'IJIMIVE'RS S-1 A-4 L- PJLO ,i --AA CHA :; GY- A A0 Y V;.ot 12,fl. , pp. 339 6 -- - Ma -28, 19 ; --,, - . . ' 'em'''' -Al-.-A C LEONA KIN&H111 A. - f .< .>. -XCR4 BE I-A~~~~~~~~~~~~7 " v v J t _~~~~~~~~~~~~~~~~~~~~1 A--Al :X~~~~~~~~~~~~ a X X f0 Rt Re _ di j~~~~~~~AA A - :'.r~~~~ ~ ~~~ A'^ ' A A A . . . ..'l-- : . - * . .. .. . 7, . . - -r sVsrI -. ' ,i ?' '' ZS ,7.s'At_' .o E 2 . - . .- t s -; -- Jr .; < a ayn t t? - - bs b X _? ' t > . r 0 ,-T -- 0'?.'-SP '? ;= '-'f''\ _ }-E ,L ; * s d S \ - F K X z o - L F ;, Y -; - ' < *<,,. .* < .* - a? t v (.,, e * __; _ 8 g > _ ? '^, -'. . . - rf i- E ' ' S -- s 1 S ' . - - . ,. .!; - - , ;.o. 8 X . r. - m FJ- ;wL W;eX '' 4 ' "'' a<, -Sa A .,; ', ' ..1 ' as ' As . . . ,: j' _ e . /e i ,; dr t \ tt- ;: A , X, .4 . 4''N 7 ,-. r f . t- e e ) . : J i , 4 i s-, z ^ t _ 6 F s J ;} > {; ,.... .f * - y Jr ,-,, D *4 ;. _ 1-?? -A I -71 ", I - I - _.- -:,? I 1- , f'- - -;-% I .- . . 1- - I I i?i * ' A A _r A ?4, - . r. .,,- r . St. . , ' h C l. ' I- iI I ,I . I" . I : * '. Yr . I I , t .w ", , .NIV T OF.. OAFOP P LOATiONA ONO DPAET3~ZNT 0 ANTlVOOLOY X W f f * ;- t Q ; _ s ; 0Air it ; ' ; *GY The follwipbi .cio dealigg h logcal and Aeth-oogia ue isud; -mdthe directi6onof Deatn o t logy are aient in ecangoforthe publi- cations of irhropologic4I deparneats aadAmuseuis, ai&I f jouirnals devoted' to fgenea antfiropology or to arihieology ad ethnology. ty are for sale at sthepris stated. Exdchanwge;Ahou]4 be dreced to- The ;change DepartmAt, Unhiversi- Libray, Beeley, Califoriu,l (. 8.. A Al orders and riwittasues- hould bbu addreed to tUn ity of Caforzia Prss -- --;-;-*;--0 Eluropea t for.-N theie.i Aneric Arhaeoloa>,. audEnolog ?:1s4ical Phi- *I087j IductioW Modern oilogy)PilOPh, a Seti Philoogy, Ott arrsowin,- -Loipsig. F Zor te seiean ta-,t Geoo Ptog y Phyiolgy, Solo and l :r-; iCan Archaeology and Ethnlogy, Z laeder -& VSohu ~BBrin . -- la, -AI~IC;AN AECI.bROLOGY AND ET.NOLOGY.-A ier, Editor.- - lces V'-- -- -tohlmne 1,-$4.25; Voluimea2-to 11 Inclusive, $3 each; Volme2A. following R 15?. each- - . -. ;-7 ; 'C' -d as U "-v. OaI.if2P, bl..Am.bArch. iEl- . : - F ries VoL 1.<-~ 1. Life and -C orfi thpa, b -Pn Eatle.G-drd P 1-88;^ - .. .~~lae l.$ 8eteX iber,1903. .......... .......-.....-...~.P.25.............= -2~ u <2 -Texts, by P1lI I Gd. p 89448. Mach 14 - 3.50 - -;- : 2. The Lanageott- oas ot CaliorniSoutizof\ San Tricl,-coj b- - , A L. KUrobe. P.t 29-_0 t a rap Jiui 1904 -m m0 _. Types ,dia ut in - ti A. L.Ke. . 81403. - ;- :.--. - ;: Jun. 1- 904. .~.......... .~..-. .....* .. . ...... ..2 0 ' . - :: ->- . -Easke l)e3gnsi of; the Indan of ortweer Caliornia, i -A. L.. : . Nr eb r Pp *0~l4 =lte 21521 Jan ary 19 6B 1i 0 5 .~.............................. ;- .7 1!-The outs Luage of South a alifoni, by A. . Eroeber. =3. ; - T ne -p 8 7f-359 . Uc -; -Q M-X-: Torpff0holo0gy of ~th Eupa LIanguage, -by Pug* Eale ffod rd. -f- 344pp -un, -r b PP..J.....:.--.;.- - VoL.4i; 1.-ho -Earliest b eti tco - J r.- - -- -.- . oigbra docuer peservd n Sai a~d Japan,- by &lix 2Nutall. --a / 0- ; <.- ~2. CoiWbuto to thXe Physlcal Authrwpology oQf C liorilii .0se on ;.l-- .. . ,,;- ,- ~s ectionS inX the Departmsqn,;-gt of AntbroSoog of t,. -~iest of- - ~- - -- - California, rd 4a the U. S. National M * n, by rd1k. -- -b-> -P a-i-X . 49Pp . 904, with 5 talAeG;$t -0. d6ma. J i9:. ... . r-6L M. T-e Shoshlevan of1 alif byAL xeber. Pp. o16r & Y~bruary, 1907 .w.. . .......... a1a0 o,_ s t'- _ ~sm * tand" d- td -W \ * - - :.ida Myh frtwm -ot C-ntrM Calfoni, by. A. L. -.ober.i-p. The WasloLangWneof lst*nb Californa an 6-a,byA. L. t '-' ,roeber . IAthp -514 e 1907 4 -4 4 475 2 8.-The 'alo of thendia-o fa by A. rbe Pp. 819 -; r 0 - - nde 6 ap 857 -374 . 0-VoL .5. 1. ThePhnlgy of-te L iu & afgae PartI, 1Te. tadividnl ud,. - -bi Pln Eadle Gd-dr Pp. 1to- pla 8 March, 1q0 - r X - Navah ... + Eo Mybs :aers and d SnS,o wit lTexI an -T bnslet -s * W- bnt2 -ateO - it 4 by Pliny Eale (*dt Pp. 2143 ; - - 0 >8, nato *bTexts, PlnEarle- Gddard. p 05.238,platO 9.D?eemb -:r - . *------b+--t^-- ----- - . - w - . .... . x - 0 ;4L The -Materl C7ltr of th Wlat La >ke'i anZdr0 Mo- Inis -of --v-- *0 f-9 ]> 92, pla'tea 19.5 ffestu1e, 19.( ...............--..... 41 - .- - ugust IZ ,..:..; 1 * 0 .~~~~~ne -pp. 0 881f-; 84.X -,r'' .' Vol.8 1@ . Thetno-GopIy o te oroam NI~eIgIdias,.by Sam- 0'.V -t ~ ~~ A1r0 - ~ret ?p. 142 ru-p 1- Febrasry ,1,O .,.-.p* 8.5>0 - , ' ' 2.N T,he-^Ge4graphy adPialecz of ailL oknians, by Saniel Alfred -- j~~~aret 'P-p. 53'3.36, ma Z l - S - ** f ,, ft ,LX ,, ,X, ,beito,> sw -e- -f 0~~~3O th Edeo of th OcuatOO Ceti (E*S- 8egor (by> th Mss;w. ok \ - S - -;.-,, . Inias, by A. Ii ~-fos~ web . *o p.> *S6- 380.- r- oe.-- 2w and S --, ol -. , * e . - In 0d, p. 8 0C- ETHNIC GROUPS. (Referred to in the text) 1. Mohave. A 2. Luisefio. 3. Southern Yokuts, YauelmanL 4. Southern Yokuts, Yaudanchi. 5. Kawaiisu. 6. Tiibatulabal. 7. Central Miwok. 8. Northern Paiute. 9. Washo. 10. Southeastern Wintun. 11. Eastern Pomo. 12. Yuki. 13. Yurok. 14. Karok. 15. Hupa. 16. Wiyot. 17. Chimariko. 18. Costanoan. 19. Salinan. 20. Chumash. I. a _,a~ 0 qtN4 44 % UNIVERSITY OF CALIFORNIA PUBLICATIONS IN AMERICAN ARCHAEOLOGY AND ETHNOLOGY Vol. 12, No. 9, pp. 339-396 May 28, 1917 CALIFORNIA KINSHIP SYSTEMS BY A. L. KROEBER CONTENTS PAGE Introductory ........................................................... 340 Mohave ------------------------------------------------------------------------------------------- 340 General Fcatures -347 Luiselo --348 General Features -351 Yokuts -352 Relations of Miwok and Yokuts ...........................6.............. ........... . 356 Northern Paiute - --....._358 M ar riage6 1 . .......................................................................... ......... 358 Washo ............- ......... 362 Relations to Northern Paiute ......-----------363 Relations to Other Systems ..... 364 Tiibatulabal and Kawaiisu ..... 366 Shoshonean Systems ----- 366 Wintun ................-3........ . 368 General Features . 369 Pomo ...---------------------------------------------------------.....---..----.....................................................370 General Features and Relations ..........................................................-................ 371 Yuki ----- 372 General Features and Relations . 373 Yurok ..................- . 374 General Features -375 Three-Step Relationship .................. 376 Classification of the California Systems - 378 Kinship and Type of Culture .380 Kinship and Social Institutions ..........................- ...... 382 Summary ............................... . 385 Theoretical Considerations .................................--5.................... 385 340 University of California Publications in Am. Arch. and Ethn. [Vol. 12 INTRODUCTORY The following systems of relationship designation were collected at intervals during the last fifteen years, but mostly before 1908, in the course of various field studies of the California Indians. They are in most cases unsupported by genealogies or concrete records; probably the majority of the lists are not wholly exhaustive; and in a few instances the data may not be entirely correct. I have long hesitated to publish this material. But it happens to represent all ethnic provinces and parts of the state, except the northeastern corner, and therefore permits of distributional inferences; and it furnishes a basis for the consideration of certain theoretical problems; in addition to which, information on kinship in California has become a need in wider comparative studies. I therefore present the data, trusting that they will be of service in spite of their imperfections. MOHAVE The Mohave system is an elaborate one. It contains a considerable number of terms; and the principles according to which these are applied are sometimes complex. N- or ny- denotes "my." A faint initial h-, of the same meaning, has been omitted from most words not beginning with n-. Parent Class N-akut-k, father of a male. N-a 'ai-k, father of a female. N-tai-k, mother. H5-uma-i-ch, man's son. Compare humara, child. J7uchi, man's daughter. Ith'au, woman's son or daughter. Iki-ch-k, man's stepfather; reciprocally,1 man's stepson; also, father's mother's brother, mother's mother's brother, and reciprocally a man's sister's child's 1 Reciprocity is logical or conceptual between terms that are complementary in meaning; as, Mohave namoik, mother's younger sister, and inoik, woman's older sister's child. Reciprocity is verbal only in Zunfi nanna, grandfather and grandson, because the complementary concept to grandfather is not grandson but man 's grandchild. Reciprocity is conceptual and verbal in Yokuts t 'uta, mother's mother and woman's daughter's child. Reciprocity is conceptual and approximates verbal completeness in Luisefio tu', mother's mother, and tu '-mai. woman 's daughter 's child, in which -mai is a diminutive. Terms which are conceptually and verbally reciprocal may be designated as self-reciprocal. Conceptual reciprocity without verbal identity is commonest between relatives separated by one generation, most frequently in the uncle class, but also in the parent and parent-in-law groups. Verbal reciprocity, identical or derivative, is usual only between relatives that are of the same generation or separated by two or more generations, especially those in the grandparent and brother-in-law 1 Kroeber: Calif ornia Kinship Systems child; also a man's younger brother's son's son and a woman's younger sister's son's son; also a man's son's son 's son, and father 's father 's father 's father, that is, great-great-grandson and great-great-grandfather reciprocal in the male line. Iki-ch-k is a term used chiefly by males of males; it never denotes a person of one's own generation; and it always implies remote kinship-a lineal relative four generations distant, a collateral relative two generations away, or