As a linguist who spends much time researching Australian Aborigin; languages, I have often been informed by people I have met in m travels that ‘You must have an easy job – it must be pretty simple figuring out the grammar of such a primitive language.’ If you go further and ask your travelling companions over a beer or six why they hold this belief, you encounter a number of sub-myths:
There is just one Aboriginal language.
Aboriginal languages have no grammar.
The vocabularies of Aboriginal languages are simple and lack detail; alternatively, they are cluttered with details and unable to deal with abstractions.
Aboriginal languages may be all right in the bush, but they can’t deal with the twentieth century.
I’ll deal with each of these individually below. Two of these myth are dealt with elsewhere in this book, and I shall deal with those in rather less detail.
The first white arrivals in Botany Bay came equipped with an Aborignal vocabulary recorded by Captain Cook and others in Cooktown, north Queensland but soon found this was of no more use in com municating with the owners of the Botany Bay region than a Lithuania phrasebook would be in London: Captain Cook recorded the Guugu Yimidhirr language (giving us the word kangaroo in the process), while the language of the Sydney region was Dhaaruk, only distantly related. In fact, Aboriginal Australia displays striking linguistic diversity and, traditionally, around 250 languages, further subdivisable into many dialects, were spoken over the continent. Many Aboriginal communities would prefer to count these dialects as distinct languages. If we did this, we would have to elevate this figure to about 600.
Some languages are, of course, more closely related than others. In Western Arnhem Land, for example, such languages as Mayali and Dalabon are as closely related as English and Dutch, so that ‘I will eat fish’ is ngangun djenj and ngahnguniyan djenj respectively. Others, such as Ilgar, are only very distantly related (more distant from Mayali and Dalabon than English is from Bengali, although Mayali and ligar are spoken only a couple of hundred kilometres apart), so that ‘I will eat fish’ in ligar is anyarrun yihab.
With so many languages spoken by a population of at most three quarters of a million, you can easily work out that the average language would only have a couple of thousand speakers. But, of course, people’s social universes were much larger than this. This meant that by adulthood it was normal to be multilingual; this was made easier by the fact that most people married spouses with a different language to their own, so that children grew up speaking both the mother’s and the father’s languages, as well as other languages their grandparents, for example, may have spoken. For example, my Ilgar teacher, Charlie Wardaga, learned Ilgar from his father, as well as Marrgu from older people in the area he grew up in, Garig and Manangkari from other relatives, Gunwinygu from one grandparent (and he took a Gunwin-ygu-speaking wife and frequently sings at ceremonial gatherings where Gunwinygu is the common language) and Iwaidja through living in the Minjilang community where it is the dominant language.
In the first difficult weeks when I was beginning to learn the Kayardild language of Bentinck Island in Queensland I experienced the usual language-learner’s nightmare of failing to understand most of what was said. One of my more considerate teachers, Pluto Bentinck, would help me by repeating each sentence, working his way through all possible orderings of its words: dangkaa bangaya kurrija, dangkaa kurrija bangaya, bangaya dangkaa kurrija, dangkaa kurrija ngada, and so on. Given that dangkaa means “the/a man’, kurrija ‘see(s)’, and bangaa ‘the/a turtle’, how could he put the words in any order without changing the meaning from ‘the man sees the turtle’ to ‘the turtle sees the man’?
Speakers of a language like Kayardild have this freedom because the identification of who does what is carried out by so-called case markers on the ends of words: the -ya on bangaya marks it as the object of the verb and hence the thing seen, while the -a on the end of dangkaa marks it as the subject and hence the seer. So while it is true that words can be put in any order, it does not indicate lack of grammar – grammar, as a code for expressing meaning, can take many forms in different languages, and here (as in Latin or Russian) the work is done by word endings rather than word ordering (see Myth 10: Some Languages Have No Grammar). You should be able to work out for yourself six ways of saying ‘the turtle sees the man’; see answer 1 at the end of this chapter.
This system of case endings is so efficient that it allows parts of sentences to be specific in ways that aren’t always clear in English. Consider the sentence ‘The man saw the turtle on the beach.’ Who is on the beach – the man, the turtle or both? Kayardild expresses each of these meanings differently – where it is the turtle on the beach, the ‘associative’ suffix -nurru is added to ngarn- ‘beach’, and the resultant ngarnnurru receives a further -ya to link it clearly to the object, giving dangkaa bangaya kurrija ngarnnurruya (or any other of the 4 × 3 × 2 possible word orderings). If it is the man on the beach, -nurru is used again, plus -wa to link it with ‘man’ (a cannot directly follow u, so a w is inserted): dangkaa bangaya kurrija ngarrnnurruwa, and the other orderings. And if both are on the beach, a different suffix -ki is used, giving dangkaa bangaya kurrija ngarnki, and so forth.
