Archive 4 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

23/07/2007The LPD Pron Pref Survey 2007 (x)#042
22/07/2007Transcription Excercise 2#041
20/07/2007The LPD Pron Pref Survey 2007 (ix)#040
14/07/2007The LPD Pron Pref Survey 2007 (viii)#039
12/07/2007OH! & Allegedly Unstressable Schwa!#038
01/07/2007Twenty Questions#037
29/06/2007Transcription Excercise 1#036
28/06/2007The LPD Pron Pref Survey 2007 (vii)#035
26/06/2007The LPD Pron Pref Survey 2007 (vi)#034
24/06/2007The LPD Pron Pref Survey 2007 (v)#033

Blog 042

The 23rd of July 2007

The LPD Pron Pref Survey 2007 (x)

Item 27, in another surprising inclusion, deals with the adjective applicable which also featured in both the previous Wells questionnaires in 1988 when the first set of respondents showed a 77% percent preference for stressing its second syllable and again in 1998 when the new set displayed the same preference by 84%. It looks as if Wells now wants to know whether there's anyone left who favours the 19th century earlier stressing. His 2008 proportions were unchanged at 85% to 15%.

Daniel Jones in the EPD recognised the new stressing from his first edition in 1917 but never gave it precedence. Neither did Gimson from 1967. Roach and Setter from 1997 have agreed with Wells and so does ODP. Murray in the OED in 1885 gave only the forestressed form and, with typical feebleness about attempting to come up to date, the second edition of 1989 was unchanged. It's going to be interesting to see what OED 3 does about pronunciations. The early stressing is holding out better in the US as LPD2 confirmed quoting the 64% preference of a poll there. 

Item 28 kilometer also featured in the 1988 and 1998 polls with  preference for stressing the second syllable scoring respectively 48% and 57%. In 2008 it was much the same at 63%. This seems to chime with my feeling slightly more comfortable with the later stressing than I once did but I certainly vacillate with this as I do with many other words. The US preference is of course the one we seem to be edging towards. Unlike LPD, EPD now gives it preference. Not so ODP. 

Item 29 is mischievous asking whether respondents prefer it (i) stressed on the first or (ii) the second syllable or (iii) the latter plus an extra sound so that it rhymes with devious. This is another re-visiting: LPD2 had a graphic which represented a greater than 70% British preference [maintained at 65% in 2008] for the traditional fore-stressing (and only slightly less American enthusiasm for it) but a surprisingly substantial vote for stress on the second syllable especially amongst younger respondents – so much so that it was mildly surprising to note that the warning triangle signifying "considered incorrect" had not been withdrawn from the entry.

No doubt this problem accounts for the item's reappearance in the questionnaire. The triangle was certainly justified for the devious type which, like "grievious" for grievous has never had much currency among the better educated. I would hazard the guess that they both arose from a misplaced inclination to "restore" a sound taken to have been lost from the correct version (or of course merely simple confusion on the part of many users). Many people use a pronunciation of previous which might be conveyed by the spelling "prevous" and I suspect most of us say what could well be indicated by the spelling "prevously". Thus these two incorrect -ious items may well be what we might call misguided yod-dropping reversals.

The very last item of the 2007 poll revisited the adverbial form necessarily which in the 1998 poll had shown 72% preferring neces`sarily to fore-stressing. This is the version in my opinion rightly given precedence in LPD. Indeed I gave it precedence in 1972 in my CPD and in the ALD third edition of 1974. OBG agrees and remarks that the other "is becoming less common". However EPD in which Jones didn't recognise the existence of the later stressed version till 1956, still retains the precedence of the fore-stressing, as does ODP. Craigie in the OED in 1906 made no mention of a later stressed version but for once OED2 in 1989 at least got it in – though again only in second place.

What I should have liked to see was whether anyone favoured what I find is the newest development in this matter namely the increasing use since the last decade of the twentieth century, by GB speakers with no obvious Scottish affiliations, of the ash vowel /æ/ instead of /e/ in the ending -arily

Blog 041

The 22nd of July 2007

Transcription Excercise 2

Here's a "model" transcription of our first exercise.

