Causal and non-causal explanations in theology: the case of Aquinas's primary–secondary causation distinction

Abstract

The basic question of this article is whether Thomas Aquinas's doctrine of divine providence through his understanding of primary and secondary causation can be understood as a theological causal or non-causal explanation. To answer this question, I will consider some contemporary discussions about the nature of causal and non-causal explanations in philosophy of science and metaphysics, in order to integrate them into a theological discourse that appeals to the classical distinction between God as first cause and creatures as secondary causes to explain God's presence and providence in the created universe. My main argument will hold that, even if there are some philosophical models of explanation that seem to allow one to suggest that, at least partially, this doctrine could be seen as a non-causal theological explanation, there are other models that offer seemingly stronger reasons to see this doctrine in full as a causal theological explanation.

This article asks whether Thomas Aquinas's doctrine of divine providence is best understood in the terms of causal or non-causal explanation. Alexander Reutlinger (2017b, 9) claims that ‘one finds a strikingly common theme in philosophy of science, philosophy of mathematics, and metaphysics: an increasing attention to non-causal explanation’. This article is a first gloss at extending this conversation to the domain of philosophical theology and discussions concerning theological explanation. As a first exploration of this theme, this article aims at opening new ways of approaching and analysing the relation between the divine and the creature from different theological traditions in dialogue with contemporary philosophy of science and metaphysics.

To answer this question, I will consider some contemporary discussions about the nature of causal and non-causal explanations (mostly in terms of grounding) in philosophy of science and metaphysics, the understanding of which varies from author to author, to integrate them into a theological discourse that appeals to the classical distinction between God as first cause and creatures as secondary causes to explain God's presence and providence in the created universe. My main argument will hold that, while there are some philosophical models of explanation (Ylikoski 2013; Reutlinger 2016, 2017a, 2017b) that seem to allow one to suggest that, at least partially, this doctrine could be seen as a non-causal theological explanation, there are other models (Wilson 2018) that offer similarly strong reasons to see this doctrine in full as a causal theological explanation.

I will first offer a brief account of how theological explanation is discussed in current literature, followed by a short discussion of why the notion of dependence might be the key notion to arrive at an answer for the central question of the article. Later, I will succinctly present Aquinas's distinction between primary and secondary causation and explain why it is relevant for current debates in the theology of divine action, moving afterwards to discuss different contemporary approaches to causal and non-causal explanation that might bare some weight into the discussion about theological explanation. I will end with an assessment of Aquinas's distinction against these approaches in order to decide whether one should consider this doctrine a causal or non-causal explanation.

Theological explanation today

The notion of theological explanation has seldom been explored in the recent theological literature. In fact, it is usually assumed that God is interpreted in theological explanations to have a causal role. For instance, in a discussion about the bodily resurrection of Christ, Geivett (2006, 93) explicitly claims that ‘the best causal explanation for a bodily resurrection is a miracle – an act of God’, concluding that ‘divine agency must be judged at least an intelligible category of causal explanation’ (Geivett 2006, 94). There are, however, a few scholars who have recently discussed the issue of theological explanation. Nevertheless, they all also analyse theological explanation as a causal type of explanation. These works are those of Paul Allen (2005 and 2017), Michal Oleksowicz and Piotr Roszak (2021), and John Bishop (2018). Allen, drawing from the works of Ernan McMullin (1924–2011) and Bernard Lonergan (1904–1984) on scientific realism and theological functional specialities respectively, argues that ‘theological explanations can be construed analogously to scientific explanations’ (Allen 2005, 2), in that both ‘purport to explain the existence of unobservable entities’ (Allen 2005, 3) through causal explanations. As unobservability is critical to the natural sciences, ‘it is also critical to the way theology accounts for God, an unobservable feature of reality’ (Allen 2005, 11). Oleksowicz and Roszak (2021) present several models of causal explanation in the sciences as they appear in the new mechanical philosophy, arguing that there is a plurality of causal explanations at work in the natural sciences today grounded in an acknowledgement of a metaphysical plurality in natural causation. This image, they hold, corresponds to what happens in the theology of, for example, Thomas Aquinas, who, holding to a metaphysical plurality of causes at work in nature (both natural and divine), presents a plurality of theological causal explanations, which they call a theological polyphony.

