1. A Clear and Distinct Idea of Active Power

In his Essay, Locke writes: “bodies, by our senses, do not afford us so clear and distinct an idea of active power, as we have from reflection on the operations of our minds” (Locke Reference Locke1812, II.21.4). This observation is symptomatic of a broader controversy that animates the debate about the nature of causation across the seventeenth and eighteenth centuries. This controversy revolves around two opposite poles, which are both reflected in Locke's observation. On the one hand, it becomes increasingly difficult for early modern authors to consider material bodies as genuine efficacious causes (pace the scholastic view, which is in the process of being dismantled during the period). On the other hand, human beings do seem to have immediate and internal access to what an efficacious causal power looks like, by simply reflecting on their volitions and capacity of bringing about voluntary effects. Locke himself, for instance, maintains that “the idea of the beginning of motion we have only from reflection on what passes in ourselves; where we find by experience, that, barely by willing it, barely by a thought of the mind, we can move the parts of our bodies, which were before at rest” (II.21.4). If one puts together the idea that bodies do not seem to provide a reliable instance of genuine efficacious causal powers, while the human experience of voluntary actions does, it is possible to conclude that causation is in fact best understood as agent causation, namely, as the kind of efficacious causality exercised by free intelligent agents.

In this paper, I argue that Samuel Clarke offers a particularly thorough development of the idea that agent causation is the only acceptable model of causation in nature. In Clarke's account, proper causal efficacy can be attributed only to intelligent entities endowed with the freedom of choosing whether or not to bring about their effects. Freedom of choice is the only genuine causal power that makes it possible, for the agent that possesses it, to be a genuine initiator of causal processes. In Clarke's account, the idea of a necessary agent becomes a contradiction in terms, while the idea of a necessary connection between the agent and the obtainment of certain effects turns out to be untenable. Any causal chain can be traced back to the free decision of some intelligent agent who initiated it in virtue of its power of freely deciding whether or not to do so. Moreover, since only intelligent entities can be properly understood as efficacious causes, beings deprived of intelligence are not agents or efficacious causes at all and their effects are thus the result of the immediate action of some intelligent being operating upon them. In what follows, my goal is to fully unpack Clarke's account of agent causation and explore its broader implications. Before engaging in this task, however, it is important to clarify why an investigation of Clarke's account of agent causation significantly contributes to three related areas in the current debate on the history of early modern philosophy and science.

First, recent years show a growing methodological discussion about the need to broaden the canon of early modern history of philosophy and science by taking into account those authors who have been so far understood as “minor figures” and have been considered ancillary to elucidate the position of the “major figures” (Shapiro Reference Shapiro2016; Beaney Reference Beaney2018). Although Clarke was surely an authoritative voice in his own time, today he is still most often portrayed as Newton's spokesman. However, when it comes to fundamental questions in metaphysics, philosophy of science, and moral philosophy, he often develops these issues with a systematicity and degree of analysis that is rarely present (and that sometimes has no clear parallel at all) in Newton.Footnote ¹ This paper offers the first systematic reconstruction of Clarke's account of causation presented in its own right, and explores its historical and philosophical justifications and implications. I contend that it is methodologically necessary to first understand Clarke's position in his own terms in order to fully assess how it relates to that of other authors. In order to create a more inclusive and open-minded history of early modern philosophy and science it is necessary to allow the (until now) “minor figures” to speak for themselves, setting their own philosophical priorities and goals, rather than approaching them through the prefabricated cliché of the “spokesmen” of “canonical” authors. In section two, I thus provide the fundamentals of Clarke's own account of agent causation as they emerge from Clarke's own philosophical project.

Secondly, Clarke is usually characterized as a voluntarist, insofar as he stresses God's freedom of acting without that action necessarily being determined by his intellect. Clarke's voluntarism is most often discussed in the context of his correspondence with Leibniz. As John Henry puts it: “the correspondence reveals, above all else, a clash between the differing worldviews of the voluntarist and the intellectualist” (Henry Reference Henry2009, 89).Footnote ² However, the label “voluntarism” requires further clarification. Peter Harrison has mounted a sustained case against the “voluntarism and science thesis,” according to which voluntarism would have played a major role in the emergence of early modern science. Harrison's overall objection is that “the term ‘voluntarist’ is used to characterize such a range of positions that it, too, has probably outlived its usefulness” (Harrison Reference Harrison2002, 17). Among the problems of the voluntarism and science thesis, Harrison raises two points that are specifically connected with the voluntarist reading of Clarke. First, Clarke is often characterized as a voluntarist because of his emphasis on the fact that the created world depends on God's arbitrary will. Harrison claims that, in this context, the label ‘voluntarism’ conflates two rather distinct ideas. On the one hand, voluntarism is taken to be a synonym for expressing the world's dependency on God's will. The problem here is that this dependency “is a straightforward implication of classical theism” (ibid., 6) and nothing distinctive in itself. On the other hand, voluntarism is taken to stress the immediate involvement of God in the course of nature. However, Harrison rightly points out that this latter position is better understood as occasionalism: “voluntarism (a doctrine about the divine will) ought not to be equated with occasionalism (a doctrine about causation). Occasionalism, moreover, is quite consistent with intellectualism.” Harrison's second criticism concerning the characterization of Clarke as a voluntarist is that Clarke believes that moral laws are grounded in “eternal reasons.” According to Harrison, this view was shared by Newton and Cudworth and it “is an unmistakably intellectualist position” (ibid., 15).

I am sympathetic to Harrison's methodological call for precision in the use of broad and potentially ambiguous labels such as “voluntarism.” In section three, I argue that Clarke's account of agent causation provides a more precise ground to construe Clarke's position as a form of moderate voluntarism. Clarke's account entails a distinction between two kinds of necessities, which allows Clarke to maintain that there are “eternal reasons” that ground moral laws, while at the same time these reasons do not bound in any way God's free decision to act and create a world according to them.Footnote ³

Thirdly, Clarke's account of agent causation makes it convenient for Clarke to exploit a version of partial occasionalism (i.e. the view according to which only material beings are completely devoid of causal powers and all their effects are brought about immediately either by God or by intermediary intelligent agents appointed by God). The implications of Clarke's account of agent causation and of partial occasionalism are relevant for both our understanding of Clarke's positioning within the early modern debate on causation, and for today's scholarly debate on the impact of occasionalism on early modern theories of causation.

