Cosmopsychism and the Problem of Evil

Abstract

Cosmopsychism, the idea that the universe is conscious, is experiencing something of a revival as an explanation of consciousness in philosophy of mind and is also making inroads into philosophy of religion. In the latter field, it has been used to formulate models of certain forms of theism, such as pantheism and panentheism, and has also been proposed as a rival to the classical theism of the Abrahamic faiths. It has been claimed by Philip Goff that a certain form of cosmopsychism, namely agentive cosmopsychism, poses a threat to classical theism because it can explain features of the universe like fine-tuning without having to deal with the problem of evil. This is because, unlike the classical theist, the cosmopsychist can deny at least one of the divine attributes motivating the problem of evil, namely omniscience, omnipotence, and omnibenevolence. In this paper, I shall consider which of the divine attributes the cosmopsychist should focus on when responding to the problem of evil and shall conclude that the rejection of omnibenevolence is the most satisfactory option.

Introduction

In philosophy of mind interest in cosmopsychism, the idea that the universe is conscious,¹ is increasing, and it is now seen to be a compelling solution to the hard problem of consciousness (accounting for subjective, or phenomenal, experience) by a growing minority of scholars.² Cosmopsychism is a form of panpsychism, the idea that consciousness is ubiquitous in the cosmos.³ In the contemporary literature, panpsychism is often developed as a view in which the fundamental entities are things like quarks and protons and these have mental properties, and this is distinct from cosmopsychism.⁴

The basic principle of cosmopsychism is fairly broad, and the position can be developed in a number of different ways. A particularly influential form of cosmopsychism in contemporary discussions is priority cosmopsychism, in which the consciousness of the cosmos is ontologically prior to the consciousnesses of individual organisms. The fundamental entity here is the cosmos itself.⁵ Priority cosmopsychism is a development of panpsychism influenced by Jonathan Schaffer’s priority monism, in which the cosmos, the whole, is the only fundamental entity and is prior to its parts.⁶

Cosmopsychism has also begun to enter debates in philosophy of religion. Recent research has attempted to use cosmopsychism as a way of understanding particular forms of theism, such as pantheism and panentheism.⁷ It has also been used in opposition to certain forms, namely the classical theism of the Abrahamic religions. It has been argued that certain forms of cosmopsychism can capture key advantages of classical theism whilst avoiding some of its most troublesome criticisms.

The relevant variants to consider are forms of agentive cosmopsychism, in which the universe can be deemed an intelligent, rational agent with its own first-person perspective.⁸ Due to the cosmic subject’s possession of these features, it is thought that these models can be used to explain things like fine-tuning, the fact that the universe appears to be fine-tuned for life. Since the cosmopsychist is not committed to the cosmic subject’s possession of the divine attributes of the God of classical theism, it is also argued that the problem of evil can be avoided. In this paper, I shall consider how the agentive cosmopsychist should go about avoiding the problem of evil.

I shall begin by outlining how it might be argued that the cosmopsychist can account for fine-tuning whilst avoiding the problem of evil by denying the cosmic subject’s possession of at least one of the divine attributes used in the argument. I shall then consider which divine attribute the cosmopsychist should focus on when attempting to avoid the problem of evil. I shall argue that the denial of omnibenevolence is the preferable option.

Agentive Cosmopsychism and Classical Theism

Let us begin by considering agentive cosmopsychism in more detail. The basic thesis of cosmopsychism, that the universe is conscious, does not automatically entail that the universe is an intelligent agent. Whether one is able to provide an agentive model depends on one’s answer to the question of how other subjects relate to the cosmic subject. There are two main approaches to this question, and I shall discuss each in turn and assess their suitability for use in an agentive model.

