Friday, July 27, 2012

The Social Epistemology of Reputation

Draft. Do not quote without permission. Submitted to the journal: Social Epistemology


Abstract We monitor the informational environment and catch reputational cues, gather signals from our informants and develop our trustful attitudes in context. I present here the project of an epistemology of reputation as a way of using social configurations to acquire information. I review the definitions of reputation that exist in social sciences, stress the importance of the relational/social dimension of reputation as a property of entities and put forward a definition of reputation suitable for epistemology. I then sketch some social configurations that allow us to extract reputational information and some typical heuristics we use to navigate the social information around us.

In the first episode of The Newsroom, a new television series released by HBO on June 2012 set behind the scenes of the fictional news channel ACN, Jim Harper is a young producer who is striving to establish his credibility in the eyes of the narcissistic and charismatic Will McAvoy, anchor man and managing editor of the News Night show. He manages to find two key witnesses about a potentially disastrous oil spill in the Gulf of Mexico, but doesn’t want to reveal his sources. McAvoy insists, saying that he is too young and not credible enough yet to avoid unveiling his sources. Jim hesitates. He has been brought to ACN by his boss, McKenzie McCale, executive producer of the News Night show and Will’ ex-girlfriend, who still has a strong influence on Will. Then he says: “You don’t need to trust me. She trusts me and you trust her, that’s enough”. That is, when reality is vague, when checking the facts is difficult, the ultimate objective information you may give about your trustworthiness is your reputation, that is, the information about the social ties that flow towards you and say who you are and what you know. This article is about the role of reputation in knowledge matters, that is, how the social information around us, that defines our role in a hierarchical network of people, things and ideas, can be used to validate or invalidate what we come to know. My choice of this topic for the 25th anniversary of this journal has a double motivation: one that is anchored in the past and the other that is projected in the future. Anniversaries are occasions to think about past achievements and visions of the future. In 25 years, social epistemology has become a mainstream approach to understand how knowledge is produced, disseminated and validated. Nobody would dare to deny today, not even the most traditional epistemologists, that the way in which social factors influence knowledge processes is not an accessory inquiry that complements a more central epistemological concern about how people come to know, but it is an essential aspect of our cognitive appraisal of the world. Without considering the multiple social, cultural and historical dimensions in which our theories of the world are embedded, we simply fail to grasp the reality we want to describe. In my own work during the past ten years, I have explored the role of two special social “forces” in shaping our knowledge: trust and reputation. Our social ties - the way in which we trust each other and manage credibility - are essential ingredients of the way in which we come out believing that the world is as it is. Thus, in my research, social epistemology has been a major intellectual tool in understanding how and what we come to know. Yet, while trust as an epistemological notion [Origgi, 2004, Fricker, 2007, Faulkner, 2011] has become a mainstream concept for social epistemologists, reputation, that is, the role of social networks and rating systems in understanding knowledge processes, is still a very marginal notion in social epistemology and in philosophy in general. That is why, for this anniversary, I would like to project social epistemology in the future and open the discussion around a theme that, in my opinion, will become crucial in this area of inquiry in the following years. In a sense, my interest in reputation is a symptom of the air du temps: As information grows, our distance from reality seems to grow proportionally. Our judgements, evaluations, decisions are the more and more embedded in networked systems, such as the Web but not only, that provide us with ratings, “likes”, recommendations, as if information to be accessible should be “pre-evaluated” by various social networks and rating systems. These devices proliferate in an information-dense society in order to reduce the cognitive deficit caused by a too wide offer of information . Yet, while the explanation of the role of these reputational devices has been object of study in sociological economics [Karpik, 2007, Podolny, 2005] and market theory [Akelrof, 1972], very little has been said so far about their epistemological role in the production and acquisition of information. We are facing a major informational revolution in which the assemblage of information is social and in great part automatic and based on statistical matching algorithms about social behaviour whose functioning is opaque. The fact that Facebook suggests us friends on the basis of our profile information, or that the New York Times suggests us further readings on the basis of the content of the article we are reading are examples of this automatic and yet social filtering of information. We base our judgements on these filters and on the new categories that are generated by the networks of information. The epistemological means available to understand what happens with the organization of knowledge and its classification in such new environments are still quite limited. So, take this essay as a first step towards an epistemology of the present that takes the question of the use and misuse of reputational and rating devices seriously. In the following, I will try to sketch a definition of reputation suitable for my epistemological purposes. I will present then some theoretical background about the notion and try to situate the concept in the literature that is spread around many disciplines: economics, sociology, but also signalling theory, theory of strategic interaction, marketing studies, etc. I will then present a sort of “phenomenology” or reputation by describing some examples of social configurations that are familiar ways to use reputation for extracting information from the world. I will conclude with some remarks about the possibility of a normative stance towards the use of reputation as an epistemic device.

1. What is a reputation?

Reputation, from the verb puto in latin, meaning “counting, considering” plus the suffix re- that indicates the repetition, is the consideration of the value of an agent by other agents based on his or her past actions and creating expectations on the future conduct of that agent. Reputation is a special kind of social information: it is social information about the value of people, systems and processes that release information. Every social interaction brings forth an evaluative dimension of reciprocal judgment, a perception of who we are that we leave in the eyes of others. Every social interaction brings forth also a mastery of this presentation of ourselves, a consciousness of the image of ourselves we want to leave track of through our behaviour. As Hume says, we learn about our conduct as we learn about our bad breath: the reactions of others act as a mirror in which we discover features of ourselves we were not aware of. The evaluative dimension of our social life, the generation of opinions on each other’s actions is reputation. Reputation is commonly seen as the informational trace of our past actions: it is the credibility that an agent or an item earns through repeated interactions. Reputation is a relational property: it is the informational value of our interactions. If interactions are repeated, reputation may conventionalize in “seals of approval” or disapproval or social stigmas [Origgi 2007]. The notion of reputation has surprisingly attracted little attention in philosophy. Much of the literature that revolves around this central feature of our social and cognitive life can be found in economics, social psychology, evolutionary anthropology and theory of strategic interaction. In the following, I will present a review of definitions and uses of reputation in the various social sciences that have approached the notion, criticize them and propose my own definition that I consider more suitable for my epistemological purpose.

