Archive 25 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 for this siteRSS Feed

13/02/2010Death Reports Exaggerated (ii)#250
12/02/2010Death Reports Exaggerated (i)#249
08/02/2010A Notational Heresy#248
25/01/2010How do you say it?#247
18/01/2010Pronunciations by ANTONIA FRASER#246
13/01/2010A Decade of OED3 Pronunciations#245
11/01/2010Elision of Yod (ii)#244
07/01/2010Elision of Yod (i)#243
06/01/2010Irrational Spellings of Names#242
25/12/2009Current Wyn Dropping#241

Blog 250

The 13th of February 2010

Death Reports Exaggerated (ii)

To return to the allegedly moribund centring diphthongs, they were first clearly identified by Henry Sweet in the English revision of his Elementarbuch des Gesprochenen Englisch the Prĭmer of Spoken English of 1890 when Jones was about ten years old. DJ didnt seriously depart from that account but when he was fully in his stride he showed in certain cases an oddly idiosyncratic rejection of various of their allophones as not admissible by him into his in some ways curiously narrow perception of what he accepted as ‘Received English’. We can all agree no dou't that during the second half of the last century the ailing /ɔə/ diphthong breathed its last gasp as GB. It survives as a perfectly common ‘educated’ usage in various English regions but if you hear it from a General-British speaker it’s most likely to be provoked by an “accident” of prosody. For my part I habitually say [wᴧn, tuː, θriː, fɔː] but if I say '1, 2, 3, 4' on wide raising tones the last may develop a kind of involuntary offglide giving [fɔə] or even something more like [fɔ.ə]. Some speakers who didnt have /ɔə/ for four etc could be heard to use it with words like sure but the last time I remember hearing such a person was decades ago.

The Martin Ball 1984 JIPA article 'The centring diphthongs in Southern English — a sound change in progress', of which the author has reminded us in this discussion, had engaging diagrams which showed appealing symmetries for the Victorian GB closing and centring diphthongs. One caveat for those who may like to turn up this int'resting article of a quarter of a century ago is to take note, regarding its section on ‘Historical evidence’, that it gave first-publication dates for books it quoted but might actually give quotations from considerably later revised versions of their texts. On /ɔə/, purportedly quoting Jones (1918) but actually using a revised wording that didnt appear till 1956, it sed that he was “in his mid thirties” of the time of a remark that he actually made not when he was 37 but when he was 75. It supported what Gimson sed in his Introduction to the Pronunciation of English, (giving a page reference, 115, which was actually appropriate not for the 1962 but for the 1970 edition) by a  remark “/ɔː/ increasingly replaces earlier forms...” quoting the then current EPD “(Jones, 1977)” which was also Gimson’s work.

The same problem of dating occurred when Ball sed of /ɛə/ that “Jones (1909)” [ie The Pronunciation of English] at its page 64 mentioned that “occasionally one hears a monophthongal long ɛː”. This was actually a reference to the book’s fourth edition of 1956. Immediately after this there unfortunately appeared “However this is not mentioned in Jones (1919) or Ward (1929)”. As evidence for a suggestion that “monophthongisation of [[ɛə]] appears rarer” than that of [[ʊə]] he adduced the fact that “none of the editions of [Jones’s] English Pronouncing Dictionary note it” adding “nor does Windsor Lewis (1972)”. This was not his only inappropriate citing in evidence of my Concise Pronouncing Dictionary which was expressly a pedagogically oriented work deliberately limited in what it offered in the interests of its EFL target audience. If he’d consulted my 1969 textbook A Guide to English Pronunciation he cdve found that my opinion was that mainstream GB /ɛə/ was “generally realised as a long simple vowel (a) before consonants (b) when unstressed (c) when stressed but in a structural word”. I havent wished to revise this statement since then. The results from Ball’s modest experiment in having ten young speakers read the words hair, bear, pair, scarce and Sarah “suggests that monophthongization is ... affecting /ɛə/ in the speech of younger speakers at least...” He had only one item with /ɛə/ before /r/.

I don’t intend to repeat here remarks I made my article ‘IPA vowel symbols for British English in Dictionaries’ in JIPA Volume 33 Number 2 of December 2003 a slightly amended version of which is to be seen on this website at §5.1.9. I refer readers to the list of references accompanying that article for the 1954 transcriptions by the late Peter Strevens listed there which showed him as preferring /ɛː/ to represent the /ɛə/-type phoneme of his speech and also to the book by Laura Soames which gave transcriptions of Victorian GB for which she showed what we might now call allophonic transcriptions of words like dairy with /ɛː/. I have never understood why Daniel Jones could say of both /ɪə/ and /ɛə/ (tho oddly omitting to make such a remark in respect of /ʊə/) “... there are no phonemic variants differing to any marked extent” [sc in “RP” from his descriptions of them]. The matters of /ɪə/ and /ʊə/ will hopefully be given attention shortly.

Blog 249

The 12th of February 2010

Death Reports Exaggerated (i)

On the twentieth of Janu'ry two thous'n & ten John Maidment made another of his challenging suggestions viz that “/ɪə eə ʊə/ are doomed to extinction”. Of course he might be right. There’s no knowing, but I think some people seem to be heralding their demise a bit prematurely. There’s quite a lot I intend to say about this matter but let’s first consider John’s latest remarks at “Diphthong death again” his blog of the sev'nth of Feb'ry. This returned to the the topic with a report of what he’d noted about the speech of Dan Snow the youngish (31) presenter of a colourful BBC tv document'ry, first of a series called Empire of the Seas, of which he examined its first thirty minutes. “In the section I watched”, he sez, “there were 38 occurrences of words which in a phonemic transcription using the current de facto standard symbol set would contain the vowel /ɪə/. Dan Snow had [ɪ:] for every single one of them”. I’m sure I didnt listen nearly as carefully as he did but I cert'nly came away from to some degree trying to replicate his experience with a few markedly diffrent impressions from his.

