MCGEE] COMPARATIVE LEXICOLOGY 315* 



The Sedan numeral ''eight" is expressed by t\vo diftereut terms. The first is 

 based ou the numeral three, and the second on the digit four. The former is the 

 remaining factor of au original expression which signified by uttered elements 

 "three added to five (== the full baud) ", but the need for economy of expression led 

 to the suppression of the uttered element denoting "five", as soon as the shorter 

 "three added" acquired the usual signiflca.tion of "eight". The basis of the digit 

 is ku'pka or kapx'a, "three", with the suffix -kwitn {-kxiie, -que), jiresumably denoting 

 "added, plus". This reiiresents the usual method of forming this digit. The second 

 term, ksliox'ifku, is that which is presumably based ou the numeral "four". This 

 is the form given by M Pinart. But Sr Pimentel, citing Sr Teuochio, writes 

 this osrojonkum, which at first sight appears to be quite different from the other; 

 yet the r of the latter evidently stands for a modified x ^^<^ the j for a Xt -iDfl 

 making these substitutions the term becomes osx'oxoskum, which is approximately 

 the form in which Professor McGee and Mr Bartlett wrote this digit in the numeral 

 "eighty". Now, it is self-evident that if the element "four" constitute a factor in 

 the conibinatiou denoting " eight", it must be added to itself by addition or multi- 

 plication, and the result will be the same in either event. The final -olkn ajipears 

 also as -oiki'im, -olchkom, and -oskum in these Serian vocabularies, either in the 

 numeral "four" or its multiples. Tlie origin and signification of this ending are 

 not clear; but taking into consideration the great variations in tlie spelling of its 

 recorded forms, especially in so far as the consonant sound preceding the A-sound is 

 concerned, it may not be presumptive to adopt the s-sound (though s,y' may be 

 more correct) as that wliich represents approximately .at least the true sound, for it 

 varies from I, t, Icli, to s. And it has been seen that the final -urn is a flexion deuotive 

 of serial or consecutive counting and so not a part of the stem. Then it is seen that 

 -sk- (the last two hyphens representing uncertain vowels) is the termination requir- 

 ing explanation. Now, it is probable that this termination is identical iu meaning 

 and origin with the -siik, -shox, -alio, -srlioch, and -shroj (= -shx'ox) terminating the 

 forms of the digit "six". If this identification be correct (and there is no present 

 reason to doubt it), it signifies "repeated, again, duplicated", as was suspected 

 and stated in the discu.ssion of the forms of the numeral "six". So granting this 

 derivation to be correct, kshoxolka, then, signifies " four repeated ", which of course 

 denotes "eight". 



In the Yuman list, the first eleven forms are evidently composed of the numeral 

 "three" and a suffix signifying "added, plus, more than", but the last three of the 

 group want this suffix, a fact due perhaps to the fault of the collector rather than 

 to linguistic development. The terminations -eleepai and -shbe-k and its variants 

 have already been explained when treating of the numeral "' seven ". And the twelve 

 forms beginning with chip-Iioke (16) are variants from a common original composed 

 of the numerals " two" and "fonr". It will be readily seen that chip- in such aform 

 as chip-hoke is a contraction of a form such as tchibabk (14), "four", cliepap (24), 

 "four", as may be seen in the Yuman list of terms for the digit "four". Now, the 

 next portion of the term is -hoke, which is but a slightly disguised numeral "two", 

 as maybe seeu by reference to the schedules of the numeral "two". Compare hooak 

 (23), hudka (19), uake (2), and/iriiYi'/ii (18), all signifying "two". Now, the uext term, 

 maike-homvk-enaich (8), is a combination of maike, "above, over, more than", liomok, 

 "three", and the ending -enaich (or -kanaich), which may be either au ordinal or a 

 distributive llexion. The form nio-khamuk (14) is a combination of the prefix nio-, 

 signifying "added, above, or more th.an", and the conceptual term khannik, "three", 

 the expression signifying "three over, or added to". The next two examples are 

 evidently irregular, if not spurious. The form pakaikhin-atvach is composed of jyakai, 

 "seven", khin-, "one", and the suffix -uwacli, "added to". Now, the last, the 

 Cochimi nt/aki-vamivapai , appears to be erroneous. It contains the term mjaki for 

 ginyaki, "hand", but the remainder of the expression is composed of elements that 

 are not comparable to anything in the meager material at present accessible. The 

 Serian and the Yuman terms herein show no relationship. 



