Phonological history of English diphthongs

English diphthongs have undergone many changes since the Old and Middle English periods. The sound changes discussed here involved at least one phoneme which historically was a diphthong.

Old English

Further information: Old English phonology and Phonological history of Old English

Old English diphthongs could be short or long. Both kinds arose from sound changes occurring in Old English itself, although the long forms sometimes also developed from Proto-Germanic diphthongs. They were mostly of the height-harmonic type (both elements at the same height) with the second element further back than the first. The set of diphthongs that occurred depended on dialect (and their exact pronunciation is in any case uncertain). Typical diphthongs are considered to have been as follows:

• high, fully backing, /iu/ /iːu/, spelt ⟨io⟩ (found in Anglian dialects, but merged into /eo/ /eːo/ in Late West Saxon)
• high, narrower, possibly /iy/ /iːy/ or /ie/ /iːe/, spelt ⟨ie⟩ (found in Late West Saxon)
• mid, /eo/ /eːo/, spelt ⟨eo⟩
• low, /æɑ/ /æːɑ/, spelt ⟨ea⟩

As with monophthongs, the length of the diphthongs was not indicated in spelling, but in modern editions of OE texts the long forms are often written with a macron: ⟨īo⟩, ⟨īe⟩, ⟨ēo⟩, ⟨ēa⟩.

In the transition from Old to Middle English, all of these diphthongs generally merged with monophthongs.

Middle English

Further information: Middle English phonology

Development of new diphthongs

Although the Old English diphthongs merged into monophthongs, Middle English began to develop a new set of diphthongs, in which the second element was a high [i] or [u]. Many of these came about through vocalization of the palatal approximant [j] or the labio-velar approximant [w] (which was sometimes from an earlier voiced velar fricative [ɣ], an allophone of /g/), when they followed a vowel. For example:

• OE dæg ("day") and weg ("way") (where the /g/ had been palatized to /j/) became [dai] and [wɛi]
• OE clawu ("claw") and lagu ("law") became [klau] and [lau]

Diphthongs also arose as a result of vowel breaking before /h/ (which had allophones [x] and [ç] in this position – for the subsequent disappearance of these sounds, see H-loss). For example:

• OE streht ("straight") became [strɛiçt]
• OE þoht ("thought") became [θɔuxt]

The diphthongs that developed by these processes also came to be used in many loanwords, particularly those from Old French. For a table showing the development of the Middle English diphthongs, see Middle English phonology (diphthong equivalents).

Vein–vain merger

Early Middle English had two separate diphthongs /ɛi/ and /ai/. The vowel /ɛi/ was typically represented orthographically with "ei" or "ey" and the vowel /ai/ was typically represented orthographically with "ai" or ay". These came to be merged, perhaps by the fourteenth century. The merger is reflected in all dialects of present-day English.

In early Middle English, before the merger, way and day, which came from Old English weġ and dæġ, had /ei/ and /ai/ respectively. Similarly, vein and vain (borrowings from French) were pronounced differently as /vein/ and /vain/. After the merger, vein and vain were homophones, and way and day had the same vowel.

The merged vowel was a diphthong, transcribed /ɛi/ or /æi/. Later (around the 17th century) this diphthong would merge in most dialects with the monophthong of words like pane in the pane–pain merger.

Late Middle English

The English of southeastern England around 1400 had seven diphthongs, of which three ended in a front vowel:

• /ɛi/ as in nail, day, whey (the product of the vein–vain merger)
• /ɔi/ as in joy, noise, royal, coy
• /ʊi/ as in boil, destroy, coin, join

and four ended in a back vowel:

• /ɪu/ as in view, new, due, use, lute, suit, adieu (the product of a merger of earlier /iu/ and /eu/, also incorporating French loans that originally had /y/)
• /ɛu/ as in few, dew, ewe, shrewd, neuter, beauty
• /ɑu/ as in cause, law, salt, change, chamber, psalm, half, dance, aunt.
• /ɔu/ as in low, soul

Typical spellings are as in the examples above. The spelling ew is ambiguous between /ɪu/ and /ɛu/, and the spellings oi and oy are ambiguous between /ɔi/ and /ʊi/. The most common words with ew pronounced /ɛu/ were dew, few, hew, lewd, mew, newt, pewter, sew, shew (show), shrew, shrewd and strew. Words in which /ʊi/ was commonly used included boil, coin, destroy, join, moist, point, poison, soil, spoil, Troy, turmoil and voice, although there was significant variation.

