Conceptualizing divine trust

Abstract

Does God trust human persons? Very little in philosophy of religion has been written about God's trust, which seems striking for two joint considerations. First, many of the Abrahamic faith traditions posit that union and close personal relationship with God is the telos of human life. Second, trust seems to be an essential element in ideal, close relationships between persons. While there is much in the faith literature that emphasizes the role of trust on the human side of the divine–human relationship, there is very little on divine trust. To fill this lacuna, this article addresses the conceptual issue of how divine trust could be understood within the Abrahamic faith traditions (particularly in Christianity and Judaism). I begin by examining whether an account of divine trust can be developed alongside divine attributes like divine foreknowledge. After identifying some plausible conditions of trust within the philosophical literature, I present a couple of trust scenarios as a means of demonstrating that divine trust is not only conceptually plausible (i.e. compatible with divine foreknowledge), but that divine trust is best construed as a particular trust type – therapeutic trust. That is, I argue that divine trust aims at inspiring humanity's trustworthiness.

Introduction

Does God trust human persons? Very little in philosophy of religion has been written about God's trust, which seems striking for two joint considerations.¹ First, many of the Abrahamic faith traditions posit that union and close personal relationship with God is the telos of human life. Second, trust seems to be an essential element in close relationships between persons. To be sure, there is much in the literature on faith that emphasizes the importance of trust on the human side of the divine–human relationship, but given that ideal, close personal relationships require mutual trust and trustworthiness, it would seem that this same consideration about trust would extend to God's side of the relationship (even if it be in some asymmetric or qualified sense). Recently, however, interdisciplinary research on faith (pistis) has laid the groundwork for considering the nature of God's trust and its role in divine–human relationships. To that end, this article seeks to build upon this research and raise the conceptual issue of how divine trust could be understood within the Abrahamic faith traditions (particularly in Christianity and Judaism). That is, I raise the question of whether the contemporary philosophy of trust literature offers any accounts of trust that are compatible with divine attributes (especially God's exhaustive foreknowledge).

A detailed outline of the article is as follows. In the first section, I offer some preliminary remarks about faith, trust, and the relation between trust and close personal relationships. In so doing, I mean to both clarify the context of our enquiry and motivate interest in an account of divine trust. In the following section, I give a very brief overview of the trust literature in philosophy and the ‘reliance plus’ family of trust accounts that seem to offer broad characterization of how trust is understood. In this survey of the literature, I identify two conditions that seem necessary for standard accounts of trust (SAT): (1) reliance and (2) vulnerability to betrayal. I conclude this section by raising the two questions central to our enquiry. First, how can we understand divine trust (especially in a way that is consistent with the conditions of trust)? Second, is God's exhaustive foreknowledge compatible with the conditions of trust? In preparation for answering these questions, I move on to the third section where I review some of the various ways our trust conditions are understood in the philosophical literature.

I raise two trust scenarios in the penultimate section. Each scenario presents a different premise regarding God's foreknowledge (a foreknowledge of P's trustworthiness and a foreknowledge of P's failure to be trust-responsive). I review each scenario in turn, examining whether we can ascribe trust to God in a way that is in keeping with our conditions of trust in SAT (as detailed in the third section). In so doing, I develop an account of God's risk (compatible with God's exhaustive foreknowledge), arguing that an account of divine trust is plausible and best understood in terms of a particular trust type – therapeutic trust – whereby God trusts with the aim of inspiring a trustee's trustworthiness. At the end of the penultimate section, we raise a question in the second scenario, which we seek to answer in the final section. Namely, why say that God can trust in particular instances in which God knows a trustee will fail to prove trustworthy? Answer: God's therapeutic trust can be best understood in terms of practical aims – namely engendering the trustee's trustworthiness. As a result, the aim of divine therapeutic trust can extend beyond particular trust interactions. Furthermore, in answering why God would therapeutically trust humans, I argue that God's therapeutic trust is voluntary and proffered for humanity's sake – namely for humanity's moral transformation and reform.

Preliminary remarks about faith, trust, and close personal relationships

Jointly, I take the following considerations to be vitally important for the Abrahamic faiths (especially the Christian and Jewish traditions):

1. (I) Faith entails trust.

2. (II) The telos of the life of faith is close personal relationship (and union) with God.

3. (III) Trust is an essential element in close personal relationships.

Considering (I), some might take ‘faith entails trust’ to be a platitude. However, there is a fair amount of precedent for (I) as a well-argued conclusion in recent analyses of faith in the philosophy of religion literature.² Additionally, there is a lot of recent research in classics and biblical studies in support of making this relational and fiducial connection to the pistis/fides lexica (the Greek and Latin transliterations for ‘faith’ and its cognates).³ I do not have space to unpack those arguments here, but I will mention an example shortly. For the purposes of this article, I am assuming that faith entails trust.

There are many theological figures to reference in support of (II). Let's look at one example from the Western Christian tradition. In numerous places throughout Summa Contra Gentiles, Thomas Aquinas thoroughly develops the notion that ‘the desire of man naturally tends towards God as towards the ultimate end’.⁴ In her seminal monograph, Wandering in Darkness, Eleonore Stump makes extensive use of Aquinas's theology on this point – ‘[For Aquinas,] . . . the ultimate good for human beings is union with God . . . Union with God is shareable, and persons united with God are also united with each other’ (Stump (2010), 95).⁵ Stump emphasizes that this union is ongoing – especially as it plays out in the redemptive processes outlined in the Christian scriptures – for ‘the goal . . . is not growth in some non-relational inner state such as virtue . . . the goal is the establishment and deepening of a relationship of love between two persons, one human and one divine’ (ibid., 408). Similarly, Jordan Wessling argues that the nature of God's love appears to contain some kind of intimate relational or interpersonal component . . . and some kind of desire for union’ (Wessling (2020), 39–40). Wessling notes that ‘a fairly common way of making sense of love's union is in terms of a kind of friendship where there is mutual affection and reciprocal concern for one another as well as a sharing of purposes’ (ibid., 57–58). So, for Stump and others, the divine–human relationship is intended to be an ongoing union of love. Understood this way, the telos of faith is not some crescendo to a final endpoint; rather, faith establishes and continually facilitates an ever-deepening relationship with the divine – and it is in this end that human flourishing is grounded. Oliver Crisp likens the trajectory of this union to ‘a mathematical asymptote, where a curve is on a trajectory toward a line, though the two never finally intersect’ (Crisp (2019), 203).

