Antique RP

The problem with using the term Received Pronunciation to describe a single accent is that it has been used for more than two centuries. Accents change, and the accent described with this term has changed a great deal over that time. John Walker uses it in 1774 , though he is not so much describing an accent as an accepted pronunciation of particular words. Later in his pronouncing dictionary it’s plain that he defines accepted and received as synonyms. A.J. Ellis uses the term in 1869 to describe the pronunciation…of “pronouncing dictionaries and educated people”. The evidence we have from other sources however, makes it clear that the accent of these “educated people” in England has changed tremendously. So if we use the term RP, we’ll have to do so the same way we use “cc” in reference to emails, as a historic artifact that we’ve decided to live with, but that has lost its original associations. I’ve followed this path of convenience, but with further descriptive terms to specify a particular accent of a particular time, place, and cultural association. Hopefully that will help us avoid the problematic usage of RP referring broadly to the whatever variety is currently accented and in use by those in academic, political, and economic power. I’ve chosen three specifying labels that refer to the evolutions of the accent through three overlapping periods of perhaps 40 years each: Antique RP, Traditional RP and Modern RP.

Here’s a chart to help visualize how each of these varieties have given way to new one:

…that’s Alan Turing on the left

You may have noticed that the first section didn’t start with a title naming it as PEOPLE. Arguably it makes sense to include discussion of the history of the accent label in this section because PEOPLE is meant to include an understanding of how the speakers of an accent tend to have some shared identity, both in their own understanding and in the views of others. The name RP is certainly part of that understanding, Nonetheless, I thought it might be useful to treat that as an overview for the whole label of Received Pronunciation, before narrowing in to discuss the cultural context of Antique RP.

Part of that cultural context is the way people respond to an accent. We all have gut responses to accents that might reveal something about the ideas we associate with the accent. as an experiment in capturing those responses in yourself, take a listen to this recording, allowing yourself to become aware what images and feelings it evokes:

The accent that we’re calling Antique Received Pronunciation (ARP) is associated with the strictly enforced class hierarchy of the British Empire at its peak. While sociolinguists might call this a prestige accent, that’s not the same as saying that the accent has any inherent qualities of beauty or superiority. Being the accent of those with power, and those who hoped to get power, it might be read as conveying wealth and erudition, or arrogance and malice. Along with the fall of that empire, came a decline in the social value of ARP.

Comparing the accents of England’s royal family over the years, it’s surprising how much the accent changes in the last generation. In a family whose profession it is to maintain a symbolic continuity of social hierarchy, it makes sense to hear more stability in accent through the decades. When we reach the generation of Prince William, however, it’s clear that there would be no social value for him to use Antique or even Traditional RP.

King George V: born 1865
Queen Elizabeth II: born 1926

It is the relative tenseness, and closeness of this accent’s oral posture that are most salient. That is to say, the jaw tends to be held a bit higher than in most American accents, and the lips and tongue tend to carry a slight tension. This narrower field of play and slight, we could say ‘elastic’ tension means that its speakers are able to speak quickly and with a great deal of articulatory activity. In particular, the tip of the tongue is very active, and the front of the tongue tends to be narrower. This accent, as compared to compared to the accent of the contemporary characters, tends to have significantly less lip rounding on some sounds.

The tongue root is more advanced, and front vowels tend to be formed very front indeed. To put this another way, American accents tend to be marked by tongue root retraction and centralization in comparison to RP.

The lip corners tend to be very slightly retracted and compressed toward the teeth. However, some phonemes are realized with more rounding than their American counterparts, and so the net effect is of lip corners being pulled gently inward while making  fairly frequent flicks forward. The velum is active, closing of the nasal cavity except on nasal consonants. This can vary, though, depending on character.      


The tautness of the oral posture, as noted above, makes possible some very rapid sequences of articulation, and the general texture this produces may feel quite staccato to an American ear.

As a feature of the realization of phonemes, RP makes more distinction between “long” and “short” vowels.

RP also uses differences in vowel length for emphasis. This leads to a rhythmic pattern described by Gillian Lane-Plescia as “rattle, rattle, bing!” Pitch changes in RP tend to be far more frequent, far more simple (rather than compound) and far more sudden, occurring quickly across a single syllable in the phrase, rather than occurring gently than across a chain of several words.

