Previous Next

Aboriginal Placenames

5. The methodology of reconstructing Indigenous placenames

Australian Capital Territory and south-eastern New South Wales

Harold Koch

Figure 5.1: Canberra area modern placenames. Photograph by Grace Koch.

figure0501.jpg

Aim and overview

The aim of this paper1 is to show how and to what extent, by using linguistic and historical methods, it is possible to reconstruct placenames of Aboriginal origin in those parts of Australia where direct knowledge of the language of the placenames has been lost. The examples are drawn largely from the Australian Capital Territory (ACT) and to some extent the adjacent area to the south which goes by the name Monaro. This section discusses what we mean by reconstruction and what aspects of a placename we should try to reconstruct (meaning and pronunciation). Also sketched are the issues of transmission of a placename from a traditional Aboriginal system to a different Anglo-Australian system: ‘intercultural transmission’ involves transfer across a linguistic and cultural barrier, and ‘historical transmission’ through time within the Anglo-Australian system of geographical nomenclature. The next section ‘Historical transmission’ describes how methods from the discipline of history can help us understand the original bestowal of the name in the Anglo-Australian system as well as changes in the locational referent, pronunciation, and spelling of the name. The section ‘Intellectual transmission and the linguistic reconstruction of sounds’ shows what the discipline of linguistics brings to the reconstruction of the original pronunciation and explanation of the variant spellings, giving reasons for the massive variability that is typically found. The method of comparing and matching is applied in detail to a number of examples. Following this ‘Reconstructing the etymology’ shows how to go about reconstructing the ‘meaning’ of placenames, based on the evidence of early testimony as well as the (cautious) use of wordlists. The final section summarises the methodology that has been presented and applies it to the name ‘Canberra’.

On reconstruction

What do we mean by reconstruction?

Reconstruction is a method for gaining knowledge of the past that is not available through direct transmission. The results of reconstruction are nearly always incomplete and are always to some extent hypothetical.

Why do we need to ‘reconstruct’?

In many parts of Australia there has been a break in the continuity of Aboriginal languages and in the transmission of knowledge about the names traditionally associated with places in the lands occupied by the speakers of those languages. Nevertheless we may have available placenames that are still in use that were derived from such languages – although their pronunciation may have been altered and their local reference may have shifted. Further names may be available that are no longer in use but were nevertheless recorded in writing at some time since the coming of Europeans to that area. If our goal is to understand these names from the viewpoint of the traditional culture from which they were taken, we need to engage in the historical exercise of ‘reconstruction’.

What are we trying to reconstruct?

The typical questions asked of placenames of Indigenous origin are about their meaning and about their form:

  • • What is the meaning of the name?
  • • What is the correct spelling of the Indigenous name?

Let us first examine more closely the question of meaning. Placenames, like other proper nouns, do not have a ‘meaning’ or semantic content of the kind that most other words in a language have.

Proper nouns have no inherent semantic content, even when they are homonymous with lexical words (Daisy, Wells), and many, perhaps all, cultures recognise nouns whose sole function is to be proper (Sarah, Ipswich). Typically they have a unique intended referent in a context of utterance. (Coates 2006: 312)

The ‘meaning’ of a placename thus reduces to its locational referent, i.e. the place to which it refers – what kind of place, its geographical extent and its precise location. For example, Victoria is a politically constituted geographical area, a state, in the south-eastern part of the geographically defined political entity called Australia.

Yet when people ask about the meaning of a name of Indigenous origin, they usually are not asking about its locational referent but about what I will call its etymology. It is a question about the meaning of the elements that constitute the name, what their non-name meaning is (cf. Reed 1977 and RASA 1900).2 Thus, regarding Victoria, the question of meaning in this sense is that victoria is a Latin word meaning ‘victory’. But the question of the origin of its meaning may extend to a further question: how is this meaning related to its locational referent? In other words, how did the name come to be associated with the place? This demand may be partially satisfied by supplying the information that the state name Victoria is derived from the name of the British queen, Victoria (whose name etymologically meant ‘victory’), who was the sovereign at the time of the constitution of Victoria as a separate British colony. So there is an association behind the name, a story that explains the bestowal of a particular name. I will call this question regarding meaning the etiology of the name.

Now, with placenames of Indigenous origin, we can rarely expect to have information about the story behind the name. Nevertheless we do have a few examples of names bestowed by Europeans who in colonial times named a place after an Aboriginal person. Thus the Tarra River of Gippsland was named after the Goulburn Aboriginal Charley Tarra, who was part of the exploration party led by Angus McMillan (Watson 1997: 169). Or, according to one tradition, the name of a property in the modern Australian Capital Territory, Congwarra (near the site of the Tidbinbilla Space Tracking Station) was named after an Aboriginal elder of the same name.3 Indigenous names of this kind reflect European patterns of name bestowal.

The bestowal of placenames in traditional Aboriginal societies follows a different pattern. In Aboriginal societies whose nomenclature patterns are known we find that many (but not necessarily all) placenames allude to mythological stories. They may refer directly, with various degrees of lexical complexity. For example, a name may explicitly translate ‘magpie’, ‘belonging to magpie’, or ‘where the magpie attacked’. On the other hand, a name may refer only in a very oblique fashion. Such is the case of the Kaytetye placename ‘Artarre’ – ‘tail feathers (of emu or turkey)’, which alludes to a story where in the Dreaming two men (really carpet snakes) decorated their bodies by inserting emu tail feathers into their belts and performed a dance to attract a woman, who was really the sun, with disastrous consequences: the two men still stand there memorialised in the form of a couple of trees on a hillside near a creek. To know that the placename “means” ‘tail feathers’ is of little help without knowing the story behind the single word. In other words, the etymology without the etiology is rather useless; this is, however, what is usually given as the ‘meaning’ (e.g. in Appleton and Appleton 1992).

In long-settled areas of Australia, where the traditional languages have been poorly documented, we have very little hope of ever recovering the ‘meaning’ in the sense of the mythological story to which the name alludes – unless there is some mythological knowledge either preserved by the relevant Indigenous community or found in the historical records of the early European settlers who learned it from the local Indigenous population.

I suggest that the proper questions we should be asking, in the process of reconstructing Aboriginal placenames, instead of those given (as dot points) above, are those listed below, of which the first three concern meaning and the last two concern the form:

  • • What is its locational referent? i.e. what specific place did the name originally refer to?
  • • What is its etymology? i.e. what is the ordinary-language meaning of the word in the relevant language?
  • • What is its etiology? i.e. what is the story behind the name? or, How did a particular word come to be the name of the particular place?4
  • • What was the pronunciation of the placename in the language from which it was taken?
  • • What would be an appropriate spelling of the name in an orthographical system that accurately represents the phonology of the relevant language?

How much meaning can we expect to reconstruct? I suggest that we will typically be able to reconstruct at least an approximate locational referent for a placename, especially if it was bestowed on the basis of a named Aboriginal location in the vicinity, but that this location may not be very precise. Regarding etymology, we may sometimes be able to posit a tentative meaning, but we should not normally expect to be successful, given the incomplete documentation of many languages and the fact that not all Aboriginal placenames have a clear etymology even in the traditional culture. With respect to the etiology, I would claim that only rarely could we expect to have access to this knowledge in regions of Australia where Aboriginal traditions have been disrupted.

Until the coming of Europeans to their homelands, the placenames used by Aboriginal people each had a function within an indigenous system of nomenclature, referring to sites (and presumably their surrounding area) that had significance for the location of resources, way-finding, personal and group identity, and traditional beliefs (from which the name may have been originally bestowed). When early settlers established a homestead or even a temporary stock station and needed a name to register their land selection and to distinguish it from that of their neighbours, they took a name used by the local Aborigines for some site in the vicinity (whose specific local reference and significance was probably often unknown to the newcomers) and applied it to a feature (homestead, whole property) that had significance in their own economic and legal system. The significance of the name in the European nomenclature may have undergone subsequent changes – for instance being transferred from the name of a property to that of a town or a river, a range of hills, a parish, a trig station, a nature park, or even a satellite tracking station. The form of the placename would have been imitated by the first Europeans who learned it from the local Aboriginal people, and it most likely suffered some distortion in pronunciation in the transmission across the linguistic divide. Different Europeans may have interpreted the Aboriginal word in different ways, or even if they had the same interpretation they may have spelled it in different manners. Other Europeans may have learned the name from Europeans but reproduced it differently in pronunciation and spelling – either unconsciously or deliberately. Some may have learned the word from its written form and introduced a new pronunciation based on the spelling.

The changes in the referential meaning and the spelling of placenames within the European system of nomenclature can be traced to some extent through the record of written documentation. The function of the placename in the earlier Aboriginal system of nomenclature, on the other hand, is largely inaccessible to us now. Our best evidence for recovering information on (especially the pronunciation and location of) Aboriginal placenames comes from the period of time – which may have lasted for several decades – during which there was an overlap in the use of the land between the traditional Aboriginal residents and the initial European settlers. This conception of the transmission of placenames is illustrated visually in Figure 5.2, where an originally Aboriginal placename and its reflected European placename is put on a timeline (appropriate for the ACT – see below).

Figure 5.2: Transmission of placenames over time

Aboriginal system

European system

Time line

1825

1860

2005

The first European settlers arrived in the regions now called the ACT and the Monaro in the 1820s.5 The first European land-holders each took up a large area of land and many of them named their properties with Aboriginal names. The gold rush at Kiandra in 1860 brought great movements of people through the district. The Robertson Land Act 1861 also led to the influx of many small landholders, most of whom, unlike the earlier large landholders, would not have had learned about their land directly from the local Aborigines. In recovering aspects of the Aboriginal nomenclature, therefore, greatest reliance should be placed on those historical sources that reflect knowledge dating from the 1820s until around 1860. In general, earlier documents can be assumed to better reflect the knowledge of the first Europeans who acquired placename information first-hand from members of the traditional cultural community. Furthermore, placenames have often undergone changes over time: they sometimes shifted in their local reference and came to be used to refer to slightly different places. They have also sometimes undergone changes to their pronunciation, some of which may be due to pronunciations based on spellings rather than on imitation of the Aboriginal pronunciation.

Intercultural transmission and historical transmission

There are two processes that need to be reconstructed if we are to reconstruct aspects of the traditional Indigenous placenames from their present-day usage. We might designate these processes ‘intercultural transmission’ and ‘historical transmission’. The names were transmitted from the Indigenous language and cultural system into the early Anglo-Australian system during the period of cultural overlap. This process of intercultural transmission typically involved a lot of misunderstanding, loss of information, reinterpretation of form and meaning, etc. Subsequently through historical transmission within Anglo-Australian culture the placename may have undergone changes in locational referent, pronunciation, and spelling. It is obvious that, before an attempt is made to reconstruct the intercultural transmission of a placename, one should undo the effects of historical transmission to recover the form and sense of the placename as it was first used in the Anglo-Australian system. The interpretive process of undoing the effects of historical transmission, which is a prerequisite to the reconstructing the intercultural transmission, requires primarily the methods of the discipline of history, whereas unravelling the intercultural (and especially interlingual) transmission depends largely on the methods that stem from the discipline of linguistics.

Historical transmission

What history offers

The historical study of documents can be expected to provide information on the circumstances of the bestowal of the placename in the Anglo-Australian toponymic system, answering such questions as: which Europeans first claimed the land as their property; when and by what authority was the claim recognized; when was the land surveyed; what use was it put to; when and by whom was a topographic feature (river, hill, etc.) named; when was a town gazetted? Historical sources can also yield information about changes in the application of names to sites, supposed sources and ‘meanings’ of placenames (including whether it was from an Aboriginal or a European source), the chronology of attestations of a given placename (in possibly differing spellings) in documents. Historians may be able to evaluate the relative reliability of various historical witnesses.6

It is primarily historians who have the tools to investigate such primary sources as land title records, correspondence with Colonial Secretaries, surveyors’ reports and maps, unpublished diaries, etc. For my work on placenames I have largely relied on secondary sources produced by historians, such as: local histories (e.g. Hancock 1972; Fitzhardinge 1975, Gillespie 1991; Andrews 1998; Moore 1999), biographies (e.g. Wilson 2001[1968]; Clarke 1986), editions of travellers’ journals (e.g. Lhotsky 1979; Robinson 1998), and compilations of historical documents (Young, Mundy and Mundy 2000).

Historical evidence for changes in the locational referent

It is important to realise that the local referent of a placename undergoes changes over time. The local reference may shift somewhat, as shown below; it may also undergo considerable expansion in its scope. (The most dramatic changes of referential scope can be found in the name ‘Canberra’ – see the section ‘Changes in locational referent’.)

Given the shift in local reference, it is important that reconstruction should begin with the earliest documented local referent. The earliest application of the placename to a site within the European landscape should be closer to the Aboriginal use of the placename than later local applications of the same name to related but different sites. Nevertheless, even if we assume that the first European to apply the name to the spatial domain relevant to European interests derived it from an Indigenous source, we still cannot be certain that the local Aboriginal people from whom the name was learned applied it to the exact same referent.

I think it is a fair assumption that the Europeans who first gave an (Indigenous) name to a pastoral property would have applied the name in the first instance to the focal area where the homestead was located, then extended this name to include the whole property. The same name may later have been applied to other natural features (plain, creek, mountain or range of hills) or institutions (towns, churches, post offices, schools) in the vicinity of the focal name. It is not unusual for a name in the European system to end up with a range of related designations. For example, the name ‘Tidbinbilla’, first applied to George Webb’s pastoral run, came to be associated with a mountain range, a river, a nature park and a Space Tracking Station (see further below).

Consider the name ‘Queanbeyan’. This name was first applied to Timothy Beard’s Quinbean station, from 1828, on the south side of the Molonglo River near its junction with the Queanbeyan River, i.e. around Oaks Estate. When in 1838 a town was newly gazetted some distance upriver, the name ‘Queanbeyan’ was applied to it. The name also came to be applied to the river that flows through the town. Knowledge about the earlier application of a name is necessary for any attempt to reconstruct its local referent in the system of traditional Aboriginal nomenclature.

Another placename that has undergone a shift is ‘Pialligo’. Pialligo was the site of a 1620-hectare property selected in 1825 by Robert Campbell’s overseer, James Ainslie. This estate was later called Duntroon, after an ancestral Campbell property in Scotland. The Pialligo name was later applied to a parish and now designates a suburb located near the Canberra airport. What was its original designation? According to one historical source, “The aboriginal name of the open plain on the northern slopes of which was built the homestead of Duntroon, was Pialligo” (Gale 1927: 10). One might wonder whether the name applied to the whole area of the plain or whether there was a more specific referent for the name, e.g. a particular site in or beside the Molonglo River (such as a traditional camping ground7); Mount Pleasant itself or a particular site on its slopes; or even Mount Ainslie. We would also like to know the source of John Gale’s information. (Given the European settlers’ preoccupation with unwooded plains which could provide pasturage for their livestock, one might reasonably surmise that they would have soon have extended any Aboriginal placename near their pasturages to the whole of the relevant plain.)

‘Tharwa’ is reported to be the Aboriginal name for the hill, first named Mt Currie by Alan Cunningham in 1824 and later re-named Mt Tennant after the bushranger John Tennant, who had his refuge there during 1827 (Moore 1999: 3, 144). We cannot know whether the Aboriginal name applied to the whole hill, a site on it, a site in the vicinity, or variably to all of these. In 1837 the name was applied in the European domain to a grazing property licensed to George Webb (Moore 1999: 32), who had built a home “on the ridge just north of the present site of Tharwa village” (Moore 1999: 59). In 1882 Tharwa became the official name of a postal receiving office (from 1894 a post office) at the Cuppacumbalong property (Moore 1999: 191). It has long been the name of a village.

Another property name for which we have a hint of more specific place referent in Aboriginal toponomy is ‘Gegedzeric’, near Berridale NSW, which was established by Richard Brooks in the late 1820s. George Augustus Robinson, the Chief Protector of Aborigines in the Port Phillip District, passed through the area in 1844 and noted: “Mr Brook’s station at Je.jetheric, the name of hill near homestead” (Robinson 1998: 131). Here the property was apparently named after the hill near which its homestead was built. The name was also applied (perhaps by the Europeans only) to the adjacent creek: “Buckly is on same creek as Brooks, called Jejetheric creek” (Robinson 1998: 132). More recently the name was applied to a church, St Mary’s Gegedzerick. This case is instructive: without the testimony of Robinson (which may have been based on the word of local Aborigines, whom he met there), we might conclude that the Indigenous name applied to a feature near the homestead rather than to the creek, but we would not know that it applied to the hill.

Historical evidence for an Aboriginal origin

The historical sources may tell us whether a placename had an Aboriginal origin or was carried over from an overseas, typically European, name.8 Usually European names are obvious and are not easily confused with possible Aboriginal words, and they typically have a fairly obvious motivation. Thus Robert Campbell’s Duntroon and Arthur Jeffreys’ Acton reflect family estates in Scotland and Wales respectively, Henry Hall’s Charnwood recalls a forest in England, and Herbert’s Naas was presumably named after the garrison town in Ireland. Anna Maria Bunn is said to have named St Omer (near Braidwood) after a village in France where her mother died. In the absence of clear documentation, however, we should not be too quick in assuming a European origin. I mention two cases in point.

