The demise of ɪə as in NEAR
NEAR in RP
In the earlier standard/reference accent of British English, Received Pronunciation, words like NEAR contained a centring diphthong, ɪə. This was a vowel which glided from the lax quality ɪ to the quality ə within a single syllable. It can be heard in this clip from a 1930s Pathé documentary about beef and beer:
Such a vowel is heard relatively rarely today. Although British dictionaries still use “/ɪə/” in their transcriptions, a lax diphthong of this type is now rather old-fashioned.
In contemporary Standard Southern British (SSB), we hear tend to hear either
(1) a long pure vowel, the monophthong ɪː; or
(2) a form in which the tense FLEECE vowel is followed by schwa, which we could write as ɪjə or, with traditional symbols, as /iːə/; this form can plausibly be considered to comprise two syllables.
Many speakers use both forms. For such speakers, NEAR is what John Wells has termed varisyllabic, and the long monophthong in (1) can be thought of as derived from the disyllable in (2) by ‘smoothing’. I discuss smoothing rules in this post.
Varisyllabic NEAR in speech
First let’s compare beer pronounced as bɪə from the 1930s documentary with beer pronounced as bɪjə from the contemporary online Cambridge dictionary:
And here in a BBC documentary is UCL’s Professor Mark Miodownik saying here a carbon fibre composite is being made, with here pronounced as hɪjə:
Now the monophthongal type (1) pronunciation. Here is BBC presenter Andrew Plant saying all these Christmas cards here, with the monophthongal hɪː type of here:
Such varisyllabicity is also found in words of the WIRE, CURE and SOUR sets. Here, from the Cambridge and Oxford Advanced online dictionaries, are retire/retirement, secure/security, near/nearly. The ire, ur and ear portions sound disyllabic in the first word of each pair, monosyllabic in the second:
I said there that the relevant portions of those words sound disyllabic or monosyllabic. But it’s often unclear how many syllables there are in a given stretch of speech. One very useful kind of evidence is the way that native speakers fit words into the rhythmic patterns of verse and song. Verse forms often have a fixed number of syllables per line; and popular song, by default, sets one syllable to one note.
Varisyllabic NEAR in song
Let’s look at the varisyllabic treatment of WIRE in musical settings. Songwriter-performer Noel Coward, in his comic song Mad Dogs and Englishmen, set retire and perspire with -ire treated as a single syllable:
But other songwriters have broken –ire across two notes, with the schwa firmly belonging to the second. An example is the theme song to the James Bond film A View to a Kill (by John Barry and Duran Duran), where the line dance into the fire breaks fire across two notes — fi-re, as it would be written in a vocal score:
We can also find this kind of varisyllabicity within a single musical line. In his song Cat People (Putting Out Fire), David Bowie pronounces fire both as a single-note monosyllable and also as a disyllable broken across two notes, in the single line And I’ve been put-ting out the fire with ga-so-line, put-ting out fi-re:
The crucial point for this article is that NEAR words can exhibit exactly this kind of musical varisyllabicity that we’ve just seen in WIRE words, sometimes set to one note and sometimes set to two. (This never happens to the vowels traditionally written “/eə/” and “/ʊə/”.) Here is Paul McCartney singing the first line of his song Martha My Dear in a deliberate pastiche of old-fashioned RP, with dear pronounced dɪə and set to a single note:
But in Take That’s Holding Back the Tears, writer-performer Gary Barlow sings the word te-ars broken across two notes, with the schwa quality clearly belonging to the second:
(The final syllable of the line, -more, is sung as a cadential melisma, with its monophthong spread over three notes; Verdi often does a similar thing as an ornamentation.)
Gary Barlow is from the northwest of England, and speaks with a northern accent. Though his singing here isn’t northern-accented, his varisyllabic tears might be thought of as a regionalism. So let’s turn to the traditional territory of Received Pronunciation, and listen to some southerners educated at public schools (= exclusive private schools).
