Wednesday, 30 September 2009

Eye-rack and Eye-ran

Here’s a nice little video from YouTube. It’s an advertisement for CNN, in which Christine Amanpour purports to attempt to teach the approved pronunciation of Iraq(i) and Iran to a would-be presenter who wants to say them with /aɪ-/.

The last two seconds of the clip is nice, too. We’re supposed to say /tʃetʃˈnjɑː/.

While on the subject of mispronunciations, a few weeks ago I saw the former US presidential candidate John McCain pronouncing infectious as /ɪnˈfektʃuəs/. Put that in the same box as prenuptial (blog, 9 June 2006) and rumbustious (blog, 27 Nov 2008).

Tuesday, 29 September 2009

irritating hamburgers

Emilio Marquez writes
I would like to ask you about the rhythmic patterns of … words [such as irritating].
If Spanish speakers are asked to place the primary stress in the correct place…, they will invariably make the last syllable "pulse[d]". If they are asked to pulse lightly on any vowel between the primary stress and the last syllable, they will turn this light pulse into the primary stress of the word, i.e. they will say either IRritaTING or irriTAting, but not IRriTAting).
1) Do "worshiper" and "worshipful" resemble "hamburger", or perhaps "educate"?
2) Which are the strongest (potentially "pulsed") "sit" or "me" vowels in the words "qualificative", "prejudices", "accompanying"?
3) What is the rhythmic pattern of the word "participle" when this shows initial primary stress?

This question relates to possible posttonic secondary stress, which is bound up with the important distinction in English between strong and weak vowels.
To recap, the weak vowels in English are i, u, and ə. The strong vowels are all the rest. (Exceptionally, ɪ and ʊ can be either strong or weak. Under ə we include all the syllabic consonants, including AmE ɚ.) In irritating ˈɪrɪteɪtɪŋ the vowels are strong, weak, strong, weak respectively.
Because of this, native speakers tend to perceive the penultimate syllable, teɪt, as being more strongly ‘stressed’ than the final syllable ɪŋ. But what they want to call ‘stress’ is arguably no more than a way of saying that the vowel is one of the strong ones.
Actual rhythmic beats following the main word stress accent are all pretty optional, which is why the British tradition is not to show any secondary stress in words like this: ˈɪrɪteɪtɪŋ, not *ˈɪrɪˌteɪtɪŋ. The alternative tradition, usually followed in the States and (for example) Japan, is to recognize a secondary stress on the penultimate, írritàting.

To answer the specific questions:
(1) Introspectively, worship(p)er (ˈwɜːʃɪpə, strong-weak-weak) is rhythmically different from hamburger (ˈhæm(ˌ)bɜːɡə, strong-strong-weak). In hamburger the penultimate vowel is long; in worship(p)er it is short. Both are also different from educate (ˈedjukeɪt or ˈedʒəkeɪt, strong-weak-strong). The last vowel of worshipful can be pronounced strong or weak, so this one can go either way, like worshipper or like educate.
(2) Qualificative is a pretty rare word. I would pronounce it ˈkwɒlɪfɪkətɪv (strong-weak-weak-weak-weak). Such a long string of weak vowels is unusual, and some people avoid it by saying ˈkwɒlɪfɪkeɪtɪv (strong-weak-weak-strong-weak). Prejudices is ˈpredʒudɪsɪz (s-w-w-w). Accompanying is əˈkʌmpəniɪŋ (w-s-w-w-w).
(3) As you are aware, people are divided about participle. I give it initial stress and say ˈpɑːt(ɪ)sɪpl (s-(w-)w-w).

However… despite all this apparent hair-splitting, my advice to Spanish-speaking EFL learners is not to attempt any degree of stress after the main stress in a word. Ignore all posttonic stresses. This applies even in compounds like washing machine ˈwɒʃɪŋ məˌʃiːn. As Emilio is aware, attempts at a conscious secondary stress are likely to lead only to mispronunciations of the type ˌwɒʃɪŋ məˈʃiːn.

Monday, 28 September 2009

interpreting stress

There are two English words beginning inter- which have an unexpected stressing: interpret and interstice. Looking at the spelling you might expect *ˈɪntəpret and *ˈɪntəstɪs. What we actually have is ɪnˈtɜːprɪt and ɪnˈtɜːstɪs.

I am not clear what the historical reason for these oddities is. Consider first interpret: nearly all other trisyllabic verbs in inter- have final stress: interact, interbreed, intercede, intercept, interdict, interfere, interject, interlace, interleave, interlink, interlock, intermix, interpose, interrupt, intersperse, intertwine, intervene, interweave. The only other exceptions seem to be the initial-stressed interest and interview: both are presumably to be explained as denominative.
The best explanation that I can come up with for interpret is that the word may originally have been borrowed as a noun, from Latin interprēs, interprĕt-. Then, if we apply the SPE main stress rule for nouns, we disregard the final vowel because it is short (in the Latin oblique stem) and place the main stress on the preceding vowel, followed as it is by a strong cluster (-rpr-). This stressing is unaffected when we convert it to a verb.
Unfortunately the OED shows the verb interpret as first attested in English in 1382, two hundred years before the noun interpret ‘interpreter’, now obsolete, which has only a single citation, from 1585.

