Propositional faith and trustworthiness

Abstract

I argue that propositional faith presupposes trustworthiness, in this sense: faith that p is fitting only if the one in whom you have faith to bring it about that p is trustworthy to bring it about that p. In defence of this, I argue that propositional faith is a species of interpersonal faith and that interpersonal faith presupposes trustworthiness. I also discuss some of the consequences of my thesis for theistic propositional faith.

Introduction

It is more or less obvious that propositional faith, or faith that p, presupposes truth, in this sense: faith that p is fitting only if it is true that p. (More on the relevant concept of fittingness in a moment.) This is why it seems incoherent to have faith that p if you know that it is not true that p. If you have just watched the Celtics lose an overtime thriller in agonizing fashion, you cannot coherently still have faith that they will win the game, unless you somehow remain unsure about whether they lost – for example, whether what you saw was fake news or that officials will discover that more time needs to be put on the clock. Because faith that p is fitting only if it is true that p, faith that p is at least characteristically extinguished when you come to know that it is not true that p. If you return from the other room and announce your faith that the Celtics will win, you will normally change your mind when you discover that the game is already over.

It is less obvious that propositional faith presupposes goodness, in this sense: faith that p is fitting only if it would be good if it were the case that p. But this explains why it seems incoherent to have faith that p when you know that it would not be good if it were the case that p. If you are aware that the bags of rice being delivered to the refugees have accidentally been swapped for bags of sand, you cannot coherently have faith that the delivery will arrive safely at their camp. Indeed, it seems incoherent to have faith that p if you merely believe that it would not be good if it were the case that p. As Robert Adams and William Alston argue, a Bush supporter cannot coherently have faith that Clinton will win (Adams (1995), 12) and an opponent of universal democracy cannot coherently have faith that democracy will triumph (Alston (1996), 12). On this basis, they and others (Howard-Snyder (2013), 326; Schellenberg (2014), 82 n.) conclude that having faith that p entails desiring that p, in that broad sense of ‘desire’ on which desiring that p is equivalent to having a pro-attitude towards the truth of the proposition that p. In any event, because faith that p is fitting only if it would be good if it were the case that p, faith that p is characteristically extinguished when you come to know that it would not be good if it were the case that p. Discovering that the bags were filled with rice would normally make you lose your faith that they will be delivered safely.

Here, I will argue that propositional faith presupposes trustworthiness, in this sense: faith that p is fitting only if the one in whom you have faith to bring it about that p is trustworthy to bring it about that p. This formulation presupposes one controversial claim I aim to defend in what follows: that propositional faith is a species of interpersonal faith. In what follows, I'll defend an account of propositional faith as a species of interpersonal faith, argue that interpersonal faith presupposes trustworthiness, and discuss some of the consequences of my thesis for theistic propositional faith.

Here, ‘propositional faith’ and ‘faith that p’ will always indicate instances of faith where the relevant proposition is contingent. If the object of propositional faith is sometimes a non-contingent proposition, then we are concerned here with a particular species of propositional faith, namely, propositional faith where the relevant proposition is contingent. Assuming the proposition that God exists is non-contingent, I am not here considering faith that God exists, nor any other case of propositional faith in a non-contingent proposition. However, you might think that the object of propositional faith is always a contingent proposition, because we sometimes use ‘faith’ loosely in connection with propositional attitudes distinct from propositional faith proper, which may explain away apparent cases of propositional faith in non-contingent propositions. First, it is natural to speak of ‘faith’ in connection with religious propositions in general. Talking about a person's ‘faith’ is often just another way of talking about their religion. As a result, we may describe someone who believes (or is otherwise committed to the proposition) that God exists – whether or not they have faith that God exists – as having ‘faith’, simply because the proposition that God exists is a religious proposition. Second, it is natural to speak of ‘faith’ in connection with testimonial belief in general. As G. E. M. Anscombe points out, ‘faith’ once commonly referred to believing someone, namely, taking their word or accepting what they said on their authority, or ‘on faith’ (Anscombe (1979), 141–142). As a result, we may describe someone who testimonially believes that God exists – whether or not they have faith that God exists – as having ‘faith’, simply because their belief is testimonial.

