English-language vowel changes before historic /r/
In the phonological history of the English language, vowels followed (or formerly followed) by the phoneme /r/ have undergone a number of phonological changes. In recent centuries, most or all of these changes have involved merging of vowel distinctions.
- 1 Overview
- 2 Mergers before intervocalic R
- 3 Mergers and splits before historic coda r
- 4 See also
- 5 Sound samples
- 6 Notes
- 7 References
In most North American accents, for example, although there are ten or eleven stressed monophthongs, only five or six vowel contrasts are possible before a following /r/ in the same syllable (peer, pear, purr, pore, par, poor). Often, more contrasts exist when the /r/ is not in the same syllable; in some American dialects and in most native English dialects outside North America, for example, mirror and nearer do not rhyme, and some or all of marry, merry and Mary are pronounced distinctly. (In North America, these distinctions are most likely to occur in New York City, Philadelphia, Eastern New England (including Boston), and in conservative Southern accents.) In nearly all dialects, however, the number of contrasts in this position is reduced, and the tendency is towards further reduction. The difference in how these reductions have been manifested represents one of the greatest sources of cross-dialect variation.
Non-rhotic accents often show mergers in the same positions as rhotic accents do, even though there is often no /r/ phoneme present. This results partly from mergers that occurred before the /r/ was lost, and partly from later mergers of the centering diphthongs and long vowels that resulted from the loss of /r/.
In some cases, the quality of a vowel before /r/ is different from the quality of the vowel elsewhere. For example, in American English the quality of the vowel in more typically does not occur except before /r/, and is somewhere in between the vowels of maw and mow. (It is similar to the vowel of the latter word, but without the glide.)
Different mergers occur in different dialects. Among United States accents, the Boston and New York accents have the least degree of pre-rhotic merging. Some have observed that rhotic North American accents are more likely to have such merging than non-rhotic accents, but this cannot be said of rhotic British accents like Scottish English, which is firmly rhotic and yet many varieties have all the same vowel contrasts before /r/ as before any other consonant.
Mergers before intervocalic R
One of the best-known pre-rhotic mergers is known as the Mary–marry–merry merger, which consists of the mergers before intervocalic /r/ of /æ/ and /ɛ/ with historical /eɪ/. This merger is quite widespread in North America.[sample 1] The following variants are common in North America:
- All three are kept distinct. This occurs primarily in the Northeast, e.g. in the accents of Philadelphia, New York City, Boston, and Providence, Rhode Island.[sample 2] In the Philadelphia accent the three-way contrast is preserved, but merry tends to be merged with Murray; likewise ferry can be a homophone of furry. (See furry–ferry merger below.)
- All three are merged. This is found throughout the United States and Canada, and is almost universal in the West.
- "Mary" and "marry" are merged, while "merry" is distinct. This is also found widely, but in only about 15% of speakers in the United States overall.
- "Mary" and "merry" are merged, while "marry" is distinct. This is found in the South of the United States and as far north as Baltimore, Maryland, and Wilmington, Delaware; it is also found among Anglophones in Montreal.
The three are kept distinct outside of North America. In accents that do not have the merger, Mary has the "a" sound of "mare" (/ˈmeɪri/ or /ˈmɛəri/), marry has the "a" sound of "mat" (/ˈmæri/), and merry has the "e" sound of "met" (/ˈmɛri/). There is plenty of variance in the distribution of the merger, with expatriate communities of these speakers being formed all over the country.
Another widespread merger is that of /ɪ/ with /iː/ before intervocalic /r/. For speakers with this merger, mirror and nearer rhyme, unless mirror has merged with mere, and Sirius is homophonous with serious. North Americans who do not merge these vowels often speak the more conservative northeastern or southern accents.