ðə 'njuːˎməʊtəkɑː

  1. ðə ˊˋʤəʊnzɪz ᛁ əv 'gɒt ə njuː ˋkɑː.
  2. ˏˋhæv ðeɪ, ɪndiːd. 'wɒt ˋmɒdl ɪz ɪt, dɪə.
  3. əʊ, ˋaɪ dəʊnt ˏnəʊ.  ˋɪts ə 'peɪl ˋbluː.
  4. ˏwel, ˋwel.  'əʊl 'ted ˋɪz ə dɑːk hɔːs.
  5. ˋhaʊz ˋðæt, ðen.
  6. aɪ wəz 'əʊnli ˋtɔːkɪŋ ˏkɑːz wɪð ɪm ᛁ  ði 'ʌðə ˋnaɪt.
  7. ən i 'dɪdnt ˋˏmenʃn ɪt?
  8. nɒt ə ˋwɜːd.
  9. 'haʊ 'lɒŋ ᛁəd ðeɪ hæd ði ˋəʊld wʌn.
  10. ˏnɒt ə deɪ əʊvə 'tuː ˋjɪəz, aɪm ˋpɒzətɪv.
  11. ˋwel. 'ðæts ˊˋwʌn weɪ wi ˌʃɑːnt bi ˌeɪbl tə kiːp ˌʌp wɪð ðəm.

In the sound version the wife in line 5 has a sort of slip of the tongue losing the vowel of the final word "then" in a way that sounds very unusual. The last line creeps upwards slightly.

Here's another piece to try:

'British ˎJournalism

  1. A: 'What d'you 'think of our ˎnewspapers.
  2.  B: 'I ˏlike them very `much. `Specially the `ˏSunday ˳ones.
  3. A: ˈWhich `daily d'you like best.
  4. B: 'I ˏcouldn't `say, ˏreally. 'I like to 'look at `several, if I get the ˏchance.
  5. A: The `Times gives the `fullest ˋˏcoverage ᛁof anything `serious.
  6. B: 'M.ˏYes. But 'I like 'rather more `pictures than it ˏgenerally has.
  7. A: And it's 'not tied to any par`ticular po`litical `party, ˎeither.
  8. B: ˬOh? ˏI ˏthought it was Con`servative.
  9. A: `Not with a `capital ˏC.

Blog 040

The 20th of July 2007

The LPD Pron Pref Survey 2007 (ix)

Item 25 returns to the word yours for which a graphic was given in LPD2 showing that over 60 percent of respondents preferred it as /jɔːz/ though nearly 40 percent opted for /jʊəz/. This was in a way surprising because I very much doubt if this latter version is used so much, though one isn't surprised that a number could prefer that it should be spoken in such a way because it's more satisfying to many people to think of it as, in view of its spelling, differentiated from the word yaws. [The 2008 results were 75%/25%.]

In the OED in 1921 Onions only gave /jʊəz/. Daniel Jones in the EPD from 1917 always gave first place to /jɔːz/ but gave other versions including two that Gimson retained in his 1977 EPD revision viz /jɔəz/ and /jʊəz/ prefixing the latter with rarely probably by mistake (not corrected even in 1991) because the other form was better entitled to such a description. The current EPD and ODP are in agreement with LPD.

    Item 26 returns to the -less suffix asking whether the respondents prefer careless to end /-ləs/ or /-lɪs/ or if they can't decide between them because they feel that callous and Alice rhyme or if they prefer /-les/. I have heard individuals using this last form form but very rarely indeed (singing aside).

I haven't observed anyone consistently using a sound intermediate between /ə/ and /ɪ/ for the -less ending though such people may well exist and I myself have a few words, for example target, in which I don't really feel that I usually say either vowel. And of course one often hears a speaker utter a word like careless in non-deliberate speech with a sound that one can't safely assign to either phoneme.

