4.1. King’s College, this “small society” [14]

16Then as now, the University of Cambridge was like an administrative layered cake. Life on the university was structured by the Colleges, 31 in number since 1977. They were, and still are, multidisciplinary and autonomous institutions with their own funding, rules of recruitment, and premises. In addition to Colleges, departments gathered teachers and researchers by discipline, sometimes further divided into committees. To take an example, a researcher like Gambetta was a Junior Research Fellow at King’s College but also a member of the Social and Political Sciences Committee (SPS), a sub-division of the Department of Economics. In this mid-1980, King’s College was considered as a large College (between 60 to 80 members according to John Dunn). [15] It was a place where members “saw each other all the time and mix in all sorts of environments.” [16] Dunn explained that this was by no means a fortuitous event. The Provost (name given to the position of Director of King’s College Cambridge) was considered a true “social entrepreneur” who always worked to strengthen this collective emulation through the organisation of daily life on the one hand, and of seminars on the other. [17]

17The daily life was organized in order to maximize the cohesion of the fellows. This was done first through meals. Lunch was and remains a major structuring event in life at King’s College. It was, in everyone’s opinion, a crucial moment to make a good impression in the eyes of the other Fellows: “you don’t want to appear a fool in their eyes, you want what you do to look interesting to them and you want to pick up quickly on what they are interested in”, David Good remembers. [18] In addition to these lunches, it was not uncommon for fellows to have dinner at the College. Everywhere at the University of Cambridge, the organization of the scientific life was embedded in social life. Dasgupta shared the same experiences and pointed out that it was not unusual for the College to become a residence for “single people and fellows whose partners did not live in Cambridge.” [19]

18Attending seminars on a daily basis was also a crucial element of the social and scientific life in Cambridge. Most interviewees consequently described Cambridge and King’s College as a “very oral place”. [20] Dasgupta referred to the seminars as “a sort of group of friends getting together”, [21] a vision that is also reflected in the words of Good when he shared his encounter with Gambetta: “I knew him through a group of friends, which was also a junior reading seminar.” [22] This proximity between the fellows eventually became a joke between members of the College: “It was a very informal little society, as my friend Pascal Boyer once said, King’s College is a place where people are very well known... to each other.” [23] The intellectual consequence of a social organization geared toward multiplying social interactions from morning to night was interdisciplinary, one emphasized by Gambetta in the foreword to his 1998 book: “The College was the donor of something immeasurably more valuable than cash: a unique environment where an interdisciplinary exchange is encouraged and facilitated by daily proximity.” (p. ix). Furthermore, the financing of the seminar relied largely on King’s College and to a lesser extent on the SPS which further encouraged cross-disciplinarity. [24]

4.2. A mostly Cambridgian casting

Figure 1

Reconstruction of the calendar of the Seminar on Trust and Social Change.

Reconstruction of the calendar of the Seminar on Trust and Social Change.

19The list of organised sessions showed a particularly local casting. Out of the thirteen speakers, eleven came from Cambridge University. Only Niklas Luhmann came from the University of Bielefeld and Thomas Schelling came from Harvard.

20Among the invited participants from outside of Cambridge, one can distinguish three cases. First, those who, like Elster, share some intellectual proximity with the project but could not participate for “practical reasons”. [25] One was Albert Hirschman, who declined his invitation in September 1985, claiming to be too “overburdened”. [26] Hirschman [1970, 1977] participated to the extension of rational choice theory by building on the work of Simon and Amartya Sen to propose a constructive criticism of rational choice theory (see Cot [2013]). It was mainly in his book Exit, Voice, and Loyalty that he developed the idea that rationality was limited because it was necessarily “located in a physical space, in a space of resources or in a space of coordination possibilities” (Thireau [2015]). Another was Schelling, who criticized the standard conception of game theory as a “mathematical toolbox for economics” (Giocolli [2003], p. 111), against which he attempted to “anchor game theory in the real world by highlighting the importance of social events, institutions, convention, etc.” (Larrouy [2013], p. 1). This is in particular the objective he pursued when he developed the concept of the focal point: a game solution that people tend to choose by default in the absence of communication (Schelling [1960]).

