CC BY
8Философия религии: аналитические исследования
Philosophy of Religion: Analytic Researches 2022, vol. 6, no. 2, pp. 39-55 DOI: 10.21146/2587-683X-2022-6-2-39-55
2022. Т. 6. № 2. С. 39-55 УДК 21
M.C. Murphy
God is Not Best-Acting
Mark C. Murphy - McDevitt Professor of Religious Philosophy, Georgetown University, 215 New North, Washington, DC, 20057, USA;
e-mail: murphym@georgetown.edu
Richard Swinburne argues that God is a 'best-acting' agent, relying upon the thesis that an all-knowing, all-powerful being will be fully motivated to act in accordance with the norms of morality. In God's Own Ethics, I reject the view that such norms apply to God, offering an alternative account of the norms of divine action. Swinburne offers multiple criticisms of that alternative account: (a) that its claims about divine intention of evil are inconsistent; (b) that it makes obviously false claims about the motivations of a perfect being;
(c) that it undermines the prospects of probabilistic arguments for God's existence; and
(d) that it fits poorly with the view that creaturely perfection is analogous to divine perfection. Swinburne's criticisms are, however, unconvincing: (a) Swinburne does not notice the importance of my appeal to the distinction between intending an evil and making use of an evil; (b) Swinburne's appeals to intuition about the motivations of a perfect being are out of place in this dialectical context; (c) it does not itself call into question a certain account of the divine nature that such an account rules out a class of arguments for God's existence; and (d) Swinburne's appeal to Thomistic analogy does not reckon with the crucial distinction between those features that are ascribed to God literally (if analogically) and those that are ascribed to God metaphorically. Swinburne's criticisms of my view are thus unconvincing, though some of them do call our attention to issues that require closer examination.
Keywords: Anselm, Swinburne, God, ethics, morality, perfection
Citation: Murphy, M.C. "God is not Best-Acting", Philosophy of Religion: Analytic Researches, 2022, Vol. 6, No. 2, pp. 39-55.
© Murphy M.C., 2022
Бог не есть «действующий наилучшим образом» Марк К. Мерфи
Джорджтаунский университет, 215 Нью-Норт, Вашингтон, округ Колумбия, 20057, США; e-mail: murphym@georgetown.edu
Ричард Суинберн доказывает, что Бог, это агент, «действующий наилучшим образом», основываясь на тезисе, что всезнающее, всемогущее существо будет с полной мотивацией действовать в согласии с нормами морали. В «Собственной этике Бога» я отрицаю мнение, что таковые нормы применимы к Богу и предлагаю альтернативный подход, исходящий из норм Божественного действия. Суинберн предлагает разнообразную критику этого альтернативного подхода: он полагает, что (а) утверждения этого подхода о божественных намерениях по отношению ко злу противоречивы;
(b) что из него следуют очевидно ложные выводы о мотивации совершенного существа; (с) что этот подход в перспективе подрывает вероятностные аргументы в пользу существования Бога; и (d) что оно слабо согласуется с тем взглядом, что Божественное совершенство аналогично совершенству твари. Однако критика Суинберна неубедительна: (а) Суинберн не замечает важности моего обращения к различению между намерением совершить зло и использованием зла; (b) обращение Суинберна к интуиции
0 мотивации совершенного существа неуместны в данном диалектическом контексте;
(c) сам по себе мой подход не ставит под сомнение определенный подход к Божественной природе таким образом, что данный подход исключает класс аргументов в пользу существования Бога; (d) Суинберново обращение к томистской аналогии не учитывает ключевое различие между теми особенностями, что приписываются Богу буквально (или аналогично) или теми, что приписываются Богу метафорически. Критика Суинберном моих взглядов, таким образом, не является убедительной, хотя она отчасти привлекает наше внимание к проблемам которые требуют более тщательного исследования.
Ключевые слова: Ансельм, Суинберн, Бог, этика, мораль, совершенство
Ссылка для цитирования: Murphy, M.C. God is not Best-Acting // Философия религии:
аналит. исслед. / Philosophy of Religion: Analytic Researches. 2022. Т. 6. № 2. С. 39-55.
In "God's Moral Goodness", Richard Swinburne argues that God is a "best-acting" agent, offering a defense of this claim and criticizing an alternative view that I defend in God's Own Ethics. After some preliminary methodological points (§1),
1 will turn to the details of Swinburne's argument and his criticisms of my own position. I will first identify the gap in Swinburne's argument for the claim that God is a best-acting agent (§2), and I will briefly rehearse the central line of argument in God's Own Ethics that we should think that this gap is not fillable (§3). I will then turn to Swinburne's attempt to render implausible the claim that God is not bound by the sort of moral requirements by which we created rational beings are bound (§§4-7). I do not think that Swinburne's arguments or criticisms are persuasive, though he raises issues that merit further exploration.
§1
The view of divine action that I accept begins with the idea of God as an absolutely perfect, or 'Anselmian', being [Murphy 2017: 9]. My doing so has nothing to do with my views on the prospects of Anselmian ontological arguments for God's existence. My main reasons for starting with the Anselmian view of what it is to be God are (a) the centrality of this Anselmian notion in both classical and contemporary philosophical theology and (b) the fact that it is a non-negotiable element of my own faith tradition's understanding of God that God is absolutely perfect.
