Readers who need explanations of any of the abbreviations used may find them at Section 1 of the Home Page.
|28/08/2011||'RP' Dead but won't lie down?||#360|
|19/08/2011||The Guessing of Stressing||#358|
|25/07/2011||Less Common Assimilations (ii)||#355|
|21/07/2011||Less Common Assimilations (i)||#354|
|06/07/2011||GB Phonetics for Spanish-Speakers||#352|
|05/07/2011||Phonetics Teacher's Confessions (i)||#351|
I apologise for returning to the topic of the BBC "Archive on Four"
series program "RP RIP?". I must take back my remark that it was such a
trivial hotchpotch coz otherwise it'd've been on Radio 3. On Radio 4 on
Tuesday the 16th at 4.00 pm there was by contrast a satisfyingly serious
program on speaker recognition with an excellent contribution from
Professor Peter French our foremost practitioner of its application in
the forensic field.
But to return to Melvyn Bragg's mishmash of a rant agenst 'RP'. 'Kraut', in his 'English phonetic blog', has de·lt with the program at some length. He hasnt been inclined to be as indulgent about its triviality as I was. He justifiably complained that an article quoted from a recent edition of the Journal of Sociolinguistics which reportedly sed that 70% out of 5,000 people across Britain were "proud of their accent" (whatever that me·nt) didnt even have its year of publication mentioned (2007 by Nikolas Coupland & Hywel Bishop). Anyway, how much that cou·d mean is pretty dou·tful when one considers the fact that no notable number of people, even the best fitted to judge, can be sed to agree on precisely what constitutes GB (General British is a far less objectionable term for the regionally neutral UK accent than 'Received Pronunciation') or on who is a speaker of it in unadulterated form and who not. In Jones's day the majority of his appointees to his department werent considered by him to be speakers of 'pure RP'. If we only consider individual words it's not all that difficult to find instances of the two principal authorities, LPD and EPD, not agreeing with each other, eg at abrasive, accept, alright etc (ie 'and the like'), calm, dislike, dismantle, drawing etc, Edwardian, erupt, eschew, greasy, handicap, heinous, incisive, invasive, nothing, obtrusive, old etc, one, pastoral, refuge, restaurant, sixth, value, were, with and even the -ed and -es verb inflections and the [ɒʊ] diphthong recognised in LPD but nowhere else. They both admit as GB /ʊ/ in threepenny which can confidently be sed to be almost non-existent outside of the London area. See also our Blog 105.
Anyway, I was quite amazed to be emailed by Google's YouTube Service thus: "I read your recent blog post about the programme "RP RIP". You might not be aware that the left-wing group Chumbawamba recorded a song under a similar name. They are obviously pleased about its declining status." I've now he·rd that song. It was publisht in 2009 with that very name. Its words, beginning "Goodbye RP", are totally vacuous but its melody and the singing of them are not unpleasant. The hilarious irony of it for me is that it is sung in what can only be described as GB (okay, "RP"). Bragg didnt use the song in his program, or even refer to it, tho it seems to've provided his title for him.
His compilation contained good deal of 'music' of sorts besides other extraneous noise. It began with a cringe-inducing (for phoneticians at least) My-Fair-Lady excerpt about the 'rain in Spain' which was also reprised later. This opened a welter of mostly extremely brief quotes from a medley of scores of speakers. The average length of these clips was about forty seconds. They were frequently not allowed to be he·rd in peace. Daniel Jones's reading of an irrelevant and incomprehensible poem in medieval English was accompanied by sentimental string music. The excerpts from First World War faint scratchy dialect discs were backed by pastoral woodwind. Cheryl Coles was almost drowned out by gratuitous traffic noise. Alastair Cooke had to compete with a striking solo violin virtuoso. The Beatles' Sgt Pepper was trotted out. And so on. All this put the program firmly into the category 'entertainment'. It cert·nly was offen laughable. I shall leave mentioning what amused me to another posting.
Many readers ·ll know tht John Maidment has begun a revision of the adm·rable Speech Internet Dictionary.
If you dont know that little treasure-house already, I strongly
recommend an examination of it. It's a remarkably useful facility for
anyone int·rested in any aspects of the speech sciences. It
explains concisely and clearly all sorts of things you usually wont
find better described anywhere else online. He's invited us to offer
any advice that might make it even better by emailing firstname.lastname@example.org.
