1The position held by knowledge in economic sociology is relatively ambiguous. For instance, a reading of the contents of several French-language manuals published in recent years (François, 2008; Steiner, 2011; Le Velly, 2012; Bernard de Raymond and Chauvin, 2014) suggests that knowledge is far from considered with the same centrality as other entities or concepts such as networks, culture, devices, institutions, entrepreneurs, competition, or even quality. Each of these concepts is even covered by relatively coherent research subsets, while little economic sociology work has explicitly claimed to probe the roles of knowledge. Philippe Steiner (1998) opened up the way by studying the diversity of rationalities brought by the different economic theorizations, but apart from that author few researchers have subsequently explicitly argued the case for a sociology of economic knowledge. It is thus tempting to conclude, as Patrik Aspers did, that “economic sociology has largely neglected the issue of knowledge” (Aspers, 2009, p. 111). The thesis we shall be defending in the first part of this article is that this is far from the case if one considers for a moment how the workings of economic action have been envisaged by it. Much economic sociology work thus refers to a basic socio-cognitive challenge: interpreting situations and anticipating the consequences of actions. The different aspects referred to above, ranging from networks to quality, can thus be reread in terms of the challenges of bringing into play knowledge within the actions that traverse them. The collection that this article introduces fits in with this perspective. Based on field studies concerning subjects as diverse as the success of lean production in the car industry, the on-line appraisal of hotel services, the work of Pôle Emploi job centre staff, public timber auctions or the commercialization of pharmaceutical research, it shows the central role held by knowledge in the functioning of the economy and markets. In addition, given the ethnographical approaches taken, it shows how important it is to grasp this issue on a microsociological scale, close to concrete practices. Several results can then be highlighted, focusing on uncertainties and knowledge surrounding the qualities of goods, knowledge regimes underpinning their valuation, and lastly the performativity of knowledge. The second part of this article will then go on to present these three lines of analysis, such as they are pursued in the articles of the collection.
2Despite the diversity of its theoretical inspirations (Convert and Heilbron, 2007), all the research conducted since the mid-1980s in economic sociology shares several common points, such as the use of surveys rather than modelling, and choosing to work on the diversity of concrete markets rather than on their theoretical functioning. However, another common point needs to be highlighted here. Much of this research shares a theory of action that is sensitive to the cognitive operations whereby economic actors evaluate, calculate, act and coordinate themselves. They also share an attention to the social conditions that allow those operations. The approach defended is therefore very different from that of economists or psychologists who are attentive to the mental processes of economic actors and to deviations from a hypothesis of perfect rationality (Thaler, 1991). It is a sociological approach that reports the social supports that enable and channel cognitive operations. In sum, as Jens Beckert did (1996, 2009), it is possible to support the thesis that economic sociology is founded like a series of original responses, alternative to those of economists, to issues concerning the cognitive stakes of such operations.
3French Economics of Convention is one of the theoretical approaches that has very explicitly highlighted the cognitive challenge presiding over the implementation of economic activities. The coordination problems that economics has shed light on regarding this point provided conventionalists with a starting point for their demonstration. Situations of the prisoner’s dilemma type (if each person knows that it is in their interests to cheat, even mutually advantageous exchange cannot take place), such as those involving challenges of pure coordination (without an agreement on a common principle for valuing a good, it is impossible to trade it) show ad absurdum that it is not possible to envisage “pure” market coordination (Orléan, 1994a, 1994b). More generally, the conventionalist manifesto (Dupuy et al., 1989) argues that social actors need to base themselves on shared cognitive references, “conventions”, for engaging in the economic activities of production, trade, or consumption. Through the socio-technical devices that relay them, such as technical standards or professional qualifications (Thévenot, 1984), it is possible to agree on the orders of worth that preside over the evaluation and to increase the predictability of behaviours. In this founding reasoning, the theory of action fits in with what has often been called the “cognitive turn” of the social sciences. There is an affirmation of a central concern for the interpretative, cognitive and evaluative operations that are the basis of action, and for the social supports that enable them.