Before proceeding, we must clarify what we mean by ‘fitting’ in the present context. A fitting attitude here just means an accurate attitude.¹ In this sense, it is uncontroversial that belief presupposes truth, in this sense: belief that p is fitting if and only if it is true that p. It is in that sense of ‘fitting’, and only in that sense, that I maintain that propositional faith is fitting only if the one in whom you have faith is trustworthy. In the present sense, all and only representations that can be accurate have fittingness conditions, which are simply the conditions under which they are accurate. Belief that p is accurate if and only if it is true that p, and in that sense truth is the fittingness condition for belief. In that sense, belief represents its propositional object as true, such that it is fitting if and only if that proposition is true. In the same sense, on my view, propositional faith represents someone as trustworthy, such that it is fitting if and only if they are trustworthy. Fittingness, in the present sense, is not prescriptive: that an attitude of yours is fitting does not entail that you ought to have that attitude – consider, for example, a true belief based on insufficient evidence. Fittingness, in the present sense, is not evaluative: that an attitude of yours is fitting does not entail that it is good that you have that attitude – consider, for example, a true belief that answers some utterly trivial question.

Above, I said that it seems incoherent to have faith that p when you know that it is not true that p and that it seems incoherent to have faith that p when you know it would not be good if it were the case that p. It is obviously incoherent to believe that p when you know that it is not true that p. Your two attitudes – your belief that p, on the one hand, and your knowledge that it is not true that p, on the other – cannot both be fitting, namely, accurate. If propositional faith presupposes truth, in the present sense, then it is also incoherent, and in the same sense, to have faith that p when you know that it is not true that p, because those two attitudes cannot both be fitting. And if propositional faith presupposes goodness, in the present sense, then it is also incoherent, in the same sense, to have faith that p when you know that it would not be good if it were the case that p, because those two attitudes cannot both be fitting. The kind of incoherence involved here is the kind that is present whenever someone both has an attitude and knows of a fittingness condition for that attitude that it is not satisfied.

Propositional faith as a species of interpersonal faith

I maintain that propositional faith is a species of interpersonal faith. (Just as propositional faith is faith that p, for some proposition that p, interpersonal faith is faith in S, for some person S.) That propositional faith is a species of interpersonal faith is suggested by William Alston:

How is ‘faith in’ related to ‘faith that’? . . . It seems plausible that wherever it is clearly appropriate to attribute ‘faith that’, there is a ‘faith in’ in the background. If I have faith that Joe will get the job, I thereby have faith in Joe, of some sort. If I have faith that the church will rebound from recent setbacks, I thereby have faith in the church and its mission. (Alston (1996), 13)

Alston suggests that having faith that p is some species of having faith in someone. On my view, having faith that p is having faith in someone to bring it about (or to have brought it about) that p. ² It seems to me that whenever you have faith in someone, you have faith in them to do something, but I do not assume that here. However, propositional faith, on the proposed view, is a species of having in faith in someone to do something – namely, to bring about the truth of the relevant proposition.

The proposed view is supported by two complementary considerations. First, cases of faith in someone to bring it about that p are naturally interpreted as cases of faith that p. Second, cases of faith that p are naturally interpreted as cases of faith in someone to bring it about that p. It thus appears that you have faith that p if and only if you have faith in someone to bring it about that p. If so, parsimony speaks in favour of identifying faith that p and faith in someone to bring it about that p. If faith that p and faith in someone to bring it about that p are always found together, we should conclude that there is really one attitude here, which can be described in two different ways, rather than two attitudes, which happen always to be found together. So, in what follows, after some preliminaries, I will argue that you have faith in someone to bring it about that p only if you have faith that p and that you have faith that p only if you have faith in someone to bring it about that p.

Clarifying the proposal

The proposed view is not the view that there is a necessary connection between two distinct attitudes, faith that p and faith in someone to bring it about that p. It is rather that faith that p just is faith in someone to bring it about that p. (Alston says that, when I have faith that Joe will get the job, I thereby have faith in Joe – not that I also have faith in Joe or that my faith that Joe will get the job causes me to have faith in Joe, etc.)

Note well that interpersonal faith is not a relation: you can have faith in someone even if they do not exist. The non-existence of God would not mean that theists do not have faith in God, including faith in God to do various things.

I intend a broad interpretation of ‘bring it about’ here – this includes cases in which someone singlehandedly makes something the case but also cases in which someone contributes to something's being the case. A basketball coach who has faith in their star player to win the game need not expect them to singlehandedly or exclusively cause the game to be won.

I also intend a broad interpretation of ‘someone’ – just as you can have faith in individual human persons, you can have faith in non-human persons, other animals, groups of people, and institutions. To anticipate what I have to say about the fittingness conditions for interpersonal faith, ‘someone’ here can be anything – any being or set of beings – capable of acting and of being motivated to act, in other words, any agent, whether individual or collective.

