UCL Division of Psychology & Language Sciences

john Wells
Photo: Masaki Taniguchi

John Wells’s phonetic blog archive 1-15 June 2008

To see the IPA phonetic symbols in the text, please ensure that you have installed a Unicode font that includes them all, for example Lucida Sans Unicode or Charis SIL (click name for free download).

Browsers: I recommend Firefox (free) or, if you prefer, Opera (also free).

RSS feed for this site RSS feed


my home page


Friday 13 June 2008

Phonics plus

Another interesting paper at last weekend’s spelling conference was given by Chris Jolly, whose Jolly Phonics reading scheme, developed by Sue Lloyd and Sara Wernham, is very widely used in UK primary schools.

He has developed something he calls Jolly Phonics Extra, which is currently being trialled. The first part of this programme follows the existing scheme, based on the principles of synthetic phonics,

teaching the letter sounds, and how to ‘blend’ them together to read a new word, so c-a-t makes ‘cat’.

The second part

introduces 11 new letter shapes which need to be learnt for reading, but not for writing. From a previous trial we know that the children quickly learn these new letters and that they see them for what they are: a temporary guide to help with reading. Each new letter is similar to an existing letter and represents a sound.

This might seem reminiscent of the ill-fated i.t.a. tried out in the 1960s, but is actually very different. The revised spellings do not replace the traditional ones, as in the total immersion approach of i.t.a., but instead appear as an auxiliary text alongside the traditional one.

Once the new letters are learnt, the text in the reading books appears twice, once with normal spelling and once, against a tinted background, with revised spelling.

The revised spelling is designed to be as close to the look of existing spelling as possible, but to be phonically regular.

In the revised text, there are two further changes. Silent letters are shown in feint, and where unavoidable the spelling has been made regular, e.g. sed for said.

Children read from the normal text, but can use the revised spellings whenever they get stuck.

It’ll be interesting to see how this works out.

Chris Jolly

P.S. Nigel Greenwood points out that the device of adding a different spelling to show the pronunciation is similar to the Japanese practice of adding tiny (‘ruby’) furigana alongside or above kanji to show the pronunciation.


Thursday 12 June 2008

RP back in fashion?

Sam Woolaston is one of the Guardian’s television critics. On Tuesday he reviewed a programme starring Rupert Everett, The Victorian Sex Explorer.

The poshos are taking over — have you noticed? They run our cities, they're going to run our country.

This is a reference to Boris Johnson, the newly elected Mayor of London (blog, 9 June) and to David Cameron, leader of the Conservative party, who will become Prime Minister if the Tories win the next general election. Both are Old Etonians and RP speakers.

And they're taking over in television, too. A few years ago, you couldn't get anywhere in TV without an incomprehensible regional accent; toffs were simply figures of fun. Now look: Raef and Lucinda in The Apprentice. OK, they were fired, but that was despite their backgrounds and accents, not because of them. The public school revolution is happening, I'm telling you: it won't be long before the riff-raff is sent back to wherever the hell it came from.

Rupert Everett is not an Old Etonian, but he attended a similarly prestigious public ( = fee-paying) school, Ampleforth. He too is an RP speaker.

So perhaps RP is not in irrecoverable decline after all.

Rupert Everett

Angela Forkin writes to comment on the costs of teaching spelling (blog, 10 June):

Whilst I cannot categorically state that the spelling system in English is not a factor, what I can say is that the fundamental difference between the English educational system and that of Finland is definitely a factor. Finland, in common with most, if not all, Scandanavian countries, begins formal education at the age of 7 years, i.e.when children are developmentally ready to learn to read, write and spell (I'm thinking especially in terms of the fine motor control needed to control a pencil successfully), whilst here in England the Government persists in the notion that 'earlier is better' and insists on starting formal education at age 4/5 years.

Angela Forkin


Wednesday 11 June 2008

Spot the mistake

I’ve just been doing some exam marking. It was a phonetics written exam in which candidates, among other tasks, had to convert a short English passage in orthography into phonetic transcription.

