Further Reflections on Atheism 199further discussion of whether there is an intelligible concept of necessity
which will help the theist. For whether or not the theist accepts Anselm’s
proof, he or she is likely (I think understandably) to hold that a satisfactory
notion of God should include that of necessary existence, supposing that such
a concept is intelligible. Later in this essay I shall make further reflections on
the fine tuning argument (the contemporary form of the argument to design).
2 Anselm’s Argument
As I have said in FE, Descartes’ ontological argument was based on a defini-
tion of God as a being with all perfections. Of course if his ontological
argument were sound, as I denied in FE, God would not only exist but exist
as a matter of logical necessity. In Descartes’ ontological argument the idea of
God is said to be one of a supremely perfect being, i.e. one who has all
perfections. His argument would not be made stronger by introducing a
modal element and saying ‘allpossibleperfections’. Anselm in his Proslogion^1
argues for the existence of a being than which no greater can be conceived,
and so there is a modal element in Anselm’s proof. He also makes use of the
expressions ‘exists in reality’ and ‘exists in the understanding’. Anselm thinks
that there are two sorts of existence, merely mental existence and real exist-
ence. This way of talking invites confusion, though at this stage perhaps we
can leave open the question of whether Anselm was so confused. This will
come later. He might say that golden mountains do not exist in reality but
only in the understanding. It would be better to say that golden mountains
(or at least terrestrial ones) do not exist at all. What exist are golden moun-
tain ideas. A golden mountain idea is not a golden mountain, just as a picture
of a unicorn is a unicorn picture, not a unicorn. So we should talk of X-ideas
as existing in the mind or understanding, not of (in general) X being in the
mind or understanding. Descartes made a similar mistake in one of his argu-
ments (not his ontological argument) which was based on the (perhaps dubious)
principle that there must be as much reality in the cause as in the effect. He
thus reasoned that there must be as much reality in the cause of our idea of
a perfect being as there is in the idea. This mistake was over and above that
implicit in the notion of degrees of reality also implicit in the argument. If we
agree that the logical form of ‘is real’ is just that of ‘there is a’ we cannot say
that one thing is more real than another. We might be able to distinguish
‘necessarily exists’ from ‘contingently exists’. This is a matter to be discussed
later in this essay.
It is doubtful whether Anselm’s ‘greater than can be thought’ can be recon-
ciled with Quine’s minimalist account of modality that I mentioned on
FE p. 37. It ‘quantifies in’, i.e. it puts the word ‘can’ or ‘possibly’ inside the