He sez tht anything on any technical term to do with speech, hearing,
speech technology etc is welcome. He's now revised letters A & B
and so I've been looking at C & D to see if I have any suggestions
for him. Here are a few, with one or two digressions to other letters
prompted by cross-references. Compare them with the existing entries at
http://www.phon.ucl.ac.uk/home/johnm/sid/sidhome.htm and see what you think.
Central vowel: A
vowel for which the highest point of the tongue is markedly removed
from the periphery of the vowel area. [ə] is the completely central
vowel. The other nine with recognised IPA symbols are ɪ, ʏ, ʊ, ɘ, ɵ, ɤ, ɜ, ɞ and ɐ.
If this revision shd be acceptable it wd require a removal of the
(debatable) reference to CVs 17 and 18 as central. These entries cd
well be either accompanied by or linked to an IPA vowels diagram.
Centralised Vowel: [ˈsentrəlaɪzd] A centralised vowel is one which is located to some extent towards the centre of the vowel area in relation to its nearest corresponding cardinal equivalent. Centralisation is symbolised by accompanying a cardinal vowel with the IPA "centralised" diaeresis diacritic placed over its symbol as eg [ë] or [ö].
Checked: a reversal of the order of the two definitions of this term wd place the older one first.
Citation form: Add "also known as lexical or isolate form".
Clipping: What d'you think of exemplifying proclitic rhythmic shortening as in maintain [meɪn`teːɪn] compared with main [`meːɪn] as well as pre-enclitic reduction as in manage [`mӕnɪʤ] compared with man [`mӕːn].
Which reminds me that SID hasnt got:
Enclitic: A term in prosody which describes an unstrest word which behaves rhythmically as if it were part of a preceding word eg her in Tell her /`tel ə/. Cf the noun teller.
Proclitic: A term in prosody which describes an unstrest word which behaves rhythmically as if it were part of a following word eg the article a in the phrase a way. Cf the adverb away.
Consonant Capture: (Proposed new entry) The process whereby a syllable attracts an adjacent consonant to itself. The English word weekend altho transcribed usually only as /wiːk`end/ is widely to be heard as /wiː`kend/ at least in the UK.
Creaky voice: (Proposed addition to entry) is symbolised by a subscript tilde as in [ɑ̰]. (Such an elaborated wording might helpfully compensate for browser inadequacies, font deficiencies etc. Cf Dark /l/ etc.)
Tilde: (Proposed new entry) The diacritic mark ˜ placed in Spanish traditional orthography above the letter n to produce ñ for the palatalized sound /ɲ/ and in IPA to indicate nasalisation as in [ã, ẽ, õ] etc.
Dark /l/: The symbol is [l] with incorporated tilde [ɫ].
De-accenting: Suggest for 'likely' substitute 'normally'.
Degemination: Praps an int·resting example from English wd be lamppost which can be /`lӕmppəʊst/ but which for some speakers, at least in fluent utterance, can be /`lӕmpəʊst/.
Vocal cords: Inste·d of "(Occasionally chords)" what about saying "No longer spelt 'chords' in phonetic literature".
Vocalised: [ˈvəʊkəlaɪzd] "Produced as a vowel-like sound rather than as a consonant. An example is the use of a vowel in the region of [o] in some accents of English in place of [ɫ]".
Isnt this rather better described in terms of a completed conversion rather than a process (which term wd be applicable if the entry title were 'Vocalisation')?
Diacritic: Rather than have only two examples, both detached superscripts, might one add also any of eg è, é, ê, ë, ė, ē, ĕ, ę, ç, å, š, ř, n̩, d̪, t̚ etc?
Dissimilation: Nice entry; but funny to quote only American English when any day you can hear various speakers on British radio and tv saying /`prəʊgӕm/ or /`pəʊgrӕm/ as offen as /`prəʊgrӕm/ if one cares to lissen closely.
Email your comments direct to John or go to his website and add them to the ones already there after his post 'SID rejuvenated?' of the 8th of August.
A teacher at the University of Minnesota recently mentioned that
some of her advanced Indian students were asking if she cou·d supply
them with a list of specifically Indian stressings of English words.
Indian English-speakers are prob·bly the champions at producing what,
from the point of view of GA and/or GB speakers, are unfamiliar
word-etc stressings. Indians have a long tradition of extensive fluent
use of English acquired much more from printed than spoken sources.