What I find left wanting in so many of these accounts of love and union with God – especially those which would articulate this union in terms of ‘shared’ or ‘intimate’ experiences which bond persons to God – is that there is no discussion about the role of mutual trust.⁶ More particularly, there's no discussion about the unitive function or role of trust in these accounts of love and union with God (especially on the divine side of the relationship). Curiously though, when we turn to the social psychological literature on trust and close relationships – admittedly in human relationships – discussions are legion about the unitive role and function of trust in close personal relationships. For example, trust in close relationships bolsters loves for intimates and it prompts loved ones to adopt morally decent actions and attitudes (Preston-Roedder (2018), 185–187; Jollimore (2011), 25–26 and 151) and trust inspires trustworthiness. Trust also creates unity, strengthens relational bonds, and establishes a sort of relational solidarity ( Tsai (2017), 109, 111–112; Kuwabara (2011), 560 and 576–577). All the same, trust is not just something that benefits close personal relationships. Mutual trust is essential to those relations.

We turn now to those considerations in support of (III) – namely trust is essential for close personal relationships.⁷ In the social sciences, researchers contend that ‘there are few constructs in the field of interpersonal relationships that are more central or important to relationship functioning and outcomes than trust. Without trust, voluntary relationships are not likely to develop, let alone grow or be maintained’ (Simpson (2007), 604). In order for such a close personal relationship to form, be sustained, and grow, both persons must trust one another – the relationship must be constituted by a sort of mutual or reciprocal trust and trustworthiness (LaFollette (1996), 114–119). In keeping with the empirical research, Hugh LaFollette similarly identifies trust as an essential element in forming, sustaining, and growing such intimate personal relationships. In such relationships, persons ‘relate to each other because of who they are’ (ibid., 112).⁸ As a result, these relationships ‘are marked by a mutual desire to promote the other's interests [and needs]’ (ibid., 10). Sometimes the literature refers to these relationships as ‘thick’ personal relations, where the modifier ‘thick’ refers to the shared history, shared goals, and shared bonds of attachment that persons have developed in their interactions with one another over time. This presumably includes co-operative interactions where persons, in faith and solidarity with their intimates, take up shared-projects with the intimates (often at great cost and risk) for the sake of supporting them and demonstrating that support.

In this article, I assume that analogous considerations about mutual trust extend to the divine–human relationship, albeit in importantly qualified ways.⁹ Admittedly, there is an important asymmetry to note here between the types of beings involved (e.g. Creator and creature(s)). However, in keeping with (II) and (III) above, there are several reasons to think that mutual trust, inclusive of divine trust, plays an analogous role and function in the divine–human relationship. I suspect this is especially the case in theological models of the divine–human relationship which emphasis human involvement, participation, and even co-operation with God in God's purposes for creation (of which humanity is an important part – especially in God's redemption, salvation, and restoration thereof).¹⁰ As mentioned, there is still an important asymmetry to what divine and human persons contribute to God's purposes for creation, but insofar as the divine–human relationship is initiated and guided by divine love, we should not be surprised to find mutuality and reciprocity extending to other relational virtues like trust.

In keeping with (I) and (III), the Hebrew Bible and Christian New Testament quite clearly call for human persons to trust (have faith) in God and be trustworthy (i.e. faithful) to God. Additionally, these texts readily appeal to God's faithfulness (trustworthiness) as the catalyst for human trust in God and the call for humanity's reciprocal trustworthiness (faithfulness) to God. But what of God's trust in human persons? Teresa Morgan (2015) makes a compelling case that divine–human relationships similarly centre on mutual trust and trustworthiness and that divine–human communities are formed and shaped by relationships of reciprocal trust and trustworthiness.¹¹ In particular, Morgan argues that the Septuagint and New Testament model what she terms an ‘economy of pistis’ (mutual trust and trustworthiness) in the divine–human relationship, where trust and trustworthiness originate from God and overflow through figures like Moses and Paul to the community members, who return it to God (Morgan (2015), 187–188, 216–217). In this way, these religious texts offer implicit examples of divine trust in humanity. In focusing on the importance of the pistis/fides lexica for the Christian tradition, Morgan appeals to many examples of God's trust in Paul (and his co-workers) in the Pauline epistles. By way of illustration, she highlights repeated references to God entrusting the apostle Paul with the euangélion (gospel) and the mission of proclaiming pistis to others (Morgan (2015), 216–217). Consequently, I am interested in making sense of this notion of God's trust in Morgan's model of ‘cascading pistis’ (Morgan (2020), 213). Going forward, I will use this particular example of God's trust in Paul with the euangélion as a springboard for a conceptual discussion about how we are to understand divine trust. With this example in mind, let's segue our discussion to how trust is characterized in contemporary philosophy.

Is an account of divine trust compatible with standard accounts of trust?

Standard accounts of trust

In keeping with the philosophical literature, this example of God's trust in Paul with the euangélion would be a three-place trust relation: some person, T, trusts another, S, for something, X (e.g. that T values or desires). Succinctly stated, T trusts S for X. In the philosophical literature, some characterize trust in terms of an attitude – cognitive, affective, or some combination of both – held towards a trustee.¹² Others understand trust in terms of ‘a relation that can hold distinctively among people’.¹³ In either case, philosophers, like Katherine Hawley, posit that ‘[t]rust is standardly thought to involve reliance, plus some extra factor’ (Hawley (2014), 5).¹⁴ Andrew Kirton notes that this ‘extra factor’ is typically understood in terms of ‘betrayability’ or ‘an ability to be betrayed by the one you rely on’ (Kirton (2018), 45).¹⁵ Here, ‘betrayability’ or ‘vulnerability to betrayal’ is thought to capture the element of risk that is unique to trust, especially in personal relationships. Moreover, it is trust's ‘betrayability’ which distinguishes it from mere reliance, which, according to Margaret Walker, ‘involves only a prediction that one can count on something's occurring’ (Walker (2006), 74). From these considerations we can identify what I will refer to as the necessary conditions of trust. Coupled with Teresa Morgan's example above, we can express the conditions of trust in the following notation: God, G, trusts Paul, P, for X iff,

1. (1) G is reliant (or relies) on P for X.

2. (2) G is vulnerable to P's betrayal with respect to X.

In such interactions, G and P are persons relating to one another in a way that is conducive for either forming, sustaining, or growing a close personal relationship with one another.¹⁶ I agree with Walker that ‘trust’ picks out a particular ‘interpersonal attitude or relation, a kind of anticipation that is distinctive to how a person looks at, and can look at, other people’ (ibid.). Accordingly, interpersonal trust ‘must involve expectations of others, attributing some kinds of awareness or motivation to them that a person, but not a thing, can have’ in proving trust-responsive (ibid.).

