Journal of Linguistics and Language Teaching
Volume 16 (2025) Issue 1
A Case Study in RP – Variation and Change in the
Pronunciation of King Charles and Prince William
Joanna Dornbierer-Stuart (Birmingham City University, UK)
Abstract (English)
Received Pronunciation (RP) is the accent traditionally regarded as the model for correct pronunciation in Great Britain. Largely based on the transcriptions set out in Jones’ English Pronouncing Dictionary of 1917, it appears in most current learners’ dictionaries of British English and thus continues to influence EFL teaching. However, RP is not simply a codified standard found in dictionaries but a living variety that exhibits variation and change. A few recent studies have suggested significant changes in the accent, which begs the question of whether the model accent in our dictionaries represents the speech of modern educated British society, or whether it needs updating. This article presents a detailed case study of vocalic and consonantal variation in the speech of two native RP speakers (King Charles and Prince William) from two generations of the same family over a number of years and in a variety of speech situations. Using both real-time and apparent-time constructs within the variationist framework, data is analysed to determine whether variation signals changes that are complete or in progress. The study confirms patterns of change in RP usage identified in other studies, with a distinct shift towards non-standard features found in regional Southern British English. A discussion concludes that dictionaries should reflect this trend but that any updating should be based on descriptive data rather than prescriptive social ideals. The study should help EFL teachers to acknowledge that norms for British English are evolving and that teaching practices should be adapted to reflect contemporary language use.
Keywords: Received Pronunciation, model accent, language variation, language change, future of RP
1 Introduction
The study presented in this article investigates the unique phenomenon of Received Pronunciation (RP), the accent traditionally regarded as the model for correct pronunciation in England and Wales (Wells 2008: xix) and recently rated as the most prestigious form of spoken British English (Adams 2022). Popularised by Daniel Jones’ English Pronouncing Dictionary of 1917 and adopted by the BBC as its standard pronunciation in 1922, RP continues to be used as the model accent in current dictionaries for British English. However, as with any spoken variety, RP is subject to change as it is transmitted across the community and, as a result, some of its phonemes have started to sound likelike other phonemes. For example, [əʊ] (as in goat) commonly changes to [ɒʊ] before l (as in goal) (Hannisdal 2006: 155), but this is not yet specified in major current dictionaries such as the online Longman Dictionary of Contemporary English (LDOCE), where the pronunciation of goal is still given as /gəʊl/.
Although RP is one of the best-described accents of English (e.g., Jones 1963; Wells 1982; Cruttenden 2014; Lindsey 2019), there is scant quantitative empirical research on RP as a ‘living’ variety. The present study consequently aims to quantify recent change in RP, using both the diachronic and synchronic approach -- rarely combined in previous research -- and thereby contributes to the discussion of whether the model accent found in dictionaries can still be considered to represent standard British English in the 21st century. With the clear focus of the small case study, the present investigation analyses the speech of two living native RP speakers (King Charles and Prince William) from two generations of the same family. Drawing on the variationist framework and using general principles of variation and change, it attempts to establish whether any variation within one speaker over time (diachronic variation) or between both speakers today (synchronic variation) represents change in the accent. With concrete linguistic facts, this study can better inform whether it has become necessary to recodify, or even replace RP with something more representative of the speech of modern, educated British society.
2 Literature Review
John Wells’ classic 1982 collection Accents of English lies within the phonetic tradition of Jones. The collection contains a useful description of RP in use at the time, as well as accepted variants found within the accent. In his later work, Wells (1998) recognised a gradual change taking place in RP, which he attributes to the influence of trends from the working-class speech of London, one example being the spread of yod coalescence in words like Tuesday /’tʃu:zdeɪ/ and reduce /rɪ’dʒu:s/. Nevertheless, in his Longman Pronunciation Dictionary (LPD) of 2008, which significantly shaped transcriptions in the online LDOCE (2024), Wells still referred to British pronunciation as ‘RP’ and largely adheres to Jones’ original transcriptions (e.g., the pronunciation of Tuesday is given as /ˈtjuːzdeɪ/ and reduce as /rɪˈdjuːs/).
Alan Cruttenden, author of the current edition of the comprehensive and authoritative Gimson’s Pronunciation of English (GPE) (Cruttenden 2014: 80), claims that RP has 'evolved' into a more modern form and thus prefers to use the less socially loaded term General British (GB), along the lines of General American (GA). However, he only introduces two transcriptional innovations: /æ/ in cat has changed to /a/ and /eə/ in square has changed to /ɛ:/.
As for quantitative empirical research on RP, Harrington et al. (2000) famously revealed, in a diachronic (real-time) study, changes in Elizabeth II’s vowel patterns over 30 years – demonstrating that even the Queen’s English was affected by modern trends. In a 2002 sociolinguistic investigation using the synchronic (snapshot-in-time) approach, Fabricius found evidence for t-glottalling becoming an acceptable variant in ‘Modern RP’.
Hannisdal (2006), also using contemporary data, looked at other new variants in RP (including t-voicing, smoothing and yod coalescence). She based her study on six phonological variables and looked for these in the speech of RP-speaking newsreaders, many of whom had adopted RP as a secondary accent. She concluded that RP is very much a living variety that incorporates a considerable amount of variability and is inevitably subject to ongoing change. She believes RP is not disappearing but rather evolving and, as such, needs to be updated.
The present study is intended to complement the aforementioned quantitative studies in a number of ways. Firstly, it provides a novel and detailed list of RP features that are reported to have changed in the last 40+ years, together with the phonetic environments in which the changes have occurred. This list not only serves as a descriptive basis for the present study but is potentially a valuable resource for future research. Secondly, it uses the detailed small case study as the starting point for reflections and discussion of the larger picture. Thirdly, it establishes changes in RP through two different approaches: using the real-time approach, it presents an analysis of the pronunciation of two ‘native’ (not adoptive) RP speakers over a number of years (to reveal lifespan change), and, using the apparent-time approach, it compares the pronunciation of the older and younger speaker at contemporaneous time points (to reveal generational change). Fourthly, the study subdivides all speech samples into formality categories to explore the extent to which changes can be attributed to stylistic factors. Finally, the investigation does not begin with prespecified variables but simply uses the data to discover significant patterns of variation and possible change.
