The notation of the General British English segments 

(This is only very trivially adapted from the article which appeared at pp 59-66 of the Journal of the International Phonetic Association in 1972)

Phonetic notation of segmental sounds is of two main types, phonemic transcription and phonetic transcription proper which, to avoid ambiguity, is often referred to as allophonic transcription. The two most important discussions of types of notation for English were Jones 1956: Appendix A and Abercrombie 1964; there were further comments at Windsor Lewis 1969 Chap. III and Gimson 1970 § 5.6. Of the many different phonetic notations used for English more or less all of them in the last half century or so have employed the symbols of the International Phonetic Association's alphabet, in Britain usually faithfully observing the principles laid down by the Association for their use. In America the principles have not been so generally accepted. They were set forth importantly first in Jones, 1949 and subsequently (with some reconsiderations) in the IPA Handbook in 1999.

The present discussion will be mainly concerned with the choice of symbols for notations applied to the teaching of English as a foreign or additional language. There are two main purposes for such notations. The first of these is the teaching of the theory of the English segments in which case a sufficiently varied and 'precise' set of symbols has to be employed to admit of convenient comparison of English and other languages and also sometimes between the model variety and other accents of English. Such a notation forms a suitable basis in its phonemic form for elaboration into a general allophonic transcription. This may be called a 'comparative' transcription. Cf Abercrombie 1964.

The other main type of EFL transcription we may call the 'lexical' type since its main application is identifying for the user of dictionaries and glossaries etc the distribution of the phonemes in the lexical pronunciation of words. It is also the convenient type to require students of English pronunciation to use in transcribing specimens of connected speech. The value of exercises of this sort is that they direct attention to the important matters such as gradation and r-linking which must be mastered if learners aim to pronounce English in an idiomatic and unobtrusive, natural fashion. This is the type most suitable for those who use English as a foreign language for general purposes. For those who teach English and are thus acting as models of pronunciation for their pupils the 'comparative' type of transcription can be expected to contribute to their necessary awareness of the contrasts between the segments of their native language and those of their English target accent. For any particular community of general-use speakers an ideal notation would be as 'lexical' as possible while avoiding any symbolisation which might be particularly likely to lead to error. Regionally adapted transcriptions have been advocated by some eg  Martinet 1946 and Enqvist 1961. Another example is to be seen in Windsor Lewis 1969 which was aimed especially at Scandinavian teachers of English and therefore, to avoid the very powerful associations for the majority of Scandinavians of the symbol /u:/ which in their alphabets very often represents a very front vowel, shows /oː/ for the General British close back vowel.

(I) Comparative Notations

The best known comparative notation, if we exclude the EPD transcription from this category, is that of Gimson 1970 etc, still in Cruttenden's revisions the standard description of contemporary British English pronunciation. It was the direct descendant of what Daniel Jones called 'narrow' transcription. This he devised in 1916 and demonstrated in various contributions to Le Maître Phonétique chiefly from 1923 onwards for some years. It was at first best known in the joint and individual work of Lilias Armstrong and Ida Ward (Armstrong 1923; Armstrong-Ward 1926; Ward 1929 etc). Their use of it was phonemic as regards segment qualities but allophonic as regards vowel quantity, but those who have since taken it up have mainly used it in a strictly phonemic form. Gimson and others, at University College London especially, generally kept the length-mark as part of an invariable composite symbol; Abercrombie and others, especially at Edinburgh University, dispensed with quantity indicators altogether. See Abercrombie 1964, Arnold-Gimson 1965 and numerous items in especially the final decade of Le Maître Phonétique.

The Gimson 1970 and Abercrombie 1964 notations differed only slightly. Their most obvious contrast was in their choice of symbols for the two centralized close-vowel phonemes and the latter elements of the closing diphthongs. Gimson 1970 retained the original small-capital / ɪ / and the similar-to-small-capital / ʊ /, the latter remodelled more positively than the former into harmony with the lower-case roman type series by adaptation to an ampulla-like shape. Abercrombie (1964) followed PIPA 1949 in which Jones put forward the Association's (really only his) new symbols recommended for such vowels, the straightened italic / ɩ / and the kidney-shaped /ɷ /. (Abercrombie once remarked to the writer that the dotless roman / ı / which appeared instead of / ɩ / in Abercrombie (1964) and was excluded from PIPA 1949 was not the form he would have preferred to be used in that book.) As regards these variants, although both of the PIPA 1949 preferred symbols were perhaps aesthetically the more attractive, neither of them was more successful from a practical point of view. The older symbols seemed to be more easily identified and remembered by the non-specialist. In small or worn founts of type they seemed to have better legibility. This was apparently particularly so for / ɩ /. Jones mentioned his disillusionment with the new forms in conversation with the writer in 1965, especially over difficulties which had arisen from the slightness of differentiation between / ɩ / and / t /. A third practical advantage of / ɪ / and / ʊ / in that era was that when printers without stocks of special types had to handle copy containing them they could usually manage rather more easily to produce adequate substitutes. [Postscript: In the 1990s the Association reversed Jones's PIPA 1949 decree and re-instated ɪ and ʊ.]