They also offen have very large vocabularies. Ergo they produce so many
of these items. Stress placement seems very free in most languages of
the South-Asia subcontinent and they may use very few of the normal
English reduced weak syllables pronouncing them with the vowels that
they wou·d have if they were strest even when they dont actually put
stress on them. I dou·t that compiling a long list of the kind askt for
wd be easy and, if practicable at all, wd be of any very great
advantage. At any rate, no colleague seems to have responded with any
sort of list of the kind appealed for.
Presumably the students hoped that such a list might help them avoid employing stressings that cd be more or less of a distraction for native speakers of (General) American English. They're mostly not much of an impedance to intelligibility, offen probably no more disagreeable than the differences between GA and GB stressings that native English-speakers take in their stride. When I've had students who've produce items that are okay in US but not usual in UK usage I generally havnt thaut them worth commenting on except praps to very advanced students who might find the topic int·resting rather than worrying. By the way, GA/GB contrasts in word stressings are fairly extensively exemplified on this website at paragraphs 4 to 7 of Section 3.1.
Having sed the above, I have to confess that one uniquely strange Indian stressing comes to mind not he·rd from one of my students but uttered by a very good fr·end and colleague many years ago when we were nei·bours and both faculty members at the University of Tehran. He was the Professor of Indology and had a superb command of English. One word he used offen was "particularly'" which he wd stress very strongly on only its first syllable uttering all the others quite weakly and extremely quickly. There was really never any problem of intelligibility in any context I can recall but it was admittedly rather startling at times.
Anyway, the sorts of items that Indian students of English produce are in very few cases types unique to Indian usage. Certain unusual stressings are obviously best avoided if they're compounded with misperceptions of spelling that a fully literate native speaker of English wd not be likely to produce, for example sovereignty as /sɒv`renəti/ or uncertainty as /ᴧnsɜː`tɪnəti/. Occasionally an accompanying faulty pronunciation might cause the greater trouble than the mere stressing. A. C. Gimson, who had a very nice sense of humour, used to enjoy mentioning being told that he was a very "impotent" man where the full weakening of the vowel of the middle syllable of important from /ɔː/ to /ə/ was more of a problem than the early stressing.
Most of these unusual stressings are due to merely locally unsuitable interpretation of a spelling. They offen produce a form which is, or conceivably could be, in use somewhere in the English-speaking world or agen might in the past have been accepted usage. Others may occur as relatively idiosyncratic usages of native speakers. Some examples I've noted include advertise as /ӕdvə`taɪz/ which, like any -ize verb, is or was accepted in at least some British-Victorian, Scottish, Irish and Caribbean educated usage; antimony as /ӕn`tɪməni/; automaton as /ɔːtə`meɪtən/ character as /kə`rӕktə/; circumstances as /sɜː`kᴧmstənsɪz/; colleague as /kɒ`liːg/; detritus as /`detrɪtəs/ (an idiosyncrasy of Robert Graves); determined as /`detəmaɪnd/; development as /devə`lɒpmənt/; emergency as /`eməʤənsi/; mischievous as /mɪs`ʧiːvəs/ a current British regionalism; thesaurus as /`θesərəs/. We can hardly be surprised that students dont notice some of these: we have so many words that have two or more accepted stressings like automobile or nomenclature. The best advice to them is to avoid guessing but to keep checking with their dictionaries.
The title of this blog is a quotation of that of an hour-long
program broadcast on the evening of Saturday the sixth of August on BBC
Radio 4. Any regular reader of these blogs 'll know that they arnt
usually to be taken as entertainment but as serious comment on linguistic
matters; so be warned that I hesitated over whether to mention this
program at all. It contains a farrago of trivialities as one must
expect because serious programs are assigned to Radio 3 and this was
obvi·sly offer·d as entertainment for the gen·ral public. It was
available for a week after it's transmission from a BBC website called "Archive on 4: A look back at programmes and recordings from the BBC archives"
where it'll now only be accessible until 9:02 pm tomorrow Saturday (13
Aug) tho it'll no dou·t reappear in R4 schedules before very long.