Anyway, the thing that fascinated me about this and similar items was Daniel Jones's feeling about them. He never admitted the schwa versions of -less nor of various other similar endings into the EPD from 1917 to the 1956 so-called eleventh edition (the last for which he was responsible – there were truly only four) in which he gave his reasons for not doing so. It was because /-lɪs/, he thought, was "that most commonly employed by non-dialectal southern English speakers". He went on to add "It must be observed, however, that there do exist well-educated English people who speak in most respects with RP but who use the vowel << ə >> ... I have the impression that such speakers have had their speech influenced by contacts with people from northern or eastern parts of England ... and that these special features of their pronunciation cannot properly be regarded as coming within the limits of what should be included in this Dictionary." (p. xxix). Murray in the OED in 1888 had only given /-lɪs/.

When Gimson took over the EPD in 1967 he left things as they were until his extensive revision of 1977 when he added the schwa versions but still showed the /-lɪs/ ones first. This wasn't too surprising because they were his own habitual versions, though hardly any of his colleagues in the UCL Phonetics Department shared them, and they remained so till the 1991 edition. He did reverse the order of things with words like moderate which was again his own usage.

What rather amuses me is that my own reaction is the precise reverse of Jones's. I have the impression that, at least in people under retirement age, the use of/-lɪs/ is not mainstream General British but either a mild Londonism or markedly old-fashioned.

Blog 039

The 14th of July 2007

The LPD Pron Pref Survey 2007 (viii)

Item 23 is similar to 22 and first enquires how one prefers to say during. I'll be surprised if there're many responses in favour of /`dʊərɪŋ/ with the yod dropped. This is normal General American of course (see § 3.1.20a on this website) but in England it's an obvious regionalism confined to a few areas including parts of London and the Midlands – so far as I've noticed.

Secondly people are asked to say whether they prefer its first vowel to be /ɔː, ɜː/ or /ʊə/. All three of these versions are current and not regional. I expect most will plump for the last not least because it seems to be in best harmony with the spelling. [In the event 76% did.] It's what I have as a conscious target. Whether I and others are right in our impressions of what we actually do say is a tricky question. I suspect many of the ones who think they normally say /ʊə/ would turn out if observed carefully to often say /ɜː/. I very often hear /jɜːrə`pɪən/ for European from speakers on radio and tv. I wonder how such people'd respond to this question. I think what Wells is doing with his surveys is very worthwhile and interesting but one has to consider the results with in mind the kinds of prejudices that people tend to have. Most people probably have something of an exaggerated idea of the extent to which their pronunciations and their spellings match.

 Item 24 on poor is a fascinating one for me because I can remember agonising about which pronunciation to show first in my CPD in 1972. I decided finally for to put /pʊə/ first and /pɔː/ second but somehow my second version disappeared in the printing except that it had been kept as an alternative for the following item which was poor-house! Proof correction can be a pretty unhappy business at times. Pity the harmless drudge who undertakes it. At the time, I was being more cautious than confident but I had the EPD in front of me as revised by Gimson and it only showed /ɔː/ as a third form giving second /pɔə/ which LPD probably rightly now omits altogether coz it's become so old-fashioned when it's not a regionalism. I've always said /pɔː/ myself and now completely appropriately LPD, EPD and ODP are all in agreement in showing /pɔː/ as the current predominating version. In this case we have a Wells brilliant LPD graphic which shows vividly the changing preferences of successive age groups (at p. 593 of LPD2).

Blog 038

The 12th of July 2007

OH! & Allegedly Unstressable Schwa!

Today's Wells blog refers to the fact that he sometimes wishes that Gimson hadn't adopted /əʊ/ for the vowel of words like boat. He doesn't question that it was a better representation of the GB ('RP' if you will) diphthong's quality than the /ou/ of the Jonesian EPD. Incidentally, I have often enjoyed as an EFL teacher demonstrating to students who were brought up, as it were, on that notation how appropriate the Gimson innovation was by beginning pronouncing the /ʊə/ centring diphthong – which is notationally the exact reverse of the closing one /əʊ/ – and saying it repeatedly with gradual reversing of the force with which its two elements were uttered while making no change in their qualities. The effect is, hopefully ie if performed effectively, an auditory experience comparable to certain kinds of optical illusion in which a pattern stared at may seem to change direction such as from rising to descending steps. There's a point where you don't know whether you think you're hearing the closing or the centring diphthong and then another when you recognise that the one has turned into the other. 