On the proposed view, propositional faith is essentially interpersonal, in the same way that gratitude is essentially interpersonal. As Stephen Darwall argues: ‘Gratitude is felt . . . in response to an action by a responsible agent. It is true that we speak of being grateful for good weather, for example, but this evidently involves the conceit that the weather is a free gift, as if from God’ (Darwall (2006), 73). There is such a thing as propositional gratitude, or gratitude that p: you can be grateful that the weather is good, or that your friend picked you up at the airport, or that someone offered you a job. But propositional gratitude is a species of interpersonal gratitude. As a rough first pass, to be grateful that p is to be grateful to someone for having brought it about that p.

If propositional faith is essentially interpersonal, it is unlike other propositional attitudes, such as belief, that are not essentially interpersonal. Believing that p need not involve believing in someone. There can be ‘belief that’ without a ‘belief in’ in the background. Nor need propositional belief involve having any attitude to any person. In the same way, propositional hope (i.e. hoping that p) is not essentially interpersonal: hoping that p need not involve having any attitude towards any person. Indeed, hope, unlike belief, does not have an interpersonal species: there is no such thing as ‘hope in’ someone. For this reason, on the proposed view, faith stands to hope in the same way that gratitude stands to happiness. Having faith that p, unlike hoping that p, requires having an attitude towards someone, just as being grateful that p, unlike being happy that p, requires having an attitude towards someone. Some propositional attitudes – like gratitude – are essentially interpersonal, and some – like happiness, hope, and belief – aren't. Propositional faith, on the proposed view, belongs in the former category.

On the proposed view, propositional faith is a species of interpersonal faith. Doctrinal faith, namely, faith that a particular contingent religious doctrine, world-view, or proposition is true, seems like a counterexample to this. If you have ‘faith that the basic Christian story is true’ (Howard-Snyder (2013), 365), do you have faith in someone to bring it about that the basic Christian story is true? I propose that, in this case, you have faith in God to have brought it about that the basic Christian story is true. Or, more precisely, you have faith in God to have brought it about that the contingent parts of the basic Christian story are true. Faith in any contingent creationist religious doctrine, on which contingent truths are brought about by some supernatural or divine agent, is open to this interpretation.

No interpersonal faith without propositional faith

Suppose that I have promised to bring dessert for your important dinner party, such that you now have faith in me to bring the dessert. It is natural to interpret you as having faith that I will, in fact, bring the dessert. Indeed, it is hard to imagine how you could have faith in me to bring the dessert if you did not have faith that I would bring it. ‘I have faith in them to bring dessert, but I do not have faith that they will do it’ seems like an incoherent thing to say. In general, it is hard to imagine someone having faith in someone to φ without having faith that they will φ. When you have faith in someone to φ, you must have faith that they will φ.

If this is right, when you have faith in someone to bring it about that p, you must have faith that they will bring it about that p. But it does not seem possible to have faith that someone will bring it about that p without having faith that p. Someone cannot bring it about that p unless p, which is why it seems incoherent to say, for example, ‘I have faith that they will bring it about that a dessert will be brought, but I do not have faith that a desert will be brought.’ Thus, when you have faith in someone to bring it about that p, you must have faith that p.

Alston suggests that interpersonal faith does not require propositional faith:

[F]aith in a person presupposes that one has some positive attitude toward the proposition that the person exists and that he or she has various characteristics that provide a basis for one's faith. But it is not obvious that this attitude has to be properly characterizable as a case of ‘faith that’. (Alston (1996), 13)

However, I don't think this speaks against the idea that, when you have faith in someone to bring it about that p, you must have faith that p. Consider what Alston says about his faith in his wife: ‘I have faith in my wife; I can rely on her doing what she says she will do, on her remaining true to her commitments, on her remaining attached to me by the bond of love’ (ibid.). Granted, Alston does not plausibly have faith that his wife exists or that she has the various characteristics – her conscientiousness or her reliability, say – on the basis of which he has faith in her. However, he does plausibly have faith that she will keep her promises and remain committed to the marriage. Indeed, it would be hard to make sense of someone who claimed to have faith in their wife but not to have faith that she would keep her promises or remain committed to the marriage. On closer inspection, interpersonal faith, in this kind of case, does involve propositional faith.

