Commentary on the Linguistics of the Bardi/Jawi Native Title Claim

The full decision of the Bardi/Jawi native title case is available from the Federal Court’s web site and I have been reading it with interest. We’ll come back to Yan-nhangu here in a day or two (things are hapening here too). In the meantime, some commentary on the decision. I was curious to see what the linguistic evidence was for the claim, since I know both their report writers and Metcalfe, as far as I know, didn’t go back to Bardi country to do the report and relied on his dissertation research from the 1970s, and Clendon did his dissertation on Worrorra, from quite a way further up the coast.

I’ve copied the paragraphs directly relating to the linguistic summary below, with a comment on each. I’ve made some allowances for the report having been filtered though a non-linguist (ie French J). I’m doing this because I’ve been curious about the relationship between field linguists and native title anthropologists for a while, why for instance the work is often quite separate, and the privileged status that Land Council reports seem to be given over published research prepared over a much longer period, free (or freer) from the very pressing political issues involved in Native Title claims. I was also interested as to whether there were any identifiable mistakes or omissions in the linguistic evidence. I haven’t talked about the omissions, but there’s much more that could have been said, e.g. about place names. Native title evidence is one area where some quite theoretical and technical questions have practical implications. In this case this is especially true, as the basis for the decision to exclude the major inshore islands, including Sunday Island, was that French J was not convinced of the cultural and linguistic unity of Bardi and Jawi people at the time of European settlement, and therefore he was not convinced of the rights of “Bardi” people over those islands. That’s something that linguistics and linguistic anthropology could contribute to.

[Update since I wrote that: I’ve started going through the claimant’s evidence. There are a couple of factual points of misunderstanding through not knowing Bardi words used in discourse, or through taking definitions at face value. And they go on about ilngam fish poison and never mention banyjoord, which is the one that mainland people tended to use more. There’s some species misidentification, giido is an oystercatcher, for example, not whatever they said it was. Old man Stumpagee’s evidence was basically not used because of the difficulties understanding him, which is downright disgusting. This is what interpreters are for. – He would have been one of the main spokespeople for Jawi. And there are real issues over linguistic and cultural similarity or differences that are just glossed over. For example, there’s a focus on Bardi versus Jawi, but almost no discussion of the internal diversity of what we call “Bardi”. And there’s a very worrying parapgraphy where Nimanburu is just glossed over as a “Bardi clan” and linguistic similarities are asserted, despite the fact that the 20 mins of Nimanburu we have show numerous differences from Bardi.]

The report is in italics:

760 The applicants’ linguistic evidence was in the form of a report from Dr Christopher Metcalfe dated 13 November 2000 (X-AX). There was no cross-examination and the report was admitted subject only to one small objection. 761 A document entitled ‘A Preliminary Linguistic Report’ prepared by Dr Mark Clendon and dated January 2001 (X-AAD) was relied upon by the State. It took the form of a response to Dr Metcalf’s report. 762 Dr Metcalfe began with a brief outline of the Bardi and Jawi peoples and their languages in the literature. He referred to Dampier’s observations in 1688 and in particular his account of Aboriginal people, evidently in the claim area, running away from his vessel crying ‘gurry, gurry, speaking deeply in the Throat’. According to Dr Metcalfe this is revealing. Dampier would have, as would most Europeans from that day to this, great difficulty in pronouncing the sound ‘ng’ at the beginning of a word. The closest sound phonologically which occurs at the beginning of a word is ‘g’. The ‘ng’ sound is also produced at the back of the mouth cavity which probably led to his description of the people ‘speaking deeply in the Throat’.

Yes, this is true. I think Bill McGregor first came up with this, incidentally. It’s not just Metcalfe’s idea. Elsewhere in the native title report this is spelled “gurri”.

763 Dr Metcalfe then made the assertion that what the people were crying out was ‘Ngarri, ngarri’. For both Bardi and Jawi ngarri is a commonly encountered feared and fickle spirit being. It is responsible for many of the misfortunes which befall the people of their communities. Dr Clendon, on the other hand, said that the extrapolation of ‘ngarri’ from Dampier’s ‘gurry’ is problematic. Word-initial velar nasals such as ‘ng’ were usually not heard by English listeners. The other problem with getting ‘ngarri’ from ‘gurry’ had to do with the first vowel. If Dampier’s ‘u’ represented a short central vowel as seemed most likely, it was hard to see how he could have mistaken this for the long back vowel of ‘ngarri’. Despite the fact that the short central open vowel receives similar orthographic treatment to the long back open vowel in Bardi they are quite distinct word vowels and were not usually confused in short words in positions of primary stress by English transcribers.