The scholarly literature on Clarke's attitude towards occasionalism is scarce, but when the topic is addressed it is framed in terms of Clarke's relationship with the specific brand of occasionalism popularized by Malebranche. In his extensive study on the reception of Malebranche in England, Charles McCracken (Reference McCraken1983) does not discuss Clarke's case. John Gay (Reference Gay1963) mentions only briefly that Clarke would have been opposed to Malebranche's occasionalism because it would have “made human, as well as divine, freedom impossible.” On the contrary, Paul Russell (Reference Russell2008, 151–152) claims that Clarke advanced an occasionalist interpretation of Newton's physics for apologetic reasons. As a result, Russell maintains that Hume's “critique of occasionalist doctrine had irreligious implications that reached well beyond the specific targets he mentions in the Treatise (i.e. Malebranche and other Cartesians)” (Russell Reference Russell2008, 161), namely, it includes Clarke among its targets. The scholar who has devoted the most sustained attention to the issue of Clarke's confrontation with occasionalism is Ezio Vailati. However, Vailati's assessment remains ambiguous. On the one hand, Vailati maintains that Clarke's view clearly differs from Malebranche's insofar as Clarke admits that not only God but also minds are endowed with causal powers (Vailati Reference Vailati1995; Reference Vailati1997; Clarke Reference Clarke and Valiati1998, xxiv-xxv). On the other hand, Vailati also grants that Clarke would have endorsed some aspects of Malebranche's account of the laws of nature: “for both philosophers the laws governing bodies are extrinsic denominations, and that miracles, to the extent that they are exceptions to these laws, are viewed as mere deviations from God's general volitions” (Vailati Reference Vailati1997, 159). More recently, Yenter and Vailati conclude that despite similarities Clarke should not, ultimately, be considered an occasionalist: “Clarke was not, strictly, an occasionalist. Unlike the occasionalists, Clarke did not think that God was the real cause of interactions between finite minds and matter. Furthermore, matter has a single ‘Negative Power’ of staying at rest or continuing in motion” (Yenter and Vailati Reference Yenter, Vailati and Zalta2014). Henry supports a similar conclusion, but for a different reason. In his reply to Harrison, Henry points out that “occasionalism is intrinsically incompatible with voluntarism” (Henry Reference Henry2009, 85). The reason is that occasionalism would amount to a “circumscription of God's omnipotence based on what a human thinker decides to be impossible” (Henry Reference Henry2009, 85). Henry does not refer directly to Clarke, but rather as an example of occasionalism, he mentions Malebranche.

In section four, I argue that (pace Henry and Vailati), some form of occasionalism (i.e. partial occasionalism) is not only compatible with some form of voluntarism (i.e. Clarke's moderate voluntarism) but it also establishes a mutual argumentative synergy with it.Footnote ⁴ In order to assess Clarke's position with respect to occasionalism it is crucial, on the one hand, to fully appreciate Clarke's account of agent causation, and on the other hand, not to conflate occasionalism in general (which is a family of arguments rather than a unified doctrine) with Malebranche's own version of it. Agent causation is the key to both understanding why Clarke accepts only a partial version of occasionalism and why partial occasionalism strengthens Clarke's case for agent causation. Clarke's complex and sophisticated position demonstrates that occasionalism can be adapted and appropriated well beyond the limits of Cartesian debates and for reasons that Cartesian occasionalists in fact rejected.

This point contributes to the current efforts of better mapping the varieties of occasionalism circulating across the seventeenth and eighteenth centuries and the different argumentative strategies and justifications used to support them.Footnote ⁵ Clarke's case reveals how his concerns for securing a sharp distinction between free intelligent agents and mere material beings led him to combine distinctive themes of Newtonian natural philosophy (such as the quest for the cause of gravity and the preoccupation with securing divine providence in the natural world) with an occasionalist theory of body-body causation.

2. Clarke's Account of Agent Causation

Clarke's account of agent causation is established in his Demonstration of the Being and Attributes of God.Footnote ⁶ For present purposes, sections VIII and IX are particularly relevant since it is in these sections that Clarke argues that the self-existent being (i.e. God) must be intelligent (section VIII) and free (section IX).

Section VIII discusses the attribute of intelligence against the materialist view on which matter could be an eternal and self-existent but non-intelligent being. The bulk of Clarke's discussion here depends on his exploitation of the scholastic “causal containment axiom.”Footnote ⁷ In Clarke's formulation: “it is impossible that any effect should have any perfection which was not in the cause” (Demonstration VIII, 39). In order to argue that the self-existent being is not intelligent Clarke faces the atheist with a dilemma: either “there is no intelligent being at all in the universe, or intelligence is no distinct perfection but merely a composition of figure and motion” (Demonstration VIII, 39). The first option is clearly at odds with the fact that human beings (at least) are usually considered to be intelligent beings. The second option is the one that Clarke most forcefully aims to reject by extensively arguing that intelligence cannot be conceived of as somehow resulting from purely mechanical configurations of matter.Footnote ⁸ For present purposes, I shall leave aside the details of Clarke's argument on this point and stress that the upshot of his discussion is that an intelligent being cannot be purely material and that intelligence cannot be produced by a non-intelligent cause.

Section IX seeks to demonstrates that “the self-existent and original cause of all things is not a necessary agent but a being endowed with liberty and choice. The contrary to this proposition is the foundation and the sum of what Spinoza and his followers have asserted concerning the nature of God” (Demonstration IX, 46). While Spinoza's account of the communication of motion is targeted by the end of section VIII,Footnote ⁹ Spinoza certainly did not deny that God is intelligent, since he maintained that thought is an attribute of God. From this point of view, section VIII did not fully counter Spinoza's own position. However, Spinoza did deny that God had “liberty and choice” (to use Clarke's phrasing), since he maintained that God's freedom consists in acting in virtue of the necessity of its own nature and not being coerced by external causes.Footnote ¹⁰ In order to refute this position, the task of section IX is to show that freedom does not consist in the mere absence of external coercion. In fact, being free from the determination of external causes is neither necessary nor sufficient to define freedom. Clarke explains: “Without liberty, nothing can, in any tolerable propriety of speech, be said to be an agent, or cause of any thing. For to act necessarily is really and properly not to act at all but only to be acted upon” (Demonstration IX, 46).

It may well be the case that a necessary agent is not be coerced to act by anything but its own nature (as in the case of Spinoza's God). However, according to Clarke, this kind of necessary agent is in fact no agent at all. Proper agency requires liberty or freedom understood as a positive power of choosing to act or not to act. To use early modern terminology, this kind of freedom can be labelled ‘freedom of indifference’ and results from a positive power of the will (i.e. the power of acting or refraining from acting), rather than from a mere absence of external coercion. In the absence of freedom of indifference, the agent cannot be taken to be genuinely responsible for the effects that it brings about. Let me unpack how Clarke's argument works.

In order to appreciate Clarke's position, it is important to first distinguish between two kinds of necessity. In section III of the Demonstration (III, 14–18) Clarke distinguishes between “antecedent” and “consequent” necessity and considers antecedent necessity a form of “absolute” necessity. Antecedent necessity is grounded in the nature of the thing itself and does not depend on any supposition. According to Clarke, antecedent necessity consists in being non-contradictory:

Clarke applies the notion of antecedent necessity to God in order to demonstrate that God's nature contains the sufficient reason for its own existence (ibid., 12–13). By contrast, consequent necessity concerns the necessity with which something follows or obtains given a prior assumption or condition.

