Global Utilities

Global Navigation

Linguistics Program

Workshops

Semester 2, 2007

Workshop on the Language of Poetry and Song

Australian Linguistics Society Conference 2007 Adelaide

September 26-28

The “Language of poetry and song” was a two-day interdisciplinary workshop held at the annual Australian Linguistics Society Conference held at the University of Adelaide, September 26-28, 2007.

The workshop was an opportunity for researchers wor30 October, 2007uistics, ethnomusicology or anthropology to explore linguistic features of song/poetry with others working in this field.

Papers addressed issues such as

• How do musical form and linguistic form interact?
• Are there restrictions on the types of grammatical structures found in song/poetry?
• Can words be shortened or extended to fit song metre?
• What sorts of special vocabulary do they contain? If there are 'fillers'
• what are their metrical purposes?
• How do the intended meanings of song/poetry differ, or go beyond, the meanings in speech?
• How do the meanings and broader significances of song/poetry relate to the local social context?

Abstracts

Meiki Apted, Linguistics & Applied Linguistics, University of Melbourne

The Mirrijpu Song Language

The Mirrijpu (seagull/silver winged gull) songs are a series of traditional songs from Goulburn Island, NT, Australia, recorded by Isabel Bickerdike (University of Melbourne) and Linda Barwick (University of Sydney) as part of the Iwaidja language documentation project (a cross-disciplenary documentation team includes Nicholas Evans, Linda Barwick, Bruce Birch, Joy Williams and project manager Sabine Hoeng). After being dreamed (or ‘received’) from the spirits, the Mirrijpu songs have been passed down from father to son, and are said to be in “Mirrijpu” or “spirit language”. Although the lyrical text of these songs is highly significant and meaningful, and in most cases appears to be stable (not improvised), it has no translation. This is a common feature across many traditional Australian song repertoires (see for example Dixon 1980; 1996, Marrett 2005, and Walsh 2007) , with anything from short segments to entire songs being composed of untranslatable linguistic material. The typical (and sometimes necessarily) approach is to categorise such material as “fillers”, “vocables” or “song words”, with perhaps a brief comment that such material could be based upon borrowed, archaic, made-up, or improvised material. Given the extensive use of non-translatable linguistic material in a number of Aboriginal Australian song traditions, further investigation seems merited. The purpose of this paper is to explore the nature of the Mirrijpu song language, particularly in terms of phonetics and phonology. An overarching question of the paper is the interaction between musical and linguistic form, and after identifying the basic phonemic inventory, and considering the phonotactics of the song language in relation to findings for spoken languages of the region (ie. the languages of the composers and singers), issues such as the interaction of meter and rhythm, intonation and melody, and phonotactics and sonority will be considered. The presentation will include audio examples of key songs, and possibly also audio-visual footage of performers discussing the origin of the songs.

References
Dixon, R.M.W.(1980). The languages of Australia. Cambridge: Cambridge University Press.
Dixon, R. M. W. (1996). Dyirbal song poetry : the oral literature of an Australian rainforest people. St. Lucia, Qld.: University of Queensland Press.
Marrett, A. (2005). Songs, dreaming, and ghosts : the Wangga of North Australia. Middletown, CT : Wesleyan University Press.
Walsh, M. (2007). Australian Aboriginal song language: so many questions, so little to work with. Abstract of Paper given at the Australian Languages workshop, Pearl Beach, NSW. 17 March 2007. http://www-personal.arts.usyd.edu.au/jansimps/Ozlgs-program-2007.html
See also http://azoulay.arts.usyd.edu.au/mpsong/.

