form having general value) isrequired of the workforce, recruitment can be extremelyeconomical and can be based on a link with the forms thatare instituted, in just the same way as recruitment whichmight depend on the requirement of a standard skill basedon educational qualifications. These days it would seemthat the standard recruitment of unskilled labour iseconomically carried out by such institutions as the ANPE(Agence nationale pour l'emploi) or the Pactes pour I'emploi12which primarily supply companies that need the kind ofworkforce similar to that in Taylor's model (Mehaut, 1982).As far as the employers are concerned, "the essentialadvantage of the youngsters recruited via the ANPE isprecisely that they have no particular skills" (Faguer,1982, p. 10). This might be expressed in a different way:the ANPE as an institution can only select on the basis ofvariables which are compatible with what might be describedas its "operating model", which, in the case of thisnational institution, means the Stateregulated variableslegally provided for by contracts of employment. This

12 Employment agreements between the State and privateindustry in the form of government measures designed toencourage the employment of young people by means ofsubsidies to companies which take them on.

22

explains the many failures when the applicant turns up inperson at the company's offices. They are the results of"misunderstandings", that is, they are due to undesirablequalities (such as nationality or sex) which had not beenspecified, or properties that cannot be formalized andwhich cannot be objectified within the framework of thismodel. Preferential recruitment of this kind is thereforean example of an economic relation made possible by generalinvestment in durable and compatible forms— forms such asthe rationalization and formalization implemented inTaylor's system and the legal and administrativegeneralizations of a public institution. The relation whichis created is based on forms with general validity (such as"civil status" (age, sex etc.), educational qualifications,collective agreements, and so on) which allow the twomodels formalized above to be rendered compatible andtherefore to be linked. It can thus be seen that it is acomplete illusion to look on the ANPE in a purelyfunctional way as a public recruitment service forcompanies of small or medium size that do not have well-established personnel departments. In fact, small andmedium-sized companies work principally through investmentsof a specific kind requiring personalized recruitment andinterviews. These are conducive to the detection ofpersonal and often physical properties which do not resultfrom any establishment effort, and on them the employerbases the employee's engagement13.

13 Thus the amount of recruitment which is done on thebasis of personal acquaintance decreases markedly inrelation to the size of the establishment (Benarroch andEspinasse, 1982). As far as unskilled workers areconcerned, there are two main modes of recruitment, theformal recruitment through the ANPE mentioned above, andthe other through personal acquaintance with the employeror a member of the staff. It is probable that therespective use of these two methods depends on thedifferent degree to which formal management rules areapplied (and therefore also on how large the company is).An extreme example would be that of small companies whichwait until they have recruited an employee who perfectlyfulfils all their particular requirements before theycommit themselves to a contract (or frequently a sub-

23

However, even Taylor's model which is characterized bysubstantial investments in general and durable forms formaking use of the workforce, implies some employee trainingso as to make possible these form-giving activities. WhenTaylor sets out his principles, he continually moves fromthe techniques for setting down the "rules", "laws" and"formulae" (ibid., p. 36) on "instruction cards" given tothe workmen, to the "education" techniques which ensurethat the workmen observe these rules. And when Taylorwishes to instil the "task idea" into the workman, thelatter's education is acknowledged to be more likely tosucceed if it connects with earlier educational investmentsin training in the explicit, the objective, the rule-governed and the rational14.

Similarly, contrary to the classical analysis of thework factor, the encouragement offered by financial rewardsis not seen as sufficient to ensure the effectiveness ofthe measurable forms on which Taylor's system depends. Theworkman has to be trained in measurement (p. 94), a codeform which must exist prior to the introduction of tasks,machines, rules and so on. In order to transform individualtime into a general time valid throughout the company,equivalent to the measurement of machine time, Taylor hasto train the workmen into accepting his model, and has tocreate the links which would establish what is in hisinterest:

The average workman must be able to measure what he hasaccomplished and clearly see his reward at the end ofeach day if he is to do his best (p. 94).

contraction) to produce something so specific that Itrequires such an employee.

14 "Each of us will remember that in his own case this ideawas applied with good results in his schoolboy days(. . .). The average boy would go very slowly if, insteadof being given a task, he were told to do as much as hecould". The lifespan of this form will be noted. Today ithas become so totally naturalized that one always talksof describing tasks in what a job involves, as though nowthe idea of the task were taken for granted.

24

However, within the limits that have just beendiscussed, Taylor avoids linking his machinery toinvestments controlled by the workmen. He in fact thinksthat the workman is incapable of controlling the passage oftime, the very thing which, as has been seen, creates theopportunity for the commitment which underlines allinvestment. Moreover, the thing which Taylor fears most ofall is that the individual employee might control an areaof validity beyond the individual. He fears, therefore, thegroup of workmen.

He recommends that workmen should never be talked to asa group, that they should not work together in groups ofmore than four, and he suggests that collectiveconsultations and voting can have disastrous consequencesand should be avoided at all costs. He condemns the groupinvestments that the workmen in his factory had achieved inthe form of an "understanding". "Every new workman who cameinto the shop was told at once by the other men exactly howmuch of each kind of work he was to do, and unless heobeyed these instructions he was sure before long to bedriven out of the place by the men" (p. 49). Taylor statesthat as gang-leader he was able to get round this long-established rule because he was in no way a member of thegroup (the area over which the rule was valid), since hewas not the son of a workman and had not previously been aworkman himself. Thus Taylor strongly recommends breakingwith established social groups, however entrenched they maybe, whether such groups are informal workplace alliances orestablished trade-unions. The fact that he is so emphaticsuggests that expenditure and effort are required in orderto undo the investment in forms used for managing theworkforce and the instruments which create the socialgroups needed to establish collective personae (Boltanski,1982). Most commentaries on the introduction of Taylor'smethods do little more than set out what happens whenworkmen lose control of the work process. In other words,they are mainly concerned with the break, recommended byTaylor, in the link between the forms of work managementand groups that are formed. But a detailed examination ofthe spread of Taylor's methods in the United States shows

25

that such a break was not the only result to occur (Stark,1980). Although Taylor may have recommended rules forrelating productivity and wages on an individual basis, thespread of Taylorism tended to encourage the growth ofcollective negotiations and to strengthen the forms ofrepresentation of the social groups affected, includingskilled workers. Stark therefore disagrees with the viewthat is most strongly upheld by Braverman (Braverman,1974), that as a result of Taylor's methods an unskilledworking class took the place of a skilled working class. Infact, the skilled workers' extreme hostility towards anyattempt to abstract their skills into Taylorian formscontrolled by work engineers' was transformed intocooperation when the United States entered the First WorldWar. At that time, the flood of orders emanating from theState required regular, guaranteed and standardizedproduction, and this encouraged the introduction ofTaylor's forms as well as increasing the influence of tradeunion institutions and more general forms of negotiation(wage boards, collective negotiations, etc.) in return forgreater discipline among employees (strikes wereoutlawed)15. During this phase, skilled workers were able toimprove their wages and position, thanks to the increase inproduction, and to initiate cooperation with the engineers'groups which had itself grown up mainly through theintroduction of Taylor's methods. With the end of the War,and State orders, the most general forms of regulation,those which took place at national level, were given up andthe unions lost some ground. However, the form ofproductivity agreements between trade unions and

15 "By establishing strong ties to the State, the tradeunion leaders had gained the necessary protection toexpand their organisational base without having to alterthe organisational form of that base from craft toindustrial unionism. Union membership thus rosedramatically, but the terms on which this occurred meantthat union officials agreed to discipline theirmembership and prohibit strikes. In order to gain somebenefits for a dissatisfied rank and file, union leaderswere moved to strengthen further their connexions 'at thetop' through wage-setting government bodies and industry-wide collective bargaining" (Stark, 1980, p. 109).

