Archive 33 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

Catching Spoken Words

Users of English as an extra language offen worry unduly about not recognising what words native speakers of English have used when they listen to them on television, in cinemas or from recordings. As a recent correspondent bemoaned to me: ... "failing to catch them all, we think we cannot understand .. the native speaker .. in full".
I told her that she shd take comfort from knowing that very large numbers of intelligent well-informed native speakers of English have a good deal of the same reaction as she does. Many people prefer these days to watch dramas and movies, especially recent ones, making use of any subtitles available. The fact is that people like film directors in recent decades have increasingly come to place a greatly higher value on realism and excitement (not to say violence!) than on intelligibility, so they dont care how noisy surrounding sounds may be or how deafening 'musical' backgrounds are. They take it for granted that most of their — the statisticians tell us mainly rather young — audiences cou·dnt care less if they can't catch every word that they hear.

By the way, even subtitlers dont by any means necessarily give you exac·ly the words that've been spoken: it doesnt matter to them as long as the message is the same. Subtitlers mainly cater either to people who dont understand English at all or have hearing which is defective; also they're quite understandably at pains to keep the subtitles as short as possible. Even so, I recommended her to take advantage, whenever she can, of looking at movies on DVDs which have subtitles. I've before now on occasion been asked about a word or sentence in a film by colleagues who are very expert teachers of English in their own country when I've been able to elucidate it for them not necessarily because I can hear it more clearly than they do but because I've found myself helped to guess what it must be from the linguistic and/or cultural context. Native speakers, I suspect, generally under-estimate how offen they recognise what's being spoken from such contexts more than from clearly hearing it.

I find I also greatly prefer to watch all television news programmes with subtitles on. They're so often full of unfamiliar names of people and places from all around the world not always pronounced very satisfactorily by correspondents and presenters. And outside broadcasts in particular have very variable standards of audibility. It's so usual these days to be told of some local conditions or happenings from someone perched in a howling gale on a mountainside or beside a busy motorway. I suppose we must be grateful that, so far at least, it's usual for weather forecasters to tell us what storms or other painful things are coming our way from the calm of a television studio without necssarily having them accompanied by illustrative enactments of the horrors to come. One of my pet hates is the trick producers etc have of filming a correspondent who delivers his spiel to a presumably distant camera from right in the middle of a dense surging crowd none of whom seem to notice that he's (it's usually "he", I'm glad to say) shouting his he·d off. It  reminds me of the bit of the play Look Back in Anger by John Osborne I quoted in my book People Speaking. You may like to hear it delightfully delivered here by the actor who recorded it for me. Rant over.

Re-Occurring Content Words

A British teacher of EEL (ie English-as-an-Extra-Language), who recently first quoted something that he presumably considered to be generally accepted, viz
A content word will be spoken on its first occurrence in a stressed or prominent form, and on second and subsequent mentions it will be unstressed or non-prominent
followed it with the question
Is there any published research concerning the validity of this first mention/second mention pattern?

Accepting his preliminary assertion as broadly true, I shd say that I dou·t if anything much has been or praps even cou·d be written about whether the description is valid because it's so self-evidently true that a tendency of this kind is the common usage of the vast majority of NSEs (native-English-speakers) worldwide. What people cou·d investigate is how far this absolutely undeniable tendency is actually conformed to in ordinary spontaneous speech. The result of such an investigation might well be, I shd guess, that compliance with these "rules" was managed by the average speaker only 80 or 90% of the time. I've certainly offen he·rd the "rules" broken. Also I guess that the "scores" of "compliance" wd vary very considerably between individual speakers. Examples of "malpractice" in this respect wou·dnt be worth giving because of being so obviously in that category. I've offered some comments on this topic in my article 'Accentuation' at Section 8 of this website. This gives plenty of examples of matters such as how the basic rules may at times seem to be broken but need not necessarily be so regarded eg when what I've called "semantic re-focusings" and "animation stressings" occur. It's here.

