The problem of evil threatens theists because it draws attention to the difference between expectation and experience, between the hope embedded in faith and the reality in which we find ourselves. It confronts theism with evidence that in many ways what faith in God leads one to expect is not what one finds to be the case, in particular because what happens seems to be at odds with the faith that God is an all-aware and wholly caring agent who seeks to ensure the best for people and who has the power to put a stop to whatever might prevent it.

Three great merits of James A. Keller's study of the Problems of Evil and the Power of God are that he draws out the extent of the problem of evil, identifying a number of separate problems that constitute it, that he shows the unsatisfactory nature of many attempts to solve these problems that have been, and still are, put forward by those who wish to defend the credibility of theism as traditionally understood, and that he does this in a succinct and clear manner. In these respects his book is to be commended. While his exposition of the problems and his critical analysis of the traditional solutions may not present significant new insights (and even his own ‘solution’ to the problems, to which I will turn later, applies ideas developed in process thought), what he presents offers a useful introduction to the issues that will be of great help to people seeking to grasp what is involved in tackling the problem of evil and to reach a critically informed assessment of responses to it. It is a pity, therefore, that the cost of this present hardback edition is likely to put off students who could benefit from acquiring it.

What, then, are the problems of evil that Keller identifies and to which he finds proffered solutions unconvincing? First there is what Keller describes as ‘the traditional problem of evil: suffering and evil actions’. This ‘problem’ is generally, and justifiably, considered to be two problems. First there is the suffering that results from natural states and events. If God be an all-powerful and benevolent creator and providence, why does the natural world so often seem to contradict these aspects of the divine reality? On the one hand, natural disasters, epidemics, viruses that debilitate and destroy, creatures that can only breed and survive by killing others, and mutations that destroy the beings in which they occur, are examples of natural processes that devastate human lives and seem dysteleological. Although Keller does not spend a great deal of time on this issue, he asserts the implausibility of attempts to explain such natural evils by suggesting that they are due to some evil supernatural being or beings, or are expressions of divine discipline. On the other hand, there is the suffering that is the product of human ill will. Some people are victims of what others do. The most common response to this moral problem of evil defends the goodwill and power of God by asserting that God's activity is directed by the good of fostering and protecting human free will, and that this requires God to respect each person's freedom and responsibility even if what that person does harms other people. Those who suffer thus seem to be victims, subordinated to the good of others' use of their divinely respected freedom. Keller briefly considers attempts by William Hasker, Eleonore Stump, Thomas F. Tracy, Richard Swinburne, and John Hick to meet the natural and moral problems of evil – but in no case finds them persuasive.

The next form of the problem of evil identified and discussed by Keller is that of ‘divine hiddenness’. As he comments, ‘most people, including most theists, agree that even if God exists, it is not clear that God exists, what God's nature is, and what God's will is regarding human actions and attitudes’ (31). Although some people may claim that there can be no doubt about what they apprehend to be what God wants, most people (especially, one might add, those brought up in the critical, post-Enlightenment culture) find that such certainty is far from obviously justified. This seems at least odd if knowledge of the divine reality and will is important for the human flourishing that God is supposed to will, not least when disagreements about God may provoke vicious conflicts. It also means, although Keller does not take up the point, that the answer to the problem of evil presented at the end of the Book of Job is unconvincing.

This answer works for Job because God manifests the divine to him, thereby compelling him to recognize that he has no competence to judge in such matters. He must stop questioning because he is raising matters that are beyond his understanding. This answer does not work, however, for those who raise the problem of evil because the fact that they raise it shows that they have not (yet) enjoyed a convincing epiphany of the divine and are not persuaded by others' claims to have had some such experience. Keller examines attempts that have been made to explain the hiddenness of God by reference, for example, to divine transcendence, to human defectiveness and the dangers of knowledge, and to the need to make room for the voluntary character of authentic faith as commitment to what is not rationally certain. He judges that none of these explanations are adequate. When, therefore, in spite of their experience of the hiddenness of God, theists assume, as they typically do, that ‘God cares whether or not people have faith in God and love God’ (49), they are holding that God wants people to respond appropriately to the divine at the same time as they are faced by the experience that apparently God does not employ divine power to make it reasonably credible for them to do so. Their assumption is also troubled by the disturbing fact that there seems to be no divine action to prevent people making claims about what God desires, including instances when what is so claimed manifestly contradicts what others, with equal vehemence, assert to be the divine will. The perplexed find that they do not benefit from any decisive divine guidance to help them to perceive whether either or neither of the contradictory claims be correct.

