A mathematical approach torecovering the original Australian

Aboriginal languageChris IIIert

acknowledgementsAcknowledgements to Stuart Campbell of the University of WesternSydney, as initial supervisor, and to Bob Hodge and to VladimirDimitrov of the University of Western Sydney for their ongoingsupervision and valuable feedback on numerous things that saw theproject through including the successful College of Arts conference of2007. Derek Abbott of Adelaide University also suggested that AndrewAllison and Peter Cook could check results and also organsise an

interfaculty seminar which they did. Michael Organ the chief archivistat Wollongong University found rare materials, Keith Vincent Smith ofMacquarie University located previously unknown wordlists fromStonequarry and Argyle and Camden, Jim Smith supplied a Gundungarawordlist, Ed Duyker of Sydney University supplied a Tasmanianwordlist, Jim Barrett provided information and John Laurent of GriffithUniversity also helped. Thanks go to Daniela Reverberi for softwareand computer support and patient assistance with layout. And also thanksto the late artist Ainslie Roberts (passed away 1993) and the 19thcentury engravers W. Hatherell, W. Macleod and the photographerJ.W. Lindt, whose artwork is reproduced along with mine in theAppendix. Special thanks also go to Jenny and Phoebe Sajkovic ofWollongong, Jean Stewart of La Perouse (passed away 2009), and AllanR. Carriage of Wollongong (passed away 2011, just before the thesis

came out), and all the others who are steadfast friends.

Regarding thesis titled,

,,A Mathematical Approach to Recoyering the original Australian Language ", submitted to the

University of Western Sydney for examination,

I ceftif that this thesis does not incorporate any material previously submitted for a degree or a diploma

in any university; and that it does not contain any material previously published or written by any other

person except where due refe made in the text.

tl sie*a -Christopher Roy Illeft. 6 April 201 1

Copyright © Chris Illert 2013 of all pages up to 114, including sketches by C. Illert. Copyright

of woodblock sketches by Ainslie Roberts belong the Anslie Roberts estate. Engravings by the

19th century engravers, W. Hatherell, W. Macleod, and the photographer J.W. Lind are out of

copyright but should be properly cited. Normal copyright protocols apply to journal articles

chapters 1, 2, 3 and, 4, pages 115 to 256. If citing the journal articles, which form part of this

thesis, researchers should cite the articles in the normal way with the journal’s original page

numbering. Normal protocols and courtesies of copyright and scholarship apply.

Chris Illert contact:

Address: PO Box 595

Moss Vale. NSW. 2577.

Australia

Email: illert@1earth.net

abstract

This text is submitted as a thesis by publications. It consists of four articles already published in the Journal of Applied Statistics, making up sequences of an argument, preceded by an Over-arching statement. The thesis applies mathematical concepts and reasoning to aspects of Aboriginal languages and in this way throws new light on some problems that have hitherto proved intractable for Aboriginal linguists. Mathematical forms of analysis have not previously been much used in mainstream linguistics in general, or applied to Aboriginal languages, with the major exception of the work of George Zipf (1949), whose application of Power Laws to language phenomena has influenced researchers in many other fields while being ignored in Zipf’s own home discipline of Linguistics. The thesis uses Power Laws allied to other mathematical ideas and operations, including Lagrange forms, van der Waals effects, Huygens principle and Snell’s law, to illuminate basic aspects of Aboriginal languages. Mathematical methods can provide new ways of treating data and drawing conclusions, and produce a revolutionary new picture of the original forms of the early language. They can be used to trace major processes of change over the 60,000 - 70,000 years currently estimated as the time Australian Aboriginal people have lived in Australia. This thesis shows how mathematical analysis can be a powerful tool and resource for linguistics. It is able to reconstruct a proto-form of Aboriginal language from a much greater time-depth than linguists have believed is possible for any language. This takes the scientific study of language closer to the probable time when human language itself first emerged.

Uur: i\____-_Y-J

twinklinq

gul:i:;hininlt

0ur: I ,-------Y-

twinklin(t

drrtn -0:i .\_--_Y________.

star

-0uI

Jnara mul-muru: n\------y-----J \--\|/---J\-------y----------l

uhova thc - Eurlh

amidst

wonder

minin i - nindi\----Y r-J "\- - ,

what are - you

bu -gun: g-,,

extremely high

/-.-----\

Jrura -bura: bura -fin\____y_________--l

dark ness

sketch by Ainslie Roberts

contents A mathematical approach to recovering

the original Aboriginal language

over-arching statement: prologue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i methodological basis . . . . . . . . . . . . . . . . . . . . . . . . ii the argument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv limitations and ways ahead . . . . . . . . . . . . . . . . . . . . . vii

Appendix : protoAustralian orthographic & phonetic basis . . . . . . . . . . . . . . . . . . . . 1 occurrence of consonants . . . . . . . . . . . . . . . . . . . . . . . 8 octahedral symmetry and the occurrence of vowels . . . . . . . . . 12 data mining - analysis or synthesis? . . . . . . . . . . . . . . . . 22 words ã*** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 words Ö*** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 words Ä*** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 words �Ï*** . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

chapter 1: agglutination & fusion C. Illert, 2003. Journal of Applied Statistics 30(2): 113-143 . . . . . . . . 115

chapter 2: entropy maximization signature C. Illert & A. Allison, 2004. Journal of Applied Statistics 31(1): 73-104 . . . 147

chapter 3: linguistic boundaries C. Illert, 2005. Journal of Applied Statistics 32(6): 625-659 . . . . . . . . 179

chapter 4: phonological evolution C. Illert, 2006. Journal of Applied Statistics 33(9): 989-1030 . . . . . . . 214

glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 typographical errors . . . . . . . . . . . . . . . . . . . . . . . . . 258

picture by engraver W. Macleod, 1888

i

over-arching statement

prologue

This text is submitted as a thesis by publications according to the regulations of the University of Western Sydney. Following the terms of these regulations, four published articles making up the sequences of an argument have been submitted, accompanied by an over-arching statement. In the present case this prologue plus its appendix, outlining the building blocks for the study, together form the over-arching statement. This leads to an unconventional format for the over-arching statement, in terms of the regulations. The prologue outlines the argument and rationale of the set of articles, but at 3000 words it is shorter than indicated in the regulations. However the appendix, which contains basic arguments as well as the basic set of roots, in total more than 30,000 words. Thus the ‘over-arching statement’ serves the main function of that part of the thesis, but is either much shorter or much longer, depending on how this part of the text is defined. All four articles have been published in the Journal of Applied Statistics as a linked set. They are reproduced as published, in the thesis that follows. There are special difficulties of formatting with this work as thesis, flowing from the software requirements for producing mathematical symbols both in the original articles and in adapted versions for the sake of the thesis. It has meant that any errors in typography, sometimes introduced during the publication process, are intractable for correction, and the accompanying text of this prologue and its appendix, which together form the over-arching statement, has been difficult to change in response to comments and criticism during the process of examination. However, I trust that readers will be able to follow the argument, even with the difficulties of format. The thesis applies mathematical concepts and reasoning to aspects of Aboriginal languages in order to throw new light on some problems that have hitherto proved intractable for Aboriginal linguists. Mainstream linguistics, applied to Aboriginal languages and more generally, has not used mathematical forms of analysis to any great extent, even though the work of George Zipf (1949) has clearly pointed the way. Zipf’s application of Power Laws to language phenomena has been taken up and developed further by mathematicians, and applied to many other

ii

phenomena in many other disciplines, but largely ignored in Zipf’s own home discipline of Linguistics. This thesis will show that Power Laws can illuminate basic aspects of Aboriginal languages. This concept can be combined with or complemented by a number of other mathematical concepts. This combination of concepts can then provide new ways of treating data and drawing conclusions, to produce a revolutionary new picture of the original forms of the early language. It can trace major processes of change over the 60,000 - 70,000 years currently estimated as the time Australian Aboriginal people have lived in Australia. methodological basis The appendix to the prologue sets out the ground for the four specific published studies that follow, applying mathematical analyses to different aspects of the form and development of Aboriginal languages. It combines detailed, specialist descriptions and analysis of forms of Aboriginal language with the development of mathematical arguments that are also advanced. The result is a text that cannot be easily followed by the majority of its intended readers, yet this double difficulty is unavoidable, given the aim of the thesis to combine the two areas of knowledge, mathematics and Aboriginal linguistics. To deal with this problem I will set out here in lay language the basic methodology, which is developed more fully in the prologue to the appendix. The appendix contains essential materials and premises for the work that follows, but in greater detail than general readers might wish, at the outset of their journey. The appendix begins with the challenge that has proved a barrier for previous linguistics studies, a belief that data for Aboriginal languages in south-eastern Australia at the time of contact are so unreliable as to be unusable. Yet these languages are 200 years older than languages spoken today, many of which were subjected to a variety of influences, and these south-eastern languages were located in parts of Australia furthest from the presumed point of entry of Aboriginal people. They were less exposed to influences from visitors from the north, and hence likely to be closer to the original forms. If linguists were able to access and analyse these languages they would be a uniquely valuable corpus.

iii

The thesis uses mathematically-based data mining methods to propose a practical solution to this problem. It shows how a simple phonological system can be recovered from this apparently fragmentary and corrupted data. This consists of 35 historic word-lists plus 9 early census-lists. This material was recorded by English speakers with no knowledge of Aboriginal languages and not even the rudiments of linguistic training. Yet this body of data, in this imperfect form, is sufficiently extensive for this purpose. Statistical analysis of this can be used to find a pathway through the different orthographies in the different sources, in order to reconstruct the likely forms of words recorded in these lists. From this analysis it is possible to propose a basic phonological system of 18 consonants and 4 vowels to make up a basic system for many Aboriginal languages. The resulting system has a good fit with different systems proposed by leading linguists such as Dixon, Yallop and Eades. This analysis is then given a practical use, to provide an orthography to be able to normalise the many different forms of spelling that plague the early records and make this data otherwise so difficult to use. The analysis then makes a leap, to see if it is possible to combine the repertoire of sounds with an equivalent repertoire of meanings. The basis for this exercise is the set of 6672 Aboriginal words plus 445 place names, a total of 7117 words from around 8 distinct languages from south-eastern Australia. Using this data plus power law analysis a smaller set of hypothetical roots can be proposed, which can then be reduced even further by applying the principle of octahedral symmetry. The principle of octahedral symmetry is a powerful hypothesis about the way this set of roots is structured. It produces a manageable small number of proposed roots and defines their internal relationships, in a pattern which can be identified from this linguistic data set and analysed further. The bulk of the appendix then describes and illustrates the basic patterns of sounds and meaning that make up this set of basic roots. Ultimately the forms elicited by these methods turn out to be the basic words from all language zones from north of Sydney, to well west of Wagga Wagga, to west of Melbourne, down south to Tasmania (which was a continuous piece of land in all but “recent” times). This preliminary finding provides strong support for the hypothesis that these roots are common and structural in historical forms of Aboriginal languages, and that the generative principles outlined are indeed active and productive.

iv

From a linguistic point of view, these two achievements, the proposal of a single underlying phonology and the hypothesis not simply of the set of basic roots but also their underlying generative morphological principles, are so original that the thesis would not need to do more, and so controversial that more empirical work would be asked for to convince the many doubters. Some empirical data is supplied, but since this is primarily a theoretical thesis in applied mathematics, the remainder of the thesis takes this initial input and elaborates and explores some implications, using and developing a number of mathematical tools to do so. the argument The 1st Chapter (JAS 2003) begins to use these foundations to examine issues of lexigenesis, (word formation) specifically in Aboriginal languages of south-eastern Australia. From a linguistic point of view it examines the problem of how diversity and change can be acknowledged and accounted for in terms of the simple, powerful set of principles proposed. To do so it proposes that the primary unit for analysis should be the morpheme, small elements of language, not the word or lexeme, and corresponding to this it develops morpho-statistics rather than the more common lexi-statistics. With this concept and tool it argues that the current assumption that all the languages of the Sydney basin were different languages was an artefact of this assumption. At the morphemic level, the relationships between the languages become much closer. There is far less diversity to account for in the theory than is usually supposed, and language boundaries need to be re-examined. The existing diversity at the level of words can then be explained as the operation of two major transformational processes, condensation and permutation, acting on a repertoire of basic root-morphemes. The chapter models processes of lexical condensation leading to the development of new morphological forms, using principles from Zipf on power laws in linguistics, and also Hamming Means and recursive arithmetical additions, to reconstruct the site of deleted or condensed phonemes, and their scale and distribution. From this analysis we are able to show that roughly half of the root morphemes are hardly compressed at all, and the rest are compressed according to a Zipf power law curve.

v

However, the chapter then considers the cases where the data departs significantly from Zipf law predictions. It explains these as due to the role of a long tail in Zipf distributions, and to a process whereby there is a consistent collapse of lexemes (words) into morphemes, in a phase transition process analogous to the change from liquid to gas. This process is modelled according to the van der Waals effect. So the on-going process of lexical and morphemic change is governed by the interaction of the two principles, Zipf’s and van der Waals’, in a wave-like pattern of change. Mathematically the chapter then proposes to establish Lagrangian forms as a means to relate these two models. It generalises this Lagrangian to include a van der Waals effect as a perturbation to the Zipf distribution, as the most efficient (optimal) way for new morphemes to occur over time. It uses discrete-optimisation techniques and the new form of morpho-statistics developed in this thesis to show how roots might have developed in these languages. The remaining three chapters use mathematical methods to examine a number of issues in Aboriginal linguistics, mainly using the principles and corpus established in the first two chapters. Chapter 2 (JAS 2004, pp. 151-155) draws on an insight of Andrew Allison to show how to fit Poisson Optimisation on data from two data sets; Eades “Dhurga” (1976) and Mathews “Thurrawal” (1903) (actually ngunuwayal from Lake George). Andrew Allison suggested the relevance of Poisson Optimization for the problem, and carried out the relevant calculations of these two data sets. However, the article itself was conceived and written entirely by myself as first author. But herein lies a problem. The article was based upon the untested ideas of Eades having studied “Dhurga” (1976) and of Mathews having studied “Thurrawal” (1903). What if one of them were mistaken? This is why chapter 3 studies language regionally, discovering that Mathews’ “Thurrawal” (1903) is actually the neighbouring dialect “Ngunawal” from around Lake George. The distance separating the languages is small but the consequences are enormous, and backed up by a “southern Ngarigu” wordlist by Hercus (1969) and a “Ngunawal” wordlist by Mathews (1904). From this analysis it is possible to detect major changes in the long history of the sounds of Aboriginal language. One is an increase in the number of consonants, which provides alternative evidence to confirm

vi

the view of Australian linguists like Dixon. The other is to detect a major shift in the Zipf distribution of first consonants. This is so marked that it can be shown to correspond to the ‘mortality law’ to predict individual and species death by Gompertz. The chapter also uses this analysis to contribute to a debate amongst linguists, as to whether Aboriginal languages were always ergative, without clearly marked accusative forms. The chapter uses phonological inferences to argue for an early accusative form. The final two chapters, 3 and 4 (JAS 2005 and JAS 2006) are a pair, which use mathematical analysis to address two related linguistic problems. One is the issue of linguistic zones and boundaries. The other is the issue of the origin of these boundaries and what they can tell us about the deep past of Aboriginal languages and history. Chapter 3 uses the analysis and extends it to all six language groups. It proposes a basic typology of languages Type A and Type B, rigorously defined in terms of different solutions of a quadratic equation arising from Snell’s law, using a series of ‘linguistic signatures’. From this we can see how the waves of language interacted with each other over a deep period of time (about 60,000 years). This is a knowable dynamic process, of Gillieron wave propagation, acting through some well known physical rules such as Huygens’ principle and Snell’s law. The 4th Chapter (JAS 2006) continues to apply Snell’s law to deduce the likely minimal sound systems of Aboriginal languages, before the last Ice Age and after it. Following the work of some Australian Aboriginal linguists (e.g. O’Grady and Fitzgerald) it also uses phonological evidence to reconstruct deictic systems so ancient that they may indeed derive from very early Aboriginal language before the last Ice Age. It establishes the various levels of language and the proposed source that it came from off to the north-west. Again these boundaries are used as evidence from which to reconstruct likely events over deep times, of about 60,000 years. To try out all our mathematics Mary Everitt (pages 234 and 241) and William Dawes (pages 234, 246-7 and 250) are given brief translations, and retroflexives arising in Victoria are discussed (pages 253-254). These forms of analysis, and the interpretations and conclusions they give rise to, offer hope of re-opening some important questions about the origins of Aboriginal languages that are currently regarded as closed by linguists because they have not seen how they could be studied given the available data. Mathematical analysis as exemplified by this thesis is

vii

presented as a powerful new tool and resource to linguistics. Although the data comes from Aboriginal languages, the methods can be applied to all languages. As argued in the thesis, the proto-form of Aboriginal language reconstructed here comes from a much greater time-depth than any other language family as reconstructed by contemporary linguistics. This would take the language closer to the probable time frame for the origins of human language itself. This gives an even greater significance to the outcomes of this study. limitations and ways ahead This is a pioneering work which operates on a grand scale, bringing together two areas of thought that are not usually combined. Inevitably some tasks are less well done than others for the moment. Here it is appropriate to indicate areas where I am aware of further work that is needed, or where further theory and analysis may be applied to come up with different conclusions. That situation is always the case with scientific research. It would have been desirable to test and validate some of the mathematical techniques on artificially constructed datasets, in order to develop the principles more generally. In the same vein, it would have been more convincing if I had applied the techniques to other languages, in order to verify independently that they function usefully in other complex, real-life data sets. However, there is a limit to what one individual can do, to both build in independent data collection and also model development. I leave these two tasks to those who follow me. Australian linguists will undoubtedly also wish that I had done more extensive empirical work on current and past languages. In my own work I draw on a number of these linguists, and I have found that some are better and more reliable than others. Even though I have argued that poor data can be mined to produce valid conclusions, I would be the first to say that good data is better than bad. In a fair distribution of labour I leave it to professional linguists to provide it. I recognise that some of my use of data can appear circular, because I use mathematical tools to produce data, which I then use as evidence for the concepts that produced that data. While it is true that my data is not therefore ‘pure’, this is also the case with data from mature sciences, whose data is always found and shaped extensively by the dominant concepts in a field, yet is accepted within the given field. That is because

viii

the data, even if it is carefully selected and processed, still speaks for itself and is able to refute specific hypotheses. Given that my work crosses two basic paradigms, mathematics and linguistics, it is necessary for me to declare my position on the differences between the epistemological and ontological world views of the two. Do I assume that the mathematics I apply to language phenomena works because these phenomena are basically the same? That Snell’s law, for instance, can be applied to linguistic developments because there are fundamental affinities between language and optics? I see this as a fair question, which could lead to further empirical work which could itself then provoke further illuminating questions. However, as a mathematician I would also insist on the right and duty of mathematicians to do the mathematics well, and not determine beforehand how or why it fits the whole of reality. Isaac Newton is a model for this as for much else. He provided the mathematics of the solar system without being able to say what gravity was or how it and other forces could work. That mathematics has been refined, but it still holds, in the sphere of its operation, and beyond it. But is that mathematics or the version of the world it projects beyond improvement? This is not the case even for Newton, and in my case far less so.

Appendix on

proto-Australian

Aboriginal woman and chil.dren, with Aunte tendingthe fire, on the beginnings of a Great Walk.

picture by the engraver W. Hatherell, 1888

Appendix

"The language of the blacks was not macle for white man's tongue and thnt iswhy it sounds like blasphemy to hear him try to pronounce an Aboriginalword. ...Similarly bktck man's words shoud never have been put on paperfor there is nothing in our alphabet as we understand its sounds which wouldmake the written word any nearer the original than a feeble parody ... ahaunting echoing softness might give way to an unbelievable drama and therewere dread words which made your spine creep with horror even though youdidn't know their meaning ...".

Bernard O'Reilly (1949),"Green Mountains & Cullenbenbone"

orthographic & phonotactic basis forsouth-east-Australian lexicography

In the late 19th and early 20th centuries, when traditional AboriginalIanguages were dying and falling into disuse throughout south easternAustralia, the journals of the Royal Societies from various AustralianStates and England published hundreds of ethnological and linguisticcontributions including scores of "word-lists" collected mainly bygentlemen amateurs, with no special linguistic training, whoexperienced difficulty with the vowels and various novel sounds ofAustralian languages, frequently failing to even hear or record word-initial consonants, and employing all the worst orthographic traditionsof English spelling -- including poor phonemelgrapheme correspond-ences and "orthographic-fossil" symbols and clusters of symbols that aresilent and/or inappropriate -- with the result that Aboriginal words wererarely ever spelled in the old written records anything like how theywere spoken.

Some were better than others but by modem standards they werecollectively an orthographic hash employing word-initial vowels,distinguishing between c and k,, p and b, proposing fricatives and aletter s, even digraphs such as qu, ch, sh, ck, th, oo, ee, oe, aII ofwhich are ridiculous. Consider, for example, thirty different spellingsof the same Aboriginal name by a range of historic literary sources asin TABLE 1. L And, whilst scores of "wordlists" full of such ambiguousspellings were collected on the outskirts of main cities, importantlinguistic regions elsewhere were often overlooked. This problem isaddressed in Chapter 3 by filling geographical gaps with data ofequivalent quality from early census forms ("Aboriginal BlanketRegisters") on the assumption that phonotactic frequencies should bemuch the same in names (on censuslists) and words (on word-lists).

page 1

" bangare" , Robert Brown, 1802

" bangaree" , Lieut. James Grant," bong-ree", David Collins, 1798

" bongaree" , Matthew Flinders, 1802/3"bongaree", Phillip P. King, 1827

" bongarri" , Rene P. Lesson, 1838

" bongarie", Robert Brown, 1802" bongarry", Charles Rodius, 1844

"bongaru", Philip G. King, 1804

" bonjare" , Robert Brown, 1802

" bonjaree" , Robert Campbell, 1838

" bonj ary", Dr. John Hanis, 1801

"boon-ga-ree", P.P. King, 1819

" boongaree", Mathew Flinders, 1802

" boongarie", George Howe, 1818

"boongarree", P.P. King, 1817

1801 | "boon garry", Hist. Rec. Aust, 1917

" boongary", Lachlan Macquarie, 1815

" boongree" , David cotlins, 1799

" boungaree", Lieut C. Menzies, 1804

"boungari", J.D. d'Urville, 1834

" buggary" , Jules Lejeune, 1824

" bungaree" , Dr. Joseph Arnold, 1815

"bungari", J.D. d'Urville, 1830

" bungarie" , Rev. J. McGarvie, 1829

" bungarree" , The Australian, 1828

" bungarrie" , Alexander Berry, 1838

" bungary" , Samuel Smith, 1802

" bungery" , col. w. Patterson, l8o1"bunjaree", Capt. J. Wallis, 1821

Appendix

TABLE 1.L: thirty dffirent historic spellings of the same Aboriginalname compiled by K.V. Smith, 1992, pages 168-9.

The resulting statistical analysis of 35 historic wordlists collectivelycontaining 6672 Aboriginal words, and 9 early census lists collectivelycontaining 445 traditional names, identifies eight distinct languageregions extending throughout south-eastern Australia - in excellentagreement with traditional tribal boundaries, to the extent that they areknown. Such analysis is made possible by the economy and efficiency ofthe following assumed orthography [Eades, 1976; Yaltop, 1982; Dixon, 1980;Austin, 19971, comprising 18 letters denoting 4 "vowels" and 14"consonants", which eliminates the sorts of ambiguities displayed inTABLE 1.1 whilst also simplifying and standardising the spellings ofwords recorded in decrepit old literary and archival sources.

a,

n,b, d, d,4, g, i, 1, m,

!, n, J7, r, u, w, Y, u... where U = ayA or perhaps 2i2

onhography & phoneticisation copyright O C. U-Bnf, zOO:

This orthography recognises four "vowels" a, i, u and u though thelatter, actually a cluster, is not strictly a vowel but it behaves like oneand will henceforth be referred to as "the long-vowel". The 4 sounds

page 2

€r E2 €r

gundujorq

gorigu

+++ +

++

++

;+

TRANSECIION THROUCI]

TH€ AUSINALAN ALPS

Chapter 1

NEW

SOUT H

WALES

aWoggo Woggo

AlburY- o/

V ICTOR1A

.^\-

C,iqR9\o"j

-+_.===0 50 100 150 200

Ki\ornetre9wlike ng in the word "singing", the 7 sounds like ny, the { sounds likedj , whilst d is a stoppy sound intermediate between d and t sometimeswritten by other authurs as dfi, which is tolerable, though the digraphfh must be strictly avoided because English speakers starihissing g, fik"snakes, when they see it - which is inappropriate for present purposes.

Symbols in this proposed orthography, other than the ',long-vowel",constitute a standard usage of the International Phonetic Alphabet (IpA)- bearing in mind that, with only four contrasting "vowels'i (as compar-ed to twenty in English), each was able to attain a much wider ranga ofphonetic realisation in traditional Aboriginal language. Thus we need nosymbol specifically denoting the English oo vowel sound (as in theword "foot") which, along with other cherished English vowel sounds,are all adequately described in our orthography by the single symbol u,leaving us free (for reasons of typographical utility) to commandeer theredundant keyboard character U for portmanteau use as above, todescribe a vowel cluster about which Watkin Tench (1789) commented:

page 3

Appendix

" Some of their simple sounds, were dfficult of pronunciationto mouths purely English: diphthongs often occur: one of themost common is that of E, or perhaps, ai, pronounced notunlike those letters in the French verb har, b hate".

The existence of this proposed "long vowel", in fundamental words withvarying numbers of letters, is best illustrated by such examples as

gun ------)\---.v.-/

- very

"gayan" ("bis" ) "koiyun " ("many" )

D.K. Eades, 1976, dhurga Rev. L.E. Threlkeld, 1834, awabakal

" gaian " ("large") "klan" ("centipede" )R.H. Mathews, 19O3, Rev. W. Ridley, 1875, kamilaroi"thurrawal" = ngunawal

"giyan" ( " centipede" )"kaiun" ", "kaiyung " ("large") S. McNicol & D. Hosking, 1994,

L. Malone, 1875, wodi wodi wiradjuri

"kian" ("big, large") "kaian" ("old")I. Nathan, 1849, wodi wodi J. Rowley, 1875, wodi wodi

andmUfA -------)

\-!-/

- large, greatly

" muiry "" meheaaruu " ("large") ("large, an augmentative in general" )" meheearong " (" full moon" ) W. Dawes, 1790, eoraJ. Dawson, 1881, J. Larmer, 1853, dhurgakuum kopan noot

,,man,, or ', m,tfi,, (,,Iarge,,)" maiaura " (" emu" ) J. Rowley, 1875 & 1878, wodi wodiR. H. Mathews, 19O2,

gippslandkurnai " muree" ("great, large, long, extended" )D. Paine, 1794, eora

" mutfl " ("very" )Rev. W. Ridley, 1875, kamilaroi (" compararive of much" ), english

" Ilnra" ("great river or sea"), latin

page 4

and

wurula\--------.V------

-- three , several ,' Iots of

" werulla " ( plural)A. Mackenzie, 1872,

" thurrawal" = wodi wodi

" wowulli " ( " three" )L. Malone, 1875, wodiwodi

------J

l) Rowley and Malone agree upon2) Rowley and Nathan agree upon

Appendix

" wuraia] " or " woorall"(" lyrebird" - [tail- feathers] plural)R. H. Mathews, 1902, gippsland kurnai

" waran " (" three" )

Rev. L.E. Threlkeld, 1834, awabakal

the "long vowel"

t 875

, L. Malone 1875

1849

, contradicting Nathan, contradicting Malone

From the Wodi Wodi examples of gun,

(" aia" . l. R owtey

O - l"

aiu" o, " aiyu"

["ir". I. Narhan

can be deduced on the balance of probabilities as follows:

therefore u - aia is demonstrated in the Wodi Wodi language zone.Additionally Dhurga aya and "Thurrawal" ata corroborate thisfinding, whilst the uncorroborated offerings of Threlkeld, Ridley,McNicol & Hosking (respectively oiyu, ia, iya) are all believablemishearings of aia. Also Dawson's eheaa within mura is convincing.

In larger words such as mura and wurula individual phonemes tendto be less clearly recorded due to the jumble of sounds. Nevertheless,from these three simple examples, it appears that at least somefundamental south-east-Australian words have the phonotactic structureCVCVCV... (consonant -vowel- consonant -vowel - consonant - vowel...)and, whilst they mostly seem to comprise four or less letters, Chapter 1

argues that we must nevertheless assume the existence of at least somesix-letter ancestral words (such as wurula) in order to account for thelarge Zipf Mean which is empirically observed in historic Aboriginalwordlname lists compiled in recent centuries.

Other phonotactic combinations may also seem possible. For example,English has sequences ranging from CV ("see") and CVC ("hat") toconsonant clusters such as CCYC ("stop") and even CCCVCCCC

page 5

Appendix

("strengths"). However the alternation of consonants and vowels is sofundamental that some world languages are phonotactically self-rectifying, automatically inserting vowels to break-up consonantclusters:. eg the English word illustration in Japanese becomes"irasutoreshon ", whilst best or ves/ both become "besuto".

Additionally it is a statisticai universal regardless of language, speakeror subject matter, that "just over 60Va of everything we say will bemade of consonants, and just over 40Vo of vowels. About a third of allthe syllables we use in everyday speech will have the structure of CVCas in 'cat' ... tatxdl the 50 most commonly used words in language willmake up about 45Vo of everything we write" [Crystal, 1997,page86].

It is known that "typically, every word in an Australian language mustinvolve at least two syllables. It must begin with a single consonant andcan end in a consonant or a vowel ... there is just one vowel in eachsyllable ... tandl between each pair of vowels there must be either oneor two consonants" lDixon,1980, pages 127 & 1591. Accordingly the phono-tactic structure of words in Australian languages is taken to beCV(C)CV(C) where parentheses, (...), include optional elements. Withmany languages conforming exactly to this norm, and others divergingonly in minor ways, our earlier-discussed CVCVCV structure is one ofthe simplest possible cases.

Most consonant clusters are probably not fundamental - being formedmostly through agglutination and fusion of more fundamental words orelements. For example it is argued, in Chapter 5, that the retroflexivesrn, rd, rl, respectively arose from simple vowel-loss as follows: fUIl-) q , rud -+ Q , rul -+ f . Also we will later see that (in additionto the velar nasal 4) SE-Australian languages had an n:g sound (as inthe English word finger) typically arising from condensation of therepeat-word gun:g(Un). Dixon [1980, page 158] also suggests a doublerhotic ff, and a double lateral ll. But, just as "the articulation of theoccurrence of a vowel in a particular word is often affected by thephonetic value of consonants flanking it" [Dixon, 1980, pages 129-130], sotoo can different sounds arise from consonants such as r or 1 beingflanked by long or short vowels: for example mura sounds quitedifferent 6 mara, and gula sounds quire different to gala. So thisstudy, in line with others [Eades, 1976, page 35], finds the minimal systemof one lateral and one rhotic sufficient. Thus many of Dixon's"optional" consonant clusters are probably artefacts, arising fromagglutination and fusion, best ignored in our search for ancestral roots.

However even our minimal orthography, and the proposed phonotacticstructure of fundamental words, is still insufficient to eliminate all

page 6

Appendix

ambiguities. We also need an internally consistent system of formallogic, based upon some set of fundamental ("root") words, with whichto found the study of morphology. Probably anysuch successful systemof formal logic would correspond in some degree to actual ancestrallanguage initially captured in idioms and, through subsequentagglutination and fusion, preserved within modern words - as below:

ILLERTbased on the internallyconsistent system of formallogic to be outlined herein

TROY (1994)wrong vowels andunnecessary consonants,based on arbitrary opinion

bulu-i balu-wi

bu(Lu)-i bu-yig(ura)-wu(ru)-i ga-wI

TABLE 1.2'. contrast between reconstructions

Thus morphology offers a systematic way of thinking about philology,based upon agglutination and fusion of ancestral ("root") words" Indeed,without some set of fundamental roots (the linguistic equivalent of theperiodic table of atoms) there can be no convincing reconstructions ofsyntax (the linguistic equivalent of chemical reactions). Linguists such asEades (1976), Osmond (1989), Troy (1993), Besoldt (2004) and Steele(2006), attempting syntactic reconstructions of extinct SE-Australianlanguages with little understanding of morphology, found they had nosystematic way of determining where lexical items ended and syntaxbegan. In linguistics as well as chemistry, without atoms all is phlogistonand aether.

In the following ancient roots are proposed based solely upon Australianlexical data, but no apology is offered for also occasionally includingsuggestive Indo-European words. Such philological speculation is partof the historical tradition in this field, with important early contribut-ions having been made by G.F. Angas (1847), J.F. Mann (1840's), S.Bennett (1867) and J. Fraser (1892). And if the proposed proto-Australian roots are anything like 75,000 years old, dating halfway backto the origin of modern human speech, worldwide vocabulary must tosome extent be derived from them. Indeed, any general absence of"candidate" Indo-European cognates might constitute a counterargument that our proposed roots either might not be proto-Australianor, at least, are nothing like 75,000 years old. Thus it would benegligent not to at least speculate about possible cognates from differentcontinents when dealing with words likely to have originated at thesetime depths.

pageT

Appendix

"... it is not ctt all clear that there is such a thing as a universal stock of basicvocabulary.-."

Collin Yallop (1982), page 33

occuffence of consonants in fundamentalfour-letter south-east-Australian words

Assuming the previously-given set of fourteen "consonants" and four"vowels", let us consider all four-letter words commencing withconsonants and constructed from alternating consonants and vowels: iewords of the form CVCV. Such words can be represented by a four-vector

w = (Wr= consonant, WZ = vowel, 1t3 = consonant' Wq = uo*el) t

the total number of "possibile" words being

14x4x14x4-3135However, if we make the simplifying assumption that onlyappear in the y113 position but never word-initially, the

"possible" words reduces by an order of magnitude to

I and I cannumber of

12 x4xZx4 = 384.

If we further assume that proto-Australian had a simpler sound-system,with two less consonants [Dixon, 1980, pages 158 & 1'7 4-l'18], the totalnumber of "possible" ancestral root-words of this form is

10x4xzx4-320.But not all of the i0 possible consonants can occur word-initially withequal frequency N = 32 because, in real-world lexicons, word-initialconsonants are constrained by an Entropy-maximising "power-law" that

can be approximated by aZipf ctwe32l(h+1), for integers ft e [0,11-1], where,Flis the total number of consonants capable of appearingword-initially. Thus, whilst there may naively seem to be 320 "possible"ancestral root-words of the form CVCV, in reality only those occurringunderneath this curve are permitted by the physical principle ofEntropy-maximisation which is discussed more fully in Chapter 2.

page 8

Appendix

t .ru24

T\h20

16

12

I

4

0 h-+o123456789

Relative frequencies of respective word-initial consonants, atanged indecreasing order, must follow an Entropy maximising "power law".

The area under this "power law" approximates the maximum number ofancestral four-letter words, of the prescribed form, actually permittedby Entropy

dh = 32xln(10) = 73h+l

however even this is probably an overestimate, as a Zipf curve onlyapproximates the decreasing exponential required by rigorous theory(see Chapter 2), so 60 fundamental words may be closer to reality.

Although this "power law" dictates the frequency spectrum of word-initial consonants, it does not tell which consonants occur word-initiallywith which frequencies N7x . This can be roughly estimated by empir-ical word counting from actual word-lists and, by chosing languagezones with differing values of Il , extrapolating back to protoAustral-ian language for which we assume a simpler consonant inventory withsomething like FI= 10. And we mainly need to do this for word-initialperipheral consonants because words made from them constituteapproximately 60Vo of any lexicon - a significant sample which can beexpected to be more representative than any remaining words based onrarer or less certain word-initial consonants (see data-tables on pages204 - 2r3).

Nh = 32 ...forhe t0, H-llh+l

where -F1 = l0 for proto-Australian

32

t;

page 9

source frequency (7o)

word-initial-1

frequency (7o)

word-initial--f

freqtency (Vo)

word-initial-fl

Sydney "eora"W. Dawes, 1790

0 0 0

Illawarra "Wodi Wodi"Lizzy Malone, 1875

0 0 0

Jervis Bay "D'hurga"D.K. Eades, 1976

0 0 0

Twofold Bay census1839 "blanket register"

,.1

J 0 0

Gippsland "Kumai"L. Hercus, 1969

4.3 2.6 0

Port Dalrymple, TasmaniaJ.P. Gaimard, 1826

I 1.8 6.5 0

Research Bay, TasmaniaJ..I.H. de Billardiere. 1793 12.5 1 1.1 0

Appendix

TABLE 1.3'. Relative occurrence of some rare word-initialconsonants in historic SE-Australian wordlists, showing increasing

usage toward the south. Calc ulated from dnta tables on pages 204-213.

ln her "Dhurga" lexicon, much of which was late-collected at Jervis Bay,Diana K. Eades (1976) found no instances of word-initial-1 or word-initial-r. However the French Astrolabe expedition, at Jervis Bay in1826, recorded the nouns "leroko" ("huitre" = oyster), "loungan"("langrayen (oiseau)"), and "lerinn" ("oeit" = eyes), thereby establish-ing a likely northern limit for word-initial-1. This suggests an importantlinguistic boundary between "northern" coastal languages (such as"Eora" from Sydney, also "Wodi Wodi" from Illawarra) and "southern"languages (such as "D'hurga" along the NSW south coast, "Kurnai" fromGippsland in Victoria, also "Tasmanian").

Another difference seems to involve the important nasal n which, due toits rarity, has zero frequency in all sources in the above table. Howeverone convincing example of word-initial-Il occurs in R.H. Mathews'(1901) "Thurrawal" which documents nurT ("yonder") as "nharra" inlanguage from the Canberra region. There also seems to have been anexpression nura:Wa(fa):fl documented by the surveyor R.L. Campbell(1862) as "yarra:wa:' ("the name of the Kangaloonforests"), and alsoby W. Dawes (1790) as "qarra:watn" ("a great distance off").Conversely waf a;nuf a survives today as the place name"woro:nora". All these examples of the nasal n come from "northern"coastal language, whilst there is little evidence for it in the "south".

page 10

Appendix

Chapter 2 shows how to precisely determine values of F1 by usingentropy maximising "power laws" that best fit historic wordlists fromdifferent language zones. But for present purposes, assuming that"northern" language such as Wodi Wodi employed n word-initially butnot l or r, whilst "southern" Ianguages such as D'hurga and Kumai hadboth 1 and r word-initially but not n , we can estimate that H = 12 forWodi Wodi and H = 13 for D'hurga. Hence the following extrapolation:

TABLE 1.4: Calculated from the data tables on pages 209 and 212.

And, taking our earlier estimate of something like 60 "possible"ancestral four-letter words permitted by entropy, we might expect tofind lurking in historic Aboriginal word-lists approximately

)n60 x "" = 12 commencing with D?

100

t7.660 X ^ ' '" = 11 commencing with !

100

t4.260 X -"- = 9 commencingwithS

100

q460 x 1' = 6 commencingwithb

100

Thus before studying historic Aboriginal word-lists on a larger scale wecan already anticipate 12+ll+9+6 = 38 "tikely" ancestral roots, eachwith the consonants 1 or r in the y13 position and peripheral consonantsin the yyl position, approximately as above. This is more than half ofthe 60 "possible" roots of the prescribed form permitted by entropy.

word-initial

consonant

tueqtency (Vo)D.K. Eades, 1976

D'hurga (Il = l3)

frequency (7o)L. Malone, 1875

wodi wodi (H - 12)

linearly extrapolated

freqtency (Vo)protoAustralian (Il = 10)

m 14 i6 2A

0 4.4 8.8 r7.6

g 26.5 22.4 14.2

b 19.3 t6 9.4

page 11

Appendix

" Orgmizntbn need not be ascribed totally to ... operators, inductors,eyocators and regulators. Space ... contains its own constraints ... [thnt]dictate the organfuation of all the elements thnt can take possession of it ...the possible confi.gurations dre strictly limited.., The nature of space itself issuch a; to preclude so much, and permit so little, that there can only ever be ahandful of answers to a multitude of questions'"

p.S. Stevens (1g74)

octahedral symmetry and"vowels" in fundamental

the occurrence offour-letter words

Most Aboriginal languages throughout Australia have just three basicvowels, "a system that we tentatively assign to protoAustralian -

although there is an area in the central north where five-vowel systemspredominate" fDixon, 1980, page 178]. Additionally about half of allAustralian languages distinguish between long and short vowels and,from " comparison of cognates in widely dispersed languages ... thereare sufficient cognates for us to be quite certain that protoAustralianhad a contrast between long and short vowels in initial syllables" loixon,1980, pages 131 &2121.

In saying this, previous researchers were mainly considering sounds thatcan be represented by a simple doubling of basic vowels indicating extralength - ii, aa, uu. The present text proposes a novel clusbr al'a whichmay possibly have been the ancestral basis for most, if not all, of thevalid long vowels observed in modern languages - with the contractionsa(ia). ai(a) and \a)ia being phonetically distorted in various ways.Existence of this proposed ancestral long-vowel cluster is certainly asimplifying assumption with which to proceed. It may just be a deviceuseful for generating a representative sample of fundamental ancestralwords (some subset of the protoAustralian lexicon) or, altematively, itmay actually be true. Whatever the case, the Principle of Parsimony(also known as Ockham's Razor) mandates that "the simplest explanationthat fits all of the facts is probably true". Thts we need to consider thespectrum of lexical possibilities of the form CVCV, following from theminimal set of three short vowels and one proposed long vowel, beforedeciding whether to even think about options of greater complexity.

For specified consonants in the l4l r and, Wz positions in ancestral four-letter words, the set of all possibilities resulting from, say, the vowel uinthellz position can be generated from Pythagorean multiplication ofmonosyllables (1 * 4 - 4) whose different vowel combinations withinresulting disyllabic words can be represented as a tetrahedral array.

page 12

Appendix

1

u

W2-U

(1)

Example Ibur... is the basis ibr fbur"possible" words that can bedisplayed tetrahedrally -

ExampLe 2

mul... is the basis lbr fbr.rr.

"possible" words that can bedisplayecl tetrahedrally -

Wq = any oJ fourpos,sible vowels

(4)

bura

complete set of fourpossible disyllabic words

(4)

buru

mulu

muli

page 13

Appendix

Wq = uny o.f .fuurltos,sible wtv,e l,s

(4)

two-tie red tetralrcdron(intenriily. ciontaining ." :l:':l:!3:), i,,rlr'.ii,iing sixtccn possible wordr

( 16)

Wz = any of .fourpossible vowels

(4)

page 14

Appendix

More generally, for specified consonants in the 14l I and Wt positions in

four-letter words of the prescribed fbrm, the full set of vowel

combinations caube generated by Pythagorean multiplication of mono-

,fffuUf", (4* 4 = 16J and represer.rted as a two-tiered tetrahedron (see

iepf) containing two conjugate octahedral arrays that can be separated

Oe ='tO i Ol ona studi;individuallv (see BELOW)' Displaying all

iorriUf. iou. letter worcls in this geometrical fashion enables the entire

ip... "f f.-i.al possibilities to be visualised and systematically studied'

For example if, as suggested herein' the vowel i hod g"n"tally been

reserved as a present-tense marker tor verbs' then it would have been

"""i"rl"g and therefore inappropriate fbr. it to have commonly

."""tr"f,i" ad.lectival roots and leiical iterns based on them' Hence we

*""r.i-""i- u. 'surprised to fin6 the entire top-tier of this two-tiered

i.i.-ur*0."" unclerutilized by fundamental words from actual historic

wordlists. Conversely *..outtt argue that such an underutilized slab of

#;;;i"pgr$br" i.ii.ur ,pu"" *o-uld, in itself, .constirute

evidence that

the vowei i served some other function in historic language'

octahedron with reversed( "t oniupule" ) vowel combinations,, pre sient in 14 six P,,ssible words

(6)

two-tiered tetralrcdron( intcrn(tllv cttnluinirty dtt odIIh(dn'n)'

rcpresiting lt'n pttssil'lc w'tnlt

( 10)

page 15

Appendix

Consider a matrix A , capable of interchanging voweprescribed form, as follows

A-ll.w:

1000000100100100

ls in roots of the

W1

Wa,

1rl3

W2

... for W2 - 1,Yo

W1

W2

WZ

Wq

Its characteristic polynomial is given by the determinant

il^r-A ll -)"-r 0 0 0

0 L 0-r0 0 L-r 0

0-1 0L

- (h-r)'(1; -1) = o

implying a characteristic matrix equation of the form

(A- I)' (A'- I) V = gwhich completely summarises the symmetries of this lexical space.

To understand the meaning of these component factors consider rootscontaining identical vowels (:vlz=Wn), existing at the extreme verticesof the two-tiered tetrahedron, for which we have the TetrahedralInvariance

Aw = w

which is a mathematical way of saying that hypothetical roots such as

mulu, mulu, mala, munt,quru,1ara,4ulu, dulu, buru, bara,wuru, wara, J1uru, Iara., etc, are all unaltered by vowel inter-changement.

page 16

Appendix

In contrast roots containing different vowels (W2 + W4), occurring at

the octahedral vertices, are changed by the matrix operation

Aw = w* ... for Wz # ll+

effectively'Jumping" from one octahedral array to the other. For suchroots, in these octahedral arrays, a second application of the matrix Ais needed to restore their original form

Aw* - A2 w = w

This last equation, whichto the other then back,Octahedral Symmetry.

Thus the characteristic matrix equation is simply a product of terms,respectively representing Tetrahedral Invariance and OctahedralSymmetry, following from the very nature of the proposed "di-syllabic"four "vowel" lexical space in which nothing else is possible.

Octahedral Symmetry has clearly been a defining phonotactic propertyof lexical space since protoAustralian times but could it also havelinguistic significance - specifically could there exist pairs of adjectivalroots of the prescribed form, ! and [* (counterparts in opposing

octahedral arrays), that are semantically linked? If so they would alsoprobably be of protoAustralian origin, hence widespread throughoutknown Aboriginal languages, for the simple reason that lexical spacewould have started filling long ago, during protoAustralian times,leaving little room for an entire deep structure such as OctahedralSymmetry to "independently arise" in language much later, as anafterthought, and to spread widely across language boundaries in recenttimes. Homology (having the same origin in protoAustralian) would befar more likely than homoplasy (coincidentally having the same form),given that an entire deep symmetry is involved. Thus, if "universal"conjugate roots exist, they are likely to have protoAustralian origins.

The best place to initially seek surviving protoAustralian roots would beamongst the most unchanged ancient oral traditions on the Australiancontinent - arguably in SE-Australia where for tens of millenniaAboriginal people were geographically isolated and separated fromhuman populations with different origins and certainly weren't tradingwith Asian or Pacific neighbours. In contrast, languages in Australia's

describes roots jumpingcharacterises what will

... for W2 * Wq .

from one octahedronhereafter be called

page l1

Appendix

TABLE 1.5: eight proposed conjugate adjectival pairs of the prescribed form,collectively suggesting that Octahedral Symmetry may also have linguistic significance,compiled mainly from First Contact sources in the Sydney-Illawarra region of SE-Australia. The UU/UU combination is most common and convincing, followed byUa/aU. In some of these proposed words undetermined vowels are indicated by *.

ol1ru ----> " gllrroo" ' 1",nin" I{---,"- T . Watling 1790, Eora

concave

OU1U ----> " gorai" ' r" 1or" t

",------"- J . Rowley 1875. Cowpasrures= fat,

convel

" kltrra",J. Rowley 1875,

gura ---->

toward" ki ri" , (" throat")

L. Malone 1875, Wodi Wodi

garu

from

( " gAr""", t" to coush' t )I W Dawcs t790 Fnra I

---- I I

l"gorey", 1"juice" = squintl

I Capr. L Hunter 1790. Eora )

gul*

vital

( "kuller", t"eood"t \I J. Molon, t875. Thurawal

I

---+ I I

l"galta(n)", 1"rhe sun" tIf Si. -t. Eorls t770. Eora )

g*lu= duLl,lethargic

( "callooa", ('ro ctinb"t \I Cop,. J. Hunter 1793. Eora I

->l I

| " gooleur" , r'koata bear" t I

f eealr a Dawson 1850's. Kurnai )

buru ____> "barru", (',t"a')

=;;;;: R H' Mathews 1e03, Nsunnawat

quiet, sLow

buru ____> "burrai", ("quick")\ . a-r R. H . Mathews 1904, Wiradjuri

= vr80rous,noisy, quick

" paira"McCaffrey 1920's, Wodi Wodi

"bourra"D. Collins 1798, Eora

bura ----> baru- eitherside

( "peerah", r"ch""k"t )I M. A4unina 1863. Wodi Wodi

I--+ | |

l"bera", r"ioue" = cheektI

\1.P. Cainard 1826. Turuwal )

mu]fu1 ----> " maiAl", (" stranler" )

J. Rowley 1875, Cowpastures-- unfamiliar,

strange,unexpected

mulu __-_> " mullA" , (,,husband,, )r--v.=J W. Dawes 1790, Eora

usual,expected

" mai",J. Rowley 1875,

" mia" , (" the eye" )D. Paine 1794. Eora

mu(l*) ---->

= visible,accessihle

m*lu= invisible,ohstucted

" mileea", 1" *op"1W. Dawes 1790. Eora

''warlWUfi ----) lgood as in "ro lind" )

=iii,. W. Dawes 1790, Eorawru

= wrong,rmproper

( "wirra",,'ooa"t )I J. Malone 1875. Turuwal I

--+ | |

| "wEirdh", ra'a"t I

It . Narhon t849t. We4i tYo4, )

page 18

Uf ... as in fU ... as in Ul ... as in lU ... as in

"wr" . guru"urr" gura"uf' gura"aIT" bUrU"air" bura"ourr" bura"ar" wuri

"ni" guru"ree" garu"tey" garu"rrai" buru"rah" baru"ra" baru"ffa" wiru"rah" wiru

"u11" gul*"a11" gul*"aiaL" mulu

"lhooa" g*1u"leur" g*lu"lla" mulu"leea" m*lu"1i" m*lu

Appendix

TABLE 1.6: in keeping with the Pinciple of Parsimony we view doublerhotics and double laterals, accompanied by biTarre Roman vowel and semivowelcombinations, as orthographic gibber best replaced by the indicated combinations ofsymbols from our fficient minimal Aboriginal orthography which specificallyexcludes o, oo, e, ee, h ma maybe also !. This makes all the more sense asthe words in TABLE 1.5 are mostly from the SydneyJllawarra region where,arguably, there would have existed a single basic sound system - not several.

far-north and north-western regions contain recognisable SE-Asianinfluences and, having evolved rapidly in recent times, are less likely toresemble original ancestral language brought to Australia 75,000 or soyears ago, across the SE-Asian landbridge and oceans, ultimately fromEurasia. To the extent that SE-Australia was most remote and isolated,it might be expected to have retained the most ancient linguistic featuressuch as the proposed conjugacy of meaning associated with OctahedralSymmetry. Thus it would be best to initially find some evidence forprotoAustralian conjugacy here, as in TABLE 1.5, before expanding thestudy area across different language zones throughout a much largerregion of the continent.

Although SE-Australian language was the first-extinguished it was alsothe first-recorded, hence the most "authentic", with written accountsdating back to the 1770's and 1790's. In philology, as with any branchof taxonomy, the oldest original primary sources (not the latestcompilations) must be given the greatest weight and priority. And,although we can grumble about the quality of these earliest of allwritten sources, language from the Sydney-Illawarra region is by farthe most extensively documented, by multiple sources, precisely becausethis was where First Contact occurred. Such an abundance of lexicalmaterial from a single region readily enables us to identify likelycandidate pairs of conjugate words as in TABLE 1.5 but it also resultsin the multiple contradictory historic spellings in TABLE 1.6, that aremore a consequence of diversity of source material than genuinephonotactic information, in a language zore that can reasonably beexpected to have had a single basic sound system.

page 19

Appendix

Europeans since First Contact, unable (udging by TABLE 1.6) toconsistently and efficiently represent the Aboriginal "long-vowel" soundin combination with either a rhotic or a lateral, appear to havecompletely missed the Octahedral Symmetry represented by theconjugate adjectival pairs suggested in TABLE 1.5. But there may alsobe something else happening as well. We note that whilst the words inthe left column of TABLE 1.5 (with a "long-vowel" in the first"syllable") may be relatively uncontroversial, those in the right column(with a "long-vowel" in the second "syllable") may be far less expected.Dixon [1980, pages 131-2] wrote

"About half the languages of the Australian continent have adistinction between long and short vowels ... there are almostalways restrictions on long vowels - they may not occur insuccessive syllables, or only one long vowel may be permitted perword. In many languages long vowels occur only in the initialsyllable of a word ... Comparison of cognates in widely dispersedlanguages indicates that protoAustralian did have long vowels inthe initial syllable of a word; there is no evidence ... at otherpositions " ,

and again [Dixon 1980, page2l2]

"... there are sfficient cognates for us to be quite certain thatprotoAustralian had a contrast between long and short vowels ininitial syllables; but there is no evidence that it showed longvowels in latter syllables of a word".

Whilst the proposed minimal orthographic representations ur and ulare convincing in TABLE 1.6 we need at least a partial explanation as towhy previous researchers have so long been "blind" to the existence oflong vowels in second and subsequent syllables. From the data inTABLE 1.6 we see that the representation /u is also convincing but,notwithstanding Octahedral Symmetry, word-final ru seems to actuallyhave been abbreviated down to either rai(a) or ray(a) in traditionalspoken language. To further study this proposed tendency we can referback to TABLE 1.1 which supplies thirty different historic spellings ofa single Aboriginal name, probably buqgaru, with no less than tendifferent spellings of word-final ru

...ru ----> ...rai(a) -->

t,fe,, tt fitt

" ree" t'ffi"t'rree"

" rie"" ry" " rrie",, rry" ,' rlt,'

page 20

,A.ppendix

Logically, at most, only one of these historic spellings might be true.But, given their mutually contradictory natures, all are likely to bewrong - making our proposed contraction the simplest and bestrepresentation of word-final ru. In making this judgement we note theexistence of exactly the same sort of orthographic gibber in historic"word-lists" regardless of whether we are dealing with spellings of thedifferent words in TABLE 1.6, or of the same word as in TABLE 1.1.

This apparent contraction of word-final ru in spoken language in noway undermines or contradicts the Octahedral Symmetry suggested byword-pairs in TABLE 1.5 but it does, perhaps, explain why Dixon andothers have long been unable to recognise word-final occurrences of the"long vowel" in fundamental if not universal multi-syllable words -hence why they entirely missed the underlying Octahedral Symmetry inthe lexicon. It may also happen that the "long-vowel" interactedphonetically with other vowels and certain consonants, in actual spokenlanguage, resulting in similar contraction of the ui sound withinexpressions such as the verb given in TABLE 1.2,

bulu-i ---. bulai(a)-i --> "boolaay- ee" )

name bu4garu,and the u4 sound as in the TABLE 1. 1

and also in the important negation nut. When one actually tries to saythese combinations, it becomes clear enough why contraction occurs andthere is a predictability to it. Also it was suggested earlier that such con-tractions of the proposed ancestral "long-vowel", may be the basis forother long-vowels and even retroflexives in recent Australian languages.

Whatever the case there is no need to apologise for logical simplificat-ions, such as Octahedral Symmetry based upon a minimal orthography,made in accordance with the Principle of Parsimony. Nor do we need toslavishly "prove" simplifying models or assumptions, as simplicity andefficiency are real-world universals and there is nothing gained bystarting with assumptions of undue complexity. Rather the burden ofproof lies more squarely upon those advocating additional layers ofcomplexity, on top of simple models and assumptions that generallyseem to account for all of the known facts. If we were to take seriouslydouble rhotics, double laterals, bizarre vowel combinations, andmultiple phonologies in a single region, then all those layers ofcomplexity would require far more justification and proof than oursimple finding that the lexicon is governed by Octahedral Symmetrywith the proviso that word-final ru, and sometimes also u within wordsin conjunction with other vowels and certain consonants, was probablymodified in spoken traditional language.

page 2l

Appendix

"The scholar is a quarryman, not a builder, and aII thnt is required of him isthat he should quairy cleanly. He is the poet's insurance against factual error"

Robert Graves (1963)

"..'. I would say that the real scientist, if not the schol'ar in general, is no.

quarryman, but'is precisely and' exactly a builder - a buil.der of facts atulobsenations into cbnceptual schemes and intellectunl models that attempt to

represent the realities ... Who would deny th^t a role by far the greater ispfayed by the original thinker and critic who discems the broader outlines of^the'plan, who sy"nthesises from existing knowledge through detection of the.

false anrl itlumiiation of thi true relationships of things a theory, a conceptual"model, or a hypothesis'capable oftest? The creativity of scientific writing lies^

precisely hei6 ... to be no mere qwffryman but in some measure a creator oftruth and understandinq "

Bentley Glass (1964)

data mining - analysis or synthesis?

In the study of poorly understood extinct languages we hear researchers

complain of lexicons "down to the last 300 words and 18 verbs" [Nugent

19311 and languages elsewhere " that have disappeared without a trace"

[Crowley 1997, pages 289-291]. But, clearly, these researchers lacked morethan the desired vocabularies. They also lacked an adequate understand-

ing of morphology - of how actually "known" lexical items wereconstructed from more basic elements and, also, how to expand

vocabulary by creating valid "entirely new" words as needed. For toolong Australianists have viewed lexical items from extinct languages as

irreplaceable structureless empirical facts, mere objects to be shuffledabout within sentences in accordance with syntactic rules, items capable

of actually being

Significant effort also needs to be invested in studying the structure and

composition of lexical items themselves. A start has been made withrespect to orthography and phonotactics but, in order to study lexicalitems properly, we need to have some ideas about how to actually make

them from more fundamental components. And we also need to be able

to more accurately determine their true meanings which are usually justas poorly documented in historical sources as the lexical items

themselves. Furthermore, literary and archival sources are grossly

incomplete and quite unlikely to contain "free-form" examples of everyadjectival root as clearly documented as those cited in TABLE 1.5.