For example I he'rd the first occurrence of the word series as having, at least as much as not, the classical diphthongal value. So also the words beer, fear and frontier. On the other hand, I completely concurred with John’s comment that Snow was particularly inclined to use [ɪ(ː)]. However, some of the [ɪ] occurrences were so short that I could see no point in classifying them as /ɪə/ rather than /ɪ/ eg at experience, experienced and a later occurrence of series. (Some readers may remember my recent remark about the comparable commonness of /ɪ/ in the first syllable of seriously in my Blog 247.) The same goes for the Spencer second occurrence of we’re [see below], a very common GB weakform of the word not recorded by British lexicographers tho ODP does list the word’s other weakform /wə/.  On more than one occasion I was undecided whether I was, when hearing a particularly centralised [ɪ] in year(s), perceiving something to be regarded as /jɪə(z)/ or /jɜː(z)/. I had a similar problem with the odd token of here. Plenty of people vary between those two values of years etc. I do. I listened to the same thirty minutes but didnt find half the tokens he did.

I listened too to the Earl Spencer (45, Eton and Magdalen) perhaps understandably unctuous encomium at his sister Diana’s funeral which John sed “contains only 10 occurrences of the target. The speaker uses [ɪə] for all of them”. I found a dozen tokens and thaut they contained only a minority of diphthongal occurrences of /ɪə/ which of course supported his contention more than his own findings. I’m not really very surprised at these differences in our impressions. For many years I’ve found the same kind of result when I’ve listened myself to recordings of speech that colleagues have transcribed minutely. For that reason I hope, if he’s going to do any more of these stimulating mini investigations, he supplies us only with recordings as accessible as the Spencer and not ones like BBC-iPlayer items that are going to be snatched from our grasp within a few days or denied to overseas readers altogether.

Spontaneous speech presents us with constant problems of identification. So many of the distinctions we’re optimistically trying to judge are in fact of a gradient nature notably, as we’ve seen, the differences between simple vowel and diphthong and between long vowel and short vowel. With these we can often find ourselves hearing items that sound exactly midway between the two possibilities. Also, since we’re discussing the development of a particular variety of English pronunciation viz General British, there’s the question of whether a speaker whose usages we’re examining falls inside that category or outside it. I find that Dan Snow (educational background St Pauls and Balliol) has so many noticeably southeastern characteristics that I hesitate on how to classify him in that respect.

Blog 248

The 8th of February 2010

A Notational Heresy

Readers of my Blog 245, the first of several I plan to celebrate OED3’s completion of its first decade, will have seen in its last paragraph a reference to the preposterous correlative use for General British pronunciations of /ᴧɪ/ in price alongside /aʊ/ for mouth, something that anyone who really respects the principles of the IPA cardinal vowels system must regard as indefensible. For the benefit of those who may not be completely clear on the point at issue, it shd be pointed out that it referred to a recommendation made in the early nineties to OUP by the distinguished dialectologist Clive Upton. He gave an explanation of it in the Oxford Dictionary of Pronunciation for Current English of 2001 at p. xii “[ʌɪ] is the diphthong of nice, try. The start-point for the unmarked BR diphthong is judged to be now characteristically in the area of the vowel but (half-open back centralized), rather than the extreme low front position [a]. The notation [ʌɪ] is that first used for RP by MacCarthy.” His expression “is judged to be now” appears to be based entirely on his personal observations. In a fairly recent somewhat protracted correspondence with me on the matter he never offered the evidence of any instrumental research findings or of any leading scholars in the field as providing any corroboration of his view.

In the book he referred to, The Teaching of Pronunciation (1978), in which the late P. A. D. MacCarthy (on the only occasion he ever used this notation /ʌɪ/ for the “RP” PRICE diphthong) included a diagram at its p. 95 which showed clearly that in his judgment the starting-point for this “RP” diphthong was best diagrammed as — so very far from Upton’s “half-open back centralized” — fully open almost fully front and indeed more front than for the MOUTH diphthong notated by MacCarthy in that book as /ʌʊ/. It’s incomprehensible how Upton could have imagined that quoting MacCarthy’s use of /ʌɪ/ for the PRICE diphthong in the particular context of that one book gave him support for his choice of that notation for it in a vitally different context. He induced the reference publications branch of the Oxford University Press to adopt this strange choice despite the fact that there was no support for the move to be had from any of the principal authorities on the subject of “Received Pronunciation”. Even within the Oxford University Press there has never been any sympathy shown for this Upton practice by those responsible for significant publications like the greatly respected Oxford Advanced Learner’s Dictionary. The same applies to the important series of dictionaries produced by the Cambridge University Press and to innumerable others.

In 2001 Professor John Wells of University College London commented on this matter in an article on “IPA Transcription Systems for English” on his personal website that was later printed (at pp 3-8) in the 2001 Phonetics Group Bulletin No. 9 of Santiago, Chile (editor H. Ortiz-Lira). There he observed:
Price. The standard notation might seem to imply that the starting point of the price diphthong is the same as that of the mouth diphthong. In practice, speakers vary widely in how the two qualities compare. In mouth people in the southeast of England typically have a rather bat-like starting point, while in price their starting point is more like cart. In traditional RP the starting points are much the same. Upton's notation implicitly identifies the first element of price with the vowel quality of cut -- an identification that accords with the habits neither of RP nor of southeastern speech (Estuary English). His choice of [ʌɪ] is really very unsuitable. In a phonetic blog at 21 January 2010 he referred again to “Clive Upton’s unfortunate choice of symbol for the price vowel in the OED/COD” citing his 2001 article and my article ‘IPA vowel symbols for British English in dictionaries’ which appeared originally in the Journal of the International Phonetic Association Vol. 33 No. 2 pp 143-152 and may be seen on this website at §5.1. 

In personal communications from various scholars over the last couple of years I have found absolutely no-one who expressed approval of the Upton practice. Among the comments I have received have been the following:
“... I agree with the theoretical grounds of your criticism” from Professor Peter Roach. “I support your dislike of /ʌɪ/ but /aʊ/... it is very illogical to do this for one but not the other” from Professor Alan Cruttenden. “It is an odd decision, indeed...For people who know the IPA and its history in dictionaries, it is sending a message that there is something wrong with the long-standing tradition of using [a] here - which is a real puzzle” from Professor David Crystal. “I thoroughly agree with you on the matter of the symbolisation of the price vowel in the OED. I deplore the use of ʌ for the first element of the diphthong...” from Professor Beverley Collins.

