Berkman Center for Internet & Society
Current Topic
Lecture Hall
Change Password
Forgot Password

by Cynthia Hardy, Nelson Phillips & Tom Lawrence


The literature on trust is burgeoning and the topic has been considered by economists, psychologists, and sociologists, as well as management theorists (see Sitkin and Roth 1993; Hosmer 1995; Mayer, Davis, and Schoorman 1995; and Kramer and Tyler 1996 for reviews). While there is agreement on the importance of trust, there is little consensus on how to define it (Mayer, Davis, and Schoorman 1995; McAllister, 1995). Consequently, trust has been equated with individual expectations, interpersonal relations, economic transactions, and social structures (Hosmer 1995; McAllister 1995). In this section, we will develop a communicative approach to examine trust in the context of interactions between organizations. We commence by critiquing two commonly used definitions of trust found in the literature on interorganizational relations: trust as predictability and trust as goodwill (see Zucker 1986; Ring and van de Ven 1992).

2.1 Trust as Predictability

The first definition relates trust to predictability, which is defined as the probability with which an actor assesses that another actor will act in a certain way (e.g. Luhmann 1979; Lewis and Weigert 1985a; Zucker 1986; Gambetta 1988). Luhmann (1979), for example, argues that we make pre- dictions (or have expectations) concerning the behaviour of others. If we are confident that our predictions will come to pass, we trust them: Trust exists in a social system in so far as the members of that system act according to and are secure in the expected futures constituted by the presence of each other or their symbolic representations. (Lewis and Weigert 1985a: 968, original emphasis removed)

Trust reduces complexity by ensuring that the social system is based on mutual expectations about actors' future behaviour, encouraging social actors to select specific options of social action and reaction. The basic function of coordinating social interaction is thus achieved and cooperation is the result (see Lane and Bachman 1996).

This approach to trust often ignores the fact that power is, in fact, a functional equivalent of trust, serving also to ensure predictability in coodination (Lane and Bachmann 1996). When writers identify interdependency between partners as a factor in trust, they need also to consider relative dependency and its power implications. For example, Provan and Skinner (1989) found that high dependency of individual farm-dealers on a single equipment supplier reduced opportunistic behaviour by the dealer, increasing the likelihood of co-operation. This situation conforms to the definition above: the suppliers could be confident that the dealers would act according to expectations. But it is not clear whether the dealers co-operated due to a trusting relationship or because they had no choice (i.e. supplier power forced them to). Hasenfeld and Chesier (1989: 515) draw the former conclusion regarding the dependency relationship between social service providers and their clients: Functional trust emerges when there is considerable interdependence, and thus some degree of power balance between the service provider and the client, nor does it neutralize power's advantage; rather trust is a derivative of power-dependence relations.

The situation is further complicated when writers turn their attention to the creation of trust, usually through the institution of relevant procedures (Zucker 1986), or the display of the appropriate symbolic representation (Lewis and Weigert 1985a). Such actions could equally well be interpreted as the management of meaning (Pettigrew 1979; Phillips and Brown 1993). Where power is used unobtrusively to ensure the co-operation of others (e.g. Hardy 1985a). For example, the idea that the symbols employed by the medical profession create trust in the minds of patients has also been seen as a strategy by doctors to retain power over their clients (e.g. Barber 1980). Similarly, definitions that focus on trust as confidence in the reliability and integrity of another actor (e.g. Rotter 1967; Moorman, Deshpande, and Zaltman 1993) are problematic because reliability and integrity may be a facade if predictability of behaviour is the only condition.

In other words, many of the complexities of interorganizational interac- tions may be resolved, not by trust, but 'by implicit or explicit powei relations among firms' (Granovetter 1985: 502): 'We've got to trust them' means in fact: 'We don't trust them but feel constrained to submit to their discretion.' This simply describes, of course, a power relationship. (Fox 1974: 95)

It is, however, important to differentiate between trust-based interaction, and power-based interactions since there is a substantive difference between cooperation achieved through power differentials that render some partners unable to engage in opportunistic behaviour, and a willingness to voluntarily sacrifice the benefits of opportunistic behaviour in order to cooperate with a trusted partner. While the first kind of co-operation may produce predictable behaviour, it will not produce the kinds of advantages that have been associated with trusting relationships. Therefore, a concept of trust that cannot distinguish cooperation derived from the political realities of powerlessness from co-operation based on reciprocity and good will; nor differentiate between the self-serving management of meaning to create a facade of trust and attempts to create shared meanings on which a trusting relationship can be built, is helpful to neither researchers nor practitioners.

