A century ago Daniel Jones, the great describer and codifier of the Received Pronunciation of English (RP), was ten years old. He was born in 1881, so by 1891 we may assume that his own pronunciation (on which he doubtless based his description) was fixed in its essentials. Insofar as teachers of English Phonetics or English as a Foreign Language base their teaching on Jones's classic works (1917, 1918), they are operating with a model that is now a hundred years out of date. In practice, of course, those who are in contact with current English as now spoken have modified Jones's model, consciously or unconsciously, in this or that detail. A number of recent articles have attempted to list the phonetic changes that have taken place in RP since Jones's day (Ramsaran, 1990b; Wells, 1990b). Recognition that changes have occurred is attested by the often-heard assertion that `no-one speaks RP any more'. While this is true in the sense that no-one speaks Jonesian RP, it is only to be expected in view of the facts (i) that Jones's generation are by now all dead and (ii) that languages change. It is more helpful, and in my view more accurate, to say that RP is still alive as ever, but that it has undergone various changes.
This paper discusses several of these recent and current sound changes in RP, asking how far they can be attributed to the influence of non-standard accents (including in particular Cockney, the working-class speech of London).
The continued existence of RP (and indeed of what I have elsewhere termed 'near-RP' Wells, 1982: 279, 297-301) may be inferred from the observation that various popular pronunciation features of English English remain clearly outside RP and have been firmly resisted by it. Examples include h-dropping, g-dropping, certain realizations of /æ/, and certain vowel weakening phenomena. I discuss them in turn; there are more extensive treatments in Wells (1982: 253-256, 262, 291-2, 227).
2.1 H-dropping is the failure to pronounce [h] in positions where RP-speakers would pronounce it: thus ["{m@] for ["h{m@] hammer, [edZ] for [hedZ] hedge, [bi"aInd] for [bI"haInd] behind, [aI "{v] for [aI "h{v] I have. We are not concerned here with the loss of historical /h/ in words spelt wh-, e.g. [wQt] what, [we@] where; nor with the omission of /h/ in the weak-forms of function words, as ["tel Im] tell him. In connection with the last-mentioned, however, it is worth noting that the apparent current trend towards pronouncing [h] in many weak-forms, thus ["tel hIm], may well be due to a middle-class hypercorrective reaction against non-standard h-dropping.
2.2 G-dropping is the pronunciation of the -ing ending with an alveolar rather than a velar nasal, thus ["wO:kIn] for ["wO:kIN] walking, ["ri:dn=] for ["ri:dIN] reading. The name refers to the spelling device sometimes used to signal this pronunciation, thus walkin', readin'. Again we may detect a bourgeois reaction against this lower-class usage in the squeezing out of g-dropping from its former prevalence in aristocratic speech (huntin', shootin' and fishin').
2.3 The mainstream RP realization of /æ/ in that man has changed in the course of this century, as has often been noted. The tendency towards an opener quality may be seen as a reaction against the Cockney [E] type, leading to an /V/-like [a] which just happens to resemble variants long associated with provincial accents (northern, Welsh, Scottish).
2.4 Non-standard accents commonly weaken you to /j@/. But in RP this form is generally avoided, at least in prominent positions such as utterance-final. Thus Cockney has ["s@i j@] see you, often with you having a very open, /V/-like final vowel (sometimes shown in spelling as yer or ya). RP, on the other hand, has ["si: ju], where [u] stands for a vowel that may be somewhat shorter than that of the strong-form /ju:/, but in quality is certainly no more central than the [U] of good. The same applies to your and you're, where RP generally retains /jO:/ (or the older /jU@/), eschewing the vulgar /j@/. Once again we note the disappearance of the old U-RP /j@/. A similar tendency, though perhaps not so marked, is to be observed in the weak syllables of words such as ambulance, educate, manufacture, where RP continues to prefer [U] over the popular [@].
The popular weakening to [@] of the final vowel in window, pillow, and similar words, is similarly resisted by RP, where these items end in unstressed [@U]. Pillow and pillar may be homophonous in Cockney, but not in RP.
Let us now turn our attention to some of the recent developments and ongoing tendencies in RP. We shall consider in turn the decline of weak /I/, glottalling, l-vocalization, intrusive /r/, yod coalescence, and lexical changes.