Blog 247

The 25th of January 2010

How do you say it?

I wonder how many readers have noticed the existence of the website supplied free by Google with the title: It’s described as a “A free online Talking Dictionary of English Pronunciation”. It tells you “When your entry appears in pink, mouse over to hear it pronounced”. You can conveniently repeat what you hear very quickly. You can list up to six entries separated by semi-colons like this: cat;cart;cut;cot;caught;coat. This is a very important facility, the comparing of items. It sez it has currently over 128,000 entries and these regularly, unlike their commercial equivalents, include non-irregular plurals and very generous helpings of derivatives, notably adverbs. You can mainly only hear the headwords of the EPD and the LPD and of the Oxford and Cambridge Learners Dictionaries and you have to buy those books to get their CD’s. What’s more, such discs are useless if you’re not a Windows customer.

I consider this Talking Dictionary to be a very valuable tool for EFL users aiming at using British pronunciations. Its creator Tim Bowyer says about its genesis: “I first created howjsay back in 2006 as a way of answering my students' constant questions beginning with "How do you say..." ” He’s British and “currently (June 2009) based in Chiang Mai, Northern Thailand”. He’s also responsible for the website “” with items of int'rest to EFL students some of which have partly foreshadowed the present work. There’s oddly no explicit indication of the fact that this Talking Dictionary for the most part gives only British pronunciations. They’re all pronounced by the compiler himself. No phonetic symbols are used anywhere, all words appearing only in their ordinary spellings. The truth is that many of his readers wd benefit from being able to see phonetic transcriptions of the entries but one can’t blame him not adding them to his gigantic labours.

He provides int'resting occasional rather than comprehensive comparisons with American usages. At circumstance, immediately and patriot he has alternative versions that are “also British”. After /ɪ`miːdiətli/ he gives the “also British” form /ɪ`miːʤətli/. As he sez the word British here he uses a medial /t/ that has such complete lack of audible aspiration that it may sound quite American to many people but in fact it’s not an unusual, if perhaps a rather “casual”, British style of articulation. By the way, no attempt is made in American versions to use any not typically British value of /t/. At garage, /`gærɑːʤ/ is followed by “American  /gə`rɑːʒ/”. His /lə`bɒrətriː/ is followed by “American  /`lӕbrətɔːriː/”. For this word so much extra nasalization is he'rd that the speaker seems to be giving an American version in respect of voice quality as well as phonemics. At /`ӕmətə or `ӕmətʃə also American `ӕməʧur/ he utters that final /r/. No American versions are offered for words like ask, corollary, dance, difficulty, past. 

Most of the pronunciations provided are very suitable EFL targets eg /`praɪvɪt, `vӕljubl, `ɔːdnri, feɪvrət, eʤə`keɪʃn, `juːʒəli/. At some items a very common variant is omitted eg /`gӕrɑːʒ/. At room only /rʊm/ is given when /ruːm/ is certainly the predominant GB current usage. On occasion the version offered may, on the other hand, even sound more idiomatic than what the dictionaries show. For example, try putting alongside each other the three items hospital, pyramidal and orbital. They all sound okay, a witness to the fact that, despite the dictionaries, the first of them most usually has in British usage final /-dl/, like the second and not like the third. Another example of a form not acknowledged by the British dictionaries is found at seriously which has for its first vowel /ɪ/, a version not uncommon in the UK. LPD gives no mention of its existence nor of that of /`kɒnsəkwensɪz/ the only plural given here and much the most usual current GB usage. Quite offen alternatives are given eg at often but not at all at temporarily, perhaps understandably. LPD & EPD tend to give the false impression that /`tempərərəli/ is an ordinary usage, if not the predominant form, and omit the very common GB form /`temprəli/.

Final -y seems very often to be stronger than is typical of GB (ie /iː/ rather than /i/) probably in the int'rests of clarity. Happy has /-iː/ and so have /`ɪndiːz, `ᴧndiːz & `efɪkəsiːz/, but many polysyllables have /-ɪz/ eg  /`kɒntrəvɜːsɪz, `sekrətrɪz, `semətrɪz/. The odd word has both endings eg /`rɪəli &    ́rɪəliː/. Derbyshire as /`dɑːbɪʃə/ is a more explicit model than LPD and EPD offer. His French items are at times too French to be realistic models for users of English eg at [ɒ̃bɔ̃`pwӕ̃, ˈsӕ̃ ˎsɒ̃s & ˈmõ ˎblɒ̃]. Mostly more realistic are /kwӕn`trəʊ/ and /`ʒɒ̃rə & `ʒɒnrə/ (to which he adds the odd-sounding “French ʒɒ̃x”) ˈvӕ̃ teɪ ˎᴧ̃, krɛʃ, kʀɛp & kʀiː də `kɜː. His Italian may be, as at /ˈӕdӕ`ʤet.tɔː/, similarly, too Italian (and his Italian’s not invar'ably accurate eg zӕbӕ`njoˑne). Not really to be recommended for exact imitation are Llangollen as xlӕn`gɔːxlən or lӕn`gɔːθlɪn, and Llanelly as hlӕ`neliː . His voice is at times reminiscent of the pleasant fruity nasality of the late Clement Freud. It’s not surprising that, with such a profusion of items, one or two here and there are not perfectly enunciated or recorded eg Shrewsbury sounds like /`ʃəʊzbriː/ and his final /l/ of meteorological sounds vocalic, not a lateral closure. Aside from such minor blemishes this is a very remarkable production giving the EFL student a lot to be grateful for.

