UCL Division of Psychology & Language Sciences

John Wells
Photo: Masaki Taniguchi

John Wells’s phonetic blog archive 16-30 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


Monday 30 June 2008

Eszett — the answer

On Friday I posted an extra blog entry concerning the Guardian’s strange report that the German letter ß

has been officially accepted as a lower and upper case alphabet figure by the International Organisation of Standardisation (ISO)

Given that ß has been a recognized German letter since the year dot, and that the ISO would have no standing in matters of German orthography, I asked if anyone could explain what this garbled account was actually about.

Despite the fact that this concerns writing systems rather than phonetics as such, more readers emailed to me on this than on any previous topic. Thank you, all of you. Between you, you have solved all aspects of the mystery.

The answer to the conundrum is that in Unicode version 5.1.0 several new characters have been added, including U+1E9E LATIN CAPITAL LETTER SHARP S. These Unicode changes came into force on 4 April this year, and the ISO accepted the changes on 23 June. Hence the Guardian report.

Here is the relevant bit of the Unicode standard 5.1, with newly recognized characters shown against a yellow background.

The Unicode commentary adds:

In particular, capital sharp s is intended for typographical representations of signage and uppercase titles, and other environments where users require the sharp s to be preserved in uppercase. Overall, such usage is rare. In contrast, standard German orthography uses the string "SS" as uppercase mapping for small sharp s. Thus, with the default Unicode casing operations, capital sharp s will lowercase to small sharp s, but not the reverse: small sharp s uppercases to "SS". In those instances where the reverse casing operation is needed, a tailored operation would be required.

How it is that my email client (Thunderbird) can display the new character correctly I do not know, since I can’t find in any of my installed fonts. But it does. One of you told me that a Duden published in the GDR was entitled

and that’s what was displayed.

You pointed me to the University of Bremen’s Bremer Sprachblog where there is a discussion of the matter (in German); to a newsfeed item on for 25 June; to a brief English article in Wikipedia, and a much longer one in German.

But you don’t need to worry. The capital ß

ist nicht Bestandteil der offiziellen deutschen Rechtschreibung.
[is not a component of official German orthography.]

ISO notwithstanding.


Friday 27 June 2008 (extra)


An article in today’s Guardian makes the strange claim that

following a century-long struggle for recognition

the German letter ß

has been officially accepted as a lower and upper case alphabet figure by the International Organisation of Standardisation (ISO)

But the lower-case letter, at any rate, has for decades featured in ISO/IEC 10646, which goes hand-in-hand with Unicode. The letter ß is part of the Latin-1 Supplement that caters for the special characters needed for western European languages. Its Unicode number is 00DF.

So what is this about? Can anyone tell us?

Could it be about an upper-case version? (But as I understand it, the upper-case equivalent is SS.)

Or is it to do with the characters permitted in URLs, which have till now been restricted to the basic ASCII set?

Either way, the Guardian report is less than clear.


Friday 27 June 2008

The letter c yet again

One of yesterday’s matches at Wimbledon was between Aleksandra Wozniak and Caroline Wozniacki. The commentators had an amusing time trying to make sure we knew the difference between ˈwɒzniæk and ˌwɒzniˈæki.

In Polish these surnames are respectively Woźniak ˈvoʑɲak and Woźniacki voʑˈɲatski. So they are quite a lot more different from one another in Polish than in English.

But Aleksandra is Canadian and Caroline Danish. Although both are native speakers of Polish, in their adopted countries they have to be content with the foreigner’s version of their names, which doesn’t include (in Caroline’s case) any awareness that c might denote ts.

Nigel Greenwood reminds me that further to the southeast there is a different problem with the letter c, namely that in Turkish it stands for a voiced affricate . Most foreigners don’t know this. So Incirlik air base (in Turkish spelling, İncirlik) gets mispronounced as ˈɪntʃ- and former prime minister Ecevit as ˈetʃ-. In Turkish, he says, they are inˈdʒirlik and ˈedʒevit.