That great pioneer in the description of the phonetics of English, Daniel Jones (1881-1967), made the earliest clear statement I know of on this topic, in the simple comment which appeared originally in the first edition (1918) of his Outline of English Phonetics and was identicly repeated in all subsequent editions, in a section on "Sentence Stress", from 1918 §657 to 1956 §965 "... when a sentence contains a word that has been used just before, that word is generally not stressed." That's not quite how one usually expresses the idea today. One brief way of putting it is that "speakers in a discourse usually avoid re-accenting relatively immediately re-occurrent words or synonyms for them ". We may define an accent (or 'accentual stress') as one consciously and voluntarily accorded by a speaker to a particular word or syllable. It shd be noted that a pitch movement which is a latter element of a complex tone, even if so delayed that it is effected on a subsequent word, does not constitute an accent.

I regretted that I didnt feel that I cd offer any ans·ers to his question satisfactorily because it wasnt clear to me how far he was seeking a reply to an intellectual enquiry or if he was actively looking for ways of helping his students. I noticed that some people who responded to his appeal mentioned books by Halliday and Brazil that purported to be aimed at the EEL learner but in which I dont remember ever coming a cross a single mention of any real-life EEL students' problems. If one wants to sharpen students' lis·ning perceptions, then I'd say the only procedures I'd recommend wd be the same ones as one wd use in trying to help them with their own speaking performance. If students can understand when stresses etc are appropriate in their own speech they seem likely to be able to appreciate their uses by NSEs. The best way for any practicly-oriented teachers to get good advice is to give their advisers concrete examples of what goes wrong in the performances of their students.

Demarcation Revisited

Recently we suggested that to convincingly estimate the number of speakers of unadulterated GB/RP is hardly possible becoz the demarcation between that and other varieties of pronunciation is too offen unclear. Illustrations of this were braut to our atttention when Graham Pointon mentioned the John Wells blog entitled "One" which referred to "a trend towards a preference for wɒn over wʌn" among the respondents to his questionnaires. In that blog Wells remarkt:
 "The fact that more younger people than older report a preference for ɒ in one and for æ in chance can be seen as a greater willingness on the part of northern respondents to report a preference for their own pronunciation in cases where it is known to deviate from the perceived norm (RP: wʌn, tʃɑːns)".

One's reminded of the Oxford Dictionary of Pronunciation so-called "model" Upton put forward in his arbitrary re-definition of the term "RP" to include the Northern regional short-vowel form /ʧans/ as an 'RP' variant while not according similar recognition to eg once, one, none, nothing in their variants with /ɒ/ or to other various widespread North-of-England educated usages. See more on this topic in the present website at Section 7.4  and Section 12.5 §§ 21 & 22.

Graham sed he'd "like to propose an additional explanation" for the apparent changing views of the Wells younger respondents regarding acceptance of this kind of item. This was that "the group of words that John classifies as [his list #] 59′ .. can easily lead to confusion in the minds of SBE speakers". This group, in which GB (ie General British aka SBE aka RP) speakers may have either the trap or the bath vowel, Wells exemplified by about 40 items such as "chaff, graph, Basque, elastic, exasperate, Glasgow, lather,  substantial,  transport, plaque and Cleopatra".

Graham also asked "Could it not be .. that the sound change may be reversing itself .. and, perhaps because of the influence of American English .. but also from the influence .. of Northerners, .. be raising the uncertainty levels of Southerners..?" I agree that these may perfectly possibly contribute something to the process under discussion. However, I consider that the most important factor by far involved in preferences such as for /a/ over /ɑː/ in words like chance is simply analogy. In the great majority of English words containing the spelling <a> followed by a consonant in a strest syllable the <a> corresponds to /a/, so the speaker's natural inclination is to make these words conform to the regular pattern, producing what are generally called "spelling pronunciations".

Graham added "I have heard TRAP pronunciations in many supposedly BATH words from speakers of otherwise clearly southern varieties of British English, who were born and brought up in the southern part of England". One notices that he mentions no examples of the reverse pattern of using the version that runs counter to the notional value of the spelling. When (rarely) that pattern does occur, I suspect the speakers of hyperadaptation to the variety perceived as more prestigious than their childhood customary ones. This one finds in such words as substantial and circumstances which, I hope reliably, I associate with people like Baroness Thatcher who wou·dnt've been likely to use /ɑː/ in them in her Grantham days.
 

