It is true that in AG the players share a first preference for a)bp. But, as indeed Sen says, they do both need assurance, and the prospect of the enforcement of a contract might be how the assurance was secured. This points up an important difference in the possible role of sanctions, a difference relevant to political theory. The addition of sanctions may change the utilities for a basically egoistic agent, or they may provide assurance for an agent who is basically cooperative. The latter, a very important possibility, is connected to a point I shall consider later: that the disposition to cooperation is typically cost sensitive.

XREF: This contrasts two distinct roles of sanctions in game theory and political theory — deterrence for egoists versus assurance for cooperators — which connects to Hobbesian vs. Rousseauian accounts of the social contract. DEFINE: Distinguishes between the utility-changing role of sanctions for egoistic agents and the assurance-providing role for cooperative agents, a key conceptual distinction in the author's framework.

Each needs to know that the other has the AG preference schedule. Moreover, each needs to know that the other knows that he (the first) has that preference schedule. If that is not so, each can conduct, and can replicate, reasoning that will lead him to believe that there is a risk that the other will defect, and hence he will defect himself, … Without this knowledge, or something that will stand in for it, the cooperation will unravel.

XREF: This is the classic game theory concept of common knowledge (Lewis/Schelling), closely tied to the iterated prisoner's dilemma and coordination games.

The baroque structure of these intrigues, in which an agent might act to create the impression that a second agent was acting so as to create the impression that the first agent was engaged in a conspiracy (for instance, that one), must surely have resulted in a situation in which no one really understood what anyone was doing; and this lack of understanding will also have affected the situation to be understood. Since such a situation is structured by the agents' intentions, and their intentions are conditioned by their increasingly confused grasp of others' intentions, there comes a point at which there is no truth of the matter about what they are doing.

DEFINE: Introduces the idea that action-truth depends on the coherence of agents' mutual intentions, so recursive confusion dissolves objective answers.

Besides that possibility, however, there are motives to cooperation that are not egoistic, but are more restricted than the kind of thing represented by the simplest psychological generalization of the preference schedule OR. There are general dispositions to cooperate which are conditional, and in particular are sensitive to costs: an agent will cooperate if it does not cost too much, but beyond this threshold he will not. The threshold may itself change over time. Moreover, failing perfect knowledge of the other agents' dispositions, he may take a risk on a cooperative initiative if the cost of being deceived falls below a certain threshold. In these various … (R3) A disposition to cooperate is cost sensitive.

DEFINE: Introduces the R3 disposition: cooperation is conditional and cost-sensitive, falling off beyond a threshold.

Two agents cooperate when they engage in a joint venture for the outcome of which the actions of each are necessary, and where a necessary action by at least one of them is not under the immediate control of the other.

DEFINE: Provides a precise, testable definition of cooperation rather than a vague colloquial sense.

Cooperation is a symmetrical relation: if X cooperates with Y, Y cooperates with X (if that seems unnatural, read 'cooperates' as 'engages in a cooperative venture with'). Depends on, however, is a nonsymmetrical relation: if X depends on Y — in the sense that his getting what he wants out of the venture? depends on Y doing his part — it may or may not be the case that Y depends on X.

DEFINE: Distinguishes the logical symmetry of cooperation from the nonsymmetrical nature of dependence, a key disambiguation for the author's argument.

One thing is immediately obvious: there will not be any cooperation if the only general motives there are for joining cooperative ventures are such that no one can (knowingly) be the dependent party.

Cooperation requires trust in the sense that dependent parties need some degree of assurance that non-dependent parties will not defect: as I have already said, this need not take the form of 'thick' trust, individually based belief, but it must take some form or other. The next step after this is to claim that, in general, people will not trust others enough to bring about cooperation unless their assurance is to some extent well based: that is to say, unless people are also in general motivated, one way or another, not to defect if they are in a nondependent position.

The motivation to cooperation is egoistic, since it lies essentially in the fear of the sanctions of the sovereign. However, the motivation to cooperation is also a macromotivation, because what the sovereign's sanctions are designed to bring about (among other things) is a motivation, in general, to cooperate. It is not an accident that, from egoistic considerations, the agent on these various occasions cooperated: he knew that what he had egoistic reason to do was, precisely, to cooperate.

DEFINE: Clarifies the distinction between egoistic (sanction-fearing) and macromotivation (system-designed) cooperation motives.

An egoistic micro-motivation, lastly, is fairly self-explanatory: I have such a motivation to cooperate when, in those circumstances, I see it as in my egoistic interest to do so, but this is not because I am generally motivated, as in the Hobbesian system, to cooperate as such. This does not rule out the possibility that I may regularly find myself in such situations, or even that there may be some structure as a result of which I regularly find myself in such situations: to produce such a structure might be a triumph of the Invisible Hand. It will be a mark of this kind of motivation, however, in a pure form, that there will be no motivational momentum from one case of cooperation to the next.

DEFINE: Distinguishes egoistic micro-motivation from Hobbesian generalized cooperation as a technical concept.

Without some general structure, egoistic micro-motivation merely returns us to the Prisoner's Dilemma. So if this kind of motivation is to do all the work, we need an Invisible Hand structure that will bring it about that a tendency to defect is offset by some other egoistic micromotivation (and will also allow this to be adequately known, etc.) But one needs extraordinarily strong and implausible functionalist assumptions to believe that such structures could be self-regulating and selfpreserving, and there is abundant evidence to hand, generated by a body of relentless social experiment, to show that they are not.

XREF: Ties to Prisoner's Dilemma theory and critiques of free-market self-regulation that assume rational self-interest produces collective good.

This will presumably have to be something like 'thick trust' - reliance on and between a group of persons in terms of their known dispositions, personal relations, and so on. It is not easy to see how this could be extended beyond a very small society. The only apparent way of taking it beyond that is to identify a particular set of people, a government or elite group, in whom thick trust is invested, and whose activities, example, or regulation serve to sustain cooperative activities throughout the society; but if this is conceivable at all, I find it impossible to see how such a structure could be the basis of our arrangements without an arrant degree of traditionalist or charismatic mystification.

In fact, that there is no one problem of cooperation: the problem is always how a given set of people are to cooperate. In a modern state, people are motivated by some version or degree of each of the sorts of motivations that have been distinguished, and their actions are often overdetermined, quite often by all four. The right question to ask, at this level of generality, is not which of these simple models could be made to do all the work, but rather what combinations of such motivations make sense.

The answers to such questions are important because so many non-biologists believe that the social behaviour of animals consists of nothing but strife and manipulation. If animal nature is the result of the competitive process of Darwinian evolution, then surely it must also be ruthlessly competitive. Such a conclusion does not follow from an acceptance of Darwinism. It is also a travesty of what biologists have observed and what most now believe.

In such discussions the meaning of 'interest' is not in much doubt. I shall amplify this point since it is so easy to conflate the consequences of a dynamic process with the way in which those consequences are generated.

Male polygynous primates that fight with other males for females have much larger canines than male primates that are characteristically monogamous (Harvey, Kavanagh, and Clutton-Brock 1978). Character-environment correlations of this type are known as adaptations. They catch our attention because they seem so well designed for the job they perform.

DEFINE: This highlight introduces the concept of 'adaptations' — correlations between an organism's traits and its environment that appear designed for a function.

Much the most coherent explanation for the evolution of such phenomena is still Darwin's. Indeed, Darwin's proposal is much better seen as a theory about the origin of adaptations than as a theory about the origin of species.

DEFINE: Recasts the standard understanding of Darwinian theory, distinguishing the origin of adaptations from the origin of species as the theory's primary explanatory target.

By analogy, Darwin referred to the differential survival of the characters that adapt an organism to its environment as natural selection. The Darwinian formulation emphasizes that the precise way in which a character enables an individual to survive or breed better is part of the process. The way in which the character works is not the same as the long-term changes in the frequencies of genes that are required for the expression of the character.

DEFINE: Distinguishes Darwin's functional conception of natural selection from the population-genetic view of evolutionary change.

The idea of 'kin selection' is an extension of the intuitively obvious point that animals will often put themselves at risk and do things that are bad for their health in the production and care of offspring.

DEFINE: Clarifies the concept of kin selection by grounding it in the intuitive behavior of parental care.

Even when both parents care for young, their interests do not coincide. They certainly have a common interest in their offspring’s survival, but they have diverging interests in as much as each one might be able to increase its reproductive success by spending time seeking extra mates elsewhere.

Given that each individual can affect its reproductive success in more than one way, if it reduces the effort devoted to one form of reproduction it must increase the effort devoted to the other by a specified amount in order not to reduce its overall reproductive success. Withdrawing care of offspring must be matched by stepping up the search for an additional mate.

Therefore, animals capable of taking note of conditions should evolve so that they change the amount of time they allocate to care of the young in response to changes in those conditions. The model is simple, but the expectation is matched by many observations of birds increasing the parental care devoted to their offspring when their mate deserts or disappears.

XREF: This is classic parental investment theory, connected to Trivers and Clutton-Brock. The observation about mate desertion inducing increased care is a textbook finding in behavioral ecology.

Not only do such grasshoppers match their black environment, they prefer to settle on such backgrounds. It is obvious that grasshoppers with a conditional response to the environment in which they are reared are more likely to survive than those that are inflexible. They would also do better than those that go black on a probabilistic basis, matching the likelihood of becoming black to the probability of savannah fires. Clearly, many grasshoppers that go black probabilistically would do so at the wrong time and be highly conspicuous to predators as a result.

No animal behaves in the same way irrespective of conditions. Nor, of course, do humans. Departures from the norm are not necessarily pathologies generated by abnormal conditions. They may well be highly adaptive responses to particular ecological conditions.

Or if members of a group are not familiar with each other, no mutual aid occurs until they have been together for some time. As familiarity grows, individuals come to sense the reliability of each other. Furthermore, the expectation of an indefinite number of future meetings means that deception is a much less attractive option. The major message is this: in cooperative behaviour, conditions matter a lot.

XREF: Connects to game theory and the iterated prisoner's dilemma, where repeated play makes defection less attractive.

Firstly, evolutionary process must be distinguished from its consequences; differential survival in the past does not necessarily mean social competition now. Secondly, in biology the outcome of differential survival is the ‘interest’, and no motive is required for that interest to be maximized by a given course of action.

DEFINE: Clarifies the technical distinction between evolutionary process and its current consequences, and between biological 'interest' and conscious motive.

Central to these subjects' experience of rule breaching" is the role played by another's reaction to one's own action, in confirming the validity of the prior experience on which that action was based. As many authors have observed, without such confirmation the social world

DEFINE: Clarifies a mechanism in social rule-breaching: external reactions confirm the foundational experience that motivated the action.

To take an extreme example, Seligman, Maier, and Solomon (1971) report that dogs subjected to random and unpredictable shocks of a relatively low intensity suffer much more than those who are receiving higher but predictable charges. Typically, the first group of dogs becomes listless, their physical condition worsens, and their normal interests diminish when compared with the group receiving the greater but predictable punishment.