4This concern is also found in neo-institutional sociology which, most particularly through the work of Neil Fligstein, has also greatly contributed to the development of economic sociology. In the introduction to the major collective work they coordinated, Paul DiMaggio and Walter Powell (1991) highlighted the way in which they perceived the link between action and institution. Unlike the older institutional approaches of American sociology, and following the inflections engaged in notably by Garfinkel, Goffman or Giddens, they claim not to start out from interiorized values and norms, but from formal rules, cultural conceptions and routines that present themselves to actors like institutions. These institutions are described as being sources of constraints, but also as cognitive resources enabling actors to interpret the situations in which they find themselves and develop their own strategies. In this perspective, even the formal rules that supervise ownership or regulate competition can be grasped in their cognitive dimension. Taking a classic example (Durkheim, 1893), the legal rules for preventing the risks of fraud in the drafting and performance of contracts are constraints, but they are also a source of greater predictability of behaviours, without which it would be difficult for economic actors to engage with each other.
5Even though it does not refer to the same theoretical frameworks, the research carried out in France on market mediation devices also displays a marked interest in cognitive operations. Whether it be a matter of explaining how consumers arrive at their choices, or how producers and distributors draw up their offers, this work has also reported on the cognitive equipment provided by many socio-technical devices such as labels, packaging, guides, financial calculation formulas, market studies, focus groups run with consumers, etc. (Callon et al., 2013; Cochoy, 2002; Dubuisson-Quellier, 2006; Karpik, 1996, 2010). For the sociology of market mediation, cognitive operations are distributed: action on the market involves the actors immediately present, with their incorporated knowledge, but also the actors and knowledge enshrined in the socio-technical devices they mobilize. This means that supply and demand need to be perceived not as prior constituted realities that it would be a matter of linking via the market, but to see market mediation devices as also acting on both supply and demand. For example, the Michelin Guide, like the buyer’s guides drawn up by consumer or environmental associations act upon demand by educating and equipping consumers, but also on the supply, by establishing a repository of practices supposedly expected by an increasing number of consumers (Dubuisson-Quellier, 2013; Karpik, 2000; Mallard, 2000). For the same reason, and extending the consideration of these devices to the knowledge tools that economic theories amount to, a new study area is opened up, not so much for considering these theories in their ability to describe and analyse, but also their ability to establish or transform the economy and the markets (MacKenzie, Muniesa and Siu, 2007). In particular, the theories established in marketing, along with the devices that relay them, play an active role in configuring commercial spaces (Araujo, Finch and Kjellberg, 2010; Cochoy, 2014; Kessous and Mallard, 2014).
6Given the centrality of the cognitive issue in economic sociology, it is interesting to take another look at the work undertaken in order to show how much importance it assigns to knowledge. Unlike what an initial formulation might suggest (Zukin and DiMaggio, 1990), “cognitive embeddedness” should not be studied distinctly from structural, political or cultural embeddedness. Apart from reserving this term in a barely sociological perspective of studying the cognitive capacities of human beings, there is no justification for approaching the cognitive issue separately from networks, rules, culture or devices.
7At least since the work by Mark Granovetter (1973), it has been established that networks, the recurrent relations that radiate through the economy and the markets, have impacts in terms of knowledge. An initial challenge is the circulation of information. The type of networks and the position within them are sources of unequal access to information usable on the market, concerning for example consumer tastes or supplier reliability (Uzzi, 1997). The figure of the entrepreneur can thus be considered as that of a person able to take advantage of his interface and controlling position between spheres producing different types of information (Burt, 1992). The benefit may also come from a positioning enabling knowledge of the different types of evaluation taking place in different areas of the social world. This idea, which is recurrently found in the entrepreneurship literature (Zalio, 2009), is particularly well illustrated by the description of the work of second hand and antique dealers described by Hervé Sciardet (2003). For instance, second hand dealers will look for bargains at flea markets or will visit the recently bereaved to seek items they know they will be able to sell on rapidly to antique dealers specializing in a given type of product (old books, toys, etc.), antique dealers who may, in their turn, sell on at a higher price to collectors and/or abroad.
