[Note: this transcription was produced by an automatic OCR engine]

40

Purungu, Karimarra and Panaka are the best represented categories. According to the
lines of regression, their origin or point of entry into the desert seems to be i.r1 the region of the
Karadjeri, in the Pilbara, and their diffusion is from north-west to south-east into the desert.
We find the Karimarra and Panaka sections as far south-east as the Southern Aranda of Central
Australia, where the local forms are Kamara and Pananga Elkin 193 1 :7 1, and the Karimarra
section among the Alyawara, with the local form being Kimarra Yallop 1969.

Paljeii is found along the coastal region. Its diffusion seems to progress on a north—south
axis without penetrating far into the Western Desert. Tjarurru and Yiparrka are found only in
the southern part of the observed region. Their absence in the northern part, in particular in
the Pilbara and Kimberley regions, indicates that their origin is not to be found in the same
region as Purungu, Karimarra and Panaka. Milangka forms a corridor, with a route of diffusion
similar to that of the categories whose origin is in the Pilbara Pumngu, Karimarra, Panaka,
but it does not reach as far north. Burgulu is a category without cohesion and is not well
represented. It is absent from the Pilbara and the central and southern Western Desert, and is
limited to the south-west of the desert with a single presence in the north the Yulbaridya.

An example will demonstrate that it is justified to consider the lines of regression as a
first step in the ELABORATION of routes of diflirsion. In the Pilbara region, McConvell 1985a:1 1
distinguishes the term Paljani among the Ngarluma from the term Palyirri, a transcription
of the same section among the Kariera. For McConvell, the diflerence reflects, “a regular
sound law of Lenition and Regressive Vowel Assimilation”. He proposes two distinct routes
of diffusion for these two terms. Paljarri moves into the desert as far as Central Australia,
and Palyani moves north along the coast. The difference between these two terms was not
marked here and they are grouped in the same category, but the lines of regression confirm
McConvell’s second thesis as they follow the coast. If the Alyawara of Central Australia are
also included—their sections are Kimarra, Pijarra, Kngwarriya and Pula cf. Yallop 1969,
Pijatra being a transformation of Paljarri—the line of regression 1 draws a route from west
to east, from the Pilbara into the centre of the continent, confirming also the second route
described by McConvell, while line 1’ maintains its north-south path Map 7. The body of
data is too shallow to produce a significant statistical conclusion. It is, nevertheless, surprising
how well the lines of regression coincide with McConvell’s conclusions.



ins II7 /19 I21 I23 125 I27 I29 l3I I33 I35 I37 139

Map 7: Distribution of Paljeri including the Alyawnra