a man one generation removed who is not a blood relative at all. A mani's stepmother and a woman's stepson are denoted by unyi, whose full range of meanings is given under terms of the parent-in-law class. What a woman calls her step-parents, or what either a man or a woman calls a step- daughter, I did not learn. Brother Class Inchien-k; older brother; older sister; father's younger brother; woman 's father's sister's son or mother's brother's son, that is, male cross-cousin of a woman; man's father's father's father's younger brother's son 's son's son, that is, a man 's male third cousin in the pure male line of descent, sprung from the younger of two brothers; also, a man's son's son's son, that is, his great- grandson in the male line. The last two meanings are evidently connected, since third cousins are great-grandchildren of brothers. Isu-ich-k, younger brother; man's older brother's son (and daughter?); man's male third cousin in the male line, sprung from the older of a pair of brothers; father 's father 's father. In the last two senses isu-ich-k is reciprocal to inchien-k. Inya-k, younger sister; man 's father's sister's daughter or mother's brother's daughter, that is, female cross-cousin of a man, reciprocal to the corresponding usage of inchien-k. Oyavakiau-k, man 's paternal half brother or half sister. Tav 'alyvi-kc, man 's maternal half brother or half sister. If any separate terms for a woman 's half brothers and sisters occur, they have not been recorded. Grandparent Class N-apau-k, father's father. N-akweu-k, mother's father. N-akau-k, mother's mother; also her sister. N-amau-k, father's mother; also her sister; also the father's father's sister. If the last meaning is not an error, the generic meaning of n-amau-k is: female relative of grandmother generation on the father's side. It might be inferred that n-akweu-k analogously denoted males two generations older on the mother's side; but the relationship of mother's mother's brother is expressed by iki-ch-k, whose primary meaning seems to be step-father. A 'ava-k, son 's child, and therefore reciprocal to n-apau-k and n-amau-kc jointly; also, a woman's father's brother's son; man's father's brother's son or daughter; woman's brother's or sister's son's child. classes, but occasionally between brothers and sisters also. The foregoing, at least, are the tendencies in California, with exceptions occurring chiefly in the extreme southern part of the state. On the whole, the distinction seems to be adhered to in other regions also, but precisely to what degree remains to be determined. Reciprocity that is verbal but not conceptual is very rare or wanting in California. In general, therefore, it may be stated that reciprocity is always conceptual in this area and frequently verbal also. 1917] 341. 342 University of California Publications in Am. Aroh. and Ethn. [Vol. 12 Ahko'o-k, woman 's daughter 's child, reciprocal to n-akau-k; also, woman 's mother 's sister 's son and man 's mother 's sister 's son or daughter; woman 's sister 's daughter 's son; and, presumably, by analogy with a'au-va-k, any child of any nephew of a woman, though this wider meaning was not recorded. Ahkyo-k, man 's daughter 's child, reciprocal of n-akweu-k. It is not certain that this term is distinct from the last. There is a curious change of generations implied in the primary or simplest meanings of the terms used to denote relatives beyond the grandfather. Thus in the pure male line: Grandfather, n-apau-k, is father's father. Great-grandfather, isu-ich-k, is younger brother. Great-great-grandfather, iki-ch-k, is stepfather or grandfather's brother. Uncle Class N-avi-k, father's older brother; also, father's father's younger brother; also, of two male second or fourth cousins related wholly in the male line, the de- scendant of the younger brother calls the descendant of the older brother by this term, reciprocally to ivet-k; but as between third cousins the corresponding terms are isu-ich-k and inchien-k. The father's younger brother is called one's own older brother. N-athi-k, mother's older sister. N-amoi-k, mother's younger sister. N-akwi-k, mother's brother. N-api-k, father's sister. Ivet-k, man 's younger brother's child or woman's younger sister's child, and thus reciprocal to n-avi-k and n-athi-k jointly; also, male second or fourth cousin related wholly in the male line and descended from the older of two brothers- reciprocal in this sense to n-avi-k. A man 's older brother 's child is called isu-ich-k, " younger brother," recip- rocal to inchien-k, older brother or father 's younger brother. Inoi-k, woman 's older sister 's child, reciprocal to n-amoi-k. Evany-ic, man 's sister 's child, reciprocal to n-akwi-k. Emarepi-k, woman 's brother 's child, reciprocal to n-api-k. Parent-in-Law Class Nya-halye 'au-k, man 's daughter 's husband, wife 's father; that is, self- reciprocal term for father-in-law and son-in-law used by males only. Unyi-k expresses all remaining relationships in this class, besides several others. It denotes: woman's father-in-law; woman's son-in-law; mother-in-law; daughter-in-law; husband's brother or sister; brother's wife; man's stepmother; woman 's stepson. Itmumavenya, said to mean "who eats with you," is used in place of unyi-k after the death of the connecting relative, at least in the cases, and they constitute the majority, when this person was a male. Brother-in-Law Class Amily-k, wife's brother, man's sister's husband; that is, self-reciprocal term used by brothers-in-law. This term is also used by men to denote the husband of any collateral female relative. Inya-huvi-k, wife's sister; woman's sister's husband. Self-reciprocal. 9Kroeber: California Kinship Systems The remaining four of the eight relationships in this class are expressed by the blanket term unyi-k. The Mohave make the general statement that a man calls any female relative by marriage unyi-k, and is so called by her. This is nearly true: the only exception is inya-huvi-k. All affinities by marriage are expressed by the foregoing four terms, whose range, however, is very unequal, as a summarization reveals: Male connections of a man (of his own generation amily-k; lof another generation nya-halye 'au-k. A woman's husband and her sister call each other inya-huvi-k. All other male connection of a woman All other female connections of a man qunyi-k. All female connections of a woman Husband and Wife Ichu-ich, husband. Nya-ha'aka-ch, wife. Cousins I obtained three terms for cousins: Dhohumi-k, man's father's brother's son. Chasumav-k, woman's mother's sister's daughter. Chakava-k, man's father's sister's son or mother's brother's son; that is, a self-reciprocal term between male first cross-cousins. The Mohave terminology for cousins is as interesting as it is complex. Besides the foregoing three specific terms, there are four others from the brother and grandchild classes; but parent and uncle terms, which are found in certain other Californian languages, and among a number of Eastern tribes, are not employed. The following tabulation brings together all the data. Children of Brothers Male calls male, dhohumi-k; Ifemale, a'ava-k, son's child. Femle callt male, a'ava-k; e ma e ifemale, not obtained; analogy suggests a 'ava-k. Children of Sisters Male calls male, ahko'o-k, daughter's child; female, ahko 'o-k. Female calls male, ahko 'o-k; female, chasumav-lc. Children of Brother Call Children of Sister Male calls male, chakava-k; Ifemale, inya-k, younger sister. Female calls male, inchien-k, older brother; Ffemale, I Children of Sister Call Children of Brother Male calls male, chakava-k; tfemale, inya-k. Female calls ~male, inchien-k; Ifemale, ? 1917] 343 344 University of California Publications in Am. Arch. and Ethn. [Vol. 12 Briefly, the children of brothers call each other "son 's child," except that a special term is used when both are males; and the children of sisters call each other "daughter 's child," except that another special term is used when both are females. Cross-cousins of opposite sex denominate each other "younger sister" and "older brother," according to sex; the "younger" and "older" seem quite fixed irrespective of the actual ages of the persons or the age or sex of their parents; that is, a man Is female cross-cousin is always his younger sister, and a woman 's male cross-cousin is always designated as a brother older than herself. Male cross-cousins denominate each other by a special term. For female cross-cousins there is unfortunately no information. The basis of this remarkable plan is that cross-cousins call each other brothers and sisters, parallel cousins designate each other as grandchildren, and specific cousin terms are restricted to the cases in which all the persons involved in the relationship are of the same sex or in which the children of brother and sister are of the same sex. The terminology used between remoter cousins is equally extraordinary. This has been obtained only for the male descendants of two brothers. Brothers Older Younger l l First cousins 1 2 Second cousins 3 4 I I Third cousins 5 6 Fourth cousins 7 8 1 calls 2: dhohumi-k; 2 calls 1: dhohumi-k. 3 calls 4: ivet-k, man's younger brother's child; 4 calls 3: navi-k, father's older brother. 5 calls 6: inchien-k, father's younger brother; 6 calls 5: isu-ich-k, man 's older brother 's child. 