Introduction

Across the management, social science, and business ethics literatures, and in much of the philosophy literature, trust is characterized as a disposition to act given epistemic states. Trust is said to be an attitude, inclination, or willingness to act and accept certain risks (the disposition), given a set of beliefs or expectations about the trusted party (the epistemic states)–beliefs or expectations that the risks of a trusting action will not materialize, and by extension, that the risks are justified by the potential benefit. Work in management draws on the sociology and economics literatures, where the cost/benefit aspect of these beliefs and expectations is emphasized (see, for example, Gambetta 1988a; Coleman 1990; Williamson 1993; Coleman’s work is discussed below). So-conceived, trust is a form of strategic behavior or rational economic decision making in situations that involve risk and vulnerability.

To take the most prominent example in the management literature, Mayer et al. characterize trust as, “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the [trustor’s] ability to monitor or control that other party” (1995, p. 712).Footnote 1 Mayer et al. present their work as an extension of Gambetta’s, who characterizes trust in explicitly cost/benefit terms: “When we say we trust someone or that someone is trustworthy, we implicitly mean that the probability that he will perform an action that is beneficial or at least not detrimental to us is high enough for us to consider engaging in some form of cooperation with him” (1988a, p. 217). In the business ethics literature, Brenkert characterizes trust in similar terms, as “an attitude, disposition, or inclination to act in certain ways in light of various beliefs one has both about oneself and others. Typically these beliefs concern one’s own vulnerability and the restraint the trusted agent is prepared to exercise not to take advantage of that vulnerability” (1998, p. 295).Footnote 2 Brenkert is not making a subtle distinction between attitudes, dispositions, and inclinations; his definition is meant to be broad, and differences between those three are not important in this context.

These characterizations are best thought of as epistemological because epistemic states—beliefs and/or expectations—distinguish trust from dispositions to act in other situations (for Brenkert), and from a general willingness to take risks and be vulnerable (for Mayer et al.). The epistemic states are therefore a necessary component of trust so-conceived, though note that as we use the term, “epistemic states” is broader than justified true belief. Moreover, trust so-conceived is not a moral concept and it is not (necessarily) part of a moral relationship, a point Brenkert makes explicit: “trust is not a principle, let alone a moral principle, but an attitude or disposition to behave and respond in certain ways, viz., to accept certain risks of harm or injury from another agent on the basis of a belief (for which there is some degree of uncertainty) that the other does not intend to do harm to one (or those one cares about), even though he/she could” (1998, p. 298). So the epistemological conception of trust is the amoral conception referred to in the title of this paper.

In the first section below, we argue that epistemological characterizations of trust prove to be conceptually inadequate.

In second section, we present and defend our alternative account of trust as a moral phenomenon: when A trusts B to do something, A invites B to acknowledge and accept an obligation. When—or if—B accepts the invitation, B takes on an obligation, and in that way trust creates an obligation and forms, or deepens, a trust relationship. When accepted, a trust-invitation therefore has a moral effect.

In the third section, we discuss the role of trust in business organizations. Ghoshal et al. (1995) characterize the task of managers in terms of cultivating trust. We show that the epistemological conception of trust cannot do the work Ghoshal requires, but our moral conception can support his project.

Epistemological Characterizations Cannot Account for the Possibility of Moral Wrong in Trust Relationships

In her early and very influential paper on trust, Baier distinguished between trust and reliance, claiming that trust can be violated or betrayed, but reliance can only be disappointed, “disappointed” in the sense that outcomes failed to match expectations, not with moral overtones, as when we say to a child, “I’m so disappointed in you” (1986, pp. 234–235). We will refine Baier’s distinction between trust and reliance below, and we will argue that Baier’s own characterization of trust cannot account for the possibility of betrayal, but we set those points aside for the time being. With Baier we take the possibility of betrayal to be a central, defining feature of trust. And though Baier does not make this further point explicit, betrayal is a moral wrong; it is the violation of a moral obligation or moral relationship; so in effect Baier claims that possibility of moral wrong is one defining feature of trust relationships. And with Baier, we take this to be a minimum requirement: any account of trust must be able to account for the possibility of moral wrong.

But Mayer et al.’s definition cannot account for the possibility of moral wrong. For them, A trusts B to do X means that A makes him- or herself vulnerable by depending on B to do X. Person A is willing to do so on the basis of three expectations: that B is willing to do X, that B is able and competent to do X, and that B is likely to do X even if A cannot monitor or force B to do so. But if B fails to do X, then one of A’s expectations was wrong, one of A’s expectations was disappointed. And even though one of A’s expectations was disappointed, we cannot say that B, the trusted party, has done something morally wrong in not fulfilling his or her obligation, and we cannot say that A was wronged—because there was no obligation. For this reason Mayer et al.’s definition of trust is conceptually inadequate. The same criticism applies to Brenkert’s view: person A could make him- or herself vulnerable given certain beliefs, and those beliefs could turn out to be wrong, but beliefs-being-wrong does not amount to betrayal. On both formulations, a willingness to be vulnerable given a set of epistemological states cannot account for the possibility of moral wrong, and so is not sufficient for establishing a trust relationship.Footnote 3