Antecedent necessity is a relation of necessity which holds between properties of a thing and the nature or essence of it. In the rest of my discussion, I shall refer to antecedent necessity as “conceptual” necessity in order to stress that this kind of necessity focuses on the necessary connection between a thing's essence, or concept, and its properties, or consequences. The causal containment axiom mentioned above (e.g. the fact that material entities can only cause material effects and cannot produce intelligence) entails conceptual necessity. A violation of the causal containment axiom would lead to there being something in the effect that either does not depend on the nature of its cause or that is created ex nihilo, which Clarke considers contradictory (Demonstration II, 12). Conceptual necessity, however, does not entail that certain entities necessarily exist or that certain effects necessarily obtain. When referring to the necessity with which something exists or some effect obtains (given the required causes and conditions), I shall use the term “obtainment” necessity. Obtainment necessity may be regarded as a specification of what Clarke calls consequent necessity, insofar as it focuses on whether a certain effect or consequence actually exists or obtains necessarily in reality.Footnote ¹¹

The interplay between conceptual and obtainment necessity is at work in the following passage:

Clarke's account of freedom of indifference is consistent with the existence of both external coercions preventing the agent from bringing about its action, and with the internal necessity imposed on the agent by its own nature. This latter kind of necessity is a conceptual necessity imposed by the nature of the agent itself.Footnote ¹² This conceptual necessity becomes particularly relevant in the case of a self-existent agent that is not acted upon by external causes (such as Spinoza's God). Clarke's argument in section VIII relies heavily on the causal containment axiom, which entails conceptual necessity. It is not viable for Clarke, thus, to deny conceptual necessity as such, since this would jeopardize his whole argument for considering the self-existent cause intelligent and denying that matter can think. However, what is viable for Clarke, is to deny that conceptual necessity entails obtainment necessity. This denial is the kernel of Clarke's account of agent causation.

In Spinoza's ontology, conceptual necessity entails obtainment necessity.Footnote ¹³ Clarke has two main arguments for denying this implication. The first argument builds on his empirical considerations about the fact that the world, as it presently exists, is contingent and this contingency is incompatible with the existence of a self-existent necessary agent. The second argument shows instead that the very idea of accepting that effects necessarily obtain in virtue of their cause generates a paradox of infinite regress.

Concerning the contingency argument, Clark demonstrates in section III that the material world is contingent. As he writes:

Clarke assumes that if “we can conceive [matter] either not to be or to be in any respect otherwise than it now is,” then we can prove the contingency of matter. Clarke takes as mutually connected the contingency of the existence of some entity and the contingency of the way in which it exists. Since the material world and all its parts could exist in many other different ways, or could even not have existed at all, matter is intrinsically contingent. In section IX, Clarke restates this claim about the contingency of the natural world, in opposition to Spinoza's view. Clarke explicitly asserts: “all things in the world appear plainly to be the most arbitrary that can be imagined, and to be wholly the effect not of necessity but of wisdom and choice” (Demonstration IX, 49).

Clarke's contingency argument consists in showing that obtainment necessity, especially when ascribed to the self-existent cause, is incompatible with the contingency of the world (which is observed empirically and thus established independently of considerations on the self-existent cause). If the self-existent cause is the cause of the world and its frame, and if the self-existent cause produced these effects with obtainment necessity, then nothing in the world could be contingent, because the production of the world and its configuration would have been a necessary effect of the self-existent cause's action. However, Clarke stresses that it is (a posteriori) evident that the world and its frame are contingent. Hence, the contingency of the world cannot be accounted for by assuming that the self-existent cause has produced the world with obtainment necessity. There may well be conceptual necessity in the world, but this does not entail that the world itself is brought about with obtainment necessity.

Clarke's regress argument concerns the paradoxes involved by the notion of obtainment necessity itself. If one accepts obtainment necessity, then all effects produced are produced necessarily and no cause can avoid necessarily producing its own effects. This entails that there cannot be any beginning of the causal process. In this argument, ‘beginning’ should not be understood as a mere temporal determination, but more specifically as the point of discrimination t, before which the relevant causal operations do not obtain and it is not necessary that they obtain, while after t the relevant causal operations do obtain. If one accepts that the causal chain of dependent beings unfolds with obtainment necessity, then finite things always produce their effects necessarily when they are brought about by their own causes, which in turn produce them necessarily. In this case, there is no beginning of the causal process because every cause in the process is the necessary effect of a previous cause and the necessary cause of a subsequent state. Hence, there is no node in the causal chain that actually begins the chain itself, which leads to an infinite chain of dependent beings.

In section II (Demonstration II, 11–12), Clarke ruled out the possibility of an infinite chain of dependent beings, since the existence of such a chain would ultimately violate the principle of sufficient reason (Demonstration I, 10). Clarke demonstrated that in order to block the regress it is necessary to posit the existence of a self-existent being. Section IX shows that positing a self-existent being that is not also a free agent could not provide a genuine beginning to the chain of dependent beings, and thus could not solve the regress paradox.

Consider the scenario in which the self-existent being operates with obtainment necessity. Since the self-existent being is eternal, it has been necessarily producing its own effects since eternity. Hence, the self-existent being itself never began the causal process and the very idea of creation would be absurd (which is in fact Spinoza's point).Footnote ¹⁴ However, if the self-existent being never began the causal process, then the infinite regress of causes cannot actually be halted and the whole causal process would be an infinite and ungrounded hierarchy of causes. The infinite chain of dependent beings would remain a “one entire endless series of dependent beings” (Demonstration II, 10). The existence of such an endless series of dependent beings is, in Clarke's terms, “an express contradiction and impossibility” (Demonstration II, 11).

As Clarke writes:

To avoid an infinite regress of dependent causes it is not enough to simply posit a self-existent being, it is also necessary to assume that this self-existent being is a free agent. The only way to avoid the regress paradox and make sense of the contingency of the world is to admit that the self-existent being has the power of beginning causal processes. The power of beginning a causal process amounts to the power of determining a moment t, with respect to which the agent is directly and solely responsible for some effect obtaining or failing to obtain. This power is the power of freely choosing whether or not to bring about the effect. In virtue of this power, there is no necessity for the causal chain to obtain or unfold before the agent's own decision of bringing it about. Hence, the causal chain has a genuine beginning in the agent's free decision. In the case of God, this power allows God to be the creator of the world.

According to Clarke, if and only if an agent is endowed with freedom of choice, then the agent can efficaciously cause its own effects. A free agent has the power of determining by itself whether or not to bring about a certain effect. The kind of effect that the agent brings about may well be necessarily connected with its own nature (i.e. it may be a matter of conceptual necessity). However, the agent's power of either beginning the causal process or not beginning it, makes the agent the genuine starting point of the causal chain. Independently of how strong and vast the determinations acting on a free agent may be, if the agent is endowed with freedom, this agent can choose not to bring about the effect that it is determined to produce. Hence, if the agent freely decides to bring about the effect, it is the agent itself that begins the ensuing causal process, not the whole array of determination that acted upon it. Even if the external causes are such that they ultimately prevent the action from taking place (as in the example of the prisoner mentioned above by Clarke), the agent would still be the genuine initiator and solely responsible for the causal chain that was set in motion.

This is the reason why Clarke equates the lack of freedom with causal passivity: “if there be no liberty anywhere, then there is no agent, no cause, mover, principle, or beginning of motion anywhere” (Demonstration IX, 52). An agent without liberty is no agent at all, because for any given effect, it would not have been the agent itself that set in motion the relevant causal chain. Only if an agent is endowed with liberty and freedom of choice, can the agent account for the production of its own effects as the result of its own free choice. Only in this way, can an agent genuinely possess the “power of beginning motion” (or more generally the power of initiating causal processes) that constitutes the benchmark of causal efficacy.

Clarke's account of agent causation concludes by defining the notion of “agent, or cause” (Demonstration IX, 46), identifying its essential property as a “principle of acting, or power of beginning motion, which is the idea of liberty” (Demonstration IX, 54). Without liberty or choice, there are no causal powers that may be genuinely efficacious or compatible with the causally active role of agents.Footnote ¹⁵ Having demonstrated that liberty is an essential attribute of God (and thus that liberty is possible), and given that God is omnipotent, Clarke concludes in section X that liberty and agency are freely bestowed by God to created intelligent spiritual substances (Demonstration X, 57). Clarke's account of agent causation thus encompasses both God and created spiritual agents (including, but not limited to, human minds).