3.1 Decline of weak [I]. The vowel [I] is becoming less frequent in weak syllables. Where Jones said [%vIzI"bIlItI] visibility, we now mostly say [%vIz@"bIl@ti]. Traditional RP [I] is yielding ground on the one hand to [i], in final and prevocalic positions, and on the other hand, in preconsonantal positions, to [@]. By [i] I refer to the /I ~ i:/ neutralization that may eventuate as [I], as [i:], or as something intermediate or inconsistently fluctuating, as finally in ["h{pi] happy, ["kQfi] coffee (final), and prevocalically in ["ve@ri@s] various, ["reIdieIt] radiate (Windsor Lewis 1990; Wells 1990a: 476). The opposite trend of development, preconsonantal [@] for traditional [I], is found particularly in the endings -less, -ness, -ily, -ity and adjectival -ate, and to some extent also in -ed, -es, -et, -ace. As late as 1967 Jones's EPD continued to state that [-l@sn@s] in carelessness 'cannot properly be regarded as coming within the limits of what should be included in this Dictionary' (1967: xxxi). In his 1977 revision of EPD, however, Gimson characterizes the trend towards [@] as 'firmly established' and 'an acceptable variant' (1977: xvi). By 1990 in my LPD I judged it the principal variant, relegating the traditional [-lIsnIs] to second place.
Although the trend towards final/prevocalic [i] might well be attributed to the influence of Cockney, the same cannot be said of the trend towards preconsonantal [@]. In Cockney, indeed, [I] is still often to be heard in the endings in question. For the source of this [@] we can look rather to provincial England, particularly the north and east (including Essex, just outside London), and perhaps even more to North America, Australia, and Ireland. The older [I] should perhaps realistically have counted as one of the oddities of RP as seen from the perspective of World English.
3.2 Glottalling is the switch from an alveolar to a glottal articulation of /t/, whereby [t] > [?] in a range of syllable-final environments. This is by now very firmly established in casual RP before obstruents (football ["fU?bO:l], it's quite good [I?s kwaI? gUd]) and is increasingly heard before other consonants (atmosphere ["{?m@sfI@], partly ["pA:?li], Gatwick ["g{?wIk]). Among younger RP-speakers it can even be heard finally before vowels (pick it up [pIk I? Vp]) or in absolute final position (Let's start! [le?s stA:?]). Intervocalically within a word, it remains firmly excluded from RP (cf. Cockney city ["sI?i]). Nevertheless, the increased use of glottal stops within RP may reasonably be attributed to influence from Cockney and other working-class urban speech. What started as a vulgarism is becoming respectable.
3.3 L vocalization is the development whereby the 'dark' allophone of /l/, [5], loses its alveolar lateral nature and becomes a vowel of the [o] or [U] type. L vocalization is accordingly restricted to the environments appropriate for [5], namely preconsonantal and word-final (except where the following word begins with a vowel). Examples are [mIok] milk, ["mIdo] middle. As Gimson points out (1980: 203), there are plenty of RP-speakers who use it in labial environments, as [maI"seof] myself, ["teIboz] tables. I am beginning to wonder whether my earlier judgment (Wells 1982: 295) that "on the whole ... L Vocalization must be considered only Near-RP or non-RP" is now in need of revision.
Cockney clearly has much more l-vocalization than does RP, and uses it in environments where it is more noticeable. In particular, Cockney uses it where RP would have a laterally released alveolar plosive, as little, middle, and across certain word boundaries where RP would usually have the 'clear' allophone, as for example if....
3.4 That intrusive r is very prevalent in RP is evident to any objective observer. It involves the insertion of an r-sound at the end of a word ending in a non-high vowel (usually one of /@, I@, A:, O:/) where the next word begins with a vowel, as in put a comma[r] in, the idea[r] of it, Leamington Spa[r] and Warwick, I saw[r] it happen. It represents the generalization of the non-rhotic linking-r principle (["bet@] better, but [%bet@r "Qf] better off) to cases that are phonetically analogous but historically (and hence orthographically) different. Rather than the earlier rule deleting [r] except before vowels, both RP and Cockney now have a rule inserting [r] across a hiatus involving a mid or open vowel in the left-hand environment. (A striking instance recently heard from the British Prime Minister, John Major, was the Fontainebleu abatement with [-bl3:r @-].) R-intrusion can also operate within a word before a suffix, as magenta[r]ish, Kafka[r]esque, withdraw[r]al, saw[r]ing.
