Intrusive r


There are many articles available on the internet about the so-called ‘intrusive r’, but as a visitor to this blog has written asking me to correct people who “persist in inserting an extra [R] between the [W] and the [I] (of ‘drawing’) making the word into [DRAWRING] ! WRONG !!!!!”, perhaps yet another will not be out of place. I mentioned the word drawer and its confusion with draw in a previous post (here), but now for something longer.

Speakers of the English language can be divided in many ways. One of these is into the two classes of ‘rhotic’ and ‘non-rhotic’.  By this is meant that some varieties of English pronounce all orthographic ‘r’, (the rhotic group), while others do not (the non-rhotic speakers).

Non-rhotic accents occur in most of England, the whole of Australia and New Zealand, South Africa, and some parts of the United States, notably New England and the South, although the position in US English is complicated and changing. Scotland, Ireland, Canada and most of the United States are traditionally thought of as being rhotic.

Non-rhoticity in English means that /r/ as a phoneme occurs only when the following sound is a consonant. Every pre-consonantal and pre-final orthographic ‘r’ is dropped. In accents of England, this has been noticed for over 200 years. The consequence of this r-dropping is that words such as idea and near, saw and sore, farm and calm now rhyme (/aɪˈdɪə/~/nɪə/, /sɔː/~/sɔː/,  /fɑːm/~/kɑːm/). The majority of words ending in these vowel sounds (/ə, ɔː, ɑː/) similarly have no following orthographic ‘r’, but a minority do end in a written ‘r’. When a word such as near or sore is followed by another word, or a suffix, that begins with a vowel, the orthographic ‘r’ is pronounced, as it would be if the ‘r’ occurred at the beginning of a word, immediately before a vowel. So we get the phrase near and far /ˈnɪər ən ˈfɑː/, where the first orthographic ‘r’ is pronounced, or reversing the words: /ˈfɑːr ən ˈnɪə/. By analogy with these words, the rhyming words which happen not to end in ‘r’, acquire an /r/ sound in the same situations. Hence, the idea of … becomes /ðɪ ˈaɪdɪər əv …/ and drawing becomes /ˈdrɔːrɪŋ/.

My complainant says that “I consider it to be ignorance of the learning of pronunciation.” I suspect that he, like every other native speaker of English, learned his pronunciation from the people around him. As non-rhotics have been around for at least two hundred years, the current perpetrators of this horror (in his view) must have learned their errors from their parents, siblings and friends, just as he has learned his (presumably) rhotic ways from his.


  1. Non-rhoticity in English means that /r/ as a phoneme occurs only when the following sound is a consonant.

    Is a vowel, presumably.

    Also, note that both AAVE dialect and AAVE-influenced standard American English, which are non-rhotic, do not have either linking or intrusive /r/: /ˈfɑː ən ˈnɪə/, with hiatus, is normal.

  2. There was a time when intrusive r after ɔ was unthinkable in RP, even between words, although after a schwa it was considered perfectly acceptable (and its lack was equally acceptable). But times have changed, and most speakers of contemporary RP wouldn’t blink at this feature. So far as I know, it has never been stigmatized in English regional speech, but I’m willing to stand corrected if I am mistaken.

    Intrusive r is highly stigmatized in American speech – even where it is in common use, as it still is in the New York Metropolitan area, Boston, and – again, so far as I know – other parts of the American northeast.

    The visitor is, in my view, mistaken. So long as the intended audience can understand what is being said, the speaker’s English – no matter what the accent – is not “wrong,” only different. There are too many forms of English around the world, and too many accents among speakers of these Englishes, to be quite so didactic about someone else’s accent.

  3. It should be clear, I hope, that despite the stigmatization of intrusive r in American speech, and despite the fact that I don’t have intrusive r in my own speech, I don’t agree with those who perpetuate the stigma. It can be very difficult to let go of old prejudices, in language as about anything else, but prejudice is all it is.

  4. John Cowan – Thanks for the correction! Stupid mistake. Obviously written too quickly.

  5. In certain cases of intrusive R such as “Asia(r) and Africa” there is a need for _some_ articulatory gesture to mark the word boundary between the schwas of “Asia” and “and”. If intrusive R is not used, this could be a glottal stop, an increase in intensity, a small change in tongue position, etc.

    However, in a case such as “draw(r)ing” or “law(r) and order” there is no such need. A pronunciation of /’drɔ:.ɪŋ/ is completely possible from a euphonic point of view. This makes me very interested in AmyStoller’s claim that intrusive R in such words is more recent, at least in RP, than that between schwas.

