Ward's Inconsistent Triad

The evil-god hypothesis postulates the existence of an omnipotent (all-powerful), omniscient (all-knowing), and omnimalevolent (all-evil) god and stands in stark contrast to the good-god hypothesis, which states that there exists an omnipotent, omniscient, and omnibenevolent (all-good) god. When the evil-god hypothesis is used to challenge a theistic position, it is called an evil-god challenge. Law points out that theists typically consider the evil-god hypothesis absurd, and his evil-god challenge invites the theist to justify their belief that the good-god hypothesis is significantly more likely than the evil-god hypothesis, when it appears that they are broadly similar in reasonableness.

In his article, ‘The Evil-god Challenge – A Response’ (Think 40), Ward aims to quash the evil-god challenge by demonstrating that the existence of an evil god is inconceivable. To do this, he claims that the following three propositions entail an inconsistent triad: (i) God is omnipotent, (ii) God is omniscient, and (iii) God is omnimalevolent. Ward details several arguments to support this claim: (1) the argument from knowledge of suffering, (2) the Platonic argument, (3) the Kantian argument, (4) the mental pathology argument, and (5) the argument from worship. Although I will address each of these arguments, the focus of my attention will be on Ward's first argument, which he defends most thoroughly and which I believe to be the strongest and most interesting of the five. After offering refutations to each argument, I will ultimately conclude that there is no inconsistent triad; an omnipotent, omniscient, evil being is indeed possible. From this point forward, the term ‘god’ denotes an omnipotent, omniscient being whose moral character has not been specified, ‘evil-god’ refers to an omnipotent, omniscient, omnimalevolent being, and ‘good-god’ refers to the omnipotent, omniscient, omnibenevolent deity of classical theism. In the next section I will address Ward's last four arguments, and in the section that follows I will focus on his first argument in detail.

Arguments 2–5

Ward's second argument employs Platonic ideas in an attempt to establish the necessary interconnectedness of god and goodness. Describing the ‘usual Christian view’ of goodness as objective and part of divine nature, Ward contends that god necessarily possesses whatever qualities are intrinsically good. That is to say that there are objective values of goodness, and god embodies them all. If this is the case, then evil-god cannot exist because goodness is intrinsically bound up with the concept of god, and to speak of evil-god is as self-contradictory as speaking of a four-sided triangle. Ward then claims that evil is of what he calls ‘intrinsic disvalue’, which means that no rational being could value it for its own sake. A wholly evil being would necessarily have to embody all disvalues (Ward gives the examples of weakness, ignorance, and depression) and must desire itself not to exist. As an omnipotent being, this god would have the ability to act on its desire and would will itself out of existence.

Ward claims that goodness is of objective, intrinsic value and it is bound up with the concept of good-god. In response, we can make two alternative claims: (i) perhaps there is no such thing as objective, intrinsic goodness, and (ii) there is an evil parody of Ward's claim: evilness is of objective, intrinsic value and it is bound up with the concept of evil-god. Claim (i) assumes an anti-realist approach to morality and contradicts Ward's account of goodness by claiming that there are no objective moral facts. Ward does not present an argument for the objective, intrinsic value of goodness or an argument against its alternatives and, without these, his Platonic argument cannot be successful. Claim (ii) suggests that evil-god is the ultimate form of evil, and so whatever evil-god derives pleasure from (the pain and suffering of its creation) has intrinsic value simply because evil-god desires it. Evil-god could very well be a sadist that enjoys the pain of others and, therefore, the pleasure gained from the pain and suffering of sentient beings would be intrinsically valuable. Although I do not claim that this evil-god parody is sound, it is a parallel that undermines the force of Ward's argument nonetheless.

Let us move on to Ward's Kantian argument. According to this argument, ‘an omnipotent being would choose to act rationally not irrationally if only in order that it could discern efficient means to its goals and understand how to build a universe’. After establishing that god is rational, Ward outlines the Kantian notion that a rational being knows that any principle that applies to a being in a particular situation will apply equally to any being of that type in a similar situation. He uses this supposition to infer that a perfectly rational being would not trigger the undesirable condition of pain in others if that being found pain undesirable itself. Ward concludes from this that a rational god must be a good god.

I have two responses to Ward's Kantian approach. Firstly, god and human are very different types of being. Ward seems to be presuming two claims, (i) that god cares for humanity and (ii) that god views humans as similar to itself, both of which are unfounded. Human preferences could be completely unimportant to god. Besides, even if god is perfectly rational, this does not mean that god desires for all intelligent beings exactly what god desires for itself. For example, it would be absurd to conclude that because god desires to continue to be omnipotent, god would want all intelligent beings to have this quality. God could be a unique being and one that would not impose its own preferences onto those it created. In other words, god does not follow the divine golden rule, that is the rule of treating people as god wishes to be treated. A second strategy to countering the Kantian argument is to deny that it is necessarily rational to do good. By adopting the view that evil-god is an ethical egoist, i.e. a moral agent that should act in a way that promotes its own self-interest, it can be maintained that, for evil-god, rationality is acting in accordance with its own egoistic preferences. As such, evil-god would be inclined to maximize the suffering of sentient beings. It would be a mistake to assume that the attributes of rationality and goodness are inextricably linked, thus evil-god could be rational and evil without contradiction. It could be objected that the intrinsic goodness of an action gives an agent immediate good reason to perform that action, whereas the evilness of an action does not provide an agent with the same motivation. In response, I would argue that because evil-god's desire for evil would be divine, the desire itself serves as an immediate motivating factor for evil actions. Evil actions would be intrinsically valuable simply because evil-god desires them. Besides, even if the objection stands, this does not demonstrate that rationality and evil are necessarily mutually exclusive, only that goodness and rationality may be linked in a way that evilness and rationality are not.

Ward's penultimate argument is an argument from mental pathology and relies on the supposition that an entity that gains pleasure from cruelty is pathological, which I take to mean mentally unhealthy or diseased. According to Ward, one agent would only take pleasure in another's pain when the former is treating the latter as ‘sub-personal’, or a lesser being than itself. Ward explains that the treatment of another as sub-personal is based on false or inadequate belief, which is a flaw that an omniscient being would not have.

If we accept Ward's contention that cruel behaviour manifests as a result of one entity believing itself to be superior to another, then I contend that it is reasonable to allow the possibility of god behaving cruelly. If ever there was an appropriate example of a relationship where one being is superior to another then surely it would be between the creator of the universe and its creation. It is entirely plausible to maintain that god regards humans as sub-personal because, to god, they are! Of course, denying the intuitively appealing anthropocentric view that humans are intrinsically valuable and not just the pleasure-enhancing instruments of an omnipotent, omniscient being will not be a popular stance to hold, but it need only be demonstrated that this state of affairs is possible to overcome Ward's inconsistent triad. It seems reasonable to assume that god could be not only perfectly powerful and completely knowledgeable but superior to humans and, accordingly, misanthropic. If this is the case then, given Ward's assumptions, there is nothing pathological about god treating humans cruelly.

The final argument Ward proposes relies on the phenomenon of worship. People, he argues, seek god as an object of perfection to offer their worship to and an icon to deliver forgiveness and grant opportunities for moral redemption. Ward states that this search ‘entails that their God be powerful and knowledgeable enough to deliver them from evil’. But, he argues, there is no analogous search for an evil-god.

I have two responses to the argument from worship, which seems to be the weakest of Ward's quintet of arguments. The first says that we cannot establish the existence of good-god simply by observing the practices of people. Empirically, Ward is correct to say that people seek a supreme power for moral atonement and salvation, which leads them in turn to religious belief and worship. However, these observed practices establish nothing about the true nature of god, only the true nature of people. The second response claims that the empirical fact that people worship good-god is compatible with the existence of evil-god. It is possible that evil-god, given its nature, would want people to worship a non-existent benevolent deity futilely, because it would gain pleasure from seeing them dedicating their lives to its imaginary antithesis.

Having addressed all but one of Ward's arguments for the inconsistent triad of omnipotence, omniscience, and evil, I now turn to the most forceful of his approaches, which explores the nature of god's omniscience.

Argument 1

Ward's first attempt to demonstrate the impossibility of evil-god aims to establish that an omniscient being would not intentionally cause suffering to others due to a particular type of knowledge god must necessarily possess: knowledge of suffering. Ward's reasoning can be presented as follows: god is omniscient; therefore, god knows everything. If god knows everything, then god has ‘perfect empathy’: the quality of experiencing the exact mental states each individual person feels (including mental states of suffering). As a rational being, god must desire to experience pleasurable mental states and have an aversion to painful mental states (including those resulting from its perfect empathy). As an all-powerful being, god can realize these desires and would therefore cause happiness and not harm. To sum up, as a result of its perfect empathy, god would choose humans to feel pleasure and not pain because it feels the effects of the treatment alongside them. Here lies the apparent inconsistency between omniscience and omnimalevolence and Ward's basis for rejecting the possibility of an evil god.

To argue against this conclusion, I propose four challenges. The first two challenges highlight problems with the quality of perfect empathy that Ward ascribes to god: an identity problem and a moral problem. The third challenge aims to establish that even if we accept god's perfect empathy, this does not entail that god is good. The fourth challenge is a version of the problem of suffering that is particularly problematic for the concept of god that Ward puts forward.

The first problem faced by perfect empathy is an identity problem. To empathize perfectly with a being, x, another being, y, must not merely know exactly what x’s experience feels like, but experience it themselves, which appears impossible for two ontologically distinct beings. It seems that unless god and god's creation are one and the same (as is the case with pantheism), the personal phenomenal experiences, or qualia, felt by entities within creation cannot also be felt in exactly the same way by its creator. In other words, god cannot have phenomenal knowledge of every single experience.

Besides the identity problem, there are moral complications when considering the compatibility of perfect empathy with god's other attributes. If an omnipotent, omniscient being has a relationship of perfect empathy with another being, it must empathize exactly and completely with the other. Let us consider the extent of this claim for a moment. Good-god would have to experience precisely what it feels like to desire to murder someone in cold blood; to desire to rape an innocent child; and to desire to end its own life. Surely that is not possible for a god from whom maximal goodness stems. Similarly, evil-god would have to share in the feelings of altruistic behaviour, love, and compassion, which is against its nature. This view of omniscience is problematic for classical theists and evil-god hypothesizers alike. Furthermore, if god desires to reduce the amount of pain and suffering humans experience solely because it wants to avoid the corresponding pain, rather than because it cares about human welfare, this implies that god is selfish – a conclusion that does not seem to be compatible with the characteristic of omnibenevolence that classical theists ascribe to god. Surely if god is the source of goodness, this goodness should not be performed by god simply for its own benefit, while using human agents as a means to an end.

Of course, philosophers have puzzled over the nature of god's omniscience for centuries, and there is not the scope to firmly establish the nature of this divine attribute here. Suffice it to say, there seem to be difficulties with the concept of perfect empathy that should be investigated before ascribing this quality to god. But let us assume along with Ward that god does possess perfect empathy. My third challenge is intended to demonstrate that even if we allow a perfectly empathetic god, this being could still be evil without contradiction.

Empathy, I argue, entails neither sympathy nor compassion. A moral agent could understand what it feels like to be painfully scourged and not enjoy the feeing itself, but this by no means prevents them from gaining pleasure from whipping their dog, or even their child. In fact, it is conceivable that, in certain cases, empathy may cause the agent enjoying the suffering of another to enjoy it even more. A sadistic torturer may have been tortured themselves and empathize with their victim accordingly, but this experience may serve to heighten their own awareness of the victim's pain and consequently their enjoyment of the torture. Many people experience schadenfreude, the pleasure gained from seeing others suffer or fail, from the satisfaction of seeing an enemy receive their comeuppance to gratification from watching a murderer executed. Similarly, evil-god's enjoyment of suffering could be intensified by its perfect empathy.

It may be objected that the previous examples are not relevant to this debate as they refer to beings with limited, rather than full, empathetic knowledge, and that a perfectly empathetic entity would not be subject to these feelings. In response, it could be proposed that evil-god values its own pleasure above all else. Perhaps evil-god is akin to a divine ethical egoist and cares only about the maximization of its own pleasure, derived exclusively from the pain of others. For a sadist, desire is the suffering of others; for a masochist, self-suffering brings pleasure. Evil-god could be a sadist, enjoying others’ pain; a masochist, enjoying its own pain; or indeed (as this model implies) a sadomasochist, enjoying both the pain of others and its own pain – the ultimate schadenfreude, if you will. Arguably, therefore, god may desire suffering while retaining perfect empathy.

Ward might respond by saying that being a sadist or a masochist is incompatible with omniscience, because subscribing to either sadism or masochism is irrational and omniscience entails rationality. In response, I would argue that in order to show that sadism and masochism are irrational, one must prove a particular type of moral realism – it cannot be derived from logic alone. For moral realists, moral facts actually exist and are independent of an agent's beliefs, and this claim should by no means be taken for granted. Remember, to deny Ward's inconsistent triad we must only show that a being possessing omniscience and omnimalevolence is possible.

My final challenge is a version of the problem of suffering that has undesirable consequences for Ward's model of good-god. If Ward is right to say that good-god must necessarily be both perfectly empathetic and perfectly good, then the fact that we do suffer suggests that good-god cannot exist. Surely good-god would avoid every single instance of human suffering, so as to experience only pleasure as a product of its perfect empathy. This problem of suffering is especially forceful against Ward's concept of god, which includes the attribute of perfect empathy and leads to a potentially insurmountable inconsistent triad.

Overall, it seems that even if we overlook the potential problems with ascribing the attribute of perfect empathy to god, there are still difficulties with the claim that god's perfect empathy necessitates god's goodness. It can be concluded, therefore, that Ward's argument from the knowledge of suffering fails and that no inconsistent triad is established based on god's omniscience and omnimalevolence.