Archive 32 of JWL Blog

Readers who need explanations of any of the abbreviations used may find them at Section 1 of the Home Page.

RSS Feed

OED3 Phonetics

"Kraut’s English phonetic blog" had at the fifteenth of this month an item "Brit." vowel symbols in OED. Here's a footnote to it for anyone who might like a little background historical information.

The Oxford English Dictionary (OED), originally known as the Philogical Society of London's New English Dictionary (on Historical Principles), abbreviation NED, began publication in 1884 and was completed in 1928. In 1933 its title became officially what it had been informally for forty years. In 1989 the second edition (OED2) appeared with 140,000 pronunciations which had undergone conversion to International Phonetic Association symbols. When James A. H. Murray, its first and principal editor, decided in 1882 on the symbols he used for pronunciations, the IPA didnt even exist. The computerisation of OED2 was such a huge task that only a very limited amount of attention to phonetic matters was possible in addition to the many complications of the conversion to IPA symbols. Updatings of the pronunciations themselves were of necessity a low priority concerning "...for the general user, most of them .. merely small nuances...".

Kraut's choice of an example, the word incomprehensibility, we find in OED2 as /ɪnkɒmprɪhɛnsɪˈbɪlɪtɪ/. In 1908 Murray in NED (aka OED1) transcribed it as (inkǫmprĭhensĭbi·lĭti). His account of the consonants calls for no comment except praps for the very minor point that there's no indication of the epenthetic /t/ that many speakers employ between the second /n/ and its following /s/. Most dictionaries sensibly dont bother to indicate these totally predictable possibilities explicitly at ev·ry word concerned. Current LPD and EPD both do show them using italic t's of which LPD's is superscript.

Of the eight vowels of incomprehensibility, the ones in the four strongest syllables in-, -com-, -hen- and -bil- received only a strai·tforward exchange of one symbol for another having exactly the same significance: the vowels of sit, got, yet and agen its. The other four vowels, those of -pre-, -si-, -li- and -ty had symbols Murray described as conveying, in each case except the last, two possible values (and perhaps intermediate ones between them) which might be used in different styles of utterance. These styles he termed as either "syllabic" ie (possibly extremely) deliberate, or "rhetorical" ie declaimed etc, or as used "in singing". The values for -pre- (prĭ) apparently ranged between /iː/ thru /ɪ/ to /ə/; for -si- (sĭ) and likewise for -li- (lĭ) wd be from /ɪ/ to /ə/; that for the weak fourth syllable -ty (ti) was (in a way that's now become old-fashioned) me·nt to convey precisely the same value [ɪ] as in the first syllable in- and the tonic one -bil-. The principal strest syllable was indicated in Murray's system by the raised point (·) after its vowel. In modern practice subordinate stresses are shown but Murray quite reasonably chose not to mark them if he regarded them, as they usually were, "sufficiently indicated by the clearness of the vowel".

The editors of OED2, understandably in view of the tremendous task that the conversions entailed, decided agenst replicating Murray's delib·rately ambiguous symbols for English, using only [ɪ] "as in pit " and [iː] "as in bean". (In "unassimilated forren words only", [i] was also employed.) In general their phonetic symbols were chosen "in accordance with general present-day custom" of the day. Such a procedure has been less fully followed in the OED3 revised entries. Also the very sparing provision of alternative pronunciations has been improved upon. So far OED3 has de·lt with the words from M to (much of) R and many items outside that range especially from the "first half of the alphabet, where .. entries are noticeably more in need of revision than later on".

Kraut's choice of incomprehensibility shows clearly how good the new font of symbols looks as /ɪnkɒmprɪhɛnsɪˈbɪlɪtɪ/. What we have for this entry at the moment in OED3 is due in time to be updated to /ɪnˌkɒmprɨˌhɛnsɨˈbɪlɨti/, and if it's made to match the ODP (Oxford Dictionary of Pronunciation) version of the word, it'll be accompanied by an alternant version with a secondary stress on the first syllable, as in LPD and EPD. ODP, like them, also shows the first syllable as variably/ɪŋ-/. OED3 [ɨ] is a dual symbol indicating that /ɪ/ and /ə/ are current alternants. More detailed comments on 'IPA vowel symbols for British English in Dictionaries' may be seen in my article of that name at Section 5.1 in the main part of this website.

