“It remains to mention some of the ways in which people have spoken misleadingly of logical form. One of the commonest of these is to talk of 'the logical form' of a statement; as if a statement could never have more than one kind of formal power; as if statements could, in respect of their formal powers, be grouped in mutually exclusive classes, like animals at a zoo in respect of their species. But to say that a statement is of some one logical form is simply to point to a certain general class of, e.g., valid inferences, in which the statement can play a certain role. It is not to exclude the possibility of there being other general classes of valid inferences in which the statement can play a certain role”