‘Woden’ was the name given by Dr James Murray to a previously unnamed property in the Jerrabomberra area (near present-day Hume) that he purchased in 1837. The name is identical to the name of a pagan Anglo-Saxon god (German Wotan, Scandinavian Odin). To my knowledge there is no direct evidence that Murray named the place after the Germanic god – in spite of the confident assertion that he did: “This name was plainly the choice of a scholar as Woden was the Nordic God of wisdom, among other things. Dr Murray was to spend his life in the pursuit of wisdom” (Wilson 2001[1968]: 81). Unless there is documentary evidence, such as a diary entry, to confirm this origin, I believe we should treat it with scepticism, especially since it does not match the pattern of other European name bestowals. If a plausible Aboriginal source were available, it should be given equal consideration, given the Surveyor-General’s policy that Aboriginal names should be bestowed where possible (see Windsor this volume). Below (in the section ‘Etymological suggestions based on wordlists alone’) I suggest an Aboriginal source, which I claim deserves consideration equal to that of the European source.

Historical evidence for changes in pronunciation and spelling

Testimony of European pronunciation not reflecting that of Aborigines

For some placenames there is documentary evidence that the name used by Europeans does not accurately reflect the pronunciation of the Aboriginal word from which it was derived. Thus George Bennett’s 1834 comment “Yas (or according to the aboriginal pronunciation, Yar) Plains” (Bennett 1967[1834]: 165) alerts us to the fact that the final s of European pronunciation is a substitute for some kind of r-sound – probably a trill. Statements of early settlers such as “Giningininderra (to give the locality its full aboriginal nomenclature) Plains” (Samuel Shumack, quoted in Gale 1927: 83), “Boroomba should be Booraroomba” (Mowle 1891: 2), or “Giribombery (alias Giridibombera)” (Lhotsky 1979: 71) likewise alert us to the fact that at least some Europeans realised that their usual pronunciation of a placename did not accurately reflect that of the Aborigines.

Changes in spelling and pronunciation

Some of the variant spellings of a placename reflect a change over time in the pronunciation of the name by Europeans. Good historical method requires that only the spellings that represent the earlier pronunciations should be the basis for reconstructing the original pronunciation. There are several ways in which the form of placenames may be subject to change over the course of time. The name may become shorter, more ‘euphonic’, more English-like, etc.

One kind of simplification is the dropping of reduplicated syllables. The early Canberra property whose name has stabilised as ‘Ginninderra’ – although a post office opened there in 1859 used an official spelling ‘Gininderra’ (Gillespie 1991: xix) – is first recorded as having the first two syllables reduplicated; this reduplication was subsequently lost by a simplification of the name. The name is spelled ‘Ginninginninderra’ in the 1833 NSW Calendar and General Post Office Directory (Watson 1927: 37); ‘Ginnin Ginninderry’ [Creek] on Thomas Mitchell’s 1834 map of NSW (Fitzhardinge 1975: Illustration 1 opp. p. 16); ‘Ginnin-Ginnindera’ in 1834 by Lhotsky (1979: 65). These spellings all point to a six-syllable word something like ‘Ji.nin.ji.nin.de.ri/a’ (there is some doubt about the last vowel). There is some evidence for an early pronunciation with just five syllables, part of the second one being dropped out:9 ‘Ginginninderra’ was used in a 1831 letter by Catherine (Mrs G. T.) Palmer to the Surveyor-General (Gillespie 1992: 6), and ‘Jin Jin in derring’ in G. A. Robinson’s 1844 journal (Robinson 1998: 206). The antiquity of the pronunciation ‘Ji.nin.ji.nin.de.ri’ is guaranteed by an Aboriginal king plate that bears both the name of John Langdon, the original grantee of the land (after 1828), and an inscription “Mickey King of Gin and Gin and Derry” (Gillespie 1992: 9). It appears that the four-syllable form of the name was institutionalised only with the 1859 post office, some 30 years after the placename came to be used in the European mapping system.10 Another example of a reduplicated name that has been shortened is ‘Lacmalac’ near Tumut: “the Aboriginal name for Lacmalac was ‘Melacmelac’” (Snowden 2004: 39).

Another kind of shortening of placenames involves dropping out a single syllable, especially when two adjacent syllables were partially alike. The example cited below (see below in ‘The danger of not using early sources’) of ‘Coolalamine’ being replaced by the shorter form ‘Coolamine’ illustrates this process. A similar reduction has taken place with ‘Booroomba’, a property established by James Wright for his father-in-law, William Davis, in 1842 (Fitzhardinge 1975: 8), whose Aboriginal source had an extra syllable – ‘Booraroomba’ (Mowle 1891: 2) or ‘Boorooromba’ (Wright 1923: 31).11 Given a longer and a shorter version of the same name, we should consider the longer one to be more authentic. Thus for the placename ‘Jerrabomberra’, which was spelled ‘Jerrabombera’ on Sir Thomas Mitchell’s 1934 map of NSW (Fitzhardinge 1975: Illustration 1 opp. p. 16), we have from Lhotsky’s journal of the same year “Giribombery (alias Giridibombera)” as the name of Mr Palmer’s farm (Lhotsky 1979: 71). A decade later G. A. Robinson spelled it ‘Jerry Bunbery’ (Robinson 1998: 203). I would take the longer form (and the forms ending in a) to be intended as the proper name, with a pronunciation probably *Jeridibanbera.

An example of a placename becoming over time more ‘euphonic’, i.e. easier for English speakers to pronounce, is Bumbalong, a property near Michelago on the Murrumbidgee River. This was first recorded as “the Bunbilling run” in the 1830s and 1840s, but from the 1860s the property and local parish name were spelled ‘Bumbalong’ (Moore 1999: 119, 121). Note that the non-homorganic consonant cluster nb has been replaced by the more user-friendly sequence mb. Further, it seems that Australian placenames favoured the word-final sequence -along (cf. also Adelong, Binalong, Bukalong, Cambalong, Cuppacumbalong, etc.).12 One specific form of euphonisation is turning a long placename into a reduplicated structure, consisting of a shorter sequence that is repeated. Thus Tilba Tilba’s earlier form was recorded by Commissioner John Lambie in 1839-40 as ‘Tolbedelbo’ (Andrews 1998: 134) and by Stewart Ryrie in 1840 as ‘Tulbedelbo’ (Andrews 1998: 186). It appears that the second half had its vowels altered (e to i and o to a) and then the whole was repeated; since such a form is obviously easier to remember than the original, the substitution is a kind of simplification. Suggan Buggan, on the Snowy River in Victoria, would seem to have undergone the same process, judging from several early reduplicated spellings – ‘Soogum boogum’ from G. A. Robinson (1998: 130), ‘Chungan Bungan’ from John O’Rourke, interviewed in 1910 (Young, Mundy and Mundy 2000: 176) – beside the ‘Toogunbuka’ of Ryrie’s 1840 Journal (Andrews 1998: 175).

Aboriginal names sometimes came to be spelled as if they were English words, and sometimes even acquired a rationale for the English term (by a process that linguists call ‘folk etymology’). Thus Terence Aubrey Murray’s property was named ‘Collector’ apparently after the name of a waterhole on Collector Creek. His biographer says of the name of the farm:

Besides an obvious anglicizing of the ancient aboriginal word of Colegdar, there is another likely reason for the hastening of its debasement. With geographic logic, it became the place where wool-draymen collected so that they might form a train for mutual help… (Wilson 2001[1968]: 39, cf. p. 34).

Similarly the South Coast name ‘Bodalla’, earlier ‘Botally’, was sometimes spelled ‘Boat Alley’ (Gibbney 1989: 82), which appears to give it an English meaning.13 Equally pseudo-English is a spelling such as ‘Queen Bean’ for ‘Queanbeyan’ (Robinson 1998: 114).

Another example of the Anglicisation of spelling is ‘Delegate’ in the Monaro. Table 5.1 gives some early spellings, all of which make it clear that this was not the English word delegate that it later came to resemble. In fact the one spelling with initial Dz suggests that the placename actually began with an un-English sound – probably a laminodental sound.

Table 5.1 Early spellings of the placename Delegate

Spelling

Date

Source

Reference

Dilighet

1834

Lhotsky: heard at Limestone Cottage (Duntroon)

Lhotsky 1979: 70

Dziliket

1834

Lhotsky: copied from a manager Bath at Cooma

Lhotsky 1979: 105

Delicate

1838

Eyre, resident at upper Molonglo

Eyre 1984: 169-170

Dilligett

1840

Ryrie Journal II

Andrews 1998: 169

Delecate

1841

John Irving: “sketch of My Station on Delecate River Mannara”

Andrews 1998: 126

Delegat

1844

Robinson

Robinson 1998: 134

Delegat

1846

Lambie: itinerary

Andrews 1998: 128

Deleget

1860

J. C., letter to editor, Melbourne Age, 16 February 1860

Moye 1959: 5

Aboriginal placenames may be partially adapted to English. Thus beside early spellings of Tidbinbilla as ‘Tidbinbilly’ and ‘Tin.min.bil.le’ we find ‘Timman Billy’ and ‘Timan Valley’ (see (7) below). The last examples suggest that the name was sometimes treated as a sequence of two words, in which the substitution of Valley for Billy made it look like a plausible English placename.14 Another example is Wambrook, a property in the Monaro. This was recorded as ‘Womerob’ by Lhotsky in 1834, who copied the name from a manager Mr Bath at Cooma (Lhotsky 1979: 105). In 1839 Land Commissioner Lambie rendered it as ‘Wambrooke’ (Andrews 1998: 131). By the late 1840s it was being spelled ‘Wambrook’, according to Hancock’s “Alphabetical list of squatting runs in Monaro” (Hancock 1972: 53). I suggest that it was probably ‘Wamirub’, with the middle vowel sometimes omitted; then English speakers inserted a b in the mr sequence for ease of pronunciation (as in number vs. German Nummer) and the final b was changed to k to make at least the latter part of the word look like an English word, even a plausible placename ending in brook.

A clear case of the changing spelling of a placename is ‘Gungahlin’, now the name of northern residential district within Canberra. The name was first applied by William Davis to a parcel of land which he acquired in 1861 adjacent to his Ginninderra estate and on which he built a substantial homestead in 1862.

In 1862, William Davis called his new home Goongarline… Later it was sometimes spelt Gungahline. During the Crace family’s occupancy of the homestead and for some years afterwards, however, Gungahleen was the spelling used and the official name given to the school that was established in the area was Gungahleen School. (Gillespie 1991: xix)

The historic order of the spellings of the name is known: ‘Goongarline’, ‘Gungahline’, ‘Gungahleen’, ‘Gungahlin’.15 We can assume that the earliest form represents the closest approximation to the pronunciation of the Aboriginal word at its source.

The danger of not using early sources

The dangers of ignoring early sources, and general guessing at the etymology of a placename on the basis of inadequate knowledge, can be illustrated by the example of ‘Cooleman’. This is the name of an early property and associated plain, mountain, creek and cave in the upper Murrumbidgee region. John Gale, long-time editor of Queanbeyan newspapers and author of a history of Canberra (Gale 1927) surmises that the name was derived from coolamon, a term for a wooden dish which came into English from an Aboriginal language of southeastern Australia.

This plain [Old Coolaman Plain] is very extensive, formed of undulating ground, well-grassed and abundantly watered – a large creek flowing through it, besides containing springs and gilghi holes, from which latter circumstance, probably, it derived its name – coolamon being the aboriginal word for a large water vessel which these gilghies somewhat resemble in shape. (Gale 1985: 27)

Gale does not state from which language he knows the word. Nor does he claim to have derived this information from local Aboriginal tradition. For Gale’s etymology to work, it must contain a word from the local language. This is not documented. In fact the word coolamon/guliman is known from Kamilaroi in northern NSW (Ash et al. 2003: 94; Dixon, Ramson and Thomas 1990: 184). This word apparently entered the early NSW pidgin that developed as a contact language between Aborigines and early settlers and from Pidgin English became widely known among English-speakers, including Gale (for NSW Pidgin see Troy 1994).

The worst problem with Gale’s etymology, however, is that he did not start from the likely original form of the placename. The name originally contained an extra syllable, and the final syllable apparently rhymed with English coal mine rather than Coleman, according to the evidence of T. A. Murray, who established the station in 1838 (Gale’s visit was much later, about 1875).

The main station and its out-stations which Murray established at Cooleman Plain he called Coolalamine. This was the native name by which he knew the pastures. Later the station became known as Coolamine, while the physical features of the locality came to be spelt differently. Thus two spellings survive: Cooleman Plains, Cooleman Mountains and Cooleman Cave, but Coolamine station. Neither reproduces the ancient Aboriginal sounds, though Coolamine station preserves all but one syllable of it. (Wilson 2001[1968]: 109)

Gale’s etymology is purely hypothetical and is based on insufficient knowledge. It illustrates the dangers of (a) not starting from the earliest documented pronunciation and (b) relying on knowledge of Aboriginal words which may not have existed in the relevant local language.

Intercultural transmission and the linguistic reconstruction of the sounds

Reconstructing pronunciation: Methods from linguistics

Anyone who approaches the task of reconstructing placenames is immediately faced with the problem of variant spellings – beyond those that reflect European simplifications and adaptations over time. How are the variants to be explained? Is there a rule of thumb that can be applied mechanically, such as always preferring the very earliest spelling? The answer is not simple, since there are a number of different causes of the variability, which is a natural consequence of linguistically untrained people attempting to both perceive the sounds of a foreign language and represent them in written form using a spelling system that only imperfectly represents the sounds of a very different language – English.

The discipline of linguistics, especially its sub-branch phonology, yields insights into these factors. These transmission factors include:

  1. a. variable pronunciation of the phonemes of the Aboriginal language
  2. b. unfamiliar phonemes of the Aboriginal language, which are variably perceived by European recorders
  3. c. unfamiliar sequences of sounds of the Aboriginal language, which are variably interpreted by European recorders
  4. d. variable degrees of attention paid to phonetic details by European recorders
  5. e. differential use of the English spelling system to represent the same sound
  6. f. different dialects or languages of the European recorders

Each one of these factors can contribute to the different spellings of a placename. Linguistic reconstruction involves undoing the effects of these variables, accounting for the variants, and positing the most likely sequence of phonemes that the different writers were trying to represent using the spelling system provided by English.

Variable Aboriginal pronunciation

A fundamental principle of linguistic organisation is that the sounds of speech are organised to ‘contrast’ with one another. The principle of contrast means that words can be distinguished by the fact that they differ with respect to one or more of these distinctive sounds, which linguists call ‘phonemes’. Thus English pit and pet differ only by the ‘vowel height’ of i and e, whereas pit differs from bit only by the fact that the b is voiced but the p is voiceless (and aspirated). Languages differ in which distinctions of sound they make use of. Thus most Aboriginal languages do not use the difference between i and e or that between p and b to distinguish between words; rather i and e are part of the range of what is treated as a single vowel phoneme and p and b are likewise variant pronunciations of the same phoneme that just requires closure of the lips and the nasal passage (unlike m which also uses closed lips but allows air to flow through the nose). The vowels u and o are likewise not distinguished, nor the consonants t and d, k and g, ch and j. Thus, to take a hypothetical example, a word that is phonemically /bindu/16 might be pronounced, and recorded by English speakers, as [bindu], [bendu], [bindo], [bendo], [pindu], [pendu], [pindo], [pendo].17 In this case there are objective differences in the pronunciation of the same Aboriginal word – these differences being irrelevant to the speakers but noticeable to English hearers because English makes use of the phonetic differences. This mismatch between the Aboriginal and European languages accounts for a lot of the variant spellings. Spelling variants such as Gudgenby and Cudgenby illustrate the irrelevance of g vs k (which can also be spelled by c in English) and are an example of English spelling over-differentiating the sounds of the Aboriginal language.

Un-English Aboriginal phonemes differently perceived

Aboriginal languages typically make use of pronunciation distinctions which are not made in English.18 Thus there are typically two r sounds, one that is a lot like the normal English sound and the other that is a trill as in the Scottish pronunciation of English. Both would probably be spelled with an r by English speakers and the distinction would not be shown in writing. Here English spelling under-differentiates the sounds of the Aboriginal language.

In many languages there is a significant difference between an English-like t or d (which are treated as equivalent in the Aboriginal language) and one that is pronounced with the blade of the tongue against the upper teeth. It might sound something like the dth in width. European recorders may hear the sound as (a) a normal d or t, (b) a voiced th as in this or bathe, or (c) something un-English that requires a special notation such as dh. A placename such as ‘Tharwa’ or ‘Thredbo’ suggests that this dental sound was noticed and treated as the (somewhat different) English th sound. On the other hand, this sound seems to have been noticed only by Stewart Mowle for Brindabella, which “should be Berindhabella” (Mowle 1891: 2), and Tumut, which “should be Dumudth” (Mowle 1891: 2).19

Even more trouble for English speaker was caused by a palatal sound that was pronounced something like j or ch, but with a shorter phase of audible friction when the oral closure is released (if English ch sounds like t + sh, the Aboriginal sound is more like t + t + sh). The proper effect can be gained by pressing the tip of the tongue against the back of the lower teeth and then saying ch, with the blade of the tongue making contact with the roof of the mouth. This sound may be treated as an English voiceless ch or voiced j (which might be spelled as g as in George or dg as in bridge). On the other hand it may be heard, without any friction, simply as t or d or even as k or g. If it is perceived as different from any English sound, it may be represented by a combination of letters such as ty or dy or dj or tch or djh. The un-English nature of this phoneme was recognised by Stewart Mowle, who spelled it djh (as in djhatu ‘moon’) and the Polish scientist, John Lhotsky:

several syllables also commencing with the letter j would be far more accurately rendered by the Polish dź, which has no exact equivalent in the English tongue. (Lhotsky 1839: 157)

For sounds like this it is the variety of spellings which alerts us to the likely presence of an un-English sound. Thus the placename ‘Michelago’ has variant early spellings with the second consonant represented as k, ck, or c, in ‘Micalago’ (all of which point to a perception of k), tt (by G. A. Robinson, who obviously perceived a t), and c (‘Micilago’), cc (‘Miccelligo’), ch (‘Michaligo’) – which suggest identification with English ch. An Aboriginal sound which would account all for these different perceptions would be the palatal stop described above. Another placename for which a palatal stop can be inferred is ‘Tidbinbilla’ – see the discussion at (7) below and note Stewart Mowle’s (1891: 2) comment: “Tinnimanbilly should be Tchinbinbille”.