26

industrialists was still commonly found in the textileindustry, the railways and the machine tool industry(Stark, p. 112).

The history of Taylorism demonstrates the advantages tobe gained from putting forward a theoretical framework inwhich economic analysis can take place and which allows atechnique form to be related to a method and to a socialgroup. It has been seen that Taylorian investments in form,although they were mainly carried through by the companies,established highly "equipped" forms with an extended areaof validity, such as the definition of a standardized task,the coding of time, the formulation of rationalizedrelations by experts, etc. It can therefore be quite easilyunderstood that such forms articulate extremely well withmore general forms established or validated at nationallevel, such as scientific rationalism, legal formalism,bureaucratic regulations, and national representation forsocial groups in arenas established for collectivenegotiations. The same may be said of the articulation thathas been noted between the spread of Taylor's methods andthe waves of standardized orders coming from the State intime of war.

Similarly, it is not difficult to understand how thespread of Taylorism meant a phasing out of specificinvestments in the customs and rules of work built up byskilled workers at time as more nation-wide forms of unionrepresentation were strengthened.16

16 Stark continues Montgomery's work (1976) and points todifferent "levels of control" among groups of skilledworkers who did indeed benefit from the differences inrange of validity and the degree to which forms wereobjectified that have already been discussed: "The firstlevel of control was achieved through informal organisationat the local level" (my italics). The "mutualist ethicalcode", as Montgomery terms it, included rules of conductabout production quotas, relations with the boss and withworkmates. Certain groups of skilled workers who wereunion members extended and consolidated such rules withinoccupations. "In the last decades of the century, theprovision of the craftsman's moral code developed a

27

6. WHAT DETERMINES FORM INVESTMENTS IN COMPANIES

Though he does not realize it, Taylor unwittinglydefends a model which is diametrically opposed to his ownwhen he writes, in the Principles, that the worker in thefactory,

instead of using every effort to turn out the largestpossible amount of work, in a majority of cases (...)deliberately plans to do as little as he safely can(...). [Yet ] wherever an American workman playsbaseball or an English workman plays cricket, it is safeto say that he strains every nerve to secure victory forhis side. He does his very best to make the largestpossible number of runs. The universal sentiment is sostrong that any man who fails to give out all there isin him in sport is branded as a 'quitter' and treatedwith contempt by those who are around him (p. 13).

Thus Taylor offers a glimpse of a model quite foreign tohis thinking, a model which is based not only on the rulesof the game but on the qualifications of the players, andone which leaves a large amount of room for the play ofhabits and practices which are hardly objectified except astactics. In this model the company depends on investmentswhich are at least partially borne by its employees. It may

higher form of craft control in the enactment andenforcement of union work rules which the craft unioniststermed 'legislation'. These work rules Indicate theestablishment of a second level of control forged throughconnexions between groups of workers within the samecraft in different locales." Finally, the unioninstitution helped to formalize and generalize theserules: "The third level of control emerged when sympathystrikes involving members of different trade unions wereconducted to support attempts to enforce these work rulesor to win union recognition. Skilled craftsmen were thusforging important connexions across locales andindustries" (Stark, 1980, p. 99).

28

reward qualifications validated by the State by, as itwere, renting them out, or may make economies by usinginvestments with more limited validity that are less easilycapitalized on, but not less productive. Thus largecompanies such as multinationals with a very clear publicimage, whose workforce management resembles the "game"model and facilitates investments specific to the company,tend to avoid recruiting staff who hold nationallyvalidated qualifications which are therefore valid outsidethe company. Though they do not follow Taylor in believingthat skills and qualifications serve no good purpose, theyprefer to realize family investments and find thatrecruiting people known to them is an appropriate way ofdoing so. But they do, therefore, encourage in-houseinvestments such as a company hierarchy with its own jobtitles, carrier planning etc. (Desrosieres and Pialoux,1983, pp. 81-2). The presence of long-standing employees isa good indication that such a model has been adopted(Eymard-Duvernay, 1981).

When a management model is changed, this provides anexcellent opportunity to examine such investments. ThusGouldner's classic account, which applies Weber's model ofbureaucratic organization to an industrial company, isbased on a study of what happens when a new boss takes overand tries to undo the forms of the previous organization("the indulgency pattern"). The organization in questioncould have been described as having forms and practicalrelations which were not strongly objectified, not heavily"equipped", and, contrary to all that Taylor's principlesrecommend, as making great use of specific investments inrelations outside the company, such as religiouscommunities, voluntary firefighting organizations, sportsteams, and so on. Certain specific properties were evenless formalized: they were formed through repetition andhabit and gave rise to what are commonly called "roots". Ashas been said elsewhere, seniority is a good measure ofthis kind of specific, weakly objectified, form (Eymard-Duvernay and Thevenot, 1983a). The company studied madeextensive use of relations formed in this way as well as offamily connexions which were systematically taken intoaccount in recruitment. Such connexions could cut across

29

the establishment's hierarchical divisions, and thecompany's investment in specific forms was not great. Thecompany regulations were neither known nor observed andtime had not been established as a unit of account formeasurement and control. The division between the company'stime and the employee's time was not marked by analternation of work periods and breaks.17 The change ofmanagement involved a number of costly investments in formsspecific to the company. The limits of the company wereestablished in a more hard and fast way as was thedistinction between what belonged to the company and whatwas the employee's private property. This led, for example,to the sacking of a workman (although he had been with thecompany for twelve years) who was accused of stealing acase of dynamite. It also involved recruitment based oncriteria that were "impartial" with regard to family orother connexions, the introduction of daily and weeklyreports and of company regulations, the strict applicationof clocking on and off, etc. Unlike his predecessor, thenew manager was a graduate, and this accorded with Weber'smodel of legal authority in which jobs which are thought ofas "offices" are articulated with the qualities of theholders which are "guaranteed by diplomas certifyingtechnical training" (Weber, 1964).