PS Tim Bowyer explains his choices like this:
The aim of howjsay is to address the pronunciation needs of English-speaking 'world citizens' (for example a middle-class person living in Bangalore, as much as one living in Birmingham or even Baltimore).  With this audience in mind, I believe Standard British English to be generally at least as acceptable as American English, with the exception of those speakers, probably a minority, who want exclusively American pronunciations.  So I try to ... take some account of American pronunciations, while generally (but not always) offering British English as the principal model - in short, whatever seems to be the predominantly accepted usage globally


Blog 246

The 18th of January 2010

Pronunciations by ANTONIA FRASER

All my adult life I’ve been in the habit of making sets of notes on the pronunciations of individuals that have int'rested me. My personal term for these speech snapshots is “phonetigraphs”. I didnt coin that word but, finding it rather handy, I took it from Our Oral Word as Social and Economic Factor by M. E. DeWitt (1928) a curious, eccentric book on which some comments may be found at my Blog #010. These notes have most often been made of people on whom plenty of speech-background information has been available, mainly speakers of GB (General British ie unlocalised-England-type speech). Lady Antonia Fraser fits nicely into that category: well-known school, Oxford college and parents. She was born in 1932 and her close personal contacts and main localities of residence are well documented. My spur on this occasion was five 15-minit samples of her current speech in her reading as BBC Radio 4 ‘Book of the Week’ an abridgment of her account of her life with her late husband the playwright Harold Pinter entitled “Must You Go?” made available for listening at any time during the week following their original transmission (but unfortunately no longer).

As is to be expected she exhibited a variety of usages most of which are mainstream (at any rate for her generation) but some of which might be called modernisms and others traditionalisms. An item that struck me immediately was her regular use of /ɔːf/ for off. The lexical group that this word belonged to in Victorian GB (which cd be called the cross set) consisted of words like broth, cloth, cough, cross, lost and often which mainly contained “o” before a voiceless fricative consonant. It mus'nt be imagined that ev'ry word with that sequence joined in the development: on this side of the Atlantic at least there’s no evidence of eg coffin ever belonging to the cross group. EPD1 in 1917 lists /ɔː/ only as subvariant in coffee. The change wd seem to’ve developed around the end of the eighteenth century and to’ve become no longer mainstream GB by the end of the first quarter of the twentieth. It’s not undergone in GA the return from /ɔː/ back to a shorter-type vowel that occurred to it in GB.

By 1937 the EPD was showing eg /krɔːs/ only as a subvariant. Things like that don’t change suddenly. The commonest words were last to give in to the change back to /ɒ/. The only explanation for the reversion I know of is the one I offer at 3.7.III.15 on this website. LPD rightly gives /ɔːf/ as an existing subvariant and it isnt as unusual as people may think. In fact its history is complicated by the unusual process by which /ɔːf/ for most who use it doesn’t have the predominant /ɔː/ value but is rather something of a lengthened [ɒ]. Those who do use it with a mainstream or closer value of /ɔː/ tend to sound markedly old-fashioned-posh. Our speaker’s /ɔː/, while not completely consistent, doesnt really put it into this last category.

Among other slightly old-fashioned usages for her background is her tendency to use quite a few tokens of /ɪ/ eg in Helen and the -less and -ness suffixes as in darkness, helplessness, recklessness, sadness and breathlessness. In the last of these at least, the medial syllable is, as one may well expect, schwa. Schwa also occurs in rather modern style medially in celebration and def'nitely and in the initial syllables of resounding and eclipse. Another older-type usage is the yod-keeping at times at absolutely which nevertheless she may treat with quite modern syncopation as /ӕbsjuːtli/ as well as /ӕpsəluːtli/. She also sez /sjuːtəbl/ for suitable but lured has no yod.

I noted /wɔːmpθ/ very clearly for warmth and /mə`kɪzməʊ/ for machismo which suggests that Greek is more familiar to her than Spanish. Here are some of her usages that were mostly quite unremarkable mainstream current forms but arnt necessarily acknowledged as being so in the descriptive texts and pronouncing dictionaries. These include /`ɒbvɪsli, mӕnɪʒ tə..., deɪ`tɔːnt, t wəz..., ðɪ wəz ə..., sᴧdni, əʊni & tests, tests, test/ ie obviously, managed to, détente, It was (actually a full house), there was a, suddenly, only and Tests, tests, tests. If not on the first two articulations of tests, at least on the last one she produced the totally commonplace version of the plural of test with its final /s/ not produced.

An int'resting rather posh but faded feature (far from mainstream if it ever cdve been so categorised) is her tendency to low rhoticity. She had no r-links in bore in mind, an hour at and were installed. Also slightly more posh than mainstream are her smoothings at flowers, hours, powerful and casual /`kӕʒʊl/. For one moment I didnt recognise /`ʌmpɑːz/ as umpires perhaps coz the /ɑː/ was rather short. She has some notably “modern” cavalier delivery of some items like /ʌn`rekənaɪzd, kjʊə[ʊ]ri`ɒsti, `aɪsleɪtɪd, ɪk`strɔːdni, tu ə`luːsɪneɪt/ and /drɔːŋrəm/ ie unrecognised, curiosity, isolated, extraordinary, to hallucinate and drawing-room. On one occasion apparently was /ə`pӕ[ː]ntli/ and on another probably was so rapidly articulated as to be unclear but was quite like /`prɒl(l)i/. These were manifestations of the attractively easy and unfussily fluent stylishness of her delivery.

Blog 245

The 13th of January 2010

A Decade of OED3 Pronunciations

Amongst the greatest of our British national treasures, the glorious Oxford English Dictionary is now better than ever and steadily, if inevitably rather more slowly than one naturally tends to wish, undergoing a new admirable revision under the directorship of its Chief Editor John Simpson. Since their first appearance in March 2000, the on-going Online revisions now for the first time incorporate regular and full representations of what’s usually understood by the well-known term 'General American' (pronunciation).

The Cambridge EPD acknowledges that its American entries are “similar to what has been termed 'General American' ” but tends to use the term 'Network English' and offers no illustrations of what might be differences between the varieties. OED3 declares “The pronunciations given are those in use among educated urban speakers of standard English in Britain and the United States. While avoiding strongly regionally or socially marked forms, they are intended to include the most common variants for each word.” Regarding the surprising word “urban” it’s hard to imagine anything meaningful that it adds to the description. LPD uses the term 'General American' referring to it as the speech of those “who do not have a noticeable eastern or southern accent”.

Both OED2 (the second edition) and the new OED3 (third, revised edition) as so far completed, ie M to much of R and great numbers of miscellaneous non-consecutive items, are accessible online for most people in the UK who have a public-lib'ry ticket. The pronunciations shown in the second edition, as with OED1, represented only British usages. It referred to them as “those in use in the educated speech of southern England (The so-called ‘Received Standard’)”. This description was how the great Robert Burchfield headed the key to pronunciation symbols in 1972 in the first of his four massive volumes of OED Supplements. ‘Received Standard’, not an expression much favoured by British linguisticians in the latter twentieth century, was a term introduced in 1913 by the remarkable Oxford scholar Henry Cecil Wyld. It was no dou't one of the influences that led Daniel Jones into adopting the unfortunate label 'Received Pronunciation'.

The OED2 pronunciations, which appear (as they did in OED1) between round brackets, replaced the orginal 1884 pre-IPA symbols formulated by the revered James Murray. He omitted to specify in the dictionary exactly whose speech he was recording but indicated more or less obliquely that he had in mind mainly, tho not exclusively, the educated usages of “natives of the south of England”. For example they, he commented, wd regard as “the same sound” two vowels which his notation distinguished in eg the words fur and fir. He also had for words like bath, pass and dance a cover symbol (ɑ) to be interpreted as either (æ) ie /æ/ or (ā) ie /ɑː/. Also he had for words like soft and salt a cover symbol (ǫ̀) to be interpreted as either (ǭ) ie /ɔː/ or (ǫ) ie /ɒ/. This last practice, as with the use of differential fir/fur symbols, has become out of date and is discontinued in OED3.

For OED3 we’re told: “Each pronunciation in the revised text is given in the International Phonetic Alphabet (IPA), according to a revised model of Received Pronunciation devised by Dr Clive Upton of the University of Leeds, and the scope of this information has been extended to include a ‘standard’ U.S. pronunciation based on a model devised by Professor William Kretzschmar of the University of Georgia.” These so-called “models”, which are in some degree quite controversial, are discussed in the review of the 2001 Oxford Dictionary of Pronunciation to be seen at our Section 12 Item 5 on this website. Unfortunately the very satisfactory set of phonetic symbols of vowels of OED2, which have long been in wide general use, have been subjected to several changes of dubious value discussed at our Section 5 Item 1. It appears that Canadian items will use a set of symbols not yet specified but to be seen to be beginning to be used at the entry maidenerBrit. /ˈmeɪdnə/, U.S. /ˈmeɪdnər/, Canad. /ˈmeidnɜr/”.

It’s difficult not to feel some regret that the transcriptions adopted werent planned more to reflect agreements between the two varieties rather than to emphasise differences. Sometimes irritations are occasioned when transcriptional contrasts seem to be more motivated by the preferences or habits of the one transcriber rather than the other and not a genuine reflection of an audible difference between the two varieties. For example US entries have /i, ɑ & u/ but British ones have /iː, ɑː & uː/ tho no important contrasts of length are involved. /ᴧ/ wdve functioned perfectly well as a representation of the short open-mid central vowel for both. /æ/, being a cover vowel for a range of US values, cd well be used also for the corresponding British phoneme as is done in EPD and LPD. There seems no point in stress-marking post-tonic strong syllables for one variety and not for the other eg as at minutiae. It’s quite unrealistic to show in countless words like alien the penultimate vowel as /i/ in one and /ɪ/ in the other. See also our Blog 174. There are words like able given as ending /-bl/ for one but /bəl/ for the other. There’s not much point in representing the price diphthong by /ᴧɪ/ in one and /aɪ/ in the other. The most regrettable decision of all is the preposterous correlative use for British pronunciations of /ᴧɪ/ in price alongside /aʊ/ for mouth, something that anyone who really respects the principles of the IPA cardinal vowels system must regard as indefensible. It’s to be hoped that in time this matter may come to be reconsidered. It’s a massive task to transcribe so very many and such various words but in its treatment of pronunciations OED3 is to say the least more thoroughly and satisfactorily equipt than ever before.

Blog 244

The 11th of January 2010

Elision of Yod (ii)

Part (i) of this discussion de’lt mainly with what were referred to as “tonic’ly strest syllables”. Non-tonic syllables tend to behave differently. Perhaps rather paradoxically, because one might’ve expected to lose yods from them, weaker syllables often retain yods. In GA words like avenue, revenue, retinue, altitude, attitude, institute, hypotenuse etc all seem to show two possible degrees of weakness, the more usual ending without, and the presumably weaker with yod, thus eg /`ævənuː & `ævənju/. GB has only the latter type. Yods are mostly maintained in both GB and GA in words with a single unstrest final syllable like Agnew, Cardew, continue, curlew, granule, module, prelude, value, venue and purlieu. This last may also be /`pɜːrlu/ in GA but any British such yodless form is so unusual that it no dou't qualifies for its LPD “§” which categorises it as alien to GB. Capsule and pustule in GA have at least variants with weakenings to second-syllable schwas. GA and GB have /`ferəl/ as a possibility for ferrule.

A notable contrast between GA and GB is the characteristic absence from GA in words like tune and dune of the yods they must once have contained. Speakers of varieties of GA that havnt dropped such yods seem to be in a minority of something like 10%. (See the survey results given in LPD3 at eg due.) GB has never entirely lost the yods in such words. It retains reflexes of them in the form of variant pronunciations like /`ʧuːn & `ʤuːn/ which alternate with /tjuːn & djuːn/. There’s currently no certainty regarding which version can be sed to be predominant. In non-tonic syllables only, GA has affricate variants besides yod-dropping and yod-retaining versions. For example, compare GB aperture as in LPD /`ӕpəʧə, `ӕpətjʊə, `ӕpəʧʊə/ with GA as in Webster Online /`ӕpərʧʊr, `ӕpərʧər, `ӕpərtjʊr, `ӕpərtʊr/.

GB and GA have no words that begin */rjuː-/ or */rjʊə-/ or in which either /ljuː/ or /ljʊə/ is preceded by any consonant. Hence there’re no yods in rude, rural, truth, blue, glue, plume, sleuth etc. In varieties of Cymric (ie Welsh-language-influenced) Welsh accents, and perhaps still in recessive pockets of Yorkshire dialect, items like /kljuː/ for clue and /fljuː/ for flu may occur as reflexes of a dialectal [ɪu] diphthong converted to [ju] in adaptations aiming at GB. Some very minor and predictable dropped yods occur where front close vowels as-it-were absorb them in items like beyond, re-union and see you which very easily become /bi`ɒnd, riː`uːnjən & `siː uː/. There’s a little more along these lines in our Blog 149; Blog 126 mentions the yodless babytalk form of you.

