MED Magazine - Issue 21 - July 2004

A wish beyond words
by Richard Cauldwell

Preamble

These days, dictionaries do so much more than they used to: information on frequency, information on formality, and providing example sentences are just three of the 'new things' that dictionaries do. But my favourite 'new thing' is the soundfile. Favourite, because until recently the information about the soundshape of a word was given only by phonetic symbols. Now that they are given by recorded soundfiles, there is a listening dimension to the dictionary. This is a wonderful development, as it reduces the learner's dependence on having to create the sound with their inner voice from their (perhaps limited) knowledge of phonetic symbols.

It would seem churlish to ask for more, but this is what I am going to do. My wish is that these soundfiles should be a little more helpful in relation to the teaching of listening — that they represent everyday speech a little bit more accurately. But how best to do this?

Listening's lag

I think we would all agree that learners' listening skills often lag behind their other skills. Even when highly competent in writing, reading and speaking, advanced learners often struggle to understand everyday speech when they first encounter it. One learner, Ying, describes the problem thus: '...sometimes when a familiar word is used in a sentence, I couldn't catch it. Maybe it changes somewhere when it is used in a sentence' (Goh 1997, p366).

Ying's analysis of her difficulty focuses on words-in-sentences. Can dictionary soundfiles help? Not yet, and here's why. The soundfiles given in dictionaries are citation forms these are very useful reference points for teaching pronunciation but they are rare in everyday speech. These citation-form soundshapes are an important component of the 'familiar' forms that Ying refers to: they are those that are given in pronunciation keys in classrooms, textbooks and dictionaries.

Citation forms are isolated, with little or no vowel reduction, maximally stressed, preceded and followed by a pause, with pre-pausal lengthening, and falling tone. These forms are rare in everyday speech, which, by contrast, is streamlike: in the stream of speech, the soundshapes of words change as they blend with neighbouring words, and as decisions made by the speaker assign them different degrees of importance. Some words are relatively clear, others become almost imperceptible, almost all have soundshapes that differ to some extent from the familiar ones given in a dictionary. Only the minority have the familiar citation form. Thus the very clarity of citation forms (a virtue in a pronunciation key) misrepresents everyday speech and is an obstacle in the teaching of listening. Let us look at an example to illustrate this point.

Drama and that kind of thing

Consider the string of words drama and that kind of thing. By clicking on the speaker icon, and on the individual words in Example 1, you will hear the pronunciations taken from the Macmillan English Dictionary (MED) CD-ROM.

Example 1

There are many things to note about the MED pronunciations (male and female voices; different pitches) but I want to focus on the pronunciation of the words and and of. Two features of the soundfiles are praiseworthy steps in the accurate representation of speech: first, the 'd' of and is dropped; and second, the vowel quality of both words is reduced. But in all other respects, they are citation forms: they are still isolated, maximally stressed, with falling tone and pre-pausal lengthening.

These clear pronunciations misrepresent speech: of and and almost never occur in isolation, bounded by pauses with a falling tone and pre-pausal lengthening. But, as we listen to an extract of spontaneous speech, with typical representative examples of and and of it will become clear why dictionary makers do what they do.

Between words

Click on the speaker icon, then on the individual words of Example 2. This sequence of words (a speech unit) and the examples below come from Streaming Speech (Cauldwell, 2002). The three digits at the end of the speech unit indicate that they were originally spoken at over 280 words per minute.

Example 2

As you click on the individual words, note how the soundshape of a word affects its neighbours. Drama has an 'r' colouring at the end (called 'intrusive r'), the second word and has an 'r' colouring at the beginning, loses its final 'd' and the 'n' sound heavily influences the 'th' of that. The words kind of are very tightly linked together, with kind being clear, and of being much less clear. If it were possible to separate them, the resulting soundfile for of would sound like a very faint puff or grunt.

Example 3 presents each pair of words in the speech unit, with a brief commentary on what happens to the soundshape of each word in relation to its neighbour. The point to notice is that the soundshape of each word merges with and/or changes the soundshapes of its neighbours. This is one way in which citation forms, familiar forms, are defamiliarised in the stream of everyday speech.

Example 3

Wouldn't it be nice to have information about how the familiar soundshapes of words are defamiliarised in normal speech? This would help with Ying's listening problem.

A wish beyond words

It is obviously not a practical proposition to present soundfiles of puffs and grunts as guidance on the pronunciation of words such as and and of. It is not even the case that they would be true representations of these words in everyday speech because, as we have seen, 'truth' would require information that is merged with the soundshapes of neighbouring words (the 'n' of and influencing the 'th' of that). It is my belief however that it would be useful to give, in addition to the citation form, soundfiles for the word in its most common environment.

My wish therefore is to have information about the soundshapes of words when they are streamed together with their most common collocates. And the easiest way to do this would be to have recordings of the example sentences, which are likely to contain the most common neighbouring words. This may be impractical at present with the limitations of space on a CD-ROM, but as dictionaries move to a DVD format (as they surely must) the provision of extra soundfiles of the example sentences will become a real possibility. A churlish wish perhaps, but increasingly possible, and it would greatly aid learners such as Ying, who want more information about the soundshapes of words-in-sentences to help them become more effective listeners.

References

Streaming Speech: Advanced Listening and Pronunciation for Learners of English, R.T. Cauldwell (Birmingham, UK, 2002). For a sample unit, please visit: http://www.fab24.net/examples/streamingspeech.htm
'Metacognitive awareness and second language listeners', C. Goh (ELT Journal Volume 51, Issue 4, October 1997, pp361-369)


Copyright © 2004 Macmillan Publishers Limited
This webzine is brought to you by Macmillan Education