Finally, it is important to stress that freedom, in Clarke's account, is the continuous exercise over time, moment by moment, of the agent's own power of choosing either to act or not to act. As Clarke writes:

A free agent cannot choose only once and forever. A free agent is free because its power of choosing is constantly exercised over time. Each moment the agent is faced with two mutually contradictory possible states (to act or not to act) and the agent must constantly choose which one will obtain. This entails that the effects that follow from the agent's choice remain constantly dependent on the agent's own choice. Clarke's account of agent causation excludes the possibility of there being only one original act of choice that sets in motion the chain of causes, which then unfold necessarily and without further intervention. Rather, the fact that the exercise of freedom requires a continuous reiteration of the act of choice entails that causal chains are constantly dependent upon the agent that began them (even after their proper beginning), since the agent's choice of beginning a causal chain needs to be constantly confirmed and may be freely changed at any moment. As will become clear in the next two sections, the continuous character of the power of choosing has important implications for Clarke's understanding of how God (and finite intelligent agents) dictates the course of nature.

3. Clarke's Moderate Voluntarism

Voluntarism, as a theological position, is usually contrasted with intellectualism. Both voluntarism and intellectualism deal with the way in which intellect and will are related in God. Intellectualism holds that God's will is determined by God's intellect. Malebranche and Leibniz, for instance, develop different versions of this intellectualist claim. Voluntarism, on the contrary, makes God's will somehow independent of God's intellect. While God may be morally determined to act on the basis of what he understands to be the best, God's freedom is in no way restricted to acting in any particular way, or to keeping the same course of action for any determinate amount of time. Descartes, Boyle, and Newton, for instance, are usually taken to be representative of different strands of voluntarism.Footnote ¹⁶

For present purposes, I maintain that Clarke's account of agent causation offers independent philosophical support to Clarke's voluntarism and simultaneously helps to further qualify it. From a strategic point of view, the philosophical machinery entailed by his account of agent causation is particularly effective in the context of Clarke's apologetic refutation of atheism and deism. Since Clarke's opponents are not receptive to purely theological arguments, Clarke's philosophical account of agent causation paves the way for a defense of voluntarism while remaining on a purely philosophical footing.Footnote ¹⁷ From a philosophical point of view, moreover, the same account contributes to qualifying the kind of voluntarism endorsed by Clarke as a moderate voluntarism.

The philosophical junction between Clarke's account of agent causation and his voluntarism is established in section IX of the Demonstration. Clarke's account of agent causation entails that for an agent to be properly called an agent (i.e. causally efficacious), it is essential that the agent is free to choose whether to act or not. This freedom of indifference is what endows the agent with the power of beginning motion. Conversely, by denying freedom of indifference, the agent is reduced to a mere necessitated and passive agent which is, ultimately, no agent at all. Clarke argues that it is necessary to consider God as a free agent in this sense. The kernel of Clarke's voluntarism is thus that God is ultimately free to choose whether or not to bring about an effect. Any one of God's actions is arbitrary, in the sense that it follows uniquely from God's free decision of pursuing that course of action. Such a decision does not put any constraints on God's absolute freedom of changing it. God must be free in this way in order to be an agent. Hence, Clarke's voluntarism is not just a theological presupposition, but is the only viable philosophical option left by Clarke's rigorous understanding of causation in terms of agent causation.

In turn, Clarke's account of agent causation clarifies the scope of his voluntarism. As discussed in section two, Clarke's account of agent causation does not reject conceptual necessity, but rather it rejects that conceptual necessity entails obtainment necessity. Clarke's account is compatible with the existence of conceptually necessary connections, although it does not allow that these connections constrain an agent's freedom of choosing. Descartes’ doctrine of the free creation of eternal truths is often taken to represent a radical form of voluntarism, since it entails that both conceptual necessity and logical connections or laws depend on the mere will of God. In this respect, Clarke's voluntarism (in contrast) can be qualified as moderate. Clarke maintains that God is somehow bound by the laws of logic (Demonstration X, 55; XII, 88). Moreover, Clarke does not deny the existence of conceptual necessity, and instead crucially exploits the causal containment axiom to support his case against thinking matter. All that Clarke rejects is that conceptual necessity entails obtainment necessity.Footnote ¹⁸

Qualifying Clarke's voluntarism as moderate and understanding how it makes room for conceptual necessity is important in order to appreciate the overall consistency of Clarke's philosophical project and how he reacted to objections raised by his interlocutors. In the rest of this section I shall focus on two relevant cases: (i) the objection raised by William Carroll against Clarke's doctrine of the “eternal reasons” of things; and (ii) the discussion of the principle of sufficient reason and the issue of miracles in the context of Clarke's correspondence with Leibniz.

(i) In the last two sections of the Demonstration Clarke establishes that “there is, therefore, such a thing as fitness and unfitness, eternally, necessarily, and unchangeably in the nature and reason of things” (Demonstration XII, 83). This claim is the basis for Clarke's ethics, which is developed in the Discourse.Footnote ¹⁹ However, there seems to be a contradiction between the discussion of contingency of matter and the material world and the treatment of “eternal reasons” holding among things and generating “unalterable” relationships of good and evil.

At the beginning of the Discourse, Clarke engages in a set of replies to William Carroll, who attacked his Demonstration with a series of Remarks (published in 1705). Clarke summarizes one of Carroll's major criticisms as follows:

Clarke is not much impressed by this criticism. As he swiftly rebuts:

Clarke's reply to Carroll is built on the exploitation of the distinction between conceptual necessity and obtainment necessity. Clarke's claim concerning the “eternal reasons” that link the nature of different things concerns conceptual necessity. Given the nature of a sphere and a cylinder, for instance, it is necessary that they stand in a certain relationship, just because the very nature of being a sphere and being a cylinder entails so. However, Clarke emphasizes, this does not entail that one has to admit that there are any really existent spheres or cylinders. Conceptual necessity does not entail obtainment necessity.

The “eternal reasons” that Clarke mentions originate from the relationship among ideas. However, God is a free agent. The very nature of freedom, according to Clarke, consists in constantly having the power of withdrawing action. Hence, God is not bound in any way to create the world (or conserve it in the same shape) because the very nature of freedom of indifference entails that there is no obtainment necessity that bounds God's will to create a certain world or maintain it. This entails that the “eternal reasons” among things are “eternal” insofar as they concern God's ideas. However, God's free and arbitrary decision to create a world according to his ideas (and to continue to maintain it) could be withdrawn at any moment (in principle). In this way, Clarke develops his strong account of moral values based on eternal relationships (and entailing conceptual necessity) without undermining his argument about the contingency of the created world (which concerns obtainment necessity).Footnote ²⁰

(ii) The moderate nature of Clarke's voluntarism emerges at two crucial junctions in Clarke's correspondence with Leibniz: in connection with the discussion of the principle of sufficient reason (PSR) and with the issue of miracles.