Turning to interstice: the OED gives an alternative stressing (initial), but eveyone else shows it only with penultimate stress. All other trisyllabic nouns in inter- have initial stress: interchange, intercourse, interface, interlude, internet, interplay, Interpol, interval, interview.
My first thought on interstice was that it is what we get if we apply the SPE main stress rule for nouns (and this one really is a noun). Again, we disregard the final short vowel and place the main stress on the preceding vowel with its following strong cluster (-rst-).
The only problem with this solution is that it doesn’t explain why the same rule doesn’t apply to internet and interval, which by this logic ought to be ɪnˈtɜːnɪt and ɪnˈtɜːvəl. (They are the only nouns in our list with a short vowel in the final syllable.)
As so often with Chomsky and Halle, you find that the rules work beautifully some of the time, but you have to tie yourself in knots to explain why they don’t apply to all cases that apparently meet the structural description.

I’m not going to mention interment, in which there is an etymologically different inter, not the prefix but a different prefix (in-) plus stem (-ter-).

Friday, 25 September 2009


In response to Wednesday’s blog Eric Armstrong wrote:
Today you said
"If it is essential to symbolize the central quality explicitly, then we have diacritics available: [ä] or [ɐ̞] or [ɑ̈]. But it’s better to state such details once and for all in the transcriptional conventions, not repeatedly in a transcribed text."

My question for you is: does [ä] = [ɑ̈] ? Do those modified-by-a-diacritic symbols represent the same thing, meeting as they might in the middle? To me, I've always assumed that [ä] is more front, while [ɑ̈] is more back, that they *don't* represent the middle of the bottom of the chart. But that's just my assumption. What's the convention? Or am I confusing [a̙] and [ɑ̘] with those centralized versions, which should really be in the centre?
Or should I say [a̠] and [ɑ̟]? (that's a minus under the first and a plus under the second, if they didn't make it through the email unscathed.) Can you help clarify? I bet other blog readers might like to know...

The 1989 IPA Convention in Kiel that I mentioned yesterday also attempted to clarify what was meant by “centralization”. There it was decided that the superposed umlaut means centralization in the sense of moving horizontally towards the centre line of the vowel diagram.
But the IPA never attempted to define how much centralization is involved. Common sense suggests that [ä] would generally be still somewhat front of the centre line, and [ɑ̈] somewhat back of it. They would straddle the proposed [A].

A new diacritic, the superposed small ×, was introduced to indicate a “mid-centralized” vowel, i.e. one that was modified towards [ə] — not only fronter/backer than implied by the base vowel, but also opener/closer. So the current official IPA doctrine is that [u] with the × diacritic means the same as [ʊ], whereas [ü] shows modification merely in the direction of [ʉ].
I have to say, however, that I have never seen this “mid-centralization” diacritic in actual use. Has anyone?
And I notice my browser won’t render it properly.

Thursday, 24 September 2009

two symbols frequently confused

I have just had to remind someone not to confuse the phonetic symbols ɤ and ɣ.
The first, ɤ, is the symbol for a back close-mid unrounded vowel, cardinal 15. This is the vowel heard in Mandarin Chinese 刻 [kɤ] ‘carve’.
The second, ɣ, is the symbol for a voiced velar fricative. This, or the corresponding approximant, is the consonant heard in the middle of Spanish fuego [ˈfweɣo] ‘fire’, Greek εγώ [eˈɣo] ‘I’, etc.

Confusion of these two symbols was something I often had to correct in authors’ manuscripts when I was the editor of the Journal of the International Phonetic Association. Since other copy editors may not have been so symbol-obsessed as me, the confusion is found uncorrected in quite a few printed books.

At its 1989 Kiel Convention the IPA discussed this issue. At the time the vowel symbol was usually printed with straight sides (although on the line, x-height), making it very similar to the consonant symbol (which descended through the line). See this scan of the IPA Principles booklet (1949 edition). The recommendation of the Kiel Convention was to change the sides into a curly “rams-horn” shape, which is what we use today.


If you have installed Doulos SIL, you should be able to see what they look like in a serifed font:

Wednesday, 23 September 2009

German [a(ː)]

Lipman asked (blog comments, 21 Sep):
About [ɐ] - what would you say is the first vowel in German aber? The word is usually transcribed as [a:bɐ], but that [a] is probably a convention as well - it's not identical to mainstream English or even Italian /a/. I'd say it's in between [a] and [ɑ]. Northern German as [a] and South Eastern German has [ɑ], both of which are different from the vowel in Standard German.