Modern English

16th century

By the mid sixteenth century, the Great Vowel Shift had created two new diphthongs out of the former long close monophthongs /iː/ and /uː/ of Middle English. These diphthongs were /əɪ/ as in tide, and /əʊ/ as in house. At this time, the English of south-eastern England could thus have had nine diphthongs.

By the end of the sixteenth century, the inventory of diphthongs was reduced as a result of several developments, all of which took place in the mid-to-late sixteenth century:

• /ɛu/ merged into /ɪu/, so that dew and due became homophones.
• /ɛi/ (from the vein–vain merger) became monophthongized, merging with the /ɛː/ of words like name (which before the Great Vowel Shift had been long /aː/). For more on this, see pane–pain merger, below.
• /ɑu/, as in cause, became monophthongized to /ɒː/.
• /ɔu/, as in low, was monophthongized to /ɔː/. This would later rise to /oː/, merging with the vowel of toe; see toe–tow merger, below.

This left /ɪu/, /ɔi/, /ʊi/, /əɪ/ and /əʊ/ as the diphthongs of south-eastern England.

17th century

By the end of the seventeenth century, the following further developments had taken place in the dialect of south-eastern England:

• The falling diphthong /ɪu/ of due and dew changed to a rising diphthong, which became the sequence [juː]. This change did not occur in all dialects, however; see Yod-dropping.
• The diphthongs /əɪ/ and /əʊ/ of tide and house widened to /aɪ/ and /aʊ/ respectively.
• The diphthong /ʊi/ merged into /aɪ/. In literature from this period there frequently occur rhymes such as Mind–join'd in Congreve, join–line in Pope, child–spoil'd in Swift, toils–smiles in Dryden. The present-day pronunciations with /ɔɪ/ in these oi words result from regional variants which had always had [ɔi] rather than [ʊi], perhaps influenced by the spelling.

As a result of these changes, there remained only the three diphthongs /aɪ/, /aʊ/ and /ɔi/.

Later developments

In the 18th century or later, the monophthongs /eː/ and /oː/ (the products of the pane–pain and toe–tow mergers) became diphthongal in standard English. This produced the vowels /eɪ/ and /oʊ/. In modern-day RP, the starting point of the latter diphthong has become more centralized, and the vowel is commonly written /əʊ/.

RP has also developed centering diphthongs /ɪə/, /eə/, /ʊə/, as a result of breaking before /r/ and the loss of /r/ when not followed by another vowel (see English-language vowel changes before historic /r/). These occur in words like near, square and cure.

Present-day RP, then, is normally analyzed as having eight diphthongs: the five closing diphthongs /eɪ/, /əʊ/, /aɪ/, /aʊ/, /ɔɪ/ (of face, goat, price, mouth and choice) and the three centering diphthongs /ɪə/, /eə/, /ʊə/. General American does not have the centering diphthongs (at least, not as independent phonemes). For more information see English phonology (vowels).

Variation in present-day English

Long mid mergers

The earliest stage of Early Modern English had a contrast between the long mid monophthongs /eː, oː/ (as in pane and toe respectively) and the diphthongs /ɛi, ɔu/ (as in pain and tow respectively). In the vast majority of Modern English accents these have been merged, so that the pairs pane–pain and toe–tow are homophones. These mergers are grouped together by Wells as the long mid mergers.

Pane–pain merger

The pane–pain merger is a merger of the long mid monophthong /eː/ and the diphthong /ei/ that occurs in most dialects of English. In the vast majority of Modern English accents the vowels have been merged; whether the outcome is monophthongal or diphthongal depends on the accent. But in a few regional accents, including some in East Anglia, South Wales, and even Newfoundland, the merger has not gone through (at least not completely), so that pairs like pane/pain are distinct.

A distinction, with the pane words pronounced with [eː] and the pain words pronounced with [æɪ], survived in Norfolk English into the 20th century. Trudgill describes the disappearance of this distinction in Norfolk, saying that "This disappearance was being effected by the gradual and variable transfer of lexical items from the set of /eː/ to the set of /æɪ/ as part of dedialectalisation process, the end-point of which will soon be (a few speakers even today maintain a vestigial and variable distinction) the complete merger of the two lexical sets under /æɪ/ — the completion of a slow process of lexical diffusion."