Linda Barwick, Sydney Conservatorium of Music, University of Sydney

“A dance like this didjeridu corroboree but without the didjeridu":
commonalities between djanba and wangga/lirrga song texts at Wadeye (NT)

The repertory of some 100 Murriny Patha djanba songs composed and performed at Wadeye (Port Keats), in the Daly Region of northwest Australia since the early 1960s is superficially similar to Central Australian and Kimberleys songs. The singers are accompanied only by percussion (clapsticks played by senior performers, and body percussion of various types performed by the ensemble). The texts are isorhythmic (repeated text strings are always performed with identical rhythmic patterning), and in many cases exhibit the AABB couplet text repetition pattern that is widely found throughout Central Australia and the Kimberleys. The mixed-gender performance group, with men and women alternating in leading the descending and tonic repetition sections of the repeating melodic cycle is characteristic of Kimberleys song styles such as junba and balga.
On closer examination of the songs, however, we can see several structural characteristics that set djanba songs apart from Central Australian and Kimberleys styles. There is little flexibility in setting text to melody: with negligible exceptions, each melodic section contains exactly the same number of text lines presented in exactly the same configuration of repetitions and set over exactly the same parts of the melody (for this reason, I classify the songs as 'strophic' rather than 'cyclical'). The corpus of some 100 songs exhibits a wide variety of text repetition patterns within each verse, the majority of them 'asymmetrical': AABBB (djanba 29), AAbAAA (djanba 11), AABCCC (djanba 14) are just a few examples. Finally, the descending melodic sections (performed mainly by men at the beginning of each verse) and the tonic repetition sections (performed mainly by women at the end of each verse) are frequently differentiated by different clapstick accompaniment patterns, and always by significant changes in dance movements.
These characteristics (strophic structure, variety of asymmetrical text patterns within the verse, and variety of clapstick patterns with concomitant dance changes), while seemingly anomalous when seen from the perspective of Central Australian and Kimberleys styles, are normal features of the didjeridu-accompanied public dance song genres of northwest Australia (wangga, lirrga, and kunborrk). Jack Sullivan's statement, quoted in my title, points to a keen awareness of these structural characteristics on the part of knowledgeable elders from the neighbouring Kimberley region. The birth of djanba in Wadeye took place alongside the creation of new repertories of wangga and lirrga by various Marri language speakers at Wadeye, and it is also possible to trace various thematic similarities in the subject matter of songs across the three genres. This paper will examine several examples of djanba and compare them to wangga/lirrga song texts to demonstrate the argument outlined above.

Claire Bowern, Rice University

Linguistic Elements of Bardi Ilma.

I review some linguistic features of Billy Ah Choo’s ilma, a public song cycle dreamt in about 1965 and frequently recorded over the last 40 years. Data are 8 performances of the ilma made between 1966 (for Alice Moyle) and 2003, when I discussed the language of most of the verses in detail with the niece of the composer. The ilma is based around the life, death and post-death actions of Little Wiggan, particularly the time he was carried out to sea on a raft while hunting turtle.
Included here are comments on differences between song and spoken language in phonology (e.g. high vowel centralisation, clash in song rhythm and lexical stress), in morphology (e.g. the lack of tense marking on verbs), in lexicon (the use of ‘special’/‘flash’ and archaic words in song language) and in semantics.
I conclude with some discussion of ilma and narrative style and the varying painting of imagery in song and speech. A very characteristic feature of ilma is what we could call telescoping -- each verse is a description of an aspect of an event in the narrative which stands symbolically for the episode. For example, in Nawanarda “lightning” the ilmiidi describes the lightning as the storm mounts, the tide turns and the protagonist is carried out to sea.

Dr Chantal Crozet, ANU School of Language Studies

French songs as social and political discourse

This paper argues that in France the success of rap music and hip-hop culture is directly linked to France's long tradition for using popular song writing and interpreting as a mean to giving voice to voiceless social groups in the public arena. Based on the analysis of a selection of songs from the troubadours period to the latest hip-hop and slam (urban poetry) productions, the paper shows how the centrality given to text (rather than music) in French songs across centuries and genres provides valuable language/culture data for understanding the nature of social and cultural change in France.

