1. Trusting is a Psychological Phenomenon

Whether trusting is taking place or not depends on what is going on in the mind of an agent. Any discussion of trust should try to capture or account for our experience of trusting and our intuitions with regard to trusting, although such an account should be consistent with psychological and neurological findings.

2. Trusting Needs Betrayal (But Not Necessarily in a Bad Way)

Vulnerability to betrayal demarcates trust from nearby notions like reliance. Having said this, a perception – remember trusting is a psychological phenomenon – on the part of the trusting agent that her trust has been breached won't necessarily lead to a feeling of betrayal. It's just that someone will only ever feel betrayed, if they trusted. And it's only if someone trusts that they might end up feeling betrayed.

3. Trusting Differs From Reliance

To see how this vulnerability to betrayal follows from a trust relation and not other nearby alternatives, consider the following case:

Eusebio won't feel betrayed, though he might perhaps feel annoyed or even angry. It also seems right to say that Eusebio doesn't trust his alarm clock but merely relies on it. What this highlights is that in cases in which there isn't the possibility of feeling betrayed, then there is no trust relation and vice versa.

4. No Objects (Perceived as Such)

Why does it seem appropriate to think that Eusebio relies on, rather than trusts, his alarm clock? For one thing, an alarm clock is an object that Eusebio perceives as such. Intuitively, objects that are perceived as such cannot be the objects of trust.

If someone said that they trusted their car, rather than relied on it for example, then intuitively we would have grounds to think that they were misusing ‘trust’ or perhaps that they were making strange and false attributions to their car. While sometimes people might say that they trust their car, the thought is that if they were to speak carefully, then they would not do so any more than they would refer to their car as she or he, or generally make attributions to their car suggestive of the perception of agency and so a will. Furthermore, there's no scenario in which we would say that a trusting agent can be betrayed by his or her car.

Karen Jones makes a similar point, writing that we can only trust things that have wills. I don't quite agree as I take it we could trust something that we take to have a will, but which actually doesn't. This is obvious if we remember that trusting is a psychological phenomenon. So people in the future could form trust relations with androids, if they were to take them to have wills even if they don't actually have wills.

5. Add Conscious or Unconscious

Annette Baier (Reference Baier1986), the chef de cuisine of the philosophical trust world under whom we all labour, writes that an agent may or may not be conscious of trusting another agent. That an agent may be conscious of trusting is obvious. The claim that an agent may unconsciously trust another agent is plausible. An agent might, after something goes wrong, feel betrayed. She might do so without having realized that she trusted the person she feels betrayed by.

Similarly, an agent may be conscious of being trusted or unconscious of being trusted. Again, that one may be conscious of being trusted is obvious. It also seems right to say that an agent may be trusted without being conscious of being trusted. Consider the following case:

6. Add Involuntary

That an agent may simply find that he trusts another suggests that trust, at least sometimes, occurs involuntarily. Whether trust in every case is involuntary is more contentious. Let's first consider a case.

We can imagine the tourist considering the offer and his situation. We might be tempted to say that, being morally upright but having an empty stomach, he has no good alternative except to decide to trust the stranger. The defender of the view that trusting can only be involuntary, however, can plausibly offer the rejoinder that the agent may simply act as though he trusts. He needn't know whether he actually trusts and it may be that he doesn't trust. This might become clear if something goes wrong and he finds himself feeling unsurprised rather than betrayed or let down. For dissenting opinions on this, see Paul Faulkner (Reference Faulkner2007) and Richard Holton (Reference Holton1994).

7. Adding Voluntary Soufflés: Up Your Soufflé

One reason to worry about voluntary trusting is the implications it seems to have for agents like us. If voluntary trusting were possible and if we were to receive, say, testimony x, then we could simply decide whether to trust the testifier and so we could decide whether to believe her testimony. But now it looks like we have a route to believing at will, but it's very implausible that we can believe at will. I can't just believe that there's an elephant making apple strudel in my room, no matter how much I want to, in the way I can simply raise my arm. Actually the involuntariness of belief is a good model for the involuntariness of trust. We can do certain things to try to bring about trust/belief, but normal human agents can't believe/trust by a mere act of their own will.

