Changes in British English pronunciation during the twentieth century

It is as well to be aware that many changes over the years need not be taken to be irreversible but may be found rather to be swings of the pendulum.

I. Consonants

1. The consonant system of three pairs of plosives, four pairs of (supraglottal) fricatives, a pair of affricates, a trio of nasals, a glottal fricative, and four approximants (a pair of liquids and a pair of semivowels) is the most stable part of the English phonological system, having remained the same for centuries.
2. No doubt to a degree more easily accepted than previously in the atmosphere of greater social permissiveness since the sixties, relaxation of the force of articulation of intervocalic /t/ has been more widespread in colloquial styles as can be seen in occasional informal spellings such as gedd off and gerr off for get off .
3. On the other hand glottal reinforcement and even replacement of syllable-final /t/ seems more in evidence. It is completely ordinary for large numbers of especially somewhat younger speakers to use [ʔ] instead of [t] in large numbers of items such as atlas, outright, catwalk, get you, apartment, chutney etc where the /t/ precedes a lateral, approximant or nasal consonant. However, intervocalic replacement remains proscribed as dialectal, as in the jokey advertising catchphrase "a bi’ o’ be’er bu’er" ie a bit of better butter. Note that eg [geʔ] off can occur (as eg in the speech of Princess Diana) but only if a paralinguistic [ʔ] begins the second word thereby excluding the first [ʔ] from the intervocalic category. Compare the way schwa forms of the and to etc may often be heard (eg in to eat the apple) before words with initial-phonological-vowel structure but uttered with the (paralinguistic) consonant [ʔ] which of course is not an item in the phonological consonant inventory.
4. Failure to realise that paralinguistic and prosodic processes can cut across phonological "rules" has resulted in many complaints about broadcasters including the onetime favourite one that they have "wrongly" stressed prepositions.
5. The j-sound /ʤ/ is increasingly simplified non-initially to that of the middle consonant of pleasure /ᴣ/ especially in foreign or archaic words, eg adagio, liege, management (a type not in dicts), raj and words ending -fuge such as subterfuge, centrifuge though not often refuge and rarely if at all huge, gauge, savage etc. The placename Rugeley may at times be heard with /ʒ/ possibly influenced by association with rouge whose obvious extraneousness has meant that it's always had /ʒ/ unlike the equally exotic gamboge which seems largely to have acquired /ʤ/ as increasingly does garage when not so fully anglicised as to be /`gӕrɪʤ/. Sometimes simplification is to /d/ eg in dangerous or legislation (also a type not in LPD). Even the specialist pronouncing dictionaries can hardly be expected to record all such tendencies. In extraneous words /ʒ/ is often adopted in an attempt to produce a more "accurate" rendering of foreign sounds. Examples of this commonly attempted mistakenly are Azerbaijan, Beijing, Borgia, doge, Perugia, Sergio and word-initially with Gigli which is usually/`ʒiːli/.
6. Greater permissiveness has no doubt also encouraged the tendency to affricate the sequences /tj/ and /dj/ so that (a generation ago mainly only word-internally) they are so like the ch and j sounds that they now freely interchange with them. The current predominant forms of actually and gradually were never acknowledged in the Daniel Jones dictionary to be so until after Jones's death (in 1967). It's only recently that eg /`ʧuːzdeɪ/ has become widely recognised as fully permissible for Tuesday. LPD1 in 2000 labelled it as not "received" but this stigma was withdrawn in the 2008 LPD3. H. C. Wyld in 1921 at p.215 of his History of Modern Colloquial English had said that writing  "chewsdy" for Tuesday expressed "nothing different from the normal pronunciation" tho in his Universal Dictionary of 1932 (which did not aim at very detailed information on variant pronunciations) he showed the word only as pronounced /`tjuːzdɪ/.
7. The word actually may also commonly be heard with further weakenings of its articulation including having the ʧ as ʃ. This shows a parallel to the historical development of words like action which must be presumed to have reached their current form via an intermediate one with /ʧ/. This kind of development is commonly heard but not predominant with other words such as picture. (In GA it seems to be more common than in GB.)
8. More permissive attitudes seem also to account for the increased prevalence of so-called "intrusive r" (as in law /r/ and order etc: see Windsor Lewis 1975 §3.6) and in the yodless forms of many words like suit, sewer, superb etc though eg yodless assume, resume etc are definitely a minority usage with little sign of their spreading in England (though they seem to predominate in Scotland). Certainly very common words like supermarket now sound very old-fashioned if spoken with a yod. 
9. The BBC, or rather certain factions within the Corporation, have tended to put something of a brake on many developments partly because influential individuals exert pressure from time to time (the classic case being that dreadful old bully its one-time Director General John Reith) and partly because its highly-influential advisory Pronunciation Research Unit finds it leads to a life freer from tiresome ill-informed listener's complaints of "falling standards" if they eg promote `controversy and discourage con`troversy although there has been evidence for generations that the latter stressing was preferred by the majority of younger speakers.
10. The velar plosives /k/ and /g/ have been known to be subject to weak fricative articulation since Gimson first pointed out the fact in 1962 but it's a phenomenon that goes largely unrecognised, no doubt because it doesn't happen initially in strong syllables.
11. There are signs of a fast increasing minority tendency for General British speakers to favour the use of schwa plus (unsyllabic) consonant where previously syllabic consonants were the norm, eg in cotton, garden, bottle and struggle and even increasingly in such items as assembly, doubly, gambling, cackling etc for which it is doubtful that they ever previously contained a syllabic consonant. These last would strike many as strange but there can be no doubt about their increasing proliferation even though it doesn’t seem to have been much commented on as yet. There evidently has to be a related word with a syllabic consonant to trigger this so that eg duckling, madly, ugly, Wembley etc are not usually affected but eg buckler, burglar, butler, inkling, spindly, stickler etc may well soon be increasingly heard with this anaptyctic schwa by some GB speakers.