Some candidates were native speakers, some were not. The examiners are extremely liberal in allowing all sorts of, shall we say, less usual pronunciations. NSs, in particular, are encouraged to transcribe their own accent (but must say what it is). NNSs normally attempt RP, but there too the examiners are very flexible.

Here, however, are some transcriptions offered by candidates but penalized by the examiners. These must be wrong, and candidates ought to know that they are wrong. Readers might like to check that they can see where and why they’re wrong.

Each is a single word, extracted from the passage in which it was embedded.

James dʒeɪms
laws lɔːwz
wrong wrɒŋ
talking ˈtaʊkɪŋ
magazine mægˈzɪŋ, mægˈzɪn, məgəˈziːn
managed ˈmænɪʒd, ˈmænɪdʒɪd
dangerous ˈdængərəs, ˈdeɪŋgrəz, ˈdeɪŋdʒərəs
think ðɪŋk
news njuːs
ignoring ˈɪgnɔːrɪŋ
conservative cənˈsəvətɪv

Are you ready to be a phonetics examiner?


(This is self-referential)


Tuesday 10 June 2008

What does it all cost?

The spelling conference last weekend (blog, 6 June) attracted a fair bit of attention in the media, including an article in the Observer newspaper.

One problem with radio, television and newspaper coverage of conferences is that the reporters are typically alerted in advance about what might be of interest. They then put together a story about what is expected to be said at the conference, what will be reported, what the reaction of the public might be. They cannot wait until the conference happens and then report what was actually said, because by then the story may have gone cold and they will have been beaten to the draw by their rivals.

So it was in the middle of last week that Anushka Astana talked to me and other people about the conference. Her story (excellent in its way) was already written by the time the ‘new research’ was reported to the conference.

You wouldn’t guess it from the Observer article, but the stated topic of the conference was actually The cost of English spelling. What is the cost to the British economy of having a difficult-to-learn spelling system?

The only presentation that really attempted to address this issue seriously was a paper by Zuzana Kotercová of the University of Coventry, ‘The economic cost of current English spelling’. She started from the claim (how well based I do not know) that English schoolchildren take three years to reach the same level of literacy skills as Finnish schoolchildren achieve in three months.

In Finnish, learning to read and write requires no more than learning the letters of the alphabet and the sounds they correspond to: then you just sound out every word as it is written, and write down every word in the way it sounds. In English, this approach gives you a good start (and is now known by the buzzword “synthetic phonics”). Children can also take in their stride conventions such as the effect of silent e (hop - hope, rat - rate). But all too soon the learner is then faced by hundeds of words whose pronunciation is not what you get by sounding out the spelling (money, two, find) or in which a rule that they’ve just learned (such as the silent e) doesn’t apply (have, give) — and words whose spelling you would not get right if you based it just on how you say them aloud (friend, head, knife, climb).

One of the other conference presenters claimed that approximately 25% of learners cope well with this extra learning burden, 50% manage but with a lot of extra time and effort, while 25% do not cope and end up disheartened and ultimately functionally illiterate.

To return to Kotercová’s work: she started by estimating the cost of the extra teaching time required to achieve literacy in English. She investigated four Coventry primary schools and calculated the number of hours of additional teaching expended on spelling instruction as opposed to general literacy. The average annual cost per teacher came out at £556. Multiplying by the number of primary teachers in England, we arrive at a lowest estimate of just over a hundred million pounds per annum: the additional expense we incur in England through the inadequacies of our spelling system.

This figure does not include the additional cost of remedial literacy teaching in secondary schools, which must be substantial. Nor does it take into account the lower lifetime earnings expectations of the functionally illiterate when compared with the literate. Nor does it address the similar additional costs incurred in other English-speaking countries, still less the additional hours of teaching and learning required in the case of those millions who attempt to learn English as a foreign language.

It would be particularly difficult to quantify the adverse effects on the intelligibility of EFL learners — who naturally assume (unless taught otherwise) that broad must have the same vowel sound as road and that son must rhyme with on.

