Wright's Private Language Argument Refuted

(currently unpublished) copyright to myself. Feel free to link but do not repost.

ABSTRACT: Wright’s Private Language Argument Refuted

Crispin Wright has developed a novel take on the private language argument presented by Wittgenstein in Philosophical Investigations PI 258-60. Wright’s private language argument is ingenious and important, and it has not, to date, been refuted. In this paper I refute Wright’s argument, pointing out that it commits the fallacy of equivocation (it trades on an ambiguity in his use of the phrase “reason to believe”). While there may be a cogent private language argument presented in PI 258-60, Wright has failed to hit upon it.

Wright’s Private Language Argument Refuted

Introduction

Suppose our sensations are necessarily private and inaccessible to others. That would seem to introduce the possibility of a subject introducing a private sensation-language the meaning of which would necessarily be unknown to others. Suppose, for example, that I define “S” by reference to my private sensation. I focus my attention on the inner phenomenon, thereby impressing on myself the connection between sign and sensation. I might then use “S” to record facts about my private mental life. For example, I could use it to record in my diary those days on which I have that particular sensation. The meaning of “S” would necessarily be unknown, and necessarily unteachable, to others.

In Philosophical Investigations (PI) §§258-60, Wittgenstein appears to present an argument against the possibility of someone introducing such a “private language”. However, it remains controversial both what the private language argument is, and whether or not it is cogent.

In Does Philosophical Investigations §§258-60 Suggest A Cogent Argument Against Private Language? Crispin Wright presents an argument suggested to him by §§258-60, an argument that Wright believes is, in fact, cogent. That argument has not, as yet, been refuted. I aim to refute it here.

Wright’s private language argument

Wright begins by making a case for two key principles. The first principle lays down a condition on a sentence being apt for the expression of a fact. According to Wright,
sentences of a given family are apt for the expression of a fact only if:

(a) X believes what “P” expresses,

and

(b) What “P” expresses is true.

have an appropriately contrasting content where “P” is any of (appropriately many of) the sentences in question. (1986:228)

Wright’s second principle is that (a) and (b) will have appropriately contrasting content only if:

the information conveyed by (a) should differ from that conveyed by (b). And it is plausible that, for a large class of examples, two items of information differ just in case there can be such a thing as reasonably regarding oneself as possessing one and not the other. (1986:230)

These two principles have some prima facie plausibility, and I shall, for the sake of argument, concede both here. Wright then proceeds as follows. From the second of the above principles, Wright derives the interim conclusion that

we are entitled to regard (a) and (b) as conveying different items of information only if someone could have reason to believe one but not the other and could be aware of the fact. (1986:231)

The final step in of Wright’s argument is to show that, where “P” is a sentence of a putative private language (such as, “I am having S again”), it is not possible that someone could have reason to believe (a) but not (b), or vice verse. Wright continues:

The argument will be that when “P” is, putatively, a sentence of a language which no two people could reasonably believe they share, that is not a possibility.

It will suffice to consider four cases. Letting “A” range over believing subjects, we have

(i) A is aware of possessing both reason to believe (a) and reason to doubt (b);
(ii) A is aware of possessing both reason to doubt (a) and reason to believe (b);
(iii) A is aware of possessing both reason to believe (a) and no reason to believe (b)
(iv) A is aware of possessing no reason to believe (a) and reason to believe (b). (1986:231)

Wright believes that, whether A is X him- or herself or some third party Y, if “P” a sentence of a private language, none of the above four cases can obtain. It thus follows, according to Wright, that “P” cannot state fact. A language in which one might state facts concerning one’s own private sensations is impossible.
Wright deals with cases (i) to (iv) in turn, in each case providing one or more arguments for why, if “P” is a sentence of a private language, that case cannot obtain. I shall not run though all these arguments. It will be instructive, however, to run through Wright’s argument that, where “P” is a sentence of a private language, (iii) cannot obtain.

Let’s begin by supposing that A is not the putative private linguist but a third person Y. Wright’s argument that Y cannot be aware of possessing both reason to believe (a) and no reason to believe (b) is as follows.

True, Y does not know which aspect of X’s psychological state “P” concerns. However, Y does know that “P”, as used by X, concerns some aspect or other of X’s psychological state. And it is, generally, a feature of psychological states that the subject of such a state be some sort of authority about it. So if, for example, Y has reason to suppose X believes he is experiencing S again, then Y also possesses some reason to believe X is experiencing S again. As Wright puts it:

[I]f one may at least suppose that “P”, as used by X, concerns some aspect or other of X’s psychological state, then one is bound to take reason for (a) as supporting (b); and note that the point is not dependent on crediting X with Cartesian authority for his psychological states – it is enough that he be any sort of (fallible) authority about them, that his opinions about them count for something. (232)

Wright concludes that, where “P” is a sentence of X’s putative private language, Y cannot be in state (iii).

What if A is X – the putative private linguist him or herself? Can A then be in state (iii)? No, says Wright. X’s belief that “P” is true gives X grounds for believing “P” is true, for, again, “P” concerns X’s own psychological state, something about which X must surely be considered some sort of authority.

Refutation of Wright’s private language argument

Wright’s argument is not cogent. As indicated above, Wright begins by appealing to the principle that
two items of information differ just in case there can be such a thing as reasonably regarding oneself as possessing one and not the other. (1986:230)

But notice that this principle allows (a) and (b) to convey different pieces of information even if it is true that possessing reason to believe one immediately provides one with some reason to believe the other. For the situation may be that, while the existence of reason to believe (a) inevitably provides one with some reason to believe (b), the degree of support provided to each belief may differ, raising the possibility of it being reasonable to regard oneself as possessing one piece of information but not the other.

Here’s an analogous case. Consider the two pieces of information: Fred is a Christian, and: Fred is religious. Ceteris paribus, evidence supporting one of these beliefs inevitably provides at least some support the other. Yet it is entirely possible reasonably to regard oneself as possessing the latter piece of information but not the former.

To show that the information conveyed by (a) and (b) does not differ, Wright must show that it is impossible for one reasonably to regard oneself as possessing one piece of information but not the other. Wright fails to do this. He points out, no doubt correctly, that if, for example, A has reason to believe (a), then, because “P” concerns the putative private linguist’s own psychological state (something about which the private linguist must be considered some sort of authority), A will also have some reason to believe (b). But note that this fact is entirely consistent with it being reasonable for A to regard him or herself as possessing the information conveyed by (a) but not the information conveyed by (b). Thus it is also consistent with (a) and (b) conveying different pieces of information.

For example, suppose Y knows both that X defined “S” by reference to one of X’s psychological states and also that X has not given any thought to his definition until months later, when X is suddenly convinced he is experiencing S again. This delay may makes it rational for Y to entertain serious doubts about the reliability of X’s memory concerning how “S” is should be applied. In these circumstances, Y’s knowledge that X believes that he is experiencing S may give Y some reason to suppose that X is experiencing S again, as Wright maintains. But of course it may not be reason sufficient to make it reasonable for Y to regard him or herself as possessing the information that X is experiencing S again. Thus Y may reasonably regard him or herself as possessing the information that X believes he is experiencing S again, but not the information that X is experiencing S again. In which case, if sentence “P” is “I am experiencing S again” (as uttered by X), Wright has failed to give us any reason why the information conveyed by (a) and (b) cannot differ.

The fallacy of equivocation

Let me identify precisely where Wright’s private language argument goes wrong. It commits the fallacy of equivocation. Wright argues from the principle that

two items of information differ just in case there can be such a thing as reasonably regarding oneself as possessing one and not the other. (1986:230)

to the interim conclusion that

we are entitled to regard (a) and (b) as conveying different items of information only if someone could have reason to believe [my italics] one but not the other and could be aware of the fact (1986:231).

But what does “reason to believe” mean in the interim conclusion? The phrase “reason to believe” is ambiguous. It might mean, for example reason sufficient to make it reasonable to believe, or it might just mean some (possibly inadequate) reason to believe.

Now notice that Wright’s interim conclusion follows from his principle only if “reason to believe” means the former, not the latter. For we might still reasonably regard ourselves as possessing one item of information but not another despite that fact that possessing reason to believe one thing inevitably gives us some reason to believe the other.

Wright then claims that, where “P” is a sentence of a putative private language, the impossibility of cases (i) to (iv) suffices to show that (a) and (b) must convey the same information.

But notice that in cases (i) to (iv) “reason to believe” presumably means only some (possibly inadequate) reason to believe. Otherwise, for example, Wright’s case (iii):

A is aware of possessing both reason to believe (a) and no reason to believe (b)

comes out as the case in which A is aware of possessing both reason sufficient to make it reasonable for them to believe (a), but not reason sufficient to make it reasonable for them to believe (b) – a case which, as we have seen, is possible, i.e. because, while having reason to believe (a) immediately gives A some reason to believe (b), the reason A possesses for believing (a) may be much stronger than for believing (b), thereby making it reasonable for them to believe (a) but not (b).

In short, Wright’s private language argument trades on an ambiguity in the use of “reason to believe” - on an unwitting slide from one way of using that phrase to another. It commits the fallacy of equivocation.

There may be a cogent private language argument contained in PI 258-60. What is now clear. I believe, is that Wright has failed to hit upon it.

PLANTINGA'S BELIEF-CUM-DESIRE ARGUMENT REFUTED

PLANTINGA’S BELIEF-CUM-DESIRE ARGUMENT REFUTED
Stephen Law

FORTHCOMING IN RELIGIOUS STUDIES. NOTE THAT CAMBRIDGE UNIV. PRESS NOW OWN COPYRIGHT.