1.2. Reputation as an expectation and as a signal

 The notion of reputation in social sciences has been mainly treated in economics. In Adam Smith’s liberal social theory, reputation is seen as a way of coordinating activities in a decentralized social space of transactions. According to Smith, in a free society, markets coordinate diffused knowledge in an asymmetrical way: people have a partial view of what other people know and how they will act. Also, given that most transactions occur over a lapse of time, parties have to trust each other that they will satisfy their reciprocal interest. These informational and temporal asymmetries call for efficient means of storing and retrieving information about possible partners in interactions. Reputation is more than pure information: it is evaluated information, that is, a shortcut of the many judgements and interpretations that people have cumulated about an actor. That is why people are interested in keeping a “good” reputation by signalling to potential business partners their trustworthiness. In his Theory of Moral Sentiments, Smith famously states: The success of most people almost always depends upon the favour and the good opinion of their neighbours and equals; and without a tolerably regular conduct these can very seldom be obtained. The good old proverb, therefore, that honesty is the best policy, holds, in such situations, almost perfectly true. (TMS, 63) That is, according to Smith having a good reputation pays in the markets. If you are a reliable player, your business will thrive because your potential clients will form positive expectations about your actions in the future based on your records. Reputation is thus, in this tradition, a cognitive notion: it is the expectation about your future actions that agents form given your past behaviour. Your past records are a signal of your future intentions, a signal on which people can rely if they decide to interact with you. In market theory, reputation is seen as a signal or a set of signals. The Nobel Prize economist George A. Akelrof [1970] has shown that quality uncertainty is such a risky feature of markets, that reputation is needed: “Seals of reputation” in a market are labels, certifications, guides, that is, all the devices that tend to reduce the informational asymmetries between buyers and sellers. A rational agent, according to Akelrof, has an interest in embodying these devices in order to compensate the cognitive deficit of the informational asymmetry. Reputation as a signal is studied within a broader body of work that goes under the name of Signalling Theory [Gambetta 2009]. Signalling Theory aims at solving a fundamental communication problem: Given an interaction in which interests diverge between the two parties, how can a party be certain of the qualities of the other party? Honest signallers will try to signal their good qualities (trustworthiness, accountability, strength), but dishonest signallers will try to do the same, by mimicking high-quality signals. One may try to “make up” a reputation and by signalling past actions that never took place. For example, one may signal his or her wealth by inviting you in a luxury apartment that in fact has been borrowed by a friend. Thus robust signals, that is, signals that are difficult to fake (like having an accent, or having a physical sign of strength) are considered as the most reliable [Zahavi, 1981]. In this tradition, reputation is thus a strategic feature of interactions: it is the set of beliefs/expectations crystallized by the past interactions or elicited by a present interaction, a way for the perceiver to ground his or her trust in the perceived and a way of the perceived to send the appropriate signals in order to enhance the chances to be trusted. Yet, the notion doesn’t say anything about the possible biases in the construction and the maintenance of a reputation. As an expectation, reputation is seen as a transparent record in which all that is known about the perceived agent is what has been revealed through the past interactions or what has been intentionally signalled in a present interaction. Yet, as we shall see, reputation is a much broader social property, that is mutually constructed by the perceiver, the perceived and the social context. As we will see, judgments of reputation involve always a “third party”, that is, a community of peers, experts, or acknowledged authorities that we defer to for our evaluations. Reputation is in the eyes of the others: we look at how others look at the target and defer, with complex cognitive strategies, to this social look. I will come back to these questions later, after having gone through some other uses and definitions of reputation that we find in social science and can be relevant for my purposes.