A further form of the problem of evil is posed by claims about miracles. This is particularly germane to those theists who seek to warrant their faith by reference to various events that are held to have come about, often unexpectedly, because of the power and benevolence of God. As Keller indicates, these events may have ‘epistemic’ character in the sense that they refer to events in which God is claimed, particularly in relation to supposed ‘revelatory periods’ (61), to have brought knowledge of the divine reality and will to certain people, or they may be ‘practical miracles’ in the sense that they identify certain, often dramatic, events as being brought about by divine agency to help people, such agency being always unpredictable and frequently seemingly arbitrary in its activity. In both cases, however, theistic faith faces the problem that if God has acted in these cases, it seems deeply puzzling as to why the divine benevolence and power has not been utilized in other cases, apparently equally, if not more, deserving. For God to help some but not others appears to be unjust.

For Keller, attempts to hold that God does perform miracles while absolving God of this charge are unconvincing. There is, furthermore, the problem of identifying which, if any, events are divinely produced miracles. Such events do not come with an obvious manufacturer's label stating ‘Made by God’. To hold that we know that miracles occur because this has been revealed is not a satisfactory reply, since some prior miracle is then needed to warrant claims that there has been such a divine revelation. In the end, theists are faced with a fundamental dilemma. In view of the widespread ignorance and disagreement about the nature and will of God, and in view of the widespread suffering found in the natural and historical realms, the more theists claim that in some instances God has acted benevolently to ameliorate conditions, the more the credibility of their faith is threatened by the moral problem of why a benevolent and powerful God, who can act in such ways as they claim to perceive in cases of miracles, has not acted equally effectively in the vast number of cases manifestly not so benefiting from divine activity.

In developing his analysis, Keller now turns to various other attempts to explain the discrepancy between what faith leads one to expect and what is found to happen. These explanations assert that human beings are unable to understand, and should not expect to understand, ‘why God permits (or perhaps even in some cases causes) many of the apparent evils that occur’. In considering what he dubs ‘the ignorance defense’ of theism (73), he examines various forms of this response, including ones presented by William Rowe, Stephen J. Wykstra, and William P. Alston. Again he is not convinced that a convincing response to the problems of evil can be found here. The basic problem with the ignorance defence is that it attempts to show that ‘certain arguments against traditional theism do not succeed because the arguments presume that humans know more than humans do’. As Keller points out, the more that these supposed defenders of theism stress human ignorance in such matters, the more they undermine ‘their own positive claims about God’ (89). People cannot claim to be confident about some things about God and then, when questions arise in connection with those things, claim that they are ignorant in such matters. To meet questions about theism by claiming ignorance is liable to result in questions being asked about the possibility of any and all theistic assertions.

Before turning to developing his own solution to the problems of evil, Keller points out that these problems arise because traditionally theism has assumed that God is to be regarded, among other things, as omnipotent. For some it is on the basis of revelation that this quality must be ascribed to God. In a chapter entitled ‘Identifying, interpreting, and certifying revelation’, Keller questions the convincingness of the supposed revelatory justification of this traditional theistic assumption, considering in the course of his examination views put forward by George Mavrodes, Richard Swinburne, and Alvin Plantinga. Christian theists typically maintain that the Bible is ‘a vehicle for conveying particular revealed truths’ (109), whether or not they also hold that it is authoritatively augmented by the teaching(s) enshrined in Christian tradition. Unfortunately those who claim that in the Bible (and perhaps also in tradition) the faithful are provided with divinely ‘revealed truths’ and thereby seek justify their view of God, including their understanding of divine power, have the problem of explaining why, if this be the case, ‘these alleged revealed truths are not stated clearly and clearly divinely certified’ (109).