Aboriginal languages have a parallel palatal nasal sound, which sounds something like the ny in canyon or the ni in onion, but with the same tongue configuration as for the palatal stop. This sound caused trouble especially at the beginning of words. A placename in the Monaro was spelled both as ‘Umeralla’ and ‘Numeralla’; these presumably aim to represent initial sounds yu- and nyu- respectively. I infer that the word began with nyu- but some recorders missed the nasal part and only heard the y. I argue below (in ‘Different Aboriginal combinations of sounds differently interpreted’) that the placename ‘Namadgi’ probably began with this ny sound, some recorders hearing only the nasal (‘Namadgi’, ‘Namwich’) and others only the palatal glide y (‘Yammoit’). Again, it is the variable spellings which alert us to the palatal nature of the sound.

Another sound that gave trouble to English speakers was the trilled r when it occurred at the end of the word. G. A. Robinson wrote it r.r – his full stop usually indicates syllable division. The placename ‘Yass’ presumably ended in this trill, which was partly devoiced. This sound is probably what lies behind Hume and Hovell’s spelling ‘Yarrh’ (see Watson 1927: 18). According to Mowle (1891: 2), “Yass should be Yarr”.20 When the trilled r occurred before another consonant, some recorders heard a short vowel between the consonants. Variant spellings such as ‘Burobong’, beside later official ‘Burbong’, for a property near the Molonglo River east of Queanbeyan, alert us to the trill, which is spelled rr in most Aboriginal orthographies – thus we reconstruct the name as *Burrbang.

Different Aboriginal combinations of sounds differently interpreted

Another factor leading to variable spellings is the fact that the sound in Aboriginal languages may be ordered into sequences that are unlike those of English. English has the velar nasal ng sound (as in sing and singer) but it never occurs at the beginning of a word, as it does commonly in Aboriginal languages. Word-initial ng is typically misheard by English (and other European) recorders. Sometimes it was ignored altogether; other times it is represented as another nasal n, ny, or m; or it may be heard as k, w, or y. Since Aboriginal words rarely begin with a vowel, and often begin with ng, and ng was a consonant that was often missed, one can suspect that placenames recorded with an initial vowel – such as Adaminaby, Arable, Adjungbilly, Ajamatong, Amungula – actually began with ng. Thus for ‘Ironmungie’ (with anglicised spelling) near the Victoria-NSW border, with early spellings ‘Inemongee’ and ‘Eiemmondgy’ – apparently perceived as *ayanmandyi – we would plausibly reconstruct *Ngayanmandyi. See also Ulladulla in (6) below.

The combination of n with a following b/p or g/k is more common in Aboriginal languages than in English. European recorders sometimes substituted the more natural sounding sequence mb for nb. Thus if we find a variant spelling with nb it is likely to be a more authentic reproduction of the Aboriginal pronunciation. Hence G. A. Robinson’s (1998: 203) spelling of ‘Jerrabomberra’ as ‘Jerry Bunbery’ may be more accurate with respect to nb than is the usual mb.

When a palatal consonant such as dy/ty or ny followed a vowel a or u, an automatic transitional vowel i was often heard and reflected in the spelling, while the palatal consonant was written as if it were simply d/t or n. The spellings thus suggest an analysis ait, ain, uit, uin – with a diphthong plus consonant – instead of just a vowel followed by a palatal consonant (in our spelling system): ady, any, udy, uny. This is the source of spellings like ‘Kalkite’ (-ady), ‘Jindabyne’ and ‘Goongarline’ (-any), ‘Wamboin’ (-uny), and ‘Yammoit’ (vs ‘Namwich’, with the glide i interpreted as the main vowel and the vowel u as a glide w, for what I reconstruct – at (5) below – as *Nyamudy for the placename usually spelled ‘Namadgi’). The name ‘Wamboin’, which is probably not a local placename but taken from the Wiradjuri word wambuny ‘kangaroo’, shows the vowel u, which has a common variant pronunciation as o followed by a palatal nasal, with a transitional glide that is interpreted by English speakers as combining with the preceding vowel to form a diphthong oi, while the palatal nasal ny is interpreted as a simple n.

Sequences in Aboriginal languages of vowels (a, i, u) plus glide (y, w) plus another vowel are often interpreted by English speakers in terms of diphthongs such as ai, au which are sometimes followed by another vowel and sometimes not. Thus a sequence aya is interpreted ai.a, i.e. a diphthong plus vowel, and may be spelled as ia (as in via), or ya (as in Ryan) – as ‘Pialigo’ vs ‘Byalegee’ for the name now spelled ‘Pialligo’. A sequence like awu is likely to be perceived as just the diphthong ou, as in ‘Bendoura’, a property near Braidwood (Clarke 1986: 59). A sequence iya is perceived as simply i.a, without the intervening glide, as in ‘Quinbean’, the earliest spelling of ‘Queanbeyan’.

Differences in attention to phonetic detail

Some of the spelling variants are the result, I suggest, of the differential attention that recorders devoted to phonetic details, even though they were capable of hearing the proper sounds. Thus Stewart Mowle was exact in noting especially the dental pronunciations (spelled with dh or th) missed by other recorders, insisting on ‘Doomut-th’ for ‘Tumut’, and ‘Berindhabella’ for ‘Brindabella’, and ‘Condhoware’ for others’ ‘Condore’. Yet he was apparently inattentive in his version of ‘Queanbeyan’ to both the place of articulation of the nasal (nb rather than his mb) and the quality of the vowel in the unstressed second syllable. His spelling ‘Cuumbean’ presumably is intended to convey a pronunciation something like [‘ku am ‘bi an], with four syllables and the stress on the first and third. Other people’s spellings – ‘Quinbien’ and ‘Queenbeenn’ (Cross 1985: 1) and ‘Queen Bean’ (Robinson 1998: 14) – suggest rather [ku in bi an], or a form that combines the first two syllables into [kwin] and/or the last two into [bin]. Since English tends to reduce a vowel in an unstressed syllable to an indistinct sound as in the, speakers of English when hearing an Aboriginal word, would not pay attention to the quality of the unstressed vowel and would repeat it to themselves as the indistinct vowel, and write it down accordingly. The same tendency can be seen in Mowle’s ‘Arralumna’ (and modern ‘Yarralumla’) vs. the common earlier spelling ‘Yarrowlumla’ (also ‘Yarrolumla’) and probably in the difference between the a in the second syllable of ‘Jerrabomberra’ vs the version with i recorded by the Polish scientist Lhotsky in ‘Giribombery’.

Different spelling options provided by English orthography

Even if different European recorders intended to write the same sound – and perceived this sound as one shared with the English language – there is still scope for variable spellings, since the English spelling system does not provide a one-to-one relation betweens sounds and letters.

There is a one-to-many relation between certain sounds and their orthographic representation. Several of these will be illustrated, with examples from placenames. The sound [k] can be spelled as c, ck, or k. This accounts for the variability of the first letter of ‘Kongwarra’ vs ‘Congwarra’ and ‘Kowan’ vs ‘Cowen’. Similarly the [j] sound can be spelled as j, dg, or g (before i or e). This accounts for the difference between ‘Jerrabomberra’ and ‘Giribombery’, ‘Ginninginninderra’ and ‘Jin Jin in derring’, ‘Murrumbidgee’ and ‘Murrumbeeja’ (J. J. Moore letter of 16 December 1826 quoted in Fitzhardinge 1975: 5). The long vowel [u:] can be represented by u or oo; thus the explorer Hovell’s 1824 diary says of the Tumut River: “The natives called the river Doomut or Tumott, which is the aboriginal meaning the ‘camping place’” (Snowden 2004: 16, italics added). Similarly a long [a:] can be spelled with either ar or ah, as reflected in early spellings of Gungahlin as ‘Goongarline’ and ‘Gungahline’. The indistinct short vowel shwa [ə], which in English is typical of unstressed syllables, may be rendered as a or er, as seen in ‘Cuppacumbalong’ vs ‘Cuppercumbalong’ and ‘Cuppercumberlong’. The sequence [yu] can be spelled yu, u, or eu: this accounts for the different beginnings of the names ‘Uriarra’ vs an 1844 spelling ‘Yule Yarra’ (Salisbury 2000: 265, from St John’s burial register); ‘Eucumbene’ vs ‘Ucumbean’ (Hancock 1972: 51-53) vs ‘Yuiquimbiang’ (Lhotsky 1979: 105); and ‘Eurobodalla’ vs ‘Urabadella’ (Andrews 1998: 133). We’ve seen above how the Aboriginal language sequence [aya] was perceived as [ai.a] and spelled with either ia or ya in ‘Pialligo’ vs ‘Byalegee’.

The English spelling system also involves many-to-one relations between sounds and letters. This results in ambiguous spellings whose intended sounds may be irrecoverable. Thus the sequence of letters ng can represent (a) the single sound of a velar nasal [ng], as in singer; (b) the sequence of sounds [ng+g], as in finger or hang-glider; or (c) the sound sequence [n + g], as in ungrateful or in-group. Thus from the different spellings ‘Goongarline’, ‘Gungarline’, ‘Gungahleen’, ‘Gungahlin’ we cannot tell which of the three possibilities was intended by the early recorders. The same applies to ‘Bungendore’ and ‘Bungendow’.

Dialect differences between European recorders

More difficult to assess is the contribution of the different dialects spoken by the Europeans who wrote down the placenames. The only relevant difference in consonants between English speakers would have been between those from the north and west of the British Isles who pronounced an r at the end of a word (as in car) and before a consonant (as in cart) and those who didn’t but interpreted r in this position as an indicator of the length of the preceding vowel. An r-less speaker might nevertheless perceive an r sound in an Aboriginal word and might still spell a name *Dharrwa (if that was the original form) as ‘Tharwa’. But if the placename was only *Dhaawa, with a long vowel but no r, only an r-less speaker would spell it with ar as a representation of the long a. If we don’t know the dialect of the first recorder, and there are no variant spellings, we are unable to decide whether the Aboriginal name contained an r. In ‘Goongarline’, which was later spelled ‘Gungahline’ and ‘Gungahleen’ (and now ‘Gungahlin’), the ar marks a long vowel later spelled with ah; at any rate the sequence r + l does not occur in most Aboriginal languages. The combination er in unstressed syllables is pronounced like the u of fun only in r-less dialects of English; speakers of such dialects sometimes used er to indicate a vowel in unstressed syllables of Aboriginal placenames, such as ‘Cuppercumberlong’ vs ‘Cuppacumbalong’.

Perhaps of more consequence for placenames is the variation with respect to the vowel spelled u. Speakers from the northern parts of the British Isles pronounced the u of butter like the oo of book in Southern British English. For northern speakers presumably u would never have been used to represent the vowel of sun, but might have been used to spell the Aboriginal u vowel for which others English speakers would have used oo. Thus a placename spelled ‘Burra’, if recorded by a northern dialect speaker, might be the same as the one given as ‘Boo.rer’ by G. A. Robinson as the home of a Limestone Black (Robinson 1998: 204). (Applying this principle may be difficult, since it may not be known which European first recorded the name, which part of the British Isles they came from, exactly what dialect they spoke, and, even if these are known, what were the particular features of their dialect during the time they were growing up.21)

The spelling of a placename, once it is established, may influence the pronunciation by people who learn it from a map. Thus a name with an original soon-like pronunciation can, if it is spelled with u, come to have a sun-like pronunciation. This has happened to the placename which was first spelled ‘Goongarline’ but changed to ‘Gungahleen in the usage of later owners (the Crace family) and the local school (Gillespie 1991: xix).

A dialectal (especially Irish) pronunciation of the usual English diphthong ai as oi is obviously what accounts for the spelling of ‘Jindabyne’ (also spelled ‘Jindabine’, ‘Jindebine’, ‘Jindibine’, ‘Jinderbine’, ‘Genderbine’) as ‘Jindaboyne’ (see (1) below).

Comparing the variant spellings

Collecting variants

Whereas for historical reconstruction of the transmission of a placename through its period of European use it is early records which are most valued, for reconstructing the name across the linguistic divide it is rather the number of variants that is important. Especially for reconstructing pronunciation there is safety in numbers, since different recorders made different mistakes in hearing the sounds of words that were in a foreign language as well as using different strategies for spelling the perceived sounds of the placename. We are especially interested in collecting as many different spellings as possible, provided they are each likely to represent different European perceptions of the same word as heard from Aboriginal speakers (and not just a replication of a name learned from other Europeans who had already decided on a pronunciation).

Matching sounds

The basic procedure is one of matching. We line up the sequence of sounds used in the alternative spellings of the word so that sounds that occur in the same relative position in the word are lined up in a single column (some spellings may leave out a sound or use two letters to represent one sound). For each of the resulting sets of ‘corresponding’ sounds we propose an original sound from which each of the attested reproductions is plausibly explained in terms of either a faithful replication or an alteration resulting from one of the transmission processes described above in ‘Reconstructing pronunciation: Methods from linguistics’. Ideally we should aim to propose some reasons for each deviation from the reconstructed original sound. A fundamental principle used here is one taken from the study of the copying of manuscripts, which is expressed in Latin as lectio difficilior potius ‘the more difficult reading is preferred’. That is, we start from a sound from which the other versions are explicable in terms of simplifying ‘scribal errors’ (as they would be in manuscript copying) or understandable mis-replications of aural perception, spoken imitation and orthography representation. The procedure is illustrated in the following paragraphs.22 The reconstructed word is given in the final line.

When it comes to reconstructing the form of a placename, there are several targets. (The proper spelling is not strictly one of these – since spelling did not exist in the oral culture from which the name was derived.) Do we reconstruct the pronunciation or the phonology, i.e. the sequence of phonemes which were relevant in the source language? I suggest that we try to reconstruct the pronunciation, and if we know enough about the phonology of the language, we can then also reconstruct the phonemic representation of the word.

System of sounds used in reconstructions

The set of presumed consonant phonemes (distinctive sounds) for south-eastern NSW languages is given in Table 5.2, with a suggested spelling of each phoneme. Some of the ways that consonants may be spelled in early sources are given in Table 5.3. The likely vowel phonemes are given in Table 5.4, along with English words which illustrate their normal values. The mid vowels (e and o) are included, even though they are likely to have been non-distinctive variants of the high vowels i and u respectively. If there was a distinctive set of long vowels these can be spelled ii, uu, and aa, with values of English beat, boot, and Bart. The stress (or emphasis) was usually on the first syllable. It is assumed that every syllable begins with a consonant sound.

Table 5.2: Consonant system and one possible standard way of spelling the sounds

lips

tongue tip

tongue blade & teeth

tongue blade & palate

velar

Stops

b

d

dh

dy

k

Nasals

m

n

nh

ny

ng

Laterals

l

(lh)

Trill/tap

rr

Approximant

w

r

y

Table 5.3: Consonant spellings used in early sources

Stops

b,p

d,t

dh,th,dth,t,d

ch,tch,dg,dgh,djh,jh,gh,j,t,d,c,k,g

k,c,ck, g,gg

Nasals

m

n

nh,n

ny,n,ng,ni,ne

ng,n,m,Ø

Laterals

l

l

Trill/tap

rr,r,l

Approximant

w,u

r,rr

y,i

Table 5.4: Vowels of Aboriginal languages, with English words illustrating their values

Front

Central

Back

High

i pit

u put

Mid

(e pet

o port)

Low

a part

Examples of reconstruction

This section will illustrate reconstruction of pronunciation with several examples discussed in detail. We begin with the relatively easy example of Jindabyne. Variant spellings are given in (1). We will work across the columns of corresponding letters.

(1)

J

i

n

d

a

b

i

n

e

J

i

n

d

e

b

i

n

e

J

i

n

d

i

b

i

n

e

J

i

n

d

er

b

i

n

e

G

e

n

d

er

b

i

n

e

J

i

n

d

a

b

y

n

e

J

i

n

d

a

b

oy

n

e

*Dy

i

n

d

a~i

b

a

ny

First, the difference between J and G simply reflects two different options for spelling the Aboriginal dy sound, which is identified by all recorders with the English j sound. In the second column, the difference between i and e reflects variability in the pronunciation of the Aboriginal phoneme i, whose pronunciation ranged over i and e. The variable spellings of the second vowel reflect the fact that it was in an unstressed syllable, was probably pronounced with less prominence, but more importantly, was in a position where English would have an indistinct shwa vowel. The a and er spellings reflect this indistinct vowel and probably indicate lack of attention on the part of the recorders. One might be tempted to conclude that the spellings with i and e therefore reflect a more accurate observation of an Aboriginal i phoneme. Our confidence in such a conclusion is weakened, however, by the fact that the spellings with i and e are from G. A. Robinson, who also used the spelling ‘Jindabine’. At the end of the word, the spelling ine (and the later yne) obvious indicates an intention to spell a diphthong ai followed by an English-like n; this is likely to reflect an Aboriginal vowel a followed by a palatal nasal ny with a transitional glide that sounds like i. The spelling with oy (in RASA Manuscripts 1900) presumably reflects a substitution of oi for ai that is characteristic of speakers of certain English dialects, especially that of Ireland.23

Example (2) gives spellings of the early name of Duntroon station, which is now a nearby suburb spelled ‘Pialligo’. The difference between initial p and b simply reflects the variability in what was a single phoneme in the Aboriginal language. The letter i in this position before a vowel is likely to signal the sound of the English diphthong ai, as in violin. The y of Stewart Mowle’s (1891: 2) ‘Byalegee’ points in the same direction. Since Aboriginal languages do not have diphthongs, this sequence is interpreted as the vowel a followed by a consonant y. The difference between single and double l would not be relevant in an Aboriginal language; in English it is just a signal that the preceding vowels is not to be pronounced as the ei diphthong of pale. We can reconcile the vowels of the third syllable if we assume the a stands for a sound ei or better a Scottish-like e: as in lake and both i and e for i: as in leek. The recorders were probably aiming for the quality rather than the length of the vowel sound, so it need not be reconstructed as a long vowel. The consistent g presents no problem, but since k and g are not distinctive the phoneme could be represented by either letter; using k in the spelling system allows us to reserve g for the ng sound, while using nk and ngk unambiguously for combinations of a nasal and a velar stop. The representations of the final vowel are hard to reconcile; the most consistent spelling is o, which would suggest the phoneme u which can be pronounced as the o of halo. The spelling with a, as in sofa, could indicate lack of attention to the vowel of an unstressed syllable and substitution of the English indistinct vowel that would be used in this position.24 Mowle’s ee is a problem. If it were based on a misreading of a handwritten u, it would fit with the o spelling. Otherwise it may represent an English habit of using i interchangeably with the indistinctive shwa vowel (of sofa). If we plump for the spelling o as the truest reproduction and treat both a and ee as careless replications of an unstressed word-final vowel, we can reconstruct the phoneme u. Our second last line indicates the reconstructed phonetics and the last line represents this in terms of the likely phonemes spelled according to our conventions. The slash in p/b and i/e indicates variants in the Aboriginal system.