Anyway, Wells mentions two "drawbacks" to Gimson's choice of symbol. First he says: The letter ə had hitherto been used only for the weak vowel and the less prominent element of centring diphthongs. It certainly can't be denied that practically every book that describes GB claims that /ə/ is never stressed. See Abercombie 1964:25 "confined to unstressed syllables" and even Cruttenden-Gimson 2008 at pp 95 & 154 "... /ə/ occurs only in unstressed syllables ..." It's true that it makes a very reasonable rule of thumb for learners to ban stressing of /ə/. However, the fact is that the vast majority of GB speakers have a small number of words in which they do have a stressed schwa. And there is nothing un-English about such a vowel: it's common in various accents.

For just LPD includes, explicitly not as a weakform, the pronunciation /ʤəst/. In EPD Jones had such an entry from 1956. LPD also has /məz/ as a form of Ms, the only one recommended by OBG. LPD lists the strongforms /bɪ'kəz/ and /bə'kəz/ of because tho it does dub them "irregular". Although LPD only lists, beside /mə'sjɜː/, /'mʊsjɜː/ and /'mʊsjə/ as initially stressable forms, I'm sure /'məsjə/ before a name isn't in the least unusual. I also regard /'hələʊ/ as a perfectly normal pronunciation of hullo, at least before a name. LPD also admits that "some speakers use stressed /ðə/ as a strong form, rather than the usual /ðiː/". I certainly would think it perfectly normal if, for instance, anyone who disliked the expression "Look it up in the dictionary" to say "I prefer to speak of looking words up in /ə/ dictionary rather than /ðə/ dictionary" if only because "/ðiː/ dictionary" could well suggest "the best etc" when that was not meant. There the indefinite article a similarly received stressing. LPD has a stress mark of sorts /(ˌ)dʌnəʊ/ at dunno (which I think can as easily be heard as / `dənəʊ/ ) and, though at gonna it calls it "contracted weakform", it shows it as /(')gənə/ and has similar treatment of the entry go.

There are plenty of other examples of GB stressed schwas. A most striking one came to my notice in the sixties when I'd first begun systematically noting the usages of radio and television newsreaders and presenters. I found that a very clear majority of them used the pronunciation /`θrəpəns/ for threepence — so much so that I felt I had to give that version first place when I came to compile my CPD (Concise Pronouncing Dictionary) which appeared in 1972 just after that word began being much  less commonly he·rd when Britain had converted to a decimal currency. Among other items that I've been able to check from re-playing recordings etc have been the Queen clearly saying doesn't as /`dəznt/. King George V saying minutes as /`mənɪts/ was noted by Ida Ward in 1939 in her Phonetics of English (1972 edition p.220). 

Many examples are to be noted among "re-stressings" of weakforms. In fact schwa weakforms are all strest at times by many speakers (with the exception of those of her). The list of the most important three dozen of them consists of those of a, am, an, and, are, as, at, but, can, could/nt, do, does/nt, for, from, gonna, had/nt, have/nt, has/nt, must/nt, nearly, saint, sir, shall, should/nt, the, was/nt, will, and would/nt. Jeremy Irons as Charles Ryder in Brideshead Revisited sed "You dont want to spend the rest of your life with Kurt, / ́də ju/ ". I've come to suspect that people nowadays prefer that style to / ́duː ju/ in less formal situations.

Contexts of liquids and nasals around /ɪ/ produce many occasional examples of its conversion to schwa including /`brəljənt/, /`brətɪʃ/, /`ʧəldrən/, /`grənɪʤ/ (Greenwich), /`mələtri/, /`məsnt/, /rə`məmbə/, /`səlvə/, /`wəlsn/ (Wilson), /`twəlv/. Words like ill, still, will etc regularly have /ə/ for their vowel in most types of South African English.

All these examples have been given without any intention to attribute them solely to the very widespread drift to be heard in recent generations towards weakening or removing the traditional rounding of the /ʊ/ phoneme so that for very many  younger speakers it has merged with their /ə/. It's been around a very long time in expressions like good and for goodness sake. I've also mainly refrained from giving items that might be said to be unaccented because they're spoken with merely animation stresses as with eg /`əv ˎkɔːs/ (Of course!) or /'ðə ˎfuːl/ (The fool!).

