Metaphor as a feature of sentence-readings

Professor Sadock's position is that metaphor is not a linguistic problem at all, because the mechanisms underlying metaphor exist independently of language. The use of “a lion on a warrior's shield,” he says, manifests “the same analogical urge that functions in the issuance and apprehension of verbal metaphors.” But it is vital here to distinguish metaphor, on the one hand, from similes and other nonmetaphorical analogies, on the other. There is certainly no special linguistic problem about such explicitly figurative sentences as (1) or (2):

1. (1)He is as brave as a lion.

2. (2)He is like a father to her.

The linguistic problem arises instead because of implicit figuration, as in (3) or (4):

1. (3)He is a lion in battle.

2. (4)The child is father to the man.

The problem is to explain how such sentences can have the meaning they do, when they are false or nonsensical if taken literally. In other words, the analogical urge operates sometimes in literal speech, sometimes in metaphorical. It can produce both and so cannot explain the difference between the two.

My own position, as against Sadock's, is that the fundamental problem about metaphor is a problem for our theory of langue, not for our theory of parole. But I speak here only about metaphor, and not about other figures such as irony, litotes, allegory, or simile.