Leibniz introduces PSR in his second paper:

Leibniz's formulation of PSR is broad enough to encompass both the existence of things and their nature. In the course of the correspondence, Leibniz's strategy consists in relying on PSR (which Clarke does not formally reject) in order to show, for instance, that the Newtonian doctrine of absolute space is incoherent (LC Reference Leibniz, Clarke and Alexander1956, Leibniz's Third Paper, 26). For present purposes, I shall leave aside the details of the argumentation against absolute space and how it relates to PSR.Footnote ²¹ I focus instead on how Clarke deals with PSR in relation to God's will. As Clarke writes:

This formulation reflects the formulation that Clarke previously provided in the Demonstration (I, 8). According to Clarke, the “mere will of God” counts as a sufficient reason for the existence of certain things and an explanation as to why they should be created in a certain way. This is a restatement of Clarke's rejection of the inference from conceptual necessity to obtainment necessity: the existence of things depends entirely on God's free decision of creating them, independently of any further predetermining reason. Clarke's account of agent causation applied to God entails that God's free will is always the necessary and sufficient reason for the existence of things, and the same freedom prevents anything else from being considered equally necessary or sufficient. Clarke reaffirms here the point made in the Demonstration: God must have freedom of indifference in order to be an agent at all. PSR must be thus accommodated with God's freedom of indifference in order to avoid fatalism (i.e. Spinoza's position and that of his followers).

It is worth noting that Clarke's understanding of PSR does not rule out the possibility that, in some cases, a sufficient reason might be something more than the “mere will of God.” As Clarke says, the “mere will of God” is “oft-times” the sufficient reason of things (especially when it comes to their existence), but this does not preclude that some sufficient reasons may have a further grounding in other considerations.Footnote ²² Clarke's view about the “eternal reasons” that link the nature of things may offer a good example in this direction. The sufficient reason as to why certain configurations of the world are better than others can be accounted for by the way in which, in these configurations, the nature of different things fit together in virtue of the “eternal reasons” that connect their natures. Clarke is not explicit about this kind of sufficient reason in the context of his correspondence with Leibniz, because this would not be a matter of disagreement with him. However, Clarke does rule out (as in his reply to Carroll) that any of these considerations may constrain God's ultimately free decision of following them or not.

Leibniz immediately notices that Clarke's use of PSR aims to secure God's indifference, which Leibniz strongly rejects. As Leibniz writes:

According to Leibniz, accepting any form of indifference (in either creatures or in God) amounts to denying PSR altogether. Since Clarke's account of agent causation is built on the idea that agency requires indifference, Leibniz is eager to stress the incompatibility of Clarke's account with PSR. Clarke's appeal to the “mere will of God” is not intended to show that the “mere will of God” is necessarily the cause of “eternal reasons” among things. Rather, his point is more moderate: it aims to safeguard God's indifference, especially concerning God's act of bringing about the existence of things. Nonetheless, even Clarke's moderate voluntarism is still too strong for Leibniz, and it is too strong because of Clarke's acceptance of indifference and Leibniz's firm rejection of it. For present purposes, thus, it is important to stress that the correspondence does not prove that Clarke supports a radical version of voluntarism, but rather that the genuine crux of the controversy with Leibniz is Clarke's account of indifference, which is at the core of his account of agent causation.Footnote ²³

Another connected point of disagreement between Leibniz and Clarke concerns miracles. Leibniz raises the question of whether natural operations are commensurate with the nature of things that are supposed to be their causes. A few years earlier, Leibniz had explicitly charged Malebranche's occasionalism of reducing all natural operations to a kind of perpetual miracle insofar as occasionalism denies that natural operations depend upon the nature of things.Footnote ²⁴ Leibniz understood Malebranche as holding that the nature of created things does not maintain any traces of God's laws and thus God has to constantly reiterate his commands in order to produce regular effects over time. In his De Ipsa Natura (1698), Leibniz explains: “it is not sufficient to say that God, creating things in the beginning, willed that they follow a certain definite law in their change [progressus] if we imagine his will to have been so ineffective that things were not affected by it and no lasting effect was produced in them” (G 1875–90, IV 507/AG 1989, 158). In writing to Clarke, Leibniz contends: “that which is supernatural, exceeds all the powers of creatures. . . . I maintain that the attraction of bodies, properly so called, is a miraculous thing, since it cannot be explained by the nature of bodies” (LC Reference Leibniz, Clarke and Alexander1956, Leibniz's Third Paper, 30). As in the De Ipsa Natura, Leibniz focuses on whether or not a certain effect follows from the nature of the thing that is supposed to be its cause. Against Malebranche, Leibniz argued that occasionalism entails God's perpetual intervention. Against Clarke, Leibniz tries to push a similar argument, by insisting that since gravity or attraction cannot be accounted for by the nature of bodies, gravity too entails a perpetual miracle on God's part.Footnote ²⁵

In his reply to Leibniz, Clarke maintains that “nature” is an empty word:

Clarke disagrees with Leibniz's definition of what counts as a miracle. Contrary to Leibniz, Clarke does not accept that the distinction between the domain of the natural and that of the miraculous depends on whether a certain effect can be accounted for by the nature of its supposed natural cause. Echoing Augustine, Clarke rather maintains that the distinction between the natural and the supernatural lies merely in the frequency with which God himself brings about a certain phenomenon.Footnote ²⁶

Once again, it is important to distinguish what this claim denies from what it does not deny. Clarke uses “nature” to refer to the whole of nature, the course of nature, and powers that can be attributed to Nature. “Nature” here is taken to be a collective term for the whole of the natural world and the causal relationships that it entails. Arguing against this notion at the beginning of the eighteenth century is not a novelty per se.Footnote ²⁷ However, the argument can be developed in different ways.

Malebranche (whose position is often in the back of Leibniz's mind during the correspondence with Clarke) most explicitly rejected the (scholastic) tenet that finite things can have their own individual natures, which can express some degree of causal efficacy in their own right. The course of nature as a whole is nothing but the result of God's direct ruling of creatures. According to Malebranche: “I derive nothing from my nature, nothing from the imaginary nature of the philosophers. Everything comes from God and His decrees” (OC.xii-xiii.165/JS 121).Footnote ²⁸ The nature of things has no causal efficacy and entails no causal implications independently of God's laws. As mentioned, Leibniz criticized this claim in his De Ipsa Natura.Footnote ²⁹

However, in his reply to Leibniz, Clarke does not refer to the individual nature of things but to the whole course of nature understood as a collective term. This means that, in order to counter Leibniz's view, Clarke does not need to rule out that things have natures that may somehow account for particular phenomena.Footnote ³⁰ Clarke does not enter at all in the discussion of this issue with Leibniz. Rather, he simply points out that whatever nature things may have, since they depend on God, and God's will is free and unbounded, the “course of nature” is absolutely contingent. Once again, Clarke does not seem to push his voluntarism to the extreme of rejecting that, in some relevant sense, different things may have their own individual natures, which are connected by “eternal reasons.” Such a rejection would not square with Clarke's moral project. Clarke's account of agent causation provides him with a much easier and safer way to counter Leibniz.

The laws of nature concern the actual and regular obtainment of certain effects in the natural world. The course of nature is the result of God's choice and thus falls within the domain of God's freedom of indifference. As mentioned at the end of section two, Clarke maintains that the essence of freedom consists in the agent's continuous exercise of its power of choosing. In the case of God, God's establishment of certain laws of nature cannot be done instantaneously and thereby hold permanently. Rather, insofar as it is the result of a free choice, it has to be reiterated at every moment and remains open to be freely changed at any time. Hence, whatever nomological regularity may be observed in nature, such a regularity cannot be regarded as obtaining necessarily, since it continuously depends on God's freedom of choosing otherwise. Clarke's account of agent causation shows that there is no obtainment necessity in the natural world. Insofar as the world depends on a free agent, the world is necessarily contingent.