II. Vowels

1. The most noteworthy changes of the past century have occurred to vowels but not to the vowel system. Though some have prophesied the early demise of the traditional diphthong /ʊǝ/ of words such as cure, it still persists firmly though admittedly in a smaller number of words than half a century ago when it probably prevailed in moor, poor, sure, your and you're. It seems quite possible that spelling consciousness will maintain /ʊə/ indefinitely in most other words. The last unquestionable systemic change to the General British set of vowels was the loss of /ɔǝ/ the former diphthong of words like four which had disappeared from general use before the thirties (despite the OED2 representations), leaving the General British accent with a less neatly balanced set of centring diphthongs.
2. Th diphthong /ɔə/ appears in LPD (the Longman Pronounciation Dictionary by J. C. Wells) only with a sign (originally † later §) that signifies "widespread in England among educated speakers but ... nevertheless judged to fall outside RP". In OED2 (the second edition 1989 of the Oxford English Dictionary) it is shown as the only version of all words such as four and boarder though its compilers observe that the latter "in most varieties of southern British pronunciation has become identical" with border. At its Key to Pronunciation they said that "The pronunciations given are those in use in the educated speech of southern England (the so-called Received Standard)". The Concise Oxford Dictionary of 1990 gave identical versions of the two words. The Daniel Jones English Pronouncing Dictionary of 1991 still showed boarder with a diphthongal version in square brackets which indicated that it was current according to the editors, who moreover did not avail themselves of the option exercised at various other items to indicate it as "rare" or "old-fashioned". It has in fact always been extremely unusual to come across it as the usage of any radio or television newsreader whose speech did not also fairly markedly in other ways suggest the influence of some region of Great Britain. The new edition of 1997 abandoned the diphthongal representation entirely. OED3 Online (ongoing 2000-) likewise contains no GB /ɔə/ variants. 
3. In matters of (phonotactic) distributional changes it can no longer be said that GB /r/ occurs only before vowels. Many words, not only ones acknowledged to be so in the works of Daniel Jones like barrel and barren, but ones like authority, temperature, embarrassing, terrible, borrowing etc now have lexical forms for many speakers where this old rule no longer operates, as was pointed out at some length in Windsor Lewis 1979. Another rule now often broken is that closing diphthongs didn't occur before /r/ in the same morpheme but schwa intervened. This is not now true of eg Cairo, gyro, Irish, thyroid etc to give only examples where the newer version now predominates.
4. The most famous recent tendency of realisational change has affected rather few British speakers but among them a couple of very well known ones. His weakening of the latter elements of the two back-closing diphthongs as in home and now by Prince Charles was first caricatured by entertainers in the late seventies: they represented him as saying ite and abite arind the tine for out and about around the town. He plainly avoided this tendency in later years. His mother, though also often humorously imitated, has relatively rarely been caricatured as having the same tendency. Yet her Christmas broadcast of 1983 had a sentence beginning "I found it fascinating..." which sounded indistinguishable from " I find..." etc.
5. The major and almost universal change of the third quarter of the century was the lowering and backing of the "ash" vowel /æ/. In the fifties the average value was not as close as it seems to have been in the early decades of the century but the new opener value was as yet mainly the style of débutantes etc. However, by the mid seventies it had become so normal that the kind of quality Gimson diagrammed as the norm in the first (1962) edition of his Introduction had already by then become quite out of date. Old movies with eg John Mills as a young RAF officer saying things like eg that bad chap suggesting " thed bed chep" can now excite considerable mirth. This met a minor advancing movement of the cup vowel head on and sent it into reverse, though even now speakers are to be heard with regard to whom one may be uncertain whether the name they just uttered was eg Branson or Brunson.
6. The major realisational change of the first half of the century was the striking lurch forward of the first element of the diphthong of home etc. Only a small minority followed this most of the way to the front and by the sixties it went into reverse with anything more than slightly front of centre now sounding old-fashioned-posh. There is no trace of any reference to this occurring in the 19th century. Queen Elizabeth the Queen Mother belonged to the first generation to exhibit it. Her daughters had it but with some inconsistency. There was little trace of it in their children. Very few Americans have ever shown any such tendency. Gimson gave the lead very modestly to acknowledging the abandonment of the Victorian back-vowel value by substituting in 1967 in his first revision of the EPD a schwa symbol [ə] for the [o] that Jones had used in the first sixty years of the EPD to represent its first element.
7. The only other major realisational movement of the century had been gaining momentum by the last decade and is by now unremarkable. This is the tendency for the final unstressed vowel of words like happy to be identified more closely with the /i:/ of see etc. The main descriptions we have of the traditional Victorian/Jonesian and Gimsonian "Received" Pronunciation identify this sound with the vowel of sit, kit etc. However, they do not present an adequate picture even of nineteenth-century usage (see Windsor Lewis 1990). At any rate, for a long time more and more younger people have been using a quality far too close to be identified with the vowel of sit etc. The Oxford English Dictionary of 1989, and the EPD (the English Pronouncing Dictionary) of 1991 continued to use that symbol for it but not so the latest dictionaries the New Shorter Oxford English Dictionary 1993, The Oxford Advanced Learner's Dictionary 1995, the Concise Oxford Dictionary 1995 and the online third edition of the OED.
8. The mould was broken in 1978 by the Longman Dictionary of Contemporary English whose pronunciation editor, Gordon Walsh, sneaked in /i/ with the excuse that it was to be interpreted as either /i:/ or /ɪ/ according to whether the user was adopting a British or an American model of pronunciation. In truth it conveniently provided recognition of the fact that most speakers of minimally regionalisable English of England had come to aim at a quality for their final unstressed happy vowel which was too close to be associated with their /ɪ/ value.
9. When LPD was published in 1990 the value of the /i/ symbol was not directly equated with /ɪ/ but diagrammed with a latitude that overlaps its range and that of /i:/. It is now hardly to be challenged that a majority of non-elderly speakers of the most general varieties of the English of England have average values for /i/ near to their range of values for their /i:/. This is not to say that any but a minority of them use the "heavy" /i:/ value often accompanied by a degree of diphthongisation that is commonplace in the USA and predominant in Australasia and South Africa.