3 The Study
The investigation starts with clarification of the term RP as both a social and linguistic phenomenon, followed by a detailed listing of those features of RP that have reportedly changed, using sources such as Wells (1998), Hannisdal (2006), Cruttenden (2014) and Lindsey (2019) (Section 4). Section 5 lays out some general principles of variation and change and introduces the two constructs used in this study to measure phonological change in RP. Section 6 outlines the criteria used to select the informants and the methodological procedures employed to collect, analyse, quantify and evaluate the spoken data. In the original study, the data was analysed in accordance with the questions laid out below, and initial suggestions were made as to the cause of any variation identified. In this article, the findings have been summarised and RP scores for each speaker and each variable appear in a single table (Section 7). Section 8 discusses more generally what is changing in RP in the two speakers analysed, the possible source of the changes, the current status of the accent and possible future developments. Section 9 looks at the implications of the findings in this study for the continued survival of RP as the standard model for British English and suggests areas for future research.
The specific research questions for the present study are as follows:
What features have changed in the pronunciation of the two speakers over time (lifespan change)?
How does the pronunciation of the younger speaker compare with the pronunciation of the older speaker today (generational change)?
Which changes are affected by style?
Are these changes related to language-internal processes and/or language-external (social) factors?
With respect to the two speakers analysed, which changes are well established, and which are still in progress?
The following two sections serve to introduce our object of study (RP) and lay out some general principles of variation and change.
4 Received Pronunciation
4.1 What is RP?
The term Received Pronunciation (RP) was first used in the late 19th century by phonetician A. J. Ellis to refer to the distinctive form of pronunciation favoured in professional and royal circles in London (Fabricius 2002: 117). In the early 20th century, Daniel Jones adopted the term in his English Pronouncing Dictionary (1917), by which time it was associated with the speech of those educated at the best public schools in London and the South-East. Once codified and taught explicitly, RP spread geographically among the higher social classes, becoming a non-regional social accent that lower social classes aspired to. The term Received Pronunciation is still frequently employed in discussions on accents on BBC radio and in British newspapers, implying that it remains a relevant phenomenon in modern British society.
Trudgill (2000) suggests that only 3% to 5% of the population speak RP, although more recent estimates put the figure at 2% (Barton 2018). Estimates are confounded by the fact that the accent is changing. Milroy (2001) notes that with the democratisation of public and professional life, RP has become open to influence from below, creating a new, more popularised standard British English used by higher-status and lower-status speakers alike. According to Lindsey (2019), many traditional features have been replaced with features based on southern British accents of the middle or lower classes, so that cat /kæt/ now sounds like [kaʔt] or [kaʔ] and whole /həʊl/ sounds like [hɒʊɫ] or [hɒʊ]. As a result, Lindsey claims that RP is disappearing.
Wells, on the other hand, insists RP remains relevant. In his 2008 Longman Pronunciation Dictionary, he still specifies the model pronunciation for British English as RP. In a similar vein, Cruttenden (2014) notes that, although spoken in its purest form by at most only 3% of the British population, there is no other accent that is more geographically widespread, and the broadest British accents are generally modified towards RP.
According to the Sutton Trust Speaking Up report (Levon et al. 2022), accent is still a major indicator of socio-economic status, with RP “the dominant accent in positions of authority across the media, politics, the civil service, courtrooms and the corporate sector” (Levon et al. 2022: 4). It seems that a hierarchy of accent prestige still prevails in British society, with RP at the top.
4.2 Describing Received Pronunciation
RP has a very special status in that it is both a social phenomenon (being a model accent related to class, elitism, correctness and conservatism and therefore resilient to change) and a linguistic phenomenon (being a living variety that is apt to variation and change). This creates an obvious contradiction which, over time, causes linguists to disagree about what RP is. Moreover, a phonological description of RP will become obsolete if it is not regularly updated.
Despite the numerous descriptions of RP, there is no universally valid definition of the accent. According to Hannisdal (2006: 20-21), a number of these descriptions discuss new trends and changes in RP without clearly defining the basis for observations. There is, however, a general consensus among linguists that there is a common core of phonological features that distinguishes RP from all other accents of English, and it is certain that many people in Britain speak with an accent broadly similar to that described in dictionaries such as LDOCE. Since any linguistic variety is, to a certain degree, an abstraction, it was decided that an adequate starting point for this study would be the norms set out in the current online LDOCE (2024).
The next section provides a comprehensive list of vowel and consonant sounds in RP that have reportedly changed in the last 40+ years, together with the phonetic environments in which the changes have occurred. These innovations have been reported in a variety of sources, including Wells (1982, 1994, 1998), Fabricius (2002), Hannisdal (2006), Cruttenden (2014), Collins et al. (2019) and Lindsey (2019). The list was intended to help recognise and classify variants of RP in the data gathered in the present study. Only variants with enough tokens were finally chosen for analysis. Further details of the methodology are outlined in Section 6.
4.3 Reported Innovations in Received Pronunciation
Table 1 below lists the range of vowel sounds and vowel combinations (diphthongs and triphthongs) in RP that are reported in various sources to have undergone change. To help specify the vowels, keywords from Wells (1982) and Schneider et al. (2004) were used. Most of the reported innovations are word-internal, e.g., [əʊ] → [ɒʊ] before dark l (as in goal, gold), but some occur across word boundaries, e.g., the smoothing of diphthongs and triphthongs. The reported innovations fall into four main groups:
I. Vowel quality changes in monophthongs
II. Vowel quality changes in diphthongs
III. Vowel quality changes before dark l
IV. Monophthongisation/smoothing of diphthongs/triphthongs
Table 1: RP Vowels with Reported Innovations
Table 2 below lists the range of consonant sounds and consonant clusters in RP that are reported in various sources to have undergone change. As with the vowels, most of the innovations are word-internal, e.g., [dj] → [dʒ] before /u:/ (as in dune), but some occur across word boundaries (e.g., linking and intrusive r). The reported innovations can be divided into seven main groups:
Yod coalescence
Glottalization
L-vocalisation
R-liaison
Assimilation of /s/ to /ʃ/
T-flapping
G-dropping
Table 2: RP Consonants with Reported Innovations
The next section lays out some general principles of variation and change, with examples related to the sound system and RP, and introduces the two constructs used in this study to measure phonological change in RP.