In the case of the company (an "industrial monopoly")studied by Crozier as an example of bureaucraticorganization, which therefore derives from the previousmodel, the opposite transformations were to be found. Formsthat were characteristic of the Weberian bureaucratic type,

17 "Your free time is your own (. . .). When there's workto do they expect you to do it, but otherwise they leaveyou alone" declared one worker interviewed (p. 47).Similarly, an exhibition space served as a sort of"informal" private area reserved for the use of employeessuffering from the effects of accidents (includingaccidents which occurred in their homes) which allowedthem to have a higher income than if they spent time athome recovering. In the same way, there was no cleardistinction between company property and the employees'private property as was indicated by the fact thatmoonlighting was widely practised in this company.

30

because of their "generalized formulation" and their"written recording", were phased out and replaced byspecific forms valid within an occupational group: "Themaintenance workers have succeeded in getting the plans ofhow the machines work and the maintenance guidelines removedfrom the shop floor, and have got an agreement thatmaintenance policy will be based on individual adjustmentsthroughout" (Crozier, 1963, p. 90) (my italics).

Unfortunately, Gouldner's frame of reference isrestricted to work relations and does not allow thepreceding observations to be linked to others which concerntraditional economic variables. He only remarks in passingthat big investment in new materials (1.5 million dollarsin 1950) was accompanied by the establishment of new codeforms. However, if one accepts the idea that creating arule is as much of an investment as purchasing a machine,then one is led to re-examine, in the light of this model,some company functions which are traditionally contrastedwith production functions.

Thus the creation of a personnel department is aninvestment which implies that jobs will be classified, andrules or objectives for recruitment, job mobility,retirement, etc., will be established. One detailed examplemight be taken from the study (Eymard-Duvernay andThevenot, 1982) of what happened when a cement group thatwas very capital intensive and produced a very standardizedproduct, partially took over a company which made ready-mixed concrete for use on building sites. This companytherefore acted as an agent or, in some respects, awholesaler dealing with different customers scatteredthroughout the country. The cement group, through theactivities of one of its executives who was appointeddirector of the company, tried to match the company's formsof activity to those of the cement producers who suppliedit. The study referred to examined some of these form-giving activities which involved "working on the labourforce", such as investing in new "tools" (as the directorhimself put it) or getting a works agreement on workinghours and their equivalent forms.

31

The new director also tried to replace the informalnegotiations between vendors and purchasers by formal rulesbased on objective criteria. The fact that cement is a verystandard kind of product was helpful in drawing up suchrules,18 as were the large number of orders for the standardproduct which conformed to standard norms. This thereforeprovided a further example of the articulation between theestablishment of generally valid forms and the governmentorders and contracts already referred to in the abovediscussion of the conditions which encourage the spread ofTaylorian methods. On the other hand, personal habitsarticulate badly with such forms, and this was the casewith one client, a site foreman, who had ordered someconcrete and "guessed" the amount that had been objectivelychecked by the vendor. Some material investments werecarried through, such as the computerization of productionprocesses and of deliveries to outlying establishments (theproduction centres), and these necessitated accompanyinginvestments, in form. Moreover, investment in computers isof a mixed kind, since it involves installing machines andalso introducing rules and forms (especially the code form)that are compatible (Thevenot, 1983b). The job of technicalassistant (agent technique) was, therefore, created withresponsibility for ensuring that the highly regulatedmanagement of planning, and deliveries and contacts with

18 Conversely, Crozier observes that in the company hestudied, "the lack of homogeneity and the variablequality of the raw material" had unpredictableconsequences (which would halt the machines) (Crozier,1963, p. 130). If one accepts Crozier's proposition that"each individual's power depends on how unpredictable hisbehaviour is and to what extent he can make theachievement of common objectives uncertain" (id. p. 10),it might be said that variability, which marks the limitsof the formal laws of the company model, demands the useof more specific and more personalized forms in order tomake possible the interpretation the stabilization ofthese variable events, and their articulation with thismodel Thus variability makes such investments influentialand also gives great influence to the occupational group(in this case the machine adjusters) within which theirvalidity was constituted.

32

customers, worked smoothly together. The job was thereforecreated to improve the company's negotiating position whichthe new director considered too dependent on the vagariesof customer demands. Both the general training of those whowere recruited and a commercial training programme wereintended to increase the degree to which negotiations couldbe regulated. These new technical assistants were notrepresented by the two trade union branches in the company,the Confederation General du Travail (CGT) branch to whichthe manual workers and the drivers belonged, and theConfederation General des Cadres (CGC) branch whichincluded the supervisors and the managers. So the technicalassistants created a Confederation Française Democratiquedu Travail (CFDT) branch. In addition, at the request ofthe CGT and the CFDT, jobs which included the word"technical" in their title were shifted from the firstelectoral college (composed of manual and clericalworkers)19 to the second electoral college (composed ofsupervisors). The director considered that with theseelectoral divisions and the unions they corresponded to(the CGT had all the seats in the first college but only afifth of those in the second) he had a "more balancedstructure".

Models which view capital immobilizations as the onlyproductive investments that can complement or besubstituted for the use of a workforce, cannot account forcases where it is not the capacity to produce but ratherthe opposite which determine employment—except in anindirect way, through the construction of "delays inrecruitment" and the "cycle of productivity". This was thecase in the company studied which wanted new equipment(bigger trucks) so as to increase productivity. Thisinvestment— i.e. the truck acquisition programme — couldonly be carried through by observing a long-establishedcustom that the personnel department had changed into arule, that job changes should occur as a result of naturalwastage rather than redundancies. The setting up of anaccounts department and the introduction of managementmethods are even clearer examples of specific investments

19 An electoral college for elections to the variousrepresentative bodies such as the Comite d'entreprise.

33

designed to create relations and equivalents whichguarantee regular functioning and links with compatibleinvestments outside the company.20 However, the operationswhich consist in "making the rules work" have nothing to dowith such investment measures. This distinction generallyimplies that the holders of such posts have lower salariesand are different kinds of people as regards both theirsocial origins and their educational qualifications.21 It isalso interesting that for those executive posts whichinclude among their responsibilities investment in formssuch as the constitution of rules and objectives as well asthe articulation of these investments with forms that arevalid beyond the company (finance, management, accounts),the most highly qualified executives are recruited. Theseare people who are able to capitalize on their education ina national market and are therefore accustomed to usingstandard variables. The foregoing analysis also makes itpossible to reinterpret the distinction in the socialorganization of businesses between the "substantive norms"which govern the way a company works and which result from"decision-making procedures" and the "procedural norms"which govern the formulation of the substantive norms (Fox,1971, pp. 28-30).

