Logical Objections to Theism
Abstract: This chapter looks at a range of objections to theism that one might class as 'logical'. Some of these objections aim to show that theism involves an internal logical contradiction. Others aim to show that theism is at least logically incompatible with other beliefs to which the theist is also typically committed. Also included are objections grounded in the thought that theism is nonsensical or meaningless. The chapter provides both an overview of this broad terrain, including a map of possible responses to different kinds of objection, and then a number of examples.
This essay is in two parts. In Part One, I map out several varieties of logical objection to theism, provide some illustrations, and then set out a number of response strategies that may be employed by theists in defence of their belief. In Part Two, I examine in more detail some of the best-known examples of logical objections to theism, including (i) objections associated with the doctrines of divine simplicity, divine personhood, and divine foreknowledge, (ii) the logical problem of evil, and (iii) some verificationist and falsificationist objections.
PART ONE: A MAP OF THE TERRAIN
By theism I mean belief in an omnipotent, omniscient, perfectly good being. Of course, these are not the only attributes associated with God. Many insist that God is, in some sense, simple. Many also require that God be a person - an agent who acts (in creating the universe ex nihilo, for example) and with whom one might enter into a personal relationship.
There are a variety of arguments that I think might reasonably be included under the title 'logical objections to theism'. I will categorise them as internal, external, and nonsense objections.
If someone claims to have discovered in the rainforests of Brazil a triangle possessing not three but four sides, mathematicians won't mount an expensive expedition to confirm whether or not that object exists. We can know, from the comfort of our armchairs, that there is no such thing. A triangle is, by definition, three-sided. To assert that there exists a triangle that is not three-sided is therefore to involve yourself in a straightforward logical contradiction, and contradictions cannot be true. It appears we can similarly know, from the comfort of our armchairs - just by unpacking the concepts involved - that neither are there are any married bachelors or female stallions out there waiting to be discovered.
Internal objections to theism turn on the thought that we can similarly rule out, from the comfort of our armchairs, the truth of theism. The suggestion is that theistic belief involves a logical contradiction. Thus we can know - by means of a priori, armchair reflection alone - that theism is false.
Internal objections to theism come in two varieties: those that maintain that theism requires the instantiation of particular divine attributes that are logically impossible, and those that maintain theism requires the co-instantiation of divine attributes that cannot logically be combined.
A familiar alleged example of the former sort of inconsistency is what I shall term the Riddle of the Stone objection to God's omnipotence. That objection runs as follows. If God is omnipotent then he can bring about any state of affairs. But if God can bring about any state of affairs, then he can create a stone so heavy that even he cannot lift it. But then it follows that, having created such a stone, God can't do everything: he can't lift it. Thus, some are quick to conclude, the very idea of an omnipotent being involves a logical contradiction: God's omnipotence logically entails his non-omnipotence. Thus we can rule out an omnipotent God from the comfort of our armchairs, by means of logic alone.
The second variety of internal objection maintains, not that any particular divine attribute involves or generates a logical contradiction by itself, but rather that certain divine attributes in combination do so. An example is the supposed logical incompatibility of divine simplicity and divine personhood. For those who define God as being both simple and a person (and not all theists do), paradoxes such as the following arise. The concept of simplicity seems to require that God be non-temporal (for otherwise, being extended in time, God would have temporal parts). The concept of personhood, on the other hand, appears to require that persons be temporal (for a person must be capable of possessing psychological states, such as beliefs and desires, which must have temporal duration; moreover, persons can perform actions, which require a temporal framework within which to take place). But the claim that there exists a being that is both temporal and non-temporal involves a straightforward contradiction. Thus there's no such simple person, and so no God, so defined.
I turn now to externalist objections. The objection is not that there is some internal logical inconsistency involved in theistic belief per se, but rather that theistic belief is logically incompatible with certain other belief or beliefs the theist holds. Probably the best-known external objection is the logical problem of evil.
Theists typically accept the existence of at least some evil. The logical problem of evil involves the thought that the theist's belief in the existence of evil is logically incompatible with their belief in God. It's argued that a God, being omniscient, would know of the existence of such evil, and, being omnipotent, would have the power to prevent such evil. Moreover, being perfectly good, God would not want evil to exist. Thus, if evil exists, then God does not. So theists who also believe evil exists might seem to be caught a straightforward contradiction.
A third kind of objection to theism that I include under the umbrella of 'logical' objections are those that maintain that God talk is not false, but meaningless or nonsensical. Note that both internal and external objections conclude that theism is false. But what if theism fails in a more radical way, by failing to get even as far as asserting something capable of being true or false? That is what the proponents of nonsense objections maintain.
Why might we conclude that God talk is neither true, nor false, but meaningless? Most obviously, because we find it fails to satisfy our preferred criterion of meaningfulness. The criterion best-known for having been applied in this way is the Verification Principle (closely associated with A.J. Ayer and the logical positivists). Here's a simple version:
(VP) A statement is meaningful if and only if it is verifiable.
Under what circumstances is a statement verifiable? According to Ayer in his Language, Truth and Logic, there are just two ways a statement might be verified.
First, the statement might be analytic (roughly: true in virtue of meaning, in the way e.g. bachelors are unmarried, and triangles have three sides, are supposed to be). Such statements, being true in virtue of meaning, can potentially be verified a priori, by reflection alone, from the comfort of our armchairs.
Second, Ayer says a statement is verifiable if there are possible observations 'relevant to determining that statement's truth or falsehood' (2001: 38).
Note that verification, in Ayer's intended sense, is a fairly weak notion. There is no suggestion that for a statement to be verified its truth must be establishable conclusively. It's sufficient, thinks Ayer, that observation might supply grounds either for supposing the statement is true or for supposing it is false.
Scientific statements, such as that all actions are accompanied by equal and opposite reactions, all emeralds are green, and so on, are not analytic. Nevertheless, that statement about emeralds, while incapable of being established conclusively (there's always the risk of a future non-green emerald cropping up), can at least be supported by observational evidence. The statement can also be disconfirmed observationally (by an observation of a non-green emerald). Hence, on Ayer's view, though such scientific claims aren't analytic, they are verifiable, and so are meaningful.
However, Ayer famously goes on to rule out a great deal of talk as meaningless, including moral talk. On Ayer's view, moral statements, being unverifiable, are literally meaningless, and thus incapable of being true or false. Ayer also maintains that the statement that God exists is unverifiable; it is, he concludes, meaningless.
Before we look in more detail at various examples of these three varieties of logical objection - internal, external, and nonsense - I want briefly to map out some possible response strategies. Theists don't typically shrug, admit their mistake, and wander off converts to atheism when presented with such objections. They have a developed a range of strategies designed to deal with such criticisms. I briefly set out below the general form such manoeuvres take so we will then be able more easily to navigate our way through further examples.