There is nothing in my way of proceeding that commits me to the viability of onto-logical arguments for God's existence. Aquinas, for example, famously rejects Anselm's ontological argument [Aquinas, Summa Theologiae Ia 2, 1, ad 2], but Aquinas is an Anselmian about God in the only sense that is relevant for my account: he thinks that God is absolutely perfect [Aquinas, Summa Theologiae Ia 4, 1]. One who takes God to exist on the basis of the Church's testimony can be an Anselmian in the only sense that matters to my project, if such a person takes the testimony of the Church to be that God is absolutely perfect or takes the testimony of the Church, along with some further justified premises, to entail that God is absolutely perfect1.
Swinburne suggests not only that ontological arguments are bound not to be informatively sound. He also suggests that there is something shaky about Anselmian methodology in determining what features should be ascribed to God, for this methodology requires us to rely on "highly contestable moral intuitions" (P. 7). I am uncertain about the 'moral' part; I don't think that my views on the greatness of lions depend on moral intuitions, and I don't think that my views on the greatness of God depend on those, either. But perhaps some of these intuitions are moral, and some are other sorts of value intuitions, and all of them are indeed highly contestable. We have of course more than these intuitions to work with: just as moral philosophers are disciplined in their theory-building not only by common sense reflective moral judgment but also by formal constraints on the domain of the moral (e.g. the moral exhibits impartiality, universality, supervenience on the non-moral, reason-givingness, etc.), Anselmian philosophical theologians are disciplined not only by their value judgments about perfection but by formal constraints on maximal perfection and, for many of us, by the Church's testimony about what God is like and what responses are fitting with respect to God2. The task of shoring these indi-vidually-all-highly-contestable judgments into consistency and relations of mutual support and then bringing the product of that attempt into contention with the products of other attempts to shore up such packages of intuitions and formal constraints looks like a typical philosophical endeavor, both in its promise and its pitfalls. It cannot be that Swinburne is criticizing Anselmian theorizing about divine perfection for being difficult in the way that other philosophical tasks are difficult, but I don't know what else the worry is supposed to be.
See, for example, [van Inwagen 1998: 49-59] and [Leftow]. In Divine Holiness and Divine Action, I argue that acceptance of Scriptural testimony that God is supremely holy commits one to the view that God is absolutely perfect [Murphy 2021: 45-59].
I discuss this Anselmian methodology in Murphy [Murphy 2017: 10-12, 19-21].
So I am an Anselmian about what it is to be God, and I think that there is a being that answers to this description, and my task in God's Own Ethics was to figure out what we should believe about the norms of action of such a being, including whether those norms of action are the norms of action by which we humans are bound. Swinburne and I come to different views. He says that God's norms of action are our norms of action. I deny this.
§2
Put to the side thoughts about an Anselmian being, and instead consider a 'Swinburnean being', an agent with unlimited knowledge and power. Swinburne thinks that we can argue from being a Swinburnean being to being a morally perfect, 'best-acting' being. Here is the argument, which is familiar from earlier work of Swinburne's3:
So [a Swinburnean being] will know which of the actions available to him are morally good, and which are morally bad and which are morally indifferent; and which are better than other actions and which are worse than other actions.
To believe an action to be morally good to do entails having some motivation to do it. One would not have the concept of moral goodness unless believing an action to be morally good inclined (=motivated) one to do it. Likewise to believe an action to be morally bad to do entails having some motivation not to do it; to believe an action morally better than another action entails having greater motivation to do the former than to do the latter; and to believe an action morally worse than another action entails having greater motivation not to do the former than not to do the latter... We humans of course do not always do what we believe to be good, when we have the power to do so, because we are subject to unchosen inclinations of a non-rational kind to which we sometimes yield. But an omnipotent being would have no such inclinations, and so in any situation would always do the best possible action where there is a best action available, and an equal best action (one which is equally good as some other action or actions, all of them being better than any other actions) where there is no unique best action. I will call such a being a "best-acting" God (P. 8-9).
Let's stick with morally best actions and Swinburne's argument that a Swinburnean being will perform them. That will save words without obscuring any of the differences between Swinburne and me.
In restating Swinburne's argument, I will make use of expressions of the form "X would best 9, morally speaking". Should I act on my urge to belittle my colleague? I would best not, morally speaking. Should she lie about her accomplishments (and thus improve her chances to get hired)? She would best not, morally speaking. I use "X would best 9, morally speaking" rather than the less cumbersome "it would be morally best for X to 9" because this less cumbersome expression is ambiguous. It could refer, as the more cumbersome "would best 9" expression clearly does, to a demand on X's choice - roughly, that X ought to choose 9-ing, that 9-ing is the right decision for X to make, that 9-ing is the proper outcome
See, for example, [Swinburne 2016: 200-221].
of X's deliberation. It could, by contrast, refer to an impersonal claim about moral goodness being realized, that what is best from an impartial, moral point of view is that X 9. This is not an appeal to what is right for X to do, or what X's correct deliberation issues in, even if there is some connection between the impersonal best-ness of an agent's acting a certain way and acting that way's being the best choice for the agent to make.
Since the argument does not get off the ground unless we assume that what is being invoked is the notion that some choices are the correct ones for God to make in light of the moral norms that apply to God, I will use the more cumbersome expression to mark that intended meaning.
