John Wells’s phonetic blog

How can you tell that this visual pun was created by a north American? Because it doesn’t quite work in BrE or, therefore, in Australian/NZ/South African English.

Why not? First, and obviously, because in BrE barium has the stressed vowel eə while bury has e. So for us Brits barium sounds different from bury ‘em. But in AmE they are homophonous.

But secondly, I think, because we don’t use the form ’em for them as freely as Americans seem to. Our weak form of them is generally ðəm, with the initial consonant retained. As the OED (1891) comments under ’em,

The emphatic form of the pronoun was early superseded by THEM pron., but the unstressed form continued to be used, being regarded as an abbreviation of them. In literature it is now obs. or arch., but is still common in familiar speech.

Obsolete or archaic… yes, but not really "still common in familiar speech" (or so it seems to me). Rather, as far as I am concerned it seems to be generally restricted to a few set formulaic expressions such as If you can’t beat ‘em, join ‘em and give ‘em the money. Beyond that, them can lose (or assimilate) its ð in the same way as that, the or they when following z in phrases such as is that, was the, claims they, sees them.

When I was a small child we spent family holidays in a friend’s unimproved rural Yorkshire cottage called, for some reason, Buryemwick ‘Bury them alive’. But if asked what to do with dead cats, for example, we’d never have said “bury ‘em”, but rather “bury them”.

It would also be possible to take the final əm of barium as representing not ’em but him (‘im) — but only in an accent of English that has lost the contrast between ɪ and ə in this position, as in often the case in AmE but generally not the case in English English.

A purist might complain that you can’t pronounce a phoneme as such (still less a symbol), but only one or other of the allophones that manifest the phoneme — for example the r in red is different from the r in tread and the r in dread, so that you can’t adequately illustrate them all with one single sound clip; but by now we are used to teachers of EFL wrongly calling phonetics ‘phonology’ and speech sounds ‘phonemes’. No matter.

When you click on the symbols, however, you find something more disturbing. Clicking on a vowel or diphthong symbol is fine: you hear the corresponding sound (though their ʌ sounds a little bit odd to me). But when you click on a consonant symbol, in some cases you get the corresponding sound, perhaps lengthened (true for ŋ, s, z, ʃ ʒ, tʃ and r, for which you get ɚ), but in other cases the corresponding sound plus schwa, thus pə, bə, tə, də, kə, ɡə, mə, nə, fə, və, θə, ðə, dʒə, lə, jə, wə.

It is particularly confusing to treat different nasals differently, nə but ŋː, different liquids differently, and different affricates differently.

No one should be taught that the w in well, for example, is pronounced wə, because that would make it wəel.

It’s the usual problem: people making pronunciation practice materials don’t get advice from a phonetician, but imagine that every native speaker is an ipso facto expert.

Remember, when it’s language, people never check. They never call a linguist. They just make stuff up.

Anyhow, we mustn’t complain too forcefully, seeing that the site is free and also contains jolly games such as ‘Phonemic Hangman’ and ‘Phonetic Pelmanism’.

There’s a letter in today’s Independent about English spelling.

The rules governing pronunciation in English are more complex than, say, German or Spanish. Most of us learn to overcome these obstacles when we learn to read at primary school, but some do not. The education of this latter group is blighted by the twin evils of complex pronunciation rules and non-phonetic spelling in the language in which subject material is written.

The writer is using the term “pronunciation” here in a sense different from any of the senses listed in standard dictionaries, e.g. the Concise Oxford:

1. the way in which a word is pronounced, esp. with reference to a standard.
2. the act or an instance of pronouncing.
3. a person’s way of pronouncing words etc.

The verb “pronounce”, in turn is defined as

1. utter or speak (words, sounds etc.) in a certain way.
2. utter or deliver (a judgement, sentence, curse, etc.) formally or solemnly…

But what the writer means by “the rules governing pronunciation” is clearly the supposed “rules” determining a word’s pronunciation on the basis of its spelling. And as we all know, many of the spelling-to-sound rules of English, such as they are, are subject to numerous exceptions and irregularities.

Indeed, the writer goes on to argue that “part of the solution to the problem of illiteracy could be spelling reform”.

If a book is described as dealing with “pronunciation”, we expect to find an account of phonetics, not of spelling-to-sound complexities. David Crystal’s recent book (blog, 31 Oct) is correctly subtitled “the singular story of English spelling”, not “…of English pronunciation”.

Non-phoneticians do tend to get confused about the difference between sounds and letters, pronunciation and spelling, phonetics and orthography.

Just yesterday a PR person for Scrabble emailed me and then phoned me to ask for assistance with a story about “the most commonly mispronounced words in the English language”. She said she was concerned with native speakers, not EFL. It was not a matter of social/geographical features such as glottal stops. On my further questioning it turned out that she did not know whether she was thinking of articulatory difficulties such as a speech & language therapist would deal with, or with words of contentious pronunciation, where speakers have different views on what is correct. She said she’d get back to me when she’d thought further about it.

On reflection, though, I have an awful feeling that what she may really be thinking of is none of the above, but rather words that have “mispronunciations” associated with misspellings. Things such as *mischievious mɪsˈtʃiːviəs instead of mischievous, for example, or … *mispronounciation.

If she rings back, how much should I charge as a consultancy fee?

Here’s what it says in the recently published (and generally pretty good) textbook Better Pronunciation for Communication (Shanghai Foreign Language Education Press, 2012) by Liu Sen of East China Normal University. I read the manuscript of this book before publication, and managed to persuade the author to remove nasal release and lateral release as subcategories of ‘incomplete plosion’ and treat them separately; but she didn’t feel she could drop the term ‘incomplete plosion’ altogether, as I suggested, since it was, she said, so well established in English phonetics terminology in China.

Now, seeing the textbook as printed rather than in manuscript, I realize I ought to have intervened more forcefully.

Of the four examples given, numbers 2, 3, and 6 each have a plosive that is more likely to be elided than given no audible release; in number 5 NSs would be more likely to use a glottal stop or a no-audible-release assimilated [p]. A hold has "double the usual time" only in the gemination type, of which there are otherwise no instances here.

In numbers 3 and 4 of these examples NSs would probably elide the highlighted plosive; in numbers 1, 2, and 5 the plosive would have an ordinary oral release (though masked by the following fricative); in number 6 we might alternatively use [ʔ].

You can see why I chose as the topic for my workshop session in Shanghai ‘The characteristics of English plosives, and in particular the various ways in which they can be released’. Here is my take on what in my opinion is mistermed ‘incomplete plosion’ (even though, as Alex Rotatori pointed out, my colleague Patricia Ashby has used this term in a recent textbook).

I’m safely back home in England now after my trip to Japan and China.

I didn’t sleep much on the outbound flight from Amsterdam to Osaka. Once the meal was cleared and the lights dimmed I tried to, but kept finding myself watching the flight information screen, which alternated between a map of our current position and the details of our altitude, distance to destination, ground speed, etc. The language of latter cycled between Dutch, English and Japanese.

I noticed that whereas in Dutch and English kilometre was abbreviated to km, as you would expect, in the Japanese it was abbreviated to キロ, which is the katakana for ki-ro.

I mused on the fact that in English kilo ˈkiːləʊ is used as an abbreviation for ‘kilogram’ but not for ‘kilometre’. Is this true for AmE as well as for BrE? And why do we say unabbreviated kilogram with kɪ- but its abbreviation with ki:-? (Is this also true of AmE?) And why, as an abbreviation for ‘kilometre’, but not for ‘kilogram’, or for that matter ‘kilohertz’, do we say just K keɪ? To these absorbing questions I do not know the answers.

Then I further noticed that in giving our altitude ‘metre’ was written unabbreviated in Japanese, as メートル me:-to-ru. Its form suggests that Japanese must have borrowed this word from French (mètre mɛːtʁ(ə)). If it had come from English, it would presumably have taken the form ミーター mi:-ta:.

I believe that in Japanese, as in Chinese, ‘metre’ can also be written with the Chinese character for ‘rice’, 米 , which in Chinese is mǐ.

The usual stress pattern for two-word names is double stressing (main stress on the second word), as in ˌHenry ˈSmith, ˌMerton ˈPark, ˌLyme ˈRegis. This also applies to the Latin names of plants and animals: ˌQuercus ˈrobur, Diˌcentra specˈtabilis, ˌEquus ˈzebra, ˌPasser doˈmesticus.

But in cases such as Vulpes vulpes, Troglodytes troglodytes and Glis glis this usual pattern collides with the deep-seated Germanic principle of deaccenting repeated material. So do we keep double stressing, or do we deaccent the specific and shift the main stress onto the generic?

In the case of the edible dormouse, discussants on a recent television programme went for the single stressing, pronouncing it furthermore as if it were a single word, a common noun, a ˈɡlɪsɡlɪs. Furthermore they treated this word as invariant for number, like sheep (see screenshot above). —Well, you’d hardly expect them to know that the Latin plural of glīs is glīrēs. That’s strictly for us classicist showoffs.

What do we do when referring to people whose forename is identical with their surname? What stress pattern do we use for someone called Morris Morris or Graham Graham? Do we find such names awkward? No, I think we cope and give them the usual double stressing. Same with New York, New York. So why is Glis glis different?

_ _ _

May I remind would-be commentators that I no longer allow anonymous or pseudonymous comments. Please sign with your true name. Unsigned comments will be deleted.

_ _ _

In a few days I shall be leaving for a visit to Japan and China. So this blog will be suspended now for the rest of November. Next posting: 3 Dec.

• Above all, they should not set students the dispiriting task of learning the spellings of lists of difficult words presented out of context.
• The “short word rule” for content words accounts for the doubled consonants of inn, egg, add, odd, ill and the final e of eye, owe and bye. Compare function (non-content) words such as in, up, to, if, as, by.
• Pay attention to stress, which explains the doubling of the consonants in preferring, preferred but not in proffering, proffered.
• Be aware of the morphology (or that of the Latin origin), so as to understand, for example, the single b of aberrant (ab + errant) as against the doubling in abbreviate (ab+brev-). This even explains accommodate (ad (ac) + con (com) + mod-).

I would add the mnemonic value of related words, as when definition reminds us that definite is not *definate, while substantial and residential remind us how to spell the endings of substance and residence.

As far as reforming the system is concerned, Crystal declares baldly that “there can never be a simple solution to the problem of English spelling”. On the other hand he twice refers to the fact that Google shows the non-standard spelling rubarb to be increasingly common online. “If it carries on like this, rubarb will overtake rhubarb as the commonest online spelling in the next five years.” Then dictionary makers will “eventually have to recognise that a change has taken place” (as they already have in the case of miniscule replacing minuscule).

By this logic, dictionaries of the coming decade will also have to recognize seperate, tounge, accomodation and so on ("misspellings" very frequently encountered online), and abandon such distinctions as lose — loose, rein — reign, sight — site, to — too, your — you’re, its — it’s (all often confused on the web). Or perhaps ever more intelligent spell checkers and speech-to-text technology will prevent this from happening after all.

I think it’s important to recognize that planned, systematic reform is not truly impossible. Consider the case of the chemical element sulphur. That’s how it was standardly spelt, at least in the UK, until twenty years ago. But in 1990 the International Union of Pure and Applied Chemistry decided to adopt the spelling sulfur, and two years later the Nomenclature Committee of the Royal Society of Chemistry followed suit. In 1992 the Qualifications and Curriculum Authority for England and Wales recommended the f spelling, which is accordingly now found in textbooks and GCSE exams. I think it’s better for everyone to have an official change like this, so that we know where we are, rather than unofficial and chaotic rubarb-style changes.

Another similar example is the immunosuppressant drug of which the British Approved Name was formerly cyclosporin but is now ciclosporin. What used to be the correct spelling is now considered a mistake; what used to be a mistake is now correct. It might be better simply to allow both versions.

Unofficial changes do sometimes succeed, too, as with today, tomorrow, tonight, which have replaced the hyphenated to-day, to-morrow, to-night of my schooldays.

We could consider, for example, getting the QCA to make an official decision that all words with rh may alternatively be spelt without the h, just as we allow likeable alongside likable and (in Britain) organise alongside organize. That would take care not only of rhubarb but also of rheumatism, rhythm and rhino.