It’s not true that “they usually weren’t heard” – they appear as all sorts of different transcriptions: m, n, gn, g, w, nothing. Some transcribers never hear them, others do. Yes, ngaarri has a long vowel. That’s why it’s spelled ngaarri, not ngarri. They aren’t written with the same orthographic treatment. and yes, English speaking transcribers are normally ok at getting length distinctions in initial syllables – they tend to write ‘u’ for schwa and a, ah or ar for /a:/ (for ar, remember we’re talking British English). But we assume that Dampier wasn’t sitting down, notebook in hand, with a mob of Bardi people screaming at him. That’s a pretty optimistic picture. Our intrepid explorer and scientist, despite imminent attack from spears and despite the noise of ambush calmly pulls out a moleskin notebook and notates their clearly spoken utterances to the best of his ability. Come on. They’re shouting at him. He’s loading his musket as they’re yelling. Later on he remembers what they were yelling and writes it down, but isn’t quite the Indiana Jones fieldworker we’d like him to be and so failed to note that the language had contrastive vowel length. Isn’t that much more plausible? Nice try, Mark.

Actually, there are a couple of instances of Bardi long a being mistranscribed as “u” by the Native Title anthropologists. They spell gaardga as Gurdka, for example (bloodwood, para 72). in para 73 gaalwa is galwa. They systematically fail to note initial long i – “inalabulu” for iinalabooloo, ina for iina ‘firesticks’.

Since no one seems to have canvassed the alternatives, let’s do that. They might have been shouting ngarri – “too much”. Or maybe they saw his pith helmet and said “ngaari” “white cockatoo” (yes I’m being facetious as well as anachronistic – the Bardi for pith helmet is niiwidi-nalma). Or maybe they were calling him “gaarra”, “uncle”, or asking for “gaari”, “grog”. Or saying “garra” “keep on going”. I prefer the ngaarri story. And note, incidentally, that I too completely failed to remember the exact vowel quality that Dampier transcribed when I retold the story on this blog yesterday.

764 Dr Metcalfe set out early attempts to record the languages of Bardi and Jawi. WH Bird, who was a schoolteacher on Sunday Island for some time, produced materials of value in 1914 and 1915 especially in relation to the Jawi language. The original of his 96 page account of the vocabulary, expressions, sentences and answers to questions about culture and social structure of Aboriginal people on Sunday Island is held by the Museum of Western Australia. It is the fullest early account that there is of Jawi language and culture. He referred to a number of other writers in the field and said that the most recent comprehensive survey of the languages of the Kimberley region was written by Dr W McGregor under the title ‘Handbook of Kimberley Languages’.

What, no mention of Laves? and Bird produced “materials of value”? The original manuscript must be vastly superior to the published accounts then. The Laves collection seems not to have been used at all, which is incredibly sloppy.

765 The Bardi/Jawi languages are ‘agglutinative’. They verbally ‘glue’ or ‘affix’ dependent word parts onto stems. The stem usually carries the meaning of the word or expression. The word parts attached to the front of the stem are prefixes. Those attached after are suffixes. There are also infixes which are embedded within word parts. A number of these sets of affixes are complex in themselves.

No they aren’t agglutinative, only if you have an odd interpretation of Bardi morphology. A better analogy to “glueing” the morphology together is that they “knead” them into each other. e.g. the locative of “alang” “south” is olongon, from alanga-goon. And ingirrij could come from five different verb stems, we can’t tell because the morphophonemic changes resulted in the deletion of the root. I somehow missed the infixes in my grammar. Oops. … oh, hang on, reduplication looks infictive if you look at it the wrong way and assume that the language is agglutinative.

766 Each verb can be characterised as a sentence-in-miniature. In a transitive sentence the subject is expressed by a prefix and the object by a suffix to the stem. So the word for ‘you are looking at them’ in Bardi is ‘min-j-al-a-nj-irr’. That declarative sentence is transferred into a question ‘are you looking at them’, it can be rendered ‘min-jal-ard-a-nj-irr’. Combinations of prefixes, suffixes and infixes lead to different meanings.

No, the -ard is in the wrong place. It should be minjalanjardirr “are you LOOKING at them?” It’s one of two ways of forming polar interrogatives. “Combinations of prefixes, suffixes and infixes lead to different meanings.” yes, rather like English really. re-lock-ing doesn’t mean the same as un-lock-ed.