I agree. I think the German vowel in aber is open, unlike the not-fully-open [ɐ] of STRUT, and is backer than cardinal 4 [a] but fronter than cardinal 5 [ɑ]: in fact, just about halfway between cardinal 4 [a] and cardinal 5 [ɑ].
Here is Mangold’s vowel chart for the oral monophthongs of Standard German (Dudenaussprachewörterbuch6, p. 37).

Some people want the IPA to approve the use of small-cap [A] for this quality. I do not favour this proposal, because no language as far as I am aware distinguishes three fully open unrounded vowels, [a] vs. [A] vs. [ɑ]. Just as the symbol [t] has to represent sounds that may be aspirated or unaspirated, dental or alveolar, according to the language, so we must demand flexibility in vowel symbols such as [a].
If it is essential to symbolize the central quality explicitly, then we have diacritics available: [ä] or [ɐ̞] or [ɑ̈]. But it’s better to state such details once and for all in the transcriptional conventions, not repeatedly in a transcribed text.

Tuesday, 22 September 2009

O to be a polyglot

Here is an impressive video of an Englishman going by the name of Torbyrne, who can speak sixteen languages. (Thanks to Giridhar Rao for this.) He says
This is a video with me speaking some of the languages I have learnt over the years with captions in English. In this video I speak English, French, Spanish, Welsh, German, Macedonian, Swedish, Italian, Serbian/Bosnian/Croatian, Portuguese, Czech, Catalan, Russian, Dutch, Romanian and Albanian.

As far as I can judge from the languages I know myself, his pronunciation is excellent, in that he definitely achieves the goal of sounding native-like. However his general ability to speak clearly and articulately does arguably leave room for improvement.
Nevertheless, this should be an inspiration to us all: particularly on the day that the Guardian reports on the closure of UK university language departments.
UCL’s professor of German says
Monolingualism radically diminishes Britain's ability to compete in the international marketplace and disqualifies the British from many high-level posts that require linguistic fluency. It threatens our ability to look beyond our front doors. Foreign culture can only truly be accessed through a foreign language. Not having that exposure results in an inability to be a global citizen and limits otherwise intelligent people to cultural parochialism. At that point, we have abandoned, paradoxically in the age of globalisation, the desire to reach a certain level of intellectual development and the wish to truly count as world citizens.

Monday, 21 September 2009

ən əˈnʌðə θɪŋ

Serkan Yanikoğlu asks
What is the difference between the vowel in the word "cup" and the first vowel in the first syllable of the word "about" in terms of phonetics? Why do they have to be shown in different symbols? Do they sound the same?
If we are referring to RP and the kind of pronunciation usually taught to EFL learners who take BrE as their model, then the vowel of cup (aka the STRUT vowel) is considerable more open than that in the first syllable of about (aka “schwa”). That is, it is more similar to Turkish [ɑ].
They are represented by different phonetic symbols, [ʌ] and [ə] respectively, because they sound different and, more importantly, because there are a few pairs of words or phrases in which the difference is important for the meaning. For example, unorthodoxy [ʌnˈɔːθədɒksi] sounds different from an orthodoxy [ənˈɔːθədɒksi]. A hubbub is always a [ˈhʌbʌb], never a *[ˈhʌbəb]. A large untidy room [ə ˈlɑːdʒ ʌnˈtaɪdi ˈruːm] is different from a large and tidy room [ə ˈlɑːdʒ ən ˈtaɪdi ˈruːm]
There are only a few such cases, though, because the STRUT vowel belongs to the strong vowel system and is often stressed, while schwa belongs to the weak vowel system and is never stressed.
In other varieties of English the picture is different. There are many English and Welsh people who pronounce the STRUT vowel pretty much the same as schwa. So do many, perhaps most, Americans. For them, we could reasonably write both vowels as [ə]. So it would be reasonable for EFL learners not to worry too much about the contrast. Millions of NSs manage without it.
Your textbooks will tell you to pronounce above as [əˈbʌv] and another as [əˈnʌðə]. But millions of people in Birmingham, Swansea or Miami say [əˈbəv] and [əˈnəðə(r)]. You can too, if that’s easier.
But not if you want to speak RP.