7. STATE FORMS

The preceding sections have attempted to outline ananalytical framework which would allow a broad range of

20 In his detailed study of work organization in seven massproduction companies Durand notes that the two companieswhich have no methods department are those in which "allproduction has been computerized and automated" and onewhere "all organization is totally empirical" (Durand,1978, p. 19), that is a company which makes use of themost general forms with no investment in specific forms,and a company which has undertaken no investment in formswhatsoever.

21 Such rules are most highly regulated and automated inthe "functional model" as opposed to the "professionalmodel" (Benguigui, Grise and Mont3 , ).

34

material or formal "equipment" to be related to a singlemodel of investment in forms that are characterized bydifferent parameters including their area of validity. Suchan attempt is of interest primarily because it makespossible an understanding and a comparison of phenomenawhich are usually placed in a variety of differentconceptual frameworks. Since these conceptualizationsderive from a number of different disciplines, they tend toconstitute objects which appear to be naturally verydifferent: product, trademark, machine, capital, accounts,skill, qualification, collective agreement, etc. Taylor'smachinery, as set out in his handbook, can easily bereinterpreted within a framework of analysis such as thiswhich offers a good account of the conditions under whichit will succeed or fail. Taylor's equipment sets up durableand standard relations that are hardened into material andformal implements and therefore become more productive. Butit is only profitable if it is based on the use of formswith general validity, forms which are often public andinexpensive because they have already been constituted bythe State, the measurable forms of the objects used and theanonymous forms, such as the civil status, of theindividuals involved. The study of companies which do notconform to the Taylorian model but make use of their ownmore specific forms, which are particularly evident in theway they recruit and more generally invest in theirworkforce, suggests that great importance should beattached to the area of validity of the forms studied.Therefore, what will be discussed in this final section arethe effects of extending this area of validity and theconstitution of more general forms which will be referredto as State forms in order to indicate that their validityhas been extended, even if they are not State forms from astrictly legal point of view. By examining how generalforms are created and therefore how equivalents areproduced, titles (titres) of monetary or other nature whichregulate credit conditions in the interaction of people orcollective bodies, it is possible to reveal the investmentswhich take place prior to the creation of a market and totake a further look at the "imperfections" thatneoclassical theories criticize or would wish to alter.

35

Numerous examples of investment in general forms areprovided in the work of anthropologists on tribal councils"[that] are governed by conventions and persist in time"(Richards, 1971, p. 1), considered as "the machinery bywhich [group] decisions are reached". The implementssupplied by established rules and repeated habits tend toproduce an anonymous relation that is durable and materialinstead of an ephemeral personal opinion. Richards liststhe component parts of this machinery: there must be afixed meeting place, an unchanging spatial disposition ofmembers and their groupings, a speaking order, standardrhetorical forms of address and formulae for expressingapproval, the language and gestures of adjudication, andprocedures for producing and recording a decision. Theyield from these investments is clearly related to the formtaken by the components of the machinery and the endproducts which, because they achieve general anonymity,prevent renegotiation of the decisions of the dulyconstituted body and sometimes even confrontations, i.e. theeffective mobilization of individuals in the field(Boltanski, 1983). But in the case of less formallyconstituted councils, the alteration of decisions isavoided by very informal discussions consisting of a seriesof "mutually exclusive statements", without any argumentthat might bring out contradictions, which frequently leadto ambiguous decisions (Bloch, 1971, pp. 50-52). Byexamining the somewhat loose relation between the way suchcouncils work and the actions that result from them, thebasic conditions for the production of State forms becomevery clear. Thus, when a decision has not been formallyreached by the council, it is the number of members whosupport the proposal which influences the action taken,even though there is no explicit mechanism for countingthem such as the vote. It may also be an individual councilmember's "influence" which sways the decision so that thecouncil becomes the place where relative influence isassessed and reassessed. Finally, because there is completeabsence of investment in representation, results depend oncounting individuals who turn out on the appointed day insupport of one or another of the proposals.22 Because this

22 "The discussion happens as a series of mutuallyexclusive statements. One person makes a speech the gist

36

council machinery tends to abstract from the personal andproduce general decisions, it articulates with theconstitution of collective personae—in other words, withthe way individuals are represented. This is a fundamentalrelation which has already been emphasized, both withregard to the classification of the employees within acompany, and with regard to the occupational and tradeunion groups involved in Taylorism, and it is central tothe working of councils: "the principles by whichmembership of any council is fixed must be directly relatedto the forms of social differentiation in the society"(Kuper, 1971, p. 15). Kuper adds that this is a two-wayrelation: "social status may be affected by performance incouncil" (ibid.). In certain cases the principle ofrepresentation in the council is unified, and this mayresult in several councils co-existing each of which isformed according to different principles of classificationsuch as age and lineage (Jones, 1971). But often this workhas not been accomplished and heterogeneous principles ofclassification (women, strangers, elders, sects, trades,

of which is a proposal to come on a particular day. Thismay well be followed by another speech which seems to bein support of the proposal and full of praise for it, butin fact contains, hidden within the mass of sentences, acounter-proposal for another day. There is no argumentand it is very difficult to realise that the statementsare contradictory. What is more, no decision seems to bereached at the time. However, if a large number of peoplemention one time rather than another, everyone knows thatthis is the right time. O en, however, the matter is leftin the air and the chosen day is understood to be thatproposed by the more influential man. In this way thesecontradictions may be tests in a power struggle betweendifferent individuals. Very often it is not clear to theparticipants at what time they should turn up. If this isso, on the first day mentioned the supporters of one sidewill start to gather. If they are few, they will soondisperse. If, on the contrary, they are many, they willbe joined by waverers and then the whole thing snowballs,until ultimately perhaps even the proposer of thealternative day will be drawn in" (Bloch, 1971, pp. 50-51).

37

etc.) are to be found within a single council (Robertson,1971, pp. 149-155).

Thus the working of State machinery such as this isindissolubly linked to the production of qualifications (titres),that is, personal qualities which make equivalents possibleand therefore allow an individual holder to beindependently treated. Bourdieu's work has thrown light onthe difference between these qualities, which make possibleboth equivalents in a unified market and economiccalculation, and the "embodied" properties on which dependthe "good faith economies" of gifts and exchanges betweenindividuals (Bourdieu, 1980). His analysis of the build-upof confidence necessary for friendly transactions in whichthe purchaser may attach more importance to the choice ofvendor than to that of the product whose quality isinseparable from that of the person selling it (ibid. p.196), has been extended to include the system Orqualifications (titres) guaranteed by the education system(Bourdieu, Boltanski and de Saint Martin, 1974; Bourdieuand Boltanski, 1975; Bourdieu, 1978).