21Second, there was the case of Luhman, who was not interested in renewing rational choice theory but who had worked on the notion of trust before the seminar (Luhman [1979]). Gambetta probably sensed that Luhmann’s sociological work on trust was not incompatible with rational choice theory, because he defined trust as an interpersonal phenomenon rather than an institutional one (such as self-confidence, or confidence in the future).

22Dasgupta represents yet another type of case. Intellectually, Dasgupta showed interests in both the movement for the renewal of rational choice theory, and in game theory. He had participated in the book Utilitarianism and beyond [1982], directed by Amartya Sen and Bernard Williams. He had also collaborated with Kenneth Binmore on two books Economic Organizations as Games [1986] and The Economics of Bargaining [1987], the latter of which was then considered as “the major exposition and exploration to date of the meaning and implications of the Nash programme” (Petith [1987], p. 995). Sociologically, Dasgupta was the only member of Cambridge that Gambetta did not know personally before inviting him. The latter remembers:

23

“It was very unusual. A young man calls me on the phone […] He said that he would like to meet me, he was an Italian young Research Fellow at King’s College by the name of Gambetta, as you know, Diego. [...] He came to see me and explain to me that he wanted to start an all-year seminar on Trust, and he asked me if I would want to participate in this seminar.” [27]

24Although Dasgupta belonged to Cambridge University, his invitation is more similar to the ones addressed to the external scholars than to the internal ones.

Figure 2

Age of the seminar speakers in 1985.

Age of the seminar speakers in 1985.

25Other invited Cambridge scholars included “important figures from King’s College” [28] like Williams (Provost of King’s College), Bateson (who succeeded to Williams in 1988), Ernst Gellner. [29] Other often younger colleagues included those that Gambetta considered as friends: Lorenz, Good, Pagden. The seminar operated as a socialization space, especially in King’s College. The schedule of the seminar, from 5 to 7 p.m., was not chosen at random, it allowed a post-seminar socialization: “There was always a drink reception afterward. Quite often, several of us have dinner together. The idea was the seminar was the place where the discussion started but it went through the evening.” [30]. This post-seminar socialization was the will of Gambetta, whom Lorenz described as a true “social and scientific entrepreneur” [31]. Indeed, it was Gambetta who asked the SPS for an extension of funding to pay for the dinner for thirteen people after each seminar. [32] It was also Gambetta who proposed to the speakers to transform the seminar into a collective book as early as January 1986, when the seminar was not yet halfway through.

26The casting was thus influenced both by the Cambridge socialization tradition and the intellectual agenda that Gambetta thought should unify contributions. Philosopher Williams proposed to use game theory in order to reflect on the “motivation [s] to cooperate” (Williams in Gambetta [1988], p. 4) and on how one can derive the social conditions for cooperation from these motivations. Ethologist Patrick Bateson drew a parallel between cooperation among animals and among humans. He proposed “an evolutionary path through which trust may have emerged as an aspect of awareness and an essential element of human cooperation” (Gambetta [1988], p. xi). Social psychologist Good presented some experiments on the Trucking Game. Originally invented by Deutsch and Krauss [1960], it considered two manufacturers transporting some product to a sale point by road. Two roads could be used, but the shortest route could only be used by one truck at the time, so that defining a transportation schedule required some cooperation. It is within this framework that Good analysed “the implications of a wide spectrum of social-psychological evidence for the notion of trust and for the rationality of its manifestations” (Ibid.) Dasgupta focused on repeated interaction within the framework of game theory. The aim was to show that trust is a particular resource that grows as it is consumed. Political scientist Dunn presented a comparative study of different political systems and the multiple forms that trust could take within them. His analysis led him to believe that rationality alone cannot explain trust.