The New-Look OED

At the beginning of this month the great OED unveiled an amazing new re-organisation of its style of presentation and not just that but the offer of all sorts of valuable facilities it has never provided before. The great new benefit to those who are particularly interested in pronunciation matters is that the phonetic transcriptions of each word no longer suffer from the odious mixture in which any symbol that wasn’t to be found on ordinary keyboards was shown as a painfully jejune fixed-size glyph of tiresome illegibility which, when a word was copied and pasted, disappeared altogether. The new font is a pleasing Bigelow & Holmes’s Lucida Sans type. One surviving horror is a reminder of the first issues of OED on disc: the tildes of nasal vowels are not on top of the letters but trailing after them eg vingt-et-un is /vɛ~t e œ~/. Another seems to be the frequency with which the IPA symbol ɜ is replaced by 3.  By another minor glitch, altho the dreaded glyphs have gone from the pronunciations given at each entry, at the 'Key To The Pronunciation' they have not yet been replaced by phonetic symbols but by what one presumes are printers’ identifying names for them: thus inste'd of [ɪ] we see ‘shti’.  
Anyway, for the generality of users these won't matter a jot because the re-organisation provides an excellent new feature which means they need never agen make the tiresome search for that key if they want to check on how to interpret some phonetic symbol. This is because one only has to click on the phonetic transcription of a word and up pops an information panel with the kind of sequence of contents seen at the word pronunciation:

p - r - ə - n - uh - n - s - ih - ay - sh - n

IPA         sounds like
p              p      as in pine
r               r      as in run, terrier
ə              ə      as in another (schwa)
ˌn             n      as in nine (secondary stress)
ʌ             uh     as in butter
n             n       as in nine
s             s        as in see, success
ɪ             ih       as in pit
ˈeɪ          ay       as in bay (main stress)
ʃ             sh       as in shop, dish
n             n         as in nine

The first line respells the word with an ordinary letter (or couple of letters) corresponding to each of the word’s phonemes. There’s a fair amount of inessential repetition eg /n/ is glossed three times but that’s no problem with the amount of space available online. A curious thing about these boxes is that each main or secondary stress mark (ˈ or ˌ) which precedes the consonant that begins a syllable is identified after the example of that consonant and within fairly pointless-seeming brackets. Another puzzling kind of decision was to represent the /ᴧ/ vowel of butter as 'uh'. There isn’t a single normal word in the English language which contains the spelling -uh- with the sound value /ᴧ/. (Interjections like huh and duh don’t count coz they don’t have precise phonemic values.) The vowel /ʊ/ is equally pointlessly so represented in the respelling of put etc. Just as odd is the similar choice of 'ih' to represent [ɪ]: no word exists with ih representing [ɪ] by itself — ie with the ‘h’ not belonging in a different syllable. Anyway, this syllable conveyed by 'ih' here wd be more properly equated with the final vowel of happy — a keyword also employed in this system. Other syllables conveyed by 'ih' represent free variation between /ɪ/ and /ə/: these have the twin  deliberately ambiguous keywords roses and business.

There are other occasions on which this quite good basic idea is also less than ideally applied. For example the rather uncommon word articulate is used as a keyword to convey how the medial -u- syllable in words like reputation is spoken when the latter word is much better known than articulate whose Google incidence figure is less than 14 million whereas reputation is registered with 190 million. When a word that is a keyword is glossed, it may drolly have only itself as the ‘explanatory’ word, as we find at articulate. It can be rather embarrassing when this happens for a forren word. Fortunately some of the keywords are supplied in pairs. It’s just as well that the stressed vowel of cure is glossed by jury as well as by itself. The oe of boeuf is glossed by coeur as well as itself. Keyword items with themselves only as glosses include pet, pit, pot, goose, mouth etc and, worse than just awkwardly, loch and fin (de siècle). This last French vowel [ɛ̃] was re-spelt as 'ah'! Other curiosities include the glossing of American /u/ by 'oo' as in French douce and of the diphthong /ɔɪ/ as beginning with o as in French homme. The US variant of air as /ɛ(ə)r/, which is [ɛə], is glossed as having '(ə) as in beaten' but when the OED2 representation of the latter syllable of that word with potential schwa has been braut in line with items like OED3 motheaten which is glossed only as 'Brit. /ˈmɒθˌiːtn/, U.S. /ˈmɔθˌitn/' this cross-reference’ll have to be to something other than beaten which will presumably then have no schwa. We can, of course, be confident that all these little teething troubles'll be sorted out in due course but such items do illustrate some of the endless problems the dou'ty OED editorial team have to grapple with.

Remembering PHILIP LARKIN

Defining and Delimiting an Accent

In his Blog of Wednesday the 24th of November 2010 on ‘intrusive r in EPD’ John Wells began very sensibly by saying that the question of the status of that incidence of /r/ has to depend on “how you choose to define RP. Do you consider it (i) the implicitly agreed model of good BrE speech? (ii) a codification intended mainly for EFL pedagogical purposes? or (iii) an objective description of how people at the top of BrE social stratification actually speak?”

