CC BY-NC-ND
55
Новые тексты
FRENCH ECONOMICS OF CONVENTION AND ECONOMIC SOCIOLOGY
S0ren Jagd
Department of Social Science, Roskilde Universit, Denmark e-mail: jagd@ruc.dk
Paper presented at the International Conference Economic Sociology:: Problems and Prospects University of Crete, Greece September 8-10, 2004
Introduction
The French economics of convention1 tradition has now developed into an important interdisciplinary approach for the study of economic action. Quite recently the question of the relevance of the notion of convention more generally for economic sociology has been taken up by economic sociologists [Biggart, Beamish 2003; Dobbin 2004] as has also the question of the relation between French economics of convention and economic sociology [Swedberg 2003, Thévenot 2004]. It should also be mentioned that David Stark for several years [Stark 2000] has pointed to the relevance for economic sociology of the plurality of orders of worth model by Luc Boltanski and Laurent Thévenot which may be seen as the sociological entrance to the economics of convention tradition.
This paper attempts to take a small step in sorting out potentially common themes for economic sociology and French economics of convention by looking more closely at a few recent texts from the economics of convention tradition discussing, in slightly different ways, differences and similarities between economics of convention, institutional theory and economic sociology. First, I give a very brief overview of some of the basic arguments from economics of convention. Second, I present an argument for a potential common aim of economics of convention and economic sociology as to 'denaturalise' the institutional foundation of markets and of money presented by André Orléan in two recent papers [Orléan 2002; Orléan 2003]. Third, I present a text by Christian Bessy and Olivier Favereau François [Bessy, Favereau 2003] that describes the analysis of institutions from the economics of convention perspective. Fourth, I present a programmatic text, written collectively by five of the key members of the economics of convention tradition (François Eymard-Duvernay, Olivier Favereau, André Orléan, Robert Salais, and Laurent Thévenot), for a conference on convention theory held in December 2003 in Paris [Eymard-Duvernay et al. 2003]. Last, I discuss how these papers may contribute to direct economic sociology towards new research questions.
1 Although a research tradition in economic evolutionary game theory in economics also has got the label economics of convention, for the reason of simplicity I have decided to use this general term, economics of convention, to label the special French interdisciplinary research tradition in focus here. To avoid any misunderstandings I point out that I do not intend to deal with the game theoretic version of economics of convention in this paper.
Economics of Convention: A Short Presentation
The official 'birthday' of convention theory was, according to Orléan [Orléan 1994], a conference on the labour market held in November 1984 with the title "Les outils de gestion du travail". The material from this conference was edited by Salais and Thévenot [Salais, Thévenot 1986] under the title Le travail. Marchés, règles, conventions. As the title suggests the concept of conventions was central from the very start. An important empirical preoccupation was the study of how labour was qualified through the application of rules, norms and conventions [Salais, Baverez, Reynaud 1986]. The insight obtained in these studies lead to a generalisation of the importance of the qualification of all goods before they can be exchanged on the market. Qualifications of persons and products are then a key notion of convention theory as they form the basis for the emergence of rules, or conventions. Following this perspective the analysis of economic transactions cannot dispense from an explicit consideration of the institutional framework qualifying the goods exchanged.
Though the tradition gradually was established during the 1980s, it was the publication of a special issue of the French economic journal in France Revue économique in 1989 on Economics of Convention that made the group known in wider circles. A special feature was the programmatic introduction written in collaboration by Jean-Pierre Dupuy, François Eymard-Duvernay, Olivier Favereau, André Orléan, Robert Salais, and Laurent Thévenot [Dupuy et al. 1989].
The genesis of convention theory was based on a gradual development of different research programs into a broad research program based on a common reference to the concept of convention. It is not possible here to go into detail in the complex history of the development of convention theory. Instead we will summarise the most important element forming convention theory as presented by Olivier Favereau [Favereau 1995; Favereau 2002]. According to Favereau convention theory can be defined as part of a broader institutional tradition. This tradition is defined by two key principles: 1) Micro- and macroeconomic regularities can be causally linked to institutional rules; 2) The system of institutional rules is endogenous to the theoretical model. Institutions are then both essential and endogenous.
Convention theory is seen as based on two central sets of propositions. The first takes its departure in an internal critique of neo-classical economic theory and of new institutional economics focussing on the problems of a too restrictive application of methodological individualism. While both neo-classical theory and new institutional theory coincide in viewing actors as following their interests (which might imply following rules) for convention theory individuals follow rules (which may not exclude that they follow their interests) [Wilkinson 1997]. The argument is that individual rationality alone is not sufficient to explain social interaction between actors when uncertainty is present. A more 'realist' understanding of individual rationality must explicitly explain how the actors co-operate by drawing on some sort of shared rules. Convention theory may be seen as an attempt to 'change the language of economic theory' [Favereau 1995: 512] to explicitly deal with the institutional context of economic action. This line of research was inspired by earlier work in different areas of economic theory such as Frank Knight's and John Maynard Keynes' discussion of uncertainty, as well as the problems of non-cooperative game theory, and the work on the theory of incomplete contracts.
A second proposition is the need of an active interaction between economics and the other social sciences. Convention theory seeks to draw the consequences of both the cognitive turn and the interpretative turn in the social sciences. Convention theory focuses on how coordination is established between co-operating actors by reconstituting the mental representations of the actors under study. Convention theory aim then at constructing a 'microeconomics of understanding' opposed to a 'microeconomics of explanation' as in neo-classical and new institutional economics. The theoretical import from the other social science is more precisely directed at the sociological analysis of co-operation and the political philosophical analysis of justification of actions.
