Friday 16 March 2012

English places

Comments on yesterday’s blog addressed the fraught question of proper names in pronunciation dictionaries. I thought it might be useful if I tried to say what my policy was in LPD, at least as concerns place names in England.

I must confess that I did not set up a set of principles before starting work. Rather, what follows is a post-hoc attempt to express the principles I think I generally followed.

Let’s start from the difficult fact that in England everything is complicated by social class factors. A hundred years ago, certainly fifty years ago, and still to a large extent today, most English people spoke and speak with a local accent. Broadly speaking, the lower your social class, the more your pronunciation diverges from RP; the higher your social class, the closer to RP. Whereas RP speakers can be found in all parts of the country (or could when I were a lad, when not only the local landed gentry but also the vicar and the doctor probably spoke RP or something very close to it), “local” implies non-RP. The local accent typically includes various features that are regarded as non-standard and have traditionally been considered unworthy of mention in normative reference works such as dictionaries. (Note to nonNSs: when I were a lad is a stock phrase with non-standard were for was, used for comic effect.)

So let’s agree, for the purposes of argument, that most people who live in Hull call it ʊl. But in RP it’s unquestionably hʌl. We can leave it to the sociolinguists to determine the precise details of who uses which of these pronunciations and under what circumstances, and to what extent there are also intermediate forms such as hʊl, həl and perhaps also ʌl.

I think there is very general agreement in England that we don’t want dictionaries to tell us about h-dropped pronunciations. In any case, they can immediately be inferred once we know that in working-class England all hs are subject to being dropped.

I think it’s also generally agreed that we don’t want to be told about the use of ʊ in all names in which RP has ʌ; again, this can be inferred by rule, once we know that in the north of England the FOOT-STRUT split does not apply, which means that there is no /ʌ/ and which makes Hull a perfect rhyme for full (as against RP etc in which hʌl has a different vowel from fʊl).

Those of us who, like me, grew up in relatively high-status families in the north have been aware of this variability since childhood. In the Lancashire village where I lived until I was a teenager, our neighbours drank from kʊps and ˈdlasɪz (ˈɡlæsɪz); but in our house we had kʌps and ˈɡlɑːsɪz.

(As an aside, it wasn’t until I went to boarding school in the south that I discovered that fuck has ʌ. We didn’t swear in our family, and the only people I had heard using this word until then had pronounced it fʊk.)

The name of the village was Upholland. (See the photo. Nowadays people seem to write it as two words, Up Holland. In those days we didn’t.) Yes, our neighbours called it ʊpˈɒlənd. But we, of course, called it ʌpˈhɒlənd. It is the latter that you would expect a dictionary of place names to show; the former can be inferred from it, but not vice versa.

Our nearest town was Wigan. We called it ˈwɪɡən, although quite a few of its inhabitants called it ˈwɪɡɪn (which caused some amusement and was part of a comical exaggerated local accent).

So it is with the BATH words. Our neighbours might have a baθ (actually, many had just a zinc tub in the kitchen, brought out as needed); we had a bɑːθ. (Recall that a is the local form of the TRAP vowel, in RP classically æ.)

One of my uncles lived in Grasmere, in Cumbria. His wife, my aunt by marriage, also happened to be aunt to the Attenborough brothers, the actor Richard (Lord A.) and the TV naturalist David. (You’ll have heard how they speak: both are native RP speakers and grew up in Leicester, in the linguistic north.) They, and we, called the place ˈɡrɑːsmɪə. For many local people, though, the first syllable had the TRAP vowel.

That’s the background from which I felt confident in saying that the RP name of Castleford, Yorks., is ˈkɑːsl̩fəd, though the great majority of locals certainly give it the TRAP vowel in the first syllable. For Doncaster, on the other hand, all three possibilities -kəs-, -kɑːs-, -kæs- seem to me to be at home in RP, though probably in that order of preference.

Moving to London, I naturally gave the pronunciation of Wapping as ˈwɒpɪŋ, ignoring the ˈwɒpʔɪn you hear from many locals. Likewise Harlesden: ˈhɑːlzdən, not ˈɑːozdən.

And so to Newcastle. The RP form for all places with this name (in my view) is ˈnjuːˌkɑːsl̩. However, the size and importance of the city of Newcastle-upon-Tyne is such that I felt it appropriate to mention in LPD that locally it is -ˈkæsl̩, with a different stressing and different second vowel. (Here, of course, æ subsumes a.) There is a similar note at Carlisle.
_ _ _

This blog will now be suspended until the end of March. I may see some of you in Kyiv/Kiev, where I shall be speaking at a conference at Kyiv National Linguistic University and also visiting the Taras Shevchenko National University — or at BAAP in Leeds. Next posting: 2 April.

