Two Theories of Analogical Predication∗

How can we, poor, finite beings that we are, speak of an infinite God? Thomas Aquinas formulates a well-known answer: we predicate things of God analogically. But the doctrine of analogical predication, long influential in both Catholic and Protestant thought, has fallen on hard times. Its advocates have been accused of circularity, begging the question, and simply being uninformative. (“God’s faithfulness is not the same as the dog’s faithfulness, but it’s similar.” “How is it similar? How is it different?” “I don’t know.”) They have even been accused of agnosticism for being incapable of explaining how true statements about God and knowledge of God are possible.1 I am arguing that those complaints rest on interpreting ‘analogy’ in a familiar but ultimately inadequate way. In this paper I shall elaborate two versions of the doctrine of analogical predication. One rests on the idea of structural similarity.2 We can legitimately draw an analogy between a representation and what it represents. Properties of the representation provide grounds for inferring corresponding properties of what it represents. This is an aspect of a more general phenomenon: Two things may have similar structures, or be elements in similar structures, in a way that supports analogical inferences and, correspondingly, analogical predication. The second rests on the closely related idea of approximation or idealization. We can construct models that approximate or idealize a more complex domain, and use properties of items in the models to infer corresponding properties of items in that domain. Both understandings of analogy show that analogical predications can be