One cannot trust chaos. If nothing connects with anything else or everything with everything else, it becomes impossible to build generalizations. In other words, a single system cannot, by itself, generate higher generalization or trust. Their accomplishment presupposes an environment which already possesses structure, though not the same degree of order, and the same limitations on what is possible, as the system itself. But what structures the environment is none other than the existence of other systems in the environment.
It is through systems of a special kind, namely human beings, that there comes into the world that enlargement of complexity on which trust is focused – this is freedom of action. It is not surprising, therefore, that trust is extended first and foremost to another human being, in that he is assumed to possess a personality, to constitute an ordered, not arbitrary, centre of a system of action, with which one can come to terms. Trust, then, is the generalized expectation that the other will handle his freedom, his uncanny potential for diverse action, in keeping with his personality – or, rather, in keeping with the personality which he has presented and made socially visible.1 He who stands by what he has allowed to be known about himself, whether consciously or unconsciously, is trustworthy.
For, in addition to its immediate significance as regards situation and purpose, every socially comprehensible action also involves the actor’s presenting himself in terms of his trustworthiness. Whether or not the actor has this implication in mind – whether he is aiming at it, or consciously disclaiming it – the question of trust hovers around every interaction, and one’s self-presentation is the medium by which decisions about it are attained. Every actor is experienced by the others, not only as a causal process, by which causes are transformed into effects, but simultaneously (so as to extend some generalizing control over this process) as a complex of symbols. And he senses this, knows the symbolic implications and expressive value of his every action and inaction, for the most part, much better than he knows their effects and so, more or less consciously, gauges his behaviour accordingly.2 His motives may be of widely different kinds – he may be concerned to appear trustworthy; he may make an effort to remain true to himself and to live in a way which affords him self-respect; he may act spontaneously, according to the facts of the situation and, to that extent, in a socially naive way, by allowing his personality to function as an unconscious selection mechanism.3 In every case, despite the variety of possible structures of motivation, a similar result is forthcoming – a selective presentation of self, which provides other criteria on which to build up trust and to establish norms of anticipated continuity. To some extent, therefore, foundations of trust in the social order can be formed largely independent of the fluctuations in, and differences between, individual motives.
Since all communication, indeed every perceivable form of behaviour, says something about the person who is behaving, communication – even merely being seen by others – is a risky undertaking which requires some kind of safeguard. An individual’s behaviour always gives away more information about himself than he can reconcile with his ideal self and more than he consciously wants to communicate. Thus, his mere appearance presumes some minimum trust, trust that he will not be misinterpreted, but that he will be accepted by and large in the guise he wishes to appear. There are people who experience this prerequisite site of trust to such an extreme degree that they have difficulty merely being – let along doing anything – in the presence of other people. Their sphere of action is limited in proportion to their lack of trust and, furthermore, inability to show trust limits their chances of winning trust. The possibilities for action increase proportionately to the increase in trust – trust in one’s own self-presentation and in other people’s interpretation of it. When such trust has been established, new ways of behaving become possible – jokes, unconventional initiatives,4 bluntness, verbal short cuts, welltimed silences, the choice of delicate subjects etc. When trust is tested and proven in this way, it can be accumulated by way of capital.
So far as personal trust is concerned, therefore, the foundations of trust in a society are adjusted according to the prospects and conditions for self-presentation and the tactical problems and dangers involved in it. This mechanism transforms societal-structured conditions into sources of trust, which means that it is not to be viewed and understood merely in terms of personal strategies, as it is by Goffman. The individual personality’s system of reference is only one possible area of interest for functional analysis. On the other hand, strategic consistency is one indispensable element in the total context and the extent of the functionalization of self-presentation, the extent of psychological sensitivity, and the extent to which people are prepared to cooperate tactfully in delicate areas of self-presentation. These aspects are important variables regarding the conditions of trust in society. This determines the style in which personal trust can be created and maintained.
We cannot go here into the varied manifestations relevant to this context. We cannot even produce a classification or survey. A few short examples of analysis are all that is possible. Limiting ourselves to the question of learning, and leaving aside the extensively researched problems of infant socialization, we will consider a few of the social conditions which make it easier to bring about relationships of trust.