Let’s put aside for the moment the consideration that very many people have for quite some time been becoming increasingly dissatisfied with that label ‘RP’ whether as initials only or in its full form of ‘Received Pronunciation’. The term ‘intrusive’ can also be sed to be something less than fully satisfactory. Of course in a blog one may reasonably permit oneself a degree of light-h·arted informality of expression that one might hesitate to adopt in a less ephemeral publication. However, I still feel that these three formulations are rather surprisingly out of line with what one expects from a distinguisht speech scientist of John’s well-deserved high reputation. I suppose we may take them for more or less caricatures of how very many people who, despite extensive education having precious little sophistication in linguistic matters, are offen inclined to express themselves.

The first ‘definition’ is strongly redolent of the unscientific style and outlook very general a century or more ago. You find it in Jones’s least mature publications, including notably The Pronunciation of English (1909), which so embarrassed him in later life that he cd say to his colleague Olive Tooley “Every copy should be burnt” (see The Real Professor Higgins 1999 by B. S. Collins & I. M. Mees p. 65). Walter Ripman, his fr·end and senior by a dozen years, wrote in his English Phonetics (1931 p.9) “really good speech is called ‘received standard’ ” using the terminology promoted by another outrageous snob (even if great scholar) H. C. Wyld!

As to the second ‘definition’, whatever some people may consider privately to be the de facto state of affairs, no-one I know of has advocated such a restricted formulation — certainly not Jones, the original promulgator of the very expression ‘Received Pronunciation’— as a fully fledged institutional label or any of his many followers in the field of English phonetics.

The third ‘definition’ is quite unuseable if only because of the indefinability of the expression ‘people at the top of BrE social stratification’. How is the line to be drawn between the top and the middle? I feel sure that no-one who has made serious use of the term has ever wisht to limit its application to the aristocracy. If we’re talking about speech features that can reasonably be suggested to be exclusively characteristic of that very limited number of speakers whose usages are markedly sociologically conspicuous in that they are positively indicative of upper-class status, then I dont know of any respectable speech scientist who has advocated such an interpretation of the term ‘Received Pronunciation’ either. In any case such things tend to be what one may call ‘in the ear of the hearer’ in the way that beauty is sed to be in the eye of the beholder.

For the last half century I’ve abjured the label ‘RP’ (and its expanded form) in favour of ‘GB’ (‘General British’) tho I take the two terms to be virtually synonymous. However the approach I adopt in defining them is importantly different from that of seemingly the generality of the espousers of the term ‘RP’. It strikes me that in adopting a term which explicitly mentions one alone of the only two axes on which the definition of such a speech variety need be delineated they are according the sociological axis priority of importance. I’m very much inclined on the contrary to accord the major significance in my definition of ‘GB’ to the geographical axis. In any case, it’s impossible to state with certainty of any individual speakers that they are completely free from any features which are (a) of unchallegeable social acceptability and (b) indicative of some degree of geographical affiliation.

These considerations expose the futility or at least extremely low credibility of offers of estimates of what proportion of the population might be categorised as “pure” RP/GB speakers. We read suggestions by respected scholars like “three to five percent” but they’re hardly worth consideration. Linguistic “purity” is a futile concept. As Cruttenden so rightly puts it in the Foreword (at p. xiv) to the Seventh Edition of Gimson’s Pronunciation of English “... the percentage of speakers of RP cannot be claimed, as it often is, to be only in single figures; it is almost certainly much higher (there are no reliable figures anyway).” If we did have any remotely defensible figures, they’d be only of the most trivial significance. The true significance of GB/RP resides in the fact of its being the common denominator of all the accents of Great Britain or at least being perceived as something very like that. Huge  numbers of people, including a high proportion of movers and shakers in British public life, have usages which differ from totally unadulterated GB/RP to degrees which are of extreme insignificance.

Questions on Vowel Diagrams

A correspondent has put some questions to me that have occurred to him while thinking about some of the thirty IPA diagrams in the seventh edition (2008) of Gimson’s Pronunciation of English as excellently re-cast by Alan Cruttenden. The classic explanation of the IPA authorised diagram is to be found at pages 10 to 13 of the Handbook of the International Phonetic Association (1999). It has the stylised form of a rectangle with its lower left quarter removed. Tho the IPA-recommended descriptive terms which specify frontness versus backness and closeness versus openness were in origin purely articulatory, the diagram isn’t to be directly equated with a drawing precisely mapping tongue positions but rather to be taken as an ‘abstraction’ that symbolises auditory values.