No propositional faith without interpersonal faith

Suppose that I have promised to bring the dessert for your important dinner party, such that you now have faith that I will bring the dessert. It is natural to interpret you as having faith in me to bring the dessert. Recall Alston's claim that, if I have faith that Joe will get the job, I thereby have faith in Joe. This suggests that having faith that S will φ entails having faith in S. But that is not quite right. Suppose that you have promised to bring Jojo home from school, such that I have faith that Jojo will be home by dinnertime. I do not thereby have faith in Jojo. It is you in whom I have faith, in this case. So it is not merely the fact that I have faith that you will bring the dessert that makes it plausible that I have faith in you to bring the dessert. Nevertheless, that is the plausible thing to say about the present case.

Propositional faith can involve interpersonal faith without taking the form of faith that someone will φ. Suppose that I have promised to leave the dessert for your important dinner party on your back steps before eight o'clock, such that you now have faith that the dessert will be there before eight. It is natural to interpret you as having faith in me to deliver the dessert on time. (It is certainly not natural to interpret you as having faith in the dessert to be delivered on time.) So, even when faith that p does not take the form of faith that someone will φ, it is sometimes easy to find someone relevant such that the person who has faith is plausibly interpreted as having faith in that person to bring it about that p.

We can here note the need for the parenthetical part of the present claim, on which the person who has faith that p may have faith in someone, not to bring it about that p, but to have brought it about that p. If it is now eight o'clock and you are about to look on the back steps, and have faith that the dessert is there, you do not plausibly have faith in me to bring dessert, but to have brought dessert. Just as you can have faith in someone to do something in the future, you can have faith in them to have done something in the past.

Many cases of propositional faith are plausibly interpreted as cases of interpersonal faith in God, or in some other supernatural or divine agent.³ Suppose that your friend Smith is gravely ill, comatose, and now receiving nothing but palliative care, and yet you have faith that Smith will recover and be well again. You do not have faith that Smith will do something, since they are comatose, nor that Smith's doctors will do something, since there is nothing more they can do. But at least one clear possibility – one version of the story that would be familiar and make perfect sense – is that you have faith in God to save Smith's life.

However, isn't there another possibility? Isn't it possible that you have faith that Smith will recover, without your having faith in anyone to bring it about that Smith recovers?

Let us set aside cases in which faith involves representing nature, or fate, or some other superficially impersonal force or power, as an agent. A fatalist, who conceives of fate as an agent – for example, as Fortuna, the goddess of fortune and personification of luck – and who believes it is their destiny to become a famous musician, might have faith in fortune to bring it about that they become a famous musician, and thereby have faith that they will become a famous musician. A spiritual naturalist, who thinks of nature as an agent – for example, as Mother Nature – and who has resolved to live alone and unprepared in the wilderness, might have faith in nature to bring it about that they survive, and thereby have faith that they will survive. In the same way, a non-theist might well have faith in some such non-divine agent to bring it about that Smith will recover, and thereby have faith that Smith will recover. That is consistent with the idea that you have faith that p only when you have faith in someone to bring it about that p.

Let us also set aside cases that involve the kind of conceit suggested by Darwall in the case of being ‘grateful’ for good weather. I take Darwall's point to be that we sometimes use the language of ‘gratitude’ loosely, in cases that are not cases of gratitude proper. An atheist at a Thanksgiving dinner may have occasion to say, ‘I am grateful that I still have my health’, even if there is no one to whom they are grateful. This doesn't suggest that gratitude is not essentially interpersonal, but rather that we sometimes use the word ‘grateful’ loosely. In the same way, the fact that an atheist fan might say, ‘Oh, God, the Celtics are going to lose again!’, does not show that atheism is compatible with the existence of God. You might likewise speak loosely of your ‘faith’ that Smith will recover, in the absence of anyone in whom you have faith to bring this about. Given that our vernacular has a theistic history, we can expect these kinds of cases, in which people say things that make sense, strictly speaking, only if theism is true, without any such commitment.

None of this yet suggests that you could not have faith that Smith will recover and yet not have faith in anyone to bring it about that Smith recovers. However, two considerations speak in favour of the idea that, when you have faith that p, you must have faith in someone to bring it about that p.

First, the idea that faith that p requires faith in someone to bring it about that p provides one explanation of why faith is commonly associated with religion. Faith in someone to bring it about that p is available to those who believe in God or other supernatural or divine agents in a wider range of circumstances than it is to those who do not. Knowing nothing about your religious beliefs, if you were to say that you have faith that Smith will recover, keeping in mind that there is no worldly individual in whom you plausibly have faith to bring it about that Smith recovers, it would be reasonable to interpret you expressing religious faith of some kind – namely, to interpret you as expressing faith in some supernatural or divine agent to bring it about that Smith recovers. The idea that having faith that p requires having faith in someone to bring it about that p explains why this interpretation is reasonable, and in general why expressions of faith are often reasonably interpreted as expressions of religious faith. This is not to deny that there is such a thing as secular faith. However, intuitive cases of secular faith – such as the case of your faith that I will bring the dessert for your dinner party – are cases in which the person who has faith is plausibly understood as having faith in some non-supernatural, non-divine agent.

