If we ask people who know no philosophy or theology whether they have any view on the problem of evil, we might expect either to be asked what we mean or to encounter an assumption that we’re wondering whether they have ideas about how to reduce the evil in the world. How to do that is of course a problem of evil. In most philosophical and theological contexts, however, the problem of evil is that of whether two kinds of view can be reconciled: that there is an omnicompetent God, in the sense (minimally) of a being who is omniscient, omnipotent, and perfectly good, and that there is a very great deal of evil. The literature on this problem is vast, and this paper does not address it. For Amir Saemi, who is my central concern in this essay,¹ there is what he calls “the new problem of evil,” which “arises from the divinity of [Islamic] Scripture and seemingly prescribed evil” (p. 9). More specifically, (1) Scripture is here taken to be “the words of an omnibenevolent, omniscient, and omnipotent God”; (2) there are actions [such as wife-beating] prescribed or permitted by the best interpretation of a Scriptural passage that seems immoral according to our independent moral judgments; and (3) those judgments “reliably represent moral values, moral duties, and moral permissions” (p. 23). I assume that this poses a problem for conscientious believers in the scriptures in question, and that there is some kind of evil seemingly entailed by the existence of such scriptural passages, whether they are Islamic or belong to any other religion. On that assumption, there is a problem of evil here – at least that of an existentially painful position for certain religious people.

This problem is posed, and pursued in detail and in original ways, by Saemi, but the existence of the problem does not entail, and Saemi apparently does not treat it as posing, a challenge to the very existence of the God of perfect being theologies.² It is, however, certainly a problem for morally conscientious believers in scriptures that, like the Qu’ran, are apparently best interpreted as permitting such wrongdoings as wife-beatings. Saemi devotes a great deal of space to describing how major figures in the history of Islamic ethics might have resources for dealing with this problem – which I propose to call the scriptural problem of evil. If it is more serious for some scriptures than for others – and Saemi notes passages from Leviticus that raise it for certain kinds of readers of the Hebrew Bible – the problem is not limited to any particular religion. This paper will consider only briefly some of the Islamic resources Saemi brings forward that appear to offer possible routes to a solution. Instead, I prefer to explore Saemi’s own outline of a solution.

Quite naturally, Saemi begins with discussion of how thoughtful Muslims might deal with the “new problem of evil,” which arises from a tension among:

The tension referred to arises, as Saemi sees it, in part from a doctrine concerning the authority of testimony: “on the traditional understanding of Hadith, Hadith is part of Scripture,” and central to it is the idea that “the epistemic status of the Prophet’s words would be the same as that of the Qu’ran” itself (7). Moreover, “Hadiths [“the collection of the sayings of the Prophet Mohammad” (6)], if authentic, are the words of an infallible morally perfect Prophet” (7).

There is certainly a tension in the conjunction of these three theses, and the tension could be an instance of the old problem of evil understood as how to reconcile the amount and kind of evil in the world with the existence of an omnicompetent God. To be sure, there is no logical problem – if we assume that what is “seemingly prescribed” need not be actually prescribed or that, however reliable our moral judgments are, it does not follow that they are true. Should we allow these assumptions? They certainly go with the notion of reliability, which, properly understood, concerns the proportion of truths to falsehoods for judgments of a certain kind (say, vision-based) and is normally taken to allow for a high proportion that still falls short of entailing the truth of the judgments.

It should be noted, however, that Saemi says that by “our independent moral judgments” he means “moral judgments obtained through careful a priori moral reflection and sufficient empirical observations, independent of Scripture” (p. 9). This claim clearly does not entail the truth of all “independent” moral judgments reached by such reflection, but presumably he would argue that it does entail a high objective probability of truth. This formulation leaves room, however, for the view (to be considered shortly) that (1) the a priori reflections may yield necessary moral truths, but (2) moral judgments dependent on empirical contingencies are not necessarily true.

One challenge that remains when this interpretation of independent moral judgment is taken into account has been called the evidential problem of evil. This has been addressed in detail by philosophers of religion in the decades since Alvin Plantinga very plausibly argued that there is no logical inconsistency between the existence of evil and that of an omnicompetent God. The problem is serious, especially where natural evils are concerned, and much literature deals with the case that, even where evil includes such natural disasters as innocent people being burned to death in forest fires, it does not follow that the evidence against God’s existence is stronger than the evidence for God’s existence – or anyway strong enough to preclude rational belief that God exists.

The evidential problem of evil is not addressed in Saemi’s book except by implication. What he concentrates on is a religiously internal problem, particularly for religious people (not necessary Muslims) whose scriptures portray God as omnicompetent but also represent, as infallible and in accord with God’s command (or at least God’s will), testimony that has seemingly immoral content.