A merger found to some degree in most accents of North American English, especially those that pronounce /r/ as a labialized retroflex approximant [ɻʷ]. An unstressed /rəC/ or /rɨC/ can elide the /ə/ to become /rC/, deleting a syllable as a result. The most common is /rər/ or /rɨr/ reducing to [r] anywhere in a word, producing possible homophones like error-air, horror-whore, mirror-mere and terror-tear. This can be an uneven merger, with mirror-mere merging while most other pairs do not. In rapid speech, this can also cause words ending with /r/ to merge with their respective inflected forms suffixed with -er, producing homophones like near-nearer and sure-surer, though speakers may subtly enunciate the difference to keep the words distinct, such as [ˈnɪr] vs. [ˈnɪrː ~ ˈnɪr.r̩]. Other sequences of /rəC/ usually reduce non-word-finally, producing possible homophones like coroner-corner, Morrigan-Morgan and Oregon-organ. Word-final sequences tend not to reduce, but there are exceptions such as orange [ˈɔrndʒ]. The mirror-mere merger is usually strongest in the Western United States, and most absent in the Northeastern United States, the Southern United States and from non-rhotic North American accents in general.
The merger of /ʌ/ before intervocalic /r/ with /ɝ/ is also widespread in American English apart from the Northeast and the South of the US. Speakers with this merger pronounce hurry to rhyme with furry. Other pairs such as stir it versus turret are split in all accents outside North America.
The merger of /ɛ/ and /ʌ/ before /r/ (both neutralized with syllabic r) is common in the Philadelphia accent. This accent does not usually have the marry–merry merger. That is, "short a" /æ/ as in marry is a distinct unmerged class before /r/. Thus, merry and Murray are pronounced the same, but marry is distinct from this pair.
Historic "short o" before intervocalic R
Words that have /ɒ/ before intervocalic /r/ in RP are treated differently in different varieties of North American English. As shown in the table below, in Canadian English, all of these are pronounced with [-ɔr-], as in cord (and thus merge with historic prevocalic /ɔːr/ in words like glory because of the horse–hoarse merger). In the accents of New York, Philadelphia, and the Carolinas (and traditionally throughout the South), these words are pronounced among some with [-ɑr-], as in card (and thus merge with historic prevocalic /ɑːr/ in words like starry). In the Boston accent these words are pronounced with [-ɒr-], just like in RP. Most of the rest of the United States (marked "Gen.Am." in the table), however, has a mixed system: while the majority of words are pronounced as in Canada, the four words in the right-hand column are typically pronounced with [-ɑr-].
|RP and Boston /ɒr/|
|NYC, Philadelphia, and the Carolinas /ɑr/|
|Gen.Am. /ɔr/||Gen.Am. /ɑr/|
Even in the Northeastern accents without the split (Boston, New York, Philadelphia), some of the words in the original short-o class often show influence from other American dialects and end up with [-ɔr-] anyway. For instance, some speakers from the Northeast may, for example, pronounce Florida, orange, and horrible with [-ɑr-], but foreign and origin with [-ɔr-]. Exactly which words are affected by this differs from dialect to dialect and occasionally from speaker to speaker, an example of sound change by lexical diffusion.
Mergers and splits before historic coda r
The near-square merger is the merger of the Early Modern English sequences /iːr/ and /ɛːr/ (and the /eːr/ between them), which is found in some accents of modern English. Some speakers in New York City and New Zealand merge them in favor of the NEAR vowel, while some speakers in East Anglia and South Carolina merge them in favor of the SQUARE vowel. The merger is widespread in the Anglophone Caribbean.
The fern–fir–fur merger is the merger of the Middle English vowels /ɛ, ɪ, ʊ/ into [ɜ] when historically followed by /r/ in the coda of the syllable. As a result of this merger, the vowels in fern, fir and fur are the same in almost all accents of English; the exceptions are Scottish English and some varieties of Hiberno-English.
- The 〈ere〉 spelling in the word were (as conjugated form of to be) is observed to be the fern vowel in Scottish English and in Hiberno-English.