27Other presentations likewise dealt with trust, but in more empirical and historical ways. Historian Anthony Pagden focused on Spanish rule in eighteenthcentury Naples. He described the ways in which trust can be intentionally destroyed for the purpose of domination. Political scientist Geoffrey Hawthorn dealt with the question of whether and how vicious circles unconducive to economic development can be broken and overridden using examples from India or South Korea. Social anthropologist Ernest Gellner argued that anarchy, the absence of a centralized authority, can be productive rather than destructive for trust and cohesion. Economist Edward Lorenz reported how trust was built up between French shipbuilding firms and their subcontractors. Finally, social anthropologist Keith Hart discussed how migrants in Ghana identify safe economic relations. The contributions of the seminar were therefore either focused on the rationality of trust, or on how institutions influence trust relationships. The intellectual consistency of the seminar can also be deduced from the list of scholars that Gambetta selected to attend the sessions.

28It is interesting to note that Gambetta arranged people on this list either by College Department or by discipline. If we exclude “protocol guests” from King’s or from SPS, economists were over-represented. Two of these economists were especially thanked for their “frequent and illuminating contributions to the discussion” (Gambetta [1988], p. x) in the foreword of the 1988 book: Frank Hahn and Ugo Pagano. An honour that was by no means insignificant since they were, with Tony Tanner, a researcher in literature, the only ones to be mentioned in this way. Frank Hahn had also participated in the collective book Utilitarianism and beyond directed by Sen and Williams [1982]. In this book, Hahn’s approach is linked to that of Dasgupta, both “pointed to specific difficulties with utilitarianism and these would require relaxing the un-compromising narrowness of the utilitarian vision, but possibility not a rejection of the entire approach” (Sen and Williams [1982], p. vii). In his thesis supervised by Robert Rowthorn, who was also invited to participate in the Trust Seminar, Ugo Pagano was very critical of economists who considered rationality as a simple relation between means and ends. To him “if this relation is interpreted in the unnecessarily restrictive way that means cannot also be ends, prescribing and ascribing rationality to others, this can turn the economist into a modern vision of a conservative priest – a role to which I profoundly object.” (Pagano [1985], p. 176). Having established the institutional framework of the seminar and the way Gambetta formed its cast around its research program, we will now focus on the content and effects of the seminar.

Figure 3

Trust Seminar Guest List

Trust Seminar Guest List

5.1. Game theory as a sociological method

30Gambetta had never used game theory before these seminars. Meeting Dasgupta was decisive in his adoption of this method. As he himself recounted in retrospect: “Partha Dasgupta was important because he made me understand how you can use game theory in a very intelligent way. [...] he had a way to reason about economics which I found very very inspiring.”. [33]

31The influence of game theory in the seminar was clearly reflected in Gambetta’s concluding chapter, “Can We Trust Trust.” Although it was written by Gambetta, it can be studied as a quasi-collective work to the extent that Gambetta explained that he wished to “reconstruct what seem to me the central question about trust that the individual contributions presented in this volume raise and partly answer” (Gambetta [1988], p. 213). He also took care to specify in a footnote that “in the essay converge not just several of the ideas which contributors have published in this volume but also those which were patiently expressed in conversation, and thanks to which my reflections on trust were shaped and reshaped” (Ibid.). In this article trust is defined as “a particular level of subjective probability with which an agent assesses that another agent or group of agents will perform a particular action, both before he can monitor such action (or independently of his capacity ever to be able to monitor it) and in a context in which it affects his own action (see Dasgupta and Luhmann in particular, this volume)” (p. 217). This approach was vastly different from what Gambetta proposed in his own chapter – “Mafia, the Price of Distrust” – in which there was neither a question of trust as a subjective probability nor even a mention of game theory. Other attendants were equally enticed by this new method that they had not previously encountered. One example was Robert Rowthorn who, according to Dasgupta: “changed his opinion on how to do economics [...] [and] got taken by the game theorist’s approach.” [34].

Figure 4

Number of citations per chapter in order of appearance in the book.

Number of citations per chapter in order of appearance in the book.