The now fairly widespread analogy between the process ofeducational expansion and monetary inflation isparticularly productive in French since the word titre canmean not only a certificate, a title, an entitlement and aqualification, but also a degree of "quality" (alcoholcontent, gold standard, etc.). Thus as long as thedefinition is kept firmly in check, the use of the wordtitre may permit not just useful illustrations but fruitfulcomparisons. If one refers back to the definition ofinvestment that has been put forward in this paper, it isclear that a person who has a title/entitlement as a resultof a certificate does not have the same form as a financialtitle, since that form remains specific to the person whobears the title. However, since the State has institutedqualities which, although they are attached to persons,have the durability, the area of validity and theobjectivity of law, this has economic effects arising fromthe establishment of equivalents similar to those observedwith financial titles. Indeed, it is because the formalizedtitle is so stable that it can enter into temporalrelations of commitment and anticipation, just as it is

38

because the extent of its validity is guaranteed by theState that it may serve to establish equivalents betweenpeople who are otherwise very different. Thus"qualifications inflation" serves as a convenient metaphorto describe their explosion only to the extent that nosatisfactory account of the work of establishing this Stateform is proposed. But if the characteristics of this formare taken into account, this may explain effects similar tothose observed with monetary forms. Thus small employers'frequent mistrust of educational qualifications, which isreferred to in the various employment agencies in terms ofa "disparity" between the skills and qualificationsrequired and the content of the education and trainingreceived, may be explained by the huge gap betweenpersonalized forms of recruitment and those which are moreobjectified .

Under the first model, the employer whose company is runindividually or on a family basis is concerned to maintain,in personal form, all the relations he uses to obtain andretain the services of an employee. Thus the propertieswhich mean that an employee is taken on are recognized fromsigns which are not highly coded, are often physical, andmay not even necessarily be made explicit. If an employee wereto be recruited on the basis of his qualifications (sur titre), this would betantamount to giving him credit for (au titre) a qualification instituted by theState which creates an equivalent. In such a case, the powerrelation between employer and employee is mediated throughthe qualification equivalents that the State creates.Because of similarities with the effects of investment ingeneral forms which serve as equivalents the process beginsto resemble those that have been analysed in relation tomoney (Aglietta and Orlean, 1982; Orlean, 1982).23 Indeed,Hicks' analysis of the effects of automatic borrowing

23 See especially their analysis of "central money","homogeneous money" and "private money": "From amechanistic point of view money either is or is not; webelieve, however, that it is always in the process ofbecoming. It is a socialization process which mayexperience various degrees of extension, and at each ofthese levels it appears in a qualitatively differentform" (Aglietta and Orlean, 1982, p. 82).

39

facilities via bank credit might well be applied herethrough a comparison of highly standardized debt forms,regulated by the central banks, and the more private formsoffered by the security markets (titres financiers). What hasbeen described here as the costly investment in stableforms with general validity has the effect of creatingspatial and temporal equivalents.

Aglietta and Orlean discuss this question from a similarpoint of view when they analyse the effects of equivalence(which they call "monetary arbitration") on propertyrelations between social agents and the inflationaryprocesses that result from them.24 Within the limits of thecomparison outlined above between an individual who haseducational qualifications (scolairement titré) and a financialsecurity (titre), it may be said that by means of diplomasand certificates, the' State institutes forms with titleswhich make it possible for individuals to be equivalent,and also allow links with other code forms to be set up.This occurs, for example, when qualifications are"recognized" in the job classifications drawn up incollective agreements, and such a link is an excellentillustration of the fact that those who hold qualifications

24 "Inflation is a situation in which the centralauthorities try to support the strategy of the hegemonicsocial strata which have become indebted. They do this byincreasing the monetarization of deficits and thereforetransforming the previously established arbitrationbetween creditors and debtors. Creditors are thereforeplaced in a difficult position: they see their monetaryreceipts increase but not their economic power, i.e.their influence over productive activity. Indeed, thetransaction A (j)—U (i), which means that the creditor jcould use his financial capacity to acquire, through thepurchase of the debt U (i), an interest in the productioncentre i, is prohibited by the structure of thecentralized relation i/X/j. As has been seen, the endresult of monetary mediation is the separation ofcreditors and debtors; thus no institutional procedureexists to allow such a transaction to take place. This isprecisely the reason why centralization is able to ensureregulation" (Orlean, 1982, p. 98).

40

accredited by the State are able to get them generallyrecognized by employers. The result is that the powerrelation between the private employer and his employee isdistorted in just the same way as the creditor-debtorrelation is distorted through automatic bank credit.

It can therefore be understood why small employersrefuse to accept the power shift implied by the recognitionof educational qualifications and why they are reluctant torecruit employees on the basis of their qualifications (surtitre). If State qualifications were applied everywhere—if,in other words, no reduction in the value of qualificationsoccurred because of such refusals—whole social strata ofqualification holders entitled to something from the Stateand to whom the State "owed" something, would have theirnegotiating position in employer-employee relations greatlystrengthened as a result of the institution of the Statequalification (titre) form.

But when faced with these "State credits" that they donot control and which undermine their specific power torecruit a workforce, employers may adopt an approach morecomplex than purely and simply refusing to accredit Statequalifications. Thus larger companies which have undertakenlarge-scale, specific investments in forms with regard totheir workforce (specific job descriptions, workenhancement methods, company training and promotionprogrammes, etc.), do not recruit individuals who haveState vocational qualifications with general validity,since this would cancel out the effects of the company'sown specific investments. They may, however, take onholders of the baccalaureat as manual workers, therebyensuring that the qualification loses some value. But theyalso ensure that it is made equivalent to investmentsspecific to the company when they guarantee theseyoungsters with devalued qualifications a quick promotionto the posts of foreman or technician provided they doeverything the company expects of them. The oppositeprocess occurs, however, through "ratification"(homologation) procedures which allow job-specific

41

educational qualifications to become more widely valid.25

Homologation does not make possible such a strong form ofidentity as the equivalence26 procedure controlled by theMinistry of Education, which involves a personal representativeof the Ministry sitting on examining boards. Homologationtakes place through negotiations by a less formal committeewhose members include "personalities" as well asministerial representatives. Its object is to achieve widerrecognition for qualifications which are relatively limitedin validity such as those awarded by private schools(though not by companies). Such qualifications are lessanonymous than the most official educational qualificationssince they are awarded on criteria which include aconsideration of the holder's earlier properties (the levelat which he was recruited and his work experience) as wellas the training period that is being assessed, and they do

25 On qualifications losing value between 1973 and 1979,see Affichard (1981). On the development of theclassification of qualifications and for an analysis ofhomologation procedures see Affichard (1983).

26 The equivalence is obtained through an administrativeprocedure which allows (often foreign) qualifications tobe recognized for the purposes of exemption andadmission.