Perhaps the most analytic development of the different candidates for theological explanation is that of Bishop (2018). For him, ‘theological explanation purports to explain all that is in terms of God as the ultimate explainer’ (Bishop 2018, 143). Bishop analyses four candidates for viable theological explanations of God as the ultimate explainer: (1) God as first cause; (2) God as foundational, trans-scientific explanation; (3) God as ultimate intentional explanation; and (4) God as euteleological explanation (his own proposal). I will not discuss each candidate, as it goes beyond the scope of this article to do so. I will, however, recover the basic features that a theological explanation should possess, according to Bishop, in order to have proper theological explanatory power. What is common to theological explanations in general is that ‘theological explanation on any Interpretation is bound to have unique explanatory features and to fall short of full comprehensibility’ (Bishop 2018, 156), meaning that even if theological explanations could bear some similarities with scientific explanations, there will be bigger disanalogies present that will complicate things epistemologically speaking. Another general feature common to all candidates for theological explanations that Bishop analyses is that they are all, to some extent, causal explanations. For the first and third candidates, Bishop claims ‘that the Universe is a divine creation, which surely requires that the Universe has an ultimate efficient cause of its existence’ (Bishop 2018, 154). He states something similar for the second candidate, saying that ‘the God-role in the theist traditions has to be filled by that which is uniquely and in principle not causally dependent on anything else’ (Bishop 2018, 146), but on which everything else depends (one may add). As for the fourth candidate, Bishop also describes it in terms of causes, both efficient and final. He first explains that euteleology represents ‘a viable interpretation of theological explanation that takes Reality's directedness upon the supreme good to be inherent’ (Bishop 2018, 155) also affirming ‘that Reality contains within it concrete realizations of that supreme telic good’ (Bishop 2018, 155). To make this statement more intelligible, he further explains that:

an efficient cause is just that which explains why its effect is actual. There is thus conceptual space for efficient causes that are not literally producers of their effect: the idea of the concrete realization of the Universe's telos being its ultimate efficient cause is thus in principle admissible, even if such a coincidence of efficient with realized final cause is explanatorily unique. (Bishop 2018, 156)

Leaving aside Bishop's peculiar understanding of efficient causation, what matters for my purposes is that for all four candidates he understands theological explanation as causal, in the same vein as Allen and Oleksowicz and Roszak.

These three cases of theological explanation emphasize, to a greater or lesser degree, the differences between scientific and theological explanation, as well as the fact that they are all causal explanations. For Allen, even if theological explanation deals with the greatest of all unobservables, namely God, it still models how scientific explanations work in causal terms. For Oleksowicz and Roszak, given the plurality of causes in the natural world, which is explained by a plurality of types of causal explanations present in the new mechanical philosophy, one could affirm a plurality of causal theological explanations, as they appear in the works of Thomas Aquinas. Finally, Bishop understands that theological explanations deal with the ultimate explainer and that, in all cases he analyses, this ultimate explainer is to be understood in causal terms.

The notion of dependence and theological explanation

Within this context, models of divine action and providence that attempt to explain how God acts in the natural world work with basic notions of physical and divine causation. One major problem for these models is that of distinguishing such types of causation. Conflating physical and divine causation creates issues for traditional understandings of the divine attributes, particularly those of transcendence and omnipotence. Consider the model of quantum divine action (Russell 2018). In this model, God is taken to cause where physical causes do not, namely, at the quantum level of reality when a wave-function collapses. A typical objection to this model claims that God causes as natural causes cause (Silva 2015, Dodds 2012), given that God cannot cause where natural causes cause, requiring an ontological gap in natural causation for God to cause whatever God wants in the world. This conclusion implies that divine power and divine transcendence are somewhat limited, something that a classical theist would not accept.

Classical theists of a Thomist ilk usually attempt to solve this problem with the distinction between primary and secondary causes, a key notion for these alternative models. This distinction allows theologians and philosophers of religion to distinguish the different ways in which God and the natural causes actually cause and what they are required to explain. Among many understandings of this distinction (Farrer 1967, Silva 2014), Thomas Aquinas offers a compelling model of the relationship between primary and secondary causes.