So Swinburne's argument is this. A Swinburnean being, knowing everything, will know what a Swinburnean being would best do, morally speaking. Knowledge of moral propositions entails motivation - or, at least, X's knowledge that X would best 9, morally speaking, entails that X will be motivated to 9. To be motivated to do something entails doing that thing in the absence of obstacles or motivations to the contrary. But a Swinburnean being, being omnipotent, will not be gripped by contrary motivations or prevented by obstacles from doing that which that being is motivated to do, which includes doing what a Swinburnean being would best do, morally speaking. So a Swinburnean being is a best-acting being.
There are multiple points at which one might quibble with this argument, resting as it does on multiple contentious theses of metaethics and moral psychology. But as far as I can tell there is only one issue here with respect to this argument on which Swinburne and I must disagree. Swinburne's argument here requires that there be some truths of the form "God would best 9, morally speaking" for God to know. I do not think that there are any such truths. I of course concede that if there were any such truths, then God would know them, and then Swinburne's argument would succeed. But there are no such truths for God to know. Swinburne has to give persuasive grounds for thinking that there are such truths about what a Swinburnean being would best do, morally speaking, and he has not.
One might deny that Swinburne's argument requires that there be any such truths. After all, God could be a best-acting God trivially: if there are no truths about what God would best do, morally speaking, then obviously God does not fail to do what God would best do, morally speaking. But this is not acceptable for Swinburne. Swinburne relies upon God's being best-acting to make claims about what a Swinburnean being would likely do (P. 9), and that requires not just the claim that God does not fail to do anything that God would best do, morally speaking, but also the claim that there are some things that God would best do, morally speaking, and so we can expect a Swinburnean being to do those things.
§3
Swinburne does not report in detail why I deny that there are any truths about what God would best do, morally speaking, though he correctly remarks that my denial is due to the massive differences between God and created rational beings like us. Let me quickly rehearse my position, framed in relation to Swinburne's views.
Swinburne's assumption that moral beliefs entail motivation is most plausible on a strong version of a thesis about moral rightness called 'moral rationalism': the view that what it is for some action to be all-things-considered morally required for an agent to do is for there to be decisive reasons, of a certain morally relevant kind, to perform that action [Murphy 2017: 23-29]4. So there are truths about what God would best do, morally speaking, only if there are decisive reasons, of a morally relevant kind, to perform some action.
Further, in order for these decisive reasons to be of a morally relevant kind, they must be in the vicinity of the reasons that fix the moral requirements that we are under. In particular - and this fits with Swinburne's assumptions about what a Swinburnean being would be expected to do - these reasons must be reasons to promote and to prevent setbacks to the well-being of rational creatures generally, and these reasons must be requiring (that is, they must be reasons that are to be acted on unless there are reasons to the contrary)5.
Both of these presuppositions - that (1) there are truths about what God would best do, morally speaking, only if there are decisive reasons, of a morally relevant kind, to perform some action and that (2) for the reasons to be of a morally relevant kind, they must be requiring reasons to promote and prevent setbacks to the well-being of creatures - should be accepted by Swinburne. If he does not endorse the former, he lacks any plausible account of why knowledge of moral propositions' truth would translate into motivation. If he does not endorse the latter, then he leaves open the possibility that while God is moved by something that can be labeled "moral truths", it still could well turn out that what God would best do, morally speaking, is not something that we can predict on the basis of the moral norms that bind us humans, and so would be useless for Swinburne's probabilistic arguments.
But we should doubt that both of these presuppositions are true, and indeed I think we should judge their conjunction to be almost certainly false. For it is a platitude that the reasons of different rational agents can be different, and the more different the rational agents, the more different we should expect those reasons to be. God's difference from us creatures is overwhelming - God's knowledge and power being maximal, God's being the source of all else that exists (and thus) God's having no source.
See also [Murphy 2019: 97-100]. For moral rationalism, see, e.g., [Smith 1994]. Swinburne's commitment to moral rationalism is plain in the formulation of the 'best-acting God' argument that appears in The Coherence of Theism: "Not having any non-rational desires, [the Swinburnean being] will act on reason alone, and so will do what is probably the best or equal-best action"; this assumes that an action's being the best, morally speaking, entails its being the action that there is decisive reason to perform.
I label these the requirements of "familiar welfare-oriented moral goodness" [Murphy 2017: 24]. For the distinction between justifying and requiring reasons, see [Ibid.: 59-60]. The distinction is drawn from the work of Gert [Gert 2004: 19-39]. Swinburne's assimilation of requiring reasons to obligations (P. 13) is inexact and misleading. (E.g., one could plausibly reject the view that a Swin-burnean being is bound by any obligations while holding that that being must be best-acting; I don't think one could plausibly reject that view that the Swinburnean being has requiring reasons to tend to creaturely good while maintaining that the Swinburnean being is best-acting).
