SMITIISONIAN I .NSTITUTION

BUREAU OF AMERICAN ETHNOLOGY

Linguistic Classification of

Cree and Montagnais-Naskapi

Dialects

By TRUMAN MICHELSON J

Chapter froD1 BuLI~ETJN 123, Bureau of American Et:hnoJogy

Anthropological Papers, No. 8

UNITED STATES

GOVERNMENT PRINTING OFFICE

WASHINGTON : 1939

1989 ·.t

LINGUISTIC CLASSIFICATION OF CREE AND MONTAGNAIS-NASKAPI DIALECTS

By TRUMAN MICHELSON

· In 1912 I had an opportunity to study the Cree of Fort Totten (North Dakota), and in 1920 had a chance to study the Cree of Files Hill, Saskatchewan, Canada. In 1923 I observed the Montagnais of Lake St. John and Lake Mistassini at Pointe Bleu, Quebec. In 1924 at· the Northwest River I studied the dialect of Davis Inlet fro1n a11

Indian there, and gained a · little kn.owlerlge of the dialect of the Northwest River. The American Council of l...,earned Societies made it possible for me in the summer and early fall of 1.935 to do field­work among some of the Algonquian Indians in the vicinity of James and Hudson's Bay. I visited Moose Factory, Rupert's House, Fort George, and the Great Wl1ale River. However, I was able to do a little work on the Albany Cree and Ojibwa owing- to their presence at Moose Factory; and I did a few minutes work with an East Main Indian whom I stumbled across at Rupert's House; shrularly I worked for a, few 1ninutes on the 'Vee.nusk dialect as an Indian from there chanced to co1ne to Moosonee at the foot of James Bay. Owing to a grant-in-aid made by the American ColUl­cil of Lear11ed Societies it 'vas possible for me. to again visit the ,,James and Hudson's Bays region in the spring, summer, and early fall of 1936. The results of the previous expedition were checked up as 1nuch as possible and additional data ga.tl1.ered. I visited Moose Factory, Fort George, Attawapiskat, and Weenusk; but it was possible to get data at first hand as well as by correspondence on Rupert's House, and .first-hand information on tl1e Albany Cree, owing to the presence at Moose Factory of some Indians who came from the Albany River during n~y stay at ~1oose Factory; data on Lac la Ronge was obtained at Moose Factory from an Indian who had just come from there; data on the Ghost River (Chepy River) were obtained from the Hudson's Ba.y Co.'s post n"lanag~r 'vhom I met at Moose Factory; data on Trout I....~ake were obtained from a missionary at Weenusk. By correspondence wjth Hudson's Bay Oo. post managers and some missionaries data on the Cree dia-

69

70 BUREAU OF A:\IERICAN I~THNOLOGY £BuLL.12a

lects of Cmnberland House, Norway House, Oxford House, God's Lake, Island Lake, Montreal Lake, Stanley, and Pelecan Narrows "\Vere gathered. Prof. John M. Cooper of the Catholic University of America has ge11erously supplied some fresh data ou Tete de Boule from his field work in 1937, as has Prof. Jatn.es Geary of the same institution on Algonquin proper from his trip among them in 1937. The Atnerican Council of Learned Societies again made a generous grant-in-aid and tl1us enabled tne to do field-work in the summer of 1937 among the Montagnais-Na.skapi Indians of the northern shore of the St. Lawrence River. 1 was at Natashquan, Seven Islands, Moisie, and Bersimis, but was able by personal contact to get data. also on St. ~Iarguerite, Godbout, Shelterb·ay, and Sheldrake. By good luck I met an lndia11 at SeYen Islands ""~ho had just come from the northeastern corner of Lake Kaniapiskau in the heart of tl1:e Labrador peninsula; and also met an Indian from Davis Inlet on the northern Labrador shore. From a study of this material as well as sotne contained in documentary sources, it: follows that the state­ments n1.ade by me previously in reporting my first. expedition to the James and Hudson's Bay region are sustained.1 t It can not be too strongly emphasized tl:tat. east of Hannah Bay Cree leaves off and Montagnais-Naskapi begins. Mistassini, Waswanipi~ Rupert's House, East Main, Nichigun, Fort George and the Great Whale River bands are y-dialects (i. e., dialects in which original l is replaced by y) of Montagnais-Naskapi to 'vhich the dialect spoken at the northeast corner of La.ke Kaniapiskau n1.ay now be added. That the Fort George and Great Whale River bands distinctly form. a subgroup within tl1is larger one is confirmed.2 Similarly, not,vitl1stauding statements to the contrary, Tete de Boule is a Cree dialect in which original l is replaced by r. According to Prof. John M. Cooper there is also an l-dialect spoken there; this is confirmed by Mr. Frankland, post manager of the Hudson's Bay Co. at Seven Islands, but formerly at Obidjuan, as well as by a statement of Joseph Kurtness, a Mistas­sini Indian whom I met at Lake St. John in 1923.8 However, witltin historic titnes there has been a migration from the neighborhood of the Albany River; so that the present location of this l Cree dialect may be recent. The accompanying map (fig. 12) shows my latest information regarding the l and n dialects in the neighborhood of the Albany River. As explained previously, Romanists tlse the n­dialect and Anglicans the l-dialect. However, it may be noted that the Roman Catholic "Catechisme" of 1854 is essentially in the "t­dialect. I do not know if Mrs. Corcoran, tl1e wife of an officer of the Hudson's Bay Co. at the Albany post, assisted in this particular work or not, nor do I know her exact linguistic affiliations. Some

t S e e Appendix, p. 86, for footnotes.

MlCHIDLSON] CH.EE AND MONTAGNAIS-NA.RKAPI DIAI .. ECTS 71

other Roman Catholic works of the early period and even a little later are also in this 7-dialect. The "Englisl1" river of my previous re­ports for a long ti1ne (I do not kno'\v tl1e precise date) has been called the "Churchill River" and is not tlte present "'English River." " At Lac La Ronge a th-dialect (one in which th replaces originall) oc­curs. Th.e new data sl1o'v that the Cree dialects at Cumberland House, Norway House, Oxford Ilouse, Trout Lake, and God's Lake are also n-dialects.IS It may be n1.entioned that tl'le 1~-dialects of tl'le Albany River, Attawapiskat and Weenusk, are all closely related; there are, however, some differences in idiotn and vocabulary. I have not sufficient data to know how closely tl1.e "ne,v" dialects re­semble them. At 1\Iontreal Lake, Stanley, and Pelican Narrows the Cree dialects are y-dialects, i. e., dialects in which. original l is re­placed by y. The dialect at Island Lake is apparently mixed Cree and Ojibwa and th.e proportion of mixture is said to be high. How­ever Ojibwu ( -Algonkin) can also be sh.o"'"n in varying degrees in a number of Cree dialects (e. g., at Weenusk Ojibwa ciink 9 is in current use for ciihk; so too in the Moose Factory dialect there are some such influences both lexically and grammatically). Accord­ing to information furnished by Prof. Jol1.n ~1 .. Cooper, the Tete d(~ Boule Cree dialect shows a few lexieal and gra1nmatical borrowings from Algonkin proper. Conversely, Algonkin proper sh.ows some distinct lexical borrowings from Cree, as proved by Pro:f. James Geary's notes which he has kindly shown me. Also, some Plains Ojibwa material published by the late Alanson Skinner sl1ows un­deniable lexical borro,vings fron1. Cree.6 Despite some published contradictions;' the area designa.ted on the n1.ap is the n.-diulect area. of Afontagnais-Naskapi. In all these djalects o£ which I have any independent kno,vledge terminally -to ( -t8) appears as -t. 'Vl1ereas the extreme southeastern dialects are sharply set off, going west of Natashkwan the intervening dialects beginning with Mingan and continuing perhaps as far as Godbout (Bersimis certainly is a new a.rea) are mixed, l and n fol"'llS both occurring> as well as -t from -ta ( -ts) ; but the treatment of medial -slci- agrees with that of the Bersimis dialect (and I..~ake St. Jol1.1'1) which is a clear-cut l-dialect with certain features of its own; as do son'le otl1er features. Natives feel that ~Iingan to Bersilnis is a linguistic unit; and all things considered I agree.

Examples of sucl1 mixtures are nil "I," n.amll1.veltl}n "I like it,"· but ntcenim "my father-in-law"-all at ~{oisie and from the same informant (percontra 11-tcelin~ at Godbout, which is historically justi­fied); or legu "sand," •ni'n, ni'l "I," tc'i/n tci'l "thou," na'p~wut "men," ka'te.peltak "he who controls it" (the designation of the all-higlt

t See Appendix. p. 86, for footnotE's.

·--· Ov'/8}1/A -AL(JON()VIN \

FIGUU 12.-Dlstribution and interrelations of the Cree and Montagnal.s-Naskapl dialects.

}EiiCHliJLSON] OREE AND MONTAGNAIS-NASKA.PI DIALEdTS 73

god)-all from a single informant of Shelter Bay (ko,'tipeldahlc, Godbout). Observe a moderately consistent ·n.-speaker at Seven Islands gave la'lamisut ''thunder'' with false l in both cases (false because not historically justified; 11or does it occur in the Lake St. John l-dialect) . In this connection it should be noted that all the Montagnais-Naskapi l-dialects as well as the mixed n-group of which I have any knowledge, have variants of kassinu "all'' which obvi­ously is a transformation of an early loan from then-dialects (Plains Cree kahkiyaw: which shows that *kaasilru should be expected). All this goes on to show, if taken in conjunction with what has been said above, that a strictly genealogical classification of Algonquian languages breaks down. Nevertheless the accompan.ying map shows the essential distribution and interrelations of the Cree and Mon­tagnais-Naskapi dialects. In this connectio11 it should be added that although it is commonly supposed that Cree is always more archaic than Mo11tagnais-N askapi, actually this is not so; it is true that Cree in many cases is more archaic, but Cree is not universally so; it follows, therefore, that neither is derived from tla.e other, but both have so mucl1 in comm.on that they both must come essentially from a single source within the Algonquian stock. True transition­dialects between Cree and ~{ontagnais-Naskapi are unknown to me, if they actually exist. 'I'he boundaries seem very sharp outside of a. possible few and altogether insignificant cases where either direct or indirect speech mixture is plausible. Characteristic of all Mon­tagnais-Naskapi dialects is the palatization of k, some vocalic har­mony, the weakening of some vo"\vels and diphthongs, etc. The Montagnais of LeJeune (in the Jesuit Relations) shows mixture; and some "\Vords, and even 011e w l1ole sentence, instead of being Montagnais is Algonkin proper. It should be borne in mind that khi (and some variants) actually designate a sound usually transcribed by tc (which in Montagnais-Naskapi may be primary or secondary): without this knowledge we should be obliged to assume some words were really Cree (Tete de Boule~) which would mean a large shift of population had taken place. This hypothesis is probably unnecessary.8

This paper throughout presupposes knowledge of Bloomfield's remarkable paper on the sound system of Central Algonquin (Language, I, pp. 130-156, 1925) and the literature therein cited. Attention may also be called to my ''Prelimjnary Classification, etc .. " (Twenty-eighth Ann. Rept. Bur. Amer. Ethn.) and to other scattered papers, mostly in the American Anthropologist and International Journal of American Linguistics.

t See Appendix, p. 86, for footnotes.