- But the two accents disagree on how to pronounce the 〈ear〉 spelling (representing historical */er/ no longer independently distinguished) in words like dearth, earl, early, earn, earnest, Earp, earth, heard, hearse, Hearst, learn, learnt, pearl, rehearse, search and yearn:
The square–nurse merger is a merger of /ɜː(r)/ with /eə(r)/ that occurs in some accents (for example Liverpool, Dublin, and Belfast). The phonemes are merged to [ɛ:] in Hull and Middlesbrough.
This merger is found in some varieties of African American Vernacular English. (For example, in Chingy's song "Right Thurr", the merger is heard at the beginning of the song, but he goes on to use standard pronunciation for the rest of the song.[sample 3])
Labov (1994) also reports such a merger in some western parts of the United States 'with a high degree of r constriction.'
In older varieties of Southern American English and the West Country dialects of English English, words like ear, here, and beard are pronounced /jɝ/, /hjɝ/, /bjɝd/, meaning that there is no complete merger: word pairs like beer and burr are still distinguished as /bjɝ/ vs. /bɝ/. However, if the syllable begins with a consonant cluster (e.g. queer) or a palato-alveolar consonant (e.g. cheer), then there is no /j/ sound: /kwɝ/, /tʃɝ/. It is thus possible that pairs like steer-stir are merged in some accents as /stɝ/, although this is not explicitly reported in the literature.
There is evidence that African American Vernacular English speakers in Memphis, Tennessee, merge both /ɪr/ and /ɛr/ with /ɝ/, so that here and hair are both homophonous with the strong pronunciation of her.
Tower–tire, tower–tar and tire–tar mergers
The tower–tire and tower–tar mergers are vowel mergers in some accents of Southern British English (including many types of RP, as well as the accent of Norwich) that causes the triphthong /aʊə/ of tower to merge either with the /aɪə/ of tire (both surfacing as diphthongal /ɑə/) or with the /ɑː/ of tar. Some speakers merge all three sounds, so that tower, tire, and tar are all homophonous as /tɑː/.
The tire–tar merger, with tower kept distinct, is found in some Midland and Southern U.S. accents.
In East Anglia a merger with the [ɜː] of shirt is common, especially after palatal and palatoalveolar consonants, so that sure is often pronounced [ʃɜː]; yod-dropping may apply as well, yielding pronunciations such as [pɜː] for pure. Similarly in American English sure is often pronounced /ʃɝ/. Other American pronunciations showing this merger include /pjɝ/ pure, /ˈkjɝiəs/ curious, /ˈbjɝo/ bureau, /ˈmjɝəl/ mural.
In Modern English dialects, the reflexes of Early Modern English /uːr/ and /iur/ are highly susceptible to phonemic merger with other vowels. Words belonging to this class are most commonly spelled with oor, our, ure, or eur; examples include poor, tour, cure, Europe.[dubious ] Wells refers to this class as the CURE words, after the keyword of the lexical set to which he assigns them.
In traditional Received Pronunciation and General American, CURE words are pronounced with RP /ʊə/ (/ʊər/ before a vowel) and GenAm /ʊr/.[dubious ] But these pronunciations are being replaced by other pronunciations in many English accents.
In southern English English it is now common to pronounce CURE words with /ɔː/, so that moor is often pronounced /mɔː/, tour /tɔː/, poor /pɔː/. The traditional form is much more common in the northern counties of England. A similar merger is encountered in many varieties of American English, where the pronunciations [oə] or [or]~[ɔr] (depending on whether the accent is rhotic or non-rhotic) prevail.
The pure–poor split is a phonemic split that occurs in Australian and New Zealand English that causes the centring diphthong /ʊə/ to disappear and split into /ʉː.ə/ (a sequence of two separate monophthongs) and /oː/ (a long monophthong), causing pure, cure, and tour to rhyme with fewer, and poor, moor and sure to rhyme with for and paw.
Where the /ʊə/ becomes /ʉː.ə/ and where it becomes /oː/ is not very predictable. But words spelt with -oor that originally had /ʊə/ become /oː/ perhaps by influence of the words door and floor which rhyme with store in all dialects of English except a few older northern British dialects.