4
And there is more. It is also a platitude that reasons, especially reasons to look to the good of another, call for some explanation. The history of moral philosophy is a history of attempts to give such explanations. But of the philosophically live, historically important positions - Hobbesian, Humean, (plausible) Kantian6, Aristotelian - all of these take as indispensable to their account of the good reasons to look after others' good that we humans have some key feature, and for each of these views, God lacks the key feature that explains why we humans are bound by moral norms. Hobbesian explanations rely on our being vulnerable and dependent; but God is not vulnerable or dependent. Humean explanations rely on our exhibiting some psychological dispositions that do not arise from our status as rational; but God does not have any such psychological dispositions. (Plausible) Kantian theories proceed from each of us being merely "one person among others" [Nagel 1970: 88]; God is not merely one person among others. Aristotelian theories explain our reasons by appealing to our specific kind, which fixes conditions of our excel -lence in agency; God does not by nature belong to our kind. If we look at the terrain of otherwise plausible theories of requiring reasons to promote others' good, we will see that there is very little room occupied by theories that would entail not only that we but also a being like the Swinburnean being would have such reasons for action.
The best hope for such a view, I say, is one on which it is held just that rational creatures have intrinsic value, and due to that intrinsic value, every possible rational agent has reason to promote their good and prevent setbacks to it. I want to say two main things about this view. First, it is a contentious and controversial position about reasons for action, one that has its own difficulties and stresses, and the notion that it can be taken for granted in a consideration of the reasons for action that God has is not plausible. Second - and this is a point where my being an Anselmian about God makes a difference - it is itself especially implausible in light of an Anselmian understanding of God, on which God should be conceived not only as the greatest among good beings but as the ultimate source and explainer of the goodness of any created thing [Murphy 2017: 60-62]. The notion that the agency of the source of the goodness in all good things is constrained and required by creaturely goodness, which is itself explained in terms of the divine goodness, seems very implausible. At the very least, God stands in such a different relationship to creaturely goods that it is hard to see why we would suppose that divine action is practically necessitated by it in the same way that our action is, which is what Swinburne's view would suppose.
The central thesis about divine ethics defended in God's Own Ethics is that the good of creatures provides the Anselmian being, an absolutely perfect being,
Implausible Kantian views are those that hold that universal norms of morality derive from some formal requirement on rationality, or just announce without argument that moral norms are universal and bind every rational being. I take no stand on how far these views are Kant's position, but they are not plausible and, in the case of the latter view, beg all the live questions, so I will not pursue them further. I do discuss them a bit in [Murphy 2017: 46-48]. By contrast, plausible Kantian views are those that rest on anti-arbitrariness arguments taking as a baseline how one has to value oneself as a rational being; these are discussed in [Ibid.: 55-58].
with justifying but not requiring reasons to promote that good and to prevent setbacks to it. What this means is that while promoting the good of creatures, or preventing setbacks to their good, is something that makes divine action rational - God can rationally act for the sake of it - it is something that God does not need any reasons not to do. The good of creatures provides God with opportunities for action that God can choose to act on, but God need not, and does not need any reason not to, act on those opportunities [Murphy 2017: 67-75]. We thus cannot predict simply from God's being perfectly rational that God will be motivated to promote any particular creaturely good or to prevent any particular creaturely evil.
There is a further thesis that I defend about divine action, which is that even though God does not have requiring reasons to promote creaturely good, nevertheless God never intends evil - that is, finally bad states of affairs - whether as an end or as a means. The argument is not that God is bound to respect creaturely good, due to its value. It is that intending evil places God in an intimate relationship with evil, making that evil's obtaining a success condition for divine action in a way that is not fitting for the action of an Anselmian being. It is an unfitting response to God's goodness, not to creaturely goodness, for God to be intimately related to evil in the way that intending evil requires. In the book I suggest, in passing, that we can think of this as an explanation that appeals to divine holiness [Ibid.: 100]; I have since argued for that view at greater length [Murphy 2021: 150-152]. Because I believe that the Anselmian being exists, I am thus committed to the view that while the Anselmian being is, for every existing evil, in some way part of the explanation of that evil's existing, not a single one of the evils in this world is intended by the Anselmian being.
This is a bare sketch of the argument of God's Own Ethics as it bears on the claim that God is not bound by moral norms. I do not fundamentally appeal to intuitions about divine perfection in holding that God is not bound by moral norms, though I do supplement these considerations with the thesis that if we are to think of God in Anselmian terms, as an absolutely perfect being, we would expect God to have greater discretion over the created world rather than less, and that should lead us to expect God to be far less constrained by requiring reasons than Swinburne's view entails.
It is clear, then, that Swinburne and I disagree about whether the norms of morality that bind us humans also bind God. This disagreement, in turn, leads us to disagree about the threat posed by the argument from evil against God's existence. Swinburne is committed to the view that states of affairs involving creaturely evils are such that God has good reason to prevent their obtaining, and thus we need some account of the reasons that God has that are sufficient to make rational God's allowing those evils to obtain. It follows from Swinburne's account of divine motivation that God will be motivated to prevent these evils, and thus we should expect them not to obtain - contrary to our experience - unless we can provide some account of the reasons that God would also have that would lead God ultimately to act on different motivations and thus allow these evils to occur. Providing such an account is classic theodicy, and Swinburne takes it to be crucial to the defense of theism (P. 10-11). On my view, by contrast, the argument from evil does not get off the ground, for God's reasons to prevent creaturely evils are justifying only.
What this means is that while the prospect of preventing some creaturely evil is something that makes rational God's preventing that evil, God needs no reason not to prevent creaturely evils, and so we should not have any expectation that God would be motivated to prevent them or any expectation that our world would thus lack such evils [Murphy 2017: 104-109]. Swinburne takes creaturely evils to give God requiring reasons to prevent them, and so God will prevent them unless there are good and sufficient reasons not to; I take creaturely evils to give God only justifying reasons to prevent them, and so God might well fail to prevent them, even if God has no reasons not to prevent them. On Swinburne's view, classic theodicy is central to rational theistic belief; on my view, classic theodicy is an unnecessary enterprise.
§4
Swinburne says that there are three major difficulties for my position, and then adds an un-numbered fourth (P. 12). The first is that he thinks that there is some difficulty in my view that God does not intend any of the evils of this world. The second is that, given an Anselmian view, it is implausible that a perfect being could fail to prevent setbacks to creaturely well-being without having any reason for refraining from preventing those setbacks. The third is that a view like my own undermines any prospect for probabilistic arguments for God's existence. The fourth is that a proper view of the analogical relationship between the goodness of God and the goodness of creatures should lead us to think that we and God share a common set of moral norms. We will consider these in turn.
First, on my claim that the Anselmian being does not intend the evils of this world. Swinburne's worry here is not the more abstract concern that some folks have that theism commits one to thinking of everything that happens in the world as intended by God. Rather, the dialectical situation is this. I admit that, as a matter of revelation, God intended the existence of embodied rational creatures, and I admit that, as an empirical matter, the process by which these embodied rational creatures came into existence involved natural selection, with its essential early loss of life to creatures, which states of affairs I concede to be evils. Swinburne thinks my view problematic.
Strangely, although [Murphy] admits that "natural selection involves creaturely evils", he claims that since "God obviously intended natural selection as a means to bring about rational animals", that does not entail that God intended "evils". He claims that God foresaw the evils, but did not intend them. Yet, since as Murphy admits, God could have brought about rational animals without natural selection, this seems implausible. Maybe God intended natural selection as a means but not as an end, but he still intended it (P. 12).
If I granted what Swinburne says that I grant, then it would indeed be true that God intended natural selection - including the evils that constitute it - as a means, and that would be a problem: I would have to deny theism or deny that God never intends evils. But I do not grant what Swinburne says that I grant. When I wrote that "God obviously intended natural selection as a means to bring about rational
animals" [Murphy 2017: 117], I put that claim in the mouth of an imagined objector, not in my own voice. My own view denies that God intended the evils of natural selection, even as a means.
My view relies upon a distinction articulated by Kamm between intending an evil and making use of an evil [Murphy 2017: 116-120; Kamm 2007]. There are ways in which one can, without intending an evil, take for granted its coming into existence and incorporate it into one's plans. I do not intend my death if, knowing that I will die, I make use of the fact of my death in purchasing life insurance in order to provide benefits to my heirs. This distinction holds even if the state of affairs of which one makes use was caused by one's own action. I would like a second fifth cup of the delicious coffee on offer, but I know that it will keep me up late that night; I decide that I will take advantage of my impending sleeplessness by getting some administrative work done. This is making use of a foreseen evil of some otherwise choiceworthy option, not intending an evil as a means. My thesis, which Swinburne's remarks here do not address, is that the evil of this world is not intended by God, even if some of it is such that God made use of it for God's own worthwhile purposes. This is the account I propose regarding how we should understand God's practical stance with respect to early creaturely death in natural selection: God makes use of the evils of early creaturely death in bringing about the existence of embodied rational creatures.
There is an independent criticism in the passage quoted above that is worth bringing out and addressing. Swinburne suggests that it is implausible that God did not intend the evils of natural selection, because there were other ways that God could have brought about the existence of rational beings. Perhaps the idea is something like this. Suppose that I want to go over and pour myself a cup of coffee, and there are many seemingly equally easy ways for me to cross the room, only one of which involves my stepping on your toes - an outcome that is obvious if I take that path. I take that path. One might think that it is obvious that I must have intended to step on your toes.
I endorse the prima facie plausibility of this reasoning in the case of my stepping on your toes, but the cases are insufficiently analogous for us to draw the relevant conclusion regarding God's intention of evil. The reasons that I have not to step on your toes are requiring reasons, reasons that make it the case that other things equal I should go get my coffee without stepping on them. It is reasonable to suppose that if I - a rational, aware person - step on your toes while going to get coffee in the circumstances described, then I likely intentionally did so, in light of some other reason that I was trying to act on. But we cannot extend this reasoning to the divine case. An Anselmian being does not have requiring reasons to look to the well-being of creatures. We thus cannot make use of the plausible interpretive rule that if a rational being is aware of a requiring reason to change course and does not act on it, that rational being must be intending to act contrary to what that reason requires. This interpretive rule is not apt in the case in which the reasons in play are merely justifying.
§5
Here is the whole of Swinburne's second point against my view: "[Murphy's] account of what the perfection of a perfect being would consist in seems implausible - surely a perfect being would not allow his creatures to suffer for no good reason" (P. 12).
It is hard to know how to respond to this because there is no argument given. If the view is simply that we should expect that a perfect being would be under the same norms that we are, and those norms entail that one should not allow one's creatures to suffer for no good reason, then a perfect being would not allow that being's creatures to suffer for no good reason, then I have given my response: we have many reasons to doubt that a perfect being would be under the same norms that we are. Until Swinburne actually engages with the difficult problems of moral philosophy regarding the source and explanation of reasons for action, and deals with the application of such views to the issue of the reasons of a perfect being, then he has not held up his end of the exchange.
