Kartiya, kartipa – Barry Alpher

[From our kartiya in Washington, Barry Alpher]
In a query to David Nash’s posting (4 October) on munanga ‘white person’ in languages of Arnhem Land, Joe Blythe asks “So what about kartiya [the term for ‘white person’ in a number of Ngumpin-Yapa languages]? Any ideas?”

Here are a couple.

At least three languages attest kartiya: Walmajarri, Gurindji, and Warlpiri (in the form kardiya). Mudburra attests kardiba in the same meaning, and Gurindji attests kartipa as a variant of kartiya. (Note that in view of the Gurindji change *rt > r [Pat McConvell, pers. comm.; see under *kartu below], both of these Gurindji variants must be reckoned as loans.)

Of especial interest about these forms is first that none of these languages (in the sources that I have consulted, although a comment from Felicity Meakins indicates otherwise) attaches the sense ‘ghost’ or ‘corpse’ to them, despite the likelihood that these are the original senses of the terms for ‘white person’ in a great many Australian languages. (The term Warlpiri kardiya-jarrimi, glossed as a euphemism for ‘to die’ [Swartz dictionary], does however attest an association with death). Secondly, the variants with –ya and –pa suggest that the karti part might originally have had an independent existence.
There are two plausible origins for kartiya. The first is Warlpiri kardirri ‘white’ (not a term for ghosts). Adaptation of this to form the kartiya/pa terms would have required the interpretation of the –rri part as a suffix of some kind.
The other plausible origin is a ‘person’ term, *karti, as suggested by the following data:
*karti

Warumungu                            kartti ‘person, Aborigine; man; fully initiated man’; also kartikiji ‘son (to male), daughter (to male)’; kartunguñu ‘wife’

Djinang                                   garditi ‘sister’ (it is possible that the *i of the second syllable is a development from *a; see the Dabi form below)

*kartu

Adnyamathanha                      artu ‘woman’

Nukunu                                   kartu ‘wife’

Ngadjuri                                  atuni ‘woman, wife’

Narangga                                 gaRtu ‘wife’ (“R” is indeterminate as to glide, tap, trill; also cited as kaRtu,gaTu)

Kaurna                                    karto ‘wife’

Wirangu                                  gardu ‘man, fully initiated man, young initiate’

Warumungu                            kartunguñu ‘wife’ (see *karti)

Warlpiri                                   kardu ‘mother; mother’s brother’ (also, dyads such as mother and child, uncle and nephew, etc.)

Gurindji                                   karu ‘child’ (lenition of *rt to r is regular)

Walmajarri                               kartu (used also to refer to other women of the same subsection as a man’s wife)

Ngarluma                                kartu ‘man; male (as of animal)’; plural kartupura.
Forms which are obviously related but which come from languages about whose phonology or phonological history too little is known to permit assignment with confidence to one or another of these etyma include these:

Barngarla                                 karteti ‘wife’

Dabi                                        gardak ‘sister’

Dalabon                                   g3rdvg3rt ‘woman’
That some of these forms refer to human beings in general and others to kin categories, and some are specific to the male gender and others to the female, does not argue against their cognacy, since other sets of cognates, like *yapa and *yapu (variously ‘older sister’, ‘older brother’, ‘Aboriginal person’) exhibit the same range.
The cognacy of the –u stems (*kartu) with the –i stems (*karti), while as yet undemonstrated, seems to have a parallel in other Pama-Nyungan human-categorisation forms, like *yapa and *yapu (above) and (among others) *ngama, *ngami, *ngamu (with senses ‘breast, milk’, ‘mother’, and ‘mother’s brother’): this is a matter of morphology.
I can offer no immediate solution, however, for the choice between ‘white’ (Warlpiri kardirri) and ‘person’ (*karti and *kartu).

---

Thanks to David Nash and Jane Simpson for suggestions.