2.2 Trust as Goodwill

This definition argues that trust is more than predictability: it also includes goodwill - mutual expectations of reciprocity - between partners (e.g. Ring and van de Ven 1992) which, in turn, lead to co-operation rather than conflict or opportunistic behaviour. According to Zucker (1986), this definition draws (directly or indirectly) on the work of Parsons (1969) whose conditions for the creation of trust included the existence of common values that can be translated into common goals. Trust is generated by realizing these conditions through relevant symbolic representations (Lewis Weigert 1985a; Zucker 1986).

While the introduction of the condition of goodwill avoids conflating efforts to create trust with self-interested attempts to manage meaning; many writers subscribing to this definition assume the ex ante existence of convergent goals or a sense of community (e.g. Ouchi 1980; Ring and vai Ven 1992). The question remains, however, whether such assumptions are useful in the context of complex interorganizational relations in which partners may have different values, pursue different goals, be subject to pressures from different constituencies, and are socialized into differ cultures (e.g. Rainey 1983; Waddock 1989). In this situation, it is more likely that stereotypic views of the 'other' will inhibit the creation of community(Sebring 1977; Gray 1985), making it unlikely that common goals will pre-exist and making it difficult to develop them. Moreover, in new relationships between different organizations, the norms and values about what constitutes trustworthy behaviour often differ markedly between partners. In such circumstances, how does goodwill arise and, if it does, how can it be signaled to other partners?

Many writers have avoided this issue of building trust altogether, arguing that it arises either spontaneously or not at all (see Sabel 1993). Other writers, drawing on the work of Granovetter (1985), link trust to the social relations within which business transactions are embedded. In other words trust exists as a result of frequent interactions and previous trusting relationships (e.g. Dasgupta 1988). Reliance on trust by organizations can be expected to emerge between business partners only when they have successfully completed transactions in the past they perceive one another as complying with norms of equity. (Ring and val Ven 1992: 489) Butler and Gill (1995) argue that this approach tends towards the tautological, begging the question: how do organizational partners achieve a trusting relationship in the first place? Furthermore, even if we accept that trust builds slowly through repeated interactions, we are left with the question of what is it that changes through multiple interactions to create trust? In other words, what is the foundation of a trusting relationship?

In summary, these two common definitions of trust - trust as predictability and trust as goodwill - fail to address the complexity of trust in interorganizational relationships. The first definition is problematic because, although it acknowledges the possibility of conflicting interests, its emphasis on predictability makes it difficult to distinguish between attempts to reduce power differentials and attempts to exploit them. The second definition hinges on the existence of common goals or a sense of community that does not always characterize the situations in which interorganizational trust is supposed to develop. Finally, neither definition has much to say about how to create trust. To address these problems, we turn to the conceptualization of trust as a process of communication.


We believe that trust is a meaningful concept only when conditions concerning both predictability and goodwill are met, and when attention is paid to the communicative foundation of trust. Trust can be said to exist between partners when relations involve a high degree of predictability, on all sides, that the others will not engage in opportunistic behaviour. This definition is not unlike that of Sabel (1993) who sees trust as the mutual confidence that no party will exploit the vulnerabilities of another (also see Gulati 1995). We, however, place power at centre-stage by arguing that trust rests on reciprocal communication, and does not involve communication undertaken in order to sustain asymmetrical power relations or to exploit a position of power. In this way, predictability arises from shared meaning; while goodwill arises from the participation of all partners in the communication process whereby this shared meaning is created. Thus, we can distinguish between partners who are trustworthy and those who merely claim to be so while manipulating the confidence of the other.

In an interorganizational relationship, trust grows out of a communication process in which shared meanings develop to provide the necessary foundation for non-opportunistic behaviour. Accordingly, trust can be conceptualized as a communicative, sense-making process that bridges disparate groups (e.g. Lewis and Weigert 1985a,b; Zucker 1986; Sabel 1993). This approach emphasizes the shared meanings that partners use to signal trust and trustworthiness to each other. Sociologically, trust is conceptualized as a reciprocal orientation and interpretive assumption that is shared, has the social relationship itself as the object, and is symbolized through intention action. Sociological trust is not derived from, nor reducible to the psychological states of atomistic individuals. (Lewis and Weigert 1985b:456) Trust is therefore an intersubjective social 'reality' that cannot exist, regardless of the good intentions of partners, unless the symbols used to signal trustworthiness have meaning for all parties.