5 Language Variation and Change
A basic assumption of this study is that RP is not merely a model accent preserved in dictionaries but also a living variety subject to variation and change. According to Croft (2000), there are two primary mechanisms involved in language change. First, due to language-internal factors (e.g., ease of articulation), language users introduce innovations into the linguistic system, creating variants (e.g., singin’ instead of singing). Second, due to language-external factors (e.g., social prestige, identity building), certain variants are chosen in preference to alternative forms and spread through the community. Early descriptions of RP already recognised a large number of variants, such as “the common use of the glottal stop in place of t and k in words like department, tightly, quickly” (Jones 1957: xxx). Around fifty years ago, with the recognition that all the different pronunciations in a language are an important source of language change, descriptions of RP started linking variants with new trends in RP and many linguists (e.g., Wells 1982) ascribed the success of certain variants over others to changes in British society and the growing influence of non-standard varieties. Nevertheless, quantified empirical research was lacking. As long noted by Labov and colleagues (Weinreich et al. 1968: 188), while it may be true that all change involves variability, not all variability results in change.
According to Hannisdal (2006: 45-46), a single causal explanation for language change would be unrealistic. Change is usually attributed to a mix of internal (language-inherent) and external (social) forces acting on language as it is transmitted across the community. In addition, speech style (the systematic variation in an individual's speech depending on context, audience and formality level) is an important factor, often regarded as the bridge between internal and external forces, since a speaker’s adjustment in style involves both linguistic choices (e.g., pronunciation) and social considerations (e.g., audience and setting). The following section provides a brief outline of these three mechanisms of change (linguistic, stylistic and social), bearing in mind that all dimensions interact.
5.1 Mechanisms of Change
5.1.1 Linguistic Constraints
Linguistically motivated variation is grounded in the internal structure and mechanics of a language. It is multifaceted, encompassing structural, physiological, psychological and functional constraints, which often overlap and interact. Starting with structural constraints, inherent properties and rules of the linguistic system itself (e.g., stress patterns and syllable structure) can lead to significant phonological changes. For example, Wells (1982) noted that in RP, yod-coalescence did not occur in dune /dju:n/, tune /tjuːn/ and assume /əˈsjuːm/ (before a stressed syllable), but that it had become commonplace in soldier /ˈsəʊldʒə/, nature /ˈneɪtʃə/ and pressure /ˈpreʃə/ (before an unstressed syllable ending in /ə/).
Moving to physiological constraints, some sounds and sound combinations are easier to pronounce than others. For example, almost everyone today, including most RP users, would simplify the pronunciation of handbag to ‘hambag’ (and, similarly, handcuff to ‘hangcuff’) through the processes of omission and assimilation (Aitchison 1991: 129), though these are still given as /’hændbæg/ and /ˈhændkʌf/ in LDOCE (2024). Sometimes LDOCE (2024) recognises such pronunciation changes; for example, handkerchief is given as /ˈhæŋkətʃɪf/.
Regarding psychological constraints, it seems to be human nature to look for inherent patterns in language and tidy up irregularities. For example, the Middle English word for the noun pea was pease but it was gradually assumed that pease was plural, and a new singular pea came into being (Aitchison 1991: 145). At a similar time, the past of the verb catch became caught, following the analogy of teach-taught, just as today the past of sneak is becoming snuck, in accordance with stick-stuck. This tendency to regularise might also explain the gradual diffusion of a particular sound change through the lexicon (known as lexical diffusion). Wells (1998) noted that in RP, yod-coalescence, already commonplace in words like soldier and nature, had begun spreading to words like tune and reduce.
Lastly, it is often claimed that change in one part of the sound system necessitates change in the rest of the system so as to keep the distinction between different words clear (as reportedly happened with the long vowels of Middle English in the Great Vowel Shift). This is known as the “principle of maximum differentiation” and involves a “functional” constraint on phonological variation (Meyerhoff 2018: 15). Lindsey (2019: 17-18) notes that the qualities of most vowels in standard British English have shifted in recent decades in an anti-clockwise direction, such that the [e] in dress has changed to [ɛ], the [æ] in trap to [a], the [a] in price to [ɑ] and the [ɔ] in thought to [o], etc., but that the phonetic symbols most widely in use hardly reflect this:
Figure 1: Reported Anti-Clockwise Shift in Standard British English Vowels Since the 1950s (based on Lindsey 2019: 18)
5.1.2 Stylistic Constraints
Stylistically motivated variation refers to changes in usage and pronunciation as a speaker shifts along the scale of formality depending on the speech situation and social environment. Stylistic variation is closely related to both internal linguistic constraints (since more casual speech usually leads to a faster delivery and/or reduced articulations) and external social constraints (since style varies in different social contexts). In RP, for example, t-voicing in words like British is a weakening process which becomes more prevalent as formality of the speech situation decreases (Wells 1982: 259). On the other hand, a change in style can affect the degree of standard versus non-standard features. As Meyerhoff (2018: 201) notes, there are parallels between stylistic and class stratification of variables. Generally, the variants found in the speech of higher social classes tend to be the variants chosen in more careful styles of speaking.
A core sociolinguistic concept of change is that speakers select variants due to prestige. Pioneering sociolinguist William Labov (1972: 178-180) distinguished between changes from above the level of awareness (usually involving variants from more prestigious varieties) and changes from below the level of awareness (usually involving variants introduced by a lower class). Hannisdal (2006: 46) notes that since RP is the highest variety socially, any changes in this accent must be ‘from below’. An alternative to the prestige model is Eckert’s (2000) Speaker Design Model, whereby speakers use variables to associate and dissociate themselves with others and construct a social identity. These models provide valuable insights into how individual choices can influence broader societal norms that may gradually reshape a social construct such as RP.
5.1.3 Social Constraints
Socially motivated variation is external to the language itself and pertains specifically to the way language is used in social contexts. Social constraints relate not only to sociodemographic factors, such as socioeconomic status, ethnicity and gender, but also to attitudinal factors and the influence of language contact. According to Kerswill (2001), with education opening up and social mobility increasing in the 20th century, RP has become the property of a wider section of the population. At the same time, with prejudice towards regional varieties fading, people no longer have to hide their local accent as they climb the social ranks. As a result, non-standard varieties are exerting an increasing influence on RP.