The question of the formation and consolidation of personal trust enables us to transfer our abstractly formulated problem of the reduction of complexity to the dimension of time, and to show that the problem of forming trust relationships must be solved gradually, step by step. The ‘principle of small steps’5 replaces simpler forms of adaptation to the environment when the environment also operates in a contingent fashion or is too complex for adaptation at one stroke. For these, systems need time.
It is a necessary precondition that the situation permits selective steps, meaning behavioural choices, and that behaviour is not already determined either institutionally or historically. Thus the first basic prerequisite for building up personal trust is that human actions are perceived in general as personally determined.6 Trust is founded on the motivation attributed to behaviours. Conduct on the experience of which trust depends must appear as an expression and reaffirmation of personality. However, only such actions as are treated institutionally as ‘free’ are imputed to personality.7 Firstly, there is freedom, in the, as it were, presocial sense of other people’s uncontrollable potency to act – the source of the need for trust. Secondly, there is institutionalized freedom, namely freedom bound up with and moderated through the social order, freedom as a complex of actions or aspects of actions for which one is personally responsible, which is the source of the ability to learn trust. In order for trust to emerge and fulfil its function, freedom must be transferred from one form into the other.
The question of which actions or aspects of actions are considered significant as expressing personality is less a matter of pure causality than of what clarifies, normalizes or delimits the causal social context. Social expectation determines whether or not actions are attributed to personality. The institutionalization of such expectation, however, is a selective, simplifying process which reinforces what is selected. A few examples will serve to clarify this.
An action will not be reckoned as attributable to personality if it is recognized as the result of a direct instruction from a superior (even if the instruction amounts to putting a signature to a proposal made by the subordinate). Therefore, subordinates who want to show themselves to be trustworthy must strive to exhibit the utmost industriousness, conscientiousness and readiness to carry out tasks loyally, well beyond what is customary practice. And, on the other hand, the same reason gives rise to the opposite strategy, namely, of rising to higher positions in order to reach a status in which, according to accepted opinion, one is regarded as personality-visible and can make free decisions (even if structure and circumstances virtually predetermine those decisions).8 In general, roles, not causal laws, serve in social life as the basis for judging any behaviour to be voluntary, i.e. intentional, or involuntary, i.e. unintentional.9 Further to this, the outcome of any complex technological process – and this includes what results from actions by superiors – appears to be relatively impersonal. The greater the combination of recognizable causes, the more difficult it becomes to isolate who originated the action. Who was responsible for Hiroshima? Appraising the behaviour of motorists clearly reveals a combination of complex mediate effects, the visible demands of circumstances, and personally attributable behaviour. A mechanically generated high speed, although responsible for everything, is not a matter of personal choice but ‘normal’ practice. On the other hand, unexpected reactions are attributed to the person because of the considerable demands made on concentration and skill. Acting according to the norm is usually inconspicuous and weak in expression, and therefore is not a suitable base from which love and trust can be generated.10 What proof of being deserving of trust can anyone give by saying that he has never been in prison? Against this, deviant behaviour, initiative or criticism are attributed personally. They can only occur if the social order can at the same time offer security, for such behaviour assumes the existence of relationships of trust and strengthens them by utilizing them. In market exchange, where deals can be concluded immediately because of the guarantees afforded by financial and credit systems which bridge differences in time, it is true that personal interests are reflected in the choice of goods, but such deals do not reflect character traits relevant to the question of trust. The same does not apply to presents and favours, which call for expressions of thanks; they arise spontaneously, and run the risk of not being reciprocated.
In a situation which is relatively open-ended, double contingent, and where both parties can operate selective choices which influence one another, it is possible to set in motion a process which forms trust by tackling the problem of the reduction of complexity over several discrete steps, i.e. sequentially.11 The process itself has, on the one hand, a framework of conditions by which it is made possible and, on the other, requirements for steps to be taken to strengthen the process of mutual selection which brings trust into being. We must maintain the distinction between conditions of possibility and conditions of realization, at least for purposes of analysis.