Thursday 15 March 2012

St Asaph

The formal status of “City” has just been conferred by the Queen on the village/town/city (pop. 3,500 or thereabouts) of St Asaph in north Wales.

On the ITV evening news yesterday the newsreader introducing a report about this pronounced the saint’s name as seɪnt æˈsæf. The reporter given the job of visiting the place and interviewing locals called it seɪnt ˈæsæf. But the local councillor being interviewed said sn̩t ˈæsəf, which is what you find in the BBC Pronouncing Dictionary of British Names as well as in LPD and in EPD (now CPD).

I think our pronunciation dictionaries are right to ignore guesses or ad-hoc spelling pronunciations used by people who are unfamiliar with a proper name and rather to show, or at least prioritize, the pronunciation used by those who are familiar with the name and mention it every day of their lives.

Asaph (or Asaff or Asaf or Asa) was a sixth-century bishop, the first bishop of the diocese that now bears his name. However the Welsh name of the city named after him is different: it is Llanelwy, church on the (river) Elwy.

One of the odd byproducts of my study of Welsh some 30-40 years ago is that I retain in my head numbers of paired English-Welsh placenames (Shrewsbury = Amwythig, Swansea = Abertawe, Bridgend = Pen-y-bont ar Ogwr, Milford Haven = Aberdaugleddyf etc., not to mention London = Llundain), though I do not recall ever having sat down consciously to commit them to memory. So when I heard this mention of St Asaph, I immediately thought “Llanelwy”, its Welsh name. Put it down to déformation professionelle professionnelle (= can’t help it, it’s my job).

Savouring the pronunciation of Llan- ɬan makes me think of voiceless alveolar lateral fricatives. Yesterday I attended the funeral of an elderly Montserratian at a church in northeast London. According to the service sheet, the minister in charge, whom I hadn’t met before, was the Rev Dr J. Zihle. The service lasted some two hours, and as I listened to him I thought about this surname. Given that he was a tall black man and sounded African, I thought it unlikely that it was the Czech name Žihle ˈʒiɦle with a lost diacritic. Nor could it be Icelandic, which has hl but no z. So when I spoke to him after the service, I asked him whether he pronounced his name ˈziːɬe. “Yes,” he answered, delighted: “you’re the first English person I’ve met who could pronounce it correctly”.

The voiceless alveolar lateral fricative ɬ, the sound that is spelt ll in Welsh, is spelt hl in the Nguni languages of southern Africa. Forty years ago I attended an introductory Zulu course at SOAS.

I asked him whether the name was Zulu, Ndebele, or Xhosa (pronouncing the last-mentioned with its proper aspirated voiceless lateral click). He confirmed that he was a Xhosa. Then he had to hurry off, so I wished him hamba kahle ˈhamba ˈɠaːɬe ‘go well’, which is Zulu rather than Xhosa, but he understood — Zulu and Xhosa are mutually intelligible — and gave the appropriate reply sala kakuhle ‘stay well’. More déformation professionelle professionnelle. Mustn’t show off. But can’t resist.

Wednesday 14 March 2012

no (audible) release

Plosives are traditionally analysed phonetically as having three stages or phases:
(i) the “approach” phase, during which the primary articulators move into contact;
(ii) the “hold” (or “compression”) phase, during which the contact is maintained, blocking the escape of the air stream, so that air pressure builds up behind the closure; and
(iii) the “release” phase, during which the primary articulators move apart, allowing the compressed air to escape explosively.

In the case of “nasal release” (as in the d of hardness) it is not the primary articulators that move apart, but the soft palate that moves away from the rear wall of the pharynx, allowing the compressed air to escape nasally; in the case of “lateral release” (sadly) it is the side rims of the tongue (or more generally the side part of the primary articulators) that move, allowing the compressed air to escape laterally. Logically, we can also speak of “nasal approach” and “lateral approach” for the variants of stage (i) in which it is not the primary articulators that move into contact, but the soft palate or side part of the articulator respectively.

In the case of “geminated” plosives, as in Italian fatto, the hold phase of the plosive lasts for an extra long time. If you prefer to describe this event as consisting of two entities, corresponding to each of the phonemes involved (here, /tt/), you can say that the first has no release and the second has no approach. In that sense you can also say that in the English phrase good dog the first d has no release, the second d no approach. But at the physical level we really have just a single plosive with the same three phases as any other plosive, the hold phase simply lasting longer than usual.