Consider this example: last year the present authors expected the Seattle Mariners to win the World Series, and made a bet based on this expectation. In doing so, we made ourselves vulnerable by depending on the Mariners. Applying Mayer, Davis, and Schoorman’s definition, we therefore trusted the Mariners. But this use of the term “trust” makes no sense (given the conceptual requirement just described): even though our expectation was disappointed, we cannot say that the Mariners did something wrong by not winning the World Series, and we cannot claim to be betrayed or wronged in the process (because there was no moral obligation or moral relationship, but we will come to this positive conception of trust in the next section). So making ourselves vulnerable given a set of epistemological states is not sufficient for establishing a trust relationship.

To make the critical point more concrete, consider the following well-known example, taken from Coleman (1990), who takes the example from Wechsberg (1966): a ship owner needed £200,000 to pay for repairs made at an Amsterdam shipyard. Without the funds the ship would remain docked in Amsterdam over a weekend, and this would cost the ship owner £20,000 in expenses for the crew. The ship owner called a merchant bank in London—Hambros Bank Ltd.—which immediately agreed to lend him the money, and arranged for payment through a bank in Amsterdam. Coleman asks “why the Hambros man placed trust in the ship owner on the basis of such insubstantial security” (1990, p. 93), as no contract was signed and there was no paper documentation of the transaction. According to Coleman, the banker extended credit, making himself and his firm vulnerable, on the basis of beliefs about the trusted party and about the potential profit. Trusting the ship owner was in effect a bet, which depended on the Hambros banker’s cost/benefit analysis: “If the chance of winning, relative to the chance of losing, is greater than the amount that would be lost (if he loses), relative to the amount that would be won (if he wins), then by placing the bet he [the trustor] has an expected gain; and if he is rational, he should place it. This simple expression is based on the postulate of maximization of utility under risk” (Coleman 1990, p. 99, Coleman’s emphasis).

Coleman here gives a thoroughly economic analysis of the transaction, and in the process he uses trust in the epistemological sense as a label for the banker’s decision. This example is more complex than the previous one because borrowers are generally thought to have obligations to pay back loans. This obligation would be in place, but our point is that the ship owner had no obligation to the banker as part of the trust relationship. The trust relationship as defined on the epistemological conception does not involve—in an intrinsic way—or even make reference to any sort of obligations, and the trust relationship does not create or put obligations in place. So if the ship owner chooses not to pay back the loan, we cannot say that the banker is betrayed or wronged; all we can say is that the banker lost his bet. On Coleman’s analysis the whole transaction is therefore no different from our bet on the Mariners to win the World Series.Footnote 4 At best talk of trust adds nothing, it merely serves as a label for acting on a calculated risk. But given Baier’s claim that trusting action requires the possibility of betrayal, then, further, Coleman is wrong to use trust in the epistemological sense, because trust so-conceived does not account for talk of betrayal from the perspective of the trustor (the banker).

To be sure, there is a further question about whether Coleman’s analysis of the transaction in economic terms accurately represents the behavior of the banker as a psychological explanation. The banker might have reasoned just as Coleman suggests, in which case talk of trust is inappropriate. Or, it is possible that the banker took the ship owner to be making a moral commitment to the bank in arranging for the loan, and the banker went ahead with the transaction because of that moral commitment. If so then talk of trust would be substantive and meaningful in this example, though we cannot develop this line of thought until our view is outlined in the next section. We will have this same question about psychological explanations any time behavior is explained in terms of trust, but this is not a dispute we need to settle; our goal in this paper is only to clarify what it means to say that trust was involved in an example of this sort, and so to distinguish between two possible accounts, only one of which—we argue—should be called trust.Footnote 5

The Moral Conception of Trust

Definition and Discussion

On our view, when A trusts B to do X, A invites B to acknowledge and accept an obligation to do X. When—or if—B accepts the invitation, B takes on that obligation. In that way trust creates an obligation and forms a trust relationship. This can occur in the context of some pre-existing relationship between A and B (a moral or non-moral relationship): in that context, trust-acceptance creates a new obligation and, as a result, deepens and possibly transforms the existing relationship. Trust-acceptance can also create obligations when there is no pre-existing relationship; we give an example of this below in our discussion of so-called swift trust. In both contexts we can think of trust-invitations as seeking out a moral relationship (we return to this idea in the next section), and in both contexts trust-acceptance has a moral effect.

