April 7, 2010
by Graham

Spanish rhythm

Attending BAAP last week, I was very pleased to find there was a whole session devoted to rhythm, which I have written about before (here). One of the general conclusions was that the perception of rhythm in a language depends on the native language of the perceiver. English has a relatively complex syllable structure, while French and Spanish have far fewer consonant clusters, so that their vowel onsets are closer together. The syllable durations of French and Spanish, measured in milliseconds, are therefore shorter and more similar that those of English syllables.

Poetry, as the most rhythmical form of lanuage, may be able to help. English verse lines are measured in stresses, and one of the papers at BAAP measured the rhythm of the most regular form: the limerick. This is either a verse of five lines, with three stresses in each of the first two, and the last, and two stressed in the third and fourth lines; or it is a verse of four lines, each of four stresses, the final one of the first, second and final lines being silent, and an internal rhyme in the third line. The rhythm of a limerick can easily be tapped out as it is being read, and the silent stress is obvious.

French verse is measured in syllables: an Alexandrine has 12 syllables, with a caesura after the sixth (in a classical poem). Any line that does not have 12 syllables is not an Alexandrine. Each line can therefore be tapped out with twelve beats.

Spanish verse seems to be a hybrid: a Spanish Alexandrine has 14 syllables. There is no necessary caesura, but a condition of its being a true Alexandrine is that the thirteenth syllable be stressed. Here are eight lines from a poem by Antonio Machado to illustrate this:

Adoro la hermosura, y en la moderna estética
corté las viejas rosas del huerto de Ronsard;
mas no amo los afeites de la actual cosmética,
ni soy un ave de esas del nuevo gay-trinar.

Desdeño las romanzas de los tenores huecos
y el coro de los grillos que cantan a la luna.
A distinguir me paro las voces de los ecos,
y escucho solamente, entre las voces, una.

In each case (allowing for synalepha), the thirteenth syllable is stressed. The only places where there is no synalepha are in the third line between ‘no’ and ‘amo’, and in the last line, between ‘solamente’ and ‘entre’. This makes a total of between 13 and 15 syllables for the individual lines.

If Spanish is syllable-timed, how do we achieve equal lengths for these lines?

March 22, 2010
by Graham

Homing in

At about 8.20 this morning, on the Today programme, Nick Robinson (the BBC’s Political Editor) said that the opposition parties would want to ‘hone in’ on the revelations that Stephen Byers (former Minister for Transport) had been caught claiming to be able to influence government policy. Here he is.

The OED gives examples of hone in, where the sense is clearly ‘home in’, going back to 1965, but while the explanation for the slippage given there (confusion with hone in the sense of practise or refine a technique) is possible, it still seems to me to be a very strange development, and certainly one I would recommend that any foreign learner of the language should avoid as to many people it would still be thought of as a mistake.

March 15, 2010
by Graham


Listening to “Start the Week” on BBC Radio 4 this morning (15 March) I was astonished to hear one of the participants use this word and  pronounce it /kɒgˈnaɪtɪv/.  This is not given by any of the current dictionaries of pronunciation, nor by the OED, and I’ve been wondering what model he was taking in order to arrive at it.

There’s decisive, with its noun decision, as cognitive has cognition, but decisive has the verb decide, for which cognitive has no equivalent (cognize?).

Attrition does give attritive /əˈtraɪtɪv/ but OED notes this as a rare word. Excitive /ɪkˈsaɪtɪv/ – latest quotation 1862. Incitive /ɪnˈsaɪtɪv/ – 1888.

And out of 200 words ending -itive, these are the only ones with pronunciation /-aɪtɪv/.

March 1, 2010
by Graham

Bleck Het

John Humphrys has been sounding off about the English language – again. Why is it that reporters – journalists – believe that because language is their stock in trade, therefore they know all about it? This particular article, published in the Mail on Sunday (28 February) is headed ‘Thet men in the bleck het (… or how we finally learned to stop hiding behind ludicrous accents)’.