A key notion to understanding this model is that, for Aquinas, a cause is always that upon which something depends for its being or becoming something else: ‘the effect, thus, must depend on its cause’ (De Pot., 5, 1, co.),¹ allowing for a wide variety of causes, both natural and theological. Hence, God is said to be something upon which natural things depend. The way in which things depend upon God is, Aquinas teaches, in their very being and existence: God is the primary cause of things because what God causes is the very existence of all things: ‘the preservation of things in existence is nothing else than the influence of the being of the thing, namely, that God, as long as a thing is, causes the being of the thing’ (Super Sent. II, 15, 3, 1, 5). In addition, God also causes them to be causes as he causes the creature's actual causing: ‘God causes certain effects through mediate causes . . . because, in order to ennoble the mediate causes themselves, he communicates to them the dignity of causality’ (Super Io., 1, 4.). Thus, the secondary cause cannot do anything unless by way of the primary cause: ‘every agent acts by the divine power. Therefore, He is the cause of the action of all things’ (SCG III, 67). This language of causes seems to lead to affirming without any concerns that Aquinas is offering a causal explanation of the being and causing of things.

In fact, Lauren Ross, when explaining that in interventionist accounts of causation, causes are factors that make a difference to their effects and that ‘causal explanations are often characterized as the explanation of some effect by appealing to its causes’ (Ross 2021, 85), appeals to ‘dependence’ in a rather indirect way. For her, ‘intervening on and changing values of the cause produces systematic changes in the value of the effect, or alternatively, varying states of the effect depends on varying states of the cause’ (Ross 2021, 85). Interestingly, Ross comments that one needs to consider ‘how changes in values of the effect variable depend on changes in values of the cause variable’ (Ross 2021, 85), adding a bit later the example that ‘chemists suggest that variations in these [atomic] properties depend on variations in these [atomic] features’ (Ross 2021, 93). This use of notion of dependence to analyse causal explanations seems to suggest that this notion plays a specific role as a deciding criterion for accepting scientific explanations as causal. If this is the case, then, a theological explanation that makes use of the notion of dependence will also be, then, a causal explanation, theological rather than scientific.

Jaegwon Kim offers an analysis of the relations of dependence and explanation, presenting the two most basic and important ones, namely, those of causal dependence between states or events and what he called the mereological dependence, that is, the relation of dependence between the whole and its parts. Dependence, for Kim (as arguably for Aquinas as well), is ‘asymmetric and transitive, and can generate relational structures of dependent events, states, and properties’ (Kim 1994, 68). Kim suggests, then, that these relations of dependence ‘can serve as explanation-grounding relations’ (Kim 1994, 67). Hence, ‘explanations track dependence’, that is, ‘the relation that “grounds” the relation between an explanans and its explanatory conclusion is that of dependence’ (Kim 1994, 68). There are certainly numerous differences between Aquinas's and Kim's notions of dependence (Aquinas refers to dependence upon God for the creature's existence, for instance, while Kim speaks about worldly relations). Still, an intriguing coincidence is that both notions open the path for the existence of a variety of kinds of relations of dependence, and in that sense, one may explore ways in which they can be fruitfully brought together. For instance, one may argue that Kim's analysis might lead one to affirm that the relation of dependence to which Aquinas is referring should be understood in terms of grounding, rather than of causation. Alternatively, one may also argue, Aquinas's notion of causation as dependence might suggest that Kim's notion of dependence as grounding could simply refer to a larger set of causal relations.

Reutlinger (2017a), for his part, suggests that in contrast to causal explanations in the sciences, metaphysical explanation, which usually refers to a dependence relation termed as grounding, is better described as non-causal. Similarly, Ylikoski argues that constitutive explanations, namely, those that ‘explain how things have the causal capacities they have’ (Ylikoski 2013, 278), are conventionally understood as non-causal explanations. Both these strategies to understanding non-causal explanations could be used to describe those relations to which Aquinas is referring, making his theological explanation a non-causal explanation. Alastair Wilson (2018), however, offering a parallel analysis between grounding and nomological, that is, scientific, causation, suggests that grounding (and, one might argue, also what Ylikoski understands as constitutive explanations) is rather a metaphysical type of causation. Following this different path, Aquinas's distinction could be understood as a theological causal explanation. The remainder of this article will analyse the outcome of taking each of these different strategies in relation to the question of whether Aquinas's proposal should be considered a causal or a non-causal theological explanation.

Aquinas's primary-secondary causation distinction

Aquinas's metaphysics of being is the very foundation for understanding his views on how one can say that God is at work in the natural world; thus, a few comments on this regard would be useful from the outset. Aquinas teaches that the relation between natural beings and God is that of participation in existence: what has existence by its own essence, and thus is its own existence, namely God, can give it to other beings (Davison 2019, 35–78): ‘The first cause does not have participated being in any way, but it is pure being; therefore, all what is participated derives from that which subsists by its own essence’ (In De Causis, l. 9). Now, this metaphysical notion of participation requires three elements: (1) the possessor of the perfection that is participated; (2) a participant subject that possesses that perfection partially or in a restricted manner; and (3) that the participant receive the perfection in dependence on the higher source (Clarke 1995, 93). This third element implies the fundamental relation of dependence of creatures on God, both for their origin and their likeness to the divine being. Thus, Aquinas is adamant in affirming that ‘the creature depends upon God and not the other way around’ (De Pot. 3, 3, co.).