Some may only have been recorded within larger, more complicated,phrases or expressions. To recognise and study them, in

_such_environments, will therefore require useful hypotheses and models ofthe the structure of lexical items including compound adjectives, verbsand nouns, also numerals and arithmetical expressions.

page 22

repeated ADJECTIVES

gar(u) gar(u)ORIG|NAL B

"gaff:qaff"( "slime left behind inthe

)Itrait of stugs ancl worms" )

C. Richards 1902, Wiradjuri

(as infrom fromexcrement) las in excremenl)

= SLIMY_EST( s up erlat iv e adj ectiv e )

rJUluuiii6 g

nuluuiai[iig

"gnalluingulla"(" mushroom" )

D. Collins 1798, Eora

ORIGINAI._ B

_ WRINKLY_EST( s up e rlat iv e adj e ctiv e )

huru - buruORIGINAL_ B

= SLOW-EST( superlative adj ective )

" birra:birra", ("razy" )McNicol & Hosking 1994, Wiradjuri

buru - buruSOUTHERN_ B f"burra:burri", 1"be quick") \

l " booroo:boory a ", r " gun" t I

\ Petrit & Dawsor /850's, Kumai )( s up e rlativ e adj ect iv e )

guru- guruORIGINAL_A

HT]YGENS_CIRCLE

(" jarra:jarri". ( "tean" ). DharrukI

l" birra:birra", ( " t i sht. not heavy" )| "Thurrawal" = Ngunnawal

I O. t. Mathews: 1901, 1903

__ THIN.EST( s up e rktt iv e adj e c tiv e )

gutu- guruARIGINAL-B " purri'.burri ",

("whale" )Mickey Munima 1863, Wadi Wadi

= FAT-EST( superlative adiective )

Appendix

TABLE 1.7: examples of repeated (superlatively inflected) adjectives,including some that suggest conjugacy of superlative meanings.

Starting with the simplest case it is proposed that adjectival repetitiondenotes superlative inflection, in which event the examples in TABI-Etr.7 suggest conjugacy of meaning extending also to superlative adjectiv-es. In studying such examples it is important to appreciate that nouns areinappropriate as ascribed meanings for adjectival expressions. Insteadwe need to seek qualities of the respectively cited nouns: eg the " trail ofslugs and worms" may be the SLIMY-EST substance, a "mushroom"may have the WRINKLY-EST gills, a "gun" may be the FAST-ESTthing, and a "whale" may be the FAT-EST of creatures. Also, if buru-buru does have the superlative meanings FAST-EST or NOISY-EST,then conjugacy would have buru-buru meaning SLOW-EST, TIRED-EST, LAZY-EST or QUIET-EST, consistent with the imprecise historicascriptions "lazy" and "bye and bye". Likewise, 1f guru-guru has the

page 23

Appendix

TABLE 1.8: examples of undertying adjectival meanings projected acrossproposed word-classes that are distinguished by simple (verb and noun) markerswhich, despite being shown as suffixes, can equally appear as prefixes. Evidencein support of these and other-such words follows later in this Appendix.

superlative meaning FAT-EST, then conjugacy would have guru-gurumeaning THIN-EST, consistent with the imprecise historical ascriptions" lean" and "light, not heavy".

To further expedite matters it is helpful at this stage to ignore all of thearbitrary verb tense-markers and baroque pronoun-systems proposed byR.H. Mathews, in his numerous century old publications, that seem tohave been taken far too seriously by generations of subsequentresearchers [Eades 1976, pages 54-55; Troy 1993; etc]. Instead, later in thisAppendix, a back-to-basics analysis of primary "raw lexical data" fromnumerous historic literary and archival sources (including R.H.Mathews' own writings) will provide scores of examples, from acrossall SE-Australian language zones, clearly demonstrating universal usageof the vowel I as a present-tense marker on verbs. Likewise overwhelm-ing evidence will also be presented in support of the inanimate andanimate noun-markers proposed in TABLE 1.8.

For present purposes it is sufficient to explain that if fundamentaladjectival roots existed, as proposed in TABLE 1.5, then they did so

ADJECTIVEpresent-tense

VERBinanimateNOUN

anlmateNOUN

1gun=irn6-, uit"l

radiant

gula- i-=iifia{radiating

gula- nradiant-thinp

(eg Sun or Moin)

gula- AIuminous-creature

(eg glow-worm)

gula=visdquarrelsome

treacherous

gula- i-=_,:+= arqurng,

fighting,conlurlng

gula- nharmful-thing

(eg barbed spear )

gula- 0harmful-creature

(eg locust, sorcerer)

guru= .fat, convex,

more

guru- 1

= .fattening

gutu- nconvex-thing= arc, arch

(eg Rainbow )

guru- afat-creature

(eg whale, possum)

mU]U= visual,

visible

mulu - i= looking,

seetnS

mulu - nvisual-thing

= eye

mulu- !Isharp]eyeId]-

creature(eg eagle-hawk)

page 24

Appendix

TABLE 1.9: examples of the kind of conjugacy of meaning thnt couldreasonably be expected to flow from adjectival-pairs to correspondiig verb-pairsor to corrcsponding noun-pairs based upon them. Evidence suppot"ting thesi andother-such conjugate-verbs and conjugate-nouns follows later in this Appendix.

within the context of language and their respective meanings canreasonably be expected to have carried through to simple verbs and tosimple nouns made from them - as in TABLES 1.8 and 1.9. Further-more any conjugacy of meaning could also reasonably be expected tohave carried through from adjectival-pairs, to corresponding verb-pairsand to corresponding noun-pairs, as suggested in TABLE 1.9. Thus theexistence and true meaning of any proposed adjectival root needs to beascertained within a broader context that also includes relevant verbsand nouns for which semantic conjugacy, if true, is a powerful tool withwhich to clarify and more precisely determine the meanings of poorlydocumented lexical items and their constituent components.

conjugateADJECTIVES

gura= toward,

int o(as in absorbent)

ga.ru

- .from,out-of

(as in sound)

bulu

bulu= down. flat(as in sick, ill)

; conJugate , conJugate i conjugateI inanimate i animate

i NoUNS i NoUNS

" gua-n

" gura-rJ

" =kto:thlnc 1 =riiii7rs/b*r"d

'. (as in nouth, iar,\ | -Crealure

" l;;,;," i;;i;:";;;;l

" (as in petican):;, li,iit-'iiir. iiii ) i suckl!-creature

ll

L______lli garu-n ' garu-ai -t;;:tnn{ i =""i'-;,*6"t = saurce ' (as in kookaburra);: (,:' j:,!ip!:,: \ ispiiry)iq,i,ty-i,iiiu,e| \ Jountaln, spnn9 ) ,, '(as'in iamei or squid )tltl

i bulu-n ', bulu-gI

', emerging/ projecting ', commenc iig-c reatu re| -lhinq, ' ( as in newborn baby \i (as in btossom, \ il": I flower, I 't

t cht(k JromaneSS )', lwoman's breasts ) i

', bulu-n i bulu-g', = dow.nward-thing '. = dying-creaturet I as rn dtqesttve-t rac\\ ',

i I river; or Tallen tee ) i

, present-tensei VERBSt-, gwa- It+', = (in)coming,t recelvtngi ( as in hearing, )i I eating, drowning, I

' inhalinp. absorbins I, \ "' "Jl---------', garu - 1

', = {oing, .giving,t emftttng'. ( as in singing, li speaking, squirting, I

i I infecting )i-

" bulu- it_', = risinq, projeui.ng ,I emerqtng, slarnng', I as in blossoming )

' \.flowering, birthing )

I

t bulu- il!--...Y+

' = falling, stoppingi ( as in sinking \

', \drowning, dlin? )

page 25

Appendix

We might also expect any proposed adjectival roots to have commonlybeen used as descriptors, in association with noun-stems whose dual andplural cases could respectively be represented either by repetition orelse with the ancient quinary system of numerals (based upon therecursive arithmetic) discussed in Chapter 1 :

miru = 0dula - |

bulala - 2wurula - 3

nurula = 4murula = 5

For example singular nouns were sometimes represented by the numeraldula , in conjunction with either an animate or inanimate noun-marker,as in the standard animate singular (masculine?)

dula: g\.-----.-vJ

= creature,person

" pallii: 11" , (" mate orphan" )

"dulla:", (n"")Lizzy Malone 1875. Wodi Wodi-

"dullai: ", 1" Aborisinat man" )

J. Rowley 1875, Cowpastures.

(coastal ORIGINAL - B)

"dhulli:", ("man")

R. H. Mathews 1901. Dharruk(ORIGINAL - A)

... to which adjectival descriptors could be attached (singular noun-stems being assumed unless otherwise stated) as in

ORIGINAL_A

ORIGINAL-BSOUTHERN _B

"dali: n", (" husbend" )

D.K. Eades 1976, Dhurga.

" kuulae: " , (" *or, Aboriginal race" )

J. Dawson 1881, Chaap Whuurung.

(SOUTHERN _ B)

gura - n:(dula)\-v-J

into creature= " suckling"

(animate [implied singular] noun)

and

" giro : ng: " , (" chitd" )

J. Rowley 1875, Cowpastures

" yunl:n i " , 1" young man" )

L. Malone 1875, Wodi Wodi

ORIGINAL_B

.nura - n:@ula)noif lii-*nl ---p;6o, - ORIGINAL_B

(animate [implied singular] noun)

page 26

Appendix

and

mulu - n:(dula)--- .vtsuoL creatureORIGINAL_ B

" maliya'.n: "(" wedge tailed eagle" )

Rev. J. Gunther 1892, l iradjuri

and

mulu, - A:du(la)---"---:-!-----.-w--J

related, person

familiar

ORIGINAL_BSOUTHERN_ B

Tt]RUWAL

" mulla:ngi",(" wife, general term" )

J . Dawson I BBI, Peek Whuurung

(SOUTHERN _ B)

" mul i:da" , (" old woman" )

J. Malone 1875, Botany Bay.

(TURUWAL)

" mu]la:', tt , (" a man or husband" )

W. Dawes 1790, Eora

(coastal ORIGINAL- B)

" krJuruu i lrli " , (" teat" )

J. Dawson 1881, Chaap Whutong

" bur: n'.dd", (" penis" )

R. H. Mathews 1903,

" Thurrawal" = Ngunnawal

( " narru: fl: " . 1" moon" 1 \

lPertit & Dawson 1850's, Kurnai

I

l "g urra'. n : ", (" daytisht" t I

I Rev.lV. Ritlley t875. Kamitaroi )

and

bur(a) - n:du(la)\.-------!.J

between fiegs lHUYCENS_CIRCLE

(animate [implied singular] noun)

(animate singular noun)

... likewise also with the standard inanimate singular

garu - n:(dula)from

(inanimate Iimplied singular] noun)

and

Uuru - n:(dula)t-_!----.J\--------------]

cyclic thing

SOUTHERN_B

SOUTHERN _ BARIGINAI,_ A

(inanimate I irnplied singular] noun)

(inanimate singular noun)

page 2'7

Appendix

and

thingSOUTHERN_ BORIGINAL_ B

were sometimesincluding usually

pair of things(dual noun - stem)

ADELAIDE

" mal'.tai iitya'. " , (" cheeks" )

Teichelmann & Schurmann 1840,

Kuarna

and

.m7lu - .du(la: n) - d@la: n)pair of things

(dual noun - stem)

HUYGENS CIRCLE

" mele: kei ::t | " , (" eye[s]" )Dr. J. Lhotski 1835. Ngarigu

muf U - nidula -------:-:::=::=:::--:---)

small

" ma: n:da ", (" bush" )D.K. Eades 1976, Dhurga.

" poroi :: tcholl " , (" bush" )J. Dawson I88l, Chaap Whurrong .

(souI-HERN - B)

" p(Dpo :: " , (" baby" ), Greek

Over a vast region of Australiaconstructed using repetition of thesingle adjectives, as follows

dual-nounssingular noun,

" murray '.: tooloo "D. Collins 1798, Eora." marry:: diolo "Capt. J. Hunrer 1795, Eora.(" five" = finger)(coastal ORIGINAL - B)

" mi:'. tti ", ("little" ), AwabaknlRev. L.E, Threlkeld I 834

(inanimate singular noun)

mal(a) - du(la: n) - du(la: n)bulgy

(adjective)

visual(adjective)

-- CHEEKS (dual noun)

= EYES (dual noun)

page 28

du(la: n) - du(la: n) - gula - m(ura)pair of things

(dual noun - stem)

ORIGINAL_ A

Appendix

and

greatly radiant = " pretty"(adjective)

lro^" o7 hisroric farnsread\

\on LYREBIRD Creek NSW J

":: ja:: kular: "(" lyrebird" )

W. Russell 1914, Gundungara

and

SOUTHERN_B

... or, using the numeralstandard animate dual

g:bula(La)creature- s (two)

(animate dual noun)

very " paddly"(verb)

" dya: n:dhu:: k ii ", (" canoe paddte[s]" )

R. H. Mathews 1902

Cippsland Kurnai

bulala, in concise expressions such as the

"0 iAla " , ("ustwo")

" 4 :dlu " , (" we two only" )

WilLiam Dawes 1790, Eora

(coastal ORIGINAL - B)

"4: k'th " , (" each other" )

M. M. Everitt 1900, Gundungara

(ORIGINAL_ A)

= PAIR OF LYREBIRD TAIL - FEATHERS (dual noun)

du(La): n - du(Ia: n) - g(un - i:4uru)pair of things

(dual inanimate noun - stem)

= PAIR OF PADDLES (dual inanimate noun)

ORIGINAL-BORIGINAL_A

page 29

Appendix

or

bul(ala):n -vRrGNrL_I-)creature- s (two)

(animate dual noun)

hence the standard inanimate dual

n:buh(ala)

" eel: oong ", ("we two")

M.M. Everitt 1900, Gundungara

gurainto

SOT]THERN_B

" jerra : ni gill " , ("tousers" )

N.D. Pettit & J. Dawson 1850's

Kurnai

" m: i i: lla ", {" th" "y"'" 1

J. Oxley 1818, Wiradjuri

" Ie: ri: n ", ("eyes")

J.P. Gaimard I826

Jervis Bay

\-------------- ---.-..-------J

thing - s (two)

(inanimate dual)

and

rn(ulu): i\_________."__--______J

seetnS

(n:bula)la

either side[of the nose ](ba)ru - n

\-_-.-.v-l

thing - s (two)

and

ORIGINAL_B

TURUWAL(bula)la -

Over the same vast regionrepetition of singular nouns

gula:4harmful - creature harmful - creature

= [swarm of] LOCUSTS(animate plural)

tripple nouns were made simply fromand their two adjectives

and

dulu:nORIGINAL_B

" gula: ng: gula: ng "(" large locustIs]" )

R.H. Mathews 1903

"Thurrawal" : Ngunawal

"dula: ng'.dula: ng "(" exhausted. sinking, \| ,rody ro rru^blr" )

Rev. J. Gunther 1892,

lliradjeri

HUYGENSCIRCLE

dulu:nshabby - creature shabby - creature

= [bandof] REFUGEES(animate plural)

(inanimate dual)

pair o.f things (inanimate dual)

page 30

Appendix

and

Inuta:\0 1 -\---------i/-

blob

mula:pblob

ORIGINAL_ B

" mulla:: mullu: ng " , (" pteiades" )

Lizzy Malone I875, Wodi Wodi.

" moloo:t molot ng ", (" Pleiades" )

D. Collins 1798. Eora= [constellation of] STARS(animate plural)

likewise

QurU:n- gUrU'.n-\..-.-------.!-!-----------1-

afc arc

= RAINBOW

... alternately

0iwurula

9(UrU: n)g-='_-]

arc

" gura: n: gun'. g'. ''(" rainbow" )

R, H. Mathews 1908, NgariguHUYGENSCIRCLE

= creatures, people(plural animate)

" gn: iour",(" ascidie" : a colony of ascidians)

J.P. Gaimard, 1826, jervis bay(TURUWAL)

" n:barr", ("flock or crowd" )M.M. Everitt, 1900, gundungara

(ORIGINAL_ A)

"4: yellu", ("we three only" )":eedra", ( " Aboriginal people" )W. Dawes, 1790.

" i eArah " , (" people" )D. Southwell, 1788.

(coastal ORIGINAL- B)

ORIGINAL-ATURUWAL

ORIGINAL-B

in the standard concise expression

ORIGINAL_B

" nirrd i wry: " , (" shell" )

L Nathan I849, Wodi Wodi

This selection of plurals, all of the form adjective times noun-cluster(s),should confirm that there are probably only 38 possibilities. In fact thelast two-score or so clusters is exactly what we expect. There is noevidence for diminishing vowels or exotic consonants.

lara - wuru(la:n)\----vJ!.-.----v-]

coiled things

coils, whorls (plural inanimate)= [snail] SHELL

page 31

Appendix

"So then always that knowleclge is worthiest ... which considereth the simplefotms or dffirences of things, which are few in number, and the degrees andcoordinations whereof make all this variety.,,

Francis Bacon

muru / mura

The First Fleet officer William Dawes observed that " murry " was usedas an augmentative in general - thereby denoting both size and degree.Consider the example involving small thing (= bustr, baby, finger etc)

muru - n:dula= smatt -:1fr;

$yf,\Mk-_t,/ singular )

--\noun - stem/

or

"mainida", ( " bush" )D.K. Eades 1976, dhurga.

hence a definite Yes

" murray:tooloo "D. Collins 1798, eora.

" marry::diolo "" poroi::tcholl ", ("bush" ) Capt. J. Hunter 1795, eora.J. Dawson 1881, chaap whurrung . (" .five" = .finger - siTed)(SOUTHERN . B) (coastal ORIGINAL - B)

" lt(.Dpo:: ", ("baby"), greek "mi::tti", ("tittle" ), awabakalRev. L.E. Threlkeld 1834

dul(a: n) - muru

"tah:: moor ", (" bronlewing pigeon" )"tu'.: mDrIe", ("short") ,,thal;: more", (,,shaltow,,)W. Dawes 1790, eora.,,too r muroo,,, (,,short,, ) ::::'r':,X"in"'.("

a take in Manchester" ??)

D. CoLlins 1798, eora.(coastal ORIGINAL - B)

-----------)ORIGINAL_ B

"ed, : tttvo(1", ("bush"), sreek" Xo( :: lt o tK\o6o t " , ( " bush" t. greek

nun -- ------:--), J T]NIVERSAT.

_ YES

page 32

" ngaayn", wiradjuri

McNicol & Hosking 1993.

"4E", wadiwadiLizzy Malone 1875 .

(ORIGINAL- B)

" aod', thurruwul

J.P. Gaitnard 1826.

(THURUWAL)

Appendix

" yuin" , dharug

J. Rowley 1875.

(oRtGtNAL- A)

" nai", dhurga

J. Larmer 1853.

" naa", kurnai

Rev. J. Bulmer 1876.

(SOUTHERN - B)

1S the basis for slightly-Yes

.muru, - Uun..Y-v-slrqhtly yES

---------------)ORIGINAL_BORIGINAL_A

or

" murroo I ng ", (" ?ood" )

J. Larmer 1853, awabakal.

" more: mme ", ("yes" )D. Collins 1798, eora.

" marroo : k ", (" good" )J. Rogers 1899, eora.(ORIGINAL- B)

"nga: murra ",

R.H. Mathews 1901,

" thurrawal" = ngunawal( HUYGENS CIRCLE)

"ngja:m", ("bush")

W. Bothong 1902, dhurga(SOU'rHERN - B)

" mOO : gOOn ", (" maybe/ perhaps" )

M.M. Everitt 1900, gundugara.

" mufra: Qa ", ("maybe")

R. H- Mathews l9O1 , dharig.(ORIGINAL- A)

" muru: bah ", (" uaa" )R.H. Mathews I908,

"dharig " = wolgalu(HUYGENS - CIRCLE)

HUYGENS CIRCLESOUTHERN_B

"gi: m", ("misht")Evelyn (" Granny" ) Ferguson 1964,

Bunjalung

"noon: moo", ("maybe")

ancient egyptian

_ MAYBE/PERHAPS

Uun - muru= MAYBE/PERHAPS

page 33

Appendix

i,(bu)luPast

present

"bu: ny: i: l'La", ("struck")

W.J. Enright 1900, port stephens

"bu: n'y: i: ", ("st,ik"")

W.J. Enright 1900, port stephens

" ba : nga : : baou " , (" witt paddte" )

W. Dawes 1790, eora

bu(lu\ -downward

'rJu(ru) -flailing

1

= HIT, STRIKE i,buQu) future

A multiply recorded word-phrase is given above and expanded below. R.H. Mathews

t1904, ci6dby Besold (20d4) page 531 states that the imperative is marked with a final iinflected to the verb stem. This is wrong, as it is clearly indicitive mood.

bul(u)zm(ura) - 1

am / are(present )

- fiu(n):mutu - -)"bul: m: i:: ", ("strike" ).

"bul: m:a: nga: murta",(" I may strike him'' ).

R.H. Mathews 1901,

" thurrawal" = ngunawal( IMPERATIVE MOOD)

\_?

maybe

= am striking( present tense INDICATM MOOD)

= maybe am striking( present tense CONDITIONAL MOOD)

and

bul(u)zm(ura) -

i:(bu)ludid (past)

i:bu(lu)!_/-

shall ( future)

- IJu(n)zmuru\_,maybe= strike

(rMPEP-4TMMOOD)

(past/ future tense INDICATIVE MOOD)

= maybe did / shall strike( past / future tense CONDITIONAL MOOD)

"bul: m: i: a:: ", (" did strike" )

pasl "bul: m: i: a: nga: murra", (" may, have struck" )

R.H. Mathews 1901, " thurrawal" = ngunawal

"bul: m:: a:: ", ("shqtt strike" )

" bul : m:: * : nga : murra " , (" misht strike" )

R.H. Mathews 1901, " thurrawal" = ngunawal

luture

page 34

Appendix

whilst, conversely, the expression for definate No

nu0NO

" naiyug ", ("no" )L. Malone I875, wodi wodi.

"nang", (" no" )D. Collins 1798, eora(coastal ORIGINAL- B)

forms the basis for slightly-No

" mariaa: gin'. nee ", ("bad")" muria: gin: nea ", ("hoax, lie" )N.P. Pettit & W. Dawson 1850, kurnai(SOUTHERN _ B)

There is also a verb for

gulatishining

\--------\.-----

= shininq, radiating

"kiII: i", (" shine, spark" )RH. Mathews 1901, dharruk."gill;e", ("moon" )Rev. W. Ridley 1875, kamilaroi(ORIGINAL- A)

muru: QUn - nUQ ----------------

UNIVERSAL

" ning ", (" no" )M.M. Everitt 1900, gundungara(ORIGINAL- A)

"aiann", (" no" )J . P. Gaimard 1826, jenis bay(:TURUWAL)

ORIGINAL_ASOUTHERN_B

" moo: gun''. ning ", ("maybenot")M.M. Everitt 1900, gundungara(ORIGINAL- A)

ONGINAL*AORIGINAL_B

" kaL: i ", ("water")" kiil: ", (" urine" )McNicol & Hosking 1994, ngiyampaa." kil: " , (" urine" )R.H. Mathews 1904, wiradjuri(ORIGINAL- B)

NO\...------------

_ MAYBE

which is combined in one's largest (middle) finger

( "kullar:: heaar: " \| 1" r", ord 1ins"r"1 I

It. Oo-ro, 188t, peek.nurrorsl

\ (ORICINAL- B) )gulazG) - mur(a-n) ----)

\--v-l

- shiningt---------.--n--.

I arg est - one

page 35

mura- n. - gul(a):i

" mu:n:ya", (" .finger" )J- Dawson 1881, chaap whurrong(ORIGINAL- B)

" mee:n::hi",(" one of the sisters at katoomba" )P. Stone, SMH 1931,

C.M. Ward, 1949,

S. Gurr & G. Harrowsmith, 1949(ORIGINAL_ A)

Appendix

or

\----------V----

-- shiningORIGINAL_AORIGINAL_B

" mota: n:gull: y"," murri: n: gull: y " ,

(" Chief of the Nattai Tribe, 1800- 1858" )

" moye:n:gull;y ",(" portrait by Major Mitchell" ) .

" mya: n:garl: ie ",W. Russell 1914, " Iast chief' oI gundungara.

" murru: n: gurr: y ",W.A. Cuneo 1893, gundungan

(ORIGINAL_ A)

There is also an expression large orbiting thing (see pagesTT &78)

mura= large

----------------ORIGINAL- B

These examples of muru anddo not possess the Octahadral

" meheaaro: ng: "( " full - moon" )J. Dawson I881kuurn kopan noot.

mura gwe plausable meanings, but theySymmetry. We will return to this later.

I ar+ est - one

wip- du(Ia:n)

" murray: yanna: dah: "(" full - moon" )P.G. King 1793, eora

orbiting- thing - MOON

turtlepicture by Ainsley Roberts, 1962

page 36

There is an adjective

mulu\------v--------

- unfamiliar

hence the verb

mul(u): i\----------.v----

= startling

Appendix

mulu / mulu

ORIGINAL_ B

" maula", ("man" )D. Paine 1798, eora ,

" miyal ", (" a stranger" )D. Collins 1798, eora ,

"maial", ("stranger")J. Rowley 1875, cowpastures

" mal] t ee " , (" he was startled" )

D. Collins 1798, eora

--------------.-ORIGINAL_ B

and the objects of such influence

At the other extreme we have

mulu\-Y-l

= fnmiliar, exp ected

" moolo", (" hence, thus" )

Rev- J. Bulmer 1876, kurnai

(SOUTHERN - B)

" moolo " , 1" shaddow" 1

D. Bunce 1859

(" mullu' p " ", ("lighrnins") I

lR.H. Mathews t908. ngarrugu I

l" ."u' p : Inet!: p ", r" tishrninss' tllR.H. Marhews t904. ngunawel )

mulu:b(uhu) - mulu:b(uhu)un exp ected - above un exp ected - above

ORIGINAL _ BSO.]THERN _ B

" malayafr ", (" friend, acqua int ance" )

Rev. J. Gunther 1892, wiradjuri

(inland ORIGINAL - B) .

" mul]a " , (" a man or husband" )

D. Southwell 1788, eora ,

W. Dawes 1790, eora ,

D. Collins 1798, eora

( coastal ORIGINAL - B )

= lightnings

HUYGENS CIRCLE

page 37

Appendix

and its augmented form

mttra - mulugreatly- familiar

= "married or related"

as in spouse

= married(adjective)

" mulla: ma;: ng " , (" husbarul" )

D. Southwell 1788. eora.

" molli: mi; nQ ", 1" husband" )

J. Rowley 1875, cowpastures.

" fi1ull:Lrazt " , 1" married" )

" mali:: n " , (" husban(I" )

D. Collins 1798, eora.

( coastal ORICINAL - B )

" mu] t.t da t " , (" old woman" )

J. Malone 1815, botany bay.

(TURUWAL)

,. l" burra: mullina ". t 'marrv" tl- oNC-tNAIi I R.H. Morhr., 190a, w,iradiuri I

I lintand )RIGINAL - B) )

Qula, g - mulu z Inutz\-v-l

person ------------)

ORIGINAL _ B

or

" da t molai: " , (" name - sake" )

J Rowley 1875' cowpastures ,,dali: nt; Ifiafa ,,, (,,husbantt,,)(coastal ORIGINAL- B)

R. H. Mathews 1903,

" :t ll73]i : fi1a " , (" wikl woman" ) " thurawal' = ngunawal

( HUYGENS CIRCLE )D.K. Earles )9769 , dhurga

(SOUTHERN. B)

mu]u : Ir7utz du(la) ' 4\:----:,-j----j, ORIGINAI' - B

Derson SOUTHERN - B' TURUWAL

= married SOUTHERN - B(adjective) HUYGENS CIRCLE

HUYGENS CIRCI,E

" mulla ttt ng " , (" wife, general term"

J. Dawson 1881, peek whuurong

(SOUTHERN _ B)

" milla tt,, ng " , (" wfe" ).

" mul ., dha : " , (" otd. woman" )

R.H. Marhews 1903,

" thurawal" = ngunawal

(HUYGENS CIRCLE)

... where we starl to expand the words used on page 27.

page 38

There was agreement

Appendix

----------------+ORIGINAL_BQU(n: gun)- mul(u)

or

----------------+ORIGINAL_B

l"goz: mul ", 1" frienctship" l\I D. Cotlirt 1798. eora

I troortotORIGINAL- B) )

"ma: kunt 9 ",("lover or sweethectrt" )W. Dawes 1790, eora.

" mauz gohn: " ,(" lover or sweetheart" )D. Collins 1798, eora .

( coastal ORIGINAL - B )

" mail : bra t n t ", (" ofren" )Re v.J. Bulmer 1876, kurvai

SOUTHERN - B

----------)event

bur(u)- i: mulu - g, (dula)

quietly - persuaded

= UNFAITHFUL(adjective)

\-v-l

person

___________---__ jORIGINAL _ B

" pir:: milla z fie z "," marriage by stealing awoman and compulsion"

J. Dawson 1881, chaap whoorung

or

Uur(u), i, d(ara)- mulu - 4, bula(La)

running - relationship

= ELOPING(adjective)

ORIGINAL _ B

" yuur :: t : Iflalz : ng : bula" ,

(" marriage by elopement" )J. Dawson 1881, chaap whoorung

mu(lu)- gun: g(un)extremely- familiar

= "int imate"

and various methods

mul(u)-bura n,(dula)

familiar - betvveen

= OFTEN ( sex )(adjective)

or

two people

page 39

and possibly

Appendix

"mo:izdiaz ", 1" gir" it to me" )-----J J.F. Mann 1840's, kuringai(coastal ORIGINAL - B)

mu(Iu),i - du(La.n)

@

c*

bo

a

;o€ss

A.

q3

(-)

'::a:',').i,. |r.

\..]l

Aboriginal woman hunting "sugar-gliders" : bU::!:WU: : WUfU::!:b:i: WU::b:i:IJ . Do you notice that the spelling is all the same?

picture by engraver W. Hatherell, 1888

= anticipating thing

i.'.

$ttl

'I

page 40

Appendix

mulu / malu

There is an animate noun

mulu - 0\______!/_______

visual creature

- [sharp]eyed

and the inanimate counterpart

mu(lu) - nq-J

thingv Is u0I

-------------JUNIVERSAL

" maliya : n " , 1" wedge - tailed eagle" )

Re v. J. Gunrher 1892. wiradjuri.

" malya z fi " , (" eagle hawk" )

Mc. Nicol & Hosking 1993, Wiradjuri

(inland ORIGINAL - B) .

" mullioi n", (" eagte" )

Re v. W. Ridley 1875, kamilaroi

(ORIGINAL_ A)

" malya z n " , (" eagle hawk" )

R.H. Mathews I903,

" Thurrawal" = ngunawal

( HUYGENS CIRCLE)

"md: ", (" eyes" )

J. Malone 1875, bo tan y bay

(TURUWAL )

" mai: " , (" eye" )

J. Rowley 1875, cowpasture.

"mia: ", (" the eye" )

D. Paine 1794, eora.

" mEigh: ", (" eyes" )

I. Nathan 1849, wodi wodi.

( coastal ORIGINAL - B).

----------------UNIVERSAL

" mit ng", (" eye" )

N.D. Pettit & W. Dnwson 1875, kumai

" mooeh: ", (" eye" )

Re v.F.A. Hagenauer 1876, kurnai

(souTHERN - B)

" mehi " , (" eye" )

R.H. Mathews 1903,

" Thurrawal" = ngunawal

(HUYGENS CIRCLE) .

R. H. Matthews 1903, dharkinyung

(oRrGtNAL - A)

page 47

mu(lu):n dur(a)'i

= SeeLnS, looking

and its augmented form

mura= greatly

\----v----------,

eye

hence

mu( lu , nl - buru\-----v- ,.eyc = tazy

al so the verb

Appendix

" moor :: boora " , (" one eyed" )

----J D. Southwell 1788, eora

( coastal ORIGINAL - B )

\---1------J

throughing

" pA : n: dEir : E ", (" to see" )

----) I. Nathan 1849, wadi wadi( coastal ORIGINAL - B)

also exotic expressions such us

" mili" , (" eyes when shut" )

R. H. Mathews 1903, dharginYang

(ORIGINAL _ A) .

" Thurrawal" = ngunawal

( HUYGENS CIRCLE) ,

mul(u t n)\---v-----------

conspicuous I thing ](noun stem)

--_--)UNIVERSAL

g(un), wu(rula - i:Ilara)

" g : wo !. meil: " , (" Teather" )

-------+ Capt.J. Hunter 1793, eora

( coastal ORIGINAL - B )

At the other extreme we have

malu= obstructive,

blocking

" malu " , (" tired, lazy" )

Rev. J. Gunther 1892, wiradjuri.

(inland ORIGINAL - B) .

" mileea " , (" stop" )

W. Dawes 1790, eora

(coastal ORIGINAL - B)

" mura i mia :i. " , (" stare at" )R.H. Mathews 1904, wiradjuri

(inland ORIGINAL - B)= lookinT

page 42

and an augmented form

Appendix

mura -)= greatly

malu.(i)\------i/----------'J

= obstruct Iive ]

" mullar :i aura" , (" blind' )

R.H. Mathews 1903 ,

" thurruwal" = ngunawal( HUYGENS CIRCLE)

or

ma(lu ' i)\-v------/

= obstruct Iive ]

" meEa : ma:. (" not unders tan d" 1

---------------t' W. Dawes 1790, eora( coastal ORIGINAL - B )

malu '(i) -\-------------- ----J= blind,

obstructivesheilding

" mu.t gin", (" blind" )

R. H. Mathews 1904, wiradjuri

( inland ORIGINAL - B) .

D.K. Eades 1976, dhurga

(SOUTHERN _ B) .

gun ----)= very

" mullo:: ko" , 1" delay, d"f"r" )D. Bunce 1859.

" maL zz kar " , 1" shieldfor warding offblows" )

J. Dawson 1881, kuurn kopan noot, peek whurung

(coastal ORIGINAL - B) (SOUTHERN - B)

" mela. tho I n ", (" shield" )L. Atkinson 1853 , midthung

( HUYGENS CIRCLE)

hence the repeat-word (with repetition) emphasising degree

\-----\r.-------

thing

mu(ra) -\--v-l

= greatly

and

also

malu\______.\i_/

= obstructivesheilding

" kn: ee I m- kni ee, m ", (" btind" )------) W. Dawes 1790, eora

( coastal ORIGINAL - B )

- BLIND- EST (superlative adjective)

page 43

Appendix

= obstructive,blockage

dula - ma(Lu) - n ------J

" mul i warrie'. " , 1" Iong water" )

----) C. MacAlister 1907 , midthung( HUYGENS CIRCLE)

" obstruction - s" = [river]caterdct - S

Looking at these examples, of unobstructe d mulu and obstactles malu,does give a clear picture - but not in accordance with OctahedralSymmetry. We will return to this later.

or( " nin i ma i n ", (" shietct" )I

I J. Rowley 1875. cowpastures .

| " itr._ . mo . n ". (" shiettt" )

I t Tench 1788. eora .

\ (coasral ORICTNAL- Bl

also

mal(u\- wuru(la: n)

i

Aboriginal men hunting "kangaroo" - QUII'.QUfUi(f) o, bUfU:!picture by engraver W. Hatherell, 1888

thing

page 44

mula,(n) -

= [ constellation of ] STARS

(multitudinous plural)

" mulla t, mullu . ng " , (" Pleiades" )

" * mullyu : n" , (" morning star" )

R.H. Mqthews 1903, " thurrawal'= ngunawal

( HUYGENS CIRCLE) .

" milla,, milla , " , (" pipe ctay" )

Rev.W. Ridley 1875, kamilaroi

(ORIGINAL _ A)

Appendix

mula / malu

ORIGINAL-BSOUTHERN-B

mula: n\-----v.-------blob

There is a word

mula\__-v---__J

= droplety,blobby, dusty(adjective)

" moola" , (" sick or to vomit" )

D. Collins 1798, eora .

" mi la ", (" btood" )

J. Rowley 1875, cowpastures

(coastal ORIGINAL - B) .

" muula " , (" ulcer, sore, boit" )

" muula( ng) " , (" to vomit" )

McNicol & Hosking 1993, wiradjuri

( inland ORIGINAL - B) .

and the repeat-noun

" mula " , (" head tice" )

D.K. Eades 1967, dhurga .

" mulloo " , (" pipe clay, dust, earth" )

N.D. Pettit & W. Dawson 1850's, dhurga .

6OU:THERN- B) .

mela( ncholy ), mala( dy ) , ( flou ) miLL : english

lteLa(y7d,tu), 1tolo(vorl) pn),o(e ): greek

mala( ise) : french

" moloo t rrtolo : ng " , (" pleiades" )

D. CoLlins 1798, eore .

" mulla ,, mullu, ng " , 1" Pleiades" 1

Lizzy Malone 1875, wod.i wodi

." mulli ,, mt Ia t " , (" the Pleiades" )

A. Mackenzie 1873

" m(tlh: mltlh,: ", (" Pleiades" ),

L Nathan 1818, wodi wodi

(coastal ORIGINAL- B) .

" muula I muula; ng ",(" to be very sick" ),

McNicol & Hosking 1993, wiradjiri

(inknd ORIGINAL - B) .

___-------)UNIVERSAL\_v-_____________

blob

page 45

Appendix

\-v-----------,

= greatly

= crealure(noun-stem)

0

we need to imagineto air ma1u. Thus

" mulli ,t nt gan; Q", (" sirl" )S. Mowle 1838, ngunawal .

" mulla: Dtedrl:e ", (" girt" ), ngarigu

" milla t I7e zz " , (" vulva" ), " thurawal"

R.H. Mathews 1908, I903

(HUYGENSCIRCLE) .

and

mula\-----v---J

= vomitiferous,spewifurous(adjective)

also

muladroplety

At the opposite extremefundamentally different

" mula a fi1& " , (" vomit" )R.H. Marrhews 1904, wiradjuri

( inland ORIGINAL - B)

- gun: g(un) -----J\---_-.v--l

extremely

" mulla z IlQ zz " ,

(" wfe, general term" )

J. Dawson 1881, peek whumurg

(SOUTHERN_ B) .

that solid bubbles mula arewe have creatures

malububblv

= creatufe(noun-stem)

n

" malu t ng I gan z g", (" platypus" )

R, H. Mathews 1904, ngunawal

( HUYGENS CIRCLE) ,

- gun: g(un) ------)

" ma]]u; 17 t earl z e", (" platypus" )

R.H. Mathews 1902, kurnai .

(SOUTHERN, B)

" mu]la I n I gan: g ", (" platypus" )

M.A. Fitzgerald I89l, eora

(coastal ORIGINAL - B) .

\-_____________ ____J

extremely

= MENSTRUAL

= BUBBLEIFEROUS

page 46

and the juveniles

Appendix

0 --------->l-!-/

creature(noun-stem)

" .ji ' mmialo I ng", (" plarypus" )

N.D. Pettit & W. Dawson l85O' s. kumai

(SOUTHERN - B)

i;:;,'-._.:q--"-;--.--..-- -,_,R/a-.r. \:j

- c;: r'i-;- -i

a 8 o an n a *' #:,!"8, f (::("'{,{!ftn k i n g s p e a r s

Qi - malu:-v-

= tiny bubbles

" dya , malu ' ng " , (" platypus" )

R, H. Mathews I908,

" ngarrigu" = wo lg alu

( HUYGENS CIRCLE)

The name of the dumbbell shapedceremonial ground, at the Great1888 Shoalhaven Bunan, wasrecorded by R. H. Mathews a's

" wurra:wurru:dtha:ng" =w u rui wu ru da(r u) - n@ r a).

page 47

Appendix

mala / mulu

There is a word

mala -------J!_--_\,-/

= bulgy,inf lated

(adjective)

as ln

" malyo" , (" a swetling in general" )

" malyo ( nendi) " , (" to swell, be pregnant" )

" malyo ( partanna ) " , (" futt of joke, funny" )

Teichelmann & Schurmann 1840, kauma

( ADELAIDE - region)

" lldAd" , sreek

and also perhaps

mala\_v._____J

= buley(adjective)

----+GREEK

ADELAIDE

an augmented form

" mal'.tai iitya'. " , (" cheeks" )

Teichelmann & Schurmann 1840,

Kuarna

" pa),a: Kxd"

very

i:(gara)\-v-------------/

= strokingvery inf lated

= erect

mal(a) - du(La: n) - du(La: n)bulgy

( adjective)

as it appears in

1gun: man

pair of things(dual noun - stem)

= CHEEKS (dual noun)

gu(n)

- d@la: n)\--------v'----

thing

_ PENIS / CLITORIS / NIPPLE

page 48

" gan-----+

Appendix

: millu: tth :: i '. ", (" masterbation" )

R. H. Marhews 1903, dh arg inyung(ORIGINAL - A)

or

" carre i

------+

gun)ima(la) - d(ula:n)ga(ra: i\------------.v------l

-- stroking\-------Y------

thing

" da ::: mu'. t : " , (" masterbation" )

---J R.H. Mathews 1902, kumai(SOUTHERN - B)

d(ula: n\------------.v-----_----J

thing

to be

1-fl

" na '. ty '.'.: mil '. i : i" , (" masrerbation" )------) R.H. Mathews 1904, ngunawal

( HUYGENS CIRCLE)

- mala\------Y-l

inf late

mille : ba : da : " , (" to soak orwash in water" )Capt.J. Hunter 1793, eora(coastal ORIGINAL - B)

erect

or

= stroking

ga(ra) - gun): mal(a)\_______v_______

erect

and also

gura

int o

- bu(Iu):dar: i\-Y-l

arise

= PENIS / CilTORIS /NIPPLE

= PENIS / CLITORIS / NIPPLE

= ABSORB (verb)

page 49

Appendix

At the opposite extreme there iscompactify or skrink"

mulu mu(ra) -------+

shrivelled greatly

an augmented form meaning "/o

" mulu : ma " , (" shut a clasp knfe" )W. Dawes 1793, eora

(coastal ORIGINAL - B)

Thus we anticipated approximately 12 words of the form lnx**, 21d

immediately have 10, though muru/muta and mulu/malu do notfollow Octahedral Symmety. We will return to this later.

Aboriginal woman with pretty headwarefrom photogaph by J.W. Lindt, 1888

= compctctified (terb)

page 50

Appendix

guru / guru

guru ----------------!--v'------J UNMRSAL-- fat, convex,

more

gun= veU

" guurny" , (" fat, possibly kidney fat" )" garfa" , (" testicles" )

McNicoI & Hosking 1994, wiradjuri

( inland ORIGINAL - B) .

" gorai" , (' fat" )J. Rowley 1875, wqdi wadi

" CaUrA" , (" whale" )

D. Paine 1794, eora

" curra " , (" more" )

D. Southwell 1788, eora

" gdre " , (" more" )

W. Dawes 1790, eora

(coastal ORIGINAL - B) .

" gu'ra " , (" testacles, scrotum" )

ngunowal 1904. ngarigu 1908

" gufta(n) " , (" fat" )

" thurrawal" = ngunawal, 1903

R.H. Mathews

( HUYGENS - CIRCLE) .

hence to degrees of fatness ranging from "possum"

" gun '. gara " , (" possum" 1

--------) D.K. Eades 1976, dhurga

(SOUTHERN _ B)

" garuy" ,

(" the emu - apple, Owenia acidula" ),

" ghori ", (" fat" )" guru't, (" round" )

" kurai " or" kerai", (" lobster" )

" gwra", (" spider" )

Re v. W. Ridley 1875. karnilaroi .

" katau", (" testacle" )

" kurai " , 1" ;at'1R.H. Mathews 1901 , dharug .

" alrro", (" testacles" )

Major MitcheLl , gundrngara .

(ORIGINAL _ A)

" kauri" , (" circle, compass" )

Teichelmann & Schurmann 1840

kaurna, (ADEI.AIDE).

" kherrny", (" testicles" ), ancient egyptian

" yUpA " , (" round" ), greek

" giro" , (" circle" ), italian

The next candidate is

guru- fat

page 51

Appendix

gurc - (mu)ra- \_-_v____J \--v-________'

= fat = greatly

" euruz ra ", (" possum" )

" gu z mara " , (" possum" )

D.K. Eades 1976, dhurga

( SOUTHERN _ B) .

" KU(Uez fa ", (" opossum" )

J. Malone 1875, botany bay

(TURUWAL) "

or

.mura - (gu)ru -----+...........1-

= grea y = .fat

to "whrtle"

and

ORIGINAL _ BSOUTHERN - B

TURUWAL

" kdrao : rd" , (" possum" )

L. Malone 1875, wodi wodi

" kutru, rd" , (" possum" )

M. Munima I863, wodi wodi

(coastal ORIGINAL - B).

" mLrrfe ', ra " , 1" whale" 1

J.Inrmer 1853, dhurga

(SOUTHERN * B)

= FAT-EST(superlative adjective)

and the repeat-noun

$urU : n\______v_______

arc

HT]YGENS_CIRCT,E

" purri I burri " , ("whate" )Mickey Munima 1863, wodi wodi

" burri : burri " , ("whale" )R.H . Mathews 1903 ,

"thurruwal" = ngunawal

Q(utU : n)\--________v_/

arc

guru - guru ----1F1GNTL:E-----)

gUrU: n\----Y---------J

arc

guru i n\-.v--------..J

arc

arcs (plural inanimate)

-_ RAINBOW

HUYGENS CIRCLE ',

SOUTHERN_ B

ORIGINAL-B

page 52

" QUfiI z rr z QUra z a:: Q . . .

I rarnDow )

R.H. Mathews 1908,

" ngarrigu" = wo lg alu

(HUYGENS CIRCLE) .

guru -------+

= s lim, less,cotLcave

= SLIM-EST(superlative adjective)

The gregarious soldier crab mictyris longicarpuswell known for forming large armies of perhaps1000 on sandy tidal flats during low tide. Itburies itself and emerges by eruptions oftumbled sand ready to feed.

At the opposite extreme of meaning we have

"gurroo( ee) ", 1"thi, man's name")T.Watling 1790, eora .

"jerri ", ("thin")J. F . Mann 1840's, eora( coastal ORIGINAL-B)

hence the adjective (with repetition implying degree)

guru - guru -------+

" j arra : j arra", ( "lean" )dharug, (ORIGINAL-A) .

"birra: birra ", ("Iight, not heary")"thurrawal", ( HUYGENS CIRCLE) .

R.H. Mathews: 1901, 1903

guru - galu= fat = hdppy,

helpfull

Appendix

" tar : n t tar : n z pSfo = t ... ",

kuurn kopan noot

" taera t, kAa ,, QUur:: k ... " ,

chaap whuurong .

" tuura:71ut; 17 r, ... " ,

peekwhurrung . (" rainbow" )

J. Dawson 1881, melboume

(SOUTHERN - B) (ORIGINAL _ B)

page 53

@ €hru @llarr

Appendix

garu / gura

The most common root is

garu= FROM,

out-of , sound(adjective)

ORIGINAL * BHUYGENS CIRCLE

" gara" , (" tell" )

R.H. Mathews 1903,

" thuryawal" = ngunawal .

" kuru", (" wind" )

Dr. J. Lhotski 1834, ngarigu .

( HUYGENS CIRCLE) .

hence the augmented form

garu _ gun\-----1/-/ t-!-',

- "noisY" = Yery

" karu r gan(0 " , (" common magpie" )

" kurru, gaia" , (" shout" )

R.H. Mathews I903,

" thurrawal" = ngunawal .

(HUYGENS CIRCLE) ,

and the repeat-word

gar(u) - garu

" yap tyapa",(" g arg le" ), greek

" curry", (" wind" )

J.F. Mann 1840's, eara .

" gOrA", (" wind" )

J. Rowley 1875, cowpastures

( coastal ORIGINAL - B )

" kheroo ", (" the voice" ), ancient egyptian

" garr(tlus", (" chatterins, babbling" ), latin

" kara " , (" .from" ), japanese

ORIGINAI, _ B

HUYGENS CIRCLE

" karrai t gn", (" shout" )

W. Dawes 1790, eora

(coastal ORIGINAL - B)

" Qarr: Qaff",(" s lim e lefi behind in the trail of slugs and worms" )

C. Richards 1902, wiradjuri .

(ORIGINAL _ B) .

----------------ORIGINAL _ B

= SLIMY_EST

page 54

Appendix

as well as

oarfu\ - i ----------------." '. ' oRlctNAL - B-- (our)GorNG,

emanating(verb)

and

oaru - ng___-..,-, sourHERN - B

= soutce, teat,spring(noun)

which is the basis for

"gor: ey", (" juice" = squirt)Capt.J. Hunter 1793, eora .

"garr: ee ", ("to cough")W. Dawes 1790, eora.

( coastal ORIGINAL-B)

"kuuruu: m", ("teat" )J. Dawson 1881, chaap whurrung

(SOUTHERN_B)

Iun - (dula: n) - ga(ru): wur(u): i

ORIGINAL _ AORIGINAL - B

TURUWAL

" yoon',,, go I wa.rt r I ee"" yoon' ,,' Q : aef : ee "

(" a giver, to give" )

M.M. Everitt 1900, gundungara

(ORIGINAL _ A)

"rJ t: A. :: i " , (" give" )

J. Malone 1875, botany bay

(TURUWAL)

some-thing going-away = GIVING / OFFERING

as ln

" ngung zzz Q2:: ", (" give" )

R.H. Mathews lQO4. wiradjuri

( inland ORIGINAL - B).

"{) ;zzzW;i", 1" gtve" I

W. Dawes 1790, eora

(coastal ORIGINAL - B) .

some-( thing) 8otn8-away

page 55

ofsome

run ga(ru. i)\-------V--------

- Soing

" :;to :. !2 | ", (" to give" )

J. Rowley 1875, cowpastures

(coastal ORIGINAL - B)

",,dll ,,92, ", (" give" )

R.H. Mathews 1903,

" thurrawal" = ngunawal

(HUYGENS CIRCLE)

Re v.J. Bulmer 1876,

" yllg ; wA :': n:: "R.H. Mathews 1902,

(SOUTHERN _ B) .

of

" wo, ka,, k, i,, fi ", (" give" ),

kurrn kopan noot

" wo,, ka t Q,: Q€: " , (" give" ),

chaap wuurong

J. Dawson I 881

"u,ca,,thot:", (" gave I" )" ll, ca, 17 : a ;:, ", (" I give" )

Re v.J. Bulmer 1876, kurnai

(SOUTHERN _ B)

Appendix

- wu(ru)\---------'!-/

away

ORIGINAL _ BSOUTHERN _ B

HUYGENS CIRCLE

kurnai .

, (" give" )

kurnai

-thing

- dula, n -

= GIVING /OFFERING

wu(ru)t ga(ru) - nzdu(la) - i - t@n)

ORIGINAL _ BSOUTHERN - B

HT]YGENS CIRCLE

" z Q3 : Ilrd, i, fi ", (" from" )

R.H. Mathews I901 ,

" thurrawal" = ngunawal

( HUYGENS CIRCLE)

" we z a; 17::; " , (" give" )

D. Paine 1794 , eora

( coastal ORIGINAL - B)

_ GIVING /OFFERING

page 56

Appendix

Aboriginal man sharing culture with the kidspicture by Ainsley Roberts

- "-"^^r-:)^t - t

and also (see page 63)

wir(u)\""""""v--'_J

improprietous,

unfortunate

" yo r ka :: IlLn z 98 t ", (" eive" )R.H. Mathews 1907, birdhawal

( coastal ORIGINAL - B )

...--- \\' tl)

wu(ru): ga(ru z i) - pun - du(Lat n)

wu(ru): ga(ru t i) tflun: du(la: n)

----)

ORIG]NAL _ B

" parr, bu : g ::g :! : ", 1" I hne lostit" )

W. Dawes 1790, eora( coastal ORIGINAL - B)

= GIVING /OFFERING

GIVING / OFFERING

_ LOSE _ ING

page 57

Appendix

At the opposite extreme we have

gura\_-__v---_J, =TOWARD,

in-to, cavitous(adjective)

kurui - gara : kara, (molluscan shell);

koro (inner ); japanese.

(oafuy )rcapt, (" snail" = tubuktr shell);

X€LpoD( pyoE), (" surgeon" ); sreek.

gure, (" mouth" ); rumanian.

" ku ra " , (" throat" )

L. Malone 1875, wodi wadi .

" kurra", (" ear" )

J. Rowley 1875. cowastures .

" goora", (" drown" )

P.G. King 1793, eora .

" goorA ", (" sunk" )

D. Collins 1798, eara

(coastal ORIGINAL - B)

" gora I i ", p.c. King 1793 .

" gora t i " , Capt.J. Hunter 1793 .

" Qot : eE ", D. Southwell 1788 .

" Qoor: ee " , W. Tench 1788 .

" gor z ey " , D. Paine 1794 .

(" the ear" = the receptacle for sound ), eora

( coastal ORIGINAL - B )

"kurti", ("theear" )

R.H. Mathews 1901, dhamtk

(" ORIGTNAL - A" ) .

" gerrd : mah", ("devil" )I. Nathan I849, wodiwodi.

"go: mira ", 1"o hol""1Capt.J. Hunter 1793, eora .

( coastal ORIGINAL-B )

----------------ORIGINAL _ B

al so

oura-I ----------------jI.......\,-J UNIVERSAL

= (rN)coMrNG,receiving, "hearing"

(verb)

" Qur: i", (" ear" )

R.H. Mathews 1903,

" thurrawal" = ngunnawal .

( HUYGENS CIRCLE)

"erzi",(" canoe" : receptacle for people )

R.H. Mathews 1902, kumai

(SOUTHERN _ B)

I arg e-cavity= mouth, hole

and

guta- mura\--.v_-_.._____.___

_______---___-_-+ORIGINAL - B

page 58

also the repeat noun

or more conscisely

Appendix

-------)

QUrz : n-bul(ala) -------+

or

gu(ra) , b(ulala) -

" ko b. bara t " , (" head" )

R.H. Mathews 190l, dharruk.

(oRrcrNAL - A)

"ttbraiz n", english

" ka,, besa z ", spanish

"K:6LzItO tg", greek

"no'. gra: ", ("nose" 1

J. Rowley 1875, cowpstures( coastal ORIGINAL-B )

" ni : gir: " , 1" nose" )J. Malone 1875, botany bay

(TURUWAL)

nORIGINAL _ A

ORIGINAL _ B

"ku:b:bera:", (" head" )

W. Dawes 1790 eora

" ca ;: bera : " , (" head" )

P.G. King 1790 , eora ,

D. Collins 1798, eora .

" ko, b, bina , ", (" forehead" )

J. Rowley 1875, cowpastures .

(coastal ORIGINAL - B)

OI

bu(ra) , b(ulala: n)': QUra ------+

"bu: k:: kurae", ("between" )kuurn kopan noot, peekwhunong .

"bu : k :: karyu ", ("between")chaap wutong

( SOUTHERN _ B ) ( O RIGINAL_ B )

n : SUra-(bulala\ -------)

also brain 1= [thing] between [the] ears)

= behveen

bura

"kooru: n: kooru: n", (" pelican" )"kur: kur: ", ("a mouth or beak")

J. F. Mann 1810's. kuringai( coastal ORIGINAL-B)

" jerra: n: gill", ("trousers")Pettit & Dawson 1850's, kurnai

(SOUTHERN_B)

lUfl z 11- QUI7: 11

-- beak- pair

= tube-s

= tube-s

tube-s -- "ears"

= brain

page 59

Appendix

w1J7:w1J7:0 ar

Rs,

hence to think / remember

.wry1= wring

s

-?

" tt bulala t " , (" forget" ??)

R.H. Mathews 1901, dharruk.

(ORIGINAL _ A)

= tube

.mura - gura,(n) -= greatly

ORIGINAL _ B

" wlnnt garah::z ", (" to think"

W. Dawes 1790 eora

(coastal ORIGINAL - B)

also gastropod/snail shell (= greatly coiling tube)

1:IJZIA

= coiling

--------------+TURUWAL

-- brain

page 60

Appendix

"ma'. kour:: i: ngre", ("murex tuberculex, pI. 36")"me: koura:: ra ", ("actine" -- bubble shell, BuIIa quoyii)"mara::'.: ", ("nerite noire, pI. 65" = black periwinkle)

J.P. Gaimard 1826, jervis bay-_------JTURUWAL

or

gun, g(un)\--v-l

= extremely

as lnsome

.nun

QUfEzn - iq(ara)\""""""\'-__--..,

= coilingORIGINAL - B

TURUWAL

" coun: g I coura*n9",(" shells" )

M. Munima 1863, wodi wotli .

" gun I gam r: ng" ,

(" the volute Cymbiolena magnifice" )

T. Watling 1790, eora .

(coastal ORIGINAL- B)

also the important form

s ome -thing coming -near

" koun r: kourou, n;: ",(" pls. 52- 53" , cone shell Conus sanguinoL) .

" koun ,,t QorJforJ ;:: " ,

(" pl. 43" , casque frangd shell Cassis fimbriata)

" ; koro ::: ng" ,

(" peign, pl. 36" , spilcy cone murex )

J.P. Caimard 1826, jervis bay

(TURUWAL)

= RECEIVING

thing / item

/ REQESTING

(out(a) ' i\

li, gur@)) - du(La: n) - . ruin......1'-

= RECEIVING / REQUESTING / NEEDING

----------------9UNIVERSAL

" kit ; pur ,, ta,:: g' nin" , PW

" keut t kar t; ta r. wan" , KKN

J. Dawson 1881, (" ask" )

" ngun t da:, dha:. " , kurnai

R.H. Mathews 1904, (" bee" ) .

" ngan Ioo t k:.: ", wallithica

D. Matthews 1874, (" to want" ) .

(SOUTHERN _ B)

" yoon,: !'r ee * " , (" give" )

M.M. Everitt )900, gundungara

(ORIGINAL_ A) .

R.H. Mathews 1904, ngunawal

( HUYGENS CIRCLE) .

" t Qzf t i;: n" , (" to obtain" I" : Q ,: 1 ::: " , 1" please ( pray)" )

W. Dawes 1790, eora

( coastql ORIGINAL - B )

= tube

co mrn I

page 61

IUn - pin: gur(a) - du(la : n) - i

" dyung :: ga z dya z: i" , (" request" )

R-H. Mathews 1904, ngunawal

( HUYGENS CIRCLE)

" mo ,, i t dia : " , 1" give it to me" )

J.F. Mann 1840, s, eora

" maha ;=t n" , (" take" )

J. Rowley 1875, cowpastures

" mahzzz: n" , (" take" )

D. Collins 1798, eora

" kara zz r ma " , (" steal" )

R. H. Mathews 190I , dhatuk

(ORIGINAL _ A)

" kur , t ,:: ba" , 1" take" IR.H. Mathews 1902, kurnai

(SOUTHERN - B)

Appendix

or

ORIGINAL _ B

HUYGENS CIRCLE

"antnezeer:::",P.G. King 1793, eora

( coastal ORICINAL - B)

also the more severe notion of " enhanced recievement" as in

mu(ra - eura)zi - du(la) ' n

\-v-,thing

----------------+ORIGINAL _ B

or

" manl :t: n" , (" to take or get" )

D. Paine 1794, eora

" maa z;; n" , (" take" )

W. Dawes 1790, eora

D. Southwell 1788, eorcr

(coastal ORIGINAL - B)

----------------)UNIVERSAL

" carfahrtt, trl\", (" stealing" )

D. Collins )798, eora ,

P.G. King 1793, eora

" kara :zzz ma" , (" steal" )

J. Rowley )875, cowpastures

(coastal ORIGINAL - B) .