Caroline Wozniacki


Thursday 26 June 2008

Welsh ll

In my blog for 17 June I mentioned that Amy Stoller had drawn my attention

to a BBC web page where you can hear audio files of the pronunciation of various Welsh place names. Trying it out, I was immediately struck by the way the speaker pronounces the Welsh ll (e.g. in Llandeilo, Llangwm, Llanddewi Brefi, Benllech). Instead of the standard voiceless alveolar lateral fricative ɬ, he uses a voiceless palatal fricative ç. Jack Windsor Lewis (here) has beaten me to the draw in pointing this out.

I was delighted to receive a reply from BBC CymruWales, whose Ceri Davies wrote

Thank you for your comments about the Welsh pronunciation guides on I am not a specialist in the Welsh language, but happen to be a native Welsh speaker. To my ear, and to the ears of my Welsh-speaking colleagues here at BBC Wales Education & Learning, we cannot discern a problem with the pronunciation of the letter "ll" on BBC Wales' pronunciation guide. We have listened carefully, concentrating on "Machynlleth" and "Benllech" in particular. The audio ... sounds good to us.
It's interesting to consider the different ways in which "ll" might be pronounced, but (unless someone is actually saying a soft "ch" in error) all techniques sound acceptable to us. I wonder however whether in future we should bear in mind a particular nuance of pronouncing "ll" when recording audio intended for Welsh learners?

Martha Figueroa-Clark of the BBC Pronunciation Unit (who holds an MA in Phonetics from UCL) comments

Following my conversation with Ceri ... I have listened to many of the soundfiles repeatedly and, to my ear, there are still clear instances of realisations with the unexpected palatal fricative as well as the alveolar lateral fricative (Llanwrtyd Wells, Llanelli and Machynlleth in particular sound palatal to me). I find it very intriguing that these differences do not seem to be discernible to native Welsh speakers' ears.

So there we have it. The use of a palatal fricative rather than an alveolar lateral fricative, which is immediately obvious to us four phoneticians (Amy, Jack, Martha and me), is beneath the radar as far as (some) native Welsh speakers are concerned.

I wonder if there is a sound change in progress, and that ɬ is indeed giving way to ç. If so, and the palatal succeeds in driving out the alveolar lateral, speakers of Welsh will no longer be able to boast of having a really exotic sound in their consonant system. They’ll be no more able to lay claim to exclusivity than the Germans, and the use of the true alveolar lateral fricative will be left to Zulu and Xhosa.

In the words of the Zulu farewell, ˈhámb̤a ˈg̊àːɬé hamba kahle! Go well!

PS. OK, and some American Indian languages such as Kwakiutl. And some others.

Machynlleth: maˈxənɬeθ or maˈxənçeθ?


Wednesday 25 June 2008

The letter c again

It’s Wimbledon fortnight, and our television screens are full of tennis.

Two weeks ago I watched a tennis match on the television in which one of the players bore the name Troicki. The Eurosport commentator called him ˈtrɔɪki.

But he’s Serbian, and in Cyrillic his name is written Троицки. It ought to be pronounced, in anglicized form, as ˈtrɔɪtski. (I think it would be unreasonable for him to insist on our retaining the Serbian stress, troˈitski.)

English people seem to have a complete blind spot about the letter c having the value ts in eastern European languages. See Liberec and Vaclav (blog, 3 January).

On the other hand, the final -ić in the names of various Serbian, Croatian and Bosnian players does not offer a problem to the commentators. The diacritic may be omitted in newspaper reports, on the scoreboards and in television captions, but there is no hesitation in giving Ančić, Bogdanović, Djoković, Ivanović, Jugić-Salkić, Karlović, Ljubičić and Tipsarević a final . (Ljubičić gets his medial in English, too, although Ivanišević, like the late Milošević, usually misses out on his ʃ.)

“Djokovic” is properly Djoković or Đoković, or in Cyrillic Ђоковић. (He’s also a talented impressionist — see here. You can call it djoking around, though you wouldn’t be the first to do so. Thanks to Biljana Čubrović for this link.)