First of all there has to be some cause for displaying trust. Some situation has to be defined in which the person trusting is dependent on his partner, otherwise the problem does not arise. His behaviour must then commit him to this situation and make him run the risk of his trust being betrayed. In other words he must invest in what we called earlier a ‘risky advance’. One fundamental condition is that it must be possible for the partner to abuse the trust; indeed, it must not merely be possible for him to do so but he must also have a considerable interest in doing so.12 It must not be that he will be likely to toe the line on his own account – in the light of his own interests.13 In his subsequent behaviour the trust put in him must be honoured and his own interests put to one side. For this shelving of interests to serve as confirmation of trust, it must attain a certain level – it must present itself as a sort of missed opportunity, and not merely as the temporary postponement of a betrayal of trust. All these elements in combination might constitute the first sequence in building up trust. From this we can already draw one important conclusion – that the process demands mutual commitment and can only be put to the test by both sides becoming involved in it, in a fixed order, first the trusting and then the trusted person.
However, we have as yet only covered one initial possibility, and, even then, only the course it outwardly follows. There are, in addition, cognitive and normative aspects which cannot be arranged at will into some arbitrary sequence but must follow a definite order. It does not suffice for the process simply to take some particular course. The participants must know the exact situation and they must know from one another that each knows it. The building up of trust therefore depends on easily interpretable situations and not least, for that reason, on the possibility of communication.14 In some cases a post hoc inference will suffice; it might even be that ignorance of particular aspects does no harm, i.e. does not obstruct progress towards trust, but this can hardly be decided without empirical research. However, the process may become more liable to break down where relevant circumstances are not known, even if the requisite behaviour has been followed in the initial phases. It can also become more liable to break down through an excess of knowledge, such as when the participants mutually infer that the process is being employed in order to build up trust. For in such instances, motives are unavoidably put in question, and such questioning can easily turn into distrust.
Finally, we must add normative requirements to the cognitive ones we were able to deal with only in rather vague terms. Questions of the ethics of trust will not be dealt with until Chapter 12. Here we are only concerned with the questions whether and how the process whereby trust is generated can become the object of norms. This brings us up against a difficulty of a rather special kind. As we have seen, at the start of the process lies a risky advance, which because of its wagerlike nature does not lend itself easily to being disciplined by norms, corresponding in this sense to heroic or saintly acts. The same thing can be said of the act whereby trust demonstrated by one individual evokes a corresponding response on the part of another. In both cases the formulation of norms would simply shift the problem to another level, with no serious guarantee that trust would actually ensue. It is not possible to demand the trust of others; trust can only be offered and accepted. It is for this reason that demands to trust cannot start trust relationships off. This can only be done by what we have called an advance. The initiator may confer trust, or perhaps utilize an opportunity arisen by chance, to show himself trustworthy (for instance by returning a lost object which he has found). The truster sees in his own vulnerability the instrument whereby a trust relationship may be created. Only his own original trust offers him the possibility of putting it forward as a norm that his trust is not to be disappointed, and thus bringing the other over to his side. However, the impossibility of formulating trust in a normative sense does not totally exclude it altogether from the normative sphere.
We can use an old concept and qualify this peculiar advance with normative consequences we have been talking about as a supererogatory performance. One may call supererogatory a performance which does not flow from some previously assumed duty but manifests itself as meritorious, and attracts respect. In spite of recent attempts,15 the relationship between such a performance and duty remains logically and analytically obscure. However, what is of interest to us here is purely its function in the genesis of trust. It entails somehow an excess on the normative which has some parallels with the excess on the information discussed previously; it constitutes a surplus performance, which nobody can require, but which (for the self-same reason) engenders claims – as benefactions for instance engender a claim to gratitude. Trust relationships are not imposed, but are normalized subsequently. Thus, the function of this supererogative can be seen in the fact that it transforms conditions for emergence into conditions for maintenance. And this is exactly what is required if trust is to emerge.
If supererogatory qualities function as a rule for generating claims which can become the object of norms, they cannot, of course, be brought into play arbitrarily, but rather only where such claims are acceptable. Help and gratitude are one case; trust another. Obviously such limitations cannot entirely exclude abuse. One can create a bond by showing trust, as one can by giving a present.16 The only way to counteract this is at the beginning, when the relationship is still at the voluntary stage, by refusing more or less tactfully to participate in the process.
We are, however, still at the beginning. We have described one typical basic unit in the process of building up trust. Usually several are involved. For in the first round no one plays for high stakes – neither with great personal commitment nor against powerful interests. Deception and deviant behaviour on both sides are instances which illustrate very aptly the sequential character of the general process, its problems and its tactics.