Many writers have shown how social action hinges on interpretive, communicative, sense-making activities (Berger and Luckmann 1967; Geertz 1973; Schutz 1970), as meaning is created through specific communicative events, or conversations (Collins 1981). These conversations invoke a common reality or 'myth' (Collins 1981) which promotes action based on the tacit understandings and meanings that underlie it. In the case of trust, a myth is created which facilitates the sharing of information, subtle reading of signals, and informal interactions that signal trustworthiness and which, in turn, lead to predictable social action; reduce the likelihood of conflict and opportunistic behaviour; and obviate the need for more formal, bureaucratic controls (e.g. Ouchi 1980; Granovetter 1985; Bowen and Jones 1986; Gambetta 1988).

Such a process of myth creation is, however, problematic for two reasons. First, the creation of a common myth is difficult when symbolic meanings are not consistent between individuals with different values and back- grounds, or when intended symbols have no meaning at all. The task of finding the appropriate presentational signals of trust may not be easy, particularly when a relationship is new and involves disparate groups. To build trust between partners who share no common meanings, partners must create shared meaning where none existed before.

Second, such processes of communication run the risk of being 'hijacked' by those who would manage meaning for their own vested interests (e.g. Pettigrew 1979; Pfeffer 1981; Phillips and Brown 1993; Hardy 1985a). In other words, the claims made in these conversations may be systematically distorted in ways that produce domination. When organizations or politics are structured so that their members have no protected recourse to checking the truth, legitimacy, sincerity, or clarity claims made on them by established structures of authority and production, we may find conditions of dogmatism rather than of social learning, tyranny rather than authority, Manipulation rather than co-operation, and distraction rather than sensitivity. In this way, critical theory points to the importance of understanding practically and normatively how access to, and participation in, discourses, both theoretical and practical, is systematically structured. (Forester 1989: 239-40) In this case, the myth that is created is one of 'spurious' trust (Fox 1974): conversations have been deliberately exploited to convey a facade of trust, without any consideration of reciprocity, but in order to consolidate power. Spurious trust may produce predictable behaviour, reduce opportunism, and result in co-operation since, as we have already argued, trust and power are functional equivalents. But these advantages are one-sided, accruing to the dominant partner only and, as a result, the more creative outcomes associated with trust are unlikely to ensue.

To guard against this situation, Habermas (1984) recommends that those engaged in dialogue strive for a communicative ethic. In practical terms (see Keller 1981; Payne, 1991) it means that conversations should ideally include all stakeholders, on an equal basis, with the freedom to represent their interests and participate in a fair and open dialogue, unfettered by coercion, manipulation, secrecy, concealment, or deception. More powerful stakeholders should be responsive to the arguments and interests of less powerful counterparts (Bowen and Power 1993). They should not appeal to a priori principles but foster a truly dialogic process in which all participants contribute equally (e.g. Freire 1992). Such [high trust] relations do not exclude the challenge of plans by those affected; their submission of new information; their querying of the wisdom of means. But the shared purpose converts what would otherwise be power bargaining based on distrust into problem solving based on trust. During the 'working through' process which clarifies the available acts and probable consequences the participants trust each other and decide courses of action to which they commit themselves. Such a pattern of relations represents a high order of achievement in human affairs. (Fox, 1974:358) Synergistic outcomes are thus produced through the contribution of all participants in a communicative process that builds the necessary foundation for a trusting relationship.

The discussion so far raises a number of important points. First, trust results from a communicative process in which shared meanings either exist, or are created through a reciprocal relationship that involves all partners. Second, the spontaneous emergence of trust is unlikely in new relationships between disparate groups because such shared meanings will not exist. Third, trust can be built by using symbolic power to create shared meanings. Fourth, such management of meaning need not necessarily be undertaken with reciprocity in mind and may, on the contrary, be intended to maintain or increase power differentials. In such a case, it is not trust, but a facade of trust that results. Finally, the co-operation and predictability associated with trust is not always the product of voluntary self-sacrifice, but can also result from a lack of alternatives that follow from being the weaker partner in a relationship of asymmetrical power. While the use of power (e.g. the control of resource dependencies, the manipulation of participation, and the management of meaning; see Lukes 1974; Clegg 1989; Hardy 1994a) may increase predictability and serve the purposes of the dominant partner, it also reduces synergy. Consequently, it is useful, analytically and practically, to distinguish such cases from situations where all partners are willing to co-operate voluntarily and contribute to synergistic outcomes (Fox 1974; Lewis and Weigert 1985b; Sabel 1993). In the following section, we explore these issues further by contrasting two empirical examples of interorganizational co-operation.

Berkman Center for Internet & Society