    10. Among other realisational matters are the following:

III. Individual Words

Our final changes are miscellaneous ones that have occurred over the years to a wide variety of lexical items. Probably the biggest group of these involve the re-introduction of a sound formerly lost or weakened but which the spelling retains. The obverse of this tendency (11 below) is for a pronunciation to be avoided because it is felt to be inappropriate in respect of its spelling: this "inverse spelling influence" probably accounted for the general disappearance in words like loss of the long vowel "proper" for words like horse.

Constant contrastive use of a word can focus attention on a previously weak syllable resulting in its acquisition of a strengthened value. On the other hand increasing familiarity with words that were formerly perceived as learned items has often led to weakened versions. In the case of many compound words, as the consciousness of the independent identities of their constituents has decreased, they have tended to be accorded a single initial accent instead of two or more major stresses. In some cases, general phonetic processes may be seen at work such as "Vernerism" as I have chosen to call the tendency for post-initial stressed syllables to be begun with voiced-type consonants. (The Danish linguistician Karl Verner 1846-1896 famously observed certain voicing processes in relation to word stress in Indo-European.)

Among the very considerable numbers of individual words that have undergone changes the following may be taken as typical examples in some of the more noteworthy categories. Some items may belong to more than one of these categories.

At any item the sign ˂ indicates that the form shown is not the predominant usage. 

