Causal powers and concurrentist accounts of cooperative action

Concurrentism was first articulated in the context of scholastic Aristotelian theories of active and passive causal powers. Although the notion of causal powers fell out of favour in the wake of the sorts of criticisms raised by Malebranche and Hume, there has been a resurgence of such approaches to causation in the contemporary literature. Since concurrentism is primarily concerned with active causal cooperation between God and secondary causes, it will be helpful to begin by briefly considering the general conception of active causal power assumed by most of its late medieval and early modern proponents. Andrew Platt has recently summarized the key elements of that conception in terms of the following schema:

a exercises an active power φ to bring about a property P iff

1. (1) φ is an intrinsic feature of a;

2. (2) If a were in contact with a suitably disposed patient b, then (barring miraculous intervention and under the right circumstance) a would cause b to have P in virtue of having φ. (Platt (Reference Platt2020), 36)

Several aspects of this conception of causal power merit brief comment. First, causal powers are distinguished from their manifestations; the causal power is an intrinsic feature of the causal agent, whereas the manifestation of that power is found in the patient.Footnote ² Second, a manifestation can be thought of as roughly equivalent to the effect brought about by the exercise of an agent's power; however, since many effects are complex in nature, it may be more accurate to think of a manifestation as a particular contribution to an effect. Third, distinct causal powers are differentiated by their manifestations; as Stephen Mumford puts it, ‘each power is essentially, or necessarily, related to manifestations of a specific kind’ (Mumford (Reference Mumford, Beebee, Hitchcock and Menzies2009), 269).Footnote ³ George Molnar concisely summarizes the close connection between the second and third points:

Fourth and finally, the contemporary literature on powers has focused much attention on their dispositional nature, which is captured by condition (2) in Platt's schema. At least some of an agent's powers are intrinsic features that manifest themselves only under certain conditions, which are variously referred to as their activating, initiating, triggering, or enabling conditions.Footnote ⁵ Since a given power's activating conditions might never be met, it is entirely conceivable that an entity could possess dispositional causal powers that it never actually exercises.

According to concurrentism, although created agents have genuine active powers, they require God's concurrence to bring about any effects. However, concurrentists have struggled to explain clearly why secondary causes require such concurrence, as well as to offer a clear account of how God and secondary causes cooperate. Freddoso, who has offered some of the most detailed discussions of concurrentism to date, frequently contrasts the generality of God's contributions with the particularity of secondary causes; he claims that ‘the secondary agent acts by its created or natural powers as a particular cause of the effect, whereas God acts by his uncreated power as a general or universal cause of the effect’.Footnote ⁶ However, it is not entirely clear how this contrast is to be understood. It might be tempting to interpret it as the claim that God emanates a sort of general, indeterminate force or energy that secondary causes channel and direct in some particular way or other. But although Freddoso seems to endorse such an interpretation in one of his earliest discussions of concurrentism,Footnote ⁷ in his later, more detailed discussions he warns against it, noting that it would be a mistake to think of God's concurrence as ‘an “indifferent” influence that is somehow particularized by the secondary cause’ (Suárez (Reference Suárez2002), xcviii). He goes on to explain that according to concurrentism, ‘God's action and the secondary cause's action are one and the same action, and so just as the actions of secondary causes are obviously multifarious in species, so too God's concurrence varies in species from one circumstance to another’ (Suárez (Reference Suárez2002), xcviii). Indeed, Freddoso often describes God as ‘tailoring’ God's causal contributions to those of secondary agents in each particular instance, so the characterization of God's contributions as universal or general clearly cannot be meant to imply that they are indeterminate or non-specific.

What, then, is the point of the general–particular contrast? From what I can tell, it involves two closely related aspects. First, God's causal contributions are ubiquitous; any given secondary agent contributes to only a limited number of particular effects, whereas God cooperates universally or generally in every causal interaction. Second, and relatedly, Freddoso emphasizes that God and secondary causes exercise different causal powers. This point must be interpreted with care. I noted above that causal powers are typically differentiated by their manifestations and that each kind of power is understood as necessarily related to a particular kind of manifestation. However, if we interpreted the claim that God and secondary causes exercise different causal powers along these lines, it would imply that God and secondary causes make fundamentally different kinds of contributions to an overall effect, and that is an implication concurrentists reject. Indeed, according to Freddoso, dividing the effect into distinct parts that are separately attributed to God and secondary causes is one of the chief ‘pitfalls’ concurrentism must avoid (Freddoso (Reference Freddoso1994), 144). Concurrentists like Freddoso and Grant insist that God and secondary causes cooperate by both bringing about the entire effect, and Grant has recently introduced the hyphenated ‘co-operate’ as a technical term to refer to this concurrentist notion of cooperation without any division of labour (Grant (Reference Grant2019), 39). Thus, I suspect that when Freddoso asserts that God exercises a different causal power that is general or universal, he means to deny that the typical one-to-one correlation between powers and manifestations applies in the case of God's power. God's power is universal or general in the sense that it is not limited to producing any particular kind of manifestation, but rather, can produce any sort of manifestation whatsoever. However, in each exercise of this general power, God's contribution is tailored and adapted to the particular nature of the secondary cause, such that their two powers (one general, one particular) jointly produce one and the same manifestation.