If the view is just that the position I describe is so implausible that we just need to pause to register that fact, then I would respond that it does not seem so to me, nor would it seem such to, say, Aquinas and Scotus7. (Nor to Aristotle. Aristotle's God was not a creator, but Aristotle took to be absurd the notion that God would even have thoughts about the limited beings of this domain [Aristotle, Metaphysics XII]).
Here is another way to put it. Swinburne may be reporting how my claims about divine reasons seem to him, even upon consideration of the relevant arguments. In which case, I take it, we had better just focus on those arguments' quality, and not the seeming that resulted from Swinburne's thinking about them. Or Swinburne may be reporting his priors, his initial seeming in the face of the claim that a perfect being could allow that being's creatures to suffer for no good reason. In which case, I would want to know more about why we should place much trust in that particular sort of seeming. There is an instructive comparison here to van Inwagen's modal skepticism, on which modal judgments about matters far removed from ordinary life are not much to be trusted [van Inwagen 1995: 19-21]8. I think that we ought to have a similar sort of hesitancy about ascribing norms of action to agents who are far outside the ordinary domain of human life.
§6
The third difficulty for my position is that "it has the consequence that there cannot be a cogent probabilistic argument from the observed features of this world to the existence of God - because those features will only make it significantly
For discussion, see [Davies 2011; Adams 1987].
Interestingly, van Inwagen's initial formulation of modal skepticism is in the context of a reply to Swinburne's The Coherence of Theism.
probable that there is a God if, if there is a God it is significantly probable that the universe will have these features (which is a consequence of my first criterion for the success of such an argument)" (P. 13). Because the Anselmian being is not necessarily morally motivated, it does not follow from that being's existence that that being would be disposed to bring about a world that exhibits the morally positive features that this world exhibits. (This is the dark side of what I advertise as a consequence of my view that should be welcome to the-ists - that it undermines both logical and evidential arguments from evil [Murphy 2017: 103-109]).
It is not obvious why this is an objection to my position. For it should be an open question what sort of arguments one could possibly offer for an Anselmian being's existence, and there is nothing obviously objectionable about a view that entails that certain sorts of arguments for that being's existence are unavailable. (It is not an objection to someone's claiming that a perfect being is immaterial that this would rule out arguments for a perfect being's existence based on someone's having touched the perfect being).
Perhaps the objection is more ad hominem. I, Murphy, believe that the Anselmian being exists, and I am willing to assert it. So I must take that proposition to have the right sort of positive epistemic status for me aptly to assert it. But once I reject the sort of view that Swinburne has of God's reasons, affirming instead the view of God's Own Ethics, then I will lose access to a whole class of arguments that God exists, all those based on probabilistic reasoning based on features of the observed world. I might still have ontological arguments, but Swinburne takes these to be entirely unpromising.
My views on this sort of worry are very unsettled. Two preliminary points. First, as I note above, I do not have settled views on the availability of powerful philosophical arguments for the existence of an Anselmian being. So I am not sure how concerned I should be even if it did turn out that both of these major kinds of philosophical argument are unavailable. Second, I think that it would be good to reflect on the example of Aquinas and his mode of reasoning to God's existence. Swinburne conflates Aquinas's style of reasoning to God's existence and Swinburne's own style of probabilistic reasoning. Aquinas, though, held a view of God's ethics similar to my own and thus, unsurprisingly, does not argue for God's existence on the basis of any claims about the particular evaluative features of the world that we would, on Swinburne's view, expect God to be motivated to bring about. Instead Aquinas argues on the basis of claims about very general features of the world - motion, contingent existence, order - that metaphysical-explanatory principles entail that given these features, there must be a being that is their source9. It is possible, as Swinburne does, to treat Aquinas's and Swinburne's styles of argument
One of the five ways does appeal to goodness [Summa Theologiae Ia 2, 3], but it does not assume that God must be motivated to bring about creaturely goodness. Aquinas is clear that creaturely goodness does not have that sort of status with respect to divine motivation: "The will of God cannot be investigated by reasoning, except for those items that it is absolutely necessary for God to will. Now... such items do not include what God wills in regard to creatures" [Aquinas, Summa Theologiae Ia, 46, 2].
as instances of a single genus - arguments to God from observed features of the world - but this, it seems to me, obscures key differences between them. It is not at all clear that Thomistic metaphysical theorizing to God's existence requires views on divine motivation of the Swinburnean sort, and so at least some sorts of non-onto-logical-argument theorizing to God's existence remain open. I will not comment on how successful such arguments are or might be, but it seems to me that work done by (e.g.) Pruss [see, e.g. Pruss 2009] and Koons [see, e.g. Koons 1997] have kept the prospects of arguing to God without appeal to substantive claims about divine motivation very much alive.
I confess to being uncertain about how to deal with more radical versions of Swinburne's probabilistic challenge to my view, those that do not claim simply that my view eliminates the prospects of certain sorts of arguments for God's existence but also hold that my view of God actually undermines any the prospects of being any more than an agnostic about God's existence. Climenhaga has recently forcefully defended this view. He argues that given standard Bayesian models, the skeptical theist's thesis that we cannot assign a probability to God's allowing the evils of this world other than 'somewhere between 0 and 1' has the implication that the skeptical theist is committed to no-better-than-agnosticism about God's existence10. This is so at least if we endorse Climenhaga's 'Explanationism'11: on his view, we must employ Bayesian methods to fix the probability of God's existing because, given theism, the holding of every other contingent state of affairs depends on God's existing; and so no other way of coming to know the probability of God's existing can be available to us.