8A second challenge concerns the question of trust (Granovetter, 1985; DiMaggio and Louch, 1998; Karpik, 1996, 2010). Faced with uncertainty surrounding the quality of goods and/or asymmetries of information inducing a risk of fraud, relational networks may be a mine of distributed information accessible to and usable by stakeholders. For example, for the purchase of cultural products, personal relations are often judged to provide more relevant advice than professional critiques. Likewise, once agents have been involved in trade, know each other or believe they share certain values, further trading can take place without a heaviness of legal precautions and controls. This does not mean believing that networks prevent fraud: M. Granovetter, in an important comment, highlights that taking part in a network may even be an exceptional resource for swindling the other members. On the other hand, networks offer the additional trust that makes it possible to infer honesty and engage in trade.
9This increase in the predictability of the characteristics of goods and of people’s behaviours is also enabled by turning to certain “market institutions”. Like Sophie Dubuisson-Quellier and Pierre François (2011), we group together under this term the socio-technical devices that the sociology of market mediation has brought to light, and the formal rules and shared cultural conceptions that neo-institutional sociology studies. Logos guaranteeing the respect of certain criteria (Karpik, 1996) such as the legal rules and cultural conceptions establishing acceptable competitive practices on a given market (Abolafia, 1996) constitute forms of knowledge. For all that, it is also important to clearly grasp how incomplete that knowledge is. In the neo-institutionalist perspective, the cognitive frameworks whereby the strategies of economic actors are established depend on places and moments. For example, the strategy designed to break up conglomerates, the dominant feature or American finance in the 1980s, was backed up by theoretical reasoning and statistical data, but there was nothing to guarantee that it was the best possible strategy. That strategy was also imposed because, at the time, the cognitive arguments in favour of it were the most available ones (Davis, Diekmann and Tinsley, 1994). Likewise, market devices can also be perceived as much by the information they transmit as by the information they hide (Cochoy, 2002; Mallard, 2000). For example, an “Organic Farming” label guarantees the respect of certain types of production, but says nothing about many other characteristics, to such an extent that some producers or consumers argue in favour of alternative types of certification (Teil, 2013). Over and above hidden or masked information, it is also conventions that back up evaluations that are more or less enshrined in these systems. For example, an industrial convention centred on sanitary traceability says nothing about the geographical origin and the terroir of agricultural products (Eymard-Duvernay, 1993). Lastly, as shown in the work by Marion Fourcade (2011) on methods for the economic assessment of the damage caused by two oil spills in France and the USA, the calculation systems also have to be considered in their relation to culture. Some very different cultural conceptions of nature, such as the wild space in the USA and the lived space in France, have led to different evaluation methods being adopted, resulting in enormous differences in the levels of compensation practised. This case thus confirms to what degree market institutions enable and give sense to economic calculation, but also how much they do so in a socially bounded way.
10The knowledge conveyed by networks and market institutions ultimately constitutes an issue for competitive dynamics. The notion of “conception of control”, which is central to the analyses of Fligstein, illustrates it well. Fligstein (2001) defines it as both “local knowledge” explaining the relative positioning of actors in a given market and defining adequate competitive behaviours in that context, but also as “local culture” imposed by the dominant actors of that market, with a view to avoiding the most head-on types of competition. Likewise, when Michel Callon and Fabian Muniesa (2005) consider market actors as “distributed calculating agencies”, this leads them immediately to highlight that the different market actors possess neither the same calculation powers nor the same abilities to impose calculation tools. Networks and market institutions bring asymmetries of power, they serve the interests of different actors unequally. The case of devices whereby funeral businesses capture their clientele of bereaved families is exemplary in this case, so much does it show that the latter barely have any resources to counter the cognitive benchmarks mastered by the former (Trompette, 2007). Conversely, other cases bear witness to the tensions and disputes associated with an imposed market system. Thus, the standards certifying the organizational processes of businesses tend to make them substitutable, and thereby serve the interests of their clients, who can lead them more easily into competition. Nevertheless, certified businesses do not give up without a fight. They dispute the fact that these standards are enough to establish their quality and endeavour to impose other cognitive references, first and foremost their brand, on which they would like their trading relations to be based (Segrestin, 1997).