With its simple shape it provides information in a way that can be easily grasped. Those who use the diagram in forms that suggest mathematical precision are misapplying it. This is the case when the size of a vowel position indicator used seems to imply that it’s possible to discriminate over a hundred unrounded and another hundred-plus rounded (non-nasal) vowels. I’ve seen extreme cases in which the vowel indicators were smaller than a thousandth of the diagram’s area. The fact is that probably not even fifty of each of these two (rounded and unrounded) types can be discriminated by most people even with prolonged intensive training.

One of the questions was “Why does Cruttenden place the indicators over the line representing the most fronted tongue position”. My answer has to be that a difference amounting to only half the width of one of the indicator dots wd be too trivial to be given consideration — whether representing a single utterance or even the typical value of a phoneme. Within the performance of an individual speaker there may be departure from a phoneme ‘bullseye’ even while repeating the same lexical item in isolation. In running speech as opposed to deliberate utterance there may be a good deal of variation prompted by various segmental and prosodic contexts.

 A writer recording phoneme values has the choice of offering only bullseyes or representing some degree of variation. In EPD17, by using dot indicators that cou·d be much larger without distortion of the facts (at pp viii & ix), Roach opts for bullseyes. In LPD3 (at pp xxiii to xxv) Wells adopts fingerprint-type ‘smudges’ for each phoneme indicator quite effectively conveying a degree of variation. Wells’s choice is praps less likely to be misinterpreted by the less sophisticated reader. Cruttenden’s Figure 15 (p.113), the example cited by my questioner, in addition to identifying current relatively mainstream values of the British ash /ӕ/ phoneme provides also regional and sociological variants. It seems fitting that he uses forms larger than minimal bullseyes and of uniform size. My own choice for displaying the phoneme sets of 132 languages (see Section 9.2 of this website) was for indicators of uniform size and shape (colour-filled circles) as large as seemed practicable. (When colours have not been available for the purpose, I’ve preferred to contrast the shapes of rounded and unrounded vowel indicators.)

My questioner also remarked “I had always thought that the four outer lines of the vowel quadrilateral mark the boundary of the space within which vowels can be articulated.” This was not an unreasonable asumption for him to’ve made because it’s a simplification that most of us have preferred latterly but it was never the practice of Daniel Jones the designer of the Cardinal Vowels diagram we now use. Nor was it the practice of Gimson until for latter purposes he adopted simplified versions showing areas rather than dots. The first of these were the rather amateurish-looking ones he used at p. xv of his 1977 revision of the EPD. In the IPA Handbook most of the contributors to the 29 illustrations of the phoneme systems of various languages used bullseye dots which they kept within the peripheral lines. Cardinal vowels are usually placed upon the lines bordering the diagram.

Some Deep South Usages

On holiday recently I re·d The Help Kathryn Stockett’s 2009 novel about black women servants in the Deep South in the 1960s. It’s not a great work of literature but a quite enjoyable read. The language they use seems to be to quite an extent faithfully reflected and cert·nly has some striking linguistic features. I was able to hear a convincingly-re·d five-minute excerpt from the book at an Amazon USA advert which seems now to’ve been withdrawn but there’s a UTube interview with her about the book at http://www.youtube.com/watch?v=ZjIowZrH4iM

Phonologically interesting are things like the fact that they use no /əv/ weakform of ‘of ’, only /ə/ which occurs even before vowels and is regularly spelt “a” (which can take some getting use’-to) eg at ‘terrified a her own child’ (p.1) presumably something like [terəfaːd ə ə ʔəʊn ʧaːl]. A plural ‘womens’ is frequent; so is ‘peoples’ as in ‘so many peoples is here’. The most startling item is praps the constant use of “on” appar·ntly as a reduction of ‘going (-to)’, eg “I’m on have that dream” just possibly [oʊn] as if [gəʊn] with initial [g] elided — at least that’s the only thing I can suggest. Praps hearing the full audio of the book wd elucidate this. Certainly there is a version of “going-to” spelt “gone” which is presumably pronounced /goʊn/ (not [gɒːn] as in the GA past if ‘go’). The word ‘every’ regularly appears as ‘ever’ [ev(ə)] even in compounds like ‘everbody’.

