na. 
 
 V^. ^a; 
 
 ♦w^ 
 
 7 
 
 IMAGE EVALUATION 
 TEST TARGET (MT-3) 
 
 Photographic 
 
 Sciences 
 
 Corporation 
 
 // 
 
 ^ .1^4e. 
 
CIHM/ICMH 
 
 Microfiche 
 
 Series. 
 
 CIHIVI/ICMH 
 Collection de 
 microfiches. 
 
 Mr.^ 
 
 Canadian Institute for HistoiUal Microreproductions / Institut Canadian de microreproductions historiques 
 
 fc—lllil fllIMM— I 
 
Technical and Bibliographic Notes/Notes techniques et bibliographiques 
 
 The Institute has attempted to obtain the best 
 original copy available for filming. Features of this 
 copy which may be bibliographically unique, 
 which may alter any of the images in the 
 reproduction, or which may significantly change 
 the usual method of filming, are checked below. 
 
 L'Institut a microfilm^ le meilleur exemplaire 
 qu'il lui a 6t6 possible de se procurer. Les ddtails 
 de cet exemplaire qui sont peut-dtre uniques du 
 point de vue bibliographique, qui peuvent modifier 
 une image reproduite, ou qui peuvent exiger une 
 modification dans la mdthode normale de filmaga 
 sont indiqu6s ci-dessous. 
 
 D 
 
 Coloured covers/ 
 Couverture de couleur 
 
 I I Covers damaged/ 
 
 
 D 
 
 D 
 D 
 
 D 
 
 Couverture endommag^e 
 
 Covers restored and/or laminated/ 
 Couverture restaurde et/ou pellicul6e 
 
 Cover title missing/ 
 
 Le titre de couverture manque 
 
 I I Coloured maps/ 
 
 Cartes g6ographiques en couleur 
 
 Coloured ink (i.e. other than blue or black)/ 
 Encre de couleur (i.e. autre que bleue ou noire) 
 
 I I Coloured plates and/or illustrations/ 
 
 Planches et/ou illustrations en couleur 
 
 Bound with other material/ 
 Reli6 avec d'autres documents 
 
 Tight binding may cause shadows or distortion 
 along interior margin/ 
 
 La reliure serr^e peut causer de I'ombre ou de la 
 distortion le long de la marge int6rieure 
 
 Blank leaves added during restoration may 
 appear within the text. Whenever possible, these 
 have been omitted from filming/ 
 II se peut que certaines pages blanches ajout6es 
 lors d'une restauration apparaissent dans le texte, 
 mais, lorsque cela dtait possible, ces pages n'ont 
 pas 6t6 filmdes. 
 
 D 
 D 
 D 
 D 
 D 
 D 
 D 
 D 
 
 Colouied pages/ 
 Pages de couleur 
 
 Pages damaged/ 
 Pages endommagdes 
 
 Pages restored and/or laminated/ 
 Pages restaurdes et/ou pellicul6es 
 
 Pages discoloured, stained or foxed/ 
 Pages ddcolordes, tachet6es ou piqudes 
 
 Pages detached/ 
 Pages d6tach6es 
 
 Showthrough/ 
 Transparence 
 
 Quality of print varies/ 
 Quality indgale de I'impression 
 
 includes supplementary material/ 
 Comprend du tuatdriel suppldmentaire 
 
 Only edition available/ 
 Seule Edition disponible 
 
 Pages wholly or partially obscured by errata 
 slips, tissues, etc., have been refilmed to 
 ensure the best possible image/ 
 Les pages totalement ou partieilement 
 obscurcies par un feuillet d'errata, une pelure, 
 etc., ont 6t6 filmdes d nouveau de faqon d 
 obtenir la meilleure image possible. 
 
 D 
 
 Additional comments:/ 
 Commentaires suppldmentaires; 
 
 This item is filmed at the reduction ratio checked below/ 
 
 Ce document est filmd au taux de r6duction indiqu6 ci-dessous. 
 
 10X 14X ItX 22X 
 
 12X 
 
 16X 
 
 20X 
 
 S 
 
 26X 
 
 30X 
 
 28X 
 
 32X 
 
The copy filmed here has been reproduced thanks 
 to the generosity of: 
 
 Library of Congress 
 Photoduplication Service 
 
 L'exemplaire filmi fut reproduit grdce d la 
 ginArositA de: 
 
 Library of Congress 
 Photoduplication Service 
 
 The images appearing here are the best quality 
 possible considering the condition and legibility 
 of the original copy and in keeping with the 
 filming contract specifications. 
 
 Las images suivantes ont At6 reproduites avec le 
 plus grand soin, compte tenu de la condition et 
 de la netteti de l'exemplaire film6. et en 
 conformity avec les conditions du contrat de 
 fiimage. 
 
 Original copies in printed paper covers are filmed 
 beginning with the front cover and ending on 
 the last page with a printed or illustrated impres- 
 sion, or the back cover when appropriate. All 
 other original copies are filmed beginning on the 
 first page with a printed or illustrated impres- 
 sion, and ending on the last page with a printed 
 or illustrated impression. 
 
 Les exemplaires originaux dont la couverture en 
 papier est imprimAe sont film6s en commenpant 
 par ie premier plat et en torminant soit par la 
 derniire page qui comporte une empreinte 
 d'impression ou d'iliustration, soit par le second 
 plat, salon le cas. Tous les autres exemplaires 
 originaux sont filmfo en commenpant par la 
 premiere page qui comporte une empreinte 
 d'impression ou d'iliustration et en terminant par 
 la dernlAre page qui comporte une telle 
 empreinte. 
 
 The last recorded frame on each microfiche 
 shall contain the symbol ^^> (meaning "CON- 
 TINUED "), or the symbol V (meaning "END"), 
 whichever applies. 
 
 Un des symboles suivants apparaftra sur la 
 derniAre image de cheque microfiche, selon le 
 cas: ie symbole — ► signifie "A SUIVRE", le 
 symbole V signifie "FIN ". 
 
 IVIaps, plates, charts, etc., may be filmed at 
 different reduction ratios. Those too large to be 
 entirely included in one exposure are filmed 
 beginning in the upper left hand corner, left to 
 right and top to bottom, as many frames as 
 required. The following diagrams illustrate the 
 method: 
 
 Les cartes, plaiiches, tableaux, etc., peuvent dtre 
 filmte A des taux de rMuction diff^rents. 
 Lorsque le document est trop grand pour dtre 
 reproduit en un seul cliche, 11 est filmd A partir 
 da i'angle supArieur gauche, de gauche d droite, 
 et de haut en bas, en prenant le nombre 
 d'images nAcessaire. Les diagrammes suivants 
 iliustrent la m6thode. 
 
 1 
 
 If 
 
 2 
 
 3 
 
 32X 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
ON 
 
 NUMERALS 
 
 IN 
 
 AMERICAN INDIAN LANGUAGES 
 
 AND THE 
 
 INDIAN MODE OF COUNTING. 
 
 By J HAMMOND TRUMBULL, LL.D. 
 
 I From the Transactions 
 
 of the Am. Philological ABSOciation, 1874.1 
 
 HARTFORD, CONN. 
 
 1875. 
 
 If. iltttliimr V"-^ 
 
 
 - W - 
 
 w 
 
r 
 
 . ", 
 
 
 j-mntiMSiiiiUtu 
 
 -mtm 
 

 .T7 
 
 ON NUMERALS 
 
 IN AMERICAN INDIAN LANGUAGES, 
 AND THE INDIAN MODE OF COUNTING. 
 
 That "all miinerals are derived from the fingers'" is 
 as generally true for languages of the new world as for 
 those of the old. The North American Indians have, 
 with comparatively few exceptions, adopted decimal systems, 
 reckoning the fingers of Iwth hands. Some South American 
 trihes have not advanced heyond a qimmri/ ; and a few are 
 said to be poorer even than this. The Brazilian Tupis had, 
 at one time, no names for numbers higher than 3,'^ and the 
 Abipones of Paraguay, as Dobrizhoffer states,^ could not count 
 beyond 4, giving to that number the name of ' the ostrich's 
 toes,' geyenknute. Some nations, particularly those of Mexico 
 and /Central America, and the Eskimos, have reckoned by 
 twenties instead of tens of^ves, counting toes with fingers 
 for the base of their numeral system. The Ti.le Indians of 
 Dariep reckon in this way : 2U ia ' a man,' i. e. all his fingers 
 
 1 " Alle Zahlwortcr gehn aus von den Fingern der Hande." — Grimm's Gescli.-. 
 der deutschen Sprache, i. 167. 
 
 '•'The fact that the Tupis lost their names for 4 and 5, after the coming of 
 Europeans, is worth noting. J. de Lery, who was in Brazil in 1557, writes that 
 
 the " Tououpenambaults usque ad numcrum quinque verbis notare, hoc 
 
 mode: augepe I, mocouein 2, mossapiU 3, oioicoudic 4, ecoinbo 5." — Hist Navig. in 
 Brasiliam, 1586, p. 272. (In the 5, we recognize po 'hand.') Jos. de Anchieta, 
 in his Tupi Grammar, 1595, says: "Os numeraes nao chegao mais quo ate 
 numero de quatro: nt oiepS 1, mocdi" 2, mogapir 3, oyoirundic 4." Eckart, a Jesuit 
 missionary in Brazil, 1753-57, gives the same names for 1, 2, and 3, adding: 
 " Non plus ultra Brasili hodie numerant," though he had seen names for 4 and 5 
 (monherondyq, ambd) in ' an ancient grammar by Father Anchieta' ; " sed uterqae 
 hie numeruB mode jam exolevit." — Specimen Ling. BrasilicK, 1778. 
 
 * DobrizhofTer's account of the Abipones, ii. 1 68. 
 1 
 
 'im 
 
 )eh 
 

 § 
 
 J. If. Trumhttll, 
 
 and toes, 100 is T) men,' and so on.^ Gallatin has given a 
 good account of these vigesimal systems in his " Notes on the 
 semi-civilized Nations of Mexico," etc.,"' tlie substance of 
 which was incorporated l)y Pott in his Ziihlmethode (Halle, 
 1847). Mr. CJallalin had i)reviously observed, in a note to 
 his Comparative Vocabulary of fifty-three North American 
 nations, " that all these had resorted to a decimal numeration." 
 More recently, Busclinmnn has shown' that the system of the 
 Athapascan family is clearly decimal, exhibiting traces of the 
 vigesimal in two languages only — the Umpqua of Oregon 
 and the Kinai ; while of the languages of his Sonora group 
 (including the Comanche, Paiute, Pima, and Shoshoni), seven 
 have the decimal and five the vigesimal system, one (the 
 Taralnnnara) possessing both.' In some dialects, indications 
 of a former vigesimal system, abandoned for or in progress of 
 change to a decimal, may be observed. 
 
 The derivation of numerals from the fingers admitted, an 
 answer to the question, Iti what ordrr are the fingers counted? 
 l)ecomes a necessary preliminary to the investigation of any 
 table of numerals. Which finger marks ' one ' ? Is it the 
 little finger — or, as in the designation of numbers by educated 
 deaf-mutes, the thumb ? And, in passing from 5 to 6, i. e. 
 from one hand to the other, is the sequence from finger to 
 finger — thumb to thumb, like the Zulu' — or thumb to little 
 finger, like the Veis ? 
 
 Nearly all the information given by Gallatin and Pott on 
 these points relates to the Eskimo numerals. In the language 
 of " the Eskimos of Hudson's Bay, the names of the numerals 
 8, 9, 10. mean respectively, the middle, the fourth, and the 
 
 * See Lull's Darien Vocabulary, in the Am. Philol. Associiition's Transactions 
 
 for 1873, p. 103. 
 
 6 Transactions of the American Ethnological Society, vol. i. (1845). 
 
 B Worttafel des Athapask. Sprachstamms, §§114, 115, 157. 
 
 ' Grammatik der Sonorischen Sprachen, Abth. 3, p. 141. 
 
 8 " The Zulu, counting on his fingers, begins in general with the little finger of 
 his left hand. When he comes to 5, this he may call edesaiita ' finish hand ' ; then 
 he goes on to the thumb of the right hand, and so the word tnlisilupa ' taking the 
 thumb ' becomes a numeral for 6."— Tylor's Primitive Culture, i. 228. " The 
 Vei people and many other African tribes first count the fingers of their left hand, 
 beginning, be it remembered, from the little one, then in the same manner those 
 of the right hand." — Id. 227. 
 
 H 
 
 <• 
 
 ^ * 
 
 tmtimtm 
 
On Numerals in American Indian Languages. 
 
 3 
 
 %♦ 
 
 ii 
 
 J^" * 
 
 little finger.'" Pott, transferring this from Gallatin, infers 
 (Zahlmethode, JJOl) that the thunih of the second hand 
 designates 6, i. e. 1 + 5 of the first hand. The account given 
 hy Cranz,'" of the Eskimo mode of counting, is quoted by 
 Pott as the starting poin*: of his work : " Their numerals fall 
 very short. However, they can with difficulty make a shift 
 to mouiit as high as 20, by counting the fingers of botli hands 
 and the toes of both feet. But their proper numeration is 
 five : attau><elc, 1 — arloik, 2 — j^ingojuak, 3 — sismmat, 4 — 
 tdlimat, 5. If they nnist go further, they begin witli the 
 other hand, counting upon their fingers. The sixth [i.e. the 
 thumb] they call arhennek, but the rest, till 10, have no other 
 names but, again, ' two,' ' three,' ' four,' ' five.' Tlicy call 
 ' eleven ' arkaugat, and ' sixteen ' arbarsanget, and these 
 -teens they count upon their toes. Thus they muster up 20. 
 Sometimes they say instead of it, ' a man,' that is, as many 
 fingers and toes as a man has ;" etc. 
 
 That the fingers of the two hands were counted by other 
 North American nations in the same order as by the Eskimos, 
 several writers inform us : 
 
 " The Dakotas, in counting, use their fingers, bending them 
 down as they pass on, until they reach ten. Then they turn 
 down a little finger, to remind them that one ten is laid away, 
 and commence again. When the second ten is counted, 
 another finger goes down, and so on." ' " The Aubsdroke or 
 Crows [who are of the Dakota stock] like all the Indians 
 with whom I am acquainted," says Dr. P. V. Hayden, " use 
 their fingers in counting, bending them down temporarily 
 against the inside of the hand •■■.s they proceed," etc.* 
 
 Mr. Say, describing thel....-?n sign-language, says: "To 
 indicate the digits, they clenci the hands and extend the 
 little finger of the left hand for one, the ring finger for two," 
 
 'Gallatin's "Notes on the Semi-Civilized Nations of Mexico," etc. («< supra), 
 p. 49. 
 