As a complement to the prestige model, Bell’s (1984) Audience Design Model attributes this to language contact and the process of speech accommodation, whereby speakers adjust to the speech of their addressees to gain approval. More recent studies (e.g., Fabricius 2002) have shown that features of a generalised non-standard south-eastern British English (so-called Estuary English) are entering RP. Estuary English (EE) itself can be seen as the result of accent levelling over geographical space, whereby the middle-class RP speaker accommodates “downwards”, and the working-class local accent speaker moves “upwards”, creating a “middle-of-the-road” variety (Hannisdal 2006: 53).
Having looked at a number of mechanisms behind language change, we now need to consider how to measure change.
5.2 Stable Variation versus Language Change
Language variation (i.e., different ways of saying the same thing) can remain relatively stable over time. For example, the (ing) variable (/ɪŋ/ versus /ɪn/) has existed in some varieties of English for centuries, with the variants serving as markers of social identity and following predictable social patterns (Meyerhoff 2018: 163). Language change occurs when one form starts to become more prevalent than the other. In order to discover what is changing in RP, it is important to be able to measure frequency of use over time. There are different ways of determining a change in progress, each with limitations. Since using multiple methods can help validate findings, this study attempts to measure the prevalence of different variants in RP over time using two constructs: the real time approach and the apparent time approach.
5.3 Real-time versus Apparent-time Approach
A real-time study (as outlined in Sankoff 2006) takes the diachronic perspective and follows speakers at different life stages. The problem with this approach, apart from the practical difficulties in administering such a study, is that speakers may show little innovation in their speech later in life, since most innovation tends to take place in younger years. This means not all ongoing language change in the community will be captured.
In the apparent-time approach (as adopted by Piercy 2011), the frequency of a variant is determined at a single time in speakers of different generations, i.e., synchronic data is used to simulate real-time change. The problem here is that with social changes and movements of population, it may be questionable whether subsequent generations represent the same speech community.
To compensate for these limitations, the present study uses both approaches and focuses on two speakers from the same family. Firstly, it looks at changes in the speech of each speaker over time (to determine lifespan change) and, secondly, it compares the current speech of the two speakers (to determine generational change).
With much of the theoretical background to this study covered, the next chapter turns to the methodological procedures used for obtaining, processing and evaluating the data on which the present investigation is based.
6 Methodology
The empirical basis for the present study is 50 excerpts of speech produced in various styles by two RP speakers over the past 55 years (from 1969 to 2023). The footage amounts to 60 minutes and includes approximately 2,200 words in total. The following chapter discusses the criteria used to select the informants and the methods used to collect, analyse, quantify and evaluate the spoken data.
6.1 Research Ethics
Several ethical considerations were carefully taken into account. Regarding participant consent, no approval for the collection, analysis and publication of the data in this study was necessary, since use of publicly available statements by public figures for the purpose of research falls under ‘fair use’ and does not require direct authorisation from the individuals concerned. With respect to ethical approval, the research is purely linguistic and involves neither personal nor sensitive conclusions (e.g., inferring emotions, health conditions or personality traits from speech patterns). As for any potential conflicts of interest, there is no external funding, media collaboration or political agenda (e.g., critiques of monarchy, class issues, nationalism).
6.2 The Informants
In accordance with modern variationist research, this study favours judgment sampling (selecting participants based on specific criteria to represent a particular variety) over random sampling (selecting participants randomly to represent the whole population). Since RP is spoken by a minority and is socially rather than regionally defined, this makes broad random sampling impractical. Another key consideration is how to identify a relevant population for analysis. RP speakers today come from a much broader range of social backgrounds than in the past, making it difficult to define the accent in social terms. Relying solely on linguistic criteria, however, risks pre-defining RP rather than analysing it objectively. To address this, the study employs a small-scale case study approach, selecting speakers based on the researcher’s judgment to yield meaningful insights into the accent. This investigation thus prioritises controlled, detailed linguistic analysis over statistical generalisability.
The participants chosen for this study are probably two of the best-known and most media-present native RP speakers from two adjacent generations of the same family. HM King Charles and HRH Prince William are ideal informants since they are undoubtedly prime examples of present-day speakers of ‘the Queen's English’. In addition, they have the same gender, social and geographical background, similar boarding-school and university educations and comparable occupations, and, very importantly, there is an abundance of publicly available footage of them in a variety of situations over several decades. By controlling all speaker variables other than age, such as gender, socio-economic background, education level, the study seeks to isolate the effect of age on language variation and thus reveal how linguistic features may evolve across generations within a specific speech community. Change itself can, of course, be influenced by both internal factors (within the language system itself) and external factors (such as sociopolitical developments, technological advancements and interactions with other speech communities).
6.3 Data Collection and Analysis
The first step was to watch a number of British TV documentaries on King Charles and Prince William, including historical footage, and to identify and record suitable excerpts of each informant (up to 30 per informant) speaking at different points in their lives and in different speech situations. The recordings were subsequently ordered into three life phases:
Late teens/20s
30s/40s
50s-70s
and each life phase was ordered into four formality groups:
Scripted speech (very formal)
Interview conversation (semi-formal)
Natural conversation with the public or media (fairly informal)
Natural conversation with people the informants know well (very informal)
This produced on average ten excerpts per life phase for each speaker and three excerpts per formality group within each life phase:
Table 3: Number of Excerpts per Life Phase and Formality Group for Each of the Two Speakers
The recordings were transcribed orthographically, to enable easy reviewing of the excerpts, and phonetically, using fairly broad phonemic transcription but including details of reduced vowels and consonants, as these were considered important markers of style. With the aid of the previously prepared tables of reported innovations in RP, relevant features were identified and transcribed more narrowly to indicate variant realisations (e.g., /ŋ/ = [ŋ] or [n]), and then all variants were counted. The following is an example excerpt:
Orthographic
I love the cry that I’m a political tool. The idea of being a political tool is so marvellous.
Phonetic
aɪ ‘lʌv ðə ‘kraɪ ðət aɪm ə pəˈlɪtɪkəl ‘tu:l ði aɪ’dɪə r əv ‘bi:ɪŋ ə pəˈlɪtɪkəl ‘tu:l
[aɪ] [aɪ] [t] [aɪ] [t] [ɫ] [u:][ɫ] [aɪ] [ɪ:][ɹ] [ŋ] [t] [ɫ] [u:]
ɪz səʊ ‘mɑːvləs
[∅]
Notable features
[aɪ] x4
[t] no glottalization x1 (that I’m)
[t] no flapping x3 (that I’m, political)
[ɫ] x3 (l in second tool is followed by a vowel, hence excluded)
[u:] pre dark l x2
[ɪ:] smoothing x1
[ɹ] intrusive r x1
[ŋ] x1
[∅] elision of /ə/ x1
Since the study is limited to segmental features of RP (vowels and consonants) and does not consider suprasegmentals (stress and intonation), it was not necessary to note intonation group boundaries, except in the case where it might influence the realisation of speech sounds (e.g., phrase-final t-glottalling). Hesitations (em, er, etc.) were not transcribed phonetically.
Auditory analysis was deemed sufficient to distinguish variants as many of the variables deal with categorical distinctions that are unproblematic to discern. Vowel shifts are more problematic but can be discerned with the aid of an audio vowel chart and the trained linguist’s ear. Using the above methodology, sufficient tokens (between 21 and 167 per variable) were found for the following variables:
Table 4: Variables and Number of Tokens Identified
6.4 Quantification and Evaluation
The next step was to present the primary data in a structured format to find patterns of variation and change. Relevant features were scored and charted for each speaker over time (to determine lifespan change using the real-time construct), and the current speech of each was compared (to determine generational change using the apparent-time construct). To obtain scores for features, the number of traditional realisations of each variable was divided by the total number of realisations of the variable (traditional and modern), in the tradition of Macaulay (1977, 1997). Using this calculation, a high score (closer to 1) represents a more traditional RP accent and a low score (closer to 0) a more modern accent. For each variable, overall RP scores were calculated for each speaker as follows:
Table 5: Overall RP Scores for the (e) Variable
In addition, scores were calculated for each life phase and also for each speech situation within each life phase for each speaker, as in the following example for King Charles:
Table 6: RP Scores for Life Phases and Speech Situations for the (e) Variable
Quantification was followed by verbal analysis and interpretation. Results were analysed against current sources to discover whether known trends can be confirmed or new trends determined. Finally, RP scores across all life phases for each speaker and each variable were entered in a single table (Table 7), revealing lifespan change for each speaker as well as generational change between the two speakers.
7 Results
The results revealed several innovations shared by both speakers, as well as significant variation between each speaker. Detailed results for each variable are provided under this link.
Table 7 provides a summary of RP scores for King Charles and Prince William for each variable over time, revealing lifespan change for each speaker as well as generational change between the two speakers:
Table 7: RP Scores for Eleven Variables over Time for King Charles and Prince William
The average RP scores over time clearly show that accent as a whole does not vary greatly within the adult lifespan (from 18 years old) of each speaker. For King Charles, the average RP score decreased slightly from 0.67 in his late teens/20s to 0.61 in his 50s-70s. For Prince William, it remained more constant, increasing very slightly from 0.25 in his late teens/20s to 0.26 in his 30s/40s. Conversely, the average RP scores (in bold) vary considerably between father and son. A paired t-test across all variables in the most recent life phase of each speaker revealed that the older speaker’s scores were consistently and significantly higher than the younger speaker’s scores across all variables (p < 0.0001; Cohen’s d = 9.02), revealing clear generational change. From these results, it might be concluded that most change occurs pre-adulthood (during childhood and/or early teenage years), when presumably influence from outside the family (e.g., from school and the media) and receptiveness to innovation is at its greatest. Compared to earlier RP speakers, King Charles has aligned the vowels (a) and (e) to modern habits and has adopted smoothing of /eə/ and /ɪə/. Prince William has additionally adopted the new price and face diphthongs as well as word-final prevocalic and phrase-final t-glottalling and intervocalic t-flapping, but he remains undecided about l-vocalisation and g-dropping.
8 Discussion
Empirical analysis of eleven phonological variables in two native RP speakers (father and son) over a period of 55 years, with all speaker variables other than age controlled as far as possible, suggests that the accent has not remained static. While the analysis of l-vocalisation and g-dropping showed more conservative tendencies, the analysis of diphthongs and smoothing revealed the opposite.
In this study, variability in native RP seemed to be internally, stylistically and externally motivated. With respect to the two speakers analysed, most vowel changes in the study proved to be well established and could be explained by language-internal processes. The use of modern [a], [ɛ], [ɑɪ] and [ɛɪ], for example, might be explained as part of the general anticlockwise shift in RP vowels (Section 5.1.1), and smoothing of /eə/ to [ɛ:] and /ɪə/ to [ɪ:] could be attributed to the general tendency to monophthongise the RP centring diphthongs. This study revealed that in this special case, /ɪə/ in phrase-final position was more resistant to change, signalling connected speech was the driver of this type of change. The increased use of [ɪ:] by Prince William within certain words (e.g., weirdly, disappeared) further suggested a more recent process of lexical diffusion. Speech style seemed to have little influence on vowel changes, suggesting social factors were less relevant (since style varies in different social contexts). The only exception was the /ɔ:/ variable in all words: modern [o:] has not yet been fully adopted by Prince William, possibly due to higher salience and stigmatisation.
In this corpus, changes involving consonants were more obviously attributable to stylistic and social factors. For example, stylistic differences in Prince William’s speech for t-glottalling revealed that casual speech is potentially leading change. While King Charles used word-final [ʔ] 35% of the time, spread over all styles, Prince William used word-final [ʔ] in more than 80% of cases, with an incidence of 61% for speech styles (1) and (2) and 96% for speech styles (3) and (4). In addition, King Charles’ use of [ʔ] is limited to preconsonantal (as in What kind) and phrase-final (as in on this planet) environments, never occurring prevocalically (as in not only), whereas Prince William has started using [ʔ] in the more salient prevocalic position, which is typical of Estuary English (EE), the new generalised variety of southern English based on the popular accents of south-east England. This seems to suggest socially motivated change from the ‘lower’ classes, itself the result of language contact and accommodation (Section 5.1.3). The adoption of l-vocalisation and g-dropping by Prince William might also be attributed to these processes, although neither are yet commonplace in his speech.