Besides the historical losses of yods as in Ferrer (earlier Ferrier), Villiers /`vɪləz/ and other items already mentioned, there are many that are currently generally not recognised but are nevertheless extremely frequently to be he'rd. Among the most common types are those regarding words in which compressions may follow a previous stressed syllable, especially if it can be sed to end with an /r/, such as serious and variable which are often to be heard as /`sɪərəs & `veərɪbl/. It seems that, in common words with more syllables to be articulated in the time the speaker allots to the word, like seriously, previously and invariably the reduced forms /`sɪərəsli, `priːvəsli & ɪn`veərəbli/ are much the more common versions to be he'rd, despite the lexicographers ignoring of them.

George Eliot in Adam Bede (1859) put cur’ous into the mouth of one of her characters (not a ‘gentleman’): it certainly isnt sanctioned by current lexicographers but one wonders how often it occurs quite unnoticed. This applies also to auxiliary as /ɔːg`zɪləri/. Like /staɪ`pendri/ for stipendiary, it’s warned against by an LPD traffic-style triangle but it’s accepted by Webster Online. So many people are so vague about such things that the “hyper-corrective” form of ancillary /æn`sɪləri/ as /æn`sɪljəri/ is common enough to be listed in LPD tho labelled “§” as alien. Other examples one may give are experience as /ɪk`spɪərəns/, immediately as /ə`miːdətli/ (less common than /ə`miːʤətli/), conciliatory as /kn`sɪlətri/. Coal miners used to mainly refer to collieries as /`kɒlriz/. Very common, among weather forecasters at least, is the reduction of areas to /`eərɪz/. Also of common occurrence, but no dou't disapproved of by purists who happen to notice them, are numbers of polysyllabic words with, in their full forms, medial /jə/ etc from which the yod or the whole syllable may be elided. Examples are /`æmbləns/ for ambulance (LPD “§”), communist often made homophonous with commonest, and manufacture (LPD approved). Perhaps too casual for dignified use are what we may spell as merc’ry, partic’lar and reg’lar.

There are a few anomalous items whose irregularity isnt easy to explain including notably the yodless forms of lugubrious and recuperate. Communication as /kəˈmuːnɪ`keɪʃn/ is less common and, with its loss of yod in a non-tonic position, perhaps a little less surprising. The acronyms BUPA /`buːpə/, FISU /`fiːzuː/ and scuba (praps not universally recognised to be derived from self-contained underwater breathing apparatus) are similarly irregular or variably so. The trade term Aquascutum is given in EPD with an italic ie omittable yod tho not so in LPD or ODP. The only other yod losses of this sort one encounters are likely to be dialectal unless they are attributable to use of a relatively not completely naturalised borrowing from a forren language as for those who might  favour the yodless variant of tuna a Spanish loan. Traditional-dialect speakers of East Anglia have more dropt yods than probably any other forms of English. Readers of Dickens may find him representing their speech by spellings like “pecooliar” eg in Great Expectations.

Blog 243

The 7th of January 2010

Elision of Yod (i)

Elision occurs most notably to the four or five (the latter if you subscribe to the late Ian Catford’s categorisation) approximant consonants of English, /l, r, j, w/ and /h/ ie /el, ɑː, jɒd, wɪn/ and /eɪʧ/. Resuming the Maidment thread I now turn to yod-dropping. This goes back even to the pre-history of Old English. For example, by inference from other Germanic dialects it’s evident that OE had a suffix [jo] whose yod, when added to a word with a back vowel, caused its (final consonant and) vowel to become fronted and then itself became elided. This was how we came to have different vowels in sale and sell, tale and tell etc. The past forms of the verb to hear must have reached their present yodless form heard /hɜːd/ via an intermediate (so far as I know) unrecorded */hjɜːd/. The form /hjəː/ for hear and here was listed in every edition of the EPD that Jones was responsible for (up to 1956). It was a misjudgment by Gimson to remove them in 1964 and, altho he’s been followed in doing so by subsequent pronunciation lexicographers, they’re still to be he'rd quite commonly without attracting attention as old-fashioned.

Prob'bly the oldest couple of items in our vocabulary that still retain a letter that signals the former yod’s presence are carriage and marriage which I have no recollection of ever hearing as anything but /`kærɪʤ/ and /`mærɪʤ/. It need not be imagined that the tendency to lose a yod in such contexts no longer exists. Words like carrying, hurrying and worrying are often to be heard, especially phrase-internally, as /`karɪŋ, `hᴧrɪŋ & `wᴧrɪŋ/. The words miniature and parliament are overwhelmingly often /`mɪnɪʧə/ and /`pɑːləmənt/ but they both have rather archaic or pedantic variants occasionally to be heard like /`mɪnjəʧə/ and /`pɑːljəmənt/. The effect of the variant /`pɑːlɪmənt/ to some extent depends on how it’s spoken: if briskly it sounds hardly different from the normal version. This applies similarly to words like championship which is often he’rd in a form perfec'ly realistic’ly transcribable as /`ʧampɪnʃɪp/. The word accompanist nowadays usually only /ə`kᴧmpənɪst/ is no longer very often spelt accompanyist: it has of course lost a yod.