A personal relationship on the basis of mutual favours usually opens with small-scale activities. Kindnesses, offers of help, and small gifts which cost nothing are proffered in a form which leaves room for tactful rejection. The relationship can only be deepened when acts of friendship have been reciprocated, when there is a spark of grateful recognition, and when the relationship has stood the test of everyday coming and going. It can then involve larger gifts, and, because each person trusts the other, can sustain a long, drawn-out imbalance. At this stage of development, immediate and exactly gauged reciprocation becomes a sign of distrust, because the person acting in this way presents himself as someone wanting to extricate himself as soon as possible from the bonds of gratitude. Such behaviour makes it impossible for the donor to display trust and for the recipient to prove himself when burdened by trust in the manner suggested.
A similar deepening of trust, demanding even greater delicacy, sensitivity and tact, can be observed when illegal – or semi-legal – matters are concerned.17 Where involvement in espionage, in homosexual relationships or in drug traffic is concerned, or even in the case of far less extreme instances – such as when secretaries exchange indiscretions, or newcomers are being initiated into the usages common in workplaces18 – a series of tests has first to be gone through without those involved being betrayed, allowing them to retreat indignantly behind the screen which behavioural norms provide, or to explain it away as some harmless misunderstanding. The true character of the relationship only unfolds slowly when the candidate appears to pass the tests – and for this he must himself carry out counter-tests or even involve himself to such an extent in the situation where he too would then be guilty. Once mutual trust has been safely established, it would be blatantly tactless – if not a quite disastrous lapse – if one of the participants wanted to return to the learning stage and to use the cautious strategies which were sensible at that early juncture.
If we may now generalize, it is clear that such learning processes are only complete when the person to be trusted has had opportunities to betray that trust and has not used them.19 It is impossible to eliminate this risk during the process of learning. It can, however, be divided up into small stages and thereby minimized. But there are two qualifications to be added to this rule.
One must not infer that during the learning process trust can grow steadily and expand uninterruptedly into more important and consequential matters. To that extent experimental research in social psychology, which is concerned solely with minimal and trifling risks, affords a misleading basis for generalization.20 Changes in what is at stake and shifts in the balance between anticipated advantage and possible loss build into the learning process thresholds which make for qualitative differences – for instance, when the person who trusts is risking not only himself but also others, his family, his work, or his country. In so far as this happens to be the case, there are additional conditions for consensus formation and responsibility adjustment. A second example concerns the opposite of the conditions necessary to build up trust. In completely risk-free role relationships (for example, if the participants are shielded from all personal consequences by their membership of an organization) there are hardly any available starting points for the development and stabilization of personal trust. In this regard even establishing acquaintanceship is severely limited.21 Although trust is even then an inescapable basis for the conduct of existence, it is not primarily in the form of personal trust. It seems that personal trust is only formed where it is needed. This is the case – even today – if the individual personality takes on a social-structural relevance, if the configuration of interaction relationships is arranged in the social system in such a way that it can only be arranged thus and now has to be restructured after that individual ceases to be involved. Equally characteristic of modern society are those borderline cases of personal trust which occur not at the level of large organized systems but within simple, everyday interaction systems. Such instances are frequently short, involve different participants, are impersonal and are unlikely to be repeated, but they nevertheless not infrequently involve considerable risk. A clear example is the problem of risks and trust involved in a taxi ride.22 Due to lack of time and lack of background knowledge the participants have to depend on highly standardized ‘tests’ of the normality of the situation – and on a sufficiently normalized environment which makes the risk, although serious, appear unlikely. In this context it is particularly worth noting that a breakdown in the environmental premises cannot be cushioned by building up trust within a small system, but at best involves reorganizing the way one approaches the taxi.
Questions of how far personal trust is still needed today, in which social systems, and in which functions, would make a subject for extensive empirical research. Such research would, I suppose, indicate very quickly that the need to orientate oneself to other people’s individual characteristics is as strong now as ever it was in all areas of social life where repeated contact takes place,23 and that the myth of mass society has its origins in some kind of optical illusion. On the other hand there is no doubt that modern differentiated social orders are much too complex for the social trust essential to ordinary living to be created solely by this type of orientation towards persons. It is all too obvious that the social order does not stand and fall by the few people one knows and trusts. There must be other ways of building up trust which do not depend on the personal element. But what are they?