XREF: Relates to learned helplessness research and the broader psychology of perceived control. Also connects to why predictable pain is psychologically preferable to unpredictable mild stress.

In general, they have found that those who are more willing to trust other people are likely to be equally trustworthy in that they are less likely to lie, cheat, or steal. They are also less likely to be unhappy or maladjusted, and are typically more liked by their friends and colleagues.

QUESTION: Does this correlation hold across cultures, or is trust willingness culturally conditioned?

The first substantive definition of 'trust' in the Oxford English Dictionary is 'confidence in or reliance on some quality or attribute of a person or thing, or the truth of a statement'. A later variant stresses the economic usage, defining it as 'confidence in the ability and intention of a buyer to pay at a future time for goods supplied without present payment'. Under either of these definitions, or the others offered in this volume, trust is based on an individual's theory? as to how another person will perform on some future occasion, as a function of that target person's current and previous claims, either implicit or explicit, as to how they will behave.

DEFINE: Provides the Oxford English Dictionary definitions of trust alongside a synthetic definition involving prediction of future behavior from prior claims.

Pruitt and Kimmel (1976) examined the extensive literature on the Prisoner's Dilemma game, and concluded that one important consideration for subjects is whether or not they view the interchange from a short-term or long-term perspective. In particular, if subjects believed that they would need to interact with each other after the study was concluded, and therefore the usual social and temporal isolation of such interactions was removed, their behaviour became considerably more cooperative.

XREF: Connects to Axelrod's work on the iterated Prisoner's Dilemma and the shadow of the future — repeated interactions sustain cooperation even among rational actors. Builds on: "Cooperation among strangers requires time and repeated future interaction."

The finding of interest from this work for the argument here is that the most machiavellian of the players were most able to exploit the situation to their greatest advantage, and do much better than chance would

XREF: Links to evolutionary psychology and Dark Triad research on strategic social manipulation.

By itself, a study such as Wason's might be dismissed on the grounds that it represents a fairly artificial laboratory task in which the subjects have little interest, but similar failures to seek disconfirmation have been observed in a number of different spheres. For example, Mitroff (1974) indicates that NASA scientists have a strong confirmation-seeking bias in their work, and Mynatt, Doherty and Tweeney (1977) report a similar bias amongst subjects attempting to discover the laws of an artificial universe simulated on a computer. Snyder and Swann (1978) conclude that subjects seek evidence of extroverted behaviour rather than its opposite when trying to evaluate a description of someone who has been described as extroverted. Tajfel (1969) makes the case that such a failure to seek disconfirmation, and an inability to recognize disconfirmation for what it is when it is presented to the individual by chance, are important factors in the maintenance of racial prejudice.

XREF: Connects to Wason's card selection task and confirmation bias literature in cognitive psychology. SEED: Potential article on how confirmation bias manifests in institutions — from NASA science to racial prejudice.

Generally, it might be thought that there are two factors at work. The first is the impact of depersonification. In a number of experiments, it has been noted that subjects who cannot see or hear the person to whom they are reacting tend to inflict far more extreme punishments upon one another. The most famous study demonstrating this is Milgram (1963). In this, most subjects were willing to give electric shocks, which they had some grounds for believing could be dangerous and possibly lethal, to another person when that person was unseen and in another room. Most were not willing to do so when seated right next to the recipient of the shocks. Clearly, anything which increases contact between parties will reduce this tendency.

XREF: Connects to Milgram's obedience experiments and broader psychology of deindividuation, also relevant to online harassment/anonymity research. QUESTION: How does depersonification's effect interact with accountability structures, social norms, and digital interfaces today?

Within any human society, the context in which a communicative act is offered is critical to its interpretation. This is true for the full range of such acts, from simple conversations to informal and formal contracts: even the most detailed of contracts cannot make explicit everything which must be understood for its interpretation. The beliefs which are held about another person at the time of communication are a significant part of what that context amounts to. Therefore, anything which contributes to a greater sense of knowing the other can contribute to the participants' belief in the possibility of mutual understanding.

However, the fundamental flaw in this work might be described as a failure to realize that expressed attitudes are communicative acts, and that they can only be satisfactorily understood in the appropriate context. This point is reflected in the fact that the most successful contemporary models of the attitudebehaviour relationship (e.g. Fishbein's) are those which try to build the context into the equation which relates one to the other (see Schuman and Johnson 1977).

XREF: Connects to attitude-behavior research and the context-dependent nature of survey/measurement responses.

Despite its apparent fragility and our many attempts to do without it, it is clear that, in many societies where it is well established, trust is remarkably robust. Fraud is widespread in all the advanced economies, and is extremely costly to companies and individuals alike, but rarely

QUESTION: What mechanisms allow trust to remain robust despite pervasive fraud? Worth investigating the resilience factors.

As was quickly discovered, the ‘brute force computing’ tactic, as the unbounded rationality assumption was known, was a failure because the decision space to be searched for any moderately realistic problem was of unmanageable proportions. For example, even contemporary machines take an inordinate amount of time to examine all possible two-move sequences in the opening stages of a chess game, because the number of branches in the relevant decision tree undergoes a combinatorial explosion. Not surprisingly, this tactic was quickly dropped, and much effort has subsequently been devoted to the development of more sophisticated and realistic alternatives. If we are to understood the development and maintenance of cooperative and trusting relations, similar unrealistic assumptions will have to be abandoned too.

XREF: Connects to the broader argument that rational-choice/unbounded rationality assumptions about cooperation must be abandoned, mirroring the chess example.

Trust is central to all transactions and yet economists rarely discuss the notion.! It is treated rather as background environment, present whenever called upon, a sort of ever-ready lubricant that permits voluntary participation in production and exchange.

I am using the word 'trust' in the sense of correct expectations about the actions of other people that have a bearing on one's own choice of action when that action must, be chosen before one can monitor the actions of those others.

DEFINE: Provides a precise operational definition of trust as expectations formed before verification is possible.

But the point is that his enemies did not know this. They thought that Yudhishthira could never be motivated to deviate from the path of truth. This is why the lie worked.

QUESTION: What does this suggest about strategic deception elsewhere—does credibility amplify the impact of rare deviations?

If the incentives are ‘right’, even a trustworthy person can be relied upon to be untrustworthy. ‘Every man has his price’: repugnant though it is to our sensibilities, the cliché captures the view that no one awards an infinite weight to his own honesty.

Our opinion is thus based partly on the theory we hold of the effect of culture, class membership, family line, and the like on a person’s motivation (his disposition) and hence his behaviour. The opinion which is publicly formed and held is this person’s reputation. Our problem in essence is to infer the person’s qualities from such data.

DEFINE: Reputation defined as publicly formed opinion inferred from qualities evidenced in behavior, filtering through theories of culture, class, and family.

From our point of view the interesting feature of the story is that Odysseus had the option of (quite literally) binding himself now to a future course of action (by excluding the option of trying to get to the Sirens). He exercised that option, and this enabled him to choose a further option (that of listening to the Sirens) which he would not have chosen had he not been able to so bind himself. … Odysseus’s customary self had, as game theorists would call it, a first-mover advantage over his bewitched self.

XREF: This is the classic Ulysses contract concept from game theory and precommitment strategies, with parallels in behavioral economics and rational choice theory.

At this equilibrium it is common knowledge that type 1 people are always honest and that type 2 people always cheat a little. But at no encounter does either party know who he is dealing with. Behaviour according to type is therefore what emerges at this equilibrium.

DEFINE: Explains how a pooling equilibrium works: type labels are common knowledge even though individuals never know who they're dealing with. Builds on: "Reputation is a public inference from observed conduct to character."

I begin by considering a 'dishonest' salesman, I formalize his disposition by the hypothesis that y exceeds a. To have an interesting problem I assume further that a is positive, that is y > a > 0. … The potential customer reasons backwards: 'If I enter the showroom and make a purchase the salesman wil sell me a lemon and my pay-off will be —6, which is less than zero — my pay-off in not buying a car at all. So of course I should not enter the showroom'. And he will not. We have thus a unique equilibrium of the game where no transaction occurs and where each party enjoys a pay-off — of nil. This is worse for both than the (non-equilibrium) outcome where a reliable car is sold and where the customer and the salesman enjoy @ and a respectively. A mutually beneficial transaction is thus unrealized.

QUESTION: How does signaling or reputation mechanisms break this lemons-market equilibrium in practice?

I conclude that if the seller's reputation for honesty is high — specifically if p exceeds 8 — the potential buyer will engage in a transaction. If it is low — specifically if p is less than 6 — he will not enter

DEFINE: p represents the probability threshold for seller honesty reputation that determines whether a buyer engages in a transaction.

Trust, as economists have noted before, is a public good, a social lubricant which makes possible production and exchange (see for example Akerlof 1970; Arrow 1974). … The point is that if your trust in me increases it certainly benefits me. But if there are good reasons for this increase in trust it benefits you as well. This latter benefit I do not take into account when I try to build up my own reputation. This is the source of 'market failure' and, in particular, why there might typically be an underinvestment in trust formation.

XREF: Connects to public goods theory, externalities, and the free-rider problem familiar from economics of reputation and social capital.

If the salesman is known to discount his future benefits at a low enough rate then it is the case that the outcome, where at each period the customer enjoys 1 —@ and the salesman a — that is, where at each date the customer enters the showroom and the salesman hands over a reliable car — is an equilibrium outcome even if there is no possibility of binding commitment of any sort.

DEFINE: Explains that a repeat-game equilibrium yields cooperation (reliable car each period) even absent enforceable contracts, contingent on the salesman valuing future payoffs sufficiently.

But while this example fits in part the notion of trust as I have defined it, it captures only one aspect of the idea of trust, namely that customers can trust the salesman to hand over a reliable car. It cannot convey the idea of the salesman trying to acquire a reputation for honesty — trying, through his behaviour, to alter people’s perception of what kind of person ‘he really is’. For notice that in the foregoing infinitely repeated game customers do not revise their opinion of him. They know he is dishonest (in the sense of having the pay-off structure of figure 4.1) and behaves honestly only because of the punishment (loss of customers) that will be meted out to him should he never renege.

DEFINE: Distinguishes bare trustworthiness from active reputation-building, where behavior changes others' beliefs about one's true character.

I think this is the right sort of framework to capture the idea that people invest resources for the purpose of building a reputation for honesty. The gains from honest behaviour are built directly into the model. The fact that the dishonest salesman builds his reputation by going against his short-term interests is an important feature to highlight. This is his investment.

DEFINE: Clarifies the economic concept that honest reputation is an investment: a salesman gains reputation by acting against immediate self-interest. Builds on: "Reputation-building involves altering perceptions of one's character"

The first, and to my mind most serious, weakness of this model is that there is nothing which enables the honest salesman to distinguish himself from the dishonest one. Everything the former can do the latter can do as well. The dishonest type is the cancerous cell which 'acts' like the healthy one. It is here that the role of commitments can assume importance. If the stake is high enough the honest salesman will search for ways of persuading customers that he really is honest; he will be willing to spend resources to distinguish himself from his possible counterpart. This is not included in the foregoing model of reputation acquisition.