This notion of dependence of creatures on God is what Aquinas understands as the basis for his doctrine of creatio ex nihilo, which should be understood in terms of God's causing the complete existence of the beings that constitute the universe by participating in God's being. The key notion, as I mentioned, is that of dependence. The very fact that God creates the world means that it is completely dependent upon God, namely, that there is nothing that does not have its source in God. Aquinas explains this idea in his Questions on the Power of God, explicitly referring to his metaphysics of being:

there is a being that is its own being, which follows from the fact that there must be a being that is pure act and in which there is no composition. Hence, from that one being all other beings that are not their own being, but have being by participation, must proceed . . . Thus, reason demonstrates . . . that all things are created by God. (De Pot. 3, 5, co.)

Aquinas describes this dependence in causal terms by explaining that God is the efficient cause of the whole universe, in the sense that God directly and immediately causes the universe (S. Th. I, 44, 1). I will leave the treatment of God's formal and final causality of the universe for another occasion, focusing on his efficient causality, following on Aquinas's explicit treatment when dealing with divine providence (for instance, in the Questions on the Power of God). Thus, the universe exists because of God's direct efficient action of creation. Aquinas goes even further and claims that this complete dependence upon their creator of all the created beings refers not only to all their being but also to the actions of created creatures: ‘God is the cause of every action, as much as any agent is the instrument of the divine power acting’ (De Pot. 3, 7, co.). This idea is the key, then, to understanding how God could be said to be active within the created universe, that is, how God could be said to be provident. The basic argument runs as follows: for created beings to have powers to act, those powers should also be created and hence also dependent upon God for their existence. Then, given that creatures are kept in existence through their participation in the divine being, creatures are also said to have powers to act as long as God gives them their powers to act.

In his analysis of creation as the complete dependence of creatures upon God, the efficient cause of all things, Aquinas introduces the difference between primary and secondary causality: God is the primary cause of things because God causes their very existence, without which things would not be. Hence, the secondary cause cannot do anything if it is not by way of the primary cause causing it to be, to exist: ‘God works through all secondary causes . . . all their effects may be traced back to God as their cause’ (SCG III, 75). Ultimately, as Aquinas explains, that which causes what is first required is the primary cause, namely, God (In De Causis, l. 9). There are several features that differentiate these two kinds of causing. First, for Aquinas, ‘the primary cause is more influential in the effect of the secondary cause than the secondary cause itself’ (In De Causis, l. 1). Given that everything that the secondary cause is is caused by the primary cause, then, its power to be a cause and to produce something is given by the primary cause. So, even if the secondary cause is the real cause of its effect, properly speaking, the primary cause is primarily the cause of the effect of the secondary cause. Second, since the secondary cause does not cause except by participating in power of the primary cause, ‘the effect does not proceed from the power of the secondary cause except because of the power of the primary cause’ (SCG III, 70).

These considerations open the path to understanding the ways in which God can be said to be providentially active in the created universe. In his Questions on the Power of God (De Pot. 3, 7, co.), Aquinas argues for four different ways in which God's causality penetrates most intimately the causality of created natural things, and hence can be said to be providentially active in the world. First, something can be understood as giving something else the power to act, as in the divine act of creation: every operation consequent to a certain power is ascribed to the giver of that power as an effect to its cause. Since all power of any created efficient cause whatsoever is from God, as I have just explained, then God is said to cause every action of created things, because he gives natural things the powers by which they are able to act, as from the first principle of all perfection. Second, God may be said to be the cause of an action by upholding the created natural power in its being, as in the constant participation of God's being in the creature. Every action that cannot continue after the ceasing of the influence of a certain efficient cause belongs properly to that efficient cause, as a remedy that preserves sight is said to make a man see. God not only gave existence to things when they first began to exist but preserves them in existence. Thus, God is always preserving those powers in them and hence causing them to be. This simply means that if this divine preservation were to cease, every natural causation would also cease. Therefore, every natural casual power of a thing finds its ultimate cause in God's causality in this second sense as well. I have called these two ways the ‘foundational moments’ of God's acting in the created world (Silva 2014). The other two ways I termed the ‘dynamic moments’. This distinction will be useful to assess whether Aquinas's explanation of God's providential activity in the created world could be seen as a causal or a non-causal theological explanation.