Second, the idea that faith that p requires faith in someone to bring it about that p explains why it makes sense, other things being equal, to ask, of someone who has faith, in whom they have faith. Imagine a dinner party host who admits that they have not prepared dessert, but who says that they have faith that dessert will be served. It would be natural to ask them who they expect to prepare or deliver or otherwise produce the dessert. Compare, again, gratitude. Expressions of gratitude are naturally understood as expressions of gratitude to someone, and when it is not clear to whom someone who says they are grateful is grateful, it makes sense to ask them to whom they are grateful. By contrast, expressions of happiness invite no analogous question. The idea that having faith that p requires having faith in someone to bring it about that p explains why we can ask similar questions of someone who says they have faith. And, by contrast, expressions of hope invite no analogous question.

Dan Howard-Snyder describes a case in which he has faith that he will complete a long and difficult backpacking trip and a case in which a hiker sickened by snakebite has faith that help lies ahead (Howard-Snyder (2013), 364–365). Both cases can easily be understood as cases of religious faith, in which the person who has faith has faith in God to bring about the truth of the relevant proposition. But they also can be understood as cases of secular faith: Howard-Snyder may have faith in himself to persevere through the long journey and overcome the inevitable challenges and setbacks; the snake-bitten hiker may have faith in other hikers to provide the assistance he needs to make it to safety.

I claim that faith that p requires faith in someone to bring it about that p. Above, I insisted on a broad interpretation of ‘someone’. The case of the snake-bitten hiker suggests another dimension of breadth. Sometimes, when you have faith in someone to do something, you know in whom you have faith, such that the question of in whom you have faith has a more or less determinate answer. However, interpersonal faith does not require such determinate knowledge. Suppose you are a sailor in distress, adrift at sea. You might have faith in someone to rescue you, without knowing in whom you have faith. You just have a vague idea of someone rescuing you: perhaps another sailor, perhaps the Coast Guard, perhaps a fishing vessel. In this case, the question of in whom you have faith does not have a determinate answer. Propositional faith, however, seems possible in this case: you might have faith that you survive, constituted by your faith in someone, you know not whom, to rescue you. (Like Philip Henslowe in Shakespeare in Love, when asked how things will ‘turn out well’, you might simply think: ‘I don't know. It's a mystery.’)

In many cases, the most plausible interpretation of propositional faith as interpersonal faith will not be straightforward. For example, suppose a Bush supporter has faith that Bush will win. In whom do they have faith to bring it about that Bush wins – in Bush and his campaign, in the voters, in God? For another example, suppose a proponent of universal democracy has faith that universal democracy will triumph. In whom do they have faith to bring it about that universal democracy triumphs – in certain politicians or activists, in the citizens of the world, in the long arc of the moral universe that bends towards justice? We need to know more about the details of these particular cases to provide a plausible interpretation. Knowing that someone has faith that p does not yet determine in whom they have faith to bring it about that p. However, I maintain, it is always the case that they have faith in someone to bring it about that p.

Interpersonal faith presupposes trustworthiness

So far, I have argued that propositional faith is a species of interpersonal faith. Now, I'm going to argue that interpersonal faith presupposes trustworthiness, in this sense: having faith in someone to φ is fitting only if they are trustworthy to φ. It is a familiar idea that the concept of faith and the concept of trust are connected in potentially interesting ways. My claim here alleges one such connection.

What does it mean for someone to be ‘trustworthy to φ’? I am going to assume a particular account of trustworthiness, due to Karen Jones (2012; cf. Baier (1986), Holton (1994), Jones (1996)), although I think what I say here would apply were certain competing accounts of trustworthiness assumed.⁴ On Jones's account:

B is trustworthy with respect to A in domain of interaction D, if and only if she is competent with respect to that domain, and she would take the fact that A is counting on her, were A to do so in this domain, to be a compelling reason for acting as counted on. (Jones (2012), 70–71)

Here, I assume that B is trustworthy to φ, relative to A, if and only if B is competent to φ and would be motivated to φ by the fact that A is relying on them to φ. I follow Jones in taking trustworthiness to require competence, namely, a relevantly reliable ability to do the thing you are trustworthy to do. Someone who is unable reliably to φ cannot be trustworthy to φ. And I also follow Jones in taking trustworthiness to require a particular motivational disposition, namely, a disposition to be motivated by the fact that someone else in relying on you. I require merely being motivated by the fact that someone is relying on you, rather than treating the fact that they are relying on you as a ‘compelling reason’, but this difference will not matter in what follows. All I require is that the trustworthy person be disposed to be ‘directly and favorably moved by the thought that you are counting on her’ (ibid., 6).