Builds on: "Reputation-building requires sacrificing short-term gains for honesty."

The second weakness is that the strategies I sketch for the various participants are not the only credible ones. It follows that the outcome I describe is not the only viable one. There are other, equally credible, sets of strategies that result in outcomes which do not have this form of reputation acquisition, where transactions are not undertaken for a great many periods. In short, such repeated games possess more than one equilibrium outcome (see Fudenberg and Maskin 1986). There is therefore no guarantee that the one I have outlined will prevail. It is here that historical accidents may matter.

XREF: Relates to game theory concepts like the folk theorem and multiple equilibria in repeated games.

Trust as a human passion may rest on close familiarity or massive social distance. Many have trusted their Queen (or Stalin) as implicitly as ever they have trusted their spouse or favourite sibling. The essence of trust as a passion is the confident expectation of benign intentions in another free agent.

It is helpful to present the deficiencies of anarchism as a theory not by stressing its exceedingly skimpy and implausible conception of the causal determinants of collective social life and its vicissitudes (a charge which can scarcely be denied), but rather by underlining its blatantly capricious views about the incidence of trust in human social relations. … What really marks anarchism out as a political theory is the judgement that dispersed coercive power either is in itself, or would be if instantiated, or could readily be caused to become, quite sufficiently trustworthy for the living of acceptable human lives. At its most optimistic anarchism simply consists in the universalization of trust towards all humans who are not themselves bearers of concentrated coercive power.

This putatively realist perspective can be expressed in notably unapologetic terms — as it is, for example, by Thrasymachus in Plato's Republic, to some degree by Machiavelli in The Prince, and subsequently by Nietzsche. Thus presented, it is a perspective on power from the viewpoint of the powerholder (or of claques of intellectuals who more or less grovellingly identify themselves with the powerholder).

XREF: Connects the realist tradition from Thrasymachus through Machiavelli to Nietzsche as a coherent lineage of power-theory.

Where such institutions already exist and happen to be operating successfully it is reasonable for individuals to feel a stolid indifference towards the exertions which have brought them into being and to see them merely as occasions for current confidence. But where they have yet to be established the need for direct and exigent forms of trust is altogether more importunate.

XREF: Connects to institutional theory and social capital literature, where the difference between formalized trust and emergent trust parallels Ostrom's work on commons governance.

Much of human life, plainly, is shaped by forces which no human beings at all at the time of happening genuinely understand; and even if a few select human beings did happen to comprehend those forces, they could scarcely significantly affect them.

It is a view focused upon, and restricted by, human capacities: by the possible consequences of possible intended actions. It is within this (world historically, somewhat mole's-eye) view of human circumstance that the incidence of well-founded or ill-founded trust occupies such a key position. It does so because within this perspective what happens to a human society does depend both on the intelligence and practical skill of at least some of its members and on the use they elect to make of their freedom of action.

Fundamentally, trust is a device for coping with the freedom of other persons (Luhmann 1979; and this volume). It is because of the phenomenological plausibility of this conception of trust that it plays such a prominent role in shaping collective life.

DEFINE: Trust is defined as a coping mechanism for the irreducible freedom of other people, not merely a risk calculation.

Rather, it depends (and in ways which perhaps extend with comparative ease to some other species) on the presence or absence of relatively well-founded expectations about the conduct of others — expectations which can serve, until too harshly violated, as premises for partially cooperative conduct of one's own. What the root conception consists in is a picture of the conditional meshing of mutual interaction.

DEFINE: Clarifies the 'root conception' of cooperation as conditional meshing of mutual expectations, not mere rule-following.

No one can know how another human being will act in the future. Trust is a policy apt for conditions where knowledge is unavailable, as in the case of the free acts of another person (in Locke's view) it will always remain. Trust does not have to be any more credulous or sentimental than the judgement of those who decide how to allocate it, though it will in practice, naturally, not be any less so either.

XREF: Connects to the distinction between risk (probabilistic, calculable) and uncertainty (unquantifiable) from decision theory — trust sits in the latter domain.

For him the state was only an organizational system through which some human beings are enabled to act on behalf of (or against) others. … Men and women need in their rulers a power of agency which they can themselves only very marginally control. The subjects of every state are committed permanently by their political subjection to acts the character of which they cannot know and the consequences of which may embellish or devastate their lives.

XREF: Relates to Hobbes's Leviathan and theories of state sovereignty where subjects delegate agency they cannot control.

Trust in the relation between ruled and ruler is not a supine psychic compulsion on the part of the former. Rather, it is an eminently realistic assessment of the irreversibility of a political division of labour and a sharp reminder, from the former to the latter, of the sole conditions that can make that division humanly benign.

DEFINE: Reframes 'trust' in political relations as a contingent, reciprocal assessment rather than an emotional compulsion — a distinctive conceptual clarification.

Locke did not expect well-founded trust to be actualized at all frequently even in an essentially legitimate political society. But he thought that in political communities which were fortunate enough to be essentially legitimate the problems of political agency for all socially and politically active groups were always problems of how best to construct, reproduce, or repair structures of well-founded mutual trust.

Indeed what a modern state actually is, in large measure is a strongly institutionalized division of political labour, though not one which in any sense occludes the exercise of coercive force. The interesting question in modern political theory is not whether there is going to be a political division of labour (there is). It is the question of what forms that division is going to take and what are going to be the consequences of its taking those and not other forms.

There is no extant model of the most minimal plausibility which suggests that such legitimacy could be sustained other than momentarily and in extreme crisis without at least freedom of political speech and association and the right to organize and choose political representatives on the basis of the opinions which citizens actually hold.

But by the same token it would also need to distinguish more sharply than is customary in modern political ideologies between trust in the good intentions of more or less professional politicians and trust in their practical capacities. Trust in either might often be an agreeable start. But if it is necessary to choose between the two, it is wiser in most circumstances (the more so, the more democratic or legitimate the polity in question) to opt for trust in practical capacity. Politics is not on the whole good for the character; and it is unlikely that there really are sound reasons for viewing the intentions of most of those who have devoted decades to it with unreserved trust. But this provides no reason fox welcoming the chance of being ruled by the well intentioned but hopelessly ineffectual.

XREF: Inverts the common democratic intuition that voters should trust leaders' good intentions and moral character.

Here, especially, we can see how misguided it would be to hope to replace trust as policy (the properly sceptical choice of human political expedients) with trust as passion (in its characteristic modern form, an unreflective confidence in the efficacy and decency of existing institutions: state, party, government, union, or firm).

DEFINE: Distinguishes trust-as-policy (skeptical, deliberate) from trust-as-passion (unreflective confidence in institutions like state, party, government, union, firm).

The first is that the knowledge of how to make these instruments of destruction is ineliminable in principle except by their large-scale use. This knowledge and the possibilities which it embodies are with us for as long as anything we could sensibly think of as our way of life, or civilization itself, is going to continue. The problem is not how to rid ourselves of this knowledge. We cannot unlearn it. The problem is to stop it getting rid of us. The second reason is that the level of threat, and its erratic but rapid intensification, are themselves a product of obtrusive structures of well-founded mutual mistrust.

Building the structures of mutual trust that would be required to diminish the urgency of the risk cannot be done by burying one's head in the sand, still less by encouraging others to do so. It can only be done by focusing more clearly on the scale of the risk itself and establishing a clear mutual understanding of the priority of the need to reduce it over the need to diminish other — often in their own terms equally realistic and indisputably grave — anxieties.

To begin with, we have to avoid confusion between familiarity and trust. … Familiarity is an unavoidable fact of life; trust is a solution for specific problems of risk.

DEFINE: Establishes a clear definitional boundary between two commonly conflated concepts — familiarity and trust.

Familiarity is an unavoidable fact of life; trust is a solution for specific problems of risk. But trust has to be achieved within a familiar world, and changes may occur in the familiar features of the world which will have an impact on the possibility of developing trust in human relations. Hence we cannot neglect the conditions of familiarity and its limits when we set out to explore the conditions of trust.

Soon after we are born we begin making distinctions. An observer might think we were applying the beautiful logic of George Spencer Brown (1971). We arrive (and find ourselves included) in an unmarked

XREF: Connects to Spencer-Brown's Laws of Form and the logic of marking/unmarking, a foundational idea in cybernetics and systems theory.

Symbols are not signs, pointing to something else. They presuppose the difference between familiar and unfamiliar and they operate in such a way as to enable the re-entry of this difference into the familiar. In other words, symbols represent the distinction between familiar and unfamiliar within the familiar world. They are forms of self-reference using the self-reference of form. In fact, symbols have developed as the successors of myth, replacing it first by symbolic interpretation and later by pure symbolism.

This discovery of 'risk' as a general feature of life — roughly replacing what had previously been known as fortuna — does not facilitate the task of religion, but it adds another dimension to human experience.

DEFINE: Clarifies the historical evolution of 'risk' as a concept that replaced 'fortuna' as a general feature of life.

The distinction between confidence and trust depends on our ability to distinguish between dangers and risks, whether remote or a matter of immediate concern. The distinction does not refer to questions of probability or improbability. The point is whether or not the possibility of disappointment depends on your own previous behaviour. Since 'risk' is a relatively new word, spreading from whatever sources into European languages via Italy and Spain only after the invention of the printing press, we may suppose that the possibility of making this distinction is likewise a result of social and cultural development.

XREF: The confidence/trust distinction is Luhmann's core social systems theory concept; risk as a modern cultural category also appears in Beck's risk society literature. DEFINE: Clarifies a definitional distinction between confidence and trust based on whether prior behaviour accounts for possible disappointment.

Seen from this point of view, political and economic liberalism attempts to shift expectations from confidence to trust. Insisting on freedom of choice, liberalism focuses on the individual responsibility for deciding between trust and distrust with respect to politicians, parties, goods, firms, employees, credit, etc. And it neglects the problems of attribution and the large amount of confidence required for participation in the system. Mobilizing trust means mobilizing engagements and activities, extending the range and degree of participation. But what does this mean, if people do not perceive a condition of trust or distrust but a condition of unavoidable confidence? They will not save and invest if they lack trust; they will feel alienated if they lack confidence.

DEFINE: Distinguishes trust (individual, discretionary) from confidence (systemic, unavoidable) as the conceptual core.

Moreover, we have to acknowledge that the relation between confidence and trust is not a simple zero-sum game in which the more confidence is given the less trust is required and vice versa. Such a theory would neglect the structural complexity of social systems as an intervening variable. But a social evolution which achieves increasingly complex societies may in fact generate systems which require more confidence as a prerequisite of participation and more trust as a condition of the best utilization of chances and opportunities.