I will add several brief caveats to this ‘reliance plus’ model of trust. The goal here is to start with a relatively broad conception of trust, including both the attitudinal and relational accounts of trust in our gloss, so that we can work to a more specific account of divine trust. By beginning with this broad gloss, I have left unaddressed many of the initial considerations that are included in making sense of the sort of thing that trust is. For instance, I haven't narrowly construed trust in terms of belief or an affective attitude of confidence.¹⁷ I haven't made judgements about whether trust is best understood as a disposition or an action. I am also using our notation, ‘G trusts P with X’ so as to be as theoretically neutral as possible. As such, I leave it an open question whether the aforementioned notation is equivalent in meaning to ‘G entrusts X to P’ (i.e. whether the act of reliance in trust (entrusting) always presumes that the trustor has an attitude of optimism that the trustee will prove trustworthy).

Towards a coherent account of divine trust: identifying the questions central to our enquiry

With our conditions of trust outlined, we can transition to our central conceptual question – how are we to understand divine trust? I take our aim to be as follows: we will ascertain whether we can ascribe trust to a divine person in a way that is consistent with our two conditions of trust as we have broadly construed them in our standard accounts of trust (SAT). This will entail making sense of whether divine attributes like aseity, impassibility, and omniscience are compatible with the conditions of SAT.¹⁸

Let's briefly consider the first two attributes and our first condition of SAT – can God be said to rely on a human person? On some strict variations of classical theism, the notion that God is in a relation of reliance would be a major point of contention (if not outrightly denied). John Peckham (2021) conveys the strict classical theist's position rather perspicuously:

Divine perfection means that God is the greatest possible being. God exists necessarily and is who he is entirely of himself (a se), without dependence on anything else relative to his existence or otherwise (pure aseity and utter self-sufficiency). God ‘exists independently of all causal influence from his creatures’; creatures cannot impact God or his actions.¹⁹

If God were to be reliant on something or someone, there would be a sense in which God would depend on some being outside of Godself (to do that for which they were being trusted). For the strict classical theist, however, reliance and utter self-sufficiency are inconsistent. This view is partially motivated by a concern to safeguard divine completeness, unity, and perfection of being. For example, James Dolezal (2019) argues that ‘God cannot be the one whose greatness is beyond measure, and who is the absolute Creator on whom all creatures ultimately depend, if it turns out that He himself depends on his creatures’.²⁰

I will readily concede that the ‘strictest’ versions of classical theism quickly present an impasse for making a plausible account of divine trust (at least with respect to the way we have characterized trust so far). Consequently, I suggest we look to models which are ‘next of kin’, but still in the same family of views – that is, we will look to develop a plausible account of divine trust within neo-classical theism. How are classical and neo-classical models of God differentiated? R. T. Mullins (2020) notes that proponents of neo-classical models of God often reject or nuance one or more of the classical attributes.²¹ As Kevin Timpe observes, a common motivation for proponents of neo-classical theism is the common criticism that ‘it is impossible for God to [necessarily] possess all of the great-making attributes to the highest degree in the way . . . [that] classical models [intend to demonstrate]’.²² So, neo-classical models can be understood as revisionary attempts to reconstrue and group as many of the great-making attributes as are consistent. Like classical theists, neo-classical theists still aim ‘to give an account of the divine nature that preserve the central elements of perfect being theology’.²³ Granted, each are driven by different theological and philosophical intuitions about divine perfection.²⁴ Nevertheless, neo-classical models of God are in some nearby taxonomic space to that of classical models.²⁵

In keeping with certain versions of neo-classical theism, we can qualify our understanding of the divine attributes in such a way to offer a portrait of God that has parity with how we think about persons in close relationships. For example, some versions of neo-classical theism affirm that

God does not depend on or need anything with respect to his existence and essential nature (aseity and self-sufficiency), [but] God engages in genuine relationship with creatures that makes a difference to God (contra pure aseity). Further, . . . God experiences changing emotions (contra strict impassibility) and that God is immutable with respect to his character and essential nature but changes relationally (contra strict immutability) and therefore affirms some form of divine temporality (contra strict timelessness).²⁶

In these neo-classical models, God emerges with a dynamic mental life.²⁷ God cares (voluntarily) about creation (and God's relationship to it). As such, any account of divine trust, affectively construed, would be compatible with any neo-classical model that portrays God as having desires (analogous in some sense to human desires).²⁸ The specifics of this divine mental life would need to be worked out in accordance with however the great-making divine attributes are understood, and, at best, these models will be speculative. Important for our purposes is the following: Within neo-classical theism, there are variations of either qualified impassibility or qualified passibility that are congruent with these more moderate understandings of God's aseity and God's relationship to creation.²⁹

Neo-classical models also typically affirm a traditional understanding of God's omniscience, especially with respect to God's exhaustive foreknowledge of the future (thus distinguishing these models from open theistic models). Our account of divine trust will similarly assume that God's omniscience entails some form of exhaustive foreknowledge.³⁰

Regarding omniscience, however, is God's exhaustive foreknowledge compatible with the conditions of SAT – especially condition (2)? Well, prima facie, it seems unclear how we can posit both that ‘God trusts P for some X ’ and ‘God knows whether P will be trustworthy with respect to X ’. In a sense, it seems that God's foreknowledge mitigates (or potentially undermines) the riskiness of trust. Moreover, in trust between human persons, uncertainty seems like a background condition that constitutes the risk human persons face in trust situations (and relationships more generally). In such cases, trust is proffered as an imperfect solution. As Thomas Simpson says of co-operative situations requiring trust, ‘as well as a risk of loss, there is uncertainty as to whether those I interact with will be co-operative’ (Simpson (2012), 558). Here, it seems like part of the risk in trust extends to not knowing whether a trustee will prove trustworthy. This lack of certainty heightens the risk because there are no guarantees that the trustee will be trust-responsive once the trustor enacts their reliance, rendering themselves vulnerable to the trustee. In the case of divine trust, how could God, foreknowing Paul's trustworthiness in a given situation, be said both rely on Paul for something and risk Paul's betrayal? To answer these and other similar questions, we will next review how our conditions of trust are understood in the philosophy of trust literature.³¹

Reviewing the conditions of trust

SAT condition (1)

