a heavenly haven

Peter Roach wrote

I wonder if you noticed the interview on TV yesterday on Channel 4 News with Christine LaGarde, French Finance Minister. My wife and I were marvelling at her excellent English (including her pronunciation) but wondering why she kept referring to "tax heavens" instead of "tax havens". We then remembered that the French for this is "paradis fiscal". I hope somebody has put her right now.

A lovely example, which segues neatly from yesterday’s post!
Mme LaGarde’s English is indeed much too good for her to confuse eɪ (as in haven) with e (as in heaven). But with heaven being a near-synonym of paradise, one can understand the confusion.
It suggests that she has learnt the English expression tax haven by hearing it used in context, rather than by encountering it in the course of reading.
Strangely enough, I was discussing this topic just a few days ago (27 March) in my other blog. English haven has moved away from its historical meaning ‘harbour’ to its current meaning, a refuge.
Where English speaks of a tax haven for those reluctant to pay taxes in their home country, French, German and Spanish speak of a paradise (paradis fiscal, Steuerparadies, paraíso fiscal).
But the Slavonic languages go a different route: in Russian they call it an оффшорная зона (ʌfˈʃoɾnəjə ˈzonə, offshore zone), and in Polish terytoria offshore. The Japanese follow English with タックス・ヘイヴン (takkusu heivn).
At least, that’s how the various Wikipedia articles are entitled in the respective languages.

raw fish

Masaki Taniguchi tells me he was on a train in Japan recently when he got talking, mostly in English, to a Swiss lady who was touring Japan with her family. She was from Lausanne and her first language was French. She mentioned that she loved sushi.
Sushi comes in many varieties. While westerners usually like most of them, not everyone is enthusiastic about those that include uncooked fish.
Masaki said to the Swiss lady “Do you like raw fish?”.
Her unexpected reply was “Oh, you speak German?”.
It so happens that the (standard) German roher Fisch ʁoːɐ fɪʃ sounds very similar to the English raw fish ɹɔː fɪʃ. (Don’t ask me how it would be pronounced in Swiss German.)
This reminds me of a more complicated multilingual example, probably apocryphal. A visiting Frenchman in Enɡland went to buy an ice cream. The vendor asked him what size he wanted. The Frenchman understood the question but answered in French, “à deux boules” (with two scoops) a dø bul. Fortunately communication was not impaired, despite the interlocutors’ ignorance of one another’s language, because the English vendor heard it as “a double” ə dʌbɫ.

ginkgo

In the front quad at UCL there are two ginkgo trees, which every autumn shed their distinctive leaves on the ground.
According to a Wikipedia article (which I have no reason to doubt), the Latin (and hence the English) specific name of the tree, Ginkgo, results from a combination of folk etymology and misreading.
All the OED can tell us about the etymology of ginkgo is

[Jap., f. Chinese yinhsing silver apricot.]

In Chinese characters and Hanyu pinyin this Chinese etymon would be written 銀杏 yínxìng. When the tree was introduced into Japan from China the Chinese name was borrowed into Japanese with the pronunciation ぎんなん ginnan. (The Japanese pronunciation of Chinese words and “readings” of the kanji in which they are written is a topic way beyond my knowledɡe.)
But the same Chinese characters can also be read in Japanese as ginkyō, which is where the folk etymology comes in. Apparently Engelbert Kaempfer, the first Westerner to see the species in 1690, wrote down this form, with its pronunciation, in his Amoenitates Exoticae (1712). But his y was misread as a g, and the misspelling stuck.
Our pronunciation follows the spelling, so that we call this tree ˈɡɪŋkəʊ. If it had not been for Kaempfer’s bad handwriting, we’d presumably be calling it ˈɡɪŋkiəʊ.
Because of the pronunciation, people also often misspell it as gingko.
Confusingly, there is also a Japanese word ginkō, pronounced with -ŋ-. But it means ‘bank’.

The specific part of the scientific name Ginkgo biloba transparently means ‘having two lobes’, a reference to the shape of the leaves. You’d think that it would be pronounced in English as ˌbaɪˈləʊbə, since this is what we get for the Latin prefix bi- in bisexual, bifurcation, bipolar and other words. But in practice people who talk about the supposed medical benefits of Ginkgo biloba extract generally seem to say bɪˈləʊbə.

a controlled rolling grunt

Anyone who studies phonetics at a British university, and no doubt in many other countries too, has to learn to recoɡnize and produce a number of “difficult” or “exotic” consonants, among them the one that we transcribe ʕ, which is classified on the IPA Chart as a voiced pharyngeal fricative. Here’s what it sounds like (from Ladefoged’s website).
For a language that includes this sound, people usually think first of all of Arabic, where the sound associated with the letter ع (ʿayn) has traditionally been classified by phoneticians as a voiced pharyngeal fricative and written ʕ. (The IPA symbol was chosen to be reminiscent of the top half of the Arabic letter.)
However, Robin Thelwall argued in 1990 (JIPA 20.2:37-41) that the Arabic sound is not actually a pharyngeal fricative but a pharyngealized glottal stop. I think he is probably right. When pronounced by native speakers of Arabic, it often seems to involve, as well as a constriction in the pharynx, a momentary cessation of the vibration of the vocal folds.
The Hebrew alphabet, too, includes a letter ayin (ע), which in some kinds of Hebrew is pronounced in the same way. Apparently this was its historical pronunciation, but nowadays many Israelis just pronounce it as a glottal stop, ʔ (which also has its own letter in the Hebrew alphabet, aleph א).
The foregoing discussion assumes that you, the reader, have enough familiarity with phonetic terminology and classification to be able to follow it. I hope you do. Those who don’t are forced into inventive but incoherent descriptive attempts such as this one that a correspondent came across in a wiki about Hebrew. He sent it to me as a “gem for your collection of examples of the complete inability of the phonetically naive to describe speech sounds”.