Many if not most instances of trust-invitation and trust-acceptance are implicit, and trust-acceptance could occur in a number of ways. And, after the trust-invitation is accepted, B could fail to act on his or her obligation because of carelessness or indifference. Or B could accept the invitation and deliberately exploit A’s vulnerability. In both cases the obligation is violated, and in that way person A and the trust relationship between A and B are both betrayed—though we would most likely view the second case, deliberate exploitation, as more blameworthy. So our view can account for the possibility of betrayal, though if B failed to act on his or her obligation to do X for some good reason, we might not think B’s failure amounts to betrayal. Or we might admit degrees of betrayal.Footnote 6

This formulation emphasizes the relational aspect of trust. A more complete definition would also address the trustor’s decision to make this invitation and to rely on another, making him- or herself vulnerable. Holton (1994) emphasizes this other aspect, describing trust as something we can decide to do, though his account of what trust is, what we decide to do when we trust, is quite different from the view defended here. On Holton’s view, trust is a kind of stance we take: when we trust we rely on a person, and we “have a readiness to feel betrayal should it be disappointed, and gratitude should it be upheld” (1994, p. 4). As Holton acknowledges, this treats trust as a primitive, as a particular kind of reliance distinguished by our disposition to feel betrayed should the other party not act reliably. So Holton’s view offers no real analysis of trust, meaning that it offers no explanation of why we should feel betrayed in some situations of reliance but not others. Our project is very much to account for this sense of betrayal, to explain what makes it appropriate in terms of the obligations that trust-invitations and trust-acceptances put in place.

The moral conception of trust therefore expands—and clarifies—our understanding of what it means to be trustworthy. On the epistemological conception trustworthiness is something we assess: a person or organization is trustworthy if we believe that person or that organization is unlikely to exploit the trustor’s vulnerabilities. But when a trusted party betrays a trust relationship, we commonly think that party has done something wrong, not just something unexpected. The moral conception of trust can account for this moral evaluation of the trusted party’s behavior.

Although epistemological states are not sufficient for forming a trust relationship, even in conjunction with the relevant disposition, the epistemological states in question will most often be present as antecedent conditions or components of trust—meaning that those states will often support and be part of a willingness to trust or trusting action (on the qualification “most often” see footnote five). For example, I might trust a co-worker in the moral sense to do X when—or even because—I hold expectations about the coworker’s likely behavior. My willingness to trust will change as such beliefs and/or expectations change. Moreover, there will be a set of second-order expectations and/or beliefs on the moral conception of trust: the to-be-trusted party will hold expectations and/or beliefs about whether a particular trust-invitation was meaningfully offered, the trustor will hold expectations and/or beliefs about whether the corresponding trust-acceptance was sincere, and both parties will hold expectations and beliefs about the actual circumstances. These expectations and beliefs are second-order in the sense that they are about the trust-invitation and the trust-acceptance. These second-order beliefs and/or expectations will also be dynamic. Our characterization of trust therefore leaves room for expectations and beliefs to turn out to be incorrect, and this will most often make possible betrayal. So we are not claiming that there are no epistemological states involved in trust on the moral conception. But at the same time, our negative claim is that holding these epistemic states not sufficient for establishing a trust relationship; these states along with the disposition to make oneself vulnerable therefore do not explain what it means to trust; and the moral conception of trust presses us to expand our understanding of the epistemological states involved, to include assessments of the sincerity of the trust-invitation and trust-acceptance.

Moreover, even though we reject the epistemological characterization of trust as inadequate, work on the epistemological antecedents and components of trust is nevertheless important conceptually, in clarifying the expectations and/or beliefs that can support a decision to trust on our moral conception. So the existing literature on the epistemological conception of trust will complement our own conception by addressing such questions. And social science work on the epistemological antecedents and components of trust will also be quite important for practical reasons, for understanding the conditions that support trust in practice, especially in the economic context. For example, Burchell and Wilkinson (1997) conducted a set of surveys to determine how British, Italian, and German businesses decide whether to trust other firms, and to determine how those businesses try to establish themselves as trustworthy transaction partners. They identified a set of considerations that varied somewhat across the home of the businesses, considerations that included the presence of personal contacts, satisfaction with past performance, and long-term experience with the potential trading partner. We do not mean to suggest any criticism of this project, though we would reformulate the finding: the questions Burchell and Wilkinson asked concerned either (i) the evidence businesses use to determine whether to have confidence in their interaction with other firms, meaning the questions concerned the evidence businesses use in making decisions about the appropriate level of risk and monitoring. Or (ii) the questions Burchell and Wilkinson asked concerned second-order judgments about how transaction partners determine whether trust-invitations are genuine, trust-acceptances sincere, and whether the trusted party is competent to fulfill its obligation.