(2)

p

i

a

l

a

g

a

p

i

a

l

a

g

o

p

i

a

ll

a

g

o

p

i

a

l

i

g

o

p

i

a

ll

i

g

o

b

y

a

l

e

g

ee

*B/P

ai

a

l

i/e

g

o

phonetic

*B

a

y

a

l

i

k

u

phonemic

In (3) the different spellings of the Monaro pastoral runs (for which see Hancock 1972: 47, 53) are likewise confined to the vowels. The last syllable uses two notations to indicate what was perceived as a long a (as in bar or bah); the perceived length may simply reflect a real long vowel phoneme in the original language or merely indicate that the vowel is not the reduced vowel found a similar position in English words such as member. The first vowel is apparently perceived either as the o of tow or the u of too, followed by an indistinct vowel, possibly spelled with e or not perceived at all. The uncertainty of the Aboriginal vowel is shown by the V (for vowel) of the reconstruction. Since vowels do not occur adjacent to one another in Aboriginal languages, we posit a glide w after the o/u sound, which is spelled u in the orthography we are using. The most likely sequence of phonemes in *Muwinba, but *Muwunba and *Muwanba are also possibilities.

(3)

M

ow

e

n

b

ar

1840

Lambie

M

o

e

n

b

ar

1844

Robinson

M

ow

e

n

b

ah

1848-50

M

oo

n

b

ar

1848-50

*M

o/u

V

n

b

a:

phonetic

*M

u

w

i

n

b

a

phonemic

Example (4) gives spellings for the property called ‘Micaligo’ and village now spelled ‘Michelago’. Here the first spelling can be taken as a mis-hearing of the first nasal consonant and an interpretation of the name as a combination of English nickel and eagle. It shows that the first and third syllables were stressed; the second and fourth show the indistinct vowels of English unstressed syllables. The consonants m, l and g are consistent in the other sources. The first vowel is consistently spelled as i in miss; the third vowel suggests variability between vowels with the quality of i: (of leek), i (of lick), e (of leg), ei or Scottish e: (of lake). The last vowel shows o (of leggo). The second vowel shows the most variation in spelling; one recorder, G. A. Robinson even spells it four different ways (unless his handwriting has been misinterpreted). This is partially caused by the lack of attention paid to unstressed vowels. The best guess would be the front (i/e) vowel, since this is heard by the Pole Lhotsky, Lambie, Eyre, and Robinson with his e and y spellings. The greatest problem is the second consonant, for which Robinson heard t, Lhotsky and Ryrie the k sound (even if spelled with c before a), and the Deputy Surveyor apparently ch. We cannot be certain whether Lambie’s c (before i) and Eyre’s cc (before e) is meant to indicate the k or the ch sound or something un-English. This variability is indicative of the palatal stop phoneme dy of Aboriginal languages, which occurs here in its voiceless variant ty.

(4)

N

i

ck

e

l

ea

g

le

1833

NSW Calendar

M

i

k

e

l

e

g

o

1834

Lhotsky

M

i

c

i

l

a

g

o

1839

Lambie

M

i

ch

a

l

i

g

o

1839

Deputy Surveyor

M

i

c

a

l

i

g

o

1840

Ryrie

M

i

tt

a

l

a

g

o

1844

Robinson

M

i

tt

e

l

a

g

o

1844

Robinson

M

i

tt

y

l

a

g

o

1844

Robinson

M

i

tt

o

l

ay

g

o

1844

Robinson

M

i

cc

e

ll

i

g

o

1840s

Eyre

M

i

c

a

l

a

g

o

1848-50

list of runs

*M

i

ty

i/e

l

i/e

g

o

phonetic

*M

i

dy

i

l

i

k

u

phonemic

Example (5) matches several terms that may refer to the name now known as ‘Namadgi’, as in the Namadgi National Park – gazetted in 1984 and covering the southern 40 percent of the ACT (Garnett and Hyndes 1992: 5) – and Mount Namadgi, a bogong moth aestivation site where there are “well defined Aboriginal stone arrangements” and where “evidence … of Aboriginal activities is the most visible within the high country area” (Garnett and Hyndes 1992: 11) – which was not so named until around 1980 (Alder 1989: 34). In May 1829 Assistant Surveyor R. Dixon “met … several tribes from Moneroo and Nammage” (quoted in Flood 1980: 9, 301). In 1833 Lhotsky saw the “Namadgi range” from Duntroon.

This name was also used to designate the local Aboriginal group. About 1827 William Edgar Riley witnessed a “Corobborie at Tuggranon Isabella Plains”; his write-up of the event begins with the words: “The Namitch tribe of natives was assembled here” (Lamb 2006: 256). A list of Aboriginal people receiving blankets at Janevale (Tuggeranong) in 1834 gives ‘Namwich’ as the name of the tribe of 60-70 people whose district includes the “mountains beyond the Murrum-bid-gee” and the Limestone Plains (quoted from Jackson-Nakano 2001: 55, who identifies Namwich with the name ‘Namadgi’). G. A. Robinson, touring through the area in 1844 (see Mackaness 1941, Robinson 1998) called the group the Yammoit-mittong or “the Yammoit Tribe”; since he generally used mittong as a term designating ‘group’, this suggests that ‘Yammoit’ is the name of the place with which they were associated. He also once referred to the ‘Nam mit tong’ tribe (Robinson 1998: 204); this spelling may hide a version of the same name, perceived this time as *Nammit, with one of the two mi syllables omitted by mistake from an intended *Nammit mittong. He also lists, in his census of the group met at Yarralumla on 12 September 1844, the home of a Jemmy Bo.lore.re as ‘Yamoke tower’; this name, I suggest, may consist of an element tower, probably something like dawurr25 added to the same placename that he spelled ‘Yammoit’ in the group name.

(5)

n

a

m

i

tch

Riley 1827

n

a

mm

a

ge

Dixon 1829

n

a

m

w

i

ch

Janevale blanket list 1834

n

a

m

a

dg

i

Lhotsky 1833

y

a

mm

o

i

t

Robinson July 1844

y

a

m

o

ke

Robinson Sept.1844

n

a

mm

i

t

Robinson Sept. 1844

*Ny

a

m

o

(i)

ty

(i)

phonetic

*Ny

a

m

u

dy

phonemic

The variability of the first letter suggests the palatal nasal sound ny, with either the palatal (y) or the nasal (n) feature of sound being perceived, but not both. The spellings of the last consonants suggest a palatal stop sound that was heard as an English ch, j, t, or k. The i-quality that accompanies palatal sounds was perceived either after the sound (in ‘Namadgi’) or as a transition from the preceding vowel sound u/o (in ‘Yammoit’ and ‘Namwich’); in the latter the sequence ui was interpreted as if i is the main vowel and u is the glide w. The middle a of ‘Namadgi’ would reflect the lack of attention paid to the vowel of an unstressed syllable, as would the i of ‘Nammit’, if this form is real.

Example (6), Ulladulla, illustrates, among other things, the principle of lectio difficilior. Surveyor Florance’s first record of the harbour (also called Wasp Harbour), based on a survey of May 1828, was spelled with an initial w. The first settler, Thomas Kendall, wrote in a letter of July 1828 to the Surveyor General that it was “called by the natives Nulladolla”; he heard an initial n. Alexander Macleay, writing to Surveyor General Major Mitchell in September 1828, also spelled it with an initial n, ‘Nulla Dollo’. Hoddle’s survey of December 1828 gave the name as ‘Ulladulla’, with no initial consonant. This variability points persuasively to the velar nasal sound ng, which does not occur in this position in English, was frequently not heard at all, or perceived as just an n or as something else, for example w by Florance. The ng sound is the most difficult of the various initial sounds we might reconstruct, and is therefore the sound most likely to have been replicated incorrectly by various substitutions or by omission. The variability between Florance’s rr and the ll of other recorders suggest that the Aboriginal sound in this position in the name may have been something unfamiliar to English speakers, such as a tap r, which was perceived as an l perhaps under the influence of the earlier l in the word. A tap r would have been a variant of the Aboriginal rr phoneme, which might also be pronounced as a trill. The first vowel appears to be the rounded vowel of toll or tool (o and u were usually freely varied in Aboriginal languages), according to Florance (my source attributes both spellings to Florance). If this is correct, the u of Thomas Kendall, who was from Lincolnshire, and Alexander Macleay, who was from Scotland, must represent the north-country English accent in which dull rhymes with pull. Both Lincolnshire and Scotland are north of the line from Merseyside to The Wash which marks the southern limit of the consistent pronunciation of u as oo (Wells 1982: 251, 336).26 The third vowel of this placename must be a kind of a-vowel, spelled o by Kendall and Macleay but u (as in dull) by Hoddle. Florance’s e suggests a slightly different sound, probably the sound of derrick. This must have been an optional variant of a, since the other recorders’ spelling with o or a points rather to just the a sound. I suggest that the a was optionally moved to the e position under the influence of the preceding consonant. But for this to have been the cause the sound was unlikely to have been an ordinary d, but rather a laminodental dh, produced with the body of the tongue high in the mouth. This appears to be confirmed by the reported Aboriginal pronunciation “Ullatha Ullatha or Ullada Ullada” (Shoalhaven Holidays n.d.: 43), where th and d indicate different perceptions of a kind of voiced th.27 Finally, the last o of Macleay’s ‘Nulla Dollo’ may represent a misreading of Kendall’s a, especially since it appears that Macleay purchased his land without having visited it. The modern pronunciation of the name with the first u as in gull must be a based on the spelling. The fact that a derivation of the name from holey dollar was considered indicates that the first vowel of ‘Ulladulla’ was once pronounced more like goal than like gull, which supports the reconstruction given above.

(6)

W

o

ll

ah

d

e

rr

ah

May 1828

Florance

Nulladolla 1988: 4

W

oo

ll

ah

d

e

rr

ah

Nulladolla 1988: 5

N

u

ll

a

d

o

ll

a

Jul.1828

Kendall

Nulladolla 1988: 7

N

u

ll

a

D

o

ll

o

Sep.1828

Macleay

Nulladolla 1988: 12

U

ll

a

d

u

ll

a

Dec.1828

Hoddle

Nulladolla 1988: 5

U

ll

a

th

a

U

ll

a

d

a

*Ng

u

l

a

dh

a

rr

a

There are many problems with reconstructing the name which has been institutionalised as ‘Tidbinbilla’ – see (7). For the first consonant, variants with j and tch show that a palatal sound was missed by those who rendered it by plain t; Mowle’s spelling with tch indicates that he sensed that it was somewhat different from the sound of English ch – if English ch consists of the stop t followed by a fricative sh (t + sh), the Aboriginal sound had a longer stop phase, t + t + sh). Wright’s variant e for the first vowel (as well as for the second vowel) shows the inherent variability in the pronunciation of the Aboriginal phoneme i, which had e as a possible pronunciation. For the second vowel, the evidence of several witnesses points to i (as in English bin); spellings with a that – according to church records (Salisbury 2000: 156, 158, 258) – were used in early names of the property established by the Webb family, ‘Timman Billy’ and ‘Timan Valley’, would then reflect lack of attention to the quality of a vowel in an unstressed syllable. The same may apply to spellings with a as the final vowel, since there are several earlier spellings that suggest rather the i sound. The biggest issue for reconstruction is the second and following consonant: there appears to be support both for a pronunciation with db and for one with nm, with simplification by omitting the n or inserting a vowel between the consonants for ease of pronunciation (in the form ‘Tinnimanbilly’ condemned by Mowle). If we accept Mowle’s ‘Tchinbinbille’ as the ‘correct’ spelling of the Aboriginal name, we should reconstruct *Dyinbinbili; but then it is hard to explain how this came to be pronounced by Europeans with db and nm rather than just nb and mb (as in ‘Canberra’). It is possible, it seems to me, that an n in this position, after a stressed vowel, could variably be pronounced as dn (as happens in other Aboriginal languages); thus the name would have been heard as either *Dyinminbili or *Dyidnminbili. The latter form would cause processing difficulties for English ears and possibly lead to interpretations such as the common *Dyidbinbili or even Mowle’s *Dyinbinbili. The case for n being optionally pronounced dn is strengthened by the report that local land-owner George de Salis consistently pronounced the name as Tidnambilly (Jackson-Nakano 2005: 41), and by variability between womme and wobme given by Robinson (2000: 193-194) for ‘beat’ in his Maneroo wordlists.

(7)

T

i

d

b

i

n

b

i

ll

y

1841

Census, Moore 1999: 220

T

i

n

m

i

n

b

i

ll

e

1844

Robinson 1998: 205

T

i

mm

a

n

B

i

ll

y

1850

Salisbury 2000: 156

T

i

nn

i

m

a

n

b

i

ll

y

1840s

Mowle 1891: 2

J

e

d

b

i

n

b

i

ll

a

1850s

Wright 1923: 38

J

e

d

b

e

n

b

i

ll

a

1850s

Wright 1923: 61

Tch

i

n

b

i

n

b

i

ll

y

1840s

Mowle 1891: 2

T

i

d

b

i

n

b

i

ll

a

1800s

county map, Moore 1999: vi

T

i

dn

a

m

b

i

ll

y

de Salis

*Dy

i

dn

m

i

n

b

i

l

i

phonetic

*Dy

i

n

m

i

n

b

i

l

i

phonemic

Reconstructing the etymology

Prerequisites

Attempts to answer the question of what a particular placename ‘means’ should only be attempted after its pronunciation is reconstructed. Otherwise a meaning is supplied for a phantom name! An example of etymologising from a false pronunciation was given above in ‘The danger of not using early sources’ with respect to Cooleman, for which John Gale’s starting point (roughly Kulaman in our orthography) lacked a syllable and had the wrong final part of the name attested earlier as Kulalamany. Likewise, denials of proffered etymologies are also sometimes made on the basis of a faulty starting pronunciation. Thus Appleton and Appleton (1992: 55) cast doubt on the supposed Aboriginal origin of the name ‘Canberra’ on the grounds that “no such word has been recorded in any Aboriginal language”. This conclusion may need revision in the light of our reconstruction in the final section.

There a basically two kinds of evidence for positing the meaning of a placename. The first is early testimony based on information provided by the relevant Aboriginal people. The second is comparison of the reconstructed name with attested vocabulary of the relevant language. The greatest certainty is achieved when both kinds of evidence are available and support each other, as in the case of Jimenbuen discussed below.

Assessing the reliability of testimony

It is not easy to know whose testimony about meanings of placenames can be trusted. I suggest two factors that add to our confidence. In the first place, reliance should be placed on the word of the family of the earliest settlers who lived in the named area, provided that they interacted with the traditional Aboriginal occupants of the land – such as the Crisp family of Jimenbuen. Secondly, considerable trust should be accorded to visitors who got their information directly from Aborigines. Thus George Augustus Robinson, the Chief Protector of Aborigines in the Port Phillip District, travelled through south-eastern NSW in 1844 and talked to several groups of Aborigines, using Aboriginal go-betweens. Unfortunately, he does not seem to have been very concerned about the meanings of placenames. This seemed rather to be a preoccupation of scholarly-minded people around the turn of the century (1900), when this information was no longer readily accessible (see note 2).

We need to try to discover the source of the meanings reported. If we cannot trace them back to reliable witnesses, we can remain justifiably sceptical. Certain further criteria can aid us, however. We should be suspicious of offered meanings which describe the function of a place as a ‘camp’ or ‘meeting place’. Thus we discount meanings such as Wright’s “meeting place” for Canberra (his ‘Kamberra’) (Wright 1923: 58) and “the place where all males are presumed to be made young men” for Tidbinbilla (his ‘Jedbinbilla’) (Wright 1923: 38). Even more implausible as an etymology is the functional description of William Davis’ Gungahlin as “white man’s house” (quoted approvingly by Gillespie 1992: 253) or ‘Kundul’ (now Kurnell) as “place where Captain Cook landed” (Hon. G. Thornton in RASA Manuscripts 1900). Furthermore we can be suspicious of meanings that simply describe the topography of the place – ‘water’, ‘hill’ – since known Aboriginal placenames typically do not directly indicate these obvious features. Thus some of Mitchell’s (1926: 35) reported names in the Monaro, such as ‘camping or resting place’ for Adaminaby and ‘camp’ for Woolway, can be discounted.

On the other hand, placenames that are reported to refer to an animal species are more likely to reflect the mythology-based naming strategies that characterised Aboriginal nomenclature. Thus Mitchell’s (1926: 35) ‘big fat kangaroo rat’ for ‘Jimen Buen’ (as he writes it) is exact (see ‘Etymologies based on testimony and wordlist data’ below). His ‘plenty ants’ for ‘Ironmungie’ may likewise accurately include a reference to some kind of ants (the -mungie part, though, is found in several other placenames in the lower Snowy River area). Some of his other suggestions include creature names, but also unnecessary references to ‘(resting) place’. Thus his ‘Chakola’ “place for lyrebirds” (ibid.) is just ‘lyrebird’ in several languages of southeastern New South Wales; his “Boonyan (now Bunyan) ‘pigeon’s resting place’” (ibid.) cannot be confirmed, but at any rate would be just ‘pigeon’; his “resting place of the native companion” for ‘Billilingera’ (Mitchell 1926: 76) is likewise not confirmed, since none of the attested Ngarigo words for ‘brolga/native companion’ has any form resembling the placename. (It is possible that the placename alluded to a story about brolgas without naming the bird; i.e. Mitchell’s information may reflect the etiology, if not the etymology, of the placename.) On the other hand, the suggestion that Kurruducbidgee near Braidwood, also known as Larbert and spelled ‘Kouraduck-bidgee’, means “river of native companions” (Bernard McLean in RASA Manuscripts 1900) receives confirmation from Dharrawal wordlists, where ‘native companion’ is given as gooradawak (Ridley 1887: 418) and guradhawak (Mathews 1903: 277).28 In addition to creature names, body parts are known to play a large role in Aboriginal place-naming strategies. A suggested meaning ‘my elbow’ for the South Coast site ‘Turlinjah’ (E. C. Branch in RASA Manuscripts 1900) is plausible, and can in fact be confirmed by wordlist data in the form of Robinson’s (2000: 166) Biggah (Bega) tal.leen.jer (probably to be reconstructed as dhaliny-dya ‘elbow-my’) beside Mathews’ (1901-1902: 68) Dhurga dhurl´-leeng (probably dhaliny).

Comparing placenames to known vocabulary

This procedure consists of trying to match the reconstructed placename to ordinary vocabulary items that have been recorded for the language of the area in which the placename occurs. Several limiting factors make this exercise less than totally satisfactory. In the first place are the uncertainties and ambiguities of pronunciation that may remain even after the reconstruction has been made. Secondly, the wordlists that have been recorded probably suffer from the same clumsy spellings, etc. and are likely to be at least as dubious in their phonetic interpretation as the placenames. Thirdly, the amount of available vocabulary is likely to be only a fraction of what was in the language when it was fully spoken. Finally, there is the possibility that the placename lacked a known etymology, since not all placenames are expected to be transparent.

How much can we expect to etymologise?

We should not be too optimistic about how many placenames we might find meanings for. To see why let us explore some calculations. In no language do all placenames have a transparent meaning – although it must be admitted that languages probably differ in the degree of transparency of their placename vocabularies (cf. Walsh 2002). Let’s assume that only 50 percent of placenames can be expected to have a transparent meaning. Now let us assume for a language like Ngarigo of the Monaro that some 200 placenames can be found that are of apparent Aboriginal origin. We would expect even in the ideal circumstances to find meanings only for 100 of them (using the 50 percent figure). We would like to compare the placenames to the general vocabulary of the language. How many words would that involve? A fair estimate might be about 3000. But we do not have a complete vocabulary. Combining all the available vocabularies might yield only about 300 words. Hence we have only 10 percent of the general vocabulary to compare with 200 placenames, of which we might expect only 100 to be in principle relatable to general vocabulary. The consequence of having only 10 percent of the general vocabulary available is that we can expect to find etymologies for only 10 percent of the placenames that have general meanings, that is 10 placenames. (If the percentage of transparent placenames was 100, we could at best expect to find meanings for only 20 placenames.)

We could extend our results by looking at the vocabularies of neighbouring languages. There is usually considerable overlap between the vocabularies – especially for flora and fauna – of adjacent languages. A term that is undocumented in one language may have been recorded for a nearby language. See the example of Arable below.

Etymologies based on testimony and wordlist data

The greatest certainty about the meaning of a placename is gained from the combined evidence of reliable testimony and confirmation from a wordlist from the area. An example of an eminently successful etymology of this type is Jimenbuen, a property near the Snowy River in the Monaro region. Here we have available reminiscences from the family of the earliest settler family. William Crisp, the son of the first settler, Amos Crisp, reported that his father with his brothers and sisters in the 1830s

passed through Queanbeyan and Cooma and across the Snowy River to Jimen Buen … the name of which was later changed to Jimenbuen. ‘Jimen Buen’ is said to mean, in the local Aboriginal dialect, ‘big fat kangaroo rat’. (Andrews 1998: 109, quoting from Crisp 1947: 20)29

This interpretation can be confirmed from the available linguistic sources. We have fairly meagre wordlists for the Ngarigo language of the Monaro: the fullest documentation is by R. H. Mathews (1908: 338), who records ‘kangaroo rat’ given as dyimmang. As late as 1962 the professional linguist Luise Hercus was able to record some remembered words of Ngarigo from Aboriginal people living in Orbost, Victoria; she gives djimung as “kangaroo-rat, Bettongia sp.” (Hercus 1986: 244). So the first half of the modern placename is confirmed.30 What about the rest? It so happens that the word for ‘fat’ (substance and/or property apparently) is given for a number of sites in and around the Monaro: Bulmer (1887) gave it as bewan and Mathews (1908) as be:-wan.31 This word can be represented in phoneme transcription as /piwan/ or /biwan/ (depending on one’s orthographic preference for voiceless or voiced symbols); it was apparently pronounced as [byuwan].32

Now consider the order of the two elements. In most Aboriginal languages the qualifier word (corresponding to an adjective in English) is placed after the word it qualifies. So the meaning ‘fat kangaroo-rat’ would be expressed as ‘kangaroo-rat fat’. (The same order would be used if the meaning were ‘fat of the kangaroo-rat’.) Note that this placename is a phrase, not just a single word.

So here we probably have the etymology of the name, i.e. the meaning of its elements. It means ‘fat kangaroo-rat’, or possibly ‘fat of the kangaroo-rat’. But we still lack knowledge of the etiology of the placename, i.e. the story that would explain why this particular name was given to this place. It is also worth noting that we do not know – although perhaps Amos Crisp did – what specific site on or near the property named Jimenbuen (or its homestead) was associated with this name. Was it a waterhole in the Snowy River, a certain hill, or what? Thus, of the three desirable elements of the meaning of the placename, we have the following results: (a) the etymology is clear; (b) the etiology is completely unknown; and (c) the spatial referent is known only imprecisely. The form was presumably dyimang biwan, with stress on the first syllable of each word, with the consequence that the vowel of the second syllable was spelled with an e that reflects a weaker pronunciation of /a/ in an unstressed syllable. The early recorders apparently missed the fact that the first part of the name ended with ng rather than n. An English spelling which would more accurately reflect the reconstructed name would be ‘Jimang Buan’.

Etymological suggestions based on wordlist data alone

Even in the absence of testimony about the meaning, it may still be possible to make plausible guesses about the likely meaning on the basis of relevant wordlists.33 Here are some examples.

A property north-west of Canberra that in the nineteenth century was held by pioneer Henry Hall had the name ‘The Mullion’; ‘Mullion’ was also used as the name of a parish in County Cowley (Moore 1999: vi, 96). This is plausibly derived from maliyan or malyan, the word for the wedge-tailed eagle, commonly called ‘eaglehawk’, which is attested for a number of languages of the region, including Wiradhuri (mul.le.yan in Robinson 2000: 178, McNicol and Hosking 1994: 90), Ngunawal (mul.yun and mul.le.yal in Robinson 2000: 208 and 209 respectively, mulleun in Mathews 1904: 304), Wolgal (maliang in Howitt 1996[1904]: 102), Dharrawal in the Illawarra (mulyan in Mathews 1901: 130), and Yuwaalaraay, Yuwaalayaay, and Gamilaraay in inland northern NSW (maliyan in Ash et al. 2003: 106).

Burrinjuck, near Yass, which is sometimes given the folk-etymological spelling ‘Barren Jack’, seems to be Ngarigo ‘crayfish’, which is attested as barrinjook (du Vé 1887: 430), barranjerk (Bulmer 1887: 558), pur.run.juc (Robinson 2000: 195).

Carrott, a pastoral run in the Monaro near Jindabyne (Hancock 1972: 51) appears to reflect the well-attested Ngarigo word karrid ‘cold’.

For ‘Woden’, I would like to propose a possible Aboriginal etymology. A plausible Aboriginal source of the name ‘Woden’ is the word for ‘possum’, wadyan (or possibly wadhan), which is attested ten times for the languages of Aboriginal people extending from Yass through Queanbeyan and the Monaro to Omeo in Victoria. Some spellings suggest that the first vowel could have an o-quality (wod.jun, woy.jun); the second vowel is sometimes written with an e-vowel (watjen, widgen); and it is not unheard of for a palatal dy sound to be heard as a plain d (cf. terms for ‘black duck’ given as bud.en.bal vs boojangbung and boothunba, or bindi vs. binjey or bindhi for ‘belly’). If ‘Woden’ is indeed derived from an Aboriginal language, it is possible that its spelling was influenced by that of the Germanic god’s name (as the spelling of ‘Canberra’ was probably influenced by the English placename ‘Canbury’).

A famous snow-belt property in the upper Murrumbidgee region is called ‘Curango’,34 which is known to be a shortened form of ‘Curangorambla’ (which survives in the name of a creek and a mountain range). This name is plausibly reconstructed to *Kurangkurambla, which could be interpreted as *kurang-kurang-bula, with the second ng assimilated to the labial pronunciation of the following b, and the first vowel of bula omitted. The first part then looks like the word kurang-kurang which is given as ‘rainbow’ in wordlists for Ngarrugu (Mathews 1908: 337), and the (Ngunawal) language of Yass (Robinson 2000: 210) and as ‘Milky Way’ in the ‘Limestone’ vocabulary (Robinson 2000: 270). The final element bula is a suffix meaning ‘two’ in many languages. Thus the name might mean ‘two rainbows’.

Etymological suggestions based on linguistic evidence other than wordlists

A property in the Monaro, near Jindabyne, called ‘Biggam’ (Hancock 1972: 51) was spelled ‘Biggon’, by Lhotsky in 1834 copying a list from a manager named Bath (Lhotsky 1979: 105), and ‘Bigga’ by Land Commissioner Lambie in his 1839 census tours (Andrews 1998: 132). Lhotsky (1979: 106) elsewhere, but not in his vocabulary (Lhotsky 1839), mentions that the Monaro Aborigines eat biggon “yams … a root of a sort of Sonchus”. This term, which he spells the same as the placename, therefore has a chance of being the source of the name, given this snippet of lexical information, even though it is not given in a source that is devoted to language.

An even better case for an etymology based on a word not found in wordlists is Arable. Arable is a property in the Monaro, west of Cooma. It should be noted that this spelling is identical to an English word, which we can assume has influenced the spelling. Other early spellings include: ‘Arabel’ recorded in 1834 (Lhotsky 1979: 105), ‘Arrable’ recorded by Land Commissioner John Lambie in the 1839 census (Andrews 1998: 124), ‘Arrabel’ recorded in 1844 (Robinson 1998: 173). This name can be correlated with a vocabulary item mentioned by the naturalist George Bennett, who was told by Aborigines in the Tumut River area in the 1830s that fat crows that fed on bogong moths were called arabul (Bennett 1967[1834]: 272-273). Allowing for the indeterminacy of the vowel of the final syllable, this seems to be the same word – even though none of the wordlists for languages of the region give this as the standard word for crow. We are left with a plausible hypothesis that the placename Arable might have derived from some association with bogong-eating crows. This deduction appears to derive support from some un-sourced testimony, since an editor’s note in Lhotsky 1979 (note 168 p. 230) says that Arable “has been stated to be Aboriginal for ‘crow’”.

Choosing linguistic sources for placename etymologies

Up to this point we have ignored the question of which language or what linguistic sources should be consulted in looking for etymologies of placenames. Obviously one should look first to the language known to be spoken in the immediate area. But in south-eastern Australia the documentation of languages is patchy, and it is not always easy to determine the geographical extent to which a particular language was traditionally spoken. Another issue is the fact that linguistic varieties in an area tend to have considerable overlap in their vocabularies. Hence it may be hard to judge which samples are to be considered dialects of the same language – where the linguistic criterion for separating ‘dialects of a language’ from separate languages is usually based on a quantitative measure of difference in grammar and vocabulary, with differences in grammar usually being given more weight than vocabulary differences. Finally, the traditional names of language varieties (dialects, languages) are often not known.

For the region of Canberra and the Monaro, we need to consider three potential groups of source material, as indicated on Table 5.5. We have wordlists, and a short grammar by Mathews, for the Yass/Ngunawal language. For the Monaro we have six sources of vocabulary but no grammatical information. In between, in the Canberra area, we have a disparate set of sources, the longest of which, the recently accessible ‘Limestone’ vocabulary of Robinson, was recorded at Yarralumla from people identified as Yammoitmittong, or Yammoit mob – which I interpret as ‘Namadgi people’. The Queanbeyan wordlist, thought to be from the famous early Canberra personality Nellie Hamilton, has long been seen to reflect the same language as the Monaro wordlists, that is, Ngarigo. The Robinson material supports this conclusion. Although we lack a grammatical description, we do have some personal pronouns, which are grammatical words. The Queanbeyan and ‘Limestone’ vocabularies agree with the Monaro sources in having the unique form ngayamba for ‘I’, whereas the coastal languages have ngayaka, ngayalu, or ngayadha, and Ngunawal has a completely different form kulangka, which is shared only by the closely related Gundungurra language of the Southern Highlands.

Table 5.5: Linguistic sources for the Canberra and Monaro area

language

source name

date collected

reference

Ngunawal

Yass

1844

Robinson 2000

Yass

-1887*

Curr 1887: No. 196

Ngunawal

-1904

Mathews 1904

Canberra

Molonglo

1830s

Eyre 1845

Limestone

1844

Robinson 2000

(Canberra)

c.1840

Mowle 1891

Queanbeyan

-1887

Curr 1887: No. 195

Ngarigo

Menero

1834

Lhotsky 1839

Maneroo

1844

Robinson 2000

Moneroo

-1887

Bulmer 1887

Moneroo

-1887

du Vé 1887

Ngarrugu

-1908

Mathews 1908

Southern Ngarigu

1960s

Hercus 1986

* This notation means ‘before 1887’.

The etymology we propose for a placename can differ according to which linguistic source is used. ‘Mugga Mugga’ is the name of a hill in the southern part of Canberra. It was also the name of a nineteenth century home on the slopes of the hill (inhabited for many years by the Curley family) and still open to tourists as Mugga Mugga Historic Cottage. If we start from the name ‘Mugga’, we can easily reconstruct a form *Maka. A word maka is attested in as a generic term for snake in Curr’s and Mathews’ Yass/Ngunawal vocabularies and Larmer’s (Larmer 1898) Bateman’s Bay, as ‘black snake’ in Robinson’s Maneroo, and ‘carpet snake’ in Mathews’ Thurrawal from the Illawarra. But Lhotsky’s Menero vocabulary gives it as ‘lizard’. Robinson gives ‘lizard’ as the meaning of the reduplicated maka-maka in the Yass language. Meanwhile the generic term for snake in the Monero language, according to several wordlists, including that of Queanbeyan, is dyidyukang or dyidyikang. In searching for an etymology of ‘Mugga Mugga’, do we rely more on vocabularies of Ngunawal/Yass or on those of Ngarigo/Monaro? Should we start from the simple form maka or the reduplicated form maka-maka? Depending on these choices we could have the following meanings: snake, a specific kind of snake, or lizard. Unless we can find testimony from early sources or knowledge preserved in the Aboriginal community (independent of the wordlists), we cannot have any certainty.

Partial etymologies: Identifying elements without meaning

It is relatively common to find parts of words that recur in a number of placenames within an area. It is likely that these elements had a meaning. Although we might not be able to recover the meaning of these elements, we may nevertheless be able, by comparing names of partially similar form, to conclude that certain names consisted of two elements and to identify sets of placenames that had related meanings. Some recurrent final parts of names in this area are: -bili (Tidbinbilly/Tidbinbilla, Adjungbilly), -berra (Canberra, Yeumberra, Jerrabomberra), -dra (Kiandra, Kydra), -beyan (see next paragraph), -bi (Gudgenby, Goodradigbee) and, further south, -dibby (Wollindibby, Gelantipy), -byra (Ingebyra, Coocoobyra), -adbo or -edbo (Thredbo, Bredbo, earlier Braedbow, Byadbo).