In the usual English pronunciation in cases such as actor ˈæktə, logged on ˌlɒɡd ˈɒn, the two adjacent plosive articulations (here, velar and alveolar respectively) overlap in time. This means that the release of the first plosive is “masked” by the second. The auditory signal shows us a velar approach, then a long hold, then an alveolar release. The first plosive is sometimes labelled “unreleased”, though a better term would be to say that it has “no audible release”. (The second plosive could also be called “unapproached”, or better “with no audible approach”, but people don’t often mention that.)

The IPA provides a special diacritic, [˺], U+031A, meaning “no audible release” (Handbook of the IPA, p. 182, 204, and the IPA Chart). We could write this pronunciation narrowly as ˈæk˺tə, ˌlɒɡ˺d ˈɒn. (Unfortunately, U+031A doesn’t display properly, so in its place I’ve used U+02FA, which looks the same.)

The Unicode Standard, unfortunately, glosses U+031A as “IPA: unreleased stop”. And it is true that Cantonese final stops, English k in actor, etc are sometimes termed “unreleased”. I think this terminology is inappropriate.

Daniel Jones used to speak of “incomplete plosive consonants” and to say that such consonants “have no plosion”. What he meant was that they have no audible plosion. We no longer use his terminology, saying nowadays that they have no audible release.

Burmese, Cantonese, and Thai are examples of languages in which syllable-final plosives regularly have no audible release. This applies not only utterance-finally but also utterance-medially no matter how the following syllable begins. The explanation is usually a supervening (overlapping, masking) glottal closure, giving [t˺ʔ] etc. But — utterance-finally, at least — it could theoretically also be because the lungs relax and so stop creating an airstream.

Magda Zając, a PhD student at the University of Łódź, wrote to me about “unreleased” stops.
Do you think an unreleased stop can sometimes appear in intervocalic contexts? E.g. in phrases like: that itchy goose, that area was, she measured out a lot.

Clearly, the ts in these phrases can sometimes be pronounced as glottal stops, ʔ. But they can also be pronounced in a way that involves an alveolar articulation with complete closure, yet without the noise burst associated with alveolar release.

One of Magda’s teachers had told her that some of the sounds that she took to be glottal stops were “actually unreleased /t/s”. But another of her teachers had told her that he didn't think it was possible to have an unreleased stop intervocalically. So who was right?

I think they were probably t with no audible release. An overlapping glottal plosive masked the alveolar release.

I said
Physiologically, you can't have a fully “unreleased” plosive between vowels. In any plosive there is a compression stage, during which the air stream gets compressed behind the closure. Either this compressed air is ultimately released (by the removal of the oral closure, or alternatively by the removal of the velic closure as the soft palate moves) — or the initiator of the airstream (the lungs, usually) ceases to provide the pressure.

In utterance-final position you might have a truly unreleased plosive, with the lungs ceasing to maintain the air pressure. But not in mid utterance!

If it sounds like an unreleased plosive, perhaps it is really an oral
plosive with a supervening glottal closure. After the oral closure is
complete the glottis closes, thereby holding the air pressure from the
lungs. When the oral closure is released, the closure is inaudible because there is no air pressure behind it.

So I would call it "no audible release", not absence of release.

Magda is not satisfied.
Yes but what I don't quite understand is the following: What, exactly, is the difference between a word-final plosive with inaudible release and a word-final unreleased plosive? Because if I understand correctly, in both cases the air compression becomes weak and no relase stage can be heard. So can these terms be used interchangeably with regard to word-final plosives?
I'm really sorry for the pestering, but I really can't get my head around the difference between "unreleased" and "no audible release", especially since at times it seems that both terms are used to refer to the same phenomenon (e.g. in Gimson's pronunciation of English in the section about the release stage of English plosives).

I don’t know what further I can say to make my view clearer. It is true that in the Gimson book (seventh edition, 2008, ed. Cruttenden) the term “unreleased” does make one, isolated, appearance. Let’s put this down to an uncharacteristic lapse on Cruttenden’s part.

Let’s boycott the term “unreleased”.

Tuesday 13 March 2012


Kensuke Nanjo made an interesting suggestion about English tonicity.
I've been wondering if it's safe to say the word "else" always takes the nucleus in such sentences as "What else did you say?", "Who else did you see?", "Is anyone else coming?", "Nobody else has finished it", because the word "else" indicates that the rest of the sentences conveys old information.