Ayin is not pronounced the same as Aleph. Ayin has a gutteral sound
applied to it, a gutteral sound void of tonality almost a controlled
rolling grunt.

My correspondent commented

I don't mean to mock people for knowing nothing about phonetics, but the sheer desperate inventiveness (and uselessness) of the description was striking.

pronunciation of r

There are some emails that I receive that ask simple questions for which the answers are complicated.

Hello Sir
I am k***** [name removed]. My native language is not English so I have a problem in the pronunciation of word "r". I want to know that where we should pronounce it and where not and what are rules

thanks
K*****

[Obviously s/he means not the word r but the letter r.]
Where should the EFL student pronounce the sound r? What advice should one give? It all depends.
• For a simple life, and if your model is American English, pronounce r wherever the letter r is written.
• However, if your model is RP or a similar form of BrE, or Australian or New Zealand English, or to fit in with those around you in Africa (for example), then you should pronounce an r-sound only if the sound that follows is a vowel sound. So there should be an r-sound in red, arrive, very, tree, address, purity, but not in hard, firm, north, persuade, standard, modern. At the end of a word — as in better, far, near — you should not pronounce r if the word is on its own or at the end of a sentence; but you may pronounce one if the word is followed, without a break, by another word beginning with a vowel sound.
More importantly, what is your purpose in learning English? If you just want to understand and be understood, then pronounce all the rs. If you want to fit in with native speakers in some particular place, then you must learn to do as they do. If you want to pass school examinations for which the examiners require that some particular type of pronunciation be used, then you must fit in with their requirements.
It is important to learn not just where to pronounce r but also how to pronounce it. If, as your name suggests, your language is Arabic, which uses a tapped r-sound (ɾ), it is worthwhile trying to acquire an English-style approximant r (ɹ).
I am afraid an answer along these lines would disappoint K****, who would doubtless prefer a clear short answer. And anyhow his/her command of English would probably not be high enough to understand fully what I say.
Probably a better answer would have been

Go and ask your teacher. Or ask your friends.

Unicode Phonetic Keyboard

Mark Huckvale saw that I had been unable to get his Unicode Phonetic Keyboard to work in Vista Home Premium (blog, two days ago). But the fault turns out to have been mine, not his.
He wrote:

To enable the language bar you need more than one input language
installed. So add another language first, then you can find the
language bar to enable the phonetic keyboard. To add a language, see this:
http://windowshelp.microsoft.com/Windows/en-US/Help/52e368fa-bd32-4749-955e-331f9130889f1033.mspx
Let me know if this works!

Yes, it did! Here is what the instructions said.

1. Open Regional and Language Options by clicking the Start button , clicking Control Panel, clicking Clock, Language, and Region, and then clicking Regional and Language Options.
2. Click the Keyboards and Languages tab, and then click Change keyboards.
3. Under Installed services, click Add.
4. Double-click the language you want to add, double-click the text services you want to add, select the text services options you want to add, and then click OK.

Alongside the existing English (United Kingdom) I proceeded to add Russian, Greek, and Chinese. Having downloaded the Unicode Phonetic Keyboard I activated it as an option under United Kingdom, and now the language bar on my desktop (see picture) enables me to switch between the standard keyboard and the phonetic keyboard with a single click of the mouse.
In one mode I can type TIN; exactly the same keystrokes in the other mode produce θɪŋ. Result! (The keystrokes follow SAMPA very closely.)
For anyone using Windows XP or Vista with a UK keyboard, this is undoubtedly the most convenient input method. You can also use it to type IPA directly into a comment box for the blog.
And the multiple language capabilities of Windows are great fun. In Chinese mode, if I type shan (mountain) a menu bar immediately offers me a list of possibilities, starting with 山 (mountain), the one I want.

imaginary secondary stress

Intrigued by Justin Watkins’s posting on Facebook of a video clip of himself jumping into a pool under the headline “cenote plunge”, I looked up the word cenote. Where does one most conveniently look up a presumed technical term these days? In Wikipedia of course, where I duly found

A cenote (pronounced in Mexican Spanish [seˈnoˌte], in Iberian Spanish [θeˈnoˌte] and in English [səˈnəʊˌteɪ] …) is a sinkhole with exposed rocky edges containing groundwater.