It consisted of a very large number of very brief clips from a wide variety of speakers, few of them much longer than about a minute, linked by presenter Melvin Bragg who was described as examining whether "the 400-year reign of Received Pronunciation (RP) is finally over". He was introduced by a continuity announcer who quoted the program's title with the interrogative intonation / ˈɑː `piː| ɑːr ́aɪ piː/. He began by illustrating his own, what you might call capitulation to 'RP', going from his original General Cumbrian (he comes from Wordsworth country) and a town sub-variety laced with gipsy slang to his present "near RP" as Wellsians might categorise it. He's far from given in to 'RP' entirely as we see from his pronouncing his home town Wigton as /wɪ(k)tn/ which isnt given so even in the BBC Pronouncing Dictionary of British Names leave alone the big three (EPD, LPD, ODP).
Among the welter of snippets the most lively and int·resting ones come from Professor Lynda Mugglestone of Oxford (author of Talking Proper). She's an enthusiastic speaker of 'non-RP' tho her originally Lancashire accent has plenty of influences from it. At one point she went so far as to say provocatively that "Received Pronunciation is gradually being marginalised". As you'd expect of a program with the title it has, a cert·n amount of ref·rence is made to Daniel Jones who launched the unfortunate term 'RP', apologetically but ill-advisedly, into circulation in 1926. Bragg sed: "At the beginning of the 20th century, Daniel Jones strove for precision, tracking back to reveal how English had evolved over the centuries. He was passionate to communicate the fluidity of pronunciation — and for him how RP was clearer, simpler than the dialect spoken in the 14th century as he described in Our Changing Speech". This last reference was to a pair of BBC talks deliver·d by Jones in Feb·ry 1949. These remarks of Bragg's were an embarrassingly garbled representation of Jones's aims and attitudes, evidently concocted in desperation at the absence of any other recording of Jones more, or rather at all, relevant to his theme. The talks were published, as a text with accompanying tape recordings, by the Danish National Institute for Educational Media in 1985. Further quotations included recordings of John Reith, a BBC Brains Trust session, Beryl Bainbridge, Joan Bakewell, certain politicians and further comments many of which were so hilariously silly it'll be hard to refrain from quoting from them.
I'm always very reluctant to refer to anyone's "mispronunciations". After all, so many of the pronunciations we cheerfully regard as 'correct' started life as mistakes. There's no historical "justification" for the last <l> in the word syllable any more than the one in could etc. The word syllabus, OED sez, "appears to be founded on a corrupt reading...in some early printed editions...of Cicero". Apostrophe (the sign) is dodgy too: shou·dnt have its final <e>. The confusion self-deprecation has driven out the proper self-depreciation in the last generation or so. OED quotes some 1886 botanists "The name of Jerusalem Artichoke is considered to be a corruption of the Italian Girasole Artic[i]occo, or Sunflower Artichoke". Merchandise has no bizness being spoken with /-s/, nor hermitage or humble-pie with /h/, nor sound with /d/ and so on.
Dodgy pronunciations are of many kinds. One that came up recently was /hiːnəs/ for heinous a version Murray didnt record in OED in 1898 and OED3 hasnt come round to reconsidering yet. It was recognised in 1988 in Gimson's 14th edition of EPD as revised by Ramsaran. LPD1 in 1990 and since has specifically identified it as not a 'received' pronunciation, but EPD and ODP give it recognition and my sympathy is with their view. It's been around as a subvariant for a long time, since Sheridan's 1780 Dictionary at least. I he·rd /iː/ from excellent BBCtv newsreader Susanna Reid a day or two ago.
Even people with the most brilliant intellects quite offen exhibit idiosyncratic pronunciations, especially where tonic placement is concerned. The other day I he·rd a distinguisht speaker say auto`maton. More than once recently I've he·rd Jim Al Khalili say an`timony and `phlogiston, Jonathan Dimbleby say `scintilla, Andrew Marr say trage`dian, Alastair Sooke say `cadaver and I've he·rd sev·ral people, including the BBC's Ed Stourton, say bi`opic for `bio-pic. More examples of this kind can be seen at our Blog 049.
There's sometimes a dilemma for people who find that the traditional English version of a loanword has a different stressing from the one its own language has. An extreme case occurred the other day when a British academic with a chair in America presented an hour-long tv program he'd written where the name Bolivar turn·d up repeatedly. He began with references to `Bolivar and half way thru the program, with a grinding change of gear, turned over to Bo`livar. I imagine it was noticed but it was no dou·t felt to be not worth the expense to re-record the sound-track. Anyway, I think consistency is a vastly overrated 'virtue', as my own writings suggest.