The problem Saemi poses is serious and is a problem of evil of a philosophically significant kind. Saemi brings out how it is particularly significant for Muslims. We might therefore wonder whether any Islamic philosopher he considers – and he carefully presents many significant Islamic philosophical views – has a solution.

On first reading about the Mu’tazilites, who “developed a rationalistic theology holding that reason is a new source of religious knowledge” (p. 30) and “argued against divine, command theory, holding that there are intrinsic moral properties that can be discovered by reason” (p. 30), I was optimistic about finding a resolution of the Scriptural problem of evil. But this school “died out, mainly due to political factors” (p. 30), though the Mu’tazalites “survived in the form of the… Shia school” (3). It appears to me that there is no solution implicit in the points here attributed to the Mu’tazalites. First, that some moral properties can be discovered by reason does not imply that it is more reasonable to rely on a rational judgment that wife-beating is impermissible than on “religious knowledge” that God permits it – especially, if one can restrict the circumstances of its permissibility, e.g. to avoiding brutality and causing serious injury. Second, for some divine command theories, it is possible to recognize non-theological routes to discovering moral truths. Some of these theories may allow that, in some cases, one might have better evidence of a moral truth on the basis of “independent” moral judgment than for taking a conflicting Scriptural pronouncement to be authentic.

Another possibility for dealing with the problem of evil Saemi brings forward is suggested by some of what he calls attention to in the writings of al-Ghazali, who said, e.g., that

One might wonder whether here al-Ghazali anticipates the need for reflective equilibrium in deriving moral judgments from scripture, but that is not an idea explicitly treated by Saemi. He does say, however, that for al-Ghazali, “there are cases in which we are allowed to understand Scripture nonliterally” (p. 65) and that the reason al-Ghazali allows this is that, for him, “God does not make false claims. If we know that the literal meaning of a statement is impossible [that, taken literally, it is necessarily false], there is no possibility for God to make that statement true” (p. 65).

This seems to me important and to foreshadow Saemi’s own view, though, unfortunately, al-Ghazali apparently does not allow for our knowing the falsehood of some of the Scriptural pronouncements that call for actions that are wrong (pp. 65-66). Saemi tells us that, for al-Ghazali, “Reason has no power to discover the consequences of actions, and thus in [specific] moral matters reason has no power whatsoever to discover the truth” (p. 66). On the face of it, this conception makes no room for our knowing a priori that some acts are at least prima facie wrong (a possibility I’ll later suggest Saemi himself clearly leaves open). (Empirical impossibility would presumably be contingent and thus alterable by an omnicompetent God.)

As represented by Saemi, then, al-Ghazali leaves us with a problem concerning just how non-Scriptural moral knowledge is possible, especially if we are (causal) consequentialists. What we can see here, however, is an opening for some kind of fallibilistic approach to Scriptural interpretation and, where a moral statement is plausibly attributed to God or the Prophet and is not plausibly thought true on our independent judgment, we might wonder whether there is room in other parts of Al-Ghazali’s work, for disagreement about what God actually commands or permits.

One might speculate that, in some moments of his work, al-Ghazali anticipates a kind of reflective equilibrium but fails to give it the scope it should have if one qualifies his consequentialism. ‘Abd al-Jabbar al-Hammadhani (935-1025), by contrast, seems to foreshadow a version of the idea of prima facie duty developed by W. D. Ross.³ Abd alJabbar is quoted as saying:

Saemi interprets this (plausibly, so far as I can tell from the overall account in pp. 7091) with the reading:

This has a striking resemblance to the view of Ross. On his view, knowledge of the wrongness of wrong-making properties, such as promise-breaking, is intuitive and indeed self-evidential (in the sense that the proposition known is not only true but also self-evident); but our (presumptive) knowledge of the wrongness of a particular promise-breaking, which may of course be inferential, is not self-evidential.⁴ The possibility opened up here is that we may view some Scriptural passages as indicating ascription of prima facie duty rather than overall duty and on that basis give priority to what seems a priori knowledge of prima facie duty over textual knowledge of Scripture. This seems indeed the direction Saemi takes, and I’ll return to the matter shortly.

If any Islamic philosopher is a source of support for rejecting unqalified deference to Scriptural testimony, it is apparently al-Razi, whom Saemi describes as a “pessimist about moral testimony” (p. 137), which of course includes moral affirmations by prophets or indeed anyone claiming to have clerical or religious authority. In discussing al-Razi and the status of moral deference, Saemi brings out two general ideas that figure in his proposed solution to what he calls the new problem of evil.