Finally, on the stressing of schwas, I again consciously haven't adduced as examples the fair number of French expressions in which many of us employ them including de rigueur, which LPD shows as \ (ˌ)də rɪ 'gɜː\, René, Renault, Renoir, première etc. By confusion, one supposes, with première, the word premier is quite often to be heard as /`prəmiə/ or /`prəmiɛə/ these days.


Blog 037

The 1st of July 2007

Twenty Questions

Twenty-Questions ie / `twenti kwesʧnz /

In his Phonetic Blog of Friday 6 July 2007 John Wells referred to the fact that in his LPD he mentions "the /ˈtweni/ form [of twenty] as a BrE possibility, but with a warning triangle [⚠] against it. Yet I am aware that some people who qualify generally as RP speakers do use it" and asks "Ought I to remove the warning triangle?"

He refers to similar elisions of /t/ in want to and trying to and three other items as usages that "Londoners" employ.

Elsewhere on this website at §3.1.22b I've remarked: In rapid counting /`tweni/ or indeed /twəni/ can be heard in casual GB speech but these simplifications of /-nt-/ have as counterparts in GB only the fairly markedly colloquial forms of going-to /gənə/ (before vowels /gənu/) and want-to /wɒnə/ (before vowels /wɒnu/) often written gonna and wanna.

My impression is that these are not exclusively Londonisms, nor regionalisms at all for that matter, but are widely used weakforms that are felt to be relatively casual in style. I think one can add /`traɪnə/ ie trying to (most usually in a disyllabic form eliding the middle vowel phoneme /ɪ/) to this small group.  I know I've heard /gənə/ – and I don't think a strongform /`gəʊnə / would sound much less acceptable, by the way – at least from plenty of (GB) broadcasters not using scripts, but perhaps many speakers only use them when they feel that for them they constitute what I dubbed in my blog of the 28th of December "linguistic slumming" (ie conscious as opposed to inadvertent incorrectness). I notice LPD accepts /gənə/ as "informal" and at the spelling wanna also says "informal" with the further comment "not standard in Br E". Also I see that it has an entry wannabe with no triangle or qualification – quite rightly because that plainly is "linguistic slumming" as its spelling proclaims.

What should we say in reply to his question, then? Perhaps it would be reasonable to label it additionally as a "casual weakform" especially because one can easily imagine twenty-first, twenty-one etc being said with the weakform but it's surely much less likely that twentifold or twenty-twenty (vision) would ever be uttered quite so casually. The LPD "warning triangle" is clearly explained as meant to alert those who would not like to be considered incorrect. So if they're not so confident of their command of the usages that they aspire to employing as to ignore such triangles anyway, then surely they should continue to benefit from the caution he offers. So our advice is "Keep it!"  Or, if he doesn't keep it on the first pair we've mentioned, we suggest he does so on the second.

Blog 036

The 29th of June 2007

Transcription Excercise 1

I'm thinking of trying the experiment of offering some exercises in the transcription of short passages of conversational English that I may put up here from time to time and then a couple of weeks or so later provide a model version for those who've attempted the exercise to compare with their own effort. I propose to explain mistakes people make by inviting them to send me a copy of their effort by email. See "Send the author an email" at the bottom of my front homepage and click on it. Please note that their transcription must be made in the body of the email because I never open attachments from correspondents not already known to me. Also, I shall only be willing to look at what they send to me if they precisely comply with the accompanying instructions which will be as follows:

Exactly copy onto your transcription the tone marks shown taking careful note of their indications of rhythms.

Don't leave any space between any tone mark and the syllable it precedes.

Transcribe only in the LPD segmental symbols (ie not copying any stress markings from LPD).

Use spaces between words exactly as in ordinary spelling (not LPD spacings).

Give the pronunciations you consider most suitable for EFL learners to adopt.

Give no alternatives whatsoever. Include the title.

