IXTRODrCTIO'. U
eat, Gothic ita, Latin edo, are from the same root. The Malay is ma-kan, of which the ma is also pa, ba, and with this corres- ponds the Melanesian (Efate) ba-mi, ' to eat.' Kow, it seems to me likely that in primitive speech there were, alongside of each other, three root-forms, ba, ad, and kad, of which ba and ad passed to the West and produced the Greek pha-go, and e(s)thio, the Latin edo, the English eat, while kad spread to the East and is the source of all the other words ; ba in a less degree accom- panied it, and gives bami (Efate), -ma-fa (Samoa), and the Malay ma-kan. This root ba seems also to exist in Australia, for one dialect has has a-balli, ' to eat.'
In the Samoan tau-tc (a chief's word), the tau is an intensive and therefore, in this case, honorific, prefix, and the te is our root ta ; it thus corresponds with the Tasmanian te-ganna.
In various parts of British Kew Guinea, words for ' eat ' ax'e bai, uai, mo-ana, kani-kani, an-an, ye-kai ; and for 'food,' kai, kan, ani-ani, ai-ai, mala-m, ala, wa-la. All these come from the roots ba and ka, kan; with an-an (an for kan) com- pare the Dra vidian un, 'to eat.'
Thus I dispose of the Awabakal root ta, 'to eat '; and, if the analogies given above are well founded, then I am sure that our Australian blacks have a share with the rest of the world in a common heritage of language.
When the radical syllable, ta, is removed, the i*emainder of our sample word is -killi-ko, and both of these are formative. On comparing ta-killi-ko with other Awabakal vex'bs, such as um- ulli-ko, wi-yelli-ko, nm-olli-ko, and with the Wiradhari verbs and verbals da-alli, d-illi-ga, b-illi-ga, it is obvious that the essential portion of the affix is -illi or -alii, the consonants before it being merely euphonic. In the Dravidian languages, similar consonants, v, y, m, oi, d, t, g, are inserted to prevent hiatus, and in Fiji and Samoa there is also a great variety of consonants used to introduce suffixes. Then, as to the -illi or -alii, I find exactly the same formative in Gond — an uncultured dialect of the Dravidian ; there the infinitive of a verb has -ale o?* -ile ; and in Tamil, the verbal noun in -al, with the dative sign -ku added, is used as an infinitive ; in Canarese the -al is an infinitive without the -ku. In all this we have a close parallel to the Awabakal infinitive in -alli-ko, -illi-ko, for some of our dialects have the dative in -ol, -al.* Our formative, when attached to a verb-root, makes it a verbal noun, as bun-killi, 'the act of smiting '; hence the appropriateness of the suffix -ku, 'to,' a post-position.
The -ko in ta-killi-ko is equivalent to the English 'to' with verbs, except that it is used as a post-position in Awabakal, where it is the common dative sign. It also resembles, both in form and
- 8ee page 49 of Appendix.
�� �
