Archive 34 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

A Festschrift for John Wells

This month has seen the delayed publication by the English Phonetic Society of Japan of a 235-page volume combining the 14th and 15th issues of that Society's periodical 'English Phonetics' in the form of a "Festschrift Commemorating the Retirement and the 70th Birthday of Emeritus Professor John C. Wells" edited by Professor Masaki Taniguchi of Kochi University. After a brief account of Wells's main publications it contains six "Congratulatory Messages" and twenty-one articles of about ten pages each three of which are in Japanese.

A particularly enjoyable item, by the current President of the English Phonetic Society of Japan Masaki Tsudzuki on 'Some Phonetic Rules Regarding the Adaptation of Foreign Loan Words into Japanese', includes the comment on certain Japanese compound words containing the word 'American' that the word's first vowel was 'not heard by Japanese and was thus left out'. Such an 'aphesis' /`ӕfəsɪs/ is not unknown among native English speakers — tho it's probably mainly confined to demotic and childish use. Even so, I don't expect to see any eyebrows raised when I say in casual conversation eg “/`mɛrkənz dəʊn seɪ ˏðӕt/”. We learn from this article also that both ear and year may be heard from Japanese speakers as /`iːɑː/ reminding one of the losing of initial yod from /jiə/ by GB speakers who so offen say things like /nekst `iə/ for next year. Another rather amusing item was 'dommai' an expression of sympathy (especially directed at someone bungling an action at baseball) derived from 'don’t mind' that, in undergoing Japanese transformation comes; out like the reduced articulation of a sleepy or drunk English native speaker /dəʊm maɪ/.

Several of the articles are on matters particularly of British pronunciation. The first of these is a discussion by John Maidment on 'Transcription and the changing vowel system of English' in which he sez that "It may well be time ... in our transcription of GBE [General British English] to abandon /ӕ/ in favour of /a/". ; This change has, of course, already been implemented by Clive Upton in OED3 and its sister publication ODP. So has the suggestion to 'update' from /eə/ to /ɛː/ on which Maidment warns, "To retain the standard symbol for very much longer runs the risk of causing confusion, both for native speakers learning phonetics and, especially, for non-native speakers who consult pronouncing dictionaries or who make use of published accounts of the features of GBE". The leading EFL textbook Practical Phonetics and Phonology (2008) by Collins and Mees has used /ɛː/ since 2003. Stimulating, if rather alarming, are his speculations on whether we shd soon be changing /ɪə/ to /ɪː/ and back vowels to perhaps more strictly IPA-conforming notations such as /ʉ, ᵿ, əᵿ, aᵿ/ or even /ɨː, ᵻ , əᵻ, aᵻ /. That last one has been he·rd occasionally from the Queen and Prince Charles but he was for a while so frequently caricatured as saying 'ite and abite' for 'out and about' etc that he has since evidently avoided the usage.

L-vocalisation in British speech is the subject of two contrasting items. One, by Mitsuhiro Nakamura, concentrates on a 'parametric' approach focusing 'on the continuous activities of the organs of speech'. The other, 'The l-vocalization trend in young London English speech — growing or declining?' by Patricia Ashby, reports an investigation that notably challenges certain commonly accepted assumptions in relation to this development.

The sixteen-page article 'The stops that aren’t' by Michael Ashby and Joanna Przedlacka, whose title consciously echoes that of a 1980 JIPA paper by the volume's dedicatee, reports on the many lenitions that can occur to British stop consonants which their acoustic analyses have shown to "commonly deviate from the norms assumed for them". Their searching approach has produced evidence against dismissing occurrences of lenited stops as merely instances of casual articulation. I once received a slight shock on playing back a high-quality studio recording of my very careful saying of the words "People Speaking" introducing the recordings accompanying my book of that name. It was perfectly clear that what I'd sed was not [spikɪŋ] but [spixɪŋ]! Not that it sounded at all unusual.

'Acoustic features of the rise-fall in English' by Nobuo Yuzawa is another item on British English. The rest of the papers are, in various ways, mainly about English in relation to Japanese learners. However, among unusual topics are Hiroshi Miura's 'Frisian monophthongs and English vowel change' and a multiply-authored study of 'Finnish learners’ auditory perceptions of English pronunciation by Japanese speakers'. The only item which has no direct connection with English phonetics is the nevertheless rather int·resting description of how South Korean phoneticians have devised an enthusiastic·ly received course to help North Korean 'defectors' to the South to acquire a less unpopular accent for their new linguistic environment. Other prosodic matters also receive notable attention. Unfortunately, on this occasion it's not possible to deal with more than these brief samplings from this very worthwhile volume.

Ticklish Words (vi) SCHEDULE

When the OED's second Editor Henry Bradley came to sch- words in 1910 he listed schedule as pronounced /ˈʃɛdjuːl/ and also /ˈʃɛdəl/ and, in a very unusual nod in the direction of American usages, added "U.S. /ˈskɛdjuːl/". As to /ˈʃɛdəl/, it must've now been obsolete for well over half a century. He gave a long historical note explaining that, tho the earliest spellings, dating from the Middle English period, were things like cedule and sedule, from the middle of the seventeenth century the preference had become for a form that reflected the word's ultimate Graeco-Roman origin viz schedule.

There are some parallels between the development of schedule and those of the originally mainly ecclesiastical term schism. This was traditionally only pronounced /sɪzm/ but has since been increasingly converted to the more spelling-harmonious /skɪzm/ by recent generations as Robert Burchfield remarked, in his supplementary OED volume of 1982, in one of his rare additions only on pronunciation matters, "The pronunc. (skɪz(ə)m), though widely regarded as incorrect, is now freq. used for this word and its derivatives both in the U.K. and in North America".