Two elements of convention theory will be further discussed below. First, I discuss the role of the notion of uncertainty for understanding the problem of co-ordination between actors, and second, I discuss the crucial role of conventions in co-ordination of actors.
Uncertainty and the problem of social co-ordination
Convention theory points that neo-classical economics have encountered severe methodological problems regarding the understanding of co-ordination based on the analyses of isolated individual actors. It is proposed that in order to understand the problem of social co-ordination it is necessary to rethink the concept of human action.
The methodological problems faced by neo-classical theory in understanding social co-ordination can be summarised into 'three different obstacles in obtaining co-ordination faced by non-cooperating actors' [Orlean 1994: 17]. These obstacles are illustrating different types of uncertainty (uncertainty caused by the subject matter, socially caused uncertainty and uncertainty related to the future) faced by actors in the process of co-ordination.
First, there may be two or more points of equilibrium. The classic example is the choice of driving either in the left or in the right side of the road. At the initial stage both solutions may be a point of equilibrium. The existence of multiple equilibriums is an example of a market failure in the sense that without any sort of institutional mechanism it is not possible to reach a stable solution to the problem of co-ordination.
Second, actors in non co-operative games face fundamental strategic uncertainty in situations where the outcome of the game depends on the action of other actors. Without any external common reference the process may go on ad infinitum without leading to a solution of the co ordination problem. As argued by Orlean [Orlean 1994] the logic of horizontal strategic interaction is not sufficient to produce co-ordination.
Third, the exchange of goods of which important aspects of its qualities are not known in advance of the exchange creates other forms of uncertainty. It is argued that neo-classical theory can be applied only in situations where there is no uncertainty concerning the quality of the product. An analysis of the price agreed upon by the involved parties is here sufficient. If, on the contrary, there is some degree of uncertainty concerning the quality of the product the analysis must include the specific relation between the enterprise and the customer. The argument draws explicitly on the analysis of incomplete contracts in economic theory describing situations where the goods considered are not fully defined in advance of the exchange. The product is thereby exchanged on the basis of a consideration of the organisations involved in the exchange involving consideration of reputation, experience from earlier interaction etc.
In economics of convention it is assumed that these uncertainties are rather common. This leads to an interpretation of social action as generally unstable and uncertain. An important aspect in understanding the condition of individual action is the explicit analysis of the specific uncertainty of actors regarding the identification of the situation and the interpretative work that is necessary to identify the situation as a common situation. As described by Salais and Storper: 'coordination between economic agents take place within a context of pervasive uncertainty with respect to the actions and expectations of others' [Salais, Storper 1992: 171].
This perspective leads to the need for reconstructing the notion of human action. At the centre of interest is 'the situation in its temporality, the individual's uncertainty about the identification of the situation, and the interpretative effort that is required to determine, together with others, the situation as a shared and common one' [Wagner 1994: 274]. Society is in this perspective not an encompassing social order but a collection of multiple agreements, as well as disagreements. Instead of presupposing the existence of a social order, or the tendency of social relations to achieve
such an order, the perspective of convention theory is to 'turn the production of agreement and coordination itself into the key issue' to be studied.
The transformation from disagreement to agreement presupposes social labour to 'interpret situations, to mutually adapt interpretations, and to determine modes of agreement in common' [Wagner 1994: 275]. Many situations can be handled without, or with limited, agreement on how the situation should be understood. For example walkers who happen to be in the same place only need a limited common agreement on their situation. Buyers and sellers of a certain commodity have a larger need for common agreement, as have citizens involved in political decision-making and parents.
Instead of focusing on the conditions for and the process leading to stability and predictability convention theory involves a focus on disputes between actors on the judgement of specific events/situations and a perspective on social categories and conventions as socially constructed and therefore subject to historical transformations. The emphasis on disputes is based on the assumption that these make particular visible the resources and competences mobilised by actors, and the arguments brought forward by actors in the disputes gives the possibility for understanding the criteria applied by actors in the judgement of situations.
An important contribution of convention theory is to draw attention to the consequence of the limits of the traditional neo-classical model of the atomised economic actor. Understanding economic action as social action presupposes an understanding of the institutionally established frames of meaning drawn upon by actors. The empirical description of the institutional context of action, the conventions drawn upon by actors, becomes then a crucial element in understanding the conditions of agreement between actors.
Conventions and co-ordination of social action
According to convention theory it is necessary to reconsider the concept of institution and the issue of the stability and coherence of social practices. As described above co-ordination between economic agents takes place within a context of pervasive uncertainty with respect to the actions and expectations of other actors. Conventions emerge as responses to such uncertainty [Salais, Storper 1992]. The focus on convention theory has been on 'identifying conventions that have a wide spatio-temporal extension and may thus fit a standard definition of 'institution' [Wagner 1994].
Convention theory's contribution, according to John Wilkinson, may be its original elaboration of the notion of rules as the basis of actor co-ordination. For convention theory rules 'emerge within the process of actor co-ordination. ...they represent a response to problems arising within such coordination and should be understood as mechanisms of clarification which are themselves also open to future challenge' [Wilkinson 1997].
Conventions emerge as responses to and as definitions of uncertainty. Conventions may be seen as 'attempts to order the economic process in a way that allows production and exchange to take place according to expectations which define efficiency' [Storper, Salais 1997: 16]. Conventions may become incorporated in routines and 'we then tend to forget their initially hypothetical character' [Storper, Salais 1997: 16]. Conventions refer to the simultaneous presence of three dimensions, each having a different spatio-temporal extent and overlapping in complex ways: rules of spontaneous individual action, agreements between persons, and institutions in situations of collective action [Storper, Salais 1997: 17].
