I have chosen to start with this quotation for several reasons. It applies to language laboratories but has a broader purpose, to show the proper relationship of machines and man in general. In view of the extravagant claims that have been made for the language laboratory of the future, I wanted to set the reader's mind at rest immediately. There is no question but what machines of many kinds have an increasing role to play in education at all levels, but they are not going to take charge. They will upgrade the teacher, but not replace him. Just as in commerce and industry machines have changed the nature of man's labor, so in intellectual activities machines are making themselves felt. Electronic computers, which can do routine arithmetical operations with the speed of light, compress years of mathematical computations into hours, or minutes. This is well-known. But what of language? Language is a more personal part of the human being. The tongue is the mirror of the soul, if I may paraphrase a common saying. Everyone knows that his native language is a most intimate expression of his inner being, not to be tampered with without risk. The mere thought of a computer processing human language seems to threaten a violation of privacy. Nevertheless, computers are translating language, crudely, to be sure-no better than an elementary student-but the important point is that they are doing it as well as an elementary student. Computers are doing indexing, thousands of pages a year, with savings comparable to those from computation. Computer programs have also been written to make automatic abstracts of articles (not very satisfactorily). Others style text and compose pages by means of tape-operated photo-composing machines or typesetting machines. As we witness these encroachments on the citadel of language, it is comforting to retreat behind the thought that all these applications deal only with the finished product, as it were, of someone's mind, and this is true. In all these cases words are treated just as though they were strings of mathematical symbols. The semantic part of language remains untouched by machines. Yet in each case, a machine has taken over part of a job that used to be done by a man. The quality of the output may not yet be very good, but it is improving; and as it improves, more and more of man's routine intellectual operations can and will be done by machines; machines are cheaper, faster, and their mistakes are usually more obvious. Teaching is not a routine operation, though drill is. Teaching machines are drilling machines, but the program represents a high degree of teaching skill.