First, Thom Yorke of Radiohead, whose family moved settled in Oxfordshire when he was ten; he later attended the public school Abingdon. Yorke speaks with BATH-broadening, and has a somewhat conservative j in purs[j]uing. His song The Eraser repeatedly includes the phrase the more that I/you appear, and on every occasion he breaks -pe-ar across two notes, with the schwa quality clearly belonging to the second:
Coldplay’s Chris Martin (pictured above) attended the public school Sherborne. (He also has a First Class degree from UCL.) Last year’s Coldplay hit Every Teardop Is A Waterfall neatly demonstrates the varisyllabicity of NEAR within a single line, with the word tear occuring first as two syllables tɪjə and then as one tɪ(ː). The line is Ev’-ry te-ar, ev’-ry te-ar, ev’-ry tear-drop is a wa-ter-fall:
The evidence from speech and from word-setting in songs show us that NEAR words like tear and appear can show the same varisyllabicity as WIRE words like fire and retire. If NEAR is varisyllabic, it should be analyzed in the same way as WIRE. Just as WIRE can be analyzed as ɑjə (or, in conservative transcription, /aɪ.ə/), smoothable to ɑː, so NEAR should be analyzed as ɪjə (in conservative transcription, /iː.ə/), smoothable to ɪː.
The old-fashioned transcriptions in the major dictionaries, of course, don’t reflect this. They write a word like beer as “/bɪə/”. This transcription is phonetically more appropriate to the pronunciation we heard in the 1930s Pathé film than to contemporary speech. It also states unambiguously that a word like beer is a single syllable; this representation has no way of explaining the contemporary varisyllabicity.
The old-fashioned dictionary transcriptions misleadingly suggest that NEAR “/ɪə/” has the same structure as SQUARE “/eə/”. Allegedly they’re both centring diphthongs, starting with lax vowels and gliding towards schwa within the same syllable. This was true, of course, in old-fashioned RP. But if it were still true, then NEAR and SQUARE should pattern in a similar way. If, as we’ve seen, NEAR words can be broken across two musical notes, with ə assigned to the second, then the same should be true of SQUARE words. That is, analogous to Coldplay’s
Ev’-ry te-ar, ev’-ry te-ar, ev’-ry tear-drop…
we should expect to find the likes of
Ev’-ry sta-ir, ev’-ry sta-ir, ev’-ry stair-way…
with [e] on the first note of sta-ir and [ə] on the second.
But such forms are totally non-existent. SQUARE never exhibits varisyllabic behaviour. On the contrary, in contemporary SSB, SQUARE has simply transformed into a long monophthong, ɛː, so that it belongs to the same category as NURSE əː, START ɑː and NORTH oː.
Interestingly, there’s an increasing number of speakers who don’t have varisyllabic NEAR. For these speakers, NEAR has simply become a long monophthong ɪː, like SQUARE ɛː, NURSE əː, START ɑː and NORTH oː. These speakers never say nɪjə, bɪjə and hɪjə, but only the smooth forms nɪː, bɪː and hɪː.
So, in SSB today, we hear NEAR commonly as ɪjə and as ɪː. Some speakers use both; for them, NEAR may be considered varisyllabic, like WIRE and SOUR. Others seem to use only ɪː. Either way, the transcription of NEAR as a monosyllabic lax diphthong “/ɪə/” is now out of date.
When I first wrote this post in 2012, I underestimated the prevalence of younger speakers for whom NEAR is not really varisyllabic: in other words, who use ɪː a lot but ɪjə not so much. I implied that NEAR is simply varisyllabic in SSB. If this were true, then we could in fact dispense with a distinct NEAR phoneme: just as WIRE is composed of PRICE + schwa, we could say that NEAR is FLEECE + schwa, reducing our vowel inventory by one. In conservative transcription, we could replace all instances of /ɪə/ in dictionaries with /iːə/. This now seems untenable. For many younger speakers, it would be more accurate to replace all instances of /ɪə/ with /ɪː/ – so that the phoneme survives, but in monophthongal form, like modern SQUARE ɛː. This suggests that varisyllabic NEAR will disappear from songwriting as younger composers feel it to be purely monosyllabic.
Contemporary Standard British, as I’ve analyzed it in another post, allows a simpler statement than R.P. of the input to smoothing-compression. By recognizing the diphthongal nature of contemporary FLEECE and GOOSE, we can simply say that the process applies to diphthongs.
Does smoothing apply to FACE, as in the surname “Mayer”? Or to GOAT, as in “boa”?
Yes. As I say in the post, contemporary smoothing applies to any diphthong plus schwa. Check out my smoothing post, to which I’ve added a quite a lot more audio data since I first wrote it. Examples of FACE and GOAT are included. (And remember that smoothing is optional, but more likely when further syllables occur within the word.)