The formal notion of convention drawn upon by convention theory was developed by David Lewis [Lewis 1969: 58]:
'A regularity R in the behavior of members of a population Pwhen they are agents in a recurrent situation Sis a convention if and only if it is true that, and it is common knowledge in P that, in any instance of S among members of P,
(1) everyone conforms to R;
(2) everyone expects everyone else to conform to R;
(3) everyone prefers to conform to R on the condition that the others do, since S is a coordination problem and uniform conformity to R is a coordination equilibrium in S.'
There are nevertheless important differences between the way Lewis and other analytical philosophers use the notion of convention and the analysis of the concept in convention theory. While Lewis argues that ambiguities about the action of other actors disappear through the emergence of 'common knowledge' of conventions among actors. It is then assumed that all ambiguities concerning other actors then disappear. In convention theory it is not assumed that conventions automatically emerge and the existence of a 'total transparency' and absolute reflexivity is doubted. Situations may be differently identified in many different ways. Conventions emerge as 'something like rebuttable hypotheses put forward by actors, which then become second nature through practice'. Conventions are 'subject to many possible sources of change, ranging from their failure in the face of external tests to a reinterpretation of circumstances by actors themselves' [Storper, Salais 1997: 18]. The importance of conventions is then not to make all ambiguity and uncertainty concerning other actors disappear totally but only the 'presence of a 'collectively recognised' reference' [Orléan 1994] which stops, temporarily, the speculation concerning the intentions of other actors.
A short summary of the some of the most important works are presented below (fig. 1).
One of the important contributions from convention theory has the analyses of the types of conventions applied in different areas of society. One of the seminal contributions was Luc Boltanski and Laurent Thévenot's work attempting to classify the most important general types of conventions, as six important 'worlds of justification' [Boltanski, Thévenot 1991].
Another important contribution, more explicitly focused on business enterprises was the concept of 'conventions of quality' developed by François Eymard-Duvernay [Eymard-Duvernay 1989; Eymard-Duvernay 1994].
A third important contribution started out from studies of the labour market by Robert Salais pointing first to 'conventions of labour' and further elaborated together with Michael Storper into four 'worlds of production' [Salais 1989; Salais 1994; Salais, Storper 1992; Storper 1996; Storper, Salais 1997].
I now turn to a presentation of three recent contributions discussing in different ways the way economic of conventions are related to discussion in economic sociology and institutional theory more generally.
Figure 1. The French Economics of Convention School.
The French Economics of Convention School
Programmatic text: Special issue of Revue économique on "Economics of conventions". 1989. Vol. 40. No. 2. See especially: Dupuy, J.-P., F. Eymard-Duvernay, O. Favereau, A. Orléan, R. Salais, L. Thévenot. Introduction // Revue économique. 1989. Vol. 40. No. 2. P. 141- 145.
Mission and goal: To develop a theory of the role of conventions in the coordination of economic action. Empirical analysis of the plurality of conventions involved in the coordination of economic action, their variation and their dynamics.
Field of research: Studies of different conventions in economic action, in goods-, labour-, and financial markets.
Important researchers: Luc Boltanski, François Eymard-Duvernay, Olivier Favereau, André Orléan, Robert Salais, Michael Storper, Laurent Thévenot.
Important works: Salais R., Thévenot L. Le travail: marchés, règles, conventions. Paris: Economica, 1986; Special issue of Revue économique on "Economics of conventions". 1989. Vol. 40. No. 2; Boltanski L., Thévenot L. De la justification. Les économies de la grandeur Paris: Gallimard, 1991; Orléan A. (ed.). Analyse économique des conventions. Paris: PUF, 1994; Storper M., Salais R. Worlds of Production. The Action Framework of the Economy. Campridge, Mass.: Harvard University Press, 1997; Boltanski L., Chiapello E. Le nouvel esprit du capitalisme. Paris: Gallimard, 1999; Favereau O., Lazega E. (eds.). Conventions and Structures in Economic Organization. Cheltenham: Edward Elgar, 2002.
Important addresses: FORUM (Fondements des Organisations et des Régulations de l'Univers Marchand), Université Paris-X, Nanterre (François Eymard-Duvernay, Olivier Favereau); Groupe GSM, École des Hautes Études en Sciences Sociales (EHESS), Paris (Luc Boltanski, Laurent Thévenot); CEPREMAP, ENS Bld. Jourdan, Paris (André Orléan); IDHE (Institutions et Dynamiques Historiques de l'Économie), ENS Cachan (Robert Salais).
André Orléan: Denaturalizing the Social Foundations of markets and money
In this section I present an argument for at common ground between economics of convention and economic sociology made by André Orléan [Orléan 2002; Orléan 2003]. According to André Orléan one of the most characteristic traits of the traditional approach in economics is to 'naturalise' the social environment by postulating that the economic world is constituted by natural elements, most importantly goods and money, that may be unambiguously interpreted by actors. These elements do not need to be constructed since they are 'already there'. This essential aspect of the traditional economic approach can be illustrated by the help of two hypotheses determining the way traditional economics is structured as a discipline. Orléan labels these two hypotheses as the nomenclature-hypothesis and the probability-hypothesis.