  6. If anyone wants to go into the history etc of this topic they might like to look at,uk where the sixth article in Section 3 is “Linking /r/ in the General British pronunciation of English” which has quite a lot on the phenomenon in the twentieth century. A very recent study of the topic in the speech of television British newsreaders was part of a 2006 thesis by Dr Bente R. Hannisdal available for download as a pdf.

  7. @Amy Stoller:
    Do you happen to have any (written or spoken) evidence of stigmatisation of intrusive /r/ in American English?

    @John Cowan:
    Would you call Martin Luther King a member of the AAVE group?

  8. Actually, /r/ can occur pre-consonantally in “non-rhotic” accents. You cannot be /sɪərjəs/! Oh yes I am.

    Oh and by the way, isn’t “ferrule” a nice word?

  9. John M,
    Rhotic and non-rhotic are terms I never use coz they seem too sweeping etc. I prefer to describe accents as high-rhoticity and low-rhoticity. The lowest-rhoticity accent of English is deep-south US. As to pre-consonantal /r/s in the General British accent, I gave numerous examples in my article#4 at Section 3 of my website (

  10. Jack,

    A very sensible policy which I shall adopt.

  11. Hey guys,
    can anyone give me an example of an intrusive ‘r’ after the /a:/ sound?

  12. How about “Panama and Mexico”: /ˈpænəmɑːrəm ˈmeksɪkəʊ/?

  13. awsome thanks! I was confused because I thought the final /a:/ sound was actually an /ǝ/. But I guess you are right.

    If you wanna help me with another problem I would appreciate it (even though it has nothing to do with the intrusive ‘r’). I’m not a native English speaker so pronunciation sometimes gives me a hard time.
    The problem is that I have to find three ways in which the pronunciation of the word cease differs from sees. My suggestions are:

    1.The final “s” in sees [si:z] is a voiced /z/ sound, whereas the “s” is voiceless in cease [si:s].
    2.The duration of the vowel /i/ is slightly longer in sees than in cease.
    3.The lip position of the /i/ vowel is slightly more spread in sees than it is in cease.

    If anyone would like to comment on them please fell free to do so.

  14. Paul – Your first two differences are quite right, but I don’t feel that I make any distinction in lip position between the two words. If you “have to” find three ways in which they are different, I imagine this must be for an assignment. Please let us all know the “correct” answer when you get it!

  15. As a native speaker of English English, I understand the intrusive R. What baffles me is that the same people who say lawRandorder also say WestminstaAbbey. When the Piper Alpha oil rig collapsed, I don’t think a single BBC professional ever said anything but PipaAlpha.

    So why do they miss it out when it should be there?

  16. @mike:

    Both linking and intrusive R are optional features of the accents that allow them.

    They are both less frequent when adjacent to proper names (as in your “Westminster Abbey” and “Piper Alpha”). The study that found this is here (linked to from the intrusive / linking R Wikipedia page).

  17. Thanks dw; she makes some plausible speculations in 6.3. Although both my parents were RP speakers, I grew up in Scotland, where hardly anyone said PipaAlpha, so London BBC was very noticeable.

    In discussion I have found some English RP speakers unable to hear the distinction until exaggerated. Brain patterning is obviously important.

  18. Very interesting!
    But why, Why are there very few instances of intrusive /r/ in other vowels such as the vowel in bird ?

  19. @Kat:

    The vowel of “bird” nearly always derives from words that historically contained /r/, and are therefore spelled with R. It’s therefore not possible for them to feature intrusive R (if a phonetic [r] occurred, it would instead be classified as linking R).

    The only exceptions would be recent loanwords and proper names from languages such as French and German, whose rounded front vowels are often approximated in Southern British English by the vowel of “bird”. For example, the name of French football (soccer) player Frank LeBoeuf was usually rendered /ləˈbɜ:f/ in the English media. However, for such a word to supply an example of intrusive R, the vowel would have to be word-final, and I’m having a hard time coming up with any examples.

    In American English, the FACE, GOAT or GOOSE vowels are usually used instead for such loanwords.

  20. dw – the only word I can find is “milieu”, which LPD gives as /miːljɜː/ (with stress on either syllable) for the principal pronunciation.

  21. And I’ve definitely heard /miːljɜː/ with linking r, probably as often as not. (Sorry, JWL, no chapter and verse.)

  22. This is what I gather from my English phonetics professor:

    In non-rhotic accents of English, r only occurs prevocalically – it only occurs when a vowel follows it.