Abstract

In Warrant and Proper Function, Plantinga develops an argument designed to show that naturalism is self-defeating. One component of this larger argument is what I call Plantinga’s belief-cum-desire argument, which is intended to establish something more specific: that if the content of our beliefs does causally effect behaviour (that is to say, semantic content is not epiphenomenal), and if naturalism and current evolutionary doctrine are correct, then the probability that we possess reliable cognitive mechanisms must be either inscrutable or low. This paper aims to refute Plantinga’s belief-cum-desire argument.

Introduction
In the final chapter of Warrant and Proper Function , Plantinga argues that, if both:

(N) naturalism – the view that there are no supernatural beings

(E) evolution - current evolutionary doctrine

are true, then the probability that:

(R) our cognitive faculties are reliable and produce mostly true beliefs

must be either low or inscrutable.
Plantinga argues, further, that this argument furnishes anyone who accepts N&E with a undefeatable defeater for any belief produced by those faculties, including N&E itself. Hence, N&E has been shown to be self-defeating.
One part of this larger argument is what I call Plantinga’s belief-cum-desire argument. The belief-cum-desire argument is designed to show something more specific - that if the content of our beliefs does causally affect behaviour, and N&E, then the probability of R cannot be high.
My objective here is to refute the belief-cum-desire argument.

Plantinga’s belief-cum-desire argument
Suppose some hypothetical rational creatures much like us evolve on a planet a lot like Earth - they ‘hold beliefs, change beliefs, make inferences, and so on’ . Suppose:

(C) causal efficacy – the content of beliefs causally affects behaviour

is true. What is the probability of R/N&E&C specified with respect to these creatures – what is the probability that their cognitive faculties are reliable?
The probability, says Plantinga, is not as high as you might initially be tempted to suppose. For it is not belief per se that is adaptive, but behaviour. And behaviour is caused by combinations of belief and desire. But then, claims Plantinga, for any given adaptive action (action that enhances the creatures ability to survive and reproduce),

there will be many belief-desire combinations that could produce that action; and very many of those belief-desire combinations will be such that the belief involved is false.

Plantinga illustrates like so:

So suppose Paul is a prehistoric hominid; a hungry tiger approaches. Fleeing is perhaps the most appropriate behavior: I pointed out that this behavior could be produced by a large number of different belief-desire pairs. To quote myself: ‘Perhaps Paul very much likes the idea of being eaten, but when he sees a tiger, always runs off looking for a better prospect, because he thinks it unlikely that the tiger he sees will eat him. This will get his body parts in the right place so far as survival is concerned, without involving much by way of true belief. . . . . Or perhaps he thinks the tiger is a large, friendly, cuddly pussycat and wants to pet it; but he also believes that the best way to pet it is to run away from it. . . . or perhaps he thinks the tiger is a regularly recurring illusion, and, hoping to keep his weight down, has formed the resolution to run a mile at top speed whenever presented with such an illusion; or perhaps he thinks he is about to take part in a 1600 meter race, wants to win, and believes the appearance of the tiger is the starting signal; or perhaps . . . . Clearly there are any number of belief-cum-desire systems that equally fit a given bit of behavior.’

So adaptive behaviour can be produced by many belief-desire combinations and, ‘in many of these combinations, the beliefs are false’ . We cannot, concludes Plantinga, estimate the probability of R on N&E&C as high. And of course, if we cannot estimate that probability as high for these hypothetical creatures, then we cannot estimate it as high in our own case either.
The above argument that the probability of R given N&E&C cannot be high has some superficial plausibility. Plantinga is surely correct that:

(i) it is behaviour that evolution selects for rather than beliefs per se.

He is also correct that:

(ii) for any piece of adaptive behaviour, there are many belief-desire combinations that might produce it, on many of which the belief or beliefs in question are false.

However, I will show that, appearances to the contrary, it does not follow from (i) and (ii) that we cannot reasonably estimate the probability of R on N&E&C as being high. Indeed, I shall go further, and sketch out some reasons for supposing that the probability of R given N&E&C must, in fact, be fairly high.

Refutation of Plantinga’s belief-cum-desire argument
Consider two possible scenarios:

(a) we have evolved certain false beliefs and certain desires that, in combination, result in adaptive behaviour
(b) we evolved certain unreliable belief-producing mechanisms and certain desires that, in combination, result in adaptive behaviour.

Perhaps, on N&E&C, (a) is not so unlikely, for the reasons Plantinga cites. Suppose I have an innate belief that tigers are cuddly and that best way to pet a tiger is to run away from it. If I am also equipped with an innate desire to pet tigers, this results in adaptive behaviour.
But what about (b)? How likely is it on N&E&C that our belief-producing mechanisms are unreliable? Consider the question: what particular set of desires would a species need to evolve in order for the beliefs generated by such an unreliable mechanism to result in generally adaptive behaviour? Let’s look at some examples, beginning with the cognitive faculty of reason.

Example one: the FAC
Consider the fallacy of affirming the consequent (FAC). The FAC is an unreliable form of inference. It sometimes produces true conclusions, but often false.
Suppose evolution hard-wires a species of hominid H to be highly prone to the FAC. Suppose a member of this species, H1, concludes using the FAC that jumping out of planes is not safe. Another member, H2, concludes using the FAC that jumping off tall buildings is safe. They might reason like so:

H1’s inference:
If jumping out of planes is not safe, jumping out of balloons is not safe
Jumping out of balloons is not safe
Jumping out of planes is not safe

H2’s inference:
If jumping out of planes is safe, then jumping out of planes wearing a parachute is safe
Jumping out of planes wearing a parachute is safe
Jumping out of planes is safe

If evolution hard-wires a desire into species H to make H2’s resulting belief that jumping out of planes is safe adaptive – e.g. a powerful desire to commit suicide - that same hard-wired desire will result in the likely death of H1.
What set of desires must evolution instil in species H to render adaptive the potentially-mal-adaptive consequences of applying the FAC? There is no such set of desires!
The FAC sometimes produces false beliefs, but sometimes true. Is that the reason why evolution cannot render the FAC adaptive? Could a method of inference that consistently produced false conclusions from true premises be made adaptive by pairing it with an appropriate set of desires? No, as I explain below.

Example two: Counter-induction
Consider two hominids A and B. A reasons inductively and B counter-inductively, like so:

A observes that whenever other hominids eat, they usually continue to live, and when they stop eating, they die. He concludes that if he eats, he’ll probably continue to live, and if he stops eating, he’ll die.

B observes that whenever other hominids eat, they usually continue to live, and when they stop eating, they die. He concludes that if he eats, he’ll probably die, and if he stops eating, he’ll continue to live.

When A applies his form of reasoning to true premises, he is likely to end up with a true belief. B on the other hand, is likely to end up with a false belief. His method of inference consistently produces false beliefs.
However, if evolution equips A with a desire to live, and B with a desire to die, B’s false belief produces the same adaptive behaviour as A’s true belief. Counter-induction has delivered a false belief, but it has not produced behaviour that is mal-adaptive.
So far, it seems that Plantinga is correct: given evolution equips A and B with the right desires, the behaviour produced by their belief-forming mechanisms is equally adaptive.
But now suppose A and B engage in further reasoning, applying their respective methods of inference like so:

A observes that other hominids that forage and hunt get food to eat, and those who don’t get none. A concludes that if he hunts and gathers, he’ll get food to eat, and if he doesn’t he’ll get none.

B observes that other hominids that forage and hunt get food to eat, and those that don’t get none. B concludes that if he doesn’t hunts and gather, he’ll get food to eat, and if he does, he’ll get none.

Now, A’s reasoning helps him survive. Given his desire to live, these two inferences together will lead him to hunt and gather. That’s adaptive behaviour.
The problem is, given the desire required to get B’s first counter-inductive inference to produce adaptive behaviour, B’s second counter-inductive inference is now likely to produce mal-adaptive behaviour. Given B’s desire to die, plus his false belief that eating will kill him, his second counter-inductively generated conclusion will no doubt lead him not to go hunting and gathering. B will probably starve to death.

Plantinga is correct that for any piece of adaptive behaviour, there are many belief-desire combinations that might produce it, on many of which the belief or beliefs in question are false. But it does not follow that the probability of R given N&E&C cannot reasonably be estimated as high. The members of a species equipped with unreliable belief-forming mechanisms such as the FAC or counter-induction will end up with all sorts of combinations of false beliefs the potentially mal-adaptive consequences of which cannot be made adaptive by evolution hard-wiring that species with some particular set of desires.
In fact there are two difficulties here.
First, there is the problem of novel beliefs. An advantage of procedural reasoning is that it allows for creatures able to problem solve and adapt, within their own lifetimes, to a changing environment and novel situations. An adaptive inferential mechanism is likely to applied in new ways. But then evolution cannot then anticipate what desires will be required to render adaptive the innumerable potentially mal-adaptive conclusions likely to be drawn. If B draws the first counter-inductive conclusion, his desire to die renders his conclusion adaptive. But if B happens to go on and draw that second conclusion using the same unreliable form of inference, that same desire now renders the conclusion mal-adaptive.
The second problem is that not only can evolution not anticipate which desires creatures will need to render the conclusions of such unreliable inferences adaptive, when it comes to unreliable forms of inference, there just is no set of desires that will render the mechanism adaptive. A set of desires that renders one set of conclusions adaptive will render another set of conclusions generated by the same mechanism mal-adaptive.
On the other hand, evolution can make reliable forms of inference adaptive in a straightforward way, by equipping the species in question with desires for those things that enhance its ability to survive and reproduce. In which case, the probability that reliable forms of inference will evolve, as opposed to an unreliable forms of inference, looks to be high.