Walters (2001) reports the survival of the distinction in the Welsh English spoken in the Rhondda Valley, with [eː] in the pane words and [ɛi] in the pain words.

In accents that preserve the distinction, the phoneme /ei/ is usually represented by the spellings ai, ay, ei and ey as in day, play, rain, pain, maid, rein, they etc. and the phoneme /eː/ is usually represented by aCe as in pane, plane, lane, late etc. and sometimes by eCe and e as in re, cafe, Santa Fe etc.

Homophonous pairs

/eː/ /ei/ IPA ade aid ˈeɪd ale ail ˈeɪl ate eight ˈeɪt With wait–weight merger. bale bail ˈbeɪl blare Blair ˈbleə(r) cane Cain ˈkeɪn clade clayed ˈkleɪd Clare Claire ˈkleə(r) bate bait ˈbeɪt Dane deign ˈdeɪn With wait–weight merger. daze days ˈdeɪz e'er air ˈeə(r) e'er heir ˈeə(r) ere air ˈeə(r) ere heir ˈeə(r) fare fair ˈfeə(r) faze fays ˈfeɪz flare flair ˈfleə(r) gale Gail ˈɡeɪl gate gait ˈɡeɪt gaze gays ˈɡeɪz glave glaive ˈɡleɪv Homonyms. grade grayed ˈɡreɪd graze grays ˈɡreɪz hale hail ˈheɪl hare hair ˈheə(r) haze hays ˈheɪz lane lain ˈleɪn laze lays ˈleɪz made maid ˈmeɪd Mae May ˈmeɪ male mail ˈmeɪl mane main ˈmeɪn maze maize ˈmeɪz maze Mays ˈmeɪz page Paige ˈpeɪdʒ pale pail ˈpeɪl pane pain ˈpeɪn pare pair ˈpeə(r) pear pair ˈpeə(r) phase fays ˈfeɪz phrase frays ˈfreɪz Rae ray ˈreɪ raze raise ˈreɪz raze rays ˈreɪz razor raiser ˈreɪzə(r) re ray ˈreɪ sale sail ˈseɪl sane sain ˈseɪn sane seine ˈseɪn sane Seine ˈseɪn spade spayed ˈspeɪd stare stair ˈsteə(r) suede swayed ˈsweɪd tale tail ˈteɪl there their ˈðeə(r) there they're ˈðeə(r) trade trayed ˈtreɪd vale vail ˈveɪl vale veil ˈveɪl vane vain ˈveɪn vane vein ˈveɪn wade weighed ˈweɪd With wait–weight merger. wale wail ˈweɪl wales wails ˈweɪlz Wales wails ˈweɪlz wane wain ˈweɪn waste waist ˈweɪst wave waive ˈweɪv waver waiver ˈweɪv whale wail ˈweɪl With wine-whine merger.

Toe–tow merger

The toe–tow merger is a merger of the post-Early Modern English vowels /oː/ (as in toe) and /ou/ (as in tow) that occurs in most dialects of English. (The vowels in Early Modern English itself were /ɔː/ and /ɔu/ respectively, but this changed with the Great Vowel Shift.)

The merger occurs in the vast majority of Modern English accents; whether the outcome is monophthongal or diphthongal depends on the accent. The traditional phonetic transcription for General American and earlier Received Pronunciation in the 20th century is /oʊ/, a diphthong. But in a few regional accents, including some in Northern England, East Anglia and South Wales, the merger has not gone through (at least not completely), so that pairs like toe and tow, moan and mown, groan and grown, sole and soul, throne and thrown are distinct.

In 19th century England, the distinction was still very widespread; the main areas with the merger were in the northern Home Counties and parts of the Midlands.

The distinction is most often preserved in East Anglian accents, especially in Norfolk. Peter Trudgill discusses this distinction, and states that "...until very recently, all Norfolk English speakers consistently and automatically maintained the nose-knows distinction... In the 1940s and 1950s, it was therefore a totally unremarkable feature of Norfolk English shared by all speakers, and therefore of no salience whatsoever."

In a recent investigation into the English of the Fens, young people in west Norfolk were found to be maintaining the distinction, with [ʊu] or [ɤʊ] in the toe set and a fronted [ɐʉ] in the tow set, with the latter but not the former showing the influence of Estuary English.

Walters (2001) reports the survival of the distinction in the Welsh English spoken in the Rhondda Valley, with [oː] in the toe words and [ou] in the tow words.