The data for t-flapping pointed to both language-internal and external processes. While King Charles’ results show t-flapping to occur exclusively across word boundaries in very rapid speech (as in got a, lot of), indicating the importance of reduction processes in motivating linguistic variability, Prince William’s results show t-flapping in the more salient word-internal environment (as in pretty), again typical of EE and, on a broader scale of influence, American English.
These results clearly show that socially motivated changes in the younger speaker are no longer motivated from the ‘upper’ classes but from the ‘lower’ classes, specifically from the increasingly prevalent EE accent of the south-east. This amounts to the gradual incorporation into RP of features that until recently were associated with non-standard speech. With respect to the findings in this study, this leaves a dilemma for our definition of RP. Is, independently from our corpus, EE the new standard English that might replace RP in our dictionaries? Wells (1998) defines EE as “standard English spoken with an accent that includes features localisable in the south-east of England”. He also claims that through increased mobility and wider dialect contact, Estuary-like features are spreading beyond the south-east and losing their localisability. According to Crystal (1995: 327), there is another reason for this spread: EE is increasingly dominating the media, with the result that RP is no longer the accent that non-RP speakers wish to emulate. Giles et al. (1990) reported that traditional RP, a symbol of the Establishment and once deemed a key to success by the middle classes, has lost its prestige among younger generations and is considered by many as unfriendly, aloof and arrogant, while Coggle (1993: 85) reported that EE is perceived as modern, informal, socially neutral and high on ‘street-cred’. However, according to the more recent Speaking Up report (Levon et al. 2022), bias against working-class and regional accents has not gone away, with RP rated as the most prestigious accent in 2019.
Lindsey’s online CUBE (2024), one of the most progressive pronunciation dictionaries for current British English, uses “Standard Southern British” (SSB) as its default accent. The aim is to represent a form of contemporary, educated British English that is neutral and accessible for those learning the language (Hancock 2019). Nevertheless, according to Wells (1982: 14), an important characteristic of RP is its non-localisability within England. It is highly questionable whether SSB sounds non-localised to a northerner. Shouldn’t there also be a northern Standard English? And what about Scottish and Welsh Standard English? Let us not forget that RP, originally based on upper-class southern British, was never a mainstream accent and has never been used by more than 5% of the population (Milroy 2001), but with its widespread adoption in national newscasts, government, courtrooms, boardrooms and universities, it has become “a sort of glue, a force for uniting the country” (Barton 2018). Even if the social conditions that created RP have weakened, it still serves as a valid standard that is understood (if not spoken) by all speakers of British English, whether southern, northern, Scottish or Welsh, native or non-native.
As for RP transcriptions, these must now be seen as abstract rather than concrete representations. Being phonemic rather than phonetic, there is considerable scope for variation when it comes to the precise phonetic realisation of each sound. In the majority of cases, current variants in RP (e.g., t-glottalling, t-flapping, l-vocalisation and g-dropping) are allophonic and do not alter the basic phonemic distinctions of English, thus making changes to the basic phonemic inventory unnecessary. Where phonemes or phoneme combinations have clearly been replaced by other phonemes, as has happened in this corpus with /æ/, /e/, /aɪ/ and /eə/, it may be wise to update dictionaries. In fact, these four phonemes have been updated in some more recent pronunciation dictionaries, notably The Routledge Dictionary of Pronunciation for Current English (Upton & Kretzschmar 2017, hereafter RDP) and Lindsey’s online CUBE (2024). Where, however, features are present only in casual or rapid speech, citation forms will continue to exist in more formal styles and written language and are therefore necessary for language learners to understand and master, both native and non-native.
It is interesting to note that the online Oxford English Dictionary (OED 2025) takes Upton’s model of RP as its pronunciation model for British English. It is careful to specify that RP for OED is not aimed at prescribing correctness but, as Upton states in RDP (2017), it describes
that accent which will be most widely acceptable, as well as most intelligible, to native British English speakers, and to which the speech of very many of them will in turn approximate closely. (Upton & Kretzschmar 2017: xiv)
The model has been constructed to avoid association with any particular region or social class, although the accent is clearly more prevalent in south-east England and among educated speakers. The main points of departure from traditional RP transcription conventions include /a/ in trap words, /ɛ/ in dress words, /ʌɪ/ in price words and /ɛ:/ in square words, all of which were found to be well established in Prince William’s speech. Nevertheless, OED (2025) does not include the innovation /ɪ:/ in near words (which was found to be very well established in Prince William’s speech). More traditional /ɔ:/ (in all words), dark l (after vowels) and /ŋ/ (in unstressed -ing syllables) remain in OED (2025) (still commonplace in Prince William’s speech). Logically, OED (2025) does not represent glottal stops or flapped t in transcriptions because they are not phonemically distinctive. These features do, however, constitute significant changes in progress, as revealed by Prince William’s more casual speech. According to Fabricius (2002), RP needs to recognise differences between degrees of formality. For example, in reading-passage style, t-glottalling occurs only pre-consonantally, whereas it is spreading in prevocalic and pre-pausal environments in spontaneous speech. OED (2025) does, at least, recognise a reduction in the pronunciation of many short words when unaccented, as often occurs in connected speech; for example, and is given as /and/ or /(ə)n(d)/, of as /ɒv/ or /ə(v)/.
Inspired by our study, but not as an undue generalisation, we might wonder about the future of RP. With time, new variants will inevitably enter the accent, some of which could warrant changes to the basic phonemic inventory of RP. Presently in the south-east, younger speakers are being influenced by a new accent, “Multicultural London English” (MLE), which is said to be replacing Cockney (Lindsey 2019: 46). For instance, the price and face diphthongs, pronounced as [ɑɪ] and [ɛɪ] by Prince William, have fronted vowels and are smoothed in MLE ([a:] and [e:]). According to Cheshire et al. (2011), the accent has spread to young working-class whites; time will tell if MLE is the next source of change for RP.