 1" History of Greenland (English translation, i. 225). The Greenland numeral 
 system is more clearly and accurately exhibited by O. Fabricius, Gronlandsk 
 Grammatik, 58-63. 
 
 1 Kiggs, Dakota Grammar, p. 36. 
 
 * Contributions to the Ethnography and Philology of the Indian Tribes of the 
 Missouri Valley, u. 396. 
 
 4 
 
 \ ! M 
 
,,;fe. 
 
 P- 
 
 4 J. H. Trvnihull, 
 
 and so on to "the tlmuil) for five, . . . the thumb of the right 
 hand for six," etc. " When enumeratin^r a small number, 
 where a considerable exertion of the memory is requisite, the 
 Indians extend the left hand with the palm upward, whilst, 
 with the index of the ri^lit, the fingers are successively bent 
 in to the palm, l)eginning as before with the little Jin(/cr, and 
 the greater difficulty in recalling to mind the numbers or 
 events, tiie more ap[)arent rosislauce is offered to the intlexion 
 of the finger.''^ Prince Maximilian von Wiud ' gives a similar 
 description, observing that " wenn man an den Fingern 
 alizilblt, so fiingt man an der lin/coi ILtnd an." Mr. Swan, 
 in his account of the ]\Iakalis of Caj)e Flattery (Straits of 
 Fuca), says of their mode of counting: "They commence 
 with the little finger of the left hand, closing each finger as 
 it is counted ; then pass from the left thumb, which counts 
 five, to the right thuml), which counts six, and so on to the 
 little finger of tiic right hand, which counts ten."' 
 
 The Muskokis (Creeks), Mrs. A. E. W. Robertson writes, 
 '• turn the back of the hand towards the face and, beginning 
 with the left hand little finger, end with the right hand little 
 linger." In continued intercourse with the whites, the 
 Creeks, like some other tribes, have learned to indicate 
 numbers by holding up instead of bendiwj dotvn the fingers : 
 but, as Mr. Say observes, '' when any considerable exertion of 
 the memory is requisite " Indians naturally recur to the 
 earlier mode. Major J. W. Powell informs me, that the 
 Yutes commonly answer the question "how many?" by 
 raising the fingers, but he has seen Indians of that nation, 
 when alone, reckon numbers by turning down the fingers 
 successively, from left to right, in the manner described by 
 Say. 
 
 Whether an Indian marks ' one ' by a thumb or a finger 
 does not seem at first sight a question of much interest to 
 students of language. It is, however, one of the thousand 
 questions which every philologist must be prepared to answer 
 
 ^Long's Expedition to the Rocky Mountains (Philadelphia, 1823), i. 388. 
 
 * Ueise in das Innere von Nord-America, Bd. ii. 6SU. 
 
 ''Indians of Cape Flattery (Smithsonian Contributions, vol. xvi.), p. 100, ncte. 
 
 ; -i^ltf<^t.lMll<fTni-nl 
 
-^f^'mep-v^'ti iT-TTir; ' ;r''j'^'' 
 
 *■!* 
 
 On Nuxicrah in American Indian Languagt's. 
 
 5 
 
 /f 
 
 i 
 
 before ho is fully competent to discuss the sulijcct of Mr. 
 Robert ElHs's lately published volume " On Numerals as 
 Sij^iis of Primeval Unity among Mankind" (London, 187.'5). 
 Mr. Ellis thinks that he has detected " a great number of 
 coincidences, aU'ecfing' not only numerals, but also the names 
 of the memliers of the body from which those numerals are 
 derived, in languages far removed from each other in position," 
 and he presents these coincidences as " tho result of primeval 
 affinity — indications of unity of origin in human speech and, 
 probably, in the human race " (p. 4). He assumes that 
 " the names of numerals commonly carry in themselves the 
 proofs of their own great antiquity" (f). 2). For the Indo- 
 European and Semitic languages this assumption is perhaps 
 well grounded ; for the American, it is untrustworthy and 
 iinsustained by evidence, exce[)t — for reasons to 1)C mentioned 
 presently — as regards names for the first three numerals 
 in languages of the same linguistic group. Admitting tho 
 original unity of American speech, it is yet certain that its 
 division into widely sejjarated families must have preceded 
 the origin not of numerals oidy, but of the verbal or nominal 
 roots from which names of numerals in the several families 
 were derived. Even in the same linguistic group these 
 names, as compared with other portions of the vocabulary, 
 carry no indications of high antiquity, but rather the contrary ; 
 and in dialects of the same language names for the same 
 number are often radically unlike. Compare, for example, 
 the Algonkin ' fives ' : Massachusetts napanna taJishe, Micmac 
 ndn, Chippeway ndnan, Abnaki barenesku, Delaware palenach, 
 Illinois miaranui, Blackfoot nisito. Such dissimilarity is more 
 apparent and more general in numerals above 'five,' which are 
 with few exceptions composite. The Arikaras or ' Riccarees ' 
 of the upper Missouri speak nearly the same language as th.d 
 Pawnees and, probably at no very remote period, belonged 
 to the same nation. Their numerals correspond with the 
 Pawnee numerals, to ' five,' inclusive ; but here the likeness 
 ends, not merely the names but the primary conceptions of the 
 higher numbex'S differing in the two dialects. One Yuma 
 dialect of the Colorado, the Mojave, repeats 1, 2, 3, in the 
 2 
 
 L 
 
SUmli. 
 
 ii nyw ii f^— 
 
 ■MMM 
 
 !•■ 
 
 6 
 
 J. IL T,-umf>,ill, 
 
 :\ r 
 
 iianicB for fl, 7, and H, niul marks as ' next to ton ' ; another, 
 the Cuchan, near akin, regards 6 and 1), rcspoctively, as a 
 pair and a triplet of' threes,' and H as a douI)led 4. All these 
 in some sense " geliii ans von den Fingern," Imt the same 
 finger of the same hand or the liand itself may be — and in 
 fact very often is — ditilL'rently nunied, or the nnniber it marks 
 is difTorently expressed, by tribes speaking dialects of the same 
 langnage ; nor may we expect always to find names either of 
 ' hand ' or ' finger ' in the numeral. 
 
 In the investigation of the origin of American numerals 
 and in inferences as to their antiquity, two facts must be 
 borne in mind : 
 
 1. The primitive mode of indicating numbers by the/H^«T« 
 is still in use. The tiaine is not completely independent of 
 the sign, and, consequently, the constancy of the name in 
 passing from one dialect to another is less assured. When 
 an Indian marks ' five ' by showing or bending down all the 
 fingers of his left hand, the vocal utterance — whether ndnan 
 or bareufsku — is of secondary importance. In the Indo- 
 European languages the vocal was long ago substituted for 
 the digital expression. " It was no easy task for the 
 linguistic faculty to arrive at a suitable sign," as the 
 exclusive designation of a number, " and when the sign 
 was once found, it maintained itself thenceforth iu use 
 every where, without danger of replacement by any other, 
 of later coinage."' But this is necessarily true only of 
 languages iu which the earlier sign — by show of fingers — 
 is obsolete. 
 
 2. The origin of names for ' one,' ' two,' and probably 
 * three,' in all languages, preceded formal numeration. 
 Pairs, couples, doubles, were known before ' two ' was 
 counted on or marked by the fingers. The conception of 
 duality dates from the first conscious separation of the 'not-I' 
 from the 'I': and, with the first perception of ditferences 
 in the 'not-I' — as 'this' and 'that,' 'here' and 'yonder,' 
 ' thou ' and ' he,' ' before ' and ' after,' came the notior nd 
 J iduie of ' three,' as something ' beyond,' ' besides,' or 'above' 
 
 "Whitney, Language und the Study of Language, 195. 
 
N 
 
 ! 
 
 ■ 
 
 On Numeraln In American Indian Lnnguagea. 7 
 
 (tar, tri, trniis, tros, tries') the primary distinction; and 
 tlioroupon, the exclusive and inclusive dual, ' tliou-ho ' (and 
 not ' I'), ' I-tliou ' (and not ' lie ') ; after this, the conception 
 of phmdit//, and numeration. Some nations, as we have 
 seen, never advanced beyond the ' three.' Others (to bo 
 mentioned hereafter) only found their way to ' ten ' by help 
 of 'pairs' and 'triplets.' Hence, as Mr. Gallatin observcct 
 of American languages generally, " there is much confusion 
 and but little regularity in the formation of the names 
 expressing the higher numbers," even in nearly related 
 dialects. 
 
 Mr. Ellis's first group of coincidences, and the one he 
 regards as most important of all, includes North American 
 words "of which different names for 'finger' supply the 
 elements." These words, ho thinks, " sufficiently illustrate 
 the manner in which names for 'finger' and 'hand' are 
 employed to form numerals ; and by showing, moreover, that 
 hand miiy = fingers = finger-finger {\\\\\d\ last would be the 
 rude plural of j^w^/fr), they explain how 'hand' and 'two' 
 may be the same word, as in the Omaha nomba which has 
 both these meanings" (p. 6). He goes on to detect in the 
 Basque language terms for ' finger,' ' one, i. o. finger,' and 
 'five = hand = fingers=finger-finger,' that correspond nearly 
 with terms derived from North American languages, and 
 finds coincidences with one or another of these in European and 
 Asiatic names for ' thumb," finger,' 'palm,' 'five," six," arm,' 
 'ten,' etc. (pp. 13, 14). He suggests the probability that "the 
 Aryan languages virtually cutain the forms svas and saz for 
 ' five,' as the Basque contains zaz and as the North American 
 languages contain forms like azbaz, such as Natchez i»peiihe 
 ' hand.' " And he argues (.p. 18) that " if the resemblances 
 between all these s fives, as they may be called by way of 
 definition, were sufficient to imply affinity wherever they 
 were detected, such affinity could be no other than a primeval 
 one," — an inference the justice of which no one is likely to 
 question. Even those much-vexed Etruscan dice of Toscanella 
 are made to testify to primeval unity ; for why may not much 
 [conjectured by Mr. Isaac Taylor to stand for] 'one,' be 
 
 if 
 
-« ^T.'»3Vi^^"t'" 
 
 wmmmm 
 
 8 
 
 J. H. Trumlmll, 
 
 I 
 
 connected Avitli "California (Sekumiic) ma ' liand '" and 
 " Coniaiiclie »j(wa ' linnu,' ' aim,' " as well as with Siamese 
 mec ' hand,' Armenian mi and Greek ^kt ' one,' and African 
 (Melon) moe ' finger ' ? 
 
 Rigidly examined, these and a host of other coincidences 
 which Mr. Ellis with much ingenuity presents, would prove 
 to be less remarkable than they seem to him. It is not my 
 purpose, however, to discuss them in detail, or to seek for 
 them, collectively, any other explanation than the one which 
 I am assured in advance "is not satisfactory" — namely, 
 that so far as they are not imaginary, they " are merely 
 accidental." I propose instead to make some observations 
 on the composition and primary meaning of Indian names 
 for numbers, and first, to t oint out such relation as I can 
 find between some of these and names for the hand and 
 the fingers. The examples will be taken chiefly — but not 
 exclusively — from two great families of North American 
 speech, the Algonkin and the Dakota, because, in these, 
 published grammars and dictionaries facilitate etymological 
 research and afford means of noting differences, phonetic 
 and radical, between names in one and other dialects of the 
 same stock. 
 
 I. In some languages we find only one name for ' hand ' 
 and 'fingers' collectively; and generally, for designating 
 the fingers individually, names are formed from the word for 
 'hand,' with a descriptive prefix, e.g. the third finger is 
 ' middle of the iiand.' 
 
 Pott (Ziihlmethode, 234 ff.) has given illustrations from 
 American languages of the recognition of a likeness bp^^woen 
 me7i and trees, and of figures of speech drawn from iv. liie 
 arms are ' limbs' or ' branches' qf the human 'trunk' ; the 
 hands and fingers are 'branches' of the arms; the fingers 
 ' sprouts ' or ' leaves ' ; the thumb a ' spur ' or ' oii-[;]ioot.' 
 Sometimes the fingers, collectively, are a ' roAV of branches,' 
 or a ' fence.' Compare 
 
 Dakota . nape ' hand ' ; napsvkdza ('small piece of hand ') ' finger.' 
 
 Iowa nuwe ' hand ' ; nawepa {' hand jioint ') ' finger.' 
 
 Chippeway -nindj 'hiuul' ; hhilt 'in a row'; -ikwan 'branch'; binakxanindj 
 ' linger,' ' (one of) a row of branches of the hand.' 
 
-3^ 
 
 '^m^' 
 
 %/ 
 
 On Numerals in American Indian Languages. 9 
 
 Massachusetts -nulch ' hand ; ' pnclii ' divided ' ; pochatiik ' a brr.nili, or division ' ; 
 
 pochanutch 'a finger.''' 
 Cree (Western) -tchihhiy 'hand'; yiyiki 'forlced,' 'branching'; yujikUchitch.ln 
 'finger.'' 
 
 In some of the Algonkin languages, the name for ' hand ' 
 seems to be formed from a verbal root meaning ' to seize,' 
 ' to lay hold of : anun ' he lays hold of, catches,' anutch 'the 
 layer hold of, the seizer'; -nutch (with pronominal prefix) 
 'hand.' In the western Cree, -tchiuMy (in composition, 
 otcW) 'hand' is from the same root as the Mass. -tchan 
 'nose' (Chip, odjanj), whicli is found again in uio final 
 tchdn of Cree 'finger,' meaning 'projecting,' 'point,' 'vertex.' 
 The names for 'nose,' ' nead,' 'fore-arm,' 'hand,' in the 
 Dakota are apparently related one to another, their common-' 
 root denoting ' pointed,' ' a projection, vertex, or extremity.' 
 Compare with Dakota pe ' pointed, sharp,' pe ' top of the 
 head,' pa ' head,' paha ' hill,' pa-sH ' beak or bill,' ' snout 
 of an animal,' aj^d ' a part,' ap^ ' a leaf,' ' a fin,' etdpa ' the 
 right hand,' ishpd ' the fore-arm ' ; and Iowa ndwe ' hand,' 
 ndwe ' leaf,' naw^pa ' finger,' pa ' nose,' pa-thukh ' beak.' 
 
 'Compare Hawaiian lima 'arm' and 'hand'; manamdna 'branching,' 'a 
 brancn' (rednpl. of mann 'to be divided,' 'to branch'); mammami lima 'fingers.' 
 