Responses to internal objections
When presented with an internal objection, theists may:
(i) maintain that the alleged logical contradiction is merely apparent,
(ii) drop the divine attribute(s) causing the problem.
There are two ways in which theists might adopt strategy (i). First, they might adopt (i) by (a) attempting to show that there's some error in the reasoning of those who claim there is an internal contradiction. The theist may insist the critic has slipped up, logically speaking, and concluded there's a contradiction where in truth there is none.
Alternatively, a theist might adopt (i) by insisting that (b) while, if what the critic took the theist to mean were correct, then the theist would indeed be guilty of contradicting themselves, the critic has misunderstood. The critic's logic may be faultless, but they have nevertheless failed properly to understand what the theist means by at least one of the terms involved. Suppose I say, 'There's a bank nearby, but there's no bank nearby'. I appear to have contradicted myself. However, if it turns out that by my first use of 'bank' I meant riverbank, and by the second, financial bank, then the contradiction is revealed to be merely apparent. Theists may similarly suggest that what they mean by terms like 'simple' or 'omnipotent' as applied to God is not what the critic assumes. Once what the theist means by such terms has been clarified, any supposed internal contradiction can be shown to be merely apparent.
Here's an illustration of strategy (i)(b). Some theists respond to the Riddle of the Stone by claiming, first, that, in saying God is 'omnipotent', they don't mean God that can do the logically impossible. God can do anything possible, but making a four-sided triangle, or a stone so heavy that even an omnipotent being could not lift it, is not a logical possibility.
But why suppose that God's inability to bring about such logically contradictory states of affairs entails he lacks omnipotence? One suggestion would be to say that logical impossibilities are no real limitation on God's power, for it is not as if there is some possible state of affairs - the existence of a four-sided triangle, say - that God is somehow prevented from realizing. Rather, the expression 'There's a four sided triangle' fails to pick out any possible state of affairs.
So, some may conclude, when properly understood, the claim that God is omnipotent generates no contradiction.
Alternatively, a theist confronted with the Riddle of the Stone might adopt option (ii) and drop omnipotence from the list of divine attributes. That provides a straightforward solution, though it's a far less popular move.
Responses to external objections
When presented with external objections, theists have a further option. As with internal objections, they may:
(i) maintain that the alleged logical contradiction is merely apparent,
(ii) drop the divine attribute(s) causing the problem,
However, they may also:
(iii) drop the external belief(s) causing the problem.
Consider, for example, the logical problem of evil outlined earlier. Theists might respond by: (i) insisting that the supposed contradiction involved in believing that there exists both God and evil is merely apparent, or (ii) dropping one of the divine attributes causing the problem - the logical problem can be neatly solved by dropping any one the three omni-attributes (so, for example, the theist might drop omniscience, maintaining that God is indeed omnipotent and omnibenevolent, but is ignorant of the evil it would be in his power to remove.). However, theists also have option (iii). It is the addition of the theist's external belief in the existence of evil that generates the logical problem of evil. So the theist might avoid contradiction by dropping, not their theism, but rather their belief that evil exists. For example, a theist might respond to the logical problem of evil by suggesting that, though there might appear to be evils, the evils are in fact illusory.
Pseudo-profundity and embracing contradiction
Notice that there is a further way in which theists may respond to both internal and external objections to theism. They may choose to embrace - and perhaps even make a virtue of - the supposed contradiction.
If you want to appear profound, there are several fairly tried and trusted recipes you can follow. One of the most effective is to contradict yourself regarding one of life's big themes, such as life and death, meaning and purpose, war and peace, and so on. Here are three examples I made up:
Sanity is a kind of madness
Life is often a kind of death
The ordinary is extraordinary
Such sentences are interpretable in all sorts of ways and can easily appear profound. In George Orwell’s novel Nineteen Eighty-Four, two of the three slogans of the Party have a similarly contradictory character:
War is peace
Freedom is slavery
Ignorance is strength
Life is often a kind of death
The ordinary is extraordinary
Such sentences are interpretable in all sorts of ways and can easily appear profound. In George Orwell’s novel Nineteen Eighty-Four, two of the three slogans of the Party have a similarly contradictory character:
War is peace
Freedom is slavery
Ignorance is strength
If you are an aspiring guru, the attraction of making such contradictory remarks is two-fold. First, they make your audience do the intellectual labour for you. You can sit back, adopt a sage-like expression, and let your disciples figure out what you mean. Secondly, such remarks are interpretable in numerous ways. This gives you enormous wiggle room if someone dares challenge you (for you can imply your critic is a crude, overly-literal thinker who has failed properly to grasp the true meaning of your remark).
The thought that contradiction is a sign of profundity often crops in religious contexts. Non-believers usually take what look like straightforward contradictions within a religious doctrine to indicate falsehood. The faithful, on the other hand, may take those very same contradictions to indicate genuine insight. Indeed, religious folk will sometimes make a point of appearing to contradict themselves, saying things like 'God is, and yet he is not', 'God is one, and yet he is many', and 'God is good, and yet he is not.'
There's no denying that seemingly contradictory remarks can sometimes express something profound. No doubt we can find some kind of truth even in Orwell's poisonous examples. However, given the formulaic way in which contradiction can be used to generate the illusion of depth and profundity - that's to say, to generate pseudo-profundity - it's wise not to be too easily impressed.
Responses to nonsense objections
Responses to nonsense objections include:
(i) rejecting the criterion of nonsense/meaningfulness on which theism comes out as nonsense/meaningless,
(ii) maintaining that, whether or not the proposed criterion of nonsense/meaninglessness is correct, theism meets it.
In response to Ayer's verificationist challenge to the meaningfulness of God talk, then, theists may challenge the principle of verification on which Ayer relies, and/or insist that theism is in fact verifiable in the required sense. As we will see, both kinds of responses to Ayer's challenge have been made.
PART TWO: SPECIFIC EXAMPLES
As promised, I now turn to some examples of logical objections to theism. What follows is merely a selection. There are many more such objections.
God is widely supposed to be, in a certain sense, simple. The doctrine of divine simplicity is characterised neatly by Eleanor Stump (1997: 250) as involving four claims:
1. God cannot possess spatial or temporal parts.
2. God cannot have any intrinsic accidental properties.
3. There cannot be any real distinction between one essential property and another in God's nature.
4. There cannot be a real distinction between essence and existence in God.
The first condition is straightforward, and rules out God being extended in space in the way that, for example, physical objects are. God's simplicity is also widely supposed to require that God be eternal rather than temporal. If God were spread out across time, then he would have temporal parts, with one part occurring before another.