It is this motivational aspect of trustworthiness that distinguishes trustworthiness from mere reliability. Theorists of trust typically distinguish between trust and mere reliance by noting that we rely on, but do not trust, inanimate objects and machines: a climber relies on, by does not trust, their rope not to break; a driver relies on, but does not trust, their car not to break down.⁵ Trust involves the motivations of the trustee; reliance does not. For this reason, inanimate objects and machines can be reliable and unreliable, but not trustworthy and untrustworthy. However, we also can distinguish between relying on a person and trusting them. In one kind of case, relying on someone does not involve their motivations at all: a tennis ace might rely on their opponent, who they know has a bad backhand, to miss a backhand volley, but they do not trust their opponent to miss the volley. In another kind of case, relying on someone does involve their motivations, but not the motivation distinctive of trustworthiness: a detective might rely on a criminal, whom they know to be especially rapacious, to return to the scene of the crime to recover something valuable, but they do not trust the criminal to return to the scene of the crime.

The distinction between trust and mere reliance is psychological, rather than linguistic. It would be no violation of ordinary language to say that you trusted your climbing rope to hold fast or that your car was trustworthy; and objects, along with people and animals, are said to be ‘trusty’.⁶ The distinction is not to be found in ordinary language, but by reflection on the difference between relying on someone because you expect them to by moved by the fact that you are relying on them, which is essentially interpersonal, and relying on someone (or something) for other reasons. Likewise, the distinction between interpersonal faith and mere reliance is psychological, rather than linguistic.

To say that interpersonal faith presupposes trustworthiness, as I do, is to treat interpersonal faith as equivalent to interpersonal trust. It is to say, in effect, that to have faith in someone to φ is to trust them to φ. I think this is a welcome implication of the view that interpersonal faith presupposes trustworthiness, for it is unclear what having faith in someone to φ without trusting them to φ, or trusting someone to φ without having faith in them to φ, might involve.

Richard Holton (1994, 67) suggests that trusting someone, by contrast with relying on them, involves treating them as a person. On my view, something like this is also true of interpersonal faith. Having faith in someone to do something requires – in the sense that your faith is not fitting otherwise – their being disposed to be motivated by the fact that you are relying on them to do it. Climbing ropes and cars are not so motivated. But neither are the tennis ace's opponent and the detective's quarry.

Why think that trustworthiness, so understood, is necessary for fitting interpersonal faith? Because this explains why it seems incoherent to have faith in someone to φ if you know they are not trustworthy to φ. Moreover, because having faith in someone to φ is fitting only if they are trustworthy to φ, interpersonal faith is at least characteristically extinguished when you come to know that your trustee is not trustworthy. First, consider cases in which you know someone is not competent to φ. It seems incoherent to have faith in me to bring dessert for your important dinner party if you know that I am unable to bring dessert, for example because you know that I have just boarded a non-stop flight to Hong Kong, and if I were to call you from the airport and renege on a prior promise to bring dessert, you would normally stop having faith in me to bring dessert and start looking for other options. Second, consider cases in which you know that someone, although they are competent to φ, would not be motivated to φ by the fact that you are relying on them to φ. It seems incoherent to have faith in me to bring dessert if you know that I would not be motivated to bring dessert, for example because I am your sworn enemy, and if you discovered that I had become your enemy, having previously been your friend, you would normally stop having faith in me to do things.

However, suppose I am both competent to φ and would be motivated to φ, but would not be motivated to φ by the fact that you are relying on me to φ. Imagine that you are a gangster and I am an undercover police officer who has infiltrated your gang. Prior to discovering this, you had faith in me to collect your debts. Do you still have faith in me to do so? You know that I am perfectly able to collect your debts. And you also know that I am motivated to collect your debts – doing so is essential to my undercover mission and I am deadly afraid of what you would do to me if I did not follow your orders. But you now know that I am not motivated to collect your debts by the fact that you are relying on my to collect your debts. The aforementioned theorists of trust would say that although you may still rely on me to collect your debts, you no longer trust me to do so. I want to say the same about faith: although you may still rely on me to collect your debts, you no longer have faith in me to do so. When you discover that I am not disposed to be motivated by the fact that you are relying on me, your attitude towards me must change. We can capture that change by saying that you no longer have faith in me, because it is incoherent to have faith in someone to φ if you know they would not be motivated to φ by the fact that you are relying on them to φ.