Aaron Corn (Department of Linguistics, University of Sydney) and Neparrnga Gumbula

Following the voices of ancestors

Yäku (Names) are the most important hereditary property held by the Yolngu (People) of NE Arnhem Land other than country itself. They mark the identities of individuals and land-owning groups, record the original observations of ancestors, and largely comprise the esoteric languages traditionally deployed by appointed Yolngu Elders in ceremonial and diplomatic contexts.
This presentation will explore the role of hereditary Yolngu Manikay (Song) as the root medium through which Yäku and their esoteric meanings are expressed and taught using Gupapuyngu Yolngu recordings directed by Neparrnga Gumbula at Djiliwirri in 2004 and 2005. It will demonstrate the vital role of Yolngu Elders in maintaining traditional knowledge and esoteric discourses that can only be taught through Manikay, and the continuing centrality of such endeavours to Yolngu cultural survival and social cohesion.

Georgia Curran, Australian National University

Understanding the meanings of Warlpiri songs

Many researchers (Strehlow 1971; Merlan 1987; Marett 1994) have pointed out that the language used in Aboriginal songs is often unintelligible to a person who does not have some degree of specialised ritual knowledge. It has been suggested that this is a way of confining this powerful knowledge to its holders and those they authorise. The process of obtaining knowledge of these oral traditions also involves learning dances or actions and designs as well as the language and musical aspects needed for the performance. In this paper, I will examine verses of a Warlpiri women’s song cycle to highlight some of the special features present in this song that need to be learned to understand it. Firstly I will note words which differ from everyday speech either due to their specialised ritual nature or the use of archaic terms. Secondly I will point out instances of extended meanings of words in songs. And, thirdly I will note the heavy use of metaphor. Lastly, I will show how much understanding a Warlpiri song relies on the exegesis that is obtained at another point in time from when the song is sung. Thus for a Warlpiri person (and equally for a linguist or anthropologist) a full comprehension of the significance of songs is not confined to the learning of words, dances and designs in their ritual context but also involves gaining additional knowledge in other social contexts.

References
Marett, A. (1994). "Wangga: Socially Powerful Songs?" The World of Music 36(1): 67-81.
Merlan, F. (1987). Catfish and Alligator: Totemic songs of the western Roper River, Northern Territory. Songs of Aboriginal Australia. M. Clunies Ross, T. Donaldson and S. Wild. Sydney, University of Sydney Press. 32: 142 -167.
Strehlow, T. G. H. (1971). Songs of Central Australia. Sydney, Angus and Robertson.

Allan Marett

The effect of rhythmic modes on the structure of texts for the wangga songs of Jimmy Muluk (Belyuen NT).

In a recent paper (Barwick 2005-6), Linda Barwick showed how when lirrga songs from Wadeye are presented in different rhythmic modes (that is in different tempi and with different relationships between clapstick-beating patterns and text), the metrical requirements of the rhythmic modes can precipitate changes in the structure of the song text. In my book on wangga (Marett 2005), which is a related genre to lirrga, I was unable to observe this type of relationship between rhythmic modes and texts, mainly because my sample presented very few examples of songs that were sung in different rhythmic modes. Jimmy Muluk, who is the subject of this paper and whom I did not study in my book, was recorded a number of times by Alice Moyle in the 1960s. From her recordings it is clear that Muluk had a much more complex practice with regard to rhythmic modes than the more recent wangga performances studied in Marett 2005, and that this practice is akin to that studied by Barwick for lirrga. This paper will demonstrate the extent to which Muluk’s presentation of songs in a number of different rhythmic modes affects the structure of song text and will compare the results with Barwick’s work in order to broaden our understanding of the effect of rhythmic mode on text structure for songs of the Daly region.

Barwick, Linda. "Marri Ngarr Lirrga Songs: A Musicological Analysis of Song Pairs in Performance." Musicology Australia 28 (2005-6): 1 – 25
Allan Marett. Songs, Dreamings and Ghosts. The Wangga of North Australia. Wesleyan University Press (2005). ISBN 0-8195-6617-9.