1. Re-introduction of a consonant: forehead /`fɒrɪd→`fɔːhed/, Hertford /`hɑːfəd→`hɑːtfəd/, often /ɒfn→ɒftən/ (but contrast soften), Sandwich /`sӕnɪʤ→`sӕn(d)wɪʧ/, schism /`sɪzm→skɪzm/.
2. Restoration or introduction of a lost or stronger vowel: alphabet/-bɪt→bet/, bollard/-ləd→lɑːd/, boycott/-kət→kɒt/, breeches/`brɪʧɪz→`briːʧɪz/, consequences/`kɒnsɪkwənsɪz→ensɪz/, Elgar/-gə→gɑː/, fortune/-ʧən→ʧuːn/, magistrate /strɪt→streɪt/ metaphor/-fə→fɔː/, portrait/-trɪt→treɪt/, Somerset /-sɪt→set/, steadfast /-fəst→fɑːst/, synod /-əd→ɒd/, vacation /və-→veɪ-/. 
3. Notional spelling-value adoption: issue/`ɪʃu→ɪsju/ (though most recently in this word the tendency seems to have gone into reverse); laudanum/`lɒdnəm→`lɔːd-/, Lombardy/`lʌmbədi→lɒm-/, nephew /`nevju→`nefju/ (first recorded as predominant in LPD 1990),  retch  /riːʧ→reʧ/, strafe/strɑːf→streɪf/,  year/jɜː→jɪə/.
4. Re-modelling to presumed or actual original-language value (including "Continentalisations"): acoustic, armada, banal (EPD 1917 ei, ӕ ˚), bulimia˂, Cecilia, cortège, crêche, deity, demise˂, forte, Majorca, Lyons˂, Marseilles˂, memorabilia˂, Munich˂, niche, nihilism˂, quietus, Quixote, schizophrenia, Seville, spontaneity, strata, tête-à-tête, trauma, via, virago.
5. Re-modelling (anglicising) of items no longer perceived as extraneous: envelope, garage, gigolo, profile, questionnaire, restaurant, rucksack, sauna, ski, trait.
6. Re-modelling on the analogy of related or similar words: bastard/bæs-→bɑːs-/ (only /ӕ/ in EPD1), contrast ( /-ӕst/ (dropped since Jones's day) /→-ɑːst/,  `cervical→cer`vical, o`besity (OED1 in 1902 gave only /-es-/→ -`biːs-/), re`monstrate → `remonstrate  (after the pattern of `demonstrate), salve/sɑːv→sӕlv/, scenic /`senɪk→`siːnɪk/, se`cretive→`secretive, umbi`lical→um`bilical, varicose /-kəʊs→kəs/.
7. Elisions due to speeded articulation from increased familiarity: actually/-ʧʊəli→-ʧəli/ , deteriorate /-`tɪərɪəreɪt→`tɪərjəreɪt/, government /-vənmənt→vəmənt/, manufacture /-njʊf→nəf-/ obviously /`ɒbvɪəsli→`ɒbvɪsli/, particularly /-kjʊləli→/-kjəli//,  seriously /-rɪəs-→-rəs-/, temperature /-pərəʧə→-pəʧə//, temporarily become indistinguishable from temporally, usually /`juːʒʊəli→/`juːʒli/, vaccuum /-kjuːəm→-kjuːm/, vulnerable /`vʌlnrəbl→`vʌnrəbl/.
8. Globalising or unifying) by converting to earlier tonic syllable : quan`dary → `quandary, so`norous→ `sonorous, super`vise→`supervise, va`gary → `vagary.