The second claim involves a familiar philosophical distinction between essential properties - roughly, those an entity must possess - and accidental properties - those an entity possesses but might have lacked or might come to lack. Physical objects are widely thought to have both essential and accidental properties. For example, it's widely supposed that it's essential to this table that it be made of wood - a table not made of wood would not be this very table. However, it's not essential to this table that it be painted red, or be in my living room - these are merely accidental properties of the table. Some properties of my table are also merely extrinsic - they can be changed without a change in the object (being in my living room is an extrinsic property of this table, it can be relocated without any change to it). Other properties are intrinsic - a change in the property involves a change in the object (a change in the length of one table leg would involve an intrinsic change in the table). The doctrine of divine simplicity requires that all God's intrinsic properties be essential to him. So, while it may not be essential to God that he possess the extrinsic property of currently being thought about by me (presumably, God might easily have lacked that particular property), the property of omnipotence, being intrinsic to God, is essential to him.
The third claim requires that all God's essential properties be identical. So, for example, God's essential properties of omnipotence, omniscience and perfect goodness must, in truth, be one and the same property: a single property that, say, we are merely conceptualising or thinking about in different ways.
The fourth claim exploits another philosophical distinction - between essence and existence. The essential properties of thing typically do not include existence. For example, this table's essential properties do not include existence - the table might not have existed and it will one day cease to exist. God's essence, on the other hand, includes existence. Indeed, given the third claim, his property of existence must be identical with his other essential properties - of omnipotence, omniscience, perfect goodness, and so on. God's essence is existence.
The doctrine of divine simplicity might be thought logically inconsistent with other doctrines regarding God, such as the doctrine of the Trinity. How can God be both a simple being and yet triune? And how can God, who lacks spatial and temporal parts, be Jesus, who had spatial and temporal parts (arms and legs, and a birth and death, for example)?
Further, the claim that God is simple has also been accused of being logically contradictory per se. One of the more obvious objections runs as follows: surely, we can logically separate out God's properties of omniscience, omnipotence and omni-benevolence from him as their logical possessor. But then God is himself logically and ontologically complex, not simple.
However, there is a long theistic tradition that insists that God's omni-properties are not, in fact, logically distinct properties of him, in the way that, say, my height and my weight are logically distinct properties of me. God does not merely possess his omni-properties, he is numerically identical with them. God and his omnipotence are one and the same thing; God and his omniscience are one and the same thing, and so on. But then and, given the transitivity of identity (if a is identical with b and b with c, then a is identical with c) then all these omni- properties are also identical with each other. Hence God is, after all, logically simple.
Perhaps more problematic is the following external objection: God, while simple, is also is widely supposed by theists to share at least some properties with his creation. For example, God possesses knowledge and power, but then so do I (even if not to the maximum, as God does). The difficulty is: if power is a property that God and I share, then surely God cannot be identical with that property. And so God's possession of that property - power - does entail that he have some logical complexity after all: we can logically distinguish God from some property of his. It might seem that the only way we can salvage the doctrine of divine simplicity is by denying any commonality between God and his creation.
One response to this external objection is to insist that while God is indeed perfectly powerful, and I too am powerful to some limited degree, it doesn't follow that there is, then, some property - power - that we share. Miller (1996) argues that perfect power is not power. Perfect power is a kind of limit, as is zero power. And zero power is not power. Similarly, the lower limit in the case of speed - zero mph - is not a speed. Miller suggests that perfect power should not be thought of as maximum amount of power (what he calls a 'limit simpliciter') but as a 'limit case' like zero power or zero speed.
Graham Oppy (2003) outlines a different response to this external objection, suggesting that in correctly describing individuals a and b as being F, it does not follow that there exists some single corresponding property in the world joint possession of which by a and b makes both 'a is F' and 'b is F' true. For the predicate 'is F' may not pick out an objectively existing property. Suppose, for example, that 'is F' is defined like so: something is F if and only if it possesses either property G or property H. Suppose both a and b are F. It doesn't follow there is one property they share - for it may be that a is F by virtue of being G and b is F by virtue of being H. But then similarly, while both God and I are powerful, God may be powerful by virtue of his possessing (or being identical with) a property P1, while I am powerful by virtue of my possessing some other property P2. In which case, while we can both be correctly described as 'powerful', there need be no property we share.
That concludes my brief sampling of the many internal and external logical objections that might be raised in connection with the doctrine of divine simplicity. But note there is also the potential for a form of nonsense objection to be raised against the doctrine. For example, in his Philosophical Investigations (1998) Ludwig Wittgenstein attacks the notion of absolute logical simplicity, which played a crucial role in his earlier philosophy (Book I, sections 45-48). Our talk of what is 'simple' and 'composite' has its home in settings in which we describe, for example, a chessboard as a complex made up of squares. But is each chess square simple? Within the context of the game of chess, perhaps. However, in other contexts each square might be thought of as made up of two rectangles, or of a larger shape from which another has been subtracted. A chess square's cream colour might also be seen as a composite of yellow and white. Each chess square is bounded by four straight lines. And each of those lines might in turn be viewed as a combination of mathematical points. Conversely, a mathematical point might be seen as the intersection of two lines. Talk of 'simple' and 'composite' is highly diverse and has its home in such varied linguistic contexts. Wittgenstein thought that to abstract away from all such linguistic contexts or 'language games' and try to apply the terms 'simple' and 'composite' in an absolute way - to talk about a thing or things that are simple, period - is to end up talking nonsense:
But what are the simple constituent parts of which reality is composed? -- What are the simple constituent parts of a chair? -- The bits of wood of which it is made? Or the molecules, or the atoms? -- 'Simple' means: not composite. And here the point is: in what sense 'composite'? It makes no sense at all to speak absolutely of the 'simple parts of a chair'. (Philosophical Investigations, I, section 47)
It's arguable that the attempt to define God as being absolutely 'simple' involves a similar drift into nonsense.
The classic problem of divine foreknowledge is an external objection. The objection is that the divine attribute of omniscience is logically incompatible with the theist's further belief that God has imbued human beings with free will - that's to say: the freedom to act as they freely choose, without their action being compelled or determined by anything outside themselves. The Jewish philosopher Maimonides (1135-1204) produced a classic statement of the objection:
Does God know or does He not know that a certain individual will be good or bad? If thou sayest 'He knows', then it necessarily follows that [that] man is compelled to act as God knew beforehand he would act, otherwise God's knowledge would be imperfect…(1996: 99-100)
The suggestion seems to be that, if God knows Tom will do x tomorrow, then necessarily, Tom will do x tomorrow. But then Tom lacks the freedom to do otherwise. Whether or not Tom performs some future action x is never up to him - it is determined by factors outside himself: by what God knows he will do.
There is also an internal objection to theism generated by divine foreknowledge. If God, being omniscient, knew yesterday that he will do x tomorrow, then God can't do anything other than x tomorrow. But if God can't do other than x tomorrow, then he is not omnipotent: his power is limited by his own foreknowledge.