Other cognitive faculties
The considerations sketched out above suggest that N&E&C should lead us to estimate the probability that our cognitive faculty of procedural reasoning is reliable as fairly high. But of course, procedural reason alone furnishes us with little, if any, knowledge. Other cognitive faculties – mostly notable perception and memory – must also come into play.
How reasonable is it, given N&E&C, to suppose that these other faculties are reliable? If there is no good reason to suppose they are reliable, then there’s no good reason to suppose our various faculties working in conjunction constitute a reliable belief-forming system. My car may have a reliable carburettor, but if other parts are unreliable, the car as a whole remains unreliable.
So let’s now look at the cognitive faculties of memory and perception. Has Plantinga shown that, given N&E&C, the probability that these other faculties are reliable cannot be high?

Memory
Suppose hominid species H is equipped with an unreliable memory. Hominid H1 has at time t1 true beliefs B1 and B2. But, because H1’s memory is unreliable, she later believes the falsehoods not-B1 and not-B2. Is there a desire or set of desires with which evolution might also equip species H that will render adaptive the behaviour produced by these two resulting false beliefs? Very probably. If B1 is the belief that if you eat you will live and B2 the belief that if you don’t eat you will die, these beliefs will result in adaptive action if H1 desires to die. However, because H1 previously believed B1 and B2, she would previously have not eaten, which is mal-adaptive behaviour. There is no set of desires that will make both the input and output beliefs of this unreliable faculty result in adaptive behaviour. But then unguided evolution cannot equip species H with a set of desires that will make the input and output beliefs of this unreliable faculty generally adaptive. Evolution can, on the other hand, equip a species with a set of desires that will make the input and output beliefs of a reliable faculty generally adaptive. It appears, then, that N&E&C will therefore strongly favour a reliable memory faculty over an unreliable faculty.

Perception
How likely is it, on N&E&C, that evolution would produce a species with a reliable perceptual-mechanism-cum-desire combination, rather than an unreliable-perceptual-mechanism-cum-desire combination?
Fairly likely, I suspect. Here are two categories of unreliable perceptual mechanisms:

(1) Unreliable mechanisms producing mostly false beliefs.
(2) Unreliable mechanisms that produce significant proportion of, but not mostly, false beliefs

Let’s begin by considering perceptual or quasi-perceptual mechanisms of type (1). Such mechanisms fall, in turn, into two categories:

(1a) Unreliable mechanisms producing mostly false beliefs but in a systematic, predictable way.

and

(1b) Unreliable mechanisms producing mostly false beliefs in a random, unpredictable way

An example of (1a) would be a perceptual or quasi-perceptual mechanism that, whenever the subject is presented with a tiger, produces the belief there is a rabbit present. There is consistency to the error. An example of (1b) would be a perceptual or quasi-perceptual mechanism that, when the subject is presented with a tiger, may the first time produce the belief there is nothing present, the next time the belief a rabbit present, the next time the belief there is a chair present, and the time after that the belief there’s a side of beef present, etc., but rarely if ever the belief that there is a tiger present. While we can predict that the subject will make an error about there being a tiger in front of them, it is not possible, even given knowledge of the erroneous beliefs previously produced when a tiger was present, to predict what erroneous belief will now be produced on this occasion.
Can unguided evolution make an unreliable mechanism of type (1b) produce adaptive behaviour by combining it with an appropriate set of desires? It is hard to see how. If there is a tiger present and the mechanism makes me believe there is a rabbit present, my mistaken belief can still result in adaptive behaviour if evolution has given me a desire to run away from rabbits. But if the erroneous beliefs are being generated in a random way, there will be no particular desire or set of desires with which evolution might equip me that will make the random false beliefs generated by this mechanism adaptive.
What about a mechanism of type (1a)? Does the pattern to the errors produced by the mechanism mean that evolution can render the mechanism adaptive by combining it with an appropriate set of desires?
That suggestion might seem plausible when we consider a very simple example of adaptive behaviour, such as running away from tigers. If the mechanism systematically produces the belief that a rabbit is present whenever a tiger is present, all evolution need do is instil in these subjects a powerful desire to run away from rabbits.
But the suggestion becomes far less plausible when we consider more we consider sophisticated patterns of adaptive behaviour of the sort we actually exhibit.
Suppose, for example, that to reach food you need to survive, you need to engage in some team activity with other members of your species – e.g. negotiating some tricky terrain that includes a narrow ledge and a poisonous snake. Someone has to distract the snake while someone else crawls carefully along the edge and leaps over the snake at the exact moment it is distracted.
Now try to imagine a perceptual mechanism of type (1a) that produces mostly false beliefs about your surroundings, but beliefs that, when paired with certain desires with which evolution has pre-equipped your species, will result in the required adaptive behaviour from you and your team mates.
You must not believe there is a snake and a ledge and some food and some team mates with whom you must co-operate. And nor must your team-mates. You, and they, must have mostly false beliefs about your environment, but beliefs that, nevertheless, when paired with desires with which evolution has collectively furnished you, lead you to act in tandem with your other team members to retrieve and eat the food.
In fact, setting aside the challenge of imagining such a mechanism, it is a difficult enough challenge to construct just a set of mostly false beliefs and hard-wired desires that would result in the complex sequence of actions required. Perhaps it is not impossible. Perhaps your (1a) type mechanism causes you to believe that instead of food at the end of the ledge, there’s a little man who will give you a tickle stick if you walk carefully along a white line, jumping in the air after 15 seconds, and then reach down and take the stick. Perhaps you believe that eating the tickle stick is the best way to get tickled. If we pair this false belief with a desire to be tickled, your resulting sequence of actions might yet be adaptive. You might successfully negotiate the narrow ledge, leap over that snake (though who is going to distract it?) and then eat the food.
However, even if we can come up with a mostly-false-perceptual-belief-cum-desire combination that would, in this situation, result in adaptive action, it is still more difficult still to come up with a belief-forming mechanism of type (1a), which, when paired with an appropriate set of desires, will result in sophisticated patterns of adaptive behaviour generally of the sort of which we are capable. If the next time the food lies beyond a chasm that can only be negotiated if you and your team place a tree trunk across the gap, then the false belief ‘There’s a little man who will give you a tickle stick if you walk carefully forward along the white line, jumping in the air after 15 seconds…’ combined with that powerful desire to be tickled will send you and your team mates straight over the cliff. That is not adaptive behaviour.
It is not yet clear that there is any set of desires that, when combined with an unreliable perceptual mechanism of type (1a), will generally produce sophisticated patterns of adaptive behaviour of the kind we actually exhibit.

Let’s now turn to perceptual mechanisms of the second sort:

(2) Unreliable mechanisms that produce a significant proportion of, but not mostly, false beliefs

Such, as it were, hit and miss (as opposed to consistently miss) mechanisms may also be of two kinds: those producing false beliefs in a random way – there being no pattern to the errors, and those in which the errors are, in certain respects, systematic.
We have already seen in the case of mechanisms of type (1b) that a mechanism producing erroneous beliefs in a random way is not a mechanism that evolution might pair off with a particular set of desires such that adaptive behaviour will result.
But what of a hit and miss mechanism in which there is a pattern to the misses? An example would be a mechanism that was reliable with respect to the shape of objects but systematically unreliable with respect to position. Equipped with such a mechanism, a creature might believe, correctly, that there is a square object in its vicinity, but it will be mistaken about where that object is located.
Are there potentially many such perceptual or quasi-perceptual mechanisms that, while systematically producing many false beliefs, will still result in generally adaptive behaviour given evolution pairs the mechanism with an appropriate set of desires? And, if so, is there a significant probability, on N&E&C, that we have evolved such an unreliable mechanism, rather than a reliable mechanism?
Here is a sketch of two reasons why the answer to these questions is unlikely to be yes.
First, we have seen that it is difficult to envisage type (1a) mechanisms that, given N&E&C, will result in sophisticated sequences of team activity of the sort required to retrieve the food from that snake-inhabited narrow ledge. I cannot see that it is significantly easier to envisage a type (2) mechanisms of that sort. Try, for example, to imagine a type (2) mechanism producing mostly correct beliefs about the shape of objects but systematically incorrect beliefs about their location that will result in such successful sequences of team activity – I have tried, and failed. If someone claims there are many such potential mechanisms, the onus is on them to provide a series of examples to illustrate the point. I am unable.
Second, even if there are many such potential mechanisms, is there a significant probability, given N&E&C, that we have evolved such a mechanism rather than a reliable mechanism? Perhaps not. Consider, again, a mechanism that is reliable about the shape of objects but systematically unreliable about their position. The most obvious way such a mechanism might evolve is in two stages: first evolving a mechanism that is reliable about both the shape and the position of objects, and then engineering a mechanism that systematically reassigns positions to objects, but in such a way that, given the desires with which the species is also equipped, still results in adaptive behaviour. But why would that second level of engineering evolve given the reliable first level is already producing adaptive behaviour? What would be the pay off, for evolution, of now adding a sophisticated location-reassignment mechanism and changing the desires so that adaptive behaviour still results? If there is unlikely to be such a pay off, it is unlikely such a systematic-error-producing mechanism would evolve. Evolution will stick with the reliable mechanism.
In fact, even if N&E&C had equipped us with an unreliable perceptual faculty or faculties of type (2), it still would not follow that probably many of our beliefs are false. We have seen reasons to suppose that N&E&C will favour reliable as opposed to unreliable faculties of memory and procedural reasoning. If a species also possesses perceptual faculties that are partly reliable and partly, but systematically, unreliable, there arises the possibility – perhaps the probability – that the members of this species will be able to figure out that they are, to some extent, being systematically misled by those faculties. In which case, they may well adjust their beliefs accordingly. Their beliefs would now reliably reflect reality, despite the fact that they possessed unreliable perceptual faculties. If R is the reliability of their cognitive faculties acting in tandem, the probability of R might still be high, even if it was more probable than not that they possessed unreliable perceptual faculties of type (2).

