You might have heard that the British Library wants people to record themselves reading Mr. Tickle. And you’re probably wondering why. We have an extensive collection of recordings that capture a variety of 20th-century voices. You can hear several examples in our Evolving English exhibition and on our Voices of the UK CD. The exhibition, which opened on 12 November, gives us a unique opportunity to capture contemporary English voices from around the world. This will enhance our collections and enable researchers now and in the future to describe accents of English and track changes in our pronunciation.
So how do we compare accents? UCL phonetician John Wells devised a system of using a single word to refer to groups of English words that contain the same vowel sound. He calls these word-groups lexical sets and uses a key word, such as BATH to identify them. The BATH set refers to the pronunciation of the vowel in the word bath and other words that share that same vowel, such as laugh, ask and dance. It’s a particularly important vowel set in the UK as we can divide speakers into northerners and southerners depending on their pronunciation of BATH words. We also associate different BATH variants with American English, Australian English and South African English. On the whole consonants vary less dramatically than vowels and are more resistant to change over time, but several local pronunciations exist. The <th> sound in words like think, for instance, has recognisable alternatives in London, the Republic of Ireland, Jamaica and on the Indian subcontinent and pronunciation varies considerably among non-native speaker groups.
Professor Wells recommends 24 lexical sets as the basis for comparing accents of English worldwide. Perhaps the simplest way to compare accents is to record speakers pronouncing a word list that includes all 24 vowels and a complete set of consonants. Although this method is useful, experience shows that it doesn’t always produce the most ‘natural’ pronunciations. When reading a list we often use a more careful, precise speech style. Pronouncing words in isolation also reduces the prominence of connected speech processes like intonation and rhythm, which are such a characteristic feature of many accents. Spontaneous conversation provides the most ‘authentic’ language and the most detailed data, but researchers need hours of recording to ensure each speaker under observation uses every vowel and consonant sound.
A good compromise is to use a reading passage of continuous prose. Previous studies have used iconic texts, such as ‘The Parable of the Prodigal Son’ or ‘Aesop’s fable of the North Wind and the Sun’. For the VoiceBank we decided to use a children’s book – step forward ‘Mr. Tickle’. Dumbing down you might think? In fact a familiar text with intentionally straightforward language allows speakers of all ages to read confidently, including non-native speakers of all abilities. Pilot studies with the text confirm it also encourages a relaxed, informal speech style. Crucially, the Mr. Tickle text here contains almost all of Wells’ lexical sets, a comprehensive set of consonants and several connected speech processes. This will ensure the recordings we collect will be valid for research now and in the future. Our exhibition includes a representative sample of voices from all parts of the English speaking world. We’d be tickled pink if your accent, your town, YOUR English became part of our story.
Somehow I missed this when it first appeared.
It’s a really useful description of a research methodology for investigating variations in pronunciation.
(Got that, Helen?!)