Performances of early plays offen show actors using anachronistic pronunciations. One famous actor startled me recently by saying /`eɪgeɪt/ not /`ӕgət/ for agate in quoting Shakespeare. If he'd not confined his attention to texts with the spellings modernised, he'd've found that the word, whether faithfully reflecting Shakespeare's own handwritten versions or not, cou'd be spelt by contemporary printers variously as aggat, agget and aggot etc. There's a wonderful but sadly overweight and unwieldy volume of Shakespeare's complete works, produced by Oxford University Press in 1986, entirely in the spellings in which they were originally printed. It's the kind of thing such people might do well to look at from time to time.
There are undeniable mistakes as when a famous biologist cd be he·rd
to say /`erənəs/ for erroneous or the late dear Humphrey Lyttelton
saying not /ə`raɪ/ but /ɔː`raɪ/ for awry (one
has even he·rd it on occasion as /`ɔːri/ from the odd less
sophisticated person). Finally, there are the dog's-bre·ckfast shots at
forren words that are so very common. The specially egregious example
that triggered the present ramblings occurred the other day when a
female newsreader, whose name I prefer not to mention, in an item
referring to a performance by the Israeli Chamber Orchestra at the
recent annual Wagner festival at Bayreuth /baɪ`rɔɪt/ pronounced that name as /beɪ`ruːt/ which is, of course, the usual pronunciation of Beirut.
To the three General British examples of the type of anticipative
assimilatory devoicing given in the previous blog on this topic we
may add various others. The collocations have to and had to dont receive the special treatment accorded to used to by LPD etc but in a similar way they are offen pronounced as /hӕf tu/ and /hӕt tu/. Breadth, length and width
have the alternant assimilated forms /bretθ, leŋkθ & wɪtθ/ recorded
in LPD, EPD and ODP (OED3, not revised for them yet, has only /d/
forms). LPD has assimilated forms of the archaisms didst, shouldst and wouldst
(/dɪtst, ʃʊtst & wʊtst/) which one imagines are largely arrived at
by extrapolation rather than c·nfirmed by any quantity of observations.
That leads one to wonder why couldst and hadst
dont get the same treatment. Jones in EPD always (1917-63) specified
that the weakform of should /ʃt/ "only occurs before breathed
consonants" (a qualification dropt by Roach et al). 'Breathed' /breθt/
was an older term for 'voiceless'.
The phrase of course is very offen /əf `kɔːs/. The coinage term fivepence (archaic since 1971) is recorded in all the big three with its alternant /faɪfpəns/ and LPD also has /`faɪfpəni/ for fivepenny. There are still among many GB-speakers some relatively unnoticed simple words that exhibit this feature. In my Concise Pronouncing Dictionary of 1972, despite its avowed restrictions, I felt that certain items were so very frequently in use that they ought to be put on record. They included eg /`ӕpsəluːt, `ӕpsns & əp`sɜːd/ for absolute, absence and absurd. I have a suspicion that certain people are particularly inclined to adopt the assimilated version when a word is being uttered very emphatic·ly. One of the most easily observable such speakers at the moment is the current prime minister David Cameron who very regularly assimilates a favourite word of his absolutely to /`ӕpsəluːtli/. Anyone who has a copy of the wonderful Channel 4 1981 production of Brideshead Revisited can hear Jeremy Irons say You were a[pː]solutely right, Lunt. It's my opinion that pronunciations like /əp`sɜːv/ for observe or /`ɪpsn/ for Ibsen are unlikely to sound remarkable to most people today. (That's how Norwegians say his name as it happens.)