Conclusion
Regarding his hypothetical creatures ‘ a lot like us’ evolving on another planet ‘a lot like Earth’, Plantinga claims that, for any given piece of adaptive behaviour they engage in,

there are many belief-desire combinations that will lead to the adaptive action; in many of these combinations, the beliefs are false. Without further knowledge of these creatures, therefore, we could hardly estimate the probability of R on N&E and this final possibility [C] as high.

The word ‘therefore’ is not justified by what precedes it. While it may be true that for any piece of adaptive behaviour there are many false-belief-cum-desire combinations that might produce it, it does not follow that we cannot reasonably estimate the probability, for either Plantinga’s hypothetical creatures or for ourselves, of R/N&E&C to be high. This is, not least, because, when we turn from beliefs to belief-producing cognitive mechanisms of the sort with which we are equipped (e.g. reason, perception, memory), it is no longer clear that there are many (indeed, any) unreliable versions of such mechanisms that, by virtue of unguided evolution pairing them with certain hard-wired desires, will nevertheless result in the sort of sophisticated patterns of adaptive behaviour we exhibit.
So Plantinga’s belief-cum-desire argument fails. Indeed, I have sketched out some reasons for thinking that the probability, on N&E&C, that our cognitive faculties, operating in tandem, are reliable, is actually pretty high (though I certainly do not claim to have established that here).
However, there remain two concessions to be made.
The first is: we should acknowledge that to refute the belief-cum-desire argument is not to refute Plantinga’s larger evolutionary argument against naturalism. The belief-cum-desire argument is intended to work in tandem with certain other estimates of probability to deliver the conclusion that the probability of R given N&E is either low or (given the questionability of these various estimations) inscrutable. The other key estimates are that, on N&E, the probability of C is low (because semantic epiphenomenalism is likely to be true), and (ii) that, on N&E and not-C, the probability of R is low. Now, as a matter of fact, given Plantinga’s estimates of these other probabilities, his conclusion that the probability of R given N&E must be low still follows (though the probability is now not as low as it would have been given the probability of R given N&E&C was high). So, we while we may have succeeded in refuting the belief-cum-desire argument, we cannot claim to have refuted Plantinga’s larger argument that the probability of R given N&E must be low or inscrutable. I make no such claim.
Secondly, Plantinga offers at least one other argument for the limited conclusion that we cannot reasonably estimate the probability of R/N&E&C to be high. Someone’s beliefs about the world might be largely false because they mistakenly think e.g. everything is conscious, and they refer to things in such a way as to attribute consciousness to them, so all their beliefs are of the form ‘that conscious so-and-so is such a such’. Despite being largely false, this individual’s beliefs could still be adaptive. Thus, even if N&E&C is true, this individual’s cognitive faculties could still be unreliable (in the Plantingian sense that they produce largely false beliefs). So, given N&E&C, we cannot be confident we have not evolved such unreliable mechanisms. This alternative argument is, again, not one I claim to have refuted here.
So the aims of this paper have been comparatively modest. Nevertheless, the belief-cum-desire argument, even if not indispensable to Plantinga’s larger project, nevertheless constitutes one of the most interesting and initially intuitively appealing parts of Plantinga’s larger case, and its loss is significant.
I note that in its most recent incarnation (in ‘Content and Natural Selection’ ), Plantinga’s evolutionary argument against naturalism no longer includes the belief-cum-desire argument – perhaps because Plantinga, too, has come to realize its failings.

THE EVIL GOD CHALLENGE - forthcoming in Religious Studies

POSTSCRIPT:

My Paper "The Evil God Challenge" is now available online at the CUP journals page http://journals.cambridge.org/repo_A72V8TEm

This is the final, published version, appearing in Religious Studies shortly. PDF VERSION.

THE EVIL GOD CHALLENGE

Stephen Law

NB Note COPYRIGHT RESTS WITH CAMBRIDGE UNIVERSITY PRESS. This is forthcoming in Religious Studies.

Abstract


This paper develops a challenge to theism. The challenge is to explain why the hypothesis that there exists an omnipotent, omniscient and all-good god should be considered significantly more reasonable than the hypothesis that there exists an omnipotent, omniscient and all-evil god. Theists typically dismiss the evil god hypothesis out of hand because of the problem of good – there is surely too much good in the world for it to be the creation of such a being. But then why doesn’t the problem of evil provide equally good grounds for dismissing belief in a good god? I develop this evil god challenge in detail, anticipate several replies, and correct errors made in earlier discussions of the problem of good.



I: THE EVIL GOD CHALLENGE

Let’s call the central claim classical of monotheism – that there exists an omnipotent, omniscient and supremely benevolent creator – the good god hypothesis. Typically, those who believe this hypothesis, while perhaps insisting that it is a ‘faith position’, nevertheless consider it not unreasonable. Believing in the existence of God, they maintain, is not like believing in the existence of Santa or fairies. It is much more reasonable than that.
In response, critics often point out that, even if most of the popular arguments for the existence of God do provide grounds for supposing that there is some sort of supernatural intelligence behind the universe, they fail to provide much clue as to its moral character. Suppose, for example, that the universe shows clear evidence of having been designed. To conclude, solely on that basis, that the designer is supremely benevolent would be about as unjustified as it would be to conclude that it is, say, supremely malevolent, which clearly would not be justified at all.
Critics may add that there is, in addition, ample empirical evidence against the existence of such supremely benevolent being. In particular, they may invoke the evidential problem of evil.

The problems of evil

There are at least two problems of evil. The logical problem begins with the thought that the claim:

(1) There exists an omnipotent, omniscient and maximally good God

is logically inconsistent with the claim:

(2) Evil exists

Under ‘evil’ I mean to include both suffering and morally blameworthy actions. The argument then proceeds as follows. Clearly, (2) is true. Therefore, (1) is false.
Note that the amount of evil is irrelevant this version of the argument – all it requires is that there is some, no matter how little.
Perhaps the logical problem of evil does not pose such a great challenge to theism. To deal with it, it would suffice to show that an all-powerful, all-knowing and maximally good God might allow some evil for the sake of some greater good.
A second problem – the evidential problem – rests not on the thought that (2) is logically incompatible with that of (1), but on the thought that (2) provides us with good evidence against (1). The amount of evil does now become relevant. Even if we acknowledge God might have reason to allow some evil, surely there can be no good reason for quite so much? We can sharpen the problem by noting that God will presumably not allow any gratuitous suffering to exist. There must be a good reason for every last ounce of it.
Many argue that not only is there little reason to suppose that the God of classical monotheism exists, the sheer quantity of evil that exists provides us with overwhelming empirical evidence that he doesn’t. Those theists who maintain that belief in God, if not proved, is at least not unreasonable, are mistaken. Far from being a question reason cannot decide, the claim that the God of classical monotheism exists seems to be straightforwardly empirically falsified.

Theodicies

Faced with this objection, theists respond in various ways. They may suggest we possess good grounds for believing that, not only is there a creator, this being does indeed have the properties attributed to him by traditional monotheism. I will return to that suggestion later. They may also suggest that the problem of evil can, to a significant extent, be dealt with. Many theistic explanations of evil have been offered, including the following:

Simple free will solution. We are not blind automata, but free agents. As a consequence of God having given us free will, we sometimes choose to do wrong. Suffering results. However, free will allows for certain important goods, such as the possibility of morally virtuous action. God could have created a universe populated with puppet beings that always did as God wants. But the behaviour of such puppet beings lacks the dimension of moral responsibility that makes our actions morally virtuous. By cutting our strings and setting us free, God inevitably allowed some evil. But this evil is more than outweighed by the important goods that free will allows.

The ‘character-building’ solution. This is, to borrow John Hick’s phrase, a ‘vale of soul making’. We know that a bad experience can sometimes make us stronger. People who have suffered a terrible disease sometimes say they gained greatly from it. Similarly, by causing us pain and suffering, God allows us to grow and develop both morally and spiritually. It is only through our experiencing this suffering that we can become the noble souls God wants us to be.

Second order goods require first order evils. Theists may remind us that God had inevitably to include quite a bit of suffering in his creation in order that certain important goods could exist. Take, for example, charity. Charity is a great virtue. Yet we can only be charitable if there exist others who are needy. Charity is a so-called second order good that require first order evils like neediness and suffering (or at least their appearance) to exist. The second order good outweighs the first order evils, which is why God allows them.

When offered in response to the evidential problem of evil, such explanations are sometimes called theodicies. It is on the evidential problem of evil and on theodicies that I focus here.
Of course, as theodicies, these explanations have obvious limitations. For example, even if the simple free will solution succeeds in explaining the evil we bring about by our own free action, it fails to explain so-called natural evils – such as the suffering brought about by natural disasters. Arguably, all three theodicies fail to explain why there is quite so much suffering in the world.
True, other, sometimes more sophisticated, explanations have been also offered, as we shall see. Some believe these theodicies, if not individually, then at least collectively, largely take the sting out of the evidential problem of evil. The problem, they suppose, may not have been entirely solved, but it has at least been brought down to manageable proportions.
Still, there remains an acknowledgement by many serious-minded theists that it is certainly isn’t easy to explain quite why omnipotent, omniscient and supremely benevolent being, would unleash so much horror on the sentient inhabitants of this planet over hundreds of millions of years. This leads some to supplement these explanations with a further appeal – to mystery. God works in mysterious ways. Because God is infinitely intelligent and knowledgeable, his divine plan is likely to be mostly ‘beyond our ken’ . In which case, the fact that the reason for much of the evil that exists is beyond our understanding is not good evidence for his non-existence.
As I say, the three theodicies outlined above have been challenged. I too intend to challenge them, and also several others, but in an unusual way. I intend to take a step back and question the character and plausibility of such explanations collectively, by means of an analogy.

The evil god hypothesis

Consider a different hypothesis. Suppose the universe has a creator. Suppose also that this being is omnipotent and omniscient. But suppose he is not maximally good. Rather, imagine that he is maximally evil. His depravity is without limit. His cruelty knows no bounds. There is no other god or gods – just this supremely wicked being. Call this the evil god hypothesis.
How reasonable is the evil god hypothesis? I have already pointed out that, certainly in their simplest versions, most of the popular arguments for the existence of God fail to provide any clue as to our creator’s moral character. In which case, to the extent that they support the good god hypothesis (that’s to say, not very much, if at all), they also support the evil god hypothesis.

The problem of good

Still, isn’t there overwhelming evidence against the evil god hypothesis? I am referring, of course, to what might be called the evidential problem of good. The problem is that of explaining why an omnipotent, omniscient and supremely evil being would allow quite so much good into his creation. Why, for example, would an evil god:

• Give some of us immense health, wealth and happiness?
• Put natural beauty into the world, which gives us pleasure?
• Allow us to help each other, thereby reducing suffering and increasing the amount of things evil god despises, such as love?
• Give us children to love who love us unconditionally in return?
• Equip us with beautiful, healthy young bodies?

Surely, if a supremely evil being is going to introduce sentient beings into his creation, it will to torture them and have them do evil. Surely he won’t allow love, laughter and rainbows. Nor will he permit us to perform the kind of selfless and courageous acts that ennoble us and reduce the pain and suffering of others.
So, yes, the world contains much evil. But there is also a great deal of good - far too much good, in fact, for this plausibly to be the creation of such a limitlessly powerful and malignant being.
Notice how the evidential problem of evil mirrors the evidential problem of good. If you believe in an omnipotent, omniscient and maximally good god, then you face the challenge of explaining why there is quite so much evil in the world. Similarly, if you believe in an omnipotent, omniscient and maximally evil god, then you face the challenge of explaining why the world contains quite so much good.

Some reverse theodicies

Of course, few, if any, of us believe the evil god hypothesis. Prima facie, not only is there little reason to suppose such a being exists, there also seems to be overwhelming evidence against his existence. When presented with the evil god hypothesis, most of us immediately dismiss it as absurd, typically because we consider the problem of good decisive.
But notice that, just as there are moves theists make to try deal to with the problem of evil, so there are similar moves we might make to try to deal with the problem of good. Here are some examples:

1. Simple free will solution. Evil god gave us free will. Having free will means we sometimes choose to do good, which evil god hates. However, it also introduces the possibility of evil acts for which agents can be held morally responsible. An evil god could have created a universe populated with puppet beings that he ensured always behaved unpleasantly. But the behaviour of such puppet beings lacks the dimension of moral responsibility that transforms such acts into actions of the most depraved and despicable kind. To maximize evil, an evil God will want us to perform cruel and selfish acts of our own volition.

In response to this first suggestion, some may object: ‘But why is a world such as this, in which we possess free will, worse than a world in which we possess no freedom and are simply compelled to cause endless misery to each other? Surely the latter would be far more evil. So why didn’t evil god create it?’
But this is to forget that a world in which we are compelled to maximize suffering is a world in which no morally evil actions are performed. And moral evil is a particularly profound and important form of evil (as even theists typically acknowledge). Just as, from the point of view of a good god, a world lacking morally good actions is gravely deficient, so similarly, from the point of view of an evil god, a world lacking morally evil actions is also gravely deficient.
In response, it may be said: ‘But still, a world in which there is a free will is far preferable to us than a world in which we are compelled to cause each other endless misery. The second hellish sort of existence would be far worse. And thus preferable from an evil god’s point of view. So why didn’t evil god create it?’
There is some plausibility to this response. Notice, however, that much the same kind of worry can be, and has been, raised about the standard free will theodicy. Dostoyevsky’s Ivan Karamazov, for example, asks whether our freedom isn’t bought at an unacceptably high price if it results in the torture of innocent children. Surely, Ivan and others suggest, given the choice between creating a heavenly world in which we are made noble and virtuous and enjoy a profoundly joyful existence, and a world in which, as a result of our having been given have free will, humanity, as a consequence, endures endless war, murder, rape, torture, the Holocaust, and so on, a good god would choose the former (certainly many of us would much prefer to occupy the former heavenly world; indeed, many theists hope and pray they will eventually do so).
So, while there may be a difficulty here for the free will solution to the problem of good, that does not reveal it to be any less plausible than the standard free will solution to the problem of evil, given this kind of worry is common to both.
Here are two more solutions:

2. The character-destroying solution. Hick was mistaken: this is a vale, not of soul making, but of soul-destruction. Evil god wants us to suffer, do evil and despair.
Why, then, does an evil god create natural beauty? To provide some contrast. To make what is ugly seem even more so. If everything were uniformly, maximally ugly, we wouldn’t be tormented by the ugliness half as much as if it was peppered with some beauty.
The need for contrast also explains why evil god bestows upon a few lavish lifestyles and success. Their happiness is designed to make the suffering of the rest of us even more acute. Who can rest content knowing that they have so much more, that they are undeserving, and that no matter how hard we might strive, we will never achieve what they have (and remember, too, that even those lucky few are not really happy).
Why does evil God allow us to have beautiful children to love and that love us unconditionally in return? Because we will worry endlessly about them. Only a parent knows the depths of anguish and suffering that having children brings.
Why does an evil god give us beautiful, healthy young bodies? Because we know that out health and vitality will be short-lived, that we will either die young or else slowly wither. By giving us something wonderful for a moment, and then gradually pulling it away, an evil god can make us suffer even more than if we had never had it in the first place.

3. First order goods allow second order evils. Some evils are second order evils requiring first order goods. Take jealousy. I cannot feel jealous unless I perceive others to have something worth being jealous of. Evil God had to allow a few of us to have goods (or perceived goods) so that jealousy might exist.

Let us call such attempts to explain the problem of good reverse theodicies. If these reverse theodicies leave you unconvinced, remember that, like a defender of the good god hypothesis, we can also play the ‘mystery’ card. Being infinitely intelligent and knowledgeable, evil god’s supremely ingenious and diabolical plan is likely to be largely beyond our ken. In which case, the fact that we can’t understand why there is so much good in the world if he exists is not good evidence of his non-existence.

The symmetry thesis

The three reverse theodices introduced above to deal with the evidential problem of good obviously mirror the three theodicies we looked at earlier. In fact, other theodices can be mirrored in this way too (see below). This suggests an interesting way to challenge theism.
How persuasive are our three reverse theodicies? Intuitively, not at all. Rather than being taken seriously, they usually provoke amusement among theists and non-theists alike. But this raises the question: if the reverse theodicies are feeble and ineffective, why should we consider the standard theodicies any more effective?
We may also raise a larger question. In terms of reasonableness, isn’t there a broad symmetry between the good god hypothesis and the evil god hypothesis? Take arguments supporting the two hypotheses. I pointed out earlier that many of the popular arguments in support of the good god hypothesis turn out to provide much the same sort of support (i.e. not very much) for the evil god hypothesis.
Moreover, when it comes to dealing with the evidence against the respective hypotheses provided by the enormous quantities of both good and evil that we find in the world, we can construct similar kinds of explanation. In particular, the three theodicies offered to deal with the evidential problem of god are mirrored by the reverse theodicies outlined above.
I shall call the suggestion that, in terms of reasonableness, there is indeed such a rough symmetry between the good god and evil god hypotheses, the symmetry thesis.

The scales analogy

Suppose the reasonableness of the good god and evil god hypotheses is in each case indicated by a pointer on a set of weighing scales. Depending on how each of our two scales is loaded – considerations adding to reasonableness being placed on the left of each scale; considerations subtracting from reasonableness being added to the right – the pointer on each scale moves from highly reasonable through a range of positions (fairly reasonable, not unreasonable. etc.) to highly unreasonable.
Certainly, we find that many of the popular arguments loaded by some theists onto the left side of the good god scale can just as effectively (or ineffectively) be loaded onto the left side of the evil god scale. We also find the weighty problem of evil on the right side of the good god scale is mirrored by the hefty problem of good on the right side of the evil god scale. And we find that three theodicies we have seen used by theists to try remove or lessen the weight of the problem of evil on the good god scale (perhaps we might think of them as large helium balloons that can be attached to the problem to lighten the load) are mirrored by reverse theodicies that might be used to reduce the weight of the problem of good.
The symmetry thesis says that, when we load the scales correctly with all the available evidence and other considerations pertinent to the reasonableness of a belief (incidentally, I make no commitment to evidentialism here ), the two scales settle in roughly similar positions.
Now most of us, theists included, consider the evil god hypothesis highly unreasonable. We suppose there is little of any substance to place on the left had side of the scale, and that, when the boulder that is the problem of good is added, the scale lurches violently to the right, not withstanding the effects of any reverse theodicy helium balloons we might then try to attach. Yet adherents of the good god hypothesis typically suppose the good god scale far more evenly balanced. To believe in a good god, they think, is not like believing in fairies, Santa or, indeed, an evil god. When this scale is properly loaded and the pointer observed, they say, we find it points to ‘not unreasonable’ or even ‘quite reasonable’.
In short, those who embrace the good god hypothesis typically reject the symmetry thesis. The challenge I am presenting to those who believe in the God of classical monotheism, then, is to explain why, if belief in an evil god is highly unreasonable, should we consider belief in a good god significantly more reasonable?
We might call this the evil god challenge.