8. Optimism of Competence

Carolyn McLeod (Reference McLeod and Zalta2011) writes that R can't be said to trust someone, say T, to Φ, if he does not believe that T has the competence to Φ. This seems right. I can't be said to trust you to buy me a good wine, if I don't belief that you know anything about good wines.

As McLeod points out, and this also seems right, trust may be compatible with the trusted person presently not being competent in the relevant respect. Parents may trust that a child will make them cappuccinos in their old age without believing that the child is currently competent to do so.

9. Think Well of Others – A First Pass

McLeod (Reference McLeod and Zalta2011) writes that paradigmatically ‘trust involves being optimistic’ that the trusted person will do something for us. The idea, as it is presented, is that if someone is easily suspicious of others, then that suspicion can forestall trust relations from forming. However, what is necessary for trusting is not just an absence of suspicion in the way mentioned, but that candidate recipients of trust are well regarded. But how exactly?

10. Goodwill

Now for the main course. Baier (Reference Baier1986) first put forward a goodwill view in ‘Trust and Antitrust’. Below I attempt to improve on Baier's pièce de résistance.

The ‘will’ in ‘goodwill’ indicates that a trusting agent may trust a perceived agent, though this is understood liberally enough to include entities such as states and corporations. The ‘good’ in ‘goodwill’ need not mean a moral will or agency, nor a will or agency perceived as moral by the trusting agent. Rather the sense of good involved is that the will is regarded as good by the trustor given some end, the end being whatever the trusted agent is trusted to do. For example, a bank robber may trust his accomplice to be waiting outside the bank in a getaway car. In doing so it seems obvious that he need not suppose that the will of his accomplice is good in a moral sense.

11. Good as Approval

The ‘good’ of goodwill in cases of trust should be understood as approval of what is believed to be the agent's will with regard to the relevant matter by the trusting agent who cares about that matter. The bank robber may trust his accomplice to be waiting outside the bank in the getaway car, if he approves of what he takes to be his accomplice's will in this matter, which he of course cares about.

Such approval also has an affective aspect. A trustor has a positive affective attitude towards the will of the trusted agent about that thing which the trustor cares about. Thinking of goodwill in this way, as we shall see, helps us avoid problems that Baier's account has faced.

12. The Belief Component

A motivation for the approval condition is that it captures the sense in which if a person trusts another, then there is a sort of endorsement of the will of the trusted agent regarding the matter cared about. When the parent trusts the tutor to educate his child, then the parent not only thinks that the tutor has the competence to do so, but that he also has the right will to do so – the teacher not only has the know-how, he's also a dedicated professional. This belief that another agent has a good will about something we care about, where ‘good’ is reducible to approval on the part of the believing agent, indicates that the trusting agent takes the trusted agent to be ‘on the same page’ or ‘on the same side’. This being on the same page or same side is with regard to the will of the trusted agent in relation to whatever it is they're trusted to do. This is, in fact, exactly what we would expect when one agent trusts another. We wouldn't say that one agent trusts another in a particular matter if he doesn't approve of the other agent's will in that matter.

It's in the nature of trusting that even when the trusting concerns the performance of some action, the trust may have a relatively general form and so imply like general beliefs. When John trusts Elena to look after his cat while he's away, he believes that Elena has a good will towards looking after his cat Omakase. This involves him believing or being disposed to believing that she will feed Omakase at appropriate times and in appropriate amounts, that she would take Omakase to the vet were he to become seriously ill, and so on. Approving of Elena's will towards looking after Omakase can mean, however, that he believes or is disposed to believe that she won't do certain things for Omakase. For example, if John were to consider a scenario in which the building Omakase is in is burning down and Elena only has time to save either Omakase or a group of small children, then John would believe that Elena would save the group of small children. Were the scenario realized and Elena saved Omakase ahead of the children, then John might be sorry that he trusted her to look after Omakase. (Of course, this is not to say that a trustor may be sorry they trusted another for only moral reasons or that moral reasons have to be a reason why an agent might be unhappy with trusting another.)