The case of trust is very different and requires quite another type of self-reference. It depends not on inherent danger but on risk. Risks, however, emerge only as a component of decision and action. They do not exist by themselves. If you refrain from action you run no risk. It is a purely internal calculation of external conditions which creates risk. Although it may be obvious that it is worth while, or even unavoidable, to embark on a risky course — seeing a doctor, for instance, instead of suffering alone — it nevertheless remains one's own choice, or so it seems if a situation is defined as a situation of trust.

Builds on: "Risk, unlike trust, depends on one's own prior conduct."

These considerations lead to the hypothesis that the printing press has completely changed our modes of coping with the unfamiliar. The distinction of familiar and unfamiliar becomes blurred; the religious technique of reintroducing the unfamiliar into the familiar through symbolization loses its former potence. There is no longer any need to distinguish symbolon and didbolon, good and bad forces, at a cosmological level. These schemes are replaced by the question of whether knowledge and power are used in a positive or a negative way, given particular interests. Thus confidence and eventually trust are the decisive issues, and familiarity survives as a purely private milieu® without function for society as a whole.

XREF: Relates to McLuhan's media theory and Ong's work on orality and literacy, both of which connect communication technology to cognitive and social change.

A second major change comes about when the predominant type of social differentiation shifts from stratification to functional differentiation. People® are no longer placed in a fixed social setting but must have access to all functional subsystems of the society on which they simultaneously depend. Structures become contingent; the law can be changed, if not by statute then by judicial practice (Horwitz 1977). The fluctuations of the economy are no longer thought to be limited by the boundaries of 'just prices'. Science surprises the public with new discoveries and new theories as a matter of routine. !° Essential structures and territorially bounded cultural entities are largely displaced by timelimited entities such as fashion and style.

DEFINE: Clarifies the distinction between stratification and functional differentiation as two modes of social organization.

Trust remains vital in interpersonal relations, but participation in functional systems like the economy or politics is no longer a matter of personal relations. It requires confidence, but not trust.

DEFINE: Distinguishes trust (interpersonal) from confidence (systems-based participation), a key conceptual distinction from Luhmann. Builds on: "Modern societies shift people from fixed roles to multiple functional systems"

The large functional systems depend not only on confidence but also on trust. If there is a lack of confidence there will also be a diffuse sentiment of dissatisfaction and alienation or even anomie. This may have no immediate impact on the system. If trust is lacking, however, this changes the way people decide about important issues. Trust, as may be recalled, is an attitude which allows for risktaking decisions. The development of trust and distrust depends on local milieu and personal experience.

These conditions may be extended by television culture, for instance in the case of political leaders. The testing and control of trust and the continuing perception of those symbolic events that eventually destroy it require a relatively concrete setting. They depend on a previous structural reduction of complexity. Moreover, they require a visible relation to one's own decisions about accepting a risk.

I want to explore what could be thought to be three ironic and even disquieting truths about trust.! I want first to argue that a socially extensive trust, that is to say, something more than trust between friends, cannot be created except in and by what I call 'aristocracies'; second, that having been so created, there is good reason to believe that the aristocrats will undermine it; and therefore, and third, that if trust is to be maintained, this will have to depend on conditions which are external to the social arrangements in question or which, if internal to them, might at first sight seem to be conditions that pre-empt anything one would want to call 'trust' at all.

QUESTION: What does the author mean by 'aristocracies' here, and how does trust get created within them only to be undermined?

There are at least three reasons to believe that a socially extensive and enduring trust is not an easy thing to create. The first of these is clear in Bernard Williams's contribution to this volume.

XREF: Relates to social capital literature and institution-building, where trust is often treated as a by-product rather than a constructible asset.

I do not by this mean, or do not only mean, a society bound together by a militaristic code of honour and propped up by a toiling mass in bondage and felt boots. I mean more generally a society which turns on a code, in which the quality of persons is measured by the extent to which they observe this code, in which there can, that is, be said to be 'persons of quality' — a society, in the familiar phrase, of 'virtue and honour'.

XREF: Connects to discussions of honor cultures and social stratification — the notion that personhood and worth are defined relationally through code-observance, echoing work on honor-shame societies and Weber's status groups. QUESTION: What exactly distinguishes this looser 'code-based' society from the militaristic honor-bound one being set aside? Where does the boundary fall?

I trust my friends, but just as there can be treachery in a kiss, so too there can be betrayal in an honourable mien. It has repeatedly proved too much to ask of persons of quality that they always remain so in those societies, which are all the societies there are, in which people will be presented with conflicting claims. … I see no reason to trust trust just as trust. And so, I see no reason to trust its markers. Given the syntactical and sartorial corollaries of aristocracies, I see no reason to trust someone who speaks and dresses too well; and given the sartorial corollaries of the new socialist person, I see no reason to trust one who dresses with too ostentatious a drabness either.

XREF: Connects to skepticism about surface markers of trustworthiness — resonates with ideas about signaling theory and the unreliability of social performance.

The Adam Smith I have in mind is the one illuminatingly reconstructed by Istvan Hont and Michael Ignatieff (Hont and Ignatieff 1983: 1-44). This is the Smith who argued that if one were to go over to an economy of high wages and a high division of labour, one would not only generate more wealth; one would also insure the poor against the capricious charity of the rich; one would replace an unpredictable provision by persons of quality with a predictable if minimal quantity; one would replace trust, which was not self-reinforcing, with a mutual interest in the results of the pursuit of self-interest, which would be.

XREF: Connects to the author's broader engagement with Smith's thought and the Hont-Ignatieff reconstruction.

Moreover, although these self-reinforcing systems do depend on other systems to support them - that was Hirschman's point — they do not in fact guarantee that those ancillary systems will work. They will work, of course, if pecuniary incentives are already in place. But it is part of the point of the argument that such incentives may not exist.

XREF: Connects to Hirschman's notion of unbalanced growth and dependency arguments, likely relevant to development economics discussions.

And most severely of all, even if the backward linkages do link, either in the way that Hirschman envisaged, or in some other way, they can do so in such a way as actually to strengthen those social and political relations — the hierarchical relations of patronage and so of 'corruption' that Hirschman, like Adam Smith, so disliked — which it is their purpose to override.

XREF: Turns Hirschman's and Adam Smith's optimism on its head — both disliked patronage and corruption, yet linkage-driven development may reinforce such hierarchies.

If the devices are not — the second of my three ironies — undermined by the aristocrats themselves, they can be undermined by a more predictable environment. And they can be overriden by a power — even an alternative aristocracy — which is independent of the interests which maintain them.

For this reason, as Cumings reports the colonial administrators saying at the time, what could be achieved with incentives in Taiwan required coercion in Korea.

QUESTION: What structural or cultural differences between Taiwan and Korea made incentives viable in one and not the other? Worth investigating the colonial context behind this asymmetry.

The most fundamental reason [why despotic rulers are unable to achieve their goals] is found in the lack of reliable information. ... The flaw of the system is that all acts tend to have an immediate political significance, which means that information degenerates into informing and so becomes worthless for planning purposes. Or else the information is offered that the informant believes his superiors want to hear, even if they insist on information that reflects the world as it is rather than the world as they would like it to be.

XREF: Connects to Scott's 'Seeing Like a State' on legibility and high-modernism, and to the general literature on how institutions distort information upward.

Exploiting the persistent fear of imminent invasion from the DPRK, a fear encouraged by the Americans, and assiduously gathering information on a large number of putative subversives in a wide variety of places in society, the KCIA had in turn exploited the preexisting distrust in the society to the government’s advantage. Indeed, the KCIA’s power may have exceeded that of any comparable agency in any country in the world, even, pace Elster, the KGB.

DEFINE: Explains the KCIA's operational logic and its unique historical position relative to comparable intelligence agencies like the KGB.

Nevertheless, even if it is conceded that the government of the ROK has in this way overcome what might otherwise have been an economically debilitating vicious circle, it cannot be said that it has created trust. On the contrary. And it is indeed perfectly possible that if and when the present pattern of rule changes in the country, as Chun Doohwan intends it to do in 1988, existing cooperations and the successful economic organization which depends on them could dissolve or even collapse.

XREF: Contrasts with later developmental-state literature on South Korean industrialization, and foreshadows the 1988 democratic transition questions. QUESTION: Interesting question: did Korean economic cooperation actually dissolve after Chun's rule change in 1988, or persist through institutionalization?

Yet there is at the same time no good reason not to think that its economic organization is — like Japan’s, which has recovered in several fundamental respects the character it had assumed in the 1930s — one which has already generated a sufficient interest in its continuation.

QUESTION: What does it mean for an economic organization to 'generate interest' in continuing? Is this about institutional self-perpetuation, or stakeholder investment?

I have argued that a socially extensive trust cannot be created except in and by what I have called 'aristocracies'; that having been so created, there is good reason to believe that the aristocrats will undermine it; and that if trust is to be maintained, this will have to depend either on conditions which are external to the social arrangements in question or on strategies which are internal and which, although perhaps directed by what could be called an 'aristocracy', might at first sight seem to preempt anything one would want to describe as 'trust' at all.

Builds on: "State-led economic cooperation lacks durable trust and may collapse."

In their different ways, the failures of backward linkage even in high-technology systems, the changes in agricultural practice and political affiliation in northern India, and the developmental successes of South Korea, each suggest that whatever may eventually result, a self-reinforcing and functional trust may not be able to generate itself, but may require something analogous to an 'aristocracy' to impose the initial conditions of its generation.

The second qualification, most evident from the example of Korea, is that it may be possible deliberately if deviously to create trust, which cannot always therefore be seen as a 'by-product', as Elster claims; and that at least one way of doing so is the way that Elster, after Tocqueville, does not believe is possible. If this in turn is right, it would seem to follow that in some conditions some 'aristocracies' — a political party, for instance, or an army — can create trust after all, and can do so by creating distrust.

XREF: Challenges Elster and Tocqueville's view that trust is only ever a by-product, relevant to any prior engagement with their work on social capital.

Trust — fides — was also, for anyone reared (as all eighteenth-century Italian theorists were) on Machiavelli, the necessary condition of all political association. Trust, wrote Doria in his Della vita civile of 1710, 'is the sole sustenance of states and leads to their stable maintenance' (1852: 352). Fede here is a purely social and a visibly public virtue. It is, in Genovesi's telling phrase, fede pubblica.

DEFINE: Fides/fede pubblica is defined as a social, public virtue — the foundation of states per eighteenth-century Italian theory.

In the sense Doria used it when describing what he, and indeed most European contemporaries, regarded as the single defining characteristic of Spanish culture, it implied the existence of a self-love vested in, and consequently at risk through, the persons and objects -and most particularly the persons viewed as objects, that is, the women -by which a man (for a woman's honour is merely a repository for the man's) is surrounded.

DEFINE: Clarifies the concept of honour in Spanish culture as male self-love dependent on women, who serve as repositories for male honour.

But there is also that other kind of honour which derives from Aristotle's account in the Nichomachean Ethics of 7.un (book 1, chapter V, 5). This is public, in that it requires a witness, and thus implies a relationship between two or more parties and with persons outside the kin group.

XREF: Connects to Aristotle's Nicomachean Ethics, specifically the distinction between honour that depends on public recognition versus internalized forms of esteem.