Friday, 18 September 2009

More Latin

Jason Nedecky writes further:
You mention Westminster School held on to the older Anglo Latin
pronunciation. From Wikipedia on the coronation of a British monarch:
"King's or Queen's Scholars of Westminster School can exercise their
right to be the first commoners to acclaim the sovereign, shouting
their traditional "vivat"s as he or she enters the coronation theatre."
In 1953, those Westminster scholars still said [vaɪ væt]. I suppose
they were holding on to that pre-1900 pronunciation.
I think that’s traditional for everybody. See LPD.
In a somewhat-related idea: the capital city of the province of
Saskatchewan is only pronounced [ɹə ʤaɪ nə] in Canada.
While an Anglo Latin exists in the Church of England in those
canticles, it it interesting to note that one hears only that Italianate "Church" Latin in full choral works of English composers. I am thinking of Vaughan Williams Mass in G minor and Dona Nobis Pacem, Elgar's Ave Verum corpus, Britten's War Requiem, and the list goes on. …And I am speaking of BrE recordings here (King's College, LSO, David Willcocks, Adrian Boult, etc.) Maybe this is along the lines of
Paul's comment about his experience of a more Italianate Latin learned
in his school.
Someone commented on the Orff piece. It is my experience that because of the mixture it has of Latin with older German and French, one has to decide at the outset what system(s) to employ. Even then, it does not always work — I am thinking particularly of "iam amore virginali totus ardeo" in which the soprano and later baritone sings with the "Ragazzi." So often the soloists have already opted for a German [vir gi na li] and the boys come in to rehearsal having learned [vir ʤi na li], and it's back the drawing board.

We might also mention the Agnus Dei, which for me at school was ˈæɡnəs ˈdeɪiː, but which one often nowadays hears sung as ˈænjʊs-, with a clearly Italian-influenced treatment of gn. (Again, see LPD. EPD doesn’t mention this possibility, though.)

Thursday, 17 September 2009

weak ɪ

Giuseppe Macario wrote
Although /'pɜ:fɪkt/ and /'ba:skɪt/ is what I have found in my British
dictionaries, I get the impression that the vowel pronounced by most
native speakers of English, no matter where they live, is not the /ɪ/.
The first vowel in /ɪk'spres/ confuses me, too. Are my doubts

I think the point here is that English has two separate vowel systems, the strong system and the weak system. The weak vowels are much less tightly defined than the stressed ones. Although people equate the weak vowel of 'perfect' and 'basket' with /ɪ/, it may vary towards [ə] more than a stressed /ɪ/ would.

The Longman Dictionary of Contemporary English uses a special double-decker symbol to represent a weak vowel that may be [ɪ] or [ə] or something intermediate between them (see picture).
The word perfect is a special case, because in this word some speakers do not weaken the unstressed vowel, but keep it as /e/ (as in the corresponding verb, /pəˈfekt/, where the vowel in question is stressed and therefore strong).
Nevertheless, my advice to Giuseppe is that he can safely treat the unstressed vowels in perfect (noun), basket, and similar words as being just like /ɪ/ (as in KIT). That’s how I believe I pronounce that vowel myself, and that’s what I hear millions of other people say.
A note of warning: Italian learners often have trouble perceiving and making the difference between English /iː/ (FLEECE) and /ɪ/ (KIT). The unstressed vowel we are discussing is certainly not like English /iː/.

Wednesday, 16 September 2009

Benedicite, omnia opera

Lipman’s comment on yesterday’s blog and the document he linked to, together with emails received, suggest that it would be helpful to go a bit further on the school pronunciation of Latin.
Jason Nedecky writes
Having sung a lot under BrE choral directors here at the Anglican Church of Canada, I have become somewhat accustomed to [anglicized Latin]; I am thinking particularly of those canticles of the Book of Common Prayer adopted for Mattins: the "Benedicite" [ˌbɛ nɪ ˈdaɪ sə tɪ], the "Jubilate Deo" [ˌʤuː bɪ ˈlɑː teɪ ˈdeɪ əʊ], the "Venite" [və ˈnaɪ tɪ], the "Te Deum" [ˈtiː dɪəm], and so on.
The other day, I was watching a documentary on the Coronation of Queen Elizabeth II, in which the famous anthem "I was glad," set by Parry contains that passage of "Vivat!" acclamations, adopted and Anglicised on that momentous occasion, as, from what I can hear: [ˈvaɪ væt ɹɪ ˈʤaɪ nə ɪ ˌlɪ zɪ ˈbeɪ θə].

The canticle names Jason cites are actually a mixture. Yes, at school we called the Benedicite ˌbenɪˈdaɪsəti (see LPD), but as a Latin word meaning ‘bless (ye)!’ we would have said ˌbenɪˈdiːkɪteɪ.

The Westminster School document referred to by Lipman describes the way English was pronounced in English schools up to about 1900. This way of pronouncing Latin derives from the reformed pronunciation introduced by Erasmus and his followers five hundred years ago. While this Erasmian reform removed some anomalies, its timing was unfortunate.
The advantages of the new pronunciation in England were soon to be diminished by an accident of linguistic history. For the reforms came at a time when the extensive changes from the Middle English to Modern English vowel system were still incomplete; and so any reforms in Latin or Greek pronunciation underwent these changes… the Latin vowels ā, ī, ē, for example, became diphthongs as in English name, wine, seen.
[Allen, Vox Latina, App. "The pronunciation of Latin in England"]

That is why we have aɪ rather than iː in the penultimate syllables of Benedicite and Venite as names of canticles, and similarly in hundreds of Latin loanwords in English, from appendicitis to virus.