1.3. Reputation, esteem, recognition and all the passions of glory

 Although Smith endorses a vision of reputation as advantageous cooperation, in other passages of the Theory of Moral Sentiments, he puts forward another view of reputation, based on the human passion for sympathy, and more related to other concepts such as honour, or recognition: “Nature, when she formed man for society, endowed him with an original desire to please, and an original aversion to offend his brethren. She taught him to feel pleasure in their favourable and pain in their unfavourable regard. She rendered their approbation most flattering and most agreeable to him for its own sake; and their disapprobation most mortifying and most offensive” (TMS, 116) And again: “It is the vanity, not the ease or the pleasure, that interests us” (TMS, 50) In this second sense, reputation joins another family of philosophical/sociological notions such as honor, glory, esteem, recognition, as the fundamental human social passion to be esteemed and considered by the others. In his Elements of Law and Natural Politics, Hobbes defines “honour” is a similar way: “And according to the signs of honour and dishonour, so we estimate and make the value or WORTH of a man” In the same line, that is the line that sees reputation as a primary motivation not reducible to more fundamental economical motivations such as interest and accumulation of capital, Geoffrey Brennan and Philip Pettit [2004] have developed an “economy of esteem” or kudonomics (from the Greek work kudos, or renown) based on the idea that: “Among the things that may be expected to move people most reliably and forcibly is the desire to be thought well of by their fellows and the aversion to be regarded badly. People […] live under the regime or law of opinion and that regime moulds all that they are and do. It provides a gravitational field that silently shapes their dispositions and their doings, exerting a long insistent influence that can be resisted at the cost of considerable effort and self sacrifice” [Brennan and Pettit: 23] The theory of reputation in classical and modern economics is thus ambivalent between an interpretation of reputation as a mean to achieve an interest and as an end in itself. It is worth noting, although I won’t explore in depth this tradition, that in this second sense, reputation is almost synonymous of recognition, a fundamental concept in social philosophy that, from Hegel to Axel Honneth [Honneth 1992], is used to explain the motivations of the self in terms of a “struggle for recognition”, that is, a struggle to be recognized by the others along a series of dimensions. This permanent struggle among subjects for the recognition of their identity generates - according to the authors - the most fundamental pressures to societal changes. A tradition that puts together elements of the theory of reputation as a signal with a conception of reputation as a fundamental motive of social life is the theory of social distinction, a line of thought that that goes from the works of Thorstein Veblen [1905] to those of Pierre Bourdieu [1984], especially in his critique of taste and aesthetic judgement as a way of a social class to distinguish itself from another. These sociological theories can be seen also as a critique of the Marxist conception of the class conflict, which is seen in this tradition as a product of a struggle for reputation and status and a symbolic positioning in the social hierarchy more than a consequence of economic determination. Veblen writes that in the pursuit of accumulation “the struggle is substantially a race for reputability on the basis of invidious comparison” [Veblen, 1905: 32]. I will come back to these positions in the discussion of reputation as status, but let me sketch briefly how reputation as a social resource has been modelled in rational choice and evolutionary theory. 1.4 Reputation as an evolutionary strategy: indirect reciprocity The economic notion of reputation as an expectation about the future behaviour of an agent has been modelled, in the rational choice tradition, as a repeated game [Axelrod 1984]. Repeated games have been introduced in order to study cooperative strategies. As Axelrod has famously shown, the deceptive equilibrium of the Prisoner’s Dilemma game can be overcome. If the game is reiterated, people prefer to cooperate instead to defect because they are able to see possible gains in the future for having adopted a cooperative strategy. In a sense, by repeating the game, people earn a reputation of “nice” players that encourages the parties to reciprocate. Reputation has been considered in evolutionary theory as an important ingredient for solving the social dilemma of human cooperation. Not only people can overcome the Prisoner’s Dilemma, as mentioned above: they go beyond pure tit-for-tat strategies and show reciprocal altruism: individuals exchange favours [Trivers, 1971], but the delay of reciprocation may be long, so it is important to keep track of who did what to whom [Boissevain 1974]. The informal economy of reciprocal altruism is sustained by reputation [Nowack & Sigmund, 2005]. The social dilemma reputation is supposed to solve is the following: In a selfish-gene driven world, what can be the evolutionary advantage of complex biological systems such as human societies that base their survival on cooperative and altruistic interactions? Indirect reciprocity is easily captured by this motto: “I help you, someone helps me”. If I help you, I earn a reputation of being an altruistic person and people in the community will be willingly to help me in the future because they know that I will return the favour. The reputation that is valuable for indirect reciprocal exchanges is not “earned on the spot”, compared to the model I presented in 1.2.: it depends on social communication and on how gossip about my good behaviour circulates within the community. This “second-hand” information about our reputation distributed through communication and gossip is crucial for the stabilization of a shared social image on which future interactions and expectations can be based [Cavazza 2012]. In this perspective, reputation is not just an expectation: it is a socially distributed representation that monitors the social identity of individuals. This vision of reputation as a social representation allows me to introduce another concept of reputation in use in social sciences, that of social image.


1.5. Reputation and the “looking-glass self”: the social image

No analysis of reputation is richer than Erwin Goffman’s fine work on ethnomethodology and symbolic interactionism. His 1956 book on The Presentation of the Self in Everyday Life could be read, from the first lines, as a treatise on reputation: “When an individual enters the presence of others, they commonly seek to acquire information about him or to bring into play information about him already possessed. They will be interested in his general socio-economic status, his conception of self, his attitude towards them, his competence, his trustworthiness, etc.[…], many sources of information become accessible and many carriers (or 'sign -vehicles ’) become available for conveying this information.” [Goffman 1956: 11] Goffman’s interactionist approach to sociology was a way of describing - with a rare depictive mastery – how we construct our social self in public, how we manage to signal some qualities and hide some flaws of our personality and how our interlocutors use all these aspects of our public performance to get information about us and to interact in a strategic way. Having a reputation, earning or losing it, is the outcome of a social interaction in which our conduct modifies the distribution of opinions in the social environment. Goffman is both one of the fathers of the theory reputation as a signal, especially in his work on strategic interaction [Goffman 1969] and a contributor to the social psychology of reputation [Emler 1990], that is, of the study of how our social identity affects our choices and behaviours. One can describe his books as series of studies in “reputation management” in which he analyses in vivo the properties of the interactions that affect and are affected by our way of dealing with our social image. The Presentation of Self in Everyday Life lists all the possible “moves” (performance, front, dramatic realization, idealization, etc.) that are available to the subject in a social setting. Goffman sees these settings as theatrical scenes in which face-to-face interaction follows the rules of a dramatic performance. What is important for our purposes here is his idea is that our social self is not just a “parade”: its phenomenological richness is causally efficacious in determining the outcome of an interaction and what the others will end up knowing about us. Social psychologists have studied the role of social self in the development of our personality. At the beginning of the XXth century, the American sociologist Charles Horton Cooley introduced the concept of the looking-glass self [Cooley 1902] to describe the process through which a person’s self grows out of society’s interpersonal interactions and the perceptions of others. As he says: “I am not what I think I am and am not what you think I am. I am what I think that you think I am”. The way in which we interiorize the others’ gaze and use it to construct our own feeling of who we are is still lively subject of research in social psychology . The limits of this approach is that it mainly deals with the rich phenomenology of face-to-face interactions, but doesn’t say a lot about how reputations constructed in interaction last or decay, modify themselves in time and are used and misused for further cognitive purposes beyond the immediate needs of the ongoing social interactions. We will see in the following that reputation matters even when we are not face to face and even when it is used to evaluate items that are not individuals, such as objects, ideas, etc.

