SOME EEMAKKS ON THE AUSTEALIAN LANGUAGES. 245 



•exists from Tasmania in the extreme south, right on to the Gulf 

 ■of Carpentaria. If you ask me why there is only one word for 

 ' two,' while the words for 'one ' are so numerous and different, I 

 reply that, among other languages, and especially in the Turanian 

 family, there is a similar diversity in the words for 'one'; and 

 the reason is this that, wherever there is a considerable number of 

 words for 'origin,' 'commencement,' 'before,' &c, there will be a 

 similar variety in the words for ' one,' which are formed from them. 

 But the range of ideas for ' two ' is somewhat limited ; the only 

 ideas possible are 'repetition,' or 'following,' or something similar. 

 Let me show you this by a few examples. The Hebrew shenaim, 

 'two,' is a dual form, and is connected with the verb shanah, 'to 

 repeat;' the Latins also say ' vigesimo altero anno' to mean in 

 the 'twenty second year;' but alter is 'the other of two,' and in 

 French and English it means to 'change;' and secundus in Latin 

 comes from sequor, ' I follow.' Thus we shall find that words 

 for 'two' are the same as words for 'follow,' 'repeat,' 'another,' 

 'again,' 'also,' 'and,' and the like; and most of these ideas are 

 usually expressed by forms of the same root-word. 



As to the form of the word bula, we have here no friendly 

 karaji to tell us whether the -la is radical or not. I" think that 

 the -la is formative, and that it indicates the dual number, the 

 bu- being thus the root. The Tasmanian bu-ali (Milligan writes 

 it pooalih) is probably the nearest approach to the original form, 

 the bu being the root and the -ali the dual affix; these would 

 easily coalesce into bula. In the Tasmanian pia-wa, the pi a seems 

 to me to be a dialect form of bula, for the liquid I easily drops 

 out, and in the Aryan languages a modified u approaches very 

 nearly to the sound of i (cf. Eng., sir); in the Polynesian, i often 

 takes the place of u. Thus, bula would become bu-a, bi-a, pia. 

 The wa in pia-wa, as in marawa ' one,' is only a suffix, the same 

 as b a in our colony. All the other words for 'two' are only 

 lengthened forms of bula. 



As to the kindred of bula, I find that, in the Papuan island of 

 Aneityum (New Hebrides), the word in-mulis 'twins'; there the 

 in is the common prefix used to form nouns; the mul that remains 

 is bul, 'two'; there also um, for mu, is 'and'; in the other islands 

 it is ma, mo. In New Britain, bal-et is 'again,' bul-ug, 'again,' 

 'also,' 'another,' mule, 'again,' bula, 'another,' ' an additional 

 one' (cf ma, 'and'), bula, ka-bila, 'also' (with -bila cf. Tasm. 



is the sound of a/r; just as a Cockney will say 'idear' for ' idea/ ' mar' 

 for ' ma/ or 'planer' for 'piano/ In one vocabulary that I have seen 

 almost every word terminates with r on this principle ! 



It is evident also that the same mistaken principle vitiates the spelling* 

 of some of the words for ' one/ given on a previous page ; for instance, 

 m a r d a ought to be m a-d a, and b a r d j a should be b a-j a. 