9. Increasing perception of compound as unified: bank `note → `bank note, country`side→`countryside, deck `chair→deckchair, fountain `pen→`fountain pen, great`coat→`greatcoat, sea`side→`seaside, sponge `cake→`sponge cake, tom`cat→`tomcat, top`coat→`topcoat, week`end→  `weekend. This last word has undergone movement of stress to the front only for some speakers. For others it shows retention of final-syllable stress, yet often with unification signalled by capture of the medial /k/ onto the final syllable, as witnessed by its aspiration. (This last not in LPD).
10. Re-positioning (usually postponement) of tonic (presumably often) to facilitate articulation: `applicable→                      ap`plicable,`combatant→com`ba/ӕ/tant,`comparable→com`pa/ӕ/rable,
11. `conversant→con`versant,  con`tribute→`contribute˂,`despicable→ de`spicable˃, `disciplinary→disci`plinary, di`stribute→`distribute˂, ex`quisite˃, ho`spitable, in`tegral˂, in`ventory˂, for`midable˃, elec`toral˂, `mandatory→man`datory˂, `Monaco→Mo`naco˂,`peremptory→pe`remptory,`Seville→ Se`ville,`Uranus→
12. U`ra/eɪ/nus,`urinal→u`rinal , `Westminster→West`minster. In the case of Tra`falgar we see the "amphibrachicising" tendency found in many words such as ro`coco (Italian roco`co), Ta`ranto etc. A poem by Thomas Hardy clearly shows that he stressed the word Trafal`garas in Spanish. The verb at`tribute hasn't gone the way of `contribute and `distribute as they are now perhaps equally often heard.
13. Weakening of unstressed sit vowel to schwa or zero: Allen, Athens, Belinda /bə-/, celebrate,  cruel, cushion, evil, felicity/fə-/ foreign, goodness, Helen, Kennedy, listlessness/-ləsnəs/, peril, pollen, portrait, system, waitress, woollen. Further developments still are commented on at Blog 105.
14. Strengthening of the suffix /-ɪs/ to /(`)es/: `countess˃, `goddess, `hostess, `Jewess, `lioness, manage`ress, mayo`ress, tailo`ress. Some words in this category have only strengthened the final vowel; many move the stress to the suffix. Any may do so if positive contrast is indicated.
15. Vernerisms: absorb, absurd˂, chrysanthemum˂, discern˂, luxurious, resource, Oxonian˂. 
16. Assimilative tendencies: have to /`hӕf-/, of /əf/course, supposed /-əʊs/ to; absolute /`ӕp-/, advertise /`ӕvvə-/, obviously /`ɒvvɪsli/, hospital /`hɒspɪdl/.
17. Inverse spelling influence ie re-modelling chiefly according the notional value of o before certain consonants as opposed to the "appropriate" value of o in words like horse, orphan, etc : broth, cloth, cross, lost, often etc were converted from earlier /ɔː/ to /ɒ/. This tendency has even perhaps led  many speakers to adopt /ɒ/ in such words as auction, Austin, Australia, caustic˂, claustrophobia. 