 "My principal authorities for Aloonkin languages are: Masmchmetts, Kiiot's 
 Indian Grammar and version of the Bible; Chippewai/, Baraga's Otcliipw6 
 Dictionary and Grammar; Cree, Lacombe's Grammairc ct Dictionnaire do la 
 Langno des Oris, and (Hudson's Bay dialect) Howse's Cree Grammar; 
 Delaware, Zeisberger's Grammar, and Vocabulary ; Almaki, Haslcs's Dictionary, 
 by Pickering; Micmac, Maillard's Grammar; Dr. Hayden's Vocabularies of 
 the Dlackfoot, Shi/eme, Arapoho, and Atsina. For the Dakota, my chief 
 reliance is, necessarily, the invaluable Dictionary compiled by the Rev. 8. R. 
 Riggs and his associates in the Dakota mission of the American Board; 
 and for other dialects, Dr. W. Matthews's Uidatsa (Minitari) Dictionary, Dr! 
 Hayden's Assiniboin, Auhsuroke (Crow), Mandan, Omaha, Iowa, and Winnebago 
 Vocabularies, the Rev. Wm. Hamilton's loua Grammar ; for the Ponka numerals, 
 a primer, "Ponka ABC Wa-ba-ru" (prepared by the Rev. J. Owen Dorsey, of 
 the Episcopal mission); and for the Osaije, Prince Maximilian von Wied-Neuwied's 
 Vocabulary, compared with Gallatin's (in his Comparative Vocabulary). 
 
 The vowels arc to be sounded as in German, except u which is the short 
 English u in but, or the neutral vowel, variously represented in vocabularies as 
 a, fi, u, and V. For the n which marks a nasalized vowel, I have substituted a 
 ■superior' C), and for the gutturals — variously represented by ch, H, h, Xy 
 etc. — I have used cii or kk. The italic ch has the English sound (as in church), 
 atid th, sh, and zh (used interchangeably with j) are as iu English. 
 
 w 
 
 m 
 
 
 
iW 
 
 10 
 
 J. H. Tnimbull, 
 
 II. Counting -the fingers from left to right, the numerals 
 are distributed thus : Little finger, 1 and 10 ; Fourth finger, 
 2 and 9 ; Middle finger, 3, 8 ; Fore finger, 4, 7 ; Thumb, 5, 6. 
 
 1, 10. The fifth or little finger is variously designated in 
 American languages, as ' tiie last of tha h id,' ' the least,' 
 ' the youngest son,' ' the little daughter of the hand,' etc. 
 From one or another of these names, that of the numeral 
 ' one ' has, in many languages, been taken ; but in others 
 we find another expression for unity, ' one by itself,' which is 
 probably of earlier origin than finger-counting. A distinction 
 corresponding to that which is marked by the Indo-European 
 cardinal and ordinal, between one airiffle and one coming 
 before others, ' fore-est,' first of a numeral series, seems to be 
 universal in language. 
 
 In the Algonkin, these two names are represented by 
 
 Massachusetts pdsuk and n'liitt, nequt. 
 
 Cliippowiiy pai/zhik (f'ejig) ningoto. 
 
 Crce peiak, paiak, niktit 'aome one,' nikuton 'formcr\y.' 
 
 A note in Cotton's vocabulary of the Massachusetts 
 language distinguishes these names thus : " Nequt, a thing 
 that is pasL Pasuk, a thing in being." This note has 
 puzzled more than one writer on the Algonkin languages." 
 Cotton himself had only half caught the tiuo distinction 
 between pdsuk ' one onlt/,'' literally, ' a small thing,' and 
 n'qut 'first' or ' fore-est,' ' beginning.' The latter was used 
 when speaking of a one which had been (or necessarily must 
 be) followed by another, and in this way came its appropriation 
 to " a thing that is past," i. e. a former, thing. Hence, Mass. 
 nukkone ' old,' i. e. passed by, and the ordinal ne-gonne 'first,' 
 and ne-kutche ' the beginning,' ' it begins.' The prefixed n' 
 in easteiMi Algonkin numerals is merely demonstrative. 
 
 Pdsuk is a contraction of piasuk Q^easik, Eliot) ' very 
 small,' the diminutive of piak ' small, little."" Comp. Chip. 
 pangi ' a little,' pangishe ' very little.' The root, pi, is seen 
 
 " See Mr. Pickering's note, in his reprint of Eliot's Indiiin Grammar (2 Mass. 
 Hist. Sec. Coll., ix.) p. xlv. ; Dnponceaii's Memoire, 389, 390. 
 
 1" Abn. ii, plu. 6(-((i-, Mass, piak, a 'grain,' 'bit,' or ' bead ' of shell money; 
 whciice the name adopted by the English for unstrung ' peag' = Abn. ioa^ban-biak 
 'white beads,' Eng. 'wamponipeag.' 
 
 I 
 
 s 
 
( 
 
 "^■■■«sBiWMHR<j*«^^ 
 
 On Numerah in American Indian Languages. 11 
 
 in Cree peiak ' one ' and apis ' small,' diaiin. apids ' very 
 small.' The little finger being counted as 'one,' j)dsuk, 
 bezhik, 'the very small' has, as a result of association, been 
 substituted in several dialects for ti^qid, nikoto, as the name 
 for ' one,' but the latter reappears in tiie composition of the 
 higher numerals : e. g, Cree peiak 1, nikot-ivaxsik 6, i. e. ' one 
 over ' ; Abuaki pezuku 1, nekiid-a's 6, nekuda"nkdo 11. 
 
 The following are some of the names of the little finger, in 
 , North American languages : — 
 
 Alo. Cree iskwetchiiclumis 'lust little finger.' 
 
 Cliippewny isUkwe' nindj 'last of the hand.' 
 Ahnaki askwaiimireisi ' youngest (last born) of the hand.' 
 Massachu. mutldnonitch 'youngest son {multdsons) of the hand.' 
 
 Dak. (Sioux) shashte ? Comp. clii' stin ' little ' ; chatan' name of a fourth son. 
 Minnitari (Hidatsa) shdki-kazki diminutive of shdki 'hand.' 
 Mandnn tingkni inyka ' little finger.' Comp. Iowa i-yangke 'one.' 
 
 MuSKOKi (Creek) eirke-eclihuswuvhe 'hand's little daughter.'^ 
 Choctaw ibbak-ushi-uldi ' hand's little son.' 
 
 Pawnee sAe/s-;>i< 'finger little.' 
 
 Gallatin's vocabulary (from Parry) of the Hudson's Bay 
 Eskimo gives eerkitkoka (Greenl. eMkkuk) ' little finger ' as 
 the name for 'ten.' The Algonkin 'tens 'are related to — 
 bu,t not derived from — names of this finger. The^e will be 
 noticed hereafter. 
 
 2, 9. The Fourth finger — second by Indian reckoning — 
 is in some American languages, as it has been in many 
 languages of the eastern world, ' the nameless ' (Sansk. 
 andman, andmikd ; lAiXman. hewardia ; Tibet, mingmed). In 
 others, it is designated only by its position ' next the little ' 
 or 'next the middle' finger. In mission-Indian it lias 
 received the name of ' ring finger.' Lacombe gives Western 
 
 'For translations of this and other Muskoki (or Creek) finger names, I am 
 indebted to Mrs. A. E. W. Robeitson of the Tullahasse mission, and !o Buckner 
 and Hcrrod's Muskoki Grammar. For other languages of this group, I use ti.e 
 Rev. Cyrus Byingtuu's "English and ChocUw Definer" (1852) and his Choctaw 
 Grammar (posthumous) edited by Dr. D. G. Brinton (1870), and valuable 
 vocabularies (MSS.) of the Muskoki, llitcliitce, Coassutti, and Alabama, collected 
 by Gen. Albert Pike, in possession of the bmithsonian Institution, which 1 hope 
 will soon bo published, and with them, one of the Muskoki language, compiled 
 by the Rev. W. S. Robertson and Mrs. Robertson. Eor the Pawnee, uud related 
 Arikara, I rely on Dr. Hayden's vocabularies. 
 
* "'<J"'»W ' -^""'W * ' I I"" r 
 
 "* i -f l U 
 
 MiiiiM 
 
 ■p 
 
 ». 
 
 i. 
 I. 
 
 |- 
 
 12 
 
 -/. IL Trvmbull, 
 
 Creo atchdni-tchitchan (f vom atchaniH 'a riiip;'), and so Vou 
 Tschudi ill his WiJi-tcrbucli of the Kccliua of Porn has niui- 
 rucanu ' ring finger,' from siui ' ring.' In a few languages, 
 its name denotes ' becoming smaller ' — whether from its 
 shape, more 'tapering' than other fingers, or from its size, 
 as between the middle and little fingers, is not certain. 
 
 Dakota shdsle ii/okihe ' littlc-liiiper next-to.' 
 
 Miiiitari shaki-kazi-utidn 'that wliioli the little finger joins,' or, as Dr. Matthews 
 
 (Hidatsa Dictionary) translate*, 'hase of the little finger.' 
 Muskoki enkehochffkU stku 'hand's name-without,' 'the nameless.' 
 
 As a numeral I find the name of this finger only in the 
 ' nines,' and here only in the 
 
 Eskimo, Hudson's Bay mikkecliikkamoot ' nine' = 'fourth finger' (Parry). 
 
 Greenland niikkekrak, 'fourth finger,' liternlly 'it hccomcs smaller.' 
 Algonkin, Sliyenne na-so'toyos 'my fourth finger'; sohh'tu 'nine.' 
 
 3, 8. The ' Middle ' finger is so named in almost all 
 languages, but it not unfrcquently has the additional 
 designation of ' the groat ' or ' chief.' It gives in many 
 dialects a name — but not generally its otvn name — io the 
 numerals ' three ' and ' eight.' In the Algonkin languages, 
 of two expressions for ' in tlie middle ' or ' iialf-way between ' 
 (Mass. noeu and nashaiie, Chip, nawaii and nassaivaii), one 
 is given to the finger, the other to the numeral. 
 
 Ahnaki na"wi-relsi ' middle of hand ' ; nass ' three.' 
 
 Chippeway ndwi-nindj " " nisswi 3 ; nijwasii S. 
 
 Creo tdwi-tchitchdn 'mu\(i\c finpcr;' nisloo. 
 
 Mass. (nas/uiKc, 's/inii'e 'half-way'); nish,nishwi'\'shwi- 3; shwosukS. 
 
 Arapoho (n«i//it" 'in the middle'); nais3; naisa-loh' S. 
 
 Saul(i nissoaS; shtUishic 8. 
 
 Shycnne no'/oyos 'middle finger'; naa 3; na-nuhh'tu 8. 
 
 ■? Blackfoot nohkh, noho-ka ' three.' [Mass. noeil 'in the middle.'] 
 
 Dakota mipeoclwkai/a 'middle finger' (ochokaija 'in the middle'). 
 
 Minitari shdki-dumdtadit 'middle of the Iiand'; ddmi, ndwi 'three' 
 (dumdta 'in the middle,' nmvah'taru 'between'). 
 Muskoki (Creek) enke nurkiiphueiU 'hand's middlestander.' 
 
 Choctaw iWiui Ss/ii')X-/iin)ia ' hand's middle son.' 
 
 Pawnee skUsi-kadika ' half-way finger.' 
 
 Navajo htillah ndizi " " {hulali', eld 'hanA'). 
 
 In one dialect of the Eskimo (Hudson's Bay) the name, 
 as in the Shyenne above-noted, appears only in ' eight : ' 
 
 MUuklimut 'the middle finger,' 'eight' (Parry). 
 
i^MmSfnffffflftftf" 
 
 On Numerals in American Indian Languages. 13 
 
 4. 7. The Fore linger has been, almost universally, the 
 'showing finger' or 'index.' Names for 4 and 7 are in 
 Algonkin languages taken from it, or from the act of 
 showing, or their connection Avith it is established through 
 the demonstrative pronouns : 
 
 Eskimo (Greeiil.) t|■i■c^• ' the pointer.' 
 
 Algonkin, Ciiip. imi-nimlj ' siiowitig finger ' ; niwin ( = niouin) ' four.' 
 
 Crco ittmiliii/ani-trhilchi!/ ']>()iiiter finger'; newu 'four.' Comp. 
 
 Mtiih 'tiiiit yonder!,' ainih 'tliis one.' [As was !)efore 
 reinnrkcd, the n' prefixed to the Algonkin numerals is a 
 demonstrative panielc, and docs not belong to the root.] 
 Massachusetts ynii (Eliot; = mi) 'four'; yen 'this,' 'there.' 
 Narragiinsett yoh " ;jo ' there,' ' that way ! ' 
 
 Illinois niwi.nhii " newa, newe "voilsV, regardo \h," 
 
 iwa, iiwe "le voilik." 
 nani'sotoyos 'my fore finger'; ni'soto ' seven' ;2 compare 
 nisiwo 'that.' But Shyenne nipa 'four' has a different 
 origin. 
 yen 'four'; ti'v.na 'to touch one to call his attention to 
 
 anything' (llayden). 
 ni-sui ' four ' ; sumis ' look ! ' 
 nape tokaheya 'hand's first' (modern'!). 
 nape apazo 'hand's ))ointer' (pnzo, iipdzo 'to point to, to 
 show by pointing' — >/;)a denoting action of the hand). 
 Mdskoki (Creek) fliifce-esHipttit 'hand's pointer.' 
 
 Choctaw ibhalc-ushi-tiklM ' foremost (or eldest) son of the hand.' 
 
 (The name for 4 is not, in any language of the Chahta- 
 ' Muskoki group, taken from this finger.) 
 
 Navajo <t" 'four'; ti 'here,' 'this'; n'la'-te 'there' (/a = hand). 
 
 Apache Hi " ti' this,' ' who ' ; ti-tchi ' this day.' 
 
 5. 6. The Thuaibs mark ' five ' and ' six,' but rarely, if 
 ever, give a name to either number, in American languages. 
 In Algonkin, and in many other American languages, the 
 tl urab is the 'big,' 'thick,' or 'stout' finger; sometimes, 
 ' the chief." 
 
 Shyenne 
 
 Arnpoho 
 
 Black foot 
 Dakota 
 
 2 The Zulu corresponds with the Shyenne in taking numeral names from the 
 fingers of the second hand. " The Zulu verb komha ' to point,' indicating the fore 
 finger or ' pointer,' makes the numeral 7. Thus, answering the question, ' Hovir 
 much did your master give you V a Zulu would say, ' Ukombile ' ' He pointed with 
 his fore finger,' i. e. ' he gave me seven,' and this curious w / of using the 
 numeral verb is shown in such an example as ' amahusi akombile ' ' the horses 
 have pointed,' «. e. 'there were seven of them.'" — Tylor's Primitive Culture, i. 
 
 228. 
 