Take a listen to this :

and see if you can match it to this drawing:

Now listen to the pattern with the words:

Here are some more vivid examples from The Millionairess

Take a look at these two sets of words:









Try saying them out loud, first column A, then column B. 

If you’re an American, you probably used the same vowel (or pretty close) in all of these words.  In RP and in other London accents, the words in these two columns are pronounced with different vowels. For words in column B, where an American would use [æ], an RP speaker will use [ɑː].  Or to use the language of Lexical Sets, BATH  words and TRAP words are split in RP. For American speakers, the sets are merged

For speakers with a merger of this sort, determining which words belong to the BATH set can be challenging. A subset of our broad category of words pronounced with /æ/ will need to be identified and shifted, and running through the list of words, it’s difficult to see any rules that would help in making this distinction.

It may be useful to know that in the 18th Century, most speakers of English categorized these words exactly as a modern American would. It was only after the American Revolution that speakers of English in England began to lengthen the vowel in some cases. This process is called “pre-fricative lengthening,” and it describes a tendency for vowels spelled with “a” to last a little longer when followed by a cluster of fricatives such as “ff, ft, sk, or ss.” Eventually, the lengthened group also began shifting towards a back vowel. As a result, words like “path”, “ask”, or “grasp” are likely candidates to be part of the BATH set.  Unfortunately, this is not a perfect predictor of pronunciation.  There are words like “grant” or “can’t”  that fall into the BATH set, despite their not following this “pre-fricative” pattern, and the words “cant” and “pants” remain in the TRAP set.  So, to be sure, a pronouncing dictionary should always be consulted.

Actors who are unsure whether a word belongs to TRAP or a BATH, sometimes hit upon a compromise solution, of pronouncing uncertain candidates with a vowel poised somewhere between these two choices. This is not a happy solution however, because it eliminates the distinction that is a feature of this accent.

The process of pre-fricative lengthening does bring up another point. You’ll notice in the transcription above the vowel [ɑ] is followed with the length symbol [ː].   This indicates that the vowel lasts a bit longer.

We’ve looked at vowel length before, and so we know that the actual durations of vowels vary according to a range of circumstances. What’s  important to pay attention to here is that TRAP vowels are short vowels, and BATH vowels are long. Again, that doesn’t mean that there is a required duration , only that under the  conditions most favorable to full length (a stressed, monosyllabic word with a voiced continuant consonant following) a BATH vowel is free to get a bit longer than a TRAP vowel.

BATH → ɑː

Whenever a man asks me to marry him

TRAP → æ

Driving back to work after I’d had lunch

As we’ve already discussed, TRAP words share a similar pronunciation in RP and most varieties of American speech. There are certainly subtle variations, but we can easily think of this vowel sharing the same identity. To put it another way, the TRAP set contains all the /æ/ phonemes remaining after the BATH subset has been removed.

One confounding factor for some American speakers is the possibility (the likelihood?) that you have a split in your TRAP set, realizing some as relatively open monophthongs, and others as more close. Just compare hat, hang, hash, ham and habitat. It would be surprising to me if you used exactly the same vowel realization in all of those words. We might hear a TRAP/TRAM split, or a TRAP/TRASH split, or a HAT/HANG split. We can return to this point when discussing varieties of American accents, but for RP we should be targeting a realization that is definitively front, short, and monophthongal.

In the case of Antique RP, we should target our TRAP vowels quite high, possibly as high as [ɛ]. We might also entertain the possibility of this vowel breaking slightly so that “trap” is realized as [ɛæ̯]. This raised TRAP tends to impinge on the territory of DRESS, raising it towards [e̞]. This comparative raising and fronting of front vowels reinforces what we’ve discovered about the posture of this accent.  

You seem to take my death rather coolly.

Finally, In addition to the TRAP/TRAM split of American English, there’s another subset of (nominally) TRAP words that many American speakers of English would consider members of the SQUARE set. When the consonant following the vowel is /r/, as in words like carry, Paris, arrow, RP speakers (including J.C. Wells) analyze these words as TRAP+ /r/. Most American accents would categorize these as SQUARE words, pronouncing them with something like [ɛɚ̯]. Unfortunately, the rules governing this pattern aren’t fully predictable. A double “r” in between two vowels is a good indicator. The words arrow [ˈæ.ɹəʊ̯] carriage,  [ˈkæ.ɹɪʤ],  and marry [ˈmæ.ɹɪ]  all fit the TRAP + /r/ pattern in RP. When it comes to a single ‘r’, though, things aren’t as predictable:

TRAP + /r/

A foolish marriage…whenever a man asked me to marry him…

TRAP + /r/

Paris      [ˈpæ.ɹɪs]

marry    [ˈmæ.ɹɪ]

Arab      [æ.ɹəb]


parent   [ˈpɛɚ̯.ɹənt]

Mary     [ˈmɛɚ̯.ɹɪ]

area       [ˈɛɚ̯ɹi.ə]

Think of the /r/ as beginning the following syllable. In ARP it would also likely be a tap [ɾ]. We’ll come back to this a bit later.


→ ɒ̜ → ɔː → ɔː

→ ɒ → ɒ → ɔ̝ː

The three lexical sets covering the low back vowels are LOT, CLOTH, and THOUGHT. Looked at together, we can think of these sets as overlapping brackets allowing us to note the distribution of these vowels along a scale of length and roundness.  In Western American English, we might see all of the words in these categories pronounced with an unrounded, and not particularly lengthened [ɑ]. My own, Midwestern accent separates LOT [ɑ] from CLOTH/THOUGHT [ɒː]. For Traditional and Modern RP, LOT and CLOTH form a single, merged category of words with [ɒ] and THOUGHT is a separate category realized with [ɔ̝ː].  Antique RP hangs onto an even older form which groups CLOTH and THOUGHT together, realizing them as [ɔː], and maintains LOT as a separate, less rounded [ɒ̜].  What complicates this even more is the fact that CLOTH switched to the LOT pronunciation fairly early in the 20th Century for some speakers but was retained by more conservative speakers for longer.

What this means for us as actors is that if we used the Traditional RP form, merging LOT and CLOTH as [ɒ], we wouldn’t blow our disguise as ARP speakers. On the other hand, using the older form of [ɔː] for CLOTH, could sound unusual to a contemporary audience.  There’s a joke in The Pirates of Penzance which turns on two characters getting the words often and orphan confused. That relies on the more conservative Major General pronouncing them both as [ˈɔːfn̩] while the younger Pirate King, pronounces orphan as [ˈɔ̝ːfn̩] and often as [ˈɒːfn̩]. In most other cases, however, it might be more helpful to take treat LOT/CLOTH as  [ɒ] and THOUGHT as [ɔ̝ː].

Here’s Queen Elizabeth in 1958, clearly retaining the conservative form:

It’s possible for for some of you to see me today because of these changes. I’m not surprised that many people feel lost, and unable to decide what to hold onto and what to discard. My own family often gather round to watch television

and here’s a version with these vowels slowed:

It’s POSSible for for some OF you to see me today beCAUSE of these changes. I’m NOTsurprised that many people feel LOST, and unable to decide WHATto hold ONto and WHATto discard. My own family OFTEN gather round to WATCH television

Here are a few further examples showing a bit wider range of possibilities.

To impress upon its followers their absolute equality in Philosophy and Politics.
The quality of mercy is not strained
The awful thing is…


FACE feɪ̯s

These two diphthongs are not wildly different from their realizations in the US. They’re placed together here because they are both somewhat fronter and higher than their US counterparts.

When I was driving back to work, waiting for the lights to change, the car started to shake from side to side.
waiting for the lights to change, the car started to shake from side to side.

GOAT ɛ̈ʊ̯

In the first half of the 20th Century, this diphthong made a rapid shift forward. For some speakers, the nucleus of the diphthong went as far as / ɛ̈ /. Because the diphthong shifted back to [əʊ̯] as the century proceeded, the fronted position is both a strong marker of Antique RP, and a social marker for those at the top of the social, economic and academic hierarchy, In other words, it’s almost a caricature of poshness.

The following passage from Queen Elizabeth’s 1958 Christmas broadcast shows this strongly fronted form. I should also apologize for my editing, which has made her speech into a bizarre caricature of logic. I promise she made more sense than this before I started cutting.

Last October, I opened the new Canadian parliament to the peace of my own home, so that we can show the world everything that we know is right, what to hold onto, and my devotion to these old islands. And as you know, in the old days the monarch led his soldiers on the battlefield, as if they were old and thrown aside, as they are at this moment. But no one was close and personal.

