The wrong anti-God

The anti-God that I take seriously is the malicious omnipotent omniscient being, who, it is said, creates so that creatures will suffer, because of the joy this suffering gives It. This may be contrasted with a different idea of anti-God, that of an evil being that seeks to destroy things of value out of hatred or envy. An omnipotent omniscient being would not be envious. Moreover, destructive hatred cannot motivate creation. For these two reasons I find that rather implausible. My case holds, however, against that sort anti-God as well as the malicious one.

The variety of anti-Gods alerts us to the problem of positing any character to God, whether benign, indifferent, or malicious. There are many such character traits we could hypothesize. Why not a God who creates as a jest? Or a God who loves drama? Or a God who, adapting Haldane's quip, is fond of beetles? Or, more seriously, a God who just loves creating regardless of the joy or suffering of creatures?

I conclude that a God who is first cause and the ultimate explanation of all contingent truths should be posited as having no character at all, neither a righteous one nor a wicked one, neither a benevolent one nor a malicious one.

Analogical predication

Analogical Predication is, unfortunately, a self-referential topic. There is Aristotle's usage, which, adapted to classical theism, asserts that human goodness is a sign of divine goodness. There is Aquinas's, which Nicholas Wolterstorff (2005) interprets as saying that goodness is predicated literally but analogically of God, because the ontology underlying the predication is the identity of God with God's goodness instead of the more usual characterization of the goodness of, say, Mary as one of her attributes. The goodness being there considered is, however, the possession of value, not righteousness. What I am here concerned with is the ‘as if’ sense of analogy, which I take to be the sense that Alston (1993) attributes to Aquinas. In this sense, when we predicate righteousness of God by analogy we are saying that God performs the acts that an agent of righteous character (in a consequentialist way) who was blessed with omniscience but laboured under the burden of omnipotence would perform. Moreover, it is no accident that God performs such acts: we may rely upon them. To be sure, there are problems deciding what God will do if, of necessity, there is no best act of creation.Footnote ³ But that is not my present concern, so, for ease of exposition, I shall suppose there is a best act of creation, allowing readers to substitute for that phrase, ‘act such as a righteous consequentialist of unlimited power and knowledge would perform’.

I shall argue that we should neither predicate ‘righteousness’ nor its opposite ‘wickedness’ literally of God, but that in the absence of these literal predications we should predicate ‘righteousness’ by analogy in the ‘as if’ sense. At least we should do so with the proviso that this is moral righteousness in a consequentialist fashion.

My resort to ‘as if’ analogy when considering divine goodness is not ad hoc. For ‘as if’ analogy generalizes to other divine attributes. Thus we may say by analogy that God is F if the divine acts are those we would expect of an agent, that is like a human being who is F but who is without human limitations to power and knowledge. For instance, we might well say that God is wise in this sense, without implying that divine wisdom is like human wisdom. Divine wisdom is that, whatever it is, that explains why God's acts are (in the analogical sense) wise. I would resist any more informative answer to the question ‘What explains why God's acts are wise?’ than ‘God's nature as an agent without limitations’.

There are two objections to my reliance on ‘as if’ analogy, to which I now reply. The first is that once we start analogizing it is hard to know when to stop. Is the first cause an agent, as Swinburne asserts, or is the first cause merely something that has effects as if It is an agent? If It is an agent should we say It is motivated by reasons, or merely that it is as if It has reasons? If It has reasons should we say It has a righteous character or merely, as I am proposing in this article, that it is as if It has a righteous character? In this article I do not answer the first two questions. Instead I discuss the existence of anti-God within the scope of the supposition that the first cause is an agent who acts for reasons. The justification for this way of proceeding is that: (1) New, Murphree, and Law work within these suppositions; and (2) it is a requirement for rational enquiry into the existence of God that for the purposes of this enquiry we initially work within some suppositions about the nature of (the putative) God, even if we later give these an ‘as if’ gloss.

The second objection is that medieval and subsequent Thomist theories of analogy are subtler than my ‘as if’ version. Thus they would not merely say that ‘moral righteousness’ is predicated of God because the divine effects are like that of a righteous agent, but also suggest that the effects of a righteous human being in some fashion reflect the divine nature. Hence, it could be concluded that the ‘as if’ analogical predication evacuates religious language of its connotations in a way that the more nuanced theories of analogy do not.Footnote ⁴ My reply is to repeat that I am considering the God of the philosophers, and it may well be that this God changes to become more like the object of religious devotion.