Stephen Morey

Syntactic variation in different styles of Tai Phake songs

The Tai Phake are a community of around 2000 people living in Assam State, India. Their language has 6 lexical tones, an isolating profile, and the majority of words are monosyllabic. VO structures are the most unmarked, but pragmatic factors mean that many variations in constituent order have been recorded.
At least eight different styles of songs, from traditional to modern, are used by the Tai Phakes. The styles vary according to a number of parameters: some require a fixed number of syllables per line, some give the poet/composer considerable freedom to vary the length of lines. Some styles require the lexical tones to be realised (Morey 2007), others, it seems, do not. Some styles show a fairly strict adherence to VO structures, while others allow much more freedom of constituent order. Some styles allow for considerable elaboration in expression, including repetition and considerable compounding, while others are much more economical.
This paper will explore the relationship between the various song styles and the syntactic profile licensed by each style. We will suggest that styles allowing freedom in some of the parameters will be constrained in some of the others.

References
Morey, Stephen. 2007. ‘The realisation of tones in traditional Tai Phake songs’, paper presented at the North East Indian Linguistics Society conference, Gauhati University, Assam.

Cassy Nancarrow

Interaction of form and meaning in Lardil burdal songs

This paper explores the interaction between form and meaning in Lardil burdal songs of Mornington Island in two parts:
1. The relationship between linguistic form and meanings ascribed to particular songs, and
2. The interaction between musical form and linguistic form, including phonological differences between song language and spoken language.
Relationship between linguistic form and ascribed meanings
Many words in burdal songs are not identifiable as words in spoken language. There are some song words that have meaning only within the burdal context. A few words are borrowed from neighbouring languages.
Burdal text is highly elliptical in its grammatical structure, usually being a collection of individual words with little identifiable syntax. Meanings of songs tend to be learnt, rather than interpreted through analysis of song text.

Interaction of musical form and linguistic form Burdal is usually sung in a series of melodic descents. A song is potentially recognisable by the rhythm of its text, but never by its melody alone. A series of songs may have the same melodic pattern, changing only the length to fit in with different texts.
Words in burdal songs have the same phonemic structure as spoken Lardil. Vowel distribution overall is similar, however interesting things happen with vowels at the ends of song legs . Occasionally vowel sounds are inserted as 'fillers' on the off-beat at the start of a melodic descent.
Consonant distribution in general is similar to that of spoken Lardil, however there is a slight tendency towards the use of the peripherals /k/, /ng/, /b/, /m/ and /w/, and for all the legs to start with the same sound. Stress patterns in spoken language can be reversed.
Rhythmic and textual elements of a burdal are generally fixed in relation to each other, whereas melody and pitch are changeable. Where different rhythmic patterns occur they tend to be significant, such as representing different parts of a story.

References
Nancarrow, Cassy. 1999 Burdal Lardil: Creation and communication of meaning in Mornington Island songs. Unpublished Honours thesis, Department of Linguistics, University of Sydney.
Ngakulmungan Kangka Leman. 1997 Lardil Dictionary: A vocabulary of the language of the Lardil people, Mornington Island, Gulf of Carpentaria, Queensland, Ngakulmungan Kangka Leman: Mornington Shire Council.

Renae O’Hanlon, University of Queensland

Language and identity in Australian Hip Hop songs

Given Hip Hop’s focus on lyrical delivery, its performances are a rich source of data for sociolinguistic research. Its growing popularity in Australia guarantees a substantial pool of songs performed by Australian youth which can be analysed for their sociolinguistic features. However, those wishing to investigate face a relative dearth of existing research by linguists on the topic of language in Australian Hip Hop, or indeed in other music genres. Thankfully the situation seems to be changing (eg, O'Hanlon 2006; Pennycook 2007), though emerging work necessarily has as its starting point the solid base of US Hip Hop language literature. The current study also utilises US work (such as Alim 2006) to provide a point of comparison between the sociolinguistic features of Australian Hip Hop songs and those of the US counterpart. Australian Hip Hop is found to engage with similar issues of identity construction and performance through language. It features layers of local, national and global identities, interspersed with cultural and ethnic identities, all of which are played out through the linguistic behaviour of the performers. Additionally, the centrality of rhyme to Hip Hop and its typical style of vocal delivery (verse delivered in a speech-like manner to a strong accompanying rhythm) lead to interesting linguistic phenomena, including stress shift. Other factors, dictated by the genre, lead to a strong presence of poetic devices – in particular, simile. It is hoped that linguists will take up the challenge of boosting the level of research into these fascinating issues in the Australian context, thus contributing to a global pool of Hip Hop language studies.