7 calls 8: ivet-k, man's younger brother's child; 8 calls 7: navi-k, father's older brother. Fifth cousins, it may be surmised, call each other like third cousins. All these terms are conceptually reciprocal. It will be noted that the actual age of any cousin is immaterial. The terminology is fixed by the respective ages of the brothers from whom the reckoning starts. On this basis, and the assumption that uncle-nephew terminology is to be employed, it seems natural that the allotment of names between second cousins is on the plan that the descendant of the older brother is the "uncle"; but it is surprising that between third cousins it is the descendant of the same older brother who is reckoned the nephew. The explanation may be in the fact that inchien-k means older brother as well as father's younger brother, and that therefore I apply to my father's brother (if he is the junior) the same term which he applies to my father. Something of the idea inhering in this terminology appears to have been extended along the descending line of cousins, with the result that whatever my cousin of my own generation calls me, I call his father or my son calls him, Kroeber: Calif ornia Kinship Systems for inchien-k or navi-k; whereas for isu-ich-k and ivet-k, my father calls him or I call his son whatever he calls me. Thus: y > o, I > y: inchien-k 5 > 6, 6 > 4: inchien-k 4 > 3, 5 > 4: navi-k 6 > 5, 5 > 8: isu-ich-k 7 > 8, 6 > 7: ivet-k and, it may be suspected, 2 > 3 : ivet-k as 3 > 4 9 > 8 : navi-k as 8 > 7 Reciprocity Reciprocity is very strongly developed in the Mohave system. It is manifest in practically every class of terms. Self-reciprocal, that is, reciprocal both conceptually and verbally, are iki-ch-k, with a wide variety of meanings, but all falling into pairs that are exactly complementary; unyi-k, of which exactly the same can be said; the three other terms for connections by marriage: 'nya-halye-'au-k, amily-k, and inya-huvi-k; the three specific cousin terms; and a'ava-k and ahko'o-k as used between cousins. Conceptual reciprocity without verbal identity occurs in the terms used between parents and children; between grandparents and grand- children; between all uncles or aunts and their nephews and nieces, and between brother and sister terms as used by cross-cousins. The only irregularity is that, in the grandparent class, a'ava-k, son's child, is reciprocal to both n-apau-k and n-amau-k; and simnilarly in the uncle class, ivet-k to n-avi-k and n-athi-k. The only terms that are not reciprocal are the three for brothers and sisters, when used in that fundamental and unextended sense; and possibly those for half brothers and sisters. A similar degree of reciprocal expression seems to pervade the kinship system of the Papago of southern Arizona. Except for Yurok and Wintun, all known systems in California are more or less recip- rocal; but none are so extreme in this respect as Mohave. Relation to Clan System The Mohave possess a clan system similar to that of several other Yuman tribes. It is patrilinear, exogamic, and totemic, though its totemism is veiled: the clans themselves have no names, but all the women of one clan bear the same name, which carries a totemic implication or connotation. 1917] 345) 346 University of California Publications in Am. Arch. and Ethn.- [Vol. 12 I am unable to discern in the kinship terminology any definite reflection of the division into exogamic units. The cousin nomen- clature is an example. With unilateral descent, if the children of brothers are of the same clan, the children of sisters must normally be of different clans; yet the actual terminology is exactly parallel. The children of brother and sister, again, must necessarily belong to 4 O / different clans; yet it is only/these thht, cousins brother-sister names are applied. The frequency with which the sex of an intermediate relative is denoted by Mohave terms may seem an indication of the unilateral reckoning of descent in the clan system. But this is offset by the instances in which collateral kindred are not merged in lineal, as is often supposed to be the normal practice where exogamic groups prevail. The partrilinear reckoning of the Mohave, on the other hand, may have led to their making certain distinctions among males that are not made for females. Thus there are two words for father, only one for mother; a man uses different words to denote his son and his daughter, ? woman only one. The primary meaning of the generic term iki-ch-k seems to be stepfather, whereas the only word for step- mother is unyi-k, whose fundamental denotation is a female affinity or the affinity of a woman. The terms of the parent-in-law and brother-in-law classes reveal a marked asymmetry in favor of males. There are two words denoting the male affinities of males, and only two to express the three times as numerous female affinities of females and those between males and females. The terms which my informant, who, although a man, was assisted by several women, failed to mention are in every case those used by women or applied to them: stepdaughter; woman 's step-parent; woman's half-brother or sister; woman's female cross-cousin; second or remoter cousin, either female or descended wholly or partly from females. Since all the parallel terms for males were usually volun- teered, it appears that the Mohave think and express themselves first in terms of male lineage. There are only two cases of the finer distinction being drawn on the female side. The daughter's son and her daughter are distin- guished, the son 's children classed together. There is a term for mother 's younger as well as mother's older sister, but the father's younger brother is merged in one's own older brother, and the same for the reciprocals. 1Kroeber: California Kinship Systems GENERAL FEATURES Apart from the overwhelming inclination toward reciprocity, the distinctive features of the Mohave kinship system are the following: Relatives of the most diverse generations are denoted by the same terms. This is not on the plan of many American systems that if I call a relative, such as an uncle, by a certain name, I apply the same name to his son, grandson, and so on ad infinitum, that is, to my cousin, cousin once removed, and the latter's descendants; or that the word for grandfather is simply made to include the great-grand- father. The principle or principles followed in Mohave remain rather obscure; but the one point emerges with certainty, that the Mohave are normally at pains to use terms of the most clearly discrete signifi- cance as to generation, for their kin of adjacent generations. Thus first cousins are called grandchildren, not uncles; the great-grand- father is denominated younger brother; and so forth. It would seem that, the wider the leap, the more satisfactory the terminology; pos- sibly because an element of confusion is thereby minimized. In fact, it might almost be said that it is only in a technical and narrow sense of the word that there is ignoring of generations. As regards the distinction of collateral from lineal relatives, the Mohave are unusually precise at several points. Parallel uncles and aunts are not merged with the father and mother, nor nephews and nieces with children. Three-fourths of all cousins are designated by terms other than brother and sister. Sex of the intermediate relative is specified in practically all words into which this factor can enter: grandparents and grandchildren; terms of the uncle and aunt class; cousins; and half brothers and sisters. Some may see in this prevalence an influence of the clan system. To me it seems rather associated with the tendency toward reciprocity. Expression of both the sex of the speaker and the sex of the relative denoted tends to lead to a great multiplicity of terms if consistently carried out in a reciprocating system, especially in the grandparent and uncle terms. The Mohave solve the problem in the usual way: they express one category in the terms applied to the younger rela- tives, the other category in those for the older relatives. Both factors are specified in the self-reciprocal terms of the stepfather, cousin, parent-in-law, and brother-in-law classes, and in those used between a father and his children; whereas the term for older brother-sister, 1917] 347 348 University of California Publications in Am. Aroh. and Ethn. [Vol. 12 the word unyi-k, and a few of the nephew-niece and grandchild desig- nations-especially if their extended meanings be included-are wholly indeterminate as to sex. The distinction of absolute age within one and the same generation follows an irregular course. It occurs between brothers and sisters; is lacking for half brothers and sisters when these are specified as such; is made for parallel uncles and aunts and disregarded for cross ones, and the same for their reciprocals; is wholly wanting among first cousins; but always, though indirectly indicated, so far as the evidence goes, for remoter cousins. Affinities by marriage are never merged with blood kin. Iki-ch-k would be an exception if the stepfather relationship were counted as belonging to the former group. From the point of view of development of terminology for the several natural groups of kindred, salient features of the Mohave system are the consolidation of designations for marriage connections into a very few words, and the development of an elaborate nomen- clature for cousins, including at least three specific terms in a total of seven or eight employed for first cousins. LUISERO The Luisefio are of Shoshonean stock, but live in an entirely different social environment in their southern California home from the distantly allied Tiibatulabal and Kawaiisu of the Sierra Nevada, whose kinship systems have been described by Mr. E. W. Gifford,2 and from the still more remote Northern Paiute treated of in the present paper. The Luisefio terms are not used in their absolute forms as here given. In actual speech they occur only with possessive prefixes, such as no-, "my." The ending -mai is a diminutive. The system has been recorded independently and without dis- crepancies of moment by the late P. S. Sparkman and myself. The former's list of remote and extended applications of terms is some- what fuller. Parent Class Na', father. Yo, mother. Ka-mai, son. Shwa-mai, daughter. 