The Nomenclature Hypothesis
The nomenclature-hypothesis consists in the postulate of the existence of a list of n goods with a homogeneous quality known by all actors. This is the departure of all microeconomic introductions though it rarely is explicitly discussed, it is 'natural'. From this perspective the economy present itself spontaneously for the eyes of the most naïve and little informed observer. For the consumers the problem is to determine the quantities (xb x2, ..., xn) of the n goods available so their utility is maximised under the constraint that their expenses will not be greater than their income. For the producers the problem is to determine the quantities (xb x2, ..., xn) of the n goods produced in such a way that they maximize their profit. Their constraint is set by the available production technology.
In these two cases the logic is formally identical: the individuals are not occupied with the behaviour of other actors but only with the price of goods consumed and produced.
According to Orléan the importance of this hypothesis is not that it is a trivial and neutral description of the market economy. Supposing the existence of use values socially recognized by all actors it describes a universe already structured by common points of references that permits actors to coordinate effectively. By presenting the n goods as a fact that is observed as an exogenous fact by all actors, presenting their quality as a common knowledge among all actors, a common language is de facto constructed that facilitates market coordination and permits an exchange agreement to unfold. In other words, from the outset it is supposed that there exist a pre-given list of goods.
It is an important point that the Arrow-Debreu model functions perfectly in situations where use value are fixed and known by all actors, but only applying this model, as a tool to comprehend economic action, leads to important limitations. The most fundamental limitation is that theoretically we cannot understand an important part of market coordination because the focus is only on price-quantity relations, at the expense of other key aspects of market coordination, such as the endogenous evolution of notions of product qualities. To start with the situation where all objects are perfectly certified or where all actors have a perfectly defined utility function is to suppose that the 'market question' is already partially solved. In such a framework the Walrasian agreement appear largely as a consequence of the fact that in advance the actors ex ante have succeeded to agree on the definition of qualities of products, without specifying explicitly the process that has permitted to reach such an agreement. In this way the analysis of the formation of a market order in pushed into the shadow because it is implicitly supposed a framework already existing.
The general equilibrium, as described in traditional economic models, in not a spontaneous order. It emerges by the problematic meeting of actors. These reflections lead André Orléan to propose a new formulation of the 'market question' as follows:
How do the atomistic individuals succeed in constructing common points of references
permitting them to exchange and regulate their exchanges?
This formulation of the 'market question' point to the need of a more explicit consideration of the diverse forms of social work in and around markets: activities of codification, of certification, standardisation, and regulation.
The Probability Hypothesis
André Orléan discusses a second hypothesis playing an equally fundamental role in traditional economics, the probability-hypothesis that spells out the way traditional economics conceptualize the future. Following this hypothesis there is supposed the existence of a list of m states of the world, e, e2, ... em, describing in an exhaustive way all that may happen tomorrow. This list is supposed to be known to all economic actors. Because all the states of the world is given a probability by the actors, the list is associated with a distribution of probabilities, objective or subjective, hence the name of the hypothesis as the 'probability-hypothesis'. In the same way as we saw concerning the list of goods in relation to the nomenclature-hypothesis this list of states of the world can be formally analysed as playing the role of 'natural' mediation between the individuals in the sense where it permits the economic actors, in their relation to the future, not to be preoccupied with the opinion of other actors because all know how the future will be by having probabilities linked to all possible states of the world. The probability-hypothesis leads to the exclusion of the analysis of the interaction through which there is constructed a legitimate representation of the future, i.e. a representation that is accepted by all agents as a common point of reference.
Orléan argues that the importance of the probability-hypothesis can be illustrated by the role of financial conventions. The starting point of economics of convention in relation to financial markets is the questioning of the probability-hypothesis that supposes the future economy can be described exhaustively in the form of a list of possible event defined a priori. Following the perspective of economics of convention the existence of a representation of the future shared by all in the financial market is not given a priori, but it may, at the best, be a result of market exchanges themselves. As in the case of goods, the definition of such a representation do not pre-exist the exchange, but it is produced by the exchange.
Concerning the position of economic sociology in relation to these two hypotheses the situation is seen to be complex. André Orléan [Orléan 2003] joins the critic of Pascal Chantelat [Chantelat 2002] who argues that economic sociology identifies too closely social relations and personal relations. This conceptualisation expresses a too 'intimately' a conception of social relations which tends to reduce the social relations only to durable, continued and intense relations [Chantelat 2002: 523]. From this follows a tendency to consider impersonal and discontinuous exchanges as, at the limit, being outside the social. This is, according to Orléan, particularly visible in the analysis of trust, which is seen as personal trust, resulting from relational proximity and affect. In this conceptualisation it is not conceived that the anonymity and the discontinuity of relation constitutes a form sui generis of social relations. This social form can exist and can be reproduced even in the absence of a network of personal relations. Economic sociology has then not explicitly pointed to the embeddedness of market exchange to have started long before the constitution of networks and the re-personalisation of economic relations. Economic sociology seems then not to account explicitly for this institutional foundation of market relations.
On the basis of this diagnosis, Pascal Chantelat develops his analysis of 'pure market relations' as a specific form of social relations drawing on Simmel's concepts. In the same line the research project of economics of convention, according to André Orléan, can be seen as oriented towards 'denaturalising' the economic approach by showing that these exogenous mediations postulated in the hypothesis of nomenclature and probability in fact are social constructions, 'conventions'. The idea to decuple a social moment where the goods and the representations of the future are constructed and a strictly economic moment where the goods are exchanged and the risks are attached to future events do not work. These two processes are intertwined.