32Game theory was to have a profound impact on Gambetta’s research and his approach of trust. It was with Michael Bacharach, a game theorist, that he produced the work that he considered “the most accomplished of my contributions on this notion.” [35] The vision of the game theory of Bacharach was close to that of Schelling in that they “shared an apparently identical assessment of game theory’s limitations, and both attempted to surpass them. [...] [Bacharach] tried through his entire career to avoid traditional game theory’s pitfalls unable to account for human behaviour through its approach in terms of perfect rationality and common knowledge, by investigating the epistemic assumptions in games, the consequences of an idealized conception of player’s reasoning, and lastly, individualism in game theory” (Larrouy [2013], p. 2). Bacharach and Gambetta jointly received two research grants: one for 1994-1996 from the Economic and Social Research Council and the other for 2000-2001 from the Russel Sage Foundation. In their 2001 article, Bacharach and Gambetta combined game theory and signalling theory. The idea behind this association is that during interactions everyone had a “all-thing considered” payoff. That payoff structure differs from the one constructed with the assumption that the individual is rational and selfish. In their “all thing considered” payoff the individual considers a set of ethical and social considerations. For trust to be created between two people it is necessary to send to the other person signs, or “manifesta”, to indicate our own payoff structure. Therefore, trust is defined in line with the notion of subjective probability mentioned above as a “particular belief, which arises in games with a certain payoff structure.” (Bacharach and Gambetta [2001], p. 148).

33While the seminar space allowed the game theory method to move, in this case from economics to sociologists, the subject of this seminar – the notion of trust – did the opposite.

5.2. Trust as an economic object

34The seminar created a space that allowed participants to grasp objects that were not commonly investigated in their respective disciplines. In economics, it was with the publication of the book that trust became a bona fide theoretical object in economics. Two articles in the 1988 book deserve special attention in this regard. The first one is Dasgupta’s “Trust as a Commodity”. It was the first economic article to mobilize game theory to explicitly discuss trust. In this article Dasgupta defined trust “as a subjective probability” (p. 62) in a way that likely influenced the definition offered in Gambetta’s conclusion. This article also broadened the category of social dilemma games through adding trust games that were subsequently used by experimental economists (Camerer and Weigelt [1998]; McCabe et al. [2003]). The second chapter deserving special attention is the concluding chapter written by Gambetta “Can We Trust Trust.” This article determined how economic theory subsequently addressed this notion of trust and shaped the debates that arose from the introduction of this new object. In line with Luhmann’s contribution, Gambetta discuss trust as a relation between a person and another person (or a group of persons). This was to become the notion of trust at the heart of the debates in economics. It was for example the one referred to as “tripartite” trust by Russel Hardin [1993], in his influential article “The Street-Level Epistemology of Trust”. The latter made explicit reference to Gambetta’s work and more generally to the Trust seminar (he cited 7 of the 13 chapters of the 1988 book, which account for more than 20 % of the total references in his article). Oliver Williamson [1993] also mentioned the trust seminar in his article “Calculativeness, trust and economic organization.” He framed the whole article as an attempt to “explicate what Diego Gambetta has referred to as “the elusive notion of trust” (p. 453), a quote he took from the introduction to the 1988 book. In this article, Williamson argued that trust can, and should be, considered as an objective probability in economics. Trust then becomes the result of a calculative process, which is what led Williamson to reject the very term trust since “calculative trust is a contradiction in terms” (p. 463). Worse, as the use of the concept of trust in economics only “muddies the (clear) waters of calculativeness” (p. 471), it would therefore be necessary to substitute the notion of trust for that of risk. Williamson’s article pointed to what would become the major debate in trust research: is trust a reflection of risk or a reflection of uncertainty? These two major antagonistic positions have been at stake in numerous approaches: incomplete contract theory, reputation models, Bayesian games theory, experimental economics, transaction cost theory or the economics of convention. Several participants to the seminar took a clear stance in this debate. Like Bacharach and Gambetta in the article “Trust in Sign” [2001], we have already discussed, or like Dasgupta and Lorenz who also refused to reduce trust to risk. Dasgupta made trust an element of social capital while refusing to see the social world as stable and certain (Dasgupta [2009, 2011]). Lorenz was extremely critical of the theory of incomplete contracts which, according to him, inevitably reduced trust to risk (Lorenz [1992, 1999]).