Let's review each condition of trust and survey the various ways each condition is understood in the trust literature. With respect to condition (1), let's say more about the sort of ‘reliance’ that philosophers have in mind. To say that a trustor relies on a trustee to do X, is to say that the trustor depends on the trustee to do X. So, for T to rely on S to do X, would be for T to depend on S for X, by allowing S to take care of X as entrusted.³² This notion of ‘allowing’ squares well with features of Annette Baier's (1986) influential account of trust. According to Baier, when the trustor enacts their reliance on the trustee for some X (i.e. something they desire or care about), the trustor affords the trustee the discretionary power to prove trustworthy with respect to X (Baier (1986), 237). So, allowing the trustee to prove trustworthy amounts to affording the trustee the discretion to prove trust-responsive. The trustor cannot be said to allow the trustee to prove trustworthy (with respect to X), until the trustor enacts their reliance, ‘putting’ X in the trustee's care. In so doing, the act of trust puts something the trustor cares about (or desires) within range of the trustee's ‘striking power’ (ibid., 235). In this way, we can see how the act of reliance renders the trustor vulnerable to the trustee (in keeping with condition (2) of SAT above). Joseph Godfrey's (2012) account of ‘reliance-trusting’ expands the three-place trust relation to a schema with five elements: ‘T trusts S with X, and has in view some outcome, O, because of some basis or reason, R’ (Godfrey (2012), 39).³³ In this way, Godfrey's schema expands the three-place trust relation to include both the trustor's expectation and reason(s) for trust (O and R can also double as ‘placeholders’ for the trustor's evaluative and explanatory considerations (e.g. T desires O (for X) because of R). I briefly mention Godfrey's schema here because it connects with some features of divine trust that we will discuss in later sections. For our present purposes, however, the three-place trust relation is still the basic core to SAT, Godfrey's schema, and the account of divine trust that we will look to develop. Trust theorists who prefer articulating trust in terms of a three-place relation (e.g. Baier) implicitly join both the good entrusted and the outcome desired and expected (ibid., 406); However, Godfrey ‘distinguishes the good entrusted . . . from the outcome desired and expected in order to take account of two ranges of discretion: the range of what one cares for, and the range of what counts as a good outcome’ (ibid., 39 n. 22 and 406–407).

Where does that leave us with respect to divine trust and condition (1) of SAT? The reliance relation that God has in certain instances of trust would be voluntary. God does not rely on anyone outside Godself for anything essential to God's being. Further, if one held the view that God opted to freely create the world and relate to it in such a way that entailed God freely opening Godself up to being affected by it (in a way that doesn't forego the Creator-creature distinction), then we could say that God could opt to be dependent upon human persons in select instances (in keeping with qualified passibility, for example). Nothing outside God would compel God to enter into reliance relations with human persons. God would be only reliant on human persons insofar as God voluntarily opts to enter into those relations.

In the case of divine trust, for God to rely on Paul, P, to do X (for example), would be for God to depend on P for X, by allowing P to do X. In these cases, God depends on P and allows P to do it, just so long as God does nothing to ensure or prevent that P takes care of X.³⁴ In keeping with Baier (mentioned at the beginning of this section), we could also understand God's reliance on P for X in terms of God giving P the discretionary power to prove trustworthy with respect to X. Consequently, God's enacted reliance on a given trustee could also help make sense of God's vulnerability to P because God's reliance would put X (i.e. something God cares about) within range of P's ‘striking power’ (Baier (1986), 235). So, it is in virtue of God's enacted reliance that the P receives the discretionary power, rendering God vulnerable to P. We'll say more about how this can be articulated on Godfrey's schema in a later section. For now, however, let's turn to condition (2) and how to understand ‘betrayability’.

SAT condition (2)

With respect to condition (2), there are two obvious considerations to bear in mind. First, the trustor being vulnerable to betrayal entails that the trustor, by enacting their reliance, is exposed to the possibility of betrayal. In the case of divine reliance, however, it will become clear in our trust scenarios that this first observation can be difficult to parse out. Nevertheless, for the purposes of this article we will assume that God can only be said to trust, if God is the sort of being that can be betrayed – that is, God can only be said to trust if God's exhaustive foreknowledge does not preclude God's ‘betrayability’. Secondly, we will assume that God's betrayal makes a difference to God and God's relationship to the trustee. We will review these considerations in greater detail in our trust scenarios. However, I first want to say more about the way in which we are understanding ‘betrayability’, which falls under the purview of some broader notion of risk. In betrayal, the trustor, in virtue of their reliance, exposes themselves to whatever would befall them should the trustee fail to take care of things as entrusted. So, in the case of divine trust, God's risk (in trust) would be understood as whatever God voluntarily exposes Godself to in an act of reliance – whether it be some goal going unrealized (at least in that particular moment) or one of God's desires going unfulfilled (at least temporarily). In many cases, it may be the combination of the two. God's risk of betrayal in trust relations is a more particular type of risk though. To capture this, let's briefly survey some ways in which betrayal is understood in the trust literature.

Betrayal (Account 1)

On many accounts of trust, being ‘vulnerable to betrayal’ stems from an expectation of trustworthiness; that is, the trust that one places in another is undergirded by the trustor's expectation that the trustee will be trustworthy. This notion of expectation also aligns with the ‘outcome’ component in Godfrey's trust schema. On some accounts of trust, this expectation can be understood in terms of attitudes like the belief that the other is (or will be) trustworthy – namely, that the trustee will come through for the trustor. Some trust theorists note that accepting that the trustee will be trustworthy would suffice (i.e. the trustor could act as if the trustee will be trustworthy because, for instance, they may be unsure about the trustee).³⁵

On some trust accounts, these expectations are also said to dispose the trustor to certain reactive attitudes depending on whether the trustee takes care of things as entrusted (e.g. feelings of ‘betrayal’ or ‘resentment’ when the trustee does not; feelings of gratitude when they do). On Holton's (1994) account of trust, the trustor's reliance on the trustee involves the adoption of a ‘participant stance’ which readies the trustor to feel betrayed, for example (Holton (1994), 66). Reactive attitudes, like betrayal, are generated out of this participant stance. Holton's account adapts P. F. Strawson's understanding of ‘reactive attitudes’. For Strawson,

the personal reactive attitudes rest on, or reflect, an expectation of and demand for, the manifestation of a certain degree of good will or regard on the part of other human beings toward ourselves; or at least on the expectation of, and demand for, an absence of the manifestation of active ill will or indifferent disregard. (Strawson (1974/2008), 15)

Strawson refers to these ‘participant’ reactive attitudes because ‘[they] belong to involvement or participation with others in inter-personal human relationships’ (ibid., 10). For that reason, betrayal can feel personal when this expectation in trust arises out of personal relationships – not just as an expectation that someone will do X, but that this person has developed a certain pattern of relating to me – they are a certain sort of way towards me (e.g. they act with my good will in mind) and so we continue to expect that (in virtue of our relationship and shared history). On some accounts of trust, these sorts of expectations are a product of obligations and shared commitments that arise in close interpersonal relationships.