On the other hand economic sociology has been critical of the under-socialised conception of homo economicus in economic theory as shown particularly by Mark Granovetter's work [Granovetter 1985]. Orléan point that in Granovetter's analysis the under-socialised actor is opposed by the over-socialised actor who automatically and unconditionally follows customs, habits or norms. Granovetter notes that these two approaches converge toward the same conception of action. Paradoxically, as shown by Granovetter, the individual in the over-socialised model is not less atomised than his under-socialised cousin. The over-socialised actor acts in isolation without considering others because he has the world in his head. This perspective can be used to reinterpret the Walrasian model. Instead of considering the Walrasian world as a world without institutions, it is more correct to see it as a society strongly structured around powerful institutions, as demonstrated by the socially validated goods and a legitimate representation of the future. Furthermore, the individuals have internalised market norms so that they are led to make consumption of objects the only socially pertinent goal. In other words it is a world without envy, without a thirst for social recognition, without the attachment to social groups. Actors are uniquely preoccupied with calculating their consumption and their exposition to risk. This homo economicus is not at all a dissocialised or under-socialised human being. By adhering to such a model the Walrasian model is to be seen as a model of the over-socialised actor. To stress the over-socialised character of homo economicus permit to break with the idea of spontaneity and naturalness that often is associated to the idea of the market and interest.
Following this perspective the goods and the representations of the future are to be seen as specific social institutions. It means that by the nature of the objects in focus of the study, the conventions, the interactions, the embeddedness, the legitimacy of common points of references, the economics of conventions shares much interest with economic sociology. At the limit, this convergence is so strong, for example when the stress is put on the social context and on the idea of embeddedness, that the economics of convention ends appearing as a subfield of social economics. Orléan argues then that that a common aim of economics of convention and economic sociology could be to 'denaturalise' the traditional approach in economics.
Christian Bessy and Olivier Favereau: The Study of Economic Institutions
In this section I present a thorough outline of the analysis of economic institutions from the perspectives of economics of convention by Christian Bessy and Olivier Favereau [Bessy, Favereau 2003]. The paper opens by pointing that the blindness of economics with regard to institutions, both by marginalist and Marxist traditions, has contributed to harden the frontiers between the different disciplines in the social sciences. On the other side has the recent more explicit consideration of the role of institutions in microeconomics, both in the neo-classical version and in the diverse institutional traditions, led to closer relations with the other social sciences, in particular with law, sociology, cognitive psychology, and history.
The initial aim of economics of convention, according to Bessy and Favereau, was not to propose an economic theory of institutions, but to analyse the relation between individual action and different collective frames of action. The founding hypothesis of economics of convention was that although these collective frames of action are external to persons, these collective frames of action are created, actualised and questioned through personal action. This hypothesis is based on a more complex methodological individualism that the dominant version. The reason that the notion of convention was preferred was that the notion institution was too loaded with a holist perspective, too naturally seen as a structured collective entity.
Instead of focusing directly on institutions economics of convention have instead taken a closer look at the notion of rules. Assuming that all rules are more or less incomplete and that actors should agree on a scheme of interpretation of the rule to coordinate their action, economics of convention mobilise the notion of convention to describe such a scheme of interpretation. Following this perspective lead to two basic hypotheses for an institutional theory: first that institutions are endogenous, and, second, the importance of the reflexivity of actors confronted with problems of coordination. This lead to the point that an important role should be attached to language.
Economics of convention argues that there are three basic institutions, language, money and law: 'there is no individual rationality without language, no market economy without money, and no pluralist society without law' [Bessy, Favereau 2003: 136]. These three institutions are the basis of what we call the 'social'.
The existence of language is seen as a logically necessary condition for the emergence of common worlds. Besides calculative capacities homo economicus in this perspective also need cognitive capacities and capacities of interpretation. The emergence of money is seen a basic condition for the market society. The study of the institutional analysis of money was one of the starting points for one of the key members of the economics of convention tradition André Orléan (Aglietta, Orléan 1982). Summarising their perspective, Bessy and Favereau argue that 'if the money can be logically deduced from the plurality of goods, and the language can be logically deduced from the plurality of humans, the law can in the same way be logically deduced from the plurality of justifications' [Bessy, Favereau 2003: 140].
Bessy and Favereau present an overview of the empirical studies of institutions based on the economics of convention perspective. In the first generation of work they find two perspectives: The first perspective focused on tests of justification. Based on studies of situations of evaluation and the judgements of personal qualities, the orders of justification were emerging empirically. This empirical work was done in parallel with the elaboration of the models of cites by Luc Boltanski and Laurent Thévenot [Boltanski, Thévenot 1991]. The second perspective focused on institutional devices, such as the rules of law, in particular in the regulation of the labour market are articulated in enterprises. The work of Robert Salais is a good illustration of this type of work [Salais, Baverez, Reynaud 1986].
A second generation of research developed during the 1990s. The third perspective was searching for the mediations between the "general" and the "local" forms of coordination such as public employment policies. A fourth perspective intended to reintegrate the political in the dynamics of institutions. Examples of this type of research could be the study of unemployment in Germany [Zimmermann 2001] and also the macro-historical analysis of the changes in recent capitalism by Boltanski and Chiapello [Boltanski, Chiapello 1999].
What may be particularly interesting in this paper is the attempt to link explicitly the study of the three institutions of language, money and law. The linking of language and money have been taken up in economic sociology, especially in Viviana Zelizer's work [Zelizer 1994]. The role of law in relation to economic action has been much less elaborated on.