 1.6 Reputation and status

Consider this case. I am attending a conference. I am not an insider of the discipline, but it is a subject that could be relevant for my future intellectual projects and I am eager to learn. During the coffee break, I hang around observing people: I don’t know anybody and feel a bit uncomfortable about whom to approach for a little conversation. I try to read the badges they are wearing on their jackets, try to scrap their origins, where do they come from: Here is a Harvard Professor, encircled by a bunch of groupies who listen to his words as he were the Messiah. I remember I’ve heard his name somewhere. Maybe a paper that I have read on the subject, or maybe it came out when I started to make Google searches in order to learn a little more about the topic. Here is the classical French Professeur d’Universités, with his self-important posture, and with such a strong accent in English that nobody understands what he says. I recognize the usual habits I saw so many times in France: he has conquered the front position at the table of the desserts and won’t leave it until the end of the pause. I use my prejudices, my background information, the reputational cues I am able to infer from their academic status, their age and manners…But what informs me most is the attitude that they have towards each other: the think network of status signals that the social situation provides . Who are the established people? Who the marginal? Who are those everybody want to talk with and those to avoid at any cost? I don’t remember who said that the difference between an important person and a nobody can be measured in this way: if the first one tells a joke, everybody burst out laughing, if the second one tells the same joke, a flustered silence follows. I observe all these signals: at a certain point, the Harvard Professor, who is surrounded by a thick crowd, pronounces the name of one of his young scholars: the circle of fans turns the gaze towards the newborn star: his status leaks, as a magical elixir, from one to the other, and enlightens the young man with a new, brighter voice. All these cues are simultaneously present in my mind, as I explore the environment around me: chunks of rankings of institutions that I have registered in my mind over the years, old and new cultural prejudices that have resided in the back of my mind through my wanderings around the world, quick and dirty heuristics about the dynamics of social networks, conversational commitments and so on. What I am looking for is a readable social map of the environment, a weighted landscape that allows me to orient my first steps in this new domain of knowledge. At the end of the coffee break, as a result of these mental operations, the world around me has a new informational configuration. If anything will be fixed in my mind after the conference, it will be a new, better informed association between a bunch of new ideas and the configuration of epistemic authorities I was able to shape in my mind by using a set of very different strategies and cues. Among the most informative cues, the way in which the insiders look at each other, the invisible yet tangible social network they establish while interacting. Status signals are a reservoir of information about the hierarchical social configurations that define social life. They are a major subject of inquiry in social sciences such as anthropology and sociology. The way in which we perceived authority relations and are embedded in social hierarchies can be seen as the “quintessence” of social sciences. Anthropologists such as Louis Dumont [1971], and Maurice Godelier [1982] have argued that the social world is intrinsically structured in hierarchical relations that are more fundamental than institutions. That is, as Godelier has shown in his study on the construction of Big Men in New Guinea, hierarchical relations don’t come into existence with the introduction of social institutions. A perception of fundamental inequalities and hierarchies within a social group exists even in groups where no institutions (such as money, government, etc) exist. In his analysis of the system of casts in India, Dumont argues that the structural relations in society are fundamentally hierarchical. Karl Manhheim, the founder of the sociology of knowledge, speaks about our “hierarchical appraisal”, as if our way of experiencing the world were essentially hierarchical . Status hierarchies are thus seen in this perspective as omnipresent structural phenomenon of human but also animal societies. Social differentiation leads to the emergence of status hierarchies in animals [Bonabeau & alii, 1999; Chase & alii, 2002]. In sociology, this phenomenon is usually explained with reference to mechanisms of cumulative advantage such as the Mathew’s effect [Merton 1972; DiPrete & Eirich 2006] in which small qualitative differences between individuals get amplified via status-conferring collective procedures. More recently, analytical sociologists such as Gould [2002], Podolny [2005], Manzo and Baldassarri [forthcoming] have developed models to explain the emergence of status hierarchies by showing how reputational rankings greatly overstate quality differences between individuals and crystallize into categorizations whose effects are to further enhance these differences. That is, status differences in society neither depend only on quality differences among individuals nor on structural advantages or disadvantages. Rather, they depend on the effects of amplifications that self-reinforcing reputational rankings (in which people defer to each other’s status by following strategies of reciprocation and fairness) make emerge in interaction. These models are interesting for studying singularities in status attribution as for example, the extreme discrepancies in status recognition in certain professions (such as art, sport or science) [Menger 2002]. But, as the authors insist on, they see status as not the same thing as reputation. According to Podolny, status refers to a “diffused sense of better or worse that is indirectly tied to past behaviours, but it is more directly tied to the pattern of relations and affiliations in which the actor does and does not choose to engae” [Podolny, 2005: 13]. I will challenge his position and propose a definition of reputation that takes into account status relations. The notion of status as defined in this literature is too narrow for my purposes, because it considers only relations among actors (and not, for example, between actors and things or beliefs) and sees status as associated with an immediate perception of the social hierarchy, whereas in many contexts our status attribution depends on hierarchies that we do not directly perceive, but accept from others that we consider authorities in a domain. As a conclusion of this long review of interrelated notions, let me say that the proponents of these notions enter the battlefield of ideas by stating that “their” notion is different from all the others. Thus, Brennan and Pettit spend the first chapter of their book in distinguishing esteem from other notions, such as reputation. According to them, reputation implies the reidentifiability of the object, whereas esteem may accrue even to someone who is not reidentifiable (I mean, I can esteem someone “on the spot” for a certain action without whereas my positive attitude won’t play a role in establishing his or her social identity in the future). As I already said, Podolny too tries to distinguish his notion of status from that of reputation. But, apart from the canonical reasons that pertain to the market of ideas to distinguish notions in order to earn salience in a saturated cognitive environment such as academic research, I think that a general definition of reputation may encompass most of the relevant aspects of these ideas, given that strong family resemblances that I have tried to show among them. I think that the common denominators of these very different intellectual projects lies in the shared intuition that there exists a social dimension of our selves, motivations and judgements that is strongly relational and depends on the judgements of others in an interactive way, that is, neither on a fixed institutional “superstructure”, nor on a completely individualistic attitude, but in the relational constant comparison between ourselves and the others. Now, I would like to turn to the more constructive side of this contribution and propose a definition and some illustrations of reputation as a property that, from the observer’s point of view, can be an epistemic resource to extract information about not only the others, but also the world.

2. Reputation as an epistemic notion

Quality uncertainty and informational asymmetries have become crucial epistemological issues in contemporary information-dense societies. The vast amount of information available on Internet and on the media makes the problem of reliability and credibility of information a central issue in the management of knowledge. Informational items that do not come with some label, or seal of approval from the appropriate communities, are lost in the data deluge of the Information Age. From the evaluator’s perspective, that is, the agent who has to filter information, reputation has an informational value. But what is reputation in this informational sense and how does it relate to the plethora of definitions proposed above? Reputation is a relation between an agent or an item and the set of social and cultural representations of the agent or the item. Cultural representations are distinguished from social representations in this way: cultural representations are metarepresentational , that is, they are socially distributed representations of representations: I judge your reputation through a social representation that is not only generated by our ongoing interaction, but also by culturally stabilized representations that have previously been attached to you, that have circulated about you and that are held by people I trust or defer to. Having a reputation is thus being attached to an evaluative representation of ourselves that is socially and culturally generated and stabilized. A reputation is the shortcut of the many strategies, heuristics, evaluations that have been placed a person or an item in a certain hierarchical configuration. Given that in my definition of reputation I want to include items, having a reputation doesn’t imply being aware of having it, a condition that seems to be implied by many definitions that I have listed above. Also, reputation is not only a cognitive notion (as in the rational choice approach, which is formalized as an “expectation”): it is a social and cultural notion that has to do with the way in which the social environment is organized: it spreads through networks, cumulates through ratings and manifests various effects that are independent of our cognition.