David Crystal tells us in his autobiography that he challenged John Humphrys to justify his (JH’s) claims about what David Crystal believed. JH had to admit that he couldn’t, not having read all DC’s output. They are apparently now the best of friends. So why does Humphrys lay himself open to criticism from linguists yet again? In this article we find “The rules of RP as observed by the BBC stated that ‘class’ had to rhyme with ‘arse’ (not that such a rude word would have been permitted) and ‘that’ with ‘wet’ and ‘house’ with ‘rice’ and ‘off’ with ‘law’. And so on. In short, it meant thinking posh.”

Where does he get these ‘rules’ from? No BBC document that I have ever seen stipulated that any particular accent was enforced on broadcasters, and if they had been, neither my two predecessors as head of the Pronunciation Unit nor myself would have been employed in that position – they were both Scots, with Scottish accents, and my accent is irredeemably North Midland.

Even BBC radio announcers have long been taken from a wide range of backgrounds, with recognisably non-RP accents. Humphrys brings out the inevitable Wilfred Pickles, who was a Yorkshire-born actor, but in the 1960s, when John Snagge and Alvar Lidell, together with Frank Phillips were still reading the news, so were Roy Williams (New Zealander) and Dwight Wylie (Jamaican). They may have been in a small minority, but I don’t think they would have accepted that they were ‘token’ colonials – if they had not been up to the BBC’s exacting standards, they would not have been employed.

In any case, Humphrys is alleging that people have stopped speaking with ‘ludicrous accents’, and implies that this is a conscious decision by the individuals concerned. He does not allow that RP, like all accents, has evolved during the last hundred years, and that the evolution is not only from one generation to the next, but may also be by imperceptible degrees in the speech of a single person, even though that person is not aware of the changes. He mentions the study of the Queen’s accent which Jonathan Harrington carried out some years ago, which “concluded that her vowel sounds had undergone a subtle evolution” (Humphrys’ words). This was not to be unexpected, but Humphrys says “the posh have given some ground” as if the Queen was deliberately changing her accent, rather than reacting, as we all do, to the speech of those around us

It takes a remarkable individual who can resist all linguistic influence through a long life.

February 16, 2010
by Graham


The BBC New website tells us that this mean “together” in Dari, but there seems to be a disagreement between Radio and Television about its pronunciation. Radio 4 newsreaders are all stressing the first syllable, while their television colleagues are stressing the second.

I don’t know if Dari has a strong preference for stress placement, but whether it does or not, shouldn’t they decide on one or the other, and stick to it? Frank Gardner, the BBC’s security analyst, is in Afghanistan at the moment, and reporting for both sets of output. Is he having to either avoid the word altogether, or else adjust his pronunciation according to the programme he’s taking part in?

This isn’t a question of one or other being “correct” (although if Dari is like Iranian Farsi in its stress patterns, the final syllable is likely to be the most prominent), but consistency. Stress placement is one of the questions most frequently raised with me by non-linguists, both in English words and foreign names.

February 6, 2010
by Graham
1 Comment

Wymondham (Norfolk) and Wymondley (Hertfordshire)

Here are two place names whose pronunciation history converged for a time and then diverged again.

Wymondham is, in my experience, always pronounced /ˈwɪndəm/. I don’t have access to either of the Engish Place Name Society (EPNS)’s volumes on Norfolk, but the Oxford Names Companion gives its origin as ‘homestead of a man called Wīgmund’. The pronunciation was presumably something like /ˈwiːgmʊndhæm/. Over time, the /g/ and /h/ will have disappeared, the /ʊ/ and /æ/, being unstressed vowels, will have reduced to schwa, leaving /ˈwiːməndəm/. The first schwa then also goes the way of all flesh, and the /m/ assimilates to the following /n/: /ˈwiːndəm/. At some stage /iː/ is shortened, and hey presto, we have /ˈwɪndəm/. This must all have happened before the Great Vowel Shift started to apply.