" garra iii; " , (" catch, stop, hold or take"

C. Richards )902, wiradjuri

( inland ORIGINAL - B)

gura - d@La'n - i) - mu(ra)

.= RECEIVING / REQUESTING / NEEDING

greatly inco min g

= TAKING I STEALING / DEMANDING

= TAKING / STEALING I DEMANDING

pa9e 62

Appendix

or

mu(ra , dula) , n, (gura)

" ma :: n: i; au" , (" take" )

R.H. Mathews 1904, dhanuk

(ORIGINAL _ A)

" ma t nz i: ne" , (" take" )

W.J. Enright 1900, pt. stephans

( coastal ORIGINAL - B)

wunproprietously

.fonunateLy

----------------ORIGINAI, _ B

(i ' our(al\mu(ra)-d@la),n-l I -----+

\gur(a) ' t.) UNIvERSAL

. =TAKING / STEALING / DEMANDING

" muu t tch zz a z ka" , chaap whurong

" ma z: nn z a : kae" , kurn kopan noot

J. Dawson 1881, (" take (this), get" )

(SOUTHERN _ B)

and also (see page 57)

" ma :: n:: wLtfe ", ("to find")W. Dawes 1790 , eorah(coastal ORIGINAL - B)

- COLLECT

- FIND - ING

page 63

Appendix

garu / gura

A simple example is

garu

= squishy, fluffy

" garu", (" nasal mucous" )

D.K. Eades 1976, dhurga

(SOUTHERN _ B)

" kurru", (" clouds" )

J. Malone 1875, botctny bay

(TURAWAL).

R.H. Mathews 1903,

" thurrawal" : ngunawal

(HUYGENS CIRCLE)

as opposed to

gura

= fitm, hard

ORIGINAL _ BSOUTHERN _ B

TURUWALHUYGENS CIRCLE

" gaffu", (" marsh" )

Re v.J. Gunther 1892, wiradjuri

(inland ORIGINAL - B) .

" kurru " , (" clc,ud" )

L. Malone 18'75, wodi wodi

" (nd)qaril(g)", (" snot" )

W. Dawes 1790, eora

( coastal ORICINAL - B)

" kari" or " karri"Eucalyptus cl iv e rs ico lo r, tre e

with the hardest known wood.

T. Wood 1934 .

" jarrah", E. & J. Trautman 1980

" djarry" , G.F. Moore I842

Eucalyptus robusta / m arg inata .

( PERTH REGION )

" kuru(willang) ", (" hail" )

R.H. Mathews l9Ol , dhanuk

(ORIGINAL _ A) .

" gAralbang) ", dharawal .

" bura", dhurga .

(" stone or rock" )

D. K. Eades I 976

(SOUTHERN _ B)

ORIGINAL - A

SOUTHERN _ B

pa5e 64

Appendix

gula / gulu

The common root is

gula

= shiny, vitalradiant

(adjective)

" kull€r", (" gctod" )

J. Malone 1875, bo tan y bay

(:TURUWAL).

rcaLo, rcahLo ( female beauty);

rcal"a (well, alright, properly );

" killi ", (" spark, candte light" )

R. H. Mathews 1903, eoroh

(coastal ORIGINAL - B) .

" galah " ,

(the cockatoo, Eolophus roseicapillus )

yuwaalaraay, northern NSW .

" killu" , (" clear, shining" )

Re v.W. Ridley 1875. kamjlaroi

(ORIGINAL _ A)

ORIGINAL _ BSOUTHERN - B

" kal : gall " , (" noisy, as in Cicada " )

J. Dawson 1881, chaap whurrong

(souTHERN - B)

"gilzgil ",

(" bore, place name meaning ' flowing away'

A.E. Matin 1943

ORIGINAL _ BORIGINAI. _ A

TIJ RIJWAI,

")

ydld ( blue slcy ); greek

gulio, latin "

G.F. Angus 1847

and the repeat word

gul(a) - gul(a)

= MDIATING - EST.

EMANATING _ EST(superlative adj ective)

" gall: gall" ,

(" contagious, as in smallpox" )

Capt. J. Hunter 1793, eorah.

" Qul : Qul" , (" smallpox" )

J. Rowley 1875, cowpastures

( coastal ORIGINAL - B)

page 65

and

hence the words

gul(a): i.

\-------------- -------------l

= shining, radiating,

[Iight, vitali4t, sound ](verb)

" kill t i" , (" shine, spark " )

R. fl. Mathews 1901 , dharruk

" gill i", (" moon" )

Re v.W . Ridley 1875, kamilaroi .

(ORIGINAL - A)

gul(a), i - 4uru

and also the nouns

gula, (i)\-1-l

= radiating

= twinkling, sparkling(verb)

" kul 'r4ura", (" moon" )

J. Malone 1875, botany bay .

(TURUWAL)

Note : A.W. Reed [ ] 969, pages 6 - 8l complained

of the 22 different ways that wirad.juri sources alone

used to spell water r it is not science he claimed.

"galtitng ", ("water")

Re v.J. Gunther 1840's, wfuadjuri

( inland ORIGINAL - B)

" QUll t e: n ", (" aswamp " )

C. Mqcqlister 1907, midthung

( HUYGENS CIRCLE)

"galla:: n", ("the sun" )Sir Joseph Banks 1770, eorah

(coastal ORIGINAL- B)n -------+

thing(noun - stem)

nuru n - sul(a\:ithing

(noun-stem)

Appendix

ORIGINAL _ AORICINAL _ B

"kalt i ", ("water" )

" kiil , ", (" urine " )

McNicol & Hosking 1994, ngiyampaa

" kil' ", (" urine " )

R.H. Mathews 1904. wiradjuri

( inland ORIGINAL - B)

ORIGINAL - BTURUWAL

HUYGENS CIRCLE

. "baro: n'. gl i e", t"vein")

-------) D. Collins 1798 , eorah

(coastal ORIGINAL- B)

= throbbing

page 66

Also the augmented adjectival

Appendix

form

-------)gula:(i) - gun " yella:: gan ", (" sood")R.H. Mathews 1908 , ngarigu

(HUYGENS CIRCLE)

as ln

gun z gula, i\-------------- ---------------

very shining = NICE

"gun:gahLiin ",

n|_v-,

place / thing(noun - stem)

HUYGENS CIRCLE

"goon:garlii:ne ",historic homestead buih and named bythe sportsman william davis jr in 1860's

"the Queanbeyan Age", 30th Jan 1873 ,

and the form (see pages 85 and 95)

"kan'.gahoo::n "myrtle, a hardwood used

Sir W. Macarthur 1861,

(gunzgu)lazi- n , rmJl

the

----------------SOUTHERN _ B

providing the logical basis for GOOD / CORRECT as in

",lati:Ij r erj", (" good" )

Pettir & Dawson 1850's, kurnai

":Iratfit ", (" good")

R.H. Mathews 1902, kurnict

": 12:: nz e ", (" good " )

Re v.J. Bulmer I 876, kumai

" man:dhan : g t ", (" right"

R.H. Mathews 1908, ngarigu .

W. Govett 1828 - 32, ngunawal .

( HUYGENS CIRCLE)

0unYES

GOOD / CORRECT

" na:guntQr ", (" good")

D.K. Eades 1976, dhurga

(SOUTHERN _ B)

gun I gula. (i)

NICELY

)

)

very shining - NICE

= NICE place

page 67

Appendix

or" icol: nea ", (" sood" )

Kitty Cooper 1870's , moira(SOUTHERN _ B)

- BAD / WRONG

"cttrg tQilr ", ("bad, ugly" )

T. Rayment 1933 .

" wong ' kal , ke t i ", (" foolishness" )

" WoI1! : kdl ,, " , (" to be afool" )

Rev.L.E. Threlkeld 1831, awabakal

(coastal ORIGINAL- B)

ORIGINAL _ AHUYGENS CIRCLE

SOUTHERN - B

"gan::i.na", ("bad" )" gaI ::: nAng " , (" bad " )

D.K. Eades 1976, dhurga

"QUIItitn3.", (" batt " )

H. Hale 1846, dhurga

" COfn; €: D4 " , ("bad")J. Larmer 1853, dhurga

" ghur;: rang ", (" bad " )

W. Bothong 1900, dhurga

" tintt, ban " , (" bad " )

Pettit & Dawson 1850, kumai

(SOUTHERN _ B)

SOUTHERN _ B

" nanyn.: galon: Q : ", (" bad " ), queanbeyan

"narre:golan,g, ", 1" bad " 1, yass

"tu., k' kayil: ", 1"no"1

E.M. Curr 1887, ngunawal

( HUYGENS CIRCLE )

"tu, g, gaelee: ", 1"no"), batemans bay

"tutg,gr i", ("ro "), ullad.ulta

J. Larmer 185j, dhurga

(SOUTHERN _ B) .

NOT

or

(gun),gula.i- qun +

apposed to BAD / WRONG

gunzg(uLa)'i - nug

" eun,r i : Ite " , (" excrement" )

" glfr :z: nung " , (" bad" )

R.H. Mathews I903,

" thurrawal" : ngunawal

( HUYGENS CIRCLE )

" Qttrdt ; ba " , (" U"a" )

M.M. Everitt 1900, gundungara

(ORIGINAL_ A)

._ GOOD / CORRECT

NICE

= BAD / WRONG

page 68

Appendix

Stronger meaning was attained

= extremely shining

" queah: Qan: Q ", gundungoro

(" the name of Pr incess Betsy " )

(ORIGINAL_ A)

I. Nathan 1849, wadi wadi

( coastal ORIGINAL - B )

, " ke: kdl: ke i ", t"r*eer, nicc. ptcasan!")---------) Rev.L.E. Threlkeld 1834, aw;bakal

(coastaL ORIGINAL- B)

gula:i - gun.g(un\ORIGINAL _ AORIGINAL _ B

HUYGENS CIRCLE

" yella = Qan: e ", (" rejoice" )

R. H. Mathews 1903,

" thurrawal" = ngunawal

( HUYGENS CIRCLE)

or

gu(n) t gul(a): gu(n): i

as in EXCELLENT / INSPIRED / MAGNIFICENT

extremely - shiningLy

ORIC]NAL _ BORIGINAL _ A

HUYGENS CIRCI,E

or

R. H. Mathews I903,

" thuryawal" : ngunawal

( HUYGENS CIRCLE )

" nu: k ,,t kttnr g " , (" eood " )

L. Malone 1875, wodi wodi

(coastal ORIGINAL- B)

" ya, d' r,dun, e ", (" good " )

M.M. Everitt 1900, gundungara

(ORIGINAL_ A)

----+ " :i gui: guii Qfiit ", t"excettenr" tE. Dunlop 1840, awabakal

(coastal ORIGINAL- B)

page 69

(gula t i) z gu(n). gu(n) - qu(n)

= extremely shining

gu(D g@la: i): QUn: g(un)

YES

= EXCELLENT / INSPIRED / MAGNIFICENT

= EXCELLENT / INSPIRED / MAGNIFICENT

as opposed by FOOLISH / ABSURD

(gula, i); QUn: gu(n) -extremelt

= FOOLISH / ABSURD

HUYGENS CIRCLE

nu! - (gula: | : QUn : gu(n\

" WAII * QAflt g " , (" wrong " )

R.H. Mathens 1908, ngarigu

( HUYGENS CIRCLE)

"un:*don:g",Re v.L.E. Threlkeld 1834, awabakal

(coastal ORIGINAL- B)

ORIGINAL _ A

ORIGINAL _ BHUYGENS CIRCLE

Appendix

nu0NOT

'.i gumt aia: nunQ", (" foolish")R.H. Mathews 1903,

"thurrowal" = ngunawal

or

----------------UNIVERSAL

" um*bun z gai ",R.H. Mathews 1901, dharyak

( ORIGINAL _ A)

" cal]ooa " , ("to climb" )

Capt.J. Hunter 1793, eora .

"k la" or"kolle", ("bear")

J. Rowley 1875, wodi wodi

(coastal ORIGINAL- B) .

" guulAy" , (" net bag, fishing net" )

McNicol & Hosking 1993, wiradjuri

( inland ORIGINAL- B) .

" ka]ua" , ("to ctimb" )

R.H. Mathews 1901, dharruk

(ORIGINAL_ A) .

= FOOLISH / ABSURD

At the opposite extreme we have

gulu

= leth arg ic,

dull, fossiliTed(adjective)

" gooleur " , (" koata " )

N.D. Pettit & W. Dawson 1850,

" kullah " , 1" koata " )

Re v.J. Bulmer 1876, kurnai

(SOUTHERN _ B)

" gula ", (" native bear " )

R.H. Mathews 1904, ngunawal

( HUYGENS CIRCLE)

page 70

Appendix

as rn

There would also have been gulu-iverb which is conspicuous by itscomplicated names for Mary Everitt's

" kula:: maI7", 1"koata")R. H. Mathews l90l, dharruk

(ORIGINAL-A)

" culla:: wine",Historical Records ofNSW iii, pS21

(coastal ORIGINAL-B)

meaning "loafing" or " la4t" , aabsence, and there are quite

red handprint caves

gulu:(a) - .mu7,the

-------)l-----v---------

= " koala"

... the furry, nocturnal, tree-living, marsupial Phascolarctos cinereus. ltwas popularly viewed as a monkey till at least 1872, a sloth till at least1886, and a cuddly koala bear in the May Gibbs books since 1917.Vocalic metathesis has reversed the vowels.

murula t n : Q(ula, i) - gulu : g@ru) : QUn: g(un\

= marks

" murrolu'. n'. q::eulu: ng :'.", ("red hand stencils in cave" )-------------7 w."Ru;selt t9 t41 guniungara

Thiswith

" murrolu'. n: '. : : '. gun'. e", ("cave with hand stencils" )Jimmy Pippen 1890, gundungara

(coastal ORIGINAL-B)

sound again offers a plausable spectrum, but not in accordanceoctahedral symmetry. We will return to this later.

picture by Ainslie Roberts, 1962

{'nt

= hand (see pages 35 - 36)

page 71

Appendix

gula / galu

ORIGINAL _ AORIGINAI, _ B

There is

" kular " , (" anery " )

R.H. Matlrcws 190I , dharrak .

" gului " , (" ansry " )

R.H. Mathews 1903, dharkinyung

(ORIGINAL_ A)

hence the verb

gul(a): i\_v_/

= arg uing, fightingconjuring, sulking

(verb)

and the repeat noun

gula q gula:gt----v.-------- \-----------.v-l

harmful - creature harmful - creature

= [swarm of ] LUCUSTS(animate Imultitudinous] plural)

a word

gula\-.v-_____J

angry, harmfullquarrelsome(adjective)

" gulaa " , (" anery " )

J. Ellis 1994, wiradjuri

( inland ORIGINAL - B)

" ghoul " , (" malevolant ghost " ), english

" kul t i " , 1" an argu*ent, fight" ), nyangumarta, NW Australia

" ku] t i " , (" anger, wild" ), guugu - yimidhirr, NE Australia

" ku] , i " " (" orgry, wild, quarrelsome" ), wirri, E Australia

" kul . i " , 1" anger, wikl" ), warumungct

" kul : it " , (" temper, anger" ), madi - nr.adi, SE Australia

" kull , ia ", (" cut" ), R.H. Mathews 1903, " thurrawal" = ngunuwal, HC

----------------+UNIVERSAL

-------)"gulazng.gulaIng"(" I arg e locustfs ]" )" thurruwal" 1903 = ngunawul .

" gula I nt gula I n"(" large locust[s]" )

ngunawal 1904 . R.H. Mathews

( HUYGENS CIRCLE)

page 72

also

Appendix

mu(ra) - gula - ----)greatly harmfull

based on the adiective

(" ma , eulu ,

l"gura'galaI r" clever man"

ln.a. uo,n"-r.

TJ

creature

IlQ", ngunawal 1904

: 11", ngariga 1908 .

= sorcerer, kooradji)

( HUYGENS CIRCLE)

" kila z la" , w. Dawes 1790, eora.

" ghoola ; ra " , w. Dawes 1790' eora

" Qula z ra " , P.G. King 1790, eora .

" goola I fa " , D. collins 1798, eora .

" kila z ra", J. Rowley 1875, wodi wodi .

(" angry, cross, displeased, ill - natured" )

( coastal ORIGINAL - B)

" maiat gut IllJ ", (" spear" )--.------------\- 7 L. Malone 1875. wodi wodi

(coastal ORIGINAL-B)

ORIGINAL _ B

with repetition expressing degree of severity

gula z m(ura) gula: m(ura)

malevolanl

also barbed reed-sPear

mu(ra) z gu(la)

malevent =

or

.J7U(U ,gu(la) | mu(ra)

emptY

" reed spear"

p@ru)\-----v-l

empty

= " reed spear"

= malevolent / treacherous creature

gulaz(mu)ra

= malevolant(adjective)

" gula, rra", (" angry" )

J. Bayliss 1927, witadjuri

( inland ORIGINAL - B)

malevolant

" gulu:m.guluim",| " stranper. esneciqllv danPerous stranger. a wild black fellew" )' we rgia

(";bod;il(] black fetlow who would kill people")' madi mqdi

L. Hercus 1986, Pages 202 & 223

" nooro: ca: my ",

------+ (" I a long ] spear with one barb" )

D. Collins 1798, eorah

(coastal ORIGINAL-B)

= MURDEROUS-ESZ (superlative adjective)

page73

Appendix

At the other extreme

7gatu

= happy, helpful,harmless

(adjective)

= harmless

----_-----------SOUTHERN _ B

"galu" o,"karlo",("white facetl heron" )

R.H. Mathews 1902, knnai

and mud crab (see picture on page 53)

" gali: ga", ("uab")

= fat

or

Furthermore the opposing senses are captured in

and

galut gun\--------------_/

very - helpful

= permlSSIOn, acceptance

and I'm sure there was a verb galu-i meaning "celebrating, helpfult",as apposed to gula-i meaningapproximately 9 words of the formthough the lone pair gula/gulu doesWe will return to this later.

baru ------+

side

" gala t gan t bar'ra " ,

("wtpe the dit off the feet" )

Rev.J. Gunther 1892, wiradjuri

(inland ORIGINAL-B)

"sulking". Thus we anticipatedg***, and immediately have 10,not follow Octahedral Symmetry.

guru - gal(u) -----t= fat

gula

unhelpful"

harmful

= harmless

baru ------+

side

" gula I birra ", ("refuse, reject")

C. Richards 1902, wiradjuri

( inland ORIGINAL-B)

D.K. Eades 1976, Dhurga

(SOUTHERN- B )

" yara'. gal" , (" mud crab" )

First Fleeters 1790, eorah

(coastal ORIGINAL - B)

" gari: ga", ("ffab")

D.K. Eqdes 1976, Dhurga

(SOUTHERN_B)

= refusal, rejection

pagel4

The next candidate is

Appendix

buru / buru

ORIGINAL - B

SOUTHERN _ BHUYGENS CIRCLE

" barru " , (" tired" )

buru\----Y------J

= slow, quiet,

tired, tranquil

(adjective)

" burro ", (" stitL" ), kaurna

Teichelmann & Schurmann 1840

(ADELAIDE)

" birra", (" tired" ), wiradjuri

McNicol & Hosking 1994

( inland ORIGINAL - B)

hence the superlative adjective

buru - buru\_-v--------------

IAZY _ EST, QUIET _ EST,

SLOW - EST(superlative adjective)

as well as the augmented forms

mur(a) - burugreatly slow /quigr = " gliding"

( adiective)

" bernt(ng) ", (" mist" )

" thurrawal" = ngunawal

R.H. Mathews 1903

(HUYGENS CIRCLE)

" burcay" or"burral", dh,,so

(the name of Shoalhaven River)

W. Bothong, cited by A. Campbell 1899

(SOUTHERN _ B)

" birra: birra", 1',taay"-est 1

McNicol & Hosking 1994, wiradjuri-------) (inland ORTGINAL-B)

" burro : burro", ("bye andbye")

Teichelmann & Schurmann 1840

( ADEI-AIDE)

" mier : barrar", ("smooth")

------+ J. Banks 1770, eora

(coastal ORIGINAL-B)

page75

gun - buru -------+= very slow / quiet

( adjective)

" jan, boara" , (" stow" )

" thurrawal" = ngunawal

R.H. Mathews 1903

( HUYGENS CIRCLE)

al so

burulazy I part of ]

"br,:itgela: ", (" sleep")

J. F. Mann 1840's, eorah.

"b(trrazir ", ("nieht" )

I. Nathan 1849, wodi wodi .

" bourra zz: " , 1" stq" 1

D. Collins 1798, eoroh.

( coastal ORIGINAL- B)

and

Appendix

" koong z burry" , (" steep" )

M, Munima 1863, wadi wadi

" koon t bury" or" koong I burrie " ,

(" awhtrlpool or eddy, a bay within L. Illawarra" )

F. McCaffrey 1920, s, wadi wadi .

(coastal ORIGINAL - B)

" but , i: ooloo z " , l" night time" )

M.M. Everitt 1900, gundungara .

" bera,, ck z " , (" dead" )

W. Russell 1914, gundungara .

(ORIGINAL- A)

" moura ,, k , " , (" btotk" )

J.P. Gaimard 1826, jervis bay.

" purru :; " , (" night" )

J . Malone 1875, bo tan y bay .

(TURUWAL)

ORIGINAL_AORIGINAL-B

TURUWAL

hence the adjective

buru : (gula z i zrJuru) buru, (gula r i :rJuru)

= BLACK-EST(superlative adjective)

Ithe daily ] shine cycle

__ NIGHTTIME

page76

andfull moon

muralarge

ORIGINAL- BSOUTHERN _ B

Appendix

" boo ru ::::bootaii: ", ("btackest")

M.M. Everitt 1900, gundungara

wilt du(Ia, n) buru- g(ula) t i z (guru)

orbiting thing = MOON

ORIGINAL_ B

" yanna t dah z: para z g : i : ",("new moon" )

P.G. King 1793, eora

mid. half

Re v.F.A. Hagenauer 1876, kumat

" ngerru z tzzz 17" , (" moon" )

W. T hom as 1860, kurnai .

" nerra z:::::17" , (" moon" )

Re v.J. Bulmer 1876, kurnai

" naffu zz::z n" , (" moon" )

N.D. Pettit & W. Dawson 1850, kurnai

(SOUTHERN _ B)

ORIGINAL _ A

hence a hew (= black) moon

as distinct from half moon

" ngirra z ng 2i:it " , (" moon" )

SOUTHERN _ BTURUWAL

" inarra: I7Q ::z:z " , (" half a moon" )

J. Lingarrl 1838, ngunawal

( HUYGENS CIRCLE)

" narra ; D! zz;zz " , (" anything small" )

P.G. King 1793, eora .

(coastal ORIGINAL - B)

" murray :z:: yanna.: dah t ",P.G. Kine 1793, (" full moon" ), eora

(coastal ORIGINAL-B)

" meheeaf I ong *rt ", 1" full moon"l

J. Dawson 1881, kuum kopan noot

(SOUTHERN_B)

- g(uru: it gula)-wiJt,du(la, n)

between shine - c,tcles = black

= lunar cycles

= Lunar cvcles

pagell

or

Appendix

mura. - du(ha: n) : win- g@la t i:rJuru)Iarge

SOT]THERN _ B

" mirtae . ta ! nyuu z k :: ", (" full moon" )

J. Dawson I 881 , chaap whuurong

(SOUTHERN_B)

At the opposite extreme we have

buru\___v._____,

= quick, noisy, vigorous( adjective)

" burrai" , (" wallaby" )

Re v.W. Ridley 1875, kamilaroi

" baro" , (" quick" )

RH. Mathews 1901, dhatuk .

(ORIGINAL_ A)

" burra", (" quick, rock wallaby" )

RH. Mathews 1904, ngunawal .

(HUYGENS CIRCLE)

and the adjective

buru - buruFAST _ EST, NOISY _ EST

( superlative adjective)

" buurray" , (' fart" ), wiradjuri

S. McNicol & D. Hosking 1994 .

" burrai", (" quick" ),

" burai" , (" chitd" ),

R.H . Mathews 1904, wiradjuri .

( inland ORIGINAL- B)

" barrao" , (" make haste" )

" beria" , (" sin g" )

J. Rowley 1875, cowpastures .

( coastal ORIGINAL- B)

" burra , burri", ("be quick" )

N.D. Penit & W. Dawson 1850's, kurnai

(SOUTHERN.B )

ORIGINAL _ AORIGINAL - B

HUYGENS CIRCLE

-------+ "booroo z boorya", 1',gun',1

There may even have been a yerb buru-i which meant "hurrying,

bustling", in opposition to the verb buru-i meaning "loafing, v)asting-

= lunar c,tcles

time, sleeping"

page 78

Consider the wordtDATU

\""""\i-l

= either - side, oppo sin g

( adjective)

" birra" " (" cheek" ),

Appendix

baru / bura

ORIGINAL - B

TURUWALHUYGENS CIRCLE

" peerah", (" cheek" ),

M. Munima 1863, wodiwodi .

" birra", (" the cheek" ),

Capt.J . Hunter 1793, eon.

" brua" , (" near to" ),

W. Dawes 1790, eora .

(coastal ORIGINAL - B)

" bera", (" joue" = cheek)

J. P. Gaimard 1826, jeruis bay .

(TURUWAL)

and the word "belly-button"

= Centre,

middle

mu1,the

- .Iura,ORIGINAL _ AORIGINAL _ B

HUYGENS CIRCLE

" mllm z birri" , 1" navet" 1

R.H. Mathews 1901, dharruk

(ORIGINAL_ A)

": nyirra" , (" navel" )

R.H. Mathews 1903,

" thurrawal" = ngunawal

( HUYGENS CIRCLE)

" (dhaum)bara" , (" spread apart" 1

R.H. Mrthews I903,

" thurrawal" : ngunawal .

(HUYGENS CIRCLE)

p|tpo ('EOo ), (" arm" ), greek .

bra(cchium), (" arm" ), tatin.

bra( ziere), y,encn .

brea(St), engtish.

" moo z neero" , (" navel" )

W. Tench 1789, eora .

" min : duru" , (" navel" )

" mu : nuru" , (" navel" )

W. Dawes 1790, eora .

" moo: ndrdh" , (" navel" )

D. Southwell 1788, eora .

" moo z nuffo", (" navel" )

D. CoLlins 1798, eora

(coastal ORIGINAL - B)

page79

Appendix

hence (points either-side of centre --) "ends"

= either - side of centre" ka, buru t r1 ; ", ("end[s]")

-------) Rev.W. Ri-dley 1875, kamilaroi

(ORIGINAL_A)

":beraitberai"( " constelhtion of Orion" ),

" ga :: n ,: bir", ("uoss")

Rev.W. Ridley 1875, kamilaroi

(ORIGINAL_A)

.nu(ra)'- baru - n: (bulala)

po int[ s ][dual noun - stem]

either - side of centre

= dipole

o

x

o

o:

....'.,x,.....'

o

O.'""x"""'O +

dipole

dipole quadrupole

Thus we describe pairs of points, laying as dipoles, about the centre of

o.....'x.'.....oJzura

and the repeat expression " cross"

Iu(ra): baru - p@ra): baru ------)either - side of cente

= dipole

= quadrupole

mass JTU(A

page 80

Also we may describe a line by defining a new constant

.J1Af2: fl,= edge, border,

surface(noun - stem)

": nirri z n" , (" edse" )

Re v.W. Ridley 1875, kamilaroi

(oRrcrNAL- A)

": yiri: 11" , (" scates offish" )

R.H. Mathews 1904, wiradjuri

( inland ORIGINAL - B)

Appendix

.ml.n,' the

"matnarazn",hair omament made by sticking

kangaroo teeth in hair with gum,

P.G. King 1793, eora .

" yin: nerry t " , (" eyebrow" 1

D. Collins 1798, eora .

" z4ari: 11", (" eyebrow" )

W. Dawes 1790, eora .

" : 112ffa : fl" , (" eyebrow" )

D. Southwell 1788, eora .

(coastaL ORIGINAL - B)

or a pair of parallel lines - a double edged sword

( para-buL(ala)\ttlbu(lala) - na(ra))

ot the banks of a river

Jlzrz z tbulala) --------+

= edges, borders(implied dual noun - stem)

baru= oppo sin g

" karra ' bil"

" bu t nya"" sword's edge, literctlly the

back of the sword"

W. Dawes 1790, eora( coastal ORIGINAL-B)

" parro z nreh z"

"riverside bends/ flats at Bendeelain the Kangaroo Valley"

Charles Throsby, 30th March 1818(HUYGENS CIRCLE)

= edges

(dual noun - stem)

page 81

or

Appendix

Jlarz; <bulala) - -----J

= edges, borders(implied dual noun - stem)

baru= oppo sin g

also the repeat expression " square"

" yiri:: brai", 1"edges")

Rev.W. Ridley 1875, Kamilaroi

(ORIGINAL_A)

Ia(ra\ | baru - Ia(ra), baru -------)

oppo srn I

" , berai:z berai"("constellation of Orion" ) ,

Rev.W. Ridley I875, kamilaroi

(ORIGINAL_A)

" tbarou:rlQe:"("carce" = square)

J. P. Gaimard 1826, .jenois bay

(TURUWAL)

oppo sin g erlges

page 82

= pair of things

Appendix

We also find baru in the word for eyes

and in the formal description of a bivalved shell which (in contrast to thegastropod) is made from a pair of interacting shells - which need to be

mulu I i= looking

and plaits

baru= eitherside

I of the nose ]

- b(ulAla -

" mo ::: bura t " , (" eyes" )

L. Malone 1875, wadi wadi

(coastal ORIGINAL - B)

",t le: ri: 11" , (" eyes" )

J.P. Gaimard 1826, jervis bay

(TURUWAL)

n

-______________)

UNIVERSAL

" mz i :b :berai : ", (" "y"t" )

R.H. Mathews 1901, dharug

(oRtGINAL- A)

" mu:: b: bara: " , 1" eyes" 1

J . Inrmer 1853, dhurga.

(SOUTHERN _ B)

greatly coiling

"balaz:mtba;i, ",

(" tresse, pour les cheveux" = plaits of hair )

J.P. Gaimard 1826, jentis bay.

(TURT]WAL)

" balla :: m: b : i: " , ( place name )

surveyor H.F. White 1834, illawarra

bula(lazn) - m(ura) - ba(ru) - iz(gara)= eitherside

I of the head ]

\-v-l

= pair lof things ]

TURUWAL

page 83

Appendix

considered - hence the bivalved (pelecypod) shell

slightly spiralling

slightly spiralling

bar(u) i : mu(ru) - d@ru z n) - da(ru)' n - Uar(a)

= eithe,rside = pair of shells

(dual noun - stem)

ORIGINAL - B

TURUWAL

" p I i,t th: thu : n: Qfizr", (" oyster" )

M. Munima 1863, wodi wodi .

" b : i z: t :: to :: gni", (" oyster" )

J. Rowley 1875, cowpastures .

I NOTE : Sydney Rock Oyster, Crassostrea commercialis]

(coastal ORIGINAL - B)

" bour zzz:: da :: " , (" arche coquille" )

[NOTE: Sydney Ark Shell, Anadara trapezia ] .

bura -= between

I beaks ]

0 : (dula)\-v-l

creature

----------------SOUTHERN - B

and the notJn penis

bur(a) -= betvveen

flegs l

n: du(la)l-----v----------

thing

"::1fl2 td z: jA z: w2 ", (" telline coquilte" =Tellin sp.)

J.P. Gaimard 1826, jenis bay .

(TURUWAL)

At the other extreme we have a pelican

HIJYCFNS-CIRCI.F

" boora z 11 : ", (" pelican" )

Pettit & Dawson 1850's, kumai

" burazn:", (" pelican" )

R.H. Mathews 1902, kurnai

" bur: n: da" , (" penis" )R. H. Mathews 1903,

" Thurrawal" = Ngunnawal

page 84

Appendix

also ([thing] between labia) clitoris

bura= between

I less ]

!uLu\------.V.---'J

undulations= labia

" burru: nggal " , 1" ctitoris" 1

R. H. Mathews 1903,

" Thurrawal" = Ngunnawal

" bra : ngolo" , (" mate" = X )

Re v,J. Bulmer 1876, kurnai

----------------HUYCENS_CIRCLE "SOUTHERN-B

or (see pages 67 and 95)

.Qlu)lu. - @g taynnru.rai------)

,iilF^r,-, bgrwee4 ORIGINAL-B

= tabia I legs l

and nose

.bura. - m(ulu),i - dula:(n)= Oetu\,een = eyes thing

ORIGINAI._ B

and ltwo [things] either-side of the nose) zoslrils

m:u:dla: p: pat ", ("nostrit[s]")

Teichelruann & Schutmann 1840

ADELAIDE

" ]eu'. berer " , 1" woncu" ;1

" Iu: bra " , 1" woman" 1

G.A. Robinson & N.l.B. Plonley 1929 / j0

" paira'. m'. ee'. tah: " , (" smeuy" )

F. McCaffrey 1920' s, wodiwodi

(coastal ORIGINAL - B)

"'. m: u : dla: " , (" nose" ),

Teichelmann & Schunnann 1840

ADEI"4IDE

(bura) - m(ulu) : i - dula b(ulala) - ba(ru - n)

--------J

= either-side pair I of things ]

page 85

Appendix

bara / buru

In the spirit of pages 48-50 we likewise offer

and

bara= anbtciiferous). UNMRSAL

walk( ative )waddle( iferous )

(adjective)

,,baral,,, (,, woyaby,, ) " burrai,,, (,,wathby,, )

D.K. Eades 1976, dhurga . Rev'w' Ridley 1875' kamilaroi '

" bootan" , (" pelican" ) ('RIGINAL- A)

Petrit & Dawson 1850's, kurnai

" buran" , (" perican" ) " burrah" ' (" wallabv" )

R.H. Mathews 1902, kurnai . R'H ' Mathews 1903'

" thurrawal" = ngunawal .

(SOUTHERN _ B, ,HUYGENS CIRCLE)

buru\--./= bouncy. UNIVERSAL

hoppy(adjective)

" buru" , (" kangaroo" ) " bi rri " , (" kangaroo" )

" burung" , (" vagina" ) J. Malone 1875, bo tan y bay .

R.H Mathews 1903, (TURUWAL)

" thurruwal" = ngunawal

( HUYGENS )IRCLE) " butoo" , (" kangaroo" )

M. Munima 1863, wodiwodi .

"buru", (" kangaroo" ) " buruQ", (" aboriginal girls name" )

D.K. Eades 1976, dhurga W. Dawes 1790, eora .

(SOUTHERN - B) (coastal ORIGINAL - B)

page 86

There is a serious state

bulu\-------!__./

down, flnt,= " sick"(adjective)

Appendix

bulu / bulu

ORIGINAL_BITUYGENS CIRCLE

" belly" , (food descends through), engtish

" bil]ious" , ( feeling down), enslish .

" billiosis" , r"ti" .

" phala" , ( fruit I falls dnwn]), sanskrit .

" bulyar" or " bulier" ,

(" dead" )

L. Malone 1875, wodiwodi .

(coastal ORIGINAL - B)

" bullia", (" dead" )

R.H. Mathews 1903,

" thurrawal" = ngunawal

(HUYGENS CIRCLE)

" billya: gooin ", ("kir')Richard Dawsey 1887, mt. dromedary .

" pdllyo z p6h ", (" going to die" )

Isaac Nathan 1849, wodi wodi .

(coastal ORIGINAL-B)

" o: belia ",

(name of the Gumnut baby who sank / drowned )

May Gibbs 1921, gundungara .

(ORIGINAL_A)

balu z t ", ("the watch is sropped. dead")W. Dawes 1790, eora .

" bo: yee ", ("deadordie")D. Paine 1798. eora .

" bo, i", ("die")J. Rowley 1875, cowpastures .

(coastal ORIGINAL- B )

----------------ORIGINAL _ B

- " gravely - ill"

-------+ORIGINAL _ A

----------------ORIGINAL - B

" pralaya" , ( period of rest / dissociation), sanskrit .

and its augmented form

bulu -" sick"

also we have

i z bulul------y___-__-_-_J

falling, sin king= " stopping"

or

bulu z i\------------v-/falling, sinking= "stopping"

gunvery

the vertr

page 87

Appendix

I can not stress sufficientiy that when morphology is based uponarbitrary opinion, instead of systematic formal logic, wrong consonantssneak into the orthography. For example Troy 11994, page 711 writes

,, balui ' baluwi "W. Dawes 1790, eora _______) ,, bUyi,,

" boi"J. Rowley 1875-, ,o*portrr", . J'Troy 1995 '

... which is formally and logically wrong (refer back to TABLE 1.2).

AIso there is

bulu: i= " dYin+"

rl du(la)\------n--------

person(noun stem)

" bull , i, nQ: " , (" die" )

RH. Mathews 1903,

" thurrawal" = ngunawal

(HUYGENS CIRCLE)

n: du(la\thing

(noun st€m)

" bullo r Il: " , (" stomach" )

Re v.J. Bulmer 1876, kurnia

(SOUTHERN _ B)

bul(u): i -------)\_--v-l

falling

" bale z:t tti", (" dead" )

"bo:it ", ("die")

R.H. Mathews 1901, dhanuk .

" ballu zzt ti " , ( aie" )

R.H. Mathews 1903, darkinyung

( ORIGINAL _ A)

and

bulu t i\-V--l

= "digesting"----------------SOUTHERN _ B

" pau:: nz cia" , itatian \" pau :z n: ch" , ,ponirn I

" bdu ::: che " , g"r^on )

as well as.falling-down thing

= thing

bulu (n)down

" bulu z: bal , i ", (" fatt down" )

R.H. Mathews 1901, lharruk(ORIGINAL_A).

"bul rt b: i ", l"leaning tree" )

R.H. Mathews,falling down = " leaning"

page 88

or

Appendix

bulu -------)

down

(" sun go down" )

" bu:: n: bal ", 1"tree-ood"1D.K. Eades 1976, dhurga .

"boo:z n: bal ", (wooa)"b: i: m r bal ", 1wood1" zz n z ulla", (wood)

J. Larmer 1853, dhurga .

( SOUTHERN -B )

SOUTHERN _ B

"- but.b",(" gumnut baby' s name" )

May Gibbs 1925, gundungara

(ORIGINAL _ A)

= thing

n

ORIGINAL _ BSOUTHERN _ B

TURUWALHUYGENS CIRCIE

" kan t b z i" , (" firewood, twigs" )

R.H. Mathews 1901,

" dhurawal" = ngunawal

( HUYGENS CIRCLE)

" kam: b z i" ' (" feu" = fire)

J.P. Gaimard 1826, jervis bay

(TURUWAL)

bu(lu): i -!-v-,

falling

.falling down = " waning" or " dead"

and SUN setting (see page 99)

(7uru) | wiJlt dula,SUN"

(noun stem)

": wzfle ' dilya - tull:: tull",

the yelta people's " setting sun song"

Rev.J. Bulmer 1850's, kurnia

(SOUTHERN - B)

falling*down = " setting"

and b ey ond- que stion de ad

gun - b(ulu), i\---v.----J \---v-----------

= very = dead

" Qan z b z i", (" rtre wood" )

D.K. Eades 1976, dhurga

(SOUTHERN - B)

"kanrb:y", ("fire")

I. Nathan 1849, wodiwodi .

"kun,b:ee", ("fire")

M. Munima 1863, wodi wodi .

( caastaL ORIGINAL - B)

page 89

Appendix

mu(ru) - bul(u):i

---------rORIGINAL _ B

hence

folk(plural noun - stem)

in contrast to

" mu I bal: i ", 1"'irt "1

W. Dawes 1790, eora

( coastal ORIGINAL-B)

" bti'djee"" bzi:dya"First Fleeters, eora

(coastal ORIGINAL-B)

" bul:: m ", ("beat,fisht")

D.K. Eades 1976, dhurga

" bul t, ma ", (" fighl')R.H . Mathews 1903,

"thurrawal"=ngunawal

" muree: p: ie ", ("beat hard" )":prie",("beat')D. Paine 1794, eora

II12h :: ", ( "to strike" ), hintlos tan ne

G.F . Angus 1874,

S. Bennett 1867 .

" bultm:zie ",("to clap the haruls while dancing" )

Capt.J. Hunter 1793, eora

" bul: mu : Qafi : ", ("beat")L. Malone I875, wadi wadi .

pAl I mO z QOh: ,(" going to Jight" ),

I. Nathan I849, wacli wadi

(coastal ORIGINAL-B)

" rAL).o: pu::t", pulsate, greek

Also there is .sick

or

.b(ulu).: i z 4i - .0 z (wu.r)ul(a) ---+" b:i:dji:Q:al"" b:a:dja:g:al"

the Castle Hill tibewho had Smallpox

Fist Fleeters, eora

(coastal ORIGINAL-B)

bul(uzi) - mu(ra)l----'v------'---

greatlyfalling, stopping

(verb)

or

mura- b(ulu)ziSreatiY falling, stopping

= "beating"(verb)

Also there is a word for clapping

SOUTHERN _ BHUYGENS CIRCLE

----------------ORIGINAL _ B

slightly ill = "sick"

very slightly = " gentLy"

page 90

Appendix

creature

wuru(la:Qun)- 0 -b(ub)ri -------+\-.v--l

falling

wt(rula: !un) - b(ulu), i --)

"ba::ng:u: ", (" flying squirrel" )

Rev.J. Gunther 1837, wiradjuri

(inland ORIGINAL - B) .

" bd z: nQ : d : " , (" squirrel" )

I. Nathan 1849, wodiwodi .

" bu:: nQ z z ", ( Jlying squirrel" )

W. Dawes 1790, eora .

"but flQ zofaz ", ( flying squirrel" )

F. McCaffrey 1920's, wadi wadi

(caastal ORIGINAL- B)

and gliding creature

\--.--w---/

slowly

wotti :: m r b : i ", (" Jtying fox'' )

W. Russell 1914, gundungara

(ORIGINAL_ A)

WaffA :: n :: ", (" flying squirel" )R.H. Mathews 1902, kurnai

(SOUTHERN_B)

" wo tt bb : i t n ", (" flying squirrel")

D. Collins 1798, eora

" wu ,t b: i: n ", ("possum")

W. Dqwes 1790, eora

(ORIGINAL_B)

or

0creature

or

btt(lu ri)- n -wuru(lat gun) ----) bu:: g:wr.ti -----4

" bu:: nQ:0o : ", (" squirrel" )

J. Larmer 1853, dhurga .

" ba:z ng : U : " , (" flying squinel" )

" bu:: ng : 8u: ", (" flying squiruel" )

R.H. Mathews 1901, 1903,

" ba ::4 : Qu : ", (" flyingfo:r" )

D.K. Eades 1976, d.hurga .

(SOUTHERN _ B)

... where the correct syllable is wuru(la) as in McCaffrey's word, orwu(rula) with a quiet w as recorded by most early lists. BillyYerramagang (1820's) gave the "little squirrel" as pe]:tn::kan::g orcu:r.::ke::kan which both derive from bul(u):i:4:(wurula):gun:(wurula):gun . The option of D.K. Eades and R.H. Mathews is simply

= gliding

= glidins

= gliding creature

wrong.

page 91

bu:g:wu

the sugar-glider, Petaurus breviceps, eating nectarfrom a blossom. It is a furtive nocturnal creatureliving in family-groups, in tree-tops, where

_it

searches for insects and the nectar of blossoms. Ithas litters of two babies at a time, which live in themother's pouch for upto 8 weeks. They are very-

sociable animals, with upto 12 sharing a nest ofdry leaves inside a hollow tree, cuddling togetherfor warmth in winter.

@ Ghris @lbrr

Appendix

There is also a diectic form (pages 96-97)

from past

= "back", behind(verb)

"g: i: Ii" o,"gti,Ie",1" rhat Iyonder visible l" t .

" Qa:: li " o, " ga: le",(" this Ihere, visible ]" ) .

M. Shnrpe 1994, bundjalung

At the opposite extreme we find

buluup, projective

(adjective)

" paialla" , (' tatk' )

R.H . Mathews 1901, " thuruwal" = ngunuwal .

( HUYGENS CIRCLE)

" bela" , (" casuarina" = mangrove seedling)

.1.P. Gaimard 1825, .jervis bay .

(TURUWAL)

" baa]a" , (" sum tree" )

Kitty Cooper 1870' s, moira .

(SOUTHERN _ B)

" pLli " , buddhist suiptures .

" baal " , (" sun, stars, divini4t" ), hebrew .

" g: i: Ii", ("back")

J. Rowley 1875, cowpastures

" kara: bul"(sword's edge, literally the back of the sword),

" kata:t bul", (" the back" ),

W. Dawes 1790, eora .

" koro :: boul " , (" bact<" )

D. Paine 1794, eora

( coastal ORIGINAL - B)

----------------ORIGINAL _ B

----------------UNIVERSAL

" paialla" , (" teu, tatk" )

J. Rowley 1875, cowpastures .

" pyalla" , (" words, testimony" )

M.A. Fitzgerald 1891, eora .

" piale" , (" speak, spoke" )

D. Paine 1794, eora ;

W. Dawes 1790, eora ;

D. Southwell 1788, eora .

" byalla" , (" speak" )

Capt.J. Hunter 1793, eora .

" pilla" , (" to taush" )

P.G. King 1793, eora

( coastal ORIGINAL - B)

page 93

Appendix

and the augmented forms

bu(lu) - .mura,!=-v-

, up greatlyORIGINAL. B

TURUWAL

= rall, hiBh

" baia , mi" , (ancestral god of NSW )

A. Mqssola, 1971 ,

A.W. Howitt, 1904.

" fe z mur " , (" upper leg bone" ), Iatin.

" bau, me " , (" tree" 1. german .

" pou: mara"(" montagne" = moantain),

J.P. Gairrard 1826, jentois bay

(TURUWAL).

" baa t mirr " , (" long tatl" )

McNicol & Hosking 1994, wiradjuri

( inland ORIGINAL - B)

and

also the vertr

" boo, gont 9",(edible moth Agrotis inf usa)

S. Mowle 1838, ngunawal

( HUYGENS CIRCLE)

" balla : gon: ", (" sponge" 1,

" bul : ga : " , (" mountain" ),

N.D. Pettit & W. Dawson 1850, kumai

(SOUTHERN _ B)

" bul: ga : ", (" hitt" ),

J. Rowley 1875, cowpastures

(coqstal ORIGINAL- B)

"bulla,i"," bull: i"," booal I ye" ,

" name of tack ri sin I up the

Illqwarra Escarpment"

F. McCaffrey 1920' s, wadi wadi

(coastal ORICINAL- B)

bulu - i\..............\_/= ri sin g, emerging

commenclnS(verb)

"balti", ("baby, veryyoung" ),

" ba : i" (" blossom or flower" ),

Re v.J. Gnnther 1837, wiradjuri

(inland ORIGINAL - B)

" plto t i", 1" height, size" ), greek

----------------ORIGINAL _ B

gun. g(un)

very Yery ='t extremelY"

bulu= up (dbove)up (soaking )

up ( s tan ding )

ORIGINAL _ BSOUTHERN - B

HUYGENS CIRCLE

page94

,buJu-n:du(la)- i

" booz n r da t ", ("apple tree" ),

"baalu t n t " (" mangrove seedling" ),

W. Macat'ther I861, cowpastures

(coastal ORIGINAL - B)

" bela n " (" casuarina" = mangrove seedling),

J.P. Gaimard 1826, .jervois bay

(TURUWAL)

Appendix

and verb with noun

thing

ORIGINAI. _ BSOUTHERN _ B

TU RUWAL

" bu : n: ta | ", ("thish")

H. HaLe 1846, d.hurga .

" boo: n: da : ", ("thish")

E.M. Curr 1887, dhurga .

" booloo z,: tha : ", 1"over, above" 1

Re v.J. Bulmer 1850' s, kumai

" balla tt t : " , (" cherry tree" )

R.H. Mathews 1902, kurnai

(SOUTHERN _ B)" pntd,Lo : v :: i " , (" baloon" ), greek

"llnoD::T:i", ( " thigh" ), greek

", (" blouse" ), greek" pnLov z: (a z

and, in the far southem zone, we have

du(la, g)\-v-l

= creature

(bu)lu- i -\_-v--.-/commencLnS= " growing"

-----.._-.------SOUTHERN _ B

\-----------v..-

(noun - stem)

"1:i:ty: ", ("chitd, bor"),

R.H. Mathews 1902, kurnai .

" Ia :z the; " (" boy" ),

Re v.l. Bulmer 1876, kurnai

(SOUTHERN _ B)

"l: i:d: ", ("hai,")

Pettit & Dawson 1850's. kumai

" It i:ttt ", ("hair")

Re v. F. A. Hagenauer 1876, kurnai

" I : i z t t ", 1" hoir, hair of the head" )

Rey.J. Bulmer 1876, kurnai ;

R.H. Mathews 1902, kumai

( SOUTHERN - B)

... where a young woman goes from bura:4ulu (Huygens Circle) to(gu)lu:bura (page 85), and a nice place is (gun:gu)la:n @age 67).

= arising, upstanding

= thing

page 95

at-.i

ElJ

.9

U

E

Hump-back whales Megaptera novaeangliae, a familiar sight in coastalwaters, are migratory mammals which grow to the enormous length of 19metres and weigh as much_ as 48 tonnes at maturity. The Latin nine, ,,mega-ptera"

= great-wing, refers to their large serrated pectoral fins. These creaftrres s-pend summer inthe Antarctic feeding on kill. As winter approaches they slowlv misrate northward uo theeast-Australian coastline to give birth to calves in eueeniland's whitiundav resion. ivhenthe calves are strong enough, with a sufficienrly thick layer of lat for insulition] rhey retumsouthward again. During this period of increasing body mass. in deeper cooler waiers, thecalves sometimes experience difficulty achieving positive buoyancy. This is rather serious,for air breathing

_ rymm{q that must return to- the surface regularly, so the motherssometimes swim below their calves nudging them upward. tn ISIZ John Rowley ofHolsworthy, the son of Lieutenant Rowley, explored gunduTara terrirory to the south-west of sydney along with coastal dariwul guides, in his spare time about the campfrerecording Aboriginal words supposed_ly flom the "Georges River, Cowpasture, andAppin" uea. By May 1819 he and Charles Throsby had found a route from theCowpastures lnear present-day Campbelltown) to Bathurst by way of Moss Vale, Taralgaand the Oberon district and, as a rewrard, was granted 1000 aires al Bong Bong.

,.:::r;:;:,.illr:ii. .,lirilii!:i: :r;i..l;

,;,;,]]rj,,'' ;"r1,,':1.r,,,',,. i::;.it':. :rrl ri

bul: gu - g: i: Iu - (A)l-------1/------ \------------v_.-.-., r-\-

= hump = back [whale]

g(aru) - i:]q-;;nIrom balkward!...-----a-

= rear, bqck

Hill, bulga.Bacl<, gili.Humpback, bulga-gili.

aGhris @//ort

bul(u : i) : gu(n : gun : n : Qula),

A "mountainous" back, like this, is quite different to one which is nara:gulv =

It wasn't ti]] the 1870's, by which time Rowley had retired to scone, that his Aboriginalwordlist was finally published in Rev. W, Ridley's classic book "Kamilaroi and 6therAustralian L_anguages". A couple of these published Aboriginal words may be used tocreate an understandable name for the Hump-back whale

(" bul :: Oa i'.i'.:'. " t"hi " ) )

--------+ |":::::: g: i: li: " t"bark )

I

\lJ. nowtey. t8751 1lJundusan 1)

"hunched or bent" with age.

And the common

rising up = " uPs tan ding"

. gun - dul(a: n) ----------------\:..."_, .-"--- UNMRSAL= r",171 thing

Appendix

" goon: del; ", sordungara

Burrogogorang Chiejlain (1802 - I8I4)

(ORIGINAL _ A) .

"gun, du z ", ("tree")

R.H. Mathews 1903, " thurawal" = ngunawal

( HUYGENS CIRCLE) ,

"gunIdai: ", ( " stringy bark tree" )

Rev.J. Gtinther 1837, wiradjuri

( inland ORIGINAL - B)

" p : i: the; n,: ba ",

( " new moon" )

yelta peoples "moon song"

J. BuLmer 1850, kurnai(SOUTHERN B)

" g-oun t dah : " , (" tree" )

I. Nathan 1849, wadi wadi

" k-un t du r ", ("tree")

L. Malone 1875, wadi wadi

(coastal ORIGINAL - B)

" ghoon t doo: ", 1"tree"1

W. Bothong 1900, dhurga

(SOUTHERN _ B) .

makes two adjectives and a noun give

thing = " 7n66r"(noun-stem)

bu(lu : i) - bu(Lu) - dul(a: n) -------+

" ba :: ba: dal: ", ("tee")D.K. Ead.es 1976, dhurga .

(SOUTHERN_B)

"bo::baItoo|",mountain in upper Kangaroo ValLey

L. Atkinson, " Sydney Mail"

7 th Jan I871( HUYGENS CIRCLE)

or

\-------.v..-..-/

thing

---.Y-r$tng

page 98

Appendix

and SUN rise (see page 89)

Uu(ru) : wilt : (dula) - bu(Lu) : i - bulu -------------+ORIGINAL _ A

" SUN'(noun - stem)

ORIGINAL _ BM IDTHUNG

" co: ing , - bt i t bola", ("thesunrise")

P.G. King 1793, eora

(coastaL ORIGINAL * B)

"- b, i,b", (" gumnut baby's name" )

May Gibbs 1925, gundungara

(ORIGINAL _ A)

bla z n z da: bla z n: dd ", ("sweat")

Rev. J. Bulmer 1850, kurnai

(SOUTHERN_B)

"- b:eerbaylu"," - bay:t bay " ,

Black Mqria 1840's, midthung

( MIDTHUNG) .

Two noun-stems occur in sweat / persperation

bulu- n: da(ru) - bulu- nz da(ru)

as opposed to possum skin rug

bu([u) : i -d(aru) - bul(u) : i- d@ru)

"pziztzpti:t",(" opossum cloak or gown" )

Mickey Munima 1863, wadi - wadi

(coastal ORIGINAI - B)

Thus we anticipated approximately 6 words of the form Lr***, andimmediately have 8, variously twisted into form within the confines ofOctahedral Symmetry. The result is most gratifying, but we shallcontinue on to see if the final case is similar.

-------+

"buttd,bul ,i: ", ("opossumrug")

J. Rowley 1875, cowpastures

(coastal ORIGINAL- B)

ft stn g up

= emiss ion- exterior / skin = emission-exterior / skin

= dead-exterior / skin = dead-exterior/ skin

page 99

I

Appendix

wip:i:bu:1:bu-milt-miru - the first dawn

\\

1il

page 100

Appendix

Uuru / rlara

Some words employ AUru - periodic, hence the noun

" narruz n ", ("moon"), Pettit & Dawson 1850's

" neffa: n ", (" moon" ), Rev.J. Bulmer 1876

" ngirra t ng ", {" moon" ), Rev.F.A. Hagenaur 1876

kurnai, (SOUTHERN - B)

and the verb (with repetition implying vigour or frenzy)

{lUrltz 1 - IJUTU: 1 - IJUTU: 1

"purraz n " , (" daJlight" )

"4iru: ", (" d.arkness, nigh{')

"4urrat ", (" after" )

Rev.W. Ridley 1875, kamilaroi

(ORIGINAL _ A)

Uuru, n\_---.v-.-..J

-- periodic thing

(noun)

ORIGINAL _ A

SOUTHERN _ B

also

= wigqling, throbbing

(verb)

" neuru; neunt; n!: ",

(" red" = pul sin g blootl.)

Lizzy Malone 1875, wadi wadi

(coastal ORIGINAL - B) .

" girr I i: girr; i,:t ",(" red"= pul sin g blood),

McNichol & Hosking 1994, wiradjuri

( inktnd ORIGINAL - B) .

diz n -tiny - thing

=" clitoris"(noun)

ORIGINAL _ ASOUTHERN - B

HUYGENS CIRCLE

"g: i:g:t4: a", (" sew" )

Rev.W. Ridley 1875, kamilaroi

(ORIGINAL _ A)

" ngaier I nger # ", t" red" t

" ngarigu" = wolgalu,

" ngurr t i ::: ", (" shin"= ,,eigg1sv1

" thurruwal' = ngunawal ,

R.H. Mathews: 1908, 1903 .

( HUYGENS CIRCLE)

" ji z n: maD", ("wife')

J. Rowley ) 875, cowpastures

--)

ORIGINAI, * B.rn.n,

the

page 101

Appendix

hence

Qi:n.- g(uru):i= tiny thing = twinkling(noun - stem)

"di z n : ng I ee ", 1"rtar"1

M.M. Eveitt 1900, gundungara-__.---._-_-____

ORIGINAL - A

with plural

- Uuru'(i) - rJ(uru1i)

sparkling stars

ORIGINAL_ B

" jitn: ji:n:fiuru:2fllJ t

(" sparkling stars" )

L. Malone 1875, wadi wadi

66), and also general cyclebut perhaps not "new-moon"

also

qur(u): i\.............../= wiggly

and gula:i-0uru ="full-moon" or even

- black.

HUYGENS CIRCLE

"watery-pond" (page

"half-moon" (page71)

" ji ', lli, fiQUfr z i ", ("centipecle")

"thuruwal" = ngunuwal

R.H . Mathews 1903

At the opposite extreme we have

oara= Looped, spiral

(adjective)

" gnArrA ", ("aknot or atie")

W. Dawes 1790, D. Blackburn 1791, eora

" gnara " , (" knot o, tan gle in a line" )

J . Hunter 1793, eora

---)

ORIGINAL _ B

tiny numerous flegs l

page 102

and

Appendix

= greatly high = spiraller

"bu.mr:i:n9",J, Mason 1838, cowpastures

" bi : mar t i t n ", (" boomerang" )

J. Rowley 1875, cowpastures

" bu ; mmeru z; rte " , (" wooden sword" )

Dr.J. OLdfieLd 1828, eora

(coastal ORIGINAL - B) .

