The Language of Song: Some Recent Approaches in Description and Analysis

Click to increase image sizeClick to decrease image size

Acknowledgements

We would like to thank Jane Simpson for providing helpful comments on an earlier draft of this paper.

Notes

¹Workshop on the Language of Song held by the Australian Linguistics Society conference at the University of Adelaide on 27–28 September 2007. We thank all the participants at the workshop for their insightful comments about issues involved in working on song, and especially for the discussion during the final session which arose from the question of how to analyse songs, given that by their very nature they require expertise in language, music and specific cultural knowledge.

²Here we draw upon the work of Dell and Halle (2008).

³Following Dell and Elmedlaoui (2008) we use the term ‘enculturated listener’ to refer to the singers or listeners of songs within a culture, in much the same way one uses the term ‘native speaker’ when referring to linguistic competence.

⁴See Wild (1987) for a discussion of this in Warlpiri.

⁵Our use of ‘text’ means the syllables as they are uttered; they need not be a written version or performance of a written version.

⁶Rumsey, for example, refers to such genres in Papua New Guinea as ‘chanted tales’ (2001, 2005).

⁷See also Clunies Ross (1986), and Koch and Turpin (2008) for a recent overview of the language of Aboriginal songs.

⁸Note that a ‘line’ of song does not refer to the same thing as a ‘line’ of speech.

⁹Many Arandic songs, for example, are subject to alliteration on the final syllable of each line whereby the first couplet of a verse ends in a pattern which is reversed for the second (final) couplet of the verse, thus creating a palindrome e.g. –a, –ay; –ay, –a (Turpin 2007b).

¹⁰See Patel (2008: 96–179) for a detailed discussion of the hierarchical organization of rhythm in music.

¹¹For examples of rhythmic modes see Barwick (2003), Barwick et al. (2007), Marett (2005) and Treloyn (2007, 2009).

¹²See Fabb (1997) and Foley (1997) for an overview of the formal features of many types of verbal art.

¹³Arandic songs have a large proportion of their text in a neighbouring linguistic variety (Turpin & Green in press).

¹⁴Strictly speaking poetic meter refers to the fixing of the length of lines by counting some aspect of linguistic form, and this counting is associated with control over poetic rhythm or the placement of word boundaries, or some other aspect of the line's form (Fabb 1997; Fabb & Halle 2008). Dell and Elmedlaoui put text-setting in the purview of metricality (2008).

¹⁵Sanskrit song texts also adhere to a four-mora pattern (Deo 2007).

¹⁶This is also the case in many central Australian songs—see Turpin (2007a) for a discussion of a song style where the temporal unit encodes a syllable, and thus a long note (two temporal units) encodes two syllables.

¹⁷Rumsey (2007) considers how lexical tone is overridden by melody in Tom Yaya sung narratives from Papua New Guinea; see also Patel for a broader consideration of this issue (2008: 217).

¹⁸As an example, Central Australian and Kimberly songs often prefer verbs in line final position (Strehlow 1971: 213; Austin 1978: 531–532; Keogh 1990: 74; Ellis 1998: 435; Turpin 2005: 335).

¹⁹Rumsey identifies formulas in songs from Papua New Guinea (2001: 194, 2005, 2007). In Australian Aboriginal songs there is a tendency for formulas to occur at the ends of lines; see, for example, Barwick (2003: 32) and Koch and Turpin (2008: 180).

²⁰Compare Fabb's discussion of Iatmul songs of Papua New Guinea where verbs have more general meanings when they occur in songs (1997: 7); Koch and Turpin for examples in Aboriginal songs (2008: 180).

²¹This term was coined by Sylvia Wright (1954), in an article called ‘The Death of Lady Mondegreen’, in Harper's Magazine, November 1954.

²²Marett discusses particular songmen in Arnhem Land who have the ability to interpret songs that are in spirit language (2005: 43).

²³In relation to Lirrga songs of Northern Australia, Barwick too finds that the performer has a choice in which text meanings to highlight in any one performance (2003: 82). Ellis and Barwick note that these multiple text meanings ‘are never in conflict with each other; and … are in fact mutually supportive’ (1987: 44–45).

²⁴See, for example, analyses by Austin (1978: 529), Donaldson (1984: 247), Merlan (1987: 146), Hercus (1994: 311), Tunstill (1995: 63), Dixon and Koch (1996: 26), Hercus and Koch (1999: 81), Marett (2005) and Walsh (2007). Marett conjectures that this is a significant reason why Aboriginal songs are difficult for outsiders to appreciate (2005).

²⁵Compare with proverbs in English, or with the Apache place names discussed by Basso (1996).

²⁶Merlan's analysis of how people explain songs from the Roper River area (1987: 146), which contrasts with casual conversation and narratives, is reminiscent of this approach.

²⁷Strehlow (1970: 94) argues that the performance of a song series acts as a mnemonic for the location of important sites. When Central Australian people travelled the country on foot, the location of such places would have been essential knowledge for survival.

²⁸See http://firstaid.about.com/od/cprbasics/qt/07_cpr_tips.htm, October 2008. See also ABC News Available at: http://abcnews.go.com/GMA/OnCall/story?id=3837780&page=1, November 2007, last accessed 25 November 2008. Note that many other songs, such as Queen's Another One Bites the Dust have the correct beat for CPR, however due to their less suitable lyrics these have not been promoted.

²⁹Consider the Monty Python skit where a song text ‘Here comes another one’ is set to a different musical idiom following a request for the song to have more of a particular emotional effect, from their album Contractual Obligation, 1980 (Charisma Records).