The Journal has received this newsletter which describes a well-developed system for the computer coding of phonetic symbols. Although the account shows a bias towards the transcription of English, the system does provide for an unusually wide range of prosodic codes and of diacritic symbols, including some for which the current International Phonetic Alphabet makes no provision. Given the Journal's active interest in the subject of computer coding of phonetic symbols, and the wish to encourage further debate and comment in our pages, we are publishing this particular newsletter in its entirety. In the discussion of the symbols, it is necessary to be able to distinguish ‘l’ (lower case ‘l’) from ‘i’ (upper case ‘i’). We have therefore used the serifed typeface of the original for the citation of the symbol forms (—Ed.).

Ear training is an essential element in any phonetics or pronunciation teaching. In recent years I have been developing an ear-training method based on repeater tapes (each line repeated nine or ten times) of ‘unusual’ personalities and accents. The recordings are first transcribed by rule of thumb according to our own norm (RP English, Standard Dutch, etc.) and the idea is that the major differences from this norm are then listened to and analysed. If the observations are summarized at the end, one has a fair idea of what is characteristic of that particular type of speech. By way of illustration I shall present such a summary of 85 seconds of speech by The Rt Hon. Margaret Thatcher, the British Prime Minister, taken from an interview on Dutch television made in 1982. It will appear that we do not only analyse phonetic phenomena, but cannot help noticing in this case their semantic impact as well. Perhaps our observations may be of some interest to semanticists and psychologists as well as phoneticians.