spell it out

I’ve been reading David Crystal’s new book Spell it Out. Like everything he writes, it is entertaining and well-informed, and thoroughly to be recommended.

The core of the book is a historical account of how our spelling ended up as the infuriating mess that it is. We have the Anglo-Saxon monks adapting the Latin alphabet to the sounds of Old English, bequeathing us the gn- of gnat and ways of coping with vowel length, with final e to show a long vowel and consonant doubling to show a short one (hence hope — hoping and hop — hopping). There were the Norman clerks faced with writing an unfamiliar language, experimenting with various possible solutions and hitting on such ideas as letter doubling to show the long vowels of deep and moon. There was even the monk Orrm who tried unsuccessfully to persuade everyone to adopt a more rational and systematic orthography. Meanwhile, our basically Germanic language had to digest thousands of new words of Romance origin, which meant for example that we applied the consonant doubling principle to show the Middle English pronunciation of various French-derived words (e.g. baggage).

Contemporary handwriting meant that vv looked confusingly like the w that had taken over from OE wynn (ƿ); so doubled v was avoided where it would otherwise have been called for, for example after the short vowel of loving (compare the single v of roving); meanwhile even single v was not at that time distinguished from u, so that lov looked like lou, which led to the adoption of a final silent e to show that there was a consonant sound at the end of love, have, give — leaving the anomaly that persists to this day when we compare these words with grove, rave, drive; and leading to the homographic ambiguity of live (which can be either the verb lɪv or the adjective laɪv).

Then came major sound changes, notably the Great Vowel Shift, which left us with sets of related words in which the common element is still spelt identically but nowadays pronounced very differently, and in which medieval scribes and printers opted to follow the sense rather than the sound: crime — criminal, type — typical, cave — cavity and so on.

Then came a concern for etymology. As Latin words were borrowed they kept their Latin spellings (hence single r in florid but double in horrid). The unpronounced etymological b was “restored” in debt and doubt and likewise the p in receipt, though not, for some reason, in the exactly parallel deceit and conceit. Somewhere along the line Flemish compositors decided to introduce an h into ghost.

(to be continued)

high head, falling head

When I received an email signed “Seeka You”, I assumed that this was a rather odd pseudonym used by someone reluctant to give their real name. I was accordingly disinclined to expend time and effort on a reply. But the person in question assures me that this is their true name. Anyhow, he/she had a question about intonation, which I had earlier failed to answer when it first reached me immediately after my stroke (at which time I was in no state to answer it).

In your book English Intonation, you follow the principle that “the high falling head is used only before a fall-rise nuclear tone” and “the high level head is used before all other nuclear tones”. I’ve listened closely to RP speakers in different styles of speech, and I can't seem to avoid the impression that most of the time (70%?), they use a falling head regardless of the nuclear tone; using a high level head seems to be very formal and infrequent even in fairly formal speech. Is my perception incorrect?

To explain: The head is the piece of the intonation pattern that extends from the first accent up to but not including the nuclear accent. Prenuclear patterns (and therefore heads) are one of the less important phenomena in the intonation of English (they don't seem to encode much meaning), and I therefore relegated the whole matter away from the core chapters of my book to chapter 5, “Beyond the three Ts”.

I think the last two sentences constitute a pedagogically justified simplification of the complex reality. In recording the copious spoken examples that accompany the text we found it easy and authentic-feeling to conform to it. The footnote to the last sentence you see above reads as follows.

But I am not aware of any corpus-based analysis that would enable us to judge whether this pedagogical simplification deviates seriously from actual usage. So I told S.Y.

I don't know. It’s an empirical question to which neither you nor I know the answer.
There’s perhaps also a question of definition: how much of a downward deviation from level has to be present in a high head before we categorize it as ‘high falling’?
I certainly don’t feel that high falling heads (as I perceive them) fit naturally for me before anything other than a fall-rise.

I might have added that this is in all likelihood one of the things that varies considerably between accents. High falling heads before a high-fall nuclear tone (which S. Y. claims to be so usual) would put me in mind of a Highland Scottish accent; they don’t feel at all right for my own speech.

biopic

An interesting misinterpretation of spelling from Richard Osman, the resident expert on the BBC1 programme Pointless: he referred to a baɪˈɒpɪk, that is a biopic, a film about someone’s life, a filmed biography. It is, of course, normally called a ˈbaɪə(ʊ)pɪk, being composed of bio- plus -pic(ture).

Given bionic baɪˈɒnɪk and myopic maɪˈɒpɪk ‘short-sighted’, you can understand where he was coming from. After all, biopic looks as if it contains the suffix -ic, which regularly throws the word stress onto the preceding syllable.