When you find an error in Wikipedia you’re supposed to just quietly correct it, not publicize it. But I thought it worth while drawing attention to this over-use of the secondary stress mark by some would-be phoneticians.
Not every unreduced vowel has secondary stress! Particularly not in languages that, like Spanish, do not go in for vowel reduction. The Spanish pronunciation should most definitely be shown as seˈnote, θeˈnote, and the English, in my view, as səˈnəʊteɪ.
Few if any Spanish words have more than one stress (the claimed exceptions are usually adverbs in -mente). And none have post-tonic secondary stress.
In English we have post-tonic secondary stress in compounds such as ˈgrandˌfather and ˈwashing maˌchine. But not (in my view) in words such as ˈeducate (not “ˈeduˌcate”) or ceˈnote.

typing IPA symbols

Harry Campbell says

Finding IPA symbols to post online can be a pain; what I do is copy and paste using the handy gadget at ipa.typeit.org … how about a link to this from the blog?

OK, done. But I actually told you about this little utility on 28 November 2008, little more than three months ago. As I said then,

You are restricted to the characters shown here. There is no length-mark available, though you could use an ordinary colon.

At that time I also wrote about David Montero’s IPA character map.
Personally, I do not use either of these. On my laptop I use Mark Huckvale’s Unicode Phonetic Keyboard, but I still cannot get it to work in Windows Vista Home Premium, which is what I have on my desktop computer. So what I do is compose my text in Word, then copy and paste into blogspot.
In Word 2007 for Windows it is very easy to insert IPA symbols or other special characters. Go to Insert | Symbols | Symbol | More symbols, and find your symbol in the drop-down box. (I’ve put a shortcut on my Quick Access Toolbar to save time.)
Select a suitable font (one that includes the IPA). Find your character in the display box, scrolling as needed, and Insert. Characters you have used recently will be available for reuse.
You can set up macros (Shortcut Key), too, if you want.
Most of the IPA symbols are in the IPA Extensions subset. But æ ð ø are in the Latin-1 Supplement; ŋ œ are in Latin Extended-A, and β θ χ are in Greek and Coptic. Stress marks and length marks are in Spacing Modifier Letters, while non-spacing diacritics are in Combining Diacritical Marks.
Another way in Word is to type in the Unicode number in hex, select, and do Alt-X: hey presto, the code number turns into the character. Then copy and paste. For Unicode numbers, go to the Unicode charts, or consult my Unicode page (which will be useful if you are not sure about which symbol denotes which sound).
Lastly, you can always select, copy and paste from a web document, e.g. a Wikipedia article or one of my phonetic pages. If you want to post a comment using one of the symbols in my posting (or someone else’s), you can copy and paste that too.

More about e and ɛ

Here’s some more on the question of whether the vowel of English DRESS is best represented as e or as ɛ (yesterday’s blog). I’m sorry that the three extra points I planned to make have grown to five, making seven in all.
3. The vowel of English DRESS varies considerably. A former RP quality, now obsolete, was very close to cardinal 2 [e]: a similarly close-mid quality is to be heard in Australian English, and in New Zealand it is often even closer. The current British average quality is somewhere between cardinal 2 [e] and cardinal 3 [ɛ]. There are also local varieties in which it is fully open-mid. In American English it can be open-mid and considerably centred. If we want to symbolize all of these possibilities in a single symbol, the phonetic case for [e] is not much different from the case for [ɛ].
4. The choice of symbol for DRESS is bound up with the question of the choice of symbol for FACE. In accents where the FACE vowel is monophthongal and not distinctively long (e.g. Scottish English, perhaps some northern AmE) we must write face as fes and dress as drɛs. In all other varieties of English, where FACE is either a long monophthong or a diphthong, there will be no confusion in writing DRESS with e as long as FACE is written correspondingly as eː or eɪ (or, with Trager and Smith, as ey, or with Chomsky and Halle as ēy).
5. Furthermore, there are many accents in which the diphthong of FACE starts more open than the height of DRESS: for example, popular London and SE England (“Estuary English”) and Australian and New Zealand speech. (It was phonetically perverse of the Macquarie dictionary of Australian English to write FACE as eɪ and DRESS as ɛ.)
6. Consider now the position, say, of Japanese learners of English. In their own language they have just a single mid front vowel, ェ e. It is (in my view) entirely satisfactory for them to use this sound for the English DRESS vowel, and as the first part of an eɪ diphthong for English FACE. If we were to write DRESS as drɛs we would be implying that they need to learn to use a special (non-Japanese) quality for DRESS, different from the starting point of FACE feɪs. And instead of occasional emails from Europe asking why I don’t use ɛ, I would be faced with hundreds of emails from Japan asking me to explain why I use a funny symbol ɛ. (My readers will correct me if I am wrong, but — as far as I know — Japanese, Korean and Chinese dictionaries of English, except in Taiwan, all transcribe the DRESS vowel as e.)

7. Lastly, the tradition in English pronouncing dictionaries ever since the first appearance of EPD nearly a century ago has been to write DRESS as e. If it ain’t broke, don’t fix it.

e and ɛ

Persijn M. de Rijken, a student at the University of Bergen, writes

In your dictionary, words such as "bed" are transcribed with an [e], when elsewhere I often see the same vowel rendered as [ɛ]. Also, when I listen to various vowel recordings I find that [ɛ] comes closer to the sound I make when I say "bed". Of course, I am not a native speaker of English, and it may well be my Dutch or Norwegian that is creeping in, but I was wondering as to the nature of your choice for using [e]?