The explanation of these dynamic moments depends on how Aquinas understands the workings of an instrumental cause, as he uses this analogy to explain them. For Aquinas, an instrument, when being used as an instrument, has two different effects: one that pertains to it according to its own nature, and another that pertains to it insofar as it is being used by the primary cause, with the latter transcending the nature of the instrument. It is through the first effect, however, namely, that which pertains to the instrument, that the second effect, namely, that which is according to the principal cause, is performed, hence the use of this and not another instrument. Still, neither effect could be caused by the instrument if not by depending on the causing of the principal cause. Thus, both effects (cutting, and cutting in such a manner) are said to belong both to the instrument and to the principal cause.

These effects refer to the two dynamic moments of God acting in and through created secondary causes. Thus, the first of these two ways of causing refers to the first effect of an instrumental cause in the following way. Every created cause performs its operation according to its own nature and powers, moved by God to act, and to achieve its proper effect, for instance, when someone uses the sharpness of a knife to cut a loaf of bread to make toast, the first effect would be the very cutting of the knife. The second way of causing the action of the instrument refers to producing an effect that goes beyond the power of a particular created cause. In my example, this second effect would be cutting the loaf of bread into a particular shape, which is something that the knife by itself cannot perform. Ultimately, these two dynamic moments are possible given the immanence of the universal power of God, the primary cause, in secondary causes. For Aquinas, however, the cause of an action is that by whose power it is performed, more even than that which does it: even as the principal efficient cause, in comparison to the instrument, is more the cause. Thus, even if a created cause is immediate to its effect when considered in itself, if one considers the power by which the action is done, then the power of the primary cause is more immediate to the effect than the power of the secondary cause, since the power of the secondary created cause is not coupled with its effect save by the power of the primary cause. Therefore, ‘God is more the cause of every action than secondary active causes’ (In De Causis, l. 1).

Still, for Aquinas, the causal powers of created things suffice for causing in their own created order yet depend on the divine power to cause (De Pot., q. 3, a. 7, ad 1). God and natural efficient causes cause on two different levels. The same effect is ascribed to a created cause and to God, not in the sense that God causes one bit of the effect and another bit by the created cause. Rather, the totality of the effect proceeds from each in different ways, just like the whole effect is said to proceed from the instrument and principal cause: ‘nothing prevents one and the same action from proceeding from the first and second agent’ (S.Th. I, 105, 5, 2). It is in this respect, in the joint causing of the two orders of primary and secondary causality, that the created powers are not enough to cause their own effects and that they depend on God. Just as a craftsman gives the axe the power with which the axe actually chops the wood, God gives created things the power to cause. Aquinas argues, however, that this natural causing of secondary causes is, in a sense, also necessary, because even though God could cause the effect without the created cause, He wishes to cause by means of nature in order to preserve ‘the order of things’ (De Pot., q. 3, a. 7, ad 16). God, thus, acts through natural causes because of the immensity of His goodness, by which He decides to communicate His similitude to created things, not only in their existence, but also in their being causes of other things.

In these four ways, then, God is said to be efficiently the cause of the causing of every created cause inasmuch as everything participates in, that is, depends on, His power to be a cause. Therefore, God is the cause of everything's action inasmuch as (1) He gives everything the power to act, and (2) preserves that power in being (the foundational moments), (3) applying it to action, and inasmuch as (4) by His power every other power acts (the dynamic moments) (for a more in-depth analysis of this doctrine, see Silva, 2022). Finally, primary causation thus understood is, for Aquinas, present both in the causal relations between creatures (since God is always present in every action of the creatures), as well as in the very act of creation.

A theological causal or non-causal explanation?

The question is to what extent one can consider Aquinas's explanation of the doctrine of God's providential involvement in nature as causal or non-causal. Clearly, Aquinas is explicit in using causal language. Nevertheless, as I mentioned above, there are certain positions in contemporary philosophy of science and metaphysics that might describe non-causal explanations in similar terms as those Aquinas uses to explain at least part of his doctrine. Hence, my main goal in this section will be to assess whether any of the perspectives present in this debate offers any insights into our understanding of Aquinas's doctrine.