Climenhaga correctly notes that the force of this argument against my view requires the assimilation of the skeptical theist's position (on which God's allowing evil may have some objective probability, though we have no clue as to what that probability is) and my view (on which God's allowing evil has no objective probability at all). I am not sure whether I should accept that assimilation. There seems to me to be a world of difference between a view that takes the objective probability (e.g.) of God's creating to be determinate but entirely unknown and a view that takes God's creating to be the wrong sort of thing to even have an ob -jective probability, and I think more work needs to be done on how the denial of objective probabilities in some contexts is to be reconciled with Bayesian episte-mology. But I would reiterate that I have offered independent arguments that an absolutely perfect being's reasons for promoting creaturely goodness are no more than justifying, and these arguments remain even in the face of a puzzle about how we could know that such a being exists. I have a problem: I do not know best how to handle worries like Climenhaga's, worries that are plainly radicalized versions of Swinburne's point that my view makes trouble for certain sorts of natural theology. They have a problem: the best way to think about divine
10 Climenhaga, N. If We Can't Tell What Theism Predicts, We Can't Tell Whether God Exists: Skeptical Theism and Bayesian Arguments from Evil. Oxford Studies in Philosophy of Religion, forthcoming.
11 Ibid.
perfection suggests that God would not be necessitated or even disposed to create, and would not be necessitated or even disposed to bring about one set of crea-turely goods over another.
§7
There is a fourth argument that Swinburne offers, this time a positive argument that we should take God's ethics to be like our own. The reason is that taking God's excellence to be somehow analogous to our own requires us to think of God's moral excellence as like our own, but greater. This undermines my view that the moral norms that we are under are not norms of action that apply to God. Here is Swinburne:
Those of us who depend on probabilistic arguments for a belief in the existence of God, understand them to show the existence of a creator who is like humans in being a substance who exercises his powers intentionally to bring about effects in virtue of his knowledge of what the effects are like. Murphy and most other theists acknowledge that God possesses such properties, although they emphasise that God possesses these properties, as Aquinas writes, "in a more eminent way" [Summa Theologiae 1a 13, 3]. So the obligations to the creatures of whose existence and nature he is the full cause must have some similarity to the obligations of human parents to the children of whose existence they are only a very small part of the cause. Our understanding of moral responsibility involves the recognition that if an agent is the cause of some effect which he has an obligation to make good rather than bad, then the greater his share of causing this effect, the greater is the resulting obligation on him or her to ensure that the effect is a good one. If it depends almost entirely on one physician not making a mistake whether a patient lives or dies, then the obligation on him or her to keep the patient alive is greater than if there is only a small chance that if that physician makes a mistake the patient will die. Hence God's obligation to ensure that our lives are good must be immensely greater than our obli -gations to ensure that the lives of our own children are good. I conclude, contrary to Murphy, that God does have "a requiring reason". to prevent these intrinsically bad states of affairs, such as pains, from obtaining, unless their occurrence is a logically necessary condition of some good at least as good as the bad state is bad (P. 13).
Swinburne notes that theists of Aquinas's sort are committed to thinking of God as having the perfections that we humans have, but "in a more eminent way", and he attempts to exploit that concession to make an argument for God's sharing moral norms with us. And there is a clear sense in which I am committed to the view that God has the perfections that we humans have, but in a more eminent way. But Swinburne makes two errors in attempting to utilize this commitment to argue to the claim that God is bound by moral requirements of the same sort by which we are bound.
Grant for the moment that the fact that our perfection includes moral goodness commits the Anselmian to holding that God's perfection includes moral goodness, but in a more eminent way. (I will return to this below). The first error that I want
comment on is that Swinburne presumes that what "exhibiting moral goodness in a more eminent way" involves is God's being bound by the same moral requirements that we are bound by, but these requirements' asking for more from God due to God's having so much more knowledge and power than we humans have. Put to the side whether one from whom morality asks to do more (and who complies with morality) is morally superior to one from whom morality asks to do less (and who complies with morality). (Surely this will depend on other factors, including how difficult it is for the different agents to carry out those demands). It does not obviously follow from God's being morally good in an eminent way that what this in particular involves is those same requirements' applying to a wider range of possible actions due to God's superior power and knowledge. We might instead have expected this to involve God's being bound by a higher morality than we are bound by - a set of moral norms suitable for the exalted sort of being that God is. If one expects that the moral norms that a being is bound by will depend in some way on the sort being that it is, then we would expect not that the same moral norms that apply to us apply to God, but rather that there is a different, higher set of moral norms that apply to God, which enables divine moral goodness to be of a more exalted type than our moral goodness.
The second error that I want to comment on is Swinburne's initial move from our perfection including moral goodness to God's perfection including moral goodness, but in a more eminent way. This is not good Anselmian reasoning. Any plausible Anselmianism will have to distinguish between those perfections that can be as -cribed to creatures that are impure perfections and those that are pure perfections [Scotus, Ordinatio 1.3.1: 24]. Impure perfections are those that themselves presuppose some imperfection or kind-limitation on the part of its bearer. Being fast (in the sense of local movement) or being perceptive are impure perfections for beings who have them as perfections. Speed is a feature of beings that occupy a limited space and so can move from one place to another; perception is a feature of beings who can be ignorant of something and can come to knowledge by grasping via perception salient features of their environment. By contrast, pure perfections, like being powerful or being knowledgeable, do not presuppose imperfection or any sort of limitation. In the very article of the Summa Theologiae that Swinburne cites in the passage above [Aquinas, Summa Theologiae 1a 13, 3], Aquinas marks the difference between these by noting that while whatever is positively predicated of God is predicated analogously, we can distinguish between what can be ascribed literally to God and what can be ascribed only metaphorically [Aquinas, Summa Theologiae Ia 13, 3 ad 1]12. While being powerful is ascribed analogically to God, it can be understood to be literally true of God, while being perceptive or being fast can be ascribed to God only metaphorically, even if we describe God as maximally perceptive or exceedingly fast.
The relevance of this distinction is that it is an open question whether moral goodness is an impure or a pure perfection. One of the thrusts of the God's Own Ethics argument is that we should believe, both on metaethical and Anselmian grounds, that moral goodness is an impure perfection. If so, it is no more than
12 See, for an extremely helpful discussion of this point, [Harris 2017].
metaphorical to say that God is morally good, but in an eminent way; and if, like Swinburne, one takes it to be literal and begins to reason about divine action on that basis, then one will very likely be misled. I do not doubt that there is a pure perfection in the vicinity of moral goodness: agency, the capacity to move oneself to action on a rational basis. Perfect agency is, I agree, a pure perfection, literally ascribed to God, and in an eminent way. We thus know that God is a perfect agent, and that divine action is perfect action. But all we can infer from this pure perfection's being literally ascribed to God is that God acts perfectly on God's reasons. That does not tell us, though, what the divine reasons are that God acts on, whether they bear on the well-being or perfection of creatures, and whether those reasons are requiring or justifying, and whether they are reasons of promotion or reasons of respect.
Swinburne's fourth consideration does not, then, give us any reason to support his view that divine action is governed by moral norms.
§8
The task that falls to Swinburne is to give good grounds for thinking that a Swinburnean being will be bound by moral norms, where being bound by moral norms involves having decisive reasons for doing what those moral norms require. He has not yet given such grounds, and we have a strong basis to doubt that any such grounds could be given.
References
Adams, M. McC. Duns Scotus on the Goodness of God, Faith and Philosophy, 1987, Vol. 4, No. 4, pp. 486-505.
Aquinas, T. Summa Theologiae. Trans. Fathers of the English Dominican Province. Westminster: Christian Classics, 1948. 3057 p.
Aristotle, Metaphysics. Trans. C.D.C. Reeve. Indianapolis: Hackett, 2016. 712 p.
Davies, B. Thomas Aquinas on God and Evil. Oxford: Oxford University Press, 2011. 192 p.
Gert, J. Brute Rationality. Cambridge: Cambridge University Press, 2004. 230 p.
Harris, J.L. Analogy in Aquinas: The Alston-Wolterstorff Debate Revisited. Faith and Philosophy, 2017, Vol. 34, No. 1, pp. 33-56.
Kamm, F.M. The Doctrines of Double and Triple Effect, and Why a Rational Agent Need Not Intend the Means to His End. In Intricate Ethics. Oxford: Oxford University Press, 2007, pp. 91-129.
Koons, R. A New Look at the Cosmological Argument. American Philosophical Quarterly, 1997, Vol. 34, No. 2, pp. 193-211.
Leftow, B. Arguments proving the existence of G-d. [https://www.oxfordchabad.org/ templates/articlecco_cdo/aid/640776/jewish/Prof-Brian-Leftow.htm, accessed on 12.06.2022].
Murphy, M.C. Is an Absolutely Perfect Being Morally Perfect? In Paul Draper, ed., Current Controversies in Philosophy of Religion. New York: Routledge, 2019, pp. 93-108.
Murphy, M.C. Divine Holiness and Divine Action. Oxford: Oxford University Press, 2021. 288 p.
Murphy, M.C. God's Own Ethics: Norms of Divine Agency and the Argument from Evil. Oxford: Oxford University Press, 2017. 224 p.
Nagel, T. The Possibility of Altruism. Oxford: Clarendon Press, 1970. 148 p.
Pruss, A. The Leibnizian Cosmological Argument. In William Lane Craig and J.P. Moreland (eds.), The Blackwell Companion to Natural Theology. Malden: Blackwell, 2009, pp. 24-100.
Scotus, J.D. Philosophical Writings, translated by Allan Wolter. Edinburgh: Nelson, 1962. 396 p.
Smith, M. The Moral Problem. Malden: Blackwell, 1994. 226 p.
Swinburne, R. The Coherence of Theism, second edition. Oxford: Oxford University Press, 2016. 320 p.
van Inwagen, P. Introduction. In God, Knowledge, and Mystery: Essays in Philosophical Theology. Ithaca: Cornell University Press, 1995, pp. 11-21.
van Inwagen, P. Ouam Dilecta. In The Possibility of Resurrection and Other Essays in Christian Apologetics. Boulder: Westview, 1998, pp. 1-27.