This understanding of honour —the sense in which the term is used in such phrases as 'honouring one's bond' — is, of course, a crucial component of trust. As Genovesi said, onore e virtu (honour and virtue), in the absence of any effective coercive mechanism, can be the only guarantors of a debtor's reliability. But, as we shall see, no virtue is possible in a society of honour understood in the first sense.

XREF: Connects to broader themes of trust, reputation, and social capital as informal enforcement mechanisms in economic contexts.

The virtuous republic is, of course, a society grounded on a set of shared and stable values. These values — trust in and concern for one's fellow citizens, the triumph of public over private good - depend crucially upon the amount of information each party has.

XREF: Connects to debates about digital divides, media literacy, and how information asymmetry undermines democratic trust.

The Spaniards had succeeded in destroying the purchase of the laws on the community by subverting the customs which sustained them; they had destroyed the rule of law by setting up separate courts both for the barony and for the priesthood and by allowing into the legal system entire categories of exemptions and exceptions so that no one could predict the outcome of a case or know which part of the law applied to him.

‘For commerce’, he wrote, ‘is an art which unites men in civil society so that they give to each other mutual assistance ... in such a way that neither the distance nor any other consideration can impede this mutual aid.’ But trade can only flourish under two conditions, ‘and these are liberty and security in contracts, and this can only occur when trust and justice rule’

Counterfeiting in particular was, Genovesi pointed out, punished with the utmost severity even among ‘barbarous nations’ such as the Incas, since it threatened the social bonds which held together the entire community (1803: 102-9). In Naples, however, a man might be executed for theft, a crime whose consequences were hardly to be compared with the loss of life, but let off with a fine for an act which, in Doria’s view, ‘offends against public trust [fede pubblica] and thus against the entire republic’

XREF: Connects to modern debates about white-collar crime versus street crime sentencing disparities.

With the collapse of those conditions on which true commerce depends, trade became a question of mutual deception. Bonds and even money, since so much of it was false, were no longer freely accepted and the Neapolitans were reduced to the condition of the savages described by Genovesi who will only give with the right hand if they simultaneously receive with the left (1803: 70-1). As a consequence of massive degrees of mistrust loans were only made at exorbitant rates of interest, bankrupting all but the richest of merchants and leading to the widespread belief that a predominance of ‘idle money’ was the source of the kingdom’s ills

For Doria and Genovesi trust is the basis of the well-ordered republic. It is, in short, the motive for a man to behave towards members of the society at large in much the same way he behaves towards members of his own kin group; or, at least, it provides strong reasons why he should not ~— except in purely private concerns — privilege that group over any other.

Quentin Skinner has argued forcefully that the kind of humanistic ethics Doria and Genovesi were attempting to revive, which set this idea of the ‘wellordered republic’ at the centre of all social relations, employed a similar conceptual vocabulary to Calvinism; that — to put it the other way round — the social ethos of Calvinism can be seen to constitute a special case of classical republicanism (Skinner 1974a; 1974b). If we accept this view then many of the problems historians have had with Weber’s famous thesis on the association between Calvinism and the rise of capitalism disappear, since it might be maintained that it was not a specifically Calvinist or Puritan work ethic which encouraged economic take-off in the Protestant cities of the seventeenth century, or Christianity itself in the medieval cities of north and central Italy, but the secular ethic of classical republicanism.

XREF: Directly engages Weber's Protestant ethic thesis, reframing it as a special case of classical republicanism rather than a distinctively Calvinist work ethic.

This method is to gang up in a group, which in effect hangs up a notice saying: anyone who commits an act of aggression against any one of us must expect retaliation from us all, and not only will the aggressor himself be likely to suffer retaliation, but his entire group and all its members will be equally liable. And this notice is in effect posted by the very culture which pervades pastoral societies. It constitutes the code of honour which is familiar to all. So the gangs themselves do not need literally to put an announcement in the press or even on their tent. The culture, or specifically the obligation of feud which is inherent in it, does it for them.

Inside each such group, order is maintained by a similar mechanism: the group itself divides into sub-groups which each exercise restraint over the others.

This is a teleological or 'functional' argument, of course, open to the objection that the need for the cohesion of social groups will not automatically engender that cohesion. In the absence of a designer and creator, the favourite logical substitutes are our old friends, natural selection and rational foresight. Presumably each of them operates, though no one really knows in what proportion.

QUESTION: The unresolved proportion question — how much social cohesion comes from natural selection versus rational foresight — seems genuinely open and unexplored here.

Treason is built into the working of such a society and its culture and, if carried out properly and in due form, no stigma attaches to it. This kind of society could not possibly function without treason. Treason within it performs the same role that price changes perform in a market society. It helps to maintain equilibrium through realignment. Just as a rise or fall in price ideally redirects resources to the point at which they are optimally deployed, so a realignment — treason, if you will — restores balance to the segmentary tribal world. An outnumbered clan Cl, unable to hold its own within tribe T1, will leave T1 and join T2.

XREF: Parallels market equilibrium economics — price reallocation analogized to tribal realignment. Might connect to game theory and coalition literature.

Given that the society as a whole needs a level of production and trade which presupposes urban centres, and given that urban centres (clusters of specialists) are quite incapable of social cohesion and hence of political and military effectiveness, what is to be done? What is done is that the tribes provide cities with rulers: the state is the gift of the tribe to the city. Economically, the tribe needs the city: politically, the city needs the tribe.

XREF: Connects to the broader tension between urban specialization and social cohesion — a recurring theme in political sociology and state-formation theory.

The rulers in the citadel are the people who acquired the capacity or the tendency to trust each other, and hence to be cohesive, by being ungoverned. If you wish to govern, you must never have submitted to government yourself. Only the ungoverned condition of the savannah, desert, and mountain, engenders the kind of tribal group whose members trust each other and display cohesion.

XREF: Inverts the usual Hobbesian logic that civilization tames the unruly into governable subjects; here the ungoverned are themselves the natural rulers.

This and this alone makes such groups capable both of government and self-government. Consequently, from time to time, some of them occupy the citadel. Once they are there, and this of course is the best known part of Ibn Khaldun's theory, they gradually lose that very cohesion they acquired when they were ungoverned, and which enabled them to rule. In his schematized version, this takes about three or four generations. They lose it in the process of governing and of benefiting from the privileges of government.

XREF: Echoes cyclical theories of imperial decline elsewhere, such as Machiavelli's return to first principles or Polybius's anacyclosis. QUESTION: Does the privilege of governance necessarily corrupt group cohesion, or is this specific to Ibn Khaldun's desert/asabiyyah framework?

In the politically organized, self-administering countryside, under the cloak of a nominal adherence to an exclusive unitarianism, there is in fact a Durkheimian religion which works by providing social punctuation in space and time. The sacred is incarnate, with a local habitation and a name. It helps to define groups, define their status, define their boundaries: it reinforces their cohesion.

In the towns, at least at the top, there is something which does more closely approximate to the Islam of the learned theologians: an exclusive unitarianism with a hidden deity, which proscribes mediation, and so theoretically recognizes no clergy, and which preaches an ethic of rules rather than of loyalty. It neither expresses the cohesion of groups, nor provides an idiom of treachery.

The difference in style is well reflected in how you testify at court. The country folk testify in groups in collective oath. Under proper Muslim law, you testify as an individual. The collective oath of the tribesmen is really an expression of commitment to the group, an expression of a willingness to bear responsibility with it. It is not a testimony to matters of factual accuracy, which is the purpose of the oath offered by the individual in an urban trial.

A person's safety no longer lies in the strength of his kin group, but rather in having a friend at the nearest district office, who in turn has a friend at the provincial capital, who has a friend at the national capital. So, a sort of dyadic unsymmetrical system of patronage relations takes over in the political sphere.

The most significant cultural event in the Muslim world of the past 100 years has been a very major Reformation, a shift from mediationist religious brokerage to a symmetrical relationship between the deity and believers in conformity with the old, but previously unpractised, high theology. Suddenly religion is egalitarian and symmetrical, centralized, with an exclusive single deity and without mediators.

XREF: Parallels the Protestant Reformation's removal of priestly mediation, an explicit comparison the author draws.

In the Islamic model the central authorities were not positively hostile to trust on any theological principle, but were simply intolerant, within the very considerable limits of their capacity, of the emergence of coherent trust-engendering groups which would pose a threat to them. They would keep such groups at bay, but they could not destroy them.

XREF: Connects to broader historical analyses of how state power interacts with civil society and social capital formations.

In conclusion, cohesion and trust contribute net merely towards the establishment of a social order as such, but also towards the establishment of checks on government.

Religion provided an idiom in terms of which the various elements of the total society could communicate with each other. It provided the terminology for both tribal trust and urban fear. It could intermittently help tribal cohesion to extend its scale, and to revitalize the polity by providing a new set of rulers.

Tribal life, conditioned primarily by lack of government and the consequent need for cohesion, was the only source of a certain kind of trust, which bound men in groups, even though it was also haunted by the possibility of treachery. In the towns, there was specialization, no trust, and order based on fear.

DEFINE: Contrasts two distinct types of social cohesion: tribal trust (born of necessity, bound by group cohesion) versus urban order (based on specialization and fear rather than trust).

The use of the term equilibrium, therefore, is not just an analogy but is adopted here in the sense of a state of affairs in which all or most agents, in spite of-what they may think of the collective outcome, have not found adequate incentives to behave differently and to change that outcome in any significant way.

DEFINE: Defines economic equilibrium not as metaphor but as a precise condition where no agent has adequate incentive to change their behavior. XREF: Connects to game theory's Nash equilibrium concept and collective action problems.

Through this fictional account, by which Tocqueville manages to convey his reflections on the journey, we begin to grasp some of the elements of the peculiar process of adjustment to — rather than rebellion against — domination which to different degrees involves both Naples and Sicily: the absenteeism of an aristocracy lured by the pleasures and servility of the court; the predominance of private over public concerns, of duplicity, cunning, and favour over merit; and the fashion for crime and murder, which have become habitual and are even felt to be legitimate

XREF: Connects to theories of learned helplessness and accommodation in subjugated societies, as well as Gramsci's cultural hegemony.

Moreover, not only did the Spaniards exploit distrust for the purpose of domination, but they taught some of their subjects to do so too and to pass it on to others. The Bourbon Spanish domination, which replaced the Habsburgs in 1724 and, except for a brief interval, lasted until the Italian unification in 1861, continued to pursue the policy of divide et impera and took particular care to foster the hatred between Neapolitans and Sicilians

XREF: Connects to classic divide-and-conquer colonial strategy seen across other empires studied.

It is probably not a coincidence, though, that only 11 years (1838) after Tocqueville wrote his account, we read for the first time in an official report of the existence of a thing called the mafia (Hess 1986: 114); and, what is more, we read of it as an already established social force.

XREF: Links Tocqueville's observations on Sicily to the first documented mention of the mafia in 1838, suggesting the organization predated its naming.

