In this paper the author discusses "error-analysis"; its emergence as a recognized technique in applied linguistics, with a function in the preparation of new or improved teaching materials; and its new place in relation to theories of language learning and language teaching. He believes that error-analysis has suddenly found a new importance, and from being mainly a technique for short-circuiting the practical difficulty of bilingual comparison it has become a vital source of information about the progress of a learner towards his eventual competence in the language and a crucial component in the search for adequate theories of language learning and language teaching. (Author/DO)