It’s certainly the case that else in this meaning usually bears the nuclear tone.

•  ˈWhat ˈelse did you say?
•  ˈWho ˈelse did you see?
•  Is ˈanyone ˈelse coming?

•  There’s ˈsomething ˈelse I’d like to talk about, | as ˈwell.
•  If I ˈcan’t ˈtrust ˈyou, | who ˈelse can I trust?

(the latter two examples from LDOCE).

We need the qualification “in this meaning” (i.e. ‘not already mentioned’), because there are other senses of else (‘if not’, ‘otherwise’) in which it is usually accented, but does not necessarily bear the nuclear accent:
•  ˈHurry ˈup, | or ˈelse we’ll ˈmiss the ˈtrain.
•  They’ll ˈeither reˈduce the ˈprice | or ˈelse inˈclude ˈfree inˈsurance.

Another LDOCE example is interesting.
•  I’d like you to come, and anyone else who’s free.
If we devise a context for this example, there are two possibilities. In both, we can assume that come is given.
(i) In one of them, “you” are free. So free is a given, therefore unaccented.
•  I’d like ˈyou to come,| and ˈanyone ˈelse who’s free.
(ii) In the other, “you” are not free. (“I’m designating your attendance as being part of your duties. Other people, if not required elsewhere, should also attend.”) In this case free is NOT given, and the appropriate accentuation pattern would be
•  I’d like ˈyou to come,| and ˈanyone ˈelse who’s ˈfree.
The difference in meanings is conveyed by whether or not else bears the nucleus.

In my reply to Kensuke I said
I don't think you can make this generalization. There can be contrastive focus on some other item in sentences containing "else".
• OˈK, | you've ˈtold me who ˈelse was inˈvited, | but ˈwho (ˈ)else ˈactually ˈcame?

Like other tonicity rules, this proposed rule can be overridden by the demands of narrow/contrastive focus.

Monday 12 March 2012

the Lindsey system

Some of you will have read a thought-provoking blog posting by Geoff Lindsey (trading as English Speech Services), in which he argues for a radically revised transcription system for the vowels of the English of England.

ɑj bɪlɪjv ɪts ɪmpootnt nɔʔ dʒəs tə dɪskəs transkrɪpʃn sɪstəmz əz ə θɪjərɛtɪkl (θɪɪrɛtɪkl) ɛksəsɑjz bət oolsəw tə trɑɪ ðəm awt ən tɵw ɪnvɛstɪɡɛjt haw ðɛj wəək ɪn praktɪs. səw hɪjə (hɪɪ) ɡəwz.

wɔʔ dɵw ɑj fɪjl (fɪɪl) əbawt ɪt ɪn dʒɛnrəl?

əpɑɑʔ frəm ɪʔs dɪskənsəətɪŋ ənfəmɪlɪjarətɪj, ɑj fɑjm mɑjsɛlf rɪjəlɪj (rɪːlɪj) lɔŋɪŋ tə bɪj əlawd sɪmblz fə tɵw fəwnəlɔdʒɪkl ɛntətɪjz ðət ɑjm əkəstəmd tɵw.

• wən ɪz ʌ. sɪns STRUT əŋ commA ə nɔʔ məədʒd fə mɪj, ɑjd lɑjk tə bɪj ɛjbl tə kəntɪnjɵw tə jɵwz ə sɛprəʔ sɪmbl fə ðə foomə. ɑjd bɪj hapɪj tə mɛjk ɪt ɐ rɑɑðə ðən ʌ. bəʔ dʒɛf dɪsmɪsɪz STRUT əz ə məəkjɵwərɪjəl (məəkjɵɵrɪɪl) ən fəŋkʃən lɑjʔ ɡrɛjl (fə wɪtʃ hɪj wɪnz tədɛjz mɪkst mɛtəfə prɑjz).

• ðɪj əðər ɪz ðə happY/schwee sɪmbl i. fə happy, wɪtʃ dʒɛf wɵd rɑjt hapɪj, ɑjd lɑjk tə bɪj əlawd tə rɑjt hapi. (ɑj ɪmadʒɪn ðət “hapɪ”, ðəw fənɛtɪklɪj akjərəʔ fə mɪj, wɵb bɪj sɪjn əz ɪnsəfɪʃnʔli mɔdn əm mɛjnstrɪjm.)