This is a question quite often asked. Indeed, when I recently met the famous linguist David Stampe for the first time, he upbraided my notation system on similar grounds. Because the DRESS vowel is lax, not tense, he says, it ought to be written ɛ, not e.
You can see a brief treatment of this issue at the end of this short article.
Here is a somewhat longer analysis of the problem. I’ll do it in two parts, one today and the other tomorrow.
1. Although I usually use e for the English vowel of DRESS, there is nothing incorrect or unscientific about using ɛ instead. This is the symbol used in the Kenyon-Knott system, still in use among American dialectologists and speech people (who just call this notation “IPA”), and also for EFL in Taiwan (but as far as I know, virtually only in Taiwan). Ladefoged used it in earlier editions of his Course in Phonetics, before switching to e. Clive Upton uses it in the ODP and in his transcriptions for the OED and the Concise Oxford, though the Oxford ALD remains faithful to e. The fact that different people use different notations does not imply any difference in the sound that we symbolize.
2. An important principle of good phonetic transcription is simplicity. As Ladefoged formulates it,

A simple phonemic transcription uses the smallest possible number of different letters. [D. Abercrombie, 1964, English Phonetic Texts, London: Faber & Faber, page 19.]

Since the DRESS vowel can be unambiguously written e, an English transcription with the symbol e is simpler than one with ɛ. The same applies in those languages that have just five vowels (or just five peripheral vowels), e.g. Modern Greek, Serbian, Polish, Czech, Japanese. In these languages the mid front vowel is best, and most simply, written e. Whether its quality is closer to that of cardinal 2 [e] or cardinal 3 [ɛ] is irrelevant. On the other hand, in languages that have a close-mid front vowel in contrast to an open-mid one (e.g. French, German, a Scottish accent of English), the two symbols e and ɛ must both be used; and that means ɛ for French même, German Bett, and Scottish dress.
There is not some great phoneme system in the sky from which particular languages select their phonemes, with one IPA symbol always standing for the same thing. Languages differ. We use the symbol t for the aspirated alveolar plosive of English and the unaspirated dental plosive of Russian because it is simpler to do so. We use the same symbol l for the clear alveolar lateral of German, the variably-coloured alveolar lateral of English, and the retracted post-alveolar lateral of Korean (yesterday’s blog), because it is simpler to do that than to festoon the transcription with diacritics. It is better to state such information in the conventions that accompany a phonetic transcription rather than in the transcription itself. In a language that distinguishes dental and alveolar (or postalveolar) sounds, we obviously need to symbolize their place of articulation explicitly. In a language that doesn’t, we don’t.
For points 3, 4, and 5 please wait until tomorrow.

going up

Jonathan Marks said

I used to like the visuals in the right-hand column of the old blog. Pity they seem to have disappeared in the new version.

It’s true that the new site does not allow me to place pictures in the sidebar. But I can still put them in the body of the text. I’ll try to remember to include more than I have been doing since moving sites. There are two small graphics in today’s post, just to make sure.
Contrary to popular belief, neither the Korean hangŭl script nor Japanese kana are opaque in the way that Chinese characters are. The Korean writing system is actually alphabetic, while the Japanese katakana is a straightforward syllabary. Both of them can be spelled out symbol by symbol.
Here is the word elevator written in Korean.
It reads el-li-be-i-tŏ. When the Korean phoneme written ᄅ is between vowels it is pronounced as a tap, ɾ. But in syllable-final position, and when geminated (as here), it is pronounced as a lateral, l. As the IPA Handbook puts it,

/l/ is [ɾ] intervocalically; … /ll/ is [ll]…

Here is the same word in Japanese.
It reads e-re-be-:-ta-:.
So the differences between the two languages in the treatment of this borrowing from American English are these,
In Korean,
• English l is correctly rendered as a lateral (but a geminated one);
• English eɪ is correctly rendered as a diphthong;
• the final AmE ɚ comes out as ŏ (ʌ)
In Japanese,
• English l comes out as a tap ɾ, not as a lateral (except by chance sometimes)
• the diphthong is rendered as a long monophthong (though Japanese no longer distinguishes between ei and ee, anyway);
• the final ɚ comes out as a long aː.
In both languages, English v is rendered as b. The two languages differ in how they treat the second, unstressed, vowel: but then so do native speakers.

names of letters

I was listening to Classic FM while driving, when the announcer told us that the next item would be Johann Sebastian Bach’s Air on a gə String.
I thought at first that this was a twee joke, but it doesn’t seem to have been: everything else was perfectly serious. Why he didn’t say on a dʒiː string, as normal, I’ve no idea. (Anyhow, musicological pedants would insist on Air on the G String.)
Naming the letters of the alphabet in this way is something I’ve previously encountered only among primary school teachers and children learning to read, amongst whom it seems quite common to name the letters not eɪ, biː, siː, diː, iː, ef… but æ, bə, kə, də, e, fə….
I wonder what they do when they get to K. I think it would have to remain keɪ so as to distinguish it from C kə.
I have recently joined a male-voice choir. One of our warming-up routines involves reciting the letters of the alphabet, using their normal names eɪ, biː, siː…

(do) ˈA B C D
(re) ˈE F G
(mi) ˈH I J K
(fa) ˈL M N O P
(so) ˈQ R S
(la) ˈT U V
(ti) ˈW X Y
(do´) ˈZ.

—but although we are British and would normally call the last letter zed, to make it rhyme we have to pretend to be Americans and say ziː.

I can’t get it \right any more

Where does the nucleus go in these sentences?

It’s no use sitting at home all day. You ought to get out more.

(What’s the trouble, Mary? Why are you crying?) — You don’t love me any more.

These cases of more/any more seem to be by default unaccented. They behave as if they were adverbs of time or place: because when in final position adverbs of time or place are typically unaccented.

It’s no use sitting at \/home all day. | You ought to get \out more.

You don't \love me any more.

_ _ _
I am proud to tell you that this blog has been chosen as blog of the month by the Dutch language website taalpost.nl.

Weblog van de maand: John Wells's phonetic blog
John Wells is een van de beroemdste fonetici ter wereld en de auteur van onder meer enkele gezaghebbende uitspraakgidsen voor het Engels. Sinds kort onderhoudt hij ook een openbaar weblog over uitspraakverschijnselen, vooral van het Engels. De afgelopen dagen besprak hij bijvoorbeeld de opvallende manier waarop 'Israel' soms in Engelstalige liedjes wordt gezongen, de uitspraak van de Griekse klank ou in woorden die in het Engels zijn overgenomen, en veel meer. Lezers kunnen ook reageren.

[Weblog of the month:
John Wells is one of the most famous phoneticians in the world and the author of, amongst other things, one or two authoritative English pronunciation guides. Recently he has also been keeping up a public weblog on pronunciation phenomena, especially those of English. Over the past few days he has discussed, for example, the striking way in which Israel is sung in English-language singing, the pronunciation of the Greek sound ou in words borrowed into English, and plenty more. Readers can also post comments.]

(Thanks, guys. I hope my translation is correct.)

A glottal stop is different from zero

Masaki Taniguchi tells me that he was seeking advice at the computer help desk at UCL when the IT person told him to click on the stɑː button, or so he thought. But he couldn’t see any star button. What had happened was that he had failed to notice the glottal stop after the vowel: it was not the stɑː button but the stɑːʔ button (start button) that was wanted.
This is a nice example of the phonological function of the glottal stop in English, serving to distinguish words that might otherwise sound the same. The glottal stop, although consisting of no more than a silence, contrasts with its own absence. (OK, it may be more an effect on the phonation of the end of the vowel than just an instant of silence, but the point is the same.)

ˈstɑː bʌʔn = star button
ˈstɑːʔ bʌʔn = start button

Not all English people use a glottal stop at the end of start before another consonant. Nevertheless it is unusual to have an exploded t. If it’s not glottal it’s likely to be assimilated to the following bilabial, producing an unexploded, indeed unreleased, p̚ thus ˈstɑːp̚ bʌʔn.

Worse, if he had been listening to an American rather than a Londoner, the ˈstɑːp̚ bʌʔn would have been not the start button but the stop button.

ˈstɑːp̚ bʌʔn = start button (BrE), stop button (AmE)

Pity the poor EFL learner. It is difficult to discriminate auditorily between unreleased stops at the bilabial, alveolar, velar and glottal places.

But it's not just those learning English. Those of us who are not native speakers of Cantonese find it very hard to hear the difference between final p, t, k, all unreleased, in that language. (Hong Kong airport is called Chek Lap Kok, and each syllable ends in an unreleased/glottallized final consonant.)

ghazal

A ghazal is

“a species of Oriental lyric poetry, generally of an erotic nature, distinguished from other forms of Eastern verse by having a limited number of stanzas and by the recurrence of the same rhyme.” (OED)

It is found in the literature of Persian, Urdu and other south Asian cultures.

And how do we pronounce it? It’s not in any of our three pronunciation dictionaries. The OED says it’s ˈgæzæl. But the OBGP gives “guz-ul /ˈgʌzʌl/”. So it’s one of those words like pandit/pundit, in which an a-like vowel in a foreign language, or a ə-like vowel of Hindi/Urdu, can be mapped either onto æ, following the spelling, or onto ʌ, following the pronunciation in the language of origin.

What about the initial gh spelling? We pronounce it g in English, as in ghost, ghastly and ghoul, but what is it in the source language(s)?
As far as I know it is not a voiced-aspirated Hindi/Urdu plosive such as we find in the words ghat, dhobi, Bharat. Rather, it is a voiced velar fricative, ɣ (or perhaps somewhat further back, as far as ʁ). In Arabic script the word is written غزل and in Devanagari ग़ज़ल. Both غ and ग़ stand for ɣ. (I am sure someone will write to tell me if I have got this wrong.)
All this musing is prompted by the comment in Wikipedia that

The Arabic word "ghazal" is pronounced roughly like the English word "guzzle", but with the first, g-like consonant further back in the throat.

This is unfortunately a piece of phonetic ignorance: the gh-sound might be “further back in the throat” than g (though that is not necessarily the case). But the important thing is that it’s not a plosive but a fricative. How do we explain that in a way that would be meaningful for the layman?

Here’s part of an English-language ghazal from Wikipedia.

Where are you now? Who lies beneath your spell tonight?
Whom else from rapture’s road will you expel tonight?
Those “Fabrics of Cashmere—” “to make Me beautiful—”
“Trinket”— to gem– “Me to adorn– How– tell”— tonight?
I beg for haven: Prisons, let open your gates–
A refugee from Belief seeks a cell tonight.
God’s vintage loneliness has turned to vinegar–
All the archangels– their wings frozen– fell tonight.
Lord, cried out the idols, Don’t let us be broken
Only we can convert the infidel tonight.
Mughal ceilings, let your mirrored convexities
multiply me at once under your spell tonight.

habeas corpus

I keep noticing the spelling mistake habeus corpus (instead of habeas corpus). The explanation is straightforward: the final syllable of habeas is usually pronounced in English just like the final syllable of corpus, namely as əs. If it’s pronounced the same, there is an obvious temptation to spell it in the same way.
The word habeas is the second person singular present subjunctive active (‘thou mayst have’) of habeō, habēre, to have. Its final vowel was long, habeās, so when I was taught Latin at school we pronounced it ˈhæbeɪɑːs (or ˈhæbiɑːs). English legal Latin gets both the initial and final quantities wrong, producing ˈheɪbiəs. But then legal Latin is something of a law unto itself, or lēx in sē as we might say.
The reason our Latin and Greek teachers were so insistent on our “getting the quantity right”, i.e. distinguishing between long and short vowels, is that classical versification is based on the idea of heavy (‘long’) vs light (‘short’) syllables. This is important in the appreciation of Latin poetry (see my blog for 18 Aug 2006), but was even more important for us schoolchildren who were expected to compose Latin verse every week, usually in the form of translating a passage of English poetry.
We had a useful book called a gradus, short for Gradus ad Parnassum, ‘a step towards Parnassus (the mountain sacred to the muses)’— Ainger A.C. and Wintle H.G., An English-Latin Gradus, London: John Murray, 1890. The gradus was a dictionary with a difference. It specialized in providing sets of homonyms or near-homonyms, all with the vowel quantities clearly marked, so that we could select a particular translation that would fit the scansion we wanted. Poetry translation became a kind of jigsaw puzzle, finding words that would fit the metre. A hexameter line consisted of six feet, each of which was either a dactyl (ˉ ˘ ˘) or a spondee (ˉ ˉ). So if you wanted to put ‘money’ at the beginning of a line of verse, you could choose argentum, aurum, nummus, res, or divitiae; but not the commonest equivalent, pecunia, and not opes, bona or moneta.

I don’t suppose many schoolchildren have to do that sort of thing nowadays.

syllabic plosives

Timothy Birkett wrote:

I was put on the spot the other day when a student asked me if /p/ in potato could be marked as a syllabic consonant - assuming there was no audible schwa after it. Thinking that p couldn't be the nucleus of a syllable because it's not sufficiently sonorous, I said "no". Is this true, and if so, how would you transcribe it?

I agree that the answer is “no”, and I would transcribe it straightforwardly as p: for example, hæv əpˈteɪtəʊ.
In principle, I would say that plosives can never be syllabic. People have sometimes tried to claim that a very reduced form of thank you, kkju, has two syllables, the first consisting of an inaudible and unreleased syllabic k. And in Romanian and Japanese there are high vowels which become devoiced or disappear after voiceless plosives, giving contrasts such as Romanian lupi̥ vs. lupu̥ (or better lupʲ vs. lupʷ) — but I would say that these original disyllables were reduced to monosyllables by this process.
Nevertheless I agree that there is a fine line between devoicing a vowel and deleting it entirely, perhaps leaving behind a secondary articulation on the preceding consonant.
I have often had to correct beginners who wanted to transcribe wanted, for example, as ˈwɒntd̩, imagining that the final sequence ɪd or əd was a syllabic d. But it isn’t.

reflexes of Greek ου

I realize that in my discussion of Utopia I didn’t make my reasoning altogether clear (as Jack Windsor Lewis complains in his own blog). I said

Etymologically, More built his word from the Greek οὐ ū ‘not’ plus τόπ(ος) tóp(os) ‘place’ (as in topic, isotope).

but then continued

Theoretically it ought to be pronounced with uː

— without explaining why.
My thinking was that the Greek vowel ου ū normally maps onto English uː, as in the word acoustic, which is from Greek ἀκουστικός akūstikós (ἀκούω akūō ‘I hear’) and gives English əˈkuːstɪk, not *əˈkjuːstɪk.
I realize now, though, that there are a number of other Greek words with ου ū in which this vowel is indeed mapped onto juː, among them Muse (Greek Μοῦσα Mūsa). In Uranus (Greek Οὐρανός Uranos) the following r diphthongizes the vowel to jʊə or reduced ju; in Luke Λοῦκας and anacoluthon ἀνακόλουθον the yod is dropped because of the preceding liquid.
So it is actually acoustic that is the odd man out. As the OED comments,

The reg. Eng. representative of the Gr. would be acustic

in which case I suppose we would pronounce it with a yod. (Actually, in the 1930s the BBC Pronunciation Advisory Council was debating whether to recommend the pron əˈkaʊstɪk, still recorded in the 1963 EPD.)
Likewise, in French it would regularly have been acustique, with y, rather than acoustique with u.
There is another Greek ου word which behaves very strangely in English, namely nous (Greek νοῦς nūs), which we in Britain pronounce naʊs, though the Americans say nuːs.
_ _ _
I shall now stop posting to the old site: future blog entries will appear only here. You will see above that I have now discovered how to force my choice of font for the phonetic symbols here, just as on the old site.
I will continue to post new entries between 21:00 and midnight each Sunday to Thursday, as before; but since the date is inserted automatically by blogspot they will no longer be postdated to the following day.

Canaan and Sinai

Apropos of naïve (yesterday’s blog), Jérôme Poirrier writes:

I am under the impression that a lot of people make the ɑː to aɪ change also in a word like Naomi (as naɪˈəʊmi). As a foreigner, I am unsure of that, since there are also a great many people who do not resort to a foreign value for the a and say neɪˈəʊmi.
The corpus of such words in everyday vocabulary is extremely low; but more data should be gathered by listening to the way anglophone media pronounce chains of vowels in Hawaiian and Polynesian proper names...

OK, let it be gathered. Who’s first?

Some people may be surprised to hear this, but the traditional RP form of Naomi has initial stress, ˈneɪə(ʊ)mi. You won’t find anything else in Daniel Jones. As far as I can see, Jack Windsor Lewis was the first to record the penultimate-stressed form, in his CPD (1972). So for traditional RP (and for me), Naomi does not illustrate the point at issue, which is the unstressed prevocalic vowel(s) spelt a.

There are, however, two Biblical names that are perhaps relevant: Canaan and Sinai. These words each have initial stress. For foreign words of this vintage, long a would be expected to be read as eɪ (as in Amos, Salem, Jacob, Emmaus), not as ɑː.

Here’s the twelfth edition of EPD (Daniel Jones’s last, 1963). Focus on the penultimate vowels Canaan, Sinai.


What we find in both of these cases is a tendency to reduce weak eɪ to i, and perhaps further to j or zero.

(Who would believe that Sinai used to rhyme with tiny?)

O come, O come

It’s the wrong time of year to be asking this, but have you noticed how some people pronounce Israel as ˈɪzraɪel when singing? The hymn that I notice it in is the the Advent hymn

O come, O come, Emmanuel
And ransom captive Israel!

with its refrain

Rejoice! Rejoice! Emmanuel
shall come to thee, O Israel.

In LPD I said that Israel, in speech normally ˈɪzreɪl or ˈɪzriəl, is “in singing usually ˈɪzreɪel”.
But on reflection I certainly ought to have mentioned this further possibility with -aɪ-.

Listen, for example, to the Irish singer Enya (using slightly different wording from the usual Anglican one):

Where on earth could this treatment of the second vowel have come from? It’s not straightforwardly based on the spelling; there are all sorts of vowel sounds that correspond to the spelling a, but aɪ is not one of them.

On reflection I think that we have a tendency (perhaps ‘rule’ would be to put it too strongly) to change ɑː to aɪ before a following front vowel (a position from which it is usually shielded by a linking or intrusive r). It’s a kind of anticipatory articulation.
We see this in the word naïve. On the basis of the French it ought to be nɑːˈiːv. In practice people mostly say naɪˈiːv.
If we assume a starting point ˈɪzrɑːel, based on the spelling or the Latin or Hebrew pronunciation (real or imagined), then my proposed near-rule would make it ˈɪzraɪel. QED.
(Or is Enya actually singing her version of ˈɪzrɑːel?)

FURTHER TEST
Table (second attempt, successful)

This is column one.	This is column two.
line 1	line 1
line two	line two
laɪn θriː	laɪn θriː

utopia and dystopia

As any fule kno, the word Utopia was invented by Sir Thomas More in his book of the same name, published in 1516.

A fruteful and pleasaunt Worke of the beste state of a publyque weale, and of the newe yle called Utopia; written in Latine by Syr Thomas More knyght.

More’s imaginary land is depicted as enjoying a perfect social, legal, and political system. Later generations have extended the meaning of the word to cover any ideally perfect country or situation.

Etymologically, More built his word from the Greek οὐ ū ‘not’ plus τόπ(ος) tóp(os) ‘place’ (as in topic, isotope).

Theoretically it ought to be pronounced with uː-. But in practice it is pronounced juː-, exactly as if it were the prefix eu ‘good’, Greek εὐ. According to the OED, More himself made a pun upon this.

Vtopia priscis dicta ob infrequentiam, Nunc ciuitatis æmula Platonicæ..Eutopia merito sum vocanda nomine.

Be that as it may, it is as utopia that we know this word.

Words with eu- can have regular antonyms with dys-, from the Greek δυσ- ‘bad’. Thus we have eupeptic - dyspeptic and euphoria - dysphoria. Hence we get the relatively modern coinage dystopia, an imaginary place or condition in which everything is as bad as possible, which the OED dates to John Stuart Mill in 1868.

If the prefixes u- and eu- were not pronounced identically, we would not have had the irregular pair utopia - dystopia.
_____

TESTS OF SYMBOLS AND LAYOUT
IPA
ɪf ˈɔːl ðə ˈwɜːld wə ˈpeɪpər
ən ˈɔːl ðə ˈsiː wər ˈɪŋk
ən ˈɔːl ðə ˈtriːz wə ˈbred n ˈtʃiːz
wɒt ˈʃʊd wi ˈhæv tə ˈdrɪŋk

IPA diacritics n̩ n̥ l̩ l̥ m̩ m̥ (very poorly aligned)

Polytonic Greek
Μῆνιν ἄειδε, θεά, Πηληιάδεω Ἀχιλῆος
οὐλομένην, ἣ μυρί’ Ἀχαιοῖς ἄλγε’ ἔθηκε

Kanji 出口 入口

Arabic عيدالفطر ‘īdu l-fiṭr

Table (doesn't work)
This is column one. This is column two.
line 1 line 1
line two line two
laɪn θriː laɪn θriː

mirror, mirror

Over the nearly three years that I have been doing this blog on the UCL site quite a few people have complained to me about two points:
* that they cannot leave comments on the blog (the only thing they can do is to email me, with a comment that I might or might not pass on); and
* that the RSS feed either doesn’t work or isn’t very good.
These are consequences of the fact that the UCL departmental server doesn’t offer these facilities.
So for all this week I shall be mirroring the blog here, on Google’s dedicated blog server. You can find us not only on the UCL site but also here, and on this site you can, if you wish,
* leave comments (provided you register first, with a nickname if you prefer anonymity), and
* enrol as a “follower”, which means you will have a page on which you will be alerted each time there is a new posting on my blog or on others that you follow, with the title and first few lines of each posting.
For me the blogspot site will also have the advantage of automatic archiving rather than the hand-coded archiving that I have been doing till now.

There are several things we need to check with the new site. You can help by telling me whether they work for you.
* Most importantly, IPA symbols. I am not sure how blogspot handles fonts for Unicode characters not included in the basic font, which is Verdana. Perhaps it depends on the browser or the fonts installed in the system. The IPA symbols come out satisfactorily on my screen. Do they on yours? I have also (on my other blog, which has always been on blogspot) used Cyrillic, Hebrew, Japanese katakana and Korean hangŭl, all of which come out OK. I haven't yet tried polytonic Greek or Chinese.
* There are several details of layout I haven’t yet tested on the new site. In particular, I am not sure how it will cope with tables (which also allow one to have more than one column of text, for example).
* The blogspot editor makes it easy to embed links, images, and videos. I must see how it copes with sound files. I haven’t yet found whether it is possible to link to an image, and if so how to do so, other than messing around with the HTML by hand.

If everything turns out to work satisfactorily on this blogspot site, I may transfer to it entirely in a week or two. If it proves unsatisfactory, I won’t. If I do, I’ll put a redirect on the old url, of course.

zhoozh

Fired by the discussion of initial ʃt, Harry Campbell mentions an even stranger recent acquisition: a word we can agree neither how to spell nor how to pronounce: but let’s list it as ʒʊʒ zhoozh, as in to zhoozh something up, meaning to make more attractive, smarter, more exciting, to jazz it up.
The OED gives only the pronunciations ʒʊʃ and ʒuːʃ and the spellings zhoosh and zhush. Someone ought to tell the OED that many, perhaps most, of the people who use this word pronounce it with a final voiced consonant, ʒ. And I am not sure that I have ever heard it pronounced with uː rather than ʊ. I think the usual pronunciation is indeed ʒʊʒ, which twice violates the usual phonotactic constraints on ʒ, a consonant usually confined in English to intervocalic position, as in pleasure ˈpleʒə.
Although I know this word passively, it is not one I would actively use myself. Stylistically it strikes me as not just slang but camp slang (and I may be gay but I have never been camp). Indeed, the OED’s first citation (1977) is from Gay News, from a sentence which is written entirely in Polari, and which sounds as if it is a quote from Julian and Sandy in Round the Horne.

As feely homies..we would zhoosh our riahs, powder our eeks, climb into our bona new drag, don our batts and troll off to some bona bijou bar.

According to the Wikipedia article on Polari,

"Zhoosh" has entered English more recently, especially through the TV series Queer Eye for the Straight Guy. Its initial consonant, unique in that position in English, has led new users to generate variant spellings such as "zoosh", "soozh", "tszuj" etc. The word begins and ends with the same phoneme, the "zh" sound as in the word "measure".

...which I agree with.
In a letter to the Guardian in 2003, W. Stephen Gilbert says

you might zhush up a tired salad by adding some garnish, or stick some zhush in an article for the Guardian by adding a couple of dubious jokes.

But if he were to use that word in an article for the paper

some po-faced sub [would] remove it on the naff grounds it wasn't in the desk dictionary.

It clearly ought to be. Sorry I didn’t pick it up in time to make it into the third edition of LPD.