2 Present series, xii, 219-248, 1917. Kroeber: California Kin-ship Systems Brother Class Pash, older brother. Kes, older sister. Pet, younger brother. Pit, younger sister. Grandparent Class Ka', father's parent; also, brother of the father's father and sister of the father's mother; also, woman's father-in-law, and, reciprocally, man's daughter- in-law; also, woman 's daughter-in-law; also, man 's brother 's son 's wife and woman's sister's son's wife, that is, parallel nephew's wife.3 Ka'-mai, reciprocal to ka' so far as this denotes persons of the grandparent generation; that is, son's child, man's brother's son's child, woman's sister's son 's child. Kwa, mother 's father; mother 's father 's brother. Kwa-mai, reciprocal to kwa; that is, man's daughter's child, man's brother's daughter's child. Tu', mother 's mother; mother 's mother 's sister. Tu '-mai, reciprocal to tu'; that is, woman 's daughter 's child) woman 's sister 's daughter's child. Piwi or piwai, great-grandfather or great-grandmother, apparently in any lineage. Piwi-mai, reciprocally, any great-grandchild. Sosa, great-great-grandparent or great-great-grandchild. Yuto, a person removed one generation farther than the sosa. Taula, one generation more distant than yuto, that is, great-great-great-great- grandparent or child. The terms for ancestors or descendants from three to six generations removed are evidently convenient devices for expressing the lapse of generations, and little else. They completely ignore the factor of lineage which is denoted in the grandparent terms; are sexless; and, it may be surmised, are applied indis- criminately to lineal and collateral kindred. It would be interesting to know their etymologies. Kek, grandmother 's brother; grandfather 's sister; reciprocally, man 's sister's grandchild, woman 's brother 's grandchild; also, man 's brother 's or woman 's sister 's child 's spouse. Specific terms for kindred removed by three steps of relationship-other than of the speaker 's own generation or three generations lineally removed from him-are rare the world over. This particular term is so far unparalleled in California. Uncle Class Kmu, kamu (nu-kmu, cham-kamu), father's older brother. Kmu-mai, kamu-mai, reciprocal, man's younger brother's child. Mash, father 's younger brother; also, stepfather. Mai-mai, or me, reciprocal, man 's older brother 's child; also, man 's stepchild. Nosh, mother's older sister. Nosh-mai or nos-mai, reciprocal, woman's younger sister's child. Yos-mai (evidently from yo, mother), mother's younger sister; stepmother. Kuli-mai, reciprocal, woman's older sister's child; woman's stepchild. Tash, mother's brother. 3 SiC, in the data available, although this signification overlaps one of those given for kek below, namely, parallel nephew-niece's spouse. 1917] 349 350 University of California Publications in Am. Arch. and Ethn. [Vol. 12 Mela (compare mai-mai, me), reciprocal, man 's sister 's child. Pa-mai, father 's sister. Ali-mai or ala-mai, reciprocal, woman 's brother 's child. Parent-in-Law Class Kwa pa-na, man 's father-in-law; man's son-in-law. Self-reciprocal. The literal meaning is "my daughter's child its father. " The term therefore really denotes the son-in-law, and its apparently absurd application to the father-in- law must be due to a conventional extension under the influence of the tendency toward reciprocity. Tu' pa-na, man 's mother-in-law; woman 's son-in-law. Self-reciprocal. Liter- ally, daughter's child's father. The secondary application is again to the older person. An extended meaning is woman's sister's son-in-law. A woman calls her father-in-law ka', paternal grandparent. Possibly this stands for "my child's father's parent." The father-in-law in turn, and the mother-in-law also, apply the same term ka' to their daughter-in-law. A woman calls her mother-in-law ka' shungal, "father's parent woman," or, "father-in-law woman." It is not certain that the qualifying shungal is always added. Na-hwa, parent of child-in-law (like Yokuts makshi, Miwok maksi). The term is also said to be applied to children-in-law; and to "the nephew's" parent-in- law. The latter meaning seems inconsistent with the prevailing Luisefno principles of designating kindred. Brother-in-Law Class Tolma, woman's brother's wife or husband's sister; that is, a self-reciprocal term between sisters-in-law. Exactly equivalent to Mohave inya-huvi-k. The etymology may possibly be from to'ma, wife. Mes pa-na, all other brother-in-law and sister-in-law relationships; that is, woman's brother-in-law and any immediate affinity of a man in his own generation. There is no independent word mes in modern Luiseflo. Me or mai-mai, reciprocal to mash, denoting a man 's older brother 's child, cannot be considered the source, for me pana, " my older brother 's child 's father, " would only be a meaninglessly roundabout way of saying " older brother. " The derivation must therefore be from mela,4 man 's sister 's child. Mela pana, man s sister 's child 's father, would therefore denote a man 's sister 's husband. Evidently the phrase was then used reciprocally for wife's brother; and finally extended to include the other relationships which it denotes. Husband and Wife Kung, husband. Pewo, husband, literally, "partner"I or "mate." Shnga-ki, wife, from shunga-l, woman. To'ma, wife. Aki, co-wife. At least in address, however, " older sister " or " younger sister" is usually substituted when the personal relation is amicable. Cousins Parallel cousins are brothers and sisters. Whether they are older or younger depends upon the respective ages of their parents, not of themselves. Ukshum or yuksum, any cross-cousin. 4 Perhaps the same stem me plus noun ending -la, -1; and mes for mesh in composition (compare nosh and nos-mai), mesh being me plus another frequent noun-ending -sh or -cha. Kroeber: California Kinship Systems GENERAL FEATURES The Luisefio system closely parallels that of the Mohave. There is the same dominant inclination toward exact reciprocity, made even more striking by a greater prevalence of verbally reciprocal terms. The tendency affects practically all the terms of the grandparent, uncle, parent-in-law, and brother-in-law classes in both languages; Mohave adds parents, and Luiseiio cousins. Another fundamental common feature is the limitation of terms to designate connections by marriage. The Mohave plan is the use of a very few self-reciprocal words of narrow range plus one term that covers all other affinities. The Luise-no appear to employ no radical words at all for affinities (the special term for woman's sister-in-law is very likely a derivation from "wife"), except the somewhat generic nahwa, but help them- selves out with circumlocutory phrases which are as purely descriptive as the corresponding English ones; or by boldly extending the meaning of terms for blood kindred. The degree to which the various factors entering into kinship are given expression by the two tribes is also very nearly the same. And, finally, there are special resemblances, as in the separation of parallel uncles and aunts into those older and younger than the parent, whereas cross-uncles and aunts are not so distinguished. The one important divergence is in the terminology for cousins, in which the two systems follow radically different methods. Among special peculiarities of Luisefio is the employment of literally self-contradictory phrases of transparent mehning for many connections by marriage, as the obvious result of the reciprocal in- fluence. This trait has some analogues in Northern Paiute, though there it takes the form of a wrong implication of sex and the cause appears to be mere simplifying assimilation. In both instances, how- ever, it is purely descriptive terms that are logically misused. This point is of considerable theoretical interest. If affinitv terms which on their face denote one thing, and that alone, are used in other senses from merely psychological causes, such as tendencies toward reciprocal or simplified expression, the presumption is that terms for blood kindred are also sometimes radically altered from their original meaning under the stimulus of similar causes without any accom- panying change in form of marriage, kind of descent, or social insti- tutions. The only difference is that transparent descriptive terms allow us to prove without doubt that the extension or alteration of 1917] 351 352 University of California Publications in Am. Arch. and Ethn. [Vol. 12 meaning has taken place in a particular case, whereas when we are confronted with unanalyzable stem-words the same sort of evidence can rarely be brought. But a very high probability must remain that a certain proportion of even the most elementary and important terms of relationship the world over have derived their present significance from causes not connected with form of marriage or descent. Other unusual traits of the Luiseiio system are the occurrence of terms for lineal relatives three to six generations distant; for the whole class of cross-cousins as a unit; for a child-in-law's parent; and for collateral cross-relatives of the grandparent generation. All of these evince a distinct feeling for specific relationships removed by three steps of kinship, whereas most other Indians cover such remote re- lationships by applications of terms for nearer kindred. Again we face a feature of kinship designation that is the reflection of an abstract idea. In making the seniority of brother-sister cousins depend on the parents' ages the Luiseino follow a practice that is adhered to by a number of American tribes but which in the present state of knowledge is unique in California. YOKUTS The following system is that of the Yaudanchi tribe, belonging to the Tule-Kaweah group of the Foothill division of the Yokuts.5 Terms in parentheses are from the Yauelmani, who, though fairly near neighbors of the Yaudanchi and in frequent association with them, speak a dialect of the Valley division. Both tribes are from the southern range of Yokuts territory and in contact with Shoshonean tribes, such as the Tuibatulabal and Kawaiisu. Yokuts systems have been collected by Mr. Gifford from the Tachi, at about the center of the area of the stock, and the Gashowu and Chukehansi in the north; but these are as yet unpublished. Parent Class Natet, father; vocative: opoyo. (Yauelmani, in reference, nopop.) Nazhozh, mother; vocative: ishaya. (Yauelmani, in reference, no'om.) The initial syllable in n- in these words appears to be a prefix, originally meaning "my," which has become crystallized; while the stem of natet, nopop, and nazhozh seems to have been reduplicated and then reduced. Buehong, son; man's brother's son. (Butson.) 