Expectations when one trusts another may relate to the ways in which the trusted person will do what they are trusted to do and the circumstances in which they will do what they are trusted to do. When one trusts someone to Φ, then one approves of their will with regard to Φ where this implies that one believes or is disposed to believe that they will act in an approved-of way across a range of possible circumstances.

13. The Affective Component

The motivation for saying that the trusting agent has a positive affective attitude towards the will of the trusted agent is twofold. First, it seems to fit with our experience of trust. When we're aware of trusting someone to do something, we're often aware of having a positive affective attitude, although precisely in what way we have a positive affective attitude is usually not very clear to us. Furthermore, when we haven't been conscious of trusting, but eventually find that we had trusted but now no longer do, say after a betrayal, then we seem to experience the loss of something aside from belief. There seems to be a loss of something affective. Requiring that trusting implies a positive affective attitude on the part of the trustor gets us this result. Not only does the affective requirement accord with our experiences of trust, inclusion of the condition gets us the right results in the significant cases of trust discussed in the literature and new cases set out here. Without the affective aspect, say with only a belief aspect, what we would have doesn't seem recognizable as trust.

14. Improving on the Pièce de Résistance

While the goodwill account offered here is based on Baier's (Reference Baier1986) goodwill account, it also differs from it in a number of significant respects. The final definition Baier (Reference Baier1986: 259) offers is that ‘[t]rust … is reliance on others’ competence and willingness to look after, rather than harm things one cares about which are entrusted to their care'. In this definition we see both a competence requirement and a suggested vulnerability requirement; the trustor is relying on the other to be both willing and able to look after something with which the other has been entrusted. One significant difference is in how goodwill is spelt out in the definition.

For Baier (Reference Baier1986), the goodwill requirement is understood as a ‘willingness to look after, rather than harm’. The analysis I offer is simpler and clearer, it doesn't make reference to evaluative notions like ‘look after’ and ‘harm’. We can judge whether trust is present on the basis of whether certain descriptive facts obtain. We don't have to make sense of whether someone is looking after rather than harming something with which that person has been entrusted. In fact, this talk seems crude when we consider that we don't ordinarily talk of harming things or tasks and yet much of our trust talk is about looking after things and undertaking certain actions.

15. Jones's Resistance to the Pièce de Résistance

Jones (Reference Jones1996: 18) has criticized Baier's view for seeming to imply that one can trust at will. All that one has to do is entrust something one cares about to another and be reliant on that person's competence and willingness to look after it. Perhaps we shouldn't understand ‘entrust’ as merely giving over, but if entrusting itself is a form of trusting, then we have cause for complaint for using the term to define trust. In any case, by requiring that the trusting agent have beliefs of the sort outlined, the account that I offer avoids these problems.

16. Holton's Resistance to the Pièce de Résistance – Part I

Holton (Reference Holton1994) also criticizes Baier's account. (Actually, I'm not entirely persuaded by Holton's criticism of Baier's account, if we take Baier's (Reference Baier1986) final definition of trust as her final position on trust in the paper. Nonetheless I show how my account avoids the criticism.) First, Holton sets out a confidence trickster scenario, think of a card shark, in which the trickster operates by relying on the goodwill of others. This is intended to show that reliance on goodwill is not sufficient for trust. I'm not claiming that such a reliance on goodwill is sufficient for trust. Trickster scenarios seem to be ones precisely in which the trickster doesn't believe in the competence of another person in the relevant matter. That's why the other person can be tricked. The card shark's mark wants to use the card game to, say, win money. She engages the mark precisely because she believes he doesn't have the competence to win money through the card game.