I only have one problem with your post: I agree with your analysis of the NEAR words, but when it comes to words such as fire you say:
Older smoothing: fɑə
Newer smoothing: fɑɑ
Conservative speakers stop at stage 1, while less conservative speakers […] go to stage 2.)”
It is my impression that what you call “newer smoothing” (tower or tire = tar) is much more frequent in speakers born, say, before WW2. So I would rather say that more conservative speakers go all the way, to stage 2, while less conservative speakers remain at stage 1.
I think your excellent ear picked up that John Whittingdale’s entire rhyming with tar occurred within a posh, conservative accent (he was born in 1959). Yes, monophthongal pronunciations of that type have been around for a long time; but if you check out something like the Pathe newsreel site, a nostalgic treasure trove of old R.P., you’ll hear that [aə] was very common, [ɑɑ] not common at all. The new monophthonging is quite widespread. I’ve replaced the Whittingdale example with an example from the less conservatively accented Prof. Jim Al-Khalili; and if you check out my smoothing post there are pronunciations of variety by Simon Keenlyside (also born 1959 but far less conservative), liabilities by a speaker who I think was in his 30s, and aisle by Prince Harry (born 1984).
Excellent post. I like all your musical. It spices things up.
When playing games like charades, I have occasionally disagreed with my team-mates on whether words like “near” are one syllable or two. For me, “near” is normally [ni:ə] but “nearest” is [niərəst].
I think that this is another case of a pronunciation that’s been underreported in the past (as I said with SQUARE last time). In the Survey of English Dialects for the north, most sites are given ɪə for their NEAR vowel. There must have been more i:ə than was reported.
Charades is a great example. For structures like the syllable, language games are a widely used source of evidence.
A very insightful and helpful post as usual. Thank you.
Like many others, I was taught that the centering diphthongs were monophonemic and (therefore) monosyllabic. Nonetheless, I couldn’t help perceiving some (many?) instances as disyllabic. Later, when I learnt a little more, I realised they (especially /ɪə/ and /ʊə/) didn’t even stand the test of the so called “sonority scale”, as [ə] is a higher peak than either [ɪ] or [ʊ] and should, strictly speaking, be treated as a separate syllable. But there were other “problems” with the sonority scale as well, like speaks appearing as three syllables. I eventually just learnt to visualise them as single-syllable one-phoneme diphthongs.
As far as I remember (I can’t check right now), in some of the earlier editions of the EPD, at least in some words Jones had a secondary variant for the NEAR vowel, with a rising diphthong: jəː, in his notation. (Or it could have been mentioned somewhere in the introduction to Gimson’s editions). I’ve definitely heard /əˈpjɜː/ (əˈpjəə) in singing, but I don’t think I’ve spotted it speech. But while /jɜː/ (jəː, jəə) would be very clearly monosyllabic, /ɪə/ sometimes intuitively strikes me as disyllabic even in traditional RP, including your clip from the 1930s documentary about beef and beer (especially the second instance or beer).
Why do you believe /ɪə/ is demonstrably monosyllabic in that particular case?
I’ve just seen that in his latest posting JW Lewis mentions /njɜːli/ as an “earlier form” of nearly, so I suppose [jɜː] for /ɪə/ was current once.
So happy to find this post – when I asked my phonetics professor I got a very woolly answer that seemed to come down to ‘yes it’s one syllable’…but this has always bothered me a lot in some words. The clips are really useful in showing that contrast. So thank you 🙂
Many thanks for reading and for your kind words!
I can certainly hear the sound change you describe. My question is: do pairs such as “bid”/”beard” become homophones, considering the /ɪ/ in “bid” is usually quite long? Thank you!
Good question. “Bid” and “beard” don’t become actual homophones — they can certainly be differentiated — but in practice they can undeniably be pretty similar. The same applies to “shed” and “shared”. (Btw I wouldn’t say that “bid” is quite long in SSB, though of course it’s not as short as “bit”.)
Yes, SSB undeniably has phonemic vowel length, like Australian English. But it applies mostly to environments where the historical /iːr/ and /eːr/ could occur, so before single consonants and certain clusters. Perhaps that could be the topic of your next post. Maybe the /ʌ/-/ɑː/ should also be treated as a short-long pair. I find that STRUT is nowadays pretty consistently realized as back [ʌ] (or [ɑ], if you want to write it that way), which is far away from cockney [a].