Around 1870, however, a new reformed pronunciation was formulated by various Cambridge and Oxford scholars, and by the early twentieth century had spread to most British schools teaching Latin (though Westminster was a notorious exception). By the time I learnt Latin, then, the English Great Vowel Shift had been undone as far as English Latin was concerned (though we still continued to pronounce Latin entirely in English sounds). From the earlier ˈvaɪvæt rɪˈdʒaɪnə for vivat regina we had switched to ˈviːvæt reˈɡiːnə (and some brave souls even said ˈwiːwæt).
The earlier pronunciation has persisted in proper names and assimilated words, but has generally given way to the newly reformed pronunciation in actual Latin phrases and sentences.
Before 1870 I assume the Jubilate would have been called the ˌdʒuːbɪˈleɪti. In this case, however, for some reason the twentieth-century reforms have won out, and we all now say -ˈlɑːti (though we may hesitate between j- and dʒ-).

In my first year of Greek at school we still pronounced Greek with GVS-shifted vowels: so for example Μοῦσα ‘a Muse’ was ˈmaʊsə and τί ‘what’ was taɪ. But in my second year we switched to the newly reformed pronunciation and from then on said ˈmuːsə and tɪ.

With the influence of Catholic, Italianate pronunciation, Benedicite can now also be heard as ˌbeneˈdiːtʃiteɪ. But not in Anglican schools and churches.

Tuesday, 15 September 2009

in medias res

Consider these English words, all borrowed from Latin or from Greek via Latin. How is the final consonant pronounced?
-as: paterfamilias, Judas, atlas, Maecenas;
-es: tabes, faeces, Hades, bona fides, scabies, rabies, species, series, obsequies, isosceles, anopheles, herpes, lares et penates, Euripides, Sophocles, diabetes, Dives;
-is: cannabis, ibis, aegis, aphis, corydalis, chrysalis, annus mirabilis, syphilis, torticollis, acropolis, metropolis, epidermis, lenis, penis, Adonis, ex-libris, hubris, Berberis, iris, clitoris, basis, emphasis, psoriasis, oasis, stasis, exegesis, thesis, genesis, diaeresis, enuresis, crisis, thrombosis, silicosis, psychosis, osmosis, diagnosis, sclerosis, neurosis, halitosis, ellipsis, synopsis, catharsis, analysis, paralysis, appendicitis, bronchitis, mantis, glottis, axis, lexis;
-os: chaos, logos, pathos, ethos, thermos, cosmos, tripos, rhinoceros;
-us: syllabus, omnibus, incubus, abacus, diplodocus, focus, locus, discus, mucus, exodus, nucleus, Theseus, Zeus, sarcophagus, asparagus, fungus, typhus, polyanthus, ceanothus, radius, regius, genius, angelus, nautilus, phallus, gladiolus, calculus, tumulus, stylus, hippopotamus, isthmus, animus, humus, anus, tetanus, Venus, tetanus, genus, terminus, alumnus, campus, opus, Lazarus, virus, plesiosaurus, papyrus, Pegasus, Croesus, census, versus, narcissus, apparatus, status, f(o)etus, tinnitus, coitus, emeritus, nexus.
In the case of words ending in -as, -is, -os or -us, the final consonant is a voiceless /s/. But with those ending in -es it is a voiced /z/.
I have no idea why there is this difference. But it corresponds to how we pronounced Latin when I was at school. Dominus was ˈdɒmɪnəs, librīs was ˈlɪbriːs, but spēs was speɪz.
In classical times Latin s represented a voiceless sound in all positions (Allen, Vox Latina, p. 35): so there’s no justification there for our English voiced /z/ in series and diabetes. I wonder where it came from.

Monday, 14 September 2009

you ARE a dark horse!

Tami Date asked for my comments on the accent pattern of the underlined parts below.
(1) A: The Joneses have got a new car.
B: Have they, indeed. What model is it, dear?
A: Oh, I don’t know. It's a pale blue.
B: Well, well. Old Ted IS a dark horse. I was only talking cars with him the other night.
A: And he didn't mention it?
B: Not a word. (adapted from Jack Windsor Lewis phonetic blog #207)
(2) A: I've just got engaged to Sheila.
B: Well! You ARE a dark horse. (adapted from O'Connor and Arnold (1973: 220)

My immediate reaction was to suggest that ‘be a dark horse’ is an idiom with a fixed focus pattern, like ‘to have a bee in one’s bonnet’. A dark horse is someone who is not well known and who surprises people by winning a competition; hence, someone who acts secretively to surprise people. (LDOCE says this second sense in BrE only.)

Jack Windsor Lewis said he would hesitate to call it an idiom
partly because you can produce any number of such constructions ad lib and partly because an idiom is something I think of as not predictable by simple logic.

“Any number of such constructions?” I've been trying to think of some. I have to say I have not been very successful in this task. What we want is an accented verb to be followed by lexical material that is new in context but unaccented.
Instead of saying that someone was a dark horse you could say
That was a surprise.