In accents that preserve the distinction, the phoneme descended from Early Modern English /ɔu/ is usually represented by the spellings ou, and ow as in soul, dough, tow, know, though etc. or through L-vocalization as in bolt, cold, folk, roll etc., while that descended from Early Modern English /ɔː/ is usually represented by oa, oe, or oCe as in boat, road, toe, doe, home, hose, go, tone etc.

This merger did not occur before r originally, and only later occurred (relatively recently) as the horse–hoarse merger. This merger is not universal, however, and thusly words with our and oar may not sound the same as words with or in some dialects.

Homophonous pairs

/oː/ /ou/ IPA Bo bow ˈboʊ bode bowed ˈboʊd borne bourn(e) ˈboə(r)n borne Bourne ˈboə(r)n coaled cold ˈkoʊld coarse course ˈkoə(r)s doe dough ˈdoʊ does doughs ˈdoʊz doze doughs ˈdoʊz floe flow ˈfloʊ foaled fold ˈfoʊld fore four ˈfoə(r) forth fourth ˈfoə(r)θ fro frow ˈfroʊ froe frow ˈfroʊ froes frows ˈfroʊz froze frows ˈfroʊz groan grown ˈɡroʊn holed hold ˈhoʊld O owe ˈoʊ oh owe ˈoʊ moan mown ˈmoʊn mode mowed ˈmoʊd Moe mow ˈmoʊ no know ˈnoʊ noes knows ˈnoʊz nose knows ˈnoʊz ode owed ˈoʊd pole poll ˈpoʊl pore pour ˈpoə(r) road rowed ˈroʊd rode rowed ˈroʊd roe row ˈroʊ roes rows ˈroʊz role roll ˈroʊl rose rows ˈroʊz shone shewn ˈʃoʊn shone shown ˈʃoʊn so sew ˈsoʊ so sow ˈsoʊ sole soul ˈsoʊl soled sold ˈsoʊld soled souled ˈsoʊld throe throw ˈθroʊ throne thrown ˈθroʊn toad towed ˈtoʊd toe tow ˈtoʊ toed towed ˈtoʊd tole toll ˈtoʊl

Cot–coat merger

The cot–coat merger is phenomenon occurring for some speakers of Zulu English where the phonemes /ɒ/ and /oʊ/ are not distinguished making "cot" and "coat" homophones. Zulu English also generally has a merger of /ɒ/ and /ɔː/, so that sets like "cot", "caught" and "coat" can be homophones.

Rod–ride merger

The rod–ride merger is a merger of /ɑ/ and /aɪ/ occurring for some speakers of African American Vernacular English (AAVE), in which rod and ride are merged as /rad/. Some other speakers of AAVE may keep the contrast, so that rod is /rɑd/ and ride is /raːd/. This merger requires the presence of the father-bother merger before it can occur.

Smoothing of /aɪ.ə/

Smoothing of /aɪ.ə/ is a process that occurs in many varieties of British English where bisyllabic /aɪ.ə/ becomes the triphthong /aɪə/ in certain words with /aɪ.ə/. As a result, "scientific" is pronounced /saɪən.ˈtɪf.ɪk/ with three syllables and "science" is pronounced /ˈsa(ɪ)əns/ with one syllable.

Pride–proud merger

The pride–proud merger is a merger of the diphthongs /aɪ/ and /aʊ/ before voiced consonants into monophthongal /ä/ occurring for some speakers of African American Vernacular English making pride and proud, dine and down, find and found etc. homophones. Some speakers with this merger, may also have the rod–ride merger hence having a three–way merger of /ɑː/, /aɪ/ and /aʊ/ before voiced consonants, making pride, prod, and proud and find, found and fond homophones.

Line–loin merger

The line–loin merger is a merger between the diphthongs /aɪ/ and /ɔɪ/ that occurs in some accents of Southern English English, Hiberno-English, Newfoundland English, and Caribbean English. Pairs like line/loin, bile/boil, imply/employ are homophones in merging accents.