The metaphysical objection to concurrentism

With this background in mind, we are in a position to understand the central point of what Grant has labelled ‘The Metaphysical Objection’. Concurrentists insist that although the contributions of God and secondary causes are jointly sufficient for their effects, neither contribution can be sufficient by itself (Freddoso (Reference Freddoso1994), 151–152). For if the contributions of secondary causes were sufficient, they would have no need of God's concurring with them; and if God's contributions were sufficient, then concurrentism would collapse into a form of systemic causal overdeterminism in which the actions of secondary causes, even if genuine, are entirely superfluous. Thus, concurrentism is committed to the following claims concerning secondary causes:

However, I have suggested that concurrentism cannot offer a coherent model of cooperative action that satisfies both SC₁ and SC₂'s requirements. If the contributions of secondary causes are genuine (as SC₁ affirms), then, as I put the point previously, ‘it would seem secondary causes should be able to accomplish something without assistance or concurrence’ contrary to SC₂ (Miller (Reference Miller2011), 8). In other words, there should be some aspects of an effect that secondary causes are capable of producing without requiring God's specific, ‘tailored’ assistance – viz. those aspects that are brought about through the manifestation of their powers. I elaborated on this point by noting that there seem to be three types of circumstances in which one agent requires assistance from another in order to bring about an effect. An agent may require assistance because:

1. (a) the agent lacks a kind of causal power that is necessary for bringing about the effect; or

2. (b) the agent is, for some reason, prevented from exercising one or more of the power(s) it possesses; or

3. (c) the agent has a kind of causal power required for the effect, but not to the degree required to bring about the effect. (ibid.)

The trouble for concurrentism is that none of these types supports a model of cooperative action that satisfies all of concurrentism's requirements.

Type (a) is unacceptable because it would inevitably lead to some version of cooperation by division of labour. That is, since causal powers are differentiated by their manifestations, saying that secondary causes always require divine concurrence because they lack a required kind of causal power that only God can provide would entail that God and secondary causes make different sorts of causal contribution, accounting for different aspects of the overall effect. Since concurrentism rejects cooperation by division of labour and insists that God and secondary causes must co-operate by both bringing about the entire effect, it would seem to be committed to affirming that God and secondary causes make the same kind(s) of causal contribution.

Type (b) involves the sort of assistance required when agents face what might aptly be described as ‘Ikea dilemmas’: although a single person may possess all of the kinds of powers needed to assemble a piece of Ikea furniture, in some cases it can be physiologically impossible for one person to exercise all of the requisite powers simultaneously without the assistance of another agent's ‘extra pair of hands’. But ‘Ikea dilemmas’ are an ad hoc way of requiring assistance, and thus cannot serve concurrentism's purpose of explaining why secondary causes would always need divine concurrence.Footnote ⁸

The fact that strong concurrentists want God and secondary causes to jointly produce the entire effect suggests that type (c) might be a more suitable model. Perhaps secondary causes have all of the kinds of causal power needed to bring about their effects, but simply do not have those causal powers to the requisite degree, so God's concurrence fills in what is lacking. Freddoso discusses an example in which an agent requires assistance to lift the back end of his car over a ridge of ice in his driveway. Neither he, nor his friend, are able to get car off the ground at all lifting alone; however, acting together they are able to lift it over the ridge. In this example, both friends act upon the car in the same way – that is, they exercise the same kinds of causal powers. Freddoso goes so far as to suggest that we should not think of them as performing two separate acts, but rather as jointly performing a single unitary action. Furthermore, he claims that this model is,

Admittedly, this model does seem to satisfy SC₁'s requirement that secondary causes make genuine contributions. However, whether it also satisfies SC₁'s requirement that their contributions be non-superfluous is less clear. One obvious disanalogy between Freddoso's example and God's cooperation with secondary causes is that unlike each of the friends cooperating to lift the car, God is perfectly capable of bringing about effects without the contributions of secondary causes. Perhaps a better analogy for God's concurrence with secondary causes might be a parent's assisting a young child in lifting an object that is too heavy for the child to lift, but not too heavy for the parent. Nevertheless, the parent might limit her causal contribution, putting forth enough effort to overcome the child's lack, but not so much that the child's effort makes no contribution to the effect. By limiting her own contribution the parent can lift with the child rather than merely lifting it for or instead of the child. Thus, although the child's contribution may be entirely unnecessary in relation to the parent's ability, the parent ‘tailors’ her contribution to the child's in a way that avoids rendering the child's contribution superfluous. Interpreted in this way, Freddoso's model can satisfy both of SC₁'s requirements.