It’s observable that many words where the consonants /t, d, n, θ, s, z, l/ may precede /j/ followed in a syllable by a close back vowel tend to take forms in which a yod has been or often is elided or absorbed into an affricate. Of the standard Englishes GA (General American) is the most frequent yod-dropper. Words of the types tube, tumour, tune, tulip, tulle and  due, dubious, during, duty, deuce and new, newt, neutral, knew, nude, nuisance have no surviving yods even in learned etc words for the great majority of GA speakers whereas in GB (General British) new-types all have yods and the tube and due types a yod or its reflex. The more common words have long tended (it now seems increasingly) to convert their yods into the latter elements of affricates giving rise to homophones like dew and Jew. The less common ones are prob’bly more inclined to preserve their yods as with adduce, adieu, subdue, attune, étude etc. Undou'tedly enthusiasm and  enthusiastic very often have no yod but that’s prob'bly less true of Lithuanian and Methusaleh and pretty certainly thew, Arthurian, thurifer, Thucydides and Thule can be sed to usually have yods. During the course of the mid twentieth century pronunciations with yods of the words sewage, sewer, suet, Suez, suit, suitable suicide, pursuit, suture, pseudo- and numerous items beginning super- became more and more unusual. I shd guess in many quarters hearing the pronunciation /`sjuːpəmɑːkɪt/ wd now be an occasion for mirth. And the faded slang Super! uttered with a yod reminds one irresistibly of old-fashioned speakers like the genteel-comic characters amusingly once played by the actress Joyce Grenfell.

There are few words in current not-old-fashioned-sounding GB which are predominantly he'rd with tonic'ly stressed syllables containing either of the sequences /`ljuː/ and /`ljʊə/. Subvariant forms with yod are he'rd mainly for words that arnt necessarily in everybody’s everyday vocabulary such as elude, lewd, lucid, lucrative, ludicrous, luminous, lure, lurid, lute. Words and names like Lewes, Lewis, Lewisham, Lucas, Lucy, ludo, lukewarm, lunatic, Lusaka, Luther, Luton have no longer any truly current variants with yods.

A remarkable unique example of the reversal of the GA preference for dropping versus GB preference for yod retainment is the word figure which is invariably /`fɪgə/ in GB but only sanctioned as /`fɪgjər/ in GA. (I’m not taking account of the word fritter in this generalisation simply because spellings of it like friture have been completely abandoned since the sixteenth century. Compare dialectal 'critter' from creature) The reverse of this pattern is seen in the recognition of the existence of the GA yodless form /`foʊlɪʤ/ of foliage in the US dictionaries and ODP. Webster Online doesnt represent it as the predominant usage and even prefixes it with its cautionary ÷ signifying rather neatly that it’s a usage on which opinion is divided. Curiously the online speaker who demonstrates two versions of it gives the yodless one first. Despite its absence from the GB-recording dictionaries, the yodless form is so commonplace that one dou'ts if it ever attracts any attention when used.

Blog 242

The 6th of January 2010

Irrational Spellings of Names

The admirable Amy Stoller recently made this comment:
Davies is always pronounced ˈdeɪviz in the US, at least in my experience. Not to do so can lead to accusations of affectation. Difficult to cope with when one knows that the person in question pronounces their own name ˈdeɪvɪs. After all, they should know, shouldn't they? But there it is.
I can’t quite agree with this conclusion. As I see it a name is a word that’s part of our common language and I don’t feel that anybody can reasonably be said to "own" any word even if it does happen that they are pleased to use it as their name. I don’t see why I shou'dnt say any word as I prefer so long as I don’t positively cause inconvenience or incomprehension. So if anyone insists on using a very unhelpful spelling for their name, I don’t think they have any right to be affronted or even object if other people say it in a way that’s a rational interpretation of an irrational spelling. After all it’s within their power to decide to change the spelling of their name.

I’m not saying that they’re doing anything wrong but I sometimes feel mildly irritated and recalcitrant to contemplate the way the BBC’s Pronunciation Unit is at quite such pains to make sure their clients know just how various, especially posh, people like to have their names spoken. Large numbers of entries in their publications have mentioned alternative versions with the addition that an entry is “appropriate for” a certain  Baron, Earl, Viscount etc. This is of course a contrary view to Graham Pointon’s. In one of his blogs he sed 

While a personal name is “owned” by its bearer, a place name can be said to be “owned” by its inhabitants. This was always the BBC’s reasoning for advising the pronunciation of British place names that was favoured by local people, [actually by the local gentry not the peasantry] and why I stuck to the recommendation for Althorp that was used by Earl Spencer and his family even when the senior management of the BBC insisted otherwise. I, as Pronunciation Adviser, wrote to the present Earl Spencer (Diana’s brother) in 1992, and in January the following year, he wrote back saying [in a scrappy handwritten note]. “This [ɔːltrəp] is definitely correct. I can remember my grandfather pronouncing it like this ... it is clear that alternative pronunciations only came about recently, out of laziness (it became simpler not to correct the many who mispronounce it) ...” However, some time after this, he succumbed to the pressure, and put out a press statement saying that henceforth the house should be called ‘áwlthorp’ – as spelt.

To my mind he did the sensible thing to sanction his home’s being called /`ɔːlθɔːp/. The people behaving unsuitably were those who wanted it to be pronounced /ɔːltrəp/ but couldnt be bothered to bring their spelling of it out of the Middle Ages. The map of England is littered with thousands of names that people have been too lazy or unenterprising or misguidedly sentimental to bring up to date orthographically to the puzzlement, dismay or irritation of so many who weren’t braut up on the spot. Something similar goes for our wicked general spelling too.

I’d alre’dy written much of this when I noticed that John Wells’s blog of today drew attention to the fact that a Sky News television reporter speaking at Marlborough to local people had, unlike them, called it /`mɑːlbrə/. Well, the cigarettes are always called that and no dou't more people know something of them than of that Wiltshire town. John concluded his remarks on it with: "Mind you, just why the usual BrE pronunciation of Marlborough ... has ɔː rather than ɑː is a question to which I don’t know the answer." Perhaps it might help to recall to him that people dropped pre-consonantal /r/ not all at once and everywhere but at various rates and places. It so happened at Marlborough that there, after they dropped its pre-consonantal /r/,  the word came in for the change from /ɑːl/ to /ɔːl/. This last change apparently happened for most people before they dropped such /r/ sounds. PS 7 Jan: Commenting on the Wells blog, "Warren" made the excellent point that the early dropping of the first /r/ of Marlborough was very likely induced by the kind of dissimilative process to be seen operating in the tendency of many (high-rhoticity) Gen Am speakers to drop the first /r/ in words like farmer and surprise.