5 Present series, II, 240, 1907. Kroeber: California Kinship Systems Ahid, daughter; but also child. A man 's brother 's daughter, and a woman 's sister 's child of either sex, are called ahi, which is of course from the same stem. That there is no confusion in my notes appears from the objective cases of the two words: ahda and ahia. Brother Class Nibech, older brother. (Nibech.) Ne'esh, younger brother. Na'at, older sister. (Na'at.) No 'ot, younger sister. (No 'ot.) An old possessive prefix appears to have become incorporated in these words also. B[iHkozh, brother or sister of opposite sex from speaker, irrespective of age. Self -reciprocal. Grandfather Class Enash, any grandfather; any grandchild of a man. (Enes.) T 'uta, mother 's mother; woman 's daughter 's child. (Kamits, mother 's mother; ts 'utsa, woman's daughter's child.) Bap', possibly papa, father's mother; woman's son's child. (Bapa.) Hitwaiu (t palatal), great-grandfather; man's great grandchild. This word also means " ghost "; but the reciprocal usage indicates that, whatever its original meaning, it is also employed as a definite term of relationship. Mokoiot, great-grandmother; woman's great-grandchild. This term must be derived from mnokoi, whose present meaning is mother 's sister. Uncle Class Komoyish, father's brother. (Komoyis.) Mokoi, mother's sister. (Mokoi.) Agash or akash, mother's brother. (Akash.) Guiha, father 's sister. (Nusus.) Chayah, man's sister's child: reciprocal of agash. (Tsayah.) Napash, woman 's brother's child; reciprocal of guiha. (Napas.) Ahi, woman 's sister 's child: reciprocal of mokoi; also, a man 's brother 's daughter. Except that a man calls his brother 's son buchong, that is, son, ahi therefore denotes all parallel nephews and nieces, and is reciprocal in meaning to komoyish and mokoi together. Its connection with ahid, daughter, has already been mentioned. (The Yauelmani equivalent is not entirely clear. It may be butson, son or child in general.) Father-in-Law Class Nahamish, father-in-law. (Nahamis.) Ontip, mother-in-law. (Ontip.) Napatum, son-in-law; also, sister's husband. (Napatim.) Onmid, daughter-in-law. (Onmil.) Makshi, parent of child-in-law. Self-reciprocal. Brother-in-Law Class Nip 'ei, wife's brother. (Nipi.) Onpoi, husband's brother, wife's sister. (Onpoi.) 1917] 353 354 University of California Publications in Am. Arch. and Ethn. [Vol. 12 Itwap; brother's wife; also, husband's sister. (Yauelmani, itwap, with the same meaning, except that a woman calls her brother 's wife kitwinits.) For sister 's husband, see napatum, above. Informants mentioned that one married an onpoi on the death of one 's spouse. The two meanings of the term are not reciprocal, however. Itwap and napatum' are both reciprocal to onpoi, and both denote other relationships as well. Husband and Wif e Yiwin, wife, and yuwenich, husband, are both from the stem yiw, appearing with the formative suffix -in as the verb "to marry"; as, yewin-ji, "he married.." Yuwenich means "the marrier." Neither term seems to be used in address. The Yauelmani are said to refer to the wife as moki: compare Yaudanchi mokoi, mother 's sister. Death of Connecting Relative The following terms for affinities by marriage are altered upon the death of the connecting relative: ontip becomes unitipi; napatum becomes napitimi; onmid becomes onimidi; onpoi becomes unipiyi. The alteration is by a process that has several analogues in Yokuts grammar. A suffix -i is added which shifts the accent a syllable farther from the head of the word and changes the vowels of all but the initial syllable. The idea of severance of relationship is expressed in several neighboring Shoshonean lan- guages;6 but the means here described is peculiar to the genius of Yokuts.7 Reciprocity All five terms of the grandfather class are exactly self-reciprocal. In the uncle class there is no trace of verbal reciprocity. The cross uncle and aunt terms, however, each have a conceptual reciprocal. The reciprocals for parallel uncle and aunt are the words for children, or terms derived from them. In the parent-in-law and brother-in-law classes there are no reciprocals, except for makshi, parent of a child- in-law. A woman calls her husband's sister itwap and is so called by her; but the word is also used by a man for his brother's wife. More- over, in Yauelmani, husband's sister remains itwap, but the reciprocal is kitwiiits, if the recorded data are not confused. It is therefore necessary to conclude that the Yokuts entertain little more feeling than we for reciprocity in the brother-in-law class which is so favorable for the expression of this idea. That the word for great-grandfather means "ghost," that is, "dead person," ensures that it was first applied to the aged relative 6 Present series, xii, 241, 1917. 7 Present series, II, 178, 201, 1907. Xroeber: California Kinship Systems and that its reciprocal meaning of great-grandchild is secondary. This example renders it probable that the other reciprocal terins in this class are also children 's terms which their grandparents re- bestowed on the little ones. The generic southern Yokuts term for mother's mother and a woman's daughter's child is t'uta.8 In Yauel- mani, however, the mother's mother is called kamits. But as the presumably secondary reciprocal remains ts'utsa, it must be concluded that the Yauelmani once used this term also with the meaning of mother's mother which it possesses among the other Yokuts, and that kamits was subsequently introduced. A change of social institutions cannot be invoked as explanation, because no custom of marriage, descent, or kin function can possibly be involved. Any condition of Yokuts society that permitted the Yaudanchi t'uta-t'uta terminology would be equally well served by the Yauelmani karnits-ts'utsa termin- ology. The situation is simply that one tribe adheres to its original usage of a single self-reciprocal word, while the other has come to employ two terms that are exactly complementary. There is nothing to prevent this process of enlargement of the series of terms, or the contrary one of reduction, from having gone on indefinitely while the accompanying society remained identical. It is entirely conceivable, for instance, that the Yauelmani might in time have come to use not only six but ten words in the grandparent class in place of the original five; or that, on the other hand, they might have added verbal to conceptual reciprocity in the words of the uncle group, and thereby diminished their number from seven to four. The final outcome of such a process would be a Yauelmani system of nomenclature thor- oughly different at many points from its original form and from that of allied peoples, without any change of social system and merely through a change of psychological attitude as expressed in speech. Much the same can be inferred from ahid and ahi, two terms scarcely differentiated in sound and the first of wavering, the second of asymmetric and therefore probably also fluctuating meaning. Either the Yaudanchi once called their parallel nieces "daughters" outright, and later began to differentiate between these two kinds of relatives by altering the term when applied to one of the two; or they once possessed a special term for parallel niece (or for a woman's parallel nephew-niece) and later replaced this by the word for daughter (or child), the old sense of distinctness of the niece from the daughter however remaining sufficiently strong to prevent a 8 Compare Paleuyami djudja, present series, II, 267, 268, 1907. 1917] 355 Xroeber: California Kinship Systems and that its reciprocal meaning of great-grandchild is secondary. This example renders it probable that the other reciprocal terins in this class are also children 's terms which their grandparents re- bestowed on the little ones. The generic southern Yokuts term for mother's mother and a woman's daughter's child is t'uta.8 In Yauel- mani, however, the mother's mother is called kamits. But as the presumably secondary reciprocal remains ts'utsa, it must be concluded that the Yauelmani once used this term also with the meaning of mother's mother which it possesses among the other Yokuts, and that kamits was subsequently introduced. A change of social institutions cannot be invoked as explanation, because no custom of marriage, descent, or kin function can possibly be involved. Any condition of Yokuts society that permitted the Yaudanchi t'uta-t'uta terminology would be equally well served by the Yauelmani karnits-ts'utsa termin- ology. The situation is simply that one tribe adheres to its original usage of a single self-reciprocal word, while the other has come to employ two terms that are exactly complementary. There is nothing to prevent this process of enlargement of the series of terms, or the contrary one of reduction, from having gone on indefinitely while the accompanying society remained identical. It is entirely conceivable, for instance, that the Yauelmani might in time have come to use not only six but ten words in the grandparent class in place of the original five; or that, on the other hand, they might have added verbal to conceptual reciprocity in the words of the uncle group, and thereby diminished their number from seven to four. The final outcome of such a process would be a Yauelmani system of nomenclature thor- oughly different at many points from its original form and from that of allied peoples, without any change of social system and merely through a change of psychological attitude as expressed in speech. Much the same can be inferred from ahid and ahi, two terms scarcely differentiated in sound and the first of wavering, the second of asymmetric and therefore probably also fluctuating meaning. Either the Yaudanchi once called their parallel nieces "daughters" outright, and later began to differentiate between these two kinds of relatives by altering the term when applied to one of the two; or they once possessed a special term for parallel niece (or for a woman's parallel nephew-niece) and later replaced this by the word for daughter (or child), the old sense of distinctness of the niece from the daughter however remaining sufficiently strong to prevent a 8 Compare Paleuyami djudja, present series, II, 267, 268, 1907. 