Category Archives: Laves

Laves update – August

Work has continued on the typing, translation, and annotation of the corpus. Stuff is happening. Really cool differences are coming up, which are being tagged.


Laves update

I’m going to try to give regular updates for what’s being achieved on grants, since it’s a good way for me to do record-keeping and it may be useful for others to see what grant tracking looks like.

We started in June. Since then, here’s what’s been achieved:

  • I bought a field laptop.
  • Linda ran all my Laves photocopies with notes through a sheet-feeder scanner and broke the files into individual PDFs of the text.
  • We have an annotation scheme (we had this as part of the proposal, but it’s been a bit refined).
  • Linda’s typed more than 1/4 of the texts.
  • I’ve been correcting the transcriptions and annotating and (preliminarily) translating them.

Laves’ texts: transcription

Working on texts like those in the Laves corpus means making a lot of choices about what to type. For example, we could produce a facsimile of the originals and annotate that, or we could produce something that’s a bit further from the originals.

We’ve gone in the latter direction for this project. I deliberately decided to standardise the texts to current Bardi spelling, rather than using Laves’ spelling. I have also not retained his punctuation, and have edited his notes.

This might be a bit controversial. It does mean that it’s hard to use the texts ‘on their own terms’, and without going back to the original texts it won’t always be clear what is Laves’ analysis and what is mine. (Heidi Kneebone has made some comments about using old materials on their own terms, in relation to the Reuther Diyari materials of South Australian languages. )

It’s worth remembering that the text collection is field notes. It was probably never intended to be treated as is, without further work (and there is evidence from the materials that Laves himself started to return to the materials in 1932 or 1933). There were a couple of reasons for not just typing exactly what Laves wrote down.

First, Laves doesn’t have free translations, and his interlinear glosses are a) hard to follow and b) sometimes incorrect. Moreover, he has very minimal punctuation. Therefore there is already need for the editor to make judgements about things like phrase boundaries in translation. This is being done in consultation with speakers. Luckily there are a number of first position and second position particles in Bardi which are a good guide to where to break clauses.

Secondly, the main reason for doing this project (apart from investigating the linguistic differences between the modern language and that recorded by Laves) is to make a repository of texts which will be of value and of use to Bardi people. Since not many people speak Bardi these days, if the text collection is going to be useful, it will need to be keyed to the Bardi dictionary. That means either spelling words in the texts the same way as they are spelled in the dictionary, or providing some sort of equivalency list (which would make the process of looking up words pretty tedious, since there are many differences). I will provide a bunch of comments, both in the texts and in the introduction, but it seems a bit silly to reproduce known errors which will make the texts less easy to use. This (along with the work involved) is also the reason that I decided not to produce a text with Laves’ Bardi on one line and modern Bardi on the next. The differences are mostly quite minor.

There are two main areas of minor phono-differences. One is retroflection. Laves consistently writes many more apicals as retroflex than any other writer on Bardi. That in itself is interesting, since he is fairly accurate in other areas (like in vowel length) does not change the transcription much as he gets better acquainted with the language. There is a lot to be said about retroflection in Bardi, the extent to which it’s phonemic, and the role it plays in verb inflection (a set of contrasts that was all but dead in the last speakers, or that I wasn’t hitting on the right way to elicit). However,

The other area is in the number of vowels regularly recorded. Laves uses a backwards epsilon as well as a, i, u and o (and long o, long u, long i and long a). Phonemically, there is no long o: it’s either ow or maybe owu (Laves writes bow or o for these) or a phonetic variant of u:.

There are other differences, in lexicon, dialect, morphology, and syntax, and we are representing those. For the Jawi texts, they are printed pretty much “as is” apart from the phonological standardisation. For the other ones, there are some differences which make grammatical differences for current speakers. So, the current plan is to print the ‘normalised’ text, with extensive notes about the differences between what was written and what appears in the edition. This seems an acceptable compromise between respecting the integrity of the original text and making a book that current speakers and owners of the language can use.

Laves project web page

I’ve started a page for the Laves materials project. It’s still a draft at this stage but more material will go up as the project progresses. You can reach it here:

The Laves Texts (precontact narratives)

In the most recent round of Documenting Endangered Language Grants (offered as a joint program by the NSF and NEH) I received a grant to work on the Laves Bardi texts. Here’s the blurb:

Detailed linguistic fieldwork on the Indigenous languages of Australia did not begin in earnest until the 1960s. One of the few earlier sources is the manuscript collection of Gerhardt Laves, an American who went to Australia in 1928 and spent approximately 18 months doing linguistic research. He worked intensively on six languages, one of which was Bardi. Bardi is now spoken by approximately 25 elderly people in northwestern Australia. Laves’ Bardi records are extraordinary. They were written at a time before the full impact of European settlement had caused extensive language loss in Australia’s North-West and record a language and culture which has since changed significantly. Although the texts are legible and the language accurately represented, the materials are very difficult to read without excellent eyesight and a good knowledge of Bardi. This NSF/NEH-funded project will bring together a team of linguists and Bardi community members to work on the texts to produce an annotated edition, under the direction of Dr Claire Bowern. The linguists will digitize the texts, word process them, and provide rough translations on the basis of Laves’ notes and the Dr Bowern’s knowledge of the language. They will then work with Bardi speakers to refine the translations, discuss the grammar of the Bardi language, and work out the context of the stories. The result will be a book of the texts, with translations, annotations, and discussion.

This is an important project, linguistically and culturally, for Bardi people, linguists, and for science more generally. Linguistically, the Laves texts represent the earliest accurately recorded materials for Bardi, and preliminary work indicates that there have been subtle but numerous changes in the language over the last hundred years. We know little about variation and change in small linguistic communities like this, and within Australia the Bardi materials provide a rare opportunity to get accurate longitudinal information about language change. This in turn allows the study of change in languages with very complex inflection, which is both understudied and of wide relevance in linguistics. Culturally, the texts are very important. They provide information about pre-contact traditional law, mythology and everyday social interaction, and are a very detailed record of a way of life no longer practiced. Most of this knowledge is still held by the oldest people in the Bardi community but has not been written down, and work on describing and explicating the events in the Laves materials is a wonderful opportunity to study pre-contact culture. Finally, the texts themselves are an important language learning and cultural resource for Bardi people themselves, and repatriation of these materials is very important. The materials will likely figure prominently in future language revitalization programs. With so much of the world’s linguistic heritage in danger, this project provides a model of how early materials may be used to benefit both local communities and science.

Linda (aka Tulugaq) is also working on this project and is doing much of the typing and deciphering of the abbreviations that Laves uses (of which there will be more..).

We’ve started work and are translating and typing the texts at present. The end result will be a book of the texts and a documentation.

I’ll be blogging on various things that come up in the text from time to time.