Betrayal (Account 2)

On other glosses of betrayal or disappointment, the feeling or emotion that accompanies betrayal is not the essential feature. Reactive attitudes are merely dispositions that follow from perceiving and evaluating that one has been betrayed. However, a person can be unknowingly betrayed. For example, in an exclusive and committed romantic relationship between two people, one person can be unfaithful to the other, regardless of the partner knowing about the infidelity. On that account, reactive attitudes are not a necessary condition of betrayal. Rather, we can think of the betrayal in this exclusive romantic relationship, as the intentional acts of one partner, which undermine the commitment and exclusivity of the romantic relationship. Avishai Margalit describes acts of betrayal as an ‘ungluing’ of one's relation to the other, especially in the case of ‘thick human relations’ (Margalit (2017), 47). In any case, we still might think that a partner would be disposed to having certain reactive attitudes, should the infidelity be made known, but Margalit's account pinpoints something deeper to which reactive attitudes point in cases of betrayal. We need not pinpoint that to which reactive attitudes refer, but we have already gestured towards commitment as a candidate (in the aforementioned example of infidelity).³⁶ Following LaFollette, we will propose that insofar as betrayal undermines the commitment in the relationship it undermines the relationship itself:

for commitment is not something added to a personal relationship, it is a thread woven into the fabric of all close relationships. We do not enter an intimate relationship and then choose to be committed. To have a close relationship just is to be committed, at least in some minimal sense. (LaFollette (1996), 183)

On this second understanding of betrayal, betrayal amounts to acts (perhaps even thoughts or desires) that undermine any one of the essential, ‘good-making features of thick relations’ (Margalit (2017), 100). In virtue of this, some accounts of trust posit that betrayal only takes place within the domain of close interpersonal relationships (where there are commitments, etc.).

These two different ways of accounting for betrayal or disappointment are not mutually exclusive. Perhaps they are if we understand our second account of betrayal as offering something of a necessary condition for betrayal, while our first account of betrayal helps to pick out a unique behavioural disposition that follows from perceiving betrayal (pointing to the feeling that often accompanies betrayal). For now, though, we will leave it an open question as to how exactly these accounts fit together. As it stands, we have enough of a map to identify signposts as we navigate how we might understand betrayal in divine–human relationships. Now that we have completed our review of the conditions of trust, let's move on to our trust scenarios.

Trust scenarios, divine foreknowledge, and the ‘betrayability’ of divine trust

Now that we have a clearer picture of the various ways the conditions of trust are understood, let's look at two trust scenarios to see how divine trust might be conceptualized alongside God's exhaustive foreknowledge. As mentioned earlier, each of our scenarios will present, alongside the conditions of trust, a premise which articulates G's desire for X, which partially accounts for the risk in G's reliance on P (G risks a desire going unfulfilled). Our trust scenarios differ in their last premise, which details that God foreknows that the trustee will or will not prove trustworthy – (4) and (4′) below. The premises shared in each of our trust scenarios are construed as follows:

1. (1) G is reliant (or relies) on P for X.

2. (2) G is vulnerable to P not proving trustworthy with respect to X.

   That is,

   1. (2a) G expects P will prove trustworthy (from the accounts of betrayal in the previous section).

3. (3) G desires X (to some degree or another).

Our foreknowledge premises in our first trust scenario is:

1. (4) G knows P will prove trustworthy with respect to X.

The foreknowledge premise in our second trust scenario is:

1. (4′) G knows P will not prove trustworthy with respect to X.

In each case, we will look to demonstrate whether we can plausibly ascribe trust to God.

We have included (2a), from the previous section, to help mark out the conceptual space for detailing how G could be vulnerable to P's betrayal. With these components listed, let's survey some options for detailing how we could understand God's vulnerability in (2) and (2a), in light of God's foreknowledge and our other premises.

Divine trust – Scenario 1

In Scenario 1, how can God be vulnerable to P not proving trustworthy with respect to X, if God foreknows P will prove trustworthy with respect to X ? It seems God's foreknowledge mitigates some, if not all, facets of risk here. How is God vulnerable? In (2a), God's vulnerability in trust pertains to God's expectation that P will prove trustworthy. Given that God foreknows P will be trustworthy, it seems God could only be said to expect that P will prove trustworthy. If God can be said to rely on what God knows will be the case, would this not be a case of mere reliance (and hence, not trust) because God's reliance would be akin to a human person's predictive expectation that a certain state of affairs would unfold as anticipated? If we understood the divine mind and divine reliance as being undergirded by reasons or considerations that pertain to the interpersonal care of others (rather than ‘mere prediction’), then we might have the beginnings of a theory of divine trust. In other words, if there were separate, interpersonal considerations in the divine mind that were sufficient for enacting divine reliance (considerations which would allow for the possibility of betrayal), then such a divine reliance might reach the threshold of trust.

Should we think of God's foreknowledge in (4) as having some backward effect of constraining the sort of expectation God has? This raises the question of whether there must always be parity between God's expectation in (2a) and God's foreknowledge in (4). If expectation were to be solely understood in terms of some confidence or optimism that a certain state of affairs would occur, then God's expectation would need to be congruent with what God foreknows. However, not all accounts of trust construe it in terms of some affective attitude of confidence or cognitive attitude that a certain state of affairs will obtain.³⁷

What other options are there for how we understand the expectation in (2a)? Recall our earlier accounts of betrayal. The expectation in (2a) could be understood as a disposition to some reactive attitudes, which stem from relational commitments – especially if this reliance were understood in terms of (or in relation to) the good-making features of the divine–human relationship. One might think that foreknowledge eliminates the ability for a divine person to have or experience reactive attitudes though. That's not a problem for our second account of betrayal because reactive attitudes are not essential. Rather, they point to the good-making features of relationships that are essential, like commitments. If we understand the expectation in (2a) as being informed by relational commitments, then we will need to detail the influence of the latter on the former. First, the relational commitments would do more than predict the continuation of the relationship; the commitments would influence the type of expectations a person has on another with whom they are in a relationship. For instance, in our most ideal relationships, we do not merely expect that a person will not do things to undermine the relationship; we assume that our intimates will abide by standards of care which govern the sort of person they will be toward us. Stephen Darwall's account of trust provides a helpful articulation of this distinction between types of expectation (Darwall (2017), 36–37).³⁸ In particular, trust in our ideal, closest relationships is not merely an expectation that our friend, partner, etc. will prove trustworthy with respect to that for which we rely on them. Implicit in our trust is an expectation of this person – this person will be the sort of person who acts with our good will in mind. Elaborating in further detail, Darwall posits,

[e]xpectations that take propositions or possible states of affairs as objects, whereas the object of an expectation of is a person (or group or collective of persons). Expectations of both sorts impose standards. But where an expectation that says how things will be (or how we have reason to believe they will be), an expectation of says how someone should act or be. (ibid., 36)

For Darwall, this ‘expectation of’ is personal, pertaining to a moral standard by which we might hold another person (ibid., 36–37 n. 3).³⁹ Roughly put, there is a sort of ‘is/ought’ distinction that runs parallel to how ‘expectation that’ and ‘expectation of’ are distinguished. In the former, ‘expectation that’ pertains to the prediction of how it is that the trustee will act; in the latter, ‘expectation of’ pertains to how the trustee ought to act towards the trustor. Neither sense of expectation is entailed by the other.