Here’s the same passage with the GOAT words extended:

And here are some more central examples:


NEAR nɪə̯ SQUARE skʍɛ̞ START stɑːtv NORTH nɔːθ CURE kçɔː lettER ˈlɛtә NURSEnɜːs

All of our London accents are non-rhotic, so we can apply (almost) all the same rules as we would with MLE or Cockney. We’ll need to take a closer look at CURE NURSE and lettER, but the general rule is that if an /r/ phoneme occurs directly after a vowel in the same syllable, it isn’t pronounced.

START – One day last year
NORTH (FORCE) – It felt as if I were riding a bucking horse
CURE – …away with it by pure luck.
lettER – I never utter a libel
NURSE – and earthly power doth then show likest God’s when mercy seasons justice

CURE ɔː ʊə̯

You can hear in the CURE sample above, that Maggie Smith retains the older form of [ʊə̯]. This realization shifted rather early in the century towards [ɔː] thus merging CURE with NORTH and FORCE, and for that matter, THOUGHT.

NURSE ɜː ɛ̽

In the sample below, you can hear Queen Elizabeth pronouncing NURSE words with a vowel so far forward as to nearly arrive at [ɛ]:

We will certainly lose the trust and respect of the world who had worked together as friends. And as you know this was the first time, more personal and direct.

lettER ˈlɛtə͡ ɹɪz  ˈlɛtə tu

JC Wells laid out both commA and lettER sets to account for the fact that in a non-rhotic accent of English, commA and lettER words end in the same sound. Nevertheless, they are identical realizations of different phonemes. Words in the lettER set, like Peter or tuner sound exactly like pita or tuna , but because of the underlying /r/ phoneme, those words would behave differently when immediately followed by a vowel. Peter is would be pronounced [ˈpitə͡ ɹɪz], but Peter says would be pronounced [ˈpitə. sɛz]. In MLE, Cockney, and Modern RP, both commA and lettER would occasion this linking /r/, even though commA has no underlying /r/. This is called an intrusive R.

in this sample, you can hear Maggie Smith taking things a bit too far, and resisting the linking /r/ in two lettER words followed by a vowel.


Now, sometimes this occurs inside a word. For example, the word very is a SQUARE word, and that diphthong is immediately followed in the next syllable by a vowel. There are four features of this situation that I’d like to draw out.

First is the linking /r/. Second is the tendency for these linking or intervocalic /r/s to be realized as a tap [ɾ]. We’ve actually encountered this in the TRAP+R section. Third is the fact that words like this are actually not SQUARE, but DRESS+R! There are a few of these subsets that reveal that historically, our rhotic diphthongs started as a sequence of vowel+R:

DRESS + R and TRAP+ R became SQUARE
KIT + R became NEAR
STRUT + R became NURSE
CLOTH + R became NORTH

IN RP, and especially in Antique RP, this process is resisted in intervocalic positions, and that gives us

very ˈvɛɾɪ
marry     ˈmæ.ɹɪ
miracle ˈmɪɾɪkᶦɫ̩
courage   ˈkɐ.ɹɪʤ
foreign ˈfɒ.ɹən

Fourth… well, let’s give #4 it’s own section:

happY ˈhæpɪ

This lexical set contains the words that end with an unstressed FLEECE vowel. Wells didn’t provide us with many sets of unstressed vowels, and he did so here because there are some interesting variations. First, there’s the Early Modern English pronunciation of [ə̯i] which shows up in the rhyme in MND of eye and remedy. Something similar is still current in Yorkshire. Antique RP has [ˈhæpɪ] or [ˈhæpɪː]. This is a notable marker of the distinction between Antique and Traditional RP. The newer version with [i] is distinctive enough that Wells gave it a name “happY Tensing”.

Here are several from HRH Queen Elizabeth II

Happy Christmas. Because television has made it possible for many of you to see me. I very much hope my own family never really touches your personal lives. At this critical moment in our history, we will certainly lose the greatness of this country amicably and peacefully. things are very different in our qualities. I’m not surprised that many people feel the difficulty of a very happy Christmas.

and with these sounds slowed:

Happy Christmas. Because television has made it possible for many of you to see me. I very much hope my own family never really touches your personal lives. At this critical moment in our history, we will certainly lose the greatness of this country amicably and peacefully. things are very different in our qualities. I’m not surprised that many people feel the difficulty of a very happy Christmas.