However, it runs into more serious difficulties with regard to SC₂. If both friends are making genuine contributions to lifting the car, then each of them should be able to lift many lighter objects without assistance. The same holds for the parent and child in the modified example, and there is no apparent reason why it should not also hold for divine cooperation with secondary causes. If secondary causes require assistance because they have the requisite kinds of power, but not to the requisite degree, the model can provide no basis for SC₂'s insistence that secondary causes must always require assistance. Thus, I previously observed, ‘claiming that I have all of the causal powers needed to lift things, but that there could never be anything, no matter how light, that I could lift without assistance … seems flatly self-contradictory’ (Miller (Reference Miller2011), 9). In short, if the need for assistance is grounded in a limited degree of a causal power, it cannot plausibly be claimed to be universal.

Thus, none of the three ways of requiring assistance seems to provide an adequate model for explaining secondary causes’ universal need for concurrence. Furthermore, it seems that we can diagnose the concurrentist notion of co-operation without division of labour as the underlying source of difficulty. If God and secondary causes were each understood to contribute different essential aspects of an overall effect, then it would be easy enough to see why God's concurrence would always be required. However, given its insistence that God and secondary causes must both bring about the entire effect, concurrentism seems committed to affirming that secondary causes must have all of the kinds of causal power the effect requires. But if they have all of the kinds of causal power that are needed and are not prevented from exercising the requisite powers, then their powers should be sufficient to bring about at least some effects without assistance.

Grant's responses to the metaphysical objection

Grant's response to the metaphysical objection begins by pointing out that my argument is vulnerable to refutation if (a)–(c) do not exhaust the possible ways in which a secondary cause might be in need of assistance. In my original presentation of the objection, I acknowledged that there may be more elaborate cases in which agents require assistance, but I expressed my suspicion that those cases will all involve more complex combinations of (a), (b), and (c) (Miller (Reference Miller2011), 8). It would be nice if that conclusion could be defended with something more than a suspicion; however, I am not quite sure how to construct an argument that would prove that (a)–(c), together with more complex combinations thereof, exhaust the ways of requiring assistance. On the other hand, that position does seem intuitively obvious. If an agent had all of the kinds of causal powers required to bring about an effect, had those powers to the requisite degrees, and was not prevented from exercising any of them, would it not seem to follow that the agent could bring about the effect without requiring further assistance? If an agent met all three of those conditions, what could it still be missing that might possibly explain its needing help? Nevertheless, Grant's defence of NODUC rests upon his contention that the metaphysical objection is vulnerable on this front. Although he does not explicitly separate them, I find it helpful to differentiate between a weaker and a stronger version of his response.

When Grant begins to explain the way NODUC conceives of the relationship between ‘God, creaturely causes, the effects of creaturely causes, and creaturely causings (or causal acts)’, the very first thing he says is that ‘it is highly plausible to expect that these relationships are sui generis’ (Grant (Reference Grant2019), 37). Thus, even before considering criticisms of the theory such as the metaphysical objection, he suggests a pre-emptive defence:

Grant later appeals to this pre-emptive line of defence in his response to the metaphysical objection. He notes that the ways of requiring assistance identified by (a), (b), and (c) ‘all admit of nontheological examples’, but since he considers it likely that the relationship between God and secondary causes is sui generis, he infers that the way creatures stand in need of assistance should be sui generis as well (Reference Grantibid., 198 n. 35). Moreover, Grant seems to suggest that we should expect the first three ways of needing assistance to be unenlightening when it comes to God's concurrence precisely because they admit of non-theological instances. The apparent implication, then, is that we should regard it as a virtue, rather than a vice, if none of the ways of needing assistance we are familiar with can shed light on a fourth option. Thus, the weak version of Grant's response to the metaphysical objection simply denies that there is a need to identify any additional way(s) of requiring assistance beyond (a)–(c); given the unique nature of God's causal contributions, concurrentists can simply posit the existence of an unidentified and mysterious fourth way of needing assistance.