Given this distinction from Darwall, we could understand the expectation in (2a) to extend to an assumed (or hoped for) standard of how the person (trustee) ought to act towards the trustor in the relationship – not a prediction of whether the trustee will do as entrusted. As such, (2a) could be nuanced within our earlier trust scenario to include ‘G expects of P to prove trustworthy with respect to X’ – see (i) below. A mere ‘expectation of’ is not sufficient for demonstrating vulnerability when relying on another for something though. For example, parents can have an ‘expectation of’ their children to care for and be kind to their siblings. However, this does not entail that the parents would rely on their eldest child, for example, to take care of their younger siblings (if they wanted to go out for dinner for the evening).⁴⁰ So, an ‘expectation of’ need not entail an ‘expectation that’ the oldest child, for example, would take care of their younger siblings, should the parents leave the siblings in the eldest child's care. What else is needed alongside this ‘expectation of’ to make one's reliance vulnerable to betrayal? In order to say that the parents trusted their eldest child to take care of the child's siblings (and were thus vulnerable to their eldest not doing as entrusted), the ‘expectation of’ would also require some action on the part of the parents, which both communicates their trust and renders them vulnerable to whatever would happen should their eldest fail to take care of their siblings. This act of reliance is what actualizes the riskiness of their trust.

So, in addition to this ‘expectation of’, a trustor must enact their reliance. In many instances, an ‘expectation of’ may coincide with an ‘expectation that’ the trustee will take care of things as entrusted. However, the point here is that this needn't be the case. For instance, a parent may entrust the family's vegetable garden to their child's care for the purpose of helping the child to mature and develop certain virtues. In such a scenario, the parent need not ‘expect that’ the child would take care of things as entrusted though. They may be unsure about whether their child could take care of the garden, tending to it as required. Nonetheless, the parents can still opt to rely on the child to care for the garden, acting as if the child would do as entrusted. The act of relying on the child entails that the parents allow the child to do as entrusted (as opposed to intervening to ensure that the child does what they have been entrusted to do). In this way, the child is given the discretionary power to prove (or not prove) trustworthy. In actively relying on the child (affording the child the discretion to do as entrusted), the parents render themselves vulnerable to the child failing to prove trust-responsive.

Returning to our first trust scenario, we can add the ‘expectation of’ and ‘discretionary power’ or action premise as sub-components to (2) and (2a), rendered as follows:

1. (2) G is vulnerable to P not proving trustworthy with respect to X.

   That is,

   1. (2a) G expects P will prove trustworthy (previous section).

      That is,

      1. (i) G expects of P to prove trustworthy with respect to X.

      2. (ii) G gives P the discretionary power to prove trustworthy with respect to X. (G affords this discretionary power by enacting their reliance)

Here, God's risk (i.e. vulnerability) is reified by the combination of these two additional sub-conditions in (2a). By enacting this reliance, God's risk becomes actualized because the act of relying on the trustee gives them the ‘discretionary power’ to prove trustworthy.

Earlier we mentioned, that (3) – God desires X – also partially figures into whether God is vulnerable. Given that God's desire comes to fruition in our first trust scenario, God (in virtue of foreknowing that P will prove trustworthy) is not confronted with the desire going unfulfilled, so God is not left exposed on that front. At the same time, however, God is not the cause of P taking care of X as entrusted. God was vulnerable in the sense that he was susceptible to whatever the trustee chose to do. God's act of reliance put the ball in the court of the trustee – so to speak. God was dependent upon them to do what God desired to come to pass.

In the first three sections, we used the example of Paul of Tarsus (i.e. Paul being entrusted with the euangélion) to begin developing our account of divine trust. In accordance with that example and these additional features, God's trust of Paul with the euangélion would entail that Paul was given the discretionary power to prove trustworthy with a particular task that God desired Paul to carry out. If we understand the Paul–God relationship to be one constituted by a sort of mutual love and care, then we would understand the bonds of the relationship as being constituted by certain normative expectations that follow from relationships of mutual care – i.e. relationships of care are undergirded by obligations to support the beloved in things that they care for or desire. This relates to the virtue of support and relational solidarity in relationships with intimates. What makes God vulnerable is God's act of relying on Paul to do as entrusted. God's act of reliance on Paul, which bestows upon Paul the discretionary power (i.e. opportunity) to prove trustworthy, is not something that God's foreknowledge causes to bring about. It is up to Paul to do as entrusted. In this way, God's trust in the divine–human relationship puts something that God cares about ‘within striking range’ of human persons – an invitation accompanied by the risk of unrequited trust. In our second trust scenario, we will examine an alternative scenario in which God's desire goes unfulfilled. Let's turn to that scenario now.

Divine trust – Scenario 2: why would God trust?

In the previous section, we noted that God's vulnerability to betrayal falls under the purview of some broader notion of risk. Again, that quick gloss of the risk of betrayal hinted that there may be instances in which God's desires go unfulfilled. Our first trust scenario did not accommodate that gloss because God's desire for X came to fruition in (4). In this second trust scenario, we will review a case in which God desires X, knows that P won't prove reliable with respect X, but still (voluntarily) opts to enact reliance on P for X (allowing the desire to go unfulfilled). Our goal here will be to discern whether it makes sense to say that God's reliance can be betrayed and count as a genuine instance of trust. Let's list each component of this second trust scenario (as well as our nuanced (2a) from our first trust scenario):

1. (1) G is reliant (or relies) on P for X.

2. (2) G is vulnerable to P not proving trustworthy with respect to X.

   That is,

   1. (2a) G expects P will prove trustworthy. (previous section)

      1. (i) G expects of P to prove trustworthy with respect to X.

      2. (ii) G gives P the discretionary power to prove trustworthy with respect to X.

3. (3) G desires X (to some degree or another).

4. (4′) G knows P will not prove trustworthy with respect to X.