As Holton observes, trust and mere reliance differ when it comes to how we respond to failure: ‘We feel hurt or resentful when [people] let us down; grateful, perhaps touched, when they help. In contrast, when a machine breaks down we might feel angry or annoyed; but not (unless we are inadvertently anthropomorphic) resentful’ (Holton (1994), 66; cf. Baier (1986), 234–235). In the same way, in the case of the gangster and the undercover police officer, although you might be angry or annoyed if I were to fail to collect your debts, you could hardly resent me or feel let down. When you merely rely on someone to φ, and they fail to φ, you may be frustrated or disappointed; only when you trust someone to φ, and they fail to φ, might you feel betrayed. Likewise, on my view, having faith in someone to φ differs from merely relying on them to φ when it comes to the reactive attitudes that are possible when someone in whom you had faith to φ fails to φ. With faith, unlike mere reliance, feelings of resentment and betrayal are possible.

You might object that it is sometimes reasonable to have faith in someone whom you know is not trustworthy. Faith is sometimes forward-looking or ‘proleptic’, because having faith in someone (relatively) untrustworthy can cause them to become (more) trustworthy.⁷ Suppose that I have just taken the first steps on a long and painful journey of recovery from alcoholism. I call to ask you what flavour of cake you want for your party, but you can hear the clinking of glasses in the background and my familiar tipsy drawl. You know I have relapsed and that I am both unable and unmotivated to deliver the cake. But you have faith in me anyway, in the hope that this will contribute to my recovery in the long run. You have faith in me to deliver the cake, despite knowing that I am not trustworthy to deliver the cake. However, this is compatible with my claim here. An attitude can be reasonable even if it is not fitting. As Bernard Williams (1995) suggests, proleptic expression often involves misrepresentation, as when you blame someone who is not blameworthy, in the hope that your blame will cause them to reconsider their actions. To blame someone, Williams argues, is ‘to tell him that he had a reason to act otherwise’, yet this can be reasonable, in cases of proleptic blame, even when ‘this may not have been true’ (ibid., 42). Nevertheless, to tell someone something you know is not true is to misrepresent things to them. Telling someone that p is fitting, in the present sense, only if it is true that p; but that is compatible with the existence of cases in which it is reasonable to tell someone something that is not true. Likewise, faith in someone to φ is fitting only if they are trustworthy to φ; but that is compatible with the existence of cases in which it is reasonable to have faith in someone who is not trustworthy. Of course, you might at this point wonder whether proleptic faith is possible. You might think that, in the present case, you do not really have faith in me to bring dessert, but merely say that you have faith in me, in the hope that your saying this will have certain good effects. And that is fine, and all the better for the view that fitting interpersonal faith presupposes trustworthiness.

Is God trustworthy?

I argued that having faith that p is having faith in someone to bring it about that p. I then argued that having faith in someone to φ is fitting only if they are trustworthy to φ. Thus, having faith that p is fitting only if the one in whom you have faith to bring it about that p is trustworthy to bring it about that p. In this sense, propositional faith presupposes trustworthiness. It is thus incoherent to have faith that p if you know that the one in whom you have faith to bring it about that p is not trustworthy to bring it about that p. ⁸ And we can expect faith that p to be extinguished when you come to know that the one in whom you have faith to bring it about that p is not trustworthy.

Above, I suggested that propositional faith sometimes involves faith in God to bring something about. If interpersonal faith presupposes trustworthiness, whether propositional faith in such cases is fitting depends on whether God is trustworthy to bring about the truth of the relevant proposition. This raises some important theological issues.

First, you might think that God is not a person,⁹ and therefore not an agent, and therefore incapable of motivation.¹⁰ If God is incapable of motivation, then God is not trustworthy to do anything, in the present sense of ‘trustworthy’. It follows, on the view that interpersonal faith presupposes trustworthiness, that it is not fitting to have faith in God to bring anything about, and it is incoherent to have faith in God to bring it about that p while believing that God is incapable of motivation.