The first of these re-formulations eliminates talk of trust but nevertheless accommodates one central theme in the broader management and business literatures, namely that trust plays a central role in reducing transaction costs, because trust enables individuals and organizations to interact and trade without comprehensive legal protection, extensive monitoring and institutional support, or with less protection, monitoring and support—all of which are very expensive. On this re-formulation, Burchell and Wilkinson’s paper concerns the evidence firms need to make more accurate predictions about the likely behavior of potential transaction partners, more refined decisions about the appropriate level of legal and/or institutional protection, and perhaps more aggressive cost/benefit tradeoffs. Trust in the epistemological sense simply drops out of the analysis, it adds nothing. But as noted, this is not to deny the deeper theme, namely that transactions without or with less protection, monitoring, and support can bring substantial cost advantages. This point is very important conceptually: though we acknowledge the important role for work on the epistemological antecedents and components of trust, which we take to be consistent with and to compliment our own work, the danger here is that the social scientific account of trust crowds out or replaces normative explorations of the meaning of trust. The danger is that we come to think that the epistemological conception of trust is complete, and lose sight of the moral dimensions of trust relationships. The danger is that we take Burchell and Wilkinson—on the first formulation—to give us an empirical account of trust rather than an empirical account of the epistemological conditions that support (but do not define) trust.Footnote 7

Examples

Our moral characterization of trust is meant to be general and apply across both business and non-business contexts, though the particular obligations and the depth of the relationships will vary across cases. Co-workers trust one another because (or when) they are bound by a complex set of mutual invitations and mutual obligations, to contribute to joint projects, share both the benefits and responsibility if there are difficulties, etc. An employee trusts her manager when she—at least implicitly–invites her manager to acknowledge and act on an obligation to care for her well-being, where this goes beyond using the employee to serve a corporate function; we might think, or at least hope, that a manager accepts this invitation and takes on this obligation when the employee is hired (see for example, Soule 1998). In both examples, trust is a relationship that involves concrete obligations that the trusted party accepts, at least implicitly. Here consider a specific example involving managers and employees at a consulting company. An associate constructed an analytical model and developed a series of recommendations; the manager vetted and approved material; and then the material was shown to a senior partner, in preparation for a client presentation. The senior partner—clearly unhappy and even agitated—asked the associate, why did you do X? The manager then agreed with the senior partner, asking the associate, “Right, why did you do that?” This is a striking violation of trust: in accepting work assignments, employees rely on a commitment on the part of managers (and also co-workers) to take responsibility for their own decisions and for their part in a collective piece of work. The moral obligation and trust relationship here were largely if not entirely implicit, but the manager’s behavior was a betrayal of his obligations and a betrayal of the trust relationship. Here we can only talk of betrayal or violation if we think the manager and the employee were bound by moral obligations embedded in a working relationship. Otherwise, we could only say that the manager did not do as the employee expected, and if so, there is no basis for saying that anything wrong happened.

Applying the characterization offered here makes explicit the implicit obligations present in other kinds of examples. Taxi drivers trust customers to pay their fares, and the customers take on that obligation when they ask for a ride (Gambetta and Hamill 2005); at the same time the customers trust the drivers to take them directly to destinations, and drivers accept that obligation when they pick up riders. Customers trust businesses to communicate honestly about products, meaning that businesses accept an obligation to communicate honestly when engaging in transactions. Outside of these business examples, friends trust one another, children trust their parents to look out for their best interests, and so on. On the epistemological view, characterizations of trust focus only on the risk that B might exploit A’s vulnerability because—we think—those characterizations assume something like transaction cost theories of the firm, discussed below in the last section. But the examples outlined here show that the obligations involved in real trust relationships are much more complex and concern positive action on the part of the trusted party, as part of a relationship, which goes beyond the trustor’s concern with not having his or her vulnerability exploited. One of the advantages of our view is that we can account for these positive obligations.

Two further examples can make our characterization of trust more concrete and clarify the differences between our conception and the epistemological one. First, consider this example, from Baier: “We often trust total strangers, such as those from whom we ask directions in foreign cities, to direct rather than misdirect us, or to tell us so if they do not know what we want to know; and we think we should do the same for those who ask the same help from us” (1986, p. 234). With Brenkert, and with Mayer et al., the act of trusting strangers involves a set of beliefs, expectations, and assumptions. Because we lack specific information about the strangers in real cases of this sort, they could only analyze this example in more general terms: we trust strangers on the basis of general expectations connected to our shared humanity (or something like that). This sort of analysis concerns the possible justification conditions for trust in this case, it explains why people trust, not what trust is. This way of putting the point—this suggestion that epistemological definitions concern antecedent conditions, not trust itself—is a reference to the point made earlier, namely that the epistemological definition of trust still has much to contribute to questions about the epistemological components of a willingness to trust or to make a trust-invitation.

Instead, in this example we suggest the following: asking for directions is an invitation to a trust relationship, though a thin or minimal relationship to be sure, and if the stranger accepts the invitation, then the stranger takes on an obligation. So the act of trusting—when accepted—establishes a trust relationship between the two parties and therefore a personal relationship where none existed before, even if—as in this case—the result is a thin relationship involving a minimal obligation. As noted, trust often occurs in the context of an existing relationship or an existing moral relationship; in cases of that sort trusting actions create a new obligation, and in that way transform or deepen the existing relationship. This existing relationship could offer reasons to expect the trusted party to fulfill his or her obligation, but—again—this point about antecedent expectations does not address what it means to trust and does not account for the way trust establishes or transforms relationships by creating obligations.