One might immediately notice that (2a) and (4′) present an inconsistency. However, it is only an issue if we understand the expectation as optimism that a certain state of affairs will obtain (as we initially reviewed in the first scenario). To resolve the implicit inconsistency, we can employ our earlier distinction between ‘expectation of’ and ‘expectation that’. That is to say, God could have an ‘expectation of’ P's trustworthiness even if God knows that P won't prove trustworthy with respect to X. Again, we could say that God trusts P for X when God expects of P to prove trustworthy with respect to X and God gives P the discretionary power to prove trustworthy. So again, we can understand God's vulnerability in terms of God's act of reliance on P and expectation of P to be certain sort of person. One might note that in this case, God giving the P the discretionary power to prove trustworthy entails that God ‘acts as if’ P will prove trustworthy (at least in certain circumstances). As such, some renderings of (2a), may include (iii), rendered as follows:

1. (2a) G expects P will prove trustworthy. (previous section)

   That is,

   1. (i) G expects of P to prove trustworthy with respect to X.

   2. (ii) G gives P the discretionary power to prove trustworthy with respect to X.

   3. (iii) G acts as if P will prove trustworthy with respect to X. ⁴¹

      (Here, G acting as if is what enacts reliance on P, affording P discretionary power.)

Note, we have added ‘acts as if’ because God foreknows the trustee won't prove trust responsive. On account of that, God is vulnerable to the trustee not just in virtue of affording the trustee the discretion to prove (or in this case not prove) trust-responsive. God actually faces a situation in which God's desire goes unfulfilled in virtue of trusting P. So, God's vulnerability here is not just understood in terms of the act of reliance, but also God rendering Godself vulnerable to a desire going unfulfilled (‘expecting of’ P to be the sort of person who is trustworthy with respect to X), despite foreknowing P will not prove trust-responsive.

In response, we might anticipate two interrelated questions – even if this were a coherent construal of our second trust scenario, why construe divine trust this way? Further still, why would God act as if P were trustworthy, giving P the discretion to prove trustworthy, when God foreknows P will fail to prove trustworthy?⁴² Answer: God's enacted reliance could be an act of trust that aims at inspiring the trustee's trustworthiness. Someone may incredulously reply that trust does not inspire trustworthiness if it does not result in trustworthiness – as would be the case in our second trust scenario. For now, we will earmark that objection and address it in the next section. A more complete explanation involves an appeal to what is called ‘therapeutic trust’ – a specific type of trust which aims at inspiring a trustee's trustworthiness.⁴³ In the next section, we will provide some more details about therapeutic trust and offer an explanation for why God could trust human persons in this way (in keeping with the reason, R, for divine trust in Godfrey's (2012) aforementioned trust schema).

Towards an account of divine therapeutic trust

Therapeutic trust shares many of the same features as our standard accounts of trust. It can similarly be characterized in terms of a three-place trust relation – G therapeutically trusts P for X. In therapeutic trust, the trustor also relies on the trustee for something and risks the trustee failing to prove trustworthy. Furthermore, the account of therapeutic trust that we will review connects well with our ‘trust scenarios’ and provides an explanation as to why God would therapeutically trust in a particular instance – even when the person may (initially) fail to be trustworthy.

In contrast to standard accounts, therapeutic trust's aim is what makes it a distinctive trust type. Namely, a trustor, G, therapeutically trusts the trustee, P, for X with the aim (or purpose) of promoting the trustee's trustworthiness.⁴⁴ According to Trudy Govier,

therapeutic trust is based on the assumption that people who are explicitly entrusted with certain tasks or goods will feel an obligation to live up to the expectation of others, and [feel] guilt if they do not do so. It is based on the human desire to reciprocate goodness and to live up to what others expect. (Govier (1998), 173)

In explaining why God would opt to trust in our trust scenarios, we will understand divine therapeutic trust as having a similar basis. How so? Recall our second trust scenario. There we said that God can therapeutically trust someone – despite foreknowing they won't be trustworthy in a particular instance – because divine trust has the aim of inspiring trustworthiness. This aim need not be limited to any singular trust interaction; in fact, God's aim to inspire trustworthiness can be understood as a long-term goal, undergirded by practical reasons for therapeutically trusting human persons.

Moreover, the divine–human relationship, as a sort of close personal relationship, provides a plausible context in which God can repeatedly rely on humanity, over an extended period, with the long-term aim of inspiring humanity's trustworthiness. In such a relational context, even if the aim of therapeutic trust isn't initially achieved, the trustor could still opt to enact their reliance in future instances for several reasons. For instance, the trustor's reliance could be understood as a continued appeal for the trustee to act in accordance with the standards of care that govern and maintain the relationship. In keeping with Govier's observation above, it seems like therapeutic trust would flourish in interpersonal contexts – especially when a trustor has the recourse to appeal to the norms and standards that govern relational reciprocity and the obligations of care and goodwill that persons in close relationships have to one another. By way of example, the following considerations could undergird a person's long-term therapeutic trust. By voluntarily relying on a trustee, a trustor could hope that the benefits that accompany therapeutic trust either (i) inspire the trustee to prove trust-responsive or (ii) give rise to feelings of regret in the trustee (when they fail to prove trustworthy). In the latter case, the trustor could be hoping that the trustee's regret could motivate the trustee not to exploit trust in the future. In this way, the trustor could recognize that a particular failure to prove trustworthy need not end in vice; instead, the regret could serve as a stepping stone to inculcating habits of trustworthiness.

In the trust literature, there are often examples of long-term therapeutic trust employed as a means of demonstrating that the virtue of trustworthiness is a habit formed over long periods of time. Accordingly, we often see examples of therapeutic trust where the trustor is looking to aid the trustee's moral development or reform – for instance with ‘children, ex-criminals, petty offenders, lapsing partners, students, employees learning new tasks’ (ibid., 173). In such cases, therapeutic trust is ‘understood as an act of entrusting’ and it can be an unobjectionable means to this sort of moral development – unobjectionable ‘provided we assume that the expectations they are encouraged to live up to are reasonable and right, and the risks to third parties are kept at an acceptable level’ (ibid.).

Given these considerations, our ‘trust scenarios’ can be understood as instances of divine therapeutic trust. In our second scenario, the trustee's failure to prove trustworthy doesn't entail that divine trust is unreasonable. Instead, divine trust can be understood in terms of some long-term aim of inspiring P's eventual trustworthiness (with respect to X). As such, divine therapeutic trust can be understood as even anticipating particular instances in which a trustee fails to prove trust-responsive. Additionally, divine therapeutic trust similarly relates to the moral development of humanity. This aligns with many of the ‘redemptive’ moral arcs traced in the sacred texts of the Abrahamic faiths where humanity is portrayed as needing moral transformation, development, and reform. The portrayal of this need is not only compatible with an account of divine trust – the divine–human relationship arguably requires it. For if humanity is never afforded the opportunity to develop trustworthiness, it can be difficult to account for how humanity develops the virtue of trustworthiness – especially in relationship to the divine.