But the middle class can sustain the search for exclusivity with the same personal gifts which in other circumstances it invests in peaceful business: order, foresight, caution, and cunning (p. 97). They go so far as to practise understatement about their real power, a style they ostensibly share more with the British than the Italian mentality. This style, moreover, has won them reputation and trust.

The reasons, both theoretical and historical, have yet to be properly understood, but the evidence suggests that neither has the mafia evolved towards a single monopoly successful in submitting all others, nor has it dissolved. The characteristics of its 'persistence suggest those of a turbulent equilibrium.

Franchetti says that in Sicily those who are clever, energetic, and ambitious can only find a way to improve their social position by dedicating themselves to the 'industria della violenza' (p. 97).

XREF: Connects to broader discussions of mafia sociology and how blocked social mobility channels ambition into organized crime. Also relevant to theories of state failure and private violence provision. Builds on: "Mafia equilibrium persists without monopoly or dissolution"

A wide range of evidence (Arlacchi 1983) suggests that, on the margin, people who migrate do so also because of the mafia. Given that 'mafia-averse' people tend to migrate more frequently, the degree of opposition is likely to decrease and the power of the mafia to increase. The effects of migration are clearly unintended, because those who migrate do not do so in order to enhance the force of the mafia (although they may of course recognize those effects).

DEFINE: Illustrates an unintended-consequence mechanism: selective out-migration of opponents reduces local opposition and thereby empowers the mafia. Builds on: "In Sicily violence is the only career path for the ambitious."

The latter, through the example Hawthorn gives of South Korea, consists of an authoritarian and military rule which, in itself monopolistic, would not tolerate local monopolies. In a sense, this implies that in order to get rid of the mafia, what we need is simply another — bigger and better — mafia.

Builds on: "Mafia-averse emigration unintentionally concentrates and strengthens remaining mafia power."

Italy did once have a power which was largely independent of the mafia, or at least of the social strata supporting it, and which did contribute to a partial undoing of the mafia (Duggan 1985): this was of course the Fascist regime. In contrast, democracy, by its very nature, has to rely for consensus on larger parts of the population. Thus the temptation to come to terms with those who hold a monopoly of people's votes, regulate the dispensation of political trust, and somehow guarantee local 'law and order', has been strong, and in several instances has proved irresistible.

XREF: Connects to broader literature on democracy's accommodation of informal power brokers and the relationship between authoritarian regimes and organized crime.

The mechanisms which motivate cooperation in any form of human endeavour, as Bernard Williams explains elsewhere in this volume, comprise four basic elements: coercion, interests, values, and personal bonds. People, that is, may decide to cooperate (1) for fear of sanctions; (2) because cooperation enhances their mutual economic interests; (3) because they have general reasons, whether cultural, moral or religious, to believe that cooperation is good irrespective of sanctions and rewards; and finally (4) because they are related to one another by bonds of kin or friendship.

XREF: This four-part framework echoes familiar motivations-of-cooperation theories (coercion, self-interest, values, kinship), connecting to game theory and social contract thought.

Their combination, within a world of deep distrust, is itself a robust pillar of mafia business for, irrespective of values and cultural codes, the force of constraints and opportunities can bring about rational adaptation on the part of people living in proximity with mafioso networks, even if they are not related by kin or do not entertain strong beliefs about the social importance of adhering to the code of honour or that of silence.

Many mafiosi have begun their careers with violent acts (Hess 1973; Arlacchi 1983), but have subsequently relied on the reputation with which such acts provided them: 'basta la fama' wrote Franchetti (p. 104), in line with some of the most advanced game theory (see Milgrom and Roberts 1982, who show that, contrary to the standard economic claim, 'predatory practices' to maintain a monopoly are not irrational if reputation and future challengers are taken into account). They become persons with a reputation and are trusted, and, if interests diverge, they are trusted, in a limited but effective sense, to resort to

XREF: Connects to game theory literature (Milgrom and Roberts 1982) on reputation and credible threats, relevant to any work on signaling and deterrence.

The distinction between a damage avoided and a benefit gained is up to a point artificial. [In most cases] the line that separates them is impossible to determine, or rather it does not exist in human feeling. When evildoers intrude on and dominate most social relationships, . . . the very act that saves one from their hostility can also bring their friendship with all its associated advantages (p. 129). … The violence of extortion and the self-interest of the 'victim' tend to merge and to provide an inextricable set of reasons for cooperation: the advantage of being a 'friend' of those who extort one's money or goods is not therefore simply that of avoiding the likely damages that would otherwise ensue, but can extend to assistance in disposing of competitors, or protection against the threats of isolated bandits, and against the risk of being cheated in the course of business transactions.

We know that the absence of trust in business has a devastating effect: a high expected probability of being cheated (of being saddled with a 'lemon') may lead to the non-emergence or even to the collapse of market exchange, especially in those cases where asymmetric information - concerning the quality of the goods exchanged - is relevant (Akerlof 1970; Dasgupta, this volume). Within the cluster of people they protect, mafiosi offer a peculiar solution to the problem raised by the market of 'lemons'.

XREF: Connects to Akerlof's 1970 'market for lemons' theory on asymmetric information in markets.

The mafioso, by means which are left unclear but which presumably involve his reputation for toughness, seems capable of deterring the seller from handing over a 'lemon'. Without that protection the coachman is truly at risk, for he is indeed, as it were, saddled with a bad horse. … More difficult to interpret is the fact that the seller too gives the mafioso a tip. This could suggest, in line with Dasgupta's argument (this volume), that with respect to that particular transaction the protection the mafioso offers is really a public good which benefits both sides. If he did not act as guarantor the exchange would not take place at all, for the potential buyer would be deterred from entering the transaction for fear of getting a bad deal. The seller's tip, in other words, might reflect the price he is prepared to pay to be trusted.

DEFINE: The 'lemon' problem in markets where quality is unobservable, with the mafioso as an extralegal guarantor who makes exchange possible by deterring bad-faith behavior on both sides.

I was captivated by what seemed a paradox: on the one hand the banality of a barely differentiated Dickensian mob, of water carriers, bread sellers, shit shovellers, taxi drivers, pickpockets, and prostitutes; on the other the communal spirit of hill tribesmen whose fathers were earth priests and who expected to end their days as custodians of ancestral shrines. I was impressed by the energy and ingenuity of their efforts to enrich themselves and by the inevitability of long run failure for all but a handful. It seemed as if the economy was being made, unmade and remade from day to day.

XREF: Echoes the churn of early commercial capitalism and Schumpeter's creative destruction applied to a pre-modern urban setting.

There remains the zone of free-floating social relationships formed by choice in the expectation of mutuality. The most neutral term for this zone is association, but its strongest form of expression is friendship, the negotiated order of free individuals joined by affection and shared experience rather than by legal sanction or the ties of blood. It is in this third area of social life that trust plays so prominent a role, relatively unmediated by the formal obligations of kinship and contract.

Since bargaining depends on the ability of 'each party to walk away from the deal, this pervasive structure of credit and clientelistic particularism restricts the scope for such behaviour. It is not accidental that game theorists talk about used cars and 'lemons'. The rational negotiation of risk is less appropriate when most purchases go

I noticed that something like chains of debt were commonplace. Each migrant made and received loans (including traders' credit). If a man came into a windfall — a stolen wallet or a back-dated pay rise — he was as likely to lend the money as to pay off existing debts. This meant that when he was being dunned by a creditor, he could pass on the pressure to one of his debtors. There is an analogy with traditional bridewealth transactions where in-laws typically chase an infinitely regressive chain of debt in order to retrieve an outstanding cow. Another metaphor for this process, besides that of a chain, might be the ripples caused by a stone dropped into a pool.

XREF: Connects to general economic anthropology on debt and obligation networks, and to Mauss's gift-exchange theory.

The sanctions available to traders when dealing with defaulters were weak. One woman was reduced to entering a client's room at 4 a.m. to steal his only pair of trousers, which she then waved triumphantly in the courtyard while shouting out his crimes: 'You think Alhassan is a big man because he walks around in trousers and white shirt, but he won't pay a poor woman for the bread and cola he eats!' Public humiliation of this sort is a poor substitute for legal sanctions.

QUESTION: What alternative enforcement mechanisms existed in this trading community beyond humiliation?

Half the sample neither spoke nor wrote any English (Ghana's official language) and only one in six was literate. Yet, as a group, they were remarkable for having made a break with traditional religion; whereas more than nine out of ten Frafra migrants were pagans, 21 per cent of the entrepreneurs were Muslim and 18 per cent were Christian or some agnostic mixture, a total of almost 40 per cent in the Ghana sample. This figure rose to near 50 per cent in urban areas and 60 per cent in the southern cities of Accra and Kumasi.

Most naive operators opt for hiring a driver for fixed wages, since the prospective profit is greater. This is why they usually fail: a wage employee has no incentive to maintain the vehicle or to be honest with the takings. … The most secure method is to sell the vehicle to the driver on an instalment plan and make him responsible for maintenance, a method pioneered in Ghana by Lebanese traders.

XREF: Ties to principal-agent theory and incentive design in economics; parallels modern questions about gig-economy versus salaried labor incentives.

Many of the catering and retail entrepreneurs were single women; and few of the men had conjugal relations which allowed them to enter into partnership with their wives, since the traditional role pattern is typically both patriarchal and segregated. … The fact that descent organization did not travel to the south meant that kinship ties were much more fluid and the authority of household heads was frequently weak. Here the contrast is with Asian minority communities whose tight cohesion provides patriarchs with external support to domestic hierarchy. … One significant exception consisted of a group of Christian entrepreneurs for whom marriage meant an egalitarian partnership with a wife conceived of as friend.

XREF: Contrasts with Asian minority communities' tight cohesion supporting patriarchal authority — a useful comparative frame for kinship and gender studies.

Rates of return are nominally very high, interest rates of 25 to 50 per cent per month being normal as against the legal maximum of 30 per cent per annum and current bank lending rates of under 10 per cent per annum. On the other hand, default may be as high as four loans out of five, an expression both of the general lack of trust and of the value attached to trustworthy relationships.

XREF: Connects to informal credit markets and why microfinance and rotating savings associations emerge where formal banking is absent.

Whenever my landlord introduced me to one of his 'good friends', I knew that he was almost certainly a crook, probably from another ethnic group, a member of a criminal fraternity stretching back decades with a common background of gambling dens, police raids, gaols, diamond smuggling, drug rings, and all the rest. These were the men he turned to when he needed to trust someone, not his family or his fellow tribesmen.

To cut through the detail, I wish to propose a continuum of words connoting belief based on the degree to which they rest on evidence of the senses. Faith requires no evidence; trust is an expectation based on inconclusive evidence, is tolerant of uncertainty or risk; confidence is a strong conviction based on substantial evidence or logical deduction. These are all subjective attitudes. … Belief is a feeling that a person or thing will not fail in performance. Feeling varies inversely with evidence or proof. Thus faith is an emotionally charged, unquestioning acceptance.