Evolving English concluded

To return one more time to the topics raised by the very int·resting exhibition etc "Evolving English", a very minor one was the remark that 'RP' "does reveal a great deal about [a speaker's] social and/or educational background". Of course that's true but it's tempting to quibble that, among many others, butlers, valets, ladies' maids and other servants of the affluent such as assistants in high-class flower etc shops (an occupation Eliza Doolittle aspired to) may well at least function professionally without displaying regional features in their speech. Of course any of these may well be 'bi-dialectal'.

More seriously, we may ponder the comment that "recent estimates suggest only 2% of the UK population speak it". I've never he·rd of any serious attempt to produce even moderately credible statistics in this field where the borderlines between 'RP' and non-'RP' are too uncertain for it to be a simple matter. What we have are, I fear, pritty unscientific g·esses. Not th·t that matters much becoz the chief consideration of consequence regarding 'Received Pronunciation' is not how many speak it in unadulterated form but the hugely greater numbers of speakers, at the very least in England, who have only very few and rather insignificant speech features that differ from it.

Two further remarks were "It has a negligible presence in Scotland and Northern Ireland and is arguably losing its prestige status in Wales. It should properly, therefore, be described as an English, rather than a British accent".  Its presence in Northern Ireland is undeniably negligible. Any speakers who've spent their entire formative years there are very unlikely to be classifiable as 'RP' speakers by anyone. The diplomat 'Paddy' Ashdown gained his nickname from the accent with which he arrived from Northern Ireland at an English public school at the age of eleven. No traces of it remain. On the other hand there are not a few well-known speakers who have spent their pre-adult lives in Scotland or Wales who have very few or no noticeable Scottish or Welsh features in their speech. A few examples who immediately spring to mind are Alistair Darling MP, Ming Campbell, Lord Kenneth Clark, Neil MacGregor (of the British Museum), Donald McLeod (Radio 3 Music presenter), Andrew Marr (BBC tv news), Lord Harry Wolf; Jeremy Bowen (BBC senior Correspondent), Huw Edwards (BBC tv), John Humphrys (BBC Radio 4) and the late Roy Jenkins (former MP).

Finally a comment on the remark "The phrase Received Pronunciation was coined in 1869 by the linguist, A J Ellis, but it only became a widely used term used to describe the accent of the social elite after the phonetician, Daniel Jones, adopted it for the second edition of the English Pronouncing Dictionary (1924)". This is in danger of being misinterpreted in that Ellis plainly did not set out to propose 'received pronunciation' as formal institutionalised term any more than he did so with the synonyms he used such as 'refined' or 'educated' or 'cultivated' pronunciation. He also used the expressions 'received American' and 'received Irish'. See more detail on this matter at Section 7.3 §5 of this website. Even when he used the abbreviation 'rp' he didnt capitalise the letters.

Daniel Jones certainly did have the intention to institutionalise his term. He did so prob·bly in partial imitation of H. C. Wyld who in 1913 had proposed the term 'Received Standard' which Jones presumably must've felt he was improving upon. Jones didnt introduce his new proposal in the 1924 second edition of his EPD, which he had revised only in respect of its wordlist, but in the six-page "revised Introduction" to the 1926 third edition. It was then that he abandoned his earlier-favoured expression 'Public School Pronunciation' and converted 'PSP's to 'RP's. The sep·rate prefatory further dozen pages of "Explanations" of that third edition awkwardly retained the previous uses of the label 'PSP' which actually no dou·t to his embarrassment remained in the dictionary for another decade until its fourth edition in 1937.  

More on Evolving English

The Name Nancy STORACE

BBC Radio 3 is today coming to the end of a twelve-day lit·rally round-the-clock feast of Mozart's music which is aiming to offer us "Ev·ry note that Mozart wrote". He happens to be just-about my faevrət c·mposer so I can't remember ever so completely neglecting our speech-mainly Radio 4 for over a week. More than once in this past week Radio 3've played recordings of the 'grand' concert aria, Ch’io mi scordi di te?... Non temer amato bene. The music of this was written by Mozart as a farewell gift for Nancy Storace, a soprano with whom he performed it, playing its very unusual piano obbligato part along with the orchestra, at a concert for her in 1787 at the Vienna Burgtheater when she was shortly to leave for London. She was the famous singer who in 1786 created the leading part of Susanna in his incomparable opera Le Nozze di Figaro.