While numbers of people claim to be guided by the truths revealed in the Bible, in practice it is found that on the basis of scripture (and also of tradition) there is ‘a variety of positions that may legitimately be called Christian, for each of them may be a legitimate response today to God as manifested and understood in the Christian tradition(s)’ (111) – and some of these responses are expressed in radically inconsistent assertions. On reflection, there seems to be an incongruity in holding that an omnipotent God has given authoritative teaching to guide people, not least in guiding them to see the divine as omnipotent, when there is no clear way of showing where this authoritative teaching is located (a problem, one might add, that is made far more difficult when the sacred scriptures and traditions of different faiths are taken into account), or, if the location be agreed upon, of how the contents are to be interpreted and applied. As Keller sums it up,

Are there, however, grounds other than alleged revelatory ones for holding that the traditional understanding of God as omnipotent is justified? Keller holds that religious experience fails to warrant that understanding. Nor in his judgement is the argument convincing that maintains that omnipotence is entailed by the intrinsic perfection of the divine. While it may be agreed that God must be considered as ‘that being than which no better can be conceived’, this does not provide the basis for ‘an agreed-upon or non-controversial analysis of what the power attributed to God should include’ (120). What people attribute to God because of God's perfection depends upon the metaphysical understanding and axiology that they bring to their understanding of that perfection. Furthermore, Keller argues that considerations of the role of God as creator (in particular as bringing about creation out of nothing), claims about miraculous events in the world that thereby comes to exist, and assertions that God's power will ensure that evil will finally be overcome do not provide strong grounds for holding that God is omnipotent in the sense of having the power ‘to unilaterally bring about any logically possible creaturely state of affairs’ (121).

In view of what Keller is to present as his solution to the problems of evil, it is perhaps surprising that, while Hartshorne's Omnipotence and Other Theological Mistakes is listed in the bibliography, he does not discuss the views on divine power put forward in this and other works by Hartshorne. Nor does he take up the question of whether, since power is essentially relational, the notion of God as omnipotent in the sense of having absolutely all power (omni-potent) is inherently incoherent. One being may have power, even unimaginably superior power, over another, but that power has power over another only insofar as that other still has some power in itself to be. A being with absolutely all power would have nothing to have power over: it would be all that is.

At this point, it is tempting to conclude from Keller's study of the problems of evil and the unsatisfactory nature of attempts to solve them that theism is not rationally credible. Whatever arguments may be put forward to warrant theistic faith, they use evidence that is selective and partial. It is as if from a host of observations plotted on a graph, a researcher chooses a limited selection of them that conform to a (desired-for) pattern and then tries to find grounds for holding that the other ones can be ignored. This is a questionable and unconvincing method of reasoning in any subject, not just in science. There may, of course, be pressures upon the researcher to support a particular conclusion and it may be the case that novel ideas are faced with many apparent problems when they are first put forward, but in the long term the success of a way of understanding the evidence is that it comes to be seen as a more convincing way of interpreting all the evidence that is known than its predecessors and rivals.

In the case of theism, the problems of evil identified by Keller may well have accompanied it from the beginning and it seems clear that it has not yet found satisfactory ways of solving them. Does theism survive in spite of the evidence because people want it rather than because it is rationally supportable? One response may be to argue that theism is not to be justified by arguments from what is found in the world but by a priori considerations. The problem with this is that such reasoning is likely to be judged clever but not convincing if its conclusions fail to fit in with what is found to be the case in the world. Another response may be to hold that theistic faith is faith in spite of what appears to be the case – a religious form of desire denying reality. It may make people feel better but it is not the assent appropriate to a person seeking to follow the Lockean principle of rationality.