Here’s an alphabetic list of other expressions in the book Brits may well know of but will             cert·ny find pretty unfamiliar tho not necessarily as linguistic items: 

‘besides’ is used in the sense ‘aside from’ by the white narrator
‘booster chair’ is, it wd seem, being used where we might say ‘high chair’
‘clothes dryer’ is apparently our ‘tumbler dryer’
‘crisco’ (explained at p.43) is a brand of shortening ie vegetable fat for making cakes, bread etc
DAR is Dau·ters of the American Revolution
excusal (now a very rare word in UK)
‘fiddleheads’ refers to the fronds of a young fern sometimes used as food
‘grits’ are coarsely ground maize like polenta
‘half tester’ means a partial canopy over a bed
‘huaraches’ /wə`rɑːʧiːz/ are Mexican-Indian leather-thonged sandals
Jameso, a name of one of the negroes, reminds one of British schoolboy nicknaming usages
Jell-O is the US apparently universal equvalent of our ‘jelly’
‘kudzu’ is a climbing plant used as fodder etc pronounced /`kʊdzu/
NAACP is of course the National Association for the Advancement of Colored People (seeing     which as an abbreviation is unusual for us) is pronounced /ˈen dʌbl (ˈ)eɪ siːˎpiː/
‘nigra’ is a variant of ‘nigger’ pronounced /`nɪgrə/
‘nub’ has the meaning ‘knob’ of a small variety
‘Law’ is the expletive ‘Lord!’
LSU is Louisiana State University
‘Ole Miss’ is The University of Mississipi
‘po’boy’ is a kind of baguette sandwich
    (An OED3 June 2010 Draft Revision entry “po’, adj. and n.6 colloq. (chiefly U.S.). Brit.             /pəʊ/, U.S. /poʊ/  In use as  noun representing a colloquial [sic] pronunciation of ‘poor’ ”. I     shd think ‘colloquial’ wd be more suitably replaced by ‘dialectal’ in that draft. Cert·nly regarding the UK).
‘poke salad’ is young leaves of pokeweed used as a salad
‘polky-dot’ is ‘polka dot’
‘poufed up’ is rolled up à la Marie Antoinette’s hairstyle (OED Brit. /pʊf/, /puːf/, U.S. /puf/ )
‘relaxing room’ is self-explanatory for lounge or the like
‘sassy’ is a pronunciation of ‘saucy’ ie cheeky
Skeeter (as explained at p.57) is the nickname of the chief white woman involved given to her     because as a child her limbs were so thin. Many US folk pronounce mosquito /mə`skiːtə/:     The white southerners who gave her the name wd prob·bly be very-low-rhoticity speakers
‘sexual correction tea’ is appar·ntly a herbal viagra equivalent but odd-sounding in its context
‘stroller’ means ‘pushchair’
‘tire iron’ means a metal crowbar for removing inner tyres
‘tote’ is lug, carry etc (Ole Man River lyrics contain Tote dat barge! Lif' dat bale! Git a little             drunk An’ you land in jail
‘wazoo’ seems comparable to British ‘arse’ in signification and status

PS: I’m very grateful to Amy Stoller for some very helpful comments on the above which have saved me from sev·ral infelicities and enabled me to correct mistakes in my original version.

Pronunciations of Aung San Suu Kyi

This newly-much-used name is only to be found in a couple of not universally available pronunciation dictionaries so there may be some confusion on the part of listeners hearing it from a variety of speakers. Of the big three pronunciation dictionaries, only the Longman Pronunciation Dictionary has it (since 2008). It gives it as /aʊŋ ˈsӕn suː `ʧiː/. This is exactly what one finds had been included in that notable pot pourri the Oxford BBC Guide to Pronunciation of 2006. This latter isnt essentially the kind of scientific record of usage one sees in the LPD but a set of recommendations extracted rather subjectively from records kept for the benefit of the BBC’s employees very much in the context of their traditional aims of “correctness” and diplomatic suitability etc. These entries in LPD and OBG will have recently been much appreciated by their users newly needing to consider the name: they’re comfortingly authoritative.