Second, cases of so-called “swift trust,” as described by Meyerson et al. (1996), can be explained in exactly the same way, applying our analysis of Baier’s example in a business context. Meyerson et al. note that “temporary systems exhibit behavior that presupposes trust, yet traditional sources of trust—familiarity, shared experience, reciprocal disclosure, threats and deterrents, fulfilled promises, and demonstrations of nonexploitation of vulnerability—are not obvious in such systems. In this respect, temporary systems act as if trust was present, but their histories seem to preclude its development” (p. 167). Meyerson et al. go on to argue that temporary systems provide a different set of antecedents or justification conditions for trust, including prospects for future interaction, clearly defined roles and expectations, time pressure, and perhaps most important, mutual dependence. “[T]he “mere” process of group formation alone may provide an initial foundation for the emergence of a protean sort of swift trust” (p. 185). But from the perspective developed in the present paper, and using the example from Baier, a better conclusion seems to be this: the fact that trust is present in temporary groups shows that the antecedents listed above are not necessary for trust. So the problem, from our perspective, is that Meyerson et al. are operating with an inadequate conception of trust. Trust is not a willingness to cooperate and accept vulnerability on the basis of certain factors that suggest that the vulnerability is manageable. Instead, trust is a moral relationship that can bind even strangers willing to commit themselves.

Hollis and Fukuyama: Two Related Accounts of Trust as a Moral Phenomenon

Hollis (1998) defends what he calls a “normative conception” of trust. This normative conception is the closest we have found to our moral conception of trust in the literature, but it still differs in an important respect. For Hollis trusting actions do not put obligations in place but, instead, appeal to background moral obligations. So, if Hollis trusts another and lends out his copy of the First Critique, the borrower has an obligation to return the book because returning the book is the right thing to do. In contrast, on our view, when Hollis trusts the borrower, the act of trusting (lending the book) creates an obligation on the part of the borrower, because taking the book is an implicit acceptance of this obligation. On our account, then, the borrower has an obligation to Hollis because Hollis trusted him or her, and because the borrower accepted that obligation.

Hollis’s normative conception is too narrow for two reasons, and therefore does not amount to an alternative to our moral conception of trust. First, the two positions—ours and Hollis’s—are not mutually exclusive in the book example. As just noted, on our view the borrower has an obligation to return the book because of the obligation created by trusting and because it is the right thing to do. But the two sources of obligation can come apart: if A invites B to meet him at a certain coffee shop, and if B accepts that invitation, then (i) B takes on an obligation to meet A at that coffee shop, (ii) A trusts B to meet him there, and (iii) it is right for B to be at that coffee shop. There is no prior or general moral obligation for B to be at that coffee shop, so we cannot explain B’s obligation in this example in terms of “what is right” independent of the trust relationship. As a result Hollis’s definition cannot account for the obligation here. Instead the trust-invitation, when accepted, creates the obligation, and in that way trusting action constitutes right in this context, even where there is no appeal to general moral obligation. Because Hollis’s view cannot accommodate the obligation-creating aspect of trust, it is inadequate as an account of trust as a moral phenomenon.

Hollis’s normative conception of trust is too narrow for a second reason: it precludes normative trust among thieves, because thieves trust one another to do what is wrong. For example, Tony Soprano trusted Salvatore “Big Pussy” Bonpensiero in Hollis’s normative sense; this trust relationship between Tony and Big Pussy was not a matter of predictions or expectations. But Tony trusted Big Pussy to continue to be loyal to the mafia family, so Tony trusted him to do what was wrong (though of course, Tony did not think it was wrong, or did not want to admit that it was wrong). And Big Pussy betrayed Tony by doing what was right, by collecting evidence against the crime family for the FBI. Hollis does address the question of trust among thieves; he seems to concede that there is such trust but then denies that they are examples of “trust-within-reason” (p. 106), which means, as far as we can tell, that trust among thieves cannot be defended on the account of practical reason he develops.Footnote 8 As a result, we take Hollis’s normative conception of trust to be at best incomplete.Footnote 9

From the perspective developed here, Fukuyama’s (1995) characterization of trust is a variant of Hollis’s, in that moral content is present only in background moral obligations. Fukuyama’s characterization is inadequate for similar reasons. Fukuyama suggests that stability and prosperity in postindustrial societies depends on reciprocity, moral obligation, duty toward community, and trust. And on his view trust is “the expectation that arises within a community of regular, honest, and cooperative behavior, based on commonly shared norms, on the part of other members of that community” (1995, p. 26). Trust is, on this characterization, the result of shared ethical norms about reciprocity, moral obligation, and duty toward community, “including honesty, reliability, cooperativeness, and a sense of duty to others” (pp. 43, 46). So trust itself does no work; it is those deeper norms that shape behavior and attitudes, and talk of trust is a kind of shorthand for the presence of these social virtues. This is especially clear in one of Fukuyama’s examples. Employment in large firms in Japan is essentially lifetime employment. Fukuyama notes that this sort of arrangement would seem to invite free riding: employees are compensated on the basis of seniority rather than job performance, so “any increased benefits arising from superior performance are in effect a public good with respect to the company as a whole, giving an individual an incentive to shirk his part of the burden” (p. 188). But there is an especially strong work ethic among those employees, and Fukuyama explains this in terms of a reciprocal moral obligation: firms provide stable employment and steady advancement, while workers provide their best efforts. This sense of reciprocal obligation is deeply embedded in Japanese society as a result of its feudal traditions. Our point here is that trust drops out of the explanation, it does no real work: firms and employees interact given their mutual acceptance of certain norms about reciprocity. Talk of trust in this context is, in effect, shorthand for the presence of the social virtues that most concern Fukuyama, and trust-as-shorthand obscures the way trust can create obligations and so affect relationships.

Baier’s Account of Trust, and a Remark on the Ordinary Use of the Term “Trust”

As noted, Baier claims that one defining characteristic of trust is the possibility of betrayal. We have taken this claim as a starting point. But Baier’s own conception of trust cannot account for the possibility of betrayal, so—perhaps ironically—her conception of trust is not moral, and it is therefore inadequate as an alternative to the account developed in this paper.

Baier defines trust as making oneself vulnerable by giving to another discretionary control of some object, relying on or depending on the goodwill of that other person. When Baier talks of relying on another’s goodwill, or depending on another’s goodwill, she apparently means that we give over discretionary control of some object, on the basis of a belief that the other person will act decently and not exploit our vulnerability. This reading is perhaps surprising, but in later papers Baier summarizes her conception of trust in these terms: “I define [trust] as accepted vulnerability to another person’s power over something one cares about, in the confidence that such power will not be used to harm what is entrusted,” where we take Baier’s reference to confidence to suggest that trust involves beliefs about the low likelihood of harm (1995, p. 327, our emphasis). This reading is supported by a comment in another essay, where she characterizes trust as “acceptance of vulnerability to harm that others could inflict, but which we judge that they will not in fact inflict” (1995, p. 152, our emphasis). Baier’s definition is therefore an epistemological one, and as a result the criticisms presented in the first section above apply here: when Baier writes of relying on another’s goodwill, there is no moral obligation that can be violated, and no sense in which a relationship is constituted (or transformed if there is an existing relationship). So Baier cannot account for the possibility of betrayal. In other words, Baier defines trust in terms of relying on another’s goodwill, but it is unclear why relying on another’s general goodwill, understood as his or her decency, usual care/concern, etc., can be betrayed while relying on another’s skill at some task, on his or her regular schedule to keep myself on time, and so forth, cannot be. This is not an idiosyncratic reading of Baier. Karen Jones also reads Baier in epistemological terms, though Jones does not use the term “epistemological.”Footnote 10

Read more sympathetically, Baier’s reference to goodwill could be an elliptical reference to general moral obligations, so we could read her as offering a definition of trust like Hollis’s: when we trust, we expect another to act with goodwill, and this means doing “what is right.” But above, in our discussion of Hollis’s work, we argued that trusting action creates obligations, and a definition of trust that appeals to pre-existing obligations is not adequate because it cannot account for the creative dimension of trust.

Two further points about Baier’s approach are important in this context, to make clear the way our conception of trust is related to the ordinary use of the term. First, as mentioned, Baier’s goal is to distinguish between reliance, which, on her view, can only be disappointed, and trust, which can be betrayed. This approach is complicated by the fact just mentioned that Baier goes on to define trust in terms of reliance. But setting that difficulty aside, ordinary language is ambiguous here: it seems that in many cases relying on a person (as opposed to relying on an object, like a well-manufactured car, or on natural processes) just is to trust that person on the moral conception. We have suggested that a trust relationship between A and B is constituted when A trusts B to do X; A is now dependent on B, and B acknowledges some responsibility for the outcome. This can also be true when A relies on B: if A relies on B to do X, A is now vulnerable to B’s actions, and if B acknowledges that vulnerability and is willing to take responsibility for X and for A, then we should say that A trusts B in relying on him or her. Reliance in this sort of case can therefore be betrayed, though relying on the weather or another person’s attributes—his or her strength, intelligence, or even good will—can only be disappointed. So we take Baier’s point about trust and betrayal as a starting point, but we reject her claim that this forms a basis for distinguishing between trust and reliance.

