Archive 18 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

Aspiration in Dictionaries

At the 11th of March I was on holiday in Egypt so that I was unable to congratulate John Wells on reaching his seventieth birthday then. I do so now wishing him good helth and many more productive years. His example inspired the beginning of these “phonetiblogs” and has stedily kept my enthusiasm for them going with his constantly stimulating writings. Among his recent items have been an admirably clear and full account of the phenomenon of aspiration in GA and GB under the title VOT posted on the 14th and 15th of April. What follows now is a footnote to that account.

EAL (English-as-an-additional-language) users may need occasional help on aspiration if they’re keen to approximate closely to native speakers’ usages in this matter. They may find in using pronouncing dictionaries that there isn’t complete accord between them. However, these days they can be expected to agree on showing a word like mistake with the second syllable beginning with the /s/ thus /mɪ`steɪk/ so that it’s clear that its /t/, being closely articulated with the following /s/, will ordinarily be free of any noticeable aspiration and thus different from the /t/ of mistime /mɪs`taɪm/. Things weren’t always so: the Daniel Jones EPD up to 1977, when Gimson finally got round to doing his really extensive re-editing of it, had only /misˈteik/ and /misˈtaim/. The first pronunciation dictionary I know of to make the distinction was the J. S. Kenyon & T. A. Knott 1944 Pronouncing Dictionary of American English tho it may well have been a feature of the previous Webster dictionaries: they certainly have exhibited the practice since at least 1961. The first British pronunciation dictionary to feature it was my CPD of 1972.

It’s not always easy for the lexicographer to decide between the two representations. Quite often both types exist tho editors are inclined to ignore the fact and offer only one transcription probably usually mainly in the interests of economy of presentation. The ODP editors seem to think that there’s a transatlantic difference for dystopia having /dɪsˈtəʊpɪə/ for GB but /dəˈstoʊpiə/ for GA. They may, of course,  believe that a previous schwa conduces to the preference for the /s/ to begin the following syllable. For dyspepsia both ODP principal editors have the second syllable beginning /-ˈp-/. On the other hand, current EPD has /dɪˈspepsiə/ but LPD /dɪsˈpepsiə/ the latter of which seems more suitable to me at least if you only offer one version. For discourage LPD starts the second syllable after the /s/ while EPD starts it with it. They both only offer one type for GA and GB. ODP’s GA has /-ˈsk-/ only but for GB it gives /-sˈk-/ first and /-ˈsk-/ second. This seems to be one of the fairly rare cases of acknowledgment that both types may occur.

Various pairs of alternatives are, however, offered in EPD eg at misprint, mistrial and mistrust. On the whole the three dictionaries seem to agree tho EPD seems to be the odd one out in cases like distaste, miscast, miscarry and misconduct for which it has /-ˈs-/ which seems in those cases at least to be more likely to be a subvariant version if anything. There are other words for which a choice might well be offered with regard to which the only form you generally find is the one which in my observation isnt the predominant one. The exception here is MWO (Webster Online) and its ODP adherent at any rate at the word sixteen. Keeping alternatives to a minimum as was very desirable for my “Concise” Pronouncing Dictionary I had no hesitation in giving only the transcriptions /fɪ`ftiːn/ and /sɪk`stiːn/ for fifteen and sixteen. Even MWO plainly drew back at the sight of an English syllable represented as beginning with /ft-/!

$ Pronunciation's Origins (iii)

1. Further vowel phoneme distributions showing GA maintaining WDP values include those where GA and GB differ by exhibiting in tonic syllables either predominantly or subvariantly one or the other of /i/ and /e/. Thus /i/ is the predominant or variant value recorded in WDP for either˚, leisure, neither˚, zebra˚ and zenith. By contrast /aɪ/ predominates in GB for either and neither and /e/ is the exclusive value in leisure and zenith. Only in the latter half of the last century /e/ replaced /i/ as the favoured GB form for zebra. WDP gives the usual GA but mainly alien-to-GB /e/ for epoch, devolution, febrile, scenic and the first syllable of prede`cessor (as it is strest in WDP). GB prefers /i/ in evolution but not exclusively. Non-dialectal Scottish usage has /e/ in devolution. The only WDP form of sarsaparilla ends /-elə/ which is an MWO-listed subvariant not recognised in any of LPD, EPD or ODP.

2. Been is given only in GA style ie as /bɪn/ in WDP. In the sense of “learner’s book”, WPD gives primer only with /ɪ/. MWO gives this as usual GA, referring to the usual GB form with /aɪ/ as “chiefly British”. For the nouns process and progress WDP prefers /ɒ/ in  GA fashion to GB /oʊ/. The strest vowel in `halibut is listed only with /ɒ/ the reflex of which survives subvariantly in GA (and in minority Yorkshire regional usage). WPD has a variant /`sɑsi/ of saucy that ODP lists as a current GA subvariant. The preterite form of shone with /oʊ/ was a familiar variant to Walker, tho not the one he recommended : it is now the only usual GA form but alien to GB. The WDP /oʊ/ sole value of the latter syllable of thorough has maintained that diphthong in GA but is only /ə/ in GB. WDP prefers hover with /ʌ/.