Wymondley, according to both the EPNS on Hertfordshire and the Oxford Names Companion, has a slightly different man’s name as its root: Wilmund. The EPNS volume was published in the 1930s, and gives the pronunciation, as we might expect, /ˈwɪmli/, following the same sort of path as Wymondham. However, the only pronunciation I ever hear now (and I live about three miles from Great Wymondley – Little Wymondley is a couple of miles further away), is /ˈwaɪməndli/. I’ve just consulted a scion of a long-established Hertfordshire family, and she tells me her father used to say /ˈwɪmli/, but only in a jokey sort of way. So, the spelling pronunciation was around for most of the 20th century, and can’t be attributed simply to incomers to Letchworth Garden City (founded in 1903) and Stevenage New Town (1947). This distorts the etymology: there was never a long vowel in the first syllable, so /ˈwiːməndli/ was never the pronunciation, so far as we can tell, and /ˈwaɪməndli/’s only justification is as a spelling pronunciation.

February 3, 2010
by Graham


I thought Uttoxeter deserved a post of its own, because it also raises a transcription and dialect question.

Both /ÊŠt/ and /ÊŒt/ have been quoted, by Michael Lamb and John Maidment, as possible pronunciations of the initial sounds, and I agree. But in northern dialects the /ÊŠ/ ~ /ÊŒ/ split never happened, so the vowel used by locals unaffected by “education” (as John and I have been!) is still the equivalent of /ÊŠ/. On the phonological level, this is fine.

However, to transcribe it as /ÊŠ/ for all northern accents/dialects, as is most often done, leads to parodists of these accents/dialects using the southern English /ÊŠ/ indiscriminately in both STRUT and FOOT. If I try to recall my own “pre-educated” pronunciation, the nearest southern English vowel is actually a short /ɔː/, and this is certainly what I hear from my unreconstructed friends and relatives who have not had my “advantages”. This means that there are minimal pairs between but and bought, pun and pawn, full and fall, where the distinction is one of vowel length only. (Incidentally, when I’m tired, I may get /ÊŠ/ and /ÊŒ/ the “wrong” way round, and am quite likely to say /pÊŒt ÊŠp/ instead of /pÊŠt ÊŒp/. Birth will out!)

This is the situation for North Staffordshire as I hear it. I can’t find much in the literature about the phonetics of the speech of the Potteries. Certainly TV and radio adaptations of Arnold Bennett never have actors with convincing accents. There are several examples of this undifferentiated vowel from Garth Crooks (former footballer, now broadcaster) here.

February 1, 2010
by Graham

Abergavenny, etc

John Wells was mentioning (here) the unpredictability of the pronunciation of British place and family names from their spellings, and some are recorded in the Dictionary of Blunders. The fact that they are mentioned at all must mean that in the author’s opinion they were being mispronounced, and this may be giving us an indication that in some cases the pronunciation was actually changing at the time he was writing (the 1870s or early 1880s).

ABERGAVENNY (family name) is pronounced Abergen’-ny. (This is also still, apparently, the pronunciation of the Marquis of Abergavenny, although it is not his family name.)

BELFAST is pronounced Bĕl-făst’, not Bĕl’-fast. (The author does not specify the exact pronunciation of ‘a’ the second time. Nowadays either stress pattern seems to be acceptable, and either /æ/ or /ɑː/ for the ‘a’.)

BERKELEY STREET is pronounced Bark-ley Street, and not as spelled.

BERKSHIRE is pronounced Bark-shire, and not as spelled. (There are some British dialects in which ‘er’ is still pronounced /ɜː/.)

CARSHALTON is pronounced Casehorton. (The BBC Pronouncing Dictionary of British Names acknowledges that this had existed, but by 1971 was no longer heard. The pre-war BBC publication Broadcast English II, which covered English place names, did not include it at all.)

CHOLMONDELEY (family name) is pronounced Chum’ley.

CINQUE (the Cinque Ports) is pronounced like sank. (Not today it isn’t. The BBC recommendation is ‘sink’.)

CIRENCESTER is pronounced Cissester. (Strangely, the BBC’s original recommendation, in 1930, was /ˈsɪsɪtə(r)/. Nowadays, the spelling pronunciation has prevailed: /ˈsaɪrənsestə(r)/, and I believe it is often shortened to /ˈsaɪrən/.)