Why the first cause is literally neither virtuous nor vicious

If by righteousness we just meant tending to do what was morally right, considered in abstraction from the motives or reasons for acting, then there would be no distinction between predicating the term ‘righteous’ literally and analogically of God. So I stipulate that by righteousness I mean a character trait that is apt to result in doing what is morally good. As such it is a place-holder either for various virtues, or for all the virtues harmoniously integrated, or for some distinguished virtue such as being benevolent or being apt to love others. Likewise wickedness is either a place-holder for various vices or some distinguished vice such as being malicious. And we may note that because human beings often desire to do what is wrong it requires virtue to do what is right always or for the most part. We should not extrapolate this from the human to the divine case unless we have already posited some divine desires to do what is wrong, which I shall not do. Therefore, we should not carelessly assume that without righteousness God would do what was wrong. The crucial question will turn out to be whether God would do anything at all unless endowed with some character trait.

There are two reasons for not attributing to God any virtues or vices. Both concern simplicity, but in different senses. There is simplicity in the information-theoretic sense, required for reasonable ways of choosing between empirically equivalent theories, as in Ockham's and other razors. And there is simplicity in the ontological sense of lacking structure, as when we use the term ‘simple’ for something that has no proper parts. Now Law concedes that the hypothesis of an evil God might lack simplicity but thinks this is a fairly minor consideration. I disagree. Perhaps that is because he is considering only the information-theoretic sense of simplicity. It is simpler and so a priori more likely that God lacks any character trait than that God has one. If God has any character trait at all, then maybe we are to assume It must either be benevolence or malevolence. I don't see why we should make this assumption, but perhaps it is a widely shared intuition. So I concede it, and make no case from information-theoretic simplicity for any conclusion stronger than the following: relative to there being a God of some type, it is no more likely that God has a character trait than that God lacks one. Therefore, relative to there being a God of some type the probability of one of malevolent character is at most 25 per cent, as is the probability of one of benevolent character. The God of no character, who is neither literally righteous nor literally wicked, has a probability of at least 50 per cent. Provided I can argue that the God of no character is by analogy said to be righteous, then the traditional belief that there is a good God has a probability of at least 75 per cent relative to there being a God at all.

Ontological simplicity is a further consideration, which increases the probability of a God of no character. It is intuitive that there is an ontologically simple first cause or, if you prefer, ultimate explanation, whether it is the laws of nature or God. Ontological simplicity for laws of nature is, I guess, a matter of not being expressible as a conjunction without increasing the complexity in the information-theoretic sense.Footnote ⁵ The drive to find a unified field theory in physics attests to this intuition on behalf of naturalists. Likewise, ontological simplicity is part of the classical conception of God, and may be explicated as both denying that God has parts and as asserting that all the divine attributes are implied by just one non-conjunctive attribute. Admittedly, this is somewhat obscure, but God may be assumed simple if all divine attributes are implied by God's being an agent without qualification, where I take it that limits to the power of the first cause, whether due to ignorance or weakness, would count as qualifications. Hence a God without character is compatible with divine simplicity. To say, however, that the divine choice between possibilities is guided by some divine virtue or vice is to introduce a distinction between the divine agency and the divine character, contrary to ontological simplicity. I submit, therefore, that there is more than information-theoretic simplicity at work in denying that God is literally wicked, although I grant that God is (or was) not literally righteous either.

It might be objected that classical theists combine divine simplicity with a distinction between the divine power and the divine knowledge. So why not combine divine simplicity with a distinction between the divine agency and the divine character?Footnote ⁶ My response is that classical theists are committed to the identity of the divine power with the divine knowledge, so the only distinction they can make between them is like that between the Morning and the Evening Star, as being different ways of referring to the same thing. In the case of divine power and knowledge I find this identification plausible enough because ignorance and incapacity are both limitations on agency, so we may take the divine nature to be unlimited agency. In the case of divine agency and divine character there is no way of making the identification except by begging the question against the anti-God challenge, by insisting that being an agent without limitations somehow entails having a positive moral character.

Bringing about the good

Within the scope of the supposition that God is an agent acting for reasons, the act of creation is a choice among possibilities. I am now going to state a theory about choices, whether human or divine, which, following Leslie, I call axiarchism.Footnote ⁷ I would prefer to call it the manifest theory, but that begs the question against the anti-God theorist who may want to frame a mirror image hypothesis, antaxiarchism. I shall be arguing that our knowledge of our own nature as agents gives us good reason to prefer axiarchism to antaxiarchism.

For ease of exposition I make two simplifying assumptions. One is that we ignore moral obligations in the narrow sense and concentrate on the value of consequences of actions. The other is that there is a unique act with the best consequences. Axiarchism for both human and divine agency is then the conjunction of the following:

1. (1) Comparative evaluation and hence the judgement of which act has the best consequences is objective in the sense of being true or false independently of any mind, even God's. (Presumably it is therefore non-contingent.)

2. (2) That the consequences of act X appear to be better than the consequences of any other act being considered directly motivates the agent to perform X unless the agent suffers from akrasia. (The motivation would be indirect if it was due to a desire to do whatever appears the best.)

3. (3) Akrasia does not afflict human beings all the time, and when it occurs it can be explained by details of the human condition that should not be extrapolated to the divine condition.

The rival antaxiarchist theory replaces ‘best’ by ‘worst’. (1) is a version of moral cognitivism and in the divine case we may take God as knowing which act has the best consequences. (2) and (3) together express a version of internalism about motivation, a topic more usually discussed in the context of moral obligations (Richmond Campbell Reference Campbell and Zalta2008).

Both axiarchism and antaxiarchism combine cognitivism about values with internalism about motivation, and there is a contemporary debate over which of these is false, based on the premise that they cannot both be correct. (See Richmond Campbell Reference Campbell and Zalta2008.) That premise is itself, however, derived from naturalism. Consider, for instance, Mackie's argument that moral properties would be ‘queer’, because they would have to satisfy the internalist requirement that merely being aware of the property motivates action (Mackie Reference Mackie1977). My response is that there are neither special moral properties nor, more generally, any special evaluative properties. Instead, the burden of ‘queerness’ is born by agency. It is the idea of acting for a reason that is ‘queer’ unless somehow reduced to a neurophysiological explanation of behaviour. And axiarchism is, I submit, an obstacle to any such reduction. It follows that ‘queerness’ is merely a matter of incompatibility with naturalism.

I take it that the existence of God (including anti-God) is contrary to naturalism if by that is meant the thesis that in some fashion all truths are reduced to scientific ones. If we reject naturalism, then axiarchism has the support of our experience of our own actions when we act freely. That is why I would prefer to call it the manifest theory of action. I go further and submit that it is this experience of our own freedom that should be one of the chief reasons for rejecting naturalism. Here I follow Charles Campbell (Reference Campbell1967, ch. 2) in taking the experience of overcoming temptation as the paradigm of human freedom. We experience various desires but decide to do what we believe is for the best. A purely naturalistic account would consider this merely a case of competing desires, in which either the strongest wins or the result is random. But that is not how we experience temptation, and it is not enough for freedom.

Because theists reject naturalism and because axiarchism is based on human experience they should at least find it congenial. I say ‘at least’ because to believe in a God who is an agent acting for reasons requires: (1) that, following Swinburne, we treat agency-causation as one of the ultimate ways of explaining; and (2) that we keep our hypothesis about divine agency as economical as possible by not attributing to God desires that are independent of the divine knowledge of values (see previous section). So axiarchism is, I hold, not merely congenial to theists but a premise in an attractive argument for theism.

Granted axiarchism, we expect a God lacking in character traits not to have any preferences, except those due to knowledge of values. Hence we expect God to perform the best possible act of creation, assuming there is one. Hence ‘moral righteousness’ is applicable to God by analogy.

I anticipate the objection that antaxiarchism is a priori just as probable as axiarchism. Even if it is, we are not engaged in purely a priori reasoning. Axiarchism is supported by our own experience of our acts, especially cases of temptation. What the antaxiarchist says of human beings is contrary to our experience. Normal human beings are not motivated to bring about X because of the judgement that X is the worst possible. Even people who do very wicked things, even those we are tempted to call evil, either give in to some temptation or judge that what they are doing is good. Maybe there are some pathological cases of people who do what they judge to be the worst, but I would expect them to do so because they already have the perverse desire to do what is the worst. Hence condition (2) in the characterization of antaxiarchism is not satisfied. So they would not provide evidence for antaxiarchism. To be sure we can conceive of beings that exemplify it but I know of no actual examples.