Here's the first text:

The 'NewˎMotorcar
  1. A: The ˊˋJoneses | have 'got a new ˋcar.
  2. B: ˏˋHave they, indeed. 'What `model is it, dear.
  3. A: Oh, `I don't ˏknow. `It's a 'pale `blue.
  4. B: ˏWell, `well. 'Old 'Ted ˎis a dark horse.
  5. A: `How’s `that, then?
  6. B: I was 'only `talking ˏcars with him | the 'other `night.
  7. A: And he 'didn’t `ˏmention it?
  8. B: Not a `word.
  9. A: 'How 'long had they had the `old one?
  10. B: ˏNot a day over 'two `years, I’m `positive.
  11. A: `Well! 'That’s ˊ`one way we shan't be able to keep up with them.

If you'd like to hear a version to compare it with your transcription, go to §4.1.15.

For those who haven't got phonetic and tonetic symbols readily available to use LPD in (the body of) your email you need to have in your Unicode Character Palette Favourites the symbols listed below. MS Word and various other word processors nowadays make Unicode available.

ɪ U+026A Latin small letter cap I  
ɑ U+0251 Latin small letter alpha  
ɒ U+0252 Latin small letter alpha  
ɔ U+0254 Latin small letter open o  
ʊ U+028A Latin small letter upsilon  
ʌ U+028C Latin small letter turned v  
ɜ U+025C Latin small letter reversed open e  
ə U+0259 Latin small letter schwa  
θ U+03B8 Greek small letter theta
ð U+00F0 Latin small letter eth   
ʃ U+0283 Latin small letter esh  
ʒ U+0292 Latin small letter ezh  
ŋ U+014B Latin small letter eng  
ː U+02D0 Modifier letter triangular colon (length mark)  
ˈ U+02C8 Modifier letter vertical line 
ˊ U+02CA Modifier letter acute accent 
ˏ U+02CF Modifier letter low acute accent  
ˎ U+02CE Modifier letter low grave accent  
` is the first key at the left of the bottom line of many querty keyboards; alternatively U+02CB  
æ can be obtained by holding down the Alt key and pressing the apostrophe key (on a Macintosh); alternatively U+00E6.

PS The intonation notation is of a somewhat simplified type taking no account for example of the extent of a falling movement. Anyone who saw my blog of the 17th of June may remember that I insisted there that it's perfectly commonplace for speakers to use a high-beginning falling tone on a word that is not being highlighted for its own sake. At line 3 above the word It's is shown as uttered on a falling tone not by my choice but because I transcribed it so from the recording of the passage I gave to actors to read with no hint to them about what intonations to use. This example of an "animation" stress I hadn't noticed when I chose the passage for the present experiment.

Blog 035

The 28th of June 2007

The LPD Pron Pref Survey 2007 (vii)

Item 20 is another one that has surprised me. I've been aware for half a century that were as /wɛə/ and were'nt as /wɛənt/ are highly recessive. I seem to remember from my earliest systematic observation of the usages of BBC newsreaders in the 1960s that the only one of them who regularly used such versions was Alvar Lidell, a charming man who had some other rather idiosyncratically old-fashioned features in his elegant speech. I predict not more than 10% will favour them. [In the event it was 6%.]
Item 21 is /eɪʃə/ versus /eɪʒə/. OED 89, as so predictably, gives us (for Asian) still the unrevised picture supplied by Murray in 1887 ie only
/eɪʃən/ and the archaic /eɪʃiən/ which latter wasn't given in EPD in 1917 even! ODP and OBG both give priority to /∫ / but LPD has given it to /ʒ/ since its first edition in 1990 and EPD agrees. EPD in Jones's day didn't give /ʒ/ till 1947 and then only as a "rare" subvariant. Gimson in 1977 (tho not in 67) removed the "rarely". So we've come in line with the US (as shown in Webster in 1966). It'd be very foolish to put it down to American influence but the curious fact is we've shared a tendency. In my younger days I thought of /ʒ/ as a Northernism but now I have to agree with LPD and EPD and I'll be surprised if it doesn't get a majority vote. [It did :64%/36%.]
Item 22 is /tjuːn/ versus /ʧuːn/. I can well understand the appearance of this one: LPD gives the "RP" thumbs-down to /ʧ/ tho EPD simply gives it second place as does ODP. The trouble is that people now very freely affricate /tj/ and /dj/ in a way it seems they didn't use to as much. So now /tj/ and /ʧ/ are quite indistinguishable for most of the time. Past evidence is curiously far stronger for merger of /dj/ and /ʤ/ or is this topic a mite /tiːʤəs/? Anyway, the Wells respondents surely won't prefer /ʧ/. I hardly know which I use myself but I do know I'm not  at all consistent. PS John Wells's blog today is about American pronunciations: anyone who'd like to pursue that topic further will find 14,000 words on it and the British differences on this website at §3.1.