The first is constitutive cosmopsychism, in which there is a constitutive relation between other subjects and the cosmic subject. A constitutive relation is ‘where all the true statements about human consciousness are (wholly or partially) grounded in true statements about fundamental consciousness.’⁹ In constitutive cosmopsychism, macro-subjects (subjects with macro consciousness, like humans) are metaphysically grounded in the cosmic subject.¹⁰

What does it mean for something to be grounded in something else? In metaphysics, grounding is the explanatory relation between more fundamental and less fundamental entities.¹¹ In Goff’s words, the grounding relation ‘is a non-causal explanatory relation that holds between facts.’¹² If x grounds y, facts about y obtain in virtue of facts about x. To use Goff’s example, a party exists because there is a group of people revelling, and thus facts about the party obtain in virtue of facts about the group of people revelling.¹³ Thus, to say that macro-subjects are grounded in the cosmic subject is to say that facts about macro-subjects obtain in virtue of facts about the cosmic subject, and their experiences would be grounded in its own.

There are serious difficulties that come with formulating a constitutive model of agentive cosmopsychism. If the grounding of other subjects in the cosmic subject entails that the cosmic subject has transparent access to their experiences such that it experiences them as they do, as it does on certain prominent variants of cosmopsychism in contemporary discussions, the perspective of the cosmic subject would contain those of all other subjects.¹⁴ Miri Albahari has questioned how such an understanding can deal with macro-subjects with contradictory beliefs (one that intensely fears spiders and another that loves them, for instance). She deems incoherent the idea that the cosmic subject can harbour all these contradictory beliefs and identities within its first-person perspective.¹⁵ The cosmic subject would lack a coherent perspective and would have countless conflicting beliefs and desires, and we would end up with a cosmic subject that is not an intelligent agent but is messy and non-rational.¹⁶

A way of avoiding this issue altogether is to adopt the alternative, non-constitutive cosmopsychism, in which there is a non-constitutive relation connecting other subjects to the cosmic subject. A non-constitutive relation ‘allows for new properties to arise within the combined organic subject that are not present in the fundamental subject.’¹⁷ Other subjects can be said to derive their being from the cosmic subject, but their mental states are not grounded in those of the cosmic subject.¹⁸

This variant enables the development of an agentive model because macro-subjects are not grounded in the cosmic subject, and the cosmic subject need not experience all the emotions, thoughts and so on of other subjects as its own.¹⁹ In this model, other conscious subjects still derive their being from the cosmic subject, but since they are not grounded in it they do not make up its perspective.²⁰ Since the perspective of the cosmic subject would not contain those of other subjects, we do not have to conceptualise it as messy and unintelligent. A non-constitutive understanding is thus suitable to use when formulating an agentive cosmopsychist position.²¹ To be clear, non-constitutive models are not necessarily agentive, but they more easily enable the formulation of an agentive model.

What is the appeal of agentive cosmopsychism when contrasted with classical theism? It seems to be able to explain features of the universe like fine-tuning, things that classical theism is often invoked to explain, without suffering from some of classical theism’s key drawbacks. I shall further explain this here.

To start, we require an understanding of fine-tuning. It has been noted that the existence of life is extremely improbable. For life to be possible, there are certain parameters within the laws of physics and initial conditions of the universe that need to fall within an extremely narrow range.²² The universe is fine-tuned for life, and many scholars see this as a fact that requires explanation. We cannot simply view it as a fortunate feature of our universe.²³

God is one possible explanation of fine-tuning.²⁴ God would be powerful and knowledgeable enough to create a universe that would give rise to life, and would have the benevolent disposition required to compel him to do so. Fine-tuning might thus be construed as evidence for God’s existence. However, there are other elements of our universe that challenge belief in God. The problem of evil is the most significant. There are several different variants of this problem. One is the logical problem of evil, which runs as follows:

• 1. God is omnipotent, omniscient, and wholly good.

• 2. Because evil is in opposition to good, a wholly good God would eliminate evil as far as it can.

• 3. There are no limits to what an omnipotent being can do, so God would be able to completely eliminate evil.

• 4. An omniscient being would be aware of the existence of evil.

• 5. Evil exists.

• 6. Therefore, God does not exist.²⁵

The essential claim of this argument is that the existence of evil is inconsistent with that of God. This is not the only variant of the problem, however. An alternative formulation of this issue is the evidential problem of evil. This argument runs as follows:

• 1. If God exists, there are no instances of gratuitous evil (those not necessary to secure some compensating good or prevent a worse evil).