 'Compare Latin pollex, "vocatus quod inter cseteros polleat virtute et 
 potestate." — Isidori Origines, quoted by Pott, 'Zahlmethode,' 288. ^ 
 
 8 
 
 i 
 
 ^w 
 
% 'j^^'^^iu^'^*^*****'"' 
 
 »mmmmiiiiiiiisSSSlil&m 
 
 ■ i ! 
 
 ifi 
 
 I 
 
 J. H. Tnmhull, 
 
 Alo. Cliippcwny mitchitchi-mndj 'great tiiiKcr.' 
 Crco (Western) misi-tchitchitn 
 
 Abnaki 
 Miissiiehusetts 
 
 Dakota 
 
 MnsKOKt 
 
 Pawnee 
 
 Navajo 
 
 (C/AiV/i-vrf-iT/x/ 'chief (yrciitest) t1ni;er.' 
 
 I.elllqitaiiilch, niiieklbot oinaki'cliia {omidh-u 't;reat'), and 
 
 Sliyeiinc iiamii-ii-im'oil.-, Imivc the same meaning. 
 napulmuka 'the liiiiid's elder' (Imnhn '|)aroiit, ancestor 
 
 elder brother.' HifrK'*)- 
 enke klihki, and Clioetaw ibbak ishke, ' the hand's mother.' ' 
 stas'-si-ri^s 'lar<:e linger.' 
 hitllah ts('> ' tliiok ' or ' bij; finger.' 
 
 III. Names of miinl)cr that are not derived from the 
 fingers individually. Names for ' one ' and ' two,' as has 
 been said (p. 6), must have preceded digital numeration. 
 
 1, Tliere are, as we have seen, two expressions for the 
 numeral 'one'; namely, ^ onli/ one' and '■first (fore-cst) 
 one.' In Algonkin languages these arc represented in 
 Mass. pdsuk and ntjuf. The former exi)rcssion is sometimes 
 related to the pronoun of tlie first person singular and to the 
 demonstrative 'this'; sometimes it has the meaning ' alone,' 
 ' single,' or ' by itself.' Its root in Algonkin and Dakota 
 languages denotes 'small.' The other expression for ' one ' 
 ( = Mass. n'qut} is from a root denoting j^'writy or /ore- 
 coming, in order or time, ' beginning:' and it has in many 
 languages the secondary meanings, ' old,' ' aforetime,' etc. 
 
 In the Dakota family, one of these expressions is used for 
 the cardinal, the other to form the ordinal: e. g. Dak. tvamhi', 
 wa"zhi'-da",' wa'cha 'one'; toka'heya 'first' (from tokd 'at 
 the first'); Hidatsa (Minitari) duStsa, Ivefm 'one,' itsika 
 'first'; Iowa Tya"gke 'one,' pakranaha 'first.' Between 
 phonetic decay and dialectic growth ,= the Dakota 'ones' 
 
 * So in Mahuan (Pott, 'Ziihimethodc,' 299), and in American Maya, Huasteca, 
 Tamanaca, etc. ; and in Botocudo nipo-diik ' hand's mother.' 
 
 oihankton wa"zhi-na. The suffix, dan, Ihank. na, is restrictive; 'one on/y.' 
 "The form in counting is waiir.ha" (A. L. Riggs) or, as Dr. Hayden writes it, 
 vmnch, Tliis is furtlicr contracted in the Ponka to win, and in the Omalia to wi, 
 
 «0r rather, between "laziness and emphasis," as Mr. A. H. Sayce (Principles 
 of Comparative Philology, 16) prefers to call the two great causes of ])honetic 
 change. Compare Whitney, I-anguagc and the Study 6f Language, 70, 95. 
 In no American family of language is the operation of these princjples 
 more ai)parent and more troublesome than in the Dakota. Not merely that 
 wa'zhidan is shortened to Omaha i«j or changed to Mandan makh'ana and Iowa 
 iyangke, but in the same dialect, and from the lips of the same speaker, a name 
 
 •o*^ 
 
 man m 
 
 
mmm 
 
 ,rrr''Tfiilii 
 
 at 
 
 '*i^ 
 
 ' i ' 
 
 On Numerah in American Indian Laiiguages. 15 
 
 have becjomo so widely variant that they oamiot all be 
 confuleutly referred to a comiuou root. In several^ dialects, 
 if not in all, the numeral has lost all consciousness of its roots, 
 becoming a mere phonetic mark. Compare 
 
 Ponkn mn 
 
 Omiilii. tt'i nnd miaxtcheh 
 Mnnilan makliana 
 Osage minche 
 
 Dakota wa'M'da" 
 Assiiiiboiii washtna 
 Winiieliago izlmk'ida, hezunkera 
 Iowa liia"(ike 
 
 Hidntsa duetsa, Itietsa 
 
 and — least conformable of all — Aubsaroko hamat'. 
 
 I was at first inclined to refer the Dakota 'wa"zhi to the 
 root wi"zh 'to bend,' from the bending down of the little 
 finger in counting. Comparison of ten dialects of the same 
 family makes it more probable, if not absolutely certain, that 
 it is the equivalent of Algonkin }i(tsuk 'the least' or 'very 
 small': compare with ivazhi, u'a'nikhadan 'very little' and 
 waTicha-dan 'very little, none'; wrfVce, the name of the 
 fourth child in a family, if a daughter (remembering that 
 the thumb is 'parent' or 'elder' of the hand), and tvdnka 
 'soft, weak, tender.' With Iowa iyagke, comp. Mandan 
 ungkni-ingke 'hand's little one'; and Winneb. izhdki-da, 
 with wachek 'young." In the Assiniboin, nape 'hand,' with 
 
 may vary as noivassa, duetsa, luetsa (Minitari) 2; pilika, pirika, 10; nahwi, 
 dami, 2 ; bim, mida ' a tree,' etc. In this last-mentioned Dakota dialect, the 
 Hida'tsa (called Minitari and Gros Ventres), Dakota i/ becomes d (ya ' thou| and 
 ya 'to go' = (/e), 6 and w are interchangeable with m, and I, n, and r, with d 
 (Matthew's llidatsa Dictionary and Grammar, p. 28). 
 
 ' Since this paper was written, I have been favored by the Rev. A. L. Riggs ot 
 the Dakota mission (Santee Agency, Nebr.), with some notes on the Dakota 
 numerals, to which his father, the Rev. Stephen L. Riggs, contributed some 
 suggestions. For the grammar and vocabulary of the language, I could have 
 no higher authority; and when I have ventured to differ from Mr. Riggs's 
 conjectures as to the origin of the numerals, it has been only after thorough 
 comparison of the names in eleven languages of this family, with whatever light 
 was to be had from published and manuscript vocabularies. Of the names for the 
 lower numbers, Mr. Riggs writes: "I have thought that, as high as 'three,' the 
 names of numbers arose from sight of outward objects, as 'one' evidently does." 
 " Wanji, root twin, interjectional, 'seel'; ji [dii] is not necessary, as the form in 
 counting is wanchi (for wan-e-cha). Ji means 'separately'; dan added has 
 ■ something the foi . e of ' only.' Nonpa, 2. Root, onpa ' to lay on,' ' to add.' The 
 origin of the n will be sought in different directions, according to the theory of 
 
 the numeral It may be that it comes from 7iape. While nape is the whole 
 
 hand, in composition it may stand for a 'finger,' which is nape-sukaza = ' a single 
 hand ' " [or, ' a portion, particle of hand "i]. 
 
 1 
 
t-*****!/*" 
 
 MWFH«i«MM*m«fM 
 
 immmmmUSim 
 
 ! : 
 
 > \ 
 
 
 16 J. H. TnmhnU, 
 
 nape-washi 'finger,' i.e. ' luiml's little one,' and .r««;i/-«« 
 'one'=='a finger only.' The Dakota for 'fingers is 7iap- 
 suJcdza 'hand's sumll portions,' tVom su 'seed, gram, i.e. 
 'a particle,' which in Dakota more commonly becomes,^ ni 
 composition, chi or cho, as in Dak. chilca-da 'very small _== 
 Assiniboin c/u7ca-«« = Omaha Mn^a 'young,' Mandau -aulc, 
 
 and diminutive shuke, as a suffix." 
 
 In Hidatsa dm-tsa (otherwise luem, nowasscO there is 
 
 wider divergence from the root ; but we recover the meiining 
 
 through shaki-adutsamlke 'fingers' («/m/c/=hand), aduma 
 
 'a seed,' adutsdhi 'fiVo\nt,M tapering end or part' (Matthews). 
 Aubsaroke (Crow) hamat' has the same liican.ng. Comp. . 
 
 Mandan hdmahe 'small,' mk-hdmahe 'little child'; and Aubs. 
 
 amue ' a grain, a kernel.' The suffix -at, -ate, is the common 
 
 Aubs. diminutive, i/amaf 'one' = ' the least. 
 
 In the Chahta-Muskoki family, N\e find the two forms — 
 
 'one only' and 'the first,' represented in 
 
 Choctaw ach&fa 1 ; ' sole, single, only one.' 
 
 tikha ' the first ' ; also, ' before,' ' ancient,' of time past. 
 
 Ummona ' once.' 
 Muskoki hiim'lcin ' one.' ^ 
 
 Coassatti cha/dka and Alabama chafuhka-schle ' one. 
 
 Without attempting an exact analysis of these names, 1 
 remark (1) that Ch. ummona 'first,' Mmona 'once,' is merely 
 a demonstrative : Mmo, himak ' now,' ' at this time,' 'to-day ; 
 Mmonasi 'instantly,' &c.; obviously related to Muskoki 
 Mm'kin V and homa 'before': (2) that G\u achufa and 
 Coas. chafdka, seem, like the Algonkin and Dakota 'ones, 
 to be derived from a root meaning 'very small, 'a grain, 
 particle, or point 'j_«mip^i^^M/i^ an ^U^^ 
 
 ~7^;;mr^o7Dl^";U^ younger b'.^ili^rMOmaha sanga), contractea to 
 .un . ,11 ' a law. a nail ' (Om. sha-ge) ; cho and .u ' a kernel,' ' g-n. 'seed. 
 
 « Mrs A E. W. Robertson (wife of the Rev. W. S. Robertson, of TuUahassee 
 Ind. T^^ritory), whose knowledge of the Creek language is a. thorough as that 
 of any one now living, writes (under date of Aug. 3d 1874). I sec no 
 connection between the [lower] Muskoki numerals and the names of tod or 
 Zers unless hu.nke 1, may be a contraction of h.yu enke ' .h.s hand. In 
 fonCction, m and „ seem to run into each other: e. g ,no.et ^-0- «on. /.j- 
 becomes /.fim. before words beginning with m; as heyun mechetu. to do th s 
 becomes kmlnecketU, heyUn v.aketU ' to say this ' becomes km,naketu. In a s.mdar 
 way, Aei/un «ifce[' this hand'] might become Aumie one. 
 
 .^**»^ 
 
msssL. 
 
 -f-fgr^.:,&4^n9V¥m^^^m' 
 
 •ft-**"^ 
 
 On Numerab in American Indian Lam/mges. 17 
 
 ehush 'tip,' 'point' (e. g. ihhak-chmh 'finger nair=' hand's 
 point'), i6(i/fc-<.-A»/-««?i 'tapering,' elmld"haH 'little, not mucl.. 
 Pawnee &»ka 1, is evidently IVom a root found in i>Hl-,Uki 
 and pir-mU 'young,' /«<«Z«8'Ai ' small,' and probably in >ikcts 
 
 2 Names for two seem to come from roots denoting (1) 
 separation or distinction, as ' that," tlie other,' (2) likeness, 
 equality, or opposition, (3) addition, ' putting to' or ' putting 
 with,' (4) coupling, pairing, or the like. These names, as 
 has been said, must have preceded finger-countmg or any 
 formal numeration. They are often related to -possibly 
 may have in some languages been derived from — names of 
 natural pairs, as ' arms,' ' hands,' ' feet,' ' wings,'^etc. From 
 them or from the same roots come, by later derivation, names 
 of artificial pairs, e. g. 'moccasins,' 'leggings,' etc., and of 
 dual relation, as 'wife,' 'husband,' 'brother,' etc' And 
 here is the explanation of that connection between names of 
 the ' hand' and 'two,' which Mr. 11. Ellis regards as evidence 
 "that hand may =fi»grrs =fingcr-fiu</err and as "helping 
 to exhibit the radical affinity which unites the North American 
 languages" (p. 6). 
 
 Of natural 'pairs,' the hands have most often given 
 a name to — or received it from — the numeral; because 
 they are two, not because they " = finger-Hnger." Pott 
 (Zahlmethode, 29) notes Puris (Brazilian) core ' band,' cunrt 
 2 ; Hottentot tlcoam ' hand ' and 2 ; Sanskrit^ra ' hand^ 
 
 'Ti;7^;;^„M^l^o^^^)r^d 'husband/ kl-W ■wife,' giU 'two,' m.d ki-alsi'^ 
 ■near' i e. 'next to'; of all which the common root is found in ki-n he, 
 i e. 'another'; and Choctaw tuk-lo 'two,' tek^hi 'wife.' The connection of the 
 grammatical dual with the idea of correlation, or of collocation merely, m dlustrated 
 by a peculiarity of Keelumn speech. The regular termination of the plural is 
 -ama, but thc,-e is a special plural in -ntin, for objects belonging to or associated 
 with the noun in the singular: e. g. hhmsl 'house,' M«a«NTiN ' all who belong 
 to the house 'or are 'of the household'; and with a noun denoting aftinity or 
 consanguinity this sufKx -ntin forms a dual, inclu.ling two individuals in 
 coirelaiion : e. g. chosa 'husband,' chosantln ' husband and wife ; mama ' mother. 
 mamantin 'mother and child'; ususi 'daughter,' ususintin 'daughter and mother ; 
 pana 'sister,' panantin 'sister and brothe, '; with masi 'companion and yaua 
 'servant' it forms nouns meaning 'a pair,' masmtin being mure commonly u.std 
 for persons and yanantln for inanimate objects. -Von Tschudi, die Kechua 
 Sprache, pp. 95, 161. 
 
'«..M«^^- 
 
 -1 
 
 t ; 
 
 PI 
 
 IX 
 
 J. U. TrumhuU, 
 
 hdhn ' anil,' pakxha ' wing,' aiul netra ' eye,' uU used 
 also Ibr ' two.' The Samoycd Tawgi, also, exprcs.scs the 
 immhor 2 aiul tlie snhstaiitivo ' liund ' liy terms nearly 
 identical.'^ In Labrador Kskinio, Richardson's vocabulary 
 has iiKii/i/oJc and (igunit lor 2, nggu ' hand ' and aggait ' the 
 hands.' In the Algonkin and Dakota languages names for 2 
 and Ibr 'hands' or 'arms' seem to be nearly related, cither 
 by derivation of one from the other or of both from a common 
 root. In Algonkin dialects, compare — 
 
 Chip. 
 