" nd no- : wr!: ", ("shelt")

L Nathan 1849, wadi wadi

" gnar z he: ",

" z whorl: ",

S. Bennett 1867, engLish .

MIDTHUNG

" boo : mera ;,: ng ",

L. Atkinson 1853, midthung

( HUYGENS CIRCLE)

bu(Lu) | mura - i: g(ara\

-;---.--:-. ORICINAL _ B

also

hence

daru.-------''------

exterior

= [trse I bark

----_______-..--ORIGINAL _ B

or

w(ur)ula- Aa(ra- 4i),n - da(ru)

-----..---------ORICINAL _ B

= Ioop - s, whorl - s

Qiz(n) - wurula- 4@ra)tiny - numerous - spirals

= fibrous strands -- " stringy"

" dturA : du:: rald' z 9",("the bark to make Jish lines" )

W. Dawes 1790, eorl

:zz wiralu :4 ", (" forest oak tree")

L. Malone I875, wadi wadi

= barle

" wu]u z !d:: fi:da", (" pigeonberry tee")L. Malone 1875. wadi wadr

-- ftbrous strands

page 103

Appendix

NOTE: the coniugate to dafu = exterior/about (seepagesI03,236-7),

appears tobe dura = interior/through , as in the adjective

" dhara I marra ", (" draining" )

" dharra I mafra ", 1"absorb, swallow" )

C. Richards 1902. wiradjuri

(inland ORIGINAL-B) .

ti : ma ", ("to squeeze \rater from a sponge" )

W. Drwes 1790, eora

(coastqL ORIGINAL-B)

dura - mura -----)\------V.-=----

= greatly through

and the verb

i 'dtraa--------

through -_ (wu)rul(a)

\_-----.V-=-,

-ing

= stinging, stabbing, injecting

" i: teroo : " , (" river nile" ), ancient eglptlan

" : dire t " , 7" to say /telL" 1, italian

" :throu;Qh", english

a'lso nouns such as snake

ORIGINAL _ BHUYGENS CIRCLE

" tthir' roul",(" pLace named afier mountain leeches" )

BilLy SatldLer 1892, wadi wadi

(coastal ORIGINAL - B)

" : thura z r2 " , (" sting or stab" )

R.H. Mathews 1903,

" thuruwal"= ngunawal

( HUYGENS CIRCLE)

" : thurru : ng ", Rev.J. Bulmer 1876 .

" : toorroo; r1Q ", Rev.F.A. Hagenauer 1876

" : tooroo : ", w. Thomas 186o .

" t thuro: ", N.D. Pettit &W. Dawson 1850 .

("snake" ), kurnai

(SOUTHERN_B).

" : throo : 11 ", ("mountain leech"),

Billy Saddler I892, wadi wadi

(coastal ORIGINAL-B)

irdura 0 ---)\------1--.-/\-,.J

- stinginS, creaturesucking

page 104

and mosquito

i, dura\.____!-.-,

= stinging,sucking

Appendix

,4i,'a ---)tiny

creature

" z teura: die: ny ", l"mosquito")P.G. King 1793, eora .

": tewra: die. ng ", ("mosquito"1

W. Dewes 1790, eora

(coastal ORIGINAL- B) .

" t dir, di t k ", l"mosquito")R.H. Mathews 1902. kumai

(SOUTHERN_B) .

Em a STutlz t// A4'd2t ]f,,g

page 105

,B14lz. .rtuI/, I

Appendix

7ulu / dulu

Finally a pair which is conjegated by different word-initial consonants -

gulu\---w---J

= corrupated, wrinkledfolded, crushed

(adjective)

" gnullon " , (" forehead." = wrinkled)

D. Collins 1798, eora

---------)UNIVERSAL

" gnalloa " , (" to sit" = folded legs)

" nulla " , (" forehead" )

P.G. King 1790, eara

" alloua " , (" to sit down" )

D. Paine 1798, eora

" ni llo " , (" forehead" )

W. Tench 1788. eora

" nourlew", (" forehead" )

M. Munima 1863, wadi wadi

(coastal ORIGINAL - B) .

" nguu]uung", (" forehead" )

NcNicol & Hosking 1994, wiradjuri

(inland ORIGINAL - B

and

" nolo: nyila", (" quarrel" )

Dr.J. Lhotski, ngarigo

( HUYGENS CIRCLE)

" lnlu " , (" forehead" )

Re v.W. Ridley 1875, kamilaroi

(ORIGINAL_ A)

" nulluc " , (" tki," )

nulu - nuluJ .' , HUYGENS CIRCLE

=wrinkliest, f logged(superlative adjective)

" gleat " , (" sinews under the knee" )

Pettit & Dclwson 1850's. kurnai

(SOUTHERN _ B)

" ngulu", (" forehead" )

" ngulur " , (" sweat" )

" ngullung " , (" sit" )

R.H. Mathews 1903,

" thuruwal" = ngunawql

( HUYGENS CIRCLE)

ORIGINAL _ B

" gnallu : nE la" , (" mushroom I gills ]" )

D. Collins 1798, eora

" knulla : I'nulla","the blacks painted a criminal white,

then boondied (clubbed) him to death"

F. McCaffrey 1920's, wadi wadi

(coastaL ORIGINAL - B)

page 106

Appendix

also

" gTulolt z gdn ", ("short")

l. Nathan 1848, wodi wodi

( coastal ORIGINAL-B) .

and

daru\-1.-l

exterior= [ocean] surface

ORIGINAL _ AORIGINAL _ B

" dhulu", (" straight, direct, true" )

" dhulay", (" long, extended, straight" )

C. Richards 1902. wi radju ri

" dhulu", (" stratght" )

" thulu " , 1" wooden spear" )

R.H. Mathews 1904, wiradjuri

( inland ORIGINAL- B)

dulu ------J

extendetl

warafar

" yula t warra ", ("stretch, separate")

Rev.J. Ginther 1892, wiradjurt

( inland ORIGINAL-B) .

" wre: ckil ", ("tatt")

Rev.J. Bulmer I816, kumai(SOUTHERN_B),

wara - dul(u) ----)

.rJulu - gun oRrcrNAb_---+AL_B

= very crushed

nul(u\ - gun

... and the female marker bura:4uLu or (7u)lu:bura, the clitoris,which has been discussed (see page 85).

At the opposite extreme we find the word

dulu\---v_----J

= straight, extendedseparated(adjective)

" dyulu " , (" wooden spear" )

R.H. Mathews 190j, dharkinyung

(ORIGINAL- A)

as ln

" chul : gun I derri ", ("wave")

J .F. Mann I840's, eora

(coastal ORIGINAL-B) .= very undulating

or

page 107

hence the verb

duluzi - ,bqfu\------v-

separattng sides

= OPENING, SPLITTING(verb)

which contrasts with

Appendix

----J

ba(ru)= either side

mir(u):i - .baru

" dhala::barra","crqck, burst, break, ... as when opening a

pea- pod, cracking a nut, breaking a stick"

Rev.J. Cinrher 1892. wiradjuri

(inlandORIGINAL-B).

" wirr : i : bard ", (" shut the door" )

W. Dawes 1790, eora

(coastal ORIGINAL-B) .

stdes\---------.V-.---

vanishing

= CLOSING, SEALING(verb)

and also

dulu= stretched,

tensioned

= knot

g(ara) , " thulu : m ': ba ", t"knot")-----'-----7 Rev.J . Bulmer t876, kurnai

(SOUTHERN_B).

We see that 4 * x x finds it necessary to choose conjugates fromdifferent word-initial-classes. It is not that there is a lack of words, quitethe opposite, it is just that mathematical-law makes it impossible tochoose them for use - its the way of the Entropy Maximising PowerLaw.

picture by Ainsley Roberts, i962

page 108

Appendix

"... [in] the words of the Stigmatine Fathers 'Some people see-things that are

"ia' siy wt y ? We aream ol things that never were and say why not ?' " '

P.S. StePhens (1974)

"... in 1896, I mentioned the fotmer existence, among the Wiradthuri tribes,

of i tecr"i kmguage [the 6urbung], known only..to. the iniriated' The

tamilaroi likeiise" possess a myitic tongue, called Yauan, which is

inculcated at the Bori ceremonies'' lt is dfficult to assiSn a reason for this

laipuag,e, because it is never used in the plesence of women or children, butis iio|nn exclusively by the men when carrying out the ceremonies.ofinitiation. It is possibly a typical remnant of the language of earlier conqueringtribes in the remote past

R.H. Mathews (1904)

discussion

Now is an appropriate time to deal with the Octahedral Symmetry. It is

a good approximation, but there are signifi,cant differences on foursefarate ociasions: (i) muru/mura, (ID gulu/gula, Qil) dulu/dulaand (iv) mulu/malu. If these particular occasions can be justified, itwould be very helpful.

The following expression (pages 62-63) illustrates a problem arising fromthe past-tense of verb-stems containing g. . .

mur(a- dula)tn- pur(a) -TAKE = CATCH / GATHER / COLLECT

i,(bu)lu\_------v-J

= did

i=am/is/are

i,bulu\-------n-J

- shall/will(tense)

page 109

Appendix

Aboriginal man and fire in a canoe night-fishingpicture by engraver W. Hatherell, 1888

past

present

" maa zz nz:i : la ", ("they take or catchfish" )Anganangan, cited by W. Dawes 1790, eora .

" mAr:::z i: lla ", ("caught")

W.J. Enright 1900, eora .

" maa ;z n : g ti ", ("taken or married.")

Booroong, cited by W. Dawes 1790, eora .

" mar :zz koor. a . barley ", {"come out ard trsh" l

Long Dick, cited by J.F. Mann 1810's, eora ..luture

Whilst the future-tense mUr:::Quri:bulu and the present-tenseIDU:fl:Q:i pose no problems, there is a real problem with the past-tensewhich cannot be mu::n:Q:i:1u because the word gilu already exists(see pages 93 and 96-97). William Dawes recorded that it was customary

page 110

thedolphin

"On.seeing a shoal ofporpoises, they sing while

tlrc fish is above the water.

no -te - le - bre

no-te-le-breuntil it goe,t dov)n,

... when they sing the words

no-tee, no-tee, no-tee

urttil it ises again".

ID. Collins, 1798]

rr.Lnslarinn lotlright O Ct'i, trrrn t, tool

for p to be traditionally omitted in this special case, useing the specialtense-marker -171a,. For this to be true we also require maru, thecontraction ma(ru) and conjugate of mura, to be interperted as

mutu. Hence we have the breaking of the first octahedral symmetry

muru / muraIn the second case gulu occurs instead of galu, to avoid confusion withthe ablative suffix pa of ga(ru), hence we have the breaking of thesecond octahedral symmetry

gulu / gulaA more subtle expression will become obvious in the third case when{ulu occurs instead of dalu, to avoid confusion with the contracteddietic da(ru), thereby breaking the third octahedral symmetry

dulu I dulaThe fourth case mulu occurs instead of mula, to avoid confusion withthe contraction m(ur)Ula = 5, hence breaking the fourth octahedralsymmetry

mulu / malu

page 111

proto-Australian words

bulu= d.owh, low, flat,

dead, cessation

bulu= up, high, projecting,

conmencement

gulu= duLl, lethargic (as with

Koala), petnfed

gula= shiny, radiant, viml,

nice/good, alive

malu= blind, obstructing,shield.ing, opaque,

(eye)cataracts

mU]U= visible, accessibk, ckar,

channelling

bulala=2 {lUI7 = very muntla= 5

gula= malevoLent, deadly,

treac he rous, angry, sorcery

galu= benevolent, helpful,

honest, happy

mulu= familiar, intimate

mulu= unfamiliar, stran&e,

unexpected

buru= noisy, quick,

energetic, aw^ke

buru= quiet, slow, lazy,

tired, sleepy, stilL

guru= fat, globular, convex

more

guru= slim, conceve, Iess

malu= [air] bubbles

(as with Platypus)

mula= fli4uidl droplex, progeny,

dust

baru= eit he r-si.de, o ppo s ing

bura= between, compromise

garu= from, out-of, sound

gura= loward, into, throat,orifice, cave, valLey

mala= inflated, bulging,

enlwnced

mulu= deflated, s hriv e llcd,emaciated, diminished

buru= bouncy, hoppy

(as with Kangaroo)

bara=wqddW

(qs with Wallaby)

gura= film (as vrith stones, hail

or muscles)

garu= squishy (like nasaL mucous),

flufu (like clouds)

mura= large, greatb

MUTUsmall, slightly

gilu miru =0= behind, rear ahead, in front

TABLE 1.10: translation, phoneticisarion & compilarion, copyright O c. ILLERT,2003

Appendix

page ll2

aDou t EATS O

In= fure

nura= yonder, dismnt

dulu ryulu= strait, extended, separated = sinusoidal, wrinkled.,

dulu= shabby, untid!, bent

dula= lsingular, unity

wilu= return, come-back

nurula =4 wurula=3,plural, severa[ lots of

0uru= cyclic, oscillatory,twinkly, shimmery

nata- knotted, coiled, spiral

daru= about, extemal,

outsi.de surface

dura= through, intemal, insfule

wffu= profane, improper,

lefi, crookd

wurl= sacred. proper,

right, aligned

IArA= full, solid.

IUIa= hnLf, middle

Iuru= empty, hoLlow

mlJ1= tlze, a (singular)

run= somethings (plural)

J1u0= someones (plwal)

wata= far, distant

WUTU= cway

w1Jl= wring, twist,

rotate, spin, orbit

Appendix

translation, phoneticisation & compilation, copyright O c. ILLERT,2003

These four variations from octahedral symmetry are placed in ourabove table of words. Now we have evidence for all their existence. InChapter I we study the structure of specially long words, and inChapters 2 and 3 we test the meaning of our phonetic-alphabet toAustralian languages and evolution. Chapter 4 continues looking atchanges in modern languages and retroflexives.

In relation to sources of words, the most suggestive are mura (p. q) /muru (p.32), Qaru (p. s4) / guru (p. s$, dura (p. 104) / daru (pp.103,236-23'1) , bulu (p. sD / bulu fu. 93), and seperately mula (p. a5) ,mala @.4B), gula (p. 6s), euru (p. sl) and baru (p.79). Whetherthey are ultimate universals will not be the basis for this thesis, but Iknow that I simply could not have done it without them.

1d00060bo

page 113

Appendix

bibliography

C. ILLERT & R.M. SANTILLI, lggl "foundations of Theoretical Conchology,

second edition" , Hadronic Press, Palm Harbour FL.

C. ILLERT & C.A. PICKOVER, May 1992. "generating inegularly oscillating fossilseashells", IEEE Computer Grrtphics and Applications l2(3): 1'8-22.

C. ILLERT, July 1995. "Australian supercomputer graphics exhibition", IEEEC o mp ut e r G raphic s and Ap p lic ations I 5 (4) : 89 -9Q.

C. ILLERT, Aug.2000. "the last Shoalhaven Lore Master", Shoalhaven Chronograph22(10): 1 & 5.

C. ILLERT, June 2001. "the centenary of Mary Everitt's Gundungara grammar",Joumal and Proceedings of the Royal Society of New South Wales 134: 19-44.

C. ILLERT, 2003. "lexigenesis in ancestral south-east-Australian Aboriginallanguage", Joumal of Applied Statistics 30(2): 113-143.

C. ILLERT,2003. "early ancastors of Illawara's Wadi-Wadi people ...", Northemlllawarra Aboriginal ColLective Inc., Wollongong.

C. ILLERT, 2003. "three sisters dreaming - or did Katoomba get its legend fromKangaroo Valley", Northem IIIawarra Aboriginal Collective Inc., Wollongong.

C. ILLERT & A. ALLISON,2004. "phonogenesis and the origin of accusative syntaxin proto-Australian language", /o urnal o;f Applied Statistics 3l(1):73-104.

C. ILLERT, 2005. "origins of linguistic zonation in the Australian Alps, part 1,Huygens Principle", Joumal of Applied Statistics 32(6): 625-659.

C. ILLERT, 2005. Conespondence to Mr Robert Bruce, Dep. Environment andHeritage, Canbera - re Emergengy Listing of Westcliff Colliery Area 5, longwallblocks 31-33, under the EPBC Act - attachment l: " the traditional story of the GreatWalk down the Georges River to Lo Perouse, in 1890", oral account by Jean Stewart.

C. ILLERT, 2006. "origins of linguistic zonation in the Australian Alps, parl 2, SnellsLaw" , Journal of Applied Statistics 33(9): 989- 1030.

C. ILLERT, 200'7. " interdisciplinary aspects of Aboriginal language revival inparmership with the Aboriginal communities" , College of Arls (conference), p. I-24.See the song on page 19.

C. ILLERT, editor 2007. " SW Growth Centres Commission; Edmondson ParkSubdivision; Kangaloon Boreholes; Helensburgh Colliery; Menangle Eel Farm; SandyCreek; Wongombra ", Northem Illawarra Aboriginal Collective.

H. GOODALL & A. CADZOW,2009. "rivers and resilience, Aboriginal people onSydney's Georg,es River" , University of New South Wa1es. See pages 103, 134, I40 -141, 289 - 3r8.

page 114

Journal of Applied Statistics, Vol. 30, No. 2, 2003, 113–143

Lexigenesis in ancestral south-east AustralianAboriginal language

CHRISTOPHER R. ILLERT, School of Languages and Linguistics, University ofWestern Sydney, Australia

The 1/x frequency distribution is known to researchers ranging from economistsand biologists to electronic engineers. It is known to linguists as Zipf ’s Law (Zipf, 1949)and has recently been shown not to be a consequence of the Central Limit Theorem (Troll& Graben, 1998)—leaving an ‘unsolved problem’ in information theory (Jones, 1999).This 1/x distribution, associated with scale-invariant physical systems (Machlup &Hoshiko, 1980), is a special case of the general power law xj arising from the LagrangianL(x, F(x))ó1

2x1ñjF2 and, as j need not be an integer, some related research understand-ably involves fractals (Allison et al., 2001). The present paper generalizes this Lagrangianto include a van der Waals effect. It is argued that ancestral Aboriginal language consistedof root-morphemes that were built up into, and often condensed within, subsequent wordsor lexemes. Using discrete-optimization techniques pioneered elsewhere (Illert, 1987;Reverberi, 1985), and the new morpho-statistics, this paper models lexeme-condensationin ancestral south-east Australian Aboriginal language.

1 Introduction

Mathematical modelling has been slow to penetrate linguistics. In the half centurysince George Kingsley Zipf published his seminal book, ‘Human Behaviour and thePrinciple of Least Effort’, most work on Zipf ’s law has been published by researchersoutside the field and scattered through journals of biology, physics and computingscience.

A recent paper showing that Zipf ’s law was not a consequence of the CentralLimit Theorem was published in Physical Review (Troll & Graben, 1998), and apaper linking scale-invariance to 1/x ‘noise’ was delivered decades ago at asymposium of electronic engineers in Orlando, Florida (Machlup & Hoshiko,

Correspondence: C. R. Illert, 2/3 Birch Crescent, East Corrimal, N.S.W. 2518, Australia. E-mail: illert@1earth.net

ISSN 0266-4763 print; 1360-0532 online/03/020113-31 © 2003 Taylor & Francis Ltd

DOI: 10.1080/0266476022000023703

Chapter 1, JAS 2003

Page 115

114 C. R. Illert

1980). The full implications of Claude E. Shannon’s information theory havebarely penetrated linguistics and even the rudimentary ‘lexico-statistical’ modellingthat is practised (Crowley, 1992) was imported from the ubiquitous cladistic treesin palaeontology, taxonomy and population biology.

It is hoped that this present paper, following in the spirit of George Zipf,goes some way toward providing new and worthwhile mathematical insights ofinterdisciplinary significance, drawn from one frontier of modern linguistic theory.

2 The linguistic data

South-east-Australian language was amongst the most unchanged ancient oraltraditions still existing on Earth when the ‘First Fleet’ arrived in Sydney Cove fromEngland more than two centuries ago. For many tens of thousands of years, south-east Australian Aboriginal people had been geographically isolated and separatefrom human populations with different origins, and they certainly were not tradingwith south-east Asian or Pacific neighbours. This contrasts with languages aboutAustralia’s far-north and north-western fringe that do contain recognizable south-east Asian influences and, having evolved rapidly in recent millennia, are less likelyto resemble any original ancestral language brought to Australia 80 000 or so yearsago, across the south-east Asian landbridge and oceans, ultimately from Eurasia.To the extent that south-east Australian language was the most remote and isolated,it might be expected to have retained the most ancient features.

Although south-east-Australian Aboriginal language was the first to be extingu-ished, by the ensuing European colonization and settlement, it was documented ina literary tradition spanning the centuries back to First Fleeters such as WilliamDawes, Watkin Tench, David Collins and John Hunter. There were also a numberof important scientific expeditions to Australia, including those of Quoy & Gaimard(1826) from France and Hale (1846) from the USA, supplementing linguistic andethnological materials gathered by early missionaries such as Gunther (1830s) inthe Wellington Valley, Threlkeld (1834) north of Sydney to Newcastle, AndrewMackenzie (1875) in the Illawarra, John Bulmer (1860s) in Gippsland, andTeichelmann and Schurmann (1840) in Adelaide. By the late 19th century, variouscompilations of Aboriginal words began to occur ranging from those of Curr(1886), Dawson (1881) and Ridley (1875), to that of Schmidt (1919).

Unfortunately early English observers misheard the vowels and various novelsounds, frequently failed to hear or record word-initial consonants, and employedall the worst orthographic traditions of English spelling—including poor phoneme/grapheme correspondences and ‘orthographic-fossil’ symbols and clusters of sym-bols that are silent and/or inappropriate—with the result that Aboriginal wordswere rarely ever spelled in the old written records anything like how they werespoken. Even today words such as enough (ó inuf ), phlegm (óflem), light (ó lait),quay (ókai), yacht (óyot) etc, illustrate spelling problems that should long-agohave motivated the English-speaking world to ‘put a broom through’ the OxfordDictionary, making graphemes and phonemes correspond.

Whether or not contemporary Australianists realize it, the severity of thisubiquitous orthographic ‘static’ in decrepit archival source materials has been amain factor contributing toward the sense of hopelessness and defeatism in allaspects of south-east Australian linguistic salvage work. The prevailing view isextremely pessimistic, with every reason other than the obvious one being given toexcuse slow progress in this field:

Chapter 1, JAS 2003

Page 116

Lexigenesis in ancestral south-east Australian Aboriginal language 115

There has probably been so much language loss, language contact,and possibly also radical restructuring of languages brought about byinterrupted transmission that proto-Australian may well prove to beunreconstructable in principle, if it ever existed . . . linguists would bebetter off for some time yet continuing to chip away at the lower levels ofreconstruction, and offering the results of their work to archaeologists forinterpretation. (Crowley, 1997, p. 295)

Two centuries of failure entitle us to suspect that it is a waste of time trying tounderstand the various English spellings employed in old ‘word-lists’ based uponwhere in England the respective observers came from, whether they spoke rhoticor non-rhotic English, what their vowel systems were etc. Most were barely literate,and could not spell sensibly for that reason alone—even the educated ones spelledthe same sounds differently almost every time they encountered them! This lackof consistency is typified by the writings of Thomas Mitchell, for example, whorecorded the same words on different occasions as below:

1878 1887

footó ‘teyrah’ ‘gerra’moonó ‘yow-warra’ ‘huerra’headó ‘booaw’ ‘bua’noseó ‘nar’ ‘nga’

Rather than trying to understand such frustrating orthographic ‘static’ in decrepitliterary/archival sources, we need instead to eradicate it using a firm knowledge ofthose sounds that were actually being spoken by 18th and 19th century Aboriginalpeople.

Colin Yallop’s (1982, pp. 55–71) ‘Australian Aboriginal Languages’ contains anexcellent chapter on the sounds of Australian languages. Additionally, Dixon’s(1980, pp. 158–159, 129–130) book ‘The Languages of Australia’ provides agenerally accepted inventory of proto-Australian consonants and vowels, along withsuggested phonotactic rules.

Unfortunately, however, no Australianists has yet had the courage to aggressivelyand systematically use any such tools to clean-out the ‘static’ from centuries ofaccrued decrepit archival materials. Sources ranging from Curr (1886) to Capell(1970) and ‘the Macquarie [Dictionary of] Aboriginal Words’ (Thieberger &McGregor, 1994) invariably revert to using the Roman alphabet and inappropriateEnglish spellings, including digraphs and far too many vowel sounds, to representthe quite distinctive non-English sounds of south-east Australian language.

The present author suggests that ambiguity in the spelling of south-east Aus-tralian Aboriginal words may be reduced through the use of an ‘alphabet’ of 18symbols denoting four ‘vowels’ and 14 ‘consonants’:

a, b, d, d, , g, i, l, m, n, ŋ, n, �, r, u, w, y, υ

where υóaya or perhaps aia.This orthography recognises four ‘vowels’ a, i, u and υ, although the latter,

actually a cluster, is not strictly a vowel but it behaves like one and will henceforthbe referred to as ‘the long-vowel’. The symbols other than this ‘long-vowel’constitute a reasonably standard usage of the International Phonetic Alphabet

Chapter 1, JAS 2003

Page 117

116 C. R. Illert

(IPA)—bearing in mind that, with only four contrasting ‘vowels’ (as compared to20 in English), proto-Australian can afford to allow each vowel a much wider rangeof phonetic realization. Thus, we need no symbol specifically denoting the Englishoo vowel sound (as in the word ‘foot’) which, along with other cherished Englishvowel sounds, are all adequately described in our orthography by the single symbolu, leaving us free (for reasons of typographical utility) to commandeer the redundantkeyboard character υ for portmanteau use as above. Additionally it is well knownthat there is only one lateral, l, in all language throughout one third of the entireAustralian continent located roughly east of the meridian through Melbourne(Dixon, 1980, map 5, p. 141). In addition, although some researchers suggest theexistence of two rhotics in proto-Australian (Dixon, 1980, p. 158) there is nopresent need for separate symbols to denote them. Just as ’the articulation of theoccurrence of a vowel in a particular word is often affected by the phonetic valueof the consonants flanking it’ (Dixon, 1980, pp. 129–130), so too might differentrhotic sounds arise from r being flanked respectively by long or short vowels: forexample, mυra sounds quite different to mara. Thus, one lateral and one rhoticseem sufficient for our present purposes. It is uncertain whether the laminals d andn existed in proto-Australian, but they were certainly part of south-east Australianlanguage two centuries ago (Eades, 1976) so are included in our ‘alphabet’. Thatthis system seems to apply across numerous 18th and 19th century linguisticboundaries throughout the south-eastern quadrant of mainland Australia, a vastgeographical region maybe 1000 km in diameter, suggests that it must approximatea proto-phonology. At the very least it is a plausible and useful basis fromwhich to attempt salvage and diachronic analysis of extinct south-east Australianlanguages.

Even if this orthography is not completely correct it is nevertheless quite capableof significantly cleaning up decrepit literary and archival sources, eliminatingimpossible sounds and absurd spellings, revealing useful and important informationsuch as the previously unknown set of conjugate-pairs of ancestral adjectival-roots,presented in Tables 1(a) and 1(b), which appear to underlie all Aboriginalvocabulary throughout south-eastern Australia, at least from Newcastle to Sydneyto Canberra to Melbourne to Adelaide—yet have been entirely missed by genera-tions of Australianists for 200 years, judging by the statement ‘. . . it is not at allclear that there is such a thing as a universal stock of basic vocabulary . . .’ (Yallop,1982, p. 33).

Another notation in the present paper utilizes brackets and colons in an intuitiveway, just as we might write the English expression they’ll as they:(wi)ll using bracketsto denote the missing bit and a colon to separate individual morphemes. The needfor this notation is apparent, considering that English speakers tend to run ngtogether as in the word ’singing’ and, although a velar nasal ŋ does indeed exist inAboriginal language, there is also another quite different sound denoted n:g wherethe g is separate like the terminal k on the English word ‘disk’. To pronounce thisn:g as ŋ is seriously wrong—but how would we know without the colon? Likewisethere is an ŋ:g sound, as in the spoken English word ‘finger’, also requiring a colonas a separator.

Our precise orthography, in combination with this common-sense notation ofbrackets and colons, provides a scientific notation of sufficient precision to enablestudy of the fine-structure of condensed words, revealing what lies below thesurface of decrepit 18th, 19th and 20th century ‘word-lists’. For example, theadjectival roots in Table 1 were commonly combined together in small phrases and

Chapter 1, JAS 2003

Page 118

Lexigenesis in ancestral south-east Australian Aboriginal language 117

T 1. (a) Conjugate pairs of ancient south-east Australian adjectival roots which emerge from oldword-lists when the suggested phonology is applied

bulυ gulυ malυódown, low, flat, dead, ódull, lethargic (as with Koala), óblind, obstructing, shielding,cessation petrified opaque, (eye) cataracts

bυ lu gυ la mυ lυ, milióup, high, projecting, óshiny, radiant, vital, nice/good, óvisible, accessible, clear,commencement alive channelling

bulalaó2 murυ laó5

gula mulυómalevolent, deadly, treacherous, ófamiliar, intimateangry, sorcery

galu mυ luóbenevolent, helpful, honest, óunfamiliar, strange,happy unexpected

burυ gurυ maluónoisy, quick, energetic, awake ófat, globular, convex more ó[air] bubbles (as with

Platypus)

bυru gυru mulaóquiet, slow, lazy, tired, sleepy, óslim, concave, less ó[menstrual] droplets, progeny,still dust

barυ garυ malaóeither-side, opposing ófrom, out-of, sound óinflated, bulging, enhanced

bυra gυra muluóbetween ótoward, into, throat, orifice, ódeflated, shrivelled,

cave, valley emaciated, diminished

buru gura mυraóbouncy, hoppy (as with ófirm (as with stones, hail or ó large, greatlyKangaroo) muscles)

bara garu murυówaddling (as with Wallaby) ósquishy (as with nasal mucous) ósmall, slightly

T 1. (b) Conjugate pairs of ancient south-east Australian adjectival roots that emerge from old‘word-lists’ when the suggested phonology is applied

ŋulu dulυósinusoidal (dunes, waves), óshabby, dirty, untidy, bentfolded (legs while sitting),corrugated or wrinkled( forehead, mushroom gills),crushed (chewed food)

dυ laó1 wurυ laó3,singular, unity plural, several, lots of

ŋuru darυ wirυócyclic, oscillatory, twinkly, óabout, external, outside surface óprofane, improper, left,shimmery crooked

ŋara dυra wυrióknotted, coiled, spiral óthrough, internal, inside ósacred, proper, right, aligned

Chapter 1, JAS 2003

Page 119

118 C. R. Illert

T 2. Simple phrases involving condensation

‘Words’ recorded by Dawes ‘Words’ recorded by MannSimple phrases (lexemes) involving (1790)—with ascribed (1840)—with ascribedcondensation—with correct meanings meanings meanings

ŋu(ru):wi�:dυ (la:n) (cyclically- ‘gu:wiŋ::’ (‘sun’) ‘goo:win:dia:’ (‘star’)orbiting-thingó sun, star)

w(uru-garυ):i:(bu)lυ-�(uŋ) ‘w::e:lla:ŋa’ (‘to follow or be ‘w::ee:lo:ng’ (‘behind’)(óbehind [distant]) second’)

�uru-bu(lυ :i):dυr(a):n ‘yuru:pa::ta:’ (‘very hungry’) ‘nurru::::ng’ (‘empty’)(empty stomachóhungry)

m(i�):bυ l(u):ŋ:bυ l(u) m:ull:n:aoul’ (‘tomorrowmorning’)

(mi�):bυ(lu):ŋ:bυ(lu) ‘:bu:ng:eye’ (‘this daynow present’)

(mi�):bυ(lu):ŋ:bυ l(u) (commencement ‘:bu:n:yell (‘sun’)of dayósun rise, dawn)

‘condensed’ in different ways, as shown in Table 2, often creating several seemingly‘different’, but actually equivalent, versions of the same lexeme—see ‘sunrise/dawn’in Table 2 above.

First Fleeters encountered, what appeared to them to be, an astonishing lexicaldiversity in Aboriginal language about the Sydney region. Collins remarked that:

We often heard, that people from the northward had been met with, whocould not be exactly understood by our [Sydney Aboriginal] friends; butthis is not so wonderful as that people living at the distance of only fiftyor sixty miles should call the sun and the moon by different names . . . Inan excursion to the banks of the Hawkesbury, accompanied by twoSydney natives, we first discovered this difference; but our [SydneyAboriginal] companions conversed with the river natives without anyapparent difficulty, each understanding or comprehending the other . . .We have often remarked a sensible difference on hearing the same wordsounded by two people; and, in fact, they have even been observedsometimes to differ from themselves . . . . Collins (1798, vol. 1, pp. 506)

Europeans never did understand that ‘the same word’ was spoken with ‘sensibledifference’ on different occasions, by the same and other informants, becauseAboriginal language was to a large extent operating at the level of root morphemesas in Table 1, instead of at the level of lexemes (‘words’) as in Table 2. Hencevariable condensation of morphemes within a lexeme—a small matter to any nativespeaker—constituted a monumental difference (‘a different language’) to Europeanears. This is why the Sydney natives had no difficulty understanding their rivercousins whose lexemes (‘vocabulary’), including different names for sun and moon,seemed to European ears to be totally unrelated. Additionally, Capt. John Hunter(1793) wrote about an even more novel feature of the language, permutation ofmorphemes within a lexeme:

Chapter 1, JAS 2003

Page 120

Lexigenesis in ancestral south-east Australian Aboriginal language 119

It should be observed, that in speaking, wolle-warre frequently changesthe position of his words . . . so when walking one night from ProspectHill to Rose Hill, we frequently stumbled against roots, and he exclaimedwere wade, and wade were, ‘bad wood’, or ‘bad roots’.

as inwirυ ñ wa(ra):d(arυ :n)ñ i:(garυ)� ‘‘were:wa:d:: e:’’

óbad, evil thither—goingó journeying, travelling

wa(ra):d(arυ :n)ñ i:(garυ) ñ wirυ� ‘‘wa:d:: e:were’’

thither—goingó journeying, travelling óbad, evil

Further examples of exactly this kind of ‘morpheme-shuffling’, provided in Tables3 and 4, leave little doubt that all of these lexemes belong to a single common

T 3. Simple phrases involving permutation

‘Words’ recorded by ‘Words’ recorded by MannLeximorphs arising from permutation— Dawes (1790a)—with (1840)—with ascribedwith correct meanings ascribed meanings meanings

wi�-dυ(la:n) ‘yana:da:’ (‘moon’)

dυ la:(n)-wi� (orbiting-thing) ‘yello::wna’ (‘moon’)

wu(rυ la:gυn)-b(ulυ):i- ŋ ‘wu::b:i:n (‘possum’)

bu(lυ:i)- ŋ -wu(rυ la:gυn) ‘bu::ng:u:’ (‘squirrel’)

g(υn):wurυ(la)-bulυ:(i- ŋ) ‘g:re:bble::’ (‘oppossum’)(slowly-falling/gliding creatureóthe SugarGlider Petaurus)

g(υra):i-�(in:�uŋ)-da(rυ:n):wu(ru) ‘g:i:ni::dya::ou’(‘I have[or did] crack [the flea]’)

wu(ru):n:da(rυ -�in:�uŋ)- g(υra):i ‘u:n:de:::k:i’ (‘come here’)(ócoming-hither, converging)

T 4. Simple phrases involving permutation

‘Words’ recorded by ‘Words’ recorded byLeximorphs arising from permutation— various observers—with various observers—withwith correct meanings ascribed meanings ascribed meanings

dυ la-ma(lυ)-n ‘hila:ma:n’ (‘shield’)(Rowley, 1875)

‘ilee:mo:n’ (‘shield’)(Tench, 1798)

malυ-dυ(la)-n ‘mela:tho:n’ (‘shield’)(obstructing-thingóshield, cataract) (Atkinson, 1853)

�uru-gu(la):mυ(ra) ‘nooro:ca:my’ (‘[a long]spear with one barb’)(Collins, 1798)

mυ(ra):gu(la)- �(uru) ‘maia:gu:ŋ (‘spear’)(óhollow (reed) barbed [spear]) (Malone & Malone, 1875)

Chapter 1, JAS 2003

Page 121

120 C. R. Illert

language based upon the universal roots given in Table 1. It becomes quite difficultto argue that the lexemes in Tables 2, 3 and 4 differ, even at dialect level, onceone understands that what Europeans thought of as unique fixed ‘words’ weremerely lexemes with multiple equivalent cognates deriving from parent leximorphsthat vary according to the permutation and condensation of constituentmorphemes.

These leximorphs in Tables 3 and 4 are a conceptual breakthrough, for the firsttime enabling Australianists actually to understand the linguistic basis underlyingTench’s (1798) observation that:

Although our natives and the strangers conversed on a par and understoodeach other perfectly, yet they spoke different dialects of the same language;many of the most common and necessary words used in life bearing nosimilitude, and others being slightly different. That these diversities arisefrom want of intercourse with the people on the coast can hardly beimagined, as the distance inland is but thirty-eight miles; and from RoseHill not more than twenty; where the dialect of the sea-coast is spoken.It deserves notice that all the different terms seemed to be familiar toboth parties, though each in speaking preferred its own.

To appreciate how this revolutionizes existing Australianist scholarship we needonly consider the paper of Capell (1970) which contrasted early ‘word-lists’(actually lexeme-lists) obtained in the vicinity of Sydney by Dawes (1790a) andMann (1840s). Dawes obtained his lexemes at Sydney Cove, on the southern sideof Sydney Harbour, near the present Sydney Opera House, mainly from a youngwoman ‘budigurυŋ’ (ó‘tubby’) with whom he was living. Mann obtained his wordsfrom ‘Long Dick’, son of ‘Bungaray’ and ‘Queen Gooseberry’ of the PittwaterTribe, from Broken Bay on the northern side of Sydney Harbour. These twogroups traded and intermarried and would have met each day, in their canoes,whilst fishing out on Sydney Harbour. It is simply inconceivable that they wouldnot have been able to understand each other.

Yet, based upon a cursory comparison of their different ‘words’ (actuallylexemes), Capell (1970) concluded:

. . . the Sydney language did not cross Port Jackson . . . [it] was limited tothe peninsula on which Sydney now stands . . . [this is] evidence notsimply of a new alignment of tribal boundaries about Sydney, but of alanguage new to the Australian inventory. The fact that the Sydneylanguage—whatever it should be called—did not cross Sydney Harbour. . . has some historical importance . . . the difference between the two isat language level, not dialect level; it would be necessary for natives ofeach area to learn each other’s language, as apparently they did, or theexistence of two languages would have been noticed at the first.

In fact, a morpheme-level comparison between the Dawes and Mann ‘word-lists’, as in Tables 2 and 3, reveals that natives on both sides of Sydney Harbourwere speaking exactly the same language. One would be scratching to even arguein support of a difference at dialect level. Capell (1970) got this wrong and is byno means alone amongst Australianists. Just about all determinations of linguisticboundaries throughout south-eastern Australia, made in recent decades basedupon ‘lexico-statistical’ analyses (Crowley, 1992, pp. 168–190), are likewise wrongfor the same reason—they unjustifiably assume that language functioned at the

Chapter 1, JAS 2003

Page 122

Lexigenesis in ancestral south-east Australian Aboriginal language 121

level of lexemes (‘words’) rather than at the level of morphemes. The very name ‘lexico-statistics’ tells why this methodology, still in universal use amongst Australianists, isfundamentally flawed. The only meaningful basis for comparison of lexemes fromold ‘word’-lists would be ‘morpho-statistics’.

Dixon, who also claims little difficulty drawing firm language and dialectboundaries throughout south-eastern Australia, admits that ’the methods oflexicostatistics are not fully applicable to the Australian situation’ but asserts thatthe trick is to ‘place equal reliance on grammar. . . . any tentative conclusions fromshort vocabulary comparisons need to be verified by comparing the full grammaticalsystems . . .’ though even this has limits because ‘. . . languages change at such arate that, after more than about three or four thousand years of separation,genetic links are no longer recognisable’ (Dixon, 1980, pp. 37, 229, 237). Thus,Australianists, who generally agree that all language from the southern portion ofthe Australian continent ‘is related’, have made ’no attempt . . . to PROVE a geneticconnection, by explicit reconstruction . . .’ (Dixon, 1980, p. 222).

So, whilst Australianists cringe at the problem of extracting ‘information fromnoise’, researchers in other fields such as Stanley et al. (1999) are busy ‘. . . adaptingto DNA the Zipf approach to analysing linguistic texts, and the Shannon approachto quantifying the ‘redundancy’ of a linguistic text in terms of a measurable entropyfunction . . .’.

The present paper supplies a ‘workable’ proto-phonology and a list of ancientsouth-east Australian adjectival-roots in Tables 1(a) and (b) which, alone, are morethan has been accomplished to date by any previous researcher. Additionally, thediscovery of leximorphs (as in Tables 3 and 4) enables multiple independentconfirmations (‘sightings’) of ancient adjectival-roots, within the different syntactic-environments of cognate-lexemes from scores of genetically related ‘modern’ south-east Australian languages, in a fashion never previously possible. This enables amathematical description of ‘optimal’ lexeme condensation which, in turn, providesa reliable new basis for diachronic studies spanning previously unimaginable times(tens of thousands of years instead of mere thousands)—so much so that we mayeven have identified the first ‘lexical echoes’ from another continent in such ancientyet recognisable words as: ŋu:b:bulónipple, barυ :dóbreast, gυ :g:wυ lócockle,g:wurυ lóquail or squirrel, dυ la:nózahlen or tally, mirυ :d:mi:dυóarithmetic. But,even should these particular examples prove coincidental, it is certain that no linksever will be found if Australianists continue to accept the defeatist premise that allevidence is lost, making it pointless for scholars even to discuss ‘. . . geneticconnection between Australian languages and anything outside the continent . . .’(Dixon, 1980, p. 238).

3 Morpho-statistics

Root-morphemes, perhaps from an isolating proto-language, were originally up tofour phonemes long, as in Tables 1(a) and 1(b). The passage of time producedstable lexemes (the linguistic equivalent of ‘molecules’) through agglutination andmorphological fusion of these root-morphemes within ‘common phrases’ (calledleximorphs), which varied in complexity up to full transitive sentences formed fromthe intersplicing of a noun-phrase (containing an n or ŋ marker) and a verb phrase(containing an i marker). Examples have already been given of root-morphemesvariously condensed (as in Table 2) and permuted (as in Tables 3 and 4). To describe

Chapter 1, JAS 2003

Page 123

122 C. R. Illert

F. 1. Hamming Value H (the number of missing phonemes/morpheme).

more accurately this process of agglutination and morphological-fusion we definethe Hamming Value, H, as the number of phonemes missing from any givenmorpheme. We then take any lexeme and plot a graph/histogram of f, the frequencydistribution of its variously condensed root-morphemes, versus their respectiveHamming values over the interval H é [0, 4].

The downward sweep of this frequency distribution reflects a belief that severecondensations, especially the omission of entire root-morphemes, should occur lessoften. In this model, the total number of (variously condensed) root-morphemesin a lexeme is

Tó;4

Hó0fH

and their mean-condensation is

có;

4

Hó0Hî fH

;4

Hó0fH

ó 1Tî;

4

Hó0Hî fH

The simplest model of lexeme composition assumes that half the root-morphemes(note: a better approximation is f0ó0.45T) are not condensed at all, with theremainder being variously condensed according to the Zipf distribution

fHB

12

T

Hò1. . . for H é [0, 4]

where

Chapter 1, JAS 2003

Page 124

Lexigenesis in ancestral south-east Australian Aboriginal language 123

TZipfó ;4

Hó0

12T

Hò1ó �1

1ò1

2ò1

3ò1

4ò1

5�î12

TB T

and

cZipfó;

4

Hó0Hî

12T

Hò1

;4

Hó0

12T

Hò1

ó�0ò1

2ò2

3ò3

4ò4

5��1

1ò1

2ò1

3ò1

4ò1

5�B 1.2

which is an average condensation of slightly more than one phoneme per root-morpheme.

Example 1

sun-rise /dawn (óbirth /commencement [of the] day)

A comprehensive survey of relevant south-east Australian literary and archivalsources reveals that the lexeme for ’sunrise/dawn’, appearing in Table 2, is just afragment from a set of (at least four) equivalent leximorphs which are as follows:

ó[the] solar day which [is]

[the] sun shine–cycle which [is]

wi� :(dυ la :n)ñ i:bυ (lu) ñ ŋuru :gυ lañ bυ l(u) ñ mi� �rising up

óbeing born, commencing

�‘‘win@yoo:::a:boo:ng@::ba:mın’’ ‘‘::::bu:ng::eye:’’ (‘‘this day now present’’)

(‘‘lit., sunrise’’) ‘‘::::bu:n::yell:’’ (‘‘sun’’)

(Everitt, 1900){gunduŋara} (Mann, 1840) {eora} �and

ó[the] solar which [is] óday

wi� :(dυ la:n)ñ b(υ lu): iñ mi� ñ bυ l(u) ñ gυ l(a):ŋuru�óbeing born, commencing

�‘‘win:::b:i::::n’’ (‘‘morning’’)

‘‘:::b:i:m:bul::ong’’ (ó‘‘Tuesday’’ ???) [ J. Rowley, 1875] {gunduŋara}

‘‘:::::mun::gaal:oon’’ (ó‘‘Sunday’’ ???)(Throsby, 1821) {dariwυ l} ‘‘:::::mun:::garra’’ (‘‘sun’’)

(Bunce, 1853) {darling downs} �

Chapter 1, JAS 2003

Page 125

124 C. R. Illert

and

ó[the] solar day which [is]

ŋu(ru:gυ la)ñwi� :dυ (la:n) ñ b(υ lu)ñ i:bυ lu ñ mi� �óbeing born, commencing

‘‘co::ing:::b:i:bo:ba’’ (‘‘sun rising’’)

‘‘co::ing:::::’’ (‘‘the sun’’) ‘‘co::ing:::b:i:bola:’’ (‘‘the sun

(Collins, 1798) {eora} rise’’)

(King, 1793) {eora}

‘‘ca:::::b::bo:ne’’ (‘‘full moon’’)

(Lingard, 1838) {ŋunuwυl} ‘‘neo::wee::::::’’ (‘‘sun’’)

(Dawson, 1881) {chaap whurrong}

‘‘goo::win:dia:::::’’ (‘‘star[s]’’)

(Mann, 1840s) {eora} ‘‘q::uen::::::’’ (‘‘sun’’)

(Paine, 1794) {eora}

‘‘no::ween:th:::::’’ (‘‘sun’’)‘‘ky::un::::::’’ (‘‘sun’’)

(Bunce, 1853) {darling downs}(Rowley, 1875) {gunduŋara}

� �and

ó[the] solar day [which is]

dυ (la:n)ñbυ (lu)ñwi�ñbυ (lu): iñgυ l(a:ŋuru)�rising upóbeing born, commencing

�‘‘ta::bbaro:wrin:burr:ee:t:’’ (ó‘‘sunrise’’)

‘‘::bur::p:i:gal:’’ (ó‘‘morning’’)‘‘ta::baro:wrin:bourr:i:t:’’ (ó‘‘sunrise’’)

(Mathews, 1901) {dharruck}(Pettit & Dawson, 1850) {kurnai} �

Lexemes from this set of leximorphs have mean-condensation rates, c, rangingtypically from about 1.8 up to 2.4. Smaller values indicate the relative independenceof morphemes, and looseness of clustering, within their respective lexeme. Greatercondensation rates imply the emergence of a lexeme as a ‘word’ in its own right,with the individual meanings of constituent morphemes becoming less importantover vast time-periods. By comparing Hamming distributions of lexemes in varyingstates of condensation, we can guess the relative-age and also identify thoseindividual morphemes considered most deletable by native speakers.

Chapter 1, JAS 2003

Page 126

Lexigenesis in ancestral south-east Australian Aboriginal language 125

Example 2

sun-set/dusk (ósun going-down [distant])

Conversely, we have a set of leximorphs for sun /moon-setósun/moon

thing orbiting

bulυ ñ d(υ la:n)ñ garυ : i:�(uŋ:nυ ra)ñwi��ódown ógoing there (distant)

‘‘bee::t::go::n@ya::winyoo’’ (‘‘sun goes down’’)

(Everitt, 1900) {gundungara}‘‘pu:::ker:i:ng::’’ (‘‘sun’’)

(Mackenzie, 1874) {durga}‘‘bulla:::jarra::ng::’’ (shine’’)

(Mathews, 1903) {dariwυl}‘‘bu:::gur:i:n::’’ (‘‘sun’’)

(Hale, 1839) {durga}‘‘ba:::kuru::ŋ::’’ (‘‘sun’’)

‘‘bu:::kur:i:ŋ::’’ (‘‘hot’’)‘‘boo:::goo::ya::’’ (‘‘sunset’’)

(Malone, 1875) {dhariwυl}‘‘bu:g::gr:e:n::’’ (‘‘sun’’)

(Larmer, 1853) {durga}‘‘bu:g::gera::ng::’’ (‘‘sun, hot’’)

(Mowle, 1838; Curr, 1887) {ŋunuwυl}

� �and

sun/moon

gar(υ : i)ñbu(lυ )ñnυ (ra): �uŋñwi�:(dυ la:n)�going down there (distant)ósetting

�‘‘:::::winu::’’ (‘‘hot’’)‘‘kaa::pa:(g)nuu:nang:::’’ {peek whurrong}

(Russell, 1914) {gunduŋara}‘‘mir::ma:neo::wee::’’ {chaap Whurrong}

(Dawson, 1881) (‘‘go down sun’’) �and

orbiting shining thingósun/moon

wi�:d(υ la:n)ñ garυ : iñ bulυ ñ�(uŋ:nυ ra)ñ gυ l(a)�going down there (distant)

�‘‘:::yarra::bu:nye::’’ (‘‘go away’’) ‘‘woi:d::g:i:t:::gull’’ (‘‘Wednesday’’ ???)

( Larmer, 1853) {durga, Braidwood} (Throsby, 1821) {dariwυ l} �

Chapter 1, JAS 2003

Page 127

126 C. R. Illert

Again it is instructive to plot Hamming distributions for lexemes in varying statesof condensation:

All these words contain some version of the morphemes garυ : iógoing andbulυódown hence ‘setting’. The most deletable morphemes seem to benυraóyonder /there [distant] and dυ la:nóthing [singular]ó‘sun’.

These examples demonstrate that our estimate of 1.2 is too small for the Zipfmean-condensation. The smallest values attained in actual lexemes werecó16/9ó1.7777 (‘sunrise’) and có15/8ó1.875 (‘sunset’). For simplicity ourinitial estimate assumed H é [0, 4] . . . the Hamming spectrum of root-morphemes.

Chapter 1, JAS 2003

Page 128

Lexigenesis in ancestral south-east Australian Aboriginal language 127

But there are also morphemes which, although not fundamental roots, neverthelesscommonly occur in lexemes. Some of these morphemes contain up to six phonemes;examples being the numerals bulalaó2, wurυ laó3, nurυ laó4 and murυ laó5.This means that our earlier estimate neglected the asymptotic tail of the Zipfdistribution, beyond Hó4, and that a better estimate for the Zipf distribution isas follows:

fHóf0

Hò1for H é [0, 6]

hence

cZipfó;

6

Hó0Hî f0

Hò1

;6

Hó0

f0

Hò1

ó�0ò1

2ò2

3ò3

4ò4

5ò5

6ò6

7��1

1ò1

2ò1

3ò1

4ò1

5ò1

6ò1

7�ó 4.40714

2.59286B 1.7

which is very realistic! Thus, even though lexemes such as ‘sunrise’ and ‘sunset’may not themselves contain any of these larger (six phoneme) morphemes, themorphemes that they do contain are nevertheless condensed according to a lawthat takes into account the existence of larger morphemes. Clearly the asymptotictail of the Zipf distribution cannot be ignored. A useful formula for calculationpurposes is:

cZipf(m)ó còmñ1ñln(mò1)cò ln(mò1)

for H é [0, m]

where the Euler constant có0.5772 15664 90153 28606 06512.

Example 3

female-breasts /nipples (ópair of lactators)

A set of leximorphs, containing six-phomeme morphemes and demonstrating thefull H é [0, 6] tail of the Zipf distribution, underlies lexemes for women’s breasts(ó large [things]) or nipples (ósmall [things]):

ba(rυñ�murυmυra�ñ garυ): i:ŋ(uru):w(ur)υ l(a)ñ bula(la:n)

óeither-side ó lactating pair of [things]ó large/small

‘‘p:::e:ng:al::’’ (‘‘poitrine’’óchest, bosom)

‘‘pa::::::pilla:’’ (‘‘nipple[s]’’) ‘‘bera:::::::’’ (‘‘joue’’ócheek[s])

(Pliny the Elder) {latin} (Gaimard, 1826) {durga}

� (McCaffrey, 1920s) {dariwυ l}

‘‘b:::i:ng:ul::’’ (‘‘breast[s]’’)

(Hale, 1846) {durga} pa:::i::yil::’’ (‘‘the breasts’’)

(Threlkeld, 1834) {awabakal}� �

Chapter 1, JAS 2003

Page 129

128 C. R. Illert

and

ólactating

ŋu(ru:wurυ la)ñ bar(υ):bula(la):nñ i:gar(υ)ñ�murυ)mυra)�

óeither-side pair of things ó large/small

‘‘ngu::b:bu:n::g:’’ (‘‘female breast’’)‘‘:::pial::i:go:’’ (ó‘‘Canberra’’)

(Mathews, 1901) {darug}(Gillespie, 1984) {ŋunawυ l}

‘‘nu::p:pa:n::g:’’ (‘‘teats of animals’’)‘‘:::yal:::gar:’’ (‘‘breast’’)

(Dawson, 1881) {kuurn kopan noot}(Munima, 1863) {dariwυ l}

�‘‘na:::bu:n::g:’’ (‘‘breast, nipple’’)

‘‘ga:::bu:::g:’’ (‘‘nipple’’)(Rowley, 1875) {darug}

(Threlkeld, 1834) {awabakal}(Collins, 1798) {eora}

‘‘ni::p:pl::e::’’ {english}‘‘::bar:bu:n:::’’ (‘‘arms, oars’’)

‘‘:::n ::g :cg :’’ (‘‘source, spring’’) {greek}(Mragula’s Song, 1890s) {kurnai}

� �and

ó lactating

ŋu(ru):wur(υ la)ñmυ (rañ garυ): iñ�dυ (la)

bula(la)�:n-barυñ�uŋó large one’s

óeither-side [pair of ] thing[s]

‘‘nga::mm::i::::’’ (‘‘female breast’’)‘‘ngu::m::i::::nyang’’ (‘‘female breast’’)

‘‘nga::mm::i::::(ngurru)’’ (‘‘milk’’)(Mathews, 1904) {ŋunawυ l}

‘‘nga::mma::i:tya:::’’ (‘‘woman’’)(Eades, 1976) {durga}

(Teichelmann & Schurmann, 1840) {kaurna}�‘‘nga::::::n:bra:’’

‘‘a::m::::::gnann’’ (‘‘mamelle’’óbreast) (ó‘‘Canberra’’)‘‘kga::m:::::burry:’’ �

‘‘a:wa:::::::nhan’’ (‘‘lait’’ómilk)(Gillespie, 1984) {ŋunawυ l}

(Gaimard, 1826) {durga}� �

also the familiar place-name ‘‘Canberra’’.1

ó lactating

gar(υ : i)ñm(υrañŋuru:wurυ lañ bulala):n-barυó large óeither-side [pair of ] thing[s]

‘‘caa::m:::::berra’’‘‘ca::::::n:berry’’

(Mowle, 1840s)‘‘ca::::::n:burry’’(Moore, 1826, 1831)

‘‘ke::m:::::bery’’�(Lhotski, 1835)

‘‘ca::::::n:berry’’‘‘ca::m:::::berry’’

‘‘ka::m:::::berra’’(Department of the Surveyor General, 1837–38)

(Wright, 1907)� �

Chapter 1, JAS 2003

Page 130

Lexigenesis in ancestral south-east Australian Aboriginal language 129

Chapter 1, JAS 2003

Page 131

130 C. R. Illert

Example 4

bivalved-shell (ópair of opposing spiral[thing]s)

Another set of leximorphs containing six-phoneme morphemes, and demonstratingthe full H é [0, 6] tail of the Zipf distribution, is as follows:

óslightly/gently spiralling

bulala:(n)ñ�i:m(urυ)

gυn:wurυ l (a):gυ(n: i )�ñ ba(rυ)ñŋ(ara)�

ópair of [things] óeither-side[opposing]

‘‘le:::ro:ko:::’’ (‘‘huitre’’óoyster)‘‘tallo::::p:’’ {peek whuurong}

‘‘::kan:el::::’’ (‘‘moule coquille’’ómussel )[pl. 78]‘‘chaluu::::p:’’ {chaap whuurong}

‘‘:::wooura::::’’ (‘‘venus’’ócockle)[pl. 84](Dawson, 1881) (‘‘shell of mussel’’)

(Quoy & Gaimard, 1826) {durga, Jervis Bay}� �and

óslightly/gently spiralling

ba(rυ)ñ�i:m(urυ)

gυn:gυn:w(ur)υ l(a): i�ñ bula(la:n)ñŋ(ara)�

óeither-side pair of [things][opposite]

‘‘t:i:m:bo::nn’’ (‘‘shell of a mussel’’)

‘‘bi:i:m:bala::’’ ‘‘pa:k:kan::e:::ung’’ (‘‘shells generally’’)

(‘‘cockle, mussel, mud-clam’’) (Dawson, 1881)(Eades, 1976) {durga: Jervis Bay} {kuurn kopan noot, peek whuurong: melbourne}

‘‘pee:::w:ee:::’’ (‘‘shells’’) ‘‘::co:q:uill:e:::’’ {french}

(Munima, 1863) {dariwυ l: Illawarra} ‘‘:co:ck:l:e:::’’ {english}

‘‘:ka:k::i:::’’ { japanese}� �

Chapter 1, JAS 2003

Page 132

Lexigenesis in ancestral south-east Australian Aboriginal language 131

Example 5

gliding-possum (óslowly-falling [gliding] creature)

The gliding possums Petaurus, including the ‘sugar glider’ and its cousin the‘squirrel (g:wurυ l ?) glider’, have the name diving /gliding creature. The relevantlexemes, appearing in Table 3, are fragments from the following set of leximorphs:

creature

g(υn):wurυ (la)ñŋñ bulυ : i�slowly falling

ógliding

Chapter 1, JAS 2003

Page 133

132 C. R. Illert

‘‘:werri:m:b:i’’ (‘‘flying fox’’)

(Russell, 1914) {gunduŋuru}‘‘g:re::bble:’’ (‘‘flying fox’’)

(Mann, 1840s) {eora}‘‘:warra:n::’’ (‘‘flying squirrel’’)

(Mathews, 1902) {kurnai}� �

and

wu(rυ la:gυn)ñ b(ulυ): iñ ŋ ��‘‘wo::bb:i:n’’ (‘‘flying squirrel’’)

(Collins, 1798) {eora}

‘‘wu::b:i:n’’ (‘‘possum’’)

(Dawes, 1790) {eora} �ógliding creature

andbu(lυ : i)ñŋñwurυ (la:gυn)� ‘‘bu::ŋ:wu:’’�

ógliding creature

‘‘ba::ng:o:’’ (‘‘squirrel’’)

‘‘bu:ng:ora:’’ (‘‘flying squirrel’’) (Nathan, 1849) {dariwυ l}

(McCaffrey, 1920s) {dariwυ l}

‘‘bu::ng:oo:’’ (‘‘squirrel’’)

‘‘ba::ŋ:gu:’’ (‘‘flying fox’’) (Larmer, 1853) {durga}

(Eades, 1976) {durga}

‘‘bu::ng:u:’’ (‘‘flying squirrel’’)

‘‘bu::ng:gu:’’ ‘‘flying squirrel’’) {dariwυ l} (Dawes, 1790) {eora}

‘‘ba::ng:u:’’ (‘‘flying squirrel’’) {durga}

(Mathews, 1903, 1901) ‘‘ba::ng:u:’’ (‘‘flying squirrel’’)

(Gunther, 1837) {wiradjuri}� �

Chapter 1, JAS 2003

Page 134

Lexigenesis in ancestral south-east Australian Aboriginal language 133Statistical quantities such as c are most meaningful for lexemes with many

morphemes, say TP6, indeed the more the better. These lexemes for ‘gliding-squirrel’, made from only five morphemes, can experience significant variations inc (from 1.6 to 2 in the above histograms), scattered either side of cZipfó1.7,depending merely upon the condensation of a single morpheme (in this casewurυ la). Exactly the same thing happens for bu:ŋ:wurυ (McCaffrey, 1920) (có1.6)and bu:ŋ:wu (Dawes, 1790a) (có2). Additionally g:wurυ :bulυ (Mann, 1840s)has a value có1.2, which is significantly less than the Zipf mean.