One idea Saemi proposes is that (overall) “obligation is a function of the overall interaction of [objective] reasons” (p. 158). This needs the qualification that

The reasons accessible to us are not all the reasons there are. This opens up the possibility that it may be rational to reject testimony that in fact, for reasons not accessible to us, makes doing something objectively justified – so that (1) we can say (perhaps in hindsight) that there is (or was) a justification for doing it, though (2) it is not (or would not be) rational to do it given all our accessible evidence. That evidence will of course include some intuitive moral principles and contextual information regarding the kind of action whose moral status is in question.

The second element important here, for Saemi, is the Moses principle:

Here we can see the rationale for the solution Saemi proposes regarding the presumptive obligation to defer in moral matters. Even if we suspend judgment on the objective (factual) reliability of our own judgments,

It is in the (“ethics first”) chapter on reliability that we are given an overall statement of what Saemi calls the new problem of evil. He reminds us that he has been arguing that “our independent moral judgments are reliable in the sense that we should rationally be guided by them” (p. 172). Many philosophers would resist this on the ground that reliability is a matter of objective probability of truth and rationality in action is a matter of accessible, sufficiently good evidence. I am not sure Saemi himself needs to accept the view, as will be evident later. In any case, he combines the view with a “conscience constraint.” As initially stated, that constraint is the view that

where overall wrongness is objective in the sense Parfit (whom Saemi quotes) specifies: Parfit calls what is wrong in the fact-relative sense wrong in that the act in question would be wrong for us to do if we knew all the morally relevant facts (p. 173). This constraint Saemi takes to be usable to show that the objective, fact-relative view is incorrect (p. 176). A crucial example here is the two-drug case: the physician (Jill) has excellent evidence that drug A will save the patient and drug B will be fatal, but unfortunately the facts are the reverse. Nearly everyone agrees that Jill rationally ought to give B, which is indeed right in Parfit’s evidence-relative sense of either term, i.e. “ought” and “right” as applied to actions.

A natural objection here is that the example is insufficiently analogous to the case in which, e.g., Abraham is commanded by God to sacrifice his son. A plausible reply to that objection is that the content of the command, by violating an independently highly plausible moral principle, is strong evidence that the command is not from God. Scripture first theorists may rejoin, however, that even from the point of view Saemi is taking, Abraham, for instance, (1) has excellent (accessible) evidence, presumably through personal acquaintance, that it is God, who is commanding what seems clearly wrong and (2) knows a priori that if God commands something then, given divine omnicompetence, it is not wrong to do it. On this view, even if Abraham were, on his accessible evidence, to refuse to obey, his refusing is not, even given the authority of God or indeed of authentic Scripture, what he morally ought to do.

It appears, then, the problem of evil Saemi is posing can arise within the subjectivist perspective. The subjectivist, like the objectivist, must compare the accessible evidence that God is speaking with the accessible evidence that what God apparently commands is wrong. There is no uncontroversial answer for Abraham, just as there is not for the subjectivist. It does not follow that Saemi’s solution fails, but it looks as if he has not at this point clearly solved the problem. Specifically, the accessible evidence can be divided, as opposed to being predominantly on the side of our considered moral judgments. This is an issue I’ll return to.

In proposing a position to the Scriptural problem of evil that is reasonable for conscientious Muslims, Saemi says, that given the Moses principle and a revised version of the conscience constraint, “conscientious Muslims, due to their uncertainty about the reasons behind Scriptural injunctions, have a moral duty to act in accordance with their own independent moral judgments” (p. 197). We should note that the revised conscience constraint explicitly identifies the moral “ought” with the normative ought that is implicit in conscientious moral deliberation. Specifically, what we have is

Some readers will prefer the first formulation of the constraint. They might say that in killing the patient by giving the wrong drug, Jill did a moral wrong but is excusable. Saemi prefers to say that she did what she morally should have – followed what anyone at the time would have said was excellent evidence.

Both views seem plausible. Killing the patient is opposed by powerful accessible reasons; yet giving the fatal drug on the accessible evidence for its efficacy was the reasonable option. But at this point we should bring excuses into the picture. In that light, we can see that the main substantive issue here is how to describe what morality calls for. For Saemi, this is reasonableness; for many others, it is substantively objective in a way that allows the right act and the most reasonable act to differ. I believe that Parfit, like Ross (in his major statement of his position), Kant, and the historically most prominent utilitarians understood moral terms as objective but would grant excusability where doing something objectively wrong is sufficiently supported by the agent’s accessible evidence. The point is not that excusability is simply a subjective endorsement, or even just an affirmation of the reasonableness of doing what is a priori prima facie wrong, e.g. giving the patient a fatal pill. Excusability is neither. It is in part a moral rejection of blameworthiness. This view makes clear why the physician who kills the patient must appeal to the excusatory evidence favoring the fatal drug’s use. Saemi’s proposal would force us to call the act morally right since it was very reasonable on the evidence.