Marc Lange is perhaps one of the most outspoken scholars in support of non-causal explanations in the sciences today. In his work Because without Cause, he analyses some scientific explanations that ‘do not derive their explanatory power by virtue of describing the world's network of causal relations’ (Lange 2017, xi). Lange contends that, while it might be the case that all mathematical explanations are non-causal, some scientific explanations could be non-causal. The later are mostly reduced to explanations by constraints involving mathematical necessity, which is affirmed to be stronger than causal necessity. This means that this kind of non-causal explanation accounts for phenomena pointing to its inevitability. These explanations are non-causal because they do not offer any information about the network of causal relations but explain by means of some mathematical constraint that bounds those causal relations (Bueno and Vivanco 2019, 551). Bueno and Vivanco (2019, 552) present two problems with this interpretation of non-causal explanation: it is clear neither how the mathematical necessity explains the phenomena in question, nor how these mathematical constraints are explanatory in themselves. Regardless of these problems, Lange managed to put back on the discussion table the issue of non-causal explanations, which, for him, ‘have generally been underappreciated in the vast recent philosophical literature about explanation’ (Lange 2017, xi).

Alexander Reutlinger (2017b) has offered a nicely structured landscape of the current debate on scientific explanation, particularly into non-causal explanations. I shall not go into too much detail about the situation as he explains it, but suffice to say that he divides the main positions into three approaches: causal reductionism, pluralism, and monism. The first strategy simply assumes that there are no non-causal explanations, meaning that seemingly non-causal explanations can ultimately be reduced to causal ones. Pluralism is the view that causal and non-causal explanations refer to two different theories of explanation, and that each supplements the other. In essence, for a pluralist there are different types of explanations that cannot be reduced to each other (Lange would be an example of this kind). Finally, monism is the view that holds that there is only one theory of explanation that captures the essential features of both causal and non-causal explanations.

Reutlinger (2016) supports this third approach, suggesting a counterfactual theory of explanation (CTE) that caters to both causal and non-causal explanations. In this CTE ‘causal and non-causal explanations are explanatory by virtue of revealing counterfactual dependencies between explanandum and explanans’. Following Woodward (2003), Reutlinger (2017b, 6) explains that for CTE ‘explanation is a matter of exhibiting systematic patterns of counterfactual dependence’, a definition that allows one to bring together both causal and non-causal explanations, since in both cases one could ask ‘what-if-things-had-been-different’ questions. In summary, ‘the notion of counterfactual dependence is the broader notion of determination that one expands from causal explanations such that it encompasses the non-causal explanations’ (Reutlinger 2016, 737). Now, this relation between explanans and explanandum must satisfy certain conditions for it to be explanatory: (1) the veridicality condition, requiring that the explanans is (approximately) true; (2) the implication condition, requiring that the explanans logically entail the explanandum; and (3) the dependency condition, requiring that the auxiliary statements of the explanans support at least one counterfactual implying that had they been different, then the explanandum would have been different.

Assuming the CTE for non-causal explanations in the sciences, Reutlinger applies CTE to non-causal explanations in metaphysics. These non-causal explanations in metaphysics refer to non-causal dependencies in terms of relations of grounding or constitution: ‘grounding [or to be constituted by] is taken to be a non-causal dependence relation whose explanatory power is, thus, non-causal’ (Reutlinger 2017a, 241). The question is, clearly, how grounding could be explanatory. Reutlinger suggests that CTE could apply to the realm of metaphysical explanations as long as one accepts a salient epistemological difference from scientific explanation. For Reutlinger, both the implication and dependency conditions are straightforwardly met by metaphysical explanations in terms of grounding. The veridicality condition, however, is not, since in metaphysical explanations one does not know whether one's metaphysical commitments and assumptions are true. To face this situation, Reutlinger weakens the argument by claiming that this is not necessarily a problem, since CTE allows for ‘explanations whose explanantia we currently do not (fully) know to be true and perhaps never will’ (Reutlinger 2017a, 249). These are what he calls ‘how-possibly’ explanations, which are present in the sciences as well as in metaphysics. Can this type of explanation be also present in philosophical theology?