To avoid such erratic behaviours of lexemes made from small numbers ofmorphemes, we can average all the c values within a parent leximorph or, indeed,over the entire set of leximorphs. Between the extreme values 1.6 and 2, in theabove histograms, lies a statistically more meaningful mean có1.8. An instructiveexample like this, involving only five morphemes, raises awareness of the errorslikely to arise in this kind of analysis and how to minimize their impact.

4 Lexeme�word, ‘a phase transition’

Critical re-examination of previously given histograms for various lexemes, eachwith its own respective mean-condensation c[ cZipfó1.7, reveals that almost noneare well described by the Zipf distribution alone. They generally seem to drop toa minimum in the vicinity of Hó1 or 2, thereafter rising to a crest at Hó2 or 3or 4, eventually falling away toward zero as H�6. These distinctive undulations,which also occur in non-linguistic contexts ( Jones, 1999), are particularly pro-nounced in lexemes possessing large values of c such as

la:rυ :gυ (óoyster)

b: i:ŋ:wυ l (ófemale breasts)

(Quoy & Gaimard, 1826)

but are also apparent for smaller values of c as in the examples

dυ :bυ :wi�:bυ : i:g (ósun-rise)

(Pettit & Dawson, 1850s)

bu:d:ga:�:wi� (ósun-set)

(Everitt et al., 1900)

wu:b: i:ŋ (ógliding-possum)

(Dawes, 1790a)

Generally the Zipf distribution with mean-condensation cZipfó1.7 describes indi-vidual relatively independent morphemes, some slightly condensed, coexisting (upto a dozen at a time) in syntactic environments (called leximorphs) which mustoriginally have been ancient sentences or phrases. The observed values of camongst subsequent genetically-related lexical progeny quantify the extent to whichconstituent morphemes have become progressively and collectively condensed,through time, into increasingly bound clusters (called lexemes) which can eventuallyemerge as independent ‘words’ in their own right. In this way new morphemes canbe created to cluster with previously existing morphemes, within new hybrid lexemes,all the while constrained by a universal Zipfian law that perpetually transcendssyntactic scales and grammatical regimes across countless millennia.

The above-mentioned undulation in Hamming frequency histograms, belongingto real-world lexemes, arises because increasing condensation is achieved through

Chapter 1, JAS 2003

Page 135

134 C. R. Illert

less-frequently condensing morphemes slightly (creating a trough at about Hó1 or2), and more frequently condensing morphemes severely (creating a crest at aboutHó2 or 3 or 4). With increasing condensation, the collapse of a lexeme into aword in its own right (with constituent morphemes irretrievably bound together andno longer functioning individually) is in every sense a phase-transition analogous tothe physical condensation of gas to liquid or liquid to solid. It may be carrying theanalogy too far, thinking of morphemes as ‘elementary particles’ whose phonemesthat are lost through condensation equate to ‘lost mass’ manifesting itself asbinding-energy that holds the lexeme together, but some physical analogy is needed.

The celebrated Kinetic Theory of Ideal Gasses models phase transitions usingtwo terms: one due to Robert Boyle (1627–1691), which has volume varyinginversely with pressure (analogous to the Zipf law in linguistics), and another termdue to Johannes van der Waals (1837–1923) which describes the deviation fromideal behaviour during phase change (Starling & Woodall, 1950). In a similar spiritwe now model condensation of real world lexemes as follows:

F(H)ó iò AHòa

ò B(Hòb)2òp

George Kingsley Zipf Johannes van der Waals(1902–1950) (1837–1923)

for H é [0, 6] and cP cZipfó1.7

Where AP0 and, in order for F to collapse to the Zipf limiting case, we typicallyhave ióñT.(cñ cZipf)/6a, aó(T. cñA)/(T. cZipfñA), bóñ3. c. cZipf /(6ñ ln(7))and póA. cZipf /6. To determine the main constants, A and B, we return to thedefinition of the Hamming mean:

Tî có;6

Hó0HîF(H)

ó;6

Hó0Hî�iò A

Hòaò B

(Hòb)2òp�—�xóHò1�

7

1 �i(xñ1)òA(xñ1)xñ1òa

ò B(xñ1)(xñ1òb)2òp�dx

ó18iòA�6ña ln�6òaa ��

òB�12

ln�(6òb)2òpb2òp �ñ b

�p �atn�6òb�p �ñatn� b�p���

Rearranging for B in terms of A gives

Bó

Tcñ18iñA�6ña ln�6òaa ��

12

ln�(6òb)2òpb2òp �ñ b

�p�atn�6òb�p �ñatn� b�p��

Chapter 1, JAS 2003

Page 136

Lexigenesis in ancestral south-east Australian Aboriginal language 135

In the Zipf limit the two sides of this equation simplify as follows:

limitc� c Zipf

[B]ó0

� limitc� c Zipf

[A]ó limitc� c Zipfa�1i�0 �

Tcñ18i

6ña ln�6òa)a ��

ó TcZipf

6ñ ln(7)ó 1.7

4.05409TB1

2T• f0

which is consistent with our earlier definition of the Zipf curve.At the other extreme, as H é [0, 6], we might imagine c attaining a maximum

value of 6 (if all morphemes were six phonemes long and all were totally condensed).In such a case cñ cZipfó6ñ1.7ó4.3, hence the first-order approximation

AB TcZipf

6ñ ln(7) �1ñ 14.3

(cñ cZipf)� for c é [cZipf , 6]

can be used to numerically calculate a, i, p and B, from previously given equations,in terms of the known quantities T, c and cZipf . Thus, by varying c we are able toplot progressive stages of lexeme-condensation (Fig. 2) following the genesis ofwords through time.

F. 2. As c increases, from the Zipf value 1.7, the above numerically-generated family of optimalfrequency-distributions describe diachronic-condensation of eight typical morphemes (Tó8) within anincreasingly-bound lexime. This model applies for c values up to about 3.4, beyond which the (dashed)curves drop below the H-axis implying unphysical negative frequencies. This is discussed futher in

Example 6.

Chapter 1, JAS 2003

Page 137

136 C. R. Illert

This numerical model predicts an upper limit to ‘allowable’ values of c in real-world lexemes, approximately when có2cZipf , beyond which the trough of thefrequency distribution drops below the H-axis, implying unphysical negative fre-quencies. Unfortunately, due to extinction of Aboriginal language throughoutsouth-eastern Australia, it is no longer possible to check this prediction throughdiachronic studies involving living language but, as arithmetical lexemes may (inprinciple) be indefinitely enlarged (Mathew, 1899), we do have another avenue forstudying real-world lexemes condensing toward this predicted limit.

Example 6Numerals

0ómirυ1ó(. . .)ñmi(rυ)ñ dυ (la)ñ n �

ò zero unity thing[inanimate]

zeroòunityó1

‘‘:mi:tu:n’’‘‘:me:ta:nn’’ (‘‘one, today’’) (Mackenzie, 1872) {dυrga}

(Quoy & Gaimard, 1826) {dυrga}

‘‘:mi:tto:ng’’‘‘:mi:dto:ng’’ (Bennett, 1832) {ŋunuwυ l}

(Munima, 1863) {dariwυ l} (Curr, 1887) {ŋunuwυ l}

‘‘:me:d @du:ng’’ ‘‘:mi:ttu:ng’’(Everitt, 1900) {gυndυŋara} ‘‘::dulla:’’

(Malone, 1875) {dariwυ l}

‘‘:mi:da:ng’’(Eades, 1976) {dυrga} ‘‘::zahle:n’’ {german}

‘‘::tally:’’ {english}

� �and

mirυñ nñ dυ la��‘‘mira:n:yalla’’(Lhotski, 1834) {ŋunuwυ l}

‘‘mitu:n:dthali’’(Mackenzie, 1874) {dυrga}�

Chapter 1, JAS 2003

Page 138

Lexigenesis in ancestral south-east Australian Aboriginal language 137

2óò[1, 1]óbuñ (mirυ :dυ)lañ (mi)rυ :(dυ)lañ (n)

ò 1 1

1ò1ó2

‘‘bu::lla:::’’(Bennett, 1832) {ŋunuwυ l} ‘‘bu::lla:lla:’’(Hunter, 1795) {eora} (Larmer, 1853) {dυrga}

‘‘boo::la:::’’ ‘‘bu::lla ::la:’’(Dawes, 1790) {eora} (Everitt, 1900) {gυndυŋara}

‘‘bu::lar:::’’ ‘‘bu::lao::la:’’(Malone, 1875) {dariwυ l} (Mackenzie, 1872) {dυrga}

�‘‘pou::lar:::’’ ‘‘bu::lla:rrah::’’(Quoy & Gaimard, 1826) {dυrga} (Munima, 1863) {dariwυ l}

‘‘pu::la:::’’ {sanskrit} ‘‘bu::laa:ra::’’‘‘pu:::ru::’’ {sanskrit} (Threlkeld, 1834) {awabakal}

‘‘:ari:th:me:tic:’’ {english}

‘‘:ao_:h:k`:|:’’ {greek} ‘‘baa:::::’’ (‘‘and’’)

‘‘:ke:qog:ka::’’ {greek} (Dawson, 1881) {chaap whurrong}

‘‘nu:mero::::’’ {spanish}

� �3óò[2, 1]ówuñ (bu:mirυ :dυ la:mirυ :dυ lañmi)rυ :(dυ)lañ n

ò 2 1

2ò1ó3

��‘‘we::::::ru:lla:’’

‘‘wo::::::wu:lli:’’(Mackenzie, 1874) {dυrga}

(Malone, 1875) {dariwυ l}‘‘wa::::::ra::n’’(Threlkeld, 1834) {awabakal}

�

4óò[3, 1]ónuñ (wu:bu:mirυ :dυ la:mirυ :dυ la:mirυ :dυ lañmi)rυ :(dυ)lañ n

ò 3 1

3ò1ó4

Chapter 1, JAS 2003

Page 139

138 C. R. Illert

��‘‘ne:::::::::ra::ng’’(Bennett, 1832) {ŋunuwυ l}

‘‘na::::::::::lo:’’ (Larmer, 1853) {dυrga}

(Hunter, 1795) {eora}

‘‘nu:::::::::rro::’’(Oldfield, 1828) {eora}

�

5óò[4, 1]ómuñ (nu:wu:bu:mirυ :dυ la . . . mirυ :dυ lañmi)rυ :dυ lañ n

ò 4 1

4ò1ó5

��‘‘ma::::::::::::rry:diolo:’’ (‘‘five’’)

‘‘mou::::::::::::ree::’’ (‘‘five’’)(Hunter, 1795) {eora}

(Munima, 1863) {dariwυ l}‘‘ma::::::::::::rro:la’’ (‘‘hand, finger’’)’(Russell, 1914) {gunduŋara}

�

Chapter 1, JAS 2003

Page 140

Lexigenesis in ancestral south-east Australian Aboriginal language 139

Thus we have a quinary system, involving recursive arithmetical additions of theform ò[a, b]óaòb, as in

ò [ò [ò [ò [1, 1], 1], 1], 1]ó1ò1ò1ò1ò1ó5

mu nu wu bu

and generally

Addition leximes(numerals) T c

2óbularυ 6 13/6ó2.1666663ówurυ la 9 23/9ó2.5555554ónurυ la 12 33/12ó2.755ómurυ la 15 43/15ó2.8666666 18 53/18ó2.9444447 21 63/21ó38 24 73/24ó3.0416669 27 83/27ó3.074074

10 30 93/30ó3.1. . .. . .. . .

n 3n (3ò10(nñ1))/3n

where, as these arithmetical lexemes increase in size, their mean condensation cconverges to the value

limitn�ê �3ò10(nñ1)

3n � ó 103B 2cZipf

in accordance with the prediction of the numerical model.

F. 3. As morphemes are added (i.e. as T increases) the troughs can drop no lower, so increasingcondensation c manifests as ever higher crests in the Hamming frequency-distributions of successive

numerals.

Chapter 1, JAS 2003

Page 141

140 C. R. Illert

These distinctive numerals occurred in arithmetical expressions such as

3óò[1, 2]ó(. . .)ñ�mi(rυ):dυ la)dυ la:(mi)rυ �ñbu: la:(rυ)ñn

ò 21

1ò2ó3

��‘‘:dya:rra::la::n’’(Mathews, 1881) {dharruk}

‘‘:mi:ttala::lee::’’

‘‘:kala:r:ba:::’’(Larmer, 1853) {dυrga}

(Quoy & Gaimard, 1826) {dυrga} �and

3óò[2, 1]ó(. . .)ñbu: la:rυñmi(rυ):dυ (la)ñn

ò 2 1

2ò1ó3

��‘‘:bu:lla:ra:mi:dto:ng’’ ‘‘:bu:lla::ma:tu:ng’’(Munima, 1863) {dariwυ l} (Larmer, 1853) {dυrga}� ,

also

4óò[3, 1]óbuñwu:rυ : lañm(irυ :dυ lañn)�‘‘wo:wo:r:lyu:m::’’

(Munima, 1863){dariwυ l}ò 3 1

3ò1ó4and

4óò[2, 2]ó(. . .)ñbu: la:rυñbu: la:rυñ(n)

ò 2 2

2ò2ó4

��‘‘:b:la:oeri:b:la:oeri:’’ ‘‘:bu:lla:r:bu:lla:r:’’(Rowley, 1875) {darug} (Malone, 1875) {dariwυ l}� ,

also

5óò[4, 1]ónu:rυ : lañnñbuñ(mi)rυ :dυ (la)�‘‘u::llu:ng:b:ro:tha’’

(Larmer, 1853){dυrga}4 ò 1

4ò1ó5and

5óò[ò2, 2], 1]óò[4,1]ó(. . .)ñbu: la:r(υ)ñbu: la:r(υ)ñbuñmi(rυ):dυ (la)ñn .

ò 2 2 ò 1

2ò2ó4

4ò1ó5

�‘‘:bu:lla:r:bu:lla:r:bo:mi:ttu:ng’’

(Malone, 1875) {dariwυ l}

Chapter 1, JAS 2003

Page 142

Lexigenesis in ancestral south-east Australian Aboriginal language 141

5 The optimization problem

For the Lagrangian

L(H, F(H))ó12

g(H) F2(H)

where

g(H)ó ñ(Hòa)2((Hòb)2òp)2

A((Hòb)2òp)2ò2B(Hòb)(Hòa)2

a necessary condition for the functional

Jó;6

hó0L(H, F(H))

to be minimized is that F(H) must satisfy the Euler Equation

0ó ddH �LLLF�ó d

dH(gF)ó gFògF

hence

F(H)óiò AHòa

ò B(Hòb)2òp

for H é [0, 6]

The limit as B�0 makes ió0 and g(H)óñ(Hòa)2/A, reducing F to the Zipffunction A /(Hòa), in which event Aó1 and xóHòa gives góx2 and Lóñ1

2 x2F2

corresponding to the power-law xñ1 as discussed in the abstract of this paper.Conversely in the limit as A�0, g(H)óñ((Hòb)2òp)2/(2B(Hòb)) reducingF to the van der Waals function iòB/((Hòb)2òp). On a philosophical level, theLagrangian is a pure ‘kinetic-energy’ term describing lexeme motion/evolution inHamming space. This optimization problem shows that our proposed van derWaals perturbation, to the Zipf distribution, is the most efficient (optimal) way fordiachronic lexeme-condensation (ónew morpheme creation) events to occur.

Acknowledgements

This paper is in support of, and in part fulfilment of the requirements of a PhDby the author at the University of Western Sydney. The author is indebted to ananonymous referee for helpful comments; to Associate Professor Campbell andAssociate Professor Derek Abbott for valuable discussions; to Daniela Reverberi,and to Michael Organ the Chief Archivist at Wollongong University’s Michael BirtLibrary, for helping to locate rare literary and archival materials; also to AndrewAllison and Peter Cook of the Department of Electrical & Electronic Engineeringat Adelaide University for valuable discussions about statistics and other assistance.

NOTE

1. The veteran journalist John Gale (1913) recorded ‘a woman’s breasts’ó‘goyan:g:bera’, derivingfrom gυn:g(υn)-barυóextremely either-side. He also stated that the place-name Canberra meant ‘awoman’s breasts . . . deriving from the two eminences now known as Black Mountain and Mount Ainslie. . .’. And, indeed, the above permutations of the word ‘one’s breast[s]’ do quite match the earliestwritten versions of this place-name, even quite unfamiliar versions such as ‘pial:i:go’.

Chapter 1, JAS 2003

Page 143

142 C. R. Illert

REFERENCES

A, A. et al. (2001) State-space visualisation & fractal properties of Parrondo’s games, PreprintSeries, Department of Electrical & Electronic Engineering, Adelaide University, Australia.

C, A. (1970) Aboriginal languages in the south-central-coast of New South Wales, fresh discover-ies, Oceania, 41, pp. 20–27.

C, D. (1798) An account of the English colony in New South Wales, two volumes (London, UK, T.Cadell & W. Davies).

C, T. (1992) An Introduction to Historical Linguistics (Oxford University Press) See Chapter 5‘the comparative method’, and Chapter 8 ‘subgrouping’.

C, T. (1997) Chipping away at the past. In: P. MC & . (eds) Archaeology andLinguistics, pp. 275–295 (Oxford University Press).

C, E. M. (1886) The Australian Race, four volumes (Melbourne, Australia, Government Printer).D, W. (1790a) Grammatical Forms of the Language of New South Wales, in the Neighbourhood of

Sydney . . ., manuscript, Marsden Collection, School of Oriental and African Studies, London.D, W. (1790b) Vocabulary of the Language of New South Wales, in the Neighbourhood of Sydney

(Native and English) . . ., manuscript, Marsden Collection, School of Oriental and African Studies,London.

D, J. (1881) Australian Aborigines, the Languages and Customs of Several Tribes . . . in the WesternDistricts of Victoria . . . (Melbourne, Australia, George Robertson).

D, R. M. W. (1980) The Languages of Australia (Cambridge, UK, Cambridge University Press).E, D. K. (1976) The Dharawal and Dhurga Languages of the New South Wales South Coast (Canberra,

Australia, Australian Institute of Aboriginal Studies).E, M. M. et al. (1900) The organisation, language and initiation ceremonies of the Aborigines of

the south-east-coast of New South Wales, Journal & Proceedings of the Royal Society of N. S. W., 34,pp. 262–281.

F, J. (ed.) (1892) An Australian Language . . . (Sydney, Australia).G, J. (1830s) Grammar and vocabulary of the Wiradhari dialect in New South Wales. In:

J. F (1892) An Australian Language . . . (Sydney, Australia) Appendix D, p 56–120.H, H. (1846) The languages of Australia, pp. 479–531 of Ethnology and Philology, volume VI of

the Reports of the United States Exploring Expedition, under the command of Charles Wilkes (Philadel-phia, USA).

H, D. V. & B, P. (1981) Elements of Statistical Inference, 5th edn (Boston, USA,Allyn & Bacon).

I, C. R. (1987) Formulation and solution of the classical seashell problem, Il Nuovo Cimento,9(D), pp. 791–814.

I, C. R. (1999) Maria’s lullaby. In: K. K, There is no mystery, an artistic response to LakeGeorge, pp. 47, 48, 172 (Canberra, Ginninderra Press).

I, C. R. (2001) The centenary of Mary Everitt’s gundungara grammar, Journal & Proceedings ofthe Royal Society of N. S. W., 134(1–2), pp. 19–44.

J, B. K. (1999) The scale invariance of 1/f noise. In: D. A & L. B. K (eds) Unsolved Problemsof Noise Fluctuations, UPON’99: Second International Conference, Adelaide, Australia, 12–15 July 1999,American Institute of Physics Conference Proceedings, 511, pp. 115–123.

L, J. (1898a) Native vocabulary of miscellaneous New South Wales objects, Journal of the RoyalSociety of N. S. W., 32, pp. 223–229.

L, J. (1898b) Aboriginal words and meanings (Bateman’s Bay), Science of Man, 2(8), pp. 146–148.M, S. & H, T. (1980) Scale invariance implies 1/f spectrum, Second International

Symposium on 1/f noise, Orlando, USA.M, J. & M, L. (1875) Kamilaroi and other Australian Languages In: W. R (ed) (Sydney,

Australia, Government Printer).M, J. F. (1840s) Australian Aborigines, a few notes on their language, the language of Long Dick

. . . of the Cammeray Tribe .. . . Manuscript, Mitchell Library, Sydney, Australia.M, J. (1899a) Aboriginal arithmetic, Alma Mater (Melbourne) 4(4), pp. 15–16.M, J. (1899b) Eaglehawk and Crow (London, David Nutt).M, W., S, R. L. & W, D. D. (1986) Mathematical Statistics with Applica-

tions, 3rd edn (Boston, USA, Duxbury Press).M, T. (1878) Tangambalanga—Pallanganmiddah tribe. In: R. B. S (1878) The Aborigines

of Victoria, vol 2, pp. 67, 174–175 (Melbourne, Australia, Government Printer).M, T. (1878) Upper Murray vocabulary. In: E. M. C (1886) The Australian Race, vol 3

pp. 562–563 (Melbourne, Australia, Government Printer).

Chapter 1, JAS 2003

Page 144

Lexigenesis in ancestral south-east Australian Aboriginal language 143

M, M. (1863) Aboriginal words and their meanings. In: G. B (ed) (1899) Science of Man,2(8), pp. 141–142.

N, I. (1848) The Southern Euphrosyne and Australian miscellany . . . (London. Whittaker &Co), see word-list on pp. 134–135.

O, M. (1993) Illawarra and South Coast Aborigines 1770–1900, 2 vols (Canberra, Australia,A. I. A.T. S. I.S. ).

P, N. D. & D, W. (1850s) In: A. W. H manuscripts MS9356, Latrobe Collection,State Library of Victoria, Melbourne, Australia.

Q, J. R. C. & G, J. P. (1826) In: M. J. D D’U (ed) (1833) Voyage de decouvertesde l’Astrolabe 1826–9 (Paris, France, Ministere de la Marine), see the ‘Philologie’ volume, pp. 1–30.

R, D. (1985) Clockspring model of sprouting & blossoming, Z. angew. Math. und Phys. 36(5),pp. 743–756.

R, W. (1875) Kamilaroi and Other Australian Languages (Sydney, Australia, Government Printer).R, W. (1878) Journal of the Anthropological Institute of Great Britain & Ireland 8(3), pp. 263–265.R, J. (1875) In: W. R, Kamilaroi and Other Australian Languages, pp. 99–115 (Sydney,

Australia, Government Printer).R, W. (1914) My recollections (NSW, Australia, Camden).S, W. (1919) Die Gliederung der Australischen Sprachen . . . (Vienna, Mechitharisten-Buch).S, W. (1919) Die personal pronomina in den Australischen sprachen (Vienna Hoelder).S, C. E. & W, W. (1949) The Mathematical Theory of Communication (USA, University of

Illinois Press).S, R. B. (1878) The Aborigines of Victoria, 2 vols (Melbourne, Australia, Government Printer).S, H. E. et al. (1999) Scaling features of noncoding DNA, Physica-A, 273(1–2), 1–18.S, S. G. & W, A. J. (1950) Physics, pp. 269–283 (Longmans-Green).T, . . & , C. W. (1840) Outlines of a Grammar, Vocabulary and Phraseology

. . . (Adelaide, South Australia).T, W. (1798) Sydney’s first four years . . ., Library of Australian History, Sydney, Australia.T, N. & MG, W. (eds) (1994) The Macquarie [Dictionary of ] Aboriginal words (Sydney,

Macquarie University).T, L. E. (1834) An Australian grammar . . . (Sydney, Stephen Stokes & Co).T, G. & G, P. (1998) Zipf ’s Law is not a consequence of the central limit theorem, Physical

Review E, 57(2), 1347–1355.W, M. B. (1988) 1/f noise and other slow, nonexponential kinetics in condensed matter, Review

of Modern Physics, 60, 537–571.W, B. J. & S, M. F. (1989) On the ubiquity of 1/f noise, International Journal of Modern

Physics B, 3, pp. 795–819.Y, D. (1982) Australian Aboriginal Languages (Andre Deutsche) see chapter 3 on ‘pronunciation’.Z, G. K. (1949) Human Behaviour and the Principle of Least Effort (New York & London, Hafner).

Chapter 1, JAS 2003

Page 145

 

Chapter 1, JAS 2003

Page 146

Journal of Applied Statistics,Vol. 31, No. 1, 73–104, January 2004

Phono-genesis and the Origin ofAccusative Syntax in Proto-AustralianLanguage

C. ILLERT* & A. ALLISON***University of Western Sydney, New South Wales, Australia, **Adelaide University, SouthAustralia, Australia

(Received August 2002, revised November 2003, accepted July, 2003)

A It is claimed that a set of 62 known (Illert, 2003) ancient Aboriginal wordsconstitute a representative sample of the original proto-Australian lexicon whose maximumlikelihood (Fisher, 1912) ‘power law signature’ is determined and shown to precisely fit geneticallyrelated ‘modern’ lexicons from south-eastern-Australia. This measure of ‘sameness’ builds theconfidence required to justify inter-lexicon diachronic word-frequency comparisons which providea powerful new statistical tool capable of revealing important features of ancestral grammar.This paper supplies the first ever published modern translations of authentic traditional languagedocumented in obscure literary and archival sources which have, until recently, been lost (Dawes,1790b; Wood, 1924; Troy, 1992) or overlooked (Everitt et al., 1900; Illert, 2001) for centuries.These newly found examples of accusative syntax supported by word-frequency data may comeas quite a surprise to some linguists (Dixon, 1980; Osmond, 1989; Troy, 1992; Nichols, 1993)who, in the absence of adequate evidence, seem to have long-imagined that language from thisregion—if not the entire continent—simply had to be inherently and at the core ergative. On thecontrary we find that changing word-frequencies, from proto-Australian to modern times, supplyoverwhelming evidence of the emergence of ancient accusative prefixes which have even survivedinto recent centuries in the Sydney region. Additionally it is found that, over millennia, wordsdie-off in a lexicon, replaced by others, according to the famous ‘mortality law’ of Gompertz(1825) which also describes the likelihood of death of biological organisms within populationsand is the basis for modern actuarial science (Bowers et al., 1997). Just as disease and epidemicscan wipe out entire cohorts of creatures from a population, so too can syntactic change annihilateword-classes in an evolving lexicon.

K W: Maximum-likelihood, frequency analysis, Gompertz law, morpho-statistics,linguistics

Introduction

The Aboriginal ‘eora’ dialect from Sydney is perhaps the oldest documentedremnant language from anywhere in mainland Australia. Two centuries ago—along with ‘thurrawal’ supposedly from the Illawarra, and ‘dhur�a’ from theShoalhaven—it was part of a strip of similar language extending for hundreds

Correspondence Address: C. Illert, 2/3 Birch Crescent, East Corrimal, N.S.W. 2518, Australia.

0266-4763 Print/ 1360-0532Online/04/010073-32 © 2004 Taylor & Francis LtdDOI: 10.1080/0266476032000148966

Chapter 2, JAS 2004

Page 147

74 C. Illert & A. Allison

of kilometres along the New South Wales coastline. Adjoining them inland, tothe west, were the Blue Mountains ‘dharruk’ and ‘�undun�urra’ dialects, whichwere different from coastal counterparts. However, regardless of any differences,traditional language has long been lost throughout the entire region and it isonly by considering this entire block of dialects collectively that there is sufficientliterary and archival material to salvage and reconstruct a meaningful quantityof words and grammar.In a previous paper, Illert (2003) discussed the orthographic decrepitude of

some relevant 18th and early 19th century ‘word-lists’. Out of this old-fashionedliterary tradition came a rather jumbled paper, by Robert Hamilton Mathews(1903, pp. 275–279) containing a 428-word ‘thurrawal’ lexicon. And, morerecently, a phonetically lovely 321-word ‘dhur�a’ lexicon was recorded from ‘thelast living speakers’ by a modern linguist Diana Kelloway Eades (1976). Thesetwo lexicons jointly provide a net vocabulary of about 750 Aboriginal wordsincluding verbs, adjectives and nouns, all from roughly the same geographicalregion.Of course, there are other relevant word-lists (e.g. Rowley, 1875; Malone, 1875;

Munima, 1863; Nathan, 1848; Quoy & Gaimard, 1826; Collins, 1798; Hunter,1793); however, in order to avoid differing orthographic notations and varyingstandards of reliability, this present study focuses upon the thurrawal word-listof Mathews (1903) and the dhurga word-list of Eades (1976). We will rigorouslycompare these two ‘modern’ Aboriginal lexicons with a proposed 62-word proto-Australian lexicon most of which has already been previously published by Illert(2003, Tables 1(a) and 1(b)).The genetic relatedness of all three languages is self-evident, both of the

modern lexicons being awash with common roots, Mathews’ (1903) thurrawalword-list alone attesting to more than half of the 62 known ancestral words (seeAppendix A). In addition Eades’ (1976) dhur�a lexicon provides evidence forroots such as the conjugate pairs �aru�‘�aru’ (‘nasal mucous’) and �ura�‘bura’(‘stone, rock’), buru�‘buru’ (‘kangaroo’) and bara�‘bara:l ’ (‘wallaby’), also‘mula’ (‘head lice’), wara�‘wara:ŋa�’ (‘boomerang’), wυri�‘wari:�’ (‘child’)and wurυla�‘wula:n’ (‘money’).At present, there is good evidence for the existence of 14 proto-Australian

words commencing with the consonant m, 11 with �, ten with b, six with �, sixwith w, five with ŋ, three with dɔ, three with n, three with d� and one with n� . Allthese words, ranked according to respective word-initial consonant frequencies,can be displayed as the histogram in Figure 1—for which even visual inspectionsuggests some kind of power-law arising from entropy.It is possible likewise to count manually the frequencies of word-initial

consonants in both the thurrawal and dhur�a lexicons, and to plot similarhistograms for them too. The data for all three lexicons, presented in Table 1,seem fairly much in keeping with previous work in the field. For example, Dixon(1980, Section 6.6.2, p. 167) examined the phonotactics of 40 recent Australianlanguages, ‘virtually all those for which adequately detailed phonological descrip-tions have been provided’, finding thatthe most frequent word-initial segments are, in every language, the peripheral stops andnasals . . . [whilst] apicals are always the rarest stops and nasals in that position . . .,

Chapter 2, JAS 2004

Page 148

Proto-Australian Language 75

Figure 1. The 62 known words of a proposed proto-Australian lexicon, ranked according totheir respective word-initial consonants. These ancient roots were the building-blocks from

which ‘modern’ south-east-Australian vocabulary was constructed—see Appendix A.

. . . in war�amay for example25% of the entries in the dictionary begin with � and 12% with m, . . . [whilst] only 2% ofthe roots begin with d, and a minuscule 0.5% with n . . .,

. . . in every language

semi-vowels may occur initially . . . peripheral w begins rather more words than does laminaly, although both are common,

. . . also

[word-initial] laminal stops and nasals occur quite frequently . . . [with] the laminal nasal ny. . . well attested initially but . . . not so common finally . . . It could be that pA had a singlephoneme, realised as [ny ] initially and as [y] finally.

Word-initial-consonant frequencies in Table 1 reflect these basic phonotacticfeatures, and analysis of variance shows all three data sets to be indistinguishableto the 0.01 level of statistical confidence regardless of whether we study the rawor normalized values. Such basic measures of ‘sameness’ are important if theproposed set of 62 proto-Australian words are collectively to be viewed as arepresentative sample of (albeit ancestral) south-east-Australian vocabulary.Additionally, we will study the ‘power law’ signature1 of these lexicons to buildthe confidence that is required to justify inter-lexicon comparisons of wordfrequencies, as in Table 2, providing a powerful new statistical tool capable ofrevealing important information about ancestral grammars which differed in theway they marked the subject and object in a sentence—ranging between ergativesyntax which generally employed a suffix to mark a noun as the agent of atransitive process, and accusative syntax which could manifest as an affix marking

Chapter 2, JAS 2004

Page 149

16 , C. Illert & A. Allison

Table 1. Lexical-frequencies of word-initial-consonants. N is the total number of words withany specified word-initial-consonant, in each ol three different languages. The correspondinglexical frequencies in each languagg respectively equal to 100 x Nl62 {proto-Australian},100 x N/321 ldhurga\ and 100 x N1428 {thurrawal}, are expressed as approximate percent-

ages with a rounding error of +0.07.k

proto-Australian(Illert, 2003)

total of 62 words

dhurga(Eades, 1976)

total of 321 words

thurrawal(Mathews, 1903)

total of428 words

word-initialconsonant f(:%)

word-initial

consonant

word-initial

F(%) consonant F(%)

m

sb

11

w44n

dD

dvrI

t411

l06

6

5

J

)J

1

0

0

0

0

22.6r 7.816.1

9.7

9.78.1

4.8

4.8

4.81.6

00

0

0

s 85 26.5

b 62 19.3

m45144 36 1t.2w 26 8.1

d 19 s.9

0 14 4.4n 72 3.8

d 11 3.4y82.5t1 3 0.9

fr00r00100

b

sm

4w0vdndJl!fI

93 21.7

79 18.5

64 15

46 r0.73',7 8.6

35 8.225 5.8

21 4.9

t2 2.8

6 1.4

6 1.4

410000

Table 2. Changes in lexical frequencies olword-initial-consonants

Lrt:F*-f.: I"rf .dt ffi ...forfa,(t)eCt[-7,0J

where l:0 corresponds to present-time and ]n is the age of the ancestrai proto-Australianlexicon

Diachronic change proto-Australian -dhurga proto-Australian + thurrawal

Lf,LfnAfo

LftLf,LfqLJ',,

Lf,LJA

Lf'

^faLf.

14 -22.6:8.6o/o26.5 -11 .8:8.7%

4.4-8.1: -3.7%19.3 - 16.1 :3.2',%

0.9-9.7:-8.8%11.2-4.8:6.4%

2.5-0:2.5%3.8-4.8 - -1%

5.9 -0:5.9%0- 1.6: -1.6%

3.4-4.8: -1.4n/o8.1-9.7: -1.6%

t5 -22.6: -7 .6%

18.5 - 17.8:0.7%8.2-8.1 -0.1'%

21.7 -16.1:5.6oh1.4-9.7: -8.3%

r0.7-4.8 -s.9%5.8 - 0:5.8%

2.8*4.8: -2%1.4-0:1.4o41-1.6: -1.6%

4.9-4.8:0.1%8.6-9.7: -1.1%

Chapter 2, JAS 2004

Page 150

Proto-Australian Language 77

pronouns (and sometimes nouns as well) as the goal of a transitive process (seeDixon, 1980, pp. 285–291; Yallop, 1982, pp. 72–142; Silverstein, 1976).

The Power Law for Lexicons

At any given time, the contribution to a lexicon’s entropy, due to the relativefrequencies of H different word-initial consonants, is defined as

Só ;H�1

h�0

fhln ( f

h) (1)

subject to normalization

;H�1

h�0

fhó1ó100% (2)

and Hamming’s constraint on the mean

;H�1

h�0

hfhók (3)

which is essentially a statement of the constancy of south-east-Australian phono-logy, over the millennia, since ancient times. For entropy to be maximized,subject to these constraints, the entire functional

Jó� ;H�1

h�0

fhln ( f

h)�òj1�1ñ ;

H�1

h�0

fh�òj2�kñ ;

H�1

h�0

hfh�

must satisfy

0óLJLf

h

ó ;H�1

h�0

[ln ( fh)ò1ñj

1ñj

2h]

hence the exponential frequency distribution

fhóe�1�1e�2hóf

0ah (4)

for hó0, 1, . . .,Hñ1, where the two Lagrange multipliers are defined asj1ó1òln ( f

0) and j

2óln (a)\0.

If we select

f0ó1ña1ñaH

� fhó�

1ña1ñaH� ah (5)

for hó0, 1, 2, . . . ,Hñ1 then the normalization constraint (2) holds trivially, forany value of a, because it becomes the sum of a geometric progression to Hterms

Chapter 2, JAS 2004

Page 151

78 C. Illert & A. Allison

;H�1

h�0

fhó�

1ña1ñaH� [1òaòa2ò. . .òaH�1]ó1ó100%

The precise value of a, that best fits the proto-Australian data, can be calculatedas follows using Fisher’s (1912) ‘Method of maximum likelihood’. The ‘actuallyobserved’ word-frequencies, f

hin Table 1, were calculated from discrete countable

quantities Nhthat cannot be negative, so it makes sense to assume Poisson (not

Gaussian) error—in which event, the probability of obtaining these ‘actuallyobserved’ word-frequencies individually is

ph•f

hó(u

h)Nh

(Nh)!e��h (6)

the distribution ascribed to Simeon Denis Poisson (1781–1840) where the respec-tive means u

h, one for each data point, are required to exactly follow the ideal

‘power law’ (equation (5))

uhó62100fhó62100 �

1ña1ñaH� ah (7)

for hó0, 1, 2, . . .,Hñ1. Thus, the ‘likelihood’ of obtaining the entire set of 14‘actually observed’ proto-Australian word-frequencies collectively is

ló<H�1

h�0

phó<

H�1

h�0

�62100fh�

Nh

(Nh)! e��62�100 fh (8)

where Hó14.Taking the logarithm of both sides gives

ln (l)ó ;H�1

h�0 �Nhln ( f

h)òN

hln�

62100�ñ

62100fhñln (N

h!)�

where Hó14 and, for ‘maximum likelihood’, we require a such that

0óLln (l)Laó ;

H�1

h�0 ��N

hfh

ñ62100�

LfhLa� (9)

where Hó14. There are three ways that this could occur.

(1) we could have Nhó62f

h/100 corresponding to the case where all the observed

data-points lie precisely along the ideal curve (equation (5)) with no scatter(as is almost the case with Eades’, 1976, dhur�a data in Figure 2);

(2) or it might happen that

0óLf

hLa

Chapter 2, JAS 2004

Page 152

Proto-Australian Language 79

Figure 2. The ‘maximum likelihood’ curve fh generated from proto-Australian data, assumingthat Hó14, and plotted along with all three data sets from Table 1. This curve cuts thevertical axis at f

0ó25.49%, above the proto-Australian data point fmó22.6% but below the

dhur�a data point F�ó26.5%. Also, the tail-end of the curve, f13ó0.6%, is close to the zero-

frequency of word-initial-l in all three languages given in Table 1. Note how this ‘maximumlikelihood’ curve—generated from proto-Australian data scattered above and below it—isnevertheless closely hugged along its entire length by data points corresponding to the

‘modern’ dhur�a lexicon of Eades (1976)

as with the zero-slope associated with the string of zero-frequencies fhin the

tail of the distribution for all integers h[Hñ1 (as in Figure 3).(3) or it may happen that only the sum in its entirety equals zero—the ‘Solution

of Maximum Calculational Difficulty’—in which event we proceed bydefining

LYLa•L2 ln (l)La2

ó ;H�1

h�0 �ñN

hf 2h �Lf

hLa�

2ò�N

hfh

ñ62100�

L2fh

La2� (10)

where Hó14.

Less than five iterations of the Newton/Raphson formula

anewóa

oldñY(a

old)

Y(aold)

yields the value

aB0.7496ô0.0005 (11)

Chapter 2, JAS 2004

Page 153

80 C. Illert & A. Allison

Figure 3. The ‘maximum likelihood’ curve fh , generated assuming that Hó10, is a much betterfit to the ancient proto-Australian data from Table 1. This curve cuts the vertical axis atf0ó22.94%, close to the proto-Australian data point fmó22.6%, whilst the tail-end of thiscurve beyond the step (h[Hñ1ó9) corresponds precisely to the ‘observed’ proto-Australianzero-frequencies in Table 1. It would seem that ancestral language had a smaller alphabet

for the proto-Australian data in Table 1. Substituting this ‘maximum likelihood’value into equation (5) then gives

f0B0.2549 decreasing to f

13B0.0060 (12)

whilst, from equation (3), the mean is

kó ;H�1

h�0

hfhó13f

13ln aòf0ñf

13(ln a)2

B2.7254 (13)

The resulting ‘maximum likelihood’ curve (Figure 2) was calculated from datain Table 1 on the assumption that Hó14. Certainly ‘modern’ languages such asdhur�a and thurrawal were based upon a phonology of 14 consonants all ofwhich could appear word-initially, although word-initial-l and word-initial-r wereso rare that their ‘observed’ zero-frequencies in Table 1 are unsurprising. Oneexample of word-initial-l, la:rυ:�υ (‘oyster’), was recorded in dhur�a languageby the French explorers J. R. C. Quoy and J. P. Gaimard in 1826 when theirship, the Astrolabe, was anchored in Jervis Bay (Illert, 2003, example 4). Inaddition, important examples of word-initial-r are the condensed ‘numeral’ r:υla(Illert, 2003, example 6) and its fuller form r:d�υla, which was probably the originof the rd intervocalic retroflex stop in some ‘modern’ central Victorian languages(Illert, 2004). But proto-Australian was phonologically simpler than the two

Chapter 2, JAS 2004

Page 154

Proto-Australian Language 81

Figure 4. The changing maximum likelihood ‘power law’ for 9OHO14

modern languages in Table 1, possessing nowhere near 14 consonants. Dixon(1980, p. 177) has suggested that there were only 10 or 11 consonants in earlyproto-Australian, leading to 13 or 14 at a later stage. And we see from Figure 3that the ‘maximum likelihood’ curve generated from the value Hó10 doesprovide a better fit to the actual proto-Australian data alone.For each different value of H the ‘maximum likelihood’ calculation (equation

(11)) generates slightly different values of a, hence slightly different ‘power law’curves. Over the range of immediate interest Hó9�14, from proto-Australianto modern language, there is an approximate relationship

aB0.901ñ0.011îH (14)

Accordingly, Figure 4 shows how the ‘power law’ changed over millennia assouth-east-Australian phonology became more complex. This meaningful andlinguistically important progression is easily described by the model, and impor-tant insights into accompanying phonotactic changes can be gleaned by askingwhy the word-frequencies in Table 1 changed during this process as in Table 2.It will be argued that such facts are explained by the existence of accusative-prefixes which arose in ancient times and survived into modern language withinthe Sydney region. In support of this proposition, seven newly translatedsentences will be given as examples—four chosen from Mary Everitt’s (1900)�undun�urra grammar, which has been overlooked by scholars for 100 years(Illert, 2001), and three from fragments of the eora dialect recorded in the diariesof the First Fleet officer, Lieutenant William Dawes (1790a, b), which were lostfor 200 years and only recently found in a London library (Wood, 1924; Troy,

Chapter 2, JAS 2004

Page 155

82 C. Illert & A. Allison

1992). All seven examples are especially noteworthy being the first ever publishedmodern translations of material from either of these primary sources. Theobserved accusative syntax, supported by word-frequency data, may come as asurprise to some linguists who, in the absence of adequate evidence, seem tohave imagined that language from the Sydney region—if not the entire conti-nent—simply had to be inherently and at the core ergative (Eades, 1976; Dixon,1980; Osmond, 1989; Troy, 1992; Nichols, 1993).

The Ancient Verb-Phrase mi�

It is well known that there were ancient determiners with both indefinite andinterrogative significance:mi�ó�imówhich/what/who (singular), and�un/�uŋósomethings/someones (dual or plural). In this respect, Dixon (1980, pp. 372–377)noted

It appears that *’who’�‘what’ and also *’what’�‘who’ are common meaning shifts inAustralia . . . [and that the] indefinite-interrogative form . . . mia ‘what’ . . . is found inalmost every language in New South Wales and in some from Victoria, South Australia andQueensland (excepting the far north) . . . [though] not . . . in any western language,

and that

There are usually separate forms ‘who, someone’, with human reference, and ‘what, some-thing’, with non-human reference . . .,

but he goes too far asserting that

mia always means ‘what’, never ‘who’, and appears to inflect like a nominal; there is nohint of it being related to a monosyllabic root, at any stage of development,

which is contradicted by James Dawson’s (1881) chaap whuuron� word ‘winyaar’(‘who, whose, whom’). The problem here lies with Dixon’s methodology—sooften drawing inferences from the most transformed living languages located tothe far-west or north, instead of going directly to where the answers can mostreasonably be expected to exist, in the more ancient and least changed—thoughextinct—languages recorded along the south-east-Australian seaboard. No doubtthis is due to the relative ease of studying living as compared to extinct language.However, analysis of south-east-Australian data is still possible and quite capableof yielding meaningful information.We can assume that ancient adjectival roots did not usually contain the vowel i,

which was reserved as a tense-marker on verbs. Additionally mi� and �im seemto have been used interchangeably, suggesting an ancient verb-phrase that was aloose cluster of fundamental morphemes:

Chapter 2, JAS 2004

Page 156

Proto-Australian Language 83

m(υlυ):i ñ �un �show, indicate some[thing]

óWHICH or WHAT

‘m:i:nyaŋ’ (‘what’)‘::nyaŋ’ (‘what’)Sharpe (1994), bundjalun�

‘m:i:nya’ (‘what’), �undun�urra‘m:i:n�’ (‘what’), dharrukMathews (1901a, b)

‘m:i:nye’ (‘what’)(‘what (nom., acc.); sing., dual, pl.’)Meyer (1843)cooron�, lower murray

‘m:i:nyan�’ (‘what’)McNicol & Hosking (1994), wiradjuri

‘m:ı:n@ya’ (‘what’)có5/3ó1.6666Everitt et al. (1900), �undun�urra

‘m:i:n’ (‘which’)Govett (1836)Tarlo Dance Song, ŋunuwυl

‘m:i:’ or ‘m:i:ñm:i:’ (‘what’)Dawes (1790), eora

� �or

�un ñ i:mυ(lυ) �some[thing] show, indicate

óWHICH or WHAT

‘nan::ma’ (‘what’)Bulmer in Smyth (1876), kurnai

‘�aŋ::’ (‘what’)Ferguson in Holmer (1964), bundjalun�

‘�nun::aa’, peek whurrun�‘nun::aa’, kuurn kopan noot‘ne::aa’, chaap whurrun�Dawson (1881), (‘what’)

‘n:i:n’ (‘what’)Everitt (1900), �undun�urra

� �hence who(m)

ŋ:(d¯υla) ñ �

�(uŋ):i:m(υlυ)m(υlυ):i:�(uŋ)� �

person which/what

óWHO (singular)

‘un�::ny:i:m’ (‘who’)Mathews (1901b), thurrawal

‘::w:i:nyaar’, chaap whurron�‘�n::::aar(a)’, kuurn kopan nootDawson (1881) (‘who, whose, whom’)� �

Chapter 2, JAS 2004

Page 157

84 C. Illert & A. Allison

or

�uŋ:i:(mυlυ) ñ ŋ:d�υ(la) �which/what person

óWHO (singular)

‘n�an:i::n:de’ (‘who’), kurnai‘nyan::::da’ (‘who’), dharruk‘n�unn:::un�:’ (‘who’), tharrawal‘n�an::’ (‘who’), dharkinun�Mathews (1902, 1901a, b, 1903)

‘nan::’ (‘who’)Bulmer in Smyth (1876), kurnai

‘n�un@n:i::n�:a’ (‘who, interrogative’)Everitt (1900), �undun�urra

‘n�an::::dhi’ (‘who [is]’)‘n�an::::dhuu’ (‘who does something’)McNicol & Hosking (1994), wiradjuri

‘nan��:e:::’ (‘who’)Meyer (1843), cooron�, lower murray

� �and generally

�uŋ:i:m(υlυ) ñ �d�υ(la:ŋ)

bula(la:ŋ)w(ur)υl(a)ŋ�

which/what person[s]

person[s]óWHO (singular, dual, plural)

� �‘unna�:a:::’ (‘who?, singular’)

‘unna�:a::wula:’ (‘who?, two’)

‘unna�:a:m:ul:an’ (‘who?, several’)

Mathews (1901a), �undun�urra �Accusative Prefixes: mi�, �un/�u�, �un

In accusative languages the object of a sentence can have a prefix, such as mi�(singular, animate or inanimate) or �un (dual/plural, inanimate), to distinguishit from the subject—as in the following sentences

Chapter 2, JAS 2004

Page 158

Proto-Australian Language 85

(1) [the]-woman-requests some-tobacco

ówants, requests ówoman(present tense) (subject)

�in:�(υra):i ñ bula:n �u(n) ñ d�a(rυ):n ñ �a(ma):ŋ(ulu)

óSOME substance mouth–chewable

ótobacco (object)

� ‘yoon:�@:eeñbul@la:n n�ooñdhu:nñ�@u:n�’(ó ‘give to woman food’)

(Bessy Simms cited by Mary Everitt, 1900) {�υnd�υŋara}

(2) I-need (ógive-me) some-grease FOR-[the]-spear

�in:�(υra):i ñ ŋυ �u(n)ñ�υ(n):w(urυla):wu(rυla):ŋ(uru)

óneed, want, request óI óSOME(present tense) (subject) viscousñ[substance](ógrease/ fat/wax)2

(direct object)

óbarbed spear3 óFOR(indirect object) (dative)

�u(la):mυ(ra) ñ �i� �‘yoon:�:eeña��ee jañbo:w:wu:n �u:m@meeñ�in@nee’

� ó‘give thou me grease spearñfor;give me some grease to rub my spear with’�

(Bessy Simms cited by Mary Everitt, 1900) {�υnd�υŋara}

These examples illustrate the fact that living creatures are the natural perpetratorsof actions and that people speak largely about themselves, thus—unless it isspecifically indicated to the contrary—the subject of any sentence in an accusativelanguage is probably an un-affixed (‘free-form’) pronoun, which can usually beentirely omitted whilst still leaving the meaning of the sentence intuitively obvious(Silverstein, 1976). Thus, the two previous examples could be paraphrased as‘give some tobacco’ and ‘give some grease for the spear’. It is still obvious, fromthe human context, who the omitted recipient happens to be.Conversely, nouns are the natural objects of actions, making it also intuitively

obvious that any noun or proper noun occurring in an accusative sentence isprobably the object—even when not accusatively marked. Indeed, it is notuncommon for sentences lacking both a pronominal-subject, and also theaccusative prefix for their inanimate object, to still have a clear overall meaning.

Chapter 2, JAS 2004

Page 159

86 C. Illert & A. Allison

However, in the ‘noteworthy’ event that an accusative sentence contains apronominal-object, an accusative-affix becomes essential to distinguish it fromthe subject/agent. Thus, compact pronominal-objects are generally prefixed asfollows:

There can be no mistaking these prefixed-object[s] with unprefixed pronominal-subjects; not even in reflexive sentences, characterized by subjectóobject, whichomit pronominal-subject leaving their accusatively prefixed pronominal-objectserving a dual role

as in the sentences:

(3) [I (reflexively implied singular /)] am-beating myself (singular):

ŋu(ru):bul(υ):i ñ mi� ñ ŋυ

óamñbeating who(mñself ) I(pres. tense, verb)

ómyself

�compact, 1st pers. sing.pronominal OBJECT �

‘n�oo:bull:eeñmın@ñya’�

(‘I am beating myself ’)REFLEXIVE

(Everitt et al., 1900) {�υnd�υŋara}

Chapter 2, JAS 2004

Page 160

Proto-Australian Language 87

(4) [We (reflexively implied dual /)] are-beating ourselves (dual):

ŋu(ru):bul(υ):i ñ �uŋ ñ bul(ala):ŋ

óareñbeating someñ(selves) us/we (dual)(pres. tense, verb)

óourselves

�compact, 1st pers. sing.pronominal OBJECT �

‘n�oo@:bil@l:eeñyan�@ñeel:oon�’�

(‘We two beat ourselves’)REFLEXIVE

(Everitt et al., 1900) {�υnd�υŋara}

A number of syntactic circumstances (including a split subject) can prompt theexpansion of a compact-pronominal-object, necessitating a third accusativemarker ŋun, as follows:

Chapter 2, JAS 2004

Page 161

88 C. Illert & A. Allison

as in the sentence:

(5) [we] flanking pair [will-escort] Burung

óawayòaway (ó[we] pair of ) exterior(escort)s(split dual SUBJECT)

(. . .) ñ wu(ru):d�a(rυ) d�υ(la:ŋ):ŋun wu(ru):d�a(rυ)

óshall escort óher (singular)(fut. tense verb)

buru:ŋñ(mi)�� ‘ñyu:di dyi::ŋun yu:di buru:ŋña ’(‘we two are going to see booroon� part of the way home’)

óOBJECTñwho(m) (yemmerrawanne cited by Dawes, 1790) {eora}

or a dual/plural object

as in the sentence:

(6) We (dual) shall seek them (dual)—Ngulgear [and] Tugear.

mυ(lυ):i:bυ ñ ŋ:bula(la)

óshall/will see óuse/we(future tense verb)

�compact dualSUBJECT �

Chapter 2, JAS 2004

Page 162

Proto-Australian Language 89

óTHEM (right-expanded dual pronominal OBJECT)

ŋun ŋal�ear : (mi)� ñ bula(la):ŋñ tu�ear : (mi)�;

O1 O2

O1òO2óŋal�ear and tu�ear

‘na::bañ:bula ŋun: ŋal�ear:ñbula:ŋ(a)ñtu�iar:na’� (‘we two will go and see (or look for) ŋal�ear and tu�ear (they two)’)

(Dawes, 1790) {eora}

Of course the accusative prefix ŋυn is not used in association with subjects(either compact or expanded),

as in the sentence:

(7) They (plural)—Major Ross, Mr Clark & Mr Faddy—went to Norfolk Island

THEY (leftñexpanded plural pronominal SUBJECT)

major:ross:(mi)�, mister:clark:(mi)�, mister:faddy:(mi)�, �uŋ:mυ:i:lυS1 S2 S3 walkedó‘went’

(verb, past tense)S1 ò S2 ò S3 ó Mjr Ross and Mr Clark and Mr Faddy

ñŋ:wur(υla); norfolk:islandñ�wuóTO(allative)

�‘major:ross:, mr:clark:, mr:faddy:, yen:y::aouñn:r; n:i:’

(‘Mjr Ross, Mr Clark (and) Mr Faddy went to N.I.’)

(patye�oraŋ cited by Dawes, 1790) {eora}

Chapter 2, JAS 2004

Page 163

Corrections

90 , C. Illert & A. Allison

Hypothesis: -Lf.x -Lft,> -LftThe previous examples show how the south-east Australian language, in recentcenturies, employed the accusative markers mip, pun/pug and 4un. Frompublications such as Alexejew et al. (1968) we have long known that the mostfrequently used words, even in European languages are'the'{English}; 'de','le(a)' {French}; and'der','die' {German}. So it is clear that mip, punlpu4 and

4un would likewise have been amongst the most commonly spoken words in SE-Australian language over a large geographical area for a very long time.

Thus, in order to minimize confusion-given that these accusative markersgenerally were prefixes, respectively beginning with the consonants m,7 and 4-we might expect the relative frequencies of words beginning with these particularconsonants to have decreased across time in the general SE-Australian lexicon.Specificaliy, if mi7 and Tunlpult were used with about equal frequency innatural language, and pun less so, then we would expect proportionate decreases

in the relative lexical frequencies of word-initial-m, word-initial-p and word-initial-4-the emergence of accusativity within ancestral language leaving thisdistinctive, measurable, phonotactic signature upon the subsequent collective SE-Australian lexicon.

Because the fundamental root(s) punlpult obviously pre-existed the lexemem(ulu):i:p(u0): miL, we might expect the decrease in relative frequency of word-initial-p to be just slightly greater than that for word-initial-n, b:ut not muchmore because syntactic accusativity generally requires both markers. Also, assum-ing that compact pronominal objects are used about twice as often in normalspeech as fully expanded ones, it would seem reasonable to assume the decrease

in relative frequency of word-initial-4 to be approximately half that for word-initial-m: refer to the values of Af., Lft' U A,fa inTable2.