Leibniz's strategy is to collapse Clarke's position with Malebranche's. However, Clarke has two very good reasons to resist this move and instead develop his own version of the denial of the course of nature. Firstly, both Leibniz and Malebranche reject the notion of indifference and subscribe to some form of intellectualism.Footnote ³¹ Freedom of indifference is, however, at the core of Clarke's account of agent causation and puts him on the voluntarist side of the dispute. Secondly, Clarke's denial of any obtainment necessity in the ordinary course of nature does not commit him to reject (as an extreme voluntarist would do, pace Henry Reference Henry2009) the idea that things may have their own natures, related by “eternal reasons” (which entails a form of conceptual necessity). However, Clarke's account of eternal reasons is not a concession to the intellectualist party either (pace Harrison Reference Harrison2002 and Reference Harrison, Force and Hutton2004), since it does not put any constraints on the exercise of God's own freedom, which remains completely unbounded and indifferent. Clarke's account of agent causation and his moderate voluntarism thus allow him to resist Leibniz's charges without having to come close to, let alone endorse, anything like Malebranche's own position.

4. Partial Occasionalism in Support of Agent Causation

Occasionalism tackles two different questions: (1) what is the kind of causality that belongs to certain entities? and (2) what is the source of causal efficacy in nature? Although these two questions are connected, they must not be conflated. Often, the solution to the second question leads one to identify God's immediate causal involvement in nature as the only source of causal activity in the world, and this has often been taken to be the most distinctive feature of occasionalism. Partial occasionalism, however, maintains that God's causal intervention is only directly needed in a certain domain, which is usually identified with that of body-body causation. Partial occasionalism allows both God and certain entities to be genuine causes of certain phenomena, while it rules out that other entities are. With respect to the first question, both partial and unrestricted occasionalism consider the entities devoid of efficacious causal powers as sine quibus non conditions for the genuinely efficacious agent(s) to operate. This does not mean that non-efficacious entities are irrelevant for the causal processes, but rather that they play only a passive role and that, by themselves, they are neither sufficient nor responsible for directly bringing about the relevant effects.Footnote ³²

Late scholastics and a number of early modern authors rejected occasionalism and embraced concurrentism instead.Footnote ³³ To the first question, concurrentist accounts reply that while God is the primary cause of all effects, creatures are secondary causes, which are genuinely efficacious (or per se causes), although their efficacy requires God's concurrence to work. To the second question, therefore, concurrentist accounts reply that, at least in the ordinary course of nature, causal efficacy has two roots: God and the nature of creatures themselves.

Unrestricted occasionalism and concurrentism are mutually exclusive insofar as unrestricted occasionalism denies and concurrentism defends that creatures are genuinely efficacious. Without genuine causal efficacy rooted in the nature of things, God has nothing to concur with and is left as the sole genuine cause. To use scholastic terminology, this divergence is captured by saying that, in the concurrentist scenario, not only God but also creatures are per se causes (even if only in their own ontological order and subordinated to God's concurrence), while in the unrestricted occasionalist scenario God is the only per se cause and creatures are mere sine quibus non conditions for God's operations.

Despite this radical divergence, however, concurrentism and partial occasionalism may be combined. Insofar as partial occasionalism restricts the occasionalist account of causation to one specific domain of creatures (bodies and material entities), it leaves room in other domains (e.g. minds) to be genuinely efficacious. God has thus to be directly involved as the sole source of causal efficacy in the purely physical domain, while in the spiritual domain (and in every domain in which God decides to make minds responsible for certain bodily effects, such as in the case of mind-body union) minds also are active. Louis de la Forge is one of the first early modern authors to defend this kind of partial occasionalism.Footnote ³⁴ I contend that Clarke too develops his own version of partial occasionalism, which is distinctively shaped by his account of agent causation and his commitment to Newtonian physics.

From a systematic point of view, partial occasionalism (in the physical realm) follows as an immediate consequence from Clarke's account of agent causation. Agent causation (which is the only account that Clarke accepts) simply does not allow non-intelligent beings to act as genuine efficacious causes. Every effect attributed to non-intelligent beings must be caused by the intervention of some intelligent being. However, in terms of argumentative efficacy, Clarke can also exploit partial occasionalism itself as an independent resource to strengthen the case for his account of agent causation. If material bodies cannot be properly considered to be agents or causes, then the only kind of proper causation in nature must be ascribed to either God or to intelligent spiritual beings.

This latter strategy is fully developed in Clarke's Discourse, which follows the Demonstration and continues Clarke's fight against atheists and deists. Clarke has a good reason for adopting this strategy. As I discussed in section two, Clarke's account of agent causation relies on the argument put forward in section VIII of the Demonstration, according to which purely material beings cannot cause or give rise to intelligence. This point is controversial in itself.Footnote ³⁵ If Clarke can exploit partial occasionalism to offer an independent proof that non-intelligent (i.e. material) beings are devoid of causal efficacy anyway, he can render it more plausible that agent causation is the only viable account of causation.Footnote ³⁶

To establish partial occasionalism in the physical domain, one must rule out the possibility that bodies or material beings can be genuine per se causes. Bodies could play a causal role in natural processes in two (unexclusive) ways: (i) by contributing to the causal process in virtue of some intrinsic efficacious causal power that they would have (as traditional Aristotelian scholastic authors, for instance, would maintain); or (ii) by operating on the basis of certain laws of nature that would be intrinsically connected with the nature of bodies (as Leibniz would contend). In the Discourse, Clarke offers arguments against both these options.

Concerning causal powers, in the Discourse Clarke explains: “matter being evidently not at all capable of any laws or powers whatsoever, any more than it is capable of intelligence, excepting only this one negative power, that every part of it will, of itself, always and necessarily continue in that state, whether of rest or motion, wherein it at present is” (Discourse, 341). Some scholars (e.g. Yenter and Vailati Reference Yenter, Vailati and Zalta2014) maintain that inertia associated with matter can be responsible for the production of some positive effect. This reading would be supported by several passages, especially in Clarke's commentary on Rohault's System of Natural Philosophy (first Latin edition published in 1697) in which Clarke connects inertia and the power of resisting motion.Footnote ³⁷

I contend that this reading should be resisted and that Clarke's considered view suggests that all causal powers that might be attributed to matter are merely negative powers, which do not have any causal efficacy on their own. I argue two main points in support of this contention: (1) inertia, impenetrability or any other properties that can be attributed to matter do not have any intrinsic causal efficacy; and (2) matter cannot ground the laws of nature that are required for the unfolding of natural processes.