References
Alim, HS 2006, Roc the Mic Right: The Language of Hip Hop Culture, Routledge, New York.
O'Hanlon, R 2006, 'Australian Hip Hop: A Sociolinguistic Investigation', Australian Journal of Linguistics, vol. 26, no. 2, pp. 193-209.
Pennycook, A 2007, Global Englishes and Transcultural flows, Routledge, Abingdon.

Isabel O'Keefe, PhD, Linguistics & Applied Linguistics, University of Melbourne

The Oyster-catcher bird speaks: The language and music of Kaddik-kaddik (Oyster catcher bird) songs

This paper examines the linguistic and musical characteristics of the Kaddik-kaddik songs of the late Frank ‘Kaddik-kaddik’ Namarnangmarnang, recorded by Taylor (1981) and Hiatt (1966) and makes some observations about their interpretation and significance in the current Kun-barlang speech community. The Kaddik-kaddik songs are a set of Kun-barlang love songs, which belong to a genre of public love songs found in Western Arnhem Land. The Kaddik-kaddik songs are predominantly sung in everyday Kun-barlang language, however, some of the songs also include syllables, lexical items and whole sections sung in ‘song words’ or ‘spirit language’. This paper outlines the way in which everyday lexical items and phrases are incorporated into the song meter, and the way in which stress patterns differ between the spoken and sung versions of lexical items and phrases. A comparison is also made between the phonological structure of ‘song words’ and every day words and between the way ‘song words’ and every day words are incorporated into the musical structure. The ‘song words’ are analysed in terms of their metrical purpose within the song as well as analysing their cultural significance and the issues they pose for current speakers’
The use of everyday language in the Kaddik-kaddik songs also mirrors that of everyday spoken language, particularly in its use of “referential indeterminacy and ambiguity”, also noted in by Garde (2005) for kun-borrk songs. This paper will also examine this characteristic of ambiguity and the way in which current Kun-barlang speakers make conjectures in interpreting these texts, often with great enjoyment.

References
Garde, M. (2005). The language of kun-borrk in western Arnhem Land. Musicology Australia, 29, in press.
Hiatt, L. (1966). Songs of Arnhem Land. Canberra: Australian Institute of Aboriginal Studies, AIAS 6.
Taylor. (1981). Music of Maningrida master tape 1-4.

Tonya Stebbins, Linguistics Program, La Trobe University

Mali Baining Songs: An overview of the repertoire.

Abstract
This paper provides an overview of the different song types encountered in Mali society. The Mali community are part of the Baining ethnic group and have their homelands on the Gazelle Peninsula in East New Britain Province, Papua New Guinea. The paper includes samples of a number of traditional styles: Day dance songs, Night dance songs, Feast songs, Children's songs, Hymns, Choruses, Popular music. Distinguishing features of each style will be noted with particular reference to linguistic differences between the styles. Although much attention has been paid in the literature to Baining reticence around the secret/sacred meanings of the Day and Night dances, the repertoire as a whole, and the musical and linguistics aspects of the Day/Night dance songs in particular, have received no attention to date. There are interesting commonalities across the repertoire that are identified in the paper. Possible avenues for future research of the corpus will also be considered.