DEFINE: Distinguishes faith (no evidence), trust (inconclusive evidence), confidence (substantial evidence), and belief (an inverse-of-evidence feeling). XREF: Connects to epistemology and decision theory frameworks distinguishing certainty from justified belief.

Modern usage restricts the number of our friends to a narrow circle of intimates, narrower by far than those whom we might occasionally be inclined to trust. But among the Frafras of my study — and perhaps for medieval Englishmen — trust was closely bound up with an idiom of classificatory friendship which had a relatively broad range of application.

XREF: Contrasts modern Western friendship norms with the Frafras' and medieval English broad classificatory idiom.

I agree with the emphasis placed by both Partha Dasgupta and Niklas Luhmann (in this volume) on the importance of personal agency in defining trust. Thus Dasgupta identifies trust as an expectation about another person's actions which influences our own actions before we have full knowledge of the outcome; and Luhmann exposes to full view the distinction between personal and impersonal social organization. Clearly trust is the predication of one's own actions on actions of others which bear an identifiable risk of turning out unfavourably.

DEFINE: Distills a definition of trust common to Dasgupta and Luhmann: acting on expectations about others before outcomes are known, accepting identifiable risk.

The Frafra migrants face the problem of establishing economic forms, durable relations of partnership and hierarchy, in the city slum. Three basic models for how to go about this present themselves. The most obvious and apparently profitable of these is the contractual ethos of the civil society they think they have joined, that is individualism, rationality, market, impersonal law — the bourgeois package of city life. Here economic relations are in principle entered into freely and by choice, but the contract imposes binding obligations sanctioned by state law. There are several reasons why this does not generally work. State law does not apply except in the form of occasional punishment. Markets are erratic and the substantive conditions of rational calculation are largely absent.

XREF: Connects to Polanyi's embeddedness argument that markets require institutional and legal scaffolding not spontaneously present.

Looking for some alternative form of guarantee, the migrants turn to the antithesis of modern civilization, their own customary moral institutions founded on the identities of kinship, descent, and family, reinforced as they are by a common language and religion, by the certain obligations of birth and community. This option too is fallible. The institutions do not travel en bloc to Accra, and some of the important public sanctions of domestic hierarchy are missing: lineages, ancestor worship, the security of their homeland. The migrant community is an egalitarian brotherhood of floating young men anchored in the patronage of a few resident big men. It is not closed or powerful enough to shore up the authority and reliability of kinsmen.

This picture of comparatively rapid growth and technological sophistication contradicted established views of the role of small firms in the French economy. In particular, it was inconsistent with the dualist model which predicted technological backwardness in accordance with the confinement of small firms to unstable portions of the product market.

What I was observing, however, was different. It was a shift to subcontracting on a permanent basis for such standard operations as turning, milling, and drilling. It allowed the firms to avoid making investments in up-to-date machine tools and was frequently the occasion for a reduction in capacity, with some existing plant being sold off. While the general type of operation subcontracted was not specialised or specific to the particular firm in question, the design and specifications of the components were. Thus it was not a case of substituting in-house production for standardized components available in the market: rather, components were being machined (turned, milled, etc.) by subcontractors according to firm-specific plans produced in the design offices of the client firm.

XREF: Connects to the make-or-buy literature and analyses of vertical disintegration in production networks.

This form of subcontracting is encouraged by the rigid restrictions French legislation places on laying off workers. During the 1970s these restrictions spawned a proliferation of temporary help agencies whose employees do not benefit from the same protective legislation. Another common arrangement is the use of so-called intérimaires, permanent employees of one firm with an excess supply who are loaned on a temporary basis to a firm facing a capacity constraint.

XREF: Connects to labor market rigidity discussions and the economics of employment protection legislation.

Cost considerations, then, dictated the initial switch from in-house production to reliance on the market. Further discussion showed, however, that these evolving market relations were a far cry from those of the standard textbook, where 'faceless buyers and sellers meet for an instant to exchange standardized goods.' By 1984-85 the firms had begun to use a suggestive word to describe their relationship to these new subcontractors: partnership (partenariat). In the course of conversations in which I myself participated, other equally emotive terms evoked the nature of interfirm relations: the importance of loyalty (fidélité); the existence of a moral contract (contrat moral); and the need for mutual trust (confiance mutuelle).

XREF: Contrasts with textbook neoclassical economics where market exchanges are faceless and instantaneous. Connects to broader literature on embeddedness and relational contracting.

Trusting behaviour consists in action that (1) increases one's vulnerability to another whose behaviour is not under one's control, and (2) takes place in a situation where the penalty suffered if the trust is abused would lead one to regret the action. In economic terms this implies that the action would not be taken in the absence of trust because the expected net benefit is lower than if some alternative is chosen. In short, there is no best strategy independent of trust.

Firstly, trust presupposes decision-making in a situation of risk, where the risk is attributable to the strategic behaviour of others or to the possibility that they will behave opportunistically. By opportunistic behaviour I have in mind not only such blatant forms as stealing and lying, but also more subtle techniques such as withholding information in an effort to confuse. As Dasgupta (this volume) has noted, the possibility of such behaviour is a necessary condition for the question of trust to arise. If all people are invariably honest, doing their best to fulfil their commitments, then there is no problem of trust as I have defined it.

DEFINE: Defines trust conditionally: it presumes risk from others' strategic or opportunistic behavior, including subtle forms like withholding information. XREF: Connects to the broader trust literature, including Dasgupta's work on risk and commitment, foundational in social capital theory. Builds on: "Trust requires voluntarily increasing one's vulnerability to another party."

Secondly, the action and hence the risks are avoidable: one does not have to engage in trade with another firm (although this implies forgoing the potential benefits of trade). Being able to avoid the relation is fundamental. If you could not, you might say something like: 'I have no choice but to trust this person, institution, etc.' Clearly, there being no choice, we need not invoke trust to explain our behaviour. Of course, as Luhmann (this volume) points out, whether or not a relation is seen to be avoidable is often highly subjective, and presumably varies in accordance with the structure of institutions and political power.

I would like to suggest that the transaction cost literature (associated notably with Williamson 1975; 1985) can tell us something about the role of trust in the economy. As the term 'transaction cost' suggests, this literature is concerned with the organizational implications of the costs of making a transaction. This includes not only the cost of reaching an agreement satisfactory to both sides, but also that of adapting the agreement to unanticipated contingencies and of enforcing its terms. … The relation of transaction costs to the economy parallels that of friction to a physical system: it is often ignored in formal models but is none the less of great practical significance.

XREF: Connects to the reader's likely familiarity with Williamson's work and the analogy to physical friction as an explanatory device in economic modeling.

If transaction costs are thought of as friction in the economy, then trust can be seen as an extremely effective lubricant. To quote Arrow (1974: 23): 'It saves a lot of trouble to have a fair degree of reliance in other people's word.'

DEFINE: Introduces trust as a friction-reducing mechanism in economic terms, framing it metaphorically as lubricant against transaction cost friction. Builds on: "Transaction costs resemble friction in their practical significance"

In conclusion, then, once firms are locked into a relation, trust is essential. The need for trust can be circumscribed through the efforts of the parties to engage in comprehensive contracting. But the limits to our rationality in a world where surprise is inevitable ensure that such efforts will be imperfect at best. Trust cannot be disposed of entirely.

The first point to be made is that there is extensive use of safeguards by client firms to minimize the possibility of being the victim of opportunistic behaviour and hence the need to rely on trust. There is an Italian saying which captures much of the sentiment: 'It is good to trust but it is better not to trust.' First, client firms prefer to solicit tenders from a minimum of three subcontractors in order to preclude an opportunistic distortion of production costs. They also prefer to split an order between a minimum of two subcontractors so that if difficulties develop with one it is possible to switch to the other. This practice may, of course, entail a loss of potential scale economies depending on batch size.

The evidence suggests that reputation is important but no substitute for experience. My information pertains primarily to the client’s perception of the subcontractor and I would not want to exclude entirely the likelihood of asymmetry here. The client firms are comparatively visible, being prominent members of the Lyons engineering community and in most cases larger than the subcontractors. My interviews suggested that they are acutely aware of their reputation for reliability. When asked how they signalled their trustworthiness to potential subcontractors, they invariably referred to this.

It would be difficult to explain how trust could be present as often as it is in subcontracting if it were to depend on sentiments of friendship. If this were true, we would have to adopt in the economic sphere Hawthorn's (this volume) conclusion in the political: that any attempt to produce trust which was not in fact an attempt to produce something else, something self-reinforcing, must fail. It seems implausible to entertain feelings of friendship for someone when it is realized that he or she is acting in a friendly way solely for the purpose of facilitating an economic transaction.

XREF: Connects to Granovetter's embeddedness argument about how social relations undergird economic transactions — this appears to be a companion piece engaging the same debate about trust in economic life.

Economists generally assume that the narrow pursuit of interest results in efficient economic exchange. The aim of this paper has been to dispute this assumption. … The theoretical section addressed two issues: to determine the conditions under which the need for trust develops in relations of economic exchange … Trust is crucial when contracting parties invest in specific assets, locking

XREF: Connects to trust economics literature, Williamson's transaction cost economics, and relational contract theory.

'According to the trite observation' — Adam Smith wrote — 'if there is any society among robbers and murderers, they must at least. . . abstain from robbing and murdering one another' (Smith 1976: 86; see also Saint Augustine in Dunn, this volume). This 'trite' observation serves a double purpose: it reminds us that basic forms of cooperation are inevitable if a society is to be at all viable, but it also points out, perhaps unwittingly, that there are instances of cooperation — notably those among robbers and murderers — that we may want to dispose of rather than improve. We may want /ess cooperation (and trust) rather than more, especially among those who are threatening us, and whose cooperation is a hindrance to ours.

The condition of ignorance or uncertainty about other people's behaviour is central to the notion of trust. It is related to the limits of our capacity ever to achieve a full knowledge of others, their motives, their responses to endogenous as well as exogenous changes. Trust is a tentative and intrinsically fragile response to our ignorance, a way of coping with 'the limits of our foresight' (Shklar 1984: 151), hardly ever located at the top end of the probability distribution. If we were blessed with an unlimited computational ability to map out all possible contingencies in enforceable contracts, trust would not be a problem

XREF: Connects to trust literature and the distinction between trust and enforceable contracts — relevant to institutional economics and game theory.

The rulers of a slave society — assuming that they do not mind what slaves think — can restrict their trust in the slaves and in the viability of their society to the belief that the slaves are not going to commit mass suicide. They simply trust to the fact - not invariably borne out by historical evidence — that most humans, even under extreme conditions, have a preference ordering which ranks life before death. Here, trust must be understood in the limited sense of trusting the effectiveness of coercion as a motive for cooperation

The freedom of others, however, is not by itself sufficient to characterize the conditions in which the issue of trust arises. Our relationship with people who are to some extent free must itself be one of limited freedom. It is a freedom in the sense that we have to have a choice as to whether we should enter into or maintain a potentially risky relationship: it must be possible for us to refrain from action. If it were only others who enjoyed freedom, while we had no alternative but to depend on them, then for us the problem of trust would not arise: we would hope rather than trust

DEFINE: Clarifies that trust presupposes mutual limited freedom — the option to refrain, otherwise one hopes rather than trusts.