11The articles in this collection follow on from the work and issues we have just retraced. They extend and deepen that work through three sets of results. The first takes another look at uncertainty, the problems associated with the absence or instability of knowledge when defining the quality of goods. The second raises the debate on the heterogeneity of regimes and types of knowledge that can be brought into play by actors in the valuation of goods. Lastly, the third takes further the analysis of the performativity of economic knowledge by introducing thoughts on a type of commodification of the production of that knowledge.
12The cases studied by Vincent Cardon, Gérard Marty and Liliana Doganova confirm the relevance of this fundamental question: How can trade be entered into when merchandise qualities are uncertain? As V. Cardon reminds us, choosing and booking a hotel room in a town one does not know, without prior experience and without even being able to visit it beforehand, is a wager. It is even tempting to add that customers should completely refuse to commit themselves in such commercial exchanges, so great is the risk of being deceived by hoteliers who know that they have a constantly renewed clientele before them. The article by G. Marty on the auction of timbers from French state forests raises similar questions as to the right way to deal with the uncertainty of the quality of a good. The wood being sold is not visible during the auctions, and in cases where “standing wood” is sold, the trees have yet to be felled. Yet, for the professionals in the sector, wood quality is a complex and very variable aspect: for example, it is not limited to the definition of a species or a geographical origin. As it stands, the auction sales system seems to be profoundly incomplete. In the case covered by L. Doganova, it is not so much a matter of coming to terms with the uncertainty in space, as in the previous cases, but in time on this occasion. L. Doganova relates the conditions for assessing the market value of a product arising from pharmaceutical research, namely a molecule, likely to become the basis for medicines. The article looks more precisely at the criteria and knowledge on which companies base their decision whether or not to invest in research on a given molecule. The question is crucial, because the investments are high, and it involves making a decision to invest in research with high risks of failure, and for a future product whose potential outlets are theoretical and totally uncertain. As highlighted by L. Doganova, the case study, by considering the scope of possible futures, provides original insight into the issue of valuing commercial goods in a situation of uncertainty, going beyond questions of the asymmetry of information between sellers and buyers.
13How can one agree to provide funding amounting to millions of euros for pharmaceutical research without having a few certainties as to its results and its trade outlets? How is it possible to choose a hotel room out of dozens of possible options without having tested it beforehand? How is it possible to buy batches of timber without examining them beforehand? In the three texts, emphasis is placed on socio-technical devices: the TripAdvisor® internet site collecting and compiling consumer opinions and scores, the catalogue of the Office National de la Forêt (ONF) providing a whole range of information on the batches of timber put up for sale, and the mathematical calculation formula and databases enabling the funders of pharmaceutical research to carry out a reduction of probable futures, notably in terms of the cash flows generated.
14The cognitive dimension of such market systems is well known. Nevertheless, the merits of the texts in this collection lie in the fact that they do not simply conclude on a transmission of knowledge and a reduction in uncertainty. On the contrary, in the three cases the uncertainty remains (see also Goulet and Le Velly, 2013). For instance, the three texts show how such devices are subject to controversy, as regards their reliability, and challenges, through the existence of competing quality assessment methods. The controversy and challenges are then brought about by different stakeholders: buyers in the case of timber auctions, hoteliers in the case of TripAdvisor®, researchers and experts in the case of funding for pharmaceutical research. However, in all the cases, they bear witness to the partial and performative nature of the systems studied: partial because they operate a selection and a formatting of knowledge; performative, since they consequently configure the economy and the markets. The uncertainties remain, all the more so in that this partial nature is known and debated. Even so, the devices operate: they take part in guiding tourists and in funding research projects. As explained by G. Marty, even the ONF sales catalogue needs to be considered from this dual viewpoint: without being enough for the final judgment, it is involved in preparing the evaluation and generates effects on sales prices or on unsold goods. Lastly, we come to the particularly stimulating question at the heart of L. Doganova’s article: Why are such cognitive devices, which offer imprecise and partial knowledge, nonetheless used? The answer she proposes, contrary to an analysis in terms of irrationality or cultural rituals, is that it is their inaccuracy and their plasticity that lies at the heart of their uses. It is because they “calculate badly” or more precisely that they can calculate in several ways, that the mathematical formula is used, as a basis for exploring possible futures and for negotiation.