1917] 355 356 University of California Publications in Am. Arch. and Ethn. [Vol. 12 wholly unmodified employment of the word " daughter" for the relationship of niece. In the one event we are confronted by an incipient dissimilation, in the other by an incomplete assimilation of terms. If now we assume that the cause of this change was an alteration in the social organization of the Yaudanchi, such as a drift to or from the levirate, for instance, it follows either that this social alteration was also halting and incomplete, which is likely to be difficult to corroborate by independent evidence in the case of a prim- itive tribe, and therefore to remain a purely speculative opinion; or, if the change in social conditions was fulfilled, the change in nomen- clature lagged behind and now reflects the social evolution only brokenly. RELATIONS OF MIWOK AND YOKUTS The Central Miwok system has been presented and analyzed in full by Mr. E. -W. Gifford.9 Its special peculiarities appear to be three. First, there are five terms for three-step affinities by marriage- pinuksa, kumatsa, moe, haiyeme, maksi-which denote such persons as the wife's mother's brother, a woman's sister's son's wife, and the husband's brother's wife. The word maksi has the same significance as Yokuts makshi, and is interesting as a case of outright transfer of a kinship term from one language to another. As it is one of a class in Miwok, but so far as known stands alone in Yokuts, the latter people are likely to have been the borrowers. It is, however, necessary to bear in consideration that in as much as I did not ordinarily attempt to secure terms of this type of rather remote and indirect relationship, there is a possibility that they may actually occur in several of the systems here presented from which they now appear to be lacking. Second, the grandparent class is much reduced in Miwok. There are only the three terms: grandfather, grandmother, grandchild. The grandmother's brother is a grandfather, and so on. Third, the system is rather asymmetrical. The father's brother is a father, but there are two terms for the mother's sisters. There is one reciprocal to father's sister, two to mother's brother. There is one word denoting parents-in-law, two for children-in-law. Olo is the brother's wife, irrespective of sex, but there are two reciprocals for husband's brother and husband's sister. 9 "Miwok Moieties," present series, xii, 139-194, 1916. 1Kroeber: California Kinship Systems The 'differences from Yokuts are not serious. The Yokuts self- reciprocal word for brother-sister of opposite sex is lacking. The terms of the grandparent class differ in not being reciprocal at all in Miwok, whereas in Yokuts they are self-reciprocal. The Yokuts great- grandparent terms are not represented. Yokuts generally has con- ceptually reciprocal terms for parallel as well as for cross relatives of the uncle class; Miwok merges these parallel relatives in the parent class, except for the mother's sisters. Yokuts distinguishes and Miwok combines the father-in-law and mother-in-law. Yokuts possesses four terms and Miwok five in the brother-in-law class, and the allotment to these of the eight logically possible relationships is mostly different, Miwok proceeding on the principle that such terms are conceptually reciprocal without being self-reciprocal, and that the sex of the spouse is always denoted while that of the brother-sister is left indeterminate, whereas the Yokuts classification is more random. The cousin termin- ology, on which Mr. Gifford has full and interesting data, can unfor- tunately not be compared on account of lack of Yokuts data. Reciprocity is nearly equally developed in the two systems, the Yokuts, however, favoring it rather for blood kin and the Miwok for the less numerously recognized connections by marriage. Both systems evince much less reciprocity than either Luiseiio or Mohave. The Miwok men marry their mother's brother's daughters, but Mr. Gifford concludes very convincingly that the original form of marriage is that of a man to his wife's brother's daughter, because twelve Miwok kinship terms are in accord with this type of marriage and none with cross-cousin marriage. Unfortunately it is not known whether the southern Yokuts marry either of these relatives; nor can anything be predicted in the matter because the full significations for most of the Yokuts terms corresponding to the twelve in question have not been obtained. Another matter that is of logical bearing on the Miwok and Yokuts systems is an exogamic, patrilinear moiety organization. The northern and central Yokuts possess this organization in a form much like that of the Central Miwok. For the southern Yokuts, from whom the kinship terms here presented were collected, its existence seems im- probable. It is very doubtful, however, whether this organization has seriously influenced kinship terminology. Of twenty-nine Miwok terms used by a man, twelve refer to his own moiety, nine to the opposite, and eight do not indicate moiety; for a woman, the corre- sponding figures are fourteen, seven, and nine. 1917] 357 Kroeber: California Kiinship Systems letter, but represents a mixed vowel occurring in all Shoshonean languages and often written u. Parent Class Na, father. Pia, mother. Tua, son. Pade, daughter. Brother Class Pabi'i, older brother. Wanga'a, younger brother. Hama'a, older sister. Peni'i, younger sister. All first cousins are called brothers and sisters, whether cross or parallel. Whether they are called older or younger depends on their actual age, not on the ages of the respective parents. Grandparent Class Kenu 'u, father's father; and, reciprocally, a man's son's child. Togo'o, mother's father; and, reciprocally, a man's daughter's child. Mu 'a, mother 's mother; and, reciprocally, a woman 's daughter 's child. Hutsi'i, father 's mother; and, reciprocally, a woman 's son 's child. Hebi'i was given as father 's father 's mother, and reciprocally as a woman 's son 's son 's child. It probably has a wider meaning. It enters into composition with other terms to denote certain connections by marriage. In these com- pounds it appears to designate relationship less than three generations remote. Uncle Class Hai'i, father's brother. Pidu 'u, mother's sister; also mother's co-wife, even if not related in blood. Atsi, mother's brother. Pahwa, father's sister. All these are used alike for the older and the younger brother or sister of the parent. Each has an exact reciprocal, which is, however, entirely different in sound. Huza, man's brother's child, reciprocal of hai'i. Mldo'o, woman's sister's child, reciprocal of pidu'u; also, child of a co-wife, even if unrelated in blood. Nanakwe, man's sister's child, reciprocal of atsi. Mido'o, woman's brother's child, reciprocal of pahwa. I suspect that mido'o, woman's sister's child, and mido'o, woman's brother's child, are the same, especially since I recorded both as accented on the second syllable, which is according to rule if the first vowel is short, whereas a long initial syllable carries the accent. It is not unlikely that mido 'o has been extended from woman 's sister 's child to denote also her brother 's child, re- placing a former adatsi, which survives in composition in the name which a woman applies to her brother's wife. 1917] 359 360 University of California Pu1bli-cations in Am. Arch. and Ethn. [Vol. 12 Parent-in-Law Class Yahi, father-in-law; mother-in-law. Togo-nna, son-in-law. This word means literally "father of the child of the daughter of a man," and logically is therefore usable only by males; but it is employed by women also, who have no other designation for a son-in-law. Kenu-pia, daughter-in-law. Again a man 's term used by women also: "mother of the child of the son of a man." The word is a true compound, kenqu'-pia, not kenu'"'u pia'. Hebi-yati, literally " woman 's son 's son 's child 's father-in-law, " or "father 's father 's mother 's father-in-law, " was recorded with the meanings of father- in-law 's father, father-in-law 's mother, and father-in-law 's paternal grand- mother. In the last instance the compound denotes one 's wife 's great-grand- mother, whereas hebi itself denotes one 's own great-grandmother. I infer that hebi-yahi is applicable to a considerable range of affinities by marriage, its first element denoting that the person denoted is two or three generations older, and the second element having about the force of our "in-law"; much as we might describe an old lady as our "great-grandmother-in-law." Hebi 'i togo-nna-an epithet of two words, not a compound-was given as the reciprocal of hebi-yahi, specifically used by a woman for her