767 The use of pronouns both as free parts of speech and as affixes within an inflected verb is highly developed in Bardi and Jawi. In inflected verbs, which operate as sentences-in-miniature, the subject is expressed by a prefix derived from the free forms of the pronouns. The verb ‘to pick up’ uses the verb stem ‘andi’. The sentence ‘I pick up – present tense – them’ will be rendered ‘ngan-andi-nj-irr’. This means ‘I am picking them up’.

Not all of them are similar to the free pronouns: mi- and joo are pretty different, as are ginyingg (< *ginya) and i-. I personally think that there’s no reason to relate the two synchronically, and such stories are rather Bopp-like (as in the IEist, not the astronomer). Pronominal affixes are reconstructible to Proto-nyulnyulan, these are not inner-Bardi developments (a point potentially of relevance here, no?) Nganandinirr, without the j sequential, would be more common in isolation.

768 Time in Bardi and Jawi verbs is conveyed by two basic concepts, namely uncompleted action and completed action. There are two imperfect tenses, namely the present and the future. There are six perfect tenses. These are the immediate past, the near past, the middle past, the distant past, the indefinite past and the remote past.

um, maybe.

It’s not at all clear to me that Bardi makes use of an imperfect/perfect distinction anywhere usefully in the grammar. -n is continuative and combines with most of the other TAM affixes, but its absence isn’t necessarily a marker of perfectivity. And those past categorisations are missing one – the fictional past, which is where that analysis comes in. There’s a past imperfect, and a future perfect. Maybe this bit is a result of misinterpretation by the judge. Hope so.

769 The wealth of terms available to describe various facets of the tides exemplifies the relationship between the Bardi and Jawi people and the natural world. Intimate knowledge of the seasons and changes through the year of climate and processes in the natural world have also led to a richness of definition and description. Bardi concepts of the season are more complex than that of a simple wet-dry pattern. There are 6 seasons distinguished largely by wind and rainfall directions and intensity, the ripening of fruits and the appearance, disappearance and ‘fatness’ of fish and animals. So ‘mangal’ is the wet season, ‘Ngalandany’ is the end of the wet and literally means ‘no fruit’. ‘Iralbu’ is the period of big or king tides. This is significant because low tides are ideal for reefing. ‘Bargana’ is the cold season when people start to light night fires. ‘Djallalai’ is a short warming up season. ‘Lalin’, the build up to the wet season, is hot and humid.

And they have lots of words for snow. That literally means “no fruit” idea comes from Smith and Kalotas and I no of no Bardi morphology or lexical process that could make ngaladany literally mean “no fruit” except polysemy. This whole para is quoted almost verbatim from Smith and Kalotas, complete with dodgy spelling. why don’t people ever respect community orthographies?

770 Plants are important as seasonal indicators. The signs of the availability or optimal condition of other food resources reflect extensive knowledge of the plants of the peninsula. Examples include such sayings as ‘when gorrgorr fruits, turtle arrive’, ‘when marulal flowers, turtle go’, … ‘urulbur is in flower, stingray are fat, and turtle eggs bilal flowers, dugong appear’. Known plant foods include fruits, seeds, roots, bulbs and corms, gum, nectar and other sweet substances and insect galls. There are many plants which are relied upon as medicines and used in dealing with rheumatism, aches, cuts, sores, itchy bites, sore teeth and gums.

That paragraph makes no sense. “and turtle eggs bilal flowers, dugong appear’.” Incidentally, and completely unrelatedly, it seems that marroolal flowers when turtles arrive here at Milingimbi. Note trill, not glide.

771 The flora is of great importance to Bardi and Jawi people for food sources. The language evidences great knowledge of the environment and its application to the well-being of the Bardi and Jawi communities. Dr Metcalfe cited an extract from Paddy and Smith, E & S Paddy and M Smith (1987), ‘Boonja Bardak Korn: all trees are good for something’ Perth: WA Museum. The report set out a number of Bardi and botanical plant names according to use, part used and preparation which were contained in the Paddy and Smith writing. 772 Dr Metcalfe then turned to the particular Bardi and Jawi words which describe the relationship of the claim group and individuals of it to their country. He identified the four regional groupings and the names of the people within those regions: (a) Olonggon (b) Baniol (c) Gulargon (d) Adiola The Jawi community at Sunday Island are referred to as Inalabulu which means ‘island people’.

No, wrong. Iinalabooloo are the people from Jayirri and Jalan. Sunday islanders are Iwanyoon. And now I need to do some serious checking because I have different names for two of these groups. These were not, as far as I can tell, a primary unit of social organisation, though, except perhaps for marriage. [although I note from skimming the public version of the secret testimony that there’s a bit more info about this, and that they get the Iinalabooloo designation right. Some confusion may have arisen between the technical designation of iinalabooloo, ie the people from Jayirri and Jalan, and the productive general work, which means ‘islander’.]