Two adjacent pastoral runs in the Monaro district east of Cooma, ‘Kybean’ and ‘Kydra’ (Hancock 1972: 47) seem to involve the same root ky- (presumably kayi- in Aboriginal phonology) followed by two separate elements -bean and -dra, which recur in other placenames (cf. Queanbeyan and Kiandra). Although we do not know the meaning of the elements, we can surmise that the names consist of two elements each. Looking at ‘Queanbeyan’, we can therefore surmise that Quean- was a separable element. Phonological reconstruction suggests a form *Kuwinbiyan. The first European site with this name, Timothy Beard’s 1828 holding on the Molonglo River some distance from the township of Queanbeyan gazetted in 1838, was only a few kilometres distant from a place named ‘Kowen’ (also spelled ‘Kowan’, ‘Kohan’, ‘Kohn’, ‘Cowen’, ‘Cohen’, ‘Coen’) – the area was also called ‘The Swamp’, ‘Dirty Swamp’, ‘Glencowan’, ‘Glenbirnie’ and ‘Glenburn’ (Cross 1985: 50-51). This name can be reconstructed as *Kuwin, which is identical to the first element of *Kuwinbiyan. It is conceivable that the latter name, Queanbeyan, consisted originally of a modification of the name ‘Kowen’, meaning perhaps a specific site by the river in a more general area called *Kuwin.

Reconstructing the etiology

To be certain of the story behind the name we need evidence that comes from the Aboriginal people of the area. Unless we have some testimony to this, we remain in ignorance. At best we can suspect elements of a story from the supposed ‘meaning’ that has been offered in our sources. (What is offered as ‘meaning’ may be the etymology or literal meaning of the word, or a description of the topography or function of the site, or an element of etiological story behind the naming of the place.)

To illustrate: Mitchell (1926: 76) reports that ‘Billilingera’ (a property near Cooma) is said to refer to where brolgas stay, as noted above in ‘Assessing the reliability of testimony’. But the name is not relatable to known words for ‘brolga’ that are found in wordlists. The name may rather conceal a story about brolgas, and allude to it by means of a word that refers to some particular aspect of the story. Similarly, ‘Booroomba’ (reconstructable as *Bururumba or *Burarumba), a property in the ACT west of the Murrumbidgee, is said to mean “wallaby jumping over a rock” (Cross 1985: 68, Gillespie 1992: 253). Perhaps this ‘meaning’ simply refers to the red rock wallabies, which lived (and were shot) in great numbers in this area (Moore 1999: 71). On the other hand it may conceal a word or words referring to this event, even if not does not directly denote ‘wallaby’, ‘jump’, or ‘rock’.35

Summary and application of methodology to the name ‘Canberra’

In this final section we summarise the methodology and show how it applies to the name ‘Canberra’. This is appropriate both because of the amount of information available (providing something to say on almost every aspect of our procedures) and because of the importance of the name, being the only name of an Australian capital city that is derived from an Indigenous placename.

The evidence of history

Study of the historical record throws light on the bestowal of the placename within the Anglo-Australian context, its locational referent and its changes, the history of spelling, reports on its pronunciation, etymological source and meaning.

Name bestowal, de-bestowal, re-bestowal36

The name was bestowed by Joshua John Moore, the first pastoralist with interests in the Canberra area, whose superintendent John McLaughlin (an ex-convict from Dublin) in December 1824 moved livestock from Moore’s Baw Baw station near Goulburn to the Canberra area, where he camped at Acton ridge beside the Molonglo River, at the site of the present National Museum of Australia. On 16 December 1826 J. J. Moore wrote to the Colonial Secretary, asking to purchase the land which he had been occupying, describing its location in these terms:

The land which I wish to purchase is situate at Canberry, on the E. Bank of the River which waters the Limestone Plains, above its junction with the Murrumbeeja, adjoining the grant of Mr. Robert Campbell, snr.

In 1831, after John MacPherson was granted 640 acres immediately to the west, along the Molonglo River from Sullivan’s Creek to Black Mountain, Moore wrote in protest to the Surveyor-General about his desire to retain

the one thousand acres in my possession situate on the Molongoo River called and known by the name of Canburry, and bounded on the South by the Molongoo River, and on the West by Canburry Creek.

The property was then surveyed by Hoddle as a 1000-acre block between parallel lines running magnetic north from the Molonglo River to about Haig Park, with its western boundary on the Acton Peninsula east of the mouth of Sullivan’s Creek and its eastern boundary around Regatta Point. In 1837 Moore purchased a further block of 742 acres to the immediate north of his existing holding. In 1843 the property was purchased by Arthur Jeffreys, who renamed it ‘Acton’. The property was resumed by the Commonwealth in 1911, and the original homestead, which was built around 1830 and called ‘Canberry Cottage’, was included in an expanded Acton House, which survived as government offices until it was demolished in the 1940s to make way for the Canberra Hospital.

Meanwhile the name ‘Canberra’, de-bestowed in 1843 as the name of a property, survived as a general name for a wide area, and was re-bestowed in 1913 as the name of the site selected as Australia’s new federal capital.

Changes in locational referent

The history of the name ‘Canberra’ provides us with some of the most spectacular changes of referential scope of a placename.37 While the specific site referred to by its Aboriginal precursor is not known, the name ‘Canberra’ itself was first applied in the European domain to the pastoral property claimed by J. J. Moore in the 1820s, with its main buildings located on a ridge beside the Molonglo River at the site of the current National Museum of Australia. The name was also applied in the first decades of European occupation to the plain lying north of the Molonglo River and between the Black Mountain-O’Connor Ridge and Mount Pleasant-Mount Ainslie-Mount Majura range; Sullivan’s Creek was called ‘Canbury Creek’. The area called ‘Canbury’/’Canberry’/’Canberra’ extended eastward on this plain as far as the church of St John the Baptist (established in 1841) and a blacksmith shop and post office (near Blundell’s Cottage) on the Duntroon estate. The placename was sometimes applied as well to the area south of the Molonglo River to the area of Klensendorlffe’s farm (including the area around Albert Hall, the present parliament buildings, and as far as Red Hill which was once called ‘Canberra Hill’). ‘North Canberra’ referred to the area of Lyneham, including St Ninian’s Church and the Old Canberra Inn. Meanwhile the original Canbury property was re-named ‘Acton’ in 1843 by a new owner, Jeffreys, after his family’s historic property in Wales. Much later, in 1913, since the name ‘Canberra’ was selected as the name of the federal capital, its scope has expanded as the city has grown.38

In reconstructing the original scope of the placename, do we start from the whole Canberra Plain which was mostly occupied by Moore’s property or the focal part of this property, where its original huts and homestead were built? Given the principle that placenames tend to expand their reference to encompass and even exceed the legal entity, I would opt for the narrowest scope as being closest to the original designation. This was the conclusion drawn by an early historian, Frederick Robinson:

Canberra, therefore, stood originally for the area enclosed by Black Mountain westerwards (with Canburry Creek at its foot) and the Molonglo southwards, or more definitely (as seen from Dixon’s Survey of 1829) for the river promontory at Acton, where Moore’s huts were placed. (Robinson 1927: 4)

History of spelling of the name

As indicated in the citations in ‘Name bestowal, de-bestowal, re-bestowal’ above, J. J. Moore spelled the name of his property as both ‘Canberry’ and ‘Canburry’. The spelling ‘Canbury’ was used, along with ‘Canberry’, by surveyor Hoddle in 1832 (Murphy 1987: 14), and was widely in use in the early years of European settlement. This may have been influenced by English placenames ending in -bury, including a town called Canbury in Kent (now within the greater city of London). The Polish scholar John Lhotsky, who visited in 1834, spelled the name ‘Kembery’. The registers of St John the Baptist Church, which begin in 1845, show ‘Canbury’ at first, with ‘Canberry’ becoming more frequent from 1858, and ‘Canberra’, first used in 1857, becoming the usual spelling from 1862 (Salisbury 2000: 4). The Rev. Pierce Galliard Smith, the long-serving minister of St John’s, who was appointed in 1855, is credited with introducing the spelling ‘Canberra’ for both the church and the school attached to it, a usage which was followed in the name of the post office established in 1863 (Body 1986: 41). This spelling was also used for the subsequent federal capital.

Reported source: Is the name Aboriginal?

Historians have accepted that Canberra is of Aboriginal origin for two reasons. First, the normal practice was to name unsurveyed land in the first instance after its local Aboriginal name.

Canberra is, without any doubt, a native name. When Moore used it in 1826 only a handful of white men were in the district. All other names used by him, and in Dixon’s first survey of 1829 (except Mt. ‘Ainsley’), are obviously native – Pialligo, Yarrowlumla, etc. (Robinson 1927: 5)

Second, early testimony says so. John Lhotsky, visiting in 1834, mentioned both the Kembery Plain and the Kembery River “as it was originally called by the natives” (Lhotsky 1979: 55, 61). The Canberra historian John Gale “unreservedly accept[ed] Dr. Lhotsky’s statement that the original of the name of Canberra was a native name”, according to the argument of a letter “Canberra – a native name” by R. H. Cambage published in the Sydney Morning Herald (Gale 1927: 27). There is also the testimony of Stewart Mowle, who worked for T. A. Murray at Yarralumla from 1838.

Stewart Mowle, who came to the district in 1838 and knew the natives well enough to converse with them, has recorded that natives themselves pronounced Canbery (or Kemberry) as Gnabra … S. M. Mowle, List of Aboriginal names in the South, 1891, National Library. (Wilson 2001[1968]: 61)

An Aboriginal source is also assumed by those who claimed that Canberra was the name of an Aboriginal tribe. W. Davis Wright, who grew up at Lanyon in the 1840s, claimed that Canberra was the name of the local Aboriginal tribe, that one of their focal areas was at the Canberra/Acton site, and that the name was pronounced Kamberra (Wright 1923: 57-58). Similarly, W. P. Bluett, who got his information from John Blundell (born 1838) and Mrs John MacDonald nee Webb (born 1842), claimed that the “Nganbra-Pialligo tribe” consisted of two groups, the “Pialligo blacks”, who camped at Pialligo, and the “Canburry or Nganbra blacks”, who camped at the foot of “Black’s Mt” close to “Canburry Creek” (Bluett 1954).39

A further argument for an Indigenous source comes from the variety of spellings. If the name were English, it is hard to see why the spelling should vary so much. On other hand:

The uncertain quality of the middle and final vowels of the name, as evidenced by the varied spellings – Canberry, Canburry, Canbury, Canberra, is also easily explained from its aboriginal origin. (Robinson 1927: 5)

In spite of this evidence for an Aboriginal origin, there have still been more recent defenders of a European origin of the name ‘Canberra’ (see Murphy 1987 for a summary of theories and an attempted defence of a link with a place in the London suburb of Kingston-upon-Thames).

Reported meaning: What is its etymology?40

W. D. Wright, who, as noted above, claimed that ‘Kamberra’ was a tribal name, is also responsible for the suggestion that the name meant ‘meeting place’.

The correct rendering of their tribal name was Kamberra. Their corroboree ground was at Kamberra, as far as I can gather the exact spot being near the Canberra Church [sic41], where the Administration Offices are now erected at Acton, Canberra, and by Canberra Church toward the old Duntroon dairy. It served also as their general and best known meeting place. (Wright 1923: 57-58)

This was followed by Massola (1968), whose entry for Canberra says, “from Nganbirra, a camping place”.

The other widely known theory is that the name meant ‘breasts’. John Gale was told at Andrew Cunningham’s funeral at Lanyon in 1887 by Rev. Canon Champion of Bungendore that a very old swagman familiar with the area had heard from the blacks “that the English equivalent for Canberra was ‘a woman’s breasts’”; this was suggested by the two hills Mr Ainslie and Black Mountain (Gale 1927: 14).

Other proposals that have been offered, without supporting evidence, are Mawer’s suggestion that the name is “a corruption of the Ngunawal version of corroboree” (Mawer 1983: 13), an Aboriginal word for ‘kookaburra’ (reported in Murphy 1987: 16), or an Aboriginal word meaning ‘The Head’ (Frederic Slater, reported in Murphy 1987: 14, who later claimed it was rather Celtic for ‘Head Town’).

Reported pronunciation

The historical record hints at some differences in pronunciation of the name. The first difference affects the final vowel. When Rev. Pierce Galliard Smith, rector of St John’s church, began using the spelling Canberra, the squire of Duntroon, George Campbell emphatically used the underlined spelling ‘Canberry’ in a written note to Smith (Murphy 1987: 15). Secondly, mb was used instead of nb by Wright (see above) and by Lhotsky. Thirdly, the first vowel seems to have been regularly pronounced with the short English a of can, but Stewart Mowle insisted on a long vowel, which he spelled either aa or ar:

The native name of Canberry is Caanberra … The first syllable is long phonetically – Karnberra. (S. M. Mowle diary of 21 July 1838, quoted in a letter from his son Aubrey Mowle 17 March 1913; reference in Murphy 1987: 14).

Fourthly, the initial consonant is given as ng for his tribal name ‘Nganbra’ by Bluett, relying on the memory of old-timers (see extract above). The same pronunciation is presumably behind Stewart Mowle’s ‘Gnabra’, which he claimed was the Aboriginal pronunciation (Wilson 2001[1968]: 61) – assuming that gn is an attempt to represents the ng sound and that he forgot to write the n before b – “he wrote in his diary that the place name was pronounced Karnberra” (Wilson 2001[1968]: 186) and elsewhere wrote the name as ‘Caamberra’ (Mowle 1899: 47). Mowle’s apparently contradictory evidence makes sense if he meant that the Aboriginal name began with a nasal sound ng but Europeans pronounced it with a stop sound k, and that the first vowel was pronounced originally long as in carnage but was pronounced by Europeans with the vowel of can.

The accent was apparently on the first syllable, and the vowel of the second syllable was indistinct and often elided. Lady Denman, at an official ceremony 12 March 1913 naming the federal capital, “in a clear English voice which left beyond doubt both the name and its official pronunciation, said: ‘I name the capital of Australia, Canberra.’ The name… she pronounced ‘Canb’ra’, with the accent on the first syllable...” (Wigmore 1963: 62). Earlier evidence for the omission of the second vowel comes from an 1840 marriage witness who spelled the name ‘Canbrey’ (Robinson 1927: 5).

Insights from linguistics

Reconstructing the pronunciation

Using the variant attested spellings and the methodology proposed here we aim to go beyond the pessimistic conclusions of Watson:

The most noticeable of these [placenames of the aboriginal inhabitants] is the name of the city, Canberra. In 1826, J. J. Moore spelt it Canberry, but, in 1831, he spelt it Canburry. John Lhotsky, in 1835, spelt it Kemberry. The department of the surveyor-general, in issuing deeds, in 1837, spelt it Canberry, and in 1838, Camberry. S. M. Mowle, who resided in the district form 1838 to 1852, spelt it Caamberra. W. Davis Wright spelt it Kamberra. It is impossible now to determine the original correct pronunciation or the meaning. The official pronunciation, now adopted, is with the accent on the first syllable and as if it was spelt Canbera. (Watson 1927: 17 [italics added, HK])

To reconstruct the likely original pronunciation we begin with a matching of the variant spellings – see (8). For each set of corresponding sounds we propose an original sound and try to explain all deviations from this by plausible transmission processes of the kind discussed above.

The only segment for which we have absolute unanimity is the b. The variability of the preceding consonant – nb vs. mb – is easily resolved by the principle of lectio difficilior and the lack of attention: the nb sequence, common in Aboriginal languages but more rare in English, can be assumed to have been simplified by the substitution of mb in the usage of some speakers.

(8)

C

a

n

b

e

rr

y

1826

Moore

C

a

n

b

u

rr

y

1831

Moore

C

a

n

b

u

r

y

1832

Hoddle

K

e

m

b

e

r

y

1834

Lhotsky

C

a

m

b

e

rr

y

1838

Suveyor-General

C

a

n

b

r

ey

1840

Robinson 1927: 5

K

a

m

b

e

rr

a

1840s

Wright 1923

C

a

n

b

e

rr

a

1857

Rev. Pierce Galliard Smith

C

aa

n

b

e

rr

a

1838

Mowle 1838

C

aa

m

b

e

rr

a

Mowle 1899: 47

Gn

a

b

r

a

Mowle 1891

Ng

a

n

b

r

a

Bluett 1954

Kg

a

m

b

u

rr

y

Gillespie 1991: xviii

*Ng

aa

n

b

i

r

?

The first consonant is also explained by the principle of lectio difficilior: the sound ng is common at the beginning of words in Aboriginal languages but impossible in European languages. It is easy to see that if it were reproduced at all in English (it is often simply omitted in early wordlists), it would be replaced by a sound that does occur word-initially. On the other hand, if the Aboriginal word began with k, it is impossible to understand why anyone would report that it was pronounced otherwise. Mowle’s gn (of ‘Gnabra’), Bluett’s ng in his ‘Nganbra’ tribal name, and a spelling ‘Kgamburry’ reported by Gillespie (1991: xviii) are taken to represent the velar nasal sound usually spelled ng in Aboriginal orthographies.

Regarding the first vowel, Stewart Mowle’s various comments make it clear that this was the vowel of car rather than that of can. This agrees with what we know of Aboriginal phonologies, which tend to have the [æ] sound only as a variant of /a/ when adjacent to a palatal sound, especially y. Yet the name was popularly pronounced with the vowel of can, and this pronunciation was institutionalised by Lady Denman in the official proclamation of the name of the federal capital in 1913. The can-sound is easily explicable as a spelling pronunciation. One might ask why it was spelled Can-, to which one might ask in reply: how else could it have been spelled? Perhaps using aa: but this is rarely used in English. But why did they not spell it Carn-? This is explicable if it was first written down by an Irishman who would have pronounced r after a vowel; since ar was not available to him, he could only use the spelling with just a. In fact the name was probably first written by John McLaughlin, J. J. Moore’s Dubliner overseer (Mawer 1983: 11). A further reason for spelling the vowel with a simple a is the fact that the vowel of can in Irish English was in fact pronounced closer to the sound of Modern Australian car than to that of modern can (see Wells 1982: 129, 422). Other English speakers, such as his boss J. J. Moore, would have naturally replaced his [a] with their own [æ], making allowance for his Irish accent. The e of ‘Kembery’, written in 1834 by the Polish Lhotsky, is explicable by his accent; lacking the English sound of short a in can, he would have substituted the e of ken. In the various spellings of this vowel we see the role of the dialect or accent of the European recorders: an Aboriginal central low [a] (as in Southern British car) may have been copied by an Irishman using a vowel pronounced more to the front of the mouth, which in turn was reproduced by other English speakers as the [æ] of can, which in turn was reproduced in the Polish scholar’s accented English as the e of ken.