• dʒədʒɪŋ bɑj rɪjakʃənz ɑj tɪpɪklɪj ɡɛʔ frəm nɛjtɪv spɪjkəz, ɑj θɪŋk mɛnɪj pɪjpl wɵd wɔnt ə spɛʃl ɔw fə ðə vawəl (vaal) əv cold, soul, fənɛtɪklɪj vɛrɪj dɪfrənʔ frəm ðɛɛr əw əv coat, so.

ðə stɛjtəs əv smɵwðd vawəlz ɪz nɔʔ klɪjə (klɪɪ) ɪn dʒɛfs sɪstəm. ɑjm θɪŋkɪŋ nɔt əwnli əv ðɪj ɑjə → ɑɑ əv science, lion ən ðə awə → aa (?) əv power, sour bət oolsəw əv ðə NEAR ən CURE vawəlz (vaalz). am ɑj səpəws tə rɑjt here əz hɪjə oor əz hɪɪ? fə jury (ɪf ɪʔ rɑjmz nɑjðə wɪð story noo wɪð furry) ʃal ɑj rɑjt dʒɵwərɪj oo dʒɵɵrɪj? oo dəznt ɪʔ matə? səw fɑɑr ɑjv ɡɪvn jɵw bəwθ pɔsəbɪlətɪjz. bət ɪf wɪjə (wɪɪ) ɡənə jɵwz ðɪs sɪstəm ɪn tɛksbɵks ən dɪkʃənrɪjz wɪj nɪjd tə mɛjk əp ɑɑ mɑjnz wɪtʃ wən wɪj prɪskrɑjb əz ðə noom. ɑjd lɑjk tɵw ɔpt fə ɪɪ ən ɵɵ, bət ɑj səpəwz ðɛj wɵɡ kɔnstɪtʃɵwt ən ənnɛsəsrɪj kɔmplɪkɛjʃn fər ɪj ɛf ɛl pəəpəsɪz.

ɑj dəwnʔ θɪŋʔ ðə sɪmbl ɵw ɪz wɛl tʃəwzn. frəm ə pjɵwəlɪj (pjɵɵlɪj) praktɪkl pojnt əv vjɵw, ɑjm əfrɛjd ðə lɛs ɛkspəət jɵwzəz əv fənɛtɪk sɪmblz wɵd fɑjnd ɪʔ dɪfɪklʔ tə kɪjp ɵw dɪstɪŋkt from əw, oor ɪndɪjd bəwθ əv ðɪjz dɪstɪŋkt frəm aw. ɛnɪjhaw, ðə dɪfrəns bɪtwɪjn GOAT ən GOOSE ɪz moo ðən dʒəs lɪp rawndɪŋ ɔn ðə fəəst ɛlɪmənt əv ðə dɪfθɔŋz. ðəz ə dɪfrəns ɪn vawəl (vaal) hɑjt. səw ɑj θɪŋk ə bɛtə sɪmbl fə GOOSE wɵb bɪj ɨw.

wɔʔ dʒɵw θɪŋk?

kɔmɛnts (ɪn transkrɪpʃn, plɪjz, ɪf jɵw kən manɪdʒ ɪt) wɛlkəm.

Friday 9 March 2012


The winners of a televised competition to represent Russia in the forthcoming Eurovision Song Contest, it was reported yesterday, are the “Buranovo Grannies”.

Apart from an English-language refrain, they perform their song in their ethnic language Udmurt. This has thrown this little-known language into the spotlight.

There is not a lot of information about Udmurt available. I haven’t had an opportunity to consult that old warhorse, Vinogradov’s Языки народов СССР Yazyki narodov SSSR, ‘Languages of the Peoples of the USSR’ (Moscow: Nauka, 1967), but there is a certain amount of information available in Comrie’s The languages of the Soviet Union (CUP 1981), in Ethnologue, in Omniglot, and in Wikipedia.

All agree that Udmurt (aka Votyak) is a Uralic language, in the Permic or Permian branch of the Finno-Ugric group. Its closest relatives are thus Komi-Permyak and Komi-Zyryan/Zyrian. It is related, but less closely, to Finnish and Estonian, and even more distantly to Hungarian. There’s a family tree of the Uralic languages here.

The number of speakers of Udmurt is reportedly just under half a million.

The Wikipedia account of the language contains just a single gnomic sentence on its phonology:
The language does not distinguish between long and short vowels and does not have vowel harmony.
So what, you might think — it also doesn’t distinguish between aspirated and unaspirated consonants and does not have lexical tone. We need to know what it has, not what it doesn’t have. The vowel harmony point is worth making, though, since most other Uralic languages do have vowel harmony.