The other contrasts between the two notations were not so much typographical as dependent upon interpretation either of the PIPA recommendations (some of which were necessarily mutually incompatible at times) or of the question of what exact type of English accent was being represented.

In the case of the Abercrombie 1964 / ɛ / versus the Gimson 1970 / e / one imagines that, rather than representing a different quality norm (Gimson 1970 Fig. 11 placed the norm fractionally above the mid line; Abercrombie indicated his personal usage to the writer as about mid) the Abercrombie choice was motivated by a desire to exhibit the contrast between this vowel and the closer initial quality of the diphthong /eɪ/. Gimson 1970, on the other hand, no doubt observed the PIPA 1949 (p.7 §20) recommendation for 'broad transcriptions' to use the 'nearest appropriate roman letter'.

In the case of the contrast Abercrombie 1964 / a / versus Gimson 1970 / ӕ /, it was Abercrombie 1964 which followed PIPA 1949 more closely: see PIPA 1949 §17. Additionally there seems to be some difference of opinion on what value to take as a norm. Gimson 1970 placed it 'just below the half-open position' whereas Abercrombie, besides himself using a value apparently below semi-half-open (for this term and 'mid' see Windsor Lewis 1969 Fig. 5), commented on the tendency of this vowel to an opener value in connection with his choice of this symbol (Abercrombie1964:37). Adherence to the letter of the PIPA 1949 recommendations in Gimson 1970 would have meant in this case adoption of / ɛ / which, in a book intended partly for the general EFL public, would perhaps have been rather unhelpful. Some American and Continental phoneticians have in other contexts quite justifiably employed / ɛ / for this vowel, eg Jassem (1950). Another reason why Gimson 1970 was strictly speaking not in accordance with the Association's official recommendations was probably connected with the difficulty alluded to at PIPA 1949 §19 ¶38: 'The Association's treatment of a and ɑ as different letters denoting different sounds has not met with the success originally hoped for. In practice it is found that authors and printers still generally regard the two forms as variants of the same letter.' Jones went on to suggest that æ did not involve this difficulty.

For the initial element of the diphthong indicated in Abercrombie 1964 as /oɷ/ and in Gimson 1970 as /əʊ/ there seemed to be some lack of agreement on the norm to be adopted. This time Gimson 1970 seemed to follow the PIPA 1949 recommendations the more strictly. It seemed unlikely that Abercrombie would consider this norm to be a back vowel, and although PIPA 1949 was not perhaps explicit on the point, it seems reasonable to take the diagrams there as implying that any central vowel not in unsuitable contrast with another should be represented, whether rounded or not, by ə. In fact a better theoretical alternative to ə than the Abercrombie (1964) o (which was presumably a traditionalism) might have been e. But / eʊ /, which could be taken to represent [ëʊ], would have obviously been unpopular with many for its suggestion of a 'conspicuous' variety of RP (see Windsor Lewis, 1969: 87-8). The average usage of the younger readers of national news bulletins on British radio and television in the seventies (excluding those educated abroad) was probably more front than the centre of the vowel area.

In the previous case, if the Abercrombie 1964 choice of symbol was essentially based on an opinion of what to take as a norm of pronunciation, then one has simply to register disagreement. Cf Wells 1964a. However, in the case of the initial element of /aʊ/, there was no question of such disagreement. The Abercrombie 1964 choice of symbol was justifiable on the principle of preference for roman letter-forms even if it were not so on the basis of the quality of the norm. The Gimson 1970 selection of norms for the initial elements of /aɪ/ and /ɑʊ/ showed that differentiation of them was based on placing of the [a] just behind the 'central' vertical division of the vowel diagram (cf Windsor Lewis, 1969 p.15 Fig. 4). Here again we may suppose that the Gimson 1970 preference was influenced by considerations of its large EFL audience. This consideration may not apply to the majority of this audience but at least the present writer, when teaching in Norway, found this distinction valuable: hence its adoption in Windsor Lewis 1969. Norwegians very often attempt both /aɪ/ and /ɑʊ/ with the same rather back initial quality which sounds perfectly acceptable for /ɑu/ but for /aɪ/ is well wide of any mainstream General British  target.