But have we really identified a problem regarding divine foreknowledge? After all, we also have (admittedly fallible) knowledge of what will happen in the future. I might know, for example, that Ted will go to the shops tomorrow. But it is a necessary condition of knowing that P that P is true. Hence my knowing Ted will go to the shops tomorrow entails that Ted will indeed go to the shops. So does my knowledge of what Ted will do tomorrow entail he lacks the freedom to do otherwise?
Actually, this problem of foreknowledge has a solution. What is necessary is the conditional (if-then) statement that if I know that P, then P. It does not follow that if I know that P, then necessarily P, i.e. that it is a necessary truth that P, that things could not have been otherwise. More generally:
Necessarily: If x, then y
does not entail
If x, then necessarily: y.
Compare: necessarily, if Tom is a bachelor, then Tom is unmarried. That conditional is a necessary truth. But of course it does not follow that if Tom is a bachelor, then it's a necessary truth that Tom is unmarried - that Tom lacks the freedom to get married.
Similarly then, while it's a necessary truth that if I know Ted will go to the shops tomorrow, then Ted will go to the shops tomorrow, it doesn't follow that my knowledge of what Ted will do entails Ted lacks the freedom to do otherwise.
So have we solved the problem of divine foreknowledge? If my knowing that Ted will go to the shops tomorrow is consistent with Tom's having the freedom to do otherwise, why shouldn't God's knowing what Ted will do tomorrow be consistent with Ted's freedom to do otherwise?
If there is still a problem regarding divine foreknowledge, it seems that will be because there's something special about God's foreknowledge. Of course there are differences between human foreknowledge and God's foreknowledge. In particular, unlike us, God is infallible about what will happen in the future. So does God's infallibility entail that, if God knows Ted will go to the shops tomorrow, then necessarily Ted will go to the shops tomorrow - that Ted lacks the freedom to do otherwise?
It seems not. God's infallibility requires:
Necessarily: if God believes that P, then P.
So, necessarily: if God infallibly believes Ted will go to the shops tomorrow, then Ted will go to the shops tomorrow. However, it does not follow that if God infallibly believes Ted will go to the shops tomorrow, then necessarily Ted will go to the shops - that Ted lacks the freedom to do otherwise. Again, that inference involves an illicit slide from: Necessarily: If x, then y, to: If x, then necessarily: y.
Of course, if Ted knows today that God believes he, Ted, will go to the shops tomorrow, then Ted might, given his freewill, choose not to go to the shops and so render God fallible. Now obviously, given that necessarily, God is infallible, Ted must lack the ability to make God fallible. But does this in turn require that Ted lack freewill?
No. Whenever Ted thinks he knows what God believes Ted will do tomorrow, and Ted acts to make God's belief false, it turns out Ted's just mistaken about what God believed Ted would freely choose to do.
However, while God's infallible belief that Ted will go to the shops tomorrow does not seem to be incompatible with Ted having the freedom to do otherwise, perhaps, if we also add into the mix (i) the suggestion that God knew yesterday what Ted will do tomorrow, and (ii) a further necessity - that the past is unalterable - then a successful argument that divine foreknowledge is incompatible with free will might constructed. Some more sophisticated versions of the problem of divine foreknowledge take this form. Here's one example (let T be: Ted will go to the shops tomorrow):
1. Yesterday, God infallibly believed P
2. If E occurred in the past it is now necessary that E occurred then.
3. Thus it is now necessary that God believed T.
4. Necessarily, if God believed T, then T.
5. If p is now necessary, and necessarily (if p then q) then q is now necessary.
6. Thus it is now necessary that T.
7. If it is now necessary that T, then Ted cannot do other than go to the shops tomorrow.
8. Thus, Ted cannot do otherwise than go to the shops tomorrow.
9. If you cannot do otherwise when you act, you do not act freely.
10. Thus when Ted goes to the shops tomorrow, he will not do so freely.
It remains contentious whether any such argument is cogent. Rather than explore in yet more detail the precise form the problem of divine foreknowledge might take, let's now look briefly at perhaps the best-known response to the problem: that of Boëthius.
Boëthius (1999: Book V) attempts to deal with the problem of foreknowledge by denying the first premise of the argument sketched above. That's to say, Boëthius would deny that God infallibly knew yesterday that Ted will go to the shops tomorrow. This is because Boëthius considers God to be eternal - a timeless being. It's not that God lacks knowledge; rather his knowledge is not temporally locatable in the way ours is. Being a timeless being, God does not know anything at a time. Rather, he knows timelessly.
Of course, the view that God is a timeless being is controversial, and indeed we will see that some theists consider the claim that God is eternal to be incompatible with the claim that he is a person, for example (see Divine Personhood below). Boëthius's solution is obviously not available to those who believe both that God is a person and that God's personhood requires that he be a temporal being.
Not only is Boëthius's solution unavailable to a certain sort of theist, it appears his solution in any case fails. The original objection can easily be tweaked to deal with it. For suppose Boëthius is correct and God only timelessly infallibly knows that T. Still, it is presumably now necessary that God timelessly infallibly knows T. But then it still follows that it is now necessary that Ted will go to the shops tomorrow. But then Ted still lacks the freedom to do otherwise.
The thought that God is a person is the source of various internal objections to theism.
We noted in Part One that the suggestion that God is a person would seem to require that God possess beliefs and desires on the basis of which he may rationally act. However, as we saw above, the doctrine of divine simplicity seems to require that God be eternal - be a timeless being. And now an objection looms. Arguably, psychological states like belief and desire necessarily have temporal duration, and thus require a temporal setting. But then the claim that there exists an eternal personal God appears to generate a contradiction: God must be timeless, but not timeless.
There are various further internal objections relating to divine personhood and eternity - for example: (i) that being a person involves being capable of change, but change in turn requires that the being in question be temporal, (ii) that the possibility of performing actions (arguably a condition of personhood) requires that God be located in time, for actions require a temporal setting. (Pike, for example, argues that 'a timeless individual could not produce, create, or bring about an object, circumstance, or state of affairs' since that would involve temporally locating the agent's action (1970: 110).), (iii) that being a person requires having a mind, and 'the quickest and most direct way of showing the absurdity of a timeless mind is as follows: A mind is conscious, and consciousness is a temporally elongated process.' (Gale 1991: 52).
There are also nonsense objections that focus on the combination of divine personhood and eternity, on which the suggestion that God is a timeless, unchanging being who is also an agent capable of performing actions is held to involve, not a contradiction, but a drift into incoherence. Bede Rundle for example, says: 'I can get no grip on the idea of an agent doing something where the doing, the bringing about, is not an episode in time, something involving a changing agent.' (2004: 77)
Some theists, in response to these kinds of objection, have dropped one or other of the divine attributes in question. You might drop timelessness in order to maintain attributes required for personhood. For example, Nicholas Wolterstorff drops timelessness in order to accommodate the Biblical thought that, among things, God, as a person, planned certain things and remembers certain things: God's 'planning must occur before the planned event occurs. For otherwise it is not a case of planning'. (2007: 164). Alternatively, a theist might drop personhood in order to maintain timelessness.
