used in the customary speech of some people for four-legged animals, in that of others for horses, and in that of still others for donkeys. The third category is exemplified by the words ghā’it, za’īnah, rāwiyah (water-carrier), and mazādah (waterskin). Linguistically, ghā’it denotes low-lying ground. Since people would go there to relieve themselves, they named what comes out of a person after the place where it was deposited. Za’īnah is the name of a riding animal; they then applied its name to the woman riding it. There are other similar examples.

The point is that this customary literal usage did not become literal through a group’s agreeing to transfer the expression to a new meaning. Rather, some people used it intending that customary meaning; its use then spread, and it became customary literal usage through this use. This is why some of them added the qualification “in the language in which communication takes place” to their definition of literal usage. Moreover, they themselves recognize and state that a particular usage may come to predominate for certain expressions, so that the customary meaning becomes their best-known meaning, and, when used without qualification, they signify nothing else. Customary literal usage thus supersedes linguistic literal usage. Yet the expression is literal in this newly developed customary usage without having previously been assigned to the meaning in which it is used. It is therefore clear that defining literal usage in this way is unsound.

If they say, “By ‘the meaning assigned to it,’ we mean the meaning in which it was first used,” the reply is: How can it be known that these expressions, which the Arabs used to communicate when the Quran was revealed and before then, had not previously been used for some other meaning? If they cannot establish that no such earlier usage existed, they cannot know that these are literal usages. This contradicts what they themselves agree upon. Furthermore, this would entail that no expression could be known with certainty to be literal—a conclusion no rational person would accept.