• 2. There are instances of gratuitous evil.

• 3. Therefore, God does not exist.²⁶

This problem is not simply pointing to the inconsistency of evil with God’s existence. It is saying that evil that is not ultimately for the greater good provides us with evidence against God’s existence. The claim, therefore, is not as strong as that made in the logical problem of evil, but it is still a difficult obstacle for the classical theist.

There are two types of evil that contribute to the claims made in problems of evil. There is moral evil, the evil brought about intentionally by moral agents like human beings.²⁷ There is also natural evil, which is evil that is not brought about intentionally by moral agents but is instead caused by natural processes (famine, earthquakes, etc.).²⁸ This is important to keep in mind.

It is thought that the cosmic subject in agentive models can account for fine-tuning whilst avoiding the problem of evil. The cosmic subject also seems like a suitable candidate for explanations of fine-tuning, as it is also an intelligent agent that one might suppose is capable of establishing the relevant parameters within the range required to give rise to life.²⁹ Like God, this being has the intelligence and ability to establish a universe in which life can exist. It can thus be viewed as a rival to classical theism in explaining fine-tuning.

Furthermore, the cosmopsychist has a luxury that is often considered to be unavailable to the theist in responses to the problem of evil.³⁰ They are able to reject at least one of the key divine attributes that motivate the problem.³¹ Whilst this is not the only strategy available to cosmopsychists when responding to the problem of evil,³² it has been used in attempts to produce cosmopsychist models capable of avoiding the problem, and thus warrants discussion. I shall consign my focus to this strategy in this paper. I shall discuss each of the divine attributes in turn and, whilst engaging with the proposals made by these scholars, shall highlight which divine attribute should be the key focal point in a cosmopsychist response to the problem of evil and how one should use it.

Omniscience

The first attribute to be considered is omniscience. This plays a key role in arguments from evil because it means that God cannot be unaware of the existence of evil. There are two routes that the cosmopsychist might take in formulating a response to the problem of evil using this attribute. They could simply deny this attribute and use some lack of knowledge to explain the existence of evil. Or they could use certain forms of knowledge possessed by the cosmic subject in response to the argument. Freya Mathews has taken the latter route in her response to the issue in her model of cosmopsychism, and I shall consider this option first.

In her model, Mathews uses the extensive knowledge of the cosmic subject as a way of addressing the problem of evil. She provides a constitutive account, in which the benevolent cosmic subject produces other subjects. According to Mathews, the cosmic subject constitutes other subjects, and they are identical with it. Due to this, the cosmic subject experiences everything they experience, and thus suffers everything they suffer. Mathews takes this to be a satisfactory solution to the problem of suffering, as the cosmic subject suffers with other subjects.³³

The transparent access the cosmic subject has to the experiences of other subjects in Mathews’s model means that it faces the problems raised by Albahari of how the cosmic subject can harbour the perspectives of all subjects, many of which would be contradictory, within its own. This raises two issues for Mathews’s model. Firstly, there is the question already raised of how the cosmic subject could actually have a coherent first-person perspective when all of these conflicting contents are forming it. This is a problem in our case, as it prevents us from deeming this cosmic subject an intelligent agent.

Secondly, it would impact the way the cosmic subject experiences the perspectives of other subjects. Imagine some extreme instances of suffering, such as a parent’s grief after losing a child or someone undergoing torture. In such instances, the suffering is overwhelming. It dominates their experience. The cosmic subject, due to the sheer multitude of perspectives it would be harbouring, would not experience these at this intensity. They are experienced alongside the experiences of many other subjects, some of which are pleasant. They do not consume its perspective, like they would with the sufferers themselves. Thus, there are elements of the suffering, such as its intensity and all-consuming nature, that would not be captured in the experience of the cosmic subject.³⁴ Therefore, the extent to which it truly suffers with them is limited, as the intensity and all-consuming nature of such instances of suffering significantly contribute to their severity.³⁵

It thus seems that Mathews’s proposal has difficulties when applied to an agentive model. The transparency issue means that we cannot retain the agentive status of the cosmic subject. Furthermore, it is not clear that the solution fully resolves the problem of evil. The intensity of the suffering of a subject would not be fully captured in the experience of the cosmic subject because it would have the experiences of other subjects to counterbalance it.

What about the other route, to actively deny omniscience? A cosmopsychist model has been formulated elsewhere that will form a useful basis for this discussion. In an attempt to remedy the issues raised above of how a perspectival cosmic subject can generate all the different subjects and contain their perspectives within itself, Itay Shani has argued for an account in which the experiences of these subjects are contained within but largely concealed from the cosmic subject. Other subjects are contained in the cosmic subject’s field of consciousness, but they are ‘proverbial blind spots’ within it,³⁶ and are thus hidden from sight.³⁷

This might be of use to scholars seeking a response to Albahari’s problem, but does it work as a response to the problem of evil?³⁸ There are grounds for thinking that it might. If the cosmic subject is unaware of the experiences of other subjects, it would be unaware of their experiences of suffering. It would also be unaware of the mental contents within subjects, including evil ones such as desires to kill or torture. Thus, there are many elements of evil that the cosmic subject could plead ignorance of, perhaps explaining why it does not prevent evil.

On the other hand, evil is not only to be witnessed in the mental lives of subjects. It is also to be witnessed in its outward expression. Indeed, subjects typically only detect evil within another once it comes to the surface. Our lack of direct knowledge of the murderer’s experiences would not prevent us from seeing the killing they perform as evil. Thus, similarly, the cosmic subject’s concealment of the experiences of other subjects should not prevent it from detecting evil in their actions or in the expressions of suffering made by victims.

In response, it could be claimed that not only does the cosmic subject conceal the experiences of other subjects, but it is also unaware of their existence. However, this response is unsatisfactory as it undermines the cosmopsychist’s ability to explain things like fine-tuning and the existence of other subjects. The generation of subjects, and the fittingness of the universe for this purpose, would seem like something of an accident if the cosmic subject was completely unaware of the existence of what it is generating. The supposition of life arising accidentally is exactly what is denied by proponents of the idea that fine-tuning needs explaining. Cosmopsychism would lose much of its explanatory potency in this debate if this explanation is used, making this response unattractive.

Both solutions considered in this section do not assist the cosmopsychist in this debate. The fact that the transparency problem impacts Mathews’s model means that it has issues with providing an agentive account, and its explanation for evil and suffering has difficulties. Moreover, the adaption of Shani’s model does not fully account for evil either, as the cosmopsychist cannot suggest that the cosmic subject is unaware of suffering or evil without diminishing their ability to explain fine-tuning.

Omnipotence

The next attribute for the cosmopsychist to consider is omnipotence. It could be argued that the denial of this attribute can provide a response to the problem of evil for the agentive cosmopsychist, as it might entail that it is not within the cosmic subject’s power to prevent evil. There are, however, different ways of formulating a model of cosmopsychism in which the cosmic subject is not omnipotent. I shall consider several here.

The denial of this attribute is key in Philip Goff’s model of agentive cosmopsychism. For Goff, the universe is an intelligent agent with the capacity to recognise and respond to facts about value. Since life is of great value, it would ensure that the universe was fine-tuned for it. Moreover, the omnipotence of the universe is denied. Its limitations are recorded by the laws of physics. All that happens is determined by the rational choices of the universe, but it can only do what is consistent with the laws of physics.³⁹ It is also said to have a disposition to form spontaneous mental representations of the complete future consequences of the choices available to it.⁴⁰

Goff argues that this model is able to explain fine-tuning. He claims that during the Planck epoch (the first period of cosmological history), there were laws that determined the form of the laws and initial conditions that obtained after this period, but the cosmic subject determined the values of the parameters within them so that they gave rise to intelligent life. In claiming that this universe is not omnipotent due to the constraints imposed by the laws of physics, Goff claims to avoid the problem of evil.⁴¹

When assessing Goff’s model, it is worth bearing in mind that the laws of physics themselves are fine-tuned. Therefore, in determining the relevant parameters, the cosmic subject is also setting its own constraints and would have knowledge of the future consequences of this. Due to this, it is not entirely clear that his model is completely satisfactory with regards to its explanation of evil. Although Goff’s cosmic subject is not omnipotent and thus avoids the logical problem of evil raised against theism, one might question its ability to escape the problem of evil altogether.⁴²

Given its supremely rational nature and the impact evil has in decreasing the value of the world, the cosmic subject would be compelled to prevent as much evil as it can. This, plus the fact that the cosmic subject has fine-tuned the very laws that it is constrained by and has knowledge of the future consequences of its actions, means that there should be no gratuitous evil. The cosmic subject will know what the relevant parameters need to be set at to allow for the flourishing of life with the least amount of evil possible.⁴³

If Goff were to argue that this is the case, he must make several assumptions. Firstly, the cosmic subject must have chosen values for the relevant parameters that only permit just as much evil as is necessary for life to arise and perhaps flourish. Secondly, since it fine-tunes the very laws that it is constrained by, it must have set those constraints such that it can actively prevent as much evil itself as is compatible with the existence of life. Thirdly, it must do all in its power to actively prevent evil. If it fails on any of these counts, it permits more evil than is consistent with its nature.

Any evil of this type would not only constitute evidence against the existence of this cosmic subject; it would disprove it, because the allowance of gratuitous suffering and evil is inconsistent with its nature as it decreases the value of the world. Therefore, Goff must assume that this is the best possible world that the cosmic subject could have designed that enables life to arise. The cosmic subject must have created a universe in which there is a balance between the amount of life generated and the amount of evil that makes this universe one of optimal value. This assumption is akin to the theses defended by scholars advocating best possible world theodicies, and suggests that the problem of evil has not really gone away for Goff.⁴⁴ It is still possible for evil to pose a problem for his model if one can argue that some is gratuitous.

Moreover, the assumption that there is no gratuitous evil in this cosmopsychist universe seems controversial.⁴⁵ A scenario in which the cosmic subject could have at least lessened the amount of evil in the universe, perhaps by setting the parameters at a different value, certainly seems conceivable. Moreover, despite its limitations, there still seems to be much that the cosmic subject could still do. Could it not, for example, have eliminated Hitler and other key figures of the Nazi party through some disease to prevent the horrors of the Holocaust? Scenarios such as this seem conceivable. There are thus reasons to believe that it is capable of preventing more evil than it does, raising questions about its existence. The problem of evil is still a significant challenge for Goff’s model.

To his credit, Goff acknowledges the possibility of this kind of argument, and suggests further qualifying the powers of the universe in response.⁴⁶ However, in formulating such a response one must provide an explanation of how the powers of the universe can be further restricted in a way that does not compromise the ability of this model to explain fine-tuning. This would be a complex and challenging endeavour, suggesting that it is worth exploring other options to see whether they allow us to evade the problem more easily.

There is an alternative route one might take when denying the omnipotence of the cosmic subject to explain the existence of evil. One might claim that the cosmic subject does not have power over the minds of other subjects. Although it might be able to influence them in its interactions with them, it cannot control them. Therefore, it lacks the ability to prevent other subjects from willing, then performing, moral evil.⁴⁷

This accounts for moral evil, but it does not explain natural evil. Furthermore, because the external world is the cosmic subject, it would be the entity carrying out these natural processes, and would thus be the cause of natural evil.⁴⁸ How can we accommodate for natural evil in such a model?

When subjects witness evil, they often express some kind of reaction. There will be a look of disgust or terror etched on their faces, or perhaps they will shout and scream out of fear or anger. They may even express their condemnation through words, gestures, or physical violence.

In this model, natural evil could be conceived as the cosmic subject’s way of doing this. It is its way of expressing condemnation for the moral evil performed by other subjects like human beings. When witnessing other subjects perform evil actions, one might think that the cosmic subject reacts in a similar way to us. It experiences emotions like anger and disgust, and outwardly expresses these emotions and condemnation of instances of moral evil through things like tsunamis, storms, and other natural phenomena that can have devastating effects. It is unable to prevent moral evil, so is consigned to combatting it in this way.

There is some intuitive support for this claim. It accords with the way that we, as subjects, react to things like evil. Moreover, humans often anthropomorphise natural phenomena. We can see anger in the roar of thunder, calm in the stillness of the seas, and even happiness in the shining of the sun. In cosmopsychism, one can interpret such claims literally, saying that this is actually the way the cosmic subject communicates its emotions.

Despite this, there are difficulties with taking such an approach. It seems that if the cosmic subject were using natural evil as a response to moral evil, much of the suffering inflicted by natural evil is gratuitous in the sense that it appears unnecessary to fulfil this purpose. Natural evil seems entirely random, and can impact those who are morally innocent or who lack the capacity for moral wrongdoing, such as children and certain species of animals.

It would seem more prudent for the cosmic subject to only inflict such evils on those actually performing moral evil. If murders were instantly punished by a bolt of lightning, or stealing a bout of illness, not only would this ensure that only those causing harm were punished, but it would also allow intelligent subjects to infer that natural evil is a response to moral evil, and would thus act as a prevention for future evils. The cosmic subject in this understanding would certainly be capable of operating in this way, and the fact that it seemingly does not is of serious concern to this model.⁴⁹

Both of the options for rejecting omnipotence considered here have difficulties. It is unclear that Goff’s model actually avoids the problem of evil, and appealing to the lack of control that the cosmic subject has over other subjects does not satisfactorily explain natural evil. These explanations are thus insufficient.

Omnibenevolence

The final option one has is to deny the omnibenevolence, or all-goodness, of the cosmic subject. Omnibenevolence is a crucial component of the problem of evil because a being with this feature surely would not be content with the existence of evil. The denial of this attribute thus opens doors for the cosmopsychist in finding ways around the problem.

The denial of the all-goodness of the cosmic subject has some explanatory advantages. If the explanation for the existence of evil is not to be found in the cosmic subject, where is it to be found? The non-constitutive relation operative in non-constitutive models grants one licence to claim that certain properties emerge in other subjects, and perhaps those resulting in moral evil are some of these. However, this explanation of evil would be somewhat mysterious, and it would only account for moral evil, and not natural evil, and might thus result in similar issues to those raised against the second option discussed in the omnipotence section. It seems simpler to think that evil is already there, within the cosmic subject itself. This would mean that evil is neither inconsistent with nor constitutes evidence against the existence of the cosmic subject. The denial of omnibenevolence entails that this model can avoid the problem of evil altogether.

To be clear, we do not need to view the cosmic subject as omnimalevolent, or all evil. It need not be deemed a kind of cosmic demon, and indeed, such a view would come with its own challenges, such as explaining the existence of goodness.⁵⁰ The formulation of such a view may be possible, but it does not seem necessary in this case. We can hold that the cosmic subject has properties one would consider good as well as those that give rise to evil.

One might be hesitant about denying the omnibenevolence of the cosmic subject. In some models of cosmopsychism, benevolence plays an essential role in the explanation of the cosmic subject’s generation of other subjects. For example, in the thought of Plotinus other beings are said to emanate, or overflow, from the benevolent one.⁵¹

The idea that other subjects are produced as a result of the benevolence of the cosmic subject has found contemporary expression in the work of Freya Mathews. Mathews characterises the cosmic subject as plenitudinous, claiming that since it had already exhausted its own existence, it then created other subjects out of itself through internal self-differentiation as a result of its generosity and desire to bring them into being.⁵²

One will note that the benevolence of the cosmic subject forms an essential part of the explanation for the existence of other subjects. Benevolence is a fundamental feature of its being, as it is compelled to generate other subjects from itself. If we deny that the cosmic subject is omnibenevolent, we risk sacrificing this explanation for the existence of other subjects.

However, it seems conceivable that the cosmic subject would still be compelled to generate other subjects despite not being omnibenevolent. Perhaps, as in Goff’s model, it is a rational entity, able to respond to facts about value. One might think that this would be enough to compel an entity to generate life that, although not omnibenevolent, would still have good impulses alongside its evil ones. Due to the evil contents of the cosmic mind, one might query whether it would be sufficiently compelled to adhere to this, but its good contents make a model in which this happens a possibility.⁵³

There are some who might interject here, and claim that the evil impulses of the cosmic subject could hinder one’s ability to use it in explanations of fine-tuning. As Swinburne points out, a being with no non-rational influences will not perform an action that they judge to be worse than refraining from it. They would not suffer from weaknesses of the will, and would thus not act against their better judgement.⁵⁴

Goff himself follows this line of reasoning to claim that the cosmic subject in his model, one with a flawless capacity to respond to facts about value, would be simpler than any with some kind of flaw. Moreover, Goff also claims that his model, in which the cosmic subject is limited in its power, is simpler than a cosmic subject that has irrational desires or a flawed capacity to recognise and respond to reasons. Explaining this, and how it results in the universe as we find it, is seen to be extremely complicated by Goff.⁵⁵

I am not convinced that this provides one with good reason to accept Goff’s model over one denying omnibenevolence. Goff’s claim that his model is simpler only bears weight if its explanatory merits are equal. As has been argued in the omnipotence section, it is not clear that this is the case. The problem of evil still seems to be a significant issue for Goff’s model, and when denying omnibenevolence instead, evil is no problem at all. Furthermore, reasons can still be provided for the cosmic subject’s decision to create life when denying omnibenevolence. Therefore, this model seems able to predict the existence of the universe as we know it without having the difficulties explaining evil that Goff has.

With regards to the challenge of explaining the cosmic subject’s flaws, it may be the case that Goff does not have to provide reasons for these flawed capacities or irrational desires, but it seems that he only moves, and does not actually avoid, this issue. This is the problem of explaining the existence of the flaws within subjects that result in moral evil and, since moral evil clearly exists, one must provide an explanation somewhere.

To be clear, both the cosmopsychist model being proposed here and Goff’s model are capable of avoiding this problem. This model, in suggesting that evil, or at least the flaws that result in evil, exist within the cosmic subject, has no problem with explaining how evil can arise in the universe because the universe is not wholly good. Goff, on the other hand, can claim that moral evil is a result of the cosmic subject’s limitations. Since it is not all powerful, the cosmic subject may have been unable to generate flawless macro-subjects, resulting in moral evils.

The difficulty of explaining the flaws that result in moral evil, namely irrational desires or flawed capacities for recognising facts about value, still arise for Goff. He has to explain how the limitations of the cosmic subject result in such flaws at the level of macro-subjects, which is a different challenge to that faced by the cosmopsychist rejecting omnibenevolence, who has an explanation for these flaws in macro-subjects due to the presence of the flaws in, and macro-subjects’ generation from, the cosmic subject. It is not clear that Goff’s model is any simpler in this regard. His explanation just needs to be provided at a different level: that of the macro-subjects themselves. In fact, since Goff has to explain how such flaws arise in other subjects when they are not present in the cosmic subject, one may argue that this model is actually simpler than Goff’s.

These considerations give us reason to prefer cosmopsychist models that reject the omnibenevolence of the cosmic subject. The problem of evil offers no threat to the existence of such a cosmic subject, and we can still provide reasons for fine-tuning. Moreover, it seems to have advantages over the alternatives considered here, making it the most satisfactory solution.⁵⁶

Conclusion

This discussion has shown that one can produce a model of agentive cosmopsychism capable of explaining fine-tuning whilst avoiding the problem of evil, and that the most satisfactory way to do this is to deny the omnibenevolence of the cosmic subject. This does not mean that one cannot deny omniscience and omnipotence as well, but only that the crucial step in the explanation is the denial of omnibenevolence.

I started this paper by outlining agentive cosmopsychism, and why one might think it has an advantage over classical theism in its ability to explain fine-tuning whilst avoiding the problem of evil. I then considered different options for its avoidance of the problem of evil that focus on either omniscience, omnipotence, or omnibenevolence, before concluding that the rejection of omnibenevolence was the most sufficient of all those considered here.

To be clear, this has not been an exhaustive survey, and different agentive cosmopsychist models might be produced in the future that vindicate explanations involving the other two attributes. Nonetheless, this paper shows that the agentive cosmopsychist has a plausible way forward in this debate with the rejection of omnibenevolence.