Thankfully, Wikipedia does give us a consonant chart, a vowel chart, and information about the writing system and its relationship to pronunciation. The consonant system appears to be pretty unremarkable. Among the consonants Wikipedia lists ɲ (or ) and ɕ ʑ (or sʲ zʲ), though these appear to be just positional co-allophones of n and s z; there are also distinct ʃ ʒ tʃ dʒ. The vowel system is reportedly as follows. I wonder if there’s really a distinction between ə and ʌ — as far as I can see, both are represented by the same letter in spelling.
…which brings us to the writinɡ system. Like other languages of the Russian Federation, Udmurt is written in Cyrillic. As usual for such languages, the Russian alphabet has to be supplemented with various special letters or letters bearing diacritics. Udmurt uses just the diaeresis diacritic, with five extra letters Ӝӝ Ӟӟ Ӥӥ Ӧӧ Ӵӵ. The first two stand for affricates, Ӧӧ for ə ~ ʌ. It is not clear from our meagre sources what the effect of the diaeresis in Ӥӥ is meant to be, given that both it and plain Ии seem to stand for i. The use of Ӵӵ seems to imply that is indeed distinct from , the latter being written plain Чч.

It is hard to avoid the conclusion that in the 1930s, when Soviet minority languages were being equipped or re-equipped with alphabets and standardized Cyrillic writing systems, Stalin carefully ensured that no two languages had quite the same set of special letters. That’s why we now have an “Extended Cyrillic” section in Unicode (U+048A Ҋ to U+04FF ӿ), and a “Cyrillic Supplement” (U+0500 Ԁ to U+0527; I can only display them up to U+0513 ԓ).

As well as this provision for the minority languages, Unicode now caters for Old Church Slavonic and Old Cyrillic in “Cyrillic Extended-A” (U+2DE0 to U+2DFF) and “Cyrillic Extended-B” (U+A640 to U+A697), containing characters I can’t display to you but which if you’re interested you can inspect on the Unicode site.

Meanwhile, I’m looking forward to hearing some spoken, or at least sung, Udmurt.

Thursday 8 March 2012

I don't believe it!

Most of us will have had the experience of mentioning some well-known phonetic feature of a given language/dialect/accent, only to have a native speaker deny that any such thing occurs.

I might point out to my phonetically unsophisticated audience that in words like bad, which in Britain ranges generally in the area bæd ~ bad, Americans often pronounce something with a much closer vowel, perhaps nasalized, often with an off-glide, something like bẽəd. Brits nod their heads in agreement. But a kindly middle-aged American lady protests that she has never heard such a thing: Americans never use such an ugly pronunciation.

Or, talking to a British audience, I mention the phenomenon of smoothing, whereby for example science, notionally ˈsaɪəns, is often/usually pronounced without any kind of phonetic ɪ in the middle, just ˈsaəns ~ saːns. People protest that they would never do that themselves, although they admit that perhaps effete upper-class/vulgar lower-class people might sometimes do so.

So it is with American t-voicing. You point out that atom and Adam are (usually) homophones in AmE. Oh no, says an American, that’s not true. Indeed, when Webster’s Third International (1961) pioneered the idea that the intervocalic consonant in this word is d, this was not well received in the United States. To this day, another dictionary from the same stable, Webster’s Collegiate, tells you that this word is pronounced with t. (OK, given that the t ~ d contrast is typically neutralized here, it may be reasonable to continue to write it as t. Arguably, the adjective atomic, where the t remains voiceless, enables even illiterate speakers to know that the underlying consonant is t, not d.) More sophisticated Americans will tell you that some distinction of vowel length or vowel quality enables the distinction to be preserved in such words despite the voicing of the obstruent. Indeed, this is why in LPD I felt obliged to fudge the issue by using a special voiced-t symbol. But the ODP, to its credit, gives the AmE pronunciation of atom as ˈædəm, with no further options.

Likewise with American nt-reduction. In the positions where AmE has t-voicing, many Americans reduce historical nt to plain n. This makes winter a homophone of winner ˈwɪnɚ and dental ˈdɛnl̩ ~ ˈdɪnl̩ a rhyme for fennel.

If your audience are reluctant to trust the evidence of their own ears when you play them sound clips of people using the pronunciations you are talking about, they may be convinced by the visual evidence of spelling mistakes.

I came across a particularly neat example of a spelling mistake based on nt-reduction. Shumway “Innernational” Airport, located in Shumway, Illinois, has its own Facebook page.
Actually, though, given that the reported population of Shumway is just 202, I suspect that the whole thing is a spoof. But there are 56,800 Google hits for innernational.