So her figures are not only based on a very small sample, they must also be pretty conservative. But I hope she will inspire some team of economists and educational researchers to attempt to get a robust answer to the question.


Monday 9 June 2008

O tempora! O mores!

The previous blog was about what you might call pronunciation spellings. This one is about the more familiar territory of spelling pronunciations.

The word mores, meaning ‘the customs, social behaviour, and moral values of a particular group’ (LDOCE) is a word that only educated people know. And most of them know that it is disyllabic, pronounced ˈmɔːreɪz (or in a more old-fashioned way ˈmɔːriːz).

So it was quite a surprise to hear someone on BBC R4 pronounce it monosyllabically as mɔːz, as if it were the plural of more.

It isn’t, and it isn’t etymologically related to more. It is the Latin plural of mōs ‘custom’, of which the stem is mōr-. The nominative plural ending is -ēs. The word moral is derived from the corresponding Latin adjective.

I expect that one day soon we shall hear lares et penates (household gods) pronounced ˈleəz et peˈneɪts. But at least the new London mayor, Boris Johnson, wouldn’t commit such a solecism. He is a classicist.

I have to admit, though, that I think he’s being overoptimistic in suggesting that teaching unruly youths Latin would stop them knifing one another.

(Hey, it worked for me. I was taught Latin when young and I’ve never knifed anyone.)

Boris Johnson


Friday 6 June 2008

Greengrocers' spelling

Tomorrow I shall be attending a spelling conference. So here is a spelling puzzle.

Photos: the caravan gallery

What do these London street market signs say?

Brits ought to be able to interpret them. Non-Brits may find it more difficult.

Obo-jeans is straightforward if you’re British. This is clearly a pronunciation spelling of ˈəʊbədʒiːnz, or more elegantly ˈəʊbəʒiːnz, which is properly spelt aubergines. Americans know these vegetables as eggplants.

Monge-two reads ˌmɒndʒ ˈtuː, or more elegantly ˌmɒ̃ʒ ˈtuː, properly spelt mangetouts. These are peas of which you eat the whole pod with its contents. Hence the name mange-tout, French for ‘eat (it) all’. I think Americans call them snow peas or sugar snap peas, terms not used in BrE.

If nothing else, the mis-spellings prove the lively existence of these vegetable names in the spoken English of people who are not very literate. (Or, I suppose, it just might be a literate greengrocer having a laugh.) Despite their French origin, they are well known to the general public — they are not restricted to educated speech. They may not even be perceived as French by the ordinary user.

Educated people, though, recognize that they are French and therefore may well pronounce them in a more sophisticated, semi-French way.

In native-speaker French they’d be obɛʁʒin, mɑ̃ʒ(ə)tu.


Thursday 5 June 2008

More on "I'm going to"

Several correspondents commented on the imma contraction of I’m going to (blog, 3 June).

Lynne Murphy, an American living in Britain and author of a blog on BrE-AmE differences, writes

It strikes me that this is from American English dialectal (and often humorously applied in other dialects) pseudo-aspectual a-prefixing. I.e., I'm a-going toI'm a-gonnaI'm-a.

Phil Thompson, to whom I owe the little cartoon alongside, writes

I can report from California (although I grew up in Iowa) that aɪŋənə is very common. In my own speech Iíd say that that is the most common realization, although Iíd offer another possible form: aämŋənə. Obviously, there are many subtle variations and each speaker is continually throwing off new “isotopes”. Searching my inner linguistic soul I find that the form implied by imma is not part of my accent identity. It seems to indicate a southern or AAVE speaker and I hear it in my mindís ear as amə or aːmə or āmə. I canít quite hear the syllabic variants you suggest...

(AAVE = African-American Vernacular English.)

I was probably wrong in my speculation about possible double mm or syllabic .

The muffin cartoon spelling implies a rather less reduced form, namely amənə.