In spite of the above, it has to be acknowledged that this kind of assimilation, as Akamatsu's remark suggested, is very far from being an extensive feature of connected GB speech. The obvious unusualness to other people of the typical regional Yorkshire assimilation underlines that fact. This almost exclusively demotic Yorkshire phenomenon is apparently quite recessive and markedly absent from mesolectal Yorkshire-accented speech. It has been described in an aside on this website at ¶29 of Section 7.4 'The General Central Northern Non-Dialectal Pronunciation of England'. Examples of it noted from the notorious miner's leader Arthur Scargill of Barnsley include /`brɔːtkast/ broadcast, /sɛt/ said (before something) and /θaʊznt paʊnz/ thousand pounds. This last was relativ·ly upmarket for him because the usual basilectal Yorkshire version of thousand is without any /d/. J. B. Priestley didnt normally display any of these specially Yorkshire assimilations but he cou·d be he·rd to say /bratfəd/ praps out of sentiment or force of long held habit — Bradford was where he came from. Margaret Thatcher's press secretary Bernard Ingham was to be noted as saying goldsmith as /`goʊltsmɪθ/. Earlier forms of English too had this assimilation to some extent as we see if we compare placenames like Ratcliff and Radcliffe, both 'red cliff', and Radford and Redford, both 'red ford'.
My fr·end and former Leeds Univers·ty colleague Tsutomu Akamatsu publisht in 2009 (in Moenia,
Revista Lucense de Lingüística & Literatura, Vol. 15. Universidade
de Santiago de Compostela) an article entitled 'Yorkshire assimilation'
discussing the phonology of this phenomenon of (essentially demotic)
'Central Northern (British) English pronunciation'. He began his
account with illustrations cited from Wells's 1982 Accents of English at pages 366/7 where the phenomenon is described as occurring "when
a final voiced obstruent comes into contact with an initial voiceless
obstruent, either within a compound word or across a true word
boundary, and has the effect of completely devoicing the former
consonant". In the course of his discussion Akamatsu quoted me
at his page 132 as confirming that (however I exac·ly exprest it at the
time) I cou·d well believe that demotic Yorkshire speech wd include in
the phrase 'Bradford City' pre assimilations of both d's to /t/. I
cou·d indeed attest having very offen he·rd the first /d/ so converted
but I acknowledge that I presume that the latter one is so treated by
speakers who have the feature. On the same page he remarks that the 'Yorkshire assimilation is a unique phenomenon ... among various types of speech in British English'.
In terms of its frequency and regularity as an inter-word process I
agree completely but the statement has stimulated me to comment on the
occurrence, as a restricted usage occurring sporadically, of
assimilation of voiced obstruent consonant to a following
voiceless one in some mainstream General British more or less
intra-word-type sequences, examples of which are the following.
Newspaper is currently listed as being usually /`njuːspeɪpə/ in EPD (2006). That was the sole form listed in EPD1 in 1917 but by 1956 Jones was including a "rare" variant /`njuːzpeɪpə/. In OED1 Craigie in 1906 gave the unassimilated form alone; the assimilated one has now appeared in second place in OED3 but ODP, from the same stable, completely omits it. In 1972 in my modest-length EFL-oriented CPD I gave space only to the assimilated form. Gimson in EPD in 1977 removed the 'rare' but still gave /`njuːzpeɪpə/ only subvariant status. Subsequent EPD editions have followed suit. The first form listed in LPD1 in 1990 was the assimilated one but from LPD2 the unassimilated form has been prioritised, a decision very possibly influenced by responses to a British poll on the matter reported in that edition.
The word used in the phrase used to, where the verb element normally has the sense 'accustomed', appeared in EPD1 in 1917 as "juːst" in which the italicisation of the final /t/ was to be understood as signifying that it tended to be omitted but 'both forms' were 'of approximately equal frequency'. In OED1 in 1926 Craigie at sense 20 ('accustomed') of the verb use noted "... now only in past tense used to, with pronunciation (yūst tu, yū·stŭ)". The NED symbol (ŭ) no dou·t signified that in non-deliberate utterance the value cd be taken to be usually with schwa. OED3 has the note "In the modern period .. in the collocation used to, the final (voiced) dental of used was assimilated in British English to the initial (unvoiced) dental of to immediately following ... and subsequent assimilation of the preceding sibilant resulted in the usual current pronunciation in these senses, British /ˈjuːstuː/; compare, by a similar process, U.S. /ˈjuzdu/. Artin in Webster 1961 had labelled the unassimilated form as occurring ('when “to” follows immediately') only 'sometimes' and didnt specify the phonetic value of that 'to'. OED has had since 1986 an entry "used-to-be, n." by OED3 transcribed "Brit. /ˈjuːstəbi/ , U.S. /ˈjustəbi/".