On a broader scale, with the continued influence of American culture worldwide, the pronunciation of British speakers may be brought closer to that of Americans in the future. For the time being, however, British English certainly retains some distinctive characteristics, such as the distinct lot vowel and non-rhoticity. According to Lindsey (2019: 11-12), areas of greater American influence lie in words and expressions (recent examples being multiple instead of many and so instead of well at the beginning of an answer) and intonation (e.g., making statements sound like questions using ‘Uptalk’, a habit regularly practised by Prince William).
It seems reasonable to suggest that RP will continue to serve as the model accent to which all other British accents modify. Certainly, every person has an accent that signals some part of their social background, and every person has biases towards or against certain accents. RP should therefore continue along the path taken by Upton’s RDP (2017), which is to remain both intelligible and acceptable to the greatest number of British (and global) speakers of English possible. A renaming is perhaps not necessary since received can be understood as ‘not given by birth but received as a model’, although General British (GB), along the lines of General American (GA), may help to eradicate socially loaded connotations from the past. Above all, any standard should be practically rather than ideologically motivated, with a tendency to broaden to more tolerance towards a larger number of speakers. Ideally, it should be able to accommodate substantial variation, allowing for occasional updating when new habits constitute major phonemic shifts.
This is not to say that local dialects should be discouraged. As Wilfred Pickles (a Yorkshireman and BBC’s first non-RP newscaster) said in 1949,
May it be forbidden that we should ever speak like BBC announcers, for our rich contrast of voices is a local tapestry of great beauty and incalculable value. (Knowles 1974)
The question arises as to whether codification is contributing to accent levelling, or whether, in a century or two, the British will resort to two distinct varieties in everyday life, as in German-speaking Switzerland. All native Swiss Germans use their local Swiss German dialect at home and socially but revert to High German for reading and writing (since Swiss German is not codified) and for speaking to non-Swiss German speakers, i.e., the use of High German is practical rather than status-bearing. The huge disadvantage is that much effort has to be invested by Swiss Germans at school to learn High German, which they rarely speak as deftly as their native Swiss dialect. High German remains to them a foreign language, which actually helps keep Swiss German alive. The situation in Germany is different: local dialects are still spoken today in informal situations and at home, but speakers use a continuum of varieties ranging from the more dialectal to the more standard, according to the situation. The tendency here is that dialects are vanishing, perhaps the direction in which British English is heading.
9 Conclusion
This study has shown that there is an important distinction to be made between RP as a codified construct found in most current learners’ dictionaries and RP as a living variety that exhibits variation and change. Due to its codification, the former inevitably lags behind the latter. This study revealed discrepancies between the RP found in LDOCE (2024) and the pronunciation of two living native RP speakers from different generations of the same family. In particular, the symbols /æ/, /e/, /aɪ/, /eɪ/, /eə/ and /ɪə/ were found to misrepresent the pronunciation of both speakers, who consistently used pronunciations closer to [a], [ɛ], [ɑɪ], [ɛɪ], [ɛ:] and [ɪ:]. In the speech of the younger informant, t-glottalling and t-flapping (features not traditionally associated with RP) were found to be commonplace, suggesting changes that were well established. L-vocalisation and g-dropping (also non-RP features) were evident but not yet dominant, suggesting changes in progress.
Cruttenden’s GPE (2014) is a little more progressive, having adopted modern /a/ and /ɛ:/. By contrast, Lindsey’s online CUBE (2024) has adopted most of the new vowels mentioned above but was found to be too progressive in replacing /ɔ:/ with /o:/ in all words. Prince William himself used [o:] in only one in six cases where it was possible. Upton’s RDP (2017) has adopted modern /a/, /ɛ/, /ʌɪ/ and /ɛ:/ but has remained conservative with /eɪ/, ɪə/ and /ɔ:/. This dictionary provides the model pronunciation for the online OED (2025) (which claims to be an unsurpassed guide to English for researchers) but many of the innovations do not yet appear in the online Oxford Learner’s Dictionaries (2025), probably because these dictionaries prioritise clarity and simplicity, which inevitably result in pronunciations that are more conservative, or else resources for updating are more limited.
This detailed investigation into the pronunciation of two high-profile native RP speakers from two generations of the same family over a number of years and in a variety of speech situations confirms patterns of change in RP usage established in previous studies, notably the shift from [æ] to [a] in trap words and from [e] to [ɛ] in dress words (Harrington et al. 2000), the smoothing of [eə] to [ɛ:] in square words and [ɪə] to [ɪ:] in near words (Hannisdal 2006) and word-final preconsonantal t-glottalling (Fabricius 2002). This study further expands on these observations by quantifying recent changes in the price and face diphthongs as well as the growing trends of [o:] instead of [ɔː] before dark l, l-vocalisation, word-final intervocalic t-flapping and word-final prevocalic and phrase-final t-glottalling. These growing trends all confirm a continued levelling towards non-standard features typical of EE.
Naturally, caution should be exercised in generalising these findings to the broader RP community. In particular, this study did not include adoptive RP speakers, native RP speakers from other regions of Britain or speakers from different social strata, who might exhibit more conservative tendencies or other innovations. A key challenge in any such investigation would be defining these geographically dispersed RP subgroups and identifying a sufficient number of representatives from each, which could limit the depth and reliability of the findings. Although the sample size of this investigation is very small, the detailed quantitative and qualitative analysis in this generational case study might serve as a foundation for future research with larger datasets.
While there may be a case for broadening RP’s scope linguistically and socially, the role of RP as an authoritative codified standard ensures that it will continue to be a social force that regulates usage, especially in more formal speech. It might be a step in the right direction to recognise that a standard should not be prescribed by a single social group but should represent a speech norm collectively and subconsciously agreed upon (self-policed), which can only be captured by observing linguistic behaviour.
Acknowledgements
The author would like to thank Dr. Robert Lawson, Associate Professor in Sociolinguistics at Birmingham City University, for his valuable feedback on the study presented in this article, and also the Editor of JLLT for his enthusiasm and support.
References
Adams, Richard (2022). Bias against working-class and regional accents has not gone away, report finds. In: The Guardian. (https://www.theguardian.com/inequality/2022/nov/03/bias-against-working-class-and-regional-accents-has-not-gone-away-report-finds.; 03.11.2022).
