Now--with these preliminary cautions against the over-valuation of apparent differences--we may compare the new data for the structure of the Kowrarega and Limbakarajia with the reccived opinions respecting the Australian grammars in general.
These refer them to the class of agglutinate tongues, i.e. tongues wherein the inflections can be shown to consist of separate words more or legs incorporated or amalgamated with the roots which they modify. It may be said that this view is confirmed rather than impugned.
Now, what applies to the Australian grammars applies also to Polynesian and the more highly-developed Malay languages, such as the Tagala of the Philippines, for instance; and, if such being the case, no difference of principle in respect to tkeir structure separates the Australian from the languages of those two great classes. But the details, it may be said, differ undoubtedly; and this is what we expect. Plural numbers, signs of tense, and other grammatical elements, are evolved by means of the juxtaposition of similar but not identical elements, e.g. one plural may be formed by the affix signifying many; another, by the affix signifying with or conjointly; one preterite may be the root plus a word meaning then; another the root plus a word meaning there. Futures, too, may be equally evolved by the incorporation or juxtaposition of the word meaning after, or the word meaning to-morrow. All this makes the exact coincidence of the details of inflection the exception rather than the rule.
This doctrine goes farther than the mere breaking-down of the lines of demarcation which separate classes of languages like the Australian from classes of languages like the Malayo-Polynesian. It shows how both may be evolved from monosyllabic tongues like the Chinese or Siamese. The proof that such is really the case lies in the similarity of individual words, and consists in comparative tables. It is too lengthy for the present paper, the chief object of which is to bring down the inferences from the undoubtedly great superficial differences between the languages of the parts in question to their proper level.
In respect to the vocabularies, the extent to which the analysis which applies to the grammar applies to the vocables also may be seen in the following instance. The word hand in Bijenelumbo and Limbapyu is birgalk. There is also in each language a second form--anbirgalk--wherein the an is non-radical. Neither is the alk; since we find that armpit = ingamb-alk, shoulder = mundy-alk, and fingers = mong-alk. This brings the root = hand to birg. Now this we can find elsewhere by looking for. In the Liverpool dialect, bir-il = hand, and at King George Sound, peer = nails. The commonest root, = hand in the Australian dialects, is m-r, e.g.:
Moreton Bay : murrah.
Corio : far-onggnetok.
Karaula : marra.
Jhongworong : far-okgnata.
Sydney : da-mora.
Murrumbidje : mur-rugan.
Mudje : mara.
Molonglo : mar-rowla.
Wellington : murra.
Head of Bight : merrer.
Liverpool : ta-mura.
Parnkalla : marra.
All this differs from the Port Essington terms. Elbow, however, in the dialects there spoken, = waare; and forearm = am-ma-woor; wier, tao, = palm in Kowrarega.
To complete the evidence for this latter word being the same as the m-r of the other dialects and languages, it would be necessary to show, by examples, how the sounds of m and w interchange; and also to show (by example also) how the ideas of elbow, forearm, and hand do so. But as the present remarks are made for the sake of illustrating a method, rather than establishing any particular point, this is not necessary here; a few instances taken from the names of the parts of the human body being sufficient to show the general distribution of some of the commoner Australian roots; and the more special fact of their existence in the northern dialects:
English : hand.
Peel River : ma.
Terrutong : manawiye.
Raffles Bay : maneiya.