The Economics of Convention Programme: Values, Coordination and Rationality
The last text to be considered in this paper is the programmatic text for the conference on the economics of convention held in December 2003 in Paris. The text with the title 'Values, Coordination and Rationality. The Economy of Convention or the Time of Reunification in the Economic, Social and Political Sciences' was written collectively by François Eymard-Duvernay, Olivier Favereau, André Orléan, Robert Salais, and Laurent Thévenot [Eymard-Duvernay et al. 2003]. The basic perspective of economics of convention is first presented:
'If we agree that the coordination of human action is problematic and not the result of laws of nature or constraints, we can understand that human rationality is above all interpretative and not only immediately calculative. The agent first has to apply conventional frameworks to comprehend others' situations and actions before he/she can coordinate him/herself. This understanding is not only cognitive but also evaluative, with the form of evaluation determining the importance of what the agent grasps and takes into account' [Eymard-Duvernay et al. 2003: 1].
The relation between economics and sociology is described as an aspiration towards generalization currently visible in both economic and sociology, in the form of an attempted extension into the preferred domain of the other discipline. Economics is seen to spread to non-commercial relations, e.g. the family, power, politics, and organizations, by applying mechanisms as 'contracts' and 'games'. Likewise sociology is seen to expand into the analysis of economic action:
'Economic sociology offers a counter-attack to these extensions and intends to reduce economics to a field equivalent to the other social actions in which it specializes. The advantage of this opposing extension consists in inscribing so-called economic relations in a far wider space by highlighting their entanglement with social actions. With the common aim of denaturalizing economic relations, a rich body of research on "the social construction of markets" has emerged' [Eymard-Duvernay et al. 2003: 4].
Sociology is found to encompass a far wider range of social actions than do economic theory. Therefore, the reduction effected by sociology when it expands into the economic domain is not as radical as the symmetrical reduction. Nevertheless, the extension of sociology also raises new questions:
'The models of social action, even when they more or less metaphorically employ the language of markets and interests, imply modes of coordination that are profoundly different because based on social groups, social representations, social practices, a sense of the social, and social intercomprehension. They fail to characterize the specificity of frames of action and coordination involving market objects. Despite its fecundity, the notion of embeddedness of economic transactions in social relations attests to this reduction to models of social links' [Eymard-Duvernay et al. 200: 4-5].
Thus, the reservation against economic sociology is that instead of taking the perspective of a plurality of possible frames of action, economic sociology seems to substitute the market coordination framework with a coordination based on social (network) relations instead of being open to a plurality of forms of coordination as suggested by economics of convention.
The perspective of economics of convention is to seek an integration of the perspectives of economics and sociology while recognizing that each disciplinary tradition illuminates different aspects and different modes of coordination. That is why the economics of convention tradition has constructed a framework of analysis devoted to an issue common to both traditions: the problematical coordination of human action.
In the paper the framework developed by the economics of convention tradition is presented with a focus on three issues central to both economics and sociology: the characterization of the agent and his/her reasons for acting; the modalities of coordination of actions, and the role of values.
The Characterization of the Agent
The presentation start by posing the question: 'With what are we equipped in terms of agents or devices [dispositifs], to account for coordinated actions? The answer obviously depends on our interpretation of the word coordination' [Eymard-Duvernay et al. 2003: 5]. The notion of coordination developed by economics of convention highlights the role of collective forms of evaluation. The most public forms subject coordination to the demand for justification. This notion of coordination is not opposed to the idea of conflict. Coordination is put to test and realized against a background of failure and of conflict and criticism.
In studying the problematic coordination of human actors economics and sociology have, it is argued, concentrated on different specifications of this coordination. Classical authors in both economics and sociology are found to have remained close to the reference models from the natural sciences, highlighting equilibriums, orders and structures of social reproduction. For interactionist sociologists, uncertainty remains part of the idea of an "order of interaction", even if it is "negotiated" locally in the situation. Ethnomethodologists are particularly doubtful concerning the notion of order. For economists, the problem is concentrated on notions of uncertainty and information.
Economics of convention proposes to take into account the uncertainty weighing on the coordination of behaviours by differentiating forms of uncertainty and thus information, and then relating them to different forms of evaluation. Evaluation is then seen to be at the centre of coordination. In all coordination, whether in the market, in an enterprise, or aimed at political agreement, there is no regularity at the start of the action. In this sense, uncertainty exists for everyone. Overcoming that uncertainty requires the conventional construction of products, services and of expectations that are the mediums of the commercial interaction and productive activity of firms. The notion of convention, it is argued, enables us to characterize this moment of common construction. Conventions channel uncertainty on the basis of a common frame of evaluation that qualifies objects for coordination. Market conventions of qualification are seen as one among a plurality of such conventions of qualification.
The Modalities of Coordination of Action
Two types of pluralism are argued for. The first type analysed is a "horizontal" pluralism of conventions of qualification having a high degree of generality. The second type of pluralism concerns the distinction between several levels of convention, from public coordination to different forms of local coordination.
Concerning the first type of pluralism of general modes of coordination, it is argued that
'By recognizing that the most general modes of coordination are based on such forms, we take seriously the demands for justice and democracy that weigh on organizations, as well as the sense of fairness, of the public good or of the common good expected from the actors engaged in these coordinations. The importance of these expectations, situated at the heart of political philosophy, has been diminished considerably in prevailing economic and sociological approaches. Either they reduce all evaluations to individual preferences incorporated into prices, or they limit them to arbitrary social values in their diversity. The fact of taking the legitimacy of these forms of evaluation and their pluralism seriously modifies our understanding of both actors and organizations' [Eymard-Duvernay et al. 2003: 8].