 2.1. Is reputation a Cambridge property? The paradox of Violetta Valéry

The philosopher Peter Geach [1969] coined the term Cambridge change, to distinguish extrinsic from intrinsic changes of an entity: if at time0 I am 175 cm tall and at time1 I am 180 tall, I have changed a property of myself. If at time0 I was taller than Leo, my son, but at time1 I am shorter because he grew taller, nothing really changed in me. Given that reputation is a relational/social property, it may seem to share some features of the Cambridge properties. I can earn or lose a reputation without any change in my essence. Let me illustrate this situation by an example. In the Verdi’s opera, La Traviata the famed courtesan Violetta Valéry is the beloved mistress of the young nobleman Alfredo Germont. Their feelings are so deep that she decides to completely abandon her previous lavish life and go to live with him in the countryside. One day, while Alfredo is out, she receives the visit of Alfredo’s father, Giorgio Germont, who explains to her - in the famous aria Pura siccome un angelo - that the reputation of the family is ruined because of Alfredo’s relationship with her, and that Alfredo’s sister won’t be able to marry the noblemen to whom she is already engaged because her reputation is lowered by this relationship. Violetta finds the argument perfectly compelling and accepts to leave Alfredo. Why so? What is so compelling about a leaking of reputation from her relationship to the marriage of a person she doesn’t even know? What have changed in the life of Alfredo’s sister when she acquired a property of being the sister of Violetta Valéry’s lover? Not so much: she is exactly the same person she used to be. Yet, the leaking of reputation is not inert, like in the case of Cambridge properties: it hurts her life and her feelings and causes the breaking of her engagement. How does reputation leak from an entity to another and how effecting the leaking is? It seems obvious from the example that, although a change in reputation may seem a Cambridge change, even if nothing happens in the person that is “victim” of the change in the reputational configuration, the causal consequences are real. That is because reputation leaks through a social network of deference relations. A good person who judges you good makes you earn status. A bad person who judges you good makes you loose status. Reputation is thus not a Cambridge property. Even when we cannot always control how it leaks from an entity to another, the leaking has causal consequences. It is important to notice that reputation leaks only if the items are linked each other in a network. My reputation is untouched by the physical proximity of a lower-status stranger in the underground. But my reputation as a high-status person may be touched by joining the social network of people who take the underground. A Picasso painting exposed in the same room in a Museum with my painting greatly increases my reputation as an established artist: his reputation leaks towards me. But if the two paintings happened to be close together because a burglar stole in a rich friend’s apartment my painting that I offered her for her birthday and her Picasso, then I won’t acquire any reputation from my proximity with Picasso. Sometimes we can misinterpret physical proximity with linkage in a network. It is a common practice in many milieux to ask for a picture together with a “higher-status” person. I was invited once at a conference on the Web and Tim Berners-Lee, its inventor, was there. I immediately asked him permission to take a picture together and shared the picture on Facebook with impressive (but, of course, bogus) reputational effects. Although we were both invited to the same conference and thus, in a sense, we shared a social network, our ties are very weak and the fact that we were together on that picture is more a happy coincidence of physical proximity than a reliable trace of a strong tie between us.

2.2. Reputation and Classification

As I said, reputation serves the cognitive purpose of making us navigate among things and people whose value is opaque for us because we do not know enough about them. We use seals, scales, grades, indexes, classifications not only to evaluate them, but to create valuable categories that allow us to classify reality. Take this example. What is that scarlet piece of tissue in the shape of an A sewn on Hester Prynne's gown in Nathaniel Hawthorne's masterpiece The Scarlet Letter? Is it a symbol of her sin, a "badge of shame", an indelible sign of her community's contempt? Is it a cruel reminder of her past, a succinct history of her misdeeds? Imagine that in the same colonial New England village, you do not have just a badge for the poor Hester, but each member of the community wears a letter that represents some past records of its owner. We can also imagine sets of identical badges worn by members of the community who have similar records: sinners, heroes, drunkards, etc. Imagine that the elders of the community have the right to attach these labels to the villagers. Their judgments, based on their purported wisdom, become an easy way for the villagers to dispose of a basic classification of social types within the community that will allow them to manage their relations with others, to make inferences and predictions about their behaviour, that is, construct a basic "social map" that will help them orient in their society. Morally this may be questionable, but epistemologically it can be useful. The A sewn on Hesther’s gown is a seal of disapproval that constitutes a form of classification. In many contexts, we do not classify entities and then evaluate them: the very act of classifying them is to order them according to the reputational rankings that are already available in our culture. Let’s start with the timeline of classification and evaluation. When we think about categorization, it is a common view to see taxonomies as cognitive tools that describe objective relations and properties between classes and objects, that is, they provide an ontological structure for a domain. Evaluative tools come afterwards in order to impose a ranking on these items. Contrasting this classical picture of knowledge organization, I claim that sometimes reputation is prior to classification: in many epistemic practices throughout, we use rating systems in order to classify items. A second important aspect of reputational and rating systems is that they combine two types of information for the sake of knowledge organization and evaluation: information about content and social information, information about people and past interactions. One of the best examples of the interplay between reputation and classification is the system of classement of Bordeaux wines [Origgi 2007]. Bordeaux wines are classified according to a variety of local ranking systems whose best known is that of châteaux that was established in 1855 in response to Napoleon’s III request to rank Médoc wines for the Parisian Exposition Universelle. The ranking was established by wine industry brokers according to the past reputation of the châteaux, calculated in terms of their trading prices in the past 100 years. Thus, in order to establish a classification that has created stable cultural objects such as the Grand Crus, the Premier Crus, etc. reputation was needed. In a word, reputation is not only a way of extracting social information, but also a way of constructing it through the classifications it makes possible. The importance of being included in a classification is that it assures to the item that has been included a long-lasting reputation due to the stabilization of the classification system in a culture. This opens the question of the relation between reputation, effective value and time.

