subsequent transcription, renders some sort of improvement urgently needful; nor are these wastes the only grievance, as the introduction of a second personality into the operation of recording speech introduces a simultaneous possibility of error, and an outrageous waste of time is caused by the necessity of reading over what one has dictated laboriously to a stenographer or into a phonograph, to make sure that it is correctly transcribed. It is obviously a much more difficult matter to translate speech directly into printed words than to translate it into something which may again produce the sounds of speech. The first step would be the invention of something which would print a phonetic representation of speech—as, for instance, shorthand of the kind invented by Sir Isaac Pitman. Even this requires us to imagine machinery of a kind whose very rudiments do not at present exist. Indeed, we can only conceive such an instrument by the use of the supposition that some entirely new manipulation of sound-waves will be discovered; and if we conceive that, there is no particular reason why we should hesitate before the notion of speech directly translated into print such as we use in everyday life. If we are going to limit the possibilities of the future by the actual achievements of the present, we shall certainly fall short of any adequate notion of what a hundred years'