A similar split occurs in many varieties of North American English that causes /ʊr/ to disappear and split into /ɝ/ and /ɔr/, causing pure, cure, lure, sure to rhyme with fir, and poor, moor, Boor to rhyme with store and for. In many of these dialects, tour remains with /ʊr/, leading to a three-way split between tour /tʊr/ (although often sounding more like [tur]), pure /pjɝ/ and poor /pɔr/. In some others, tour changes to disyllabic [ˈtu.ɚ] or [ˈtu.ɪ̈ɹ].
The card–cord merger is a merger of Early Modern English [ɑr] with [ɒr], resulting in homophony of pairs like card/cord, barn/born and far/for. It is roughly similar to the father–bother merger, but before r. The merger is found in some Caribbean English accents, in some versions of the West Country accent in England, and in some Southern and Western U.S. accents. Areas where the merger occurs include central Texas, Utah, and St. Louis. Dialects with the card–cord merger don't have the horse–hoarse merger. The merger is disappearing in the United States, being replaced by the more common horse–hoarse merger that other regions have.
The horse–hoarse merger is the merger of the vowels /ɔ/ and /o/ before historic /r/, making pairs of words like horse/hoarse, for/four, war/wore, or/oar, morning/mourning etc. homophones. This merger occurs in most varieties of English. In accents that have the merger horse and hoarse are both pronounced [hɔː(ɹ)s], but in accents that do not have the merger hoarse is pronounced differently, usually [hoɹs] in rhotic and [hoəs] or the like in non-rhotic accents. Non-merging accents include Scottish English, Hiberno-English, the Boston accent, Southern American English, African American Vernacular English, most varieties of Caribbean English, and Indian English.
The distinction was made in traditional Received Pronunciation as represented in the first and second editions of the Oxford English Dictionary. The IPA symbols used are /ɔː/ for horse and /ɔə/ for hoarse.
In the United States, the merger is quite recent in some parts of the country. For example, fieldwork performed in the 1930s by Kurath and McDavid shows the contrast robustly present in the speech of Vermont, northern and western New York State, Virginia, central and southern West Virginia, and North Carolina; but by the 1990s telephone surveys conducted by Labov, Ash, and Boberg show these areas as having almost completely undergone the merger. And even in areas where the distinction is still made, the acoustic difference between the [ɔr] of horse and the [or] of hoarse is rather small for many speakers.
The two groups of words merged by this rule are called the lexical sets NORTH (including horse) and FORCE (including hoarse) by Wells (1982). Etymologically, the NORTH words had /ɒɹ/ and the FORCE words had /oːɹ/.
The orthography of a word often signals whether it belongs in the NORTH set or the FORCE set. The spellings war, quar, aur, and word-final or indicate NORTH (e.g., quarter, war, warm, warn, aura, aural, Thor). The spellings oVr or orV (where V stands for a vowel) indicate FORCE (e.g., board, coarse, hoarse, door, floor, course, pour, oral, more, historian, moron, glory). Words spelled or followed by a consonant mostly belong with NORTH, but the following exceptions are listed with FORCE by Wells (1982):
- The past participles borne (but not born), shorn, sworn, torn and worn
- Some (but not all) words where or follows a labial consonant:
- afford, force, ford, forge, fort (but not fortress), forth
- deport, export, import (but not important), porch, pork, port, portend, portent, porter, portion, portrait, proportion, report, sport, support
- The word corps (but not corpse; corps is a perfect homophone of core)
- The word horde (a perfect homophone of hoard)
The cork–quark merger is the merger of the sequences /kɔr/ and /kwɔr/, making cork and quark homophones [for those who do not rhyme quark with mark //]. This merger is generally observed to be strongest in American English. Even more minimal pairs become homophonous when accompanied with the horse–hoarse merger, which most General American accents also have. Additionally, some people may only merge "courter" and "quarter" while leaving all other pairs distinct.