A day or so ago I he·rd the admirable music presenter Donald Mcleod /mə`klaʊd/ introducing a performance of the aria by giving the background information I've just mentioned. As I wou·d've at one time, he pronounced her name as /`stɒrɪs/. How to say that name was something I discust the year before last in my blog 184. I reprise that topic now becoz I have added confidence in suggesting that the name is best anglicised as little as possible. As I pointed out then, because Nancy Storace’s surname was evidently an Italian one in origin, it had latterly seemed to me that the most the appropriate pronunciation to give it, in the absence of a pritty-well universally agreed anglicisation, had to be after the Italian manner. It's hardly surprising that some people, especially if they were aware that her first names were English (Ann Selina) as was the professional first name she went by (her pet name Nancy), shd be inclined either to believe her surname was English or anyway to anglicise it to rhyme with Horace. It could be that she or her father, an Italian double-bass player etc, submitted to some anglicisation of the family surname while in England where she lived until the beginning of her teens. However, this seems not to've been usually the case in view of the important fact that then, as the DNB (Oxford Dictionary of National Biography) article by Joseph Knight and Jane Girdham has pointed out, "Their name was often spelt Storache, for the sake of indicating its pronunciation". Nancy's family's having an Italian name must've been felt to be an advantage professionally for artists who publicly performed Italian music. In addition to the above, we can see from the phonetician Alexander John Ellis's book Pronunciation for Singers publisht in the mid 1870s that her brother Steven (aka Steffano) Storace's name was famous enough for Ellis to include in a list (at pp 242 to 246) of 'Pronunciations of the Names of Composers' where it appeared in a transcription that was equivalent to [stɔ`raːʧe].

According to DNB, after staying in Florence from 1779 to 1781, she went on to sing "in Lucca, Leghorn .. Parma, Milan .. and Venice". She was very successful in these Italian opera houses but moved to Vienna in 1783. Then, she'd been, along with two Italian singers and the Irish tenor Michael Kelly, offered a contract by Emperor Joseph II to sing at the Burgtheater. She stayed there about four years and became a very good friend of Mozart's. It seems highly likely that, though Mozart had no dou·t le·rnt some English when he was in London as a child, they wd've spoken to each other in Italian. Her surname, by the way, must've surely been derived from the Italian word storace [stɔ`raːʧe] meaning storax, a term for an aromatic gum resin.

Evolving English Pronunciation

As I commented on the John Wells blog of the 31st of December 2010 about the new British Library exhibition "Evolving English" and its asssociated survey etc, I very much applaud it as an int·resting and worthwhile undertaking but I felt obliged to point out that cert·n statements about e·rly BBC policies it's putting forth are strictly speaking factually incorrect. One of these was in a panel headed "Star item ... Broadcast English 1929" which sed that the BBC Advisory Committee on Spoken English "... recommended that all presenters avoid regional accents and use Received Pronunciation (RP)". 
At the associated blog headed "Royal Weddings and 'RP'" there is a remark that a decision was taken by a BBC Advisory Committee by which it "chose RP as the accent to be used by presenters...". 
This inaccuracy has been widely circulated for a long time even unfortunately supported by authorities who shdve exprest themselves more carefully including even A. C. Gimson who in 1962 sed "The BBC adopted this form of pronunciation ['RP'] for its announcers..." at An Introduction to the Pronunciation of English p.70. Whatever might've been argued to be the de facto situation in much later years, nothing specific·ly like that, as far as I can ascertain, was ever stated in those e·rly days by anyone in the BBC of appropriate authority, cert·nly neither by John Reith nor by the Committee's secretary Arthur Lloyd James, either using the expressions "RP" or "Received Pronunciation" or any synonym for them.

In the Preface to her BBC Pronouncing Dictionary of British Names G. M. Miller, remarked in 1971 that "the myth of 'BBC English' dies hard. It owed its birth no doubt to the era before the Second World War, when all announcers and perhaps a majority of other broadcasters spoke the variety of Southern English known as Received Pronunciation..." Even that moderate statement was not absolutely true. At least the late Joseph Macleod, who was employed as an announcer from 1938 until well into the war years, had certain marked Scotticisms in his speech. The reality was, anyway, that the BBC's announcing and newsreading staff in its earlier years were only a very small fraction  of what it was to become from the explosion of the sixties onwards. In 1934 only six were broadcasting to the UK and for some years after that the number hardly doubled.

Announcers were chosen by the markedly Scots-accented Director John Reith who wou·dnt've really known what the term 'Received Pronunciation' me·nt. He showed in a television interview with Malcolm Muggeridge that he wasnt even clear whether or not he himself had a Scottish accent. His desiderata in choosing his staff were that they came from "good" families and had attended superior institutions for their education. He appointed as chairman of his first Advisory Committee the poet laureate Robert Bridges who not merely wasnt well informed about the realities of pronunciation matters but actually had cranky ideas that led him to attack Daniel Jones's respectably based scientific judgments in that field. Reith increasingly packed his three successive committees, originally of half a dozen members and ultimately of eighteen, with inappropriate celebrities. They only ever contained a handful of scholars suitably equipped to offer sound advice. Besides Jones, at one time one of those committees contained Charles Talbut Onions one of the editors of the OED. He joined in 1930 but within a couple of years protested to its secretary Lloyd James at the 'insufficiently rigorous procedure' with which the discussions were conducted, adding that 'in no other department than language ... would the distinguished amateur be tolerated'. He bowed out of the Committee completely two years later.

Arthur Lloyd James was a well qualified linguistician who became appointed as its secretary, vetted and trained new announcing candidates and edited the Committee's 'Broadcast English' series of booklets. The nearest he got to speaking in terms of an accent the BBC expected of its staff broadcasters was to refer, in his book of collected writings The Broadcast Word (1935 p. 31), to 'broadcast speakers who regularly use varieties of Southern English that are not adversely criticized' as being 'preferred by the BBC'. The main qualification was, he sed, that the announcers shd 'speak a variety of educated English' be it that 'spoken by a Scotsman, or a Welshman, or an Irishman or an Englishman' (ibid p.26). At p. 42 he sed of the Northern-preferred short vowel in a word like dance versus the Southern long-vowel preference "both are common among educated speakers". Regarding the booklets of advice to BBC announcers on how to pronounce place names he sed "The pronunciation that we have attempted to record is not that of any particular dialect; it is certainly not South-Eastern. It is, rather, a normalized form, suitable for the whole of the English-speaking world" (p.63). The transcriptions in these booklets accordingly always showed the (non-'RP') r's that some of their users might be presumed to use habitually and so do their successor publications to this day such as the Oxford BBC Guide to Pronunciation of 2006. So far from simply requiring 'RP' from announcers, Lloyd James was at pains to proscribe the use by announcers of various common 'RP' features such as monophthongisations of the vocalic sequences of words like pure, fire and power because he considered them too unpopular with certain elements of the listening public, referring to them as "too aggressively modern"  (ibid p.21). The same applied to some types of front-of-centre beginnings to the /əʊ/ diphthong which he referred to as "distasteful to the majority of the public" (ibid p.24). The  Advisory Committees on Spoken English were abandoned at the beginning of World War II being replaced by the BBC Pronunciation Unit, a small group of specialists who incident·ly have had three main he·ds none of whom was an 'RP' speaker.

More on the Queen's Speech

Our last blog de·lt with a single word but of course it wasnt the only one of int·rest in her Chris·mas Message. Another was her p·onunciation of always as /`ɔːweɪz/. OED3 hasn· come to revising this yet so we have still Murray's century-and-a-quarter-year-old opinion (of January 1884) now given us in more familiar symbols "/ˈɔːlweɪz/ /-wɪz/" This is exactly what Upton & co showed in ODP (the Oxford Dictionary of Pronunciation) which also gave a /-wəz/ variant but as an American usage only. The Jones EPD (English Pronouncing Dictionary) always (1917-63) put /-wəz/ first. Gimson in EPD1977 put /-weɪz/ first then /-wəz/, then /-wɪz/. Roach et al 1997/2006 have kept to that order. Wells from LPD1(the Longman Pronunciation Dictionary) has only reversed the order of the subvariants to /-wɪz, -wəz/. I like that best but prob·ly cz I a·most a·ways say /ɔː(l)wɪz/ m·self. Wells alone has rightly acknowledged the common alternant you cd hear her using on Chris·mas day with no /l/ — a variant LPD conveys by italicising that letter to /l/. Gimson cd well have admitted that to EPD too: he offen used it personally.