 Cl-L'O 
 
 ■tiiiidj 
 ■nisk 
 
 linml' 
 
 
 
 two.' 
 
 (■nisk, however, being used only 
 ill composition, us kilchi-m'ak 
 'right hand'). 
 
 Muss. 
 
 ■nutch, 
 
 •nitch ' 
 
 hand * 
 
 nis 
 
 ti 
 
 cf. nisin 'copuliit,' nlchaii 'who 
 
 Almnki 
 
 -reisi 
 
 
 (( 
 
 nisa 
 
 II 
 
 gives liirtli to u child.' 
 
 Illinois 
 
 
 
 
 ninch-ui 
 
 ti 
 
 
 Miami 
 
 
 • 
 
 
 nichui 
 
 ■ II 
 
 
 Arapoho 
 
 ■ichet 
 
 
 (* 
 
 nis 
 
 11 
 
 cf. Inush ' arm,' inailuka ' the 
 other side,' nesltise 'eyes.' 
 
 Shyenne " men " 
 
 In one Algonkin language only, the Micmac (of Nova 
 Scotia), Ave find another name for 2, tabu, i. e. 'equal' ('par,' 
 ' pair') ; but that it had once a wider range, we have proof in 
 the Cree tcpa-kup, Abnaki td'ha-tca's, Mohegan tupou-ivus, 
 and Montauk (L. I.) UiTpa-wa 7, i. e. 2 + (or 2 of the second 
 luuid). The root, in the sense of 'equal,' and of 'enough,' 
 ' sufficient,' is found in all Algonkin languages : e. g. Mass. 
 (redupl.) tatup, tatuppi, .\bn. tetebi-tvi ' equally,' etc. ; Cree 
 niya-tipiycvw ' I myself,'' tiplyaw ' he him-self,' etc., tipi-new 
 ' he measures it,' i. e. ' makes it eqxial to,' tepi ' enough,' etc. 
 Mass. tatitp-pin ' a string' or ' cord ' is as near akin to Micmac 
 tabu 2, as is Engl. ' twine ' to ' twain.' 
 
 [The presence of this 2 in one Algonkin language, and 
 evidence (in the 'sevens') of its former use in others, 
 suggested a doubt as to the origin of the relation I had 
 believed to exist between ' twos ' and ' hands ' in this family 
 of speech. The authority of W. von Humboldt^ and of Pott 
 
 2 Benloew, Kecherches sur I'Origiiie dcs Noms de Nombre, p. 50. 
 "Die Kuwi-Sprachc aut'der Insel Java, Bd. i. s. 20 ft". 
 
 •i-. 
 
'who 
 
 On Nmneralu in Anertiutn Imlmu Lti>i<iiht;i,'H. 
 
 1S> 
 
 disposed iiic to reco^'iii/o this relation. \ comp.irisoii of 
 the several Aljionkin diiiloots and evidiMice sup|)lie(| l>y other 
 Anieriuan languages led nie to (luestion it, and now 1 am 
 nearly eonvineed that the eonncetion of the numeral with 
 natural duals, 'hands' or the like, is not l>y derivation of 
 one name from the other ; that the likeness, if not aecidental, 
 is a consequenf-e of derivation from a eommon root; and 
 that the primary eoneeption of the Algonkin ' two,' whether 
 expressed by Micmae (aim or Chip, nij, is that of ' sameness,' 
 ' likeness,' or ' equality,' represented in the modern Chip, iji, 
 Crco hue ' 80,' ' sueh.' 
 
 The first three numerals are, in the Massaehusetts dialect, 
 1. mi (jut, 2. ?i is, 2. nish; 
 
 in the Chippeway, 
 
 1. nin goto, 3. n z)', 3. n isnwi. 
 
 In these the prefix is, apparently, merely demonstrative 
 (Mass. ne 'this,' 'that'), and docs not belong to the root. 
 In the 'two,' we have, I think, the Chip, iji, Cree issi and iji, 
 'so,' 'so as,' 'like' — which Baraga (Otchipwe Grammar, 
 493) classes as a conjunction, and Howse (Cree Grammar, 
 132, 142) as "the relative adverb of manner" and also " a 
 generic noun." As a verb, it signifies, in the Chippeway, ' to 
 be like ' or ' the same as ' : e. g. anishinabeg nind-ui ' I dress 
 like (appear like) an Indian ' ; iji-nagwad ' it looks like ' 
 something, etc. Niji, contr. nij, 'two,' is 'this, such as' or 
 ' like ' the first ~ corresponding nearly to Micm. tahu ' par,' 
 ' that wliich pairs.' The same root is in the Chip, nidji, or 
 nidf ' like myself,' ' my fellow,' ' alter ego,' which is only 
 distinguished from the numeral by the change of pronoun 
 in the second and third persons — kidji, kidf 'thy fellow,' 
 'thy equal,' widf 'his fellow, or equal' — used chiefly 
 as adjectives, as ividf-anishinaben 'his fellow-man.' The 
 dialectic variations of this particle correspond with those of 
 the numeral ' two ' : Chip, iji and n'ij, Cree isi and niso, etc. 
 In the Illinois dialect, ninchui is 2, nichl or nigi " comme 
 cela" (Gravier). 
 
 If, then, Algonkin ' hands ' and ' twos' are directly related, 
 it is nearly certain that their relation is that of derivatives 
 
 M 
 
 iiimwui i Maia 
 
 mt^ 
 
 sir- 
 
— %/-M(4lJ« 
 
 I, 
 
 20 
 
 .7. //. Trmnhull, 
 
 from ii coniinoii root, or that tin; forinor riM^eivo tlioir name 
 IVuiii — iiiNlc'iid of fiiviiijf it to — tlie iiuincM'al. And this 
 appoars to he Iiik; of the relation of corresponding names iii 
 other Amoriean families of Hpoech.J 
 
 The Dalvota 'two' is the most constant of all tho 
 numerals, and dialectic variations nowhere disguise its 
 relation to natural ' pairs.' The ' twos ' arc : 
 
 (Sioux) Dakota ho'jxi, nuni, Omaha lUDiilKi, irumfia, Mandan 
 m'lm'pa, Osage nomlxnit/h, I'onka ndulm, Iowa ii6tvc^ Winnebago 
 nomp, Auhsaroke noinpe, IJidatsa ndpa, ddpa. 
 
 Witl) these compare: Dak. nape 'hands' and 7iapm *a 
 pair, they two,' hiTpd (a pair of) 'moccasins,' etc., Om. 
 U'lDiha ' hands,' ' fingers,' Osage tiam/ir ' hands,' Fonka mtiipr, 
 Iowa naw6-pa ' finger ' = ' hand's head, or tip,' Winneb. niihara 
 ' hands.' untnp-wiiiKirct ' fingers,' Aubs. ndpcre 'both,' Ilidatsa 
 huupa ' moccasins.' 
 
 The primary meaning of tlie root, o"pa, seems to be ' to put 
 to, with, on, or against,' 'ap-ponere' or 'op-ponere'; as a 
 verb, o"pa is ' to place or lay any thing' on or with another :. 
 comp. o'pa 'to go with,' ' to he at' or 'on,' and (contr.) ovi 
 ' with ' ;* ao'pa, contr. ao", ' to lay or place on ' (as, wood on 
 the fire); sa'pa 'over, beyond, more than,' used in forming 
 the numerals 11 to 19 (o. g. ivikchemna sa'pa topa 14 = 10 
 -f4); ha"pa 'moccasins,' aka-sa'pa 'opposite,' 'set over 
 against,' etc. Perliaps. a"pa 'day' (^a"pa-o 'dawn') is from 
 the same root. We shall find it again in topa 4. Tho 
 prefixed n' m no'pa — which in other dialects varies to w and 
 (/ — seems to be merely a demonstrative or directive, as in 
 the Algonkin numerals, and as in the Dakota verbal particle 
 na ' take it' (imperative only), and in mo" or nu" ' be it so.' 
 
 In the Chahta-Muskoki group, the ' twos ' have a similar 
 origin, in tho notion of ' coupling,' ' mating,' or ' ad-joining ' : 
 
 Choctaw tuklo, Muskoki (Creek) hokkdUn, Hitchiti tvkh'lun, 
 Coassati tdkoluo, Alub. tdkolo-chie. 
 
 The root is represented in Choctaw okla, a collective 
 
 ♦"Kocile, Gram, of Vei Language, notices that /fro means both 'with' and 
 2, ond thinks tlie former meaning original (compare the Tahiti piti ' together,' 
 thence 2)." Tylor's Primitive Culture, i. 235. 
 
 HMDP 
 
fif-'ifHittiHiiTimipi 
 
 On Numerah in Amrrlrnn Ttvlian Langungi'g. 21 
 
 pwnnuu U9e«l f » form the plural of nouuH and both tlio dual 
 and plural (J\<\ ttiii^r. ) of vcrl)S, with tlio iiioaniiins 'they 
 two,' ' tiioy,' ' people,' ' tril)C,' etc., motlified as <Jcluha " all, 
 the entire crowd, niiinlier, or quantity" (Hyin^^ton, Choctaw 
 Grani.,:l2,41). The Chuctfiwf', prefixed, probably represents 
 the "distinctive jtrcposition" ct 'here, this way,' etc. (id. 42), 
 a demonstrative. From the same root, apparently, are Ch. 
 ho'cu/iln, conjunction copulative, 'also,' ' of the same class,' 
 hltukla (^=rd-okhi) 'twice,' and the verbs (ok-chi. 'to tie,' 
 and iha-takht ' to go with,' ' to accompany.' Comp. Musk. 
 mhokolii ' twice,' hlim-hohtlat ' secondly,' etc. 
 
 Athapascan 'twos' are, more commonly, related to names 
 for ' feet ' than to ' hands.' Chepcwyan " h:h ' foot,' ' shoe,' 
 'track'" (or their plurals), is often used as a numeral for 
 2 or 'a pair.' In the Apaciie, 2 is na-ki; 'foot' or 'feet,' 
 ki-e; 'moccasins,' si-ke; Navajo na-ki 2; ik4 'foot'; kikh 
 'moccasins."' 
 
 3, Names for 'three' when not taken directly from tho 
 middle finger or 'half-way' of the hand, sometimes have 
 the meaning, 'beyond,' 'further' ('trans'), or 'greater'; 
 sometimes 'much,' 'the many' — a plural as distinguished 
 from a dual. 
 
 All the Algonkin ' threes ' are of the ' middle ' (see p. 12, 
 ante), except the Micmac tchicht, whicli seems to have had 
 the meaning of ' more ' or ' again ' ( = Delaware tehitch ' still 
 more'). 
 
 In the Dakota family, the ' threes ' exhibit wider variance 
 than the ' twos ' from the original stock : 
 
 Winneb. id", tau" 
 Iowa Ulni/i 
 Omaha thdbathi 
 Ponka thdbthin 
 Osago laubena 
 
 The etymology is obscure. Comparing the Dakota and 
 Aubsaroke forms with the Omaha, Ponka, and Osage, it 
 seems probable that -am is a contraction of a"pa — as nom is 
 the contraction of no"pa 2, and torn of topa 4. This would 
 
 6 Gallatin, Synopsis of the Indian Tribes, p. 215. 
 
 Dak. ydmni, ydmini 
 Assinib. ydmini 
 Mandan ndmeni 
 Hidatsa ddmi, ndwi 
 Aubsar. nam 
 
 * 
 
 I 
 
 siS****^ 
 
. s'l a ff-u 
 
 **'";«HS4"..*« 
 
 
 ; t 
 
 22 
 
 J. if. Trumhnll, 
 
 refei the numeral to the same root with the ' two.' The 
 prefix may he the simple verbal yn (Hidatsa (7^ 'going,' as 
 in dya ' they go together' and ' it becomes,' or more probably 
 the inseparable prei)osition i (combining with the following a, 
 as ?/rt) meaning ' next in order,' 'again.' Tiiis would make 
 yam = i-ao'pa or ?/rt-o>a — agreeing nearly with the verb 
 iydo"pa ' to lay on, to place on,' of Riggs's Dictionary. The 
 pronunciation of the numeral is marked ya'mni, which 
 suggests a reference to the verbal root mni 'spread out' or 
 mna ' gathered, collected ' ; but the other dialects show that 
 this root is not essential to the name, and if it enters 
 into the composition of the Dakota name, it is probably 
 supplementary to the principal root, so that ya'mni = 
 yam-mni.^ 
 
 The Winnebago and Iowa names have, apparently, a 
 different origin, and Winneb. ta may be the (regular) 
 contraction of taka ' great.' 
 
 In many dialects of the west and southwest, the name of 
 the numeral has this meaning of ' great,' ' much,' ' many,' or 
 the like : e. g. 
 
 Yuma (Mqjiive) Anwioco 3, ' hnmik ' \^Tcai.' 
 
 (Cuclian) /loiHufc, n'ljamuk " 
 
 Pawnee tawit, -taMj/o (suffix) 'over, ahove,' /lawa 'more.' 
 
 Ankara tawhit (wA English), terlme ' many,' tiiincheu 'jiveat.' 
 
 Navajo tahh, Ma ' much,' na-Ui-ni ' a chief.' 
 
 4, Above 3, traces of digital numeration become more 
 common, but the fact that in many languages 4 is a ' doubled 
 2,' or pair of pairs, seems to indicate that in these its 
 conception and name were earlier than finger-counting. All 
 
 8 The Rev. A. L. Riggs, i" his letter of July 27th, before mentioned, regards 
 mni as the root. He writes as follows : 
 
 " Yamni; root mni or mna. Mni is ' to gather in a eircle or group'; as i/hmni 
 wachipi ' the circle dance,' umchij/api ' assembly.' Three is the smallest number, 
 of course, that can make a group or circle. The correlate root mna is more 
 widely in use, and the meaning clcaren: ioMNA 'lo acquire or gather for one's 
 self,' Mvxyan ' to gather,' opa-mna ' a cluster,' as of young trees growing up out 
 of the root or sttini]) of an old one. If yamni comes from mna, the change of a to 
 I would be for eui)hony. If yamni comes from the sight of outward objects 
 [preceding formal enumeration], then we may find the i/a to signify grouping by 
 ca//jn(7 — 'calling' another to the two. If it springs from the finger count, the 
 origin of ya is not clear. As causative affix, it should come after." 
 
