A reference guide to the spelling of Australian Aboriginal languages in early sources
This is an open-source collaborative reference book initiated by language centres, scholars, historians, native title anthropologists and speakers of Australian Aboriginal languages. The project operates on zero funding but you can contribute by adding your knowledge to the wiki.
Piers Kelly, Dr Harold Koch, Dr Patrick McConvell Dr David Nash Professor Jane Simpson Dr Peter Sutton
Words from Australian Indigenous languages that were recorded in old sources can be very difficult to interpret. This is because standard transcription conventions for Aboriginal languages did not emerge until the 20th century. Before this time, individual scribes applied their own ad hoc spelling systems that often failed to adequately represent the languages as they were spoken. To make matters worse, many of these scribes were not even internally consistent in their transcriptions of Aboriginal words. As a result, many important ethnographic, linguistic and historical materials are now accessible only to specialist scholars who have knowledge of Aboriginal sound systems and long experience in interpreting spellings in archival sources.
Much hinges on the accurate decoding of old sources. For obvious reasons, those wishing to learn and reclaim their language need to know the correct forms if their revival efforts are to be meaningful. In native title cases the proper attribution of a local placename, tribal name, named ancestor or social category may be crucial to the outcome of the claim. Sound interpretation of spelling is also vital for reconstructing the history of a language, including knowledge about how the language changed over time, the migration patterns of its speakers and of past interactions with other groups.
In the recent past, inexpert interpretations of historical spellings have had real-world consequences. Research on a Western Australian native title claim misidentified certain placenames which, according to one linguist, compromised the evidence. Some Arrente street names in Alice Springs have been written using historical spellings that are so confused that Arrernte speakers themselves cannot identify them. In Canberra the traditional owners are divided between the 'Ngunnuwal' people and the 'Ngunuwal' people, a political dispute that has its origins in varied interpretation of historical records.
Our project aims to develop a 'Historical spelling decoder'. This will be a digital reference resource that will allow users to select the name of a scribe (eg, from a drop-down menu), type a word into a blank field, and produce an instant 'reconstitution' of the word in both the International Phonetic Alphabet (IPA) and, where available, the modern standard orthography preferred by today's language owners. In the Beta phase we will provide reconstitutions for the following historical scribes: Daisy Bates, FJ Gillen, Ursula McConnell and Baldwin Spencer. More may be added as the project progresses.
Phase one: A linguist researcher will be engaged to identify and gather all existing published and unpublished materials by linguists who have already made careful reconstitutions of transcriptions by these authors. This linguist will then assess the general principles of each scribe to determine which are generalisable. Eg, does Daisy Bates always use 'ch' for the sound 'tj'? Is a word-final 'i' supposed to rhyme with 'eye'?
Phase two: A computer programmer, preferably with a background in linguistics, will encode these principles into the 'Historical spelling decoder' and create an html interface for it.
This guide is intended for use by those reviewing historical records. No prior training in linguistics is assumed. Each chapter is dedicated to the interpretation of an individual scribe and is prefaced with general remarks about their conventions. Following this is a glossary of core words used by the scribe and a transliteration into a consistent and accepted modern orthography of the languages in question.
Even today, poorly trained researchers are known to adopt their own spelling conventions for Aboriginal terms and placenames, sometimes with disastrous or embarrassing consequences. The present volume is not designed as a 'how to' manual for proper transcription but simply as a guide for interpreting historical sources. For information on how to represent spoken Aboriginal words in writing, please contact the relevant regional language centre for advice and materials.
DISASTROUS CONSEQUENCES OF MISINTERPRETING OLD SPELLINGS comment from Tony: Hi Piers,
I think it is a good idea to do as you suggest: i.e. a guide to the orthographic peculiarities of individual authors. However, and I guess this is something most anthros are already aware of, I'm against the idea of converting things to 'phonological' renditions and similar. I think often its best to struggle through with the original data - often individual researchers will make a different interpretation, and these can be equally valid. There's nothing worse than getting the clinical sanitised version of some old data and not having the option of drawing your own conclusions,
Tony Interesting - I'll have to think about this one. I know of examples where anthropologists have drawn false conclusions about conceptual links through sloppy transcriptions (one with Warlpiri 'blood' and 'ceremony' [yawulyu and yalyu, I think] by Francois Dussart, and one by Ken Maddock with Dalabon 'liver' and 'subsection' [they're differentiated by retroflexion]'. But that's probably not 'public record' enough for what you need. I'll think about this more but David and Jane would also be good sources Best Nick try this link - I've moved the post to my wordpress blog with something like the original date.
By the way, Yolngu often leave off the underlining for retroflection so maybe I shouldn't have been dogmatic about buŋguḻ (and I can't remember how it's spelled in Burarra but I have a suspicion it's bunggul).
There were a cuple of problems with the Bardi claim data:
. species misidentification, in particular, including transcriptions resulting in two different names for the same tree.
. place name spelling obscuring which place is meant. This was really serious and I think compromised the evidence. There are a lot of places with similar names (due partly to dialect differences) so for any given place it was impossible to tell whether it was a new name or a misspelling. . arguments about dialectology which were based on folk linguistics AND dodgy data
There were some other things too, but that's enough for now..