son 's son 's son-in-law-her great-grandson-in-law. Brother-in-Law Class Adatoi, wife's brother; man's sister's husband. Self-reciprocal. Also used between men as a friendly term of address when no relationship exists. Adatsi-pia, husband's sister; woman's brother's wife. Self-reciprocal. The word means "mother of the adatsi." If the term was first used for the brother's wife, adatsi must be an old name for a woman's brother's child. This inter- pretation is supported by the fact that the actually employed designation for a woman's brother's child is probably the same, and certainly nearly the same, as for a woman's sister's child-a uniting of relationships not in accord with the plan of the remainder of the Northern Paiute system. If, however, adatsi-pia was first used for the husband?s sister, then adatsi must have meant husband's sister's child. In support of this interpretation is the similarity of adatsi- probably composed of a stem ada and the diminutive suffix -tsi-to adatoi. This word adatoi denotes a man's brother-in-law; but its former meaning may have been wider; since my informant stated that sisters-in-law sometimes called each other adatoi, jokingly he thought. If adatoi ever meant brother-in-law or sister- in-law in general, its connection with adatsi could hardly be doubted: adatoi being the brother-in-law, adatsi would be the "little brother-in-law," that is, the brother-in-law 's son, or a junior relative of the husband, such as his sister's son; and the adatsi-pia, his mother, would in the latter case be the husband's sister. It is difficult to decide between these two explanations. Yet, whichever is right, or if it be a third, the term adatsi-pia is descriptive and could originally not have applied to both the persons to whom it is now applied; for there is no group of relatives to whom two sisters-in-law can both be mothers. The term therefore once belonged to one of these relationships and has been extended to include the other, as it is now self-reciprocal, presumably through the operation of the inclination toward reciprocity. If this tendency is strong enough to cause a change of meaning of exactly descriptive terms until they become self- contradictory, its potential influence must be great, and should suffice to brinig about even more considerable alterations of ordinary non-descriptive terms, Kroeber: California Kinship Systems whose scope is readily extensible by analogy or metaphor without the prodluction of a transparent logical clash. Nenai'i, husband 's brother; woman's sister 's husband. Huza-na-pia, wife 's sister; man 's brother 's wife. The latter must be the original meaning, the former its extension-again etymologically inexact, uinless double marriage of brother and sister to sister and brother had been the rule. Huza denoting a man 's brother 's child, the huza-na-pia is of course his brother's wife. Nenai'i and huza-na-pia are mutually reciprocal, while the two other terms of this class, adatoi and adatsi-pia, are each self-reciprocal. A different grouping of the four meanings expressed by nenai'i and huza-na-pia would have made these also self-reciprocal. As it is undeniable that extensions or alterations of meaning have taken place in this class of terms, it is reasonable to consider why these changes did not operate in the direction of consistency, that is, of uniform self-reciprocity. The reason seems to be that in such case nenai'i and huza-na-pia would each have denoted both males and females. Under the existing system of Northern Paiute, however, each of its four terms of this class refers only to men or only to women, to wit; man 's brother-in-law, woman 's sister-in-law, woman 's brother-in-law, man 's sister-in-law. Reciprocity must from its very nature interfere with the consistency with which certain conceptual factors entering into relationship (such as generation, sex of the speaker, and sex of the relative) are expressed; and the reverse is equally true. In the uncle and grandparent classes of Northern Paiute terms, where the reciprocity is complete-although only logical in one case and verbal as well in the other- the consistency of employment of the three conceptual factors or categories is thoroughly violated. In the brother-in-law class, on the other hand, complete uniformity of reciprocal expression is not attained, but every term is exact in its denotation of sex of the person referred to as well as the sex of the speaker. Husband and Wife Kuma, husband. Nodekwa, wife. The terms of address were not recorded. Woho, co-wife. A woman says i-woho', "my co-wife,y" in reference, but addresses her as i-bea'a, "my friend," if they are not sisters. As an address, i-woho is an insult. Na-wo'ho is used when a man's two wives are meant: na- is a reciprocal prefix. MARRIAGE The Northern Paiute deny cross-cousin marriage, though my informant attributed it to the Shoshone on their east, who, he said, will marry their pahwa's daughter. This may, however, be only the expression of an opinion of the loose morality of the latter pepole, since Gilbert also mentioned that the Shoshone married their parallel cousins, which is scarcely possible. The brother-sister terminology for cross-cousins among the Paiute confirms their denial of the practice by themselves. Even first cousins once removed and second cousins cannot marry among Gilbert's people. This is certain for parallel cousins; but 1917] 361 362 University of California Publications in Am. Arch. and Ethn. [Vol. 12 unfortunately my records do not allow me to assert the same rule positively for second cross-cousins, although I believe it applies. First half cousins, on the other hand, can and do marry. I secured an instance of the children of two half brothers marrying. Such half cousins were common among the Northern Paiute as the result of polygamous marriages by men. There seems to have been even some encouragement of half cousin marriage, as favoring a peaceable and permanent union; although if, as often happened, the half brothers lived in remote localities, a marriage of their children was likely to be terminated by the return of one of them to the old home when ties of blood and association called. Geography was otherwise a factor of importance in these matters, on account of the varying degree of acquaintance which it imposed. My informant's father and the latter's half brother, who lived apart, arranged a marriage between a son of the former and the daughter of the latter. The girl was willing, but the young man, having pre- viously visited at her home, had got to calling her "sister," and refused to marry her on that ground. He had known her too long, he said. A man might marry a woman and her daughter-his stepdaughter, of course. This is a common practice of most of the California Indians. WASHO" The stems of the Washo terms of relationship are used with pos- sessive prefixes, such as di-, my, urm-, your, da-, his. A few words replace di- by di-m-, la-, or 1-, um- by mi-m- or m-. Parent Class Koi, father. La, mother. Malolo, parents. Ngam, son. Ngamu, daughter. Ngaming, child. Brother Class At 'u, older brother. Isa, older sister. Beyu, younger brother. Wits'ulk, younger sister. Cousins are older or younger brothers and sisters according to their own ages, not those of their parents. 11 Present series, IV, 309, 1907. 1Kroeber: California Kinship Systems Grandparent Class Baba, father 's father, man 's son's child. Elel, mother's father. Eleli, man's daughter's child. Ama, father 's mother, woman 's son 's child. Gu, mother 's mother. Guyi, woman 's daughter 's child. Uncle Class Eushi, father's brother; reciprocal, masha. Da, mother's brother; reciprocal, magu. Ya, father's sister; reciprocal, shemuk. Sha'sha, mother 's sister. Masha, man 's brother 's child; reciprocal, eushi. Magu, man 's sister 's child; reciprocal, da. Shemuk, woman 's brother 's child; reciprocal, ya. The term for woman 's sister 's child was not obtained with certainty. It may be shemuk. See below. Parent-in-Law Class Ayuk, parent-in-law. Bu-angali ("lives with "), son-in-law. Eyesh, daughter-in-law. Brother-in-Law Class Uladut, man's sister's husband, wife's brother. Self-reciprocal. Di-ngaming de '-eushi (" my child 's father 's brother "), husband 's brother. Di-magu da-koi (" my sister 's child's father "), woman 's sister 's husband. This phrase does not necessarily prove that magu is the term which a woman applies to her sister's child: in Northern Paiute there are analogous cases of a woman using a man 's term in descriptive phrases of this type. Di-mash da-la (" my brother 's child 's mother "), man 's brother 's wife. This phrase was also obtained for wife 's sister, but the latter meaning is in need of corroboration. Yangil, woman 's brother 's wife, husband 's sister. Self-reciprocal. Husband and Wife Bu-meli, husband (meli, "to make a fire"). (M)laya, wife. The vocative terms are not known. Reciprocity Reciprocity is consistently verbal and conceptual in the grandfather class, and conceptual only in the uncle class; it is not expressed in terms for relatives by marriage except in one brother-in-law and one sister-in-law term. RELATIONS TO NORTHERN PAIUTE The Washo and the Northern Paiute are the only tribes of those here considered who live east of the Sierra Nevada. Both extend 1917] 363 364 University of California Publications in Am. Arch. and Ethn. [Vol. 12 from Nevada into eastern California. Their customs are little known, but appear to be similar, though the languages are utterly distinct. Their kinship systems are practically identical. Parent class: father, mother, son, daughter. Washo adds a derivative for child and a term for parents. Brother class: older and younger brother and sister. Grandparent class: paternal and maternal grandfather and grandmother, each used reciprocally in the same form, except that two Washo grandchild terms add