15The processes of valuing goods (Muniesa, 2012; Kjellberg et al., 2013; Berthoin Antal et al., 2015) are also a cornerstone for dealing with the role of knowledge in economic activity. Valuing a good consists in carrying out a decoupling in relation to its singular properties, and a characterization designed to enable prioritization, comparison, evaluation and calculation.
16Several articles in the collection reveal the recurrence of these valuation operations. In the case studied by L. Doganova, use of a mathematical formula to update cash flows even generates the production of a unique data item; the updated cash flow value, making it possible to compare otherwise extremely heterogeneous research and offer the bases for trade negotiation. The ONF sales catalogue also sets out to generate what G. Marty calls “market objectivity”: the possibility for timber buyers to refer in their comparisons to a single standard of judgment. It has been known since the work on professional qualifications (Thévenot, 1984), that the question of knowledge regimes and judgment benchmarks supporting the assessment of people’s activity is also crucially raised in wage relations. In this collection, the article by Jean-Marie Pillon also shows this well, by exploring the workings of a very precise activity involving staff at the Pôle Emploi job centre. Faced with the large number of job seekers that each advisor had to deal with, and faced with the insufficiency of coding based solely on qualifications and professions, they were given the possibility of entering a written comment of 400 characters maximum. The comment standardized in that way was designed to bring out other characteristics of job seekers, such as their attitude to job seeking and their past professional experiences. It reflected the importance of the ordinary and singular knowledge that the advisors acquired on each particular case, in order to enrich the assessment of the potential professional value of job seekers beyond the standard analytical grids, improve the circulation of information between job centre staff members, and ensure more precise matching between job offers and seekers. Lastly, it is the same question of comparability, and the continuity of knowledge and the qualities that it helps to define, that is raised in the article by V. Cardon regarding TripAdvisor®. Indeed, this research raises the issue of mechanisms that enable a subjective assessment of the views of multiple consumers based on how they feel about their experiences, in a multicriteria score designed to provide Internet users with ways of comparing and discriminating between offers that are more synthetic, transversal or global.
17These articles also inform us about the diversity of knowledge types, or regimes, involved in assessing commercial goods and services, and about the debates that this diversity gives rise to. Are all types of knowledge equal in value? User knowledge, in much the same way as it has acquired an increasing role in the medical sector through patient recognition (Epstein, 1995; Rabeharisoa and Callon, 1999), has become an essential element in guiding the choices of other consumers. In this respect, the case studied by V. Cardon touches upon the disagreements and debates that might arise from the views expressed by users, with some professionals considering themselves wronged by certain comments that they judge unfair, or even fraudulent in a fair number of cases. The comparison is then made with another market device, itself seeking to characterize the value of hotel services: the stars awarded by the Ministry of the Economy, and the standardization associated with them. The question is once again raised about the reliability of knowledge and of the systems that concentrate and support it, due to their ability to guide the outcome of a commercial undertaking either positively or negatively. In order to limit fraud due to unfounded opinions, the article by V. Cardon recounts the attempt made by the professionals and the Internet site to certify and trace the “true” knowledge of users and, in a certain way, draw the dissemination of lay knowledge, which is often borne at the outset by a spirit of emancipation from conventional prescription systems, into supervised and stabilized formats that become a prerequisite for its survival. In this article, as in the one by J.-M. Pillon about the Pôle Emploi job centre, the professionals in the market, and users, are thus faced with the challenge of supervising and formatting singular knowledge and minimizing bias, whether of a type linked to idiosyncrasy or to falsification.
18While the ultimate aim of valuing goods seems to be akin to the good functioning of the markets, the case of the ONF auction catalogue leads us nevertheless to reconsider this assumption. For instance, G. Marty shows that this market is characterized by buyers who have such different uses for wood that it is impossible for the ONF to propose valuation criteria in its catalogue that make sense to each of them. Referring to the work by Lucien Karpik (2010), G. Marty shows that, against all expectations, the wood purchased during these auctions falls into the category of singular goods: it tends to escape all attempts at a single classification based on shared criteria.
