Let us now look at a _method_ for learning a language or science. It is not given as _the method_ in use in the schools, but only as a means of showing that the thing proposed is possible. There should be for the beginner four papers for every lesson; or four kinds of work to be done. _First_—There should be a paper _stating principles to be learned_, and adding complete references to the text-book upon which they are based, that the student may add to his teacher’s dictum, the confirmation of his own research. Let it be distinctly noted that this paper is _to contain statements of principles to be learned_, and is not to be a mere budget of directions to paragraphs and sections of a text-book. The advantages to be gained by such a method of study are too obvious to need elaboration here. _Second_—There should be a paper giving abundant and apt illustrations of those principles, derived from the best sources, adapted to the pupil’s knowledge, and different from any which have been otherwise brought to his notice. These illustrations of principles should be memorized by the student and should form the basis of the paper containing the test of the student’s work. _Third_—There should be a paper giving examples for practice in these same principles; examples for transliteration, phonic representation, or translation in case of a foreign language, examples for experimentation, classification, or analysis in case of a science or literature. _Fourth_—There should be a paper of examination or question, for the purpose of testing the student and revealing the character of the work he has done. These questions should be framed with the utmost care and skill of which the teacher is master, and should act as a quickening impulse to the student. This paper should be in a sealed envelope, and should not be opened till all the work of the other three papers has been done, and the student feels that his lesson is learned. In addition to what has been suggested, there should be required in the study of language, as soon as the student can correctly pronounce, a regular exercise in memorizing from some standard author, and daily repetitions aloud of what is thus given to the memory. In the case of English, Latin, and Greek, this seems to me indispensable. This last suggestion, it will be noticed, contains a hint that the pronunciation of a language can be taught by correspondence. It is a hint which I am prepared to assert as a proposition, and to defend as far as the Latin and Greek languages are concerned. The amount of matter given in the lesson should be enough to require one week for its preparation by a student able to devote from one to two hours daily to study.