773 Dr Metcalfe was unable to estimate the age of Bardi and Jawi languages. He relied upon William Dampier’s account of his contact with Aboriginal groups and the inference that the word ‘gurry’ should have been transcribed as the word ‘ngarri’ to conclude that there was an unusual time depth to the recording of these Aboriginal languages.

I can’t comment on this because the reasoning makes no sense to me. The Nyulnyulan family as a whole isn’t very diverse. Maybe about Romance minus French. Well, hard to say. More phonological similarity, less grammatical similarity. Pre 1829, which is what’s important.

774 Language change was mentioned. Many of the words of the ancient songs are not in current use. The differences are so profound that it can be assumed that language change must have occurred over many centuries and probably thousands of years. Bardi language maintenance programs are carried out in the school at One Arm Point which had been conducted there for many years. There has been a corresponding development of language learning courses, teaching materials and books in Bardi.

But that’s maybe partly because they are part of a specialised vocabulary. You don’t hear too many English speakers using “fain” in everyday conversation. But many more were in use when Metcalfe was doing his work in the 1970s – they’re in his dictionary but remembered rather than actively used in daily conversation these days. And daily conversation’s usually in English anyway. There are archaisms in song language in Bardi too, but there’s also metaphor, technical song vocabulary and stylistic considerations.

775 Dr Metcalfe referred to Djarindjin which he described as ‘quite different’ from One Arm Point. Although there has always been a significant section of Djarindjin population who are Bardi or part-Bardi there have also been people from other parts of the Dampier Peninsula and from further afield. Despite the different backgrounds of the population there is a program in the Bard language. Once a week a group of six, mainly women community members, teach a course in Bard to two of the four classes. There are also language and culture-based projects within the school program in the area from time to time.

It’s not that different. How different is “quite different”? As different as Djambarrpuyngu and Gupapuyngu maybe, if that.

776 Both Bardi and Jawi are prefixing, non-classifying’ languages. This distinguishes them, according to Capel’s map, from a number of other tribes. Both are members of the same family of languages of Nyulnyul. Dr Metcalfe said: ‘When comparing word lists from the languages of the Nyulnyulan family it becomes obvious that Bardi and Jawi are more closely related to each other than Bardi is to Nyulnyul. The cognate differences increase with the geographical distance between Bardi and Jawi and the other member languages of the family.’

Capell has two ls. And there are shared innovations which relate Bardi and Jawi (and Bard) to the exclusion of Nyulnyul, and that separate Jawi from Bard(i). But you can’t just do that from lexicostatistics.

777 Worms, speaking about the name of Sunday Island, had said it was inhabited by Jawi who were linguistically and anthropologically related to the Bardi. An instance of the relationship between the two languages was provided in the translation of a series of traditional stories. There were a number of Jawi words with Bardi equivalents. Dr Metcalfe referred to a story of Arriana, the Eagle-Hawk. A passage from the story showed differences in significant words but these were set in the main body of the story in which about 95% of the vocabulary is common both Bardi and Jawi.

There are some sound changes that distinguish the two, and some vocab differences, and a really cool verb form which I worked out an etymology for two weeks ago and am going to keep for the LSA or somewhere.

778 Dr Metcalfe’s report was the subject of criticism by Dr Clendon. He observed however that the size of Sunday Island and other evidence, including evidence from Bagshaw, indicated that the number of people speaking Jawi as a first language was always quite small. It was far too small to support a separate and distinct semiotic code. On that basis he could reasonably infer that Jawi traditionally had dialect status with respect to Bardi, with its much greater geographical area and more numerous speakers. Considering the group status of Bardi/Jawi as against surrounding languages, Dr Clendon noted that Nyulnyul occurs to the south and Unggarrangu to the north-east. The Buccaneer Archipelago appears to be divided linguistically into a western zone adjacent to the tip of the Dampierland peninsula and an eastern zone adjacent to the north-western part of the Yampi Peninsula. The islands of the latter zone are collectively referred to as Mayala in Bardi and Jawi. Linguistically, if not socially, according to Dr Clendon, there was a clear discontinuity between Sunday Island in the west and Mayala in the east. The existence of such discontinuity in the south was not at all clear. He noted Dr Metcalfe’s claim that Bardi and Jawi are more closely related to each other than Bardi is to Nyulnyul and did not see sufficient evidence for this. Mr Bagshaw’s 1999 anthropological report had suggested that Bardi and Nyulnyul speaking people intermarried and that it was possible, if not likely, that linguistic as well as social distinctions in the region might be blurred. “support a distinct semiotic code”.

Well, Gurr-goni’s got 60 speakers and has done for ages. How many do you need, Mark? and what constitutes “separate” here? Their word for diamond python is baninybooroo, not miyaloorroo. That’s an important emblematic difference. Jawi suffered a population collapse as a result of smallpox and flu introduced from the mainland before the mission records began – I don’t think we know how many Jawi people there traditionally were. Yes, there’s a clear linguistic break, given that Unggarranggu is Worrorran, but things look a bit fuzzier socially if you take Oowini into account, and the fact that there were Worrorran speakers on Sunday Island who also spoke Bardi and/or Jawi. yes, Bardi and Nyulnyul people intermarried, but not all Bardi people, only, as far as I know, Gularrgoon/Olonggon people (but the land council has all the genealogies). Others had preferred marriages from Ardiol, others from nyanbooronony (basically, Oowini). Some of this is inferable from the Laves texts, others from genealogies, but I notice that those families are largely absent from the list of claimants.

779 Dr Clendon saw Dr Metcalfe’s section on linguistic indicators of connection with country as the most pertinent information and presenting the most compelling linguistic evidence in favour of the applicants’ case. The lists he referred to indicated not only that Bardi has a marine vocabulary but in the case of tides the Bardi speaking people had an ongoing association with the sea. It would be unlikely that people with a hunting and gathering economy would have had such a precise marine terminology unless they lived near the coast and depended upon it for sustenance. There was a very impressive body of evidence that Bardi speaking people have had an intimate association not only with the sea but with the sea in these particular regions and furthermore that the association has been a longstanding one. That impression was reinforced by the list of meteorological terms offered by Smith & Kalotas in 1985 and reproduced by Dr Metcalfe. They clearly refer to a tropical climate and locate Bardi speaking people in the tropical zone. The list of demographic and directional terms was very significant as it was indicative of speakers’ awareness of themselves as members of a larger group to which the particular settlement pattern applied. The list also indicated speakers’ awareness of the extent of the country in which the people who used the particular terminology live. The section on connection to country in his opinion was the most interesting part of Dr Metcalfe’s report and the one most likely to be of use to the applicants.

How about the lack of inland place names?

780 Dr Clendon obviously regarded Dr Metcalfe’s report as lacking substance in important respects. He criticised Dr Metcalfe’s methodology as relying far too heavily on unsubstantiated claims made by earlier writers. Typically Worms was quoted on p 18 with respect to linguistic and social closeness between the Jawi and the Bardi. But Dr Metcalfe did not point to any evidence to support his contention. Stylistically he regarded Dr Metcalfe’s report as too apt to subjective assertion.

And Clendon’s wasn’t? As far as I can see they both did a pretty sloppy job. Sorry guys, email me if I’m being unfair and I will happily turn my wrath on the summary for being misrepresentative and will publish a full and frank apology. But until then I consider this a lousy job.

781 In my opinion, what I can take from Dr Metcalfe’s report and Dr Clendon’s report is that the Bardi and Jawi language groups are closely related and that the Jawi language may be regarded as a dialect with respect to Bardi. This may be regarded as consistent with the proposition that Bardi and Jawi could be regarded as one society for the purposes of a native title determination. It is not however, as will emerge later in these reasons, conclusive of that proposition.

Yes, nowadays. But there’s evidence that they were earlier more differentiated. And dialect status isn’t necessarily indicative of social ties, it’s much more complicated than that.

782 Certainly the evidence brings out the long-standing connection to country reflected in the detailed references in the languages to aspects of the environment. That detail is indicative of the accumulation of knowledge over a lengthy period of time. That aspect of the linguistic evidence supports, in my opinion, continuous connection of both Bardi and Jawi people with the country of the claim area since before 1829. It also supports the inference that the immediate offshore marine environment was an important element of their country.

Well, if you’ve read this far you see they got the answer right anyway. Good show.

Incidentally, the anthro evidence looks much better. They have a booroo list, including some I couldn’t get. Badly spelled. Why on earth do these anthropologists do no linguistics? Not even enough to transcribe reasonably accurately the important information they are noting down? Don’t they own a copy of the Bardi dictionary anywhere in the KLC? Are there any cultural antrhos who read this blog? Please tell me about this! There are probably interesting cultural theory implications that we could draw from it.


Comments are closed.