Second, the proper contrast is between trust and confidence, not trust and reliance. As noted, on epistemological characterizations, trust just is a disposition to act with confidence that others will not exploit our vulnerabilities, confidence that others will behave as expected, and confidence that the benefits of cooperation or participation in a shared project will outweigh the costs, where this is all expressed in terms of beliefs or expectations. In effect, then, epistemological approaches reduce trust to action-on-the-basis-of-confidence. We do not deny that the term “trust” is often used in the epistemological sense. But distinguishing between the moral and the epistemological conceptions of trust requires more precision than found in everyday use. And this is essential because—by reducing trust to confidence—the epistemological approach obscures the moral nature of trust and also the presence of substantive obligation in trust relationships, rather than mere expectations or beliefs held by the trusting party.Footnote 11

Final Clarifications

Three final points are important in clarifying the scope and application of this project. First, we do not mean to suggest that trust is a component of all moral relationships; or that all relationships involving obligation are trust relationships; or that one has a moral obligation to trust another or to accept a trust relationship. Regarding this last point, Uslaner (2002) argues that we have a moral obligation to trust in order to make the world a better place, because trust is a precondition of civic life. This general obligation to trust, what Uslaner calls “moralistic trust,” is grounded on a belief in the general goodwill of others: it is “largely based on an optimistic view of the world and a sense that we can make it better” (p. 4). But Uslaner does not defend a moral conception of trust; his conception of moralistic trust concerns the antecedent conditions, not the definition of trust or the nature of trust relationships.Footnote 12

Second, questions about the abuse or misuse of trust are important. And the potential misuse of trust (“antitrust”) is one of Baier’s central concerns: “Exploitation and conspiracy, as much as justice and fellowship, thrive better in an atmosphere of trust” (1986, pp. 231–232; see also Brenkert 1998, pp. 301–302). But questions about the use and abuse of trust are not related to the central issue in this paper: questions about the use and abuse of trust, and questions about how trust can be put to immoral ends, are distinct from questions about whether trust is itself a moral phenomenon. Discussion of the moral aspect of trust in the literature sometimes confuses these points: the fact that trust can be abused, or the fact that trusting parties could be engaged in immoral activity, does not have any bearing on whether trust itself is a moral relationship.

Third, Flores and Solomon (1998) also describe trust in terms of relationships; their central point and their examples concern the way trust is created in specific kinds of relationships. But when they argue, more generally, that “What is important here is to get away from the solipsistic view that trust and love (and emotions in general) are processes or states in individuals’ minds or “heads” rather than relationships between them in the social world” (p. 217), it is unclear what position they reject. Brenkert, for example, would situate trusting parties (on his definition) in social relationships, and argue that a trusting attitude and the beliefs that justify it depend on the specific context, the existing relationship between the parties, both parties’ emotional histories, and most likely a long list of other factors. And even if Brenkert’s approach is transactional in the sense that his concern is specific instances of trust, person A trusting B to do X, those transactions are always in some social context. So Flores and Solomon’s paper is important in emphasizing the variety of trust relationships (simple, basic, blind and authentic trust, all distinguished by their justification conditions), but it is unclear what their appeal to relationships is meant to accomplish at a theoretical level, it is unclear how their position amounts to an alternative to Brenkert’s or to epistemological conceptions more generally, and their emphasis on relationships does not amount to a moral conception of trust.

Application in Business Organizations

We conclude by outlining an application, to show that the epistemological conception of trust cannot do the work required of it, and to show that the moral conception clarifies the nature of trust relationships.

Over a series of papers, Ghoshal has argued that the transaction cost theory of organizations is ultimately self-defeating (Ghoshal et al. 1995; Ghoshal and Moran 1996; Moran and Ghoshal 1996; Ghoshal 2005). According to transaction cost theories, persons are self-interested, some are opportunistic, and “organizations exist because of their superior abilities to attenuate human opportunism through the exercise of hierarchical controls that are not accessible to markets” (Ghoshal and Moran 1996, p. 14).Footnote 13 But Ghoshal argues that hierarchical control is ultimately self-destructive or self-defeating: such control can prevent opportunistic behavior, but those controls also cause negative attitudes among employees, and those negative attitudes can promote further, unexpected opportunistic behavior. This creates the need for more hierarchical control, which increases transaction cost, and eventually the transaction cost advantage of organizations over markets is lost. There is much support for this line of thought (see Cialdini 1996; Miller 2004; Kramer 1999; ACLU 1999, which is mentioned in Ehrenreich 2001, whose more anecdotal work also supports this line of thought). Ghoshal concludes, “The manager’s primary task is redefined from institutionalizing control to embedding trust” (Ghoshal et al. 1995, p. 14). Ghoshal does not say more, he does not cite a conception of trust or explain how trust makes a difference. But in this concluding section our goal is to fill in that gap, to show that Ghoshal’s claim only makes sense given our moral conception of trust.

Applying the epistemological conception of trust, Ghoshal would be suggesting that a manager’s primary task is to embed in his or her organization a willingness to be vulnerable, given expectations about the likely behavior of the trusted party, or to embed in the organization a willingness to take risks given careful, accurate assessment about the potential benefits and possible costs. But that is what members of organizations already do on the transaction cost theory Ghoshal rejects. So the epistemological characterization of trust cannot explain how an organization built on genuine trust differs from one in which incentives and sanctions increase “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor.” That quotation is Mayer, Davis, and Schoorman’s definition of trust; we use their definition here to emphasize the point, to show that their definition cannot contribute to our understanding of an organization built on trust. For this reason Ghoshal cannot be thinking of trust as characterized in epistemological definitions.