3. The next largest group of words with vowel contrasts to the /ɑ ~ ӕ/ set which we described above in Blog 178 §§6-to-9 is the set in which GA /ɔ/ corresponds to GB /ɒ/. These similarly are largely found in the context of following /f, θ/ and /s/. Examples are coffee, cough, trough, off, office, often, soft, broth, cloth, moth, cross, lost, gloss, moss. All these had the /ɒ/ vowel in WDP but in an odd remark at soft Walker reveals that there was some currency of /ɔ/ in his remark “sawft borders on vulgarity”. These words all now have /ɔ/ in GA and did so also in Victorian GB but reverted (probably in consequence of what I’ve termed “reverse spelling influence”) to /ɒ/ fairly completely in GB by the second quarter of the last century. The word sausage may have been involved in this migration along with austere, cauliflower, false, fault, vault, falcon, laurel, Laurence, auction, Vauxhall etc.

4. The form beginning /tru-/ of truculent, the only one in WPD, is vouched for by MWO as still current (subvariantly) in GA but seems to have died out from GB in the middle of the last century (Gimson dropt it from EPD in 1977). GA has preserved the WDP sole form of vineyard as /`vɪnjəd/ which has now given way to /`vɪnjɑd/ in GB (pace the three pronouncing dictionaries). At the entry heigh-ho MWO gives /ˈhaɪ.ˈhoʊ/ as first variant which matches the only WDP version but isnt given in any of the three major British pronunciation dictionaries even in their representations of GA. For waistcoat WDP had only /`weskɒt/. GB clearly prefers /`weɪskoʊt/. I find I agree more readily with EPD’s “old-fashioned” label for /`weskət/ than LPD3 which records it as a subvariant with no comment. It’s rather surprising that MWO thinks /`weskət/ predominates for waistcoat in GA: one notes that the Webster line is not followed in this by ODP’s American editor. He takes a contrary view also with wassail which MWO represents as predominantly spoken with the weak ending which is the only kind in WPD. All other authorities agree that it usually ends /-eɪl/. He agrees with the EPD American editor in not following MWO in recording the type /huə(r)/ for whore which WPD gives as the first of two versions. At the word hundred WPD gives “hun-durd” (with a misprinted wrong number over the u for the unstrest syllable, that for /u/). It was doutless intended to convey the value in a variant given in MWO as “\-durd\”. Finally in this mixed bag of observations we see that Walker insists that /-ʃiə/ is the only type of form taken by -shire as a suffix: this is listed by the dictionaries other than EPD as the predominant American usage by contrast with the British use most often of /-ʃə/. 

5. Lastly we shd mention the final unstressed syllables most often spelt as in happy but also as in coffee, money, caddie etc. These in (at least later) Victorian GB usage had become generally [-ɪ] but that value was (pace Kenyon and Knott) and is a minority form in GA. The present GA predominant value [-i] has doutless descended from the GB precursor of the eighteenth century when it was the evident norm. Of course the pendulum has swung agen in England in the last two three decades bringing back [i] as the current GB norm. This has no dout been as a result of speakers perceiving [i] as better because clearer speech. Many of us non-RP speakers have had the pleasure of coming into fashion with no effort.
This little investigation has witnessed that, altho American-British pronunciation differences have no dout arisen in a variety of ways, a notable number have been arrived at by the course we have suggested. 

American Pronunciation's Origins (ii)

1. Considering phoneme occurrences in WDP we find that herb, humble and humour are listed with no /h/ as is usual or at least common among GA speakers by contrast with GB. No forms without /h/ are given at huge or human which have such GA subvariants.  All wh- words are listed with /hw-/ which of course is recessive in GA but has not undergone the total abandonment which has been its fate in GB in unselfconscious use. Walker’s prescriptive prejudices no dout manifest themselves in some of these items.

2. There seem to be no signs of GA’s extensive yod-dropping in WDP except that at the entry for duke we find Walker referring, no doubt significantly, to “a slight deviation often heard in the pronunciation of this word as if written “Dook” [which] borders on vulgarity”. (He adds that he regards “Jook” as “not so vulgar”). Conversely figure is sed to have “a coarse and a delicate pronunciation” ie respectively in the GA and GB styles /`fɪgjər/ and /`fɪgə/.

3. Yod coalescences with sibilants are more general than in GA or GB. WDP to an extent prefigures GA in showing them in transcriptions like /`besʧiəl/ for bestial and similar versions of celestial, cordial, courteous, Elysian, flatulent, glazier, grazier, halcyon, hosier, issue, nauseous, pendulum, petulant, tissue etc tho surprisingly not for credulous or fraudulent. He even has /(d)ʒ/ etc in audience and, at least alternatively, in comedian, euthanasia, gradiant (so spelt), hideous, ingredient, medium, obedience, odious, sedulous, tedious and trapezium. He has only /ʃ/ in conversion, diversion, excursion, perversion etc. There is no sign of /`pɪkʃə/ for picture etc. He deplores presumptuous, unctuous and voluptuous as /prɪ`zᴧmpʃəs/ etc so we see that these common if criticised versions had already appeared in his day. 

4. Among other consonant usages more familiar in GA than GB, we find greasy with only /z/ and erase, hypotenuse, parse and resource with only /s/. Vase is either /veɪz/ or /veɪs/. He records with with /ð/ “or” /θ/. Only /ð/ is given at brothel, a GA subvariant not recorded for GB. At schedule we find /ske-/ and /se-/ with the traditional GB /ʃe-/ barely mentioned and then unfavourably. The form of suggest he prefers has /g/. That of tourniquet has a final /t/ which has remained in GA but finally dropt from GB in the middle of the last century.