27. Three kinds of statuts may be distinguished: "statutsoctroyés" (conditions which are granted) which have theirsource in a "policy pursued by the employer who, as faras industrial relations are concerned, looks neither foragreement with the trade unions nor, indeed, for any formof dialogue with them", the "statuts concertés" (consultativeconditions) where the constraints of collectivebargaining agreements do not apply but where, as the CNPF(French employers' organization) puts it, "the employerstake care to consult and to ask questions in orders tobroaden the basis of decision-making"; "statuts contractés"(negotiated conditions) which involve "explicitrecognition, on the part of the employer, that the tradeunion side has a right to be consulted and to negotiaterelations at work" and where "negotiations conclude withsigned agreements" (Jobert and Rozenblatt, 1983, pp. 9-10).

42

not automatically provide the same rights (such as accessto other jobs or additional training).

In a more general way, forms which, because of theirlifespan and their area of validity, are between Stateforms and more specific company forms, are constitutedwithin the framework of collective agreements. Thenegotiating machinery and the bodies involved, as well asthe rules which are established, may in fact only be validwithin the company (through the comite d'entreprise (jointworks council), representatives of management and workersin the company, "company conditions": company rules, jobclassifications, wages, work conditions, representation,etc.), the original model for which was the 1955 Renaultagreement. They may also be valid only in relation to anoccupational sector in one region, as with the manyregional agreements which resulted from the strikes of1936; or their validity may apply to a whole industry (andmay include both employer organizations and trade unionsand collective agreements covering whole industries; withnational classifications, redundancy payments, pensions,and so on) (Jobert and Rozenblatt, 1981). As has been notedabove, when the validity of these forms is extended theirlifespan is usually extended as well. An agreement whichapplies only to one company is weaker than a collectiveagreement applying to a whole industry which amounts to apermanent "charter", whilst some fundamental rules (such asredundancy payments, pensions, and so on) are always laiddown in industry-wide agreements (Jobert and Rozenblatt,1981, p. 6). But as has also been said, the lifespan offorms also depends on the cost of their establishment, andthis is high because they are not simply decreedunilaterally, but are negotiated in all due form with unionrepresentatives who therefore take partial responsibilityfor them.27 But though establishment costs are high,

27 Three kinds of statuts may be distinguished: "statutsoctroyés" (conditions which are granted) which have theirsource in a "policy pursued by the employer who, as faras industrial relations are concerned, looks neither foragreement with the trade unions nor, indeed, for any formof dialogue with them", the "statuts concertés" (consultativeconditions) where the constraints of collective

43

maintenance costs are low by comparison with those neededfor specific investments unilaterally undertaken bymanagement which lead to much contention and many disputesand legal proceedings. Such disputes are particularlycostly in France where, unlike the United States there iSno grievance procedure which makes possible conciliationwithm the company (Bonafe-Schmitt, 1982),so that the causesof the dispute are broadened and personal grievances aretransformed into collective demands.28 The level of formalequipment of this kind will very much depend on how broadlybased trade union representation is. It is more difficultto mobilize large numbers of people effectively throughoutan industry, but if mobilization is limited to a singlecompany (and unless it is, it cannot succeed), this can actas an obstacle to unification and to action being extendedover a wider area which is necessary if the collectivepersonae (in the form of union representatives) are toachieve influence and national stature. What determineswhich level of activity is chosen is often the length oftime that trade union delegates have held their position,with those who have served longest being more expert inregional and national negotiations (Jobert and Rosenblatt1983, pp. 195-l96).

bargaining agreements do not apply but where, as the CNPF(French employers' organization) puts it, "the employerstake care to consult and to ask questions in orders tobroaden the basis of decision-making"; "statuts contractés"(negotiated conditions) which involve "explicitrecognition, on the part of the employer, that the tradeunion side has a right to be consulted and to negotiaterelations at work" and where "negotiations conclude withsigned agreements" (Jobert and Rozenblatt, 1983, pp. 9-10).

28 The new labour legislation (the lois Auroux) containsprovisions for compulsory negotiation as well asguaranteeing the employees the right to express theirviewpoints or their grievances and will thereforereinforce the tendency for the company to be the basiclevel at which disputes take place and are resolved. Thesame legislation requires rules and penalties to be givenobjectified, written form (articles L 122.40 and L 122.34of the Code du Travail).

44

In West Germany, all collective negotiation machinery isbased on State forms, in such a way that Stateintervention, in the strict sense of the word, is much lessgreat than in France. The framework of negotiation isstrictly defined, as is the date and the lifespan (four orfive years) of the contractual agreement (there are nostrikes over contractual clauses). The way in which suchagreements articulate with more specific forms of agreementis also laid down. Company agreements only relate tomatters which do not figure in the collective agreementand, conversely, the terms of the collective agreement mustrefer back to an agreement within the works council(Jobert, Rozenblatt and Tallard, 1980, pp.113-115).

Invested forms have such different characteristics, bothas regards the costs involved in establishing them and theeffects they may have, that it seems necessary to lookcritically at all models which formally relate equipmentwhich varies in this way. Although according to the "humancapital" theory (Becker, 1980) it seems interesting toextend the notion of investment when that notion coversoperations of commitment over time that are not taken intoaccount within the classical framework, merely treatingthese operations as equivalent, by means of econometricformalizations, cannot be justified by referring to amarket in which interests may meet. Indeed, the varieddegrees of formalization, durability and validity of aState educational qualification, of training gained on thejob, and of an individual physical property, prevent thesethings being related to a single monetary equivalent sinceit is too difficult and therefore too costly to do soeffectively.29

29 A concept such as "firm-specific human capital" (Becker,1980) attempts to grasp these differences. In a moregeneral way, a whole set of theoretical research has,within the neoclassical framework, tried to get beyondthe limits inherent in the perfect market model and totackle questions which have a bearing on the onesdiscussed here. Such theories would include the "theoryof the implicit contract" which is a mutual agreementbetween employer and employee in an environment of

45

Given the significance of general forms and theirusefulness in ensuring the regular and economic functioningof theoretical formulae as well as management principlesand machines, it is tempting to look on their developmentas an indication of regular progress towards greater use ofmore standard and more rational tools and the disappearanceof older more specific and less objectified forms. It istherefore important to conclude by emphasizing that this isnot the case, and that such an evolutive picture would beextremely inadequate in the light of what can be seentaking place, especially at a time of recession.

lt is impossible to deny that standardization hasgathered its own momentum by the consolidation of equipmentthat it produces. But it cannot avoid coming into conflictwith the individual qualities of those who, at any givenmoment, phase out, invest in or maintain the forms in use,and who use such forms as a resource, thereby creatingeffects of differentiation. The process of differentiationhas been studied in the life-style, practices andconsumption habits which may be found in different socialmilieux and which condition the reproduction of thosemilieux (Bourdieu, 1979). An analysis of code forms makespossible an understanding of the link between theoperations marking individual properties which make upidentity, the criteria which form collective bodies, andthe qualities which define products.30 The effect ofextending equivalents, on producers and users of morespecific forms, has been discussed above with relation toeducational qualifications (titres) and financial securities(titres). As far as the definition of products, implements

uncertainty, explaining "seniority based lay-offprocedures"; "screening" theories or "signalling"theories which discuss the conception of the coding ofhidden properties (of the employee) in order to identifycases where imperfect information about the quality ofthe person has been supplied. For a discussion of thesetheories see Eymard-Duvernay and Thevenot (1983a, 1983b).

30 For the modes of coding and decoding of characteristicswhich serve to shape social identity and to make itrecognizable, see Boltanski and Thevenot (1983)

46

and skills is concerned, investment in standard formsencourages greater circulation,31 as is shown by the factthat their sale becomes possible (through licensing,franchising, engineering, etc.). Such investment thenreduces the copying cycle, increases competition and endsup by stimulating the creation of specific forms inopposition to the standard forms, and this process can beseen in the textile and clothing industries, in steel (withspecial steels), in the motor industry (where the varietyof ranges and models prevents norms being applied to allproduction lines), etc. It would appear that it iscompanies which make use of resources of this kind whichhave best weathered the recession, and in order to describetheir economy, it would seem necessary to include in anyanalysis the fundamental characteristics of such resourcesas I have tried to define them in this paper.

Laurent Thevenot (born 1949) is Administrateur at theInstitut National de Statistique et des Etudes Economiques.After work on social categorization and on theestablishment of a new official nomenclature ofprofessions, subjects on which he has published severalarticles, he is now engaged in a study of coding behaviouras exemplified in the present article. Author's address:INSEE, 18 Bd. Adolphe Pinard, 75675 Paris Cedex 14.

REFERENCES

Affichard, J., 1981 "Quels emplois après l'école: la valeurdes titres scolaires depuis 1972", Economie et Statistique,134, 7-27.

Affichard, J.A., 1983 "Nomenclatures de formation etpratiques de classement", Formation Emploi, 4, 47-61.

31 "The more standardized the product and routinized theproduction process the easier lt Is for unsophisticatedmanagers, engineers and workers to copy" (Sabel, 1982,ch. 5).

47

Aglietta, M., 1976 Régulation et crise du capitalisme, Paris,Calmann-Levy.

Aglietta, M., Orlean, A., 1982, La violence de la monnaie,Paris, PUF.

Becker, G.S., 1973, "A theory of marriage, Part 1", Journalof Political Economy, 81, 4.

Becker, G.S., 1974, "A theory of marriage, Part 11", Journalof Polifical Economy, 82, 2.

Becker, G.S., 1980, Human capital, Chicago, University ofChicago Press, 2nd. edition (Ist. edition, 1964).

Benarroch, F., Espinasse, J.M., 1982, "Les salariésrecrutés en 1980; caractéristiques et moded'embauche", Travail et Emploi, 11, 71-81.

Benguigui, G., Montjardet, D., 1975 "Recherche sur lafonction d'encadrement", Paris, Université de ParisVII, Groupe de Sociologie du Travail.

Berger, P., Luckman, T., 1967, The social construction of reality,London, Allen Lane.

Bloch M., 1971, "Decision-making in councils among theMerina of Madagascar", pp. 29-61 in: Richards; Kuper(eds). Op. cit.

Boltanski, L., 1970, "Taxinomies populaires, taxinomiessavantes; les objets de la consommation et leurclassement", Revue Française de Sociologie, 11, 34-44.

Boltanski, L., 1982, Les cadres; la formation d'un groupe social,Paris, Minuit.

48

Boltanski, L., Thévenot, L., 1983, "Finding one's way insocial space; a study based on games", Social ScienceInformation, 22, 4/5, 631-679.

Bonafe-Schmitt, J-P., 1982 "La pratique de négociationcollective dans l'entreprise", Revue Francaise des AffairesSociales, 36, 167-181.

Bony D., Eymard-Duvernay, F., 1982, "Cohérence de labranche et diversité des entreprises; étude d'un cas",Economie et Statistique, 144, 13-23.

Bourdieu, P., 1978, "Classement, déclassement,reclassement", Actes de la Recherche en Sciences Sociales, 24, 13-23.

Bourdieu, P., 1979, La distinction, Paris, Minuit.

Bourdieu, P., 1980, Le sens pratique, Paris, Minuit.

Bourdieu, P., Chamboredon, J-C., Passeron, J-C., 1968, Lemétier de sociologue, Paris/La Haye, Mouton.

Bourdieu, P., Boltanski, L., de Saint-Martin, M., 1974,"Les stratégies de reconversion; les classes socialeset le système d'enseignement", Social Science Information,12(5), 95-107.

Bourdieu, P., Boltanski, L., 1975, "Le titre et le poste;rapports entre le système de production et le systèmede reproduction", Actes de la Recherche en Sciences Sociales, 2,95-107.

Boyer, R., 1979, "La crise actuelle: une remise enperspective historique", Critique de l'Economie Politique, 7-8,5-113.

Broverman, H., 1974, Labor and monopoly capital; the degradation ofwork in the 20th century, New York, Monthly Review Press.

49

Callon, M., Latour, B., 1981, "Unscrewing the bigLeviathan", in: Knorr-Cetina, Cicourel (eds). Advancesin social theory and methodology. Boston, Rout!edge and KeganPaul, pp. 277-303.

Cicourel, A., 1964, Method and measurement in sociology, London,The Free Press of Glencoe.

Commons, J., 1934, Institutional economics, New York, Macmillan.

Cornfield, D., 1982, "Seniority, human capital, andlayoffs; a case study", Industrial Relations, 21(3), 352-364.

Crozier, M., 1963, Le phénomène bureaucratique, Paris,Seuil.

Desrosieres, A., 1977, "Eléments pour l'histoire desnomenclature socioprofessionnelles", in: Pour une histoirede la statistique, t.l. Paris, INSEE, pp. 155-231.

Desrosieres, A., Pialoux, M., 1983, "Rapport au travail etgestion de la main-d'oeuvre; problèmes de méthodes",Critique de l'Economie Politique, 23-24, 66-85.

Desrosieres, A., Goy, A., Thévenot, L., 1983 "L'identitésociale dans le travail statistique; la nouvellenomenclature des professions et catégoriessocioprofessionnelles", Economie et Statistique, 152, 55-81.

Devouassoux, J.; Morel, B., 1983, "Le diagnostic; entre lapratique médicale et la collecte statistique", Actesdu Colloque du Conseil National de la Statistique (18-19 April 1983), Paris, INSEE.

Dubois, P., Durand, C., 1983, "Les politiques patronalesd'innovation", Critique de l'Economie Politique, 23-24, 144-161.

Dumett, M., 1975, "Wang's paradox", Synthèse, 3/4, 301-324.