It is important to realize that the emergence of accusative syntax in ancestrallanguage had these phonotactic consequences, for the collective SE-Australianlexicon, that are predictable even in the absence of data! Without actual data wemight not be able to guess absolule magnitudes of individual frequency changesbut, on syntactic grounds, we do have a good idea of their relatiue magnitudes.It has never before been possible reliably to obtain such numbers describingchange through time because, until now, there was no demonstrably representa-tive proto-Australian lexicon to be compared with modern counterparts.

The Lexical Signature of Accusativity

Normalization of the lexical lrequencies in Table I required that 100%

+6',tf,-'{t $q tf4:I i-x (fi' t l-s t rr t rs) t (flr t l-q t-fr)

(15)

=If;:(f-*fn*fr+f) + (fr+fo+ t)W+(fr+fr+fn+f, +f7)

:Tn = (F-+ F n*F o+ F) + (Fr+Fr+ F")

+(Fr+Fn+F, +F, +F7)

Chapter 2, JAS 2004

Page 164

: Proto-Australian Language 91

where these terms have been grouped into three stable lexical blocks which retaintheir relative sizes. Words commencing with peripheral nasals and stops are one-

such stable block, constituting the bulk of all three lexicons, such that

,{^t{rt-o{rt{'- F^+ Fu+ tr, ' ri" ZA4o./- . (16)

Words commencing with lamino-palatals are another stable block within allthree lexicons such that

frL{',L{t.- F-+P', (r1)

The remaindeq words commencing with apico-alveolar and lamino-interdentalstops and nasals as well as the peripheral semi-vowel, collectively constitute a

third stable block within all three lexicons such that

{" t f' I f. , fa I '" , f*' t* ir=4€€s/6 (18)NJ*' '-

^Yet below this mirage of constancy there have been\dramatic changes in lexicalfrequencies of word-initial consonants rAt--++!, within each of these threelexical blocks (see Table 2), during the transition from proto-Australian

r{.\i.$*. Both modern lexicons exhibit large net decreases in word-initial. ^.

) \ - uasals. largely balanced by increases in word-initial stops, with word-initial semi-

$* t vowels compensating for any lexical residues (see Table 3).

The transition proto-Australian-'dhurga most ciearly illustrates the reciprocalnature of relationships between the changing frequencies of word-initial peri-pheral nasals and stops

Lf-= -LfsandLfa= -Lfa ( 1e)

In addition, both modern lexicons exhibit a large word-initial decrease in thenasal 7, variously compensated by the other lamino-palatals, most clearlyiliustrated by the transition proto-Australian-'dhurga for which

Lf'= - Lfd,- Lfv (20)

Table 3. Pattern of replacement olword-initial nasals by stops, approximately as occurred in the

dhurqa lexicon, with word-initial semi-vowels compensating for residual lexical discrepancies

r _ _ - - --- -- - --- - _-- - - -t_ - -i peripheral I apico i lamino i,fll

- a- - - - - - - - - - - 1

I labial i dorsal i alveolar i interdental i palatal it----------t- ------f -

Corrections

f.* fn+ fo+ fo = Fm+ F n*F o+F o = 64Yo

fr* fo* fr= F n+F a+F y= l5%o

f, + fn+ fn + fo + fo= F*+ Fz+ F, +F, + F,

(16)

(t7)

=20.3Yo (18)

><b

m

g

a

d

n

Chapter 2, JAS 2004

Page 165

92 C. Illert & A. Allison

perhaps largely due to cumulative phonetic drift of ancestral � into the otherlamino-palatals: ny�njy�dj and ny�y. This general pattern, of word-initialnasals turning into or being replaced by stops, continues to a lesser extent (closerto the level of statistical noise) into the third main block of words (commencingwith apico-alveolars and lamino-interdentals as well as the peripheral semi-vowel) with both modern lexicons suggesting a relationship

*fw�ñ(*fn�ò*fd)ñ(*fnò*fd� ) (21)

and Table 3 providing a good way to visualize all these changes.Generally, transfers (‘leakages’) between the three stable lexical blocks

e1ó�0

ñT

fmòf�òfŋòfb dt

ó(*fmò*f�)ò(*fŋò*fb) (22)

ó�ñ0.4% (proto-Australian�dhur�a)ñ1.2% (proto-Australian�thurrawal)

e2ó�0

ñT

f�òfdɔòfy dtó*f�ò(*fdɔò*fy)

(23)

ó�0.1% (proto-Australian�dhur�a)3.4% (proto-Australian�thurrawal )

e3ó�0

ñT

fwòfn�òfdòfnòfd� dt

ó*fwò(*fn�ò*fd)ò(*fnò*fd�) (24)

ó�ñ0.3% (proto-Australian�dhur�a)ñ2.2% (proto-Australian�thurrawal )

such that e1òe

2òe

3ó0, are all an order-of-magnitude-greater for proto-

Australian�thurrawal, as compared with proto-Australian�dhur�a, suggestingthat the thurrawal lexicon of Mathews (1903) is actually more-decrepit and less-authentic (despite being recorded at most a century earlier) than the dhur�alexicon of Eades (1976).This may seem surprising given that Robert Hamilton Mathews (1841–1918),

a Fellow of the Royal Society of New South Wales and author of over 170published papers, undoubtedly had access to proficient Aboriginal linguisticinformants. However, he was a 19th century amateur obsessed with publishingpapers and prone to plagiarism (Illert, 2001), who ‘tended to doctor andnormalise his notes for publication’ (Dixon, 1980, p. 15) on occasions simply

Chapter 2, JAS 2004

Page 166

Proto-Australian Language 93

inventing linguistic material (see Appendix A), so much so that we can nowprovide the above arithmetical measures quantifying the extent of errors and‘doctoring’ within his published data.At the other extreme, Diana Kelloway Eades (1976)—an expert phonologist

with modern linguistic training—largely through sheer competence k(notwith-standing the misgivings of Breen, 1980) achieved an order-of-magnitude-betterjob with her dhur�a lexicon, despite the handicap of working in this field 75years later, at a time when

Dharawal and Dhurga are not being kept alive in any active way. It is certainly only with afew of today’s oldest generation that words from the languages are surviving at all. . . . forthe past ten years, there have only been a handful of old people who remember any Dharawaland Dhurga words. There have been no fluent speakers in this time, although a few haveremembered isolated hints of grammar from time to time. . . . as these people in fact neverspoke the languages fluently, and only use occasional words from Dharawal and Dhurga, itis unrealistic to expect anything but complete abandonment of any use of the languages overthe next decade.

Non-zero values of e1, e

2and e

3are largely due to Anglicization of Aboriginal

words by English-speakers. For example the values of *f� in Table 2 show thatMathews (1903) grossly under-recorded word-initial-�s by an order of magnitude.Sometimes he misrecorded them as bs, as in �υru:�υru�‘birra:birra’ (‘light, notheavy’) and its converse �urυ:�urυ�‘burri:burri’ (‘whale’), this latter phoneticerror echoing the distinctively wrong and previously published word ‘purri:burry’(‘whale’) (Munima, 1863) {thurrawal}. Othertimes Mathews (1901) misrecordedword-initial-�s as lamino-palatals, giving �υru:�υru�‘jarra:jarra’ (‘lean’){dharruk}, again echoing another distinctive previously-published phonetic-error �υru�‘jerri’ (‘thin’) (Mann, 1840s) {eora}. Such transfer of word-initialperipheral-dorsal stops, to lamino-palatal stops or semi-vowels, also partiallyaccounts for the excessive values of e

1, e

2and *fy with Mathews’ thurrawal data.

Another factor contributing to his excessive value of *fy is the way thatEnglish-speakers tend to ‘hear’ the word-initial lamino-interdental-stop (espe-cially when followed by a long vowel) as some exotic vowel-cluster often involvingthe lamino-palatal semi-vowel, as in d�υla:wi��‘yello:wna’ (‘moon’) (Mann,1840s) {eora} or d�υla:ma(lυ):n�‘hila:ma:n’ (‘shield’) (Rowley, 1875) {dharu�},‘ilee:mo:n’ (‘shield’) (Tench, 1788) {eora} and ‘yila:ma:n’ (‘spear’??) (Eades,1976) {dhur�a}. Such a transfer, word-initial lamino-interdental stop to lamino-palatal semi-vowel, also partially explains the excessive values of e

2and e

3with

Mathews’ thurrawal data.For her part, Eades (1976) was weakest at distinguishing apico-alveolar and

lamino-interdental stops. The value of *fd associated with her dhur�a data seemsgreater than we might expect so, probably, some of the words commencing withd in her dhur�a lexicon should actually have commenced with d� .In fairness to Mathews, his 1903 thurrawal lexicon contained a good sample

of verbs and adjectives as well as nouns, whereas the 1976 dhur�a lexicon ofEades was all nouns. To the extent that accusative markers only prefix nounsand pronouns one may expect their impact to be more apparent in Eades’ dhur�alexicon. But one is still inclined to dismiss as unrealistic Mathews’ values of e

1,

e2and e

3in the several percent range, whilst suspecting that Eades’ values in the

Chapter 2, JAS 2004

Page 167

94, C. Illert & A. Allison

range of fractions of a percent may correspond to real effects in need of furtherstudy, because all these effects are at least one order of magnitude smaller thanthe huge net decrease in lexical frequency of word-initial nasals in both lexicons

ro) f.+fo+fn+_f,r+f,drJ,,

:Lf^*Lf't+Lft+Lfn+Lf, (25)

_ ! - Zl .ln (proto-Australian-- dhurga)

[ - t S.+V" (proto-Australian --thurrawal )

balanced by a net increase in word-initial stops in both lexicons

(26)

_ IZZ.t't" ( proto-Austral ian -- dhurga)

It Z.l'1" ( proto-Aust ralian -t hurrawal)

Clearly, one fifth of the entire ancestral lexicon-mainly words commencingwith m, .1n or 4-- experienced 'a great die-off' when modern languages came intobeing. This lexical effect of truly major proportions left a residual phonotactic'signature', on modern south-east-Australian lexicons, which seems to be wellexplained by the emergence in ancestral language of mip, pu4 and qun as

accusative prefixes.

The Diachronic Optimization Problem

Our problem requiles minimization of the lexical transfer functional

PO

o:sr re2*ca:.J ,

Li*dt

with respect to word-frequencies/o.ef such that

f : {"f*(t)e ct[- 7,0]:lfa:100% V r e [- 7] 01,*

"f. <-n: f* and ltol: p-

r', t-'/-r'* ^-"trS)-Fx as defined in Table 1)

subject to the physical condition that frequency changes (as in Table 2)generate lexical Entropy

.f*:rcx-fxl" (f*)

roI fo+fn+1a,+fq+fadtJ _,

: Lfs+ Lft'+ Lfa,+ Lfq+ Afa

(21)

each

(28)

Chapter 2, JAS 2004

Page 168

Proto-Australian Language 95

The appropriate formal constraint

'ó�0

�T

;��12f 2*ò

12i2* ( f* ln ( f*))2� dt (29)

gives the full Lagrangian

Ló;�

f*òj;��12f 2*ò

12i2* ( f* ln ( f*))2� (30)

which satisfies the Euler Equations

0óLLLf*ñddt �LLLf*� (31)

provided the respective frequencies are all Gompertz functions

f*(t)óe�e��t�c� for t é [ñT, 0] (32)

where

c*óln (ñln (F*)) and i*ó[ln (ñln (F*))ñln (ñln ( f*))]/T (32)

Since the time of Gompertz (1825) this function (equation (32)) has been usedas a ‘mortality law’, specifying the likelihood of death of biological organismswithin populations, making it the basis for modern actuarial science (Bowerset al., 1997). But it likewise describes the extinguishing of word-classes withinan evolving lexicon, and its respective constants, calculated from the frequencydata in Table 1, are presented in Table 4—although it is not possible to calculateindividual values i

�without knowing the value of T.

Appendix B supplies the first ever published mathematical model describingphonogenesis over the last 120 000 or so years. During this time, humanphonologies worldwide increased in size and complexity, by an entire order ofmagnitude, evolving over tens of millennia into modern sound-systems varyingin complexity from Rotokas in the Solomons (which has 11 forms) to Khmer(which has 74). As ‘modern’ dhurga had 14 consonants, compared to 10 inproto-Australian, we estimate that the proto-Australian lexicon in Table 1 maybe about 40 000 years old—hence the time constant in our equations is T�40 000years. Its precise value will no doubt be subject to ongoing debate but ourestimate at least provides a start, based upon some genuine phonological data,in a field where researchers have so far been unable to refute even the absurdproposition that proto-Australian may only be a few thousand years old

We have no way of telling whether proto-Australian was spoken 30,000 years or longer ago,or much more recently than that. . . . [it] could have been spoken by the dingo-bearing peoplefour millennia ago . . . Although we shall probably never be able to offer any proof, it is likelythat proto-Australian was spoken a considerable time ago; probably some time before proto-Indo-European, which is dated at around 3000 B.C. (Dixon, 1980, p. 19)

Chapter 2, JAS 2004

Page 169

96 C. Illert & A. Allison

Table 4. The Gompertz constants, calculated from the data in Table 1, describing how thefrequencies of individual word-initial-consonants changed from proto-Australian times(tóñT) to the present (tó0). Infinite values arise here from the zero frequencies in Table 1but, if one is unhappy with the notions that eñêó0 and ee

ñêó1, then it is possible to replace

the zero-frequencies in the tail of the distribution with their corresponding small butfinite ‘maximum likelihood’ values from equation (5) giving iyTócyñln(ñln( f0a11)) andidTócdñln(ñln( f0a10)). Also, for proto-Australian�dhur�a, we obtain cn�óln(ñln( f

0a11))

and in�Tócn�ñln(ñln( f0a9))

proto-Australian�dhurga proto-Australian�thurrawal

f*(t) i*T c* i*T c*

fm(t) 0.2779 0.6754 0.2445 0.6420f�(t) ñ0.2634 0.2843 ñ0.0231 0.5245fŋ(t) 0.2185 1.1418 ñ0.0055 0.9178fb(t) ñ0.1040 0.4973 ñ0.1783 0.4230f�(t) 0.6936 1.5418 0.6028 1.4510f�(t) ñ0.3251 0.7830 ñ0.3059 0.8022fy(t) ñê 1.3062 ñê 1.0439fn(t) 0.0818 1.1898 0.1657 1.2738fd(t) ñê 1.0392 ñê 1.4510fn�(t) ê ê 0.1242 1.5418fd�(t) 0.1079 1.2160 ñ0.0046 1.1035fw(t) 0.0734 0.9216 0.0472 0.8954

and that, without written records, we may as well give up because

. . . proto-Australian may well prove to be unreconstructable in principle, if it ever existed . . .linguists would be better off for some time yet continuing to chip away at the lower levels ofreconstruction, and offering the results of their work to archaeologists for interpretation . . .(Crowley, in McConvell & Evans, 1997, p. 295)

Acknowledgements

This paper is in support and in part-fulfilment of the requirements of PhDs byboth authors at their respective universities. They would like to thank ananonymous referee for valuable suggestions; Bob Hodge of the University ofWestern Sydney, Derek Abbott of Adelaide University and John Laurent ofGriffith University, for helpful discussions and feedback; also Daniela Reverberifor assistance with computer software and Michael Organ for assistance locatingrare literary and archival sources.

Notes

1. Since the time of G. K. Zipf, linguists have meant by a ‘power law’ any curve that even looks like 1/x,whilst statisticians have come to use it only for curves of the form xñb. The present paper models ageometric distribution ax which, because a is a decimal number between zero and unity, looks like a1/x Zipf curve and is therefore a ‘power law’ to linguists. There seems no reason why we should notpersist with traditional linguistic terminology, so long as ‘power law’ is written with inverted commas.

Chapter 2, JAS 2004

Page 170

Proto-Australian Language 97

2. Mathews (1904) recorded a wiradjuri word for grease/fat/wax, ‘:wa:mmo’ (‘fat’), deriving from (�υn):wu(rυla)-ŋu(ru)óretarded oscillations (óviscous) [substance] . . . hence also ‘be:wa:n-b:a:n�’ (‘fat’)(Mathews, 1904) {ŋunuwυl} deriving from �υ(n):wu(rυla):ŋ(uru)-b:i:�ŋóstomach-fat, and also‘�a:wa:ŋ::al ’ (‘honey’) (Mathews, 1904, 1908) {ŋaru�u} {ŋunuwυl} deriving from �υ(n):wu(rυla):ŋ(uru-i):�υl(a)óviscous-shiny [substance]. Viscous substances like grease or honey contrast with waterwhich is a flowy (óbulυ:i:d�υra:ŋuru) shiny (ói:�υla) [substance].

3. The expression �u(la):mυ(ra)ógreatly-deadly refers to a [spear with a] barb . . . hence the words‘nooro:ca:my’ (‘[a long] spear with one barb’) (Collins, 1798) {eora} and ‘:ka:mai’ (‘reed spear’)(Mathews, 1908) {ŋaru�u} both deriving from �uru-�u(la):mυ(ra)óempty/hollow/tubular barbed[spear] or, conversely, ‘maia:�u:ŋ’ (‘spear’) (Malone, 1875) {dariwυl} deriving from mυ(ra):�u(la)-�(ura).

4. R. H. Mathews’ allegedly plural word ‘yuin:buloala’ (‘man, collectively’) is wrong. The numeral bulalaonly denotes the dual case (two people), whilst the correct suffix for several or many people is wurυla.Additionally, it is quite difficult to imagine either wurυla or bulala being misheard as ‘buloala’ by anycompetent observer. This simultaneous mis-phoneticization and mis-semantization, ‘buloala’ beingused as plural, is a rather distinctive error which originated in the story of ‘The Nut Gatherers (twowomen) at Bendthually (óbυ:n:d�υla)’, as recorded at the Shoalhaven on 17 June 1872 by the ReverendAndrew Mackenzie and published two years later in The Journal of the Anthropological Institute ofGreat Britain and Ireland, volume 3, pages 255–6 (1874).

exteriorñexterioróoutside pair

ówomen (dual) óevil(split dual adjective)

d�arυ d�arυwυl d�arυ m:i:�υlañ(bula)la wirυ:�υl

(tribal name)� ‘jerrañthurawalñdtheri m:e:�aaolañli wurra:�al’(the two evil women outside/beyond [óshunned by] the tribe)

In this story there were two women and the word for woman (singular) was m:i:�υ or m:i:�υla (‘good-looking’), hence women (dual) was m:i:�υla-(bula)la which Mackenzie recorded well enough as‘me�aaolali ’. But he did not realize that m:i:�υ and m:i:�υla were equivalent and must have thoughtthat the suffix was ‘(me�a)aolali ’ hence, with an imagined soft b and some vocalic metathesis, we havethe now infamous ‘buloala’. Mackenzie obviously rechecked with his Aboriginal informant, LizzyMalone, who then gave him the correct numerals bula(la)ó‘bular’ and wurυlaó‘werulla’ or ‘wowulli ’as published in subsequent articles, including p. 113 of Rev. William Ridley’s book ‘Kamilaroi and otherAustralian Languages’ in 1875. Thus, Mackenzie had a morphological reason for originally making thismistake and subsequently corrected it. On the other hand Mathews (1903) had no such excuse and hismisogynist error, turning a correct feminine suffix into a wrong masculine counterpart, is so bizarrethat it illustrates how Mathews ‘topped-up’ his 1903 thurrawal vocabulary by plagiarizing and cobblinglexical items—errors and all—from previously published sources. In the process he unknowingly skewedthe statistical features of his data set in various tell-tale ways.

References

Alexejew, P. M., Kalinin, W. M. & Piotrowski, R. G. (Eds) (1968) Sprach-Statistik (Munich: Fink).Bak, P. (1997) How Nature Works, the Science of Self-organized Criticality (London: Oxford University

Press).Bowers Jr., N. L., Gerber, H. U., Hickman, J. C., Jones, D. A. & Nesbitt, C. J. (1997) Actuarial

Mathematics, p. 71 (Itasca, IL: Society of Actuaries).Breen, J. G. (1980) Linguistic salvage in Australia, review of Diana Kelloway Eades’ ‘The Dharawal and

Dhurga Languages of New South Wales south coast’, Lingua, 52, pp. 179–193.Brillouin, L. (1956) Science and Information Theory (New York: Academic Press).Bulmer, Rev. J. (1850s) Collected papers, Anthropology Department, National Museum of Victoria,

Melbourne. Also see Smyth (1876) and Curr (1888).Chomsky, A. N. (1957) Syntactic Structures (The Hague: Mouton).

Chapter 2, JAS 2004

Page 171

98 C. Illert & A. Allison

Collins, D. (1798) An Account of the English colony in New South Wales, two volumes (London: T. CadellJr. & W. Davies).

Comrie, B. (1989) Language Universals and Linguistic Typology (Oxford: Blackwell).Curr, E. M. (1887) The Australian Race, four volumes (Melbourne: Government Printer).Dawes, W. (1790a) Grammatical Forms of the Language of New South Wales, in the Neighbourhood of

Sydney. Now in the Marsden Collection, School of African and Oriental Studies, London.Dawes, W. (1790b) Vocabulary of the Language of New South Wales, in the Neighbourhood of Sydney

(Native and English). Now in the Marsden Collection, School of African and Oriental Studies,London.

Dawson, J. (1881) Australian Aborigines, the Languages and Customs of Several Tribes of Aborigines in theWestern District of Victoria (Melbourne: George Robertson).

Dixon, R. M. W. (1976) Grammatical Categories in Australian Languages, Linguistic Series 22 (Canberra:Australian Institute of Aboriginal Studies).

Dixon, R. M. W. & Blake, B. J. (Eds) (1979) Handbook of Australian Languages, vol. 1 (Canberra:Australian National University Press).

Dixon, R. M. W. (1980) The Languages of Australia (London: University Press).Eades, D. K. (1976) The Dharawal and Dhurga Languages of the New South Wales South Coast (Canberra:

Australian Institute of Aboriginal Studies). The dhurga lexicon of 321 words is Section B, vocabularyfrom recent years on pp. 75–84.

Egghe, L. (1999) On the law of Zipf-Mandelbrot for multi-word phrases, Journal of the American Societyfor Information Science, 50(3), pp. 233–241.

Egghe, L. (2000a) General study of the distribution of N-tuples of letters or words based on thedistribution of the single letters of words, Mathematical and Computer Modelling, 31, pp. 35–41.

Egghe, L. (2000b) The distribution of N-grams, Scientometrics, 47(2), pp. 237–252.Elkin, A. P. (1975) R. H. Mathews, his contribution to Aboriginal studies—part 2, Oceanea, XLVI(2),

pp. 126–153.Everitt, M. M. & Mathews, R. H. (1900) The organisation, language and initiation ceremonies of the

Aborigines of the south-east coast of New South Wales, Journal & Proceedings of the Royal Societyof NSW, 34, pp. 262–281.

Fisher, R. A. (1912) The method of maximum-likelihood, Messenger of Mathematics, 41, pp. 155–160.Fisher, R. A. (1958) Statistical Methods for Research Workers, 13th edn (New York: Hafner).Gompertz, B. (1825) On the nature of the function expressive of the Law of Human Mortality, Phil.

Trans. Royal Soc. London.Govett, W. (1836) Notes and Sketches of New South Wales, London.Holmer, N. M. (1971) Notes on the Bandjalang dialect spoken at Coraki and Bungawalbin Creek NSW,

Australian Aboriginal Studies No. 32, Australian Institute of Aboriginal Studies, Canberra.Hunter, J. (1793) An Historical Journal of the Transactions at Port Jackson and Norfolk Island (London:

J. Stockdale), word-list on pp. 407–411.Huntsberger, D. V. & Billingsley, P. (1981) Elements of Statistical Inference, 5th edn (Boston: Allyn &

Bacon).Illert, C. R. (1987) il Nuovo Cimento (series D), 9(7), pp. 791–814.Illert, C. R. (1995) IEEE Computer Graphics & Applications, 15(4), pp. 89–90.Illert, C. R. (2001) The centenary of Mary Everitt’s Gundungurra Grammar, Journal & Proceedings of

the Royal Society of New South Wales, 134(1/2), pp. 19–44.Illert, C. R. (2003) Lexigenesis in ancestral south-east-Australian Aboriginal language, Journal of Applied

Statistics, 30(2): 113–143.Illert, C. R. (2004) An inter-vocalic retroflex stop in Victorian languages?, in preparation.Larmer, J. (1898a) Native vocabulary of miscellaneous NSW objects, Journal of the Royal Society of New

South Wales, 32, pp. 223–229.Larmer, J. (1898b) Aboriginal words and their meanings (Bateman’s Bay), Science of Man, 2(8),

pp. 146–148.Lyons, J. (1968) Introduction to Theoretical Linguistics (London: Cambridge University Press).Lyons, J. (1995) Linguistic Semantics, an Introduction (London: Cambridge University Press).Mackenzie, A. (1874) Specimens of Native Australian Languages, Journal of the Anthropological Institute

of Gt. Britain and Ireland, 3, pp. 247–264.Mackenzie, A. (1875) In W. Ridley, Kamilaroi and Other Australian Languages (Sydney: Thomas

Richards—Government Printer), pp. 99–101, 109–115 and 143–147.

Chapter 2, JAS 2004

Page 172

Proto-Australian Language 99

Malone, L. & Malone, J. (1875) InW. Ridley, Kamilaroi and Other Australian Languages (Sydney: ThomasRichards—Government Printer), word-list pp. 97–101 and 109–114.

Malone, L & Malone, J. (1878) Australian languages and traditions, Journal of the AnthropologicalInstitute of Great Britain and Ireland, 2, word-list pp. 262–266.

Mandelbrot, B. B. (1951) Adoption d’un message a ligne de transmission: I & II, Comptes Rendus (Paris),232, pp. 1638–1640 and 2003–2005.

Mandelbrot, B. B. (1953) An informational theory of the statistical structure of languages, in: W. Jackson(Ed.), Communication Theory (Butterworths).

Mandelbrot, B. B. (1954) Simple games strategy occurring in communication through natural languages.Symposium on statistical methods in communication engineering (Berkeley, Aug. 17–18, 1953), in:Transactions of IRE (professional groups on information theory), 3, pp. 124–137.

Mandelbrot, B. B. (1959) A note on a class of skew distribution functions, Information and Control, 2,pp. 90–99.

Mandelbrot, B. B. (1961) Final note on a class of skew distribution functions: analysis and critique of amodel due to H. A. Simon, Information and Control, 4, pp. 198–216.

Mandelbrot, B. B. (1961) Post scriptum to final note, Information and Control, 4, pp. 300–304.Mandelbrot, B. B. (1965) Information theory and psycho-linguistics, in: B. Wolman & E. Nagel (Eds)

Scientific Psychology, Principles and Approaches, pp. 550–562.Mandelbrot, B. B. (1968) Les constantes chiffrees du discourts, in: J. Martinet (Ed.), Encyclopedie de la

Pleisde: Linguistique (Gallimard).Mann, J. F. (1840s) Australian Aborigines, a Few Notes on Their Language, the Language of Long Dick . . .

of the Cammaray Tribe . . ., unpublished manuscript, Mitchell Library, Sydney, Australia.Mann, J. F. (1900) Aboriginal Place Names, Shoalhaven Telegraph, editorial, 8 September.Mathews, R. H. (1901a) The Gundungurra language, in: Proceedings of the American Philosophical

Society, XL(167), pp. 140–148.Mathews, R. H. (1901b) The Thurrawal/Gundungurra/Dharruk . . . language[s], in: Journal and Proceedings

of the Royal Society of New South Wales, 35, pp. 126–160.Mathews, R. H. (1901c) The Thoorga language, in: Proceedings and Transactions of the Royal Geographical

Society of Australasia, Queensland, 17, pp. 49–74.Mathews, R. H. (1902) The Aboriginal languages of Victoria, Journal of the Royal Society of NSW, 36,

pp. 71–106.Mathews, R. H. (1903) Vocabulary of Kamilaroi . . ., Journal of the Anthropological Institute of Gt. Britain

& Ireland, 33, pp. 259–283. The thurrawal lexicon of 428 words is on pp. 275–279.Mathews, R. H. (1904) The Wiradjuri and other languages of N.S.W., Journal of the Royal Anthropological

Inst., 34(2), pp. 294–305.Mathews, R. H. (1908) Vocabulary of the Ngarrugu Tribe N.S.W., in: Proceedings of the Royal Society of

New South Wales, 42(24), pp. 335–342.McConvell, P. & Evans, N. (Eds) (1997) Archaeology and Linguistics (Oxford: Oxford University Press).McNicol, S. & Hosking, D. (1994), in: N. Thieberger & W. McGregor (Eds) (1994) [the] Macquarie

[Dictionary of] Aboriginal Words (Sydney: Macquarie University Press), pp. 79–99.Mendenhall III, W., Wackerley, D. D. & Scheaffer, R. L. (1986)Mathematical Statistics with Applications,

3rd edn (Boston, MA: Duxbury Press).Meyer, H. A. E. (1843) Vocabulary of the Language Spoken by Aborigines of the Southern and Eastern

Portions of the Settled Districts of South Australia . . . Preceded by a Grammar (Adelaide: James Allen).Montemurro, M. A. (2001) Beyond the Zipf–Mandelbrot Law in quantitative linguistics, Physica-A,

300(3–4), pp. 567–578.Munima, M. (1863) Word-list (dated 21st Sept. 1899), Science of Man, 2(8), pp. 141–142.Nathan, I. (1848) The Southern Euphrosyne and Australian Miscellany . . . An Historical Sketch, With

Examples of the Native Aboriginal Melodies . . . (London: Whittacker & Co.), word-list on pp. 134–135.Nichols, J. (1993) Ergativity and linguistic geography, Australian Journal of Linguistics, 13, pp. 39–89.Organ, M. (1993) Illawarra and South Coast Aborigines 1770–1900, two volumes (Canberra: Aboriginal

Studies Press).Osmond, M. (1989) A reconstruction of the phonology, morphology and syntax of the Sydney language

as recorded in the notebooks of William Dawes, Unpublished student paper, Department ofLinguistics, Australian National University, Canberra, Australia.

Pettit, N. D. & Dawson, W. (early 1850s) Word-list, in: Howitt Manuscripts MS9356, La Trobe Collection,State Library of Victoria, Melbourne.

Chapter 2, JAS 2004

Page 173

100 C. Illert & A. Allison

Quoy, J. R. C. & Gaimard, J. P. (1826) Word-list, in: M. J. D’Urville (Ed.) (1833) Voyage de decouvertesde l’Astrolabe 1826–9 (pp. 1–30 of the ‘Philologie’ volume) (Paris: Ministere de lay Marine).

Ridley, W. (1873) Report on Australian languages & traditions, Journal of the Anthropological Institute ofGreat Britain and Ireland, 2, pp. 257–291.

Ridley, W. (1874) Remarks on specimens of the Midthung or Tharumba and Thurawal languages, Journalof the Anthropological Institute of Great Britain and Ireland, 3(2), pp. 249–252.

Ridley, W. (1875) Kamilaroi and Other Australian Languages (Sydney: Thomas Richards—GovernmentPrinter).

Ridley, W. (1878) Australian languages and traditions, Journal of the Anthropological Institute of GreatBritain and Ireland, 7(3), pp. 232–266.

Rowley, J. (1875) In W. Ridley, Kamilaroi and Other Australian Languages (Sydney: Thomas Richards—Government Printer), word-list published, pp. 103–108.

Rowley, J. (1878) Australian languages and traditions, Journal of the Anthropological Institute of GreatBritain and Ireland, 2, word-list published, pp. 258–262.

Shannon, C. E. & Weaver, W. (1949) The Mathematical Theory of Communication (USA: University ofIllinois Press).

Sharpe, M. (1994) Bundjalung word-list, in: N. Thieberger & W. McGregor (Eds) [the] Macquarie[Dictionary of] Aboriginal Words, pp. 1–22 (Sydney, Australia: Macquarie University Press).

Silverstein, M. (1976) Hierarchy of features and ergativity, in: R. M. W. Dixon (Ed.) GrammaticalCategories in Australian Languages, pp. 112–171 Linguistic Series 22 (Canberra: Australian Instituteof Aboriginal Studies).

Smyth, R. B. (1876) The Aborigines of Victoria, in two volumes (Melbourne, Australia: GovernmentPrinter).

Tench, W. (1789) Sydney’s First Four Years . . . Narrative of the Expedition to Botany Bay and a CompleteAccount of the Settlement at Port Jackson, introduction and annotations by L. F. Fitzhardinge,published as two books 1789 and 1793.

Thieberger, N. & McGregor, W. (Eds) (1994) [the] Macquarie [Dictionary of] Aboriginal Words (Sydney,Australia: Macquarie University Press).

Threlkeld, L. E. (1834) An Australian Grammar, Comprehending the Principles and Natural Rules of theLanguage, as Spoken by the Aborigines in the Vicinity of Hunter’s River, Lake Macquarie . . . (Sydney,Australia: Stephen Stokes & Co.).

Threlkend, L. E. (1892) Collected works reprinted as, An Australian Language as Spoken by the Awabakal,the People of Awaba or Lake Macquarie (near Newcastle, New South Wales) . . . re-arranged,condensed, and edited by John Fraser (Sydney: Charles Potter—Government Printer).

Tiechelmann, C. G. & Schurmann, C. G. (1840) Outlines of a Grammar, Vocabulary and Phraseology ofthe Aboriginal Language of South Australia, Spoken by the Natives in and for some Distance aroundAdelaide (Adelaide, Australia).

Trask, R. L. (1999) Key Concepts in Linguistics (New York: Routledge).Troy, J. (1992) The Sydney language notebooks, Australian Journal of Linguistics, 12, pp. 145–170.Wood, G. A. (1924) Lieutenant William Dawes and Captain Watkin Tench, Journal and Proceedings of

the Royal Australian Historical Society, 10(1), pp. 1–24.Yallop, C. (1982) Australian Aboriginal Languages (London: Andre Deutsche).Zipf, G. K. (1932) Selective Studies and the Principle of Relative Frequency in Language.Zipf, G. K. (1949) Human Behaviour and the Principle of Least Effort (USA: Addison Wesley).Zipf, G. K. (1965) Psycho-Biology of Language: An Introduction to Dynamic Philology (Boston, MA:

Houghton Mifflin Co. 1935; MIT Press 1965).

Chapter 2, JAS 2004

Page 174

Proto-Australian Language 101

Appendix A. Examples of Readily Recognizable Ancestral Roots in the ThurrawalLexicon of Mathews (1903)

barυ� ‘birra’ (‘cheek’óopposing sides [of the mouth]),‘dhau:m:bara’ (‘spread’óthings greatly apart),

bυra� ‘burru:n:��al ’ (‘clitoris’óthing between undulations/labia),‘bur:n:da’ (‘penis’óthing between legs),‘win:burra (‘whistle’óblow between [lips]),‘bin:burra’ (‘suck’ósuck between [lips]),‘mim:burra’ (‘kiss’óobstructing [air-flow] between [lips]),‘bura:i ’ (‘night’óbetween [sun-shine]ing[s]),‘burr:i:wurri ’ (‘morning’),

bυru� ‘barru’ (‘tired’), ‘berru:n�’ (‘mist’),buru� ‘buru’ (‘kangaroo’óhoppy/bouncy [creature]),

‘buru:n�’ (‘vagina’),bara� ‘burrah’ (‘wallaby’óambling [creature]),bulυ� ‘bullia’ (‘dead’), ‘bull:i:n�’ (‘die’),

‘bul:ma’ (‘fight’), ‘bu:m:i:lla:n�’ (‘beat’),‘�ula:t:bu:n�ara’ (‘break’),

bulala� ‘bullu’ (‘labia majora’ópair [of things]), ‘billuli ’ (‘wide’),‘yuin:buloala’ (‘man, collectively’), ???*

�υn� ‘�aian’ (‘large’óvery [big]), ‘�aian:wilai ’ (‘grow’),‘ban:�an(�)’ (‘an old man’óvery [old] man),‘karu:�an(�)’ (‘common magpie’óvery noisy),‘�ulu:�an(�)’ (‘short’óvery hunched),‘�ur:�an(�)’ (‘pigeon, bronze-wing’óvery [fluttery]),‘dhurra:�an(�)’ (‘large stream, river’óvery flowy),‘dhurra:�an:�an(�)’ (‘little stream, creek’óslightly flowy),‘�un:�un(�)’ (‘sleepy’óslightly [awake]),

�arυ� ‘�ara’ (‘tell’ó[sound] from),‘karu:�an(�)’ (‘common magpie’óvery noisy),‘kurru:�aia’ (‘shout’óvery noisy),‘jan:�aiiri ’ (‘silent’óNOT [sound] from),‘minji:�arr:i ’ (‘sky’óthe source [of light]),‘�a:mun�’ (‘talk’), ‘be:��aru:n�’ (‘vomit’),

�υra� ‘�ur:i ’ (‘ear’óreceptacle [for sound]),‘yuru:n�’ (‘jealous’ópossessive [person]),‘kara:n:�ara’ (‘steal’óenhanced receivement),

�ura� ‘�ura’ (‘hail’óhard), ‘�ura:bun�’ (‘stones’ófirm),‘walla:n:�urra’ (‘rocks’ófirm),

�aru� ‘kuru’ (‘clouds’ófluffy),�urυ� ‘�urea:n’ (‘fat; adj.’), ‘kurai:nyun�’ (‘fat; noun’),

‘�ura:uara’ (‘opossum’óchubby),‘burri:burri ’ (‘whale’órotund),‘mura’ (‘testicles’óconvex [things]),

�υru� ‘birra:birra’ (‘light, no heavy’óslim),�ula� ‘�ula:n:�ula:n’ (‘crooked’),

Chapter 2, JAS 2004

Page 175

102 C. Illert & A. Allison

‘�ulla:n�:��ulla:n�’ (‘large locust[s]’),‘kull:ia’ (‘cut’), ‘�ula:nya’ (‘kill’),‘�ula:t:bu:n�ara’ (‘break’),‘winnan:n:�ala:n�’ (‘clever-man’óshaman, sorcerer),

�alu� ‘�alu’ (‘crane, slate-coloured’óharmless),murυ� ‘murra:dha’ (‘small’ósmall thing),

‘murra:mul ’ (‘fingers, hand’ósmall five-[thing]sófingers),‘mur:uarai:n’ (‘hair on pudenda’ónumerous short things),‘�:uin:murra’ (‘twist’ógently rotate),

mυra� ‘dhulli:n:murra’ (‘husband’), ‘m:i:�a:murra’ (‘wife’),‘mur:al ’ (‘noisy’), ‘kunna:murri ’ (‘burn’),‘nhur:murra’ (‘pretend’), ‘bul:ma’ (‘fight’),‘nu��un��:baia:murra’ (‘praise’),

murυla� ‘murra:mul ’ (‘fingers, hand’ósmall five-[thing]sófingers),malυ� ‘mili ’ (‘eyes when shut’óobstructed [vision]),

‘mullar:aura’ (‘blind’),mυlυ� ‘meh’ (‘eye’), ‘mulya:n’ (‘eagle-hawk’),mula� ‘milla:n�’ (‘vulva’ó[source of] menstrual-droplets),

‘mullya:n’ (‘morning star’), ‘mulla:mullu:n�’ (‘Pleadies’),mulυ� ‘mula:ty’ (‘soft’), ‘mul:dha’ (‘old woman’),ŋuru� ‘n�urr:i ’ (‘shin’ówiggly), ‘nhuru:n�’ (‘arm’ówiggly),

‘n�u:mun�’ (‘knee’ówiggly),‘ji:lli:n�urr:i ’ (‘centipede’ówiggly little [leg]s),‘�ur:�an(�)’ (‘pigeon, bronze-wing’óvery fluttery),‘wullu:nya’ (‘swim’ólots-of flailing),‘n�ur:nun’ (‘throat’óundulatory [tube]), ‘�uru’ (‘neck’),‘buru:n�’ (‘vagina’),‘n�urru:n�:�al ’ (‘a woman’ósexually mature),‘n�auu’ (‘blood’),‘n�uru:n�:n�uru:n�’ (‘red’ópulsating/throbbing [blood]),‘�ula:t:bu:n�ara’ (‘break’),‘n�urru:n:de’ (‘hear’ó[sound] oscillations),

ŋulu� ‘n�ulu’ (‘forehead’)ócorrugated), ‘n�ulur’ (‘sweat’),‘burru:n:��al ’ (‘clitoris’óthing between corrugations/labia),‘n�ullu:n�’ (‘sit’óhunched/folded [legs]),‘�ulu:�an(�)’ (‘short’óvery hunched),

wara� ‘wurri ’ (‘far’),wυri� ‘waru:waru’ (‘straight’),

‘wure’ (‘sun’) hence ‘wuri:wure’ (‘sunlight, light of a fire’),wurυla� ‘wallu’ (‘chin, beard’ónumerous [hairs]), ‘wollar’ (‘head’),

‘wullu:n:�urra’ (‘rocks’ónumerous round/convex things),‘wullu:nya’ (‘swim’ólots-of flailing),‘�aian:wilai ’ (‘grow’), ‘wurra:i:ri ’ (‘play, dance’),

�ara� ‘nyir’ (‘scales of fish’ótop/surface [things]),‘birra:nun�’ (‘toe-nails, finger-nail’),

�υra� ‘nyirra’ (‘navel’ómiddle [thing]óbelly-button),‘nyirri:ch’ (‘navel-string’), ‘n�ura:nhun�’ (‘afterbirth’),

Chapter 2, JAS 2004

Page 176

Proto-Australian Language 103

n�υra� ‘yenda’ (‘walk’) hence ‘yendea narrea’ (‘go’ówalk yonder),‘nharra:wa:n:�ai ’ (‘lightning’óyonder distant flash),

d�υra� ‘thura:ra’ (‘sting or stab’ó[penetrating] through),‘dhurra:�an(�)’ (‘large stream, river’óvery [strongly flows] through[the landscape]),‘dhurra:�an:�an(�)’ (‘little stream, creek’ógently [flows] through [thelandscape]),‘dhurra:ma:n:�a’ (‘hole cut in tree’óhole through [tree trunk]),

dɔi� ‘ji:rrar’ (‘hair of head, fur of animals’ónumerous tiny-[hair]s),‘ji:ra:n�’ (‘leaves of trees’ónumerous tiny-[leave]s),‘ji:lli:n�urr:i ’ (‘centipede’ónumerous tiny-wiggly[leg]s),‘ji:bbir:n�’ (‘dew’),‘dyi:dyir’ (‘dry’ólittle [water] through [creek]).

Appendix B. Phono-genesis

We have seen how the proto-Australian ‘power law’ of Figure 3, extrapolatedtens of millennia forward through time as in Figure 4, actually fitted ‘modern’dhurga quite closely as in Figure 2. We should likewise be able reliably toextrapolate tens of millennia backward, from proto-Australian times, to studyhow human phonology evolved over the greater part of the last 100 000 years.Equations (8) and (9) permit this. We need only choose H, the approximatenumber of consonants in the alphabet during any given era, calculating the

Figure 5. Power-law curves, describing increasingly complex sound systems, as human languagedeveloped over the last 120 000 years.

Chapter 2, JAS 2004

Page 177

104 C. Illert & A. Allison

corresponding value of a with equation (11). The Hó2 stage (the first ‘step’shown in Figure 5) represents the earliest spoken human language, a system ofsnorts and grunts arising from mutation of the FOXP2 gene, prior to 120 000years ago when the full range of necessary improvements to the larynx and thechest muscles controlling the diaphragm came into being, and the human thoracicspinal chord increased in size facilitating the requisite precise cortical controlover breathing required for speech. We know that languages such as dhurga hadevolved a phonology of Hó14 by recent times tó0. Assuming that averagephonological complexity H, increased linearly with time t (in years), we have

tó10 000îHñ140 000

which means that proto-Australian, with Hó10, corresponds to a time-constantin equations (27) to (32) of about Tó40 000 years. The true value could becloser to 60 000 years, as proto-Australian came originally through south-eastAsia and predated human occupation of Australia. Our estimate uses the fewestassumptions but, certainly, by 45 000 years ago, proto-Australian would havespread throughout all Australia.

Chapter 2, JAS 2004

Page 178

Origins of Linguistic Zonation in the

Australian Alps. Part 1 – Huygens’

Principle

CHRISTOPHER R. ILLERT

University of Western Sydney, New South Wales, Austrlia

ABSTRACT The hitherto poorly recorded boundaries of extinct traditional south-east-AustralianAboriginal languages can now be redetermined with greatly improved precision using anentropy-maximizing phonetic-signature calculated from existing data sources, including oldword-lists and census forms, that have, until now, largely been considered informationallyworthless. Having thus determined traditional Aboriginal language zones to a previouslyunimaginable degree of geographical precision, it is argued that these boundaries should not beviewed merely as a static ‘snapshot’ but, instead, as the end-product of a knowable dynamicprocess (Gillieron wave propagation) governed by well-known physical rules (such as Huygens’principle and Snell’s Law) and operating over ‘deep’ time-scales more familiar to thearchaeologist than the linguist. Although this initial study is limited to south-eastern Australia,the new methodology provides the first real hope of obtaining a detailed understanding oflanguage dispersal throughout the entire continent over the past 60,000 years.

KEY WORDS: Lexical signature, deep linguistics, Gillieron wave propagation, Huygens’ Principle

Introduction

It is clear that there were different Aboriginal languages throughout south-eastern-

Australia two centuries ago. Charles MacAlister (1907), for example, told how ‘. . .inland people found it hard to understand the Tablelanders or the coastal tribes’. But

throughout this entire region indigenous language was extinguished so rapidly – before

studies of appropriate scientific calibre were made – that today we have only fragmentary

information about traditional linguistic boundaries. Accordingly, Terry Crowley (1997:

289–291) complained

. . . one serious problem that faces us in reconstructing the linguistic history of northern NewSouth Wales is the nature of the data that we are forced to operate with . . . the descriptivematerials range from moderately good in some cases . . . , fragmentary in others . . . , appallingin other areas . . . , to completely non-existent . . . [there are] areas for which we have namesbut no information . . . languages that have disappeared without trace. . . . Even the bestdescribed languages . . . may well represent very poor shadows of their former selves . . .much of the lexical richness has disappeared without trace. In Bundjalang, I was unable to

Journal of Applied Statistics

Vol. 32, No. 6, 625–659, August 2005

Correspondence Address: Christopher R. Illert, 2/3 Birch Crescent, East Corrimal, NSW 2518, Australia.

0266-4763 Print=1360-0532 Online=05=060625–35 # 2005 Taylor & FrancisDOI: 10.1080=02664760500079258

Chapter 3, JAS 2005

Page 179

record more than 1000 lexical items, not for want of trying . . . [and] this is the situation forone of the better described languages of the region . . . .

Norman Tindale (1974) pointed out the relevance of traditional tribal boundaries, noting

that a ‘common bond of language’ is one of the criteria that makes a group of people a

tribe, though there does not have to be a language barrier to mark a tribal boundary.

Additionally, Diana Eades (1976: 3) commented that

. . . because of the limited amount of work done in the area, we have to use all availablematerial as best as possible to work out how many languages were spoken and where . . .we must consider grammar, phonology and vocabulary, as well as geographical factorsand native speakers statements about intelligibility. It is impossible to be clear on the situationof dialects,

which raises the issue of degrees of linguistic relatedness within dialect-chains of the kind

mentioned by Crowley (1992/3/4/6: 247)

. . . on the border between Queensland and New South Wales where cognate counts in thebasic vocabulary of a number of neighbouring speech communities are relatively high andwhere the two varieties are mutually intelligible. However . . . [for] the speech communitiesat the extreme ends of this chain . . . mutual intelligibility is not conceivable.

Such geographically vast dialect chains, sometimes extending continuously for thousands

of kilometres, obviously came into being over time-scales that are simply beyond existing

linguistic methodologies. And ultimately there are the questions of deep linguistics as

articulated by Crowley (1992/3/4/6: 305)

. . . are the modern Australian Aboriginal languages all descended from a single proto-language that may have been spoken 40,000 (or even 60,000) years ago? Or are they des-cended from the language of the people who introduced the dog [dingo] about 5,000 yearsago? Or are some of the languages descended from the older, original language, andothers descended from the more recently introduced language? Until linguists are able tocarry their reconstructions further back in time than we are able to do at the moment,these questions will have to remain unanswered . . . .

Australianists do need to stop muddling the time-scales of shallow linguistics (mere thou-

sands of years) with those of deep linguistics (tens of millennia). And, contrary to the pro-

testations of Dixon (1980: 37), over deep linguistic time-scales one cannot place much

reliance upon the grammatical features of recent languages. Probably just about every-

thing currently known about Aboriginal grammar is burnt-off at time depths beyond

5000 years and therefore largely irrelevant. On the other hand, the phonological constancy

and consistency, throughout Australian languages generally, suggests that basic phonotac-

tic information conceivably just might provide a reliable account of language change over

truly deep time-scales. In principle, we might delve ever further back through time, simply

by collecting lexical data over a ‘sufficiently large’ geographical study-area. Even the size

of Australia is not the ultimate limiting factor because, long ago, proto-Australian arose

originally from other continents.

A previous paper (Illert & Allison, 2004) noted the existence of three diachronically

‘stable phonetic blocks’ (small sets of word-initial consonants), which seem to mutually

interact and compensate in accordance with the physical principle of entropy maximi-

zation, within traditional south-east-Australian lexicons. This being so, the associated

‘lexical transfer vector’ might profitably be used as a kind of linguistic ‘DNA signature’

identifying broad ancestral language families. This hypothesis is now tested using 37 his-

toric word lists collectively containing 7601 Aboriginal words, and nine early census lists

626 C. R. Illert

Chapter 3, JAS 2005

Page 180

Origins of Linguistic Zonation in the Australian Alps 627

('Blanket Registers') collectively containing 445 traditional names, gathered from knowngeographical locations chosen as uniformly as possible throughout our south-east

Australian study area (see Appendix 1).

'Lexical Signature' as a Tool for the Determination of Linguistic Boundaries

Previously (Illert & Allison, 2004) a lexical-transfer vector was defined

e : (er, Ez, €z) (1)

quantifying diachronic 'phonetic leakage' between three reasonably stable blocks ofword-initial consonants within the collective traditional south-east-Australian lexiconsuch that

+f ,dt

]- Lfi -t Lf,

where

s1fe2*e3:Q

for word-initial consonant frequencies

,f. € f = lf.al e cr[-r, o1: f.f. :100vo vt € [-r,0],

rf;t O ::fi end f?( as defined in Appendix I I

rt (n= f* and l(o): P-It is also useful to define the magnitude of this lexical-transfer vector as

s : sgn(s3)

where we distinguish between negative and positive values of s3 because sgn(e3): *indicates an expanding phonology, whereas sgn(e3): - can indicate phonologicalsimplification, and we need to know the direction of phonogenesis. All these quantities,

calculated from the raw data in Appendix 1, are summarized in Table l.In the second part of this paper 'Type-A' and 'Type-B' language will be rigoror-rsly

defined in terms of the different solutions to a quadratic equation arising from what is com-

al

a2

a3

10: J_rf ^*.f etf n-tIodt

: (Lf^ + A/s) + (A/o + A/r)

10: I f ,*fa*/,drJ-r "

: Lh * (Lfd+ Lfy)

10: I f w -l-f ,-ff a -tf ,-f f a-l.f r

lT

: Lf. * (Lft+ Lfa) + (.Lf, + Lf!)

(2)

(3)

(.4)

(s)

Chapter 3, JAS 2005

Page 181

Table 1. Components of the lexical-transfer vector

Geographical region 11 12 13 1 Language type

Australian Alps

picton census 6th August 1842 9 27 22 211.576 B

stonequarry wordlist Thomas

Mitchell 1830s

6.2 25.1 21.1 28.1 B

Argyle wordlist John D. Lang,

1840s

17.9 25.8 212.1 222.37 B

sutton forest wordlist Louisa

Atkinson, 1853

18.7 214.5 24.2 224.0 B

midthung/tharumba Shoalhaven

wordlist A. Mackenzie, 1872

3.4 20.3 23.1 24.611 B

tuggeranong census June 1834 10.4 21.9 28.5 213.565 B

northern ‘˛arigu’ (¼wolgalu)

wordlist R.H. Mathews, 1908

0.7 3.4 24.1 25.3722 B

‘thurrawal’ (¼˛unawal) wordlistR.H. Mathews, 1903

21.2 3.4 22.2 24.2237 A

˛unawal wordlist R.H. Mathews,

1904

21.5 3 21.5 23.6742 A

sthn. ˛arigu wordlist

L.A. Hercus, 1969

23.8 5.8 22 27.2167 A

Tasmania

pt. dalrymple wordlist Joseph

Gaimard, 1826

211.9 214.5 26.4 32.385 original-B

research bay, Jan 1793 J.J.H. de La

Billardiere

211.9 214.5 26.4 32.385 original-B

Western-Zone

wiradyuri wordlist

R.H. Mathews, 1904

21.8 21.4 3.2 3.929 original-B

queanbeyan wordlist Nelly

Hamilton, 1887

25.7 23 8.7 10.825 original-B

queanbeyan census 5th May 1841 25.8 24.2 10 12.299 original-B

goulburn census 1837-38-43 28.4 28.4 16.8 20.576 original-B

Eastern-Coast

sydney eora wordlist Dawes/Osmond, 1790

24.8 21.5 6.3 8.061 original-B

iyora wordlist Jakelyn Troy, 1993 22.9 21.7 4.6 5.697 original-B

cowpasture wordlist John Rowley,

1878

28.1 23 11.1 14.065 original-B

camden wordlist James Bowman,

1830s

21.2 26 7.2 9.449 original-B

bong bong census

15th November 1836

212.8 20.8 13.6 18.693 original-B

wollongong census 21st May 1834 218.7 26.4 25.1 31.948 original-B

wodi wodi wordlist Lizzy

Malone, 1875

21.4 21.7 3.1 3.803 original-B

thurrawal wordlist A. Mackenzie,

1872

212.9 24.1 17 21.731 original-B

(continued)

628 C. R. Illert

Chapter 3, JAS 2005

Page 182

Table 1. Continued

Geographical region 11 12 13 1 Language type

south coast wordlist, Richard

Dawsey, 1887

24.2 25.5 9.7 11.916 original-B

saint georges basin census, 4th July

1836

21.3 24.5 5.8 7.455 original-B

ulladulla wordlist James Larmer,

1853

220.2 23.3 23.5 31.164 original-B

bateman bay wordlist James

Larmer, 1853

22.4 27.1 9.5 12.1 original-B

wallaga lake wordlist Percy

Mumbla, 1980’s

24.3 21.3 5.6 7.179 original-B

twofold bay census 24th

September 1839

24 214.5 18.5 23.843 original-B

birdhawal wordlist

R.H. Mathews, 1907

25.8 21.7 7.5 9.632 original-B

South Coast Hinterland

berrima census 1st September

1840

216.5 0.9 15.6 22.725 southern-B

coolangatta wordlist William

Bothong, 1900

211.5 5.8 5.7 14.085 southern-B

dhurga wordlist Diana Eades, 1976 20.4 0.1 0.3 0.5099 southern-B

braidwood wordlist James Larmer,

1853

29.1 0.3 8.8 12.663 southern-B

moruya wordlist Herbert Hale,

1846

29.4 2.7 6.7 11.855 southern-B

twofold bay wordlist Johnny

Weyman, 1864

29.1 2.2 6.9 11.63 southern-B

gippsland wordlist Louise Hercus,

1969

28.1 1.2 6.9 10.708 southern-B

wemba wemba wordlist Louise

Hercus, 1969

212.5 8.7 3.8 15.696 southern-B

Blue Mountains

kamilaroi wordlist Rev. W. Ridley,

1875

2.2 24.8 2.6 5.9 original-A

dharginyung wordlist R.H.

Mathews, 1903

2.1 24.9 2.8 6.022 original-A

dharruk wordlist R.H. Mathews,

1901

1.1 22.6 1.5 3.197 original-A

gundungara wordlist M.M. Everitt,

1900

0.8 22.4 1.6 2.993 original-A

gundungara wordlist William

Russell, 1914

10.8 210.8 0� 15.274 original-A

North Coast Hinterland

turruwul wordlist John Malone,

1875

0.3 25.3 5 7.292 B

jervis bay wordlist Joseph

Gaimard, 1826

2 28.9 6.9 11.438 B

�Note: William Russell’s wordlist appears to have come from the boundary between Gundungara (13 . 0) and

Tharumba (13 , 0) zones, hence we find 13 ¼ 0.

Origins of Linguistic Zonation in the Australian Alps 629

Chapter 3, JAS 2005

Page 183

monly called Snell’s Law (Weinstock, 1952). However, the magnitudes of the numbers in

Table 1 are unimportant for present purposes. For now what actually matters are

(1) the signs of the respective numbers 1i, and(2) the empirical ‘brute-fact’ that the 46 historic word/name lists in Appendix 1 all

possess clear signatures placing them in one or other of the basic language families

shown in Table 2, and that

(3) by plotting these signatures from Table 1, according to the respective geographical

locations of each of the 46 historic word/name lists, we discover distinct language

zones extending throughout south-eastern-Australia. Interpolation of associated lin-

guistic boundaries can be manually accomplished, to an accuracy of approximately

+25 km, as in Figure 1.

Of course, hypothetically, there are eight possibile signatures

+�+�+ ¼ 2� 2� 2 ¼ 8

but þ þ þ and 2 2 2 are strictly forbidden by equation (3). Thus, having only six of

these eight signatures, there is the possibility of ‘signature-doubling’ as with the pre-

Wolgalu languages Original-A and Turuwal or the post-Wolgalu languages Midthung

and Tharumba. This follows from the conjecture (Guthrie, 1852), only recently proven

(Appel & Haken, 1977, 1978), that four colours/signatures are sufficient to colour/label any planar map such that no adjoining regions are the same colour/signature.Figure 1 is visually complicated but it has two main areas: the interior and the exterior

of an oblong region in the Australian Alps, which we hereafter call the Huygens Circle.

The four different pre-Wolgalu language zones are clearer in Figure 3(a), whilst the

four different post-Wolgalu language zones within the Huygens Circle are apparent in

Figure 3(b). We will return to this later.

Comparison with Previous Cartography

An instructive map of traditional south-east-Australian tribal boundaries was provided in

Lyall Gillespie’s (1984) book ‘Aborigines of the Canberra Region’. In contrast, one of the

earliest accounts was James Fraser’s (1892) ‘Map of New South Wales, as occupied by the

Table 2. Signatures of the various language families

sgn(1�) Language type Examples

13 . 0 original (pre-Wolgalu)

primary languages

þ 2 þ Original-A Kamilaroi, Dharginyung,

Dharug, Gundungara

22 þ Original-B Wiradjuri, Eora, Wadi-Wadi

Birdhawal, Tasmanian

2 þ þ Southern-B Dhurga, Kurnai, Wemba-Wemba

þ 2 þ B Turuwal

13 , 0

Australian Alps collision

zone (Huygens Circle)

languages

2 þ 2 A ˛arigu, ˛unawalþ þ 2 B Wolgalu

þ 2 2 B Midthung, Tharumba

630 C. R. Illert

Chapter 3, JAS 2005

Page 184

Native Tribes’ supplied in the preface of his 1892 compilation of Rev. L.E. Threlkeld’s

works. Fraser’s map is so wrong, in places, that many scholars still avoid citing it –

though it is clear that they have actually studied this 1892 map, often faithfully parroting

its errors.

Fraser’s map shows the Blue Mountains ‘Original-A’ language, in Zone VIII (‘Kurig-

gai’), correctly linking up with Zone I (‘Kamalarai’). And William Albert Cuneo

(1860–1942), writing in the Picton Post and Advocate newspaper in 1893, even referred

Figure 1. These south-east Australian language zones arise as an empirical ‘brute fact’ from the

respective signatures of the 46 historic word/name lists assembled in Appendix 1. Interpolation

of the above boundaries was accomplished manually to an accuracy of approximately +25 km

Origins of Linguistic Zonation in the Australian Alps 631

Chapter 3, JAS 2005

Page 185

to Burragorang Valley to the south-west of present-day Sydney as ‘the headquarters of the

Kamileroi tribe that claimed sovereignty over a tract of country much larger than the

present county of Camden’ (Illert, 2001). However, contrary to the maps of both Fraser

(1892) and Dixon (1980: 241, map 7), the present study shows that Blue Mountains

language, such as Gundungara and Dharug, never extended eastward to anywhere near

the Sydney coastline where Original-B (‘Eora’) language was spoken. Additionally, Orig-

inal-A and Original-B were apparently so different that an intermediary layer of ‘Turuwul’

pidgin seems to have occurred between coastal and mountain people.

This contrasts with James Kohen’s (1993: 22) now demonstrably false proposition that

‘two dialects of the Dharug language were the Eora dialect along the coast and the inland

dialect . . .’. Even the title of his book asserted that the ‘Dharug’ were ‘the traditional

owners of the Sydney region’. Twenty years previously, Arthur Capell made similar refer-

ence to the existence of ‘coastal Dharug’ people, probably also influenced by the same

error in Fraser’s (1892) map. In fact the ‘Dharug’ seem to have been a mountain tribe

who came from the north and arrived in the Blue Mountains to the west of present-day

Sydney tens of millennia after Original-B speakers, such as the coastal ‘Eora’.

There is no possibility whatsoever of ‘Eora’ being ‘a dialect of Dharug’, and we have

now empirically demonstrated the paucity of phonetic evidence supporting a traditional

‘Dharug’ presence even on the western Sydney plains much beyond the immediate base

of the Blue Mountains. The 22 þ ‘Eora’ signature was ubiquitous about present-day

Sydney, whilst the only other ‘Dharug-like’þ 2 þ signature that I have so far found any-

where near Sydney was that of John Malone’s ‘Turuwul’ pidgin. These present findings

vindicate Anne Ross (1988: 49–52) who challenged contrary views originally proposed

by Capell (1970: 21–22), Dixon (1980: 241, map 7), Kohen & Lampert (1987: 345)

which are apparently still supported by Walsh (1991) and Kohen (1993).

Another common error, one that can also probably be traced back to Fraser’s 1892 map,

follows from his Zone IX (‘Murrin-jari’), which depicted a single strip of ‘coastal language’

extending from Cape Howe to Wollongong. Ignoring the fact that this strip of Original-B

coastal language actually continued on north (well beyond Sydney, through part of

Fraser’s Zone VIII), the 1892 map gave no indication of an adjoining Hinterland strip of

Southern-B language running along the eastern bank of the Shoalhaven River parallel to

the coast. The 1976 map of Eades likewise showed a single strip of ‘Dhurga’ language

along the coast, between Jervis Bay and Wallaga Lake, extending continuously from the

sea to the eastern bank of the Shoalhaven River near Braidwood – an error parroted by

Dixon (1980: map 7, page 241). Our present analysis clearly shows ‘Dhurga’ to have been

a Southern-B hinterland language traditionally spoken at Braidwood, genetically related to

Gippsland ‘Kurnai’ and west-Victorian ‘Wemba Wemba’ but quite different to the ancient

Original-B traditionally spoken along the coastal strip from St Georges Basin to Cape Howe.

In places, Southern-B may possibly have extended eastward almost to the coast: for

example, the 1839 census at Twofold Bay showed an Original-B speaking coastal tribe,

whilst the wordlist of Johnny Weymann collected at Eden Jail on 14 October 1864

was clearly provided by a Southern-B speaker. The Illawarra Mercury, 9 March 1860,

provides the additional information that

Jackey Weyman, an Aboriginal, was placed in the lock-up at this place on Tuesday chargedwith stealing two pounds from a person . . . [and] forwarded this morning to . . . Eden forgreater security. . . . a very intelligent black, speaking English considerably better thanmany Englishmen of our acquaintance . . . [who] having had a reduced quantity of rum com-pared with the share his companions had imbibed, his head was clearer than theirs, and, whilehis victim was sleeping, it is said he picked his pocket.

632 C. R. Illert

Chapter 3, JAS 2005

Page 186

The picture emerges of Aboriginal people dwelling on the fringes of European settlements,

and learning to speak English, whilst still retaining a sense of traditional tribalism.

Weyman, a Southern-B speaker from the hinterland, seems to have had little compunction

over stealing from coastal Aboriginals but probably would not have behaved that way with

his own tribesmen. The fact that his Southern-B wordlist was obtained on the coast prob-

ably is more a reflection of his wanderings than of traditional tribal boundaries.

Although Fraser’s Zones IX and VIII both fail to indicate the parallel adjoining strips of

coastal and hinterland language found in this present study, there is independent evidence

from other primary sources. Baron Field, during his visit to Five Islands (Wollongong) on

22 October 1823, observed that ‘the Aborigines of New Holland were strictly divided into

two classes, the [hinterland] hunters and the [coastal] fishers’. Likewise, Thomas

Kendell, who settled at Narrawallee Creek in 1828, noticed that Shoalhaven Aborigines

were divided into a coastal tribe, and a Pigeon House tribe whose territory extended

inland as far west as Braidwood (Cambage, 1916; Hilder, 1982). Similar observations

were also made by A.W. Howitt (1904).

Despite causing quite major contemporary misunderstandings over traditional tribal

boundaries, Fraser’s 1892 map nevertheless contains some unique and valuable infor-

mation that is unavailable in other sources. For example it gives perhaps the only

correct account of the ‘Tharumba-Midthung’ zone’s traditional northern boundary

(through Picton), extending round to the west of Yass, perhaps all the way to the head-

waters of the River Murray – as suggested in the present study by the 1834 Tuggeranong

census, an 1853 Sutton Forest wordlist, an 1842 Picton census, and also by the 19th

century oral testimony of Alex Murdoch in his legend of ‘the first bushfire’, cited by

C.W. Peck (1925: 56–57), to the effect that

somewhere in the south, perhaps over in Victoria, there lived a great king. His tribe was verynumerous, for he imposed his will upon other tribes and welded them together. . . . He musthave come as far north as the Burrogorang – if, indeed, he did not come further – for theHunter River natives have a story somewhat similar to this

This all meshes nicely with Rev. William Ridley’s (1875: 143) observations that

Tharumba is spoken on the Shoalhaven River . . . by the Wandandian Tribe, Thurawal inanother part of the same district, south of the Illawarra where Wodi Wodi is spoken. Thurawalappears to be the same word as Turrubul and Turuwul, the names of the languages spoken atMoreton Bay and Port Jackson.

Until this present study it was not possible to answer even simple questions such as was

‘Tharumba’ different to ‘Midthung’, were they pidgins, where how and why did they

come into being? And, notwithstanding a brief late account of ‘Tharumba’ grammar by

Mathews (1902), for centuries the total known vocabulary of this language amounted to

a mere six words from Sutton Forest, the names of artefacts and objects scribbled at the

bottom of a drawing by Louisa Atkinson in 1853, supplemented by an additional 28

words recorded in the Shoalhaven in 1872 by Rev. Andrew Mackenzie. Then, recently,

Keith Vincent Smith of Macquarie University (pers. comm.) located Thomas Mitchell’s

96-word ‘Stonequarry’ wordlist and John Dunmore Lang’s 46-word ‘Argyle’ wordlist,

both bearing the distinctive Tharumba-Midthung signature þ 2 2 , thus confirming

the geographical boundaries and producing a fivefold increase in the known lexicon.

Fraser’s Zone X (‘garego’) is approximately correct for the ˛arigu zone (albeit swallow-ing up Walgalu), although to be consistent it should extend northward, along the western

bank of the Shoalhaven River up to Lake George, encompassing the ˛unawal region. Thelinguistic boundaries given by Dixon (1980: map 7, p. 241), likewise wrong by hundreds

Origins of Linguistic Zonation in the Australian Alps 633

Chapter 3, JAS 2005

Page 187

of kilometres, depict an overlarge block of ‘ngarigo’ extending continuously from Omeo

to Picton, completely engulfing and losing the quite distinctive and important ‘Wolgalu’

language and its adjoining ‘Midthung-Tharumba’ pidgin zones. These major blunders

show precisely what can happen when linguists overstress the value of fragmentary gram-

matical accounts for deep linguistic work, brashly claiming (Dixon, 1980: 37) that

. . . in every case I have investigated it is a relatively easy matter to decide whether a particu-lar tribal boundary is a boundary between different languages, or whether it just separates offdialects of one language . . . Even when dealing with long-extinct tribes of the south-east,where only scanty written records are available, there is never any difficulty in decidingwhat was one language . . . as elsewhere in Australia, there is no difficulty in distinguishingbetween language and dialect.

This view is now discredited (Clark, 1990: 3) and, whereas Dixon’s (1980: map 7, p. 241)

linguistic boundaries were largely subjective and drawn only to an accuracy of +200 km

(wrongly showing ˛arigu extending 300 km north all the way to Picton, and Blue Moun-

tains Dharug extending 100 km all the way east to the coast), our Figure 1 is at least an

order of magnitude more precise and based upon a transparent reproducible objective

methodology employing known and readily available data. Figure 1 is also in far better

agreement with traditional tribal boundaries to the extent that they are known (Gillespie,

1984; Tindale, 1974; Howitt, 1904).

The kinds of arguments that might arise from uncertainties in the boundaries in our

Figure 1 are more subtle. For example we have, within the stated error-bounds, drawn

the type-A language zone extending into Wiradjuri territory perhaps 15 km further west

than the signatures in Table 1 might suggest. This is an intentional compensation for sus-

pected skewing of post-European-contact data arising from depletion of traditional inland

Aboriginal communities, and associated large-scale eastward movement of Aboriginal

peoples to the coast, as European farmer settlers increasingly occupied the countryside.

Today a quarter of all Australians who consider themselves to be ‘Aboriginal’ live

along the south-eastern coast, particularly in the Illawarra and the Shoalhaven.

Thus, it is not particularly worrying that Nelly Hamilton (1825–1897), long described as

‘the Queen of Queanbeyan . . . the last remnant of the once numerous and powerful ngunawaltribe’ (Gillespie, 1984: 56), in 1887 supplied the local magistrate with a word-list featuring a

clear Original-B signature; the very same22 þ signature found in Queanbeyan and Goul-

burn censuses circa 1840 and also in the 1904Wiradjuri wordlist collected byR.H.Mathews.

Clearly, by the 1840s,Wiradjuri tribes had been displaced eastward, shunting theACT’s orig-

inal ˛unawal tribe ahead of them toward the coast. This is evidenced by our spectacular new

finding that the [Wandandian] ‘Thurrawal’ wordlist of R.H. Mathews (1903), with signature

2 þ 2 , is not a coastal language at all but, instead, ˛unawal language from Lake George.

Likewise, even as early as 1872, Rev. AndrewMackenzie spoke of ‘Midthung or Tharumba,

spoken by the Aborigines of Braidwood, Ulladulla, Moruya and Jervis Bay’, and R.H.

Mathews (1902) opined ‘the Tharumba language is spoken on the coast of New South

Wales, between the Shoalhaven River and Ulladulla, reaching inland to the Dividing

Range. This tongue is a dialect of Thoorga, spoken to the south of Ulladulla . . .’. Clearly,thesewere all refugees, originally from the SouthernHighlands, largely viaKangarooValley.

In addition, some of the tribal names and geographical locations are misunderstandings.

The [northern] ‘ngarrugu’ wordlist of R.H. Mathews (1908), with signature þ þ 2 , has

actually to be Wolgalu. The [southern] ‘ngarigu’ wordlist of Louise Hercus (1969) is actu-

ally true ˛arigu which, with signature 2 þ 2 , seems to be exactly the same language as

Lake George ˛unawal. The point is that, for the first time, these signatures provide a

rational objective basis upon which finally to decide such matters.

634 C. R. Illert

Chapter 3, JAS 2005

Page 188

Gillieron Waves and Deep Linguistics

The linguist Jules Gillieron (1854–1926), in a 13 volume treatise published between 1902

and 1910, pioneered the study of individual phonological changes in dialects, within a

single language zone, over small distances (tens of kilometres), over very shallow time-

scales (less than a few centuries). His view of linguistic change, today called the wave

model of lexical diffusion, emphasized that every word in a language has its own individual

history, and that sound changes do not operate simultaneously on every word in a language

over very shallow time-scales. More recently Crowley (1992/4/6: 246) has explained that

‘the wave model of linguistic change has gained respectability in modern linguistics throughrecent work on lexical diffusion . . . [it] implies that instead of sharp linguistic splits, changetakes place like waves spreading outward from the place where a stone is dropped into water,travelling different distances with different stones, and crossing with waves caused by otherstones. So, instead of a tree diagram of language relationships . . . [we have] a number ofdifferent dispersal points’.

Interesting as this all may be, there is no reason why it should not be equally valid for deep

linguistics to consider collective transfers between stable phonetic blocks, within scores of

genetically related languages, over trans-continental distances (thousands of kilometres),

over truly deep time-scales (tens of millennia). In such cases, the propagating Gillieron

wavefronts represent the movements and interactions of language super-families giving

rise not just to mere dialect-chains but, instead, to genetically related language-chains.

Accordingly we do not have to view the boundaries in Figure 1 as static empirical facts

devoid of meaning. Instead we can think of them as the end-product of a knowable

dynamic process – Gillieron wave propagation – occurring over deep time-scales.

One key feature of Figure 1 is the coastal strip of Original-B language from Ulladulla to

Cape Howe, which is separated from the larger inland body of Original-B, modern-day

Wiradjuri territory, by an intermediary zone with a Southern-B flavour typically with

word-initial-r and word-initial-l as in Gippsland Kurnai and Dhurga. The obvious dynamic

explanation is that, initially, all language throughout south-eastern Australia was Original-

B and that Southern-B subsequently arose, spreading eastward from central Victoria,

along the RiverMurray to its headwaters in theAustralianAlps, crossing the ‘Menero’ Table-

lands mifi-fiura meaning the middle [of the journey], thereafter following the Shoalhaven

River north, just as theMurray had previously been followed. Throughout this epicmigration,

depicted in Figure 2, Southern-B speakers remained river-people and hinter-landers unable to

out-compete the pre-existingOriginal-B speakerswho, over long periods of time, had already

developed specialized fishing technology for surviving along the coastal strip and also other

specialist skills for surviving in the desert regions of far western New South Wales.

Interestingly, we can even tentatively date these events. The Earth went through a ‘glacial

era’ from 60,000 to 10,000 years ago, the coldest period being from 34,000 to 18,000 years

ago. As this Ice Age drew to a close, the polar ice caps began melting, the world’s oceans

rose by 150 m, Australia’s coastlines moved inland by about 20 km and between 15,000 and

12,000 years ago the present island of Tasmania became separated from mainland Australia

by Bass Strait. Thus, notwithstanding the opinions of Crowley (1992/4/6: 304–5)

‘of course, the relationship of the Tasmanian languages will forever remain a mysterybecause the last fluent speaker of any of these languages died in 1876, and the informationthat was recorded on these languages before then was so poor that it will never be possibleto do anything useful with it in terms of linguistic reconstruction’,

Origins of Linguistic Zonation in the Australian Alps 635

Chapter 3, JAS 2005

Page 189

the clear Original-B signature (see Table 1) on the Tasmanian wordlists of J.P. Gaimard (‘Port

Dalrymple’ ¼ Georgetown, 1826) and J.J.H. de La Billardiere (Research Bay, 1793) immedi-

ately

(1) refutes Wurm’s (1972: 174) view that Tasmanian languages were unrelated to those

on the mainland;

(2) refutes previous notions that modern south-eastern languages might have been bought

to Australia a mere 4000 years ago, by the people who introduced the dingo, as

suggested by Dixon (1980: 19); Crowley (1992: 305) and others; and

(3) means that Original-B language is older than 15,000 years and that, accordingly, our

signatures chronicle truly deep time-scales.

Figure 2. Linguistic map of south-eastern Australia, at the climax of the Ice Age, perhaps 23,000

years ago. Initially all language throughout this region was Original-B which, in the course of

time, gave rise to Southern-B. In this diagram a Gillieron wave-front of Southern-B language

propagates eastward from central Victoria, possibly triggered by drought in the Willandra Lakes

region starting 26,000 years ago, progressively separating coastal and inland zones of Original-B

speakers and displacing some southward into Tasmania, which was at that time connected to the

mainland by a land-bridge across present-day Bass Strait

636 C. R. Illert

Chapter 3, JAS 2005

Page 190

Given this third point, we are able to make an informed guess as to when and why the pro-

posed eastward migration of Southern-B speakers occurred. Mungo is one of a chain of

now-dry lakes along Willandra Creek, to the west of the Lachlan River. Although they

have long been dry, a mere 36,000 years ago they were full of water, fish and shell-

fish, supporting flourishing Aboriginal communities whose remnant camp-sites are still

preserved in the Willandra sand-dunes.

But whatever happened to these flourishing communities of river-landers? Well, about

26,000 years ago, as the Ice Age approached its climax, it became colder and drier throughout

south-eastern Australia. The Willandra Lakes began to dry up and the Southern-B speakers,

understandably seeking water, migrated upstream to the headwaters of the River Murray in

the Australian Alps. Those who remained behind in the vicinity of the Darling River, in sus-

tainable low numbers, had to adapt to the extremely arid conditions that prevailed till about

13,000 years ago. So the obvious proposition is that the eastward-propagating Gillieron

wavefront in Figure 2 chronicles an overland mass migration, of Southern-B speakers,

driven from the Willandra Lakes and Darling River system by drought.1

As it moved, this wavefront progressively separated pre-existing coastal and inland Orig-

inal-B speakers, displacing some southward across the Ice Age land-bridge into Tasmania

(Figure 2). Thewavefront’s progress is chronicled by theOriginal-B speakers, withword-para-

digms similar to mainland Birdhawal and heavily influenced by Southern-B, who first arrived

in Tasmania about 23,000 years ago. For example, Cave Bay Cave, in north-western-

Tasmania, contains their 22,500 year old bone-points and needles used for making fur

cloaks, nets and baskets. Slightly more recent is Kutikina Cave, 100 km further south, the

world’s most southern Ice Age site. Occupied for 5000 years dating from 20,000 years ago,

and containing an estimated ten million artefacts, it is Australia’s richest site (Jones, 1987).

All this suggests that Southern-B language is perhaps 30,000 years old and, as it would have

arisen from Original-B, this means that Original-B must be even older. How much older?

Well, a male human skeleton intentionally buried in a grave at Lake Mungo, and still

showing evidence of ochre ornamentation, has recently been dated byAlan Thorne of theAus-

tralianNational University as 62,000 years old. This estimate is hotly contested by JimBowler

of Melbourne University who argues that it is only 40,000 years old (Grose, 2003).

Whatever the true age happens to be, it is likely that Mungo Man was a speaker of

Original-B, familiar with all of the 62 root words previously published in Illert &

Allison (2004) and most of the various leximorphs published in Illert (2003). He perceived

morning as ‘orbiting-thing [¼sun] rising up’, evening as ‘orbiting-thing [¼sun] going

down’, bivalved shell as ‘opposing pair of gently spiralling things’, and had a quinary

system of numerals based upon recursive arithmetic, all attesting to his cognitive processes

and showing just how like us Mungo People were perhaps 45,000 years ago.

Of course, Original-B speakers would have occupied all of south-eastern Australia at

that time, and not just Lake Mungo. For example coastal ‘Eora’ people from the vicinity

of present-day Sydney were also Original-B speakers, and their 42,000 year old stone arte-

facts have been discovered in gravel beds now buried under sands and silts, at Cranebrook

Terrace in the Penrith/Castlereagh area adjacent to the Nepean River, on the coastal plain

to the west of Sydney.

When the Original-B speakers who made these artefacts first arrived, in the Sydney

region, the landscape and the animals were quite different. Megafauna (animals and

birds up to 3 m tall) still roamed the countryside. There were giant kangaroos such as

the Protemnodon and Procoptodon, rhinoceros-sized grazing creatures such as Diproto-

don, carnivores such as Thylacine (Tasmanian Tiger) and Thylacoleo, and a leopard-

sized tree dweller related to possums. We are only just beginning to understand this

prehistory.

Origins of Linguistic Zonation in the Australian Alps 637

Chapter 3, JAS 2005

Page 191

Huygens’ Principle

Perhaps the oldest signs of human occupation in the Blue Mountains occur west of

present-day Sydney at Kings Table at Wentworth Falls. The location has many axe-

grinding grooves and markings on the walls of a rock shelter on its eastern side. Exca-

vations have shown that this particular shelter was in use 22,000 years ago at the very

height of the coldness and dryness of the ice age, at about the same time that Tasmania

was being colonized. Further south, in the Canberra region, the earliest definite evi-

dence of human occupation is a rock shelter in the grounds of the Birrigai Recreation

Centre at Tidbinbilla dating to 15,000 years ago. Thus, although Original-B speakers

had well and truly occupied south-eastern-Australia 45,000 years or so ago, they pre-

ferred coastal and riverland locations and it was not till about 23,000 years ago that

Aboriginal people started seriously to occupy the rugged mountainous regions that

lay between.

Eugene Stockton (1993/6: 48) has opined

We are never likely to know when Aborigines began to live in the Blue Mountains nor thedirection from which they came. A north or south approach is improbable because ofthe more rugged terrain. The likelihood of an eastern or western approach is affected bythe current debate about the original dispersal of Aborigines in Australia. Those whoargue for an Aboriginal movement around the coast, would imply an eastern approach tothe Blue Mountains, whereas those who argue for inland dispersal would suggest an entryfrom the west.

Yet, given the self-evident nature of Figure 2, let us see if there is an equally sensible

dynamic explanation for the rest of the boundaries of Figure 1 in terms of Gillieron wave-

fronts propagating over deep timescales. The 22,000 years of human occupancy at Kings

Table shelter, and 15,000 years at Tidbinbilla shelter, jointly suggest southward propa-

gation of Original-A language along the Blue Mountains chain itself during and slightly

after the climax of the Ice Age, perhaps all the way from central Queensland, at much

the same time that Southern-B was propagating northward along the Shoalhaven River

as in Figure 3(a). But whereas Original-B and Southern-B were both type-B languages

with similar signatures, it is clear that Original-A, with increasing word-initial usage of

peripheral consonants (11 . 0), was so different from the pre-existing type-B linguistic

substrate in this region that it required a boundary-layer of ‘Turuwul’ pidgin for mutual

comprehensibility.

Perhaps about 15,000 years ago when the wavefronts of Southern-B and Original-A

eventually collided in the Canberra region, creating the Wolgalu language, an encir-

cling boundary-layer of new pidgins (such as Midthung/Tharumba and ˛arigu/˛unawal) exploded into existence. Over millennia this pidgin shockwave expanded

away from the primary Wolgalu collision zone, increasingly as if from a point

source in accordance with Huygens’ Principle (Sears & Zemanski, 1966: 786;

McCarthy, 1968: 34–39), constrained on the east by landscape features such as the

Shoalhaven River as in Figure 3(b).

Most college textbooks on optics and physics provide simple accounts of the principle

of wave propagation which is generally ascribed to Christian Huygens (1629–1695). For

example, Sears & Zemanski (1966: 786) explain that

Huygens’ Principle is a geometrical method for finding, from the known shape of a wavefront at some instant, what the shape will be at some later instant. The Principle states thatevery point of a wave front may be considered the source of small secondary wavelets,

638 C. R. Illert

Chapter 3, JAS 2005

Page 192

which spread out in all directions with a velocity equal to the velocity of the [parent]waves. The new wave front is then found by constructing . . . the envelope of the[se sec-ondary] wave[let]s.

A physical consequence of this principle, one that could perhaps be more often empha-

sized, is that these envelopes asymptotically tend to become circular – eventually

radiating outwards as if from a single point – regardless of the irregular shape of

any actual extended object (such as a stone) which may have initially (by falling

Figure 3a. Linguistic map of Ice Age south-eastern Australia about 22,000 years ago. Although

Tasmania was connected to the mainland at this time, as in Figure 2, only the modern coastline

is shown above. The pre-existing linguistic substrate was Original-B. However a Gillieron

wavefront of Southern-B language propagated eastward from central Victoria, across the

Menero Tablelands then north, separating coastal and inland Original-B speakers.

Original-B and Southern-B were sufficiently similar that no boundary layer of pidgin was

required. However, from south-central Queensland, a Gillieron wavefront of Original-A

surged southward along the Blue Mountains chain. It was linguistically so different, from the

pre-existing type-B substrate, that Original-A required a boundary-layer of Turuwul pidgin for

mutual comprehensibility. In pre-Wolgalu language everywhere we notice that 13 . 0.

Origins of Linguistic Zonation in the Australian Alps 639

Chapter 3, JAS 2005

Page 193

into a pond) caused the disturbance. Thus, in the present context, although the Wolgalu

collision zone is an extended region bounded by two arcs, the surrounding pidgin wave

front (‘Huygens Circle’) became increasingly circular as it expanded outward and

away – over time ‘forgetting’ the irregular shape of the Wolgalu zone that initially

caused it. In the asymptotic limit we might have expected an outgoing circular wave-

front, propagating as if from the centre of mass of the Wolgalu zone, just as Newton’s

laws predict radially symmetric gravitational fields at large distances from massive

objects irrespective of their actual shapes. However, constrained by the north-south

Figure 3b. Linguistic map of post-Wolgalu south-eastern Australia, no more than 15,000 years ago.

By this time the polar ice-caps were melting, the world’s oceans were rising, and Tasmania was

separating from the mainland, producing the familiar modern coastline shown above. The

collision of Southern-B and Original-A wavefronts, in the Australian Alps, created the entirely

new Wolgalu language and a pidgin ‘shockwave’ which expanded outward from the collision-

zone in accordance with the principle of Huygens. All language within this circular shock-wave,

the Huygens Circle, is characterized by 13 , 0. Compare this dynamic model with the

empirically observed static boundaries shown in Figure 1.

640 C. R. Illert

Chapter 3, JAS 2005

Page 194

nature of the Blue Mountains chain and also the Shoalhaven River Gorge to the east,

the expanding pidgin wavefront (which we call the Huygens Circle in Figure 3(b)) was

quite understandably distorted into a sausage shape, extending from Omeo in the south

to Picton in the north (as in Figure 1).

The entire language region subsequently enclosed by this expanding ‘Huygens

Circle’ shockwave, including the primary Wolgalu collision zone itself, is characterized

by signatures with 13 , 0. The original Turuwul pidgin was shunted eastward to

the coast in the Shoalhaven, whilst vanishing entirely along the western side of the

Blue Mountains. From all this we see how Huygens’ Principle rationally explains the

empirically observed language boundaries of Figure 1 and accounts for important dates

independently established by archaeology.

Tree Diagram of Australian Language Super-Families

When the landscape is already occupied by various peoples speaking different languages,

migrating across each-other’s territories, competing for space and interacting in complex

ways that create new languages and pidgins, then one needs a Gillieron wave model.

However when proto-Australian speakers first entered the continent more than 60,000

years ago, it was entirely possible for languages to split and drift apart, into the vast

open spaces of a new continent, free to evolve in isolation with minimal interaction.

Under such conditions the deep-linguistic equivalent of the Brugmann & Leskien

(1870s) postulates (Bynon, 1979: Chapter 4, pp. 173–97) do hold – making a tree

diagram, based upon the signatures in Table 2, an appropriate way to classify the

ancient language super-families.

We note that Turuwul and Original-A both have precisely the same signature, þ 2 þ ,

suggesting that they are daughter-language super-families, which branched at the same

time directly from proto-Australian. This would have been the origin of type-B and

type-A language during the dawn of human occupation of mainland Australia. Following

Turuwul, the next ‘most similar’ type-B signature, 2 2 þ , is that of Original-B. And in

its turn the next ‘most similar’ signature,2 þ þ , is that of Southern-B. All these relation-

ships between pre-Wolgalu language super-families are easily depicted by a tree diagram

(Figure 4).

Making the usual assumption of roughly equal time-intervals (in this case 15,000

years) between successive nodes of our tree, we obtain the respective age-estimates:

Wolgalu (15,000 years), Southern-B (30,000 years), Original-B (45,000 years),

Turuwul and Original-A (60,000 years). These deep linguistic estimates, based upon

the relative similarity of respective signatures, fit rather well with all known evidence

and data. And the tree diagram merges believably into the previously discussed

Gillieron wave description of language change throughout south-eastern Australia

into ‘more recent’ times.

Dating Leximorphs

Example 1. 60,000 years old

When a semantic concept is common to leximorphs from Original-A, Turuwul and

Original-B, then the Tree Diagram suggests that it is extremely ancient – arising from

Origins of Linguistic Zonation in the Australian Alps 641

Chapter 3, JAS 2005

Page 195

proto-Australian perhaps 60,000 years ago. One-such example is ‘hand’ ¼ five-fingers

murU|fflffl{zfflffl}small

� n:dU la� (mu)rU (la)|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}thing�s½five�|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}

fingers½five�¼hand

�����������!NORTHERN REGION

‘marry::diolo:’(‘five’)

Capt: J: Hunter (1795) eora

‘murray::tooloo:’

D: Collins (1798) eora

‘mur::tal:’(‘hand’)

J: Rogers (1899) eora

‘m_aa::tta:ra’(‘hand’)‘mi::tti:’(‘little’)

Rev: L:E: Threlkeld (1834) awabakal

‘mu::dje:ra’(‘hand’)

J:F: Mann (1840s) kuringgai

mvr�oo, (‘baby’ ¼ small thing), greek

minutas, minime, latin

minute, minimal, english

0BBBBBBBBBBBBBB@

1CCCCCCCCCCCCCCA

or

dU (la)|fflfflffl{zfflfflffl}�murUzfflffl}|fflffl{¼small

� n�murU (la)|fflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}¼thing�s½five�

�����������!NORTHERN REGION

Figure 4. The degree of similarity between respective signatures is the basis for this tree diagram,

showing the relatedness of Australian language super-families over deep time scales

642 C. R. Illert

Chapter 3, JAS 2005

Page 196

‘ta:m::mirra’(‘hand’)

‘too:muroo::’(‘short’)

D: Collins (1798) eora

‘tu:mur i^e::’(‘short’)

Lieut: W : Dawes (1790) eora

u �aamno6, (‘bush’ ¼ small thing), greek

‘dhu:m::mar’(‘hand’)

R:H: Mathews (1901) dharug

‘yu:::mree’(‘hand’)

J:F: Mann (1840s) kuringgai

‘:::murra’(‘hand’)

Rev: W : Ridley (1875) kamilaroi

0BBBBBBBBBB@

1CCCCCCCCCCA

ormurU|fflffl{zfflffl}small

� n:dU (la)�murU la|fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}thing�s½five�|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}

fingers½five�¼hand

�����������!CENTRAL REGION

‘murrah:::mull’

‘murru:n::al’

�(‘hand’)

‘murrah:::mur’(‘fingers’)

‘:::mouree’(‘five’)

M: Munima (1863) wadiwadi

‘murrur:::mur’(‘hand’)

L: Malone (1875) wadiwadi

‘mu:n:da:’(‘bush’ ¼ small thing)

‘mara:::mal’(‘hand, finger’)

D:K: Eades (1976) dhurga

‘mara:::mol’(‘doigt’ ¼ finger)

‘mara:::male’(‘main’ ¼ hand)

J:P: Gaimard (1826) turuwul

‘murra:::mul’(‘hand, fingers’)

J: Malone (1875) turuwul

‘:::marrola’(‘hand, fingers’)

‘:::mirra’(‘½left�hand’)‘muroo:n::’(‘oblong berry’)

W : Russell (1914) gundungara

‘:::murra’(‘hand’)

R:H: Mathews (1904) wiradjuri

0BBBBBBBBBBBBBBBBBBBBBB@

1CCCCCCCCCCCCCCCCCCCCCCA

Example 2. 15,000 years old

In contrast, leximorphs restricted to the Huygens Circle are unlikely to be older than about

15,000 years: for example ‘hand’ ¼ five-small-wigglers

murU|fflffl{zfflffl}¼small

� nz}|{�murU la|fflfflfflffl{zfflfflfflffl}

¼five

� ˛u(ru:i)zfflfflfflffl}|fflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{¼wigglers

�����������!HUYGENS CIRCLE

‘moora:n:la:ng:’(‘hand’)

‘mu:n::na:’(‘hand’)

E:M: Curr (1887) ˛unawal

‘mora:::nga:’(‘hand’)

Dr: J: Lhotski (1834) ˛unawal

‘mu:n::na:’(‘hand’)

Rev: A: Mackenzie (1872) midthung=tharumba

‘murra:::ngga:’(‘hand’), R:H: Mathews

‘ngarrigu’ ¼ wolgalu, 1908,

˛unawal, 1904

0BBBBBB@

1CCCCCCA

Origins of Linguistic Zonation in the Australian Alps 643

Chapter 3, JAS 2005

Page 197

Example 3. 23,000 years old

There are also distinctive exclusively-southern leximorphs, which presumably became

established in Southern-B language and diffused into adjoining Original-B coastal

language (such as Birdhawal at Cape Howe) thence, into Tasmania, across the land-

bridge from the mainland approximately 23,000 years ago. One-such example is ‘hand’ ¼either-side five-fingered thing

barU� dU (la)|fflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflffl}¼either�sidething

�mu(rU):n:murU l(a)|fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}¼five�fingered

�����������!SOUTHERN REGION

‘pa::m::mere’(‘un’ ¼ 1½hand�)‘::a:n:me’(‘doigt’ ¼ finger)

J:P: Gaimard (1826) tasmania,

port dalrymple ¼ georgetown

‘bera::::’(‘jou’ ¼ cheeks)

J:P: Gaimard (1826) turuwul

jervis bay, new south wales

‘bre:ty:::’

‘bre:tch:::’

�(‘hand’)

R:H: Mathews (1907) birdhawal

‘pru^rra

^::::mu

^l’(‘hand’)

I: Nathan (1848) illawarra

‘pri:tt:::’(‘hand’)

Rev: F:A: Hagenauer (1876)

kurnai

‘bre:t:::’(‘hand’)

Rev: J: Bulmer (1876) kurnai

‘bre:ty:::’(‘hand’),

brabirrawulung ¼ kurnai

R:H: Mathews (1902)

bracchium, Latin

mpr �aatso, Greekbraziere, French

breast, English

0BBBBBBBBBBBBBBBBBBBBBBBB@

1CCCCCCCCCCCCCCCCCCCCCCCCA

hence

ba(rU)� dU (la:n)|fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl}¼hand�

� bula(la)|fflfflfflffl{zfflfflfflffl}�s½two�

�����������!SOUTHERN REGION

‘ka:te::boueve’(‘deux’ ¼ 2½hands�)‘kar:de::’(‘cinq’ ¼ 5½fingers� ¼ 1½hand�)J:P: Gaimard (1826) tasmania,

port dalrymple ¼ georgetown

0BBB@

1CCCA

Conclusion

It appears that Gillieron wavefronts do describe language propagation by ancient human

populations, who migrated across the land through deep time-scales, creating extended

Original-A and Southern-B regions most clearly shown in Figure 3(a). Over many millen-

nia, contiguous dialects within such zones could well have differentiated into separate

languages – such as Blue Mountains gundungara, dharruk, dharginyung and kamilaroi –

whilst the initial Huygens Circle pidgins evolved into distinctive languages even with

sub-dialects. If so we need to start thinking in terms of extended chains of genetically

related modern languages, instead of mere dialect-chains.

644 C. R. Illert

Chapter 3, JAS 2005

Page 198

At the very least, our dynamic model of Gillieron wave propagation fits the dates estab-

lished independently by archaeology and geography, explaining traditional tribal bound-

aries to the extent that they are known, and we certainly have reason to be reassured by the

fact that the boundaries in Figure 1, based entirely upon empirically obtained phonotactic

signatures, seem also to coincide with the semantic boundaries apparent from ancestral

leximorphs associated with basic nouns such as ‘hand’.

The signatures of the respective language super-families also tell stories. The 2 2 þsignature of Original-B suggests decreasing word-initial usage of peripherals and lamino-

palatals from proto-Australian times (Illert & Allison, 2004). In contrast, Southern-B, with

signature 2 þ þ , had an increasing inventory of lamino-palatals and possible oddities

such as word-initial l and r. The ++ 2 signatures, within Huygens Circle languages,

suggest understandable phonological simplifications in the confusion arising from the col-

lision of Gillieron wave-fronts.

In addition, it is most useful being able to randomly select any historic Aboriginal word-

list or census-form from anywhere in south-eastern Australia and, after a simple frequency

count and signature-calculation, being able to say with confidence what language those

people spoke and where they came from. This methodology is the beginning of rational

scientific thinking in this field.

Acknowledgements

This paper is in support of and in part-fulfilment of the requirements of a PhD by the author at

the University ofWestern Sydney. The author would like to thank the anonymous referee for

valuable suggestions; Bob Hodge of the University of Western Sydney and Andrew Allison

of Adelaide University for valuable feedback and discussions; Keith Vincent Smith of Mac-

quarie University for locating the two historic but previously unknownMidthung/Tharumba

word-lists from ‘Stonequarry’ and ‘Argyle’, aswell as aWadiWadiwordlist from ‘Camden’;

Ed Duyker of Sydney University for supplying the 1793 Tasmanian wordlist; also Jim Smith

for supplying a copy of Mary Everitt’s Gundungara wordlist.

Note

1For an excellent map refer to pages 2 and 3 of the Atlas of New South Wales, edited by R.J. Harriman &

E.S. Clifford (1987), produced by the Central Mapping Authority, Department of Lands, NSW.

References

Appel, K. & Haken, W. (1977) Every planar map is four-colorable, Illinois Journal of Mathematics, 21,

pp. 429–567.

Appel, K. & Haken, W. (1978) In, Mathematics Today, L. A. Stein (Ed.), pp. 153–190 (New York: Springer).

Atkinson, L.� (1854) ‘Plate 29’, Watercolour painting of artefacts from Sutton Forest along with corresponding

Aboriginal names ¼ Midthung-Tharumba wordlist (attributed to Louisa Atkinson by Elizabeth Lawson). In

folio PXA 579, volume 2, f.4., Mitchell Library, Sydney. The painting was prepared for an article alluded to

in the Illustrated Sydney News, 4 February 1854, but never appeared.

Berrima Census� (1 September 1840) Transcribed in M. OrganOrgan (1990) Illawarra and South Coast

Aborigines 1770–1850, Including a Chronicled Bibliography 1770–1990 (Aboriginal Education Unit,

Wollongong University), pp. 266–267.

Bong Bong Census� (15 November 1836) Transcribed in M. Organ (1990) Illawarra and South Coast Aborigines

1770–1850, Including a Chronicled Bibliography 1770–1990 (Aboriginal Education Unit, Wollongong

University), pp. 211–212.

Bothong, W.� (1897–1902) Coolangatta wordlist from Archibald Campbell’s ‘notes on Aborigines’, transcribed

in M. Organ Illawarra and South Coast Aborigines 1770–1850, Including a Chronicled Bibliography

1770–1990 (Aboriginal Education Unit, Wollongong University), pp. 469–472.

Origins of Linguistic Zonation in the Australian Alps 645

Chapter 3, JAS 2005

Page 199

Bowman, J.� (early 1840s) Camden Wordlist, in Macarthur Papers (miscellaneous), A4286, CY Reel 2358,

pp. 15–22. Mitchell Library, Sydney.

Bulmer, Rev. J. (1876) Kurnai wordlist, in: R. B. Smythe (Ed.) The Aborigines of Victoria (Melbourne:

Government Printer).

Bynon, T. (1979) Historical Linguistics, Cambridge Textbooks in Linguistics (Cambridge, UK: Cambridge

University Press).

Cambage, R. H. (1916) Captain Cook’s Pigeon House, and the Early South Coast Exploration, 24 pp (Sydney:

Samuel E. Lees).

Cambage, R. H. (1921) Exploration between the Wingecarribee, Shoalhaven, Macquarie and Murrumbidgee

Rivers, Journal and Proceedings of the Royal Historical Society, 7(5), pp. 217–287.

Capell, A. (1970) Aboriginal languages in the south central coast, new discoveries, Oceania, 41, pp. 20–27.

Clark, I. D. (1990) Aboriginal Languages and Clans: an Historical Atlas of Western and Central Victoria 1800–

1900, Monash Publications in Geography, number 37, page 3.

Collins, D. (1798) An Account of the English Colony in New South Wales, 2 volumes (London, UK: T. Cadell &

W. Davies).

Crowley, T. (1992/3/4/6) An Introduction to Historical Linguistics (Oxford, UK: Oxford University Press).

Crowley, T. (1997) Chipping away at the past, a northern New South Wales perspective, in: P. McConvell &

N. Evans (Eds) Archaeology and Linguistics, pp. 275–295 (Oxford University Press).

Curr, E. M.� (1886) The Australian Race, four volumes (Melbourne, Australia: Government Printer).

Dawes, W. (1790) Manuscripts (Marsden Collection, School of Oriental and African Studies, London).

Dawsey, R.� (1886) Wordlist # 194, from Jervis Bay to Mount Dromedary, in: E. M. Curr (Ed.) The Australian

Race, vol. 3, pp. 420–421 (Melbourne, Australia: Government Printer).

Dixon, R. M. W. (1980) The Languages of Australia (Cambridge, UK: Cambridge University Press).

Duyker, E. (2003) Citizen Labillardiere (Australia: Melbourne University Press).

Eades, D. K.� (1976) The Dharawal and Dhurga Languages of the New South Wales South Coast (Canberra:

Aboriginal Studies Press). Dhurga wordlist pp. 74–89.

Everitt, M. M.� (1901) PartialGundungara wordlist in the collected papers of R.H. Mathews (Private Collection).

Field, B. (1825) Geographical Memoirs on New South Wales (London: John Murray).

Flood, J. (1990) The Riches of Ancient Australia, an Indispensable Guide for Exploring Prehistoric Australia

(Australian Heritage Commission, University of Queensland Press).

Fraser, J. (Ed.) (1892) An Australian Language as Spoken by the Awabakal, the People of Awaba or Lake Mac-

quarie . . . Being an Account of their Language, Traditions, and Customs: by L. E. Threlkeld, Rearranged,

Condensed And Edited by John Fraser (Sydney, Australia: Government Printer).

Gaimard, J. P.� (1826) In M. J. Dumont D’Urville (Ed.) (1833), Voyage de decouvertes de l’Astrolabe 1826-9

(Paris, France: Ministere de la Marine). Port Dalrymple (north Tasmanian) wordlist in the ‘Philologie’

volume, pp. 1–30.

Gillespie, L. L. (1984) Aborigines of the Canberra Region (Canberra: Canberra Publishing).

Gillieron, J. (1902–1910) Atlas linguistique de la France, 13 vols (Paris).

Grose, S. (2003) ‘Mungo Jumbo’, Australasian Science, 24(3), pp. 18–21.

Hagenauer, F. A. (1876) Wordlist, in: R. B. Smythe (Ed.) The Aborigines of Victoria, vol 2 (Melbourne: Govern-

ment Printer).

Hale, H. (1846) The languages of Australia, in:Ethnology and Philology, pp. 479–531 (volume vi ofReports of the

United States Exploring Expedition, under the command of Charles Wilkes, Philadelphia).

Hamilton, N.� (1887) Queanbeyan wordlist, in: E. M. Curr (1886) The Australian Race, pp. 424–425.

(Melbourne, Australia: Government Printer).

Hercus, L.� (1969) The Languages of Victoria, a Late Survey, 2 vols (Canberra: Australian Institute for

Aboriginal Studies). Southern Ngarigu wordlist in vol. 2, pp. 388–402. Gippsland Kurnai wordlist in

vol. 2, pp. 381–387. Wemba-Wemba wordlist in vol. 2, pp. 213–277.

Hilder, W. (1982) In Pigeon House and Beyond, a Guide to the Budawang Range and Environs, edited by

C. Watson (Sydney: The Budawang Committee).

Howitt, A. W. (1904) The Native Tribes Of South-east Australia (London: Macmillan).

Hunter, J. (1793) An Historical Journal of the Transactions at Port Jackson and Norfolk Island (London, UK:

J. Stockdale Co).

Illert, C. (2001) The centenary of Mary Everitt’s Gundungara grammar, Journal and Proceedings of the Royal

Society of N.S.W., 134, pp. 19–44.

Illert, C. (2003a) Lexigenesis in ancestral south-east-Australian Aboriginal language, Journal of Applied

Statistics, 30(2), pp. 113–143.

Illert, C. (2003b) Early Ancestry of Illawarra’s Wadi Wadi People, Part 1 (Wollongong, Australia: Northern

Illawarra Aboriginal Collective).

646 C. R. Illert

Chapter 3, JAS 2005

Page 200

Illert, C. & Allison, A.� (2004) Phono-genesis and the origin of accusative syntax in proto-Australian language,

Journal of Applied Statistics, 31(1), pp. 73–104.

Jones, R. (1987) Ice Age hunters of the Tasmanian wilderness, Australian Geographic, 8, pp. 26–45

(plus map).

Kohen, J. L. (1993) The Darug and their Neighbours, the Traditional Aboriginal Owners of the Sydney Region

(Sydney, Australia: Blacktown Historical Society).

Kohen, J. L. & Lampert, R. (1987) Hunters and fishers in the Sydney region, in: D. J. Mulvaney & J. P. White

(Eds) Australians to 1788, pp. 343–365 (Sydney, Australia: Fairfax Syme & Weldon Associates).

Labillardiere, J. J. H.� (1800) Voyage in Search of La Perouse, Performed by Order of The Constituent

Assembly, during the years 1791-4 . . . (translated from French and published in English by John

Stockdale, London, 1800). The 1793 Research Bay (southern Tasmanian) wordlist is in the Appendix,

pp. 19–20.

Lang, J. D.� (1850s) Argyle wordlist in Papers 1838–73, volume 9, MS A2229, pages 36-7, CY Reel 900

(Sydney: Mitchell Library).

Larmer, J.� (1898a) Native vocabulary of miscellaneous New South Wales objects, Journal of the Royal Society

of N.S.W., 32, pp. 223–229.

Larmer, J. (1898b) Aboriginal words and meanings, Science of Man, 2(8), pp. 146–148.

Lhotsky, J. (1839) Some remarks on a short vocabulary of the natives of Van Diemen’s Land and also of the

Monaro Downs in Australia, Journal of the Royal Geographical Society (London), 9, pp. 157–162.

MacAlister, C. (1907) Old Pioneering Days in the Sunny South (Sydney, Australia).

McCarthy, I. E. (1968) Introduction to Nuclear Theory (USA: Wiley). See pp. 34–39 for a derivation of Green’s

Functions appropriate for describing Huygen’s Principle.

Mackenzie, A.� (1874) Thurrawal [actually Wodi Wodi] wordlist and Midthung/Tharumba wordlist, Journal of

the Anthropological Institute of Great Britain and Ireland, 3, pp. 254–255.

Malone, J.� (1875) Turuwul wordlist in: W. Ridley, Kamilaroi and other Australian Languages, pp. 99–101

(Sydney, Australia: Government Printer).

Malone, L.� (1875)Wodi Wodi wordlist, in: W. Ridley, Kamilaroi and other Australian Languages, pp. 111–114

(Sydney, Australia: Government Printer).

Mathews, R. H.� (1901) Dharruk wordlist, Journal and Proceedings of the Royal Society of N.S.W., 35,

pp. 157–160.

Mathews, R. H. (1902) The Tharumba language, Queensland Geographical Journal, 18, pp. 58–61.

Mathews, R. H.� (1902) Brabirrawulung [actually Kurnai] wordlist, Journal and Proceedings of the Royal

Society of N.S.W., 36, pp. 97–106.

Mathews, R. H.� (1903a) Thurrawal [actually Ngunawal] wordlist, Journal of the Anthropological Institute of

Great Britain & Ireland, 33, pp. 275–279.

Mathews, R. H.� (1903b) Darkinyung wordlist, Journal of the Anthropological Institute of Great Britain &

Ireland, 33, pp. 280–281.

Mathews, R. H.� (1904a)Wiradyuri wordlist, Journal of the Anthropological Institute of Great Britain & Ireland,

34(2), pp. 299–302.

Mathews, R. H.� (1904b)Ngunawal wordlist, Journal of the Anthropological Institute of Great Britain & Ireland,

34(2), pp. 302–305.

Mathews, R. H.� (1907) Birdhawal wordlist, Proceedings of the American Philosophical Society, 46,

pp. 354–357.

Mathews, R. H.� (1908) Ngarrugu [actually Wolgalu] wordlist, Journal and Proceedings of the Royal Society of

N.S.W., 42, pp. 335–340.

Mitchell, T.� (1830s) Stonequarry (‘Wollondilly River’) wordlist, in Mitchell Papers, MS A295-3, p. 425

(Sydney: Mitchell Library).

Mumbler, P.� (1980s) In L. Chittick & T. Fox (1997) Travelling with Percy, a South Coast Journey, pp. 197–199

(Canberra: Aboriginal Studies Press).

Munima, M. (1899) Illawarra wordlist obtained in 1863, Science of Man, 2(8), pp. 141–142.

Nathan, I. (1848) The Southern Euphrosyne and Australian Miscellany (London, UK: Whittacker & Co). Word-

list on pp. 134–135.

Organ, M.� (1990) Illawarra and South Coast Aborigines 1770–1850, Including a Chronicled Bibliography

1770–1990 (Aboriginal Education Unit, Wollongong University).

Osmond, M.� (1989) A Reconstruction of the Phonology, Morphology and Syntax of the Sydney Language as

Recorded in the Notebooks of William Dawes (unpublished student paper, Department of Linguistics,

Australian National University, Canberra, Australia).

Peck, C. W. (1925) Australian Legends, Tales Handed Down from the Remotest Times (Melbourne: Lothian

Publishing Co). The story of ‘the first bushfire’, pp. 56–57.

Origins of Linguistic Zonation in the Australian Alps 647

Chapter 3, JAS 2005

Page 201

Picton Census� (6 August 1842) New South Wales State Archives Office, reel 3706, frame 0445 (‘page 204’) and

frame 0446 (‘page 205’).

Queanbeyan Census� (5 May 1841) Transcribed, in: M. Young (Ed.) (2000) The Aboriginal People of the

Monaro, pp. 133–135 (New South Wales: National Parks & Wildlife Service).

Ridley, W.� (1875) Kamilaroi and other Australian Languages (Sydney, Australia: Government Printer).

Kamilaroi wordlist on pp. 18–37.

Ridley, W.� (1878) Journal of the Anthropological Institute of Great Britain & Ireland, 8(3), pp. 263–265.

Ross, A. (1988) Tribal and linguistic boundaries, a reassessment of the evidence, in: G. Alpin (Ed.) Sydney before

Macquarie, a Difficult Infant, pp. 42–53 (Sydney: NSW University Press).

Rowley, J.� (1878) Cowpastures wordlist, in: W. Ridley, Journal of the Anthropological Institute of Great Britain

& Ireland, 8(3), pp. 263–265.

Saint George’s Basin Census� (Erowal Tribe, Currumbene Creek) (4 July 1836) Transcribed in M. Organ (1990)

Illawarra and South Coast Aborigines 1770–1850, Including a Chronicled Bibliography 1770–1990,

pp. 203–204 (Aboriginal Education Unit, Wollongong University).

Russell, W.� (1914) Gundungorra wordlist, in:My Recollections by William Russell, pp. 25–27 (Camden, NSW:

A.L. Bennett).

Sears, F. W. & Zemanski, M.W. (1966)College Physics, 3rd edn (AddisonWesley). See Section 39-6, ‘Huygens’

Principle’, on p. 786.

Smith, K. V. (2003) 1770 the Endeavour lists, forgotten words from Botany Bay, AQ Journal of Contemporary

Analysis, 75, pp. 32–40.

Smythe, R. B. (1876) The Aborigines of Victoria (Melbourne: Government Printer).

Stockton, E. (1993/6) Blue Mountains Dreaming (Katoomba: Three Sisters Publications).

Threlkeld, L. E. (1830s) Collected works edited by J. Fraser (1892) An Australian Language as Spoken by the

Awabakal, the People of Awaba or Lake Macquarie . . . Being an Account of their Language, Traditions,

and Customs: by L.E. Threlkeld, Rearranged, Condensed And Edited by John Fraser (Sydney, Australia:

Government Printer).

Tindale, N. (1974) Aboriginal tribes of Australia (Berkeley and Los Angeles, USA: University of California

Press).

Troy, J.� (1993) The Sydney Language (Canberra: Aboriginal Studies Press).

Tuggeranong Census� (‘Maneroo Tribe at Janevale Homestead’) (June 1834) New South Wales State Archives

Office 4/2219.1. Transcribed in M. Young (Ed.) (2000) The Aboriginal People of the Monaro, p. 75 (New

South Wales: National Parks & Wildlife Service).

Twofold Bay Census� (24 September 1839) Transcribed in M. Organ (1990) Illawarra and South Coast Abori-

gines 1770–1850, Including a Chronicled Bibliography 1770–1990, p. 247 (Aboriginal Education Unit,

Wollongong University).

Walsh, M. (1991) Overview of indigenous languages of Australia, in: S. Romaine (Ed.) Language in Australia,

pp. 27–48 (Cambridge, UK: Cambridge University Press).

Weinstock, R. (1952/74) Calculus of Variations with Applications to Physics and Engineering, 2nd edn

(New York: Dover Publications). See ‘Geometrical optics and Fermat’s principle’, pp. 67–71.

Weyman, J.� (1875) Twofold Bay Wordlist, in: W. Ridley, Kamilaroi and other Australian Languages, p. 115

(Sydney, Australia: Government Printer).

Wollongong Census� (21 May 1834) Transcribed, in: M. Organ (1990) Illawarra and South Coast Aborigines

1770–1850, Including a Chronicled Bibliography 1770–1990, pp. 186–187 (Aboriginal Education Unit,

Wollongong University).

Wurm, S. A. (1972) Languages of Australia and Tasmania (The Hague: Mouton. Janua Linguarum. Series Critica

1). On page 174 he claims that extinct Tasmanian languages are genetically unrelated to those of mainland

Australia.

Young, M.� (Ed.) (2000) The Aboriginal People of the Monaro (New South Wales: National Parks & Wildlife

Service).

Note: Those of the above references marked with � are the actual word/name lists cited in

Appendix 1 of the present analysis.

Appendix 1

The relative frequencies of word-initial consonants within proto-Australian have pre-

viously been published and discussed by the author. These present tables – quantifying

the situation for numerous ‘modern’ south-east Australian languages – were painstakingly

648 C. R. Illert

Chapter 3, JAS 2005

Page 202

compiled from 37 historic word-lists collectively containing 7601 Aboriginal words, and

nine early census lists (‘Blanket Registers’) collectively containing 445 traditional names,

on the assumption that phonotactic frequencies in traditional Aboriginal languages would

have been much the same for proper names as for words in natural language. Despite the

repetitive larger wordlists recorded in a few language zones near major centres of Euro-

pean settlement, the more remote languages were poorly documented throughout entire

geographical regions. For example, the Sutton Forest wordlist has only six words, but it

is used because we have nothing else from there and are grateful to have any data at all

from the Midthung/Tharumba region. One can see why it is necessary to try to supplement

wordlists with census data that are often the only written information from most geo-

graphical locations in early times. Even so, notwithstanding the fact that eight of the

lists cited here have fewer than 30 words or names, it is good to actually have them –

and their signatures do seem to be sensible, suggesting robustness of our analysis even

with small sample sizes. Manually counting each word or name was a laborious task,

with the author’s judgement frequently being exercised when old handwriting proved dif-

ficult to read. Words beginning with fricatives or vowels, or digraphs such as QU or CH or

SH, were generally ignored. The respective geographical categories used herein – e.g.

‘Western Inland Region’, ‘Australian Alps’, ‘Blue Mountains Region’, ‘Coastal Hinter-

land’ and ‘Eastern Coastal Region’ etc – are only intended to be meaningful in relation

to the study area as defined in Figure 1. In these tables, N is the total number of words/names with any specified initial-consonant in each different source. The corresponding

lexical frequencies are also expressed as approximate percentages, rounded to one

decimal place, subject to the constraint of normalization which is essential if the different

data sets are to be compared with each other on an equal footing. Rounding the frequencies

(either up or down) to one decimal place precision, whilst requiring normalization of all

data sets, can produce odd results as, for example, in the Sutton Forest wordlist, where

three words out of six have to be taken as 49.9% (instead of 50%) so that each of the

other three words can have frequencies of 16.7% (instead of 16.6666666%).

Origins of Linguistic Zonation in the Australian Alps 649

Chapter 3, JAS 2005

Page 203

List of Symbols for the typesetter

Greek: ε , ε1 , ε 2 , ε 3 , ε i , α1 , α2 , α3 , α i , λ , ψ , Θ

Roman: F* , f* , f* , *f& , T , t , t , s , N , H

phonetic: ï, ó, ê, ã, Ä, Ö, Ï=,==å=, �Ã==, þ=, ä=, äl=== , ý

�

==========Ç=,==Ç=, Çl==, ‚==, =~=,==ì=, =á=,==r

mathematical:0

Tdt

− ∫ KKK= ,

*

Θ f

∂∂ &

, *∑ K=K=K , Δ f*

= ======= , < , > , = , + , - , ± , × where the subscript * denotes individual phonetic symbols as listed above and specified in the text.

Chapter 4, JAS 2006

Page 214

Origins of Linguistic Zonation in the

Australian Alps. Part 2 – Snell’s Law1

CHRISTOPHER R. ILLERT

University of Western Sydney, New South Wales, Australia

ABSTRACT In this second paper, analysing archival SE-Australian Aboriginal word/name lists,Snell’s Law is used to deduce the likely minimal sound-systems of pre Ice-Age languagesuperfamilies – some probably dating back beyond the first occupation of Australia by humans.The deduced ‘Turuwal-like’ ancestral sound-system is then used as a basis for reconstructingdeictic forms apparently so ancient that they seem to even unify ‘PamaNyungan’ and ‘non-PamaNyungan’ language within a single system of formal logic which, having apparentlyprovided the semantic basis for at least 60,000 years of speech throughout the entire Australiancontinent, deserves to be called proto-Australian regardless of whether or not it arose in SE-Asiatens of millennia before. Whatever the exact age of this reconstructed proto-Australian, presentedhere for the first time, it is an order of magnitude older than any known human language and, assuch, a ‘Rosetta Stone’ for human languages worldwide. It also provides an unprecedentedwindow into human consciousness and perception of the world up to 75,000 years ago, which isespecially significant given that humans can only have engaged in finely controlled speech andfully modern language since chance mutation of our FOXP2 gene about 120,000 years ago.These truly ancient deictic forms dating halfway back to the beginning of modern human speech,retrieved only through modern statistical analysis, provide insight into our very origins and assuch are perhaps amongst the most precious cultural treasures that humanity currently possesses.

KEY WORDS: Phonotactic signatures, archaeo-linguistics, proto-Australian, Snell’s Law

Snell’s Law across Phonological Boundaries in South-East Australia

Linguists have long needed a practical and objective method for calculating exactly how

many, and what kinds of, consonants existed in the respective languages represented by

data-samples of the kind appearing in Appendix 1. Every researcher in this field for two

centuries seems to have put forward some or other personally favoured alphabet

(Dawes, 1790; Mathews, 1900’s, Everitt, 1900; Eades, 1976; Osmond, 1989; Troy,

1993) – in recent years based upon no new evidence, the last native-speakers having

died generations ago. However the ‘lexical-transfer’ signature is indeed new. It is actually

a precision tool drawing upon archival sources including censuses and wordlists, some

containing as few as six lexical items, which generations of researchers have considered

informationally worthless. Yet, in Tables 1 and 2, clear entropy maximising signatures

Journal of Applied Statistics

Vol. 33, No. 9, 989–1030, November 2006

1Continued from Journal of Applied Statistics 32(6), pp. 625–659 (August 2005)

Correspondence Address: Christopher R. Illert, 2/3 Birch Crescent, East Corrimal, NSW 2518, Australia.

Email: illert@1earth.net

0266-4763 Print=1360-0532 Online=06=90989–42 # 2006 Taylor & FrancisDOI: 10.1080=02664760500450160

Chapter 4, JAS 2006

Page 215

seem to shine through even from the most decrepit handwritten documents dating back

centuries.

To harness the statistical power of this phonotactic information we need only realise that

when, in the course of time, an Aboriginal language increased its phonology by one more

consonant (capable of appearing word-initially) it altered the trajectory of the lexical-

transfer vector (equation (1) – see Part 1 of this paper) just as a light-ray is refracted

when crossing the boundary between two contiguous blocks of glass of differing densities.

Opticians call this Snell’s Law (see Weinstock, 1974, pp. 67–70) and, in the present

context, it requires us to minimise equation (5) subject to the constraint (3). As a first

step we combine the two expressions into a single Lagrangian term:

Q ¼ 1� l(11 þ 12 þ 13) (6)

based upon three stable blocks of potential word-initial consonants, such that

@Q

@_f m¼ @Q

@_f g¼ @Q

@_f ˛¼ @Q

@_f b¼ T11

1� lT

@Q

@_f _¼ @Q

@_f do¼ @Q

@_f lo¼ @Q

@_f y¼ T12

1� lT

@Q

@_f w¼ @Q

@_f n¼ @Q

@_f d¼ @Q

@_f n¼ @Q

@_f d¼ @Q

@_f l¼ @Q

@_f r¼ @Q

@_f n¼ @Q

@_f d¼ @Q

@_f l¼ T13

1� lT (7)

However, as these consonants do not all occur in every language, real-world minimisation

requires:

0 ¼X�

@Q

@_f �

¼ T

1(a111 þ a212 þ a313)� (a1 þ a2 þ a3)lT (8)

for integers ai representing the respective number of realised consonants (i.e. those actu-

ally occurring word-initially) from each of the three stable lexical blocks: a1 being the

number of peripherals, a2 the number of lamino-palatals, and a3 the number of others.

Choosing a Lagrange Multiplier l ¼ 1/(a1 þa2 þa3) ; 1/H, the reciprocal of the

total number of realised word-initial consonants, gives:

1 ¼ a111 þ a212 þ a313

¼ (a1 � a2)11 þ (a3 � a2)13 ¼ t11 þ s13 (9)

for integers t ¼ a12 a2 and s ¼ a32 a2, this last simplification following because of

equation (3).

Additionally if we select a1 ¼ 4 and a2 ¼ 3, as seems to be the case with Original-B

and Southern-B language, then t ¼ 1 giving the relationship:

1 ¼ 11 þ s13 (10)

990 C. R. Illert

Chapter 4, JAS 2006

Page 216

which, subject to equation (3), has an exact ‘type-A’ theoretical solution

11 ¼ s� 1þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(s� 1)2 þ s2 � 2

q� �13

12 ¼ � sþffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(s� 1)2 þ s2 � 2

q� �13

13 ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi6s(s� 1)þ (4s� 2)

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(s� 1)2 þ s2 � 2

pq(11)

Figure 5. A family of pre-Wolgalu “type-B” signatures given by equations (12), for different values

of s . (1þ ffiffiffi3

p)=2 � 1:366, assuming that a1 ¼ 4 and a2 ¼ 3, hence t ¼ 1. As phonologies require

discrete numbers of consonants we are mainly interested in those values of s that are integers. Clearly

Southern-B evolved from Original-B, increasing its phonology by a single non-peripheral non-

laminopalatal consonant (i.e. s increased from 2 to 3), in the process changing signature from

22þ to 2þþ. Subsequently, at eastern and western extremities of the Southern-B zone,

Bothong’s Coolangatta and Hercus’ Wemba-Wemba wordlists both (with 12/1 . 13/1) are better

described by assuming s ¼ 5, hence a3 ¼ 8 suggesting phonologies with two additional

consonants (probably retroflexives absorbed from adjoining Original-A zones). The actual age of

proto-Australian T cancels out of equation (8) but from the Tree Diagram, assuming a linear

relationship between s and past time, the age of each signature is estimated above

Origins of Linguistic Zonation in the Australian Alps, Part 2 991

Chapter 4, JAS 2006

Page 217

Figure 6. A family of pre-Wolgalu ‘type-A’ signatures given by equations (16), for different values

of s .ffiffiffiffiffiffiffiffi3=2

p � 1:225, assuming that a1 ¼ a2 ¼ 4, hence t ¼ 0. As might be expected from Table 5,

the special case s ¼ 2 corresponds to the distinctive Original-A signature given by equations 18.

Also, because a2 ¼ 4, it follows that Original-A had word-initial-ly (denoted here by lO), whichtends to occur in Australian languages along with the retroflexives rd, rn, rl (respectively denoted

%, �, �) originally from the continent’s north-western ‘non-Pama-Nyungen’ language zone,

subsequently in the Blue Mountains (see Figure 9), and also in western-Victoria (judging by the

large consonant-inventory of Wemba-Wemba in Figure 5)

Figure 7. A family of ‘type-B’ signatures given by equation (19), for different values of s . 0,

assuming that a1 ¼ 4 and a2 ¼ 2, hence t ¼ 2. As expected from Table 5, the special case s ¼ 1

gives the distinctive Turuwal signature given by equations (20) belonging to Turuwal. Table 3

even provides evidence of a more recent Turuwal, barely distinguishable from Original-B, with

signature given by s ¼ 2 in equation (19)

992 C. R. Illert

Chapter 4, JAS 2006

Page 218

and also an exact ‘type-B’ theoretical solution

11 ¼ s� 1�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(s� 1)2 þ s2 � 2

q� �13

12 ¼ � s�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(s� 1)2 þ s2 � 2

q� �13

13 ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi6s(s� 1)� (4s� 2)

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(s� 1)2 þ s2 � 2

pq(12)

where the arbitrary constant 1 is +, determining sgn (13), according to whether we are

considering pre or post Wolgalu language. By choosing the integer s, so as to best fit

one or other of equations (11) and (12) to the experimental data in Table 1, we can

obtain a reliable estimate for a3 ¼ sþ a2 ¼ sþ 3 which otherwise can only be guessed

at due to difficulties judging where the tail of the entropy-maximising ‘power-law’ actu-

ally truncates (Illert & Allison 2004, Figures 2 and 3).

Example 1. Original-B: a1 ¼ 4, a2 ¼ 3, a3 ¼ 5

(Five non-peripheral, non-lamino-palatal consonants capable of appearing word-initially)

For pre-Wolgalu language, substituting s ¼ 2 into equations (12) gives theoretical

values for the respective components of a ‘type-B’ unit-vector

111¼ 1� ffiffiffi

3p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi12� 6

ffiffiffi3

pp ¼ �0:578

121¼ �2þ ffiffiffi

3p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi12� 6

ffiffiffi3

pp ¼ �0:211

131¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

12� 6ffiffiffi3

pp ¼ 0:789

(13)

which may be compared with ‘experimentally observed’ values, as presented in Table 3,

calculated from the respective Original-B word/name-list signatures assembled in

Table 1.

Figure 8. The location in ts-space of pre-Wolgalu language superfamilies satisfying equation (9).

Decreasing t implies increasing a2 whilst increasing s implies increasing a3

Origins of Linguistic Zonation in the Australian Alps, Part 2 993

Chapter 4, JAS 2006

Page 219

Example 2. Southern-B: a1 ¼ 4, a2 ¼ 3, a3 ¼ 6

(Six non-peripheral, non-lamino-palatal consonants capable of appearing word-initially)

For pre-Wolgalu language, substituting s ¼ 3 into equations (12) gives theoretical

values for the respective components of a ‘type-B’ unit-vector

111¼ 2� ffiffiffiffiffi

11p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi36� 10

ffiffiffiffiffi11

pp ¼ �0:782

121¼ �3þ ffiffiffiffiffi

11p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi36� 10

ffiffiffiffiffi11

pp ¼ 0:188

131¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

36� 10ffiffiffiffiffi11

pp ¼ 0:594

(14)

which may be compared with ‘experimentally observed’ values, as presented in Table 4,

calculated from the respective Southern-B word/name-list signatures assembled in

Table 1.

Figure 9. As early as 1919 Wilhelm Schmidt was able to distinguish between Australia’s ‘northern’

and ‘southern’ language zones on the basis of word-final-sounds. Subsequent syntactic studies

amended Schmidt’s original boundary and named the much larger southern language zone ‘Pama-

Nyungan’. Our present study suggests that the ‘prefixing’ ‘non-Pama-Nyungan’ language zone, in

the continent’s north-west, may have been the heartland of the Original-A, which surged

southward along the Blue Mountains chain at the height of the Ice Age about 22,500 years ago

994 C. R. Illert

Chapter 4, JAS 2006

Page 220

Our assumption of four peripheral consonants capable of occurring word-initially

(a1 ¼ 4) seems well founded in all traditional language throughout south-eastern-Austra-

lia and, indeed, even in ancestral proto-Australian. However, the Turuwal and Original-A

language-zones in our study area seem to have possessed phonologies with unusual

numbers of lamino-palatal consonants capable of appearing word-initially (within the

range a2 ¼ 1, 2, 3, 4) as permitted by the general equation (9) which only simplifies to

equation (10) in the special case a2 ¼ 3.

We can usefully rearrange equation (9) as follows:

s ;1� (a1 � a2)11

13(15)

enabling us to use the experimental data from Table 1 to estimate s, corresponding to

hypothetical values of a2, as in Table 5. Clearly, the ‘experimentally observed’ Orig-

inal-A data sets nearly give a positive integer value for s when we assume four lamino-

palatals capable of occurring word-initially (a2 ¼ 4). In contrast, the ‘experimentally

observed’ Turuwal data sets nearly give a positive integer value for s when we assume

one or two lamino-palatals capable of occurring word-initially (a2 ¼ 1 or 2).

Figure 10. We can model the ‘Pama-Nyungan’ signature transition Turuwal ! Original-

B ! Southern-B by superimposing Figures 5 and 7. The Turuwal trajectories 1i/1 given by

equations (19) are sensible for positive values s . 0. Also for s . 1.3 the trajectories given by

equations (12) can be viewed as perturbations upon ancestral Turuwal trajectories (dashed curves)

following a causal event, about 53,000 years ago (corresponding to s � 1.3), when t decreased

from 2 to 1 (i.e. a2 increased from 2 to 3)

Origins of Linguistic Zonation in the Australian Alps, Part 2 995

Chapter 4, JAS 2006

Page 221

In the event that a1 ¼ a2 ¼ 4 then t ¼ 0 and equation (9), subject to equation (3), has an

exact ‘type-A’ theoretical solution:

11 ¼ 0:5� �1þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2s2 � 3

p� �13

12 ¼ �0:5� 1þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2s2 � 3

p� �13

13 ¼ 1

s

(16)

and also an exact ‘type-B’ theoretical solution

11 ¼ 0:5� �1�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2s2 � 3

p� �13

12 ¼ �0:5� 1�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2s2 � 3

p� �13

13 ¼ 1

s

(17)

where the arbitrary constant 1 is +, determining sgn (13), according to whether we are

considering pre or post-Wolgalu language.

Table 3. Experimental values for Original-B word/name lists, calculated from the data in Table 1.

Compare these with the expected theoretical values (equation (13))

11/1 12/1 13/1

R.H. Mathews (1904) 20.458 20.356 0.814

Nelly Hamilton (1887) 20.527 20.277 0.804

queanbeyan census (1841) 20.472 20.341 0.813

goulburn census(1837) 20.408 20.408 0.816

Dawes/Osmond (1790) 20.595 20.186 0.781

Jakelyn Troy (1993) 20.509 20.298 0.807

John Rowley (1878) 20.576 20.213 0.789

James Bowman (1830s) 20.127 20.635 0.762

bong bong census (1836) 20.685 20.043 0.728

wollongong census (1834) 20.585 20.200 0.785

Lizzy Malone (1875) 20.368 20.447 0.815

A. Mackenzie, 1872 20.593 20.189 0.782

Richard Dawsey (1887) 20.352 20.462 0.814

saint georges basin census, 4 July (1836) 20.174 20.604 0.778

ulladulla (1853) 20.648 20.106 0.754

batemans bay (1853) 20.198 20.587 0.785

Percy Mumbla (1980s) 20.599 20.181 0.780

twofold bay census (1839) 20.168 20.608 0.776

birdhawal wordlist (1907) 20.602 20.177 0.779

port dalrymple (1826) 20.367 20.448 0.815

research bay (1793) 20.367 20.448 0.815

Average 20.447 20.343 0.790

996 C. R. Illert

Chapter 4, JAS 2006

Page 222

Example 3. Original-A: a1 ¼ a2 ¼ 4, a3 ¼ 6

(Four lamino-palatal consonants capable of appearing word-initially)

Following on from Table 5, we substitute s ¼ 2 into equations (16) giving theoretical

values for the respective components of a pre-Wolgalu ‘type-A’ unit-vector

111¼ �1þ ffiffiffi

5p

4¼ 0:309

121¼ �1� ffiffiffi

5p

4¼ �0:809

131¼ 1

2¼ 0:5

(18)

which may be compared with ‘experimentally observed’ values, as presented in Table 6,

calculated from the respective Original-A word/name-list signatures assembled in

Table 1.

Table 4. Experimental values for Southern-B word/name lists, calculated from the data in Table 1.

Compare these with the expected theoretical values (equation (14))

11/1 12/1 13/1

berrima census 1 September 1840 20.726 0.04 0.686

coolangatta wordlist William Bothong (1900) 20.817 0.412 0.405

dhurga wordlist Diana Eades (1976) 20.784 0.196 0.588

braidwood wordlist James Larmer (1853) 20.719 0.024 0.695

moruya wordlist Herbert Hale (1846) 20.793 0.228 0.565

Twofold bay wordlist Johnny Weyman (1864) 20.782 0.189 0.593

gippsland wordlist Louise Hercus (1969) 20.756 0.112 0.644

Wemba wemba wordlist Louise Hercus (1969) 20.796 0.554 0.242

Average 20.771 0.219 0.552

Table 5. Estimates of integer s, corresponding to hypothetical values of a2, based upon equation (15)

and experimental data from Table 1. It seems most likely that s ¼ 1 for Turuwal and s ¼ 2 for

Original-A

s

a1 ¼ 4 a1 ¼ 4 a1 ¼ 4 a1 ¼ 4

a2 ¼ 1 a2 ¼ 2 a2 ¼ 3 a2 ¼ 4

Turuwal

jervis bay wordlist Joseph Gaimard (1826) 0.788 1.078 1.368 1.658

turruwull wordlist John Malone (1875) 1.278 1.338 1.398 1.458

Original-A

kamilaroi, wordlist Rev. W. Ridley (1875) negative 0.577 1.423 2.269

dharginyung wordlist R.H. Mathews (1903) negative 0.651 1.401 2.151

dharruk, wordlist R.H. Mathews (1901) negative 0.665 1.398 2.131

gundungara wordlist M.M. Everitt (1900) negative 0.870 1.370 1.871

Origins of Linguistic Zonation in the Australian Alps, Part 2 997

Chapter 4, JAS 2006

Page 223

So far, we have not considered the possibility of t ¼ 2 in equation (9), which would

allow a ‘type-B’ pidgin of the form

11 ¼ 0:5� (�2sþ 1þ c)13

12 ¼ 0:5� (2s� 3� c)13

13 ¼ 21ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(�2sþ 1þ c)2 þ (2s� 3� c)2 þ 4

pwhere

c ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(2s� 1)2 � 2(s2 � 2)

q(19)

Example 4. Turuwal: a1 ¼ 4, a2 ¼ 2, a3 ¼ 3

(Two lamino-palatal consonants capable of appearing word-initially)

Following on from Table 5, we substitute s ¼ 1 into equations (19) giving theoretical

values for the respective components of a ‘type-B’ unit-vector

111¼ 0:5��1þ ffiffiffi

3pffiffiffi3

p ¼ 0:211

121¼ �0:5� 1þ ffiffiffi

3pffiffiffi3

p ¼ �0:788

131¼ 1ffiffiffi

3p ¼ 0:577

(20)

which may be compared with ‘experimentally observed’ values, as presented in Table 7, cal-

culated from the respective Turuwal word-list signatures assembled in Table 1. Note how the

experimental average values 1i/1 in Table 7 are each better-fitted by equation (20) than by

equation (18). Furthermore, looking back at Table 3, the theoretical ORIGINAL-B values of

Table 6. Experimental values for Original-A word/name lists, calculated from the data in Table 1.

Compare these with the expected theoretical values (equation (18))

11/1 12/1 13/1

kamilaroi wordlist Rev. W. Ridley (1875) 0.373 20.814 0.441

dharginyung wordlist R.H. Mathews (1903) 0.349 20.814 0.465

dharruk wordlist R.H. Mathews (1901) 0.344 20.813 0.469

gundungara wordlist Mary M. Everitt (1900) 0.267 20.802 0.535

Average 0.333 20.811 0.478

Table 7. Experimental values for Turuwal word-lists, calculated from the data in Table 1. Compare

these with the expected theoretical values (equation (20))

11/1 12/1 13/1

jervis bay wordlist Joseph Gaimard (1826) 0.041 20.727 0.686

turruwul wordlist John Malone (1875) 0.175 20.778 0.603

Average 0.108 20.753 0.645

998 C. R. Illert

Chapter 4, JAS 2006

Page 224

11/1 and 12/1 given by equation (13) seem to be interchanged in the cases of the word/name

lists of James Bowman, St. Georges Basin, Batemans Bay and Twofold Bay. So perhaps they

were not ORIGINAL-B at all but, instead, a more recent kind of TURUWAL described by

s ¼ 2 in equation (19) as shown in Figure 7. This makes sense for the 1836 St Georges Basin

census collected just south of Jervis Bay where J.P. Gaimard had, a decade before, obtained

his distinctly TURUWAL wordlist. Likewise, James Bowman’s wordlist, from south-west

of present-day Sydney, comes from a region where we would have expected to find a

TURUWAL influence. In addition, isolated pockets of TURUWAL influence along the

coast, extending as far south as Tasmania, are quite explainable in terms of post-Ice-Age

coastal fishing communities migrating by canoe southward from Jervis Bay simply follow-

ing summer East Australian Currents.

General Solutions

To this point we have only studied special cases. But if we place no restrictions on the

integers s and t and commence by re-defining

c ;ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(ts� 1)2 � (t2 � 2)(s2 � 2)

q

then equation (9), subject to the constraint (3), has a general pre-Wolgalu solution of

‘type-A’

11 ¼ ð�tsþ 1� cÞ13t2 � 2

12 ¼ ð�t2 þ tsþ 1þ c)13t2 � 2

13 ¼ �(t2 � 2)1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(�tsþ 1� c)2 þ (�t2 þ tsþ 1þ c)2 þ (t2 � 2)2

p(21)

Table 8. Experimental values of ai for each of the language superfamilies, and one modern dialect,

from Examples 1 to 4. These values tell us how many, and what kinds of, word-initial consonants

existed in each traditional language superfamily

a1 a2 a3 s ¼ a32 a2 t ¼ a12 a2

Turuwal 4 2 3 1 2

Original-B 4 3 5 2 1

Original-A 4 4 6 2 0

Southern-B 4 3 6 3 1

Wemba-Wemba 4 3 8 5 1

Origins of Linguistic Zonation in the Australian Alps, Part 2 999

Chapter 4, JAS 2006

Page 225

and a general pre-Wolgalu solution of ‘type-B’

11 ¼ (�tsþ 1þ c)13t2 � 2

12 ¼ (� t2 þ tsþ 1� c)13t2 � 2

13 ¼ �(t2 � 2)1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(�tsþ 1þ c)2 þ ð�t2 þ tsþ 1� c)2 þ (t2 � 2)2

p(22)

as well as a ‘type-B’ pidgin

11 ¼ (�tsþ 1þ c)13t2 � 2

12 ¼ (�t2 þ tsþ 1� c)13t2 � 2

13 ¼ þ(t2 � 2)1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(�tsþ 1þ c)2 þ (�t2 þ tsþ 1� c)2 þ (t2 � 2)2

p(23)

As far as the previously discussed special cases are concerned: when t ¼ 0 equations

(21) simplify to equation (16) whilst equations (22) simplify to equation (17); when

t ¼ 1 equations (21) simplify to equation (11) whilst equations (22) simplify to equation

(12); also, when t ¼ 2 equations (23) simplify to equation (19).

Original-A (‘non-Pama-Nyungan’)

Previous researchers (Troy, 1994, p. 23; Osmond, 1989, pp. 4–5) have speculated about

the possible existence of a laminal-lateral ly (denoted here by lO) within ‘the Sydney

language’ and Dixon (1980, p. 158) has suggested its possible genesis along the lines

ldO ! ly dO ! lO:

Without much discussion Troy included this laminal-lateral in the phonology of her

‘Sydney language’ which, in fact, corresponds to no real language that ever existed. It

is merely a melange-lexicon that misguidedly muddles coastal Original-B (‘eora’) and

inland Original-A (Blue Mountains ‘dharug’) vocabularies – only including John

Rowley’s historic (Original-B) wordlist in the mistaken belief that it was ‘dharug’,

whilst excluding Lizzy Malone’s historic (Original-B) wordlist in the arbitrary and

equally mistaken belief that it was a form of Turuwal.

Troy’s melange-lexicon fortuitously gives the correct Original-B signature, but only by

the happy accident that the small number of actual ‘dharug’ words that she used were

swamped by the sheer size of the ‘eora’ wordlists collected by First Fleeters and their

contemporaries such as David Collins, William Dawes, John Hunter, Phillip Gidley

King, Arthur Phillip, Watkin Tench, Daniel Southwell and Daniel Paine. In our present

analysis, the frequency distributions in Illert (2005, Appendix 1) were all normalised so

no single source could have this kind of disproportionate influence simply due to its

size. In addition, our wordlists were selected from uniformly spaced locations throughout

an extended geographical region, instead of all being clustered about the single centre

(Sydney) in Troy’s study.

From Example 1 in the present paper it is clear that Sydney’s coastal ‘eora’ language

was definitely Original-B, containing three (a2 ¼ 3) lamino-palatals (as in Table 9) but

1000 C. R. Illert

Chapter 4, JAS 2006

Page 226

Origins of Linguistic Zonation in the Australian Alps, Part 2 1001

Table,9, Likely minimal systems of word-initial (peripheral, lamino-palatal and other) consonants,

associated with pre-Wolgalu language super-families, based upon the values oi in Table 8 infened

from Snell's Law

r---'.Turuwal (H :9)tl stops,4 nasals, I semi-vowel

,\tops

nasals

semi-vowel

at=Z at = 4

l

I

l

I

I

Original B (H:12)5 stops, 5 nasals, 2 semi-vowels

Original-A (H:14)5 stops, 5 nasals, 2 semi-vowels, 2 lalerals

nasals

semi-vo+vels

laleruls

-\Iop.t

nasals

semi-vo:vels

stops

at:5 az=3 ar=4

u3:6 iu2:4 d1-4

).I Southern-B (H:131

5 stops, 4 nasals, 2 semi-vovvels, I laterdl,I rholic

stops

nasals

senti-v.ovvels

loteral

rholic

at-6 ^ (1t-4a): I

I

no laminal-lateral. However, our Tables 5 and 9 demonstrate that there did exist a fourth

lamino-palatal (a2:4), which can only have been a lateral, in adjoining Original-Alanguage from the Blue Mountains furlher to the west and far north-west of present-day

Sydney. Troy's misguided inclusion of Original-A 'mountain' words, in 'coastal' lexi-cons, has contributed to the apparently false notion ofa laminal-lateral in coastal language.

Chapter 4, JAS 2006

Page 227

It is also worth noting that, at the southern-most extremity of this Blue Mountains

language zone, William Russell’s ‘gundungorra’ (which was omitted from Tables 5 and

6) had a distinctive type-A signature (with 11/1 . 13/1) which is better described by

the theoretical values

111¼ 0:663,

121¼ �0:748,

131¼ 0:085

obtained by substituting s ¼ 4 into equation (11), implying that

a1 ¼ 4, a2 ¼ 3, a3 ¼ 7

Thus, the southern-most extremity of Original-A, the zone most influenced by surrounding

Type-B language, had clearly lost its laminal lateral by the early 1900s – in contrast to

dharug dharginyung and kamilaroi off to the north, which seem to have retained it,

judging by Table 5.

In any event, Yallop (1982, p. 66, Table 3.4) has shown that languages with a

laminal-lateral usually also have a series of retroflexive consonants rd, rn, rl (respectivelydenoted %, �, �) as in the basic inventory proposed for Original-A in our present Table 9.

Accordingly, we conclude that at least some of the word-initial apico-alveolar consonants

occurring in the Original-A wordlists in Illert (2005, Appendix 1) are, or perhaps derive

from, ancient retroflexives. Australian languages with this type of consonant inventory

– including anindilyakwa, kitja, bardi, garawa, pintupi, nyungumarda, walmatjari, warlpiri

and alyawarra – feature prominently in a region of far north-western Australia; about 10%

of the total continent referred to as the ‘non-Pama-Nyungan’ language zone (Dixon 1980,

pp. 20–21 and Map 3), which was first identified by Wilhelm Schmidt (1919) through a

careful study of word-final-sounds, which Dixon described as ‘quite superficial

features . . . seldom good criteria for genetic grouping’.

Original-A (‘non-Pama-Nyungan’) is also said to be ‘prefixing’ because subsequent

researchers (Capell, 1956; Hale, 1961–3] noticed that such language also tends to use pre-

fixes as well as suffixes, unlike most language throughout the rest of Australia, which tends

to rely on suffixes alone. The unexpectedly complex system of prefixing and suffixing in

the ‘Sydney language’ recorded by William Dawes (1790), as translated in Illert & Allison

(2004, examples 1–7), could thus reflect an Original-A (‘non-Pama-Nyungan’) syntactic

influence on the isolated strip of Original-B along the south-eastern seaboard. This is

Table 10. PamaNyungan sound-shifting during the transition ORIGINAL-B ! SOUTHERN-B

1002 C. R. Illert

Chapter 4, JAS 2006

Page 228

consistent with our present study of word-initial-sounds, which establishes a surge of

Original-A along the Blue Mountains chain, as far south as the Huygens Circle in the

Australian Alps, at the height of the Ice Age (see Figure 9).

Presumably, this Gillieron wavefront propagated through south-western Queensland

(where there are modern languages such as Pitta-Pitta (Dixon 1980, pp. 160–161) with

retroflexives and a laminal lateral) originating from the ‘non-PamaNyungan’ ‘prefixing’

epicentre first identified by Schmidt (1919) in the north-west of the continent.

Pre-Wolgalu Type-B (‘Pama-Nyungan’)

Throughout nearly 90% of Australia – Schmidt’s ‘southern’ portion – we find a substrate

of pre-Wolgalu Type-B (‘Pama-Nyungan’) Aboriginal language. Our present Gillieron

wave analysis suggests that all language throughout Australia may have originally been

of this type dating back at least 60,000 years, whilst Original-A (‘non-Pama-Nyungan’)

may have arrived in north-western Australia as recently as the Ice Age (when there

were land-bridges throughout much of south-eastern Asia) eventually spreading south-

ward through Australia over the top of a long-pre-existing Type-B linguistic substrate.

If so, for ‘PamaNyungan’ and ‘non-PamaNyungan’ languages to have had common

origins, proto-Australian may have arisen in southeast Asia – somewhere between

Australia and the Indian subcontinent – up to 75,000 years ago.

The present analysis cannot go back quite that far but it does extend to the Turuwal

signature, which is arguably about 60,000 years old and phonologically similar to

proto-Australian itself, and we can usefully work forward in time from that. Having deter-

mined values ai in equation (9) which seem to best account for the ‘experimentally

observed’ signatures in Table 1 – hence the ‘likely’ minimal consonant inventories in

Table 9 we can now model the evolution of pre-Wolgalu Type-B signatures, as in

Figure 10, assuming a simple linear relationship

time (years) � 15;000� s� 75; 000 for s . 0

The intersection of our two sets of signature-curves (19) and (12), between s ¼ 1 and

s ¼ 2, corresponds to a2 increasing from 2 to 3 with the emergence in Turuwal (about

53,000 years ago) of an additional lamino-palatal consonant that modified subsequent tra-

jectories 1i/1. Graph 5 clearly shows how this injection of word-initial-y into pre-Wolgalu

Type-B language caused 12/1 to lift above the corresponding Turuwal trajectory, requiringthe trajectories 11/1 and 13/1 to both compensate by dropping below their respective

Turuwal counterparts for all s . 2, in accordance with equation (3). Even so, for

s . 1.3 the subsequent Original-B and Southern-B signatures given by equation (12)

are only slight perturbations from ancestral Turuwal trajectories (19), showing the

relatedness of all pre-Wolgalu Type-B language. Thus, word-initial-y did not exist in

proto-Australian but arose in Turuwal, about 53,000 years ago, spreading from there

into Original-B possibly – but not necessarily – within Australia.

The Semantic Basis of Proto-Australian

Table 9 supplies the distinctive minimal consonant-inventories of ancient language super-

families whose existence and relatedness have been inferred from phonotactic information

extracted from old word/name-lists. But if there actually was a single proto-language that

gave birth to some or all pre-Wolgalu language, if our Tree Diagram is even approxi-

mately true, then cognate words from different language superfamilies must have had

Origins of Linguistic Zonation in the Australian Alps, Part 2 1003

Chapter 4, JAS 2006

Page 229

common origins in adducible ancestral leximorphs (Illert, 2003) whose very existence

would provide tangible semantic evidence corroborating the claimed genetic relationships

between proposed language superfamilies. But where would we look for the requisite

archaeo-semantic-exemplars?

Deictic expressions are the linguistic equivalent of vectors, usually with the speaker as

the origin of the frame of reference. The word ‘deixis’, meaning linguistic pointing, is used

in relation to a speaker ‘pointing’ at particular places, times and individuals (Lyons, 1968,

pp. 275–280). We are engaging in the use of deixis whenever we use positionals such as

‘this’, ‘that’, ‘here’, ‘there’, ‘hither’, ‘thither’, or temporals such as ‘before’, ‘now’, ‘after’,

‘soon’. Because they are so fundamental and essential, deictics can persist relatively

unchanged through timescales so deep that all else in language may change beyond rec-

ognition. Thus, if we want to know about ancestral proto-Australian language, to the

extent that anything may be known, then we might profitably start by seeking deictics

which seem universally to underlie Australian languages.

In this spirit, Geoff O’Grady and Susan Fitzgerald (1997, pp. 352–353) supplied some

‘broad-brush’ evidence for the existence of an ancestral deictic, here (in space or time),

proposing ‘a proto-Pama-Nyungen root yIya signifying “now, today” with vocalic metath-

esis occurring in . . . [some] languages’. But, if this or any other proto-Australian word

commenced with y, it could contradict our previous finding that word-initial-y arose in

Turuwal 53,000 years ago. Fortunately the ‘broad-brush’ case for word-initial-y is uncon-

vincing in comparison with the survey below of relevant cognates from pre-Wolgalu

superfamilies throughout south-eastern-Australia. This present evidence of ancestral

word-initial-w (instead of y) is over-whelming. Additionally, the hypothesis of vowel-

interchangement (‘vocalic metathesis’) is also wrong. All relevant cognates clearly

arose from the ancestral semantic notions

some near-½place� ¼ HERE

some near-½thing� ¼ THIS

by means of the agglutination and fusion of a pair of permutable proto-Australian roots as

follows:

1004 C. R. Illert

Chapter 4, JAS 2006

Page 230

or

Accordingly, there existed a present-tense temporal

some near-present-time ¼ SOON, NOW

as in

or

Origins of Linguistic Zonation in the Australian Alps, Part 2 1005

Chapter 4, JAS 2006

Page 231

Also, language superfamilies throughout south-eastern-Australia featured the future-tense

hence there existed a future-tense temporal

some near-future-time ¼ SUBSEQUENTLY, AFTER

as in

or

1006 C. R. Illert

Chapter 4, JAS 2006

Page 232

Also, the past-tense was generally

hence there existed a past-tense temporal

some near-past-time ¼ PREVIOUSLY, BEFORE

as in

or

These newly deciphered deictics, featuring the consonant fi, for the first time enable

modern translations to cut through the orthographic ‘static’ that has accumulated over

Origins of Linguistic Zonation in the Australian Alps, Part 2 1007

Chapter 4, JAS 2006

Page 233

centuries in archival accounts of poorly understood and long extinct Australian

languages – as in the following sentences:

ORIGINAL-B (‘PamaNyungan’) example

ORIGINAL-A (‘non-PamaNyungan’) example

The antiquity of these deictics is apparent from their universality – occurring in both

Original-A (‘non-PamaNyungan’) and pre-Wolgalu Type-B (‘PamaNyungan’)

languages – from which our Tree Diagram (Illert, 2005, Figure 4) indicates that they prob-

ably arose in proto-Australian more than 60,000 years ago. That they subsequently propa-

gated into Ice-Age Australian languages, in accordance with Figure 10, is illustrated by the

following.

ORIGINAL-B (‘PamaNyungan’) example

1008 C. R. Illert

Chapter 4, JAS 2006

Page 234

SOUTHERN-B (‘PamaNyungan’) example

The Yelta People’s Moon Song, recorded by Reverend John Bulmer in the 1850s, occurs

wrongly transcribed and mistranslated on page 431 of R.B. Smyth’s (1876) book. A

correct translation is as follows: Die [moon]! Between-cycles [you] shall disintegrate.

[You] NOW waxing moon, LATER [will] wane.

There has long been folklore to the effect that Aboriginal people supposedly

‘cannot count higher than two’ (Sir John Lubbock, cited by John Fraser FRS, 1893)

and possess such poor comprehension of abstract concepts like time that they ‘will

sell their blanket in the morning – though not at night – because they forget that it

will be needed again after sunset . . .’ (Isabelle M. Pagan, 1937, p. 61). The former prop-

osition was refuted by Illert (February 2003, Example 6) whilst the latter proposition, sur-

prisingly enough, being argued by Aboriginal people currently involved in linguistic

salvage, takes the form that south-eastern languages such as Bundjalung lacked a full

range of temporal deictics. The January 1998 issue of Voice of the Land, published by

the Federation of Aboriginal and Torres Strait Islander Languages (FATSIL), ran a

front-page story

Festival’s Bundjalung presentation a world first. Audiences who attended the Sydney Fes-tival of Dreaming in September were first to see a play translated and performed in anAboriginal language. Samuel Beckett’s ‘Ngundalelah Godatgai’ (Waiting for Godot) wastranslated into Bundjalung by Elders from the Northern Rivers region of New SouthWales . . . supervised by Mick Walker . . . Beckett’s play deals with the frustration ofhis two central characters, stranded in a barren landscape and forced to grapple withabstract concepts of time. . . .With no specific words for past or present in the Bundja-lung language, the group members were required not only to translate, but to interpretthe script into equivalent Aboriginal concepts . . . in the Galibal dialect of the ClarenceRiver region.

Specifically in relation to the past tense, in Bundjalung, it should suffice to think back 40

years to when the Bundjalung Elder Ken Gordon explained that his traditional legends

Origins of Linguistic Zonation in the Australian Alps, Part 2 1009

Chapter 4, JAS 2006

Page 235

were called [tales] from-long-ago

Whereas dUra means through or interior, the conjugate adjective darU means about

or exterior; hence, the proto-Australian expression some exterior-things (eg. leaves,

fish-scales, bird-feathers, echidna-spikes, as in Figure 11)

Figure 11. Some exterior things

1010 C. R. Illert

Chapter 4, JAS 2006

Page 236

the singular case being the exterior-thing

However the absence of an affix such as mifi leaves open the syntactic possibility of a

possessive noun-stem, hence the interpretation thing’s exterior or surface-layer (e.g. skin,

fingernail, clothing, tree-bark, name, surface of the Earth, as in Figure 12)

Origins of Linguistic Zonation in the Australian Alps, Part 2 1011

Chapter 4, JAS 2006

Page 237

The fundamental root

was combined with darU in proto-Australian, creating the concept away-exterior or

surrounding-region (¼neighbourhood, vicinity), as in [you] shall-come [to] Sydney’s vicinity

(see Figure 13)

This example shows that, notwithstanding the intrinsic meanings of its constituent

individual morphemes, the cluster wuru:n:darU is a deictic idiom meaning

surrounding-region hence neighbourhood, vicinity. And this meaning can vary

Figure 12. Surface layer

1012 C. R. Illert

Chapter 4, JAS 2006

Page 238

pragmatically, as in [we] shall-row some-where

which finally explains a 215 year old mystery articulated by William Dawes in his

notebook:

. . . We think it relates to bringing Booroong to Dara, in which case it appears that they putwords sometimes between the root and the termination. They were not speaking of Dara,for since, I have heard them repeat dara in the same word when I think they could notrefer to that place. It seems to me to be peculiarly used when it is spoken of as rowing to acertain place to bring back another with you. But this is mere conjecture.

Dawes had made a major syntactic discovery, the intersplicing of positional deictic

(‘some-where’) and verb-phrase (‘shall-row’) within a sentence (see also Illert, 2003,

examples 1 to 5), and was clearly hearing the word darU . Yet philologists such as Jim

Kohen (1998, p. 240) cite these writings in support of a view that ‘ . . . the root word is

“ora”, meaning place or country’, apparently unaware that Aboriginal words simply do

not commence with nor contain vowels such as o. Kohen continued on to make the

more seriously wrong connection

. . . the prefix ‘e-e’ is listed in Southwell (1788) . . . [as] the word for ‘yes’ . . . The use of thewords for ‘yes’ and ‘no’ to distinguish tribes and subtribes has frequently been recorded in theliterature for south-eastern Australia (Tindale 1974, p. 42) . . . The Eora then were the peoplefrom the country where the word for ‘yes’ was ‘e-e’, that is the Sydney coastal region.

Figure 13. Surrounding region

Origins of Linguistic Zonation in the Australian Alps, Part 2 1013

Chapter 4, JAS 2006

Page 239

Actually, the correct word for ‘yes’, occurring in all language superfamilies throughout

south-eastern Australia, happens to be

whilst the standard animate-plural is

. . . occurring commonly within superlative adjectival expressions, as in the following

1014 C. R. Illert

Chapter 4, JAS 2006

Page 240

ORIGINAL-A (‘non-PamaNyungan’) examples

based upon the idiom

near-centre ¼ AMIDST, AMONGST, WITHIN

given that the root fiura ¼ middle/centre is well documented in expressions meaning

‘belly-button/navel’ (as in Figure 14)

Origins of Linguistic Zonation in the Australian Alps, Part 2 1015

Chapter 4, JAS 2006

Page 241

the standard animate-singular being

Clearly the supposed name of the Sydney tribe ‘Eora’ was nothing more than a contrac-

tion of the standard animate plural, with unheard word-initial-˛ and soft-w. Kohen was

confusing darU with ˛:wurU, and the word ˛Un was simply irrelevant.2 His erroneous

chain of speculation, ‘ee-ora’ supposedly meaning ‘[people from the] country [where]

“ee” [means yes]’, followed largely from word-initial consonants being consistently mis-

heard or unrecorded by early observers and unrestored by centuries of subsequent

philology.

In any event, William Dawes had stumbled upon the positional

some vicinity ¼ WHERE

as in

Figure 14. Amidst/amongst/within

1016 C. R. Illert

Chapter 4, JAS 2006

Page 242

or

hence the following.

ORIGINAL-B (‘PamaNyungan’) examples

Likewise there was a temporal

some temporal-proximity ¼ WHEN

Origins of Linguistic Zonation in the Australian Alps, Part 2 1017

Chapter 4, JAS 2006

Page 243

as in

also

Another important positional was

some-near vicinity ¼ this vicinity ¼ HITHER

1018 C. R. Illert

Chapter 4, JAS 2006

Page 244

as in

or

or

or

hence the following.

Origins of Linguistic Zonation in the Australian Alps, Part 2 1019

Chapter 4, JAS 2006

Page 245

ORIGINAL-B (‘PamaNyungan’) example

Additionally we have the

ORIGINAL-B (‘PamaNyungan’) example

In the notebooks of William Dawes (1790) we find

1020 C. R. Illert

Chapter 4, JAS 2006

Page 246

and

hence the sentence: [1] shall-long-stay here-abouts

A different but equally important strand of deictics were based upon the ancestral

semantic notions

some yonder-[place] ¼ THERE

some yonder-[thing] ¼ THAT,

as follows:

Origins of Linguistic Zonation in the Australian Alps, Part 2 1021

Chapter 4, JAS 2006

Page 247

or

as in

The fundamental root

as in

was combined with n:darU in proto-Australian to create the concept

wara-n:darU ¼ distant place

which, in turn, was combined with fiun:nUra giving the positional

some yonder distant place ¼ THITHER

1022 C. R. Illert

Chapter 4, JAS 2006

Page 248

as follows

as in the following.

Figure 15. Distant place (relative to speaker)

Origins of Linguistic Zonation in the Australian Alps, Part 2 1023

Chapter 4, JAS 2006

Page 249

ORIGINAL-B (‘PamaNyungan’) examples

which further illustrate the splitting/intersplicing of positional deictics (such as

HITHER and THITHER) about/with verbs.

Soundshifts in Southern-B

With R.H. Mathews (1901) recording nUra ! ‘nharra’ (‘yonder’) within the Huygens

Circle, supported by other observers who, quite understandably, gave word-initial-nsimply as n in surrounding language zones, there is a strong case for the common

origin of n in proto-Australian. But equally evident in modern cognates involving

THERE-THAT-THITHER are the soundshifts n ! ˛ and n ! m which probably arose,

1024 C. R. Illert

Chapter 4, JAS 2006

Page 250

about 30,000 years ago, at the birth of Southern-B language, which lost the nasal n (see

Table 9) and had to compensate. This was part of the general decrease in frequency of

word-initial-nasals in Southern-B

ð0�T

_f m þ _f I þ _fw þ _f n þ _f n dt ¼ Dfm þ Df˛ þ Dffi þ Dfn þ Dfn

¼ �23:7% ðproto - Australian ! dhurgaÞ

balanced by an increase in frequency of word-initial stops in Southern-B

ð0�T

_f g þ _f b þ _f dO þ _f d þ _f d dt ¼ Dfg þ Dfb þ DfdO þ Dfd þ Dfd

¼ þ22:8% ðproto-Australian ! dhurgaÞ

associated with soundshifts b ! d or d, and g ! d or d, which are particularly apparent

as one moves progressively inland, westward, away from the south-eastern coastal-strip

(see Table 10).

Examples of the Soundshift g ! d

Origins of Linguistic Zonation in the Australian Alps, Part 2 1025

Chapter 4, JAS 2006

Page 251

which, in its repeated form in the Southern-B language zone, was recorded by independent

observers as

Another repeated proto-Australian root occurs similarly soundshifted in rainbow

(¼[contiguous coloured] arcs):

Examples of the Soundshift b ! d

1026 C. R. Illert

Chapter 4, JAS 2006

Page 252

Another example, one which shows Southern-B influence extending into adjoining

language zones, is the already encountered temporal deictic

Retroflexives Arising from Coalescence of Phonemes ?

rUn ! �; rUd ! %; rUl ! z

We have documented a likely origin for a retroflexive nasal � in the above-cited Southern-

B version of the word ‘rainbow’. Blake & Reid (1998) have also hypothesised the

existence of a retroflexive stop % in the same NW-Victorian languages. They state:

following O’Grady’s suggested reconstruction for these sets we now think that the hypothesisthat accounts best for the data is that rt is original, that it weakened in some Victorianlanguages to some kind of rhotic, probably a glide . . . and that it switched to a laminal insome western Victorian languages.

As evidence, they offer words supposedly meaning ‘big, important, man, old man’ but

which are all probably just examples of the standard proto-Australian plural (Illert, 2002,

example 6)

Origins of Linguistic Zonation in the Australian Alps, Part 2 1027

Chapter 4, JAS 2006

Page 253

being used, in Southern-B, in the adjectival sense ‘lots of’ or ‘big’. The actual words

given by Blake & Reid are just a slightly different condensation of this same ancestral

arithmetical leximorph

In which case, the stop-sounds observed within wu:rU:dUl or wu:r:dUl or wu:r:dU or

wu:r:d are easily explained as % arising from coalescence of the phonemes rUd.Unfortunately Blake & Reid did not systematically cite the original primary sources

from which each of their respective words supposedly came. They made the common

mistake of failing to understand that, in lexical taxonomy, as with any form of taxonomy,

it is the oldest original primary source that carries most weight – not the latest compi-

lation. Their lexical taxonomy all needs to be redone in a thorough scholarly fashion, spe-

cifying the oldest primary source from which each and every one of their cited words

came. However, for now, giving them the benefit of the doubt and taking their ‘findings’

at face value, their claim to have found a retroflexive-stop in the Southern-B region of

western Victoria is consistent with our discovery of a retroflexive-nasal. Collectively,

this suggests retroflexives in SE-Australian language, mainly within Southern-B, for any-

thing up to 30,000 years.

Acknowledgements

This paper is in support of and in part-fulfilment of the requirements of a PhD by the author

at the University of Western Sydney. The author would like to thank the anonymous

referee for valuable suggestions, Dr Bob Hodge and Dr Vladimir Dimitrov both of the

University of Western Sydney, and Daniela Reverberi for software support.

Notes

1Equations 1 to 5, Tables 1 and 2, and Figures 1 to 4, occur in Part 1 of this paper.

1028 C. R. Illert

Chapter 4, JAS 2006

Page 254

2In south-east Australian languages, the condensed root ˛U(n) ¼ ‘yes’ and the pronoun ˛U ¼ ‘I, me’ provide

a possible example of homonymy – two different words that just happen to have the same form – as in

English where, for example, ‘bank’ denotes both a building (associated with a financial institution) and

also an area of ground (along the side of a river). But it is interesting how English has the homophones

�a ¼ ‘yes’ (as in ‘ay Captain’) and �K ¼ ‘I, me’ both of which just might be directly condensed from

proto-Australian: ˛Un ! ˛U ! U ! aia ! ai ¼ ‘yes’ or ‘I, me’. Furthermore proto-Australian has the

conjugate root nU˛ ¼ ‘no, not’, recorded as ‘naiyu˛’ (Malone, 1875), perhaps corresponding to ‘nen’ or‘nun’ (ancient Egyptian) with the condensation nU˛ ! nU producing such modern cognates as ‘nei’(Danish) and ‘nay’ or ‘no’ (English).

3This use of mifi instead of fiun in the expressions WHERE and WHEN (Dawes, 1790) and HITHER

(Rowley, 1875) from the Sydney region, also WHERE (Bulmer, 1850s) and WHEN (Petitit & Dawson,

1850s) from Gippsland, are probably just errors.4See Example 1 in C. Illert (2003).

References (continued from Part 1)

Baisden, F. (1998) Festival’s Bundjalung presentation a world first, Voice of the Land, FATSIL (Federation of

Aboriginal and Torres Strait Islander Languages Corporation) Newsletter, 7(January), pp. 1–2.

Blake, B.J. & Reid, J. (1998) Sound changes in languages of western and central Victoria, Australian Journal of

Linguistics, 18(1), pp. 57–72.

Bulmer, Rev. J. (1850s) In A. Campbell & R. Vanderwal, ‘Victorian Aborigines, John Bulmer’s

recollections 1855–1908’, Occasional Papers in Anthropology and History, Number 1, Museum of

Victoria (1994). The bibliography on page 87 contains a compilation of Bulmer’s writings. See also

R.B. Smyth (1876).

Capell, A. (1956) A new Approach to Australian Linguistics (Sydney: Oceania Linguistic Monographs).

Cunningham, P. (1827) Two Years in New South Wales, vol. 2, p. 23.

Endacott, S.J. (1924) Australian Aboriginal Words and Place Names and their Meanings (Melbourne: Acacia

Press).

Enright, W.J. (1900) The language, weapons and manufactures of the Aborigines of Port Stephens NSW, Journal

of the Royal Society of NSW, 34, pp. 103–118 plus plates 3 and 4.

Gordon, K. (1960s) In Notes on the Bandjalang dialect spoken at Coraki and Bungawalbin Creek New South

Wales, edited by N.M. Holmer (1971), Australian Aboriginal Studies number 31, Linguistic Series

number 11 (Canberra, Australia: Australian Institute of Aboriginal Studies).

Griffith, C. (1845) The Present State of Prospects of the Port Phillip District of New South Wales, p. 65 (London).

Hale, K.L. (1964) Classification of northern Paman languages . . . , Oceanic Linguistics, 3, pp. 248–265.

Hale, K.L. (1966) In G.N. O’Grady, C.F. Vogelin & F.M. Vogelin (eds), Languages of the world . . . , Anthro-

pological Linguistics, 8(ii), pp. 162–176.

Illert, C. (2003) Lexigenesis in ancestral SE-Australian Aboriginal language, Journal of Applied Statistics, 30(2),

pp. 113–143.

Illert, C. (2005) Origins of linguistic zonation in the Australian Alps, part 1 – Huygens’ Principle, Journal of

Applied Statistics, 32(6), pp. 625–659.

Illert, C. & Allison, A. (2004) Phonogenesis and the origin of accusative syntax in proto-Australian language,

Journal of Applied Statistics, 31(1), pp. 73–104.

Kohen, J. (1998) The Dharug of the western Cumberland plain: ethnography and demography, in: B. Meehan &

R. Jones (eds), Archaeology with Ethnography: an Australian Perspective (Canberra: Australian National

University), pp. 238–249.

Martin, A.E. (1943) 1000 and more Place Names in New South Wales (Sydney: NSW Bookstall).

Matthews, Rev. D. (1874) Kitty Cooper’s Moira Tribe wordlist, in: N. Cato (1976)Mr Maloga, Daniel Matthews

and his mission, Murray River, 1864–1902, pp. 375–377 (University of Queensland Press).

McCarthy, F.D. (1950) New South Wales Aboriginal Place Names and Euphonious Words, with their Meanings

(Sydney: Government Printer).

O’Grady, G. & Fitzgerald, S. (1997) Cognate search in the Pama-Nyungan language family, in: Archaeology and

Linguistics, Aboriginal Australia in Global Perspective, P. McConvell & N. Evans (eds), pp. 341–355

(Oxford University Press).

Pagan, I.M. (1937) Racial Cleavage, p. 61 (London: Theosophical Publishing House).

Paine, D. (1794) The Journal of Daniel Paine 1794–1797, R.J.B. Knight & A. Frost (eds) (1983), pp. 41–42

(Sydney: Library of Australian History).

Schmidt, W. (1919) Die Gliederung der Australian Sprachen (Vienna: Springer).

Origins of Linguistic Zonation in the Australian Alps, Part 2 1029

Chapter 4, JAS 2006

Page 255

Weinstock, R. (1952/74) Calculus of Variations, with Applications to Physics and Engineering, pp. 67–70

(New York: Dover Publications).

Yallop, C. (1982) Australian Aboriginal Languages, p. 66 (London: Andre Deutsch).

1030 C. R. Illert

Chapter 4, JAS 2006

Page 256

Appendix

Symbols from International Phonetic Alphabet and how they are used herein, 2. Vowels, 2-7. Number of four symbol words, 8-11. Special occurrence of the symbol i, 15. Eight proposed conjugate adjectives, 18. Four pairs of superlative inflection, 23. Adjectives, verbs, nouns, 24. Conjugate pairs of words, 25. Animate singular, 26-27. Inanimate singular, 27 -28. Dual nouns, 28-30. Animate dual nouns, 29-30. Inanimate dual nouns, 30. Triple nouns, 30-31. Constructs of ãìêr and ãrê~, 32-36. Constructs of ãräì and ãìär, 37-40. Constructs of ãrär=and ã~ärI=4l-44. Constructs of ãìä~ and ã~äì, 45-47. Constructs of ã~ä~=and ãìäì, 48-50. Constructs of Öìêr and Örêì, 5l-53. Constructs of Ö~êr=and Örê~, 54-63. Constructs of Ö~êì=and Öìê~, 64. Constructs of Örä~=and Öìär, 65-71. Constructs of Öìä~ and Ö~äì, 72-74. Constructs of Ärêì=and Äìêr, 75-78. Constructs of Ä~êr=and Ärê~, 79-85. Constructs of Ä~ê~=and Äìêì, 86. Constructs of Äìär=and Äräì, 87-99.

Constructs of Ïìêì=and Ï~ê~, 101-105. Constructs of Ïìäì=and Çìäì, 106-108. Violations of octahedral symmetry, 111. Grand table of ancient language, 112-1 13. Chapter 1

Morpheme, 115-123. Morpho-statistics, 123 - 125. Simple lexemorphs, 125-135. Lexeme to "word", a phase transition, 135. George Kingsley Zipf , 136. Johannes van der Waals, 136. Word numbers condense to 2Czipf, 141. Optimisation problem, 143.

Chapter 2

Inverse Square Law, 147 –156.

Fitting inverse square law to two basic signatures (“Tharawal” = ngunawayl and Dhurga), 164-168.

Diachronic Optimisation Problem, 168-170. Chapter 3

Fitting the inverse square law to six initial linguistic signatures, 179-184.

Fitting boundaries to languages, 185.

Comparison of boundaries with previous cartography, 184-188.

Gilleron waves and deep linguistics, 189-191.

Migration to Tasmania and collision in Canberra, 190, 194.

Huygen's Principle, 192-195

Tree diagram of Australian language super families, 195-199. Chapter 4

Snell's Law across phonological boundaries in SE-Australia, 215-229.

Evolution of likely minimal systems of word-initial consonants in pre-Walgalu language, shown in Figures 8 and 9, 227.

The making of a new consonant, “the semantic basis of proto-Australian”, 229-259.

Sound-shifts in Southern-B, 250-253.

Retroflexives arising from coalescence of phonemes, êrå=→þ=,=êrÇ=→‚=,=êrä=→ý=, 253-255.

glossary

Page 257

Chapter 4 please note

Table 3 on page 222, is associated with equations (13)

Table 4 on page 223, and Figure 5 on page 217, are associated with equations (14)

Table 6 on page 224, and Figure 6 on page 218, are associated with equations (18) Table 7 on page 224, and Figure 7 on page 218, are associated with equations (20) 

Table 8 on page 225, and Figure 8 on page 219, are associated with equations (21), (22) & (23) 

Figure 9, on page 220, belongs with “Original-A (non-PamaNyungan )" on pages 226-229 

Figure 10, on page 221, belongs with “Pre-Wolgalu type-B ( Pama-Nyungan)" on page 229 Table 10, on page 228, belongs with “Sound shifts in Southem-B” on pages 250-251

corrections made In Table 9 on page 227, replaced ĺ with Ï=and replaced Ä with ‚=

Table 9 is shown correctly on the following page.  

Chapter 3, corrections made page 181 9th line from bottom Replaced the boundary condition f* (-T) ≡ f* and f*(0) ≡ F* with f* (-T) = f* and f*(0) = F* page 204 for proto-Australian data column set only, replaced F(%) with f (%)

pages 204 to 213 for all other data column sets, italic F (%) replaced with plain F(%) 

Chapter 2, corrections made page 150: In Table 2 label, replaced Δf* = F* - f* ≡  dt with Δf* = F*- f*  = dt also please note Table 2 should be located on page 164 pages 164 – 165: In equations (15), (16), (17), (18) , & lines 13 & 15 replaced italic f* with plain f* and replaced italic F* with plain F* page 168: 4th line from bottom Replaced the boundary condition f* (-T) ≡ f* and f*(0) ≡ F* with f* (-T) = f* and f*(0) = F* 

0

*

.−∫ T

f 0

*

.−∫ T

f

typographical errors

Page 258

Correctly shown Table 9, from Chapter 4, page 227

Page 259