Blog 241

The 25th of December 2009

Current Wyn Dropping

When John Maidment introduced the topic of elisions the other day he was on that occasion inclined to concentrate on synchronic processes but, having been prompted to discuss the subject, it was difficult for me to resist contemplating the diachronic aspect. That I did with ‘historical wyn-dropping’ my last blog #240. Having got that off my chest I now turn to synchronic matters. Wyn-droppings that can be presently he'rd daily are pretty numerous tho it’s often hardly possible to decide whether a speaker’s departing from a normal form of a word or hes'tating between employing one or the other of a pair of competing forms one of which is wynless. On another aspect of the subject, no-one is an example of a word most speakers wd make /`nəʊwʌn/ when uttering it firmly whereas saying it unstrest within fluently articulated phrases they seem, to me at least, as likely to say /`nəʊən/. This is not the sort of thing that the pronouncing dictionaries make much of an attempt to record, praps reasonably so.

A word they do signally fail to deal with realistic'ly is the problematic co-operate. The only form they all offer, /kəʊ`ɒpəreɪt/, doesnt even seem to be the predominant form people use — at least in relatively unselfconscious fluent utterances. The strest vowel is quite offen ambiguous but probably the most usual versions are /kəʊ`ɔːpəreɪt/ and /`kwɔːpəreɪt/. A commonly occurring further development of the latter is the wyn-dropping form /`kɔːpəreɪt/. This in other contexts might be come across as a pedantic substitution for the normal form /`kɔːpərət/ of the word corporate. This range of forms clearly betokens the fact that speakers often tend to be vague about the difference between such words. In demotic speech I’ve he’rd the abbreviation Co-Op /`kəʊɒp/ turned into /kwɒp/ and occasionally even /kɒp/.

The French borrowing mademoiselle is usually he'rd without its wyn as some variant like /mӕdəmə`zel/ but for prob'bly the majority this’ll no dou't be one of its possible weakforms used before a name whereas they might usually pronounce the word away from such contexts as /mӕdəmwæ`zel/ or some  variant of that form.

The words toward and towards have nowadays had what were formerly their usual forms /tɔːd/ and /tɔːdz/ replaced by the obvi'sly spelling-influenced /tə`wɔːd/ and /tə`wɔːdz/ tho they do survive chiefly as informal weakforms. The form /`ɪnədz/ has now replaced /`ɪnwədz/ for inwards in the meaning “insides” (ie internal organs or intestines). In writing it’s practically always  innards and it’s usually perceived as totally informal. The earlier form /`ɔːkəd/ of awkward has now disappeared from standard usage tho it may well occur as a casual weakform used by some. Another highly colloquial usage is /`fɒrədə/ in the sense ‘more forward’ ie making progress. OED under forwarder has a 1918 quote from Wilfrid Owen who in a private letter wrote: “Am no forrader with my Chest of Drawers. The man won’t sell as it is, & says he has no time yet to work on it”.

Another colloquialism is ’un, a weakform of one as /ən/. The OED seems to regard it as dialectal but it’s not merely a localism. It may sound slangy but Google had 462m hits for good ’un. A British popular newspaper (as appeared from Mail Online) had the recent heading “Will Eastenders’ Denise marry a bad ’un?” in reference to a BBC soap opera. Its currency seems to be limited to these two phrases.

A couple of nautical wyn-dropping items are boatswain and gunwale. The alternative spelling of the former is “bo’sn”. Its spoken form /`bəʊtsweɪn/ is far less usual than /bəʊsn/. On the latter OED sez “The usual spelling is still gunwale, though the pronunciation (`gᴧnweɪl) is, at least in Great Britain, never used by persons acquainted with nautical or boating matters”. Quite! Such people say /`gᴧnl/ and often spell it gunnel.

In my childhood I normally used /`penəθ/ and /`heɪpəθ/ for pennyworth and halfpennyworth which had also the more formal versions /`peniwə(ː)θ/ and /`heɪpniwə(ː)θ/. Demotic Cardiff (and other South Wales) usages familiar to me then included ’ole ’oman /oʊl `ʊmən/ for old woman and ’on’t for won’t (attested for Cockney too and no dou't not uncommon in other dialects).

In current GB usage the best known examples of alternation between forms of a word which respectively retain and drop their wyn are of course will and would which have the weakforms /(ə)l/ and /(ə)d/. Much less noted, and indeed quite unimportant for the EFL user, is the fact that “to” has the occasional weakform /tw/ which now and agen can be he'rd to drop its wyn in fairly negligent utterances like /aɪ fgɒt t(w) ɑːsk ə/ I forgot to ask her and /aɪd əv `laɪk t(w)əv `ˏsiːn ðm/ I’d’ve liked to have seen them. Altho no GB speaker wd ever say /tə`mɒrə/ in isolation for tomorrow,  the word has weakforms including the one /tə`mɒrʊ/ as in /təmɒr`wiːvnɪŋ/ for tomorrow evening (in which syllabication may be vague) from which the wyn may be elided giving /təmɒr `iːvnɪŋ/. Similarly the word borrowing may often be he'rd as /`bɒrwɪŋ/ and even, mainly in mid-phrase, as /`bɒrɪŋ/.

LPD and EPD both record the common variant of quarter as /`kɔːtə/ which for many speakers is prob'bly really only a weakform. Only LPD registers the existence of the form /kəʊ`teɪʃn/ of quotation but then only with its “not RP” sign “§”. It may well be only a weakform in regionally-neutral speech but I don’t think it’s only a regionalism. I wonder how far the extreme weakform /ɔːlz/ is current among GB speakers for always but I don’t dou't its existence.

The word quoit has, apparently only as a subvariant form with wyn inste'd of the usual form without wyn, the pronunciation /kwɔɪt/, on both sides of the Atlantic. This form is probably an example of a spelling-influenced wyn insertion rather than wyn dropping.  There don’t seem to be any very notable differences between GA and GB in regard to wyn dropping. The OED-listed subvariant form of turquoise given in ODP for GA as /`tərkɔɪz/ doesnt seem to have survived in GB from the times when it existed in both varieties, but it’s not clear whether or not that variation shd be categorised as an example of wyn dropping. American names like (that of Diane) Warwick, which is normally he'rd as /`wɔː(r)wɪk/ at least from US speakers is likely to strike GB speakers as pretty alien.