Recent years have seen a significant upturn in research activity into Indigenous song language. We now have teams of researchers including anthropologists, linguists and musicologists engaged in longer-term song language projects. This paper seeks to outline that activity and suggest lines of inquiry for future research. To that end, I pose a series of questions to raise what are regarded as some of the more important issues to be addressed now and in the future.
Song texts may be short but they are often tricky (Barwick 2006a:57)
Twenty years ago Margaret Clunies Ross (1987:3) remarked in the introduction to an important collection of papers, Songs of Aboriginal Australia, that: 'Only within the last decade have linguists begun to pay much attention to the subject [Aboriginal song]'.
My intention is to recall the state of the art twenty years ago, then review what has been learnt about Australian Aboriginal song language over the last twenty years and to suggest some ideas for best practice in the documentation and analysis of Australian Aboriginal song language. The following examples of song texts should clarify why the title announces 'so little to work with'.
Three examples of song texts
I present examples of song texts to demonstrate some of the special difficulties in approaching language in song. The first example is from central Australia and was cited by Barwick (2000:329):
Warumungu wakiriji larrana wakiriji larra wakiriji larrana wakirifi larra mangkkuru larrana mangkkuru larra mangkkuru larrana mangkkuru larra gloss: wakiriji 'mulga'; mangkkuru 'black soil plain'; larrana and larra 'just for song' lie vocables]
This song text consists of just four different words, two of which are so-called vocables, that is, expressions which only appear in song and have no meaning. This leaves just two contentful words: wakiriji, 'mulga'; mangkkuru, 'black soil plain'. The interpretation of the song is likely to present a rich range of associations, but teasing out these interpretations is painstaking work and, clearly, one cannot come to such a song text cold and hope to glean anything approaching the richness of these song traditions. Indeed such apparently elliptical songs receive interpretations from singers during performances so that a better understanding of the song text can be built up by collecting interpretations from many performances over time.
The second example is a Murriny Patha song from the Top End of the Northern Territory (Barwick, Blythe et al. 2006) Djanba song 11 (transcription and gloss by Joe Blythe, 1 November 2005):
ngarim thakuny marramarda nyinirda karrirndurtuy bee_Sp left progeny that_place 3sS. 3_stand. Exist.dig.redup
'Kardu Ngarim, son of Thakuny, is digging holes at that place.'
With five meaningful words this song text has a little more to work with, but once again most of the following interpretation is not directly recoverable from the song text in isolation. The accompanying text (Barwick, Blythe et al. 2006) reads:
This text refers to a particular djanba ancestor, named Kardu Ngarim ('Bee Man'--kardu being the nominal classifier for humans). With the animate nominal classifier ku, ngarim is also the name of a species of native bee with a yellow body, also known as ku tiriwun. Kardu Ngarim is the son of another djanba, Mayamunggum, nicknamed Thakuny 'left-handed'. Ngarim is digging at a particular place. Most of the species of native bee found around Wadeye make ku tjithay 'honey' (or 'sugarbag' in Australian Aboriginal English). Ku ngarim sugarbag bees are one of the main totems for the Kunybinyi clan estate, which is where djanba ancestors reside and where Dimirnin clanspeople return when they die. Each clan estate has several ngakumarl ('totems', or 'dreamings'), which are associated with particular nguguminggi 'creation sites' within the estate. …