First, understood in the context of seventeenth-century natural philosophy, an efficacious (or active) causal power is taken to be what allows a certain substance to initiate a new course of action or bring about some change. Clarke conforms to this conceptual usage. As discussed in section two, when he wants to stress the real causal efficacy of intelligent agents, he stresses their “power of beginning motion.” However, inertia merely expresses the fact that matter, by itself, cannot produce anything and thus simply continues to remain in the same state until some external cause changes it. This permanence in the same state amounts to nothing but the sheer lack of causal activity to which matter is abandoned when it is not acted upon by some efficacious cause. This is why Clarke calls inertia a “negative” power (which corresponds to what, in scholastic terminology, is more usually called a passive power, namely, the capacity of receiving some change produced by a genuinely efficacious agent). According to Clarke, inertia is not the only negative power of matter. In fact, all features usually attributed to matter count in the same way as negative powers. As he writes: “figure, divisibility, mobility, and other such like qualities of matter, are not real, proper, distinct, and positive powers, but only negative qualities, deficiencies, or imperfections” (Demonstration VIII, 41). None of these features, in Clarke's eyes, mean that matter acts as a genuinely efficacious cause which is able to initiate new courses of action. In this way, Clarke endorses the widespread claim at the time that matter is utterly passive.Footnote ³⁸

Second, Clarke rules out the possibility that matter operates on the basis of certain laws of nature intrinsic to it. Clarke argues that matter by itself is incapable of obeying any laws. As he writes:

Clarke's argument relies on the empirical observation that “the very original laws of motion” apply from one moment to the next and remain seemingly constant and regular over a certain amount of time. Of course, God may change the course of nature at any time. However, the argument here concerns the empirical observation that the laws of nature do seem to remain constant for at least some time. Clarke's point is that even this minimal form of nomological constancy over time cannot be grounded in the nature of matter as such. As discussed in sections two and three, Clarke holds that matter is intrinsically contingent and thus cannot provide the sufficient reason for the regular obtainment of natural effects captured by the laws of motion.

In Clarke's account, the denial that matter can obey any law and the claim that matter is utterly passive mutually support each other. For, on the one hand, if matter by itself is utterly passive, it cannot account for the regular production of natural effects that follow according to the laws of nature, because matter could not produce any of those effects in the first place. While on the other hand, if matter by itself is incapable of obeying any laws, and natural effects occur in a lawful order, then matter by itself cannot account for these effects because it cannot account for their regular and lawful unfolding.Footnote ⁴⁰

This argument paves the way towards recognising God as the only immediate cause of natural effects in the physical realm. As Clarke argues:

Clarke's claim that God is “perpetually and actually exerting” his power “every moment in every part of the world,” conjoined with the fact that material entities cannot concur in any active way with the production of natural effects (neither through intrinsic powers nor by following natural laws intrinsic to their nature), leads to the conclusion that only God can be the genuine cause of natural phenomena associated with bodies.

Clarke's position is even more explicit when he restates his view later in the Discourse:

In the passage here emphasized, Clarke does not merely state that God is constantly involved in the operations of nature, but he rather advocates that the whole course of nature is “nothing else but” God's will of producing certain effects. This phrasing (which is similar to his reply to Leibniz) entails that God's will is the immediate cause that sets in motion and guides the course of nature.

However, in both the passages here quoted, Clarke makes room for the possibility that God does not cause all phenomena associated with bodies directly and immediately. Clarke admits that God may use “some subordinate instruments appointed by him to direct and preside respectively over certain parts thereof” (Discourse, 142–143); or that God acts “upon matter continually and every moment, either immediately by himself, or mediately by some created intelligent beings” (Discourse, 341). As this latter reference makes clear, the instruments that God can use to act mediately upon matter are intelligent beings.Footnote ⁴¹

It may be problematic to maintain partial occasionalism if intermediary intelligent agents are always responsible for the production of effects in the physical realm. To better understand the nature of this problem, consider again the two questions (outlined at the beginning of this section) that an occasionalist account of causation aims to address: (1) what is the kind of causality that belongs to certain entities? and (2) what is the source of causal efficacy in nature?

I contend that Clarke is not reticent about the first question. It is clear, from the texts quoted so far, that Clarke does not allow purely material entities to be endowed with any efficacious causal power and that he maintains that all effects attributed to these entities must be the result of some spiritual entity operating on them. Clarke's account emphasizes that, in the case of entities that cannot begin causal processes or do not have powers of bringing about certain effects (i.e. material and non-intelligent beings), some other kind of agent (i.e. intelligent and free agents) must be directly involved in the production of these effects. The causal role of material entities is that of occasional causes, which operate as sine quibus non conditions for the relevant spiritual agent to bring about its own effects. In order for gravitational phenomena to take place, for instance, material bodies must exist, have a certain mass and share spatial relationships among them. These features, by themselves, are only negative powers, meaning that they are not able by themselves to initiate causal processes. However, they determine the circumstances in which causal processes take place and are in this sense necessarily required (although not sufficient) for the real agents to operate. Intelligent agents operating on matter cannot be conceived as concurring with matter and its powers in jointly bringing about certain effects, since the powers of matter are purely negative, as Clarke stresses. This means that matter, by itself, does not have any power to determine any particular effect. Genuinely efficacious agents do not have anything to concur with. Material beings only provide the sine quibus non conditions for genuine agents to freely decide which effects they wish to bring about in these material beings.

Clarke, however, is admittedly reticent about the second question. He does explicitly stress the possibility of God allowing intermediary spiritual entities to act as genuine causes of certain phenomena. Nonetheless, I shall now argue that Clarke's considered view is that God does not rely on these intermediary spiritual entities to rule all phenomena. In the end, thus, Clarke supports the partial occasionalist claim that God is immediately and solely causally responsible for at least some natural phenomena.

Consider the following passage:

Gravity seems a universal principle that can account for “almost all the great and regular inanimate motions in the world.” Now, gravity itself is not a cause but the effect of the force exercised by some agent on matter. Clarke also stresses that the force of gravity exercised by this agent is “entirely different from that by which matter acts on matter,” meaning that it is not a mechanical force or anything like the vis inertiae that belongs to matter.Footnote ⁴² Hence, as the context of the passage makes clear, the agent that causes gravity must be an immaterial, intelligent agent. Clarke also remarks: “necessity be caused (either immediately or mediately) by something which penetrates the very solid substance of all bodies.” This line may be invoked as evidence for the fact that Clarke accepts some intermediary principle always mediating between God and nature. If gravity is a universal principle that shapes the whole of nature, and if gravity is caused by a single agent that is different from God, then God is not immediately involved in the ruling of nature. Rather, in such a case, he appoints an intermediary agent to take care of the course of natural phenomena (at least those connected with gravity).

However, from a textual point of view, the scenario in which all phenomena are ruled by intermediary agents is at odds with Clarke's remark later in the Discourse according to which “the course of nature, truly and properly speaking, is nothing else but the will of God producing certain effects in a continued, regular, constant, and uniform manner” (Discourse, 341). If the course of nature is the result of the decision of any number of agents different from God, then the course of nature cannot be taken to be nothing but God's will, since the volitions of intermediary agents, in virtue of their freedom, cannot be taken to be identical to God's own will. Although the intermediary agent's freedom is subordinated to God's own decision to bestow such a freedom to that agent, it remains a genuine freedom packed with the power of beginning (or not beginning) motion. When God appoints an intermediary agent to rule a certain phenomenon, then this phenomenon depends immediately on the intermediary agent's free decision and no longer immediately on God's own will (unless God makes the further decision of overpowering the agent or destroying it).

Given the constraints of Clarke's account of agent causation, it is also not possible to use a form of concurrentism (between God and intermediary agents) in order to simultaneously maintain that (i) God is immediately involved in ruling the course of nature and that (ii) God appoints intermediary agents to rule all natural phenomena. This is because God can immediately concur with the intermediary agent only in the sense of upholding in existence the agent's own power of beginning motion and its freedom (Demonstration X, 62). However, this concurrence cannot extend to the fact that God concurs with the agent's way of using its freedom to begin certain causal chains or refrains from beginning them. Divine concurrence with respect to this choice itself would in fact be a genuine causal interference and, given God's omnipotence, it would always result in God overpowering the finite agent. Since this would undermine the agent's own freedom of indifference, the agent itself would cease to be a genuine agent at all (by Clarke's standards). Moreover, it would not make sense to appoint an intelligent agent as responsible for some effect (which means establishing that that effect follows from the intermediary agent's own free choice), if God would then have to overpower the agent's free choice and undermine the agent's own agency. Hence, in Clarke's account, God cannot be both immediately responsible for the course of nature and appoint an intermediary agent to rule the same natural phenomena that God rules.Footnote ⁴³

Clarke's remarks about intermediary agents are meant to leave open the discussion about which intermediary agents are responsible for certain phenomena and which phenomena are ruled by intermediary agents rather than by God himself. However, Clarke also explicitly states that the course of nature is nothing but (and thus immediately depends) on God's own will. Hence, the use of intermediary agents cannot be so extensive that no phenomenon in nature is ruled immediately by God.

This conclusion is supported by the way in which Clarke's discussion continues just after the previous quote. Clarke argues:

Clarke stresses here that God's appointment of created intelligences as agents directly responsible for certain processes is more an exception to God's general immediate ruling of the course of nature. Clarke does not suggest that God appoints an intermediary created intelligence for every phenomenon in nature. Clarke takes the core of deism (and atheism) to entail that “God does not concern himself in the government of the world.” For this reason, Clarke has good reasons to refrain from accepting that God may have delegated the whole course of nature and of all natural phenomena to created intelligences, since in this case it would be hard to see in which sense one may still maintain that God governs the world, given that, insofar as free intelligent agents are responsible for certain phenomena, they are the agents that initiate and bring about the relevant causal chains.

My interpretation is also consistent with Clarke's reference, later in the Discourse, to the role of angels. Clarke writes:

This passage might suggest that created intelligent agents (angels in this context) are most probably the constant intermediaries between God and the production of natural effects. However, in the context of this passage, Clarke discusses the issue of miracles in connection with the ordinary course of nature. As already mentioned, if Clarke had maintained that angels are always the intermediary agents that mediate between God and the production of natural effects, then he would have to deny that God is immediately involved in the operations of nature altogether. Yet Clarke seems far from willing to claim that God cannot be immediately responsible for the ordinary course of nature (e.g. Discourse, 341). In fact, this is far from the most obvious and natural reading of the passage quoted above. Arguably, Clarke is restating the claim, already discussed, that the course of nature is nothing but God's decree of operating in a constant and regular way. For this reason, the course of nature is fundamentally arbitrary, although it appears regular. If intelligent agents, such as angels, are considered to be intermediaries between God and nature, and if they are considered to be responsible for some phenomena, then (Clarke's argument goes) these phenomena would still remain fully arbitrary (in the sense that they would still depend on the free decision of intelligent agents, which in this case would be angels rather than God). Clarke's point is that the course of nature remains arbitrary (i.e. dependent on the free decision of intelligent agents) even if one assumes that there are intermediary agents between God and nature.Footnote ⁴⁴ Clarke does not take a definite position on how extensive the involvement of intermediary agents is and how the division of labour between them and God works. However, this does not mean that Clarke holds that natural phenomena are always immediately caused by intermediary agents, nor is this the only or more plausible reading suggested by the texts themselves.

From Clarke's point of view, the most appealing aspect of partial occasionalism is in relation to the problem of causality, more than the source of causal activity. Insofar as partial occasionalism rules out the fact that certain kind of entities (i.e. material beings) could be the right kind of entity to have genuine causal efficacy, it strengthens the case for Clarke's own account of agent causation, which shows that God and created intelligent agents can be the only source of causal efficacy. From this point of view, what may appear to be an ambiguity in Clarke's position concerning whether God is the only cause operating in nature is in fact more plausibly the result of Clarke's instrumental use of partial occasionalism to strengthen his case for agent causation. Clarke's goal is not to demonstrate that God is the only cause, but rather to show that intelligent agents are the only kind of entity that can be an efficacious cause at all, since material beings lack causal efficacy in any relevant sense.

These considerations suggest that, from Clarke's point of view, partial occasionalism is not just an alternative to unrestricted occasionalism à la Malebranche, but that it is the only viable kind of occasionalism and that unrestricted occasionalism is to be denied on both philosophical grounds (because it denies that God can communicate the power of beginning motion to created beings) and moral grounds (because it ultimately leads to some form of fatalism).Footnote ⁴⁵ Recently, scholars (Nadler Reference Nadler2010, 123–141) have expressed reservations about the extent to which partial occasionalism can be a consistent position. Clarke's endorsement of partial occasionalism shows that, in the early eighteenth-century context, (at least some version of) partial occasionalism was not only sound, but was taken to be much less problematic than unrestricted occasionalism.

To conclude, partial occasionalism and Clarke's account of agent causation mutually support one another. On the one hand, agent causation entails that material and non-intelligent beings cannot be the right kind of thing to play any properly efficacious causal role, since they lack intelligence and freedom, which are required for genuine agency. On the other hand, Clarke's main arguments for partial occasionalism concern the fact that (i) matter is utterly passive from a causal point of view, and (ii) matter cannot by itself follow any law. Both arguments are purely negative and do not directly invoke Clarke's positive view about agent causation. However, insofar as these arguments for partial occasionalism rule out the idea that unintelligent material bodies can be genuine causes, they also make it more plausible that agent causation is the only viable model of causality and provides what Locke called a “clear and distinct idea of active power” (II.21.4).

From a historical point of view, Clarke's account of agent causation highlights three main features of the debate on causation across the seventeenth and eighteenth centuries. First, a crucial concern at the time consisted in rethinking the nature of causal efficacy and the requirements that an entity must satisfy in order to count as a causal agent in a proper sense. Instead of offering a general account of causation that encompasses both intelligent and non-intelligent agents, Clarke's account of agent causation shows that only free intelligent agents (like humans, and angels) can be genuine causal agents. Second, the issue of the passivity of material beings and the direct involvement of God's causal efficacy in nature play a crucial role in shaping this debate. Clarke is ready to trace a sharp boundary between material beings and free intelligent agents. While free intelligent agents receive their causal powers from God, who then allows them to operate as genuine (though subordinate) causal actors, material beings are utterly devoid of efficacious causal powers. Natural phenomena thus require God's immediate and continuous intervention (or that of other subordinated free intelligent agents) in order to take place. Third, occasionalism must be understood as a family of arguments and cannot be reduced to the specific variety of occasionalism developed by Malebranche. Clarke's account of agent causation is at odds with Malebranche's denial of the causal efficacy of human minds. Nonetheless, Clarke's endorsement of partial occasionalism is both a theoretical consequence of his account of agent causation and an external support for strengthening it. From this point of view, occasionalism travels and evolves well beyond Cartesian circles and plays a systematic role in the broader rethinking of the notion of causation during the early modern period.