It is hard to know what to say in response to such a bald appeal to mystery beyond the common refrain attributed to Schrödinger that at that price you can have anything; unsurprisingly, anything includes concurrentist co-operation. But if Grant's response to the metaphysical objection essentially amounts to ‘maybe there is some way of requiring assistance we simply cannot conceive of’, then NODUC is more an article of blind faith than a solid basis for developing theistic metaphysics. The fact that Grant relegates this weak response to an endnote suggests that perhaps he recognizes how feeble it is. In his main text he attempts to challenge to the adequacy of the metaphysical objection's (a)–(c) more directly by claiming to identify a fourth way of needing assistance that my argument ‘neglects to consider’. Grant proposes that an agent may require assistance because:

1. (d) Given [the] agent's power and the right antecedent conditions, it is possible for that agent to bring about a certain kind of effect provided that the effect is also simultaneously brought about by another agent on which the effect necessarily causally depends for its existence. (Reference Grantibid., 42)

Unfortunately, this alleged fourth way of requiring assistance makes little progress beyond the weaker version's blatant appeal to mystery. Unlike (a), (b), and (c), which identify clearly understandable reasons why an agent might require assistance, (d) offers no such explanation. It leaves us with the very question we began with: if an agent has all of the requisite kinds of powers to the requisite degree(s) and is not prevented from exercising them – in other words, if none of (a)–(c) apply – then why would such an agent require specifically tailored assistance from another agent? Grant's fourth way fails to answer this question; it simply asserts without explanation that assistance is needed. In essence, its response to the question of why assistance is needed reduces to an uninformative tautology: the agent requires assistance in bringing about an effect because it requires the assistance of another agent's necessary contribution. In other words, Grant's fourth way stipulates the need of a causal contribution that satisfies the demands of SC₁ and SC₂, but it does so without making any attempt to explain either what such a causal contribution might be like or why it should be needed. It seems, then, that the success of Grant's response to the metaphysical objection ultimately depends upon the adequacy of the weak version and its direct appeal to mystery.

Perhaps Grant would object that my expectations of clarification and explanation are unreasonable, and that (d)'s stipulative approach is enough to demonstrate that concurrentism's requirements can be met, even if we are at a loss to comprehend the mystery of how they might be met. Given his insistence that God's causal contributions are sui generis, he might contend that that is precisely the outcome we should expect. However, this sort of defensive appeal to the mysterious nature of divine causation is unconvincing because it misdiagnoses the source of concurrentism's difficulties. As noted above, the concurrentist notion of co-operation without division of labour seems to be the underlying problem. This sense of cooperation refuses to divide an effect into distinct parts or aspects that can be wholly attributed to one cause or another, demanding instead that each cause bring about the entire effect. Since powers are differentiated by their manifestations, such co-operation commits concurrentism to affirming that secondary causes must possess all of the kinds of powers the effect requires, and that commitment is what makes it so difficult to explain why concurrentism's secondary causes should always require assistance.

Grant's interpretation of the metaphysical objection fails to grasp the significance of this point, and as a result, some of his remarks are directed towards a straw man. For example, he responds to my comparison of concurrentism's position to the seemingly contradictory claim that ‘I have all the causal powers needed to lift things, but that there could never be anything, no matter how light, that I could lift without assistance’ by noting that:

Grant's reply here takes my reference to a secondary agent's having ‘all the causal powers needed’ to bring about an effect to be guilty of the rather obvious error of forgetting that concurrentists insist that God's concurrent power must be added to the secondary agent's contribution. But that uncharitable interpretation misses the metaphysical objection's central point: secondary causes could not bring about the entire effect unless they possessed all of the kinds of casual powers the effect requires. But in that case, although they might sometimes require assistance for the sort of ad hoc reasons represented by types (b) and (c), SC₂'s insistence that they should always require assistance is left unmotivated and inexplicable.

In short, Grant's appeal to the mysterious, sui generis nature of divine concurrence is a specious solution to concurrentism's difficulties because those difficulties simply are not grounded in the mysterious nature of divine causation; rather, they are self-inflicted by the apparently incompatible claims concurrentism makes concerning secondary causes. A convincing response to the metaphysical objection requires more than offering unmotivated and unsupported stipulations that secondary causes always require God's assistance. Instead, it requires presenting a plausible model of co-operative action that would satisfy both SC₁ and SC₂'s requirements – one that would explain why created beings with genuine causal powers would always require assistance specifically ‘tailored’ to the very contributions their own powers make to an overall effect. Unfortunately, to my knowledge, no concurrentist has ever offered a model of co-operation that satisfies concurrentism's own requirements.