Even more int·resting I found her other non-use of /l/ in only. It's totally normal, despite the failure of all dictionaries except my CPD (Oxford Concise Pronouncing Dictionary) to record the fact, for all English native speakers to use an l-less weakform of only. But she was using a strongform in stressing the word and saying (not very clearly) /əʊnnɪ/ ie replacing the /l/ with an assimilation converting it into a second /n/ — still not a terribly unusual thing to hear from GB (General British) speakers. Another not very uncommon form she used that you won't find at all in OED, EPD, LPD or ODP (tho it was an alternant in my CPD) was /ʤiːzəz/, not /ʤiːzəs/, tho with no prompting influence of any immediately following voiced sound.

Besides retaining some very conservative forms she nevertheless has joined in the flow towards replacing earlier instances of /ɪ/ with the schwas they more offen had in Victorian usage word-initially as well as word-medially. For example she began events clearly with a schwa: OED(2) and EPD offer only /ɪ/, LPD and ODP record alternant schwa. At the first syllables of become and belonging she also had the less traditional /ə/. The same applies to certain medial unstrest syllables at eg benefit and celebrate tho not at participate. Like all of us, she isnt consistent in such items.

She had the very normal elisions of medial schwas in conference as /kɒnfrəns/ and history as /hɪstrɪ/ and of as /ə/ in one of the ie /wᴧn ə ðə/ but the less usual elision of /r/ at for example which might strike many of us as a bit precious. She had something much more like the conservative /ɪ/ than the usual /ə/ at the -less suffix of countless. As to the [aɨ] that has been rightly attributed to her on occasion (as also from Prince Charles) for the /aʊ/ diphthong, she was saying only something rather approaching to it at thousands and without. She clearly had not-at-all-unusual smoothings of the "triphthongs" of required and players. One thing that struck me as notable was, hearing [ˈiɑ̝ː] at next year, that the target for year was not the /jɜː/ which she cd offen be he·rd to use in the past. Jones always showed /jɜː/ in first place but during the latter half of the last century the version more in accord with the word's spelling increasingly gained ground. Gimson retained Jones's preference at first but the more modern style /jɪə/ first appeared in EPD in 1988. The completeness of the change of people's target was confirmed in 1990 when the Wells LPD1 poll showed that 80% of his respondents favoured the more spelling-associated form. Her pronunciation here, as one hears constantly from newsreaders and others in such contexts, shows (unconscious) avoidance of a somewhat uncomfortable articulation by eliding the word's initial yod ie saying /nekst ɪə/. A final fairly trivial point is that I noticed that this at this was to become began with a dental plosive allophone of the canonically fricative /ð/ phoneme, nothing very unusual from most speakers.

An Obsolescent Pronunciation

The Queen's annual Christmas message to the nation has once again been a reminder, at least to those int·rested in the history of our language, of how fascinating it is to contemplate the usages of someone whose background influences are so fully documented. I suppose the single word that'll've most struck those who notice such matters will've been her pronunciation of often as /ɔːfn/ which occurred twice this year. This was the most common GB (General British) pronunciation in Victorian England and for a decade or so of the last century. Nowadays teachers of GB to users of English as an extra language wd rightly deprecate adoption of the /ɔː/ vowel in this word: there's not even any mention of its existence in books such as the OALD (Oxford Advanced Learner's Dictionary). The current normal vowel for the word, /ɒ/ as in got, had undou·tedly come on the scene no later than the last quarter of the nineteenth century. In the OED (from its inception in 1884) the often vowel was originally represented by a symbol that indicated neither the vowel of got nor that of short. It's always going to be a difficult matter to decide about some words whether their obsolescence has gone so far as to justify omitting them from representations of current usage. Yet the Queen's version is certainly not obsolete: it can still be he·rd from a small minority of other elderly GB speakers. A. Lloyd James, in a paper re·d to the Philological Society in London in 1932 (quoted at p.161 of his 1935 book The Broadcast Word) observed 'Young people do not now use pronunciations like lawss for "loss", crawss for "cross", cawf for "cough", although these pronunciations continue to be recorded in dictionaries. I have observed this modernism among educated young people in these parts for many years; it is the generally accepted fashion among the younger generation on the West End stage, whose pronunciations I have had exceptional opportunities of studying for ten years. But I was astonished to find recently, in a Council Central School in London (Marylebone), that the pronunciation crawss was laughed at by the children'.