5. Among stressed-vowel choices /eɪ/, which we find in ar`mada was labelled “old-fashioned” in EPD, none too soon, in 1937 but for GA it is still deemed current subvariantly by MWO (Merriam Webster Online). We find drama with the variant /drӕmə/ which is usual GA. In the following /eɪ/ is at least subvariant in GA: farrago, gala, gratis, patriot, phalanx (/`feɪlӕŋks/, placable (placate is not listed), rather (as well as having GA /ӕ/ but not GB /ɑː/), satyr, strata, vase, virago.

7. WDP regularly shows with only /ӕ/ at their alphabetical places items like after, craft, raft, staff; laugh, draught; bath, path; ask, castle, class, grass, last; plant, chance, dance, branch, slander etc. However, in his introductory “Principles” Walker claims that /ɑ/ as in glass, last etc “seems to have been for some years advancing to the short sound of this letter”. He adds, rather mystifyingly, that “pronouncing [ɑ] in after, answer, basket, plant, mast &c as long as in half, calf &c borders very closely on vulgarity”. He also sez that /ɑ/ “was formerly more than at present found before ... n ... as dance ... grant ... slander etc” and contradicts the evidence of his alphabetic entries when he sez bath, path, lath etc have /ɑ/. Another inconsistency, perhaps again suggesting incipient change in progress, occurs when slant has /ӕ/ but aslant has /ɑ/. 

8. Other words given with /ӕ/ include autograph, epitaph, lather, rather and telegraph. On the other hand contractions can’t, han’t (ie haven’t) and shan’t are noted to have /ɑ/: he has no entry aren’t or an’t.  The latter syllable of gal`lant with /ɑ/ wd seem to be embracing a French style. For wrath WPD has /ɒ/ but mentions also a minority form matching GA /rӕθ/ which is alien to GB.

9. A number of words which strikingly differ by actually appearing with /ɑ/ as alphabetical entries in WDP includes askance, chandler, command, demand, reprimand, stanch; daunt, flaunt, gaunt, haunt, jaundice, saunter, vaunt and master (unlike mast and plaster). For Maunder, raunchy and vaunt GA has subvariant /ɑ/ according to MWO.

10. There can be little dout that in Walker’s day and for generations afterwards the clearcut differences in this area we find today between GA and GB were yet to be establisht. We shd remember that even as late as 1882 the NED (OED) pronunciations committee decided on a cover symbol to represent simultaneously /ӕ/ and /ɑ/. The fact that today such words are entered in OED for British English with the two types signifies the rejection of the previous policy of showing “received pronunciation” which OED itself describes as “the pronunciation of that variety of British English widely considered to be least regional”.

American Prons' Origins (i)

1. Certain differences between the American and British pronunciations of English have originated because particular forms which they once had in common fell out of use completely or became uncommon variants on one side of the Atlantic. I propose to discuss this matter, as far as may prove feasible, along the lines on which I treated the GA (General American) and GB (General British) accents of English in my article comparing the two accents at §3.1 on the main part of this Website. The present notes are based entirely on usages documented in that most important source of information on pronunciation choices in England in the Eighteenth Century namely John Walker’s Critical Pronouncing Dictionary which delt with forty or fifty thousand words in 587 pages preceded by 89 prefatory pages of general observations he called ‘Principles’. I quote it from an obviously unchanged Thomas Nelson reprint in 1849 of its second edition of 1797. A particular value of this work, henceforth referred to as WPD, is that Walker regularly reported any different opinions from his own (which cou’d occasionally be rather silly) from a dozen or more other orthoëpists and lexicographers of the period. The transcriptions given are not as in the original text of WPD but are interpretations of them using IPA symbols as at §3.1 which of course roughly indicate phoneme types rather than precise phonetic qualities.

2. In the first place I treat matters essentially of word stress. For the words authorization, civilization, organization, solemnization WPD lists only what is the usual GA style having immediately pretonic weak stress eg /ɔθərə`zeɪʃn/. Not even as alternatives does WPD  list for these words any of the strong-vowel versions with /aɪ/ usually preferred in GB eg /ɔθəraɪ`zeɪʃn/.

3. In respect of word tonic stresses WPD often gives, as the only stressings recorded, forms with the principal accents earlier than GB most often has but which are either predominant or at least current for the words in GA. Among these are a`boveboard, `artisan, `cement, `cervical*, `circulatory˚, `controversy, `corollary, `courtesan, `despicable,  il`lustrative, `laboratory, `medullary, `miscellany, `obligatory, `partisan, `spinet, `towards, um`bilical*, `urinal. (*These two words are currently represented as predominantly later stressed than shown here in all three major pronunciation dictionaries but seem to be at the tipping point of a change to forestress as can be witnessed in the BBC recommendation to broadcasters of the stressing `cervical.)
 