If a child wants praise for some achievement you could say
You are a clever boy/girl.

Daddy will be pleased.

Compare also the possibility of placing the nucleus on an intensifying word, despite following new lexical material.
That’s very interesting.
I was extremely annoyed with them.
I know exactly what you mean.
[English Intonation, 3.33]

There’s also the clearly idiomatic expression
You are a one!
(LDOCE: BrEng, old-fashioned, to say that someone’s behaviour is amusing, strange, or surprising.)
I can’t think of any others. Jack?
All these, including you are a dark horse, are exclamations, and as such must have a falling tone of some kind (low fall, high fall, rise-fall).

Friday, 11 September 2009

But soft!

People who do not know phonetics often attempt to describe particular speech sounds by labelling them ‘soft’ or ‘hard’. The trouble is that these terms can mean quite different things to different people.

In Russian and other Slavonic languages they are well-established as meaning ‘palatalized’ and ‘non-palatalized’ respectively (which in Irish would be ‘slender’ and ‘broad’). But an English ‘soft’ C is one pronounced [s]. A Hiberno-English ‘soft’ t is a fricative.

Now consider this quote from the Wikipedia article entitled Tikka.
Tikka or Teeka or Teekka is the English transliteration for two entirely distinct Indian words: tikka with a soft initial 't' and tikka with a hard initial 't'. This often causes some confusion as to which "tikka" is meant.
Tikka with a soft initial 't' means a piece of meat, such as a cutlet.
Tikka pronounced with a hard 't' can mean a forehead mark or a needle.

So what is the phonetic difference?
After some research I was able to track down the two Hindi tikkas. The “soft” one, the one we know from chicken tikka masala (“Britain’s favourite dish”) is written टिक्का, transliterated as ṭikkā, IPA [ʈikkɑː].
The “hard” one appears to be a regional variant of the word usually written as tilak, the mark on the forehead of a Hindu.
In Nepal, Bihar and other regions, the tilak is called a tika (टिका)

…which in IPA is [ʈikɑː].
It would seem, then, that the difference is nothing to do with the t-sounds: both are retroflex and unaspirated. The difference is in the k-sounds. In one word the velar plosive is geminated; in the other, not.
I think that could have been more clearly expressed.

Thursday, 10 September 2009


I have recently been reading Pilcrow, by Adam Mars-Jones. It’s a novel in which the central character and narrator is a severely disabled boy. That may sound a bit dispiriting, but the novel is full of imaginative insights and empathy, and I enjoyed it greatly.
However at one point the author indulges in some phonetic terminology that doesn’t quite work. The narrator thinks that a fellow patient, Sarah, calls her mother Muzzie. But it turns out that in fact she calls her Mummy. The misapprehension arose because on one occasion she was being lifted back into her wheelchair, and her mother
must have been squeezing Sarah’s rib-cage in a funny sort of way, forcing the air out just as it got mixed up with an emotional sob … and then it emerged as a sort of distorted sigh. … the middle part of the word came out as a sighing ‘ZZzz’. With that extra squeeze of Sarah’s squeeze-box the consonant was mutated. A bilabial dental came out as a voiced alveolar fricative, and that’s how Muzzie got her name.

Oh dear. The ‘bilabial dental’ ought to be a ‘bilabial nasal’.
A real bilabial dental would be a simultaneously articulated [p͡t̪], [b͡d̪] or [m͡n̪]. (I’m afraid those overstruck symbols are going to constitute a severe test of your browser’s rendering capabilities.)
A pilcrow, by the way, is the sign ¶.

Wednesday, 9 September 2009

click symbols

Yesterday’s discussion of Zulu phonetics reminds me of the problem of how best to write clicks.
The click languages with the largest numbers of speakers are Zulu (with 10 million L1 speakers) and Xhosa (8 million). They both use three basic types of click (dental, alveolar/retroflex, and lateral), and have a long-standing orthographic tradition of spelling them as c, q, and x respectively.
The simple letters c, q, x represent the voiceless clicks, i.e. the clicks pronounced with accompanying [k]. They can also be aspirated (i.e. with [kh]), voiced (i.e. accompanied by depressor [g]), nasalized (accompanied by [ŋ]), or breathy-nasalized (accompanied by depressor breathy-voiced [ŋ(g)]). These are written respectively ch, qh, xh; gc, gq, gx; nc, nq, nx; ngc, ngq, ngx. (You can listen to all the series except the last pronounced by a native speaker here.)

Until 1989 these were written in IPA with the special symbols [ʇ], [ʗ], [ʖ]. I think these symbols may have been devised by Doke, whose Phonetics of the Zulu language came out in 1926.
Meanwhile missionaries and linguists dealing with the Khoi-San languages adopted a different notation system, writing [|], [!], [ǁ] for the clicks used in the Nguni languages (Zulu, Xhosa etc.) plus [ʘ] and [ǂ] for bilabial and ‘palatal’ clicks found only in the Khoi-San languages. These symbols were designed so as to be typable on the typewriters of the day: instead of | and ǁ people mostly wrote / and //, using the backspace key to create ! out of an apostrophe and a full stop, and ǂ out of the equals sign and /.
So as of early 1989 we had three notation systems. But at its Kiel Convention in 1989 the IPA, at the instigation of Peter Ladefoged, decided to go over to the Khoisanist symbols, which are what you will now find in IPA publications.
Wikipedia somewhat contemptuously remarks
At one time, the IPA was augmented with a set of Latin-based symbols for clicks, but they were never much used, and were eventually given up for the Khoisanist symbols.
However, it is not true that “they were never much used”. Generations of phonetics students from the twenties to the eighties learnt and used them (as I did myself). Everyone qualifying as a speech therapist in Britain had to know them. I myself must have taught them personally to close on a thousand students. Conservatively, I would estimate that at least 25,000 people were taught them, many more than ever worked with Khoisan languages.
I think the IPA was misguided in making this change. The old symbols were better.
Just look at these screenshots (from Wikipedia).
As you can see, the symbol for the dental click, the pipe |, can look dangerously like a sanserif lower-case l. The symbol for the lateral, the double pipe ǁ, can look like double lower-case ll. And both are also widely used for a quite different purpose, namely as intonation phrase boundary symbols.
No such danger with ʇ and ʖ.
Bring back the old IPA click symbols!
(Oh, and that’s an exclamation mark at the end, not a retroflex click.)

Tuesday, 8 September 2009

Thula sizwe

For their next Christmas show a choir I belong to is doing an intriguing song called Hope for Resolution. This involves the weaving together of an English hymn, Of the Father's love begotten and a Zulu song, Thula sizwe (‘Be still, my country’).

Many years ago I attended an elementary course of Zulu at SOAS, which included a detailed account of the phonetics. Zulu is a wonderful language for exemplifying speech sounds with different airstream mechanisms, since its sound system includes not only ejectives and implosives, but also clicks — plain, aspirated, voiced, nasalized, and breathy-nasalized. And there are alveolar lateral fricatives and affricates and even a velar lateral affricate ejective.

The words of the song, repeated several times in different combinations, are
Thula sizwe, ungabokhala,
uJehova wakho uzokunqobela.
Inkululeko, sizoyithola.

which means (as far as I can tell)
Be still, country, do not cry,
Jehovah will conquer for you.
Freedom, we will find/get it.

and is authentically pronounced (I think)

ˈtʰuːla ˈsiːzwɛ uŋɡaɓɔˈkʰaːla
udʒɛˈhɔːva waːkʰɔ uzɔɠuˈŋǃɔɓɛːla
iŋkʼuluˈlɛːɠɔ sizɔjiˈtʰɔːla

(ɓ, ɠ = voiced implosives, i.e. with ingressive glottalic airstream; = ejective, i.e. with egressive glottalic airstream; ŋǃ = voiced nasalized alveolar (retroflex) click, i.e. with an egressive pulmonic airstream for the ŋ component and a velaric ingressive airstream for the click component)

I haven’t been able to check this with a native speaker, so if anyone knows better please do not hesitate to tell me.
Zulu has high, low, and falling tones, but with the exception of thúla, -khâla and -thóla I do not know what the tones in these words are.

Zulu morphology is highly agglutinative. For example, the word uzokunqobela (‘he will conquer for you’), is made up of u- (subject concord), -zo- (future tense marker), -ku- (object concord), -nqob- (verb root, ‘conquer’), -el- (“applied” suffix, do something on behalf of someone else), -a (ending).

Certain Zulu voiced consonants, known as depressor consonants, have a lowering effect on the tone of the following vowel. Among them are the pulmonic-air voiced obstruents, but not the voiced implosives. This is relevant to the perception of the phonemic contrast between plain voiced plosives and implosives. The implosives are only very mildly implosive and the main auditory difference between them and the plain voiced plosives is that plain [b] and [ɡ] are depressors, while weak-implosive [ɓ] and [ɠ] are not. In singing, though, such tonal subtleties are naturally lost.

Our choirmaster has taught us to sing
tula sizwe uŋgabokala
ujehova wako uzokuŋkobela
iŋkululeko sizojitola
which seems close enough.
But I don’t think I shall be able to resist defiantly pronouncing the nasalized click in its proper place in uzokunqobela.