This isn't, however, quite the end of the worry. Consider a trickster case as set out by Zac Cogley (Reference Cogley2012: 36):

Here the victim can plausibly be believed to have the relevant competence, that is, to be able to help Madoff make money. Plausibly there is a sense of ‘goodwill’ which he can be said to believe that his victim possesses; a sense such that it makes sense for him to appeal to her on behalf of himself. This looks like a problem for Baier's view. Madoff can be said to be relying on another's competence and willingness to look after, rather than harm, something Madoff cares about, him making money, which he is entrusting into her care.

This, however, is not the sense of goodwill which I'm claiming has to be present if there is to be trust. It is of course plausible that he likes the will that seeks to make him money, but it's implausible that he approves of that will in that this is a will with which he is on the same page. Given what Madoff knows about his own dishonest dealings, it's not as if he would think that this is the will to have. After all, it's a will that will get his mark swindled. My account, therefore, can withstand such trickster cases.

Note that one could try to dismiss the Madoff case as a counterexample by pointing out that it's a case in which Madoff is not vulnerable to a feeling of betrayal. I won't defend this view here, but my position is that vulnerability to reactive attitudes, such as betrayal, which accompany trusting in normal agents, follows from what trusting involves and shouldn't simply be added as a separate requirement. An agent taking a drug which inhibits the feeling of betrayal wouldn't be prevented from trusting on my view.

17. Holton's Resistance to the Pièce de Résistance – Part II

The second criticism that Holton makes is that reliance on goodwill is not a necessary condition for trust. He illustrates his claim by describing a case in which one member of an estranged couple, nonetheless, trusts the other to look after their child. Holton's claim is that it's plausible to think that there could be such trust and yet think that the member of the couple who trusts the other is not relying on her goodwill towards him – the member of the couple. To adapt the criticism somewhat, we might worry that such a case shows that the goodwill condition for which I have argued isn't necessary. The thought could be that it's plausible to think that there is trust in such a case and yet think that the member who trusts the other has no belief that the other has goodwill towards him. Recall, however, the condition I'm arguing for is not one according to which the trusting agent need believe that the trusted has goodwill towards him. Rather, if S trusts R to Φ, then S believes that R has goodwill with regard to Φ, where goodwill is a will of which S approves.

As ‘goodwill’ is to be understood as approval of the trusted agent's will towards the object of trust that the trustor cares about, we can make sense of how someone could believe another to have goodwill with regard to, say, looking after my bicycle. Approval does imply that R has a positive affective attitude towards S's will, at least with regard to Φ. Is this odd if R and S represent the estranged couple in the above-mentioned example? Estranged couples can still retain some positive affective attitudes with regard to one another, including their wills, even in cases in which those attitudes are dwarfed by negative affective attitudes with regard to the person generally. Thus, the goodwill account of trusting offered here can withstand Holton's criticism.

18. The Doggy Bag

I've provided the formula for trusting. It's the belief on the part of R, the trustor, in the competence of S, the trustee, to Φ, and R's approval of what she believes is the will of S to Φ, where Φ is something about which R cares.

Now if you want others to trust you, you know what has to happen. Knowledge of this formula is valuable for a variety of ends. It's valuable for corporations, politicians, employees, and in personal matters. Let's consider its value in one example, that of the politician. Do you want to be elected? Persuade the public that what you want is the same as what they want from a politician regarding matters about which they care. But not only do you need to persuade them that you have the right will; you have to persuade them that you have the competence to deliver on what it is that they want. If some things go wrong while you're in office but you retain trust, then you'll get the benefit of the doubt. Trust is the grease that allows you to get things done with flexibility from others and without too much scrutiny.

Trust brings benefits, and may even be necessary for success in certain domains; however, it also brings risks. Betray or be perceived as having betrayed and you will incur an enmity of the sort that the person who was never trusted never receives.