The parallel usage of supposed to as /sə`pəʊs/etc, before a following (rhythmically integrated) to, was first recorded to the best of my knowledge, in 1972 in my Concise Pronouncing Dictionary. LPD from 1990 has had /spəʊst/ etc. MWO (Merriam Webster Online) has a subvariant "often -pōst" (for some senses) where it is not specified as limited to occurrence in this combination and even demonstrated in isolation. EPD, OED and ODP don't record this usage.
Recently my Japanese colleague Tami Date, who has a keen int·rest in prosodic matters, invited some comments from me. He’d been holding a workshop in which he used as material for study some extracts from the American textbook Accurate English (1993: 237) by the late Rebecca Dauer. Some items that provoked discussion were sentences he indicated as (1) and (2) in a dialog appearing in the book as follows:
Read the following selection with appropriate intonation.
Ruth is talking to Cody and his teenage son, Luke, who have just returned from a hunting trip.
“It’s ˈbad `news,” she ˏsaid. “I’m `sorry.”
(1) “Your `mother’s passed a``way”. “Grandma ́died?”
(2) ˈasked ˈLuke, | as if corˊrecting her.
To indicate the intonations used by the (rather young-sounding) American speaker, on the recording provided with the book, I’ve added to these four lines simple broad tone markings and, where normal punctuation needed amplification, tone-unit (aka intonation-phrase) bound·ry-marking vertical bars These are probably self-explanatory for most readers, tho I refer anyone who wants explanations of any of these to §8.3 on this website 'The Recognition of English Tones'.
It’s very surprising to me that at (1) “Your `mother’s passed a``way” (with Fall mark doubled to emphasise that it’s higher than the first Fall) shou·d’ve been intoned in this way. The second Fall, on 'away', strikes one as very unnatural ie improbably animated for the spontaneous announcement of such surely very sad news. When Tami requested a British colleague to record the same dialog, she intoned it as “Your `mother’s passed away” with no accent at all on 'passed away' giving it bottom pitch. Equally appropriate wdve been “Your `mother's | passed aˎway” with a low fall on 'away'. Her version was what I prefer to call a 'Happening Clause' or 'Happening Remark' (aka 'Event Sentence'). See my Blog 309.
Tami said of his workshop participants’ reactions to (2) that "all the attendees looked quite puzzled to hear the rising tone of "... as if corˊrecting her". He suggested that it might relate to the fact “that in reporting conversations, an author frequently uses expressions he said, he asked and so on. When these reporting clauses follow a quotation, they form part of the intonation pattern of the quotation itself and are affected by what precedes them”. He's perfectly right that reporting phrases often occur as tails to tones used in direct speech tho they’re usually quite short — typically of only two or three words. However, I’m afraid th·t that’s not what we have here. “Grandma ́died?” | ˈasked ˈLuke, | as if cor ́recting her cert·nly c·ntained reporting expressions in ˈasked ˈLuke etc but none dependent for its pitch values on the final tone of the preceding direct speech. We have a clause of a pair of level tones the latter of which is, as is normal, slightly lower so that both words are accented. This is followed by another separate clause (with a low prehead ) having an upper rising climax tone of its own on 'corˊrecting'. Tails to tones by definition contain no accents.
By the way, in this phrase the word 'correcting' is ill chosen. With the high rising tone it carries, it can hardly be a correction even by implication. Explicit corrections only have fully descending tones or (final relativ·ly) low rises as climax tones. What the text might better read is "as if questioning her". An expression of incredulity isnt a "correction", but it’s quite possible to take this simple high rising tone to be an example of the "checking" tones that’ve become widely noticed as they’ve spre·d in the last three decades or so initially from the Pacific-Rim areas of Australia, New Zealand and western North America.
The intonations used by Tami's British speaker were
“ It's ˈbad `news, she ˏsaid. "I’m `ˏsorry.”
“ What `happened?”
(1) “Your `mother's passed away.”
(2) “ `Grandma `ˏdied?” | asked ˎLuke, | as if cor`recting her.
We see that at (1) she employed the very idiomatic but perfectly commonly heard native English-speaker’s usage in which with such happenings clauses the subject receives the only accent and the predicate is completely unaccented. At (2) she employed a Fall-Rise complex climax tone which produces the necessary interrogative effect. She, too, chose to avoid entailing the reporting expressions but, like the American speaker, spoke them as two separate successive clauses — tho with different intonations from hers.
Here's the problematic dialog: