/Pronunciation/--"Language," said Sayce, in 1879, "does not consist of letters, but of sounds, and until this fact has been brought home to us our study of it will be little better than an [Pg167] exercise of memory."[70] The theory, at that time, was somewhat strange to English grammarians and etymologists, despite the investigations of A. J. Ellis and the massive lesson of Grimm's law; their labors were largely wasted upon deductions from the written word. But since then, chiefly under the influence of Continental philologists, and particularly of the Dane, J. O. H. Jespersen, they have turned from orthographical futilities to the actual sounds of the tongue, and the latest and best grammar of it, that of Sweet, is frankly based upon the spoken English of educated Englishmen--not, remember, of conscious purists, but of the general body of cultivated folk. Unluckily, this new method also has its disadvantages. The men of a given race and time usually write a good deal alike, or, at all events, attempt to write alike, but in their oral speech there are wide variations. "No two persons," says a leading contemporary authority upon English phonetics,[71] "pronounce exactly alike." Moreover, "even the best speaker commonly uses more than one style." The result is that it is extremely difficult to determine the prevailing pronunciation of a given combination of letters at any time and place. The persons whose speech is studied pronounce it with minute shades of difference, and admit other differences according as they are conversing naturally or endeavoring to exhibit their pronunciation. Worse, it is impossible to represent a great many of these shades in print. Sweet, trying to do it,[72] found himself, in the end, with a preposterous alphabet of 125 letters. Prince L.-L. Bonaparte more than doubled this number, and Ellis brought it to 390.[73] Other phonologists, English and Continental, have gone floundering into the same bog. The dictionary-makers, forced to a far greater economy of means, are brought into obscurity. The difficulties of the enterprise, in fact, are probably unsurmountable. It is, as White says, "almost impossible for one person to express to another by signs the [Pg168] sound of any word." "Only the voice," he goes on, "is capable of that; for the moment a sign is used the question arises, What is the value of that sign? The sounds of words are the most delicate, fleeting and inapprehensible things in nature.... Moreover, the question arises as to the capability to apprehend and distinguish sounds on the part of the person whose evidence is given."[74] Certain German orthoepists, despairing of the printed page, have turned to the phonograph, and there is a Deutsche Grammophon-Gesellschaft in Berlin which offers records of specimen speeches in a great many languages and dialects, including English. The phonograph has also been put to successful use in language teaching by various American correspondence schools.