Nick Thieberger, Department of Linguistics & Applied Linguistics, The University of Melbourne

The place of songs in the documentation of South Efate, Vanuatu

Vanuatu is renowned for its linguistic diversity, with over 110 languages, as well as Bislama, a Melanesian Pidgin lingua franca. Singing has played a central part in traditional (or kastom) life (Crowe 1981) as well as in the Christian practices - almost universally adopted by the population of Vanuatu since missionary activity started in the nineteenth century - and in stringband performance. This paper will use a performance of one song - Ririal - as a point of discussion. The song displays epenthetic phones to satisfy rythmic requirements, and is in an archaic form of the language. It was recorded in 1998 and forms part of a corpus of recordings in South Efate transcribed with time-alignment and archived with PARADISEC. This corpus contains several what I will call 'traditional' songs as well as church services with hymn singing and a stringband performance of songs in current South Efate language. The methods for building this corpus will be discussed in this presentation.

Reference
Crowe, Peter. 1981. Polyphony in Vanuatu. Ethnomusicology 25, 419-432.

Dr Myfany Turpin, University of QLD/ Charles Darwin University

A metrical approach to the analysis of Aboriginal song

It is often asserted that the language of traditional Aboriginal songs differs greatly to that of everyday speech, yet few researchers have demonstrated exactly how they differ. Two recent articles by Garde (forthcoming) and Ford (forthcoming), working on songs from different areas of northern Australia, state there are metrical requirements that force syllabic alterations to words. Garde states that the singers are aware of the song's metrical requirements, however he notes that further analysis is necessary to identify what these requirements are. This paper takes up the challenge of identifying metrical requirements in a set of Kaytetye songs from Central Australia.
Metricality, or meter, can be thought of as an abstract pattern to which text is set according to certain principles. The abstract pattern often has a set number of syllables in a line with rules stating how words are set to this pattern. For example a stressed syllable must match a strong position in the meter known as accentual meter (as in many English nursery rhymes and limericks). The nature of syllables varies across languages, and so metrical genres of different languages differ in what aspects of syllables they highlight. Syllable weight, stress and consonant/vowel patterning are different aspects of syllables that languages can contrast when setting words to meter.
In this paper I show that the 90 different text lines of the Kaytetye songs are all derived from five different rhythmic units, grouped in various ways to create 15 different metrical patterns. I then identify rules for setting words to the metrical units which are based on congruence between the number of words and rhythmic units; and between the number of syllables in the words and rhythmic units. Unlike many English genres, stress does not play a role in setting words to the metrical pattern in these songs.

Michael Walsh, Linguistics, University of Sydney

A polytropical approach to the ‘floating pelican’ song: an exercise in rich interpretation of a Murriny Patha (northern Australia) song.

This is one small piece of a research project carried out by a team of musicologists and linguists 2004-8 on song traditions at Wadeye, NT and its environs. Drawing on Peter Toner’s analysis (2001) of Dhalwangu Manikay songs from the Yolngu area which in turn relies on Paul Friedrich’s theory of polytropy (1991) an attempt is made to present a detailed account of a Murriny Patha song in the Djanba style (http://azoulay.arts.usyd.edu.au/mpsong/songs/songs/djanba.html).
In Friedrich’s model there are five kinds of trope: image (including visual, sound and musical images); modal (concerning the expression of mood: sadness, joy etc); formal (including repetition and parallelism); contiguity (including deictics and part-whole relations); analogical (including metaphor and personification). Cross-cutting these five horizontal continua Friedrich poses a vertical continuum in each category – at one end greater abstraction (more cryptic and less redundant) and at the other end greater concreteness. Each song text can be thought of as having a characteristic trajectory through polytropical space.
This detailed matrix for interconnected description strikes me as particularly suitable for Murriny Patha poetic traditions as I will seek to demonstrate for the so-called ‘floating pelican’ song: an exercise in rich interpretation.

References
Friedrich, P. 1991. “Polytropy,” in Beyond Metaphor: The Theory of Tropes in Anthropology. Edited by J. W. Fernandez. Stanford: Stanford University Press. Pp**
Toner, Peter 2001 When the echoes are gone: a Yolngu musical anthropology. PhD thesis, Australian National University.

Content Approved by: Head of School
Page maintained by: Administrative Officer
Last Updated: 30 October, 2007

CRICOS provider number: 00115M

Humanities and Social Sciences