Thursday 8 September 2011

Transcribing the sound of English

Last month, a new book on phonetic transcription was published by Cambridge University Press. The book is called Transcribing the Sound of English: A Phonetics Workbook for Words and Discourse. Its author, Paul Tench, is former senior lecturer in phonetics and applied linguistics at the Centre for Language and Communication Research, Cardiff University.

The workbook 

“is not so much a coursebook in phonetics nor a textbook on English phonology, but a training course in developing students’ powers of observation on features of English pronunciation and their skills in recording them in writing. It begins in a very elementary way but it is thorough, and eventually leads to the most comprehensive coverage of the sounds of English from words to full discourse that is available anywhere.” (p.1)

Petr Rösel has already discussed some of the characteristics of this book on his blog, and has come to the conclusion that he can recommend it to his students, “but with some accompanying comments”. 

I finished reading it two days ago and I can say that it is not bad. The book is fun to read and coverage is also quite comprehensive and clear, especially as far as English intonation is concerned. But there are, unfortunately, also some inaccuracies which I’d like to point out on here. (I find myself in complete agreement with Petr Rösel’s comments, so I’m not going to reiterate them here.)

For some reason, the workbook is riddled with transcription errors. Just to give you some examples: lovesick (p.64) is narrowly transcribed as ˈlʌʌ̥ˌsɪʔkʰ; soon (p.66) as ʃuːn; the verb (to) converse (p.68) is transcribed kənˈʌɜːs, kʰəɱˈʌɜːs; sunk (p.70) is transcribed as sʌɲk, sʌ̃ɲkʰ; (he’ll) meet you (p.93) as ˈmiːʔ u; of is ɒʌ on page 107 and əʌ on pages 108 and 123; he mightn’t care (p.119) is transcribed as hi ˈmɑitŋ ˈk kɛə; and (pleasant) places is ˈpleɪʃəz on page 120. 

There is then the word principles (p.57) which is ˈprɪnsipəlz according to the author, rather than the usual RP ˈprɪn(t)səpl̩z, -ɪp-. Here, apart from the symbol ɪ mistaken for i, it’s the non-syllabicity of the l which strikes me as unusual. I know that some speakers today tend to replace syllabic l and n in words like garden, little, bottle with schwa plus non-syllabic consonant, but this is to be regarded as a change on the verge of RP. That’s why I’m not 100% happy with transcriptions like ˈæpəl (p.52) for ˈæpl̩, ˈlaɪəbəl (p.71) for ˈla(ɪ)əbl̩, ˈɪzənt (p.119) for ˈɪzn̩t, ˈriːznəbəl (p.122) for ˈriːznəbl̩, or indeed ˈkʌmftəbəl (p.122) for ˈkʌmftəbl̩. The transcriptions proposed by Paul Tench are obviously not wrong in themselves, but the author doesn’t mention any other alternative, nor does he point out that these pronunciations are less usual in mainstream RP and are sometimes considered as “baby-like” or “childish”.

On page 98 we read:

“Notice that /t/ does not readily get elided if it would otherwise bring two /s/s together at the end of a word: ghosts /ˈɡəʊsts/”.

This is not true! I would say that pronunciations like ɡəʊsː for ɡəʊsts or ˈsa(ɪ)əntɪsː for ˈsa(ɪ)əntɪsts are quite common in rapid, colloquial RP speech. Here the t gets elided and thus disappears, but it leaves some trace of its former presence behind: the s is lengthened to compensate for its absence.  

Finally, on pages 9-10, we read that in IPA the symbol e “represents the sound in the German word Tee and the French word thé, Italian , Welsh ”. Now, the Italian word for tea is NOT : it’s . And it ISN’T pronounced te but ! The spelling for is regarded as sub-standard or non-standard and, for this reason, it’s not included in any dictionary – apart from Canepàri’s DiPI. On top of that, people who misspell and write are, for some reason, still likely to pronounce it . This is probably because this word is so common in speaking that its pronunciation is quite clear. When it comes to writing it, though, Italians may get confused, as they sometimes think that this is “not an Italian word” but an English one! Also, accent marks on Italian headwords are for some native speakers quite difficult to master, especially on monosyllabic words. 

The pronunciation te is used in Standard Italian for the pronoun te (without any accent) in expressions like Te lo dico (‘I’m going to tell you’) or in, for instance, Te vieni al cinema? (‘Are you coming to the cinema?’), where the te is a subject pronoun meaning tu (‘you’).

I very much hope all these inaccuracies will promptly be corrected by Mr Tench before the second edition of Transcribing the Sound of English comes out!


  1. Alex: Thanks for interlinking your comments on PT's workbook with mine

  2. Any comments on the recordings? For example, how do the transcriptions without syllabic /l/ compare with the recordings for the same words? It'd be a bit of a blunder if the author claimed something to be the case in the text but gave counterexamples in the recordings.

  3. Jack Windsor Lewis has sent me this comment:

    "I shd like to point out that, tho I agree that pronunciations like ɡəʊsː for ɡəʊsts or ˈsa(ɪ)əntɪsː for ˈsa(ɪ)əntɪsts "are quite common in rapid, colloquial RP speech", they have largely given way in mainstream not necessarily rapid style GB in recent decades to a reduction of the cluster produced by maintaining the /t/ but eliding the final /s/ so that the singular and plural are not distinguisht.

    Altho I find it no longer appropriate to tell students that they'd better avoid schwa in words like apple in favour of a version with syllabic /l/, I still personally react to forms like /dɪdənt/ as sounding to someone of my (elderly) generation as rather babyish. I have to try to suppress this feeling because GB-speaking fre·nds do it!"

    Thank you, Jack!

  4. @Paul Carley:

    I've been trying to listen to the recordings. The problem is they're a total mess and don't quite correspond to the example words provided in the book.
    It's like looking for a needle in a haystack!

  5. Many of the IPA errors I'm sure were introduced by the typesetters; I did see an earlier ms and these were not present.
    Bitter experience shows that typesetters of IPA (even when you've used an IPA font) can screw up almost anything. Then, when you return proofs you've sweated hours over, they correct some stuff and put new errors into things they got right the first time! If you don't get the chance to keep chasing new proofs until they get it right, then errors like these in Paul's book get through.
    In our 'Phonetics of Communication Disorders' we insisted on 4 or 5 proofs of selected pages - amazing what the typesetters did: I assume they did not know the Latin alphabet leave alone IPA