Those who have no access to or indeed knowledge of the existence of either book will mostly nowadays, one expects, seek help from the Internet. Looking online one finds a remarkable  34,000 Google responses to the query “Aung San Suu Kyi pronunciation”. The first ten pages of these I scanned yielded hardly a dozen items of any int'rest. Here they are:
Voice of America had the usual transliteration “Aung San Suu Kyi” but recommended the pronunciation AWN SAN SOO CHEE ie with the first element as /ɔːn/ not /aʊŋ/. Their accomp'nying sound file told a slightly diff'rent story — like /ˈɔː ˈzӕn ˈsuː ˎʧiː/.
Wikipedia which generally uses IPA had [àuɴ sʰáɴ sṵ tɕì] but the [ɴ]s dont have their IPA authorised value of a uvular nasal and elucidation isnt supplied. Anyway, being told the Burmese version is far less useful to most of us than being told what’s a suitable anglicisation.
Fifty Viss (the blog name of a Burmese-American university student in and from Los Angeles where he’s majoring in Biology) giving the name “Aung Kyaw” and obviously writing from an American linguistic standpoint sed int'restingly at least:
“Aung” is roughly pronounced “Oun” (rhymes with “sound,” without the ‘d’ and ’s’). Most Burmese people with the name ‘Aung’ spell it misleadingly with ‘ng’ because in Burmese, it is spelled with a silent ‘ng.’ “San” is pronounced close enough, but to be more exact, it has to be lengthened (so more “Saan” rather than “San”.) “Suu,” ... is more of a snappy and quick “Su” rather than a long-voweled “Suu.” ... The pronunciation of “Kyi” does not even exist in English, so a “Chee” is the closest approximation ... Briefly ... “San” rhymes with “sun,” except with a long vowel.“Su” ends abruptly, like French ‘zut,’ or less closely, to English “loot” (replace the ‘t’ with an abrupt stop) “Kyi” is a .. “Chee.”...
This seems to suggest [aʊn sɑːn suʔ ɕi] anglicising to /aʊn sɑːn suː ʧiː/ tho /səːn/ may apparently be an alternative of his for Americans. He added:
I do not understand why many news articles that include pronunciation keys for Aung San Suu Kyi’s name, like .. ONG-SAN-SUU-CHEE” .. even .. if they bear little resemblance to the name’s pronunciation... [T]hat has become standard .. in most media outlets like CNN.
Forv˚o’s single native speaker seemed to say something like  [ʔɑʊ ˈsã  ˈ su ˈtɕi] in successive descending level pitches after a first low one and with a rhythmic break between the two pairs of syllables.
Howjsay’s British voice gave us /ˈɔːŋ ˈsӕn | suː `ʧiː/ adding after it “but Burmese [ˈɑʊŋ ˈsan `sʊ ʧiː]”.
Yahoo recommended “orng san sue chee” .
Merriam-Webster Online transcribed \ˈȯŋ-ˈsän-ˈsü-ˈchē\ but their speaker sed something more like [ɒŋ sӕn suː `ʧiː] despite their pronunciations indicating [ɒːŋ] and [sɑːn].

Liss’ning to a number of media news reports demonstrated strikingly how unsure one can be of what one’s he'rd if an expression is uttered in an ord'nary not deliberate manner. The problems here are mostly with the word “Aung”. My impression was that only a minority were clearly saying /aʊŋ/. It’s particularly difficult to judge such syllables. I’ve noted in the past that a moderately briskly spoken /aʊn/ presumably from speakers who have very weak and/or weakly-if-at-all rounded /ʊ/ may say a word like counsel in a way that makes it sound very like cancel. One can hardly suggest that the syllable /aʊŋ/ is really difficult for English speakers. It never occurs in the lexical form of any English word but it’s perfectly ordinary-sounding and easily uttered in completely fluent articulations of sequences like crown court, town-crier, brown gloves, down-grade etc.

Most of the speakers I observed from broadcasts were saying something which either fairly clearly was, or was hard to tell apart from, /ӕŋ/. They included for example the British Ambassador to Burma Andrew Heyn, pm David Cameron, tv news presenters including Sky’s Stephen Dixon, BBC’s Andrew Marr, Emma Crosby, Anita McVeigh, Louise Minchin, Kate Silverton, Channel 4’s Jon Snow, and Radio 4 people including Charles Carroll, John Humphrys, Rory Morrison and James Naughtie. Some, including the BBC’s anonymous woman reporter speaking from Burma and John Simpson seemed to be aiming at /aʊŋ/. So did the Fiona Bruce on tv and Radio 4 Newsreaders Vaughan Savidge and Annie McKie who also later seemed to say /ʌŋ/. A few others seemed to say something in the range /ɒŋ, ɑːŋ/ or /ᴧŋ/ such as Samira Ahmed of Channel 4 and BBC reporter Adam Mynott.
The stress pattern of weak-strong-weak-strong was unsurprisingly quite often not employed.

Questions about Accenting

Tami Date has written commenting on my recent Blog 309 on ‘Accents in Happenings Remarks’ saying “I see in There’s (ˈ)someone at the `door  that the indefinite pronoun someone is not accented”.
It’s true that this word someone, as my brackets around the Alt (ie high-level) tone mark in front of it indicated, may or may not receive elective ie intentional stress as opposed to the automatic rhythmical stress that’s part of the rhythmic structure of the word. (By the way, this word, which appears in all dictionaries only as /`sʌmwʌn/, has in fact the common weakform /`sᴧmwən/, not recorded by the pronunciation lexicographers but which cd well be used here. Cf Cruttenden 2008 p. 269.) At any rate, in this case it’s more likely to receive an accent by the speaker’s choice to use an Alt in order to accord it something more than mimimal importance. But the climax (aka nuclear) tone is and can only be on door except, as I pointed out in Blog 309, if the remark is a contradiction of another speaker’s assertion such as “There’s `no-one at the door”. [By the way, if you have occasion to say such a contradicting sentence, remember that there's a distinct danger of sounding impatient in the form we've quoted. The normal English-language habit is to end all contradictions with low rises.]