Another comparable item is sceptic usually so spelt in the UK by contrast with skeleton which has a similar Greek origin. On the spelling skeptic Bradley commented that it was "adopted without comment or alternative in Johnson's Dictionary, but did not become general in England; in the U.S. it is the ordinary form". Bradley remarked of schedule that "The original pronunciation /ˈsɛdjuːl/ continued in use long after the change in spelling; it is given in 1791 by Walker without alternative; in his second ed. (1797) he says that it is ‘too firmly fixed by custom to be altered’, though on theoretical grounds he would prefer either /ˈskɛdjuːl/, ... or — ‘if we follow the French’— /ˈʃɛdjuːl/... Smart, however, in 1836 gives /ˈʃɛdjuːl/ in the body of his Dictionary without alternative, although in his introduction he says that as the word is of Greek origin the normal pronunciation would be with /sk/... In the U.S., the authority of Webster has secured general currency for /sk/".

There's no dout that /sk-/ in schedule has gained considerable ground over here in recent decades. Wells has chosen to question his respondents on the matter. The LPD3 (2008) entry sez:  ˈʃed juːl ˈʃedʒ uːl  ˈsked juːl
"The AmE pronunciation with sk- is increasingly heard in BrE. Preference poll, BrE: ʃ- 70%, sk- 30% (born since 1973, 65%); -dj- 79%, -dʒ- 21%."
EPD (2006) lists four British basic types, all bisyllabic,
 ˈʃed.juːl, ˈʃedʒ.uːl, ˈsked.juːl, and ˈskedʒ.uːl
and two American the latter of which is trisyllabic,
  ˈskedʒ.uːl and ˈskedʒ.u.əl.
ODP (2001) gives six basic British types:
ˈʃɛdjʊ̶̵l, ˈʃɛdjuːl, ˈʃɛdʒʊ̶̵l; ˈskɛdjʊ̶̵l, ˈskɛdjuːl, ˈskɛdʒʊ̶̵l;  
and one American:ˈskɛdʒəl.  
The new OED3 symbol ʊ̶̵ "represents free variation between /ʊ/ and /ə/".

The British Library summarised their results like this:
The OED distinguishes between 'shed' as a British English pronunciation and 'sked' as American English. Not surprisingly, then, all the North American voices use 'sked'. However, 25 out of 60 British and Irish speakers agree, while 35 out of 60 prefer 'shed'. We might, therefore, interpret this as evidence of recent influence from US English, but there could be other factors, e.g. the subconscious spelling association with similar words like scheme, school etc. which are clearly 'sk' for all speakers. It’s certainly plausible to imagine that schedule is first encountered in its written form rather than as a spoken form (I don’t imagine it’s a very high frequency word for young children), but perhaps there is indeed American influence at play, too.

Ticklish Words (v) SCONE

Some readers may remember that our Blog 333 was originally entitled 'Six Ticklish Words' and listed six words that the BL (British Library) 'Evolving English' scheme invited people to record their versions of (particularly if they werent inclined to record the hundreds of words of the 'Mr Tickle' story they'd've liked to've had). Their six words were "chosen for the way their pronunciation varies both regionally and internationally". In other words because their pronunciations were 'ticklish' ie 'unstable, unsteady..uncertain..' etc to quote from one of the OED's definitions of that expression. We've now come to the last of the original six tho we hope to discuss also the further two that have featured in the BL revised list of six.

The online OED3 at 'word, n.' gives its pronunciation as "/skəʊn/ /skɒn/ " and its recorded spellings as "15-18 scon, skon, (17 sconn)". It suggests that it was "Perhaps a shortened adoption of Middle Dutch schoonbrot .. ‘fine bread’ ". It adds that it's of Scottish origin and defines it as "1. A large round cake made of wheat or barley-meal baked on a griddle;...[but] more generally, a soft cake of barley- or oatmeal, or wheat-flour, baked in single portions on a griddle or in an oven".

The earliest illustration of its use is dated 1513. Before the twentieth century the quotations seem to be practic·ly all from Scottish sources and all spelt 'scone'. No early evidence of pronunciation seems available: the word wasnt listed in Walker's 1791 Critical Pronouncing Dictionary. In the Chamber's Twentieth Century Dictionary, published in Edinburgh and edited by a native of that city, A. M. MacDonald, her entry for it re·d  "skon in the South of England often pronounced skōn, (Scot.) n. a flattish, usually round or quadrant-shaped plain cake of dough without much butter, with or without currants, baked on a girdle or in an oven."

When the OED1 editor Henry Bradley came to the word in 1910 he decided, on what evidence if any it's hard to know (in respect of a word then praps little known in England), to show two pronunciations for it in the order that remains so far unrevised in OED3. It shd be realised that a dictionary as enormous as the OED can hardly be expected to immediately reflect all the offen quite rapid changes in pronunciational fashions. All the indications of its pronunciation in the Concise Oxford Dictionary from 1911 were ambiguous as to ordering of the alternatives until the last decade of the twentieth century when the non-committal diacritic over the 'o' (a macron-under-a-breve combination) gave way to IPA transcriptions with /skɒn/ placed first. H.W. Fowler in his Modern English Usage in 1926 referred to the OED ordering adding "but the sound skŏn is perhaps oftener heard".

Daniel Jones always, from Jones-Michaelis 1913 and EPD 1917, put /skɒn/ first. So did H. C. Wyld in his remarkable large single-volume Universal English Dictionary in 1932. And so did the New English Dictionary by Ernest Baker in the same year. C. T. Onions made no change from OED in editing the Shorter OED in 1933 but by 1966, when his Oxford Dictionary of English Etymology came out, he'd reversed Murray's ordering. A. S. Hornby et al's Advanced Learner's Dictionary had done so from its first edition in 1948.

Coming to the present we find that the LPD British poll respondents exprest a 65/35% preference for /ɒ/. No up-to-date British source challenges this but beyond the UK things are offen diff·rent. The LPD American entry is "skoʊn skɑːn". So is EPD's but ODP gives /skoʊn/ alone. Merriam-Webster Online gives /skoʊn, skɑːn/ in that order, the speaker only illustrating the second version. Back in 1961 Artin in the Webster 3rd International listed another variant /skoʊn, skɑːn, skᴧn/.

The report of BL's own survey begins "The OED cites a pronunciation rhyming with ‘bone’ ahead of a pronunciation rhyming with ‘gone’" when it wdve been better to quote a more up-to-date authority like LPD rather than a century-old outdated opinion. It continues "All the US voices in our survey have the former, but intriguingly 3 out of 10 Canadians use the latter. In Britain and Ireland, where pronunciation of this word is an extremely popular dinner table debate, 41 out of 60 speakers rhyme scone with ‘gone’, while 19 out of 60 rhyme it with ‘bone’. Clearly variation exists here in British, Irish and Canadian English, but apparently not in the USA."

When I was growing up in Cardiff, before and during World War 2, we certainly had the objects but we didnt call them 'scones' with either pronunciation but 'bakestones' because they were usually cooked on a 'bakestone' which was our word for a griddle. Other people in South Wales call them 'Welshcakes'. I think these are items of supporting evidence for the suggestion that, until well into the twentieth century, the object was mainly known by local terms but that, during that period of increasing countrywide uniformity in such matters, the Scottish choice won out with the general public.

Praps I shou·dnt omit to mention that Scone with an initial capital letter is a Scottish placename with yet another pronunciation, as /skuːn/. By the way, it's one of the only ten words I know which by changing its initial letter from capital to lower case produces not only a diff·rent meaning but a diff·rent pronunciation. The others are August, Ewe, Job, Natal, Nice, Polish, Reading, Smithy and Worsted. Any further suggestions?

PS Thanks to John Maidment for drawing attention to Slough/slough and to Richard Sabey for Begin, Concord (New Hampshire's capital), Mousehole, Rainier, Said, Tangier, Worms and (Guy) Forget (former French tennis player).

The Affirmative AYE

While reading John Wells's blog of the sev·nth of April about Scottish vowel contrasts, when I came to his comment on the diff·rences between the values in many Scottish varieties of English between such pairs of words as tide and tied, I was reminded of the fact that some varieties of English south of the Border have or have had one or two phonologically contrastive pairs of their own not completely unlike those. They have praps now disappeared from GB (the most general British accent) but evidence of at least their past existence can be furnished.

Murray's NED (aka OED1) listed at his 1884 'Key to the Pronunciation' the two diphthongs \əi & ai\ to convey what we may express, in something more like IPA transcription, as [ɑɪ & ɐɪ] with the key words aye ('yes') and Isaiah for the former and I, eye, bind for the latter. Murray used \ai\ at Isaian in 1900 but I've found no trace of any other word in OED with the notation \ai\. When Onions edited the boiled-down SOED (Shorter Oxford English Dictionary) in 1933 he quite understandably withdrew the symbol \ai\ tho he still, obviously by an oversight, kept it for the first pronunciation of aye in the body of the dictionary. He also omitted the rare adjective Isaian.

OED3 perfectly acceptably repeats Murray's definition of aye as "As an affirmative response to a question: Yes; even so. Common dialectally, and in nautical language; the formal word used in voting ‘yes’ in the House of Commons; but not used for ‘yes’ in modern educated speech or writing, except as an archaism".  Hardly as satisfactorily, it also repeats his 1885 parenthetic comment on the relative sounds of Aye and eye "(which many identify in pronunciation, and which differ at most only in the ‘broader’ or more back sound of aye..". The closing round bracket which this parenthesis required was omitted in NED by Murray and is still missing in OED3. The remark must surely mystify most present-day readers but in Walker's 1791 Critical Pronouncing Dictionary, which continued to be reprinted until the beginning of the twentieth century and was certainly known to Murray, these two words (for 'yes' and 'organ of vision' respectively) were recorded as pronounced something like [ɑɪ] and [aɪ].

I spent the first eighteen years of my life in the county town of Glamorgan (which, incident·ly, tho within the political border of Wales, was in a thoro·ly monoglot English-speaking area no more influenced by the Welsh language than Gloucester, Hereford, Chester or Liverpool). Among the various Cardiffisms that still linger in my speech are this very distinction. It also still persists widely in South Wales in the way that the name Dai, a mainly-demotic hypocoristic form of David is regularly differentiated from dye or Di (for 'Diana'). Some speakers may also differentiate the reduction of 'How-are-you' to Hi(ya) from high(er). 

Ticklish Words (iv) NEITHER

This word and either have been exceptional in that they've had remarkably diverse pronunciations from very early in their hist·ries. In relatively recent times in non-standard speech twenty or more were listed for each in Wright's English Dialect Dictionary (1896-1905). Going back to Old English the basic root of both was as for whether. Neither's earliest forms originated in the addition of a negative particle to either which in the first place produced a version nauther. This, consid·rably later, was modified to match either in its initial vowel.

It's not until the early eighteenth century that forms are recorded beginning [aɪ-]. OED notes it first from (Dr John) Jones 1701 (in his "Practical Phonography") which had "/ˈeːðər/ and /ˈaɪðər/". Next Buchanan (1766) "had /ˈaɪðə(r)/ without alternative" Walker (1791) said "that /ˈiːðə(r)/ and /ˈaɪðə(r)/ were both very common, but gave the preference to the former on the ground of analogy and the authority of Garrick". Smart (1849) said that ‘there is little in point of good usage to choose’ between the two pronunciations. In the NED (=OED1) in 1891 Henry Bradley sed "The pronunciation (ˈəɪ·ðəɹ), though not in accordance with the analogies of standard English, is in London somewhat more prevalent in educated speech than (ī·ðəɹ)", ["/ˈaɪðə(r)/ than /ˈiːðə(r)/" in OED's updating of the Murray transcription].

In General British usage today the /iː-/ form is common but the other is predominant. The EPD thruout its hist·ry has recorded this pattern. In America the reverse applies. ODP concurs. We may safely take it that reponses to the Wells poll results for either exactly chime with those for neither. The percentages LPD shows are UK 87/13 and US 84/16 clearly confirming the judgments for the two varieties. Some degree of contrary tendencies appear to be operating among younger speakers in both groups.

The British Library reports that in responses to its recent invitations "33 out of 60 British & Irish speakers have the ‘scythe’ pronunciation, while 27 out of 60 use ‘seethe’; 15 out of 60 American English speakers use ‘scythe’, while 45 out of 60 prefer ‘seethe’ - i.e. variation occurs in our data in both varieties and the relative frequency pretty much confirms the OED distinction for US English, but British English speakers are perhaps more equally divided than the OED suggests". This comment seems to ignore the date (1891) and the limitation of Bradley's remark to 'London'. Incidentally, in listing the two pronunciations he placed the /iː-/ form first. It'll be int·resting to see what OED3 does when this item comes to be updated.

Ticklish Words (iii) HARASS

The word harass is a relatively recent arrival into English. OED doesnt have any record of it before the early 1600s. It designates it as a borrowing of the verb French harasser which itself seems to've not been recorded before 1562 when it appeared in a dictionary (Godefroy) defined as ‘to tire .. to vex, disquiet, importune, harrie, hurrie, turmoile, torment’ etc. It's suggested that it might be a derivative of Old French harer to 'set a dog on'. That cd surely to be a Germanic word in French cognate with our verb harry which goes back to Old English. My Dictionnaire Étymologique (1938) by Albert Dauzat suggests that harasser was derived from the  French word hare, which is found as early as 1214, describing it rather unconvincingly as an onomatopoeic hunting term.

There seems to be no trace of harass being stressed on its latter syllable before Kenyon and Knott in 1944 remarked "The pronunciation həˈrӕs instead of the older ˈhӕrəs appears to be on the increase." By 1961 we find Edward Artin, pronunciations editor of the newly revised Webster 3rd International Dictionary, listing it as the predominant form. It's difficult to guess what might've braut this change about but I suspect that many immigrants from Scotland and Ireland may have arrived with such a habit already. It's much easier to account for the change that occurred in England. The Women's Liberation movement in America first popularised the term 'sexual harassment'. This expression became well known in the UK being offen he·rd with the stressing as it was used by Americans. So also the verb harass. OED has as its earliest record of the phrase's occurrence:
1973 Appleton (Wisconsin) Post-Crescent 13 Oct. 1 Katie Miller of the Division of Equal Opportunities said she does not know how many verbal complaints the agency receives on sexual harassment of employees.

OED3 Online, which has now completed revision of all letters from M to R, has yet to add any further pronunciation at harass. EPD never had the end-stress version until 1997 when Roach, Setter et al took over. ODP puts the forestress version first for British usage. So does LPD which from its first edition in 1990 has had a note: The traditional RP form is ˈhӕrəs. The pronunciation həˈrӕs which originated in the US, was seemingly first heard in Britain in the 1970s. In time it may predominate in Br E, as it already does in Am E. Meanwhile, it evokes negative feelings among those who use the traditional form. This last comment is now beginning to be true of fewer and fewer people.  Pie charts show that Wells's British poll respondents in general preferred forestress (68%) and that an American poll showed the reverse preference with only 13% preferring forestress. He commented in a blog at the eleventh of March this year that his figures regarding harass, showed "a change over apparent time (as detected by comparing results for different age groups) ..  [indicating that] .. older speakers preferred initial stress, younger speakers final stress. The same change can be detected in both AmE and BrE, with AmE leading the way by a generation or two."

AITCH Dropping and Restoring

20 Mar 2011 revised  26 Nov 2011

1. It’s curious that our name for the sound /h/ can be spelt in modern times in two ways. One of these is <h> and the other <aitch>. The latter went back to a presumable Latin form *hacca which really did have its [h] spoken. The two spellings are both pronounced /eɪʧ/ by practicly all educationally sophisticated people in ev·ry English-speaking country except only Ireland where anyone may say it as /heɪʧ/ tho there have been suggestions that a particular religious category favoured one rather than the other. There are in modern times hour, heir, honour, and honest (and their derivatives) in which, altho an aitch always features in the normal spelling, Hardly anyone of considerable education utters an initial /h/. Observation of this taboo seems to be weakening in the 21st century. The Wells LPD3 has the note. “The form heɪtʃ is standard in Irish English, but traditionally not in BrE or AmE. It is, however, spreading in Br ”, a comment I can confirm.


2. In the earlier nineteenth century the aitchless list was a little longer containing herb, homage, humble and humour (which last, labelled ‘old-fashioned’, remained in EPD until its dropping in 1997) all of which still have some sort of currency in America where they’ve added huge and Hugh to that list. Aitchless humble seems to be pretty restricted in the US to some Southerners. A conscious, usually jocular, archaistic (presumably sometimes aitchless) version of host occurs solely in the expression ‘mine host’. OED quotes its use by P. G. Wodehouse. It’s most likely nowadays to be given the /h/ despite the ‘mine’. In the twentieth century, forehead, which Murray in the OED in 1897 had recorded only as /`fɒrɪd/, was given a second form /`fɔːhed/ in 1989 four decades after Jones had first listed it as a ‘rare’ form with an aitch in 1947. EPD 2006 and LPD 2008 agree that /`fɔːhed/ is now predominant in GB. The unreprinted ODP at 2001 gave the aitchless alternant first. The main authorities OED3, LPD, ODP (EPD doesnt really have the item) give the slangily abbreviated ‘hon’ for ‘Honourable’ as used in titles as only /ɒn/. But the fact is that Nancy Mitford, in her Pursuit of Love ii. 14, wrote ‘I was a Hon, since my father, like theirs, was a lord’ with the n-less indefinite article. Reading aloud from one of her sister Nancy’s books Deborah Duchess of Devonshire clearly used the aitch in the word ‘hons’ in a program on British television transmitted in 2006.

3. An unusual historical development has left us with two forms, aitch-bearing and aitchless, of a few words. These are chiefly the names Adrian and Hadrian, Anna and Hannah, Ellen and Helen, Éloïse and Héloïse, Esther and Hester with which we may compare Alleluia and Hallelujah, yellow-ammer and yellow-hammer and the now archaic ostler and hostler. This last item was listed by Jones and Gimson in the EPD as only ever uttered aitchless. The current EPD understandably lists an alternant aitch form.



4. A minority of us in England still say at least hotel and upholstery with no aitch; very many treat abhor and adhere likewise. Exhale, exhume and dehydration seem to often have aitches. Other exh- words like exhaust, exhibit, exhilarate are normally aitchless as are also usually annihilate, inhibition, philharmonic, prehistoric and rehabilitate. GB speakers very rarely have an aitch in vehicle or vehement but, if we are to believe ODP, a form with aitch predominates in GA for the former and is common for the latter (tho Webster online lists only aitchless forms for both of them). GA speakers are by contrast distinctly more given than GB speakers to the use of the aitchless forms /æv, æz, æd/ of the verb have (whether or not auxiliary) when they’re most weakly strest. No-one pronounces the aitch in shepherd.

5. Most British placenames ending with unstressed -ham etc usually have them without /h/ from all but the fussiest GB speakers. Durham is always /`dᴧrəm/. Northampton is usually /nɔː`θamptən/. Stanhope is /`stanəp/, but Amherst and Lockhart vary and Lockheed and Rotherhithe (unlike Rotherham /`rɒðərəm/) normally have /h/. Birmingham and similar items usually end with /-hæm/ for GA speakers. Most words like petersham, Gresham and Lewisham have succumbed to having their s-h re-interpreted as /ʃ/.



6. LPD3 testifies that at home as /ə`təʊm/ (Winston Churchill’s regular pronunciation of it) still has some currency. I noticed it in Feb·ry 2011 from the composer Michael Berkeley a one-time announcer on BBC Radio 3. An oddity I often heard from the BBC radio former chief announcer the late Peter Donaldson was /`hwaɪtaʊs/ for Whitehouse: its normal GB pronunciation has a sole /h/ on the latter element. Another word whose accented syllable begins with orthographic <h> which in one of its pronunciations may not contain /h/ is the word perhaps. It's also offen heard as /pər`aps/ as well as /praps, paps/ and its canonical form /pə`haps/.



7. In gen·ral we see that where orthographic <h> begins an unaccented syllable there’s always the possibility that /h/ will be elided. It’s well known to be constantly so from all the weakform words beginning with <h>, he, her, him, his, have etc. A fact I havnt noticed any ref·rence to in the lit·rature is the consid·rable frequency with which the /h/ is elided in phrases in which the word hundred isnt accented such as `two hundred and, `three hundred and etc.



8. We dont normally lose /h/ from hilarity, humidity etc. Even hellhole and hothouse always keep both aitches. Hedgehog and household do still retain aitchless second-syllable minority forms according to LPD and also ODP (which even has one for keyhole) tho EPD only recognises one for household. Aitch is also very persistent in almost all compounds such as beehive, carthorse, childhood, fathead, shorthand and sweetheart. Cruttenden 2014 p.205 sed “Some older RP speakers treat an unaccented syllable beginning with an <h>, as in historical, hotel, hysterical, as if it belonged to the special group hour, honest et...” LPD, EPD and ODP agree only about the first two of these reflecting the uncertainty of judgements in these matters. Compare agen the EPD 1956 record of subvariant aitch-dropping for horizon. LPD3 agrees that it occurs    ‘sometimes’ but not so ODP or EPD 2008.



9. Most words not of native origin but borrowings into English from French gradually had their etymological aitches restored to them by misplaced scholastic zeal. That the aitchless forms were indeed considered “vulgar” by classical Latin speakers can be seen in verses by the Roman poet Catullus who made fun of unsophisticated aitch-droppers in precisely the same way as a present-day writer of English comic verse would do so by representing Harry as pronounced ’Arry. 



10. An odd case that escaped a revivalist aitch was able which wasnt realised to have descended from Latin habilis. In some cases there was “restoration” of an aitch which never existed in Latin as with hermit from Latin eremita which gave us also the better spelt eremite. Similarly the word hostage went back ultimately to the Latin word obsidem. The word humble in its occurrence in humble pie has an unhistorical aitch in its spelling due to confusion of the obsolete aitchless word umble (ie deer innards) with the common adjective humble (the opposite of proud). In sev·nteenth-century English and earlier the unetymological spelling abhominable had quite some currency. The word halcyon has been in the language for over six hundred years but it’s dou·tful that any normally ‘aitch-keeping’ speaker has ever used it without an aitch. Nevertheless it comes from a Greek word for kingfisher that the Greeks didnt say with any aitch: it acquired its illegitimate aitch it on its way thru Latin. Another word that lost its aitch before we got it was assassin which is related to the word hashish.



11. Regarding the rise of the aitch-dropping social shibboleth, in the sixteenth and seventeenth centuries there were no signs of such a thing. It seems not to’ve arisen until well on in the eighteenth, that century of increasing consciousness of correctitude. 

We’re indebted to Lynda Mugglestone for her remarkably exhaustive account of the extensive literature of this phenomenon in her book Talking Proper subtitled the 'The Rise of Accent as a Social Symbol' (1995, 2003) a very major proportion of which she devoted to this particular shibboleth. We can be confident that she was right when she sed (p.99) of aitch that there is ‘little room for doubt that its use in the earlier part of the eighteenth century (and before) can hardly have been invested with those social values which later came to be so commonplace in comment on language and propriety’. She quotes at page 34 from a 1762 Course of Lectures on Elocution by Thomas Sheridan (who offered advice to aitch-droppers) notably identifying him as ‘the first writer to record in terms which reveal negative sensitization to its use’. He apparently recommended them to read aitch-beginning words aloud and to ‘push them out with the full force of breath till an habit is obtained of aspirating strongly’. It’s piquantly noticeable that in so saying, he shd use that form of the indef·nite article. In 1791 John Walker in his ultimately hugely influential Critical Pronouncing Dictionary (1849 reprint p.11) warned his ‘countrymen the Cockneys’ agenst ‘Not sounding h where it ought to be sounded’.
 


12. Mugglestone quotes the late-nineteenth-cenury phoneticians Alexander Ellis and Henry Sweet as referring to correct use of /h/ as ‘a test of education’. In 1877 in his book Speech in Song, aimed at one of his less learnèd audiences (being an item in a  popular series of Novello ‘music primers’) Ellis said bluntly ‘The omission of the aspirate at the beginning of words is considered almost a crime in educated society’ (p.22). In another place (in Early English Pronunciation 1869 Vol. I p.221) he’d referred to aitch-dropping as ‘social suicide’.



13. In a recent development the name of the letter <h>, 'aitch' seems to be ceasing to be exclusively pronounced with no /h/ by chiefly younger people of some linguistic awareness. Of course the Wells opinion poll respondents are self-selected and anonymous but the poll of 1998, reported in LPD3 at its entry for the letter H, revealed that 16% of those questioned actually exprest a pref·rence for the pronunciation /heɪʧ/.


14. By the way, we may compare the treatment of words beginning with an orthographic but unsounded aitch with a similar pattern in French. There are many words in modern French spelt with an initial h tho none of them today corresponds to an aitch sound. Those derived from Latin have had an aitch letter re-inserted into their spelling from misplaced scholasticism. The reflex of Latin hora, for example, is he·rd in speech as l'heure with the preceding article in the same form as for a word in which an aitch had never existed. On the other hand, considerable numbers of words borrowed into early French from Germanic and other languages with initial /h/ in their spellings may be presumed to have preserved that sound for a while because its 'ghost' provokes such forms as the unreduced article regularly occurring in expressions like la halle. With droll paradoxicality (a lucus a non lucendo if there ever was one) the French refer to these ghostly aitches as ‘aspirated’.


15. In the earlier nineteenth century it was common to see items like our quotation from Sheridan above and in many later writers including Jane Austen. She made constant use of an before many words including eg handsome, happy, hasty, heart, heavy, hill, house and hurried. Another famous writer, Andrew Marvell, in his poem  ‘To His Coy Mistress’ early in the second half of the seventeenth century, wrote “An hundred years should go to praise thine eyes”... Readers aloud and actors always nowadays normally observe each /n/ and also use each /h/ but one wonders exactly how long that custom has existed. There are various possibilities. It seems very likely indeed that in the sixteenth and seventeenth centuries and much of the eighteenth there was, not necessarily in a predominant proportion of the lit·rate population, a good deal of licence to omit word-initial aitches, no dou·t especially when words were not accented. This practice might well have passed unnoticed by speakers in gen·ral just as the uses of weakforms of aitch-beginning pronouns are today largely unobserved and occasionally even condemned by the phonetically naïve. There’s a fair amount of evidence suggesting this might've been so even in the highest echelons of society from examination by language historians of manuscript writings etc. One notable account of such types of evidence is to be seen at pages 294-296 of H. C. Wyld’s History of Modern Colloquial English. This includes a quotation from the London-based Scottish educationist and spelling-reformer James Elphinston (1721–1809)  ‘many Ladies, Gentlemen and others have totally discarded initial h- in places where it ought to be used...
 ’

16. What may well be some support for the suggestion we’ve just described is provided by a couple of quotations from OED:
(i) Shakespeare Two Gentlemen of Verona  ii. v. 46.  ‘If not, thou art an Hebrew, a Iew, and not worth the name of a Christian’
and (ii) Milton Samson Agonistes  ‘Thou knowst I am an Ebrew’ where Milton doesnt even include an aitch in the spelling of the word Hebrew. The 1611 Authorised Version of the Bible has numerous occurrences of an before a variety of aitch-beginning words. In the case of Jane Austen, since she was born as early as 1775, it’s an open question how much she might’ve been conscious of the relatively recent tendency to  ‘criminalise’ aitch-dropping. It can’t be presumed from her use of spellings like  ‘an house’ that she necessarily used no /h/ in saying them. She, and the many others who used them, may very well have treated such expressions in the same way as many people today choose to say ones like  ‘an historical novel’ as /ən hɪˈstɒrɪkl `nɒvl/.

 

I shd like to express my thanks to Amy Stoller for valuable advice on points of American pronuncation in preparing this blog.

Ticklish Words (ii) GARAGE

Anyone who re·d my last blog will no dou·t be expecting this one to be about the word garage and so it is. However, Graham Pointon has rather pipped me at the post and left me relatively little to add to the 700-odd words of his fascinating Linguism remarks on the subject.

Anyway, this word came into English from French at the very beginning of the last century. OED3 online has four quotes from 1902 recording the first borrowings of it but it has as yet only substituted "/ˈgærɑːʒ/ /ˈgærɪdʒ/" for the Murray-style  "(gæ·rāʒ, gæ·rėdʒ)" C. T. Onions supplied in the OED 1933 Supplement. The word ultimately comes from the Nordic substratum of the modern Romance language possibly related to our word g·ard and cert·nly so to the French word gare (for 'station').

The Jones EPD in 1917 listed it as = (by which sign I borrow from the mathematicians I mean 'equivalent to in sound without exact copying of symbols used') / `gӕrɑːʒ /, already confirming the existence of the anglicised subvariant with /-ɪdʒ/. Ten years later he showed another subvariant ending with /-ɑːdʒ/. The current EPD has added to these "occasionally: gəˈrɑːdʒ, -rɑːʒ". These last two, in reverse order, it gives as the only EPD American versions. MWO (the Online Merriam Webster) reports that, unlike the US, Canada has the surprising variants /gə`rӕʒ/ & /gə`rӕdʒ/.

Graham was very amusing on the crassness of Robert Bridges on pronunciation matters. One can just imagine how embarrassing it must've been for Jones and the other genuine linguistic scholars to have to put up with the cranky ideas of the Poet Laureate who was John Reith's choice of chairman of the BBC's Advisory Committee on Spoken English. The "extraordinary" publication of the unlamented ill-conceived "Society for Pure English" edited egregiously by Bridges was truly /ɪkstrɔːdnri/.

The recent remark made by Robert Walshe of the British Library on BBC Radio 4 – that /-ɪdʒ/ is increasing in frequency at the expense of /-ɑː(d)ʒ/ – was obviously inspired by the LPD. The graphic there, besides indicating the general preference among British respondents to the Wells poll questionnaire for /ɑː/ over /ɪ/ as the second vowel (56% to 38%), vividly illustrates the increasing number of younger ones who went for the latter choice. Surprisingly, to me at least, there were 6% who actually favoured stressing the latter syllable — as of course did all the respondents to the American poll. It seems that the eagerness of American speakers to generally accent the final syllables of French words (and probably in consequence certain others such as Suez and sometimes Zagreb) can be attributed to the influence of Noah Webster or some such pontificator. It was most likely based on the dodgy idea that French speakers regularly accent the final syllables of all French words. They don't. What they do do is, when they utter a single word in isolation, stress the last syllable of that 'sentence'.

Ticklish Words (i) APPLICABLE et cetera

The British Library's "Voicebank" collection in connection with their 'Evolving English' exhibition uses a "familiar text with intentionally straightforward language [which] allows speakers of all ages to read confidently, including non-native speakers of all abilities. Pilot studies with the text confirm it also encourages a relaxed, informal speech style". They aim to "ensure the recordings we collect will be valid for research now and in the future". Participants are requested to record a nicely conversational-sounding ("Mr Tickle") children's-bedtime-type story of 612 minimum (720 maximum) words or, failing that, just the six words applicable, controversy, garage, harass, neither and scone. These all can be sed to mainly differ transAtlantic·ly. Only the third and fifth of them can be completely confidently described as "known to ev·rybody".  

applicable

has been recorded since the sixteenth century. Walker in his Critical Pronouncing Dictionary (1791/1849) listed it only as stressed on its first syllable. So did Murray in the OED in 1885.  The earliest evidence of the later stressing ap`plicable was its listing as a subvariant form in the Daniel Jones EPD1 of 1917. LPD at its first edition in 1990, offered the evidence of a poll that suggested that later stressing was more 'popular' in the UK and gave it first placing. In 2000 it gave poll evidence of substantial, tho not majority, use of later stressing in the US too. Accordingly it lists the forestrest version first for US. Another Wells poll in 2007 gave virtually the identical results. See my Blog 042.  EPD15 from 1997 has agreed with the LPD order of the stressings for UK but made no differentiation for US usage. OED3 in 2008 listed two British versions giving preference to the one with stress on the second syllable. It listed only the forestressed version for American usage despite the evidence of the LPD poll.

The change to ap`plicable that we've seen in the UK was a very common type around the earlier twentieth century exactly paralleled in the words despicable and formidable tho Jones in 1917 only at first acknowledged the stressing for`midable in that word's use as the name of a Royal Navy ship. He accorded it unqualified acceptance,  as a subvariant, in 1937. For further examples of this kind of development see this website Section 3.7.III §10. This kind of change wasnt universal. Some words have tended to go the opposite way eg distribute but not the verb attribute.

controversy

This has had a quite similar development to applicable. Walker only front-strest it. Murray in 1893 in OED1 showed it only as (the equivalent of) /`kɒntrəvɜːsɪ/ as which it still remains in OED, being only re-transcribed, not revised, in 1989, and not having as yet been included in the extensive revisions of the online OED3. It was not listed with its now-UK-predominant late stressing con`troversy in the Jones EPD until 1937 when the form was given between the square brackets Jones used-to put around subvariants. There can be no dou·t that the later-strest version was in circulation some years earlier because in 1928 the BBC publisht a booklet Broadcast English: Recommendations to Announcers Regarding Certain Words of Doubtful Pronunciation, edited by Arthur Lloyd James, in which it figured along with the decree "stress on 1st syllable". When Gimson took over editing the EPD in 1967 it appeared as a co-variant form (ie with its former subvariant-indicating square brackets withdrawn).

When the LPD first appeared in 1990 it received a special note: "Among RP speakers the kɒntr- form probably still predominates; but in Br E in general the -ˈtrɒv- form is now clearly widespread". In the British poll 56% had preferred that later stressing. JCW added "In AmE ˈkɑːntr- is the only possibility". For half a century after the Lloyd James decree the BBC continued its embargo. In 1981 the late Robert Burchfield, at page 12 of the booklet The Spoken Word: A BBC Guide, listed it with "stress on the first syllable" but softened that with a footnote saying that the stress values were "at present nicely balanced"; however adding with traditional BBC prescriptiveness "Meanwhile the traditional stressing should be observed". The latest word from the Beeb in its OBG (the cumbersomely and, really, inappropriately titled Oxford BBC Guide to Pronunciation of 2006) admitted that the later-strest version was "equally, if not more, common, and equally acceptable". Many BBC employees seemed to have the jitters over this word for decades. At one period I repeatedly he·rd some of them coming out with a galloped version /`kɒntrəvəsɪ/ which wasnt really a very natural form for anybody much. EPD, by Roach-and-co since 1997 from CUP, still puts the forestrest version first for "BBC English" pronunciation.  [PS: Within less than a month of my mini rant agenst the relatively unusual truncated version /`kɒntrəvəsɪ/ in that last sentence, I came across it enunciated gracefully by one of the best of the current BBC Radio 4 newsreaders by name Rory Morrison, incidently notable for his particularly "agreeable" choices of intonations.]                                      




                  

GB Phoneme Symbols

Students of GB ie General British English are well advised to be clear in their minds about the basic structure and relationships of the forty-four phonemes of this most neutral type of British accent.
The consonant phonemes may be grouped conveniently into three 'octets' with easily understood correlations:
p     t      k     ʧ     f     θ     s     ʃ
b     d     g    ʤ    v     ð    z      ʒ
m    n     ŋ     l      r     j     w     h
It's a good idea to memorise this pattern, especially its first line, if only in order to be sure of taking account of the very marked shortening effect of the voiceless items on any other phonemes (notably vowels) which happen to precede them in the same syllable. This is sometimes known as 'pre-fortis' shortening. Fortis, meaning 'strong' is another way of referring to the first line of consonants which are typically both voiceless and relatively forcefully articulated.

The first two lines, each of successively plosives, affricates and fricatives are precisely matching voiceless and voiced counterparts of each other. The first three items in all three lines have matching places of articulation. The remaining five items are all classified, if not universally at least according to some authorities, as 'approximants'. The first two of these approximants are characterised essentially by two different types of tongue contraction: respectively mainly lateral and mainly longitudinal. We place /l/ before /r/ because the tongue tip is in the fronter position. The final three are vocalic and listed in order of frontness/backness. The first two of them, /j & w/ differ from the vowels /i ː & uː/ only by being weaker and shorter. Calling these three items consonants is a phonological rather than phonetic classification. The last item /h/ may correspond to /iː/ or /uː/ but no more so than to all the vowels of English. It never ends syllables and takes the physical form each time of a voiceless version of any vowel it precedes. It's no less logical to group /h/ with the voiceless fricatives but that does make for a rather less satisfying symmetry.

Classifying the GB vowels isnt such a universally accepted procedure because any grouping must choose between giving priority to one or another of the features of quality or length or diphthongality. Within the category of diphthongs there are also choices to be made between the three types with different directions of movement. The result of this situation is that there are many different preferences to be found among analysts. Daniel Jones (1881-1967) preferred mainly to list them according to tongue-posture. For the non-diphthongal ones he started with the closest and continued anticlockwise around the vowel space from front to back until after the last of these he listed the central ones. After them, in somewhat more arbitrary order, he listed the twenty main diphthongs of GB.
These we list in their usual current symbols (as in LPD3 and EPD17) rather than using any of Jones's various choices he made on particular occasions:
iː, ɪ, e, ӕ, ɑː, ɒ, ɔː, ʊ, uː, ᴧ, ɜː, ə
eɪ, əʊ, aɪ, aʊ, ɔɪ, ɪə, ɛə, ʊə.
When in 1962 Gimson (1917-1985) produced a new account of the GB vowels which began supplanting Jones's as the most authoritative listing, he tabulated them as
Short: ɪ, e, ӕ, ɒ, ʊ, ᴧ, ə
Long: iː, uː, ɑː, ɔː, ɜː
Remember that these five vowels have their length marks not because they must be long but because they're characteristically rather long. In fact they can even be very short on many occasions depending on various rhythmic factors. Also the ones that dont have length marks can be stretched a lot on some occasions.
Diphthongal, front-closing: eɪ, aɪ, ɔɪ
Diphthongal, back-closing: əʊ, aʊ
Diphthongal, centring: ɪə, ɛə, ʊə

Most modern British dictionaries, including those aimed at advanced users of English as an extra language, employ the two symbols /i/ and /u/ to specify rhythmically weak allophones of the /iː/ and /uː/ phonemes. GB speakers have a small repertoire of nasal vowels employed solely for loanwords: they are in fact nasalisations of /ӕ, ɒ, ɑː/ and /ɜː/ viz /ӕ̃, ɒ̃, ɑ̃/ and / ɜ̃ /. Diagrams showing the ordinary (non-nasal) GB vowels and diphthongs can be seen at §46 of Item 3.1 elsewhere on this website.

BlogRoll

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