DUKE → djuk

There is a subset of words in the GOOSE set that begin the /u/ vowel with a /j/. This difference between /u/ and /ju/ has to do with the history of these words and how theyve changed over time. This is why “mute” and “moot” have a different pronunciation. However, we can usually tell which words are in this /ju/ subset because of the consonant that precedes the /u/. Most English speakers put “beauty” and “music” and “fume” in this category — we don’t do this with “booty” and “mood” and “food”, but that’s because they have a different history, as indicated by their spelling.

RP has a more expansive category of GOOSE+ j words. These include duke, and tune and new. Antique RP is even more expansive and includes lute and suit.



Intervocalic /t/ is unvoiced, and aspirated. This is in contrast with most speakers of American English who use an unvoiced, or at least unaspirated /t/. Of course, many words have medial /d/ in both RP and American accents, so one should be careful not to generalize the rule to the devoicing and aspirating of all such sounds. RP speakers pronounce “Betty” as [ˈbe̞tɪ] but they don’t pronounce “ready” as [ˈɹe̞tɪ]. Modern RP differs from older RP in that /t/ is often given a bit of reinforcement by a glottal stop : /t͡ʔ/

Some Calisthenics

Even after mastering individual realizations of phonemes, is still be a challenge to execute a particular sequence because of the athleticism of the tongue movements, or because the targets aimed at by the accent seem counter to our own accent intuitions. For that reason it’s useful to practice particularly challenging sequences in order to develop ease and agility.

I thought you brought hot coffee to father.

Roger’s daughter was dancing and walking on rocks.

Please understand—you cannot demand that I clasp your hand.

Can the cast eat that vast vat of pasta?  I plan to pass.


shibboleth : Wikipedia

There are a number of words and proper names whose standard RP pronunciation differs from the standard American one.  In the case of names, especially, these pronunciations are often wildly counter-intuitive.  When in doubt, ask someone who knows or look it up.  A (very partial) listː

figure                /ˈfiɡə/

clerk                  /klɑːk/

herb                  /hɜːb/

lieutenant        /lɛfˈˈtɛnənt/

controversy     /kənˈtɹɒvəsɪ/

privacy             /ˈpɹɪvəsɪ/

schedule           /ˈʃɛʤuɫ/

vitamin             /ˈvɪtəˌmɪn/

urinal                /jʊə̆ˈɹaɪnɫ̩/

garage               /ˈɡæˌɹɑːʤ/

process             /ˈpɹəʊsɛs/                          

program           /ˈpɹəʊɡɹæm/

leisure               /ˈlɛʒə/

zebra                 /ˈzɛbɹə/

(the letter) Z    /zɛd/                          

aluminum       /ˌælʊˈmɪnɪəm/

simultaneous   /ˌsɪmɫ̩ˈteɪ̆njəs/

cigarette           /ˌsɪɡəˈɹɛt/                          

lever                  /ˈlivə/

issue                  /ˈɪsju/

missile              /ˈmɪsaɪ̆ɫ/

mobile              /ˈməʊ̆baɪɫ/

futile /ˈfjutaɪ̆ɫ/

ballet                 /ˈbæleɪ̆/

brochure          /ˈbɹoʊ̆ʃə/

valet                  /ˈvælɪt/ or /ˈvæleɪ̆/           

comrade           /ˈkɒmɹeɪ̆d/

cafe                   /kæf/ or /ˈkæfeɪ̆/           

corollary          /kəˈɹɒləˌɹɪ/

ate                     /ɛt/                          

yogurt               /ˈjɒɡət/

debris                /ˈdɛbɹɪ/                          

premature        /ˈpɹɛməˌtçʊə/

glacier              /ˈɡlæsjə/                          

St. John             /ˈsɪndʒən/

Hertford           /hɑːfəd/

Pall Mall           /pæl mæl/

Warwick           /ˈwɒɹɪk/                          

Gloucester        /ˈɡlɒstə/

Leicester           /ˈlɛstə/                          

Beaulieu           /ˈbjulɪ/

Beauchamp      /ˈbiːtʃəm/                          

Caius                 /kʰiːz/

Magdalen         /ˈmɔdlɪn/                          

Belvoir              /ˈbivə/

Derby                /ˈdɑbɪ/                          

Marlborough   /ˈmɔːlbɹə/

(Caius, Magdalen are colleges at Cambridge and Oxford.  Beauchamp and Belvoir are castles, and also common street names.)