The problem of good in the literature

I am not the first to note how the problem of good might be used to generate a problem for theists. The earliest discussion appears to be in the 1968 volume Evil and the Concept of God by Madden and Hare , in which the authors devote three pages to the problem of good. After briefly sketching some reverse theodicies, Madden and Hare conclude:

[t]he point should be clear by now that the problems of evil and good are completely isomorphic; what can be said about one can be said about the other in reverse. For any solution to one problem there is a parallel solution to the other, and for every counter-argument in the one there is a parallel counter argument in the other.

In his 1976 paper ‘Cacodaemony’ , Cahn (quite independently) draws the same conclusion, claiming that

…classic arguments in defence of the view that every evil in the world makes possible a world containing even greater goods can be exactly paralleled by arguments in defence of the view that every good in the world makes possible a world containing even greater evils.

In ‘God, the Demon, and the Status of Theologies’ , published in 1990, Stein concurs with Hare, Madden, and Cahn that

[a] demonist can constrict a demonology which is isomorphic for any theodicy.

New (also unaware of the earlier literature), in his 1993 paper ‘Antitheism’ , also develops some mirror arguments for, and reverse theodicies in defence of, belief in an evil god. Finally, ‘God, Demon, Good, Evil’ , published in 1997, Daniels attempts to deal with the arguments of Hare, Madden, Cahn and Daniels by suggesting that there is a crucial asymmetry between the good and evil god hypotheses – Daniels argues that an evil god is actually an impossibility. I respond to Daniels’ objection at the end of this paper.
There are a number of important differences between my evil god challenge and the earlier challenges raised by Madden and Hare, Cahn, Stein, and New.
First, as will become clear, I reject Hare, Madden, Cahn and Stein’s central claim: that the problems of good and evil and their respective solutions are ‘exactly parallel’(Madden and Hare). The solutions are not exactly parallel. I will indicate some asymmetries between the two problems and sets of theodicies (and also asymmetries in the arguments that might be mounted for these respective gods). However, I will explain why these local asymmetries need not, and probably do not, threaten the symmetry thesis.
Second, I find fault in New’s attempt to deal with certain seemingly non-reversible arguments for a good god, and provide a better response to those arguments.
Third, I intend my evil god hypothesis to provide a more nuanced and tougher challenge to theism than the challenges raised those raised by earlier contributors to this discussion, not just by acknowledging and responding to the problem of local asymmetries, but also by anticipating and dealing with a broader range of potential theistic responses.

II: RESPONSES TO THE EVIL GOD CHALLENGE

Some may think the evil god challenge easily met. For example, haven’t we omitted several important arguments for the existence of God which are arguments specifically for good god, and which are not mirrored by any corresponding arguments for an evil god? Don’t these arguments show that belief in a good god is, after all, rather more reasonable than belief in an evil god?

Miracles and religious experience

Take for example, the argument from miracles. Miracle cures and other supernatural phenomena are regularly observed. Some are officially investigated and confirmed by religious authorities such as The Catholic Congregation for the Causes of Saints. Don’t such events provide at least some evidence for the existence of, not just a god, but a good god prepared to perform great works of good in response to our prayers?
Or consider the argument from religious experience. Religious experiences are almost always judged to be experiences of something immensely positive. Again, don’t they provides us with at least some evidence that, not only is there some sort of intelligence behind the universe, this intelligence is a force for good, not evil?
Even if such arguments are far from conclusive when considered individually, we might suppose that they contribute towards making a cumulative case for the existence of, not just a god, but the supremely benevolent god of classical monotheism.
But if this is true, then the balance of the good god scale now shifts. We have something rather more weighty to place on the left side of the good god scale, something to which there corresponds nothing that might be placed on the left side of the evil god scale. Do we now have grounds for rejecting the symmetry thesis?

New on arguments from miracles and religious experience

In ‘Antitheism’ , New attempts to deal with this seeming asymmetry by constructing mirror arguments for an evil god. He asks us to imagine a world the inhabitants of which have experiences as of an evil god (New calls them ‘anti-religious experiences’) and who note harmful or evil events that cannot be explained scientifically (New calls them ‘anti-miracles’). We have now imagined evidence for an evil good that precisely mirrors the evidence for a good god.
The problem with New’s strategy, however, is that imaginary evidence isn’t really evidence. I can’t provide evidence against a scientific theory simply by imagining some. If evidence is to count, it must actually exist.
Many theists insist we have real evidence for a good god – the evidence provided by miracles and religious experience. The problem for the symmetry thesis, the theist may insist, is that there simply isn’t anything like this sort of evidence for anti-religious experiences and anti-miracles.
New’s attempt to mirror the arguments from miracles and religious experiences fails. However, as I now explain, there is a better way of responding to the arguments from miracles and religious experience.

A better response

Do the arguments from miracles and religious experience provide better evidence for a good god then they do an evil god?
Suppose the evil god hypothesis is true. This malignant being may not want us to know of his existence. In fact, it may help him maximize evil if he deceives us about his true character. An evil and omnipotent being will have no difficulty duping human beings into believing he is good. Taking on a ‘good’ guise, he might appear in one corner of the world, revealing himself in religious experiences and performing miracles in response to prayers, and perhaps also giving instructions regarding what his followers should believe. He might then do the same in another part of the globe, with the exception the instructions he leaves regarding what should be believed contradict what he has said elsewhere. Our evil being could then stand back and watch the inevitable conflict develop between communities to whom he has now misleadingly revealed himself, each utterly convinced by their own stock of miracles and religious experiences that the one true all-good god is on their side. Here we have a recipe for ceaseless conflict, violence and suffering.
When we observe how religious experiences and miracles are actually distributed, this is more or less the pattern we find. So, even if they are genuinely supernatural, do these miraculous phenomena provide better evidence for a good god than an evil god? While a good god might create miracles and religious experiences, it is difficult to see why he would produce them in this way, given the predictably horrific consequences. Perhaps miracles and religious experiences do indicate the activity of a supernatural agency, but it is arguable that their actual arrangement fits the evil god hypothesis rather better than it does the good god hypothesis. We should not, at this stage, rule out the possibility that, if there is an asymmetry between the two hypotheses, it is because the evil god hypothesis is actually rather more reasonable than the good god hypothesis.
In reply to the above defence of the evil god hypothesis, it may be asked: ‘But why would an evil god bother to deceive us about his true character, given that full knowledge of this merciless and all-powerful torturer would actually be far more terrible?’
The answer, of course, is that an evil god will want to allow for the performance of morally evil acts within his creation. As already noted, a world lacking moral agents able to perform actions of a profoundly wicked character is a world that is seriously deficient from his point of view. So not only does evil god create a world in which we are free moral agents, he also engineers the kind of circumstances in which we are, then, likely freely to choose to do evil. Religiously motivated conflicts clearly have been, and continue to be, a major source of moral evil in the world. By means of this deception, an evil god is able to create an environment within which moral evil is likely to flourish.
One may still raise this objection: ‘But surely nothing could be worse than hell as traditionally conceived? Why doesn’t an evil god just send us straight to hell?’
However, as already noted, a mirror puzzle faces those who believe in a good god. Given a heavenly environment would be profoundly more joyful than this, why doesn’t a good god send us straight to heaven? Why are so many of us allowed to go through such appalling suffering here?
Given both god hypotheses face this kind of objection, it constitutes, as it stands, no threat to the symmetry thesis.
Moreover, we can, in both cases, attempt to deal with the objection by appealing to an after-life. We are sent to this world first, where we have the opportunity to act in profoundly morally good and evil ways (this being important to both the good and evil gods). We then pass on to an after-life: an eternity in heaven or (on the evil god hypothesis) hell, where joy or (on the evil god hypothesis) pain and suffering are maximized [and any earlier evils or (on the evil god hypothesis) goods in the first stage of our existence are compensated]. I will look at examples of such after-life theodicies shortly.

Historical evidence

Incidentally, the above response can be extended to deal with arguments for a good god based on historical evidence, such as the evidence provided by scripture (not all of which is rooted in religious experiences and miracles).
Some will suggest there is much textual and other historical evidence that might be marshalled to support belief in a good deity, but no corresponding evidence to support belief in an evil deity – and this constitutes a significant asymmetry between our two god hypotheses.
In response, we may again ask – does this historical evidence really fit the good god hypothesis better than the evil? Not if our evil god wishes to create the illusion that he is good, in order to foster the deception outlined above. It may well be in his interest to fabricate misleading evidence about his own character.
When we consider the spread of evidence supplied by the miracles, religious experiences, and also the historical evidence associated with the various different faiths, it is at least arguable that the pattern we find fits the evil god hypothesis better than the good. For, to repeat, why on earth would a good god produce these phenomena in such a way as to guarantee endless religious strife? Surely their actual, disastrous arrangement is rather more likely to be the handiwork of a malignant being?

A moral argument

Another strategy the theist might adopt, in order to establish a significant asymmetry between the good and evil god hypotheses, would be to maintain that there are moral arguments for the existence of a good god that cannot be mirrored by parallel arguments for an evil god. For example, they may argue that our moral sense could only have a supernatural origin, and that only a good god would have an interest in providing it. So the fact that we have a sense of right and wrong is powerful evidence favouring the good god hypothesis over the evil god hypothesis.
This particular argument fails, however. While it might be true that only a supernatural being is capable of furnish us with a moral sense, the fact is an evil god might well also have an interest in providing such a sense. For by providing us with both free will and knowledge of good and evil, an evil god can allow for the very great evil of our freely performing evil actions in the full knowledge that they are, indeed, evil.
Why, then, is the fact that we do possess knowledge of good and evil evidence favouring the good god hypothesis over the evil?

A second moral argument

A different kind of moral argument for the existence of a specifically good god focuses not on knowledge of objective moral value, but on its existence. Some will insist that moral facts are both non-natural and objective, and that a good god is therefore required to underpin them (or at least provides the best explanation of them).
It is at least contentious whether a cogent argument along these lines can be constructed. Notoriously, such arguments face the Euthyphro dilemma. Suppose we say that God, as divine law-maker, decrees that certain things, such as stealing and murder, are wrong. Does God decree this because he recognises that stealing and murder are, independently, wrong, or are they wrong only because he decrees them to be so? The first answer makes god redundant so far as setting up a standard of right and wrong is concerned – murder would have been wrong anyway, whether or not god exists, or, indeed, whether or not god himself happens to be good or evil. But then the objective, non-natural wrongness of murder would obtain anyway, even if there were an evil god. On the first answer, there can exist both a non-natural, objective standard of right and wrong, and an evil god. The second answer, notoriously, appears to make the moral wrongness of murder arbitrary and relative. Notice that this is a problem whichever of our two god hypotheses we favour. In short, on the first answer, there is no problem for the evil god hypothesis; on the second, there is, prima facie, equally a problem for both hypotheses. The Euthyphro dilemma thus constitutes a major obstacle to the construction of a moral argument for the existence of a specifically good, rather than evil, god.
Of course, it remains possible that cogent moral argument along the above lines might yet be constructed. I suspect that, for those who reject the symmetry thesis, this is the most promising line of attack. However, to date, it remains, even among theists, controversial whether any such argument exists.

More reverse theodicies

Let’s now return to standard theodicies and their mirror versions. Perhaps we have underestimated the range and efficacy of the standard theodicies on offer. Are there some that are not reversible?
Certainly there are many we have not yet discussed. However, in many, if not all, cases, reverse theodicies quickly suggest themselves. To illustrate, I will sketch out three more examples: (i) a reverse laws of nature theodicy, (ii) a reverse after-life theodicy, and (iii) a reverse semantic theodicy.

Laws of nature theodicy. Effective purposeful action requires the world behave in a regular way (for example, I am able deliberately to light this fire by striking my match only because there are laws that determine that under such circumstances, fire will result). That there be laws of nature is a prerequisite of our having the ability both to act on our natural environment and interact with each other within it. These abilities allows for great goods. They give us the opportunity to act in a morally virtuous way, for example. However, such a law-governed world inevitably produces some evils. For instance, the kind of laws and initial conditions that produce stable land masses on which we can survive and evolve also produce tectonic shifts that result in earthquakes and tidal waves. Still, the evil of earthquakes and tidal waves is more than outweighed by the goods those laws allow. We might think we can envisage possible worlds that, as a result of being governed by different laws and/or initial conditions, contain a far greater ratio of good to evil (that contain stable land masses but no earthquakes, for example), but, due to consequences we have failed to foresee (perhaps the absence of earthquakes is at the cost of some even worse kind of global catastrophe), such worlds will, in reality, always be worse than the actual world.

A reverse theodicy can be constructed like so:

Laws of nature reverse theodicy. Effective and purposeful action requires that the world behave in a regular way. That there be laws of nature is a prerequisite of our having the ability to both act on our natural environment and interact with each other within it. These abilities allows for great evils. For example, they give us the opportunity to act in morally depraved ways – by killing and torturing each other. By giving us these abilities, evil god also allows us to experience certain important psychological forms of suffering such as frustration – we cannot try, and become frustrated through repeated failure, unless we are first given the opportunity to act. True, such a law-governed world inevitably produces some goods. For example, by giving us the ability to act within a physical environment, evil god gave us the ability to avoid that which causes us pain and seek out that which gives us pleasure. Still, such goods are more than outweighed by the evils these laws allow. We might think we can envisage possible worlds that, as a result of being governed by different laws and/or initial conditions, contain a far greater ratio of evil to good (that contain far more physical pain and far less pleasure, for example), but, due to consequences we have failed to foresee (perhaps the greater suffering will result in us being far more charitable, sympathetic and generally good towards others), such worlds will, in reality, always be better than the actual world.

To this, some may object : ‘Very well, an evil god will produce laws of nature so we can possess the power to do evil – but surely he will also sometimes suspend those laws in order to cause us confusion and frustration and to produce evils to which the laws of nature would otherwise prove an obstacle.’
Notice, however, that both theodicies face this type of objection. A similar concern can be raised about the standard laws of nature theodicy. Yes, a good god will produce a regular universe so that are able to do good, but surely he would be prepared to suspend those laws and intervene in order, say, to thwart some particularly morally despicable act (e.g. stopping Hitler’s rise to power) or to prevent some particularly terrible natural disaster, or to help us achieve some very great good (perhaps arranging for a stroke of good fortune in a science lab that then leads to a cure for cancer). A good god would not just stand back and allow thousands of children to be buried alive in an earthquake, even if the earthquake does happen to be the result of natural laws that are otherwise largely beneficial.
After-life theodicies are also popular. Take the following version presented by Tim Mawson in his Belief in God :

Compensatory after-life theodicy. The pain and suffering we experience in this world is more than compensated for in the after-life – where we will experience limitless good. The explanation for why a good god would not simply send us straight to heaven is that it is only within a law-governed world within which we have free will (something which, according to some theists, such as Mawson , we lack in heaven) that we can enjoy important goods, including the very great good of doing good of our volition. As a consequence of inhabiting this world for a short while, we suffer, but this suffering is more than compensated for by an eternity of communion with God in heaven.

Mawson’s after-life theodicy can also be mirrored like so:

Reverse compensatory after-life theodicy. The joy and happiness we experience in this world is more than compensated for in the after-life – where we experience limitless evil. The explanation for why a good god would not simply send us straight to this endlessly cruel world is that it is only within a law-governed world within which we have free will that we can experience important evils, including the very great evil of doing evil of our volition. As a consequence of inhabiting this world for a short while, we experience some goods, but this is more than compensated for by what follows: an eternity of suffering in the company of a supremely malignant being.

It is also possible to reverse the standard semantic responses to the problem of evil. Consider this example:

Semantic theodicy. When we describe God as being ‘good’, the term means something different to what it means when applied to mere humans. This difference in meaning at least partly explains why a good god would do things that we would not call ‘good’ if done by us.

We can reverse this theodicy like so:

Reverse semantic theodicy. When we describe god as being ‘evil’, the term means something different to what it means when applied to mere humans. This difference in meaning at least partly explains why an evil god should do things that we would not call ‘evil’ if done by us.

With a little ingenuity, reverse theodicies can be constructed for many other standard theodicies too. However, as I now explain, we should probably concede that - contrary to the claims made by Madden, Hare, Cahn and Stein - in some cases, no ‘exactly parallel’ theodicy can be constructed.

Asymmetries

Take for example, theodicies founded in a particular Christian story about the Fall and redemption. When we examine Augustine’s explanation of natural and moral evils – that both are rooted in the original sin of Adam and Eve – no parallel narrative suggests itself. An attempt to construct a reverse story about a reverse Adam and Eve, who, through disobedience to their evil creator, bring about a reverse ‘Fall’ runs into insuperable obstacles. For example, while a good god might have some reason to allow the natural evils brought about by original sin to continue (for these evil consequences, being brought on ourselves, are deserved, and there remains, in addition, God’s offer of redemption) why would an evil god allow the continued existence of the natural goods brought about by the disobedience of a reverse Adam and Eve? It may be that, with some ingenuity, a rather different sort of narrative involving an evil god might be constructed to account for natural goods, but it is hard to see how it could mirror the Christian story of the Fall in sufficient detail to qualify as a reverse theodicy. Pace Madden, Hare, Cahn and Stein, it seems that not every theodicy even has a parallel, let alone an exact one.
Even where a parallel theodicy can be constructed, there may still be asymmetries. For example, if we suppose free will is itself an intrinsic good, then the reverse free will theodicy involves an evil god imbuing us with the good of free will. While an evil god may still be able to maximize evil by giving us free will, he will nevertheless have to pay a price (introducing that intrinsic good) – a price for which there is no parallel in the standard free will theodicy. Arguably, this makes the standard free will theodicy much more effective than the reverse version. The theist may insist that because free will is not just an intrinsic good, but a very great good, so very great additional quantities of evil are required to outweigh it – so great, in fact, as to render the reverse free will theodicy significantly less plausible than the standard theodicy.
So it appears that there are some asymmetries between the two sets of theodicies. However, the effect of these asymmetries appears to be comparatively minor, having little effect on the overall balance of reasonableness.
For example, given the mythic status of Adam, Eve, and the Fall, Augustine’s theodicy fails. But then the absence of a parallel theodicy does not affect the balance of reasonableness very much (and in any case, we might be able to construct a different sort of narrative to accompany the evil god hypothesis that accounts for natural goods in another way).
What of the asymmetry between the free will and reverse free will theodicies? Stein attempts to defend the thesis that for each theodicy there is an ‘exact parallel’ by arguing that free will is not, in fact, an intrinsic good.
However, suppose we grant, for the sake of argument, that free will is an intrinsic good. That requires we abandon the Madden-Hare-Cahn-Stein thesis that for each theodicy there is a reverse theodicy that is its ‘exact parallel’. But does it require we abandon my symmetry thesis – the thesis that when we load the good god and evil god scales correctly with all the available evidence and other considerations pertinent to the reasonableness of a belief the two scales settle in roughly similar positions?
I don’t believe so, for at least three reasons:
First, this asymmetry between the two theodicies may very well neutralized by another. In order for us to have a full range of free choices between good and evil, god, whether good or evil, must introduce pain, suffering and death not just as possibilities but as realities. Not only must he make us vulnerable to pain, suffering and death (to give us the option of torturing or murdering others), he must actually inflict pain and death so that we have the free choice to help alleviate or prevent it. Now if it is prima facie plausible that free will is an intrinsic good, it is no less plausible that pain, suffering and death are intrinsic evils. In which case both free will theodicies requires the introduction of intrinsic goods and intrinsic evils. While the intrinsic goods give the evil god hypothesis some additional explaining to do, the intrinsic evils give the good god hypothesis some additional explaining to do. In which case, it appears the two asymmetries balance out.
Second, even if it were true that the free will theodicy is significantly more effective then the reverse theodicy, that might not greatly effect balance of reasonableness between the good and evil god hypotheses. Suppose, for the sake of argument, that the standard free will theodicy is entirely effective in accounting for moral evils, and that the reverse theodicy wholly ineffective in accounting for moral goods (this being a far more dramatic asymmetry than even the one proposed). Thus we leave the full weight of moral good on the evil god scale, but entirely remove the weight of moral evil from the good God scale. Does this change in the balance of the two scales result in the two pointers indicating very different levels of reasonableness? Arguably not. For, ceteris paribus, there still remains an enormous amount of evil on the good god scale (such as the extraordinary quantities of suffering unleashed on sentient creatures over hundreds of millions of years before moral agents even made an appearance on the Earth). It may be argued (I think with some plausibility) that when those evils explained by the free will theodicy are removed, there remains more than enough weight of evil to keep the needle pointed firmly at ‘highly unreasonable’. The needle does not now point at ‘not unreasonable’ or ‘quite reasonable’ – it remains stuck down the ‘highly unreasonable’ end of the scale. The scale has shifted a little, perhaps, but not by very much. If that is so (and do I think it at least arguable), then the symmetry thesis remains true.
Third, let’s remember that even if the standard free will theodicy is rather more effective than the reverse theodicy, this asymmetry might in any case be counterbalanced or outweighed by other asymmetries favouring the evil god hypothesis over the good god hypothesis. In fact, we have already discovered one example: prima facie, the evidence concerning miracles and religious experience appears to support the evil god hypothesis rather more than it does the good god hypothesis.
To conclude, then, it seems that - pace Madden, Hare, Cahn and Stein - the two sets of theodicies do not precisely parallel each other. There are asymmetries. However, we have found little reason to suppose these asymmetries have much effect on the overall level of reasonableness of our respective god hypotheses. We have not yet found good reason to suppose that our two sets of scales do not, as the symmetry thesis states, settle in roughly similar positions.

Other moves

To finish, I now anticipate five responses the evil God challenge may provoke, and briefly sketch out some of the difficulties they face.

1. Significantly more good than evil
We might try to meet the challenge by showing that there is significantly more good than evil in the world. This will be hard to establish however, not least because good and evil are difficult to quantify and measure. Some theists consider it just obvious that the world contains more good than evil, but then many (including some theists) are struck by the exact opposite thought. Appeals to subjective estimations can carry little weight.

2. Ontological arguments
Might ontological arguments provide a priori grounds for supposing that not only is there a god, he is good? The most obvious difficulty here is that it is debatable, to say, the least, whether any cogent ontological argument can be constructed. The cogency of those arguments that have been offered remains unrecognized not just by non-theists, but also by many theists – perhaps the majority of philosopher-theists. They, certainly, will not be reaching for the ontological argument in order to demonstrate why the symmetry thesis fails.
New notes that some ontological arguments are, in any case, reversible. Take this example (my own – based on New and Anselm):

I can conceive of an evil god - a being whom no worse can be conceived.
But it is worse for such being to exist in reality than in the imagination. Therefore, the being of which I conceive must exist in reality.

3. Impossibility arguments
Could we meet the evil God challenge by showing that an evil God is actually an impossibility, for the very notion of an evil god contains a contradiction? Here are two examples of such an argument:

(i) Daniel’s platonic refutation’ of the evil God hypothesis
In ‘God, Demon, Good, Evil’ , Daniels suggests the resources to deal with the evil god challenge can be found in Plato’s Gorgias. Daniels believes Plato has shown that an evil god is an impossibility. His ‘platonic refutation’ of the evil god hypothesis is as follows.
First, Daniels claims we always do what we judge to be good. Even when I smoke, despite judging smoking to be bad, I do it because I judge that it would be good to smoke this cigarette here and now.
It follows, says Daniels, that no one does bad knowingly. But then it follows that if a being is omniscient, he will not do bad. There cannot exist an omniscient yet evil being. The notion of an omniscient yet evil being involves a contradiction.
I believe Daniels’ argument trades on an ambiguity in his use of the word ‘good’. True, whenever I do something deliberately, I judge, in a sense, that what I do is ‘good’. But ‘good’ here need mean no more than, ‘that which I aim to achieve’. We have not yet been given any reason to suppose I cannot judge to be ‘good’, in this sense, what I also deem to be evil, because I desire evil. Yes, an evil god will judge doing evil to be ‘good’, but only in the trivial sense that evil is what he desires. Pace Daniels, there is no contradiction involved in an omniscient being judging evil to be, in this sense, ‘good’.

(ii) The desire argument
A rather different argument would be: ‘But by bringing about evil, your evil god thereby aims to satisfy his own desire for evil; and the satisfaction of a desire is an intrinsic good. Thus the idea of a maximally evil god aiming to produce an intrinsic good involves a contradiction.’
This argument also fails. Even if we grant the dubious assumption that the satisfying of any desire – even an evil one – is an intrinsic good, the most we have revealed, here, is another local asymmetry – that, in aiming to maximize evil, evil god would have also to aim to achieve at least one intrinsic good (namely, the satisfaction of his desire to maximize evil). What we have established, perhaps, is that there are certain logical limits on God’s evilness (just as there are also logical limits on his power: he can’t make a stone so heavy it cannot be lifted). Evil god can still be maximally evil – as evil as it is logically possible to be. We have not yet established a contradiction in the notion of a maximally evil being.
There is, in any case, a more general point to be made about arguments attempting to show that an evil god is an impossibility and that the evil god challenge is thus met. The point is this: even supposing an evil god is, for some reason X, an impossibility, we can still ask the hypothetical question: setting aside the fact that so-and-so establishes that an evil god is an impossibility, how reasonable would it otherwise be to suppose such an evil being exists? If the answer is ‘highly unreasonable’, i.e. because of the problem of good, then the evil god challenge can still be run. We can still ask the theist to explain why, if they would otherwise reject the evil god hypothesis is highly unreasonable, do they not take the same view regarding the good god hypothesis?

4. Arguments from simplicity
What if the good god hypothesis is significantly simpler than the evil god hypothesis?
For example, we might suggest that a good god can be defined in a simple way, e.g. as possessing every positive attribute. As goodness is a positive attribute, it follows this god is good. The concept of an evil god, by contrast, is more complex, for he possesses both positive attributes (omniscience and omnipotence) and negative attributes (evil). Principles of parsimony require, then, that we favour the good god over the evil god hypothesis.
I acknowledge that there may indeed be asymmetries between the good and evil god hypotheses in terms of simplicity and economy. However, note that the fact that one theory is much more economical than another lends it little additional credibility if what evidence (and other considerations pertaining to reasonableness) there is overwhelming favours the view that both theories are false.
Take, for example, these two hypotheses: (i) Swindon is populated with one thousand elves, and (ii) Swindon is populated with one thousand elves, each of which has a fairy sitting on its head. The first hypothesis is more economical, as it posits half as many entities as the first. But is the first hypothesis significantly more reasonable than the second? No. For not only is there little reason to suppose either hypothesis is true, there is overwhelming evidence both are false.
Similarly, if the reasonableness of both the good and the evil god hypotheses is very low, pointing out that one hypothesis is rather more economical than the other does little to raise the probability of one hypothesis with respect to the other. The suggestion that the two hypotheses are more or less equally unreasonable remains unthreatened.

Conclusion

The focus of this paper has been on the evil god challenge: the challenge of explaining why the good god hypothesis should be considered significantly more reasonable than the evil god hypothesis. We have examined several of the most popular arguments for the existence of a good god and found they appear to provide little if any more support for the good god hypothesis than they do the evil god hypothesis. We have also seen that many of the theodicies offered by theists to deal with the problem of evil are mirrored by reverse theodicies that can then be applied to the problem of good. Prima facie, our two sets of scales seem to balance out in much the same way.
Now I do not claim that the symmetry thesis is true, and that the evil god challenge cannot be met. But it seems to me that it is a challenge that deserves to be taken seriously. The problem facing defenders of classical monotheism is this: until they can provide good grounds for supposing the symmetry thesis is false, they lack good grounds for supposing the evil god hypothesis is any more reasonable than the evil god hypothesis – the latter hypothesis being something that, surely, even they will admit is very unreasonable indeed.
While I acknowledge the possibility that the evil god challenge might yet be met, I cannot myself see how. Perhaps there are grounds for supposing the universe was created by an intelligent being. But, at this point in time, the suggestion that this being is omnipotent, omniscient and maximally good seems to me hardly more reasonable than the suggestion that he is omnipotent, omniscient and maximally evil.