In conclusion, trusting a person means believing that when offered the chance, he or she is not likely to behave in a way that is damaging to us, and trust will typically be relevant when at least one party is free to disappoint the other, free enough to avoid a risky relationship, and constrained enough to consider that relationship an attractive option.

If the pressure to act is great even when the trust threshold is lower than 0.50 — when we verge, that is, on distrust — the tension between action and belief can generate, by means of wishful thinking and the reduction of cognitive dissonance, a deceptive rearrangement of beliefs. Thus there are those who distrust entire categories of people except the member of that category with whom they have a special relationship.

Confidence, in the sense defined by Luhmann in this volume, might be described as a kind of blind trust where, given the constraints of the situation, the relationships we engage in depend or are seen to depend very little on our actions and decisions. In other words, confidence may also issue from wishful thinking and the reduction of cognitive dissonance; it would then be more akin to hope than trust. We still know little, though, about how a certain level of trust is or can be achieved and promoted.

DEFINE: Distinguishes Luhmann's sense of confidence from trust, characterizing it as blind, hope-like, and grounded in wishful thinking and cognitive-dissonance reduction. XREF: Connects to trust/risk literature (Luhmann) and could tie to cognitive dissonance theory from social psychology.

Indeed, some amount of trust must be present in any complex economic system, and it is far from inconceivable that systems with a higher level of general trust could come about. It would be risky, however, to make higher levels of trust into a cornerstone of economic reform. We may hope that trust will come about as the by-product of a good economic system (and thus make the system even better), but one would be putting the cart before the horse were one to bank on trust, solidarity and altruism as the preconditions for reform.

QUESTION: This raises the question: if trust follows good institutions, what first enables the institutions to be good without preexisting trust? A bootstrap problem worth chasing.

The first is that trust is not scarce in the sense of a resource that is depleted through use; the second that although it is often a by-product, this is not always so; and the third that there are extremely important cases where self-reinforcing arrangements acting on interests are either too costly (or unpleasant) to implement, or unavailable in the first place because trust is in excessively short supply.

XREF: Challenges common economic assumptions about trust as a scarce good; connects to debates about social capital and institutional design.

The most economical strategy of all is not that of not having to bank on trust, but that of not even having to bank on manipulating cooperative arrangements, on the assumption that rational — in the sense of optimal - cooperation evolves by itself.

QUESTION: What conditions allow rational cooperation to evolve by itself rather than requiring deliberate structuring?

Take, for instance, the cooperation of 'live and let live' which flourished between enemy soldiers in the First World War (Axelrod 1984). This might be accounted for in various ways. It may have arisen as the result of a soldier shooting — out of distraction, boredom, or nervousness -— at some clearly non-human target in the opposite trench. Or it may be that soldiers on both sides stopped shooting at regular intervals because they happened to have their meals at the same hours of the day. Such random signals may eventually have been 'interpreted' by one side as an inclination on the other towards an implicit truce, and they may have responded with other signals; first just to test for possible misunderstandings, then with increasing conviction, until the exchange slowly assumed the features of a stable cooperative abstention from mutual injury.

The probability that the other party will not act ina harmful way is raised by the understanding that mutual interest makes defection costly enough to be deterred. The soldiers' belief that they can trust each other results from an inference concerning the effectiveness of their interests as a motive for rational action in initiating and maintaining cooperation. Thus p is raised in the ' course of cooperation itself, without any assumption being made as to prior levels: 'The cooperative exchanges of mutual restraint actually changed the nature of the interaction. They tended to make the two sides care about each other's welfare' (Axelrod 1984: 85).

XREF: Echoes game theory and Axelrod's iterated prisoner's dilemma, where repeated cooperation fosters trust without prior assumptions. Builds on: "Optimal cooperation emerges without deliberate manipulation of arrangements."

Hume made the case for this process absolutely clear: When each individual perceives the same sense of interest in all his fellows, he immediately performs his part of any contract, as being assur'd that they will not be wanting in theirs. All of them, by concert enter into a scheme of actions, calculated for common benefit, and agree to be true to their words; nor is there anything requisite to form this concert or connection, but that every one have a sense of interest in the faithful fulfilling of engagements, and express that sense to other members of the society. This immediately causes that interest to operate upon them and interest is the first obligation to the performance of promises.

XREF: Connects to game theory and the prisoner's dilemma — mutual expectation of cooperation sustains social contracts. Also echoes Hobbes's account of covenants and Ostrom's work on commons governance.

Even discounting these, I would still argue that — with respect to the time span of reasonable relevance to any given generation — the spontaneous evolution of a cooperative equilibrium among humans is only just as likely as that of a noncooperative one, unless some restriction is imposed on agents' beliefs.

XREF: Connects to evolutionary game theory and questions about how cooperative equilibria arise without external constraints on beliefs — relevant to social contract theory.

If it is true that humans are characterized by a lack of fine-tuning and a tendency to go to extremes (Elster 1983), the assumption that trust will emerge naturally is singularly unjustified.

XREF: Connects to Elster's 1983 work on human nature and social theory, likely relevant to trust and cooperation debates in institutional design.

It would be best to have no memory,!® to forget that past defections may be repeated in subsequent moves. But in the case of individuals who do have some (finite) amount of memory, this is to prescribe as rational the avoidance of unconditional distrust even after the evidence would suggest it to be advisable. One inflicted 'wound', in this argument, should not be considered a sufficient reason for prolonged retaliation once the other side has been tamed by the first retaliatory response. Some degree of revealed preference for defection, in short, does not mean the absence of an interest in cooperation.

Thus tit for tat can be an equilibrium only if both players believe the other will abide by it, otherwise other equilibria are just as possible and selfconfirming. To show that trust is really not at stake, Axelrod should have shown that whatever the initial move and the succession of further moves, the game tends to converge on tit for tat. What he does do is express a powerful set of reasons why —- under certain conditions and even in the absence of trust generated by friendship, say, or religious identity — a basic predisposition to trust can be perceived and adopted as a rational pursuit even by moderately forward-looking egoists. We learn that, tentatively and conditionally, we can trust trust and distrust distrust, that it can be rewarding to behave as if we trusted even in unpromising situations.

Evolution has bestowed upon us the mixed blessing of being able to generate intentionally the as if behaviour. Knowing this, we can hardly avoid the responsibility of considering trust a choice rather than a fortunate by-product of evolution.

QUESTION: What practical tools or contexts make trust a deliberate choice rather than an instinct? Worth investigating the mechanics of intentional trusting behavior.

The strategy of economizing on trust does not of course imply that we should wait for cooperation to evolve by itself; it just claims that we should set our sights on cooperation rather than trust. We should, in other words, promote the right conditions for cooperation, relying above all on constraint and interest, without assuming that the prior level of trust will eventually be high enough to bring about cooperation on its own account. … However, attractive as this strategy may be, it begs the crucial question of what is to be done either when p is so low that conditions suitable for cooperation are not available in the first place, or when p is not high enough to sustain potentially beneficial cooperation where those conditions are too complex, costly, or unpleasant to be a conceivable alternative to trust.

What is at issue is not the importance of exploring in greater depth the causality of those forms of cooperation which are independent of trust

XREF: Relates to game theory and institutional economics discussions of cooperation without interpersonal trust, such as market-based or enforcement-based coordination.

It is constituent of the definition of personal bonds that — within the limits of character and skill — one trusts one's friends more deeply than strangers … If it were not, the motive for acting on the basis of personal bonds would vanish.

DEFINE: Defines personal bonds by their essential constituent of preferential trust, arguing this trust asymmetry is definitional rather than incidental.

I cannot will myself to believe that X is my friend, I can only believe that he is. Nor can they easily be manipulated in order to bring about mutual trust and fruitful cooperation: if X detects instrumentality behind my manifestations of friendship, he is more likely to reject me and, if anything, trust me even less

Builds on: "Personal bonds require trusting friends more than strangers."

Conceptual and theoretical arguments likewise suggest that the maintenance of trust via the extension of associations based on personal bonds might be seen to involve an element of rational pursuit. Trust, although a potential spin-off of familiarity, friendship, and moral values, must not be confused with them, for it has quite different properties (Luhmann, this volume). It is perhaps such confusion which has led to the conclusion that we should treat it as both an eminently scarce resource and purely a by-product. This does not take adequate account of our ability to act, to simulate, try out, learn, apply and codify signals and practices which may initially be predicated on unintentional states, but which could be duplicated in the as if behaviour form far beyond their source.

DEFINE: Distinguishes trust from familiarity, friendship, and moral values, which it is often confused with.

In contrast, deep distrust is very difficult to invalidate through experience, for either it prevents people from engaging in the appropriate kind of social experiment or, worse, it leads to behaviour which bolsters the validity of distrust itself … Once distrust has set in it soon becomes impossible to know if it was ever in fact justified, for it has the capacity to be self-fulfilling, to generate a reality consistent with itself. It then becomes individually ‘rational’ to behave accordingly, even for those previously prepared to act on more optimistic expectations … Only accident or a third party may set up the right kind of ‘experiment’ to prove distrust unfounded

DEFINE: Explains the mechanism by which distrust becomes self-fulfilling: it prevents the very experiments that could disconfirm it. SEED: Could ground an essay on how epistemic self-fulfilling prophecies create traps in relationships, organizations, or politics.

Trust is a peculiar belief predicated not on evidence but on the lack of contrary evidence — a feature that (as Pagden (this volume) shows) makes it vulnerable to deliberate destruction. In contrast, deep distrust is very difficult to invalidate through experience, for either it prevents people from engaging in the appropriate kind of social experiment or, worse, it leads to behaviour which bolsters the validity of distrust itself (see my paper on the mafia, this volume). Once distrust has set in it soon becomes impossible to know if it was ever in fact justified, for it has the capacity to be self-fulfilling, to generate a reality consistent with itself. It then becomes individually 'rational' to behave accordingly, even for those previously prepared to act on more optimistic expectations) Only accident or a third party may set up the right kind of 'experiment' to prove distrust unfounded (and even so, as Good argues in this volume, cognitive inertia may prevent people from changing their beliefs).

XREF: Echoes of self-fulfilling prophecy dynamics familiar in economics (Akerlof) and sociology (Merton). SEED: Potential essay angle: how self-fulfilling distrust explains persistent institutional breakdown or legacy of corruption in organizations.

The concession of trust, that is, can generate the very behaviour which might logically seem to be its precondition. … Secondly, if behaviour spreads through learning and imitation, then sustained distrust can only lead to further distrust. Trust, even if always misplaced, can never do worse than that, and the expectation that it might do at least marginally better is therefore plausible.

Builds on: "Distrust becomes self-fulfilling and self-justifying through its own effects"