(If it were British, the spelling munna would probably be taken as the Midlands/north-of-England dialect form of mustn’t, ˈmʊn(ə).)

Ryan Denzer-King writes

The contraction you commented on is fairly widespread in the American southeast (at least according to my intuition, I may be overestimating its usage). Growing up in Atlanta, Georgia, it was certainly well-known in my social circle, though when we used it it was mostly in jest, as it definitely has a lower-class connotation.

He thinks it has only a single m, and would prefer to spell it ima.

Eric Armstrong thinks there may be an intermediate stage aːm oːn I’m ’one, a southern form of I’m gonna.

John Cowan writes

The form ɑmə began as AAVE basilect and has gone upmarket in the last several decades, like a lot of AAVE basilect (it's discussed in Dillard's Black English). It's an open question whether this is nasal assimilation as you explain, or derived from a Plantation Creole future/volitional particle a ("I'm a-quit"), or some of each.
I (49, white, born just outside the NYC isogloss bundle but resident there for 30 years) have aɪmənə (not aɪŋənə, which I have never heard) most of the time, and aɪmə occasionally. I would say that aɪmənə is now a standard pronunciation, and that aɪmgənə strikes me as only suitable for formal or emphatic contexts.

With that, I’m’a declare this discussion closed. Which is just as well, since the point of the cartoon we started with has now been overtaken by events.

I’m gonna eat you!


Wednesday 4 June 2008

Sibilant genitive singulars

Alice Cheung of Hong Kong says

I'm confused about the usage of apostrophe s ('s). Is it correct to write like this:
Liz’ diary...
I noticed that many would write Liz's diary. When I look up the dictionary, I see that these examples are correct:
St James’ Palace
King Charles’ crown
I wonder if phonetics / pronunciation are the chief reasons for accepting them. If so, how about words ending with z then?

So, what can we say in answer? This topic is actually covered briefly in LPD, under -’s. There are longer and more detailed discussions in the big grammars: Pullum and Huddleston Cambridge Grammar of the English Language, p. 1595-6, or Quirk et al. Comprehensive Grammar of the English Language (Longman) p. 320-1.

Stems ending in sibilants (i.e. s z ʃ ʒ tʃ dʒ) form their plural by adding an extra syllable, namely ɪz (or for some speakers əz). They form the genitive (possessive) singular in the same way in speech, although we write this ending as ’s, using an apostrophe.

So we have

spoken: ˈlɪzɪz ˈdaɪəri
written: Liz’s diary

No one speaking standard British or American English would say ˈlɪz ˈdaɪəri. Nor do we write Liz’ diary. Similarly, it’s Liz Jones’s diary (see picture).

This is the general pattern applying to words and names ending in a sibilant.

maɪ ˈniːsɪz ˈwedɪŋ my niece’s wedding
ˈrɒsɪz əˈdres Ross’s address
sən ˈdʒɔːdʒɪz ˈhɒspɪtl St George’s Hospital
ˈdʒəʊnzɪz səkˈsesəz Jones’s successors

But there are certain exceptions, most of them being classical, literary or religious names ending in the sound z (not necessarily the letter z). With these you can optionally pronounce zero and write just the apostrophe.

ˈdɪkɪnzɪz ˈraɪtɪŋ Dickens’s writing OR
ˈdɪkɪnz ˈraɪtɪŋ Dickens’ writing
ˈbɜːnzɪz ˈpəʊɪtri Burns’s poetry OR
ˈbɜːnz ˈpəʊɪtri Burns’ poetry

This explains St James’ Palace and King Charles’ crown above.

Personally I usually follow the regular rule with these names, i.e. I would say ˈbɜːnzɪz ˈpəʊɪtri and write Burns’s poetry; I would say kɪŋ ˈtʃɑːlzɪz ˈkraʊn and write King Charles’s crown. The only cases where I might go for the irregular version would be with Greek or Hebrew names.

juˈrɪpɪdiːz ˈpleɪz Euripides’ plays
ˈsɒkrətiːz fɪˈlɒsəfi Socrates’ philosophy
ˈməʊzɪz ˈlɔː Moses’ law

There are also two special cases, which end in s rather than z. One is the fixed expression for goodness’ sake, which never has a spoken extra syllable or a written extra s. The other is the name Jesus. For this, we can write either Jesus’s love or Jesus’ love, and pronounce correspondingly.