Wednesday 7 March 2012

merrily, primarily

What effect does the suffix -ly have upon word stress?

The usual answer is none. This suffix has no stress of its own. It leaves unchanged the stress of the stem to which it is attached.

From ˈhappy we form ˈhappily, from proˈfound we form proˈfoundly, from ˈaverage we form ˈaveragely, from ˌsureˈfooted we form ˌsureˈfootedly.

But there are occasional exceptions, at least for some speakers. We all pronounce ˈnecessary with initial stress. Some of us have a weak and compressible vowel in the suffix, thus ˈnesəs(ə)ri, while others have a strong one, thus ˈnesəseri etc. In the latter case you may say that the suffix has secondary stress, but I don’t think there are any mainstream speakers of BrE and AmE who place the primary stress on the suffix. (Caribbean English may be another story.) Yet for the adverb necessarily we see that many Brits and probably all Americans shift the stress from nec- to -ar-. (Personally, I belong to the presumably shrinking proportion of Brits who don’t do this, but feel it more natural to keep initial stress and a weak -ar- in necessarily ˈnesəsrəli.)
The same applies to other -ly adverbs formed from adjectives in -ary, as voluntarily, primarily, arbitrarily. It can also apply to those from at least some adjectives in -ory, as articulatorily, obligatorily, though not apparently perˈfunctorily. Now I have heard another exception: a TV commentator talking about resoˈlutely opposing something or other. Presumably, like the rest of us, he would stress the unsuffixed adjective on its first syllable, ˈresolute. So why wouldn’t he say ˈresolutely?

I don’t know how many other speakers would agree in having this stress alternation, or what other adjective-adverb pairs there might be for such speakers that follow the same irregular pattern.

Tuesday 6 March 2012


If the IPA were to admit a separate phonetic symbol for every subtlest nuance of sound that anyone can detect, chaos would ensue as symbols proliferated uncontrollably. So for at least the last hundred years we’ve followed the rule that we don’t have a special symbol unless there is some language that makes phonemic use of the corresponding sound: for any arbitrary pair of symbols there must be a language in which a phonological contrast between the corresponding sounds is attested. In the words of the 1949 Principles,

The current Handbook (p. 27-30) expresses the same general idea at greater length and with greater sophistication. It makes the point that if we have a phonemic contrast somewhere we generally ought to have a special symbol, though not the converse that if we don’t have a phonemic contrast then we ought not to have a special symbol.

Either way, the symbol for the labiodental nasal, ɱ, seems to be a striking exception to the principle. There is certainly no widely spoken or studied language in which ɱ contrasts with both m and n. If the sound occurs at all, it is virtually always a positional allophone of one or the other or both, found only adjacent to another labiodental (normally f or v).

If you say comfort as ˈkʌɱfə(r)t and emphasis as ˈeɱfəsɪs (as many textbooks assert), then English ɱ can be considered a realization of phonemic m.

There are two reasons why matters are not so simple, however.

• In practice many speakers do not use these pronunciations consistently. Between the nasal and the fricative, epenthesizers — those who make mince and mints homophonous — will have a voiceless plosive of some sort, usually labiodental and/or glottal, thus ˈkʌɱp̪fə(r)t ~ ˈkʌɱʔfə(r)t (where “p̪” in the first version stands for a voiceless labiodental plosive).

• Rather few people have such perfect dentition that they can effortlessly create a true labiodental stop. The smallest gap between any two upper front teeth is sufficient to preclude the complete closure needed for a plosive. Equally, any nasal at the same place will not be a true stop nasal, but will have some oral escape.

In practice what we get as the realization of the phonemic m here may typically range from a bilabial m in careful or overarticulated speech, through ɱ with the reservations just expressed, or perhaps a nasalized labiodental approximant, ʋ̃, down to mere vowel nasalization in rapid speech. I expect the same applies, mutandis mutatis, in other languages.

Be that as it may, the fact remains that as an IPA symbol the letter ɱ appears to be an anomalous exception. Certainly at the time of its adoption its distinctive use in a language had not been attested.

Are things different now, when so many more languages of the world have been phonetically investigated?

The only language that people mention as allegedly having a distinctive labiodental nasal phoneme is a language of the Republic of the Congo (Congo-Brazzaville) known as Kukuya or, more fully, as Teke-Kukuya.

In this language, it is claimed, there are various conclusive minimal pairs.
I have not been able to read the original in which this data was presented. I have, however, read a review of it by the Africanist Michael Mann (BSOAS, 1976, 39:725). He does not even mention Paulian’s striking claim about the labiodental nasal.