There’s a different problem for the commentators with the Slovak player Hrbatý. In Slovak he’s stressed on his syllabic r, thus ˈɦr̩batiː. In English they call him hɜːˈbɑːti or hɜːˈbæti.

Viktor Troicki

Novak Djoković


Tuesday 24 June 2008

Two further phonetic fonts

I followed a link from John Maidment’s blog to an interesting website called Languagegeek.

This site is dedicated to the promotion of Native North American languages, especially in providing a means by which these can be used on the internet. Languagegeek provides fonts and keyboard layouts which try to cover all of the glyphs (alphabetical letters/Syllabics) necessary for writing the Native languages on the continent. Syllabics (Algonquian, Athapaskan, Inuit, Cherokee) take centre stage with type faces in several different styles. Fonts are also available for those languages which use a Roman orthography.

and, the author might have added, full coverage of the IPA (except for the new labiodental flap U+2C71).

All new fonts are high quality OpenType fonts. This means that with the correct software, your typing will look more precise and professional. These fonts have built into their programming the means by which ligatures (combinations of characters) and accents position themselves correctly. Furthermore, each font is Unicode friendly, meaning that the encoding is standard world wide.

You can download these two phonetic fonts, which are free.

I am quite pleased with these fonts. As you can see, they avoid the errors at ɝ and ʊ which we have noticed in other new fonts (blog, 5 and 12 March, 4 April 2008). The only thing I would criticize is the sans-serif ɪ which, as we have seen and contrary to the implication of the font name, really needs serifs.

Aboriginal Serif

Aboriginal Sans

The author, Christopher Harvey, also offers an interesting account of his own Southern Urban Ontario speech. Ac mae e’n siarad Cymraeg!


Monday 23 June 2008

Your tiny hand is frozen

Pronouncing Italian words is fairly straightforward, as long as you know where the stress goes.

All too often, the Classic FM announcers don’t.

More than once I have heard them say Che gelida manina with the stress in gelida on the -li- instead of on the ge-.

If you watch or listen to this aria being sung, you can hear the correct stress. Try this YouTube clip. It’s ke ˈdʒelida maˈnina.

Same rhythm as its English version your 'tiny hand is 'frozen. Duh!

In Latin we have adjectives 'acidus, 'gelidus, 'rapidus, 'solidus. The suffix -id- has a short vowel, therefore stress goes on the antepenultimate. Correspondingly, in Italian (and also in Spanish) it’s ácido, gélido, rápido, sólido; in English acid, gelid, rapid, solid ˈæsɪd, ˈdʒelɪd, ˈræpɪd, ˈsɒlɪd. I could go on: it’s just the same for the Latin and Italian words corresponding to English avid, horrid, liquid, rigid, splendid, valid and many others. Where’s the problem? Why would anyone suppose that the stress fell on a different vowel in Italian?

Perhaps the French that we British learn at school is to blame. Although French acide, rapide, solide etc. in theory have no word stress, in practice we treat them as having final stress: aˈsid, raˈpid, sɔˈlid. Taken over into Italian, this could be the source of the announcer’s mistake. (The actual item *gelide seems to have been lost in French: they say gelé.)

And by the way it’s Plácido Domingo.

Tim Owen has conducted a quick mini-poll on the pronunciation of pwn (blog, 20 June), among the members of something called The Oratory Forums. Here’s how they voted.


Friday 20 June 2008


There is always something new to learn. Today: the verb to pwn and its related noun.

According to Wikipedia, this is a leet slang term, originating as a mistyping of own and meaning to own in the sense of being able to dominate or humiliate a rival. It is used in the internet gaming culture ‘to taunt an opponent who has just been soundly defeated’ (charming!).

Facebook users who have installed SuperPoke can give someone else a virtual pwn.

Wikipedia offers a range of possible pronunciations ranging from poʊn through puːn to more imaginative suggestions such as pəˈoʊn and pwəʔˈn (with syllabic n).

The past participle is apparently pwn3d (see sidebar picture).

My own circle of acquaintances doesn’t, as far as I know, cover anyone who belongs to the internet gaming culture or who speaks leet, so I’ll have to take all this on trust.

My first feeling on encountering this word pwn, though, was that it should be pronounced pʊn, which is how it would be if it were Welsh. (Or if spelt pŵn, then puːn.)

Actually I’ve just checked my Welsh dictionary and I see that pwn is a Welsh word, pʊn, though it means ‘pack, burden’. For complicated but predictable reasons of Welsh phonology, its plural is pynnau ˈpənai.

I’ll take that as a virtual pwn.

P.S. Andrej Bjelaković, ‘a proud member of the internet gaming culture :)’, writes

For some reason I always thought of the word as being pronounced like 'pawn', i.e. with a THOUGHT vowel. I have no rational explanation for this, but all other pronunciations seem awkward.

He wonders what the preferred version is among native speakers. So do I.

P.S.2. Tim Owen says

I am a frequent user of message boards and forums. "Owned" and "Pwned" (alternatively "Pwnt") have been around fairly commonly since I first started surfing regularly in 2003. Every time I use it and have heard it (which runs into hundreds of times over the last five years), it's consistently been pronounced as the word 'own' prefixed with a 'p': 'pown'.



Thursday 19 June 2008

A 'sexy' accent?

Jim Scobbie posted this on Facebook. (Click the picture for the link to a BBC video clip.)

Just why a French accent in English is considered sexy I’m not sure. Like other stereotyped reactions to accents it presumably cannot be taken literally: there can hardly be any objective phonetic correlates of sexiness that are found in French voices and absent from English ones.

Presumably it is a relic of the late-Victorian and early-twentieth-century British attitude to France as the home of uninhibited sexuality. Dirty weekends took place in Deauville or Nice. Aristocrats or other public figures implicated in sex scandals would seek refuge in Paris. Condoms, condemned for encouraging immorality, were known as ‘French letters’. We borrowed the word risqué (first use in English 1867).

Which country has now replaced France in this stereotypical role? Thailand? Perhaps. But a Thai accent is not considered particularly sexy.


Wednesday 18 June 2008

Don’t waste a drop

Petrol — warns the newspaper as it reports the fuel tanker drivers’ strike — don’t waste a drop.

If you say this aloud, the intonation nucleus goes on drop.

'Don’t waste a 'drop.

You would think the most important word would be waste, or even don’t. Yet we place the main accent on neither of those items, but on the apparently rather unimportant drop.

There’s no kind of contrast involved. We’re not contrasting drop with bucketful or litre. So what’s going on?

Not ... a drop is a more complicated way of saying nothing. Don’t waste a drop is an idiomatic way of saying Waste nothing. It parallels other similar expressions.

I 'can’t see a 'thing. (= I can see nothing.)
I 'won’t tell a 'soul. (= I’ll tell no one.)
(Have you 'heard from them?) — 'No, | 'not a 'dickybird. (= nothing)
(Have you 'heard from 'Tom?) — 'No, | 'not a 'sausage.

I 'haven’t got a 'dime. (= I have no money.)

The nucleus that would otherwise be located on nothing or no one is retained on the dummy word that is left behind after the negative has been fronted.

This is different from what happens with the dummy words I call ‘empty words’ (English Intonation p. 150). They do not get an accent.

I keep 'seeing things.
I 'promise I won’t 'tell people.

How difficult this all is for the EFL learner!

Back to the petrol crisis. If you’re into panic buying, get down to your local petrol station immediately.

'Don’t waste a 'minute.

Actually, you needn’t bother, because I’ve just heard that the strike has been called off.


Tuesday 17 June 2008

Llantwit Major

Petr Rösel asks, on behalf of one of his students, how to pronounce the name of Llantwit Major, a small town in the Vale of Glamorgan.

Llantwit Major is in an English-speaking area, and Llantwit Major is actually its English name: in Welsh it is called something else. Nevertheless, the name looks Welsh, with the characteristic initial Llan-.

So I was interested recently to see on TV a feature about the place. It was striking that the locals all pronounced the name with an ordinary voiced approximant initial l. The visiting reporter, on the other hand, made a great effort to produce a Welsh ll, ɬ. He shouldn’t have bothered. It’s ˈlæntwɪt ˈmeɪdʒə (confirmed by the BBC Pronouncing Dictionary of British Names).

The Welsh name is Llanilltud Fawr ɬaˈnɪɬtɪd ˈvawr.

Fawr, the mutated form of mawr, just means ‘great, big’, and corresponds to the Latin Major of the English name. According to the Dictionary of the Place-names of Wales (blog, 22 February), both the English Llantwit and the Welsh Llanilltud have the same Welsh origin, llan- ‘church’ plus Illtud, the name of a monk who reputedly established a monastery there in the sixth century. English -twit, according to the authors, is traceable to an early variant Illtwyd from an Irish (!) genitive Illtuaith.

While we’re on the subject, Amy Stoller drew my attention to a BBC web page where you can hear audio files of the pronunciation of various Welsh place names. Trying it out, I was immediately struck by the way the speaker pronounces the Welsh ll (e.g. in Llandeilo, Llangwm, Llanddewi Brefi, Benllech). Instead of the standard voiceless alveolar lateral fricative ɬ, he uses a voiceless palatal fricative ç. Jack Windsor Lewis has beaten me to the draw in pointing this out, and makes various other criticisms: see his blog.

I’m not sure whether this delateralization counts as a speech defect. I have come across it once before, in the speech of a Welsh-speaking undergraduate we had at UCL a few years ago. But it is certainly not the mainstream pronunciation, and I do not think it ought to be given to the general public as a model.

Church of St Illtud, Llantwit Major


Monday 16 June 2008


The Guardian newspaper, nicknamed the Grauniad for the frequency of its misprints, has in recent years been doing much better in this respect.

Nevertheless a recent issue offered us nous spelt as nouse. This is how it appears in the on-line version.

English nous has nothing to do with the French first person plural pronoun. As a non-specialist term, pronounced naʊs, this Greek-derived word means common sense, gumption, know-how.

Presumably the Guardian journalist was misled by the spelling of such everyday words as house, mouse, louse which rhyme with it. (Aside: why do we write a final e in these words? Compare German Haus, Maus, Laus.)

In my first year of learning classical Greek, the teacher used the old schoolroom pronunciation of Greek, in which ου was pronounced as English . So νοῦς was indeed naʊs. This is the source of our usual English pronunciation of the word. But the teachers in my second and subsequent years used the modernized pronunciation, in which ου was pronounced as English .

I thought I would do a Google search on the spellings with and without final e. Nous gets 451 million hits, nouse less than one million. However, this search threw up an interesting fact, new to me: a nouse, properly so spelt, is a nose-operated mouse.


I am sorry that Jack Windsor Lewis took exception to my light-hearted and tongue-in-cheek references to RP last week (blog, 12 June). ignorant unsubstantiatable claptrap...

On a more serious issue, he asks whether dʒeɪms for James and njuːs for news would have been considered wrong if the candidate had claimed to have a north Wales accent (blog, 11 June). The answer is no. After all, I did point out the absence of z and ʒ in ‘unsophisticated accents of Welsh-speaking north Wales’ in Accents of English, p. 389.

More interesting is the question of whether ˈmænɪʒd for managed is necessarily wrong. I think it is true that in rapid informal speech the plosive element of the affricate may indeed sometimes be absent. But in the context of the exam script concerned, I am in no doubt that it reflects a weak candidate’s confusion over the symbols and ʒ, not some sort of attempt to make a sophisticated point.

Blogroll links:


To search my web pages, use this Google search.

Google Custom Search

my home page