Keller, however, does not draw any such conclusion. He is clearly intent on finding some way both to affirm the reality of God and to explain how that reality is compatible with the problems of evil that threaten to undermine theistic belief. Taking up some of the points made by process theology, his solution is based on the argument that theists should radically rethink what is to be meant by God, in particular by recognizing that ‘divine power should not be understood as omnipotence in the traditional sense’ (135). In a brief introduction to basic views of process thought, he notes that according to this way of understanding ‘the fundamental sort of creatures are momentary events’ (136). While these events are minute spatially and in temporal duration, each of them becomes what it is (i.e. as an ‘actual occasion’) by making a ‘decision’ in which it responds to the data of its inherited past in relation to the possibilities open to it. God, as perfect, not only ‘includes all the details of the past’ and envisages all possibilities but also influences each creaturely occasion as it begins to concresce (i.e. as it decides what it is to be) by presenting to it ‘a resistible inclination (in Whitehead's term, a lure) for the occasion to develop itself in a particular way’ that God's evaluation of the situation judges to be for the best (138).

Note, however, that the divine influence is resistible. As Keller summarizes the position, within limits imposed by its past,

While, therefore, God is held to ‘set the laws of nature of each cosmic epoch’ and to ‘sustain the order of nature within it’ (140–141), within each cosmic epoch God can only bring about changes by luring the entities that successively compose it to move in certain directions. And this luring is resistible. Although each concrescing entity has only an extremely limited freedom of choice (for the developments that we observe in the natural and historical worlds of human experience are the product of myriads of such choices), it can choose not to conform to what the divine evaluation considers to be for the best overall.

On the basis of this understanding of the processes of reality, Keller claims that the problems of evil that he has identified are compatible with belief in the reality of God. While, for instance, ‘God is constantly striving to make the world as good as it can be’, and while the divine experience shares all the suffering that occurs therein, God can neither coerce developments to occur as divinely desired nor ‘unilaterally filter out some or all of the bad’. The project of creation that God has embarked on means that while some sentient creatures have come to have rich experiences, others have ‘just turned out’ to have had ‘a bad life (short, painful, and so on)’ (142–143). In spite of all the divine luring, the divine response to the latter seems to me to be a case of saying, ‘Sorry, but that is the way that the cookie has crumbled for you. I tried to influence past concrescing occasions to actualize themselves differently but they did not conform to my divine will for the best, and anyway some unfortunately have to suffer for others to evolve and prosper.’

Kelly further maintains that, in accordance with process theism, the problems of divine hiddenness, the obscurity of supposed sources of revelation, and the absence of miraculous interventions are to be solved by pointing out that God could only overcome these problems by unilaterally ‘controlling certain details of non-divine states of affairs’, and that ‘these are things God cannot do’ (143). As for the question of death, Keller rejects notions of post-mortem existence and of future compensation therein. He suggests that just as persons are composed of a vast sequence of momentary actual occasions forming the history of each developing and finite self that changes as it lives, so hopes for immortality should not reside in the persistence as actual of any particular one of these moments but in the recognition that each moment of the actuality of each self is imperishably preserved for ever in the divine memory.

Keller's thesis, then, is that the process concept of God, and particularly its understanding of divine power, avoids the problems of evil that threaten the credibility of traditional theism with its understanding of God as omnipotent. Furthermore, as he makes clear in his final remarks, Keller is confident that God as so understood is properly worthy of worship as the perfect object of faith. Is his confidence warranted? Bearing in mind the ways in which the problems of evil facing traditional theism identify evidence that challenges the credibility of that theism, and that this evidence is extensive and troubling, I wonder if there is any significant and rationally justifiable difference between holding that there is no ultimate divine reality and holding that there is some such reality but, as the evidence makes clear, it is not a reality that has significant control over the course of events.