It is argued that in economics evaluation is reduced to the utility function that is assumed to be stable or only subjected to exogenous variations. Several attempts have been made to endogenize preferences, either by likening them to routines selected by the environment, or by introducing an ordering of preferences through metapreferences. In economics of convention it is attempted to go beyond that by relating evaluation to a state of individuals that depends on their engagement in their coordination environment. Compared to sociologies that assume the existence of stable determinants of social behaviours, taking into account a plurality of states of evaluation leaves room for different engagements and introduces dynamics into people's dispositions.
Markets are seen as places where the quality of goods is tested and evaluated through activities of codification, measurement, certification, regulation etc. The fact of reducing what happens in markets to the laws of supply and demand, as in mainstream economics, mean that all these social process in an around markets most often are neglected. In all types of markets: goods markets, labour markets, or financial markets, a plurality of principles of evaluation exist. This point has to be integrated into theoretical analysis.
It is argued that classical economics and sociology tend to consider the founding institutions (the market, the community) as exogenous, universal and stable. In contrast a much more dynamic perspective is presented:
'The introduction of radical uncertainties (lack of mode of coordination containing uncertainty within the limits of an order of qualification) and of critical dynamics (challenging an agreement) into analysis leads to the conception of conventions that are deformed by action and are plural and evolving. People are placed in a conventional environment (formed mainly by texts, legal corpuses, accounting units, evaluation tools) that they rearrange to remedy the lack of coordination and cooperation' [Eymard-Duvernay et al. 2003: 14].
The economics of convention makes it possible to recognize the theoretical specificity of each type of institutional market device. To illustrate the point, the special features of two types of markets -labour markets and financial markets - are briefly discussed. Concerning the analysis of the labour market, that has been central for much research in the convention tradition, open to the plurality of forms of work. This perspective also renews the role of firms. From the convention perspective, the firm organizes the articulation between goods, labour, and capital markets. The firm is at the intersection of several forms of coordination, managing the tensions that result from a situation by compromises between them. The diversity of corporate models and 'worlds of production' [Storper, Salais 1997] challenges the view of the firm as a unified and simple hierarchical mode of coordination.
The analysis of financial markets from the perspective of economics of conventions reveals the gab between the orthodox analyses of finance in terms of which securities are considered to be naturally exchangeable, like merchandise. The perspective of economics of convention is devoted to criticizing this natural state of goods, as also argued above by André Orléan. Financial markets are not reducible to a competitive mode of coordination based on market qualification of goods, like other consumer markets. Finance implies coordination by opinions, where a set of heterogeneous opinions is transformed into a reference value agreed to by all. Agents' expectations are turned towards the expectations of the other market actors. Mimetic behaviours are thus encouraged. In coordination based on a convention by opinion, in order to predict what the others are going to do, it is enough to refer to the convention. Through the game of self-validation of beliefs, there follows a relative stability of the convention that, for the agent, becomes second nature.
The introduction of radical uncertainties and of critical dynamics into the analysis leads to the conception of conventions that are plural and evolving. Actors 'are placed in a conventional environment (formed mainly by texts, legal corpuses, accounting units, evaluation tools) that they rearrange to remedy the lack of coordination and cooperation. And, furthermore, that to introduce this conventional dynamic into the analysis, the actors have to be endowed with a reflexive capacity regarding their own state, as well as a capacity to remodel forms of community life - in other words, a political capacity' [Eymard-Duvernay et al. 2003].
The Role of Values
The second type of pluralism, "vertical" pluralism, is oriented towards more situated coordination and more personal conveniences. In both sociology and economics, various researchers have focused on non-reflexive relations with the world, such as habits, routines and practices. Economic theory has, for example, proposed two local models, one with weak rationality (routines) and the other with strong rationality (contracts), both of which are considered unsatisfactory. On this basis a distinction is make between constituent conventions (Convention 1) which support the most legitimate modes of coordination and, second-level conventions (Convention 2), which encompass more limited rules intended to coordinate normalized local action plans. The analysis should then focus on the dynamic between these two levels of coordination.
Discussion: Economics of convention and economic sociology - is there a common set of research questions?
The proposal presented by André Orléan for a common project for economic sociology and for economics of convention to aim for 'denaturalising' the economic approach by showing that these exogenous mediations postulated in the hypothesis of nomenclature and probability in fact are social constructions could probably be joined by most economic sociology researchers. As argues above, by the nature of the objects in focus of the study, the conventions, the interactions, the embeddedness, the legitimacy of common points of references, the economics of conventions shares much interest with economic sociology. As illustrated by André Orléan with his analysis of financial conventions the convergence between economics of convention with the preoccupations of economic sociology may be especially clear in regard to the analysis of financial markets.
Following André Orléan's reformulation of the 'market question' it is important to focus on a more explicit consideration of the diverse forms of social work in and around markets: activities of codification, of certification, standardisation, and regulation. This type of work should be joined by economic sociology. In fact some early work was done in this area in the 1980s but there is need of a much more wholehearted devotion to analyse how markets are constituted and changed by the actors involved.
With regard to the proposals for a broad research programme as presented by Christian Bessy and Olivier Favereau the potential implications for economic sociology is more complex. What may be challenging for economic sociology is the way to link quite diverse institutions such as language, money, and law.
The implications of the last text, the plea for an interdisciplinary research program is likewise more complex. The most important point is probably the need to explicitly consider the plurality of forms of coordination. The problem of economic sociology viewed from researchers in the economics of convention tradition is to stick to limit the analysis to only a few forms of coordination, such as markets and interpersonal relations. An important point in the works from the economics of convention tradition is that most often a great variety of forms of coordination is part of economic action. Beside market and interpersonal relations, Boltanski and Thevenot [Boltanski and Thevenot 1991] have demonstrated that industrial, civic, public and inspirational forms of coordination are important. Boltanski and Chiapello [Boltanski and Chiapello 1999] have more recently demonstrated a greater role of coordination based on references to the 'network world'. The basic point is that the variety of forms of justification involved in economic action should be much more explicitly analysed.
References
Aglietta M., Orléan A. La violence de la monnaie. Paris: Presse Universitaires de France, 1982.
Bessy C., Favereau O. Institutions et économie des conventions // Cahiers d'Economie Politique. 2003. P.119-164.
Biggart N.W., Beamish T.D. The Economic Sociology of Conventions: Habit, Custom, Practice, and Routine in Market Order // Annual Review of Sociology. 2003. No. 29. P. 443-464.
Boltanski L., Chiapello E. Le nouvel esprit du capitalisme. Paris: Gallimard, 1999.
Boltanski L., Thévenot L. De la justification. Les économies de la grandeur Paris: Galimard, 1991.
Chantelat P. La nouvelle Sociologie Économique et le bien marchand: des relations personelles à l'impersonalité des relations // Revue française de sociologie. 2002. No. 43. P. 521-556.
Dobbin F. The Sociological View of the Economy // Dobbin F. (ed.). The New Economic Sociology. Princeton: Princeton University Press, 2004. P. 1-46.
Dupuy J.-P., Eymard-Duvernay F. et al. Introduction // Special issue on L'économie des conventions, Revue Economique. 1989. No. 40. P. 141-145.
Eymard-Duvernay F. Conventions de qualité et formes de coordination // Revue Economique. 1989. Vol. 40. P. 329-360.
Eymard-Duvernay F. Coordination des échanges par l'entreprise et qualité des biens // Orléan A. (ed.). Analyse économique des conventions,. Paris: Presse Universitaires de France, 1994. P.307-334.
Eymard-Duvernay F., Favereau O., Orléan A., Salais R., Thévenot L. Values, Coordination and Rationality: The Economy of Conventions or the Time of Reunification in the Economic, Social and Political Sciences // Paper presented at the conference Conventions et institutions: approfondements théoriques et contributions au débat politique. Paris. December 11-13, 2003.
Favereau O. Conventions et régulation // Boyer R., Saillard Y. (eds.). Théorie de la régulation. L'état des savoirs. Paris: Éditions La Découverte, 1995. P. 511-520.
Favereau O. Conventions and Régulation // Boyer R., Saillard Y. (eds.). Régulation Theory: The State of the Art. L.: Routledge, 2002. P. 312-318.
Granovetter M. Economic Action and Social Structure: The Problem of Embeddedness // American Journal of Sociology. 1985. Vol. 91. P. 481-510.
Lewis D.K. Convention:: A Philosophical Study. Cambridge, Mass.: Harvard University Press, 1969.
Orléan A. Vers un modèle général de la coordination économique par les conventions //Analyse économique des conventions. Paris: Presse Universitaires de France, 1994. P. 9-40.
Orléan A. Pour une nouvelle approche des interactions financières: l'économie des conventions face à la sociologie économique // Huault I. (ed.). Sociologie économique et analyse des organisations. Autour des travaux de Mark Granovetter Paris: Editions EMS, 2002. P. 207238.
Orléan A. Réflexion sur les fondements institutionnels de l'objectivité marchande // Cahiers d'Économie Politique. 2003. P. 181-196.
Salais R. L'analyse économique des conventions du travail // Revue Économique. 1989. Vol. 40. P. 199-240.
Salais R. Incertitude et interactions de travail: des produits aux conventions // Orléan A. (ed.). Analyse économique des conventions. Paris: Presse Univesitaires de France, 1994. P. 371403.
Salais R., Baverez N., Reynaud B. L'invention du chômage. Paris: Presse Universitaires de France, 1986.
Salais R., Storper M. The Four "Worlds" of Contemporary Industry // Cambridge Journal of Economics. 1992. Vol. 16. P. 169-194.
Salais R., Thevenot L. (eds.). Le travail Marchés, règles, conventions. Paris: INSEE-Economica, 1986.
Stark D. For a Sociology of Worth // Keynote address for the Meetings of the European Association of Evolutionary Political Economy. Berlin. 2000. November 2-4. Columbia University. 2000. http://www.sociology.columbia.edu/people/faculty/stark/papers/sociology worth.pdf
Storper M. Innovation as Collective Action: Conventions, Products and Technologies // Industrial and Corporate Change. 1996. Vol. 5. P. 761-790.
Storper M., Salais R. Worlds of Production. The Action Framework of the Economy. Cambridge, Mass.: Harvard University Press, 1997.
Swedberg R. Economic Sociology Meets the Economics of Conventions // Colloque Conventions et Institutions:: Approfondissements Théoriques et Contributions au Débat Politique. Paris. December 11-13, 2003.
Thévenot L. The French Convention School and the Coordination of Economic Action. Interview // Economic Sociology. European Electronic Newsletter 2004. Vol. 5. P. 10-18.
Wagner P. Dispute, Uncertainty and Institution in Recent French Debates // Journal ofPolitical Philosophy. 1994. Vol. 2. P. 270-289.
Wilkinson J. A New Paradigm for Economic Analysis? // Economy and Society. 1997. Vol. 26. P. 305-339.
Zelizer V. The Social Meaning of Money. N.Y.: Basic Books, 1994.
Zimmermann B. La constitution du chômage en Allemagne:: entre professions et territoires. Paris: Éditions de la MSH, 2001.
