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

44

be identical by relation, but also by assimilation. The difference between these two types
of identities is that, in the first type relational, sections are only identical because of their
relations with other sections, because of their structural position. They are not “organically”
merged. In the second type of identity assimilatory, two section names are conflated. They
are “organically” merged. Thus, in assimilatory identity, the two sections are simply two
names for a single section, while in relational identity the section names are associated with
discrete social units distinct people, when, for example, two dialectal groups meet but stand
in an identical structural position. The distinction will be further discussed within the case
study of the Pintupi change from a four-section system to an eight-subsection system.

IILb Second hypothesis: the zone of extension and substitution of Paljeri

The mutual substitution of Paljeri and Tjarurru in the south-westem part of the region was
noted by Bates 1985: 104-5 for the Koara. This rule may be extended to the Kalamaia and the
Waljen, as they are close to the southern extension of Paljeri and within the northern extension
of Tjarurru in this same geographical area.

The hypothesis of mutual substitution of Paljeri and Milangka in the central north—western
area, among the Njangamarda Nyangumarta, is put forward by McConve1l 1985a:l1. This
rule may be extended to the Mardu, as they seem to have adopted the Inland Njangamarda
system. We shall have to spend more time on the case of the Njangamarda. While the Coastal
Njangarnarda seem to use Milangka and Paljeri Paljan-i interchangeably,” the Inland
Njangamarda do not use Paljeri, but also use Milangka in a diflerent structural position from
its use among Coastal Njangamarda.

The hypothesis of mutual substitution of Paljeri and Yiparrka has, to my knowledge, not
been proposed. If Paljeri is replaced by Tjarunu in the south—west and by Milangka in the
north-west, yet not by Tjarurru in the south—east, then Paljeri may theoretically be replaced by
Yiparrka in the latter area I return to this hypothesis later, only noting here that if Paljeri is
identical to Yiparrka, and if Yiparrka is identical to Panaka the first hypothesis, then Paljeri
should be identical to Panaka, which is, as we saw in the spatial distribution of individual
terms, impossible.

Milangka and Tjarurru are never combined when one of them substitutes for Paljeri.
Thus, in the central western part of the region, when Milangka is used, it replaces Paljeri,
and, in the south-westem part, when Tjarurru is used, it replaces Paljeri. Bates 1925 writes
that, for the region of Mount Jackson probably for people who were later called the Kalaako
Tjarunu replaces Paljeri.

It seems, therefore, appropriate to draw a zone forming a vertical slice between the west
coast of the continent and approximately 123° of longitude, which could be called the zone of
extension and replacement of Paljeri, coded as “Paljeri —> Milanglca : Tjarurm”. Paljeri does

27 McConvell pointed out to me that this system is identical to the Yindjibamdi sections Dench 199518. The
latter, however, have reversed the generations so that Njnngumartzfs Panaka = Punlngu is equivalent to Y1ndjibandi’s
Karimama = Paljeri; and Nja.ngumarta’s Karimarna = Milangka/Paljeri is identical to Yindjibandi’s Panaka =
Punmgu.