Not all grammars of Aboriginal languages work in the same way as Kayardild, of course – any more than English and Russian work in the same way. For example, Mayali from Western Arnhem Land is a ‘polysynthetic’ language that builds up highly complex verbs able to express a complete sentence, such as ngabanmarneyawoyhwarrgah-ganjginjeng T cooked the wrong meat for them again,’ which can be broken down into nga- T, ban- ‘them’, marne- ‘for’, yawoyh- ‘again’, warrgah- ‘wrongly directed action’, ganj- ‘meat’, ginje- ‘cook’ and -ng ‘past tense’.
Australian Aboriginal pronoun systems are in some ways more explicit than English as well. The main way of showing number in Dalabon from Western Arnhem Land is through the pronoun prefixed to the verb. So we find:
biyi kah-boninj
man he-went
‘The man went.’
biyi barrah-boninj
man they two-went
‘The two men went.’
biyi balah-boninj
man they-went
‘The men went.’
But that is not all. Another way of saying ‘the two men went’ would be biyi keh-boninj. This would be appropriate if the men were related ‘disharmonically’ – i.e. in odd-numbered generations, like father and son, or uncle and nephew, e.g. be-ko keh-boninj ‘they two, father and son, went.’ The ‘harmonic’ form, barrah-boninj, is only appropriate for people in even-numbered generations, such as brothers, spouses or grandparents with grandchildren, e.g. winjkin-ko barrah-boninj ‘they two, grandmother and grandchild, went.’
In a short article like this we can only scratch the surface, but it should be clear by now that Aboriginal grammars have plenty to engage your analytic powers.
However complicated the grammars, surely the vocabularies are pretty simple? After all, there are no words for ‘neutron’, ‘virus’ or ‘terra nullius’, so that’s three down already. Assertions like this usually take one of two forms – either the languages are supposed to have a welter of detailed words but be incapable of generalizing, or they are just said to have very general words with too few to be precise. On both counts such submyths are wildly wrong.
The fine detail and nuanced observation of Aboriginal vocabularies is so great that I will only have space to consider a few words for the natural world, though one could make similar points with terms for emotions, or smells and fragrances, or ways of moving. Many plant and animal species had distinct names in the Aboriginal languages in whose territories they are found well before they had been recognized as species by Western taxonomic biology. The Oenpelli python, for example, has had the long-established Kunwinjku name nawaran but was only identified as a distinct species in the 1960s, whereupon it received the Linnean name Morelia oenpelliensis.
To get an idea of the degree of conciseness and detail in the biological vocabulary of a typical Aboriginal language, compare the Kunwinjku kangaroo terms with their English equivalents, in the table overleaf.
In addition to the various detailed terms just given, Kunwinjku also has a general term, kunj, to cover all the macropods, i.e. all kangaroos and wallabies; in English we only have the scientific term macropod to denote this category. And in addition to these different nouns, Kunwinjku also has different verbs to describe the different manners of hopping of these various macropods – kamawudme for the hopping of male antilopine wallaroo, kadjalwahme for the hopping of the corresponding female, kanjedjme for the hopping of the wallaroo, kamurlbardme for the hopping of the black wallaroo, and kal-urlhlurlme for the hopping of the agile wallaby. This focus on identifying macropods by the peculiarities of their gait is particularly interesting in the light of recent work on computer vision programs able to identify wallaby species, which had far more success doing this on the basis of their movement than their static appearance.
| Linnean and English names | Male | Female | Child |
| Macropus antilopinus (antilopine wallaroo) | karndakidj kalaba (large individual male) | karndayh | djamunbuk ( juvenile male) |
| Macropus bernardus (black wallaroo) | nadjinem baark | djukerre | |
| Macropus robustus (wallaroo) | kalkberd kanbulerri (large male) | wolerrk | narrobad ( juvenile male) |
| Macropus agilis (agile wallaby) | warradjangkal/ kornobolo nakurdakurda (very large individual) |
merlbbe/kornobolo | nakornborrh nanjid (baby) |
Quite apart from finely classifying different entities, vocabularies of Aboriginal languages often also show the ecological links between particular plant and animal species. For example, the Mparntwe Arrernte language of the Alice Springs area, where various types of grub are an important source of food, has a method of naming grubs after the bushes where you can find them: tnyeme ‘witchetty bush’ yields the tnyematye ‘witchetty grub’, utnerrenge ‘emu bush’ yields the grub known as utnerrengatye, and you can work out for yourself the name of the grub found in thenge, the ironwood tree (see answers).
Sometimes there is no term in the ordinary language to cover certain more general categories, but special language varieties learned in adulthood and used under restricted circumstances possess the more abstract terms. The most extreme example of special abstract language is found on Mornington Island, where second-degree initiates, to become full men, had to learn a special initiation language known as Demiin, which had only about 200 words and hence needed to be highly abstract. For example, the complex Lardil pronoun system, where there are nineteen distinct pronouns in the ordinary language, is collapsed to two in Demiin – n!aa ‘group containing me – i.e. I or we’ and n!uu ‘group not containing me, i.e. you, he, she, they’ (n! denotes a ‘clicked’ n-sound, for Demiin also has special sounds not used in the everyday language.)
Languages tend to have the richest vocabulary in those areas in which their speakers have been interested long enough to develop specialized terms. In the early Middle Ages it was widely believed that only Latin had a sufficiently sophisticated vocabulary to discuss law, theology, medicine and science, but as various nations began to use their mother tongues more widely, each modern European language (English, French, German and so on) soon developed its own terms. Aboriginal languages are at a similar point today – they lack many terms, but their rich grammars give them the capacity to develop them when they are needed. (See also Myth 2: Some Languages are Just Not Good Enough.)
It is natural that Aboriginal languages should have developed their vocabularies most in such realms as the Australian biota and geography, kinship and so on and not in areas that have not traditionally been a central part of Aboriginal culture – such as financial transactions, nautical terminology or nuclear physics. However, just as English has responded to the encounters between its speakers and the Australian continent by coining new terms, such as the macropod terms we discussed above, so have speakers of Aboriginal languages responded by creating new terms to deal with the proliferation of novel concepts that contact with Europeans and with late-twentieth-century technology more generally, has brought.
Making up a new word from scratch is not a usual method of doing this in any language. Instead, the usual three methods are to build up new words from the existing resources of the language for compounding or affixation (e.g. downsize in English), to borrow words from other languages (e.g. sputnik) and to extend the meanings of existing words (e.g. surfing the net). Each of these methods has been widely employed by Aboriginal languages.
As an example of compounding, Kayardild has created the words wadubayiinda for ‘tobacco’, by compounding wadu ‘smoke’ with the root bayii- ‘be bitten’, literally ‘that by means of which the smoke is bitten’, and, for ‘car’, the word duljawinda, literally ‘ground-runner’.
Many languages have borrowed their words for days and months, higher numbers, government institutions and Western medicine from English. Often the pronunciations of borrowed words are changed to the point where their original source is not recognizable: the English word ‘hospital’ ends up as wijipitirli in Warlpiri.
Extending existing word meanings has been a common solution to the problem of coining new vocabulary for automobiles. In Kun-winjku, for example, kun-denge ‘foot’ also means ‘wheel’, kun-rakmo ‘hip’ also means ‘wheel housing’, and the term for ‘to get a flat tyre’ compounds kun-rakmo with the verb belngdan ‘to settle, as of mud stirred up in water’ to give rakmo-belngdanj ‘it has a flat tyre’ (literally ‘its hip has settled’). Combinations of compounding and extension of meaning are a common way of dealing with novel concepts – when a text on nuclear physics had to be translated into Warlpiri, for example, a new compound verb was coined to mean ‘cause nuclear fission’ by using a root meaning ‘hit’ and an element meaning ‘be scattered’. The fact that Warlpiri can now be used to discuss central concepts of nuclear physics is clear testimony to the adaptability of Aboriginal languages.
Linguists would love to have primitive languages to study in order to understand how human language has evolved. But, as I hope to have shown, Aboriginal languages certainly do not fit the bill – in fact, their complexities have played an important role in linguistics over the last three decades in extending our notions of what complex organizing principles can be found in human languages.
1. bangaa dangkaya kurrija, bangaa kurrija dangkaya, dangkaya bangaa kurrija, dangkaya kurrija bangaa, kurrija bangaa dangkaya, kurrija dangkaya bangaa
2. thengatye
Good introductory books on Aboriginal languages are Language and Culture in Aboriginal Australia (Canberra: Aboriginal Studies Press, 1993), Colin Yallop & Michael Walsh (eds.), Colin Yallop’s Australian Aboriginal Languages (London: André Deutsch, 1982); more advanced but still readable is Robert M. W. Dixon, The Languages of Australia (Cambridge: Cambridge University Press, 1980); these books have many onward references. Macquarie Aboriginal Words (1994), Bill McGregor and Nick Thieberger (eds.), contains sample vocabularies for a number of Aboriginal languages and pointers to more complete dictionaries. Kayardild examples are taken from the Kayardild Dictionary and Ethnothesaurus (Melbourne: University of Melbourne, Department of Linguistics, 1992) and the Kunwinjku examples from a dictionary of Eastern Kunwinjku being prepared by Murray Garde. You might also like to check out the world’s first fully formatted hypertext dictionary produced by Peter Austin and David Nathan of the New South Wales language Gamilaraay, on:
http://coombs.anu.edu.au:8o/WWWVLPages/AborigPages/LANG/GAMDICT7GAMF_ME.HTM