On top of aiming at moral development, divine therapeutic trust can be understood as an act of care or nurture – especially in the way it can be understood as guiding the trustee's character development. How could divine therapeutic trust be an act of care? According to Govier, in therapeutic trust, the trustor can

[convey] a positive idea of what the other is and can become. . . . [P]eople who take on caring roles (parents, teachers, and counsellors being prime examples) [can] convey in their words and actions a commitment to the positive potentialities of the person cared for. This commitment emerges not only in such specific acts as pointedly relying on him or saying, ‘I'm trusting you’, ‘I'm counting on you’, and so on, but also in dispositions to interpret positively the actions and motives of the one cared for. [It] emphasizes the importance of conveying a positive image of the other person and in interpreting what he says and does in a positive light. (ibid., 173–174)

In such cases, the trust extended need not be predicated on the belief that the person is trustworthy (one may be unsure about the person's trustworthiness). Therapeutic trust can, as Victoria McGeer says, give the trustee ‘a hopeful vision of [themselves]’ and a glimpse of what they could be and ‘achieve in the context of trust’ (McGeer (2008), 247 and 250). For instance, when someone therapeutically trusts you for something and you realize that you haven't had a prior opportunity to demonstrate your trust-responsiveness to them, their trust can invoke a sense of privilege or honour (even gratitude) because they still afforded you the opportunity to prove trustworthy. An oft-cited example of this comes from Victor Hugo's Les Misérables. In the narrative, a bishop offers Jean Valjean, a fugitive and convicted thief, silver, telling him to use it to live an honourable life and become an honest, hardworking man. This act of trust is one that inspires Jean Valjean to do just so.⁴⁵ In keeping with that example, we see that therapeutically trusting another can also provide, in McGeer's words, a ‘hopeful scaffolding [that] can . . . serve as a very powerful mechanism for self-regulation and development’, inspiring the trustee to be as the trustor already sees them (ibid., 249). In turn, this may motivate the trustee ‘to be more trust-responsive so that they can maintain the respect and esteem of the trustor’ (Pace (2021), 11907).

These are examples of the benefits of trust that can obtain in cases of therapeutic trust. Some of these benefits can also feature as part of the mechanism for explaining how therapeutic trust inspires trustworthiness. For instance, therapeutic trust, in virtue of being grounded by practical reasons, is voluntary. When we couple this feature of being voluntary with the benefits that therapeutic trust affords, we can understand this trust type as having a gift-like quality (akin to an unearned favour), where the trustee recognizes that they've done nothing to merit trust. In turn, this feature of being ‘gift-like’ can similarly give rise to the feelings of obligation often associated with receiving a gift (or a favour).⁴⁶ This comparison between receiving therapeutic trust and receiving a gift points to another mechanism that can explain how therapeutic trust can give rise to trustworthiness. Namely, the benefits of therapeutic trust often generate gratitude and a desire to prove trustworthy (akin to a desire to not squander a gift).

Let's sum up this discussion. In our account of therapeutic trust, we have noted that it doesn't require a belief that the trustee will be trustworthy; thus, therapeutic trust can be undergirded by many different cognitive or affective states (including acceptance that). In this sense, it may be multiply realizable – at least with respect to the attitudes that are said to undergird it. Further, foreknowledge of whether someone will prove trustworthy does not prevent divine therapeutic trust from being extended or betrayed. Over and above that, we noted that the main difference between SAT and therapeutic trust is that therapeutic trust has the particular aim of engendering trustworthiness. Many also identify therapeutic trust as being voluntary in ways that standard notions of trust aren't. This feature of voluntariness is often related to the practical considerations that govern therapeutic trust (because it is not dependent upon a prior conceptions or expectations of the trustee's trustworthiness). As Collin O'Neil says, this feature of voluntariness is what gives therapeutic trust its gift-like benefits (O'Neil (2017), 81). These gift-like benefits in therapeutic trust uniquely generate and ground the obligations that arise from it (including moral obligations, on some accounts).

In keeping with our earlier trust scenarios (previous section), divine therapeutic trust would be rendered as follows:

1. (0) G aims to inspire P's trustworthiness.

2. (1) G is reliant (or relies) on P for X.

3. (2) G is vulnerable to P not proving trustworthy with respect to X.

   That is,

4. (2a) G expects P will prove trustworthy. (section ‘Reviewing the conditions of trust’)

   That is,

   1. (i) G expects of P to prove trustworthy with respect to X.

   2. (ii) G gives P the discretionary power to prove trustworthy with respect to X.

   3. (iii) G acts as if P will prove trustworthy with respect to X.

      (This amounts to G enacting reliance on P, as noted in the previous section.)

5. (3) G desires X (to some degree or another).

Furthermore, either ‘foreknowledge premise’ from our trust scenarios could be included:

1. (4) G knows P will prove trustworthy with respect to X; OR

2. (4′) G knows P will not prove trustworthy with respect to X.

At the beginning of this section, we noted that therapeutic trust could be a three-place relation. In the section ‘Reviewing the conditions of trust’, we also reference Godfrey's (2012) ‘reliance-trust’ schema, which added two additional elements – both an outcome expected, O, and a reason for trust, R. In the case of therapeutic trust, O could also map on to (2a) and all sub-premises. Moreover, (0) and (3) could jointly reflect R – the reason God would enact trust in someone.

Conclusion

In this article we have sought to develop an account of God's trust which is compatible with God's exhaustive foreknowledge and a broad construal of our trust conditions in SAT. In light of our trust scenarios, we also argued that God's trust is best understood as an act of reliance that aims at inspiring trustworthiness in human persons (rather than a confidence that a trustee will prove trust-responsive in a particular instance). In virtue of the fact that God's act or reliance issues forth from practical considerations (aiming at engendering human trustworthiness), we said that divine therapeutic trust always involves a reliance that God voluntarily opts to enact. In this way, God is never compelled to trust, but when God does so, it often affords certain benefits to the trustee (especially benefits that pertain to moral development and reform). Additionally, God's trust functions within the divine–human relationship to enhance the trustee (human persons). Thus, God's trust is for humanity's sake, and though it is a gift, God's reliance (therapeutic trust) is still betrayable because human persons can fail to prove trustworthy with respect to that for which God is relying on them. In many instances, the failure to prove trust-responsive can be characterized both (i) as an exploitation of God's trust (which undermines the commitments that constitute the divine–human relationship) and (ii) a betrayal of certain moral expectations God has of humanity.