COCKBURN (family name) should be pronounced Coburn, and not as spelled. (This applies to the port, and some years ago, an advert appeared in the London Underground:

Said King Charles to his court

“I enjoy a good port.”

Said a courtier game

“If I tell you the name

of the best will you make me a knight?”

The king nodded his head

and the courtier said

“Cockburn’s Port is the port for a king.

But remember to say it without the C K.”

So the court cried “Long live Harles the Ing!”)

COLQUHOUN (the name of a person) is pronounced Cǒ-hoo’n. (It still is.)

COWPER. The poet called himself Cooper, and not Cow-per.

CRICHTON is pronounced krī’ton, not krĭk’ton.

HELENA is pronounced Hĕl’-ĕ-na, not Hē-lē’na.

JACQUES is zhāk in French and jakes in English. (This is ambiguous, because the writer uses ā sometimes for /ɑː/ and sometimes for /eɪ/.)

MACLEOD is pronounced mak-loud, not măk-le’-ŏd.

MAINWARING (a family name) is pronounced Mannering.

MARJORIBANKS (a family name) is pronounced Marchbanks.

NAOMI is pronouncec Na’-o-mi, not Na-ō’-mi.

NASMYTH is pronounced Na’smith, not Naz’-mith.

PHŒBE (a female Christian name) is pronounced Fē’-bē.

ST. JOHN (a family name) is pronounced Sin’-jun.

ST. MAUR (Earl) is pronounced Sĕ-maur.

January 30, 2010
by Graham

hospital ~ orbital ~ digital

Jack Windsor Lewis has mentioned (here) the frequently different pronunciation in BE of the /t/ in hospital and orbital. I wonder if part of the reason for the /t/ of hospital becoming voiced in British English, while that in orbital does not (I know the grammar of that is a bit odd, but I’m sure you know what I mean), is a rhythmic difference. There is also a grammatical difference, of course, but a third rhythmic pattern occurs in digital, and digital and orbital have the same grammatical structure:

Hospital, Orbital, Digital

Hospital, Orbital, Digital

January 22, 2010
by Graham

More on 1880s pronunciation

Most of the pronunciations given in the Dictionary of Blunders are what one would expect for 1880, and show that the arguments about what is ‘right’ and what is ‘wrong’ are very much the same as today:

ab-do’-men, not ab’-dŏm-ěn (OED1 agrees with this, but OED2 gives initial stress before second syllable stress, although many medical people use ab-do’-men, which corresponds to the Latin quantities)

a-kū’-men, not ak’-ŭ-men (OED2 gives second syllable stress before initial stress, which corresponds to the Latin quantities)

al-bu’-men not al’-bum-en (OED2 does not recognise initial stress, although I don’t think I’ve ever heard anyone British say /ælˈbjuːmən/)

bĭtu’men not bĭt’-u-men (OED2 gives second syllable stress before initial stress, which corresponds to the Latin quantities, but current British pronunciation dictionaries all indicate second syllable stress as American)

CHIVALRIC is pronounced shiv’-ǎl-rik not chiv-ǎl’-rik (OED1 and OED2 have second syllable stress first, and also allow initial /tʃ/)

CONTRIBUTE is pronounced kon-trĭb’-ute, not kon’-trĭb-ute. (Interesting that the initial stress was being heard even then)

DECADE is pronounced dĕk’-āde, not dēk’-ade (OED1 gives only /ˈdekəd/ – as the BBC Advisory Committee was also recommending nearly 50 years later)

LAMENTABLE. The accent in this word is on the first syllable, lăm’-ĕnt-able. (Clearly, he was hearing second syllable stress )

LIBRARY. This word is often mispronounced li’-bar-ay, instead of li’-bra-ry. (No change today)

MISCHIEVOUS is pronounced mĭs’-chĕ-vŭs, not mis-chē’-vŭs, nor mis-che’-vě-us. (No change today)

Two which I can find no other ‘authority’ for are:

CAESURA is pronounced sez’-ū-rǎ (not even OED1 allows /e/ in the first syllable, nor stress on that syllable)

GLAMOUR is pronounced glā’-moor not glă’-mur. (I wonder where he got this from?)