NOTE: The following article is taken from The God Argument: The Case Against Religion and for Humanism, pp. 83-94.
The various versions of the ontological argument come down to saying in effect that deity exists by definition. It is an a priori deductive argument, that is, an argument by reason alone, and it turns on analysis of the concept of deity. This, therefore, is where the problems with defining the word ‘god’ really bite.
The classic statement of the ontological argument is given by the eleventh-century philosopher St Anselm in his Proslogion. He began by considering the concept of ‘a being than which no greater can be conceived’. If such a being did not exist, then there would be a ‘greater’ being than it, namely, one that did exist. But by hypothesis ‘the being than which no greater can be conceived’ is the greatest being there is. Therefore, it must exist. And Anselm identified this being as the deity, because of the other attributes of deity which together make it the only plausible candidate for the ‘greatest being’ – omnipotence, omniscience and the rest.
Leaving aside the undefined notion of ‘greatness’ for a moment, let us note the following. At any given moment, someone is the tallest person in London. This is a matter of logic, not of human physiology. If there are two, three or more tallest people in London who are exactly the same height, then whichever of them got out of bed latest on the day in question would be the tallest person in London, because gravity would have had slightly longer to act on the other two, shortening them fractionally. This latter fact is a matter of physiology and physics, but it connects to the logical fact that one of the Londoners – the richest, laziest or least healthy of them, because longest abed – is the tallest on that day. Finally, note that even if everyone in London were short, as a matter of logic one of them would nevertheless be the tallest.
Now consider the idea that someone is the ‘greatest’ person in London. Whatever that might mean, note that such a person need not be very great, he need only be less un-great than everyone else in London. One can now see why the Anselm argument does not get us from logic to divinity. If by ‘god’ Anselm means the least un-great individual anywhere, this is an uninteresting result. It does not improve matters to substitute the phrase ‘most perfect being’ for ‘greatest being’, as some do with the ontological argument. The universe’s most perfect being might be very imperfect, only less so than other imperfect beings, and not at all a suitable candidate for existence as a deity.
Note that the argument requires the comparative element; that is crucial; the claim cannot be ‘that there is a perfect being’, it must be ‘that there is a most perfect being’ or ‘that there is a being than which no other being is more perfect’, so that it can be more perfect than a perfect being which does not exist.
From the outset therefore there is the problem of getting from the supposed fact that something must have some property in the highest degree relative to any other similar thing, with it being the case that the highest degree is very high; and so without offering any ground for thinking that because it is the ‘most-est’ of its kind, it is a deity – let alone the traditionally conceived God.
It should also be noted, as a footnote, that there is a problem with the assumption made by proponents of some versions of the ontological argument that ‘perfection’ admits of degrees, making it a relative rather than an absolute notion. But there is a strong case for saying that ‘perfect’ is an absolute term, that is, applies in an all-or-nothing way. If something is perfect, then it is perfect, and cannot be more or less perfect than another perfect thing. It is legitimate to say that something might approach more nearly to perfection than other things do, but then by hypothesis none of the things being thus compared is perfect anyway. Yet the Anselm-type of ontological argument requires that perfection (or in Anselm’s terminology, greatness) be a matter of degree, otherwise the argument will not work.
Perhaps the ontological argument’s proponents take perfection to be a relative notion because it is an all too familiar fact that imperfection has degrees; not only are some things more imperfect than others, they can also become more and less so. But to say that something is less imperfect than something else is not to say that it is more perfect than it. It might be less imperfect in having one less flaw, and yet be extremely flawed.
The fact that ‘less imperfect’ is not the same as ‘more perfect’ is a function of the fact that the conceptual polarities ‘perfect-imperfect’ (and others like it, such as ‘mortal-immortal’) are mistakenly assimilated to such examples as ‘flat-bumpy’ and ‘calm-anxious’. In these latter cases we know what each of the contrasting concepts applies to; we can point out examples of flat things and examples of bumpy things. But in the case of ‘perfect-imperfect’ and ‘mortal-immortal’ we only know what one of the pair applies to (imperfection and mortality, of course). The opposite pole in each case is merely notional, arrived at by extrapolation from the concept we know how to apply.
So there is a problem with the Anselm type of ontological argument which relies on an existing something’s being more perfect than a non-existing perfect thing. Yet even if one leaves aside the question whether ‘perfect’ is an absolute or relative term, there is still a further problem. What does ‘perfect’ mean, as applied to deity? The formula ‘God is perfect’ in traditional theistic doctrine is intended to mean ‘omnipotent, omniscient, morally pure, without needs or appetites’ (though capable, according to the Bible, of emotions of anger and love). Arguably, though, these fall into the category of expressions which are sayable without being thinkable, like the example, given previously (p. 21), of the apparently intelligible sentence which expresses a logical impossibility.
For consider: ‘omnipotent’ means ‘all powerful’ in the sense of ‘is capable of doing anything’ or ‘is unlimited in action’. This immediately causes difficulties, well illustrated by nonsense questions such as, ‘Could an omnipotent being eat itself?’ Suppose the reply is that such a being is not the kind of being that eats, because it is immaterial. Does this mean that it cannot eat? If so it is not omnipotent: omnipotent means ‘can do anything’. If the answer is that it can eat but does not, then one can probe the coherence of the concept by asking, ‘What might it eat if it chose to?’ Alternatively, and more consistently, the first answer might be developed as follows: to say that an omnipotent being cannot eat because it is not the kind of thing that eats (compare: you cannot ‘sleep furiously’, because sleeping is not the kind of thing that can be done furiously), one is saying that its field of omnipotence is whatever is consistent with its nature. But this simply defers the difficulty again. We now need to know what its nature is to know the respects in which, within the limits of its nature, it is unlimited in power. Will this satisfy the sceptic? No, because this is already to say that the supposedly omnipotent being is only qualifiedly omnipotent – and that is a contradiction in terms.
These points are neither frivolous nor pedantic, because they show that pressure on the concept of deity quickly exposes incoherences, leaving its defenders only with the ineffability move for protecting their adherence to it. But the ineffability move cannot give us an ontological argument, which crucially depends on assertions about the nature of the deity.
Notice that these thoughts, if they cannot be answered, undermine the ontological argument even before its details are examined. But let us examine them anyway; and in a stronger form that does not require a comparative notion of perfection.
The argument’s most familiar version is given by René Descartes in the fifth of his Meditations (1641). His version has it that the concept of a non-existent ‘supremely perfect being’ is a contradiction, just as it is a contradiction to deny that the interior angles of a plane triangle add up to 180 degrees. Accordingly, because we can conceive of a ‘supremely perfect being’, it follows from the very definition of it that it necessarily exists.
The definitive response was stated by Immanuel Kant in the ‘Dialectic’ of his Critique of Pure Reason (1787), which is the section of that famous work devoted to exploring how reason can go wrong, as happens in the ontological argument. He pointed out that ‘existence’ is not a property of anything, but a condition of anything’s having a property. In Descartes’ statement of the argument, existence is a perfection which deity cannot lack, and it is therefore a property among the other superlative properties ascribable to the deity. But, said Kant, any possessor of properties cannot have its own existence as one among those properties; it must (so to speak, already) exist in order to be a possessor of properties. You cannot say of a table, ‘It is made of wood, has three legs, is round, and it exists,’ for it might have properties different from being wooden, three-legged and round, while still being a table – a metal, four-legged, square table perhaps – but it could not be non-existent and still be a table.
The point is illustrated by noting that if one thinks Descartes’s form of the argument works, a parallel version of it can be used to prove that a devil must be necessarily non-existent. It would go as follows: ‘There is a being which is the least perfect of all beings; such a being which does not exist is – since existence is a perfection – less perfect than one that does; therefore the least perfect being necessarily does not exist.’ Here non-existence is asserted to be a property of a being whose other properties are, presumably, evil, malevolence, impurity and so forth: but one wonders how a non-existent thing can be evil and impure, thus demonstrating that existing is a logically different category from, because a logically prior category to, any properties anything might possess.
A version of the ontological argument is offered by the American theist philosopher Alvin Plantinga, who does not claim that it proves that a god exists, but that it establishes that it is rational to think that a god exists.9 His argument turns on a standard way of explaining the ‘modal’ concepts possibility and necessity. Something is said to be possible if there is at least one way a world could be – a ‘possible world’ – such that it exists in that world. A world is a possible world if it is either our actual world (to be actual it has at least to be possible) or is a non-actual world the concept of which is without internal contradictions. And then we say that something ‘exists necessarily’ if it exists in every possible world – which is merely a different way of saying: a ‘necessary something’ is a something that must exist no matter what else is the case.
Plantinga’s argument is as follows. There is a possible world in which something exists that is the greatest thing there can ever be (a thing which has ‘maximal greatness’). Therefore there is such a thing. And then Plantinga says this thing is god. As noted, Plantinga (wisely) does not take this to prove the existence of a god, but claims that it makes belief in a god rational.
Another approach in this style of reasoning is to say that there is a possible world in which there is a necessarily existing x; and therefore x exists. And as with the ‘greatest thing’ in Plantinga’s version, this necessarily existing thing is identified as a god.
Neither strategy works. The second formulation turns on a technical principle in modal logic: ‘If it is possible that it is necessary that p then, by a certain rule, one can infer that p is necessary.’ Here is the explanation: anything which is possible exists, by definition, in at least one possible world. If it is possible that there is a necessary x, then there is at least one world in which x exists necessarily. But if x is a necessary being – if it must exist and cannot do other than exist – then it must exist in every possible world, including the actual world. Therefore if it is possible that there is a necessary x, there is actually a necessary x.
Leave aside the question what such a thing would be, and why – given that it is only by stipulation that a god is a necessary being – the necessary being in question is the God of theistic tradition, and ask: what reason is there for thinking that anything exists necessarily? That is, on what grounds is it claimed that it is possible that anything is necessary? In fact, the argument is question-begging, for by saying that there is a world in which something is necessary, by the definition of ‘necessary’ what is thereby being asserted is that it has to exist in every possible world. Yet with equal plausibility it can be claimed that ‘there is a possible world in which nothing exists necessarily’ – which means ‘there is a possible world in which everything is contingent’ – and if this is possible (as it surely is: our own world is such a world!) then it follows that nothing is necessary, because only if it is not possible for there to be a world in which nothing is necessary can there be any necessarily existing thing – for remember: such a thing would have to exist in every possible world.
The first version of Plantinga’s argument, which starts from the premise that ‘there is a possible world containing a maximally great entity’ is vulnerable to the challenge that one can equally start from the premise that there is no possible world in which anything is maximally great, from which it would follow that necessarily there is no maximally great thing. Are there grounds for preferring one of these starting premises to the other one? Arguably, given the problem, discussed above, questioning whether the phrase ‘maximal greatness’ means anything, it is the second premise which is marginally more intelligible and therefore sensible. At the least this shows that you have to begin by accepting that there can be a ‘maximally great something’ for the argument to have any grip; and that of course is to argue in a circle.
It would seem that Alvin Plantinga has abandoned attempts to show by argument that it is rational to hold theistic beliefs, because he now argues that there is no need to provide such arguments, on the grounds that belief in the existence of a deity is a ‘basic belief’ from which one begins, not at which one ends by investigation and argument.10 By a ‘basic belief’ is meant such as ‘the past exists’, ‘other people have minds’, ‘one plus one is two’. So Plantinga is arguing that it is just as obvious, fundamental and unquestionable that ‘God exists’.
The least of the problems with this breathtaking assertion is that the supposed basic belief that ‘the past exists’ – and so for the other examples given – can and have regularly been challenged by rational sceptical argument, and yet they are a good deal less contentious than the claim that gods, goddesses and other supernatural beings exist, or that at least one such exists.
The main problem is that calling a belief ‘basic’, so that you do not have to argue for it or provide evidence for it, is gratuitous: you can help yourself to anything you like, and of course anything follows. Choose a convenient belief, give it the most convenient content for what else you wish to believe, and then claim that it is ‘basic’ and therefore in no need of justification. This is too obviously unacceptable to need much comment. As Daniel Dennett said of this view of Plantinga’s, this is ‘Exhibit A of how religious belief can damage or hinder or disable a philosopher’.
The claim that there is a deity with supernal powers has exactly the effect of claiming that a contradiction is true: nothing can be inconsistent with the existence of such a being, and therefore nothing can test whether or not it exists. It is yet again Popper’s dictum, ‘What explains everything explains nothing,’ which shows what is wrong with that.
In Plantinga’s view, the critiques of religious belief given by the likes of Richard Dawkins and Daniel Dennett are worthless: in an interview for the New York Times he described Dawkins as ‘dancing on the lunatic fringe’ and Dennett as engaging in ‘inane ridicule and burlesque’ rather than argument.11 Apart from this, by its nature, being a case of the pot calling the kettle black, it comes richly from someone who thinks we have no need to provide argument or evidence for a belief in deity. Two things that stand out in Plantinga’s claims are, first, that theism is more consistent with science than atheism because a universe ruled by a deity is an orderly one, and therefore fit for the operation of natural laws; whereas a universe not ruled by a deity would be disorderly and not fit for description by science. This is a bizarre view: it seems to imply that unless the universe had a ruler it would be naughty, with galaxies and stars disobeying the laws of gravity and the rest. The point Plantinga misses, in common with all apologists who wish to insert a deity into the picture, is that if the laws of nature describe the universe successfully, then it blunts Ockham’s Razor to bring in an unnecessary addition to the framework of explanation for this fact.
The second point that stands out from Plantinga’s views is his claim that everyone has a sensus divinitatis but that in some people – Dawkins, Dennett, the writer of these words – it ‘does not work properly’. The Latin phrase, literally ‘a sense of the divine’ but meaning ‘an innate conviction that god exists’, is used as a booster to the claim that belief in god is ‘basic’ and in no need of justification. Again, this would be a very convenient view for the theistic cause, because it applies to everyone without exception; the non-believer is told that his sensus divinitatis is not working properly, not that he has no such thing, and this is why he cannot accept that ‘there is a god’ is as basic a belief as ‘one plus one equals two’.
I repeat the quotation from W. K. Clifford: ‘It is wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence.’ It has to be said that by Clifford’s lights Plantinga’s approach provides an example of complete intellectual irresponsibility