According to Mowle’s evidence the first vowel was pronounced long (see ‘Reported pronunciation’ above). If it were a short a-vowel as in cut, it is likely to have been spelled with the letter u. Many Aboriginal languages had distinctive long vowels; a spelling aa is how a long a is typically represented in modern orthographies.

The second vowel is spelled with e or u. The forms with -bury may have been influenced by English placenames such as Canterbury and Oldbury; at any rate they do not justify a pronunciation -boori, for which there is no other evidence. Moore’s early spelling ‘Canburry’ probably indicates the indistinct vowel that occurs in English before an r in an unstressed syllable, or simply inattention to the vowel quality in this position. On the other hand the e of the earliest spelling may indicate that Moore’s Irish overseer McLaughlin was alert to vowel distinctions before r, as in tern vs turn – unlike the majority of English speakers (see Wells 1982: 200). The reconstructed sound is thus [e], which in an Aboriginal language that did not distinguish between [i] and [e] would be taken as a realisation of the /i/ phoneme.

Most Aboriginal languages distinguished two kinds of r sound, one like the normal English r and the other a sound which varied between a trill and a tap, which is often spelled rr in modern orthographies. The variable spelling with single or double r, however, do not allow us to draw any conclusion regarding the sound to reconstruct, since there is no evidence that the European recorders were aware of the difference, and English spelling thus consistently under-differentiates these sounds. An argument could nevertheless be made in favour of reconstructing the English-type r here, from the facts that (a) the occasional absence of the preceding vowel is better explained as being absorbed into the more vowel-like r than the tap, and (b) the r is never mistaken for an l or a d, as sometimes happens with the un-English kind of r.

The final vowel presents real difficulty. The y-versions predominate in earlier sources, and support reconstruction of a vowel i. But a spelling with a was deliberately introduced by Rev. Pierce Galliard Smith. One might suppose that this reflected a growing appreciation for its Aboriginal pronunciation. Spellings with a are also used consistently by those witnesses who were most knowledgeable about Aboriginal people – Stewart Mowle, William Davis Wright and William Bluett. If the original vowel was a, how do we account for the i-pronunciations? The influence of English placenames ending in -bury is one possibility. Another is the tendency to vary the pronunciation of vowels in unstressed syllables between a neutral shwa vowel (which would be spelled with a) and a high front vowel [i] (which would be spelled with a). Note the variation between ‘Jerrabomberra’ and ‘Jerry Bunbery’ (Robinson 1998: 203), ‘Ginninderra’ and ‘Ginninderry’ (see ‘Differences in attention to phonetic detail’ above). There is the further possibility that the name really ended with a trill r (spelled rr). Then the different vowels may represent different interpretations by Europeans of the short vowel-like sound heard after the trill. The perceptions of [i] would be natural if the vowel preceding the rr was [i] or [e]; the a-spelling would represent the neutral vowel of English unstressed final syllables. It is even possible that a-forms represent a different word, the locative case form of the placename, *Ngaanbirr-a ‘at Canberra’ vs. *Ngaanbirr ‘Canberra’. It is sometimes the locative case that gets borrowed as the name of a place.42

Finally, the position of stress on the placename can be reconstructed to be the first syllable, as claimed already by Robinson (1927: 5): “Nor is there any doubt that in pronunciation Canberra should be accented on the first syllable, and almost omit the ‘e’ in the second.” This is also the official pronunciation (Watson 1927: 17), as established in the official announcement of the name by Lady Denman (see ‘Reported pronunciation’ above).

In summary, we can plausibly reconstruct the name as *Ngaanbira, with some uncertainty regarding (a) the length of the first vowel (aa or a), (b) the nature of the last consonant (r or rr), and (c) the quality of the final vowel (a, i, or nothing).

Reconstructing the locational referent

Given the tendency for the post-contact reference of a placename to expand and designate especially the ‘plains’ so loved by early pastoralists, we expect the original reference to be more localised and to be some named feature in close proximity to the focal point of the first European activity. Since the name ‘Canberra’ was first applied to J. J. Moore’s pastoral run centred on Acton ridge beside the Molonglo River and near the junction of this river and Sullivan’s Creek, earlier named ‘Canbury Creek’, we would expect the Aboriginal name to designate some place near this complex. There is testimony that there was a camping place, of Wright’s “Kemberra tribe” and Bluett’s “Canburry or Nganbra blacks” – located, according to Bluett (1954: 1) at the foot of “Black’s Mt” close to “Canburry Creek”, i.e. in the area now called ‘Black Mountain Peninsula’. There was a deep waterhole in the Molonglo River at its junction with Sullivan’s Creek (Bill Gammage pers. comm. 13 November 2006). It is likely that ‘Canberra’ reflects the name of the Aboriginal camp, which in turn may have been based on the waterhole, the creek-river junction, or even Black Mountain.

Exploring the etymology

The meaning ‘meeting place’ suggested by Wright can be dismissed as being simply a functional designation of a named place that was used as a camping place rather than being itself a name.

Assuming an Indigenous source of the name, can we equate it with any word attested in Aboriginal wordlists? Not really. Nevertheless, the claim of Appleton and Appleton (1992: 55) à propos of ‘Canberra’ that “no such word has been recorded in any Aboriginal language” now needs to be qualified. For the starting point should not be Canberra but rather a reconstructed pronunciation something like Nga(a)nbira. In fact, a form nganbirr, which is a conceivable reconstruction if we assume a final trilled r, is attested in the Gamilaraay and Yuwaalaraay languages of northern inland New South Wales, in the meaning “crosswise, across … Used to describe things that are ‘across’ something”43 (Ash et al. 2002: 122). Although these languages are rather distantly related to the Canberra language, they would surely have shared some vocabulary with it. The etymology is not very persuasive, however, given the distance of the language from Canberra and the fact that the form does not match our reconstruction exactly.

We need to explore the supposed meaning ‘breast’ reported by Gale. This cannot be confirmed from wordlists. The word for ‘breast’ (or ‘milk’) is attested in wordlists from Queanbeyan, Ngunawal, the Monaro, and Dharrawal and Dhurrga on the south coast as ngaminyang, which does not resemble ngaanbira.

One further question is whether the structure of the name ‘Ngaanbira’ might provide a clue to its linguistic analysis, i.e. whether we can identify component elements without meaning (see ‘Partial etymologies: Identifying elements without meaning’). The recurrence of the -berra part has been noted before: “Other local place-names, without doubt native, have a similar form: Jerrabomberra, Yeumberra, Bimberi” (Robinson 1927: 5). For the first of these see ‘Changes in spelling and pronunciation’ above. For Yeumberra (on the Murrumbidgee), I have found alternate spellings ‘Yeumburra’, ‘Uemberra’, ‘Umberra’, ‘Umburra’. For Bimberi I have not found any old (pre-1860) references. It appears that we can reconstruct at least three placenames in the area that end in the same sequence, which might have been a derivational suffix with a particular meaning: *Ngaan-bira, *Jiridiban-bira, *Yum-bira or *Yuum-bira. Neither the meaning of the supposed suffix nor that of the root can be further identified and associated with a meaning.

Etiology

There is no record of any story that would explain the Aboriginal placename. It is possible, however, that Gale’s reported meaning ‘breast’ (see ‘Reported meaning: What is its etymology?’ above) may contain an oblique reference to a mythological story, since body parts do often figure in Aboriginal myth-based toponymy. Nevertheless the story, if there was one, remains irrecoverable.

Conclusion

We hope to have shown that it is possible to make some informed guesses about the pronunciation and meaning of Aboriginal placenames, provided there is a reasonable amount of documentation both of the Anglo-Australian history of the place and of the local Indigenous language. Historical records must be used sensibly, and careful thought needs to be given to the processes of interlingual transmission, especially of the pronunciation of the name, and the historical transmission of the application of the name to locations in the Anglo-Australian domain. There are limits to what information can be recovered, with respect to the precise local referent, the etymology, and especially the etiological story behind the name.

References

Alder, Reg 1989, ‘Namadgi National Park’, Canberra and Historical Journal, New Series 23: 34-37.

Andrews, Alan E. J. 1998, Earliest Monaro and Burragorang 1790 to 1840, Tabletop Press, Palmerston, ACT.

Appleton, Richard and Barbara Appleton 1992, The Cambridge Dictionary of Australian Places, Cambridge University Press, Melbourne.

Ash, Anna; John Giacon and Amanda Lissarrague 2003, Gamilaraay, Yuwaalaraay & Yuwaalayaay Dictionary, IAD Press, Alice Springs.

Austin, Peter and Terry Crowley 1995, ‘Interpreting Old Spelling’, in Nicholas Thieberger (ed.), Paper and Talk: A Manual for Reconstituting Materials in Australian Indigenous Languages from Historical Sources, Aboriginal Studies Press, Canberra: 53-101.

Barrow, Graeme (ed.) 1985, John Gale’s Brindabellas and Australian Alps, Dagraja Press, Canberra.

Bennett, George 1967[1834], Wanderings in New South Wales ... during 1832, 1833, and 1834 vol. 1, Australiana Facsimile Editions no. 115, vol. 1. Libraries Board of South Australia, Adelaide [first published 1834, Richard Bentley, London].

Bluett, W. P. 1954, ‘The Aborigines of the Canberra District at the Arrival of the White Man’, paper read to the Canberra and Historical Society, 29 May 1954.

Body, A. H. 1986, Firm Still You Stand: The Anglican Church of St John the Baptist Canberra its Parish and Parishioners 1841-1984, St John’s Parish Council, Canberra.

Bulmer, John 1887, ‘No. 197: Moneroo’, in The Australian Race, vol. 3, E. M. Curr, Government Printer, Melbourne: 432-433.

Clark, Ian D. and Toby Heydon 2002, Dictionary of Aboriginal Placenames of Victoria, Victorian Aboriginal Corporation for Languages, Melbourne.

Clarke, Patricia 1986, A Colonial Woman: The Life and Times of Mary Braidwood Mowle 1827-1857, Allen and Unwin, Sydney.

Coates, Richard 2006, ‘Names’, in A History of the English Language, Richard Hogg and David Dennison (eds), Cambridge University Press, Cambridge: 312-351.

Cope, John 2006, Pioneer Parson of Early Canberra: Rev. Pierce Galliard Smith, Minister of St. John the Baptist Church, Reid, 1855-1905, Ginninderra Press, Charnwood, ACT.

Crisp, William 1947, ‘Early History and Incidents in the Life of William Crisp’, typescript, Mitchell Library A926.3/C.

Cross, Rex L. 1985, Bygone Queanbeyan, revised edition, Queanbeyan Publishing Co, Queanbeyan.

Curr, E. M. 1887, The Australian race, vols 3-4, Government Printer, Melbourne.

Dench, Alan 2000, ‘Comparative Reconstitution’, in Historical Linguistics 1995: Volume 1. General Issues and non-Germanic Languages. Selected Papers from the 12th International Conference on Historical Linguistics, Manchester, August 1995, John Charles Smith and Delia Bentley (eds), John Benjamins, Amsterdam/Philadelphia: 57-72.

Dixon, R. M. W., W. S. Ramson and Mandy Thomas 1990, Australian Aboriginal Words in English: Their Origin and Meaning, Oxford University Press, Melbourne.

du Vé, Charles 1887, ‘No. 197: Moneroo’, in The Australian Race, vol. 3, E. M. Curr, Government Printer, Melbourne: 429-431.

Eades, Diana Kelloway 1976, The Dharawal and Dhurga Languages of the New South Wales South Coast, Australian Institute of Aboriginal Studies, Canberra.

Eyre, Edward John 1984, Autobiographical Narrative of Residence and Exploration in Australia 1832-1939, (edited with an introduction and notes by Jill Waterhouse), Caliban Books, London.

Fitzhardinge, L. F. 1975, Old Canberra, 1820-1900 and The Search for a Capital, Canberra and District Historical Society, Canberra.

Flood, Josephine. 1980. The Moth Hunters: Aboriginal Prehistory of the Australian Alps. Australian Institute of Aboriginal Studies, Canberra.

Gale, John 1927, Canberra: History and Legends Relating to the Federal Capital Territory of the Commonwealth of Australia, A. M. Fallich and Sons, Queanbeyan.

— 1985, ‘Six Days on the Mountains of Cowley’, in Barrow 1985: 17-41 [first published Queanbeyan Age, 1875, reprinted with amendments Queanbeyan Observer, 1903]

Garnett, Rodney and Danielle Hyndes (eds) 1992, The Heritage of the Australian Capital Territory, National Trust of Australia (ACT), Canberra.

Gibbney, H. J. 1989, Eurobodalla: History of the Moruya District, Library of Australian History, Sydney [first published 1980].

Gillespie, Lyall 1991, Canberra, 1820-1913, Australia Government Publishing Service, Canberra.

— 1992, Ginninderra, Forerunner to Canberra: A History of the Ginninderra District, The Wizard Canberra Local History Series, Lyall Gillespie, Canberra.

Hancock, W. K. 1972, Discovering Monaro, Cambridge University Press, Cambridge.

Hercus, Luise A. 1986, ‘Southern Ngarigu’, in Victorian Languages: A Late Survey, Pacific Linguistics B-77, Canberra.

Howitt, A. W. 1996[1904], The Native Tribes of South-East Australia, facsimile edition, Aboriginal Studies Press, Canberra [first published 1904, Macmillan, London].

Jackson-Nakano, Ann 2001, The Kamberri: A History of Aboriginal Families in the ACT and surrounds, Aboriginal History Monograph 8, Aboriginal History, Canberra.

— 2005, Ngambri Ancestral Names: For Geographical Places and Features in the Australian Capital Territory and Surrounds, Ann Jackson-Nakano, Canberra.

Jones, Charles 2006, English Pronunciation in the Eighteenth and Nineteenth Centuries, Palgrave Macmillan, Basingstoke.

Lamb, Rebecca 2006, Macquoid of Waniassa: Portrait of a Colonial Sheriff, Waniassa Publications, Wanniassa.

Larmer, [James] 1898, ‘Native Vocabulary of Miscellaneous New South Wales objects’, Journal and Proceedings of the Royal Society of New South Wales no. 32: 223-229.

Lhotsky, John 1839, ‘Some Remarks on a Short Vocabulary of the Natives of Van Diemen Land; and also of the Menero Downs in Australia’, Journal of the Royal Geographic Society, London, no. 9: 157-162.

— 1979, A Journey from Sydney to the Australian Alps, Undertaken in the Months of January, February and March, 1834, (edited by Alan E. J. Andrews), Blubber Head Press, Hobart.

Mackaness, George (ed.) 1941, George Augustus Robinson’s Journey into South-Eastern Australia, 1844 with George Henry Haydon’s Narrative of part of the same Journey, Australian Historical Monographs no. 19 (NS), George Mackaness, Sydney.

Massola, Aldo 1968, Aboriginal Place Names of South-Eastern Australia and their Meanings, Landsdowne Press, Melbourne.

Mathews, R. H. 1901, ‘The Thurrawal Language’, Journal and Proceedings of the Royal Society of New South Wales, no. 35: 127-160.

— 1901-1902, ‘The Thoorga language’, Queensland Geographical Journal, no. 17: 49-73.

— 1903, ‘Vocabulary of Kamilaroi and Thurrawal Words’, pp. 275-279 of ‘Languages of the Kamilaroi and other Aboriginal tribes of New South Wales’, Journal of the Anthropological Institute of Great Britain and Ireland, no. 33: 259-283.

— 1904, ‘The Ngunawal Language’, pp. 294-299; ‘Vocabulary of Ngunawal Words’, pp. 302-305 of ‘The Wiradyuri and other Languages of New South Wales’, Journal of the Anthropological Institute of Great Britain and Ireland, no. 34: 224-305.

— 1908, ‘Vocabulary of the Ngarrugu Tribe N.S.W.’ Journal of the Royal Society of New South Wales, no. 42: 335-342.

Mawer, Allen 1983, ‘Pioneering Exploits on Canberry Station’, Canberra Historical Journal, New Series no. 11: 11-16.

McNicol, Sally and Dianne Hosking 1994, ‘Wiradjuri’, in Macquarie Aboriginal Works: A Dictionary of Words from Australian Aboriginal and Torres Strait Islander Languages, Nick Thieberger and William McGregor (eds), The Macquarie Library, Sydney: 79-99.

Merritt, John 2003 Currango Summers: A Snow Belt Pastoral Property 1851-1946, Turalla Press, Bungendore.

Mitchell, Felix F. 1926, ‘Back to Cooma’ Celebrations, The direct Publicity Company, Sydney.

Moore, Bruce 1999, Cotter Country: A History of the Early Settlers, Pastoral Holdings and Events in and around the County of Cowley, NSW, Greg Moore, Lyneham, ACT.

Mowle, S. M. 1891, Aboriginal songs (A few words of their language and mis-spelt names of places), MS 565/6/10, National Library of Australia.

Moye, D. G. (ed.) 1959, Historic Kiandra: A Guide to the History of the District, Cooma-Monaro Historical Society, Cooma.

Murphy, Greg 1987, ‘Canberra. What’s in the Name?’, Canberra Historical Journal, New Series no. 19: 14-17.

Nulladolla 1988, n.d., Milton-Ulladulla and District Historical Society, Ulladulla.