This word thus joins a list led (!) by misled (ˈmɪzl̩d instead of ˌmɪsˈled) and also containing items such as the seabed siːbd, infrared ɪnˈfreəd rays, and (my favourite) ˈsʌndrid (sundried) tomatoes.

slit fricative [t̞]

On a certain social networking site to which I belong, Tim Morley asked

Can someone explain to me, in terms of tongue position or whatever's relevant, the scouse allophone of /t/ that's a sibilant rather than a stop?
A French colleague asked me why a scouser had referred to "The Cass in the Hass", and although I believe I can perceive a difference in the cat-cass minimal pair through the scouse accent (i.e. I don't think it's just context that's allowing me to correctly interpret a homophone), I couldn't actually explain the difference that I believe I'm hearing, which kind of annoyed me to be honest!

So I said

It’s an alveolar slit fricative rather than plosive — a kind of lenition. It can be lenited in some words even further to [h]. See my blog for 15 Nov 2006.

Tim followed up with

By "slit", do you mean simply that there's "only a slit" (i.e. a very narrow gap) between tongue and alveolus? Or is there some fuller meaning that I'm not aware of?

So I clarified.

It means the space between the tongue tip/blade and the alveolar ridge is a left-to-right gap ("slit") rather than the front-to-back "groove" you get in [s]. [θ] and [ð] are dental slit fricatives, [s] and [z] are alveolar groove fricatives. The Scouse thing combines the shape of the first pair with the place of the second.

Then Tim discovered a Wikipedia claim that "no language is known to contrast a grooved and non-grooved sibilant".

Does this mean that Scouse would constitute a counterexample to this supposed universal and thereby disprove it? Hardly. I pointed out that

the nongrooved sibilant is not the default realization nor the most frequent realization of Scouse /t/. […The opposition exemplified in hit/hiss' operates] only in final or prevocalic position, and it’s not the only possibility even there. Note that in hit back (vs hiss back) you probably wouldn't get the slit fricative, but a no-audible-release [ʔ], [p] or [t].

Kevin Watson adds:

In a word like quite or internet you'd get a slit-t (of varying kinds, I've called them e.g. 'dynamic sibilants' and 'canonical sibilants' although I'm not happy with those terms) but in what or biscuit you'd get [h].

This same [t̞] articulation of /t/ in postvocalic position is found in Irish English (“soft t”). Indeed, it’s one of the most indexical features of a (southern) Irish accent, and in Liverpool obviously derives from Irish influence.

the ending -d

Some of the unsolicited queries about English phonetics that I find in my inbox are easily answered. Zheng Yong’s was one such.

One book for Chinese Primary tells [= says] that "liked" is re[a]d /laikd/. What is your point [=opinion]?

I’m sure he really knew the answer already, so I made it short and sweet.

The book is wrong.

It's wrong because liked is pronounced laɪkt. The past ending -(e)d is pronounced as ɪd (or əd) when attached to a stem ending t or d, and otherwise as d with a stem ending in a voiced sound, but as t with a stem ending in a voiceless sound. So we have t in clapped klæpt, coughed kɒft, kissed kɪst, wished wɪʃt, touched tʌtʃt, and, yes, liked laɪkt.

That’s the story for students and teachers of EFL phonetics, anyhow. It is supported, for example, by the fact that missed is pronounced exactly the same as mist (both mɪst), while passed is a homophone of past, and backed rhymes exactly with act.

Those whom people in linguistics (not EFL) call phonologists, however, may argue that the underlying representation of liked is indeed /laɪk+d/ (or, for followers of Chomsky & Halle, the more abstract pre-GVS /līk+d/). They would say that the underlying representation of the past ending is /d/, but that an obligatory rule of voicing assimilation causes this /d/ to surface as [t] when attached to a stem ending in a [-voi] segment. Or, equivalently, that there is a constraint on the value of the feature [voi] that causes [-voi] to spread from the end of the stem to the end of the word. (For a worked example of this idea as applied to the English plural ending, see here.)

Be that as it may, I can’t end this little discussion without mentioning the many West African speakers of English who pronounce the ending as [d] after voiceless stems just as after voiced ones, and operate voicing assimilation in the other direction. That is, they pronounce kissed as kizd and liked as laigd. How widespread this is in Nigeria or Ghana I can’t say, but it certainly exists. All the same, I don’t think my Chinese correspondent would consider it relevant.

with, regretful

I found myself being just a tiny bit querulous when commenting on a posting in Language Log. A reader had asked about the word with, saying

I have always used unvoiced [th] as the pronunciation of that word, and had never noticed anyone doing otherwise.

As for the voiced [ð] in this word,

I'm interested in what the distribution of this variant is, but I'm having a hard time finding it online

In reply Mark Liberman, the usually very knowledgeable writer of the post in question, said just

Short answer: I don't know. I've never heard a discussion of this point of pronunciation variation, except with respect to the varieties of English that have [wɪf] or [wɪv].

There followed a string of commentators reporting what they said or what this or that dictionary reported.

Finally I felt I must chip in:

Doesn't anyone ever consult my Longman Pronunciation Dictionary? There you will find both preference statistics and graphs for wɪθ and wɪð in both American and British English. Also a note mentioning that "in Britain, wɪθ is nevertheless frequent in Scotland" - again, with statistics.
Why do I bother, if no one reads what I write?

I suppose the problem is in the phrase “finding it online”. People now no longer look for information in books, or in libraries: they expect to be able to locate it in in Wikipedia or via Google. They don’t want the inconvenience and expense of buying a book or locating the book in a library.

So the only way I can reasonably expect to disseminate the research I carried out into whether people prefer wɪθ or wɪð is indeed to put it online, which I shall now proceed to do, Here’s the entry for with from LPD.

You’ll see that in Britain taken as a whole we overwhelmingly prefer wɪð, though the Scots, unlike the rest of us, go for wɪθ. In the States most people, like the Scots, prefer wɪθ. The graphs alongside show that the situation is fairly stable over the generations in the US, while in Britain wɪð is gradually increasing in popularity as we move from older speakers to younger.

What the LPD entry doesn’t tell you, because it’s not really germane, is that there is an archaic/dialectal form with no final consonant at all, represented in special spelling as wi’. There are also forms such as wɪv, wɪf, used by TH-fronters in the wɪð and wɪθ areas respectively, and likewise forms such as wɪd, wɪt used by TH-stoppers. So don’t be surprised if a Londoner (probably young, possibly black) says wɪv or even wɪd, or if a similar NooYorker says wɪf or wɪt. But that’s for the sociolinguists.

intervocalic semivowels

One or two footnotes to Monday’s discussion.

1. The spelling of Libya suggests that it should be disyllabic ˈlɪbjə. But all agree that in English it is actually trisyllabic/varisyllabic ˈlɪbiə. (It seems to be disyllabic in Arabic: ‏ليبيا‎ Lībyā.)

2. With foreign words containing an intervocalic semivowel there is a choice when we anglicize them. Which syllable do we treat the source semivowel as belonging to?

• In some cases we resyllabicate, taking the semivowel into the lefthand syllable to make part of a diphthong. Thus Japanese Toyota トヨタ ˈto.jo.ta becomes English tɔɪˈəʊtə. Kawasaki 川崎 ka ˌɰa sa ki usually becomes ˌkaʊ.əˈsɑːk.i. Sayonara さよなら saˌjoː.na.ˈɾa becomes (in BrE at least) ˌsaɪ.əˈnɑː.rə. Provençal jambalaya becomes ˌdʒæm.bəˈlaɪ.ə, while Spanish playa ˈpla. ja becomes ˈplaɪ.ə and papaya paˈpa.ja becomes pəˈpaɪ.ə. French Bayeux ba.jø usually becomes ˌbaɪˈɜː or ˈbaɪ.ɜː in BrE, while crayon kʁɛ.jɔ̃ becomes ˈkreɪ.ɒn.
• In other cases, though, we keep the semivowel as a semivowel and put it, as we then must, in the righthand syllable. Thus Malawi is məˈlɑː.wi (not *məˈlaʊ.i). Some people do this with Kawasaki, making it ˌkɑː.wəˈsɑːk.i rather than my ˌkaʊ.əˈsɑːk.i. I certainly pronounce Okinawa in English as ˌɒk.ɪˈnɑː.wə, not *-naʊ.ə (the initial vowel’s going to be different in AmE). There’s no choice, of course, in the case of hallelujah ˌhæl.ɪˈluː.jə, because in English we don’t have a falling ʊɪ̯.

3. Alas, my efforts to explain things clearly in the two previous posts don’t seem to have been wholly successful. Houbu Ren now writes

I don't believe I fully understand the phonological difference between million ˈmɪl jən and ˈmɪl i‿ən. Should I sound ˈmɪl jən like ˈmɪl first and then jən? or ˈmɪ first and then ljən? What would this word sound without the ‿? Would it be the same or just has to emphasize the ən, like ˈmɪ l iən ?

As explained in LPD in the text box about Compression (p. 173 in the current third edition, or p. 165 in the Chinese second edition), ˈmɪli‿ən means “two pronunciations are possible: a slower one ˈmɪl i ən, and a faster one ˈmɪl jən. The uncompressed version is more usual in rarer words, in slow or deliberate speech, and the first time a word is used in a given discourse; the compressed version is more usual in frequently used words, in fast or casual speech, and if the word has already been used in the discourse.”

In the second edition, the entry for million read ˈmɪl jən. Responding to a user’s criticism, I changed this in the third edition to ˈmɪl jən ˈmɪl i‿ən. This allows for a trisyllabic version as well as the usual disyllabic one.

Following my syllabification, you should sound it as ˈmɪl first and then jən.

If the compression symbol were not present, thus ˈmɪl i ən, that would imply that only the three-syllable pronunciation was possible. But that would be wrong, because a two-syllable pronunciation of million is certainly not only possible but also usual.