Production and perception of vowels in New Zealand popular music
An acoustic comparison of sung and spoken vowels for three New Zealand singers investigates the phonetics of pronunciation in popular music. The singers recited the lyrics to their songs and recordings of their sung vocals were also obtained, creating a dataset of paired sung and recited words. Interviews with the singers were conducted so that the pronunciation used in reciting could be compared with a more conversational style. Eight vowels were analysed in these three conditions: DRESS, TRAP, THOUGHT, LOT, START, GOOSE, GOAT and PRICE. As well as providing data for phonetic analysis, the interviews elicited information about the singers’ musical influences, and investigated the singers’ stances towards the use of New Zealand English (NZE) in singing. The results of the comparison of singing and speech reflect the singers’ various stances to some extent. Overall, however, there are strikingly few cases where pairs of sung and spoken vowels have similar pronunciations. The predominance of ‘American’ vowels in the singing of all three participants, despite stated intentions to use New Zealand forms, suggests that the American-influenced singing style is the default in this context. This finding contrasts with early research on singing pronunciation in popular music, which described the use of American pronunciation in pop music as an act of identity which involved effort and awareness (Trudgill, 1983). The results presented here support the claims of more recent studies which suggest, conversely, that it is the use of non-American accent features which requires a wilful act of identity (Beal, 2009; O'Hanlon, 2006). An important consideration in the interpretation of vowel differences between singing and speech is the role played by the act of singing itself. It has been argued that there may be a general preference for increased sonority in singing (Morrissey, 2008) which would lead to the use of more open vowel sounds. This issue is explored and some evidence is found for a sonority-related effect. However, singing inherent effects like this can only explain a portion of the variability between singing and speaking. Most of the differences between singing and speech appear to be caused by social and stylistic motivations. To investigate why American-influenced pronunciation might be the default in the singing of pop music, a perception experiment was conducted to examine the phenomenon from the perspective of the listener. Participants were played words from a continuum that ranged between bed and bad, and they responded by circling whichever word they heard on a response sheet. The perception of ambiguous tokens was found to differ significantly according to whether or not the words were expected to be spoken or sung. These results are discussed with reference to exemplar theories of speech perception, arguing that the differences between singing and speech arise due to context-specific activation of phonetically detailed memories. This perspective can also be applied to the processes which underlie the production of vowels in sung contexts. Singers draw on their memories of popular music when they sing. Their use of American pronunciation in singing is therefore the result of the fact that a majority of their memories of pop singing involve American-influenced phonetic forms.