An alternative response, other than dropping one of the divine attributes of personhood and timelessness, would be to maintain that 'person', 'action', and so on, when applied to God, mean something other than what these terms usually mean when applied to human beings. One possibility here is to insist that, when used to describe God, such language should be understood not literally but analogically. God is not literally a person, at least not in the usual sense of the term, but he is, in certain crucial respects, person-like, etc.
And of course there is also the possibility of tackling such objections head-on and maintaining that God can be literally both a person, with all that that entails, and also timeless, without contradiction. For example, with regards to the problem generated by the thought that God, as a person, must have beliefs and desires, and that these in turn require a temporal setting, some have responded by trying to show that what is essential so far as the holding of beliefs is concerned is that the being having duration, there being a kind of duration that is non-temporal. The view that God's knowledge and belief involves a form of non-temporal duration is taken by Eleanor Stump and Norman Kretzmann (1981: see esp. 446)
Omnipotence - The Riddle of The Stone, and other problems
There are a variety of internal objections regarding the omnipotence of God including, as we have seen, the Riddle of the Stone. Other internal objections include:
1. God, if omnipotent, has the power to bring it about that he is not himself omnipotent. But, being necessarily omnipotent, God must lack that power. Therefore, there exists no necessarily omnipotent God.
2. God, if omnipotent, has the power to bring about evil. But God, being essentially morally perfect, cannot bring about evil. Therefore, God does not exist.
3.God, if omnipotent, can bring about his own non-existence. But God, being a necessary being, cannot do this. Therefore there is no God.
4. God, if omnipotent, possesses the power to bring it about that another omnipotent being exists. But there cannot be more than one omnipotent being (the existence of one omnipotent being limits the power of all other beings - for example, if God, being omnipotent, can bring it about that I sneeze now, there can't be another omnipotent being able to prevent me sneezing now). But a being that lacks the power to bring it about that another omnipotent exists is not omnipotent. Therefore, there exists no such being.
There are also external objections to theism including:
5. Omnipotence and free will. Most theists believe in what's sometimes called libertarian free will. That's to say, they believe individuals can perform free actions, these being actions that are not caused or determined by anything outside of that individual, including God. An omnipotent God would have the power to bring about our freely choosing one thing rather than another. However no being has the power to bring about our freely choosing one thing rather than another. Thus there is no God.
6. Omnipotence and the unalterability of the past. Most theists accept that the past is unalterable. God might have prevented life emerging on Planet Earth, and he might now snuff that life out, but God can't bring it about that life never emerged if it already has. But if no being can bring about a different past, then there is no omnipotent being, and so no God.
7. Similarly, given the unalterability of the past, God can't bring it about that Donald Trump is eating cake for the first time given that Trump has already eaten cake. But if no being can bring that about, then there is no omnipotent being, and thus no God.
Note, by the way, that not only does God lack the power to bring about Ted's freely chooses to be kind (objection 5 above), Ted himself does possess that power. Ted also has the power to bring about his own non-existence. So it appears there are beings, such as Ted, that possess powers God lacks.
How do theists respond to these objections?
Most of us, unversed in philosophy and theology, would probably say that omnipotence consists in something like the power to do anything, as Peter Geach notes: 'The English word omnipotent would ordinarily be taken to imply ability to do everything.' (1973)
A few philosophers have been prepared to accept that God can indeed bring about anything - including the impossible. Descartes, for example, suggests God might create a square circle. However, most theistic philosophers have understood 'omnipotence' in a more restricted way.
So how should omnipotence be understood? It appears all of the above objections can be dealt with by defining omnipotence as maximal power to bring it about states of affairs - where this is in turn understood as an amount of power that it is impossible for any being to exceed (see Hoffman & Rosenkrantz (2010)).
Notice that this definition of omnipotence immediately deals with a wide variety of internal objections. Yes, God lacks the power to bring about the existence of a square circle, or another omnipotent being (so that two omnipotent beings exist simultaneously), or a different past. However, no possible being possesses such powers - because these are powers to bring about the impossible. God's lacking such power does not entail that he's not omnipotent.
Nor does God's omnipotence, defined as maximal power, require that he must possess the power to bring about states of affairs that some other being cannot. God possesses maximal power just in case his total amount of power exceeds that of any other possible being. However, that is compatible with his lacking specific powers others possess. While the Grand High Wizard's powers may vastly exceed those of any of his minions, his minions might still possess powers he lacks (perhaps they can cook, while he can't). Similarly then, God may be omnipotent, thus defined, even if it's true that God lacks powers possessed by Ted, including the power self-destruct.
What of the Riddle of the Stone? Hoffman and Rosenkrantz point out that, if God is essentially omnipotent, then his creating a stone so heavy he cannot lift it is impossible, because his omnipotence requires that there be no stone so heavy he cannot lift it. (However, a being that is not essentially omnipotent could create a stone so heavy that they could not lift it. The being would be omnipotent when creating the stone but, having created it, would no longer be omnipotent.) Again, on their understanding of omnipotence, an omnipotent God is not required to possess the power to bring about what is impossible.
So the definition of omnipotence as maximal power (as explained above) does appear to deal with the objections sketched out above. However, objections remain. For example, Sobel (2004: chapter IX, esp. 362.) argues that no being can be essentially omnipotent.
Logical Problem of Evil
As noted above the logical problem of evil is an external objection. Theists are usually committed to two claims: the claim that God (defined as omnipotent, omniscient, and perfectly good) exists, and a claim about the world - that it contains evil. The suggestion is these two claims are logically inconsistent.
Note that talk of 'evil', in this context, is of two varieties of evil - moral evils: these being the morally bad things that persons do of the own free will, and natural evils, these being the natural diseases, disasters, and so on that befall the sentient inhabitants of this planet and cause them great suffering. The suggestion is that the existence of any evil at all - be it of either variety - logically entails the non-existence of God. Thus, according to Mackie, belief in the existence of both God and evil is 'positively irrational' (1955: 200)
Alvin Plantinga's 'Free Will defence' (1974: chapter 9) provides the best-known response to the logical problem of evil.
Plantinga's response utilizes talk of 'possible worlds', a possible world being, roughly, a way reality might have been (the actual world being a possible world). Although I exist at the actual world there is a possible world in which I don't exist (in which my parents never met, for example). There are, presumably, possible worlds in which the laws of nature are different, and even worlds in which the universe does not exist (given the universe might not have existed). Necessary truths, on the other hand, including all mathematical and logical truths, are true with respect to every possible world. There's no possible world at which 2 + 2 fails to equal 4.
Plantinga's approach to the logical problem turns on the thought that God will wish to create or actualize a possible world containing free agents - agents that are not antecedently caused or determined to do what they do but are capable of freely choosing whether to do good or evil. For only such a world can contain moral goodness.
However, if God creates a world containing free agents, then he can't, as it were, compel its inhabitants to do good. If the inhabitants are to be free, then they must be free to do evil. Given this freedom, there are possible worlds in which those free agents always choose to do good, not evil.
But then why can't God just actualize such a world? Why, in particular, can't God choose to actualize one of the possible worlds in which there is a free agent, but as a matter of fact that agent freely chooses always to do good, never evil, and so moral goodness, but not moral evil, exists? Plantinga now argues it is possible God lacks this freedom.
Because God can't compel the free inhabitants of any world he has created not to do evil, it may be that in any possible world containing free agents that God might choose to create, some evil exists (i.e. evil that created by those free agents acting freely).
Why suppose it's possible that any world God might create that contains a free agent will also be a world in which some evil exists? Well it is possible, argues Plantinga, that free agents suffer from what he calls 'transworld depravity'. It may be that no matter what circumstances God might put a given free agent in, that agent will always take at least one wrong action, and so some evil will exist.
Of course, Plantinga doesn't claim to have shown that free agents do suffer from transworld depravity. However, he points out it is possible that they do - it's possible that in any world actualised by God in which a free agent exists, some evil will also exist. But if that is possible, then there is no logical inconsistency involved in supposing God has created a world containing some evil. It may be that some evil is the price God must inevitably pay to allow for moral goodness.
Mackie is unconvinced by Plantinga's treatment of the logical problem of evil. He argues that there is a plausible position on free will - compatibilism - on which free actions may nevertheless be causally determined. In which case, God can choose to create a world in which free agents and moral goodness exist, but no evil exists because those free agents are caused always to do the right thing. Plantinga responds by rejecting compatibilist accounts of free will.
There is a fairly broad consensus that the logical problem of evil has now largely been dealt with by Plantinga and subsequent developments. Sobel, however, (2004) disputes this.
The hiddenness argument
For many, God is largely hidden. We fail to find good evidence for his existence. Nor does God directly reveal himself to us (via a religious experience, say). Consequently, given this the degree of 'divine hiddenness', many of us fail to believe. But surely God, if he exists, would want us to believe in him?
An argument against theism based on divine hiddenness has been developed articulated in some detail by J.L. Schellenberg. Schellenberg's original argument in his book Divine Hiddenness and Human Reason (1993) might be summarised as follows:
(1) There are people who are capable of relating personally to a perfectly loving God but who, through no fault of their own, fail to believe.
(2) If there is a perfectly loving God, then there are no such people.
(3) Therefore, there is no such God.
Because he maintains that God would be perfectly loving, Schellenberg concludes that God does not exist. Note that premise (1) concerns the existence of what might be called inculpable or non-resistant non-believers: individuals who do not willfully aim to shut the door on any relationship with God that might be on offer, but who nevertheless fail to believe. There appear to be such people. Indeed, not only do there appear to be non-believers who aren't resistant to belief, it seems many actually want to believe.
Why suppose premise (2) is true? Schellenberg argues that a perfectly loving God will want to enter into a personal relationship with each of us. However, we can only enter into such a relationship if we believe God exists. Hence God, if he exists, will ensure that each of us believes in him (unless, of course, we are resistant). It is only if we believe God exists that the possibility of our entering into a relationship with God opens up to us.
Is this an external logical objection to theism? It can be. To qualify as an external logical objection, as I have defined it, the objection must be that the theist's belief in the existence of non-resistant non-believers is logically incompatible with their theism. Now, as I note below, not all theists accept that there are non-resistant non-believers. For such theists, then, the above argument is not a 'logical objection' to their theism. Still, many theists do accept that non-resistant non-belief exists. Is this, then, a logical objection to their theism? It depends on the status of the second premise. A variety of hiddenness arguments are possible. We might argue that the existence of a perfectly loving God logically entails there will be no non-resistant non-believers. If that is the argument, then this is indeed an external logical objection (so far as it is aimed at theists who accept non-resistant non-belief exists). However, we might instead merely argue that it is improbable that a perfectly loving god would allow non-resistant non-belief. If that is the argument, then the above argument is an evidential objection to theism. Schellenberg offers both forms of argument.
Theists might respond to logical version of the hiddenness argument in a number of ways. In particular, they may: (i) refuse to accept that non-resistant non-belief exists, (ii) reject that principle that belief non-resistant non-belief is logically incompatible with belief in a perfectly loving God, or (iii) give up belief in a perfectly loving God.
As we have already noted, some theists do deny that resistant non-belief exists. Some maintain that, in a sense, non-belief does not exist. All of us know, at some level, that God exists. We merely choose to suppress this knowledge. Alternatively, theists may allow that some of us don't know that God exists, but that our lack of knowledge is due to our own resistance. Difficulties with such responses include the fact that there's good empirical evidence some of us really are non-resistantly ignorant of God's existence, including those of us who have never even encountered theistic belief, for example.
Some theists may be prepared to accept that God is not perfectly loving. However, the most popular response to the hiddenness is to reject premise (2).
Some theists maintain, for example, that God might be justified in allowing non-resistant non-belief in order to achieve certain greater goods. Michael Murray (2002), for example, argues that were God to make his existence clearer to us, so that reasonable, non-resistant non-belief was no longer possible, then we would no longer have the opportunity to develop good characters for which we are ourselves responsible.
Others (the so-called skeptical theists) maintain that whether or not we can think of a reason that would justify God in allowing non-resistant non-belief, there could still be a reason. In fact, for all we know, there is such a reason. Our mere inability to think of such a reason fails to provide us with good grounds for believing no such reason exists (this, it's suggested, would be a faulty noseeum inference - akin to arguing that if I can't see any insects in my garage, then there probably aren't any insects in my garage). In any case, given there could be such a reason, there can be no logical inconsistency in believing both that God is perfectly loving and that non-resistant non-belief exists. In response, Schellenberg reverses this logic: if Schellenberg has, as he claims, successfully shown that belief in a perfect being is incompatible with belief in non-resistant belief, then he has shown that there's no adequate reason for God to allow non-resistant non-belief (2015: 111).
Verificationist and falsificationist objections
We have already briefly discussed A.J. Ayer's verificationist objection to theism. Ayer wields his Verification Principle to try to show that the statement
is not false but meaningless. We have also seen that responses to such nonsense objections include (i) rejecting the criterion of nonsense or meaninglessness being applied (in this case, the Verification Principle), and (ii) maintaining that, whether or not the proposed criterion is correct, the statement in any case meets it.
The Verification Principle is certainly a contentious philosophical principle. A standard theistic objection is that the principle is self-undermining. For, it's claimed, the verification principle is itself neither analytic nor observationally verifiable. Thus, by its own lights, the Principle is meaningless.
However, is the Principle observationally unverifiable? Meaning and observation do appear to be linked. The meaning of our public language is unavoidably taught and learned through observational encounters (I learnt what words like 'cat', 'house' mean by hearing them used in encounters with cats and houses, for example). But then might not some empirically confirmed theory of how language acquires meaning have the Verification Principle, or something like it, as a consequence? In which case, the Principle itself might yet be empirically confirmed.
Whether or not such a case for the Verification Principle might be made, Ayer himself provides no grounds in Language Truth and Logic (2001) for supposing his Principle is true. The Principle merely functions as an assumption. So theists can justifiably point out that Ayer's case in Language Truth and Logic for supposing that 'God exists' is meaningless rests on a principle that is both highly controversial and for which he provides no argument.
One ambiguity in the Verification Principle is: who must be in a position empirically to verify the statement if it is to qualify as meaningful? An actual human being? But then the Principle seems too strong. Consider the statement:
A dinosaur walked across that very spot exactly 200 million years ago.
This, surely, is a meaningful statement. However, it may be that no actual human being will ever be in a position to verify it. Obviously, there were no humans around 200 million years ago to verify it. And, given the amount of time that has elapsed, it might be impossible for any actual humans to verify it either now or in the future.
If, on the other hand, we allow merely possible beings, human or not, to count as verifiers, then God, as a possible being, is presumably in a position to verify all sorts of things, including his own existence. In which case the statement that God exists is not unverifiable.
Further, John Hick (1966: 195) points out that while we actual humans may not currently be able to verify that God exists, it may be that we will be able to do so in the future - after we die. Thus what Hick calls an 'eschatological' verification remains a possibility.
Finally, it's by no means clear that we actual humans cannot now verify that God exists. Might not, say, arguments from design provide us with good grounds for supposing God exists by pointing to the fact that God best explains certain observed features of the universe? Alternatively, might not the evidential problem of evil provide us with good grounds for supposing that whatever, if anything, created the universe, it is not God?
If, in response it's suggested that only a direct observation of God can verify his existence (that is not a suggestion Ayer makes, by the way), then again our Principle will be too strong as surely scientists are able meaningfully to posit the existence of all sorts of things that cannot be directly observed: subatomic particles for example.
An even tougher criterion of meaningfulness says that a statement is meaningful only if it can be falsified - that's to say, only if some possible experience might give us good grounds for supposing the statement is false. Note that the Verification Principle allows two ways in which non-analytic statements might satisfy the Principle's condition of meaningfulness: the condition is met if there could be observational grounds for supposing the statement is true, or observational grounds for supposing the statement is false. The suggestion, now, is that there must be potential observational grounds for supposing the statement is false. Observational grounds for supposing the statement is true won't do. So we are setting the bar even higher than Ayer did for a statement to qualify as meaningful.
Antony Flew (Flew, Hare, and Mitchell1(964)) seems to suggest at least something like this falsificationist principle. Flew begins with John Wisdom's well-known parable of the gardener.
Once upon a time two explorers came upon a clearing in the jungle. In the clearing were growing many flowers and many weeds. One explorer says, ‘Some gardener must tend this plot’. The other disagrees, ‘There is no gardener’. So they pitch their tents and set a watch. No gardener is ever seen. ‘But perhaps he is an invisible gardener.’ So they set up a barbed-wire fence. They electrify it. They patrol with bloodhounds. (For they remember how H. G. Wells’s The Invisible Man could be both smelt and touched though he could not be seen.) But no shrieks ever suggest that some intruder has received a shock. No movements of the wire ever betray an invisible climber. The bloodhounds never give cry. Yet still the Believer is not convinced. ‘But there is a gardener, invisible, intangible, insensible to electric shocks, a gardener who has no scent and makes no sound, a gardener who comes secretly to look after the garden which he loves.’ At last the Sceptic despairs, ‘But what remains of our original assertion? Just how does what you call an invisible, intangible, eternally elusive gardener differ from an imaginary gardener or even from no gardener at all?’
In this example, suggests Flew, something that starts out as an assertion is gradually reduced, step-by-step, to something else. The original assertion suffers 'the death of a thousand qualifications' as what might potentially have counted as evidence against the assertion is discounted until in the end, nothing at all is allowed to falsify what's asserted. At that point, suggests Flew, nothing is asserted. Flew suggests that similarly, when religious people assert such things as that God loves us as a father loves his children, that God has a plan, and so on, and skeptics point out what appears to be strong evidence against such claims, the religious tend endlessly to explain that evidence away in much the same way as the defender of belief in an invisible gardener. That's to say, they endlessly qualify their assertions so as to immunise them against any empirical refutation. But then such claims similarly end up dying the death of a thousand qualifications. Ultimately, what the theist says ends up lacking any assertoric meaning.
Is Flew's criticism fair? R.M. Hare responds with is own parable.
A certain lunatic is convinced that all dons want to murder him. His friends introduce him to all the mildest and most respectable dons that they can find, and after each of them has retired, they say, ‘You see, he doesn’t really want to murder you; he spoke to you in a most cordial manner; surely you are convinced now?’ But the lunatic replies ‘Yes, but that was only his diabolical cunning; he’s really plotting against me the whole time, like the rest of them; I know it I tell you’. However many kindly dons are produced, the reaction is still the same (Flew, Hare, and Mitchell (1964)).
Hare points out there is no behaviour of the dons that this will accept as counting against his theory. Therefore, on Flew's test, this person asserts nothing. But this conclusion seems wrong: we consider such a person mistaken. We disagree with him. If he made no assertion, no disagreement would be possible.
Young Earth Creationism provides another apparent counter-example. Many Young Earth Creationists won't let anything count against their theory that the universe is only around 6,000 years old. Whatever evidence is provided against their theory (the fossil record, light from distant stars, etc.) is, in one way or another, explained away. Should we conclude, then, that when such a Young Earth Creationist says, 'the Universe is only a few thousand years old' they fail to assert anything at all? Surely not. But then falsifiability, as a criterion of meaningfulness (or at least of meaningful assertion), is too strong.
Also, notice that some theists would not answer 'nothing' to the question: what, if anything, might show that your belief in God is false? In which case, even if Flew's criterion of meaningful assertion were correct, it would in any case fail establish that this sort of theist failed to assert anything by saying 'God exists'.
There is a vast array of logical objections to theism. By no means all have been included here. New objections will no doubt emerge. Do any succeed?
We have seen how, with some ingenuity, it is always possible to find a way round an internal or external logical objection, by, say, tweaking your definition of God, finding some fault in the logic, or, in the case of external objections, giving up one of your other beliefs. Does this mean that all logical objections fail?
In so far as such objections force theists to revise their position, no.
What these objections target are very specific conceptions of God (sometimes in combination with other beliefs) and, in many cases, the theism that involves those very specific conceptions must, on pain of contradiction, be abandoned (or else the other beliefs abandoned). The target theistic belief or belief-combination is sometimes successfully refuted.
Theists sometimes acknowledge this. However, other varieties of theism, very loosely understood, always remain on the table. So the theist can always switch to one of those other varieties. For example a theist convinced by logical objections to a personal God may, in response, switch to a non-personal conception of God. Indeed, the theist may insist that one of the other varieties of God was always what they had in mind. They've just been misunderstood, and/or have themselves been unclear about the character of the God they believe in.
From the perspective of the theist, this strategy of switching and adjusting belief in response to such objections constitutes progress in getting clear about what theism fundamentally involves. Such logical objections are helpful to theism, they say, by allowing theists to clarify the nature of God.
From the perspective of many atheist critics, on the other hand, these same logical objections are in many cases a threat to theism, and the theistic strategy of switching, modifying, and/or abandoning beliefs in order to retain at least some sort of logically consistent theistic belief looks suspiciously like, not clarification, but rather a merry-go-round of evasion.
Which of these two perspectives is the more accurate is, I suspect, the fundamental question to press regarding logical (and indeed evidential) objections to theism.
Ayer, A.J. (2001) Language, Truth and Logic New Edition. London: Penguin Books.
Boethius, Ancius (1999) The Consolation of Philosophy, Revised Edition, London: Penguin Classic.
Flew, A. Hare, R.M., and Mitchell, B. (1964) 'Theology and Falsification: a University Discussion' in New Essays in Philosophical Theology, Flew, A. and MacIntyre, A. (eds.) New York: Macmillan, 99-103. Available online at: http://www.politik-salon.de/files/theory_of_falsification.pdf
Gale, Richard M. (1991) On the Nature and Existence of God Cambridge: Cambridge University Press.
Geach, P. T. 1973. 'Omnipotence' Philosophy 48 (183): 7-20.
Hick, John (1966) Faith and Knowledge 2nd Edition, Ithaca: Cornell University Press.
Hoffman, J. and Rosenkrantz, G.S. (2010) 'Omnipotence' in A Companion to Philosophy of Religion 2nd. Edition, C. Taliaferro, P. Draper, and P. L. Quinn (eds.), Oxford: Wiley-Blackwell, 243–50.
Mackie, J.L. (1955) 'Evil and Omnipotence' Mind 64 (254): 200-212.
Maimonides, Moses (1996). The Eight Chapters of Maimonides on Ethics (Semonah Perakim), edited, annotated, and translated with an Introduction by Joseph I. Gorfinkle. New York: AMS Press.
Miller, Barry (1996) A Most Unlikely God Notre Dame and London: University of Notre Dame Press.
Murray, Michael ((2002) 'Deus Abscondicus', in Divine Hiddenness: New Essays, Howard-Snyder, D. and Moser, P. (eds.) Camrbidge: Cambridge University Press: 62-82.
Oppy, Graham (2003) 'The Devilish Complexities of Divine Simplicity' Philo 6 (1): 10-22.
Plantinga, Alvin. 1974. The Nature of Necessary Oxford: Oxford University Press.
Pike, Nelson (1970) God and Timelessness New York: Schocken.
Priest, Graham (2006). In Contradiction: A study of The Transconsistent, 2nd Edition Oxford: Oxford University Press.
Rundle, Bede (2004) Why There Is Something Rather Than Nothing Oxford: Clarendon.
Schellenberg, J. L. (2004) 'Does divine hiddenness justify atheism?' in Contemporary debates in the philosophy of religion M. L. Peterson & R. J. Van Arragon (eds.) London: Blackwell. 30–41
Schellenberg, John (2006) Divine Hiddenness and Human Reason New Edition, Ithica NY: Cornell University Press.
Schellenberg, John (2015) The Hiddenness Argument Oxford: Oxford University Press.
Sobel, Jordan Howard (2004) Logic and Theism, Arguments For and Against belief in God Cambridge, Cambridge University Press.
Stump, Eleanor (1997) 'Simplicity' in P. Quinn and-C. Taliaferro (eds.) A Companion to Philosophy of Religion Oxford: Blackwell
Eleanor Stump and Norman Kretzmann (1981) 'Eternity' The Journal of Philosophy 78/8: 429-458.
Walterstorff, Nicholas (2007) 'God is Everlasting' in Philosophy of Religion: Selected Readings third edition Peterson, M., Hasker, W., Reichenbach, B. and Basinger, D. (eds.) Oxford: Oxford University Press.
Wittgenstein, Ludwig (1998) Philosophical Investigations Oxford: Basil Blackwell.
 I assume for the purposes of this essay that contradictions cannot be true, though note that some, including Graham Priest, argue that some contradictions can be true (and, simultaneously, false). See Priest (2006).
 Note the argument is not that the evil we observe provides good evidence against the existence of God, making the existence of God less probable. That is the evidential problem of evil. See the chapter in this volume 'Evidential Objections to Theism'.
 Of course, in suggesting that the theist can solve the logical problem of evil by dropping any one of the three classical omni-attributes, I am assuming those omni-attributes are logically independent, which is contentious. If omnipotence logically requires omniscience, say, then the theist does not have the option of dropping omniscience alone. They would have to drop omnipotence too.
 This example is adapted from one provided by Linda Zagzebski in her entry to Stanford Internet Encyclopedia of Philosophy on 'Foreknowledge and Freewill': https://plato.stanford.edu/entries/free-will-foreknowledge/
 Note that, for technical reasons, the authors prefer to characterise omnipotence in terms of performing tasks, but in terms of bringing about certain states of affairs.
 Note that in later developments of the argument, Schellenberg switches from talk of inculpable belief to talk of non-resistant belief, acknowledging that one might somehow be culpable - be to blame - for ones failure to believe in God even if one is not aiming deliberately to shut the door on any relationship with God that might be on offer. It is the latter 'resistant' form of non-belief that Schellenberg maintains is incompatible with the existence of a perfectly loving God. (2015: 54-55)
 Schellenberg has suggested to me in correspondence that his (2015) volume presents an argument that belief in perfectly loving god is incompatible with belief in non-resistant believers, whereas elsewhere - e.g. in his (1993) book and in his (2004) paper - he argues only that it is actually false that God would permit non-resistant non-belief or that non-resistant non-belief provides at least evidence against the existence of a perfectly loving god.
 They may use 'God' as a label for something they first encounter through a glass darkly, as it were (Paul uses the phrase in Corinthians: 1 Cor. 13, 12). Thus the subject matter of their belief - the God they 'have in mind' - can remain a constant, even while the beliefs they hold about the subject matter may undergo considerable revision. Compare: suppose I introduce 'Tim' as a label for him - that person I now see dimly through a mist; it's still Tim I have in mind when I later admit that much of what I first believed about Tim (based on his misty appearance) was incorrect.