- Phonological history of the English language
- Phonological history of English vowels
- coil–curl merger
- English phonology
- History of the English language
- Vocalic r
- http://www.alt-usage-english.org/mmm_bc.wav Sample of a speaker with the Mary–marry–merry merger Text: "Mary, dear, make me merry; say you'll marry me."
- http://www.alt-usage-english.org/mmm_rf.wav Sample of a speaker with the three-way distinction
- http://ldc.upenn.edu/myl/thurr.mp3 Text: "I like the way you do that right there (right there)/Swing your hips when you're walkin', let down your hair (let down your hair)/I like the way you do that right there (right there)/Lick your lips when you're talkin', that make me stare"
- ^ http://students.csci.unt.edu/~kun/tln.html[dead link]
- ^ Wells, pp. 479–485.
- ^ http://cfprod01.imt.uwm.edu/Dept/FLL/linguistics/dialect/staticmaps/q_15.html
- ^ Wells, p. 480-82
- ^ Labov et al., p. 54, 56
- ^ Labov et al., p. 56
- ^ Wells, pp. 201–2, 244
- ^ Labov et al., pp. 54, 238
- ^ Shitara
- ^ Wells, pp. 338, 512, 547, 557, 608
- ^ Wells, pp. 199–203, 407, 444
- ^ Wells, pp. 372, 421, 444
- ^ Kurath and McDavid, pp. 117–18 and maps 33–36.
- ^ http://www.ausp.memphis.edu/phonology/#Vocalic
- ^ Wells, pp. 238–42, 286, 292–93, 339
- ^ Kurath and McDavid, p. 122
- ^ Wells, p. 164
- ^ Hammond, p. 52
- ^ Wells, pp. 56, 65–66, 164, 237, 287–88
- ^ Kenyon, pp. 233–34
- ^ Wells, p. 549
- ^ http://www.ling.mq.edu.au/speech/phonetics/phonology/features/auseng_features.html[dead link]
- ^ Macquarie University Dictionary and other dictionaries of Australian English
- ^ Labov et al., pp. 51–53
- ^ Wells, pp. 158, 160, 347, 483, 548, 576–77, 582, 587
- ^ Labov et al., p. 52
- ^ http://www.ling.upenn.edu/phonoatlas/Atlas_chapters/Ch8/Ch8.html
- ^ Wells, pp. 159–61, 234–36, 287, 408, 421, 483, 549–50, 557, 579, 626
- ^ Kurath and McDavid, map 44
- ^ Labov et al., map 8.2
- ^ Labov et al., p. 51
- http://www4.uwm.edu/FLL/linguistics/dialect/staticmaps/q_15.html, .
- Handbook of Varieties of English, page 125, Walter de Gruyter, 2004
- Williams and Kerswill in Urban Voices, Arnold, London, 1999, page 146
- Williams and Kerswill in Urban Voices, Arnold, London, 1999, page 143
- hoss, Dictionary.com
- antimoon.com: Discussion of regional pronunciations of quirk, cork and quark.
- Hammond, Michael. (1999). The Phonology of English. Oxford: Oxford University Press. ISBN 0-19-823797-9.
- Kenyon, John S. (1951). American Pronunciation (10th ed.). Ann Arbor, Michigan: George Wahr Publishing Company. ISBN 1-884739-08-3.
- Kurath, Hans, and Raven I. McDavid (1961). The Pronunciation of English in the Atlantic States. Ann Arbor: University of Michigan Press. ISBN 0-8173-0129-1.
- Labov, Wiliam, Sharon Ash, and Charles Boberg (2006). The Atlas of North American English. Berlin: Mouton de Gruyter. ISBN 3-11-016746-8.
- Shitara, Yuko (1993). "A survey of American pronunciation preferences". Speech Hearing and Language 7: 201–32.
- Wells, John C. (1982). Accents of English. Cambridge: Cambridge University Press. ISBN 0-521-22919-7: