38 The Loom of Language advocated the direct method* and fooled some teachers into taking it up. The most apparent reason for its vogue is that it exempts the teacher from having any intelligent understanding of the language which he or she is teaching. Common experience shows that adult immigrants left to pick up the language of their adopted country by ear alone rarely learn to speak or to write correctly; and adults who wish to learn the lan- guage of another country rarely have the leisure to waste on time- consuming instruction of the type given in urban schools where insipid pictures of rural scenes mollify the tedium of repetitive conversation Because the kind of grammar you most need depends partly on bow you intend to use a language, it is impossible to give a general recipe for writing a compact and useful grammar-book The learner who wishes to get as far as possible with as little inconvenience generally has to pick and choose from books which contain more than enough To do this intelligently is easier if we start with a general idea of how lan- guages differ. The relative importance of rules of grammar depends, among other things, on whether the language one is learning more or less closely resembles one's own or another already mastered, and if so, in what way. If we aim at learning to write a modern language, the formal grammar of conjugations and declensions explained in Chapters III and IV usually boils down to a comparatively small number of rules, far fewer than those given in most primers On the other hand, few except the more advanced text-books have much to say about other equally important rules* One class of such rules already mentioned depends on the fact that each language or group of dosely related languages has its own characteristic types of derivative words. Thus reader and builder? childhood and widowhood, reshape, rebuild^ restate and fellowship, king- shtp> illustrate four ways of building new words in English and m other Teutonic languages. Such rules may be as useful as the rules for forming such derivatives as father's If two languages are dosely related as arc Swedish and English, or Spanish and Italian, it is also helpful to know rules which tell us how * The silliness of the direct method when tried out on adults was pointed out by Henry Sweet in 1899 "The fundamental objection, rhen> to the natural method is that it puts the adult into the position of an infant, which he is no longer capable of utilizing, and, at the same time3 docs not allow him to make use of his own special advantages These advantages are, as we have seen, the power of analysis and generalizationin short, the power of using a grammar and a dictionary."