Tami continues:
However, it seems to me that depending on the context and sentence structure, indefinite pronouns can be accented as in [I’ve added tone marks thruout to his originals] :
 A: `Here you ˏare | at `last. We be`gan to `ˏwonder | what had `happened to you.
 B: `Hullo, ˏSusan. `Sorry to be ˏlate. But we ˈhad a bit of ˈtrouble | with the `car.
A: A bit of  ́trouble?
B: `Everything went wrong. The (ˈ) car wouldn’t ˏstart, ...
Here I presume that the pronoun receives nuclear accent? Yes, indeed. And would you call the sentence in question  a 'happenings remark'?
Yes, I shou'd — of a kind. But please remember that my long article on Accentuation, §8.1 on the main section of this website, begins “Highlighting of contrasts generally overrides all other tendencies to assign stress to a word or syllable”. This has the corollary “re-accenting an immediately re-occurring item” is to be avoided.  Because “went wrong” is synonymous with “had trouble” to accent it wd be the equivalent of re-accenting a repeated item. So any speaker wd be likely to deny an accent to “wrong” even if it werent a response to a kind of “what happened” question. As to “the car wouldn’t start”, the word “car” wd be likely to be unstrest if it was the only matter discusst but, if a new topic (such as the weather) was about to be introduced as well, “car” might be strest in anticipation of its contrast with “weather” or whatever word introduced that topic.
He continued:
Simlilarly,  A: `Oh, `there you are!  ˈAt ˎlast!
B: `Sorry to be ˏlate.
A: You know, it’s ˈpretty ˎcold, waiting ˏhere. `Well, now, | ex`plain yourself. ˈWhat’s been keeping you `this time?
B: `Oh, it’s one of those ˎdays. `Everything |ˈseems to have /təv/ ˈgone ˎwrong.

Or: `Everything seems to have /təv/ gone wrong.
 And also:
 A: ˈWhat’s the next ˎmove?
      B: ˋAnything can happen.
Is it possible to accent 'wrong' and 'happen'?

It has to be remembered that our textbooks’ descriptions are usually of tendencies that are so strong that it’s very advisable for the EEL user to conform with them. Yet the NS may well on occasion have a reason for departing from the most usual practice. Accenting ‘wrong’ can be resorted to in order to make the expression particul’y emphatic, but not accenting it’d be a safer model to follow. In the same sort of way, an NS might well say “ `Anything can `ˏhappen.” However, the version not accenting ‘happen’ is much more likely to occur.
Finally he sed:
A: (seeing his girlfriend crying) ˈWhat’s `wrong?
       B: ˋNothing.
       A: ˋYes, something ˋis wrong.
   Here ‘is’ is accented, isn’t it? 
Yes, indeed. Because ‘wrong’ has just been accented by the speaker and re-accenting an immediately re-occurring item (offen even if it’s only a part of a word) is normally always strongly resisted unless the word has actually changed its meaning at the re-occurrence. See §8.1 agen especially Section 12.  

Prons of DAVID ATTENBOROUGH

I share John Wells’s admiration for David Attenborough that he’s expressed in his current blog. I’ve observed this very distinguished broadcaster’s pronunciatory habits for a number of years offen wond·ring whether they might display any traces of the appar·ntly exclusiv·ly northerly (Nottingham) background of his speech-formative years. (His father may well have had some even tho he was a Cambridge graduate.) If I’d known nothing of his background but been forced to guess entirely on the evidence of his pronunciations, I’d’ve been inclined to exclude any southeastern, western or far-north-of-England influences.

Given his age of 84 (exactly four months older than me), his use of /eɪt/ rather than /et/ for “ate” is praps mildly surprising if not a Northernism but it cou’d be sed that he’s slightly more than the average speaker inclined to be influenced by the spellings of words. This is one reason why I disagree with those commenters who’ve used the term “posh” to classify his version of “sexually” with ess rather than esh. Innumerable people with nothing very socially conspicuous about their speech are similarly inclined. I consider his speech to be within the range of socially neutral mainstream GB (General British) of his generation. Praps the /ɪ/ in possible that Wells he·rd was spelling-influenced tho it’s very much a Northernism today. “For the first time” with initial /fɒ/ cd conceivably be a Northernism — it certainly occurs in Yorkshire — but the current GB /ɔː/ diphthong has commonly more variations of openness and of length than the textbooks tend to suggest.