As far as I can see, then, the problem Saemi is concerned with is approachable in both frameworks. We may take the moral “ought” to be a matter of fit with the agent’s accessible objective (“reliable”) evidence; or, we may take the more widely held view that “right” and “wrong” in the moral senses are objective but allow for excusability where wrong action is fully reasonable on adequate accessible evidence and – objectively— – exculpatory. Then blaming would itself be objectively wrong, even if it could conceivably be excusable. Whether the moral “ought” is objective or evidential in the suggested ways, we still have to determine what a conscientious rational moral agent should do when confronted by religiously authoritative commands or permissions of wrongdoing. We can say that Jill did a moral wrong but is excusable or, as Saemi seems to want to say, that since she did the reasonable thing relative to her accessible evidence, she did not do wrong at all. The crucial action, describable as killing the patient, is wrong but excusable; yet, describable as the conscientious and rational thing to do, it is reasonable. Both descriptions are true, and the issue of which one goes with the moral “ought” remains.⁶

One question I am raising for Saemi, then, is whether – quite apart from whether any major Islamic thinkers can agree on what ought is the moral one (if indeed there is only one moral ought) – they might agree that it can be reasonable to act on one’s best evidence of what is morally required and to prefer doing that to any incompatible deed dictated by Scripture. A partial rationale for this preference is that God wisely created us with this preferential option and, as perfectly good, would excuse us if we have indeed acted on good evidence. Perhaps, the pious could learn to take the view that some scriptural commands and permissions were adapted to time in place – as do some proponents of what Saemi called “the legal interpretation” of Scripture (pp. 214-223), and the goodness of God allows for reasonableness as a constraint in interpreting divine will.

The notion of reliability is important for Saemi, but, somewhat surprisingly, he does not – at least not always – treat it objectively. As noted, for him reliability belongs to what need only be highly probable: “moral judgments obtained through careful a priori moral reflection and sufficient empirical observations, independent of Scripture” (9). Careful reflection is not infallible, and observations sufficient for rationality need not be conclusive. In pursuing his position, I noted that he appeared to take some moral judgments to have generalizations as objects, such as that promise-breaking is wrongmaking (thus prima facie wrong in Rossian terms) and singular judgments, such as that I ought not to treat male and female students unequally. On the plausible view that some moral generalizations are both intuitive and a priori, they can serve as anchors in moral thinking. One could make a case that this is how, epistemologically, Kant viewed the Categorical Imperative and some generalizations he thought derivable from it, how Sidgwick viewed the principle of utility, and arguably, how some virtue ethicists might view certain of their basic general normative claims. When it comes to singular judgments, however, these are empirical even if they are intuitive. Consequentialists and many deontologists would accept this, and some have thought that singular moral judgments are never absolutely certain.

On this moral epistemology, it is not an a priori truth that you or I should not kill someone – it is a priori possible that, on earth, this will spare a whole city and, in some other realm, lead to renewed life of a better kind for the victim. So, how should Abraham have thought of God’s command? He might well have doubted whether it was God really speaking to him – one can, after all hallucinate – and one might well wonder how a perfectly good being could command killing one’s son. But Abraham might also have had evidence that the order did indeed come from God and thought that God would surely make it up to Isaac or otherwise be acting for the best.

For Parfit, reasons are facts and, like Ross (in 1930), though using different terms, Parfit took overall wrongness and rightness to be a matter of the overall weight of the relevant reasons. Wrongness and overall rightness, in Ross’s language, are “totiresultant attributes”: consequential on all the relevant facts of the situation. Both philosophers have objective list theories. For Ross, lying to people and injuring them, for instance, are a priori wrong-making; and beneficent action, say, enhancing knowledge in others, is a priori right-making. Parfit, though he speaks of objective reasons where Ross spoke of grounds of prima facie duties, granted (as I believe Ross would have) that rationality in action is a matter of whether the beliefs, and presumably also certain desires, underlying the action or calling for it are rational on the basis of accessible evidence. Is there any rational alternative to deciding what to do on the basis of such evidence? And is what is practically most rational for us in terms of our accessible reasons for action always what, from a wholly moral point of view, is right? Here Sidgwick’s dualism of practical reason looms. These problems bedevil the issue Saemi is addressing.

I find that his proposal to put ethics first—or at least moral rationality first – in dealing with the Scriptural problem of evil is plausible. Some theists will balk at taking us to have a priori principles to work with that are of a kind that theists cannot rationally deny to be of ultimate practical significance. Even a divine command ethics, which appears to be, in some form, common to the main Islamic theological positions Saemi introduces, may take certain moral principles to be a priori, necessary, and, even commanded or command-worthy by God, but not as alterable by God.⁷ If one takes moral truths to be contingent (and presumably empirical as well), one can still take them to be such as are commanded as a result of God’s nature, where even an omnipotent being cannot be expected to act contrary to its nature.⁸

Need the considerations supporting a perfect being theism go against this perspective? If it is not an “ethics first” perspective, it is as least an “inviolable ethics” one. Indeed, if the God of perfect being theology is omnicompetent, isn’t there good a priori reason to think that any rejection of a true representation of divine will tends to be a priori suspect? For many moral theorists, including some divine command theorists, it is a priori that God is morally infallible; for the moral rationalist who, like Ross, has an objective list theory of the basis cases of the right and the wrong, any command inconsistent with the relevant a priori principles is false.

The problem here is deep if, as seems highly plausible, the basic general principles that are genuinely a priori and may be considered necessary truths independent of divine commands are not absolute but, in Ross’s language, prima facie or, in contemporary terms that are apparently equivalent, are principles that each ground a reasons for action rather than, by themselves, determining final (overall) obligation. On some plausible version of Hume’s law, no non-normative description of an option we contemplate entails (a priori) that the option is overall wrong; that depends on the overall weight of reasons – as Saemi has granted – or of prima facie duties, as Ross held. Ross’s view implies that Abraham could not know that killing Isaac was overall wrong. Even if Abraham did not take knowledge to entail certainty, he might also have thought that he had better reason to think the apparent command was not from God or – more likely for the Biblical context – was intended as a test and would not be allowed to be executed. He might then be thought to be rational in doing, as he did, what God commanded.

My hope through reading Saemi’s book is that there are enough elements in the apparently pious Islamic philosophers he seeks to explicate to support this approach – or at least an ethics-essential approach – to understanding scriptures and clerical authority. I prefer that term to “ethics first”, which is easily read as giving our moral thinking a kind of priority that is not entailed by the most moderate forms of the view compatible with piety. Some of the pious may call the view I am suggesting corrupted by sin, since it would risk reposing more confidence in the truth of at least one kind of moral belief than one has in one’s receiving a genuine divine command – or in reading an infallible injunction from a text or hearing a clerical affirmation meant to convey one. But the proposed view is not biased by any one religion, clerical order, or scripture, and it is compatible with due respect for a number of them.

Indeed, do we not rely on reason as supplying the apprehensional powers needed even to understand theistic religion and its moral authority? And might we think that disobedience to religious authorities when it is rational to disagree with them – provided we meet the intellectual standards that we must master in order to be conscientious theists and reasonable moral agents – is, if not morally permissible, then at least excusable if mistaken? If Abraham was being reasonable, why wouldn’t an omnicompetent God who created him with reason as his central guiding capacity forgive him for doing wrong in that case? And if he did wrong by executing the divine command – which is logically possible if, as is arguable, omnicompetence does not entail giving only morally sound commands – why might God not withhold punishing him for that, depending on whether the action was reasonable in the light of all Abraham’s accessible evidence? If the basic standards both of right and wrong and of the reasonable and the unreasonable are a priori and necessary, then it is no impiety to treat them as constraints – for theists, God-given constraints – on the acceptance of theological and clerical authority. These standards would not be above God, but structural elements in the divine mind. This view, then, might be compatible with piety.

How much might this approach support Saemi’s conclusion regarding what conscientious Muslims should do? He indicates resources in Islamic philosophers who sought to combine their piety with natural reason. Might he have found elements that enable reconciling piety with rejection of literalism in interpreting Islamic Scriptures? Is not Scriptural literalism far more the problem Saemi is combatting than the difficulty of deciding between a subjectivist theory of right and wrong and an objectivist theory of right and wrong combined with an epistemic account of excusability? I have cited R. M. Adams as among the many Christian philosophers have done this in relation to biblical interpretation. I believe Saemi may have resources to achieve something close to this in relation to Islam, and I hope to have suggested some ways in which he can use them with good results. The suggested view, whether not it represents an ethics first approach, is an ethics essential approach.⁹