a suffix. Paiute has also a term for great-grandmother. Uncle class: four terms for parallel and cross uncle and aunt, and four exact conceptual reciprocals, which, however, bear no likeness in sound. In both languages there is some doubt whether there is a distinct term for woman 's sister 's child. Connsins: all are denominated brothers or sisters, seniority depending on their actual age. Parenit-in-law class: parent-in-law, son-in-law (a descriptive term), daughter- in-law (descriptive in Paiute only). Brother-in-law class: man 's brother-in-law, self-reciprocal; woman 's sister- in-law, self-reciprocal (descriptive in Paiute only); man 's brother 's wife or wife's sister (descriptive); husband's brother or woman's sister's husband, denoted by a single non-descriptive word in Paiute and by two separate de- scriptive terms in Washo. The two systems could not well be more similar. Two alternative interpretations are open. Either we must assume that Washo and Northern Paiute institutions are identical and that institutions are perfectly reflected in kinship terminology; or we must admit that these two systems have attained their practical identity under the partial or dominating influence of similar ways of thinking, that is, that mental or linguistic causes have been operative. RELATIONS TO OTHER SYSTEMS The Washo are in contact with the Miwok; and the Washo-Paiute system is not very different from the Miwok-Yokuts one-certainlv much more similar to it than to either the Wintun or the Mohave- Luisefio type of terminology. The greatest difference is in the cousin nomenclature, which could not well be more diverse. The Miwok terms of the grandparent class are also dissimilar: non-reciprocal grandfather, grandmother, and grandchild versus a scheme of four self-reciprocal terms each expressing the sex of the intermediate rela- tive.'2 Miwok, however, seems exceptional in this point. The southern 12 It is a striking circumstance that the Miwok disregard this consideration, although its observance would bring their nomenclature into closer consonance with their social scheme of descent, whereas the Washo and Northern Paiute, who are not known to possess moieties, discriminate according to the factor. If terminology mirrors sociology, the Miwok should distinguish paternal grand- parent and maternal grandparent instead of grandfather and grandmother. Kroeber: California Kinship Systems Yokuts, although geographically more distant than the Miwok, use terms of the exact Washo-Paiute type.13 The Miwok must therefore be regarded as occupying a distinctly anomalous position in their grandparent-grandchild terminology. This is borne out by the fact that the Wintun, Pomo, and Yuki, who tend to merge grandchildren in children or nephew-nieces, and therefore, like the Miwok, express no reciprocity in this class, nevertheless generally distinguish paternal from maternal grandparents. To the south, the Mohave and Luiseiio express both lineage and at least conceptual reciprocity; and the same seems to have been the practice of the Salinans and Chumash, so far as the fragmentary evidence allows judgment. The divergence of the Miwok system from that of the Washo and Northern Paiute at this point is therefore not characteristic of type, but due to a Miwok peculiarity. At most other points Washo and Miwok correspond fairly, or about as well as Miwok and southern Yokuts. The parent and brother classes are substantially identical. The uncle class differs in that the Miwok merge parallel relatives in parents, except for the mother's sisters, among whom they distinguish seniority. This seems another Miwok specialization, since Yokuts is more similar to Washo. The parent-in-law class is similar in that the father-in-law and mother-in- law class are merged and that there is no reciprocity. The brother- in-law relatives are differently classified; but the force of this diverg- ence is weakened by the marked difference of Yokuts from both. The Miwok terms for three-step connections by marriage, finally, are unrep- resented in both Washo and Northern Paiute; but this class seems again to present a Miwok individuality, being lacking, or practically so, in all other Californian systems, as far as we know. Just so, the descriptive terms of Washo and Northern Paiute are an evident peculiarity, since they are found only among the Luiseflo and not among any nearer tribes in California. On the whole, therefore, while Washo and Northern Paiute form an exceedingly intimate group, they also have tolerable affiliations to south central California. They are certainly at least as near and probably nearer to Miwok-Yokuts than these are to Luiseino-Mohave. On the other hand, Miwok evinces a number of specializations from which southern Yokuts is free; the latter on the whole is therefore more similar than Miwok to the Nevadan systems. 13 Except that there is only term for grandfather, though this remains self- reciprocal. 1917] 365 366 University of California Publioations in Am. Arch. and Ethn. [Vol. 12 TUBATULABAL AND KAWAIISU Mr. Gifford has described'4 the systems of these two Shoshonean tribes of the southern Sierra Nevada, neighbors of the southern Yokuts. They are similar to each other and in general type very close to Northern Paiute. The chief differences from the latter are the following: i Kawaiisu has terms for great-grandfather, son, mother, and daughter- diminutives from the stems for older and younger brother and sister. Tiuba- tulabal has a word for great-grandparent apparently borrowed from the Kawaiisu one for great-grandfather, and employs a diminutive thereof as a reciprocal. Kawaiisu applies its terms for cross uncle and aunt only to the younger brother or sister of the parent. The father's older brother is classed with the father, the mother's older sister with the mother. The Tiibatulabal scheme is like the Northern Paiute one. Both languages, like adjacent Yokuts, alter the terms for connections by marriage after the death of the intermediate relative. The means employed are suffixes. In addition, Tiubatulabal possesses a special term hoki, used be- tween grandparents and grandchildren after the death of the father or mother. Both languages possess special terms applied only to the blood father and the blood mother before the loss of any child. Tiubatulabal expresses "son" and " daughter" by a single word and "younger brother" and "younger sister" by one. The Kawaiisu terms of the brother-in-law class tally exactly with those of Northern Paiute. The Tiibatulabal ones are doubtful. No one of Mr. Gifford's half dozen informants yielded them alike. Not one of the lists reduces to the Kawaiisu scheme even when the number of terms is reduced from five to four by counting two as a single one. Mr. Gifford suggests Yokuts influence on the Tiubatulabal system on this point, and I have no doubt he is right. But I have been equally unable to make the arrangement of any of his informants fit the Yaudanchi or Yauelmani plan. There are only two conclusions that suggest themselves. Either the Tiibatulabal system has broken down at this point in the last sixty years under American and Mexican contact, or original Shoshonean and subsequent Yokuts influences have mingled and reduced the Tiubatulabal scheme to a transitional and inconsistent stage. Possibly the latter condition existed first and caused an unusual lack of resistance under the effect of our civilization. SHOSHONEAN SYSTEMS These systems collected by Mr. Gifford, with two others recorded by Dr. Sapir and included in full in the same study, and Luisenio and Northern Paiute, make six that are available from the Shoshonean family and allow a broader comparison than has been possible here- 14 Present series, xii, 219-248, 1917. Kroeber: California Kinship Systems tofore. I give first the words used to express several of the more elementary relationships. Father Mother Older sister Father 's father Mother 's moth Mother's fathe Mother 's broth Father 's sister Parent-in-law Son-in-law SHOSHONEAN STEMS OF SIMILAR MEANING Northern Uintah Kaibab Paiute Kawaiisu Ute Paiute Tilbatula na kugu, muwu1 moa moa kumu, anE pia mawil, piyu1 pie piya iimii, abul hama'a pachi paichi patsi- kuchi r kenu 'u kuno konu aka er mu 'a kagu kagu kahu utsu r togo 'o togo togu toho3 agi-st er a-tsi shinu shina-nchi2 kali pahwa paha pa paa pauwa yahi yehe yai-chi a wasu-mbi (descrip- mono muna-chi mona-tsi wiini tive) ibal Luiseiio a nal LI yO kes ka " tu' kwa tash pa-mai is (descriptive)5 (descriptive) I Blood parent before loss of any child. 2 Younger brother of the mother. s Any grandfather. 4Fathers' parent. A woman says ka', father's parent. It is clear that the stems that are used to denote the same relation- ship are very variable. The Kawaiisu, Uintah Ute, and Kaibab Paiute idioms are all of the Ute-Chemehuevi division and very close to one another. They may be said to differ only dialectically. Northern Paiute belongs to the same Plateau branch of the family, but deserves to be reckoned as a distinct language. Nearly half of its stems for kindred are different. Tiibatulabal and Luisefio are linguistically somewhat remote from the others, but certainly no more than Greek is from Latin or German from Slavic; yet the majority of their stems are new. Analogous results appear when the procedure is reversed and the meanings of identical stems are compared. Tua, tuu',u, towa, tuwa means son in all four of the Plateau dialects; tumu is son or daughter in Tiibatulabal. Nama'i, nami is younger sister in Kawaiisu, Ute, and Kaibab Paiute; nalawi is younger brother or sister in Tiibatulabal. Shinu is mother's brother in Kawaiisu, shina-nchi mother's younger brother in Ute, shina- male cousin in Kaibab. Mawu is mother's older sister in Ute, mawii mother's older sister or mother who has not lost a child in Kawaiisu, mangwu 'i- female cousin in Kaibab. