The following text was read at the VIIIth International Congress of Phonetic Sciences, held at the University of Leeds in August 1972. Its contents gave rise to quite a lively discussion, but it was never published. I think the problem still calls for a close scrutiny, particularly at a time when the Association is reconsidering its aims and procedures. Speaking from my experience, I based my argument on the problems of teaching pronunciation of second languages between the two World Wars. This was one of the more important tasks of phoneticians, especially in the UK and, to a lesser extent, in the USA; the results were excellent, often outstanding. Under the leadership of Daniel Jones, graduates from UCL acquired a thorough knowledge of spoken French, achieving a near perfection in articulation and intonation. Now, this was the avowed aim of the Department of Phonetics at UCL and other universities, which reflected on the curriculum and teaching procedures of grammar schools and high schools. An important dose of phonetic and phonemic theory was included in the course, to supply a basis for memorizing morphophonemic rules and the interface between sounds and spelling; some contrastive work was also included, contrasting English and French phonetic structures. It seems to me that this approach has now lost some of its appeal, and although phonemic transcription has gained a wider recognition since 1925, the overall aim of teachers is an acceptable approximation rather than a strict imitation. I submit this text therefore in the hope that it will be examined by the Council and the membership, to see whether it still has merit, and if so, to examine what the modern science of applied linguistics can do about it. It will be obvious to all that such teaching is closely associated with problems of transcription involving the choice and use of IPA symbols very much in the way envisaged by Peter Ladefoged in his recent circular letter.