Attenborough belongs to a generation that’s seen transitions like the changes in preponderances of the happy final vowel from [ɪ] to [i], the cities ending from [-ɪz] to [-iz] (not [-ijz] by the way), pre-pausal /ɛə/ from [ɛə] to [ɛː], /əʊ/ with a slightly fronted schwa to an unfronted one, and in some words like /sju`pɜːb/ to /su`pɜːb/ etc: in all of these he displays rather conservative habits. Similarly he has /bɪn/ for been, kilometre front-stressed and /iː/ in Kenya.
Regarding the stressing of kilometre on this one occasion, it might well be what one may call a pronunciatory throwback in that it cou·d be a momentary reversion to a form one had earlier used but later preferred to replace with another choice. In the seventies in his series Life on Earth he could be he·rd regularly saying ki`lometre.

  On the other hand, in his uses of schwa in medial vowels of items like circulatory, valuable, virtually and manufacture are more modern in style than many of his generation. A very notable spelling pronunciation he has completely regularly exhibited has been /spiːsiːz/ for “species”. It’s my prognosis that his great authority will be largely responsible for the next generation’s likely supplanting the more traditional esh version of this word with the ess form. Among other pns that Wells noted from the BBC First Life programme was believes as “bəˈliːvz (not bi-)”: it may be worth noting for readers that presumably this must be interpreted as if it were an entry from the LPD meaning by “(not bi-)” that it was not /bi-/ or /bɪ-/ or anything in between. Finally, as those who care to listen to the numbers of exerpts from Attenborough programmes available on YouTube may check for themselves, he has of-course-intermittently a number of modern forms such those with complete r-droppings from programme and problems.

A New Word on Me

On Saturday the 23rd of October 2010 I had a very unusual experience: I he'rd an orthodox English word (meaning not slang or jokey tho, as it happened an obvious loanword) which I had never encountered before in my life — at least if my memory serves me right which I’ve no particular reason to dou't. I stress that it was new in my hearing because to come across in print a hitherto completely unknown word isnt a notable occurrence. I shd g'ess on average I meet at least one a week. The word in question was used by only a single speaker, a certain Manfred Nowak, Professor of International Human Rights Protection at the University of Vienna. He spoke correct fluent English but with an obvious non-native accent probably indicating German-speaking linguistic origins (the surname is found chiefly among speakers of Czech) notable from the uvularity of his /r/s. He is referred to in Wikipedia as 'Austrian'. He was being interviewed for the British Radio 4 “Today” early-morning news programme.

The word in question was used by him three times tho not at all by his interviewer Mr John Humphrys. On one of those occasions the first vowel was not very clear but otherwise the word was fairly clearly audible as “/nɒnreɪfuːlmɒ̃/”. It was fortunately possible to confirm these impressions by re-playing the programme. The vowel of re was given the value it cdve had in a German word or in a French one spelt with é. The e is not accented in the French spelling. The final nasal vowel obviously indicated that the word had been borrowed into English from French without complete naturalisation. Unsurprisingly for such an uncommon item, it was not to be found in any of the three major pronouncing dictionaries but it was recorded in OED3, the great Oxford English Dictionary's online edition

thus:
“non-refoulement, n.
Brit. /nɒnrɪˈfaʊlm(ə)nt/ U.S./ˌnɑnrəˈfaʊlm(ə)nt/, /ˌnɑnriˈfaʊlm(ə)nt/” defined as “The principle or practice of not forcibly returning refugees or asylum-seekers to a country where they are liable to suffer persecution”. (My use of plain ɪ not barred ɪ in the British transcription is beyond my control.) The earliest illustrative quotation was from 1972. I’ve quoted from a draft entry dated December 2009. Besides this entry the word is recorded in uncombined form thus:
“refoulement, n.
Brit. /rəfuːlˈmɒ̃/, U.S. /rəfulˈmɑn/ [< French refoulement instance of water overflowing or being dammed back (1771...forced relocation of a group of people ...< refouler to push or force back, to cause to turn or flow back... 1771.” I’ve quoted from a draft entry dated September 2009. Probably the discrepancies between the draft pronunciations will be reconsidered. Cross-references led one to an obsolete spelling foul which might be sed to’ve given way to the spelling full for a verb meaning ‘trample’. Fulling was the process of treading on cloth to cleanse and thicken it whence fuller’s earth and the occupational surname Fuller. I never cease to be fascinated by such etymological ramifications.

BlogRoll

• DCblog (David Crystal)
• John Maidment
• Linguism (Graham Pointon)
• John Wells