Aitchison, Jean (1991). Language change: progress or decay? Second Edition. Cambridge University Press.
Barton, Laura (2018). Received pronunciation may be dying out. In: The Guardian. (https://www.theguardian.com/science/shortcuts/2018/may/22/received-pronunciation-may-be-dying-out-but-its-passing-is-long-overdue.; 22.05.2018).
Bell, Allan (1984). Language Style as Audience Design. In: Language in Society 13, 145-204.
Cheshire, Jenny, Paul Kerswill, Susan Fox & Eivind Torgersen (2011). Contact, the feature pool and the speech community: The emergence of Multicultural London English. In: Journal of Sociolinguistics 15.2, 151–96.
Coggle, Paul (1993). Do you speak Estuary? The new Standard English. London: Bloomsbury.
Collins, Beverley, Inger Mees & Paul Carley (2019). Practical English Phonetics and Phonology. Routledge, London.
Croft, William (2000). Explaining language change. An evolutionary approach. London: Longman.
Cruttenden, Alan (2014). Gimson’s Pronunciation of English. 8th Edition. Routledge.
Crystal, David (1995). The Cambridge encyclopaedia of the English language. Cambridge: CUP.
Eckert, Penelope (2000). Linguistic Variation as Social Practice: The Linguistic Construction of Identity in Belten High. Malden, MA: Blackwell.
Ellis, Alexander J. (1869-89). On Early English Pronunciation. Parts 1-5. London: Philological Society.
Fabricius, Anne (2002). Ongoing change in modern RP: Evidence for the disappearing stigma of t-glottalling. In: English Worldwide. A Journal of Varieties of English. July 2002.
Giles, Howard, Nikolas Coupland, Karen Henwood, Jim Hariman & Justine Coupland (1990). The social meaning of RP: An intergenerational perspective. In: Susan Ramsaran (Ed.) (2015): Studies in the Pronunciation of English. London: Routledge, 191-211.
Hancock, Mark (2019). As a model for learners, RP is outdated and needs to be replaced. In: El Gazette. (https://www.elgazette.com/point-of-view-mark-hancock/.; 10.06.2019).
Hannisdal, Bente R. (2006). Variability and Change in Received Pronunciation. PhD diss., University of Bergen.
Harrington, Jonathan, Sallyanne Palethorpe & Catherine I. Watson (2000). Nature Vol 408, 21/28 December 2000.
Jones, Daniel (1917). English Pronouncing Dictionary. London: Dent.
Jones, Daniel (1957). Everyman’s English Pronouncing Dictionary. 11th Edition. London: Dent.
Jones, Daniel (1963). Everyman’s English Pronouncing Dictionary. 12th Edition. London: Dent.
Kerswill, Paul (2001). Mobility, meritocracy and dialect levelling. The fading (and phasing out) of Received Pronunciation. In: P. Rajamäe & K. Vodelberg (Eds.): British studies in the new millennium: the challenge of the grassroots. Tartu: University of Tartu.
Knowles, M. (1974). Wilfred Pickles. Hodder & Stoughton.
Labov, William (1972). Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press.
Levon, Erez, Devyani Sharma & Christian Ilbury (2022). Speaking Up: Accents and social mobility. (https://www.suttontrust.com/our-research/speaking-up-accents-social-mobility/.; Nov 2022).
Lindsey, Geoff (2019). English After RP. Palgrave Macmillan.
Macaulay, Ronald (1977). Language, Social Class and Education: A Glasgow Study. Edinburgh: Edinburgh University Press.
Macaulay, Ronald (1997). Standards and Variability in the Production of the Vowels in ‘Standard’ English Speech in the UK. In: Journal of English Linguistics 25.3, 217-236.
Meyerhoff, Miriam (2018). Introducing Sociolinguistics. 3rd Edition. Oxon: Routledge.
Milroy, James (2001). Received Pronunciation. Who “receives” it and how long will it be “received”? University of Michigan.
Piercy, Caroline (2011). One /a/ or Two?: Observing a Phonemic Split in Progress in the Southwest of England. In: University of Pennsylvania Working Papers in Linguistics, Vol 17, Art 18.
Sankoff, Gillian (2006). Age: Apparent Time and Real Time. In: K. Brown (Ed.) (2006): Encyclopedia of Language and Linguistics. 2nd Edition. Oxford: Elsevier, 110–116.
Schneider, Edgar, Kate Burridge, Bernd Kortmann, Rajend Mesthrie & Clive Upton (2004). In: A Handbook of Varieties of English. Vol. 1: Phonology. Mouton de Gruyter.
Trudgill, Peter (2000). "Sociolinguistics of Modern RP". University College London, 8 Dec 2000.
Upton, Clive & William A. Kretzschmar (2017). The Routledge Dictionary of Pronunciation for Current English. Abingdon: Routledge.
Weinreich, Uriel, William Labov & Marvin I. Herzog (1968). Empirical foundations for a theory of language change. In Winfred P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics: A Symposium. Austin, TX: University of Texas Press, 95-188.
Wells, John C. (1982). Accents of English. Cambridge: Cambridge University Press.
Wells, John C. (1994). Transcribing Estuary English: a discussion document. In: Speech Hearing and Language: UCL Work in Progress, 8, 259–267.
Wells, John C. (1998). “Our changing pronunciation”. Summary of a talk given at the British Association annual Festival of Science. (https://www.phon.ucl.ac.uk/home/wells/cardiff.htm.; 09.09.1998).
Wells, John C. (2008). Longman Pronunciation Dictionary. 3rd Edition. John Wells.
TV documentaries
Channel 5. 2011. Prince William and Catherine Middleton – Too good to be true? April 7, 2011.
EM Productions. 2023. The New Royals: William & Catherine. October 10, 2023.
ITV. 2016. When Ant and Dec met the Prince. January 4, 2016.
ITV. 2023. King Charles III: The Monarch and the Man. May 3, 2023.
Sky News. 2022. King Charles III: A Modern Monarch. September 12, 2022.
Author:
Joanna Dornbierer-Stuart, MA MCIL
EFL qualified teacher + PhD student (pending funding)
School of English
Birmingham City University
Birmingham, UK
Email: jo.dornbierer@bluewin.ch
ORCHID: 0009-0005-5404-7423