 ;;.,j^B MW i Wi iiMWWii '«<MWWiWWW<1l«"»(»WW>li'' 
 
1 
 
 or 
 
 \ ■ ■ 
 
 On N'umerals in American Indian Languages. 23 
 
 Algonkiii ' fours,' as was seen, are demonstrative, derived 
 from the indcx-fiiiger; but in two or tlu-oe dialects tlie 'eights' 
 suggest a primitive uumeratiou by pairs. Of this mode I 
 will apeak more particularly hereafter, aud here mention only 
 the Dakota 4, formed apparently as a ' pair of pairs ' : 
 
 Dak. tdjm, cuutr. torn, Hidatsa tdpa, Mandan tdpe, Pouka 
 and Omaha ddba, Iowa tdwe, Winneb. chop, Avibsaroke shop. 
 
 There are several Dakota expressions for 'pairs' and 
 'doubles'; nrt/«'n (from nape 'hands'?) 'they two,' 'both,' 
 sakim 'tvvo together,' and from the numerals, by the prefix 
 ta, as ta-wa"zhi.. ' a i)air,' ta-no"pa ' 2 pairs,' ta-yamni ' 3 pairs.' 
 In tano"pa, or rather in the earlier ta-o'pa, ta-dpa, ' 2 pairs,' 
 we have, I think, tiie origin of topa 4. 
 
 In some languages ' all the fingers' give the name to this 
 numeral, as, apparently, in Pawnee skitiks i, = skets-iks 
 ' fingers [of^ hand.' 
 
 5. There 's much diversity, even in languages of the same 
 stock, in expressions for o and 10. In these sometimes, 
 but by no means always, is found a name of 'hand' or 
 'fingers,' or a suggestion of such name. In the instances — 
 comparatively few — in which names for 'hand' and 5 are 
 identical, or nearly so, we cannot confidently decide which of 
 the two is borrowed from the other.' 
 
 Of Algonkin ' fives ' there are two principal types : 
 
 (1.) Massachusetts napanna, meaning ' on one side,' i. e. 
 ' one of the two hands.' It is the Chip, nahane, Cree nabat, 
 but is not in either of those dialects used for the numeral. 
 In Abnaki bare-nesku, Del. palenacn, the name for 'hand' 
 is added, the expression r-^rresponding to Chip, -hane-nindj 
 ' of one hand,' as in ningoto-bane-nindj ' one handful,' nin 
 
 T<'A. V. Humboldt's plausible comparison between Skr. pancha 5, and Pers. 
 penjeh ' the palm of the hand with the fln^^ers spread out, the outspread foot of a 
 bird,' as though 5 were called pancha from being like a hand, ia erroneous. The 
 Persian penjeh is itself derived from the numeral 5, as in Skr. the hand is called 
 pancha^dkhn ' the flve-branchcd.' The same formation is found in English ; slang 
 describes a man's hand as his 'fives,' or 'bunch of fives,' thence the name of the 
 game of fives, plaj-ed by striking the ball with the open hand, a terra which has 
 madeits way out of slang into accepted language."— Tylor's Primitive Culture, 
 i. 235, note. 
 
 Mmnv-MniMMMMatMlltllM 
 
 If 
 
 «* 
 
I 
 
 '■"t'JMM^Ii^TT"*" 
 
 I ' 
 r 
 
 i 
 
 r. 
 
 I M U 
 
 u 
 
 ajaSSS i SltMtJJU^JUUIIM a 
 
 ^ 
 
 J". IT. Trnmbull, 
 
 nahane-nindj 'I am one-handed,' 'have only one hand,' etc. 
 The. Abnaki noTneda '5 times' and naJ'nalcao 'fifteen' 
 ( = 5 + ) arc from a different root, and are related to 
 
 (2.) Chip, ndnan, Cree niannoii, nli/<fnan, Micm. nan, 
 Moheg. nunon; and Sliawuno nlaUn-ui, Miami ydlan-ne, lUin. 
 miaran-ui, etc. These, thongh perhaps not all from the 
 same root, have nearly the same meaning, ' gone,' or ' spent,' 
 i.e. all the fingers of one hand. Comp. Cree nit/an ' va, 
 pars,' pi. luyank ' allcz, partez,' a " verb used only in these 
 two persons of the imperative"; niydk 'forwards,' 'onward' 
 (Lacombe, Diet. Crise); Shawn, niala, Illin. miara = Cveo 
 
 niydn. ^ 
 
 Dakota 'fives' are plainly digital: Dak. zdpta , Ora. 
 sdtan, Ponka sdta, Iowa Ihdta, Osage mttah, Winueb. mtch; 
 Hidatsa Mchm, Mandan kecwiin. Dak. zdpta" = za (for 
 mka-za ' fingers ') +j»^<a" 'turned down.'* Hidatsa Hchm, 
 from ciiu 'thrown down' or 'overturned,' with /«, the 
 intensive and frequentative prefix, ' wholly, completely,' i. e. 
 ' all turned down.' Or, if we suppose the word to have lost 
 a syllable, and restore it as sdki-c^u, we have ' hand turned- 
 down ' = Dak. za-pta". 
 
 Choctaw taldapi 5, seems to be compounded of ^aJili 
 'to finish' Or 'complete' and ahpi 'the first' = ' first hand 
 ends.' In Muskoki chagh'kih'pln, and Hitchitee clmghkupun, 
 the Musk, chunggi or chunki 'my liand' may perhaps be 
 recognized, but if so, it is nearly lost in the Musk, ordinal, 
 hlisa choUkepe ' fifth.' 
 
 Pawnee sihuks is from iksu-huks 'hands half; still more 
 contracted in Arikara ^hehii (?*.•/*«=' hand'). 
 
 In the Athapascan, la 'hand' seems to be found in Navajo 
 dst-la, Apache dsht-la 5; but only in these two of the eleven 
 languages of that family compared by Buschmann, who 
 remarks on the general resemblance of tbe Athapascan 5 to 
 the 1. Eskimo (Labrador) tedli-ma, telUmet 10, is probably 
 related to tallek 'hand.' 
 
 "This agrees nearly with the meaning given hy the Rev. A. L. Riggs (in his 
 letter of July 27th); " Zaiitu". Roots za imd ptan. Z(i is 'the hand'; thus, 
 yu-za is 'to hold,' 'to handle.' Ptan is 'turned over.' The whole of the hand 
 [i. e. all the Jinyers] is now turned down." 
 
 ftf iiiiifl 
 
 ■i 
 
:^ 
 
 On Numerals in American Indian Languages. 25 
 
 In the Shoslioni family, Comanche motvaka {mown 'arm, 
 hand, fingers,') and in another dialect mo'ovct (moo-oijet 
 'hand all'), Shosh. mandgct, Yute manigin {moo-ninch 
 'hand'), all give evidence of their manual origin.' 
 
 6. Names for this numeral in Algonkin, Dakota, and some 
 other families of language, mark it as the first that is counted 
 on tlie second hand. This is done (1.) by affixing to a name 
 for 'hand' a particle moaning 'one,' 'first,' or 'other,' or 
 (2.) by repeating the name for 1 and affixing a word meaning 
 ' again,' ' besides,' ' beyond,' ' more,' or the like, or (3.) by 
 merely expressing change ' to the other side.' Of these, the 
 second is the most common type : e.g. — 
 
 Alg. Cree niJcoto-wasik, nikut-ivassik = ■ 1 on the other side' 
 (Cree awas ' further on,' aivasd-yik ' on the other side '), 
 Chip, ningot-wdsstvi (awdssaii 'further'), Abn. nekuda"s, 
 Moll, n'guittus, Shaw, nigote-wathwi, Sauki kotoashek ; Mass. 
 nequtta-tahshe, Del. qnttasch (the affix, adtahsJw, means 
 'counted' or ' added '). — Micmac aahugOm {apch 'again,' 
 ' following' ; apchku ' going back'), and Mareschit kdmachin, 
 seem to be similarly formed. — Illin. kakatchui 6 denotes 
 'passing beyond the middle' (Icakatahe). — Shyenne nasutu 
 (jnahsoto, Abert) is ' one over.' 
 
 The Dakota presents two types — which, however, may 
 prove to be originally identical : 
 
 Dakota shdkpe IIi<latSii aka-wa, aka-ina 
 
 Assinib. shd-kpa Wiiiiieb. akewe 
 
 Om. and Ponka s/i((-/)e Aubsaioke ki-ina 
 
 Iowa sha-kwe Muiidun akd-mak 
 
 Osage sha-pah 
 
 Oto slia-kiva 
 
 Hidatsa m and M; = Dak. p. The only question is as to 
 the precise meaning of the Dakota prefix. Dakota pe is 
 'finger' or 'fingers' <hand 'points,' as in napehu-pe, 
 etc.), as is more clearly shown by Ponka 7, pe'namha ( = 2 
 fingers), and 8, pe'thabthin ( = 3 fingers). The prefix 1 take 
 to be Dak. a-kshd 'more, in addition to.' Then shdkpe= 
 a-kshd-pe='l in addition' or 'besides' (the 5); and Hidatsa 
 
 iFor other 'fives' of Buschttiann's Sonora family, including the Slioshoni, see 
 his Gramraatik d. Sonor. Sprachen, 3tc Abth. ss. 1 14, 119. 
 
Ji:*y' 
 
 -t /^i,, "^ 
 
 li^ L i ian ai ^ ii Wiiw i— w oi ' in ' t 
 
 *mmmmm 
 
 ! 
 
 \U 
 
 J. H. Trumbull, 
 
 akdwn = ak(t-ma ' ono over'; comp. Assiuib. akdn 'above,' 
 hake-cha ' afterwards,' tfec.'^ 
 
 In tbe Atlialiascaii family, Buschmann' finds 6 expressed 
 by 3x2 in five languages (of eleven compared). 
 
 7, 8, The composition of these numerals from 2 and 3 ia 
 as common in American as in other families of speech, Aa 
 independent name for cither 7 or 8 is exceptional. The 8 is 
 sometimes designated from its proximity to 10 — as 'two 
 lesi;,' 'two left,' or as -coming near' tiie end; 7, more rarely, 
 as ' wanting 3,' or the like. The common expression for both 
 numerals is formed by affixing to the names for 2 and 3, 
 respectively, a word denoting addition or repetition. In some 
 languages, an indication of ' hand ' or ' finger ' is comprised 
 in the name. The Algonkin 7 has generally the same affix 
 as the 6, meaning ' on the other side' or 'again,' The full 
 expression is preserved in Chip, nij-tvdssiui 7, nish-tvdsswi 
 8; com])are ningot-tvdssu'i 6: a contracted form, in Del. chash 
 and, with a guttural modification, in Moh. ghimt. The Oree 
 and Chippeway languages have each another name for 7 : 
 Cree tSpakup (teypuckoop, Howse), Chip, tupomvus { = tepu- 
 awassivi), the latter agreeing with the Abnaki 7, ta'bawas; 
 all formed from a ' two ' which is not now found in any 
 Algonkin language except the Micmac (see p. 18, ante). The 
 Crees have also two names for 8 ; skwdsaik {=znishu-awdsih) 
 and aiendneivu or ayendneu. The latter is peculiar. It seems 
 to be formed of iyin ' more ' and nSwu 4 = ' 4 again ' or 2x4. 
 An exceptional name for 7 is found in the Narragansett ^nada 
 (Mass. enotta of Wood's Vocabulary) ; perhaps related to 
 Mass. nahohtoeu ' second,' literally ' tiiat which comes next,' 
 or perhaps from the index-finger and act of 'showing' (Mass. 
 ndtin-au ' he shows it to,' Ciiip. enoad ' showing with the 
 fingers '). The Sauki 7, ndtvia, may have had a similar origin. 
 Illinois parare, Miami pohine 8, mean 'nearly ended,' 
 'almost done.' The composition of lUin. suatatchui, Mi. 
 suaxtetatii 7, is not clear. 
 
 i* Tlic Kev. A. L. Rijfgs has siifrgestcd a different deiivation of Duk. shakpe 
 "from s/iuti ' the nail' and kpa or k/ie 'punched out.' The prominent thumb 
 nail of the second hand is now pusliud down." 
 
 ^Sj-8tem. Worttafel d. Athap. Sprachstnmras (3te Abth. des Apache;, s. 508. 
 
 ""- 1 
 
 ! 
 
mOm 
 
 ^ 
 
 '■^' 1 
 
 On Numerals in American Indian LanguayeH. liT 
 In the Cliahta-Muskoki group wo have — 
 
 Choctiiw (i(Wo 2, and un-lidhl. 
 
 - tuchiiiaa, itn-lMcliimiS. '; ^ 
 
 Coassatti luloluu 2, lion-toLolor, 7. 
 
 Alabaiuiv tdkoldcliie 2, " hon-tnkoUi-cliie 7. 
 
 The prefix un- or /wn- ( = Choctaw o«0 means 'again.' 
 In other languages of this family, the names for 2 and 3 are 
 similarly modified by a suffix : 
 
 Musk, hokkokn 2, and . kdlU-pakcn 7. 
 
 Hitehiti tokhlnn 2, kola-}Hiken 7. 
 
 Musk, tutchenen 3, chenu-fitiken 8. 
 
 Hitehiti tohchiinim 2, idsna-paken 8. 
 
 One of Mr. Ellis's mistakes is that of regarding these 
 adverbial affixes as representatives of names for ' hand ' or 
 ' finger,' or ' five ' ; and some of the most striking of the 
 coincidences that seem to him " to exhibit the radical affinity 
 which unites the North American languages" vanish with 
 the correction of this error. He finds, for example, his "a2 
 finger" or his " J«2 finger," or tlie two combined as ''azbaz 
 'finger-fing^er':i=hand," in Delaware cottash 6, nishash 7, 
 old Algonkin (Nipissing) ninskwassoo 7, nisswassoo 8, Cree 
 "nikutwassik 6, nishwassik 7, etc.^ Whatever the Basque zaz 
 (conjecturally extracted from Basque zazpi 'seven') or a 
 possible svas of " the original Aryan vocabulary " may have 
 denoted, it is certain that in the Del. -ash, Alg. ivassoo, Cree 
 wassik, etc., we have merely an adverb meaning 'further,' 
 ' on the other side,' or the like. 
 
 In the (semi-Algonkin) Atsina dialect, 7, 8, and 9 are 
 formed respectively from 3, 2, and 1, by a suffix that denotes 
 the ' fingers ' remaining to be counted. % 
 
 In the Dakota family, there are at least two and perhaps 
 three types of ' sevens ' : 
 
 Ponka pi nanba Dak. shako-wi" Hidatsa shdpua 
 
 Omaha, pe-namba Assinib. shakii-wi Aubsar. khdpua 
 
 Osage pdnompd Winneb. shag6-wi Mandan kupa 
 
 Iowa slidhma 
 
 The first three prefix to 2, pe, pd ' fingers ' (lit. hand 
 ' points '). Of the others, I find no satisfactory analysis that 
 
'%:tmHt J * - }"'"»» • 
 
 ll 
 
 28 
 
 J. H. Trumhull, 
 
 will apply to hotli <?roups. [The Rov. A. L. Ri,t?gs has 
 suggosted, for the Dakota proper, xhake 'a nail ' and win ' to 
 bend,' with the proposition o 'in' oi- 'on' interposed, the 
 fore-finger (of the second hand) being bent upon the nail of 
 the previonsly turned thumb.] 
 
 The Ponka and Omaha ' eights ' are formed like the 
 ' sevens' — by prefixing /*« to 8 ; the ilidatsa and Aubsaroke, 
 by suffixing pe, pi, to 2, the numbers of fingers remaining 
 uncounted : 
 
 Ilidatsa rfoyia 2, dd/m pi 6, {pitikn \0). , 
 
 Aiibsar. in'ij) 2, nOpa-pe 8, (pirtika 10). 
 
 Dr. Matthews (Ilidatsa Grammar, 56) remarks that dopapi 
 probably signifies ' ten less two,' and that pi seems to be the 
 root of pitika 10. But the primary meaning of pi, pe, is 
 ' pointed ' (or as a verb, ' to penetrate '), and hence ' point,' 
 ' extremity,' ' finger,' as in Hid. icpii and iehprc = Dak. ehupe 
 in nap-chupe ' fingers,' i. e. ' hand points.' In ^Vpe ' the tail 
 of a bird,' Dak. upi. we have another modification of this 
 root ; and again in Hid. ipl-ta ' at the rear, behind,' i. e. ' at 
 the end.' 
 
 Iowa kre-ra-pa-ne 8, is clearly related (as a diminutive?) 
 to kre-jM-na 10. Dakota sha-hdo'gha" and Assinib. shakando- 
 gJiah follow the ' sevens,' the first element of the name being 
 the same in each, but I must leave both — with Mandau 
 tetu/e'e — unexplained. 
 
 9, very generally, is named as being the ' last but one ' ; 
 occasionally, as ' fourth ' of the second hand : 
 
 Ai-G. Croe kcka miirilat 'almost 10.' ^keka 'au point de.' 
 Chip. shcing-asswi (and contr. shany); comp. chdyisse 'used up,' 'all spent.' 
 Shaw. chakatswi " " " " 
 
 Mass. paskugnn ' it comes near.' 
 Del. /)cf/ifa(H^ 'coming near.' 
 
 lUin. nignlu-iiinnekki ' only one left,' lit. 'only one, no more.' 
 
 Arapoho f/iia(oWi' or sialokh' 'again last,' 'one after'; from chia 'again' and 
 
 tdkli (comp. tiikh-su 'last,' takhu-u 'after'). 
 
 Dak. Omaha, Osage, and I'onka, slidnka, Iowa shamjke. Sioux nap-chi'wanka. 
 
 Prince Maximilian von Wied notes tlic Osage as a contracted abbreviation 
 
 of (jiithona-liJieh-wininyka = 1 less 1 . Thi.s is certainly the meaninfj, but not a 
 
 tmiislation of the name. In the Sioux, imp = nape ' hand.' In other dialects, 
 
 shdnka is Iowa ii/am/ke'one,' ' little one' (and, as diminutive, cMnije), Mandan 
 
 ingka (as in ungkni-ingka 'the little finger'), Or, '■•'■ shOiga (ruAwpl. shinge- 
 
 shinge 'an infant,' very small), Sioux cht'chy. "uOa one,' and in chi-Ka-dal* 
 
 «i.,- .....,'! ...JUiWi 
 
On Numerals in American Indian Languages. 
 
 29 
 
 has 
 , 'to 
 
 'very simill ' (with whicli compare tmmd-i-dan 'very little,' wa''ske 'the 
 fourth (Rinalc) ehilil '). Sioux rhi'watika in 9, seems to oontiiin an iidilitional 
 element, which may l)e eclu' 'only' or ,W,i" 'now.'^ The mfanm- is the 
 same, in all theses iliaiei'ts, 'only one (inp'r' remains. 
 
 Hidatsa tluttm-pi im>\ Aul.saroke a„Hilniil have the same meaninj,' — 'one 
 
 finger' ; and so has Mamlan iimcii jie, from macn'nna 1. 
 
 CllAHTA-MusKOKi. Choetaw rliukali 9 = chki-rMi ' soon the end,' next the last. 
 
 [The root, cha, che, is the nearest aipproximation to a conjunction copulativo, 
 
 and may he translated ' and then,' or ' next.'] The same component is in 
 
 Alahama iln-ch.ilihili-chie (chie - finger) and Coassatti hih,-h,iha,iln. Ihxik. 
 
 nstapahkin and lliichiti ostii-jHikin, are from Musk, ostin, Hiteh. aitdkin, 
 
 'four.' 
 
 Natchkz witip'katiiiis, 1 left? 
 
 Caddo hiwiisika, 4 + hand, 
 
 Adair, sikinish, ' hands' minus ! 
 
 Pawnee d'hfik'sidi-wa, 10 minus, 
 
 Arikara nucnimwm, " 
 
 Wichita chius-skinte, 1 left'? 
 
 Kichai tanerokat, "! 
 
 SiiosuoNI shimmiromen, 10 minus? 
 
 Comanche 
 
 shcmnn'-uwum, " 
 
 " se'ermano, 
 
 Yute surrom-siwne, " 
 
 " suioiirroiimsoi/uni, " 
 
 Yuma : Cuchan hum-hamook' , 3x3? 
 
 Mojave pai'<t ' near ' 
 
 eliju-thouk ' near ' 
 
 from wlla 1. 
 " hiweit 4, stciie ' hand.' 
 " sfAiiJ 'hand.' 
 " d'luiksidi 10. 
 " nuciiini 10. 
 " chius I. 
 
 {arisko 1). 
 
 sliiinmer 10. 
 
 shiiiman 10 (Pike, MS.). 
 
 se'ermitno-wiimpnet 10. 
 
 tom-snene 10. 
 
 toamsniiinimi 10 (Powell, MS.). 
 
 hnmuok' 3 (humhook 6). 
 
 {Jtipau'ac 'near'). 
 
 {lliotvk " ). 
 
 10 The tenth finger — the little finger of the second 
 hand — gives in some languages a name to the corres- 
 ponding numeral; but more often, 'ten' is designated 
 as the 'completion' of the digital series, 'all gone,' 'none 
 remaining,' or the like. Occasionally, the name may liave 
 been taken directly from the ' hands' or ' all the fingers.' 
 
 In Algonkin languages, the 'tens' are of four types — of 
 which two are nearly related : 
 
 1. Chip, midasswi, mitasui, Illin. matatehui, Shawano 
 metathwi, Cree mitatat, Shyenne matoouto, Arapoho metaitoCK, 
 and Ataina matatasits — meaning ' no further,' ' completed.' 
 
 "TTf^ly regarded this chi" as the representative of the verbal root cAi« 
 'wanting.' To this, the Rev. A. L. Riggs objects, with good reason, that "cAi^ 
 is not 'want' in the sense of 'lack,' but always of 'aesire';" and that if it 
 made part of the name, "it should come last, as the principal verb. I do not 
 agree with him, however, as to the impossibility of getting 'one (or rather 
 ' finger,' or-- little one ') out of wanka. The other related dialects seem to testify 
 unmistakably to this meaning. 
 
 5 
 
-^■.«*«4s 
 
 •MiiMia 
 
 30 
 
 J. 11. Trumlmll, 
 
 2. Abnaki m'tdrn, Micinac m'trin, Delaware tnfrllcn, frllni, 
 Moliep;. ?»V«/iH/<=' 110 more.' 
 
 3. Massaclnisetts and Narra^axuHet pai'uk. 
 
 4. Sauki aiitl (Northern) Cliippeway kwrtrh, used occa- 
 Bionally in ra].id countinfj. This is either a contraction of 
 isku'dtch (Cree ukwry,1trh:) 'lastly,' 'at the end' (C'.nip. 
 ishhretchagan 'the last or youngest child in a raniily'),or 
 it is Nii»issing-Algonkin kaijoirctah ' no more.' 
 
 The prefix in Illin. mat-atchni. Chip, mid-asswi, Ahn. 
 m't-dra, etc., is the negative and privative particle, Ibnnd in 
 all Algonkin langnagos, though less common in Chippeway 
 than in eaKtern dialects. It is found, however, as a prefix, 
 in many Cliippoway words (e. g. nin gi^asikan ' 1 arrive in 
 time,' iiin med-n>s>iihin ' I do not arrive in time,' ' I am too 
 late ' ; n'md apdh ' I sit upon ' (a seat), nin mit-ab ' I sit upon 
 the bare ground, the snow, or the like,' ' have nothing to sit 
 upon ' ; etc.). As a verbal prefix, it has sometimes, with a 
 modified vowel, the meaning of 'ceasing,' 'leaving off,' 
 •' completing ' ; e. g. Mass. mahtu 'he ceases speaking,' Abn. 
 met-anaskiwi ' imaWy; llVm. mita-tewi ' au abandoned cabin,' 
 ni metassa 'I bury (i.e. have done with) him ' = Chip, mid- 
 dgwena ' I put him aside, or out of the way.' 
 
 The suffix dsstvi is the same as in Chip. ningoUvdsmi 6, 
 nishwdsswi 7, meaning ' further' or ' beyond.' At 10, there 
 is ' no further ' count, ' a completion.' Abn. -am, Del. -elen, 
 Moh. -anit, are forms of the same particle of comparison, 
 meaning ' more,' ' above ' ; and mid-dmvi = m''t-dra. 
 
 I have the more particularly pointed out the composition 
 of this Algonkin 'ten,' because more than one writer on 
 American languages has been struck by the likeness of Chip. 
 midasso (the ordinal) 10 and middss 'a legging.' Mr. R. 
 Ellis" observes this likeness in six or seven Algonkin 
 languages, and infers that " forms like -doswe, -tathi, -tato, 
 ■tato, etc., may be compared with Uchee (Florida) tetliah 
 ' shoes,' and tetAihah ' feet,' " etc., all contributing to show 
 that the " az finger" and the '' azhaz hand" prevail, and are 
 employed numerally, over the greater part of North America 
 
 « On Numernls as Signs of Primeval Unity, etc., p. 9. 
 
 » 
 
 mmm 
 
 r^ 
 
>?/, fi/Irn, 
 
 sed occa- 
 
 i-action ol' 
 
 ' (cuinp. 
 
 uily '), or 
 
 iwi, AUn. 
 , found in 
 Jhippeway 
 3 a prefix, 
 arrive in 
 I am too 
 I sit upon 
 liing to sit 
 les, with a 
 aving off,' 
 Ling,' Abn. 
 icd cabin,' 
 Chip, mid- 
 
 t-wds8tvi 6, 
 t 10, there 
 , Del. -elen^ 
 omparison, 
 a. 
 
 jomposition 
 writer on 
 !8S of Chip. 
 ;.' Mr. R. 
 I Algonkin 
 ■tathi, -tato, 
 •ida) tethah 
 ng to show 
 'ail, and are 
 'th America 
 
 I 
 
 Oh Nuinemh in AmeriiMn Indian Languages. 
 
 31 
 
 as well as on the eastern continent, " the vi- prefix " in 
 midanso, etc., " appearing the su as a Californian and New 
 Mexican prefix m-, which is used to convert 'arms' into 
 ' legs,' " 
 
 The learned author of " Ktudes riiilologiques snr quelques 
 Langues Sauvages" (pp. 1:?1, 132) has given an etymology 
 ofmitami which is ingenious, but to which llicre is, 1 think, 
 one iiisuperal)le objection. He derives the name from the 
 particle mi ' so,' and tasui, tano, " a {)article that expresses 
 quantity and is the eipiivalent of [the French] adverbs tant, 
 autant, comhien:' When an Indian would express ■ ten,' he 
 puts forward l)oth hands and spreads the fingers saying, 
 mi-tami ' so many.' The objection to this is, that it will not 
 ap[)ly to other Algonkin dialects, nor to other numerals in 
 the same dialect: it will not serve either for Abn. mHdra and 
 Cree mitatat 10, nor for Chip, ningotami 6, changami 9, etc., 
 in which M. Cuoq finds, not dasso ' so many,' but asiii " en 
 BUS, de ])lus." 
 
 In the Massachusetts and Connecticut dialects another 
 name is found for 10, juam/fc {piuk, piogquS, Eliot), but the 
 Chippeway mitami is represented in Mass. muttdsons 'the 
 youngest child in a family ' (mat-dm ' not after,' with -ona 
 diminutive), and in muttasp-nitch 'the little finger,' i. e. the 
 least and last. Mass. and Narrag. paiuk is, probably, a 
 similar expression, related to peauk (=pi-es-uk, dimin. of 
 pi-ak) ' least,' ' one only,' and to Cree peyak ' one,' ' alone,' 
 as well as to piko ' only,' ' no more than,' and piyis ' finally,' 
 
 ' lastly.' 
 
 The Dakota 'tens' may be reduced to two groups, the name 
 having in both the same general meaning, but not formed 
 from the same roots : 
 
 (1.) Sioux-Dnk. and Assiniboin wikchemna, wikchem'ini. 
 
 Ponka gthe-ba. 
 
 Omaha chrabcm, and g'kh'ba,'' Iowa krfpana, Oto krahbra", Osage krabra, 
 Winneb. kherapun (or kherapim-axe, Haydcn).'* 
 (2.) Mnndan pirakh, Aubsar. pirakd, Hidatsa pilika. 
 
 7 Prince Maximilian's vocabulary gives chrabene; Dr. F. V. Hayden's (in Proc. 
 Am. Philos. Society, x. 407), g'M-ha, but the second h probably is by misprint 
 for 6, since 20 is g'th'eba-namba ' two tens.' 
 
 8 In this group of Dakota ' tens ' we have a good illustration of one difficulty in 
 
 ^.MMMIIWJM I iWli lJ Miau ifc li li *' 
 
L, ^IIM 
 
 ijll 
 
 t:'' 
 
 1 
 
 ir ' 
 
 l;i 
 
 iB J. U. Tnmhidl, 
 
 At 10, the fingers fluit linve heeii liciit down arc fitrai(/htined, 
 and "the hands xjircdd out side by side."" Wlkchnima is 
 from ^'7;(« ' straiglit,' 'nnlxMit,'' and mna 's|)read ont,' with 
 the gcnvrallziwi [)refix ol' Sionx nouns, wi or «''. Ilithitsr, 
 pitika is from the \Qvh pt'iki "to smooth out, to iron ch)tlies," 
 whicli ^lattliews (Hidatsa Dictionary) refers to pokUi (fi'om 
 kiti^ 'to press to snioolhnoss with f/ir hnmlx."' Both 
 expressions "gehn ans von den Fingern," hut in neitlier 
 does a name of ' finger ' or ' liaiid ' show itself. 
 
 How slowly the savage advanced in numeration may he 
 inferred from the traces found in many languages of a mode 
 of reckoning by j>airs and triplctH. There arc sonr^ reasons 
 for believing, not only that conceptions of ' one,' ' two,' and 
 'three' (as ' this,' ' that,' and 'beyond' — or the like) were 
 antecedent to digital numeration, i-ut that the first definite 
 conception of 'four' was as a M)air of [)airs,' and that 
 multiplication of the lower inunbi.'rs often preceded formal 
 numeration to the higher. Numher begins at ' two,' and we 
 may assume — without venturing far into the 'metaphysics 
 of language' — that 2 was the first named numeral, though 
 an earlier conception may be expressed in the name given to 1. 
 Considering that eve.'y decimal system is in fact a doubled 
 quinary, and was cou^a-ucted with as constant reference to 
 
 the way of provinp — or disproving — the 'primfival unity' of American speech, 
 on no better evidence tlian is iillbrded by l)rief and ofieii inaccur.ite vocaliularies. 
 In wikchemna (discarding tlie prefixed particle), (jlluha, and Lhrapmi, tiie sanio 
 name appears uiidc three dialectical variations: kclie-vma = (ilhe-ba = kh'iapnn. 
 And the results of 'laziness' and 'emphasis' arc so nearly balanced that — tried 
 by the Indo-European standard — it would Ite hard to say which of the three forms 
 best represents the primitive roots. 
 
 "The Kev. A. L. Uiggs, MS. The derivation he suggests I'or wikch'emiia is 
 "from IV, the sign of the abstract form, ikche 'in a commou manner,' and mna 
 'gathered together.'" 
 
 ^ksha 'bent,' yuksha' 'to bend, to fold, to double'; kcha 'straight,' 'loose' 
 (un-heni), yu-kcha 'to untie, to loose,' etc. yu-kcha" 'to undcrstiii d, to compre- 
 hend' (i.e. to straighten out'!). 
 
 ''If the Ilidatsa pllaka stood alone — the more probable derivation would be 
 from ipi 'extremity, end,' as in ipila 'at the rear, behind,' and ipilakoa 'at the 
 end'; which last might have I)ccn contracted to piUika. But the meaning of the 
 name in other Dakota dialects — 'unbent' — favors pliki, notwithstanding the 
 change in accent. 
 

 •alf/ht)"ncd, 
 
 'chiiirtia is 
 
 out,' witli 
 
 Uidiilsr. 
 
 I clothes," 
 kUi (, iVom 
 <; ^ Both 
 n noitlicr 
 
 II uiiiy he 
 of a mode 
 I') reasons 
 two,' and 
 iko) were 
 st definite 
 
 and tliat 
 ed formal 
 •,' and we 
 etaphysics 
 il, though 
 ;iven to 1. 
 1 douhled 
 ferencc to 
 
 irieiin s])ecch, 
 vocalmlarics. 
 mn, tliu sanio 
 = kh'iapun. 
 . that — tried 
 le three forms 
 
 wikch'tmna is 
 ler,' and mna 
 
 gilt,' 'loose' 
 i, to compre- 
 
 ion would be 
 
 lakoa ' at the 
 eaning of the 
 standing the 
 
 On Numerah hi Amtriean Indian lAinipv M 
 
 the numlier of the handn as of the fingers, inw mtioii K 
 ptiirs would seem to 1)0 a natural expedient for \ u\i to tho 
 higher ininibcrs. 
 
 In various North American languages of the West and 
 Southwest, wo find ' fours ' fornietl from ' twos,' ' eights ' 
 from ' fours,' and, more rarely, ' sixes' and even ' nines' from 
 'threes.' East of the Rocky Mountains, traces of similar 
 numeration are uncommon. The Dakota topa 4=2 pairs, 
 has been mentioned (p. 2-5)- The Catawl»a (North Carolina) 
 purre-purra 4, apparently comes, by reduplication, from na- 
 pirra 2 ; l)ut both may have been derived from a common 
 root, found also in dn pimna 1, pukte-arra 5, and dipk-urra 
 6. In the (Algonkin) Cree, one of tlie two names for 8 
 is ayendneu, which seems to be a 'doulile 4' (see p. 26, 
 ante); and in the scmi-Algonkin Sliyenne, hJch is 1, tvtJ/m 
 ' a pair ' ; rticii 2, eniCH-anst ' 2 pairs,' ni-nish-ii<h' ' you two ' ; 
 na'a, nd 3, e-na-hanHt ' a pair of threes,' ' o pairs' (Haydon). 
 
 In the Athabascan family, Buschmann's comi)arison of the 
 numerals in twelve languages gives these results : 6 has an 
 independent name in six languages and in six others is formed 
 as 2x3 or 8x2; 8 is expressed as 4x2 in eight languages, 
 and 9 is formed on tho 3 in only one.^ 
 
 For example, in the northern Athabascan, Howse's vocabu- 
 laries^ give — 
 
 6, eike Uihey. 
 
 8, ellkee dinghe (also narki/ah-alitah = 2 less). 
 
 8, enchet'hentir (2X4)- 
 
 6, enchet'hiUy. 
 
 In the southern branch of this family, the same system 
 may be found, though less distinctly marked : 
 
 Navajo 3, t'ha, 6, has-tdr, 9, nas-tai'. 
 
 In another family, the Shoshoni (classed by Buschmanu 
 with the Sonora), doublets and triplets are common : 
 
 Comanche 3, pa-hist, 6, dyoh-pafist. 
 
 Chemelmevi 3, pai, 6, na-bai. 
 
 2, wait, 4, wat-chu'. 
 
 ^ Worttafel d. Athapask. Sprachstamms (3te Abth. des Apache), §114, n. 2. 
 * Proceedings of the I'hilological Society (London, 1850), iv. 192 ff. 
 
 Chepewyan 3, tahhee, 
 4, dinghte, 
 
 Biber 2, omihaly, 
 
 4, tenter, 
 3, idhtir, 
 
 ( 
 
 -'=-*«ww»«g*i**-ja*«'S'i 
 
 i.''^M»rt*>-»'"^^ 
 
J. H. Tnwi/.ii/t, 
 
 Slioshoni 2, wit, 4, mit suit. 
 
 Ciihiiillo 2, mfwi', 4, iiifirirhu. 
 
 Ki7.li '2.hii,he, i.hii.iha, », hiifhfthhiiatta. ' 
 
 In one Yumiv dialect, the Crclian, we have 
 
 ' n,l,iimimk\ (ijmm/iook', , hitm-himnok ; 
 
 thoufjh ill tlic Mojave, of the same group, the 0, 7, and 8 are 
 re^MiIarly Ibrmcd as 1, '2, 8 of tlic second hand. 
 
 The nunioral system of the Arikaras is peculiar, and 
 deserves special notice. Tlie Arikaras, or ' Rees ' as thoy 
 are called by the French traders, were originally the same 
 people as the Pawnees of the Platto River, their language 
 being nearly the same.' 
 
 The first five Pawnee and Arikara numerals correspond 
 nearly. From (> to 10, tin; Pawnees proceed in the more 
 common mode, by repeating 1, 2, and 8, as 'added' to 5, or 
 'of the second hand,' and naming 9 as 'less than 10.' The 
 Arikaras named 8 fi'om i', (|,y prefixing a particle), and the 
 odd numl)ers 7 and by a diminutive suffix to the name of 
 the next higher even number: thus, 
 
 6, shapis 8, tiip-slia'pia 10, niikh-ini 
 
 7, tiipslia'pistmn 9, mikh-ini-wan 
 
 And so with occasional variations, numeration [)roceeds to 
 20, which is 'a man' — for the system is vigesimal; 12 is 
 2 + 10 ; 11 is (2 + 10) viinus; 18, ndkugii-ivan, is ' lesi thf -. ' 
 
 14, ndkugW, which, again, .seems to have been formed from 
 
 15, akh'hxjlt'u ( = aM'u glt'u 'the whole foot'). In the 
 next quinate the names all come from tiie 20, iri-tau' (wita 
 'a man'), those of 16 and 18 being the less composite and 
 probably the older : 
 
 20, wlttiif 
 
 1 9, wUmi-akhko-kiiki 
 
 18, witaw-an 16, wUiilch' 
 
 17, wllutchiskiigit. 
 
 The 19 is literally ' man one-not.' Dr. Hayden's vocabulary 
 gives the numerals as high as 1000, and similar derivation of 
 
 ^Tit. F. V. Hayden's " Contriliutions to the Ethnology and Philology of 
 Indian Tribes of the Missouri Vnlley " (Philadelphia, 1862), p. 351. His 
 Ariknra vocabulary is the best and largest yet published. For the Pawnee 
 numerals, I use bi.s " Notes on the Pawnee (and other) Languages," in Proc. 
 Am. Philos. Society, vol. x. (1868), pp. 389 ff.; a.id for the Arikara, have 
 compared Prince Maximilian Wied-Neuwiod's vocabulary (Reise, T. ii. s. 465flF.), 
 and that of Geo. Catlin, in " Letters and Notes on the N. A. Indians," ii. 262. 
 
On Nurntrnls in American Lulliin Laugii<ige», 
 
 36 
 
 )mlKi, 
 
 and 8 aro 
 
 iliar, and 
 
 as tlioy 
 
 tl>c satno 
 
 language 
 
 iiTcapond 
 the more 
 ' to 5, Ol- 
 io.' Tlie 
 , and tlio 
 name of 
 
 )ceeds to 
 vl; 12 is 
 
 381 thf>'-. ' 
 
 ned from 
 
 In the 
 
 im' (wita 
 
 )sitG and 
 
 •git. 
 
 cabulaiy 
 vatioii of 
 
 liilology of 
 351. His 
 ho Pawnee 
 ," in Proc. 
 kara, have 
 . 8. 465 ff.), 
 ' ii. 262. 
 
 C' 
 
 lower from liiglicr nnmltcrs is ol)serval»lc tlirongliout, combined 
 with the common e.xpodicnts of vigesimal notation: '• 
 
 30, itain'ii («'ii//, Maxim.) M), iiil'iknuniu'' 2 |iers()iis 
 :i'i, w'ltiiHiiitihokh'ini- 30 + \'2 :w, iiiliLiiiianii-wnii 40 — 
 
 31, wilHHjiilibmriLh'mi-ivan^{'20 + \'i) - 39, /'(>//. iih((»« (iWio^'i/i -- 40, 1 not 
 
 100 is ' 6 men,' 98 is ' 5 men minus,'' and IMJ, '5 men, 1 not'; 
 and so on. 
 
 I will not add to the length of this paper l>y pointing out 
 its shortcomingH. It is olTered not as a contribution to 
 American lingnistics, but with the purpoHC of .sliowing, by 
 examples taken from u few families of American speech, that 
 it is unsafe to assume uniformity in the conception or the 
 expression of numbers, even in dialects of the same language, 
 much less in languages whose attinity is not yi-t proved ; and 
 that it is ccpially imsafe to assume that the 'hand' or 'linger' 
 always gives its own name to the number it serves to mark in 
 digital numeration — in other words, that 'two' must = 
 'hands' or 'fingers,' and 'five' or • ten ' = ' hand ' ; tliat 
 although a general corresimiulence of numeral scries in two 
 languages may justify the infen nee that both came from 
 one stock, yet no evidence of such affinity is presented by 
 occasional coincidences between single numerals in different 
 languages or between the name of any number in one 
 language and that of the 'hand' or 'finger' from which in 
 another that name might have been derived; but that the 
 value of such coincidences must depend on the analysis of the 
 names and the ascertained meaiung of their components or 
 roots. I have thought it not impossible that, from a field as 
 yet almost unworked, some of the results obtained in even so 
 partial a survey might interest comparative philologists, as 
 bearing on the question of the origin of ideas of number and 
 the beginnings of the art of counting — antecedent to digital 
 numeration. 
 
 The comparison of only a few dialects is sufficient to prove 
 that the process of mental development in the apprehension 
 of numbers has not been uniform. The Algonkin Indian and 
 the Arikara have not taken the same way from the primary 
 conception of number to the full decimal system. It is 
 
.01^*:"' 
 
 --% if<*^^, 
 
 , . *H~ r. \ 
 
 IMi 
 
 80 
 
 /. H. Trumbull. 
 
 equally evident, that one ti-ihe may have advanced further 
 tliiui another hrfore resorting to finger-counting or estal)lishlng 
 a regular sctiuencc of earlier-acquired conceptions of number. 
 The priority of the conception of 'one' to that of 'two,' or of 
 'three' to 'four' — or of the vocal expression of either 
 conception — is not determined by priority in the numeral 
 series. To one tribe, progression by 2}ciirs may have seemed 
 as nattiral as progression by units does to those of higher 
 culture; and the result would be a system — partially 
 represented by the Arikara — in which the even numbers 
 were the earlier named, and the odd numbers intercalated, 
 just as differences by halves or other fractional parts might be 
 intercalated in the Indo-European decimal system. The pre- 
 digital numerals so formed might include the 4, the natural 
 order being 
 
 2 1 4 t{ 
 
 that is : 
 
 a pair, less, 2 pairs, between (2 and 2 x 2). 
 Or it might stop at the 3, as trans 2. No evidence is found 
 that any tribe has advanced beyond 4 without digital 
 nuiiieration, and there are few numeral systems in which 
 some reference to the hand or the fingers may not be detected 
 in the name either of 3 or of 4. But when 3 = ' middle,' 
 ' between,' or 'half-way' — as in the Algonkin languages — 
 it is not possible to decide whether this meaning comes 
 directly from the 'middle finger' (half-way to 6), or from 
 position between ' pair ' and ' pair of pairs,' i. e. between 2 
 and 4. 
 
 i { 
 
1(1 further 
 itablishing 
 )f number. 
 two,' or of 
 of either 
 J numeral 
 ^e seemed 
 of higher 
 - partially 
 
 numbers 
 orcalated, 
 1 might be 
 
 Tlie pre- 
 16 natural 
 
 id 2x2). 
 s is found 
 It digital 
 in which 
 J detected 
 ' middle,' 
 »uages — 
 ig comes 
 or from 
 etween 2 
 
r%' 
 
 **.' 
 
 
 ^>P*-' 
 
 -JtaJSfr