Blog 034

The 26th of June 2007

The LPD Pron Pref Survey 2007 (vi)

Item 19 via is another example of the many words that have tended to have their spoken forms revised as a result of people's greater awareness of foreign usally modern Continental languages but probably in this case mainly the ancient one, Latin, from which it directly derives of course. Cf this website §5.12.III.4. No doubt John Wells is wondering whether the newer form is anywhere near overtaking the traditional one with the anglicised vowel. You wouldn't think so from the reference books. OBG sings from the same hymn sheet as LPD giving /iː/ as second form. But EPD and ODP have only the one version for British usage (as I did in my CPD in 1979) but I think I've been conscious of the new one for at least thirty years. LPD's entry doesn't suggest any difference between American and British usage and MWO confirms that view. EPD's American entry agrees too but ODP suggests that /viːə/ is predominant in the US. Predictably OED 1989 hadn't heard of the new form. The great Webster of 1966 had it for the US at any rate.
By the way, MWO is my abbreviation for the very handy Merriam-Webster Online: it's not exhaustive but one wishes that some British publisher could come up with anything so useful. If I use any shortened ascriptions you don't recognise you'll find them glossed in the Abbreviations and References section §1 on this main website.

Blog 033

The 24th of June 2007

The LPD Pron Pref Survey 2007 (v)

Items 17 and 18. The  first asks people to say if they prefer the stressing elec`toral. This is plainly a relatively new usage. EPD, ODP and OBG make no reference to it at all. LPD has had it since its first edition of 1990 but only with the sign that precludes it from "RP". My earliest (casual) note of it was as used by Lord (Michael) Young (1915-2002) in 1987 but I doubt if that was the first time I heard it and I have particularly noticed its use among trades unionists and Labour adherents including Moss Evans, Neil Kinnock and the former speaker of the House of Commons Baroness (Betty) Boothroyd. I've heard it also from the tv newsreader Moira Stewart. So it looks like being re-classified as respectable as man`datory has been for some time.   
     This phenomenon is the longtime tug-of-war between the unifying tendency which makes people stress the word as early as possible and the desire for a more comfortable version to articulate instead of the lingual effort required to trot out weak post-tonic syllables. It's happened with ad`versary (still a no-no in LPD), in`ventory and com`parable and to learnéd words, too, like ag`grandise and re`condite. `Cervical and um`bilical are now being increasingly so regularised. And by some Na`omi. So are irre`futable and irre`vocable, the latter even from my favourite news presenter Jon Snow.
    Item 18 impious is in the same category and even more likely to be rejected in its traditional stressing `impious  because in speech it practically disguises its connection with pious. I predict a majority for im`pious. [LPD3 did indeed give that first in 2008.]  The 1989 OED as so often failed to give anything but the traditional version.
     Oddly enough the reverse has happened to decorous! Murray in the OED in 1894 referred to 18th-century wavering between de`corous and `decorous giving both and saying "The word is not very frequent colloquially". But LPD has rightly said from 1990 "formerly" de`corous, though EPD still gives that as a second form. ODP simply omits the version with the old stressing. OED2 of 1989 gives the obsolete version first – as we've come to expect. Even the Victorian Daniel Jones never did that from 1917 on.
    Incidentally I'd like to draw my blog readers' attention to the new rather large item (about 14,000 words) amongst other recent additons and modifications on this website offering a more detailed account than usual of the contrasts between General American and General British. It was the product of my being tired of finding only skimpy accounts in so many places which gave much the same examples and
, among other deficiencies, no kinds of indications of the quantities of contrasting items involved etc. See §4.1