Tuesday 3 June 2008

Imma what?

from Pundit Kitchen

I was struck by imma in the third line of this superimposed caption. It evidently means ‘I’m going to’, and is presumably pronounced aɪmmə (or possibly aɪmm̩ə, with the second m syllabic).

No doubt the initial diphthong can be reduced, too, to some kind of a.

One of my little campaigns over the years has been to try to persuade learners of EFL that there is a respectable weak form of going to that people ought to teach and learn, namely ɡənə (or before a vowel ɡənu). We Brits tend to fight shy of the spelling gonna, thinking of it as too American. Yet it is clear that we too can and do contract the words going to (but only when they are modal — so not in sentences like I’m going to Brighton tomorrow).

One stage further is the fusion of I’m with gonna. In my own speech this can produce aɪŋ(ə)nə. The nasality of the m of I’m coalesces with the velarity of the ɡ of gonna to yield a velar nasal ŋ, in a kind of two-way assimilation. The following schwa can disappear.

But for the writer of these words put into Hillary’s mouth, presumably American, the output is not a velar but a bilabial nasal that spreads over the following underlyingly velar and alveolar consonants. This assimilation is purely progressive, like ˈrɪbm̩ for ribbon.

I don’t recall any discussion of these two I’m gonna possibilities in Linda Shockey’s Sound Patterns of Spoken English, but I haven’t got the book at hand to check. And I have no idea what proportion of speakers, American or even British, go for aɪmənə or aɪmmə rather than the aɪŋənə or aɪŋnə that I would use.


Monday 2 June 2008

The perfect voice

From the BBC website: the secret of the ‘perfect voice’.

Researchers say they have worked out a mathematical formula to find the perfect human voice.
The study, commissioned by Post Office Telecoms, asked people to rate 50 voices then analysed the results.
It found the best female voice to be a mixture of Mariella Frostrup, Dame Judi Dench and Honor Blackman. Alan Rickman and Jeremy Irons did best for the men.
The research was conducted by linguist Andrew Linn, of Sheffield University and sound engineer Shannon Harris.

Andrew Linn is a respectably trained phonetician/linguist. This is a PR company speaking.

The pair worked out their formula based on the combination of tone, speed, frequency, words per minute and intonation.
They concluded the ideal voice should utter no more than 164 words per minute and pause for 0.48 seconds between sentences. Sentences themselves should fall rather than rise in intonation.

Read more, and listen to sound clips, here.

For a pseudo-scientific formula, see here.

Perfect Voice Quality - The Formula Revealed
([164.2wpm × 0.48pbs]Fi) = PVQ
wpm - words per minute
pbs - pauses between sentences
Fi - falling intonation
f - frequency between 34.5 Hz and 12.2kHz
PVQ - perfect voice quality

Amy Stoller comments:

Must have been a slow news day. Not that I donít personally enjoy listening to Dench, Irons, and Rickman, but talk about inflated claims ...

P.S. Here is a rebuff from Prof. Linn.
This statement was not based on any research I have done or in which I have been involved.
I did not devise this formula, and indeed had asked not to be associated with the formula, which I regard as meaningless. The formula, like all the quotations attributed to me, was devised by the PR company responsible for the story.
The press release was put out without including changes I had requested, without my being shown the final version, and without my being informed that it was going out.
I had been asked to comment on the study. None of my comments or my views on the plausibility of such a formula were included in the press release. Any queries about this item should be directed to 3 Monkeys Communications, Axtell House, 23-24 Warwick Street, London W1B 5NQ.

Andrew Linn

Blogroll links:


To search my web pages, use this Google search.

Google Custom Search

my home page