BUREAU Oli"' Al\1EHIC.AN ETIINOLOGY [BULL. 123

Most of the pl"lonetic shifts of Cree (and hence largely 1\iontagnais­Naskapi) are known, and n1 their final pron1.ulgation are due to Bloomfield. It should be mentioned that a few are not and that these affect Montagnais-Naskapi as "~ell as Cree. Thus the laws of shortening long vowels are unla.lo"\vn, e. g., Cree pipun "winter'' has 'tJ, where o is to be expected, and so in all Montagnais-Naskapi dialects of whicl1 I have any knowledge. So also tl1e u for i in Cree ni1nusu7n [*ni'l'nlso,Jn expected] "my gi."andfather'' (vocalic as­similation and shortening; cf. Fox ne'1"1'te' con~e' s.A) and its corre­~pondents in 1\tlontagnais-Naskapi. Also tl1ere have been extensive analogical levelings common to Cree and Montagnais-Naskapi. Thus, the original verbs in -ii -a (kept in E"'ox) i.J.1. the paradigms have been leveled to -a; and the -ii (Cree -e) verbs leveled to -a in the first persons and second persons of the indicative. Again, verbal stems containing original ii and o in the first syllable show "change" ill participles, etc., as do also at least Ojibwa and Algonkin (Sauk, Fox, and Kickapoo do not; the "change" of *a to *ii, and *u to

· *wii is proto-Algonquian in any event). It is likely that the dif­ference of the vocalism in the "change" of Cree (M-N also) *a and Ojibwa-Algonkin is due to a phonetic sl1ift (see my discussion of Cree kiyii.sk "gull" above), but the "change" of o in Cree (and M - N) analogical. There are also some lexical traits peculiar to Cree and Montagnais-Naskapi. Thus, correspondents to Plains Cree peyalc "one" exist all over the area; so too correspondents to Plains Cree '1'nahkesiw "fox" evidently did at one time, yet at present Ojibwa­Algonkin wiiguc has replaced the 'vord in one Tete de Boule dialect. Noteworthy is Plains Cree n.ohkumis which means "my paternal uncle"; corresponds to this \Vitl"l this meaning (extended in some Montagnais dialects to n1.ean. "my maternal uncle") exist i11 all Cree and Montagnais-Naskapi dialects of which I have any knowl­edge. [Historically the worJ. should n1ean "grandmother."] I do not think that correspondents to Plain Cree nisti1n with the value of "my cross-niece'' are universal in the Cree area but they surely must he nearly so. Historically it should mean "my daughter­in-law" and it is used with tl1is value over a "\vide area. In a few Cree and Montagnais-Naskapi dialects it is replaced by a variant for "my daughter-in-law." Similarly the term (Plains Cree) nitihkwitim "my cross-nephew" structurally is peculiar to Cree and Montagnais-Naskapi; the distribution favors the asun1ption that it once was universal; but it is not now; I do not know whether wher­ever correspondents to this occur they can also be used with the sense "my son-in-law" but they certainly can in some dialects of Cree and Naskapi. I have been asked to give at least a word that i11

phonemic (not. phonetic) transcription is characteristic of the

Ml<"BELSON] CH.EE AND ~IONTAGNAIS-NASKAPI DIALEC'T~ 75

\\hole area, and occurs in no otl1er Algonquian language or dialect. This is 11ot an easy thing to do (per co11.tra it is easy to cite char­acteristic words of ~1:enon1ini which occur nowh.ere else, e. g., "i.l'el"llttew

''he, she n .atnes him, h.er ," ko' new "he, sl"le fears him, her" [the last is preferable because n.n almost identic "vord for tl1.e first occurs in a Cree dialect], Sauk, Fox, Kicka.poo, e. g., kinwawa [ph.onemic o rthogra.phy] and Peoria-1v"Iiami, e. g., k·"i1s·wa "sun'' but I think in phonemic transcription nitiihkusi·n "I am sick" is universal. On the other hand it is extremely easy to cite quantities of words which 1nust be Cree ::u1d nothing else and which occur i11. every single- Cree dialect as far as lmo'\vn. Such words in phonemic transcription are nikiwiin "I return," pihtukew "he enters," it·uhtewak "they go thither," p ·eyak "one," ni.stu "three," nipiltk "in the '\Vater," uski·niki-1cak "young men," p·ipu.n "winter," niwiwihkisten "I like the taste of it,'' -rn-i58k·"istawewak "they rush on them," niwiipam4wak ''I see t hen1 an.," itwev;al~ "they said," ekitirniikisiyan "wl1en you were in n tise.ry"; etc. If '\Ve extend this list to include also words wh.ich occur in identic form also in some though not all Montagnais­Naskapi dialects of course it will be much i11.creased. It is easy to i uere.ase the list of 'Yords of identic form which occur in several (~r~e dialects but not all are included (e. g., 'mahkesi1.o "wolf"). At this point it should be pointed out that witl1 the present material it is not possible to duplicate or approach tl1e work that has been done on some European languages and 1nore recently on. A1nerican Eng­liRh either as regards pl1onetic, morpl1ological, or syntactical dif­f~:-ences; or distribution of word8. A single person can not even accumulate the necessary rnaterials, to say 11.oth.ing of interpreting it. It goes without saying the published missionary dictionaries of l ~lains Cree and gra.m1nars, as well as Bloomfield's Plains Cree 1exts~ and the Dict.jonaire J.i""'ra.nca.is-~{ontagnais ..... et Grammaire lVIontagnaise by Geo. Le1noine and 1\{oiltagnais sans Maitre by Luc Sirois (botl1 essentially concerned with the 7-dialects; the latter deals specifically with the Bersimis dialect) have 1naterially light­ened my task. "The simplist classification of tlte Cree dialects is based upon the transformations of original *l sl1own on the map. The transforinations of the original *l in Cree and 1\fontgnais­N askapi are 'vholly independent; and sin1.ilar independent changes have occnt-red else.,vh~re (e. g., the transforn1ation of *l to n; as long as tl1is change occurs in Sauk, Fox, l(.ickapoo, Me110n1.ini, Ojibwa, Ottawa~ Potai\Vatoini, and n1.odern Algonkin which are all geographieally contiguous it is not likely that in tl1is group the ch.ange is independent but has spread).. o,ving to lack of data it is quite in1possible to tell with certainty as to wl1ether the change to r in Isle a, la Cross, Kesaga1ni ( whicl1 is virtually extinct), and

7711Rs-39--2

76 BURF~AU OF A.~IERTCAN ETHNOLOGY [BULl.~. 128

Tete de Boule Cree is independent or not. A good "key" word for the l-dialects in the neighborhood of Moose Factory is kilawiiw "ye." F,or Moose Factory niyala:l "five" is good because nf,yii-nal occurs near the Albany River. For "key" words in the other Cree dialects corre­spondents to Moose Cree kilawiiw are better than to Moose Cree kila "thou" (which is not a good key word as in phonemic tran­scription it occurs in Shawnee and Miami-Peoria [unless Voegelin is correct in writing Shawnee ii for i]) because kina in phonemic transcription also occurs in Sauk, Fox, and Kickapoo. [Similarly w"ila, nina, are best avoided; of course niya is not good; nifJa is good enough, but kiOawiiw will conform to the other "key" words.] A clas­sification based on only a single feature is insufficient, but it really works reasonably well in the present case because exhaustive diction­aries and grammars of every single dialect do not exist. To make up the deficiency partially a few 11otes are appended. The Cree dialect at Turtle Mountain (North Dakota) is a y-dialect; in my vocabulary of over 25 years ago I recorded Ot in place of ht ( wiipa8tarn for wiipahtam, "he sees it"). A single speaker gave a. sentence wiipa'llWw utiinisan "he, she sees his, her own daughter" which is obviously mixed Cree and Ojibwa. According to some published and unpub­lisl1ed sources in tl1e th-dialect tni8ku "blood" occurs in place of -mihku (Fox 7netckwi phonetically). The nun1eral 6 for nikutwiisik occurs in both Plains Cree (y-dialect) and th-dialect which in Fox is negutwiicika (phonemic transcription). The 'vord is wanting in the n-dialects of Attawapiskat and Wenusk as well as the l-dialect of Moose Factory, Swiisik "eight" has a correspondent in Fox cwiicika; swiisik is said to be the old word at Weenusk but is replaced today by niyiiniinew, and variations of tllis occur in Plains Cree, Atta,vapiRkat, and Moose Factory. (Plains Cree ayenan.e1_~_,, Moose Factory yiiniinew.) The n-dialects of Albany, Atta"":rapiskat, and Weenusk are very close to each other. Lexical differences (mostly names for articles of European origin) between th.e Albany and Weenusk dia­lects occur, such as (in phonetic, not phonemic transcription). A k,wa'pahigan "cup, basin,"W. -m/inihkwa/ ga11./ A. kapa' htciw(z:n"lid of a kettle," W. kipa' higan; A. cigahun "comb," W. pi'niihkwan; A. tiipick­wii' sun "spool of thread," W. se' stag; A. pohtenegan "thi'lnble," W. ka'slt!igwii'su'ltiipisk/ A. m.oci'twawin "scissors," W. tnii'teigan,· A. cal "shawl," W. agu'niwin; ~~- Ttdagap "dress," W. niskutiigai (obscene and obsolescent; the Albany word is now in common use; A. pa'la.tci',, "trousers," W. m-itii' s (the old "rord for "leggin," the Albany word is a corruption of English breeches; payatcis in at least so1ne Plains Cree dialects), A. pa' guyiin "shirt,'' W. pa' guyiinisiigai/ A.. sipuste­wigan ''sweater,'' W. sipegiskiiwiyan. Though the old word for "nine" ciihk (Fox ciig..t in Jones' transcription) is known at Weenusk

CllEE ..:\.ND 1\IONTAGNAIS-NASKAPI DIALECTS 77

it is ordinarily replaced by Ojib,va cling (incidentally Baraga's jangasswi is 1nade after the analogy of midass·wi "ten"). The Plains Cree dialects (y-dialects) evidently have Jnostly kekiit 'lnitiitaht. Ac­cording to Horden, this last is also known at Moose Factory. All extension of aahk also occurs according to the same authority (shaketat) '""ho also gives a for1n which can be restored in phonemic transcription as JJeyal.:uste1.o 'vhich. has exact equivalents in. at least the follo,ving 1\-Iontagnais-N asknpi dialects: Lake St. John, Bersi.mis, Seven Islands, 1\fingan, Natashquan, Davis Inlet, north.east corner of I-.~ake J(aniapiskau, 1\fistassini, Rupert's House, Fort George, and the Great \¥hale River. Clearly the word is 1)rimarily 1\Iontagnais­Naskapi and has n~a.de its way iuto ~loose Cree. The numerals for "six" and "seven" at Weeilusk (phonetically ngutwac, nicwiic respec­tively) differ co1npletely from the corresponding nu1nerals in Plains Cree, and agree si ructurally in .1.\tloosc Cree, Atta.wapiskat, and the follo,ving 1\tlontagnais-N ask a pi dialects: Rupert's House, Fort Ge,orge, Great Whale River, Mistassini, Lake St. John, Bersimis, Seven Islands, ~finga.n, northeast corner of Lake Kaniapiskau, Nata­shqua.n, Davis Inlet, Fort George, Great Wl1ale River (allowing for })honetie differences). [According to Horden Moose Cree also pos­sesses for the Plains Cree numeral for "seve11," tepakuhp / to judge from J. Howse, A Grammar of the Cree Language, etc., besides the equiYalent of this, the th-dialect also uses (approximately) nis·wasik.] Clearly the Ojib,va and Algonkin correspondents are based on close if not precise equivalents. So fro1n th.e general distribution with the means at our disposal, it can be inferred tl1at a spread westward has occurred. Note, too, the llUineral for "eight" has a similar structure at Lake St. John, Bersimis, Seven Islands, Mingan, Natashquan, and Davis Inlet. [Again, Ojibwa and Algonl{:in have close but not exact equivalents; in my opinjon the termination -1-oi has spread from the numeral for "ten.".] Here is an "isogloss" between major divisions of tl1e Montag11ais-Naskapi dialec~s ..

In view of these distributions it. is certa.inly peculiar that at Wee­nusk ntitiita' ciwag "they an. are ten" occurs, "\Yl1ich in phon.emic t.ra.nscription is mitiitasiwak \Yhich occurs in Plains Cree (y-dialect) [cf Menomini mitatahsiwAk, Fox meta'ciwAgi]. For the forms of Moose Cree given by Horden., A Gratnmar of the Cree Language, etc., London, 1881, on pp. 79, 80, Nekotwaohe-wuk ''They are six," NeswacAe-wuk "They are seven," Metache-w·uk "'Tl1ey are ten," are distinctly remini~cent. of Ojibwa and Algonkin rather than Cree (this type spread in Moose Cree). Moose Cree l1as some Ojibwa (Algonkin) loan-words, e. g., mihkiniihk "turtle," Plains Cree milJ .. kinalt..k. And it has at least one loan word from an n-Cree dialect, kahkinaw "all", if Horden is correct (Plains Cree kahlciyaw). Char-

78 BUREAU OF Al\II~RICAN ~JTHNOLOGY [BULL. 128

.a.cteristic of Moose Cree is mitau'n "completely" (nonnal Cree 'ln.itu/ni), and Wi8ahk1.vetciihk (nan"le of the culture hero; Weenusk, etc. Wisahketciihk). It should be mentioned that Weenusk Cree (an n-dialect) agrees with Moose Cree (l-dialect) in having a in such. for1ns as nipiispan [pl1onemic transcription] "he slept" (past subjunctive) wl"lere Plains Cree has h (nipiihpan); the latter seems reminiscent of Ojibwa. On the other h.and Moose and Weenusk Cree have th.e vocative plural in -tuk, Plains Cree -tik (the first agrees in Yocalism with Ojibwa, and Penobsc.ot; Fox agrees with Plains Cree); Len1.oine gives vocative pl. for Z-1\.{ontagnais as in -tuk. It may be noted tl1at wl1at Horden. designates as dubitative mood (Fox inter­rogative) in Moose Cree is living; in Plains Cree, to judge from 13-looinfield's texts this is not so. The common supposition that o ( ._f? h) for s is diagnostic especially of ~::loose Cree is erroneous. "The variation ii for ii occurs, but also elsewh.ere in the region of J ·ames and Hudson's Bays; tl1e exact distribution is unkno,vn. It should be added that not only the kinship system but actual kinsl1ip terms at Moose so coincides with tl1ose of other Cree and 1\Iontagnais-Nas­kapi dialects on both sides of the bays (excluding East Main and Rupert's I-Iouse on. both of wl1ich. I have too little data to utilize their evidence) that dissemination or at any rate influencing is clear. A special Moose feature is that in comparatively recent times nistiiw "n1y brother-in-law" with male speaker (and this is old as proved by co1nparative linguistics) has begun to be used (though not ex­<·lusively) "\vith the same 1neaning by female speakers. I repeat again that I an"l indebted to Prof. John M. Cooper for kind permission to use his Tete de Boule linguistic material. This Cree dialect is an 1'i-dialect ; tl1ere is also an l-dialect but it is uncertain whether this is the result of a comparatively recent im1nigration from the neigh­borhood of the Alha11Y River. The treat1nent of final *-wa in posi­tions in 'vhich as far as known it is lost in other Cree dialects is most peculiar. Son1.eti1nes it appears as -" (suggesting MontabTilais influ­ence or a parallelism), sometimes kept as _wa, sometimes -wa is lost. I do not kno"\'\" the rationale of this. Of course Amikwa, the name o:f an Algon.quian tribe found on the north shore of Lake Huron, Qpposite lVfanitoulin island, associated with. the Nipissing, is a par­allel for the second alternative, for in ordinary Ojibwa, Ottawa, and Algonkin dialects *-1.oa "\vould be lost in this position (Fox amehkwa, Cree a·misk, Ojib,va and Algonki11 ant.ihk in Bloomfield's transcrip­tion). Algonkin influence in n1orphology is sho'vn by -to being used in verbal-forn1s "\Yl1ere -t should be expected (e. g., rmireritam e 'toabamra'tc "l1e is glad to se.e him"), though apparently in some local­ities by some speakers -t fortns are used. Characteristic of Tete de B {H 1 1P is kli'8lt·ina "all" (I have a suspicion that this is a fusion. of

)t{ICHmLSON] CRT-til~ .AND MONTAGNAIS-NA.Sl<:API DI_-\ LECTS 79

two words for which Indo-European parallels abound). A peculiar umlaut is to be seen in nikiu. 'l.~ "otter" which comes fron:1 Proto­Algonquian *nekekwa (rigid proof of this can be given but is rather long and is so omitted; a careful study of cognates in other Algon­quian languages and. knowledge of the princip1es of Algonquian pho­nology 'viii convince any one that this is correct) and sin1.ilarly i11

ki' juk "sky" (Fox lc-i}ceg·w.A. in ~fichelson's transcription) "\\ .. hich to a certain extent is paralleled in J\Iontagnais-Naskapi. As long as kiiteito1.v~o;k (a supernatural monster) is kl1own as well as kiigito­wask, ~Iontagnais-Naskapi influence can not be denied; and in vie\\.,.. of this it may be questioned ~Thether or not such forms as at(/..:1"

"caribou" (ordinarily Cree atihk; [*atehlewa] h is sometimes on:titte<.l by Dr. Cooper "',.l1.ere it should be expected but never inserted where it does not belong Jtistorically; whetl1.er this is governed by phonetie shifts or is other·w·ise to be explained is uncertain) may not be dne to Montagnais-Naskapi influence). A. single informant stated that in northern Tete de Boule icpi1niki "above" occurred. This was contra­dicted by others. However, -k-i would be archaic (ordinary Cree i.spim-ihk, Ojibwa icpi'ln.ing, For.e Atpernegi, Proto-Algonquian *icpe­"lneiiki). Other similar for1ns '\Yith -i do not apparently exist in Dr. Cooper's n1.anuscripts; on the contrary similar forms lack the -i. The fact that the same infortnant cites rni' gis (a clamshell) in 'Vaswanipi only further n'luddles the situation, for linguistically the word vio­lates Montagnais-Naskapi phonology.

Cha-racteristic of all Montagnais-N ask a pi dialects is the change of k to to before original palatal vowels "'~hether these have been sub­sequently lost or not; the apparent exception k~ie "and" of Le1noine (l­dialect) and its correspondents in various dialects is due to the fact that the i of this word is unoriginal as shown by Cree kaye (siln­ilarly Ojibwa and Algonkin) ; the C'l1.ange of a to i in this word is clearly subsequent to tl1e shift of It! to tc before palatal vowels. [The combination *-sk- before palatal vo,vels appears in the l­and mixed n-dialects as -88- but prilnarily as -stc- (-etc) in the y­and n-dialects with so1ne subsequent changes.] The change of *-iw (earlier --i~wa) to 1-& after consonants is so universal (e. g., Rupert's House 11-inM!- "h.e dances,'' Cree ni·'Jni:r_o, Fox ni1ni?VA) that the very :fe'\v cases where it does not occur but a different change does, 1nust be due to special limiting conditions as yet unknown. Tl1.e change of *-wa to _u (voiceless 'lt), or to - 141 (labializatiou) after consonants (ex­cept after 111.- when it is lost) is characteristic of th.e area (an example is Great vVhale River ati' h.ku "caribou," most Cr(~e dialects atihk/ this obviously is more archaic than Cree; siinilarly the plural a'tihkwata js n1ore archaic than atihk1.vak because though final i is lost in both, ;yet in ~Iontagnais-Naskapi obviously it was kept until after it l1ad

80 BTJREATj OF ~-\.I\H•JRICAN ETHSOLOGY [BULL. 128

altered k to to [the animate plural termination in Proto-Algonquia.n was *-aki as shown by the evidence of Sauk, Fox, Kickapoo, Shaw­nee, ~fiami, Peoria; similarly the Proto-Algonguian locative ending *-eiilci appears in certainly most Cree dialects as -ihk, but -ihto or its transformation in Montagnais-Naskapi]. Such changes as seen in the final syllable of Bersimis tsimun "it rains" as compared with Cree kimi1.tJan are so universal (but Davis Inlet is archaic) that any devia­tions must be considered secondary. Vocalic assimilation caused by au in following syllable in certain cases is so widely and uniformly distributed that presumably they n1.ust go back to very early times (even Proto-Montagnais-Naskapi). Such cases are to be seen in Fort George iihtouk" "seal" (Cree iihkik for older *ahkikwa), ntau'k-. "otter" (Cree nikik, plural nikiwak), Davis Inlet n..awiipa.muk• "he sees me" (Cree niwapamik, Fox newiipJt.1negw .A) , Fort George nuk"stukuna'n "he fears us excl." (Cree nikustikuniin) [barring the labialization of the first syllable, for I do not know the distribution in tl1is case], etc. The assimilation seen in pup·un "winter" (Cree pipun) is also widely spread; yet in the Fort George dialect I have heard both the assimilated and unassimilated forms. Reduction of vowels, total loss, changing full-sounding vowels to voiceless ones, also are widely spread but seem to be rather different in different dia­.lects. The change of final *-iiw to -au and *-a1..v to -u evidently is old; t.here are apparently some secondary changes. I have pointed out in the 28th Ann. Rept. Bur. Amer. Ethnology that the Montagnais of Lemoine possesses some verbal-forms which have no counterpart in Cree but rather recall Fox. Owing to lack of data it is impossible to say whether these are universal or not. Lexically it may be poi11ted out that correspondents to Cree tciTniin "canoe" are lacking. As pointed out above, derivatives of *peya7~1.lstew ''nine" apparently is basic to Montagnais-Naskapi dialects and has spread to Moose Cree.

If we now turn to the major divisions of Montagnais-Naskapi dialects we may first take up the y-dialects. Besides y for original *l, correspondents to the Proto-Algonquian numerals for "five" and "ten" are diagnostic (*nyalan1.oi a11.d *TnetiitahOwi) [see my papers, The Proto-Algonquian Archetype of 'five,' The archetype of Fox meta'sui 'ten,' etc., Language IX, pp. 270-272, XI, p. 148], for they occur in no other Montagnais-Naskapi dialects as far as known. Also the numeral for "eight" in this group has no correspondents in the other major groups of Montagnais-Naskapi languages (Rupert's House nyii'naneu, northeast corner of Lake Kaniapiskau nyiinane'u, Fort George yen"'na·u', Great Whale River niyana'nau, Moose Cree yananew, Plains Cree ayeniine1.o, W eenusk Cree niyiiniinew l it should be noted that in the Fort George and Great Whale River dialects -Original *e of the Cree-Montagnais-Naskapi group (Proto-Algon-

.MICBBLSON] CREE AND .MO~T ... '-\.G.NAIS-~ASKA.Pl DIALECTS 81

quian *a) becomes a/ -e·zv becomes -a·u/ « and i is a favorite inter­change in tnany Montagnais-Naskapi dialects; but the Fort George word is otherwise unclear though clearly at least partially of the same structure as the others; the second edition of Horden's grammar gives Plains Cree ayiniinew [in Bloomfield's transcription]; the only explanation I can give is that Lacombe's transcription was partially used, for in this e=e '\vhereas in Borden's, etc., transcription e=i). It should be pointed out that in this group the words for "five" and "man, Indian," are based upon forms which sho'v the same assimila­tion as in Moose Cree nyalal "five", ililiw "person" (Prato-Algon­quian *nyiilanwi, ilen tva) ; these forms then suffer phonetic trans­formations (Rupert's House nyai'' "five," i'yu' "Indian"). The. words ni', tci', wi' "I, thou, he [she]" respectively (with the terminally glottal stop with which words ending in a vowel are usually cited) for Prato-Algonquian *nela, *kela, *wela are charac­teristic of the y-group, not occurring elsewhere. Also the word for "river" in this group characteristically is terminally like Cree sipi l according to Lemoine a form like this occurs at Lake St. John, else­where ahipu (as if we had Cree *sipow, Fox sipowi); I have recorded cipo' at Seven Islands and the equivalent at Davis Inlet. Also characteristic of the group is the retention of the diphthong in the word for "sand" (northeast corner of Lake Kaniapiskau ye' gatt, Fort George yii/ kau' ,- 1\tiingan ne' gu,- Davis Inlet neku' ,· Lemoine gives lekau and leku without naming the dialects (I suspect that the first is "coined" after a. y-dialect), Moisie-Shelter Bay legu, 1\-fingan le' go (rhetorically lengthened). As I have reported in Language and the American Anthropologist, the dialects of Fort George and the Great Whale River form a distinct subgroup in which e is replaced by a (e. g., GWR. nau "four"). I think Kiitcita'U8k• the name of the monster who slew the parents of Tcahkapes (exact form in GWR. dialect ~) is unique in phonology. Such a vocalism as in pe'sum' "sum" (pi/sum' at Fort George and Rupert's House) is characteristic of the Great Whale River dialect. So is the phonology to be seen in the final syllable of toi'pe' "ghost'' (Rupert's House toi' pai, Lemoine tshipi, Cree tcipay) . Observe also wa' tck" ''muskrat," Fort George utsk", Rubert's House utca' ck", Lemoine utshishku. The reduction in wi'na.sku "woodchuck" is, I think, pecu­liar to this dialect. A few characters of the Fort George dialect are the phonology to be seen in pai'lc" "one" (Great Whale River pai', Rupert's House pe'k", Lake St. John pe'yik", Bersimis pe""kw, Seven Islands pe'kw', Moisie pe"kw', Mingan peyuk10

', Natashquan pe'ya.k10

, Davis Inlet p{!-yuk', northeast corner of Lake Kaniapiskau pe'yak [presumably I missed tl1e labialization in the last two], the nonassimilation. in pipwn "winter" (beside pupwn; it is possible

82 BUREAU OF Al\1BRICAN E'.l'HNOLOGY [BULL. 1.23

the informant may have picked up pip-un from the Moose Cree sail­ors), also in tci'wauk'" "go ye back" (I lack a Great "\Vhale River cor­respondent; note the vocalic assimilation in au caused by "), the contraction and vocalism in tcl'lnioimi'tn "I like thee" [Moose Cree kimilwelimitin]; etc. ~.fuch of my Fort George material lacks cor­respondents elsewhere: therefore I can not tell V\T hat is peculiar to the dialect. The vocalic assimilation in atuhk'" "caribou" is presunl­ably entirely of independent origin though paralleled at Seven Islands, for the other dialects do not show it (Rupert's House ati'hJeu; the san~e at the Great Whale River, Sheldrake-Moisie ati'hkw', etc.). The name of the mythic monster Kiititosk" occurs in the same form at Rupert's House and Moisie. As cl1aracteristic of Rupert's House I have already given pe'k" "one." Characteristic is also ntcihtci "my hand", for in all other Montagnais-Naskapi dia­lects of which I have any knowledge the first to of this word is dis­similated to t. As a matter of fact ntihtoi also occurs in tl1e Rupert's I-Iouse dialect. We therefore l1ave every reason to suspect that the first forn1 comes from ~loose Cree (where the first tc re1nains; it is old, as shown by con1parative Algonquian linguistics) and tltat the dissimilated second fortn represents the archetype wl1icl1 has historic equivalents in the other Montagnais-Naskapi dialects, that is, sucl1 dissimilatio11 goes back to an aneient time. In conclusion I "\visl1 to say that in all the y-dialects h before a consonant wl1ere historically expected is very clear, with only a very fe,v cases where I have probably faultily not recorded it.

1\re turn now to the l-major group of ~fontagnais-Naskapi dialects. First we note the 1naintenance of l (which occurs sporadically in the "mixed n" group); then -sl~- becomes -88- before palatal vowels (also in the ''mixed n" group) as does -hk-. Examples are Lake St. John njlt "I" (Bodbout nil/ etc. n ·i' in the y-dialects; Moose Cree n/ila), aB8i "land" (Moisie, etc., Fort George a...fltai', Plains Cree askiy [in Bloomfield's transcription], kassi'WU' "all" (widely spread, Plains Cree kahkiyaw, Moose Cree kahkinaw). The whole l group has a special term for the numeral for "ten" which has the appropriate changes in the different dialects (Lemoine nikutulinu; Lake St. John nkutolnu, Bersimis kuteltnu [Sirois, kotelno]; Seven Islands ku'tumo'', Moisie kttt(u)no', St. Marguerite River kutuno' in the mixed n-dialects save Mingan which has a correspondent to the n­group term). The.re is a characteristic term for "five" which has, however, cognates in the "mixed n" group and (more removed) in the "unmixed n" group (Lemoine n!petetets; Sirois, petetets; Lake St. John petetetc'; Moisie pate'tato [from a speaker who regularly re­tains -ts, -tc]; Bersimis pate'tats, Seven Islands and Mingan pate'tat [the ''mixed n" group changes final *-tc to-t as in the true n-

::MICHELSO:-l] CUEE AND ~10 ... ~TAGNAIS-NASKAPI DIALECTS 83

group; tern1.inally aspiration is not phonemic]; observe also Natashquan pate'ta, Davis Inlet pate'ta in the "un1nixed n" group; the exact explanation for th.ese last forms is not at hand; since final -to nor1nally app·ears as -t it js possible that a final -t l1as been lost by dissimilation). [I have- recorded pate' tat once at Davis Inlet.] 'l"'he l-group agrees "\Vith the y-group in -ta ( -ts) HS an affricative in con­trast to both tl1e "n1ixed'~ and "urunixed n-groups" which. convert the affricati ve to a stop, -t. The question as to the retention of h before k, etc., in this group 'vhere historically expected is now taken up. My Lake St. John data is from an informant who really was a Mistassini; so that although this indicates the h "\vas kept, it is pos­sible that h.is o"\vn ~1ista.ssjni speech (a y-dialect) n1.a.y have ;'colored"­his I.~ake St. J oh.tt data. At Bersimis a single informant consistently omitted it, but k from -hk- between vowels does not become pllonet­ically -g- "\vhereas -k- from -k- in th.is position at times acoustically has the effect o.f -g-. .....t\. speaker of the ''1nixed n" group at Seven Islands consistently did the saJne. Like,vise an inforn1ant of Moisie who resides at Sheldrake. Ho"\ve-ver, an infor1nant from Godbout gave not only ntoeli1n "my father-in-la,v" but also ka'tepelda.hk "he who bosses it" (Lhe higl1 god). One informant 'vho has resided at !t.Ioisie for 51 years though origiiHtlly fron~ Bersin1is gave forms with Ill~ and It~. Another inforinant born at Sheldrake though residing at Seven Islands gave forms "\vitl1 hto, ht, and hk, but also forms k ( nitihtai' "n1.y hand," kuca' pahtaigan "th.e shamanistic shaking lodge," ni~wa' paltten "I see it," kiil':uoa' pahtahk "'the sl~ama11 of the shaking lodge," katependahk ''he 'vho is the master," ci'hk·ucic "ermiue," ati' hkw' "caribou," 1na' htcpc,-u "fox" in all of wl1ich h is expected etyinologically; but I recorded alj' k "toad", no' ku11~ "rny grandn1other," akuou' "he is sick" in all three o-f "\vh.ich hk should be expected). One inforn1.ant at Seven Islands ga Ye such forn1.s as h.·li'tapendahk "he who bosses it," kttca'pahto-iga1-"l· ''the shamanistic Hhaking lodge," u.' h.pa'n "lm1.g'' but nota "my father,'' and no'k·u1n ''my grandlnother." At Mingan -hlc- and -k- are variaphones. I mention this particularly in view of my criticis1n of Speck in the America1~ Anthropologist, n. s. 39, p. 371.

As far as I know, all of the l-group and "mixed n-group" lack any correspondent to N atasl1quan nti' hk·wati'ln (also at Romaine and St.. Augustine; I gather from Strong's paper in the American Anthro­pologist, n.s. 31, p. 277, that there are equivalents in the Davis Inlet and Barren Grounds dialects) with the meaning "son-in-law" etc. [historically tl1e word meaJ:ls "n~y cross-nephew" and is certainl)r used with th.is value in some Cree dialects, not in the vicinity of' James and Hudson's Bays, and also so1netimes with the value also of "my son-in-law"] ; they have a term wl~ich corresponds to the-

84 BUH.EAU OF Al\tERICAN ETHNOLOGY [BULL.123

y-dialects [Cree naluihkirim.]. However, the linguistic evidence postulates a Proto-Cree-Montagnais-Naskapi nitihk~tvati-m which form is confined to these languages ; Prato-Algonquian had *nefJe-iikwaOehsa "my cross-nephew," possible also a form without tlu~ diminutive suffix. I have pointed out that both the l and 11- groups (save Lake St. John) have a word for "river" which postulates an earlier sipo (w) [cf. Fox sipowi] whereas tl1e y-group has a word corresponding to Cree sipi. Owing to th.e lack of co1nparable data I can not give hardly anything to characterize the individual l-dialects. I will mention, however, that essi "land'' with initial e- occurs only at Bersimis. In the Bersiinis dialect also to appears as ts. I again call attention to the "mixed n-group"; though they show occasional l, (where original), and though tl1e treatment of medial -sk- before palatal vowels is like that of the l-dialects, termi­nally -to becomes -t as is th.e case in the unmixed n-group.

The major n group will not detain us long. Tl1ere is no data to amount to anything in much of this area. So the following is merely tentative. Besides the change of l to n, final -te became -tj' prob­ably -tt in phonetic transcription but -t in phonemic transcription (which also occurs in the "mixed n-group," ilnut, inut "Indians"); the treatment of original -sk- before palatal vo'\\ .. els was the same as in the y-dialects with changes in some individual dialects. The nu­meral for "ten" seems to be characteristic of the whole group, even though it also occurs at Mingan which seems to be of the "1nixed n-group." Observe Mingan pe'yakwV/nw, Natashquan pe'yagwan.o and pe'yagunut, Davis Inlet peyugunu.

In the Natashquan dialect a sibilant normally becomes h (I have · heard from Indians that the same thing happened at Ro1naine and St~ Augustine); compa're also Speck, A1nerican Anthropologist, n. s .. 33, p. 586 (his j is used with the value of French, not English j). I have given some examples above, but I give son1e here as it is appro­priate: nihteh "my elder brother," nihtau' "my brother-in-law (male speaker), witai' htenta1n "he wishes to lo1ow it," ami' hkw' "beaver," ihpahtau "he runs thither," mihlcwami' "ice,'' tai' htenwu' "tobacco," kuhiipahtoigan "the shamanistic shaking lodge," mahteehu'' "fox," kutwaht "six", ni' h "two." For interest I add win' kii/totahk. tcihigunu "he who has made the day." But variants occur; acini/ a7Ul ahi~ni' "stone" both occur, and I recorded nist" "three." It is possible or even probable that these variants were due to unconscious imitation by my informants of the speech of my interpreter (who \Vas of Mingan). At Davis Inlet this shift is lacking, e. g., niotio "my older brother,'' uopwiiga.n "pipe," nitcistel}na·mukw "he knows me" [Moose Cree nikiskelimik,· Fox neketkanemegw.Aj see also above], a.atici" "land." The vocalism seen in Davis Inlet tcimwa.n "it rains"

:MICHELSON] CREE AND 1\-IONTAGNAIS-NASlCAPI DIALECTS 85

differs from th'at of all other l\fontagnais-Naskapi dialects wl1ich we have so far considered (Natashqua11 toimun'; Cree kim.i1.oan. The vocalism of me'sk"' "bear" and se'toime-u "mosquito" seems not to occur in the dialects :for which. tl1ere is data. (Cree m.askwa, sakimew). Also the plural i',nwat" ''Indians" apparently is characteristic. It may be noted that terminal -hpt. sounds nearly like -f. The word for ''kettle" I recorded more than once a.Stai011 (Cree askihk, Fox .Acku'tkwA).

With tl1e materials at my disposal I l1ave now done as much as pos­sible; and it is hoped that t.his will at least serve as a stepping stone to an exhaustive classification of Cree and Montagnais-Nas­kapi dialects. Some sligl1t inconsistency in orthography and the double use of phonetic and phonemic transcription is a matter of convenience.

APPENDIX

1 American Anthropologist, n. s. HR: 685, GbG; Lant:;·uage, 12: 135, 136. s E -·ery one (including myself) has overlooked the fact that Chamberlain~

Ann. Archaeol. ltept. 1905, App. Rent. Min. Ed. Ont., Toronto, 1906, p. 123, correctly saw that 1\listassini was a ?vlontHgnais ( -Naskapi) dialect, even if he was unaware of its immediate affiliations. For those who care to foil,>~·

the course of the controversy on 1\fistassiui, et«., these additional references may be useful: F. Speck, Proceedings of the T\venty-first Congress of Ameri­canists, First Part, Tbe Hague, p. 268 [192·1], Proceedings A1nericau Philo­sophical Soc. o5, pp. 275, 276 [1926J, JAFL 28, p. 70; in a photographic repro ­duction of a map signed and dated '2G Waswanipi L-:; still classified as Cref!; in his map published in AA., u. s. 33, p. 5Gfi [" 1931], it is apparently also so classified; Davidson, .J A]j"""~L ·11, p. 262 [1925] au.nHloned llis more cautious atti­tude and said unreser\~euly ''\Va~:nvanipi Cree;'' .Jeuness, Indians of Canada~ pp. 266, 283, 423 [1932] should ul~o lJt~ consulted. I pr(•sunle Speck met a rni­grant 'Vaswanipi who spoke either Moose Creek or ,, .. aswanivi 1nixed ·with Moose Cree. At ~loose Factory I met just such a per!'oiJ, and at Fort George­another Waswanipi migrant who spoke a different 1nixture. Although the maxim is de mortuis nihil nisi bonum, I a1n compelled to state that the H1 "' 1J

and accon1panying remarks by A. Skinner, Notes on the Eastern Cree aud Northern Saulteaux~ Anthrop. Papers Amer. a.lus. N. H., IX, pp. 8-11 (1911 ; the whole volume appeared in 1912) is practieally without value; wheu it is stated that the terms are given in the dialect of the respective di\ i~ion . ..:, it sin1ply is not so. Such a ter1n as Klt-m3ti.n-'fiwuJUfJ (Fort George) is 1nix('<d Cree and l\Iontagnais-Naskapi, probably an a ttctupt of so1nc tnigrant to gi Ye the native term. The author had insufficient skill to kuo·w what linguistically is Cree and what is Moutagnais-Naskal)i; and V\1 bcn it is stated that the Fort George Indians speak the same dialect as at Rupert's House, I can say definitely that it is not so. I pass over other vagaries. Proof positive that the dialec.: t of the northeastern corner of Lake Kaniapisln1.u bf>longs to the y-group of Montagnais-Naskapi dialects, and to the grouv represented typieally by rile dialect of Rupert's House is sbo·wn by i'cpi1nihtc "up abuve,"·ickwah.tom "door." tuhkapu/ "he, she opens his, her eyes," na'peutc "ln€>u," ye'ga.u. "sand," nyai' ''five" tctsteihtcun ''he, she kno\vs it," k{i' iapeihtahk "be who bosses it," mi'tciwiihp "dwelli:ug, tent," nz.·iwe' ihta·;nu.tc "th<~Y like it/' etc.

3 Tete de Boule has mostly (erroneously) been classified as Algoukin (Algonquin). Beside the references given by n1e iu previous publications, the following are of historic interest : F. Speck, Indian Notes, Heye !vl use urn, 4, p. 251, [1927], J .£\.FL 38, p. 2 [1925], the photographic reproduction of the n1a p mentioned above, dated and signed '26 (all classified as Algonquin, or Algon­quin-Ojibwa, Ojibwa) ; Jenness, Indians of Canada [1932], p. 276 (Algonquiu) ; the classification by Davidson, Atti del Congresso Iuteruazionale degli Ameri­canisti Roma, Settembre 1926, Vol. II, Roma, 1g2s, p. 70, is not entirely clear to me ; in view of his previous classification of Tete de Boule as Algonkin (Algonq1.ti.n), this presumably is still nmiutained; Chcuuberlain, Ann. Archaeol.

86

MICHELSON) CllEE AND }IIQNTAGNAIS-NASKAPI DI.ALI~CTS 87

ltept. 1905, App. Rept. 1\iin. Ed. Ont., Toronto, 1906, p. 123, faultily classified 'r£ite de Boule as l\Iontagnais; beside the correct classification by Cooper tnentioned in my previous reports, his correct classification is in the same vohnne of the Atti, etc. (noted above), as Davidson's rather unclear one; on Speck's map, p. 565 of AA, n. s. 33 [1931] it is apparently correctly classified. A co1nplete and accurate classification of all Ojibwa-Algonquin dialects is a great desideratum. Some of the published material does not even positively tell us 'vhether a given dialect is Ojibwa or Algonquin. The orthography of Algonquin in the Jesuit Relations at times reflects the orthography of 1\-lon­tagnais. The ~·-dialect (e. g., cha.Oc'l·indantalJ·ln "ayez pi tie de nous'', .TR, ed. 'J1hwaites, vol. 16, p. 44) early disappeared or was modified. "rhen Alexander llenry, '].:ravels and adventures in Canada, Pte., New York, 1809, p. 214, says of the lunguage of the Tf'>tes de Boule "Inixture of those of its neighbours~ the Chivewuys and Cristinanx" this partially anticipates Professor Cooper's infor­mation, but of course is an error if taken in the sense that the dialect is a thoroughgoing amalga1n of the two. Possibly nonspecialists would be glad to know Cristina ux is a synonym of Cree. The older maps are quite unreliable on linguistic matters; very obviously '~native" uames are not "native.'' How and why Cree was extended to cover n1ouch of the territory east of James and Hudson's Bays is unknown to 1ne. Even at Seven Islands a native called the Fort George Indians "Crees" which linguistically is not so. So, too, 'vhen nt Lake St. John I heard whites refer to the ?tfistassini as Crees. Similarly, .John 1\IcLean, Notes of a twenty-five year·s service in the Hudson's Bay Terri­tories, in vol. XIX of the Pub. Champlain Soc., Toronto, 1932, p. 258, "The In­dians ... of Ungava are a tribe of the Cree nation designated Nascopies ... T1tPir language, a dialect of the Cree or Cristeneau, exhibits a considerable mixture of Saulteaux words .•. " [See ch. IX, vol. II of the original edition, I...~on<lon, 1849.] (Of course the mixture of Sanlteaux is untenable, to judge from Turner's unpublished vocabulary; W. B. Cabot a.pud Labrador, etc. W. R. Grenfell and others, New York, 1913, of course repeats it, substituting Ojibway for Saulteaux; this means nothing, cf. my criticism, below, of his "In Northern Labrador"). See also H. Y. Hind, Exploration in tbe Interior of The Labrador Peninsula, London, 1863, vol. I, pp. 33, 322, vol. II, p. 97, and Turner in tlte I~leventh Ann. Rept. llur. Ethn., p. 267. According to Prof. ;J. l\1. Cooper (the Northern Algonquin Supreme Being, C U A, Anthrop. Ser. 2, p. 73) iu Gorst's day, 1670-75, to judge fron1. the short vocabnl.l:lry, at Rupert's House the Indians used a Cree r-dialect, though to{lay they speak a ~Iontagnais one.. [This is in Oldmixon, now easily accessible in Tyrrell, Documents relating to the early history of Hudson Bay, vol. XVIII of the Publications of the Champlain Soc., p. 396.] A study of this vocabulary tends to show that this is an ovpr­simplification. 1\!uch is not diagnostic in a linguistic sense; 8ome words might be Cree, 1\!ontagnais, Algonkin ( -Ojilnva), such as 1noken1an, knives (really singular), rnetuH, stoC'kings (really singular). cskon, a chisel; T ·inesonec.iso, what do you call this ("""hat is it called'') Manito,wghigin, a Red-CO'at. n1ight be Cree, Algonldn or Ojib"\'\.,.a (I lack a J\fontagnais cognate) ; mekish, beads (really singular) nlight be Cree, Algonkin, or Ojibwa : the phonolog-y definitely excludes it from helng 1\iontagn.ais; Borne words might be Cree or Montagnais, e. g., assinne, shot (phonology decisive), a,sfa.n.,, come hither (shown by phonetics ; also a lexical trait) , a pit, a fire-steel (lexical trait), a notch, presently (lexical trait), pishshish, a little thing (lexical trait), petta a s1utnt.e. give me a piece ("band and give me to eat;" phonology and morph­oJogy decisive), Pe quish a con Gau. J,fo'l.oon, I eat some pudding ("I shall eat bread;" possibly Algonkin and Ojib,va also), Spog.m a. pipe (corruption of Cree

88 BUREAU OF ..:\_l\II~UICAN I~THNOLOGY [DOLL. 123

t.tspwiigan or the 1\fontagnais equivalent; the }Jhonology bars it from being Algonkin or Ojibwa), Ta.ney, where (corrupt ; pretty clearly only Cree or Montagnais), Tapoy, that true (the phonology bars Algonkin-Ojibwa) ; defi­nitely Cree are chickah-igon. n hatchet (lexical trait), picko1v, powder (the phonology is decisive), l\7 o 1uu.n.-niss e to ta, I do not understand you (corrupt; the meaning is rather, I do not understand it; the pl1ouology strongly favors Cree and nothing else), ou/lna, this (lexical trait), pihielc.em.an, a jack-knife (read Pikicke·rnan; it can not be ~fontagnais; Ojibwa and Algonki11 have a compound which ends in an equivalent), Shelca.hoon, a comb (Uree rather than Montagnais because of the phonology). Clearly 1\'Iontagna.is (and not Oree) as sho\vn by the phonology are pastosigon, a gun, and stenna-, tobacco (read ste:rn.a). Tequam, What do you say? is either Cree, in which ca:;;e Keq·uan should be read, or a mishearing for ':Pchequan, in wbieh cast? the word is l\fon­tagnais. ]}:foustod(Jwbish, a flint, seems corru1Jt; the t<~r1ninal portion doubtless should be -a.wb'isl;. Algonkin (-Ojib"va) are aru.ka;:o. lJread (rend a1·a1t:oua.; it will be recalled there was an 1·-dialect Algonkin; the 11llonology faYors Algonkin as opposed to Cree), Arrem.it-igo.sy, to SIJeak (l(•Xical trait), tHi(·kc lly, powder (the phonology is decisive). [Soth.in1.1n. reel lend, is unclear to 1ne.]. It will be recalled that the Abitibi came down to the Bay at tbe titne. I interpret tbe whole to mean that owing to the establishn1ent of the post at Rupert's House d.i.fferent Indians came there to trnde, just as at the present tin1e besides tbe l\Ioose Indians at Moose Factory, the llup~1·t·:o-: llouse ~Ioutagnais-,.;peaking Indians, or rather a group of them, Ojibwa fron1 the Albany River, etc., come there also; and that the data do not nPcessarily \Vnrrant us in holding that the geographical boundary between the Cree and ~Iontngnais has shifted within historic times, though this is possible.

6 Cotnpare Henry, loc. cit., p. 819 "Chut·cTzi.ll R .irrr •.. it was named E·nglish Ri'tu~r ;" Daniel W. Harn1on, A journal of voyagP.s and travels~ etc., Andover, 1820, pp. 167, 1.68: "The River last mentioned, is cal1ed hy the Hudson Bay People, Churchill Riv-er, and by the people from Canada, English River;" notice that on the mnp facing p. 223 of Henr).,. Youle Hind's, "X:lrratiYe of the Canadian Red River Exploring Expedition of 1855, vol. II, London, 1860, we find "Churchill or English River" ; English River ill the f';ense of Churchill River is still used in E. A. Wa tkinR, A J)ictionary of the Cree Language: ("t<: .•

J ... ondon, 1865 (p. x). See also Sir John Richardson. An Arctic Searching Ex.­peuition~ vol. I, (map facing) p. 16; vol. II, pp. 36, 37, London, 1851. La­combe on the tnap in his Dictionnaire de la Langue des Oris, Montreal, 1874, has only R. Churchill. Hunter, A Lecture on . . . the Cree Language, Lon­don, 1875, p. 2, bas English lliver. See too J'oseph Ho·w·se, A Gra1n1nar of the Cree Language, etc., London, 1844, p. 14.1 et passim, especially p. 318. The same author, loc. cit., p. 316, observed very acutely, "On the East-main side of Hudson's Bay, ( t) ch is in general use ... instead of the k (or c hard) used on tl1e "-rest-side of the Bay." Of course, the ( t) ch dialects are really Montagnais-Naskapi. When Cre-e words are cited as being at Cum­berland House, as has happened, 'vi th. th and not n. as a.t present, I do not think that this means that a shift of population has taken place; but rather, that the interpretPr or infornutnt re-ally wns not native to Cumberland House. According to Mgr. Tache, Esqui~se snr lc Nord-Ouest de !'Amerique. 1\.Iontreal, 1869, p. 82, some Crees at Isle a la Crossf' Rpenk an Z-, not r-dialect. No subse­quent writers, as far as I know, have repeated or substantiated this. When Lacombe says, loc. cit.. p. xv, "Oris du l:Albrador ... nila-kila-'\\ri.la", this really means the Cree of Moose Factory or less probably the specj al Tete de Boule dialect with l.

]'dJCRELSON] CREE A .ND MON'I'AGNAIS-~ASKAPI DIAL~JCTS 89

r; For Cumberland and Norway House Indians, see also Hunter, loc. cit . ., See J AFL 41, pp. 164, 165. For Cree influence note 1nistatiman horses,.

nic1oasttt seven. Incidentally, though Skinner applies the term Bungees [Bungis] to the Plains Ojibwa (see also J AFL 29, p. 330), previously it was applied to the Indians of the general region bet'\veen Norway House and York Factory : see Sims, J AFL 19, pp. 330, 334 ( "Bungees or Swampy Indians of Lake Winnepeg") ; see too Hind, Narr. of the Canadian Red River ExpL Exp. 1857, etc., vol. I, p. 333; and J. Stewart in Am. Arch. Rept. Ontario, 1904, Toronto, 1905, p. 89. The linguistic mixture is in part supported by the new da tn. from Island TJake ; bn t a tb oroughgoing mixture of Cree and Ojibwa for the whole area is out of the question.

Y The paper "Montagnals-Naskapi Bands, etc.'' by Speck in AA, n. s. 33 [l931], pp. 557-600. with the accompanying map, is valuable as showing the distributions of the bands. Part of this area was investigated by Speck bimself, but partially be relied upon information furnished by Alexandre Belleflenre and SylvPstre Mackenzie: see p. 574 and also his Naskapi [1935], p. 84 et passin1.. On n. 580 of "Montagnais-Naskapi Bands etc.'· we are told that the nntive names of twenty-six bands are given. On p. 584 it is stat~d that the St. 1\tlarguerite Band is the farth~st east on the coast to speak an 7.-dialect~ ~Jvidently there were misunderstandings of some sort, for contradictions occur as shown by the "native natnes" on pp. 584, 585, 686. Very fortunately n1any of the statements can be checlted by data contained in "NaRJ.mpi"; and since this appeared two years later than the article in AA, 've n1ay legitilnately infer that where the data is contradictory the data in AA is to be super~eded, and '\Vhere data in "Naskapi" confirms the state­ments i.n AA these are to be regarded as Speck's final opinions. Thus, the state­Jnents on Escoumai.ns and Shelter Bay, AA, u. s. 33, pp. 681, 583, 584 are confir1ned by data J.n "Naskapi," pp. 64, 100, 175. '.rhe contradictory data on Natashquan ("Naskapi," p. 99) is confirmed by my own data (n-dialect; terminal -tc [-ts] converted into -t (per contra 'llotac1c·wa-nwilnut'9 AA, n. s. 33, Jl. 586). That Michlgamau is an n-<lialcct. A. A., loc. cit .• p. 589, is con:firtned by the data on p. l77 of "Naskapi," but I am unconvinced that -ts' remains. That Mingan is an n-dialect (essential1y, if not wholly) is shown by my own data which confirn1s the statement on p. 584 of .... .o\..A., loc. cit.; my data makes it a dialect in which -tc ( -ts) becomes -t; against both these points note on p. 585, "J.llingan Band (Aku;andi' 'wiln.uts•. 1.vhere sornething is washed ashore people.)" The data on Musquaro, pp. 586, 587 is somewhat contradictory, but it is given in any event as an l-dialect; personally I think it much more likely to be an n- and -t dialect. The data on Kaniapiskau, p. 590 (-t~ dialect) is against the data I have presented above. That Ungava, p. 594 of AA, loc. cit., is an !-dialect is against the data from Turner and statements made to me by a Davis Inlet Indian whont I met at the Northwest River. I '\VOnder lf most of the contradictions may not be explained the following way: besides tbe Sylvestre 1\Iackenzie, chief of the Michigamau band according to Speck, there is another Sylvestre l\fackenzie at J.\.Ioisie who often goes with the· mail to Fort 1\fackenzie ; he uses the l-dialect and retains final -ts (so too Rastian Mackenzie at Seven Islands who bas been at Fort 1\Iackenzie) ; if he, and not the Sylvestre Mackenzie of Michigamau, furnished some of the _.native" names in his own dialect (or "colored" the native names with his own dialect), much would be explained. Sylvestre Mackenzie of Moisie came from Bersimis ( Basti.an also, I think) but has been at Moisie for 60 years. Similarly if .Alexandre Bellefleure really is of the Seven Islands group and· not the Ungava one, his designation of the Ungava band might easily be·

:90 BURI1JAU OF AJ\IEH.ICAN ETHNOLOGl:~ [BULL. 123

in the Seven Islands dialect; or if John Pierre of the Ungaya band but n1.arried and living at Seven Islands gave the information it is po~sible the t1esignation given is really in the Seven Islands group's dialect. The data in Cabot's In Northern Labrador n1eans nothing unless the exact provenience of his interpreters and informants is known: the data on p. 286 tnakes botb Cree and Montaguais-Naskapi exL~t east of the Georgp or Barren Ground River: Compare Speck, Naskapi, p. 56. 1\1rs. Hnhbnrd's deRignation of the Barren Ground band (A Woman's Way through Unkno,vn I~n.brn clor. New York, 1908, p. 158), l\Iusl1-a-waue-u-its, is not in the uaUve dia1~ct, uut iu the dialect of Rupert's House ; compare pp. 20, 150. [Nor is the designation lty Cabot! loc. cit., p. 198, the native one.] Similarly the Indian song recorded hy D. Wallace, The Long Labrador Trail, New York, 1907, p. 73, ls not in the native dialect: see pp. 16, 119 ("Pete" came from the neighborhood of Lake Superior). The statements of Indinns as to the liugul.;;tic affiliations of other bands than their o'vn, even 'vhen they have tnet them, 1nust be taken with a grain of salt. Thus, Bastian 1\::lackPnzle, '\\Tho bus n1et Great Whale River Indians, Chimo Inclians, those of 1\Iingan, etc., says the Great 'Vhale River Indians are Crees and say tcila! "thoun (neither of which is true), Chimo Indians tcin.a' (which I think is true fro1n other data? disregarding vowel­-quantities. as also further on), Lal{:e St. John tcila.~ (my o"rn material lacl;cs the final a'), Ber5=dmis tail (true), Seven Islands tcU (partially true), ~Iingan tcm (true), Mingan tzd.pettt• "men" (true), Moisie, Seven Islands, St. ~farguerlte nii'peuts. Bastian and Sylvestre 1\Iacl~enzie are the only ones of the a rea that I met who consistently retained -ts ( -tc) nnd did not change it to -t~· -ts is retained. at Bersimis.

8 Cooper, CUA., Anthrop. Ser. 2, p. 60, says, "Fron1 the linguh:;tic data f.,>iven by Le Jeune, we have good. grounds for concluding that ::;on1~ at least of these

·"Montagnais" Wf>re Cree-sp~nldng, using an t• dialect, etc." I regt·et to say that a do not think the evidence warrants us in making this ~upposition. It should be pointed out that Le Jenne gives one 'lDhole sentence which though supposedly In Montagnais, actually is in Algonkin Ucaie, nir, k-..h iga.toutaou im, which further­D1ore is mistranslated, for tbe second person plural is the subject). It can be proved tbat 1\lontagnais orthography at tin1es liaR inflnen<:>ed that of Algonkin. Presumably the reverse has nlso happened. Note too, i.u tbe Jesuit Relations one personal name given a~ .A.lgonkin actually is l\fontagnais (Michtaemilco1lan "Great Spoon"). For t11e valueo of k1ti, see the Jesuit Relations, ed. Thwaite~ 10, p. 117. To account for both Algonkin (which has only prilnary tc) and 1\-Iontagnais (~Yhicb ba=-: not only prhnary bnt also secondary tc) l.:hi n1ust hnve the value of tc. Note {Jma.ntiOetch (natne of a people) Jesuit R~lntions, ed.. Thwaites. 59, p. 56 but also Oma-n'ti0e1chi. at 21, p. 116; oma1niOck at 35, p. 274, ·either has k for kh ( i) [see bel owl or the word is really the Algonkin designation of the same people. [The etymology given by Gerard, Bull. 30, Bur. Amer. Ethn., under Oumamhvek is on the rigbt track, but !\ofoutagnais or Algonkin should be substituted for Cree.] In sotne other tribal uames we have a sbnilar situation. Especially noteworthy is Montagnais Khirhtent.alt "petun" ("tobacco"), .Jesuit Relations, ed. Thwaites, vol. 7, pp. 24, 26 (1634) [ed. Quebec, 1858, -vo1. I, 16.34. p. 49] recorded as tshisternau by Lemoine; :\'askapi tc' stmn(.ht, tc' stem.au is given by Speck, Nasknpi, pp. 218, 241; l\lingan tri' sternau' and F'ort George tcistii1nau', Rupert's House tctstenza.u' are recorded by 1\lichelson. This has primary, not secondary, tc as shown by Cree tc,lstema.w (l\lichelson; tsil!tiinuiw in Bloomfield's transcription; Lacombe gives tc1ziste-tnau...'), l\:lcnon1.ini nc'nenui'W, Fox ne' siimawa, the Proto-Algonquian archetype of which is *Oce' Oii'lnii/u;a. as shown by Bloom-1ield in his remarkable paper on Central ... ~lgonquian phonology, Language, vol. I.

A:f.ICUBLSONj CH.EE AND l\lON~r..-\.GNAIS-N ... <\.SKAPI DIALEC~TS 91

£'Vutkins, A Dictionary of the Cree Language, London, 1865, gives chiistatnow, i. <~., tci8t('3'1niiw, but says '"'See Kistiitnotv." I know no Cree dialect in which this lust occurs; and Bloo1nfield writes me under the date of May 17, 1937, that his Cree notes contain no variants of tsistiirnilw; nor do I know any Montagnais­Naslul.pi dialect with lc-isttJmli-tv :~· I wonder if k-istli'in.ow is not pseudo-correct Crt~e (cf. pp. xi, 201 of Watkins' Dictionary).] Beside kh(i) with the value of l(J, teh(i) is used with the same value, also ts, tz, ch (rarely), th, k(i), dk, t ( ·with certainty at least twice; perhaps more often). I have cited 6m.anti8etch •'qui sont une nation des eskimeaux" ('\Yhich is false; this is not in Bull. 30, Pt. 2 in the synonymy under Oum.am·ilvek, nor in the synonymy at the end of the vollune; the etymology given by Gerard is correct [Cree mamihl~, Algonkin miilniiig, ~.fontaguais (Lemoine) ntan"its "downstream"] but the 1\rlontagnais word should probably be restored as • nu"itni'tVA tc; note the variants Ontan~iOckhi and lhnatni-zvek of the Quebec edition for 1U4l, p. 57, 1650, p. 41 respectively. Other exan1ples are espt'l'n-i1nitclt upvv·:-n·u (Jesuit Relations, ed. Thwaites, 49, p. 66; ed. Quebec, III, 1664, ell. IV, p. 19), Montagnais ishiJintits (Lemoine), Rupert's House icpintihtc, Cree ispitni1t7r,, Ojibwa icpi;nU"ig, Fox .~:.pentegi, Proto-­AJgonquian *iczJeJneiiki_; Mitchition eagle (part of a personal name; JR. T. 59~ p. 63; the ti here means .:.: or c as proved by per,ak.BnayONeO, JR. T. 59, p. 62, ''celuy qui paroit tousiours malade"; Lemoine -nii./n.tshu.. Cree -nakusiw), Mon­tagnais metshisllu (Leinoine), Fort George nti'tic·l_t' Michelson; dissimilation for tnitcicu), Cree 1nikisi'l.v; tchise ''big" in takOa-tch·isenapeB "the little big man•' (JR. T. 59, p. 62), l\1ontagnals tsh·ishe "big, great" (Lemoine; nap~ "man," taku au "it is short'') , Cree kise-; N ou. ta-kata.nitn.ieou etch "my father they will pray" (see JR. T. 49, p. 62 JR. Q. III, 1664, ch. IV, Jl. 18) should be divided Nouta ka-; ]j'"'ox tVA ui, Plains Cree kata aya'tnihawak, Lemoine's l\fontagnais fm.ieuts, aiarni-auls (see his Diet. under Prier) ; 'vithout doubt Tchigigoutcheou the god of fine and of bad '\veather (Jit. T. 68, p. 43) is a corrupt plural and stands for 'Pchichigou..etch, identical with 1-.:hichikouekhi (JR. T. 11, p. 254 JR. Q. I, 1637, p. 46; Khichicouai, pass·i1n apparently is an obviative plural) which corresponds exactly to Speck"s tci" cagwats (Naskapi, p· 62) ; compare also Ka-Khich·igou Kltetiklti ceux qui font le jour (JR. ~r .. 11, p. 254-Jll.. Q. I. 1637, p. 46) ; J.etnoine tsh"ijikau "jour"; Plains Cree kisUcaw _; for the last word 8<t.'

Lemoine under Faire; a ·itin, aitu co1nbine<1 with ka- according to the paradigru of nt'iU:ni8hin je suis sv.ge! ilinishtt. il est ~age (Lemoine, Grammaire Montag­naise, pp. 20- -22) should be *kuaitits (uucontracted) in the third person pL animate of the participle; but this is rather peculiar; ou.i ntir·iatchi si tu nou:s veux donner [si tu no0 veux doner] (.JR. T. 7, pp. 154, 156; JR. Q. I, 1634, ch. XIII, p. 76, l\Iontagnais ui1ni.liiats in Len1oine's transcription, is another example. '.rhe word tnichoutchi "en contrechage" on p. 154 of JR. T. 7, in Le Jeu11e's normal transcription should be *1nichco'ltlch or *Tnichcottkhi, in Lemoh.u:!·:;; mi8hkuts, but the word happens to be absent from his Dictionnaire: see, how·­ever, Echange and Echanger "\Vhich prove the '\vord genuine and moreover ,.VHt­kins cites Cree mcskooch and this seems to be supported by the evidence of Ojibwa and Algonkin (Lacoinbe cites Cree ntcslMttch which is difficult, for '\\"·s e is i, L's e is e). Attik'iri-nouetchs on Bellin's map of 1755 bas a native plural with added French -s, and corresponds to A.ttikou Irin·io-uetz on La Tour's m.up o:r 1779, phonetically presumably *A tihlctvilini-w.A tc "Cnribou men." Tan tche tchichilca-Tnas6ianc [probable misprint for -yane] egO "How then can I pay for myself" (JR. T. 68, pp. 62, 63) presumably stands for tan te tcih-k.&masuylin; see Lemoine Diet. Mont. ~rsh ijikat-ihun! etc., under Payer, Lemoine, Diet. Algon­quin, Jcf,jika-n, lcijik(ts, ktjUca.rna u;anis under Payer; see also Baraga, Diet. Ot<·hipwe, II, R~i}ika. n etC;. I pay n1y debt, J<ijika.s etc. I clear Juyself of debt~ ..

92 BUREJA.U OF .A.l\IERICAN ETHNOLOGY [BULL. 1.23

Kijikawa, etc. I pay him my debt. Some other cases o~ tch are more con­veniently discussed in connection with some other variants. Some examples of -ts for the same sound as -tch (or presumably ; for in the Bersimis dialect -ta does occur) are Man·itous·iueta sorciers ou iongleurs, JR. T. 6, p. 124 (1634) but Manitousiouekht, JR. T. 8, p. 273 (1636), the Prato-Algonquian archetype ot which is •manetowesiwaki but it must not be imagined that in any Cree or Montagnais-Naskapi dialect the original final *·i of the anilnate plural, etc., .actually survived for it did not; Atticatneoets, Attikoetz, A tlika.n?egour7chi, etc. (the name of a Montagnais people; see Bull. 30, Bur. An1er. Ethn., Pf". 1 tmder Attikameg·ue, the index to Thwaites' ed. of the Jesuit Relations and that of the Quebec edition; cf. Attikamegou, the personal name of a Montagnais, JR. T. 9,

·P· 72; the presumable phonetic restoration is *.A mekw.,dikhAtc '"The Caribou­Fish People," i. e., "White Fish People"; Ottchestigouets ( Ouchestigo-uetch etc.), the designation of a Montagnais people will be discussed below. Examples of where ch has the value of tc and which are not numerous are naspich "entiere­met," etc., several times on pp. 154, 156 of JR. T. 7: this has primary to as shown by Cree nlispitc; in most 1.\'Iontagnais dialects Michelson has recorded nlispito (but nlispit' at Mingan) ; Lemoine gives niishpits; the locative iscoue­chich d'vne ftlle (JR. 'I'. 7, p. 154 is ishkueshits in Lemoine's transcription (presumably phonernlcally isk1vesihtc) : notice the locative ouascoukhi which in Lemoine's transcription is 'ltashkuts "en ciel", on the same page. The word outagouchi bier (JR. T. 7, p. 24-JR. Q. I, 1634, ch. XI, p. 49) is an haplology or haplography, presumably the latter: Lemoine gives utiikushUs.,· I have recorded Rupert's House uta'kttcihtc; Plains Uree utakusi.hk (restored from Horden~~::~

otakosi' k, w~atkins' ootakoose' k, Lacon1be's otft.lt;nslk). Another example of a locative in -khi is Ouabichtigoueiakhi . . . Kebec, JR. 'r. 7 , p. 204 (J'R. Q. I, 1634, ch. XIII, p. 86): Lemoine gives ~Iontagnais 'Ut1p-f.~hf1t1~e·ints, and in hiH Diet. Alg. gives the equivalent, Webitik·weiang. So the locative t('rntjnatiou is firmly established. The spelling dl" with the value tc is wholly isolaied: nisadkiha-u ie l'ayme (JR. T. 8, p. 36-JR. Q. I, 1635, ch. III, p. 18) which in Lemoine's transcription is nishatsh1att [Plains Cree nilulkilul.w]; observe Mon­tagnais khisakh·ttan [tu aimes cela], JR. T. 6, p. 238-JR. Q. I, 1634, cb. V, p. 29 [Plains Cree k ·isiikihtiin] which is tsh£ shatshtan in Lernoine's transcription; so too on the same page note sakltita "aime le" which in Lemoine's transcription is shatshita. It is important to here note that the orthography of Algonkin in the Jesuit Relations reflects that of Montagnais at times. So we have the Algonkin sentence napfJ.-., nisadkiha ~;n.issi kakichUdtz ... entirement i 'ayme celui qui a tout fait (JR. T. 2B, p. 88-JR. Q. II, 1648, ch. V, p. 23) which is an old misreading or misprint for ... kakichitOtz as shown by the variant n'li.ssi. 1:-,akichitO/ch celui qui fait tout (.JR. T. 24, p. 40). In Lemoine's transcription ·the Algonkin sentence is nd.pitc n·isak-iha misi ka.kijitotc. The word nap·ik will be treated below. That th was used with the value of tc is convincingly shown by the Algonkin variants ninithanisa.k mes enfants, Ninitcha·nis mon enfant (JR. T. 24, 1642-43, p. 40) which in Lemoine's transcription are flinidjttnisalo

n -inldj4nis respectively (Fox nenidtciine' s.&gi, nenidtcline's.&). The variants ot Montagnais show the same. Thus we have attimoueth [chiens] at JR. T. 7, p. 22-JR. Q. I, 1634, cb. XI, p. 49 but nama i.riniso1takh·i [mislection or misprint :for ·*i'l-inisouakhi; the singular i-1·inis·iou occurs at JR. T. 6, p. 246-JR. Q. I, 1634, ch. VI, p. 31) attimouekhi les chiens n' ont point d'esprit, JR. T. 5, p. 138-JR. Q. I, 1633, p. 13. Again, ie vois des homes n·iouapa-maoueth irinio-ueth (JR. T. T, p. 22-J'R. Q. I, 16.34, ch. XI, p. 49) which in Lemoine's transcription (in a clearly little different dialect) is ni 1l(lpa1nauts ilnuts (Moose Cree nitvlipanliitvak .Hiliwak phonemically ; Fox ne1vap..t maw.& gi neniw.&gi) : note the variant

MieHmLSON 1 CH.Eifl .i\.ND 1\IONTAGN,L\.IS-NASKAPI DIALECTS 93

i1·-bniouakhi at JR. T. 7, p. 156-JR. Q. I, 1634, ch. XIII, p. 76. Observe also ninoutinaoueth (I strike them) at JR. T. 7, p. 22-JR. Q. I, 1634, ch. XI, p. 49 \Vhich contains the same pronominal elements as above; I do not recall a Montagnais equivalent in my own notes, but Cree guarantees the word.

Another example is Achitescatoueth • • . ils passent mutuellement a la place l'vn de l'autre (JR. T. 6, p.163; JR. Q. I, 1634, ch. IV, _p.14). I do not know the modern ~Iontagnais equivalent but terminally it clearly corresponds to I...emoine's -tuts (Cree -tuwak). Examples where t is used for the so1..md in question (tc) with certainty are eca trtou, eca titou (JR. T. 7, p. 62) tais toy, tais toy (in Lemoine's transcription this is eka tshitu .. , (Cree eka kitu) and tn·itisoukou mangez, JR. T. 7. p. 176 (with primary tc; see mitshun in Lemoine's Diet. Mont. under Manger; Cree ?nitcisuk; Fox midtcieal). Note that in Algonkin ti is used with the value tc in cakOsitiik les malade, JR. '1\ 24 (1642-43), p. 42 which is again re1niniscent of lVIontagnais orthography. Thus it is that it is very difficult to judge words such as a.sti "Terre, (JR. T. 7, p .. 154-JR. Q. I, 1634, p. 75). For in l\Iontagnais-Naskapi dialects there is considerable difference in the trentn~ent of -sic- before palatal vowels. In the l-dialects, and "mixed" n-dialects the combination appears as -ss-, in the u-dialects as Btc (-etc-) save where dissimilation is obvious. Thus Cree aski·v (in Bloomfield's transcription), Lemoine's Montagnais assi ( l-dialect), Bersimis (Michelson) essi, Seven Islands (~Iichelson) assi' (mixed n-dialect), Rupert's House (Michelson) a.stci', Fort George (Michelson) a..stci', Mistassini a.sfts ·t (In which s is intermediate between s and c). But at Natasbquan (unmixed n-dialect) I have recorded asN, ahtci': the latter is certainly in accord with the same type of phonology seen in tci'htemau' "tobacco," uhka't' "his leg,'' 1l·'hpUun' "the fore arm,H mihta'peu "giant," nuthtcehu." "fox", ihkwf}'1.t "woman", u'hptvaua.-n "pipe," M·i'ht-i na"'kw "master of the fish" (according to a Mingan informant speaking somewhat mixed the correspondent is Mictiniihk10, Great Whale River Mistcinahku, 1\Iistassini Mictsina'kw, St~ Marguerite m9sana' ~ [see Speck, Naskapi, pp. 117-119, 239], Plains Cree miskinlihk "tortoise'; Lemoine messinllk "Tortue"; Watkins gives Cree mi'kina'k, an obvious Ojibwa loan-word; he also gives ntisUna'k: yet I know no Cree dialect in which this occurs; Watkins once was on the east side of James Bay, and I wonder if he may not have picked up the word there). Moreover, the 1\'Iontagnais person'al name Oestchi-nisi case (JR. T. (,.g, pp. 60, 61: here -s- must indicate -tc-; 8 is used with the value 'W, etc.) "pretends to be young" ( 1nore accurately, .. he pre­tends to be a young man") shows definitely that the Jesuits of early times were acquainted with a dialect in which -sk- before n palatal vowel became -stc-; the form is a participle-like noun showing what is technically called "change:" note Lemoine gives ·uss·inft.3hu. under Jenne, Rupert's House (J\ilichelson) 1'.<.Jtcinitcu' "young man", Plains Cree -u;skiniki:w; Cree -hk0s6·w ·'he, she pretends." There­fore in spite of the fact that Le Jeune gives na·ma nikhirassi·n ie ne mens pas (JR. T. 7, p. 56; Moose Cree eklliis7citcilo "when they lie", Plains Cree nama nikiyliskin; though Watkins gives lcinasketo "he tells lies" it is not Plains Cree; cognates exist in Algonkin and Ojibwa; Lentoine cites a cognate Montagnais which has not the suffix of Le Jenne's word) it is fair to consider his kh·ikhlsteri­ten "tu s~ais" (JR. T. 7, p. 154 JR. Q. I, 1634, p. 76) as being due to dissimilation: compare Rupert's House t 8 isUJihta.m "he knows it," nitsi'stema-u "I know him,'' Fort George ntcistcai'mau "I know him," Natashquan tc,i.'hten.ta.m' "he knows lt,t' ntci'htenten "I know it," Lemoine tsh.isselitam "he knows it," Moose Cree kilc-iskelihtfut "you (sing.) kno"v it," n'lkiskelihten "I know it/' k-iskelihta-m "he knows it," nikiskiJlimaw "I know him." We have then clear evidence of a .t least two ~fontagnais dialects being used by Le .Jenne. Notice be also says

.94 BUREAU Olf AMERICAN ETHNOLOGY [BULL. 123

. (JR. T. 7, p. 80), "ils n'ont point les letters F, L, V, consonante, X.Z. ils prononce vn R au lieu d~vn L. ils diront l\Ionsieur du Pressi pour :h:lonsieur du Plessi, ils prononcent vn P au lieu d'vn V. consonante, Monsieur Olipier pour Monsieur Oliuier." It n1ay be that the l was not quite English l and so recorded .,.. We come now to a point which is vital as to whether or not some of the "~Iontagnais" were really Cree, using an r-dialect (Tete de Boule). 'iVe ·have seen that at least two dialects were confused; and we have noted variations in orthography. We have obser,·ed that oue ·whole sentence instead of being l\1outnguai~ is .Algon­kin. The reciprocal orthographic influence o.f Montn.gnais and Algonkin on each other has been pointeu out. The question no'\v at stake is as to whether k ( i}

ever was used graphically for tc. That this must be answered a:ffirn1atively is shown by Algonkin napik "entirement" (JR. T. 29, p. 88) which in Lemoine's

. transcription is napitc (see Diet. Alg. under '.£out) which has primary final tc as shown by Cree naspitc: see above. Now na·ma lchitirin-is·in tu n' as point d'esprit occurs on p. 62 of JR. T. 7 (JR. Q. I, 1634, ch. XII, p. 57) which in Lemoine7

S transcription is nanta tshitilinish·i'n (see his Diet. l\Iout. under Sage; cf. Plains Cree iyini~iw "he is wise," lcitiyinfsin uthou urt \Vise.'~ [That Plains Cree replaces original l by 11 is a matter of common knowledge.] Contrast this with N·igousa·i kesta kitiriniss·in (JR. 'J.'. 49, p. 66-JR. Q. III, 1664, ch. IV, p. 19), mistranslated ''mon fils, tu n'as pas d!esprit." Observe the word espi'lnitch ( "up\vard") which is diagnostically nothing but ~Ioutagnais-Naskapi immedi-ately follows: note Lemoine i&1t1Ji1nit.'1, ltupert's Ilouse ic;pitnilttc, Cree ispintihk, Ojibwa icpimiiig, Fox .&'petnegi. [I have said mistranslated because as far as I know there is nothing like lcesta "ne . . . pas" in n uy Algonquian language. If k has the Yalue tc it may he taken as the equivalent of Len1oine's tshista; see his Die. Mont. under MIJnz.e; the S(~use then i~, "thou also art wise." En passant, Le Jeune made a bad break in 1\Iontag;uais grannnar by employing the first per­'SOn pl. inclusive where the exclusive form was iu order; see Jlt. T. 8, pp. 264, 265. On the value of kh (i) see above.] Observe also the careless writing of -k­~or -kh- in khikiras8ln '·tu as menty" by LeJeune (JR. T. 7, p. 166-JR. Q. I, 1634, p. 78) but nan1.a nikhi'ras:1in "ie ne mens pas" (see a little above) in both of which -88- is diagnostically l\lontagnais: if lc or kh here n1Pant k we '\-vould then have a word half Cree and half l\fontagnais. Kno·wledge that k before palatal vowels regularly becomes tc in Montagnais-Naskapi even if such vowels are subse­quently lost, is taken for granted. And Le Jenne gives 1\.lontagnais n·ikhit~poun

(ie suis saol), JR. T. 6, p. 250, though he previously gave Montagnais n.ikispoun, JR. T. 5, p. 94 (Lemoine gives [n·i] tsh·ishpiin; Watkins gives Cree kespoo he is ftlll of food; Fox ki' pu.-). Again, according to the index to Thwaites ed. of the Jesuit Relations K-iaskou (JR. T. 56, p. 190-JR. J. III, 1672, ch. VI, p. 51) is the name of an Algonkin chief, 1neaning "gull." Linguistically this word can not be Algonkin (a similar error is M1chtaernikoii.an "Great Spoon": see above) : note Algonkin gaiaek (Lemoine), Ojibwa gaiashk (Baraga ) , ·m·iei-kaJJaelc "Great Gull" (Jones ; the pl. kayackV)fLg occurs in J's Ojibwa Texts II. 178.8), Cree keyask (Watkins), ki.yusk (Lacombe), 'Venusk Cree ki11a.-sk (1\iichelson), 1\Ion­tagnais tshiashk (Lemoine), Fort George tc'i' .A. . .sk1'. The phonology certainly terminally favors Montagnais as opposed to most Cree dialects. However, the treatment of original •-wa. after consonants iu Tete de Boule Cree, as I know ~rom the manuscri1)ts of the J{.ev. Dr. Cooper 'vlio most kindly has allu\ved me to make use of them, is most complex ; sometimes it is lost as in other Cree dialects~ sometimes it appears as _wa, sometimes as -" which recalls Montagnais usage; unfortunately the Tete de Boule correspondent to ordinary Cree kiyask apparently is not in hi~ manuscripts. Now as I have said previously, from a study of Dr. James Geary's recent manuscript it is clenr there are some Cree (presumably Tete de Boule) Joan wordE in modern Algonkin (I here thank him

JdlCHELSON j CREE .\:'\D ~iO).:T.-\GXAIS-::-.;ASI~.\ PI DIALECTS 95

for his generosity for permission to use it) ; so conceivably the name might be in one dialect bnt the person designated a speaker of another. A parallel is the name Winnesbiek, a 'Vinncbago: the name patently is Algonquian ( Sauk, }'"'ox~ Kickapoo in particular; the three are closely related). Hence a decision tree from all uncertaiuty can not be made. Different is the case of nitichenf­<-assouiniki "en mon nom" (JR. T. 7, p. 154) for the locative ouascou.khi "au ciel" occurs on the preceding line (Le Jenne, JR. T. 8, p. 40, gives the locative ouakoucki which must mean -tc as sho·wn by n1sadkiha.u on the same page as well as the plural ouperigoue ouaouokhi~· all said b~T the convert Anne). And the truly Montagnai.R eharacte-r of Le .Teune~s prayer in which nif-i{•henica.ssoui-nilci occurs (·an not be doubted. ~l."he vocaliRin of the suffix ( -ouin) differs fron1 that of J~moiue's l\1ontagnais and I lm vc• no means of kno\'\,.ing whether any Montng­nais-Naskapi did or does agree with Le Jeune•s l\Iontagnais in this respect as op­posed to Len1oine's ijin fkashun. In view of the undoubted orthographic recivro­cal iuttuence of l\Iontagnais and Algonkin. it is barely possible that Algonkin influence is to be seen in this. Good examples of l\Iontagnais orthographic in-1luence on Algonkin ·khikOsls vostre fils (JR. T. 24, p. 40) and Khi{Jatoutaoxitn mentioned above, for k is nonnH11y re-tained before ·i in the Algonkin of the ..Jesuit Relations, and it is impossible that kh in tlH .. ~e t\Yu partit~ular cnses wns pronounced aR tc. [On tlle same page as l\.11 ·ik8s-is ·will be found nikacht-ipltctga·n mon sac a Petun; thiR is a misprint or 1nisllea ring for 11 ilea ell ki~.1 In view of tbe fact that at times in an<"ient documents alleged "nativeH names of peoples nre not "native" names (I regret to say that even n1.odPrn entbnologists have at times committed similar errors) it is extremely diffienlt to judge the orthography of some names of peoples. Thus though Ot.tchestigouetch, Ou.chesUyuetch. Ouf~hestigouets are patently Montagnais in orthography (the singular Otchi,<::estigO fJR. T. 59, p. 60], patently 1\Iontagnais, is not in Bull. 30, Bnr. A1ner. Ethn.; tbe expected plural would be *Otchi:;;estit!Octcll apparently does not occur) and Ouch.estigo1.".iek pregutnably so be<>ause. of Jnedia1 -ch- and -k: with the value tc as above, Oukese.r-;tigouelr is not so clear: for this orthography would coincide 'vith Cree (-st- bar L\.lgonkin) if both k's have the value of k; or have we partial Al­gonkin orthography? That the people were Montagnais here is not in que8tion ; but if OukesesUgouek is the Cree denomination, then active acquaintance with T~te de Boule must be ns~umed. Again, should both k's be taken as standing for tct [For these names and other ~~-non~"ms see the synonomy at the end of the article Oukesestigouek, Bull. 30, Bur. Amer. Ethn., pt. 2: the reference to JR. Q. 1855, 1665, 5 should he corrected to 1665, 15; the etymology proposed by Gerard is on the right track but l\tiontaguais *u.tcisfstik'",* ·utc·isistigw.Ato rather than Cree forms must be at the bases of some of the Yariaut:;;. ConRult also the indices to Thwaites' ed. of the Jesuit Relations and the Quebec ed. of 1858.] From. the variants (plurals) Attka.·rneguekhi, A .tt.·i.('Unteoets, as 'veil as A.tUkamegou, the personal name of a 1\Iontngnais, JR. T. 9, p. 72, which in form i..<o::. Montagnais rather than Cree, the peopie are definitely l\Iontagnais; it is only a question of how the k of Attikarnegouek should be interpreted. As long as we have the variation kh k "·ith the value of tc in other '\Yords, as shown above, there is no reason '\'\-by this rna~7 not also be one. Yet this particular form orthographi­cally would coincide ~vith not only Cree but Algonkin. Similarly Oupa-pina­chiou.ek and OpapinachUJekhi, Ou1nanz.iu;ek, Onz.amiOek and OtnamiOekhi; etc. (See the various indices and various articles in Bull. 30, Bur. Amer. Ethn., for the exact passages in which these are to be found.) If I understand Dr. Cooper rightly, his position is that though kh ( i) sometimes has the value tc, this is not invariably the case. The fact that Tete de Boule is a Cree r-dialect also has Influenced him.

0