2.3. Reputation and Time

A remarkable feature of reputation is its relation with time. We all know the disappointing experience of having trusted the reputation of a famous touristic spot and, once there, wondering what are we doing in such an awful place. Reputation is more resilient to time changes than the effective qualities it is supposed to represent. Ancient renown of aristocratic names can still emit a “halo” even in a world in which being an aristocrat doesn’t correspond to any effective social privileges. Institutions may benefit of a “position advantage” in a ranking system even if their prestige is much more transient than we think. I remember how people in Parisian cinemas smiled at the Woody Allen movie Midnight in Paris, when Inez, Gil’s nervous fiancée, tries to provoke Gil’s jealousy by praising the qualities of Paul, one of her ex-boyfriends, saying that he must be a genius given that he was invited as visiting professor at the university La Sorbonne. The reputation of the Sorbonne, one of the most ancient universities in the world, is related to its glorious past, and an academic outsider as Woody Allen can take this past reputation seriously even if today, if you look at academic rankings of European universities, it is clear that the Sorbonne doesn’t deserve anymore the reputation its name still has. Time is sometimes undoubtedly an advantage in the accumulation of reputation. In the case of academic reputation, that is calculated in terms of impact, that is, citations of one’s own work in the works of others, time is an important factor for earning reputation: citations take time to accumulate, and older papers have more reputation than new ones (up to a “tipping point” at which either they have earned some impact and thus get the more and more chances to accumulate reputation in the future, or they haven’t reached that threshold and disappear within the legions of the forgotten. But time creates reputational lapses that often make us rely on socially established reputational cues that are no more connected with the quality of the person or the item to which these cues are attached. The way in which different reputational devices update their evaluation is multifarious: there are evaluations that are constantly updated, such as guides of the best universities, the best wines or the best restaurants, all mainly controlled by acknowledged experts and evaluations that are based on informal social networks, or culturally embedded local values, that are not revised systematically and can thus display notorious time lapses that make us rely on cues that don’t correlate anymore to any quality. This is typical in domains in which there are no explicit shared rankings, such as for example, when we choose a doctor or a lawyer [Karpik, 1984]. Doctors’ reputation is notoriously resilient because it is based on informal social networks of acquaintances that is not regularly updated and doesn’t have clear criteria of ranking. Have you ever noticed that most friends who suggest you a doctor say that he or she is the “Best specialist in town?”. People defer to social networks in constructing these reputations, but of course networks contain many biases like locality, social class, etc, that make the information circulate in an imperfect way. Also, you don’t ask suggestions for a dentist everyday and people do not update their medical carnet d’adresses very often, so I can end up suggesting you a dentist who was really good 10 years ago even if today she has become lazy, unreliable and doesn’t deserve anymore her reputation. Same thing for holidays. How many times the person who suggests you that wonderful little hotel on the Amalfi coast was at the hotel? Maybe once, twenty years ago. She recommends you that place while today it may have become a too touristic location, with cars, noise and too many people around. Traditions, that is, shared evaluations that are spread in a cultural group and have a special authority for that group, are often based on past renown and on rare encounters: they are difficult to challenge because our deference to them is part of our cultural identity. Yet, they can be biased and make some reputations last far beyond any actual opinion. Reputations and rankings, being, as I said, cultural artefacts, that is, shared deferential attitudes about distilled opinions on a person or an item, can be institutionally enforced (like for example in education, or through public representations such as monuments, or else through act of public recognition such as prizes, medals). That is why they may last long time even if our immediate perceptions of status relations contradict them. Sometimes we may end up in a situation of “preference falsification” [Kuran 1995] in which everybody rely on a shared reputation and endorses it (a public judgement of the sort: “Hegel is the most important philosopher of all times”) while privately organizing his or her preferences in another way.

  2.4. Reputation, epistemic authority and interpretation

In a famous novel written by Jerry Kosinski, Being There - upon which a movie with Peter Sellers was based- Chauncey Gardiner is a mentally retarded gardener who, through a series of fortuitous accidents, becomes a heir of the throne of a Wall Street tycoon, a presidential policy adviser and a media icon. His simple words on how to conduct a garden have great impact on the President of United States, whom Chauncey meets in the house of Mr. Rand, the Wall Street Tycoon he happens to live with. In a conversation with Rand and Chauncey, the President asks: “And you, Mr. Gardiner? What do you think about the bad season on the Street?” Chauncey replies: “In a garden, growth has its season. There a spring and summer, but there are also fall and winter. And then, spring and summer again. As long as the roots are not severed, all is well and all will be well”. “I must admit”, the President says “that what you’ve just said is one of the most refreshing and optimistic statements I’ve heard in very, very long time […] Many of us forget that nature and society are one! Yes, though we have tried to cut ourselves off from nature, we are still part of it. Like nature, our economic system remains, in the long run, stable and rational, and that’s why we must not fear to be at its mercy […] We welcome the inevitable seasons of nature and yet we are upset by the seasons of our economy! How foolish of us!” It may seem just one among the millions of examples of misunderstanding, but let us have a closer look of what is happening here from our reputational perspective. The simple fact that Chauncey is in Mr. Rand’s house gives the President enough background information about his reputation as a source of knowledge. The President’s epistemic assessment of Chauncey’s words is based on his reputation, established by the social network they are sharing: a friend of friend who is a reliable person - Mr. Rand- is also reliable. This background assumption, together with the standard processes of interpretation, explains the President’s extra-effort in order to work out a relevant interpretation of what Chauncey said. We may also add a “confirmation bias” in the President’s interpretation, that is, he tends to reconstruct what Chauncey said in a way that confirms what he already thinks and hopes of the economic conjuncture in United States. Of course, he doesn’t simply accept Chauncey’s words under authority: he interprets them in the context of his own thoughts about US economy. What Chauncey said is just further “food” for his thoughts and helps him in coming out with a positive interpretation of the bad season in Wall Street, a message he will repeat immediately after the dialogue on television. Reputation is a powerful tool of information extraction in conversation. It serves interpretive purposes. We use our interlocutors’ reputations as a sort of charity principle [Davidson, 1984] We allocate much more cognitive effort in trying to make sense of what someone says if we value positively his or her reputation. I remember when I was a student struggling so many times in front of philosophy books that seemed to me pure nonsense and trying to make sense of the content just because I was relying on authorities who told me that that philosopher was very well reputed. I have seen lives of intellectuals change abruptly because of their inclusion in a public ranking, like “The best 50 thinkers of the late 20 years”, and their previously ignored theories becoming a subject of exegesis and collective interpretation. Reputation thus plays a role in linguistic interpretation and in the way we manage to make sense of what other people say. Given that what others say informs us on reality, this confirms my hypothesis that reputation plays a role in extracting information about the world. Of course, here also, there are many biases, like the notorious guru effect [Sperber, 2007] according to which people may go on endlessly looking for an interpretation of a nonsense if the source of the nonsense is an authority whose obscure language is perceived as a sign of depth and wisdom.