Daniel Jones (1881-1967) in all of the editions of his EPD from 1917 to 1963 always showed often as principally having /ɔː/. When Gimson's revision of it appeared in 1977 he not merely put the /ɔː/ variant into second place but labelled it explicitly as "old-fashioned". When Roach et al took EPD over in 1997 they dropt the /ɔː/ variant completely. It wasnt included in the new ODP of 2001 either. The most satisfactory procedure was surely the LPD one of including /ɔː/ forms but placing them after the /ɒ/ ones. It'd be better still if pronunciations were, as far as might seem feasible, divided into 'co-variants', ie forms which are regarded as predominant and of a very similar degree of currency, and 'subvariants' ie forms of markedly lower frequency than the others. This can be indicated concisely by following the former group with a "greater than" type of sign [>]. Readers may observe that kind of practice in articles in the main section of this website such as Section 3.1, which compares GA & GB, where the sign ˃ is placed immediately after words which have alternant versions that are the less common ones. The sign ˂ similarly so placed indicates that the pronunciation shown is a less common one of the word it follows. These take up a minimum of space unlike sometimes rather dubious labels such as "old-fashioned". The word off with /ɔː/ doesnt always strike me as old-fashioned although it seems to be a lot more likely to give me that impression the closer (more "modern") the variety of the vowel the speaker uses.

OED3 online has the praps slightly puzzling entry "Brit. /ˈɒf(t)n/, /ˈɒftən/". This, one supposes, must be intended to convey that the versions /ˈɒfn/ and /ˈɒftn/ with no schwa in either are predominant and comparably common and that the only version with a schwa of any serious currency has always a /t/. Perhaps a version with two successive bracketed consonants /ˈɒf(t)(ə)n/ wdve lookt rather clumsy but I'm not sure that it was right to exclude the variant /ɒfən/ if that was indeed the intention. Only within the account of the word's etymology is there any indication of it's being he·rd with /ɔː/ in relatively modern usage. This contained a quotation of its 1902 "NED" ie OED1 transcription "(ǫ̀·f ’n)". The apostrophe before the word's final 'n' in that original OED1 quotation is the usual OED3 strai·t rather than  curved variety. Both varieties in various fonts tend to spoil legibility by fusing with the top of the 'f'. The only solution is the (undesirable) insertion of a space such as I've adopted here. That OED1 apostrophe's function was to signalise that the following consonant was syllabic. The two OED3 translations into modern symbols, "/ˈɒf(ə)n/, /ˈɔːf(ə)n/", have bracketed schwas. The offering of two versions, the first with /ɒ/ and the second with /ɔː/, is praps a reasonable makeshift 'translation' of Murray's notation. His explanation of his awkward symbol consisting of italic 'o' with superscript grave accent and subscript ogonek as indicating "medial or doubtful length" was not one of his more felicitous efforts.

Words Like LAUNCH

"Kraut" in his blog of the 22nd of December 2010 (revised at the 23rd) sed "I'm currently looking for some evidence for the pronunciation of words such as 'laundry', 'launch', 'haunch' etc. with /ɑː/ instead of the now exclusively used /ɔː/. When did the former pronunciation fade out?...". He added that Jon Arvid Afzelius in his Engelsk Uttalsordbok of 1909 listed these items giving their pronuncations in Henry Sweet's style:
1. gauntlet gɔɔntlĭt, gaantlĭt
2. haunch hɔɔnʃ, haanʃ
3. haunt hɔɔnt, haant
4. jaundice ʤɔɔndĭs, ʤaandĭs
5. launch lɔɔnʃ, laanʃ
6. laundry lɔɔndrĭ, laandrĭ
7. paunch pɔɔnʃ, paanʃ.
These are very likely to have reflected Sweet's order of preferred usages. Afzelius knew Sweet personally and was something of a disciple of his, as his choice of phonetic symbols suggests. Otto Jespersen recorded Sweet's preference for /ɔː/ in haunt and jaundice in his Modern English Grammar Volume I § 10.55 in 1909.

The first edition of Ida Ward's Phonetics of English had a short paragraph 47b in 1929: "Words first written with au such as taunt, launch, staunch, laundry, and the place names Launceston, Taunton were formerly pronounced with ɑː; some people still consider lɑːndri and lɑːnʃ correct, and Launceston and Taunton are pronounced lɑːnstən and tɑːntən by natives of these places." This appeared in the chapter on 'Spelling Pronunciations' and was repeated unchanged in all subsequent editions. It came in a subsection of that chapter (dubiously) entitled 'Changes in vowel pronunciation due to the influence of spelling'.

Dobson's English Pronunciation 1500-1700 (second edition 1968, p. 791) had the remark "... some words (e.g. launch) still vary between the two pronunciations. In general [ɑ:] is now used when the spelling is a (dance, half &c.) but [ɒ:] where the au spelling has maintained itself (haunt, vaunt, &c.); but there are exceptions, e.g. aunt with [ɑ:], (halm also haulm) with [ɒ:] beside [ɑ:]". He might also have mentioned another rather uncommon word launce, the name of a kind of eel. OED and EPD agree on giving it as /lɑːns/ for British usage. ODP thinks it’s only /lɔːns/ here. LPD3 omits the word. The archaic or regional word maunch has a US subvariant /ɑː/ form in OED3. Shakespeare rhymed haunted with enchanted, granted and planted.

John Walker's Critical Pronouncing Dictionary of 1791 gave /ɑː/ alone for jaunt, laundress, flaunt and all seven of the words Kraut quoted from Afzelius except only haunt at which he added the characteristic comment: "This word was in quiet possession of its true sound, till a late dramatick piece made its appearance, which, to the surprise of those who had heard the language spoken half a century, was, by some speakers, called the Hawnted Tower... a plain common speaker would undoubtedly have pronounced the au as in aunt, jaunt &c..." In that work's prefatory general 'Principles' at §214 he sed that when au is followed by n and another consonant it changes to /ɑː/ as in "aunt, askaunce, ascaunt, flaunt, haunt, gauntlet, jaunt, haunch, launch, craunch, jaundice, laundress, laundry...". He then added to this list daunt, gaunt and saunter which last, like staunch, is not found in the body of the dictionary. Also he mentioned the words maund, maunder and Maundy which are all credited in OED with current US subvariants having /ɑː/, tho they're no dou·t mainly regionalisms there today. Finally he refers to the opinions on these items of the lexicographer Sheridan and the orthoepist Nares noting that the latter gave /ɑː/ for vaunt. Walker favoured only /ɔː/ for it. Some of his comments will praps seem less surprising if we take into account the fact that the earliest spellings of advance, branch, chance, chant dance, glance, prance etc all included variants with au but brawn, dawn, fawn, lawn, prawn, yawn etc are not with alternant ɑ spellings according to the OED.

H. C. Wyld's Short History of English (1927) treated this topic at §§183,184 etc referring to the original Norman-French loans having had nasal vowels which werent all lost at once with the deveopment of the diphthong au "from the nasal vowel". He cited spellings including chaunticleer, daunce, exaumple and aunt. At §259 in a note on words like daunt, haunt etc he sed "As regards the pronunciation [ɑ:] which exists also [sc. besides /ɔː/] in these words, as well as exclusively in aunt, the least unsatisfactory explanation seems to be that it goes back to a M.E. variant with [short] ɑ."

Regarding the later nineteenth century commentators, there is relatively little to add to what Jespersen mentions in the work quoted above at §§10.553 and 10.556. He had few items beyond those we've alre·dy mentioned but he referred to the NED subvariant /ɑː/ forms (which appeared in OED1 from 1884 to 1901) at avaunt, gaunt, gauntlet, haunt and jaunt. These were retained, except for jaunty, in OED2 in 1989. He suggested that the name Gaunt cd also have /ɑː/. In 1917 Jones's EPD1 had subvariant /ɑː/ at jaundice, launch, laundress, laundry and staunch. It didnt contain the word gauntlet but it had the name Gauntlett apparently suggesting that its /ɑː/ alternant was commoner than the /ɔː/. It also had Saunders with the reverse preference; LPD3 and EPD17 also have /ɑː/ alternants for that name. OED3 has staunch with the alternant spelling stanch and /ɑː/ and /ɔː/ versions in that order. I'm unable to recall where I gathered that George V (died 1936) used the form /lɑːnʃ/.

Much more int·resting detail can be seen on this topic in several of Kraut's blogs beginning in December 2010.

BlogRoll

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