It would seem that Reutlinger's analysis of how CTE can be applied to metaphysical grounding explanations could be directly applied to, at least, a set of theological explanations. Recalling my presentation of Aquinas's four ways in which God could be said to be providentially guiding the created universe, one might consider the founding moments as those that provide a grounding theological explanation for the existence and conservation of created things and their powers. In this sense, the founding moments of Aquinas's doctrine could represent a non-causal theological explanation of the doctrine of God's providence, in which each of the three conditions explained above are met: the dependency condition is certainly met, since were God to cease God's participation of being to creatures, these would cease to exist. The implication condition is also met, since the elements of the explanans, namely, God's constant participation of being, entail the existence and persistence of creatures. Finally, the weakened veridicality condition is also met, since the explanans offers a ‘how-possibly’ explanation. Following Reutlinger's example of a Humean answering a metaphysical question such as ‘how is it possible that there are laws without positing necessary connections in nature?’ (Reutlinger 2017a, 249), a Thomist could answer questions such as ‘how is it possible that God providentially guides the development of the created world’. So, it might seem that the founding moments of Aquinas's doctrine could be interpreted as offering a non-causal theological explanation. Looking at Ylikoski's (2013) ideas on causal and constitutive explanations could offer some further insight into this suggestion.

Ylikoski (2013) suggests that grounding, or rather constitution, should not be confused with causation, and hence explanations that refer to constitution should not be identified as causal explanations. Ylikoski explains that most, pace Lange, non-causal explanations are to be understood in terms of constitutive explanations, which are primarily explanations of causal capacities. That is, constitutive explanations explain how things have the causal capacities they have. Hence, causal and constitutive explanations account for different kinds of explananda, tracking different kinds of dependency. Still, assuming a monist approach to explanation, for Ylikoski both kinds of explanations refer to counterfactual dependence.

The constitutive relation between the explanans and explanandum (the causal capacities) is (1) asymmetric; (2) irreflexive; and (3) synchronous. Hence, ‘it does not make sense to talk about processes in the case of constitution’ (Ylikoski 2013, 282). This relation, then, can be understood in terms of dependence, meaning that constitution is a ‘kind of building relation’. Interestingly, commenting on this particular idea, Ylikoski affirms that ‘the asymmetry of constitutive explanation involves the asymmetry of existence’ (Ylikoski 2013, 283).

Now, the said founding moments of Aquinas's doctrine of divine providence deal with the very existence and permanence in being of the causal capacities of all created things. For Aquinas, these two moments of creating the causal power of things and conserving those causal powers are the very source and ground for those causal powers. Thus, in a theological sense, this is the most intimate constitutive relation that could exist, and hence, this would represent a non-causal constitutive theological explanation. Of course, one must return to the idea that things depend on God's participation in God's being, and are not constituted by God's own being. But still, it is through this participation in God's being that they are constituted, and with them, their own causal capacities. If we add that for Aquinas ‘creation is not a change’, in the sense that nothing changes when God participates God's being to creatures, hence putting and conserving them into existence, then this constitutive theological relation is also asymmetric, irreflexive, and synchronous, just as Ylikoski requires constitutive explanations to be. In particular, Aquinas's explanation affirms the fundamental asymmetry of existence between the creator and the creature.

The question remains, however, about the dynamic moments of Aquinas's doctrine of divine providence. With Ylikoski's distinction between causal and constitutive explanations, one might argue that, given their strong causal type of relation between the actual causation of created things and God's causing, these latter moments represent causal theological explanations. That is, for Aquinas, the dynamic moments of God's providence are an expression of God's direct involvement in creation, for which he uses a direct causal analogy, that of instrumental causation. God is the reference for the principal agent in the analogy, while all created things are referenced by the instrument. In this sense, while one cannot say that this is a diachronic relation as most causal relations are described in causal explanations in the sciences, it is, though, an explanation that accounts for a causal relation. Wilson (2018) shows the possibility of synchronous causal relations in the sciences, and these could be the analogue to understanding divine synchronous causation.

With these reflections about theological explanation led by Reutlinger's and Ylikoski's ideas in mind, one might argue that Aquinas's doctrine could be a mixed theological explanation, partially non-causal and partially causal. Stretching Ylikoski's argument about the existence of hybrid types of explanation, then, Aquinas's doctrine of divine providence exercised as a primary cause through secondary causes represents a mix between constitutive non-causal and causal explanations.

A different conclusion might arise, however, if one considers Wilson's (2018) argument. Instead of distinguishing between causation and constitution, or grounding, he suggests that grounding is a particular kind of causation, namely, metaphysical causation as distinct from nomological, or scientific, causation. For Wilson, grounding and nomological causation share a range of logical and theoretical parallels (for instance, both kinds of causation have the same logical properties and the same connections to explanation and counterfactuals, among others), a fact that can be best explained by acknowledging that these are different ways ‘for a generalized causal relation to obtain’ (Wilson 2018, 724). Thus, for Wilson, grounding explanations simply identify metaphysical causes, and hence are a different species of the same genus.