Margaret Carew: 3.14 ‘Anglicisation’ of spellings, and the backwash into contemporary interpretations
In a letter held in the NT Archives Vern O’Brien mentions an Arrernte name ‘Para Lurknga’, proposed for a park , and suggests that this could be anglicised to ‘Paralurkna’ (O’Brien 1959). Anglicisation involves representing the Arrernte word as text in a way that conforms with English spelling norms and pronunciation, without applying a consistent phonological orthographic representation based on phonological principles (Turpin 2004). In the case of Para Lurknga this involves replacing the velar nasal /ng/ with an /n/, and making it one word rather than a phrase. Incorporation of Arrernte words in this way is a marker of their status as loans into English, even though they may retain their reference to Arrernte words and concepts (Myers-Scotton, 2006:219-20). As street names these words are then used by the whole population with phonological integration as English words.
Apart from marking a word as a loan, Anglicised spelling also means that in many cases the origins of the word are obscured, and there is evidence that this can lead to a semantic reinterpretation. For example, Veronica Dobson has queried the English meaning for arenge provided by the NT Place Names Register and in Petrick (2005), suggesting that it is actually arengke ‘dog’, which is a synonym for akngwelye . According to Veronica, Arrernte people believe that Arunga Street refers to arengke ‘dog’, not arenge ‘euro’ (V. Dobson 2011, pers comm 19 May) . In this case it is likely that phonological integration has been mediated by English spelling norms. As a street name this word is customarily pronounced in Alice Springs English as a-rang-ka. This pronunciation reads orthographic ng as an intervocalic velar nasal stop cluster (a phonotactic VNSV pattern ), and approximates the Arrernte pronunciation of arengke. The Arrernte pronunciation of arenge ‘euro’, is closer to a-ra-nge, with ng read as an intervocalic velar nasal (VNV).
Erumba Street in Braitling suffers from a similar ambiguity. Petrick tentatively attributes the source as ilwempe ‘ghost gum’ (2005:64), however Veronica Dobson feels quite strongly that the name is actually derived from yerrampe ‘honey ant’, and that this is the meaning of this street for Arrernte people. According to Veronica people might say, in English, ‘oh you’re going back to honey ant street’ (V. Dobson 2011, pers comm 19 May).
In the case of some Anglicised forms, the original meaning is as good as lost. Tmara Mara circuit was mentioned above, as having derived ultimately from the phrase apmere mwerre ‘good place’. In the spelling, the prenasalised stop /pm/ is rendered as an unusual consonant cluster /tm/. The use of orthographic T is here an abstract sign of prestopping (being predictably homorganic): allowing for the unfamiliarity of prestopping to non-speakers of Arandic languages. It is consistent with the way that Spencer and Gillen spelt the word apmere (1927:90), and may have been the source for Battarbee when he established the name. The customary pronunciation of this street name now is ta-ma-ra ma-ra, with vowel insertion to break up the ‘Tm’ spelling cluster. It is my impression that, while apmere mwerre is a familiar Arrernte phrase, even for incidental learners of Arrernte, few people link this to Tmara Mara Circuit. However, it is widely recognised as being derived from Arrernte. This suggests that some names have quite empty symbolic content, and function mainly to index the Arrernte language as their source. It is an open question as to what extent other Arrernte words that have been deliberately chosen to name streets carry symbolic content within the wider language community. Even for those people who make the effort to learn the meanings of these words, I would suggest that their main semiotic function are indexical, as pointers to the existence of the Arrernte language and its speakers.
- 1 Daisy Bates
- 2 FJ Gillen
- 3 Ted Strehlow
- 4 Thura-Yura languages
- 5 Ursula McConnel
- 6 Gerhardt Laves
- 7 Nyamal
- 8 Ridley
- 9 The Marawara Language of Yelta
- 10 Lancelot Edward Threlkeld
- 11 Norman B Tindale
- 12 AW Howitt
- 13 Lorimer Fison
- 14 Baldwin Spencer
- 15 A.P. Elkin
- 16 Alfred Radcliffe Brown
- 17 Arthur Capell
- 18 R.H. Mathews
- 19 Captain James Cook
- 20 Robert Brough Smyth
- 21 W.E.H. Stanner
- 22 George Augustus Robinson
McGregor, William. 2008b. Daisy Bates’ documentation of Kimberley languages. Unpublished presentation to the Society for the History of Linguistics in the Pacific (SHLP) Inaugural Conference, ANU, 1st August 2008.
Nash, David. 2008. Deriving Mrs Daisy: the Bates records of Australian languages. Unpublished presentation to the Society for the History of Linguistics in the Pacific (SHLP) Inaugural Conference, ANU, 1st August 2008.
Pacific Linguistics. * Hoogenraad, R. (2000) The history of Warlpiri education and literacy in the context of English and alphabetic writing. Unpublished manuscript.
The Gamilaraay (Kamilaroi) Language, northern New South Wales — A Brief History of Research Peter K. Austin SOAS, University of London
The Marawara Language of Yelta
Aboriginal history, volume eight, page 56 Luise Hurcus
Lancelot Edward Threlkeld
Lancelot Edward Threlkeld
Norman B Tindale
Alfred William Howitt
- Walter Spencer.jpg
Walter Baldwin Spencer
Alfred Radcliffe Brown
Captain James Cook
Robert Brough Smyth
George Augustus Robinson
Harold has materials