RASA Manuscripts 1900, [re-issued in a CD-ROM by the Geographical Names Board of NSW in 2003].

Reed, A. W. (compiler) 1977, Aboriginal Words and Place Names, Rigby, Adelaide.

Ridley, Rev. William 1887, ‘No. 192: Botany Bay’, in The Australian Race, vol. 3, E. M. Curr, Government Printer, Melbourne: 414-416.

Robinson, Frederick W. 1927, Canberra’s First Hundred Years and After, 2nd ed. revised and enlarged, W. C. Penfold, Sydney.

Robinson, G. A. 1998, The Journals of George Augustus Robinson, Chief Protector, Port Phillip Aboriginal Protectorate, Vol. 4: 1 January 1844 - 24 October 1845, (edited by Ian D. Clark), Heritage Matters, Melbourne.

— 2000, The Papers of George Augustus Robinson, Chief Protector, Port Phillip Aboriginal Protectorate, Vol. 2: Aboriginal Vocabularies: South East Australia, 1839-1852, (edited by Ian D. Clark), Heritage Matters, Clarendon, Victoria.

Salisbury, Jean 2000, St John’s Churchyard 1844-1998, The Heraldry & Genealogy Society of Canberra Inc, Canberra.

Shoalhaven Holidays. n.d. Shoalhaven City Council, Nowra.

Snowden, H. E. 2004, Pioneers of the Tumut Valley: The History of Early Settlement, Tumut & District Historical Society Inc, Tumut, NSW.

Troy, Jakelin 1994, ‘Melaleuka: A History and Description of New South Wales Pidgin’, 2 vols, PhD thesis, The Australian National University, Canberra.

Walsh, Michael 2002, ‘Transparency versus opacity in Aboriginal placenames’, in The Land is a Map: Placenames of Indigenous Origin in Australia, Hercus, Luise; Flavia Hodges and Jane Simpson (eds), Pandanus Books and Pacific Linguistics, Canberra: 43-49.

Watson, Don 1997, Caledonia Australis: Scottish Highlanders on the Frontier of Australia, Vintage, Random House, Sydney [first published 1984, William Collins, Sydney].

Watson, Frederick 1927, A Brief History of Canberra, the Capital City of Australia, Federal Capital Press, Canberra.

Wells, J. C. 1982, Accents of English, Cambridge University Press, Cambridge.

Wilson, Gwendoline 2001[1968], Murray of Yarralumla, Tabletop Press, Canberra [first published 1968, Oxford University Press, Melbourne].

Wright, W. Davis 1923, Canberra, J. Andrew, Sydney.

Young, Michael; Ellen Mundy and Debbie Mundy 2000, The Aboriginal People of the Monaro: A Documentary History, NSW Parks & Wildlife Service.

1 This paper incorporates material presented in four workshops as: ‘Placenames in the ACT and southeastern NSW’ (Canberra, 5 December 2002), ‘Linguistic Reconstruction of Placenames’ (Canberra, 12 June 2004 and Yass, 16 May 2005), and ‘ACT Place Names: Towards reconstructing their original Indigenous form’ (Canberra, 1 October 2005).

2 The (Royal) Anthropological Society of Australasia in the 1890s circulated a questionnaire which asked for “the Native Names of places with their meanings or the reason why the blacks gave such names to the separate localities”. Some results were published in issues of Science of Man. The 1900 set of manuscripts have been microfilmed and were issued as RASA Manuscripts in a CD-ROM by the Geographical Names Board of NSW in 2003. This work will be referred to in this paper as RASA Manuscripts 1900.

3 “The Congwarra run is believed to have been named after an Aboriginal elder called Kongwarra who often camped in the locality. Some early references to the run spell the name Kongwarra and others Congwarrah” (Moore 1999: 79). We need to bear in mind, however, that Aboriginal people sometimes were named after a place.

4 A related question, how it came to be bestowed as a name in the English nomenclature system, is part of the historical study that should serve as preliminary to the kind of reconstruction I am primarily interested in here (see the section ‘What history offers’).

5 In 1825 J. J. Moore first obtained permission to occupy the land on which he set up a station probably already in 1823; his letter requesting purchase of land at ‘Canbery’ was dated 16 December 1826 (Gillespie 1991: 8-9).

6 Historians, however, are not necessarily good judges of the linguistic abilities of early recorders of Aboriginal placenames.

7 According to Bluett (1954: 1), “one group camped at Pialligo and was known to the early settlers as the Pialligo blacks”.

8 One should not overlook possible Asian sources of names, especially from British India. A recent discovery is that ‘Waniassa’, the name of Thomas Macquoid’s estate in the Tuggeranong Valley, was transferred from the name of his earlier home in the Krawang district of northeastern Java in the village of Waniassa/Wanajasa (Lamb 2006).

9 A possible explanation is that the second vowel of the sequence jinin was unstressed and so short that it may not have been heard, which would have caused the two n’s to be run together in a way that they sounded like a single consonant.

10 This example also shows that, while the shortened version is later than the longer versions, there are two longer versions. We cannot separate them chronologically and simple-mindedly declare the earliest to be the most genuine.

11 The same name seems to be behind a place on the western side of Lake George, spelled ‘Purrorumba’ on Thomas Mitchell’s 1834 map of NSW (Watson 1927: 18) and on a map of parishes in County Murray (Moore 1999: Map 8).

12 The jocular ‘Didyabringabeeralong’ reflects this preference.

13 One source even claims that ‘Bodalla’ means “a haven for boats” – and that ‘Urobodalla’ (Eurobodalla) means “another haven for boats” (D. J. Stinson in RASA Manuscripts 1900)!

14 A similar situation obtained in England, where Anglo-Saxon placenames compounded with -field were sometimes adapted to a more French-sounding name in -ville.

15 The Rev. P. G. Smith’s diaries provide an additional spelling ‘Goongahleen’ – in 1880 during the Crace era (Cope 2006: 78).

16 It is customary for linguists to represent phonemes (contrastive sounds) between slashes and variant pronunciations (called ‘allophones’) in square brackets.

17 In principle we could add variants with t in place of d; in practice, however, it is usually the voiced sound (i.e. d) that occurs after a nasal sound like n.

18 See Table 5.2 for the sounds that are expected in Aboriginal languages in this part of the country.

19 Cf. Clarke (1986: 278): “When he accompanied a parliamentary delegation in 1891 to Tumut, or Doomut-th as, with his obsession with the original Aboriginal pronunciation, he insisted on calling it”.

20 Another example of a final trill being perceived as having an s-like quality is given by Charles Throsby, in his letter of 9 September 1820 to Governor Macquarie, which mentions a “large space of water called by the natives Ber-ree-warz or Bur-rur-wars, which I saw when at Jervis Bay” (Gale 1927: 22) – possibly St. George’s Basin.

21 See the discussion of Ulladulla at (6) below for an application of this principle.

22 This procedure is also described and illustrated in Austin and Crowley (1995). Parallels with the ‘comparative method’ for reconstructing ancestral languages from differently diverged descendants are explored in Dench (2000).

23 Jones (2006: 329) quotes an 1886 comment from the English pronunciation expert, A. J. Ellis, that “Irishmen are noted for giving eye a shade of oy”.

24 Similar variation is seen in the alternation between ‘Umeralla’ and ‘Umarally’, ‘Tidbinbilla’ and ‘Tidbinbilly’, ‘Canberra’ and ‘Canbury’.

25 Cf. the property ‘Breadbatoura’ established near Cobargo on the south coast in 1835 (Gibbney 1989: 26).

26 Similarly the first Catholic priest in the area, Father James O’Doherty, from Ireland, wrote the placename ‘Mollymook’, which was also spelled ‘Mollymoke’, as ‘Mollymuck’ – using u for the same u/o sound.

27 This form introduces a complication, however, in the absence of the final rra syllable. Perhaps it was a locative suffix.

28 The term given for the expected local language Dhurga, is the slightly different form koor´-a-dhoo (Mathews 1901-1902: 69).

29 Cf. Hancock’s comment that the property name that Crisp spelled Jimenbuan meant ‘big fat kangaroo rat’ (Hancock 1972: 107 fn 1).

30 The linguistic evidence suggests that the word ends in ng rather than the n of the placename; there is some doubt about whether the second vowel is a (as given by Mathews or u as given by Hercus); Crisp’s vowel e probably reflects an English pronunciation with an indistinct vowel.

31 A reduplicated form of the word for ‘fat’ is presumably also the basis for the placename Puenbuen northwest of Bega on the south coast.

32 Diana Eades comments on the equivalent word in Dharawal, the language of Wollongong area, claiming that /byuwan/ ‘butter, fat’ is the only word that starts with a sequence of two consonants (Eades 1976: 77). I suggest rather that the word was really /biwan/ in terms of the Aboriginal phonemes, was pronounced as [biuan], which in turn was heard by English speakers as [byuan] and consequently spelled buan, with u representing the sequence [yu].

33 The Victorian Aboriginal Placename Project explicitly sought such information from professional linguists familiar with the relevant languages (Clark and Heydon 2002: 8).

34 For a history of the property see Merritt (2003).

35 The word for ‘rock’ or ‘stone’ is given as bura for the languages of Bega and Bateman’s Bay, but not the inland area.

36 The following account is taken primarily from Mawer (1983). Other detailed sources are Robinson (1927) and Fitzhardinge (1975).

37 The names of Canberra’s satellite cities, Woden, Belconnen, Tuggeranong and Gungahlin, have likewise undergone considerable expansion from former property names.

38 Of course the name ‘Canberra’ is also now also used, metonymically, for the federal government.

39 This theory that Kamberra/Nganbra was a group name has been followed in recent work by Jackson-Nakano (2001, 2005).

40 See Murphy (1987) for a survey of proposed meanings of the name.

41 He must mean the rectory, since Canberry Cottage was used for many years as the residence of the church’s minister.

42 Thus I analyse the South Coast placename ‘Bergalia’ (near Moruya) as being from the locative *Burrgili-ya, the uninflected form being reflected in Howitt’s Bugelli-Manji clan name and John Hawdon’s Burgaily squattage (see Gibbney 1989: 14, 25).

43 The word is derived from a verb nganbi- ‘lean’.

Endnotes

  1. This paper incorporates material presented in four workshops as: ‘Placenames in the ACT and southeastern NSW’ (Canberra, 5 December 2002), ‘Linguistic Reconstruction of Placenames’ (Canberra, 12 June 2004 and Yass, 16 May 2005), and ‘ACT Place Names: Towards reconstructing their original Indigenous form’ (Canberra, 1 October 2005).
  2. The (Royal) Anthropological Society of Australasia in the 1890s circulated a questionnaire which asked for “the Native Names of places with their meanings or the reason why the blacks gave such names to the separate localities”. Some results were published in issues of Science of Man. The 1900 set of manuscripts have been microfilmed and were issued as RASA Manuscripts in a CD-ROM by the Geographical Names Board of NSW in 2003. This work will be referred to in this paper as RASA Manuscripts 1900.
  3. “The Congwarra run is believed to have been named after an Aboriginal elder called Kongwarra who often camped in the locality. Some early references to the run spell the name Kongwarra and others Congwarrah” (Moore 1999: 79). We need to bear in mind, however, that Aboriginal people sometimes were named after a place.
  4. A related question, how it came to be bestowed as a name in the English nomenclature system, is part of the historical study that should serve as preliminary to the kind of reconstruction I am primarily interested in here (see the section ‘What history offers’).
  5. In 1825 J. J. Moore first obtained permission to occupy the land on which he set up a station probably already in 1823; his letter requesting purchase of land at ‘Canbery’ was dated 16 December 1826 (Gillespie 1991: 8-9).
  6. Historians, however, are not necessarily good judges of the linguistic abilities of early recorders of Aboriginal placenames.
  7. According to Bluett (1954: 1), “one group camped at Pialligo and was known to the early settlers as the Pialligo blacks”.
  8. One should not overlook possible Asian sources of names, especially from British India. A recent discovery is that ‘Waniassa’, the name of Thomas Macquoid’s estate in the Tuggeranong Valley, was transferred from the name of his earlier home in the Krawang district of northeastern Java in the village of Waniassa/Wanajasa (Lamb 2006).
  9. A possible explanation is that the second vowel of the sequence jinin was unstressed and so short that it may not have been heard, which would have caused the two n’s to be run together in a way that they sounded like a single consonant.
  10. This example also shows that, while the shortened version is later than the longer versions, there are two longer versions. We cannot separate them chronologically and simple-mindedly declare the earliest to be the most genuine.
  11. The same name seems to be behind a place on the western side of Lake George, spelled ‘Purrorumba’ on Thomas Mitchell’s 1834 map of NSW (Watson 1927: 18) and on a map of parishes in County Murray (Moore 1999: Map 8).
  12. The jocular ‘Didyabringabeeralong’ reflects this preference.
  13. One source even claims that ‘Bodalla’ means “a haven for boats” – and that ‘Urobodalla’ (Eurobodalla) means “another haven for boats” (D. J. Stinson in RASA Manuscripts 1900)!
  14. A similar situation obtained in England, where Anglo-Saxon placenames compounded with -field were sometimes adapted to a more French-sounding name in -ville.
  15. The Rev. P. G. Smith’s diaries provide an additional spelling ‘Goongahleen’ – in 1880 during the Crace era (Cope 2006: 78).
  16. It is customary for linguists to represent phonemes (contrastive sounds) between slashes and variant pronunciations (called ‘allophones’) in square brackets.
  17. In principle we could add variants with t in place of d; in practice, however, it is usually the voiced sound (i.e. d) that occurs after a nasal sound like n.
  18. See Table 5.2 for the sounds that are expected in Aboriginal languages in this part of the country.
  19. Cf. Clarke (1986: 278): “When he accompanied a parliamentary delegation in 1891 to Tumut, or Doomut-th as, with his obsession with the original Aboriginal pronunciation, he insisted on calling it”.
  20. Another example of a final trill being perceived as having an s-like quality is given by Charles Throsby, in his letter of 9 September 1820 to Governor Macquarie, which mentions a “large space of water called by the natives Ber-ree-warz or Bur-rur-wars, which I saw when at Jervis Bay” (Gale 1927: 22) – possibly St. George’s Basin.
  21. See the discussion of Ulladulla at (6) below for an application of this principle.
  22. This procedure is also described and illustrated in Austin and Crowley (1995). Parallels with the ‘comparative method’ for reconstructing ancestral languages from differently diverged descendants are explored in Dench (2000).
  23. Jones (2006: 329) quotes an 1886 comment from the English pronunciation expert, A. J. Ellis, that “Irishmen are noted for giving eye a shade of oy”.
  24. Similar variation is seen in the alternation between ‘Umeralla’ and ‘Umarally’, ‘Tidbinbilla’ and ‘Tidbinbilly’, ‘Canberra’ and ‘Canbury’.
  25. Cf. the property ‘Breadbatoura’ established near Cobargo on the south coast in 1835 (Gibbney 1989: 26).
  26. Similarly the first Catholic priest in the area, Father James O’Doherty, from Ireland, wrote the placename ‘Mollymook’, which was also spelled ‘Mollymoke’, as ‘Mollymuck’ – using u for the same u/o sound.
  27. This form introduces a complication, however, in the absence of the final rra syllable. Perhaps it was a locative suffix.
  28. The term given for the expected local language Dhurga, is the slightly different form koor´-a-dhoo (Mathews 1901-1902: 69).
  29. Cf. Hancock’s comment that the property name that Crisp spelled Jimenbuan meant ‘big fat kangaroo rat’ (Hancock 1972: 107 fn 1).
  30. 30. The linguistic evidence suggests that the word ends in ng rather than the n of the placename; there is some doubt about whether the second vowel is a (as given by Mathews or u as given by Hercus); Crisp’s vowel e probably reflects an English pronunciation with an indistinct vowel.
  31. 31. A reduplicated form of the word for ‘fat’ is presumably also the basis for the placename Puenbuen northwest of Bega on the south coast.
  32. 32. Diana Eades comments on the equivalent word in Dharawal, the language of Wollongong area, claiming that /byuwan/ ‘butter, fat’ is the only word that starts with a sequence of two consonants (Eades 1976: 77). I suggest rather that the word was really /biwan/ in terms of the Aboriginal phonemes, was pronounced as [biuan], which in turn was heard by English speakers as [byuan] and consequently spelled buan, with u representing the sequence [yu].
  33. The Victorian Aboriginal Placename Project explicitly sought such information from professional linguists familiar with the relevant languages (Clark and Heydon 2002: 8).
  34. For a history of the property see Merritt (2003).
  35. The word for ‘rock’ or ‘stone’ is given as bura for the languages of Bega and Bateman’s Bay, but not the inland area.
  36. The following account is taken primarily from Mawer (1983). Other detailed sources are Robinson (1927) and Fitzhardinge (1975).
  37. The names of Canberra’s satellite cities, Woden, Belconnen, Tuggeranong and Gungahlin, have likewise undergone considerable expansion from former property names.
  38. Of course the name ‘Canberra’ is also now also used, metonymically, for the federal government.
  39. This theory that Kamberra/Nganbra was a group name has been followed in recent work by Jackson-Nakano (2001, 2005).
  40. See Murphy (1987) for a survey of proposed meanings of the name.
  41. He must mean the rectory, since Canberry Cottage was used for many years as the residence of the church’s minister.
  42. 4Thus I analyse the South Coast placename ‘Bergalia’ (near Moruya) as being from the locative *Burrgili-ya, the uninflected form being reflected in Howitt’s Bugelli-Manji clan name and John Hawdon’s Burgaily squattage (see Gibbney 1989: 14, 25).
  43. The word is derived from a verb nganbi- ‘lean’.


Previous Next