4. Tonic stressings later than those normal in GB but predominant in GA are given in WPD as alternatives for adver`tisement and bi`tumen. WPD shows stressings later than are normal in GB but usual in GA for bu`reau, en`core, finan`cier, pre`cedent, prema`ture, pre`mier, pro`lix. GA has also the subvariant va`gary. WPD endstresses cha`grin as in GA (tho not with /ɪ/ but as /ʃə`grin/). The only stressing given for eti`quette is rather surprisingly not a GA usage but is a subvariant in GB according at least to LPD3. Similarly WPD enve`lope shows the late tonic which is presumably due to its perception, which it didnt maintain in GA, as a French loanword. Contrast `harass which remained only so strest in GB until the GA version was borrowed as a variant late in the last century. The only WPD stressing stalac`tite is that of neither GA nor GB which differ, each in its own way, respectively as sta`lactite and `stalactite. WPD includes rather few compounds: some found to match GA were `bandy-legged, `foulmouthed, `mealy-mouthed, `potbellied and `jew’s harp. The stressing post`office has not persisted in GA or GB tho it has done so in Ireland.

5. WPD supplies often dubious accounts of strong versus weak non-tonic syllables. Walker’s notations in this respect are, as Sweet sed of Bell’s, frequently influenced by “artificial elocutionary habits”. A notable exception occurs in the treatment of words with the ending -ile. These are all clearly shown only in what has become the usual GA weak style with /-(ə)l/ rather than the GB (post-Victorian) strong /-aɪl/. They include docile WPD /`dɒsɪl/, fertile, fragile, futile, imbecile, juvenile⁺, servile, steril (so spelt), tensile, textile, virile, volatile.

6. Disyllabic causative verbs ending -ate and their derivatives (when not embodying prefixes) all show in WPD only the GA usual front-stressed pattern eg `castrate, `dictate, `frustrate, `prostrate, `rostrated, `rotated, `stagnate, `truncate, `vibrate. Even e`longate, in`culcate, im`pregnate and se`questrate show the GA style. At `serrated WPD matches neither GA nor GB. Similarly neither in GA nor in GB fashion im`placable as /ɪm`pleɪkəbl/ preserves the /eɪ/ of the parent verb placate (omitted from the body of the dictionary but given at p. 78) as it persists in GA.

7. WPD exhibits in various cases divergencies of practice which persist to this day regarding eg `contrite versus con`trite, `recondite versus re`condite. It records some rivalries which have now been settled eg by con`versant almost completely replacing `conversant on both sides of the Atlantic and `aggrandize similarly giving way to ag`grandize.  Also de`monstrate has been quite abandoned in favour of `demonstrate. The same might be said of pe`remptory except that here, along with their similarly archaic wigs, the British legal profession are sed to maintain `peremptory.

8. WDP gives both `corollary and co`rollary each of which was to become the sole version respectively in GA and GB. While `controversy has remained the sole stressing in GA, it has largely given way to con`troversy in GB. In GA so`norous has resisted the change to `sonorous which has occurred in GB; de`corous has generally become `decorous, displaying a tendency whose evident newness prompted Walker to refer to it as “shocking”. He also deplored the stressing con`trary as typifying the “illiterate and vulgar” but it stubbornly persists as an informal synonym for “perverse” with that stressing no dout perpetuated by the enduring popularity of the nursery rhyme whose scansion requires it. He was nothing if not prescriptive but he was fortunately, we shd remember, meticulous as regards recording the opinions of authorities who dissented from his. Various of these changes didnt finally take place until well after the end of the Victorian era.

Hot Cross Buns

When recently Good Friday came around it set me thinking again about the stressing of hot cross bun. I had for long wondered why people stress it /ˈhɒt krɒs `bᴧn/ and not /ˈhɒt `krɒs bᴧn/. My mistake was that I’d always assumed that the compound cross(-)bun wd be stressed /`krɒs bᴧn/. This was not an unreasonable assumption but incautious because I failed to realise that the simple term cross(-)bun has in fact since Victorian times become obsolete so that I was jumping to a conclusion rather than relying on observation. One can never be certain how people treat such compounds except by observing them: the classic examples of this unpredictability are the numerous ones with Chrismas — like `Chrismas cake but Chrismas `pudding. People still buy hot cross buns but I’d never he’rd anyone talk of a cross(-)bun. You won’t find that term in the current editions of any of the three big pronunciation dictionaries. The old EPD used to have it but Daniel Jones was a true Victorian in his upbringing, all of nineteen when the old Queen died. He gave it only as /krɒs `bᴧn/ (well actually “krɔːs ˈbᴧn” in EPD1 in 1917). I vividly remember him telling me that he cd never bring himself to give up the old-fashioned /krɔːs/ in favour of the new-fangled (not his word!) /krɒs/. As for myself, I’ve never thaut of the word as anything but /krɒs/. However, I’ve just discovered that when I say hot cross bun I tend to make it /krɔːs/ in that context. I’ve wondered what can be the reason why that shd be so and I think I may have a clue.

I shd mention first th’t the difference between the two versions in my speech is not as great as it is for speakers of unadulterated GB (General British or “RP” if you like) because, unlike them, I have two different but similar long-vowel phonemes at this location. One of them is closer than Cardinal Vowel 6. That I use in eg port and boarder. The other is opener than CV6 and I use it in eg sort and border: it’s about the same quality as my (short) lot vowel. I’m very old-fashioned in this respect because this distinction seems now to’ve completely disappeared from the speech of my native Cardiff where I acquired it. It’s still maintained by quite a lot of speakers in the USA in some words and persisted among a minority of speakers in GB into the early years of the last century. Daniel Jones didnt distinguish pairs like border and boarder in his own pronunciation but recorded the variant /ˈboədə/ for the latter. OED2 in 1989 still gave more as “(mɔə(r))” but online OED3 has come up to date with “Brit. /mɔː/, U.S. /mɔ(ə)r /” in which the US “cover” representation conveys simultaneously diphthongal and monophthongal variants. My strest vowel of boarder isnt a diphthong such as EPD1 recorded but a simple vowel.

Now back to how I think I came to be at home with that long vowel in hot cross bun. One of my few memories of my early childhhood was of hearing from a bedroom window on a Good Friday morning, probably about 1932 when I was six, the voice of a vendor down in the street calling out repeatedly “One a penny, two a penny, hot cross buns”. This was done with a very memorable melody and rhythm in which the first two phrases were rather fast at the same highish pitch level. Then followed a slight pause and the last three words drawn out quite slow starting higher than the initial pitch and descending two further steps. I was very fond of imitating that chant with its resultant (physically-speaking) long vowel in cross. That’s how, I guess, I came to use, and am still capable of using, /krɔːs/ in that phrase. This chant was no dou’t a relic from the times before the first Great War when street vendors were still a common sight and some still used quite ancient traditional cries. We still had in Cardiff in the thirties a man who came round selling veg’tables with a horse and cart, a rag-and-bone man likewise equipt and an old fellow who came offering to augment our milk supply from a churn on two wheels with milk that was cheaper coz it’d had its cream skimmed off. The motive for the skimming in those days wasnt callisthenic but economic. By the way, people still buy what they ask for as “hot cross buns” but when they get them they dont expect them to’ve been warmed up in the way I imagine my long-gone vendors supplied them. They take them home and toast them themselves, as the great OED actually notes.

Open-Central Vowel Symbol?

The December 2008 issue of JIPA (the Journal of the International Phonetic Association Volume 38 Number 3) contained an eight-page article by William J. Barry and Jürgen Trouvain of Saarbrücken, illustrated with nine vowel diagrams and a pie chart, entitled ‘Do we need a symbol for the central open vowel?’. It argued that “there is a logical and practical gap in the present IPA vowel chart. The lack of a central open vowel is unsatisfactory, in particular because more languages have a single open vowel with an apparently more central than fronted or backed quality.” It was remarked that the issue had been “officially discussed” at the big IPA “Convention” held twenty years ago at Kiel in August 1989 but that the decisions of the (self-selected) group which considered the matter were “No means of symbolizing a central fully-open unrounded vowel with a special symbol should be provided. Specifically, small capital A [A] should not be recognized for this purpose.” This was said to have been decided mainly on the grounds of maintaining simplicity.

My feeling is that, if simplicity is such a powerful consideration and leaving an obvious gap in the symmetry of the chart is of little consequence, then there are other cases where simplicity could be achieved by removing pretty useless items currently admitted — not that I’m necessarily advocating doing so. Certain people have felt they wanted such a symbol and  the most popular choice for it seems to be small capital [A]. This offers no problems of incompatibility with IPA vowel symbolisations in general so there is no pressing reason not to accept it.

The  1989 Convention was both stimulating and frustrating. A week or more rather than a long weekend wdve been a lot more satisfying tho no doubt hardly feasible. The choices forced upon you between attending one or other of two or more equally interesting discussions were often rather dismaying. Anyhow I have a dark suspicion that rejecting showing any central open vowel symbols was motivated partly for some people to some extent, not unreasonably, by the feeling that it would cause the bottom of the quadrilateral to be intolerably crowded. The IPA Council in 1993 agreed to the Catford 1990 (p. 27 of Volume 20 Number 2) proposals which resulted in the 1993 revision which moved some symbols and added two (close mid and open mid) central ones so that the balance and symmetry of the vowel figure looked much more satisfying — if more complicated. It did of course still leave the anomaly of no central (fully) open vowel symbols at all. If two were inserted (as logic would suggest and would be no more unreasonable than bothering to have a symbol as useless as ɶ) then overcrowding would be a problem. 

Extensive tho this article was it didnt contain as much as one might have liked about which people have advocated the adoption of the kind of symbol under discussion and for what reasons. Among the remarks of this kind we do find the claim that [A] “does have a substantial following in the community”. In the two decades since the Convention’s decision there doesnt exactly seem to have been a clamour for it. The only mention of it in JIPA I can recall was a couple of lines within a note by Geoffrey Hunt, a member of the British Summer Institute of Linguistics “School”, in Volume 22 of 1992, entitled ‘Central Vowels in the 1989 IPA’, which announced that his School had decided to “use an alpha ([α]) as an Open central vowel”. Before that in 1975 in JIPA Volume 5 Number 2 John Wells, in an editorial note at p. 53 asking for the votes of the IPA Council of the day on the possible recognition of [A], while acknowledging his own past use of it for comparative transcriptions, commented “The IPA has always been reluctant to adopt special letters not needed in the broad transcription of any language, preferring diacritics for such non-distinctive shades of sound”. That Council rejected the proposal (8 in favour, 10 against).

I find myself puzzled when the authors say (p. 352) that they are worried about the fact that some Spanish learners of English “produce the English cat vowel with the vowel quality of Spanish gata, a condition that enforces the foreign accent instead of making the learners’ English and native vowel qualities distinct”. I find it difficult to relate this comment to the realities of most language teaching. Their comment seems to refer to a very rarefied variety thereof. Of course any literate Spanish speaker will tend to equate the vowel of British English cat with his strest vowel of Spanish gata but I see no great problem in that. So long as that speaker avoids confusing English words like cat and cut he will no doubt be using English in a fashion very effective and completely acceptable to me and, I’m sure, to the vast majority of native English speakers. I don’t find it unsuitable that a Spanish native speaker shd speak English with something of a Spanish accent: indeed I suppose I prefer it that way.

The Elephant in the Room

The Wells phonetic blog of the 8th of April was headed by an unattributed (tho perfectly easily recognisable) quotation that was an item from the advance online-only work-in-progress version of the third edition of the great Oxford English Dictionary as follows:
obstruent, n. and adj.
(Pronunciation) ...
Brit. /ˈɒbstrʊənt/, U. S. /ˈɑbstrəw(ə)nt/

This was followed by a reply to a question concerning an individual linguistic scientist’s unusual habit of pronouncing the word obstruent as “/ əbˈstr.../”. Unlike Wells’s questioner and Wells himself I was not particularly “bemused” that this shd be so. I find it a quite familiar phenomenon that many speakers from time to time happen to acquire the habit of using an idiosyncratic pronunciation of some relatively bookish word. They initially guess the pronunciation of the word and fail to notice subsequently that perhaps no-one else is treating it their way. Our language has very large numbers of words that have two or more recognised alternative spoken versions. This I suspect encourages people to, at least vaguely, feel that their individual usage is likely to be among the accepted versions of the word. This phenomenon was the topic of my Blog 49 of the 25th of November 2007 in which I gave a number of examples especially of the idiosyncratic pronunciations of some prime ministers. A couple more such items are that I myself habitually use the stressing `apparatus and that Lord Bragg regularly sez science exactly as one sez scions. By the way, the form /əb`struːənt/ may hardly exist but it embodies a perfectly natural type of English word stressing.

However, what for me was the elephant in the room was the fact that, tho it was headed with those two transcriptions of the word obstruent which look strikingly different, no-one commented on them. I wonder how many users of such reference works realise that the pairs of transcriptions they are seeing side by side do not equate at all with genuine significant differences between British and American pronunciational usages but are primarily a matter of transcriptional preferences of the different contributing editors. There are quite understandable reasons for these two styles of transcription in this case. In the new 1961 edition of Webster’s Third New International Dictionary the leading American pronunciation lexicographer of his day, Edward Artin, included an innovatory style of transcription for words like obstruent and strenuous etc. He explained in that work (at p. 37a) that he favoured transcriptions like \ˈstrenyəwəs\ because he felt ...“a very strong conviction that ... such pairs as” silhouette/Scylla wet were “exact rhymes” and that this clearly indicated in consequence that his new transcriptions of such items were more satisfactory representations of the normal pronunciations of such words.

In writing silhouette in a phonemic transcription one’s choice is essentially between /sɪl.u.`ɛt/ and /sɪl.ə.`wɛt/. Most lexicographers have preferred to use the former type no dou't because it correlates more satisfyingly with the orthography. Yet it can hardly be denied that it tends to suggest a slightly slower and/or more deliberate enunciation than the word ordinarily receives and that the latter version, despite its containing a suggestion of a schwa which can’t be straightforwardly correlated with the word’s orthographical form, satisfies the criterion of sounding not unduly deliberate and producing an effect that probably anyone would regard as normal. Of course it wd be possible ideally for lexicographers to show both versions of such words eg strenuous as /`strenjuəs/ and /`strenjəwəs/ but it’s simpler and space-saving to choose one alternative with the understanding that the other is not being simply rejected. Another possibility wd be to employ a cover symbol such as ʉ but that wd undesirably complicate matters.

There happens to be one pair of spellings in ordinary orthography that illustrates this alternation viz Aloysius and Allowishus the second of which more explicitly represents the usual version of both. LPD3 interestingly offers different transcriptions for Aloysius, showing the GA version Artin-style. Artin felt that certain similar words including some suffixed items formed from words ending with o were better represented with his ō (ODP’s oʊ) hence he gave no schwa to Croatian, ghettoize, jingoism and Genoese (ie we don't find him offering /*krə`weɪʃn, * `getəwaɪz, * `ʤɪŋgəwɪzm, *ʤenə`wiːz)  tho he did give Fa(e)roese as \ˌferəˈwēz\ which apparent inconsistency ODP has followed. ODP exceptionally gives both versions at egoism, the schwa type first, again following Artin’s lead.

British lexicographers have not embraced this innovation. Two British pronunciation dictionaries have used American editors. Hartman in the EPD has stayed within the British tradition but Kretzschmar in both the ODP and the currently ongoing OED3 is of the Artin persuasion. I think Wells and Roach chose the right alternative for their EFL oriented works. If it didnt explicitly offer itself to “learners of the language” (p.viii) ODP might have a better case for the schwa versions which wd of course be no less suitable in the British than the American entries in which they do appear. Examples readers may like to see and listen to from the Online Webster are follower \ˈfä-lə-wər \ with its appropriately matching schwa-including pronunciation sounding perfectly natural (in terms of that feature) if it were a British or American usage. Comparing it with valuer we find that too displayed with schwa \-yə-wər \ but by contrast the speaker uses the slightly more deliberate pronunciation [`vӕljuər]. Finally, it has to be admitted that this is one of those cases where the differences in transcription tend to suggest something more than the really quite small audible differences of sound. An amusing illustration of this is to be heard at the MWO recordings of the two transcriptions of zoology. They are audibly by the same speaker and don’t sound in the least different to me!

Spelling Reform Experiments

  Readers of these blogs will know that Ive for some time taken the opportunity of using, or at any rate trying out, various spellings I think miet be improovments on the traditional ones. I made some tentative suggestions in my blog 47, entitled “Spelling Reform”, on the 28th of August 2007 which was a response to a blog that had been posted by John Wells on that day. My alternative spellings have partly been a way of declaring my belief that we shall never manage to produce any reforms of our orthography if we dont learn to be pretty tolerant towards untraditional spellings. It seems very desirable to get people to feel free from the feeling that any braek with tradition can only be due to orthographical incompetence or ignorance. I’m afraid it’s quite hard for us overly literate folk to rid ourselves of snobbishness with regard to “misspellings”!

I’ve made it very clear that I doent wish to make any attempt to be consistent if only becoz I can only ascertain by experimentation what my reactions are to seeing spellings used that I miet think are possibly suitable revisions. Some of my experimental spellings Ive been happy with from the start but others have cauzed me to feel uneasy. Ive never had any hesitation in rejecting read as a past-tense form of the verb to read. In fact I dont think Ive ever in these blogs used anything other than re’d for it, tho John in the blog I mentioned, with apparent incredulity, said of me “He claimed that he himself always writes re’d for the past tense of the verb to read”. In fact I don’t think that was absolutely exactly what I sed (or ment to say) at that discussion at University College London coz, much as I shd like to do so, editors of things I publish are sure to veto such a move in most circumstances. My experiments hav revealed to me that I hav a sort-of instinctive dislike of loozing distinctive spellings or using ones with powerful suggestions of inappropriate meanings. I doent actually like reed for read but I sometimes find my dislike of it overpowered by my strong wish to avoid the ambiguity of “read”.

Ive been finding myself very reluctant to give up phonetically “unjustified” spellings in favour of more regular or logical ones in a number of cases. This has made me glad that the (formerly “Simplified”) Spelling Society that we have in this country has recommended for its “house style” that people who adopt reformed spellings shd feel free to use any traditional spellings they dont wish to abandon. I find myself often uneasy at contemplating the loss of welcome visual discriminations in various cases. I experience uncomfortably reluctant feelings about using perfectly rational spellings like rime, sum, stile, wen etc insted of rhyme, some, style, when etc. I tend to want to keep write, rite, right and there, their, they’re etc. I jib at making Johns all Jons or writing son as sun or to, too, & two all as too (or tuu or whatever).

While using my experimental revised forms I find myself tending to employ lots of otherwise-unnecess’ry apostrophes to reassure the reader that what Ive written is not a typo. I also have a reaction agenst things that look what I find to be upsettingly unlike the traditional form of a word eg wun for one, hoom for whom etc. I’m not really happy with eg the Nue Spelling whaer for where. I have even felt reluctant about trying to bring myself to use the more suitable -ize spelling insted of -ise. I’ve long preferred the latter — not very logically perhaps — as less fussy looking and, to be candid, a guard against being caut out using zed when it wasnt appropriate etymologically.

The first volume of the OED, gave axe preferentially spelt ax (as it is very largely in the US) but I was pleased when OED2 gave ax only as an alternative spelling and dropt Murray’s defensive remark “The spelling ax is better on every ground of etymology, phonology and analogy, than axe, which has of late become prevalent.” I think people probbly felt ax looked odd coz having less than three letters in a word which is not a particle is extremely unusual.

I was very interested to read the John Wells blog of the 9th of September 2008 on the occasion of the centenary of the Spelling Society — of which he is the very successful President. In it he took up a line I seemed not quite to have noticed from him before in calling for “some freeing up of the rigidity of English spelling”. I was pleased to see him saying “Lets abolish the apostrophe, or at least make it optional”. Readers of these blogs may have noticed that I feel free to omit traditional apostrophes if I wish just as I may do with superfluous final e’s of have etc. Another of his proposals then was in fact “Let’s allow people to omit the misleading final e of have and give”. He also referred, as I had done in my Blog 47, to looking tolerantly on some of the texters abbreviations.

I’m afraid I still despair of intractable problems like the lack of a schwa and of the difficulty of finding an acceptable way of differentiating /ʊ/ and /u/ — something which I can’t bring myself to feel not worth maintaining. Contemplation of the fate of the Pitman Initial Teaching Alphabet isnt encurraging. All in all you see I’m not a whole-hog reformer and that’s as much as ennything becoz I cant see full reform ever really happening, at least not in the lifetime of most people living today. Probbly the most effective step forward wd be for national authorities and examiners to be persuaded to look kindly on rational spellings like menny in junior schools. But dont hold your breth waiting for that.

Handwriting, Spellings and Sounds

praye tayke a looke atte þe uorde muumuu ande aʃke yurʃelfe jusʃte houe eaʃie it iss toe reade. notte uerie, eh? noue conʃidder howe ontill þe fyfteenthe centuarie aule bookɛs hadde too bee riten bie haunde ande avlʃoe þatte inglysshe spelinge wasse highlie uaeryabvll ande thatte thair was eeune abbsoelutlye noe reggvlarytie inne þe uurytynge off kapytvlle lettres.
 
The above bit of cod medieval writing is provided by way of introducing the topic of how present-day spelling still suffers from the effects of medieval scribes who adopted spellings more suitable for their legibility than their faithfulness to pronunciation. There were words in medieval English that might be pronounced with sounds either like [ɔ] or like [u]. No doubt many scribes preferred in various words to write the more legible version with o rather than u even if the latter letter better represented their own spoken usage. The word muumuu, had it existed then, wdve been much easier to recognise if it had been written moomoo rather than as the dozen successive linked vertical strokes of muumuu. (It’s the name of a garment the OED defines as “A woman's loose-fitting dress, usually patterned and brightly coloured [worn] ... in Hawaii”).

Thus there was a set of words in which the letter u was avoided by the scribes in favour of o chiefly in the context of other letters (originally) with vertical strokes ie  m, n, u, v and w. Hence the potentially misleading spellings today of many words including above, accompany, among, blood, borough, bosom, brother, colander, colour, come, comfort, company, compass, conjuror, country, courage,  cousin, cover, covet, done, double, dove, dozen, flood, flourish, front, glove, govern, honey, London, love, Monday, money, monk, monkey, mother, move, none, nothing, nourish, once, one, onion, other, oven, plover, prove, shove, shovel, slovenly, smother, some, son, southerly, southern, sponge, stomach, thorough, tomb, ton, tongue, Somerset, trouble, wolf, woman, won, wonder, worry, worse, worth etc and their derivatives. EFL users be on your guard!

Some words which historically used to have one of the values of the letter u have been changed over the years to new pronunciations that match the previously inappropriate spelling. These include bombard, combat, conduit, foreign, novel, solemn, sovereign etc. Others have old and recast versions alongside each other eg accomplish, comrade, constable, covert, dromedary, frontier. Some of the alternatives are mainly used on one side or the other of the Atlantic eg it’s usually £ /ᴧ/ but $ /ɑ/ for accomplish, conjuror, constable and dromedary but $ /ᴧ/ and £ /ɒ/ for grovel, hovel and hover. For mongrel it’s $ only /ɑ/ and £ only /ᴧ/ and with plover it’s $ only /oʊ/ and £ only /ᴧ/. For the rather literary word sojourn it seems to be $ only /oʊ/ and £ only /ᴧ/ or /ɒ/.

Northern England educated usage often favours /ɒ/ in one, none and nothing. Occasionally one comes across a speaker of Northern extraction who tends to adopt Southern usages while living in the South but misapplies the “correction”  by saying not only none as /nᴧn/ but by treating the prefix non- in the same way and saying eg /ˈnᴧn `sməʊkɪŋ/. Some educated Scottish speakers have the historical /ᴧ/ in novel.

[For anyone who found any of my jokey pseudo-medieval spellings a problem, they sed “Pray take a look at the word muumuu and ask yourself just how easy it is to read. Not very, eh? Now consider how until the fifteenth century all books had to be written by hand and also that English spelling was highly variable and that there was even no regularity whatsoever in the writing of capital letters.”]

Analytical issues

 On Monday the 9th of March I happened to be abraud on holiday when John Wells sed, in a blog entitled “Syllabic Plosives”,  In principle, I would say that plosives can never be syllabic.  I should say they can /ˈaɪ ʃd̩ seɪ ðeɪ `kӕn/ and that syllabic /d̩/ in particular is not really unusual in English. Other examples are evidence as /`evd̩n̩s/, hatred /`heɪtrd̩/, where had she put it /weər d̩ ʃi `pʊt ɪt/, delivered by hand /dɪlɪvd̩ baɪ `hӕnd/, who’s had they used /huːz d̩ ðeɪ `juːzd/, it’s feasible that they... /ɪts fiːzb̩l̩ ðt̩ ðeɪ.../. I shd say that my last example was quite a rarity but perhaps just about possible. The ones where the sequence /d̩n̩/ occurs are not uncommon. Indeed I believe that in normally fluent non-deliberate speech the expression hundred and  is probably most often uttered as /hᴧndrd̩n̩/ by the majority of  speakers. I’ve found practically nothing in the literature to support my observations unless I mention that William Sidney Allen, no mean phonetician, writing in Le Maître Phonétique in 1956 at page 15, transcribed the phrase reader should be with /ʃd/. My confidence in these attributions is supported by my proprioceptive observation that in producing them my tongue perceptibly  maintains alveolar contact in sequences like /-drd̩n̩.../. 

      John began that blog by declaring that he favoured transcribing have a potatoe as hæv əpˈteɪtəʊ. This I found puzzling for two reasons. One was not being sure whether the lack of space between the /ə/ and the /p/ was intentional. The other was could he mean that he was representing the /p/ as unreleased which is one of the three possibilities for such a sequence. The other two such possibilities are (i) [pə-] with a normal schwa and (ii) with a voiceless schwa which cn be represented as [pə̥t-] or as [pʰt-]. Discussion of this choice happens to have just come to my notice in looking at the new edition of the Roach textbook the subject of my Blog 169. There at p. 114 he quotes potato, tomato, canary, perhaps and today as examples of words in which “the vowel of the first syllable may disappear; aspiration of the initial plosive takes up the whole of the middle portion of the syllable...” and transcribes them [pʰˈteɪtəʊ] etc. I can’t help feeling a preference for categorising such items as containing voiceless schwa and also wondering what he takes to be the identity of the part of the syllable after its “middle portion”.

BlogRoll

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