Since this review is about Keller's thesis, it is not the place to deal with process thought's problematic affirmation (in what is sometimes called a pan-psychic or pan-experiential understanding of the briefly fleeting items of which reality is composed) of the freedom and decision-making character of the fundamental constituents of reality, but it is pertinent to question whether there is positive evidence to warrant claims that, while God can only seek to persuade and while God cannot in principle coercively determine what happens, nevertheless God does have some, however minimal, influence on what occurs. By treating God as a resistible influence on events, Keller argues that the instances of various kinds of evil can be seen to be compatible with the process concept of God. The question that then arises is why it should be held that God is a reality that exercises any positive influence on events at all. May it not be that some events are the unhappy results of ‘choices’ (whatever it be to ‘choose’ at the fundamental level of concrescing actual occasions) and some events are the happy results of ‘choices’, and most events are neither happy nor unhappy results? In other words, the crucial problem facing theism is to justify its basic notion that God is in some way agential, even if that agency is realized as a lure or an attempted persuasion.

Whereas believers use happy situations and events to support comforting assertions about God's activity, and unbelievers cite the problems of evil to challenge such assertions, the problem is to make sense of all the evidence, not just of some of it. Theists stress one selection of the evidence (think, for example, of the numerous instances of the benevolence, power and wisdom of God seen in nature by Ray and Paley and their increasingly sophisticated modern successors); unbelievers stress a different selection (such as the painful and destructive consequences of the evolution of numerous viruses and malignant tumours). But if theistic sense is to be made of the whole, then both aspects must be embraced by it, and it must also be shown that God is a significant agent in what occurs. It is not enough to say that God is always and everywhere seeking to bring about fruitful developments according to the divine valuation of the past and envisagement of the most valuable future outcomes for the present states of affairs if, and this is the big ‘if’, when things do not go that way, what happens is blamed on creaturely failures to conform to the divine lures presented to the concrescing occasions. Why hold that happy outcomes are attributable to divine influence and not to the free choices of creatures uninfluenced by divine lures? Admittedly the process metaphysic is generally understood to consider that it needs the notion of a ‘God’ presenting a range of possibilities to each concrescing occasion as well as having a preferred one that God seeks to persuade that occasion to choose, but it is not clear that the theistic interpretation of this requirement is demanded by the evidence, and is not merely, in a Berkeleyan kind of way, a means to fill up a gap in making sense of that system of understanding.

These issues lead to the basic problem of the so-called problem of evil. This problem arises for people because they presuppose that they know the nature and purpose of divine agency. This is most clearly seen when people raise questions about natural evils. Why, they ask, did God create this virus, that tumour, when they devastate human lives? A colleague once told me that for him the biggest objection to the design argument was human plumbing! Why did God have to make it something so liable to defects? But what does it mean to say that God ‘created’ these things? Awareness of the vast chains of events of a chance and necessity character that emerges from inspection of the evolutionary process, whether leading to a virus or to a Whitehead and a Dawkins, makes it hard to see why and what God is supposed to have contributed to the processes of reality. Similar problems arise when we reflect on the other aspects of the problem of evil.

Before that problem can be coherently posed as a problem, it is essential first to establish what it is rationally justifiable to mean by divine agency in nature and in history, and why it is reasonable to believe that there is such an agency in operation. Only when this has been done would it make sense to ask why the divine agency produced or permitted this or that state (regarded as evil) to come about. But, note, when this has been done, there would then be no problem of evil to solve – for an understanding of the agency of divine reality would have been established that covered what is found to happen in reality and this would mean that there were no problems of evil to challenge it. In other words, when problems of evil are raised, it is a sign that the concept of God that is presupposed is not fit for purpose.

Keller is right to suggest that theists need to go back to the drawing board. Unfortunately, he fails to show why the God who is the ground of theistic faith, whether as understood traditionally or in some other way, is to be thought of as a significant agent in what happens in the world. This, however, is a fundamental problem that needs to be tackled by all those seeking to establish a credible theism, whether they take insights from process thought or prefer some other fundamental way of understanding.