18. Adoptionof what are perceived as preferable forms by those who aspire to "clear speech": elec`toral, in`tegral, man`datory. In this category no doubt belongs a whole wide movement to rhythmically strengthen the final vowel of words like happy leading from /`hæpɪ/ or /`hæpi/ and sometimes to /`hæpiː˂/. Similarly thank you has moved /`θӕŋk jʊ ~ ju → juː˂/. The weakform of us has become increasingly avoided so `All of us as /`ɔːl əv ʌs/ is no longer necessarily perceived as contrastive as regards the us.
19. Cross-variety influence: `dispute (frontstressed until the latter 60s only by Northern and Midland speakers), `research (probably reinforced by American influence), ha`rass (a rare case of indisputable American influence from the late 60s onward, though this has long been the usual stressing in Scotland and Ireland), involve (with the -ol- as more traditionally in revolt). This  item is representative of a small group of words which were mainly Londonisms until the 70s. Perhaps in reaction against this last tendency some speakers have adopted /ɒl/ in a few words that have previously normally taken /əʊl/ eg extol, toll.

An item in the inventory of vowel phonemes referred to in Wells (1982) as "London-flavoured ... Near-RP" has undergone a re-classification to "RP" by him (something which I cannot personally confirm though do not necessarily doubt to be reasonable). Since publishing that 1982 text, Wells's subsequent observations have led him to the decision that it is no longer appropriate to classify as merely "London English" products of the split by which speakers may have a phonemic difference between eg wholly and holy with the former having a diphthong beginning much opener. Accordingly in the LPD from 1990 he has offered it as an alternative model for the EFL learner.

References

Bauer, Laurie (1994) Watching English Change Harlow, Essex, UK: Longman.

The Concise Oxford Dictionary of Current English (1911,1995 etc).

Cruttenden, Alan (2008) Gimson's Pronunciation of English. Seventh Edition. London, UK: Hodder Education. 

Gimson, A. C. (1962) An Introduction to the Pronunciation of English London: Edward Arnold. Fifth edition, 1994, edited by Alan Cruttenden 

Jones, D. (1991). Edited by A. C. Gimson and S. Ramsaran. English Pronouncing Dictionary. Cambridge: CUP. ("EPD")

Longman Dictionary of Contemporary English (1978/1987) Harlow, Essex, UK: Longman.

Oxford Advanced Learner's Dictionary (1995 etc & Online). Oxford: OUP.

The Oxford English Dictionary. Second edition. (1989) prepared by J. S. Simpson & E. S. C. Weiner. Oxford: OUP. (OED2). Third edition online (OED3) editor John Simpson.

Ramsaran, Susan (1990) 'RP: Fact and Fiction' in Studies in the Pronunciation of English edited by herself. London, UK: Routledge.

Wells, J. C. (1982). Accents of English. (3 vols). Cambridge, UK: CUP.

Wells, J. C. (1990, 2000, 2008). Longman Pronunciation Dictionary. London: Longman Group UK Limited. (LPD)

Windsor Lewis, J. (1969). A Guide to English Pronunciation. Oslo: Scandinavian Universities Press.

Windsor Lewis, J. (1972). A Concise Pronouncing Dictionary of British and American English. London: OUP.

Windsor Lewis, J. (1975). Linking /r/ in the General British pronunciation of English. Journal of the International Phonetic Association Vol. 5 No.1 pp 37-42

Windsor Lewis, J. (1979) Pre-Consonantal /r/ in the General British Pronunciation of English. English Language Teaching Journal Vol. XXXIII No. 3.

Windsor Lewis, J. (1990) HappYLand Reconnoitred. Studies in the Pronunciation of English edited by Susan Ramsaran. London, UK: Routledge.