•2 



ETIINOLOOY OF THE INDO-PACIFIC ISLANDS. 



aoiuuls Ijaviiig tlio stime proctsc meoining in unconnoctcil Inrtpfungos. 

 This calciilalion of probabililios is evidently bused on a iofrnulion 

 of wbiclr tliD syllablea are frequently biconsonsyital siml tlie vvoi'ds 

 polysjllubic. It is totally inajjplicablG to monosyiJabic liujguaj^ea, 

 or iiiileed to a comparison of uUimato roola in any formation, 

 bccanso tbese rools are generally motiosyllLibles. In tbc Kwan- 

 hvva Chinese, for cx!irn]ile, the inimbcr of words is about 48,(KXI, 

 bnt tbe sounds to express those words only amonnt, oven wiih the 

 tone flexions, to 1,2* J3, So llnit ench sonnd, on an avenige, would 

 represent 40 diflurent words if these words were ail in uhc. Bj 

 enlarging: the number of monosyllabic bingnn|T«s for comparison, 

 the number of homopljons increases. But this is not all. In thi3 

 progress of language the tones decay, become reduced in number 

 and are iiltimatcly lost. When tbe process of em:iscuIiition goes 

 on without interruption, the vowel Bomids are contracted to a very- 

 small number. In Philipine there arc only three distinct vowels, 

 0 and Uf i and e being very com mutable* Add to this that in 



rnrativc studies in onler to tnke in tang-iia^ps mth wliicli fliey ate imperfectiy 

 nffiiiniiitfid, A complete invTsfi^alionuf iJte Hlinifl liistcry i>f u single root ittMunnda 

 a tliorougii knowledge of all llit: lang;ij»f?es in tlie world «Tid no jiiiigl« iiliiiojogtst 

 cnii ever attain tlits knowledge. Hence it h only by combining tuid comparing tlie 

 labour;* of unnieroua compiirative linjiaiets that tbe ellmolojjy of rools will ulti- 

 niatdy hn pCTlRCfod. Hoof s ramify througb vocnbubirita in a vfiry complicufed 

 niunitcr cbnngiiig not only tiitir foi ms hut tlietr iix jininysi, so tiint ii h not jioMible, 

 by niovfJy tniniing over llie Iwives of a dictionttiy. fu a certtiiii wiicibtr ti piven 

 root exists iti n jiarticubir langawge ornoL \V« niusf know tite plioiiolo|rj' of tfte 

 las guage, ila pbonetii! Hii-l (il«s?arial rft alinns to otbtr iauiiuny^es, uhtl iJie k^nif of 

 analogies tbat prevail tbronuboul its gl()9,<ary tniil ensibk ns to tract; tlic mct^mor- 

 pltoses of its roots. This only man wko can pronounce wbclhur a givijn ruot exi-ita 

 cr not hi a particular language, is a s-ound conipaiativu liiimiist wbrj Ijiis devttied 

 Jnrnaelf (o a tlinrough analysis of that Iftn^na-ie. Until wniplnte nnnlytie gSossarieft 

 are prepand, tin? coniparisons f-f etbnol-'gi-sts must coniiime to if. in grcfti metisnro 

 empirical, and must lie rcn-ivcil nilh a cuntki^VriibJi; a loWii:icc for error.-i. Tlie 

 foJInwJuir eoll^pari^oIl.'^ icriiiirc a Iiirgr «llo»anct' not «»i)Iy frojj) tlit? nccisaily of llia 

 cmi hut f om the special (iifiulvnntag* s unrler wliicb tbc colliitor lahrmi's, "l bey nre 

 limited to tin; clfl^8es of wuiils int^ntimuid in the I'rcfalory Wote t" Part If. A full 

 Klimic cotnpurlaon of tlic Diaviriuii vucubuluricjJ witli lbi>&e of other fan li lies wouht 

 be a labour not for a {single iile but for the huolitgi.stsi of twvcral uciuii-utions. 



[Ikfore fending (bis serti<m to the prtss I ricrivtd by the Jn?it mail sliamcr 

 Chevalier Bunsen's i'^hiloso|diy of Fnhi-riiiil Ht^tury, io ivhicli rrnfes'ior Max 

 Wulltr has cnntriliuted two cliuptcrs on tlni >cytbtt\ Dnnirlan, 'fibt^to-l.'ltra- 

 indfan, llnd, and Malay languagia. ^ouil- ol tbe giuiuarrftl iletaib in ibk becihin 

 aisd b) tbe ne.^t cbupter bare I ilnd Ijcco anlicfpaifd hy H'rof. Miider. Whme be 

 low aupjilied daia wbii-b were nut acefs*i!>(e to tn«, I luive added a few notes whjfb 

 are diatinguiaiied by tnackeld. TJie supi leineot contoinlng tbe compaj alive vuca- 

 bulariea ba^iig Im'n printed sonti; tin e since, ! iuive not tieen able to subjoin any 

 noiea to It. 1 do not in this place oiler any rtiaiarka on the coincidenceit brtween 

 Prof, Miibfr'sTiew:* on several pointd, and tlio-^e previously publishtid by me in tbe 

 present ai'f 163 of papers. Tlmy will lieantticiently obviuita to iithnologtaitt wbft liave 

 rend my 4tb and preceding cbapltra, with the general reiuarki oa Aftonwlaa 

 ethnology contained in the vuluuie of tUls Journal lur iJ^t)]. 