19The final article in the collection, proposed by Tommaso Pardi, leads one to take a closer look at the performativity of knowledge and its capacity to order markets. T. Pardi describes the conditions under which the theory of “lean production” elaborated in a laboratory at the Massachusetts Institute of Technology and disseminated in a book for the general public (The machine that changed the world), imposed itself durably at the turn of the 1990s as the dominant interpretative framework for explaining the success of the Japanese car industry. It thus offers a detailed and fascinating description of the history of this theory through a study of the conditions for its production and success. T. Pardi shows particularly well that it would be hazardous to believe that this success is down to a greater relevance or explanatory capacity of the theory: other analyses of Japanese superiority, proposed in the 1980s, were also conclusive and the proof put forward by the authors of The machine that changed the world were lightweight to say the very least. On the other hand, in order to grasp the success of this theory, it is necessary to see how it was in phase with the inflections that were taking place at the time within the car industry and the union movement, inflections that the theory somewhat “validated” before generalization of the promoted production method, in its turn, confirmed the theory.
20At the end of his investigation, T. Pardi also highlights the transformations that American universities were going through at the same period. He suggests that increased relations with industry, through partnerships and funding, considerably modified the content of the management science research being conducted in those universities, making it much less critical of “managerial fashions” and of the simple reasoning that came with them. This absence of critical scientific discourse then participated in the performativity of “lean production”. What T. Pardi calls the commodification of academic research, echoing a critical discourse that is well developed on the commodification of knowledge (Busch, 2014), thus appears to be a vector explaining the performativity of certain types of knowledge. In brief, there is a suggestion that while knowledge sometimes makes the market, the market also sometimes makes the knowledge.
21At the end of this article, our first conclusion is that it is important to give thought to the cognitive processes that occupy markets through the tensions they harbour, due to the partial nature of the knowledge transmitted, to the debates associated with the knowledge regimes underpinning them, and the difficulties they encounter in their attempts to achieve objectivity detached from singular experiences. In particular, an examination of the different contributions provided by this collection suggests the merits of applying conceptual tools proposed in fields other than economic sociology as such, in order to understand market knowledge. We have seen here, for example, the contribution of concepts such as distribution, derived from cognitive anthropology research (Hutchins, 1995), or such as controversy, that the Science and Technology Studies have contributed to popularize (Latour, 1987). Such tools are all the more important in that they make it possible to distance oneself from the concept of information, inherited from economics but which very often serves implicitly to grasp the cognitive processes in economic sociology work. It is then all the easier to question the plurality of the knowledge formats brought into play, and the role of the associated controversies in the dynamics of market construction. From tacit know-how to the information incorporated in technical systems, from scientific articles to “airport books” such as The machine that changed the world, and including individual and collective knowledge, the types of distribution and debating of knowledge in market groups are very diverse. Studying how knowledge emerges, is transformed and switches from one format to another, how it is challenged and reinterpreted depending on the market evolution and structuring process, is an analysis challenge that this collection invites us to explore.
22Lastly, this thought process opens up specific ethical and political dimensions linked to the way markets and knowledge are interconnected. On the one hand, the increasing role played by knowledge in market functioning raises the question of its democratic control. This is seen with the contemporary debates on the challenges associated with the transparency of financial information, or privacy problems raised by the trading of consumer information, which is developing on social networking platforms. On the other hand, debates focusing on market logics involved in the funding of research and the appropriation of its results bear witness to questions regarding the economics of knowledge production and dissemination (Malissard, Gingras and Gemme, 2003; Lefebvre and Trompette, 2005; Mirowski and Sent, 2008; Roger 2013). Continued analysis of how markets and knowledge are interconnected thus proceeds from both fertile theoretical challenges and major political questioning.
The compilation of this collection follows on from crossed sessions organized by the Thematic Networks 12 (Economic Sociology) and 29 (Sciences and Technologies in Society) of the French Sociology Association, during its fifth congress held in Nantes from 2 to 5 September 2013. It was coordinated by Frédéric Goulet and Ronan Le Velly, with support from Marc Barbier, Pascale Trompette and Dominique Vinck for the Revue d’Anthropologie des Connaissances. The authors of this introductory article also thank Éric Dagiral for his contribution to the deliberations presented here. CIRAD and Montpellier SupAgro provided financial backing for the publication of the collection.