Monday, 7 September 2009

ask your favour

A very expert non-native speaker of English sent me an email beginning I want to ask your favour.
There’s no problem in understanding the phrase: after all, it sounds like what we say in English. But it's not actually what we write. We write (and think that we say) I want to ask you a favour.
It's easy to see how spoken you a can come to be reinterpreted as your. From strong juː eɪ we derive weak ju ə, readily compressed to one syllable as jʊə. And jʊə is how some British people (25% of them in my preference poll) pronounce your.
Those who, like me, pronounce your as jɔː, or weaken it to , would leave no room for confusion with you a.
— — —

There’s a job going for a Pronunciation Linguist at the BBC Pronunciation Unit. Anyone interested? Details here.
The job demands, inter alia,

• A thorough and practical knowledge of the IPA and its application in phonemic and phonetic transcription of all languages (not just English).
• A university degree in (a) linguistics or phonetics, or (b) modern languages or English with a significant linguistics component, including phonetics.
• Near-native fluency in at least one language besides English, and preferably more than one.

Friday, 4 September 2009

San Telmo

In Buenos Aires there is a barrio (neighbourhood) called San Telmo, where last week I visited an interesting open-air art and antiques market. (Who on earth buys old soda syphons? There were two stalls specializing in them alone.)
The name San Telmo naturally reminded me of the English St Elmo, patron saint of sailors, martyr, died c 303, formally known as St Erasmus of Formiae.
We all know him because of St Elmo’s fire, “an electrical weather phenomenon in which luminous plasma is created by a coronal discharge originating from a grounded object in an atmospheric electric field (such as those generated by thunderstorms or thunderstorms created by a volcanic explosion). … The phenomenon sometimes appeared on ships at sea during thunderstorms, and was regarded by sailors with religious awe, accounting for the name.” [Wikipedia].

Are Telmo and Elmo the same name? Are we talking about the same saint? Has Spanish just moved the final t of san(c)t- into the following syllable, rather as English and French moved the initial n of Arabic nāranj into the preceding article to leave orange, or OE nædre became modern English (an) adder? (There’s a name for this process, but for the moment I can’t recall it.)
Probably yes, it’s the same name: but probably not the same saint. According to Wikipedia, the Spanish “Saint Telmo or Saint Elmo” was a priest born in 1190 in Spain, never actually canonized though given the courtesy title of Saint.
Both Elmo and Telmo, I learn, are diminutives of Erasmus, though it’s not clear to me why or how the latter became the former.

Thursday, 3 September 2009

a period piece

Michael Ashby has kindly drawn my attention to a marvellous short film from the digital news archive of British Pathé.

In this five-minute film, dating from 1940, we see a Sinhalese student, played by one Dr M.D. Ratnasuriya, visiting SOAS to be taught proper English rhythm by the professor of phonetics, Arthur Lloyd James, played by himself (“Prof. A. Lloyd James, Secretary to the B.B.C. Committee on Spoken English”).

Today Lloyd James’s accent seems very dated: particularly in the TRAP vowel, that 1930s sort of tense [ɛ], in the words Jack and Paddington. He has a very open final happY vowel in carefully [ˈkɛəfʊlɪ] and immediately [ɪˈmiːdʒətlɪ]; and a correspondingly open final thankyOU vowel in will you [ˈwɪljʊ]. His THOUGHT vowel in jaw is opener and less rounded than is usual today. He uses a tapped r [ɾ] in intervocalic position, e.g. in very [ˈveɾɪ] and as linking r in your address [jʊəɾ əˈdres]. All of this is typical of the RP of that era.
Perhaps his braying NEAR vowel in the phrase what we hear [hjɐː] is to be explained as betraying his Welsh origins. His parents were actually Welsh-speaking. Collins and Mees comment
Audio recordings show that although he transformed his speech into something very close to RP, he nevertheless always retained faint traces of his native Welsh accent.
[The Real Professor Higgins. p. 275]

I’m not at all impressed with his imitation of American English.

By today’s standards his manner towards his student seems starchy and patronizing. But then today you wouldn’t see a newsvendor wearing a cloth cap and with a cigarette in his mouth.

Wednesday, 2 September 2009

muchas gratsias

Who pronounces the Spanish for ‘thank you’, gracias, as [ˈɡʁatsi̯as]?
Having flown to and from Argentina with the German airline Lufthansa, I can tell you: it’s Germans speaking Spanish as a foreign language.
In native-speaker Spanish, of course, the word is pronounced [ˈɡɾaθjas] or, in Latin America, [ˈɡɾasjas, ˈɡɾasjah].
This is a straightforward case of foreign learners being misled by the spelling. In German the letter c when followed by i or e stands for [ts], as in Circe [ˈtsɪrtsə]. So the mispronunciation [ˈɡʁatsi̯as] is a bit like French speakers of EFL saying structure with [y], using a vowel that plays no part in the English phonetic system but is what the French letter u typically stands for.

The most striking phonetic feature of Argentinian Spanish to my ears is the use of [ʃ, ʒ] for Spanish /j/, spelt y or ll. The voiceless variant is what one hears all around — for example, for Callao (street) you hear [kaˈʃao]. But all the Argentinians I spoke to claimed themselves always to use the voiced variant, [kaˈʒao], judged more elegant. That’s still not very like Castilian [kaˈʎao, kaˈjao].