If, instead, trust-invitations can create a network of obligation and inter-personal (or intra-organizational) relationships, those obligations and those relationships could support an alternative conception of organizations. This point is consistent with empirical evidence showing that trusting others can foster trustworthy behavior [see, for example, Cook et al.’s (2005, pp. 145–146) review of the sociological literature, which shows that there is a “positive effect (trustworthy behavior) from supervisors treating employees as if they are reliable even when there is relatively little information about their actual character or competence”]. And by fostering trustworthy behavior, trusting-action could foster more efficient, more productive business interaction, with cost advantages to be sure, but also create a different kind of business relationships.

To extend the point, and to suggest one further application, we can distinguish between two conceptions of organizations. Given the assumptions of transaction cost theory, namely that persons are self-interested, and some persons are opportunistic, the manager’s task is twofold: using incentives to align individual employees’ interests to achieve an organizational-level outcome, and using sanctions to prevent employees from acting opportunistically, from exploiting their positions for individual gain at the expense of the firm. This line of thought is ultimately rooted in Frederick Winslow Taylor’s foundational work, The Principles of Scientific Management. Taylor documented his success at improving productivity by offering higher pay for more work, initiating a series of studies that form the conceptual foundations for the field of management. Organizations of this sort—composed of individuals whose interests have been aligned using incentives and sanctions—are often described as cooperative. But these organizations are better described as coordinated, in the sense that management has coordinated the decisions and actions of self-interested parties to achieve an organization-level goal.

In contrast, genuine cooperation goes beyond what we have called coordination, in that cooperation involves shared ends, mutual accountability, and considerations of mutual benefit—measured on its own terms and not in terms of self-interest from participating in a coordinated scheme. The shared ends are perhaps most important here: the parties in cooperative relationships adopt ends that all can recognize as important and share, in contrast to coordinated relationships, in which individual, self-interested ends are made to converge using incentives and sanctions. In cooperative relationships these shared ends also serve as shared reasons for actions, again in contrast to coordinated relationships, in which self-interested goals give distinct grounds for action to different individuals. Cooperation is therefore far more than a set of voluntary transactions made on the basis of the economic calculation that the potential benefit exceeds the risks, and more than the strategic calculation that the vulnerabilities involved are somehow justified. The management task in a cooperative organization is therefore different from the one just described in the context of coordinated organizations; managers must foster a complex network of shared ends, mutual obligation, and mutual responsibility, as part of a broad set of relationships across employee populations. This is what Ghoshal apparently means when he describes the manager’s task in terms of embedding trust in an organization. And in this context, we can think of the trust-invitation in a broader sense: a trust-invitation seeks out a moral relationship, but in the context of a cooperative enterprise this invitation can be broader; it is an invitation to accept an obligation as part of a larger pursuit both parties can take on as an end, which amounts to an invitation to a shared pursuit that, in effect, unites individual ends.Footnote 14

This distinction between coordinated and cooperative organizations is consistent with Katzenbach and Smith’s (1993) distinction between groups and teams. On their view, a group is a “function of what its members do as individuals”: “[T]he focus is always on individual goals and accountabilities. Working-group members do not take responsibility for results other than their own” (p. 112). Using the distinction developed here, groups are collections of persons whose actions have been coordinated and whose interests have been aligned. In contrast, according to Katzenbach and Smith, a team is “a small number of people…who are committed to a common purpose, set of performance goals, and approach for which they hold themselves mutually accountable” (p. 112). Teams are therefore cooperative, and for this reason are much more productive than groups along a number of dimensions. Katzenbach and Smith’s work is intended to guide managers in the creation of teams: they emphasize the importance of setting specific team performance goals, which “facilitates clear communication and constructive conflict within the team,” and enables teams to focus on results (p. 113).

Katzenbach and Smith suggest that trust will emerge from teamwork, assuming that trust grows out of the interaction of team members as team members build cooperative relationships (see p. 116). Instead, we might think that teams require and depend on trust, because persons will only commit to shared purpose, shared goals, and mutual accountability in the context of trusting relationships. And the present paper—in particular our analysis of swift trust and also the example of trusting strangers—explains how trust can support cooperative relationships where there are no prior relationships among team members. This line of thought suggests, further, that cultivating an environment of trust could be the most important aspect of building teams.

There is more to say about this application, and in particular there is more to say about the practical question of how managers can embed trust in organizations and cultivate cooperative relationships in practice. That further question concerns tactics, but that is not the purpose of the present essay. And there is much to explore about the way that culture affects trust relationships: some cultural factors and some organizational contexts will foster trust relationships while others will impede them (see, for example, the case studies in Saunders et al. 2010). But understanding the way those factors and contexts affect trust—understood to be a moral relationship—is also a matter for future application. Our goal has been to distinguish between two conceptions of trust, and in the context of this application, to show that the epistemological conception of trust cannot do the work demanded of it: the epistemological conception of trust cannot show how an organization built on trust differs from one characterized in transaction cost terms. For this reason the moral conception is essential.