In spite of its prevalence in RP (and other non-rhotic accents), intrusive r does remain to some extent the object of overt stigmatization. It is clear that one reason for its being more frequently heard latterly is that the merger of the former /O@/ with /O:/ has caused it to spread to environments from which it was previously excluded. While RP manner and manor have been homophonous with manna for some two hundred years, sore and soar have become phonetically identical with saw only in the present century. Hence the analogy for r-intrusion after /O:/ has arisen only quite recently, which explains two things: first, why there are many RP speakers who intrude after /@/ but not after /O:/, and secondly, why intrusive [r] in withdrawal or sawing seems to trigger complaints from today's purists more readily than [r] in sonata in G.
There is a similar explanation for the fact that intrusive r is more frequent in running speech in Cockney than in RP. We have already noted the greater use of schwa in weak forms, which means that see you in a week is a context for potential intrusive r in Cockney ([...j@r In...]) but not in RP ([...ju In...]). So also with pull the window up (Cockney [...d@r Vp], RP [...d@U Vp]). Furthermore, the broad Cockney vowel corresponding to RP /aU/ in mouth is phonetically [{@], with cow and cower as potential homophones, leading to intrusive r in phrases such as how[r] it happened. (Socially intermediate London speech has [{U], with no r-intrusion; the types of intrusive r mentioned in this paragraph are all absent from RP.)
3.5 Observers have noted an increased tendency towards the coalescence of yod (the semivowel /j/) with a preceding alveolar plosive, so that tj > tS, dj > dZ. We can distinguish a number of distinct environments.
3.6 Lexical changes are those that affect specific items of vocabulary rather than all (or most) words meeting a particular phonetic structural description. The opinion poll findings reported in LPD for British English often reveal preferences differing from those of earlier generations. Thus penultimate-stressed con"troversy is taking over from initial-stressed "controversy, penultimate stress being preferred by 56% of the LPD poll panel and no doubt a much higher proportion of the general public, though probably a smaller proportion of RP speakers. For contribute, 27% of the panel claimed to prefer initial stress, a pattern not admitted by Jones even as a variant. For suit, the poll showed, /su:t/ is now preferred over Jones's /sju:t/ by a margin of 72% to 28%; for nephew, /"nefju:/ now beats the traditional /"nevju:/ by 79% to 21%. In accomplish, the variant with /V/ was preferred over that with Jones's /Q/ by an overwhelming 92% to 8%. It is difficult to demonstrate any specifically Cockney input in changing preferences such as these. There are, however, a few cases where a variant long familiar from Cockney has recently achieved greater respectability: thus either /"i:D@/ (preferred by nearly one-third of the youngest age-group among the LPD respondents) and secretary /"sekr@teri/ (a word not included in the LPD poll).
In a discussion such as this it is important not to lose sight of the fact that accents, and more generally varieties of a language, are not objective entities so much as mental constructs. As such, their definition and circumscription are open to considerable debate. Are RP and Cockney best seen sociolinguistically, as areas in a multidimensional spaceintersections of their speakers' socio-economic class, sex, and age, perhaps with contextual style? Or should we define them in a way that corresponds to Platonic notions in our collective unconscious, with RP as the ideal beautiful accent to which we all aspire but which none of us ever quite attain, and Cockney as the epitome as all that is stigmatized in the speech of the capital city? Or would we do best to accept a functional definition, with RP as no more than the codified version of English pronunciation that we teach to EFL learners as an acceptable norm?
Depending on which of these approaches we adopt, we may arrive at different decisions regarding which phonological characteristics fall within or outside RP.
The /hw/ pronunciation of words spelt with wh-, thus /hwen/ when, is part of perhaps most speakers' ideal of 'good' pronunciation; but it is not part of the actual usage of most real-life RP speakers. The smoothing of diphthongs in words such as fire, goingthus [fa@, g@IN]is certainly part of RP as defined sociolinguistically, though clearly not demanded in Platonic good speech. The same is true of intrusive r, discussed above. By Daniel Jones's codification, none of them forms part of EFL-RP, although arguably learners ought to have at least passive familiarity with the latter two.
Platonic RP might be eternal and unchanging. But my preference is for a sociolinguistic definition of RP, which entails recognizing the possibility of change. And some of the changes, it seems clear, can reasonably be attributed to influence from Cockneyoften overtly despised, but covertly imitated.
Author: j.wells@ucl.ac.uk, homepage.
Posted 1998 11 27, IPA-SAM version 1999 01 19