2.5. The epistemic critique of aesthetic judgement

While travelling in Brazil, where I was giving some lectures at the University of Porto Alegre, I had the chance to have interesting discussions and exchanges with local art critics, wine critics and art collectors. Although I like a lot contemporary art, I am not an insider, and definitely not a connoisseur of contemporary Brazilian art. An art dealer brought me to the MERCOSUR exhibition, an enormous contemporary art fair that gathers every two years artists from overall Brazil. Apart from some traditions that were clearly imported and thus easy to evaluate for me, some of the artists were impossible to classify, to rank in an aesthetic space familiar to me. In order to help me to make sense of my aesthetic experience, my friend art collector tried to provide me with information about important prizes artists had earned, international exhibitions (“This one was exposed in Venice at the Biennale last year…”) and prices of the works in last prestigious auctions. That was a way to start a conversation, to be able to compare our experiences and coordinate our judgements. But for most artists it was clearly not enough for me to have also an associated aesthetic experience. The way in which most artists were addressing the audience was too “local” for me, as if they were part of a too distant conversation that I was not able to join because of my cultural distance. At dinner after the visit, my friend art collector showed me a catalogue of an artist called Adriana Varejao, one of the top-selling contemporary Brazilian painters, whose last auction prices at Christie’s and Sotheby’s are higher than half a million dollars. In a pop-art tradition that reminds Roy Lichtenstein, she paints giant azulejos, that is, the typical white and blue Portuguese tiles that you find in churches and old buildings in Brazil. While looking at her works, I wondered whether her style was so much easier to decode for me than the previous Mercosur artists. It was as if she was referring to a conversation and commenting on traditions familiar to me also: her interlocutors are mine: they are the global artistic authorities of contemporary art, from Warhol, to Lichtenstein or Richter. When we look for indirect cues in order to express an evaluative judgement on a work of art (but also on the taste of a wine and other aesthetic experiences) one of the most informative cue is the reputation of the imaginary interlocutors of the artist. The artist alludes to other artists, comments upon them, judges them thus giving you access to a learned conversation that you are willing to be part of. The aesthetic experience is based on the discovery of sharing with the artist these rankings and authorities and thus being part of the same intellectual world. Contemporary art is not only a question of reputation in the eyes of the observer, as it is frequently said (“I like it because I know it has a good reputation”). The way in which reputation is managed in order to convey an aesthetic experience has to do with the reputations of the artist’s interlocutors, as if he or she were inviting us to share an imaginary circle of authorities whose prestige matters for us. Art works are special conversations in which the artists enacts many past and present voices and puts him or herself in a circle of recognition that we are inclined to share. Here again, the role of reputation is a much more complex process than it is generally acknowledged. It is a way of extracting the relevant information that provides us with the only possible rewarding aesthetic experience: that of being part of the same world the artist is winking at.

  2.6. The Twitter Mind: Social Networks and Collective Intelligence

Collaborative forms of sharing ratings are also relevant in the study of Collective Intelligence. I have explored the relation between reputation and the explosion of the Social Web elsewhere [Origgi 2012], but let me briefly conclude with some reflexions on our use of Web-based social networks to extract information. My point is that the success of these networks is because they are built in a way that facilitates the expression of some of our most entrenched cognitive and cultural heuristics to extract information, like “Follow the right leaders”. While Facebook is a social network aimed mainly at enhancing our looking-glass self (see above p. ??) and encouraging the management of our own reputation, Twitter is pure information extractor based on a simple status hierarchy “Who follows whom”. Choosing valuable informants, that is, being able to exploit social information in the most efficient way, may give striking cognitive advantages. That is true not in the high-tech world of contemporary social networks: the history of human cultural evolution [Tomasello 1999] seems to rely on similar strategies of following the appropriate leaders. Twitter is a reputational device that allows collective intelligent outcomes thanks to its hierarchical social networking. A more democratic network won’t have the same epistemic effects. The analysis of the pernicious effects that may bias the way in which information circulates through these highly hierarchical social networks is at its beginnings. As the study of the pernicious effects of “one to many” forms of influence on decision, like the influence of opinion leaders, is a well-established field of inquiry [Boland et al. 1989; Estlund 2007], the way in which reputation affects judgements in these “many to many” hierarchical networks and, in general, on the Web is still an open question. Conclusion People do not share information: they share evaluated and classified information that creates a “reputational stream” of shared judgements. The epistemological implications of the massive use of shared ratings in networked societies are huge: relying on other people’s judgements and authority challenges our epistemic responsibility. The reasons we trust collectively filtered ratings about an item or an agent are seldom explored. Choosing a doctor, an academic institution or a wine is a way of endorsing a tradition of values, a way of filtering information that is not always transparent and legitimate. Notorious biases in social networks - such as the Matthew effect, investigated by the sociologist of knowledge Robert Merton, according to which the nodes of a network that are more prominent have more probabilities to earn more reputation - create noise in the way reputation is diffused. Other biases need further epistemological and cognitive inquiry. As I have tried to sketch, people tend to form beliefs in order to acknowledge previously established reputations, such as voting for a certain party because a very well-reputed friend votes for that party. Also, as we have seen, reputations are resilient and may last over time even when the facts of the matters they are supposed to signal are no more there. For example, the prestige of institutions and corporations may last long time after their decay. I have defended the idea that reputation is an essential epistemological notion. Reputation is a criterion of information extraction, a fundamental shortcut for cumulating knowledge that is embedded in social networks and an ineludible filter to access facts. In an environment where sources are in constant competition to get attention and the option of the direct verification of the information is simply not available at reasonable costs, evaluation and rankings are epistemic tools and cognitive practices that provide an inevitable shortcut to information. This is especially striking in contemporary informationally-overloaded societies, but I think it is a permanent feature of any extraction of information from a corpus of knowledge. There is no ideal knowledge that we can adjudicate without the access to previous evaluations and adjudications of others. No Robinson Crusoe’s minds that investigate and manipulate the world in a perfect solitude. The higher is the uncertainty on the content of information, the stronger is the weight of the opinions of others in order to establish the quality of this content. The question of the epistemic value of these rankings, that it, to what extent their production and use by a group changes the ratio between truths and falsities produced by that group and, individually, how an awareness of rankings should affect a person’s beliefs is thus crucial. Rankings may seem to introduce a bias in judgement, that is, a precondition in the way information is categorized and classified, and the epistemic superiority of a biased judgement is in need of justification. There are some biased judgements whose value is superior because they participate in sustaining a tradition of knowledge that stabilizes positive effects in a culture. There are other biased judgements that sustain prejudicial information and should be wiped out. But, if I am right and there is no information that is completely independent of biased judgements, the aim of a social epistemology of reputation, as the one I am advocating here, should be to pry apart the biases worth keeping from those worth eliminating.

Bibliography

Akelrof, G. (1970) “The Market for Lemons”, Quarterly Journal of Economics, 84 (3), 488-500. Axelrod, R. (1984) The Evolution of Cooperation, Basic Books, New York. Boland, P.J. et al. (1989) “Modelling Dependence in Simple and Indirect Majority Systems” in Journal of Applied Probability, 26, pp. 81-88. Bourdieu, P. (1984) Distinction. A Social Critique of the Judgement of Taste. Routledge and Kegan Paul. Boissevain, J. (1974) Friends of Friends: Networks, Manipulations, Coalitions, Oxford University Press. Brennan, G., Pettit, P. (2004) The Economy of Esteem, Oxford University Press. Cavazza, N. (2012) Pettegolezzi e Reputazione, Il Mulino, Bologna, Italy. Cooley, C.H. (1902) Human Nature and the Social Order, Charles Scribner, New York. Craik, K. H. (2009) Reputation. A Network Interpretation, Oxford University Press. Davidson, D. (1984) “Radical Interpretation”, in Davidson, D. Inquiries into Truth and Interpretation. Oxford: Clarendon Press. 125-140. Elias, N., Scotson, J. (1965) The Established and the Outsiders, New York: Frank Cass & Co. Emler, N. (1990) “A Social Psychology of Reputation” European Review of Social Psychology, vol. 1, pp. 171-193. Estlund, D.J. (2007) Democratic Authority, Princeton University Press. Faulkner, P. (2011) Knowledge on Trust, Oxford University Press. Fricker, E. (2007) Epistemic Injustice, Oxford University Press. Gambetta, D. (2009). Signalling. In: P. Hedström, P. Bearman (eds) Oxford Handbook of Analytical Sociology, ch. 8. Goffman, E. (1959) The Presentation of Self in Everyday Life, Doubleday, New York. Goffman, E. (1970) Strategic Interaction, Basil Blackwell. Gould, R. (2001) “The Origins of Status Hierarchies: A Formal Theory and Empirical Test” American Journal of Sociology, 107, pp. 1143-1178. Granovetter, M. (1973) “The strength of weak ties” American Journal of Sociology, 78, pp. 1360-1380. Honneth, A. (1995) The Struggle for Recognition, Polity Press. Karpik, L. (2007) L’économie des singularités, Gallimard, Paris. English translation (2010) Valuing the Unique, Princeton University Press. Krebs, J. R., and Davies, N. B. (1998). An Introduction to Behavioural Ecology. Blackwell, Oxford. Kuran, T. (1995) Private Truths and Public Lies, Harvard University Press. Manzo, G., Baldassarri, D. (forthcoming) “Sirens and Grapes in the Emergence of Status Hierarchies”, manuscript presented at the meeting Social Mechanisms at Work, Paris, September 23-24 2011. Nowack, M. Sigmund, K. (2005) “Evolution of Indirect Reciprocity” Nature, 437, pp. 1291-1298. Origgi, G. (2004) “Is Trust an Epistemological Notion?” Episteme, vol.1, n. 1, pp. 61-72. Origgi, G. (2007) “Wine Epistemology: The role of Reputation and Ranking Systems in the World of Wine” in B. Smith (ed?) Questions of Taste, Oxford University Press, pp. 183-198. Origgi, G. (2012) “Designing Wisdom Through the Web. Reputation and the Passion of Ranking” In H. Landermore, J. Elster (eds) Collective Wisdom. Cambridge University Press, pp. 38-56. Podolny, J.M. (2005) Status Signals, Princeton University Press. Shapin, S. 1994: A Social History of Truth: Civility and Science in Seventeenth-Century England, Chicago University of Chicago Press. Smith, A. (1982) The Theory of Moral Sentiments, Indianapolis, Liberty Classics. Sperber, D. (1996) Explaining Culture: A Naturalistic Approach, Oxford: Blackwell. Sperber, D. (2000) Metarepresentations, Oxford University Press. Sperber, D. Origgi, G. et al (2010) “Epistemic Vigilance” Mind and Language, 25, n. 4, pp. 369-393. Trivers, R. (1971) “The Evolution of Reciprocal Altruism” Quarterly Review of Biology, vol. 46. n. 1, pp. 35-57. Veblen, T. (1899/2007). The Theory of the Leisure Class. Oxford World’s Classics.