    With regard to intrusive-r:

    R is inserted (intrudes) when it occurs between ə, ɜ:, ɑ:, ɔ: and another vowel.

    More on this site:

  23. I cannot see (or hear) a need for the intrusive “r”. Many commentators seem to think it is a useful articular gesture (a nice term) to demarcate between words, perhaps as in french where silent endings must become vocal in order for the next word beginning with a vowel to be clearly heard.

    In the examples of “law and order”, “Asia and Africa”, or even in “drawing room”, I can see no need for any “articular gesture” – apart from a small pause perhaps. If you say the word “law” and then the word “and”, I suppose an inadvertent, minimal “glottal stop” occurs.

    I think a “glottal stop” will occur as a consequence of the “schwa” occurring before the next word which begins with a vowel – pretty useful, I think. Perhaps I am misunderstanding the “schwa” and its purpose. Please feel free to clarify the matter.

    I am also quite intrigued by a possible american “schwa” sound. The american pronunciation of “can” and “can’t” are very similar and often leave me confused as to whether the speaker is being affirmative or answering in the negative.

    In England, the difference between “can” and “can’t” seems to be marked by the use of two different “schwa” sounds (or are they just different vowel sounds?) as in “bar” and “rant”.

    I have always believed that american English is much closer to 18th century English, and that english English may have evolved more. An evolution towards clarity (“can” v. “can’t”) seems highly advantageous for oral communication.

    I’d love to hear some thoughts on the above.

  24. Frank –
    I think you are misunderstanding what is meant by ‘schwa’. As you might imagine, the word is borrowed into English in this form from German, but it is originally a Hebrew word, and in linguistics is used to mean a ‘neutral’ vowel. This is a vowel produced when the highest point of the centre of the tongue is neither particularly forward in the mouth, nor particularly back, and neither very close, nor very far away from the roof of the mouth. In many accents of English, it is the vowel heard when we hesitate and say ‘er, er’ or perhaps ‘uh, uh’ before finding the words we are searching for. Perhaps another phonetician reading this will be able to put it better. So it is neither the -a- of “can” nor that of “can’t” in either US English or British English. Rather it is the sound of the ‘a’ in “around”, the ‘e’ in “often”, or the ‘o’ in “ribbon”.
    Moving to the “intrusive ‘r'”. I agree that there is no ‘need’ for it. It arises in those accents of English in which the /r/ sound coming before a consonant (either when the /r/ is final in a word, before a word beginning with a consonant, as in “war game”, or before silence, or within a word and followed immediately by a consonant as in “ward”) has disappeared. There are people who believe this is just slovenly speech on the part of those who have this accent, but I would remind them that no-one pronounces the ‘l’ in a word such as “would”, even though it was pronounced at one time, and no-one is ever criticised for saying /wʊd/ rather than /wʊld/. The loss of sounds in this way is a perfectly normal process in all languages.
    Following the loss of this pre-consonantal /r/ sound, which remains in the spelling, words such as “pours” and “paws” come to sound as homophones, i.e. they are identical to the ear. Remove the final -s-, and they are still identical, both being pronounced /pɔː/. Most of the words which end in /ɔː/ actually have an ‘r’ in the spelling, and when these words are followed by a word beginning with a vowel, the /r/ is still pronounced: War and Peace” /wɔːr ənd piːs/ (and this also contains an example of schwa, symbolised by the upturned ‘e’). By analogy with this returning /r/, words not containing an etymological /r/, but ending in /ɔː/, acquire an /r/ when they are followed by a word beginning with a vowel, as in “law and order” /lɔː r ənd ɔːdə/. I have deliberately written the /r/ separate from the preceding vowel to show that it is not a necessary part of either word, but something that only occurs as a feature of the junction of the two words. The same argument applies to words ending in either schwa or the long vowel /ɑː/. It also extends to the situation that occurs word internally when one of these vowels is followed immediately by another vowel, as in your example of “drawing room”.
    The “purpose” of sounds in language is to communicate between people. They have no purpose individually except perhaps to help distinguish one word from another, but the number of words which sound identical in any language makes this a dubious claim. Context is all. There are relatively few words in English distinguished solely by the presence or absence of /r/ before a consonant, and so it was quite easy for this ‘articulatory gesture’ to gradually disappear, just as initial /k/ before a consonant disappeared centuries before (words such as “knife”, “know”, “knight”). Changes do not take place universally throughout a language at the same time, or even ever, so there are still large parts of the English-speaking world where pre-consonantal and final /r/ is still pronounced.
    I hope this provides you with something of an answer.

Leave a Reply

Required fields are marked *.