The final contrast, between the Abercrombie 1964 /ɒɪ/ and the Gimson 1970 /ɔɪ/, was again hardly a disagreement on norms. From the Gimson 1970 placing of the initial element 'between the half-open and open positions' (p. 131) and diagram (Fig. 23 p.132) showing it only fractionally higher than the exact mean between those positions we can see that the choice of symbol could easily have gone in the Abercrombie 1964 direction if there had been good reason. Here one imagines that the PIPA principle of preference for the more familiar symbol was operative. The symbol ɒ was the least familiar and least desirable (particularly because least distinctive) item in the whole Gimson 1970 repertoire, so it was natural to avoid it. Abercrombie remarked privately to the writer that his choice was motivated by the convenience of avoiding the necessity of disambiguating the sequence /ɔɪ/ when it was equivalent to the Gimson 1970 / ɔ:ɪ / as in pawing versus point. Cf Wells 1964b p.31. It is undeniably more convenient to be able to avoid using a syllable-separation sign. However, there are fortunately relatively few occasions on which it is necessary either to insert the PIPA 1949 hyphen in /ɔ-ɩ/, or to extend the Abercrombie 1964 interpretation of the PIPA 1949 ruling on the use of the syllabic sign to embrace /ɔɩ/ with a syllabicity mark underneath its latter symbol. Syllable-separation or syllabicity markers are in any case unavoidable in other comparable cases such as /hə`wɑ-ɩ/ or /hə`wɑɩ/ Hawaii, /`sɑ-ɩb/ or /`sɑɩb/ sahib, /`lɑ-ʊs/ or /`lɑʊs/ Laos and /kə`kɑ-ʊ/ or /kə 'kaʊ/ cacao. In my Concise Pronouncing Dictionary, there were two other powerful reasons for choosing /ɔɪ/ rather than /ɒɪ/. Firstly, since Gimson 1970 had been fully established as the standard up-to-date description of British pronunciation, it was obviously desirable to conform as closely as possible to its version of 'narrow IPA for British English'. Secondly, since the CPD represented General British and General American pronunciation as far as possible with the same symbols, the symbolization /ɔɪ/ was more suitable, in terms of the IPA principles, for the American range of values for the first element of this diphthong.

These were the only two 'comparative' transcriptions of any currency and, to sum up their claims on our attention in an EFL context, in most respects the Gimson transcription was likely to be the more useful, and of course Abercrombie 1964 made it quite clear that its transcription was not aimed at the EFL public but at English-speaking university students. On the other hand the Gimson 1970 transcription had few if any disadvantages for such people who needed a 'comparative' notation. It could easily be used without its dispensable colons. Perhaps the least desirable complication it contained, as also did the Abercrombie 1964, although completely justifiable on a norms basis, was the use of ɛ as the first element of /ɛə/, which might well have been dropped in favour of e if only on the principles of preference for roman and familiar letters. In fact this modification to the transcription was employed from Gimson 1962 onwards.

(II) Lexical Notations

The best known earlier twentieth century transcription of all for English, the 'EPD' transcription, although so familiar in its lexicographical application in Jones 1963 etc, was not at all properly admissible to our 'lexical' category. It was even to some extent allophonic until practically all traces of this were removed from Jones-Gimson 1967. Jones 1956 referred to its being allophonic 'in one respect only' (p. 340) viz [o] for the monophthongal allophone of / ou /, but it also contained symbolizations such as ç, æː, ɱ, eə, oə etc for non-foreign pronunciations. Cf Jones-Gimson, p. viii. Jones was well aware of its lexicographical imperfections in being unnecessarily comparative. For the whole fifty years of his life which remained after the initial publication of the EPD, he seems never to have used EPD transcription in his contributions to Le Maître Phonétique or in any completely new book. He made various references to the superiority for lexical-type applications of such transcriptions as appeared first in Scott 1942 and MacCarthy 1944 and 1945. These latter types with their maximum typographic simplicity are certainly highly convenient in lexical applications for those whose mother tongue is English. On the other hand they have one very grave drawback in the EFL context: they fail to warn against, perhaps even foster, the misconception regarding the English segments to be found among EFL learners, the impression that the pairs /iː/ and /i/, /oː/ and /o/ and /uː/ and /u/ differ in length alone. This emphasizes a type of segment length contrast which is very often hardly perceptible as a signal of segment identification in connected speech. It is at least nothing like so powerful as the vital vowel length contrast which contributes to the distinction between syllables closed by fortis and lenis consonants. Cf Jones-Gimson 1967 p. viii §7 and Hill 1957 etc.