The argument just sketched assumes that, if God is not a person, God is not an agent. This assumption could be challenged. Consider the social trinitarian view that God is a group of people – a society or community or family of people. Critics of social trinitarianism argue, among other things, that the view implies that God¹¹ is not a person.¹² That implication could be challenged,¹³ but let us suppose that God is a group of people and not a person. It seems to me that groups can be agents – that groups can do things and be motivated to do things. How group agency is to be understood is, of course, a subject of controversy in the philosophy of action.¹⁴ But it seems clear enough that groups of people – including societies, communities, and families – can do things and be motivated to do things. For example, my family went on vacation last year to San Diego – and that is not merely an analogical way of speaking or ellipsis for the claim that we each went on vacation.¹⁵

However, let us grant that, if God is not a person, God is not an agent. If interpersonal faith presupposes trustworthiness, then it is not fitting to have faith in God to bring anything about.

Does the idea that God is not a person undermine the argument that many cases of faith that p are plausibly interpreted as cases of interpersonal faith in God? No. Those cases involve people who believe that God is a person – or, at least, who thus represent God inasmuch as they have faith that God will bring something about. If you believe that God is not a person, then you cannot coherently have faith in God to bring it about that p. As I argued, above, if you have faith that p, there must be someone in whom you have faith to bring it about that p. If you believe that God is not a person, then, on pain of incoherence, that someone cannot be God.

Second, you might think that, although God is an agent, they are indifferent to certain things that we care about, such that they are not motivated to bring about what we want them to bring about. Imagine that you have faith that the Celtics will win the NBA Championship, which faith involves trusting God to bring it about that the Celtics win the NBA Championship. Even if God is an agent – and even if God is perfectly good, an absolutely good moral agent, omnibenevolent, or whatever – it is easy to imagine that God is indifferent to who wins the NBA Championship, and thus not motivated to bring it about that the Celtics win. Of course, you may very much want the Celtics to win and you may be relying on God to bring it about that they do. But it is easy to imagine that those facts do not motivate God to bring it about that the Celtics win. There may be just as many fans of the Celtics’ opponents who feel the same way about the Celtics losing as you feel about their winning. It is easy to imagine that God is indifferent to our interests grounded in taste or partiality, thinking that it is none of their business to get involved or to take sides, even if we cannot but imagine that God is motivated to console and comfort the frustrated fan when the Celtics lose. In any event, if God is not motivated to bring it about that p, because they are indifferent to whether p, then they are not trustworthy to bring it about that p.

Third, you might think that, although God is an agent who cares about what we care about, they are not motivated by the fact that we care about it. Imagine that you have faith that the endangered snow leopard will survive, which faith involves trusting God to bring it about that the snow leopard will survive. It is easy to imagine that although God is motivated to save the snow leopard, they do not treat the fact that you are relying on them to save the snow leopard as a non-instrumental reason to do so. Your interest, in this case, is not grounded in taste or partiality. You may want God to save the snow leopard for the same reason that God is motivated to save the snow leopard: because it would be bad were the snow leopard to go extinct. In any event, if God is not motivated by the fact that you are relying on them to bring it about that p, because their motivation has some other source, then they are not trustworthy to bring it about that p.

Fourth, you might think that, although God is an agent who cares about what we care about, they are disposed not to goodwill but to malevolence. Philip K. Dick describes a religious experience in which he saw ‘a vast visage of perfect evil’. He continues: ‘It was immense; it filled a quarter of the sky. It had empty slots for eyes, it was metal and cruel and, worst of all, it was God’ (Dick (1987), 377). This is, of course, neither a traditional nor a popular conception of God. But it underscores the theological assumptions required, on the view defended here, if faith in God to bring things about is fitting.

All this illustrates that the question of whether God is trustworthy is distinct from the question of whether God exists. It is possible to doubt God's trustworthiness without doubting God's existence. The question of whether God is trustworthy is also distinct from the question of whether it is good that God exists or would be good if God existed (cf. Kahane (2011)). It is coherent to think that the existence of a non-personal God is better than the non-existence of God. Even if you assume that God is a maximally good moral agent, it does not follow that God is trustworthy to bring about everything you might hope for – like the Celtics winning the NBA Championship. Propositional faith that involves faith in God to bring it about that p requires not only that God exists, but that God is trustworthy to bring it about that p. The theist who has faith of this kind faces not only the question of God's existence, but the distinct question of God's trustworthiness to bring it about that p.

Conclusion

In this article, I argued that propositional faith requires trustworthiness, in other words, that faith that p is fitting only if the one in whom you trust to bring it about that p is trustworthy to bring it about that p. In defence of this, I argued that propositional faith entails interpersonal trust. An important consequence of the conclusion that propositional faith requires trustworthiness is that it is fitting to have faith that God will bring it about that p only if God is trustworthy to bring it about that p. However, the trustworthiness of God, in the relevant cases, is not obvious.