Homophonous pairs

/aɪ/ /ɔɪ/ IPA Notes aisle oil ˈɑɪl bile boil ˈbɑɪl dried droid ˈdrɑɪd imply employ ɨmˈplɑɪ file foil ˈfɑɪl fire foyer ˈfɑɪə(r) foyer may also be pronounced /ˈfɑɪeɪ/ or /ˈfwɑːjeɪ/. grind groined ˈɡrɑɪnd guy goy ˈɡɑɪ heist hoist ˈhɑɪst I'll oil ˈɑɪl isle oil ˈɑɪl Jain join ˈdʒɑɪn kine coin ˈkɑɪn Kyle coil ˈkɑɪl liar lawyer ˈlɑɪə(r) lied Lloyd ˈlɑɪd line loin ˈlɑɪn Lyle loyal ˈlɑɪəl With vile-vial merger. lyre lawyer ˈlɑɪə(r) pie poi ˈpɑɪ pies poise ˈpɑɪz pint point ˈpɑɪnt psi soy ˈsɑɪ ride roid ˈrɑɪd rile roil ˈrɑɪl rile royal ˈrɑɪəl With vile-vial merger. rye Roy ˈrɑɪ sigh soy ˈsɑɪ sire sawyer ˈsɑɪə(r) sire soya ˈsɑɪə Non-rhotic accents. Thai toy ˈtɑɪ tide toyed ˈtɑɪd tie toy ˈtɑɪ tied toyed ˈtɑɪd tile toil ˈtɑɪl try Troy ˈtrɑɪ vied void ˈvɑɪd wry Roy ˈrɑɪ

Coil–curl merger

The coil–curl merger is a vowel merger that historically occurred in some dialects of English. It is particularly associated with the early twentieth-century (but now extinct or moribund) dialects of New York City, New York; New Orleans, Louisiana; and Charleston, South Carolina. In fact, in speakers born before World War I, this merger apparently predominated throughout older Southern U.S. speech, ranging from "South Carolina to Texas and north to eastern Arkansas and the southern edge of Kentucky."

The merger caused the vowel classes associated with the General American phonemes /ɔɪ/, as in choice, and /ɝ/, as in nurse, to merge, making words like coil and curl, as well as voice and verse, homophones. The merged vowel was typically a diphthong [əɪ], with a mid central starting point (though sometimes [ɜɪ]), rather than the back rounded starting point of /ɔɪ/ of choice in most other accents of English. The merger happened only before a consonant; stir and boy never rhymed.

The merger is responsible for the "Brooklynese" stereotypes of bird sounding like boid and thirty-third sounding like toity-toid. The songwriter Sam M. Lewis, a native New Yorker, rhymed returning with joining in the lyrics of the English-language version of Gloomy Sunday.

In 1966, according to a survey that was done by William Labov in New York City, 100% of the people over 60 used [əɪ] for bird. With each younger age group, however, the percentage got progressively lower: 59% of 50- to 59-year-olds, 33% of 40- to 49-year-olds, 24% of 20- to 39-year-olds, and finally, only 4% of people 8–19 years old used [əɪ]. Nearly all native New Yorkers born since 1950, even those whose speech is otherwise non-rhotic, now pronounce bird as [ˈbɝd].

Mare–mayor merger

The mare–mayor merger is a process occurring in many varieties of British English, as well as the Philadelphia dialect and Baltimorese, where bisyllabic /eɪ.ə/ is pronounced as the central diphthong /eə/ in many words. In these varieties, mayor is pronounced /ˈmeə(r)/, homophonous with mare.

In North American English accents with the merger, it also affects sequences without /r/, where some words with the /eɪ.ə/ sequence merge with /eə/ associated with æ-tensing. Because this particular /eə/ derived from /æ/, such words are frequently hypercorrected with /æ/. The best known examples of this are mayonnaise (/ˈmeəneɪz~ˈmæneɪz/) and graham (/ˈɡreəm~ˈɡræm/, a homophone of gram).

Homophonous pairs
/eə/ /eɪə/ IPA Notes Bayer bare ˈbeə(r) North American English pronunciation of Bayer. flare flayer ˈfleə(r) flair flayer ˈfleə(r) gram, gramme Graham ˈɡreəm With æ-tensing. lair layer ˈleə(r) mare mayor ˈmeə(r) pair payer ˈpeə(r) pare payer ˈpeə(r) pear payer ˈpeə(r) prayer prayer ˈpreə(r) stare stayer ˈsteə(r) sware swayer ˈsweə(r) swear swayer ˈsweə(r) there they're ˈðeə(r)

See also

• Phonological history of the English language
• Phonological history of English vowels
• Trisyllabic laxing
• Great Vowel Shift