Doubts about ELF pronunciation – 1) schwa within words

I recently received a very interesting email from someone doing their PhD on the teaching of pronunciation for English as a Lingua Franca. The person in question is researching at a Spanish university and hopes that their work will ‘help Spanish speakers of English improve their pronunciation in the language, following the Lingua Franca Core’. Obviously I do too, and it is encouraging to find that somewhere out there somebody else is also interested in ELF pronunciation.

The main content of the email was a series of excellent questions of the sort that you can only make when you’ve thought about ELF pronunciation carefully. I want to use the questions as prompts for this blog, so what follows is the first of various posts on doubts about putting the LFC to work in real settings. The first question was about schwa within words:

1. Jenkins makes a lot of references to the lack of importance of accurately pronouncing schwa in weak forms (thus speakers could perfectly pronounce those words in their strong form remaining intelligible). However, does that also apply to schwa within words? That is to say, is it important for students to accurately pronounce schwa within a word or, on the contrary, its substitution for another vowel would be intelligible? For example, in the word “shepherd”, the RP pronunciation would be /’ʃepəd/. However, a Spanish speaker would say /’ʃeperd/. Putting aside the fact that the Spanish person would add an /r/ sound (which the LFC allows and even encourages), they would also substitute schwa for /e/. I’m not sure whether this substitution would result in unintelligible speech, thus it should be addressed in the classroom.

The simple answer here is that in ELF what applies to schwa in weak forms is also true of schwa within words. The impact on the listener is much the same and is related to the mismatch between the spoken and written forms of words in English. So interestingly, if a Spanish speaker of English were to say /’ʃeperd/, however much this might jar with some native-speakers ears, this pronunciation would be more intelligible in an ELF context, and not just to speakers of Spanish, but also to speakers of other phonetic languages. In fact, /’ʃeperd/ is probably the ‘default’ pronunciation of the word among ELF speakers, and in my own teaching here in Spain I wouldn’t do anything to modify the pronunciation of a learner who produces /’ʃeperd/ rather than the RP /’ʃepəd/.

Another thing that is worth bearing in mind when deliberating over such issues is the fact that not all NSs have an RP accent. A Scots speaker of English would usually pronounce the /r/, either as the RP alveolar approximant [ɹ], or as an alveolar tap [ɾ]. I have to admit that it’s a source of joy for me to hear Scottish commentators on BBC documentaries pronouncing ‘forest’ as /’fɒrəst/ (or even as /’fɔrɨst/), and it was understanding ELF pronunciation that finally freed me and my students from the onerous burden of suppressing /r/ from their English when it’s presence patently did not impact negatively on their intelligibility.

Another thing to bare in mind is that the vowel reductions given in dictionaries for RP are an idealized version of reality. If you listen carefully to NSs in England you discover that there are many occasions where the dictionary would predict schwa, but where in real-time speech this is missing and usually substituted by another short central vowel, or by another vowel altogether. It was fascinating to watch a well-known, London-born ELT expert some years ago, and to see how often he pronounced final -er as [ɐ] (which coupled with the substitution of /ð/ by /v/, meant that ‘further’ sounded like [ˈfɜːvɐ]!) with no apparent impact on his intelligibility. 

The last thing that I think we need to remember in this question is that it is the vowel reductions inside words that cause so many problems of intelligibility for non-native speakers listening to native-speaker English. The word ‘chocolate’, for example, is transcribed as both [ˈtʃɒk.^ələt] and [ˈtʃɒk.^əlɪt] in dictionaries. The transcription shows that the letter ‘a’ is pronounced either as /ə/ or /ɪ/. The second letter ‘o’ in the written form is pronounced as /ə/ or is simply not pronounced at all, as the use of superscript indicates. That is to say, the ‘o’ is so strongly reduced that it disappears altogether, leaving the spoken word with only two syllables. I suspect that both [ˈtʃɒk.lət] and [ˈtʃɒk.lɪt] will be less accessible to non-native speaker ears than [ˈtʃɒk.olat], which is what a Spanish-L1 speaker of English is likely to pronounce. It would be interesting to research this and get the necessary empirical data.

Of course, [ˈtʃɒk.olat] would sound ‘heavily accented’ to native-speaker ears, but native speaker value judgements of non-native speaker pronunciations in ELF settings are, to put it bluntly, irrelevant. What matters for ELF communiaction is intelligibility as judged by non-native speaker interlocutors, and not proximity to a given NS standard accent.

Advertisement