I also note that Kukuya is part of a continuum, Teke, containing eleven other languages (or perhaps dialects), for none of which the labiodental nasal phoneme is attested.

I can’t help wondering whether Paulian relied for her data mainly on a single informant who happened to have an unusual dental anatomy and who pronounced these words slightly differently from everyone else. Failing that, I wonder if this ɱ might alternatively be phonemicized as, say, /mv/ (compare the the ‘bv’ above).

Monday 5 March 2012


On Saturday morning I was at the supermarket doing my weekly shop. I had just picked up a pack of kitchen towels and put them in my shopping trolley when my attention was caught by a woman pointing out to her husband that there was a special offer on a brand I had not chosen which made them a much better bargain than the ones I had taken. So I put mine back, picked her brand instead and briefly joined in their conversation.

They hadn’t uttered more than two or three phrases, but from their speech I could tell that they were from Barbados. So I boldly said, “You’re Bajans, aren’t you?”. Yes, she replied, we are: how did you know, have you been to Barbados? And we had a brief chat about the attractions of their native island.

It would be possible at this point for me to start listing the characteristics of Barbadian English that make it sound different from other Caribbean varieties: the rhoticity, the raised PRICE vowel ʌi ~ əi, the glottal stops, the unusual rhythm. But my recognizing the accent was not a matter of taking note of each of these characteristics in turn and computing the implications. Rather, it was intuitive recognition of a gestalt, a complete unanalysed pattern of sounds. I think this is how we typically recognize accents we’re familiar with.

If you’d like to listen to sound clips of Bajan please consult the comments made on my blog entry of 23 Aug 2010, particularly those by Amy Stoller, who supplied an extensive list. (Thanks!)

As well as being the fiftieth anniversary of my MA thesis (blog, Friday), this month also sees the thirtieth anniversary of the publication of my three-volume Accents of English. You will not be surprised to know that I did not write this work by systematically starting at the beginning and working through chapter by chapter until I reached the end. Rather, I jumped around doing something here, something there, as opportunity provided. And my short section on Barbados was one of the first I drafted.

The stimulus for this was our being invited to dinner by a couple we knew in London, one of whom was Bajan, and spending the evening with them. This would have been in the early seventies. I don’t think I was quite so rude as to make notes as we were talking, but as soon as I was on my own I certainly hastened to write down my phonetic impressions of his speech. The next day I wrote them up, and gave my essay to Doc O’Connor, my supervisor for the PhD on Jamaican I had recently completed, for his comments.

At that time I had never visited Barbados. But in January 1978 I was able to spend a few days there on the way back from my first visit to Montserrat, after which I revised and extended what became section 7.2.4, Barbados, of my book.

Friday 2 March 2012

no time

Sorry — I haven’t got time to write a proper blog entry today, so instead I’m referring you to this excellent discussion by Geoff Lindsey about STRUT and various other vowels of English English. There are plentiful sound clips to illustrate it.

* * *

In other news, this month sees the fiftieth anniversary of my completing my master’s thesis. How time flies!

Thursday 1 March 2012


Megalosaurus, ichthyosaurus, stegosaurus, Tyrannosaurus rex — most nine-year-olds will be very familiar with the fact that -saurus -ˈsɔːrəs is a common ending for the names of dinosaurs.

Young dinosaur enthusiasts who live in England and other nonrhotic places are thereby well primed for jokes such as
   — What do you call a short-sighted dinosaur?
   — A “do-you-think-he-saw-us”!
where of course saw us is pronounced -ˈsɔːrəs with intrusive r.

Obviously, this joke would tend to fall flat in north America.

Rhotic speakers have to console themselves with lamer jokes like
   — What did the dinosaur say after the car crash?
though I suppose even that one wouldn’t work for people who have distinct FORCE and NORTH vowels.

They’ll have to be content with
   —What do you call a dinosaur that smashes up everything in its path?
   —Tyrannosaurus wrecks!

A correspondent saw some interesting dinosaur money boxes in a gift shop in London. Each dinosaur was given a name formed from a child’s name plus -saurus, with the pronunciation shown in an ad-hoc respelling system.

What is interesting is that the respelling system assumes intrusive r as a matter of course, so that “saw us” is seen as an adequate indication of how to pronounce -saurus. My correspondent comments that this produces the wrong pronunciation for a rhotic speaker like him.

Clearly this gift shop is not going to be catering for the Scottish, Irish or American markets.