The most valuable EFL lexical-type notation therefore must preserve the excellent features of simplicity these transcriptions possess while avoiding their undesirable emphasis on quantity contrasts in the General British vowel system. These are relatively slight in General British and even slighter in General American. A convenient way to fulfil this purpose to best advantage was to employ a set of symbols which followed Gimson 1970 as closely as possible but avoided the symbolisations necessary for the comparative context of that book which could possibly be dispensed with in lexical contexts. Such a repertory of symbols differed from that of Gimson 1970 in four ways:

(i) Elimination of colons.

(ii) Replacement of /ɒ/ by /o/.

This was in accordance with PIPA 1949 p.7 §7 which said: As sounds of the ɒ type have considerable acoustic similarity to those in the ɔ and ɑ areas, it is generally advisable to represent them by one of these more familiar symbols, or in some cases by one of the ordinary roman letters o or a.' Here we had one of the special cases referred to because ɔ was already bespoke and a would not have been more suitable even if it had not been desirable to eliminate dependence on the 'unsuccessful' differentiation of a and ɑ.

(iii) Elimination of the differential use of a and ɑ.

The only reason for preferring a to ɑ would be that it is the normal printed form of lower-case A. Since the reverse is true of handwriting forms there is very little to choose between them. For the benefit of the EFL minority who need to master a comparative as well as a lexical transcription the balance is tipped in favour of keeping the Gimson 1970 / ɑʊ / and / ɑː / (colon excepted) letter forms and adopting /ɑɪ/ in place of Gimson 1970 /aɪ/. Compare the symbolization /ɑi/ of Jones 1950 in which the two diphthongs are diagrammed at Fig. 13 virtually exactly as they were in the Gimson 1970 norms Figs. 22 and 25.

(iv) Replacement of /ɛə/ with //.

This transcription without colons and using /o/, /ɑɪ/ and /eə/ was employed in my Concise Pronouncing Dictionary of British and American English.

(There is no objection to the replacement of italic /ɑ/ with non-italic /a/ in using this notation if ɑ is not available. Whichever is used, it should always contrast clearly with /æ/.)


ABERCROMBIE, D. (1953). 'Phonetic transcriptions" Le Maître Phonétique 32-4.

ABERCROMBIE, D. (1964). English Phonetic Texts. London: Faber.

ABMSTRONG, L. E. (1923). An English Phonetic Reader. London: University of London Press.

ARMSTRONG, L. E., and WARD, I. C. (1926). A Handbook of English Intonation ('Armstrong-Ward'). Cambridge: Helfer.

ARNOLD, G. F. & GIMSON, A. C. (1965). English Pronunciation Practice. London: University of London Press.

ENQVIST, N. E. (1961). The choice of transcription in foreign language teaching Proceedings of the Fourth International Congress of Phonetic Sciences 586-9. The Hague: Mouton.

GIMSON, A. C. (1970). An Introduction to the Pronunciation of English. London: Edward Arnold.

GIMSON, A. C. (1962). 'The Transmission of language' (supplement to R. Quirk. The Use of English). London: Longmans.

HILL, L. A. (1957). 'The transcription of English', Le Maître Phonétique, 5-6.

JASSEM, W. (1950). 'Phonemic transcription of the vowels of educated Southern English', Le Maître Phonétique 10-12.

JONES, D. (1949). The Principles of the International Phonetic Association (PIPA) London: IPA.

JONES, D. (1950).The Pronunciation of English. Cambridge: Cambridge University Press.

JONES. D. (1956). An Outline of English Phonetics. Cambridge: Heffer.

JOIU:S, D. (196:1). English Pronouncing Dictionary ('EPD'). London:Dent

JONES, D., and GIMSON, A. C. (1967). English Pronouncing Dictionary. London:Dent

MACCARTHY, P. A. D. (1944). English Pronunciation. Cambridge: Heffer

MACCARTHY, P. A. D. (1945). An English Pronouncing Vocabulary. Cambridge:Heffer .

SCOTT, N. C. (1942). English Conversations. Cambridge: Heffer

WARD, I. C. (1929). The Phonetics of English. Cambridge: Heffer

WELLS, J. C. (l964a). Review of Tibbitts: Practice Material for the English Sounds (1963), Le Maître Phonétique, 13-14.

WELLS, J. C. (l964b). Review of Abercrombie (1964), Le Maître Phonétique 29-31.