50

Durand, C., 1978, Le travail enchaîné, Paris, Seuil.

Elias, N., 1973, La civilisation des moeurs, Paris, Calmann-Levy(1st. ed. in German, 1939).

Elster, J., 1979, Ulysses and the sirens, Cambridge/Paris,Cambridge University Press/Editions de la Maison desSciences de l'Homme.

Eymard-Duvernay, F., 1981, "Qualification, poste, salaire;etude sur l'industrie horlogere", in: "Les categoriessocioprofessionnelles et leur reperage dans lesenquetes", Archives et Documents 38. Paris, INSEE.

Eymard-Duvernay, F., Thevenot, L., 1982 "L'économiste etson modèle". Paris INSEE. (Mimeo.)

Eymard-Duvernay, F., Thevenot, L., 1983a, "Investissementsspécifiques et concurrence sur un marché". Paris,INSEE. (Mimeo.)

Eymard-Duvernay, F., Thevenot, L, 1983b "Lesinvestissements de forme: leurs usages pour la maind'oeuvre". Paris, INSEE. Mimeo.

Faguer, J-P, 1982 "Jeunes à l'essai; les pratiquesd'embauche en période de crise chômage". Paris, Cahiersdu Centre d'Etudes de l'Emploi, (Mimeo.)

Fox, A., 1971, A sociology of work and industry, London, Casselland Collier Macmillan .

Carfinkel, H., 1967, Studies in ethnomethodology, EnglewoodCliffs, N. J., Prentice-Hall.

Gelb, I., 1963, A study of wrifing, Chicago.

Goode, W., 1974, "Comment: the economics of nonmonetaryvariables", Journal of Politlcal Economy, 82 (2), S27-S33.

51

Goody, J., 1977, The domest!cation of the savage mind, Cambridge,Mass., Cambridge University Press.

Gouldner, A., 1954, Patterns of industrial bureaucracy, New York,The Free Press.

Guibert, B.; Laganier, J.; Volle, M., 1971, "Essai sur lesnomenclatures industrielles", Economie et Statistique, 20,23-36.

Hicks, J., 1974, The crisis in Keynesian economics, Oxford, BasilBlackwell.

Holmsthom, H., 1981, "Contractual models of the labormarket", American Economic Review, 71(2), 303-313.

Jobert, A., Rozenblatt, P., Tallard, M., 1980, "Facteursdeterminants du contenu des conventions collectives etde leur evolution; le cas du secteur de la chimie etde la pharmacie en France et en RFA (rapportexploratoire)", Paris, CREDOC. (Mimeo . )

Jobert, A.; Rozenblatt, P., 1981, "L'application desconventions col-lective dans la chimie et la pharmacie(rapport exploratoire)". Paris, CREDOC. (Mimeo.)

Jobert, A., Rozenblatt, P, Tallard, M., 1983 "L'applicationdes conventions collectives dans les branches de lachimie et de la pharmacie", Paris, CREDOC. (Mimeo.)

Johnson-Laird, P., Wason, P. (eds), 1977, Thinking; reading incognitive science, Cambridge, Mass., Cambridge UniversityPress.

Jones, G., 1971, "Councils among the Central Ido", pp. 63-79 in: Richards, Kuper, (eds). Op. cit.

Kuper, A., 1971, "Council structure and decision-making",pp. 13-28 in: Richards, Kuper, (eds), Op. cit.

52

Latour, B.; Woolgar, S., 1979, Laboratory life, and socialconstruction of scientificfacts, Beverly Hills, Calif., Sage.

Leibniz, G.W., Stark, D., 1875-1890, Die philosophischeschriften (ed. C.J. Gerhardt), 6 vols, Berlin,Weidmannsche Buchhandlung.

Mehaut, P., Rose, J., Stiglitz, J., 1982, "Systemeproductif, gestion de la main d'oeuvre et usage desmesures étatiques de formation et d'insertion", paperpresented at the colloquium on "Les politiques del'emploi", Nancy, Febr., GREE.

Orlean, A., 1982, "Inflation et souverainete monetaire",Critiques de l'Economie Politique, 18, 93-113.

Pinsky, L., 1980, "Analyse ergonomique du travail mental desaisie-chiffrement sur terminal d'ordinateur", Le TravailHumain, 43 (1), 183-184.

Pipino, L., Gigch, J. van, Tom, G., 1981n "Experiments inthe representation and manipulation of labels of fuzzysets", Behavioral Science, 26, 216-228.

Richards A., 1971, "The nature of the problem", pp. 1-12in: Richards; Kuper (eds). Op. cit.

Richards, A., Kuper, A. (eds), 1971, Councils in action,Cambridge, Mass., Cambridge University Press.

Rivard, P., 1983, "Le dilemne utilite/elucidation et sesconséquences sur une enquête statistique privée",Actes du Colloque du Conseil National de laStatistique (18-19 April 1983), Paris, INSEE

Robertson, A., 1971 "The development of town committees inAhafo, Western Ghana", in: Richards; Kuper (eds). Op.cit., pp. 130-170.

53

Rosch, E., 1977, "Classification of real-world objects;origins and representation in cognition", in: Johnson-Laird; Wason (eds). Op. cit., pp. 212-222.

Sabel C., 1982, Work and politics, Cambridge, Mass., CambridgeUniversity Press.

Spence, M., 1973, "Job market signaling", Quarterly Journal ofEconomics, 87, 355-374.

Stark, D., 1980, "Class struggle and the transformation ofthe labor process; a relational approach", Theory andSociety, 9 (1), 89-128.

Stiglitz, J., 1975, "The theory of 'screening', educationand the distribution of income", American Economic Review,65, 283-300.

Taylor, F.W., 1967, The principles of scientific management, NewYork, Norton and Company, First published in 1911 byFrederick W. Taylor.

Thévenot, L., 1979, "Une jeunesse difficile; les fonctionssociales du flou et la rigueur dans les classements",Actes de la Recherche en Sciences Sociales 2627, 2-18.

Thévenot, L., 1983a, "L'enregistrement statistique: unemesure decisive", Actes du Colloque du ConseilNational de la Statistique (18-19 April 1983), Paris,INSEE.

Thévenot, L., 1983b, "L'economie du codage social",Critiques de l'Economie Politique 23-24: 188-222.

Thomson, E., 1967, "Time, work-discipline and industrialcapitalism", Past and Present, 38, 56-97.

Volle, M., 1982, Histoire de la statistique industrielle, Paris,Economica.

54

Weber, M., 1964 From Max Weber, essays in sociology, Oxford,Oxford University Press (1st German ed. 1922).

Weiss, S., 1976, "The sorites fallacy; what difference doesa peanut make?", Synthese 2-3-4, 253-272.

Zerubavel, E., 1982, "The standardization of time; asociohistorical prespective", American Journal of Sociology,88, 1-23.