Of course, there are some differences between what Wilson calls nomological and grounding causation. First, grounding is synchronic, while most nomological causation is diachronic. Nomological causation is often understood in terms of ‘concrete physical processes that transfer marks, or mass-energy, or some other conserved physical quantity’, though they are also understood in terms of ‘production’ (Wilson 2018, 730). Grounding, on the contrary, is particularly concerned with causal dependence that does not go via causal production. In addition, grounding is supposed to be ‘more fundamental than any fact it grounds’ (Wilson 2018, 730), while ‘nomological causation is not usually seen as connected to fundamentality in this way’ (Wilson 2018, 731). Perhaps the second of these is the most interesting of Wilson's observations for our purposes: grounding relates to ‘metaphysical causal dependence rather than with metaphysical causal production’, but ‘it does need to involve characteristically causal patterns of counterfactual dependence’ (Wilson 2018, 732).

What is even more interesting is that when Wilson claims to be arguing that ‘the general notion of causal dependence is conceptually separable from its particular application to concrete objects, events, and states of affairs’ (Wilson 2018, 732), he uses the example of God's act of bringing the world into existence as a potential case of causation not among concreta. Surprisingly, he does not reflect on the fact that this might be a case of metaphysical causal production.

Now, if Wilson is correct in asserting that grounding, or constitution, is a kind of metaphysical causation, then we might be in a position to argue that all founding and dynamic moments of Aquinas's doctrine of divine providence can be construed as theological causal explanations of how it is possible for God providentially to guide the development of the created universe. The founding moments would correspond to the kind of metaphysical causal explanation that refers to grounding, while the dynamic moments would correspond to that case of metaphysical causal explanation that refers to metaphysical production. Of course, both would involve causal patterns of counterfactual dependence, since were God to cease the participation of God's being in creatures, or not to cause the actions and proper causation of created things, then things would either not exist at all or not produce at all.

In this scenario, Aquinas would have been correct in his explicit use of causal language to refer to both moments of his doctrine of divine providence. In fact, if, as Kim (1994) claims, explanations track dependence relations, and, as Aquinas claims, causal relations are dependence relations in becoming or being, then Aquinas would be offering causal theological explanations.

Conclusion

In this article I attempted a first approximation at answering the question whether there could be theological non-causal explanations. In particular, I analysed whether Thomas Aquinas's doctrine of divine providence through his understanding of primary and secondary causation could be understood as a theological causal or non-causal explanation in light of some contemporary debates surrounding causal and non-causal explanations in the sciences and in metaphysics.

Aquinas presents four ways in which God is said to be involved providentially in the created universe with regard to efficient causation. The first two I named the founding moments, since they serve as the very foundation of the existence of things and of divine providence. So, the first is the giver of a power to act, in which God gives natural things the powers by which they are causes. The second upholds natural causes in their being, God continuously upholds natural causal powers. The third and fourth ways, which I termed the dynamic moments, namely applying the natural cause to cause and causing effects that go beyond the natural causes, depend on Aquinas's understanding of an instrumental cause. The relation between a principal agent and an instrument requires that when someone applies the causal power of an instrument, this action has two different effects: one that pertains to it according to its own nature and causal powers, and another that pertains to it insofar as it is moved by the primary cause, and one that transcends its own nature and causal powers. Thus, the third way refers to the first effect of an instrumental cause: every natural cause causes according to its own nature moved by God to cause. The fourth way, finally, refers to the causing of an effect that goes beyond the power of created causes (Wippel 2007, Silva 2014).

I argued that if one were to consider the ideas of Reutlinger and Ylikoski on non-causal explanations in terms of grounding and constitution, then one could argue that Aquinas's founding moments could be understood as theological non-causal explanations, as they refer to the most constitutive explanation of created things. The dynamic moments would then account for a theological causal explanation. On the contrary, I argued that if one were to consider Wilson's ideas on grounding as a type of causation, then both the founding and dynamic moments of Aquinas's doctrine of divine providence should be understood as a theological causal explanation. What is certain is that, if one accepts that what Aquinas is trying to explain is the creature's relation of dependence upon the creator, then his is certainly an explanation, since, as Kim argues, explanations track dependence relations.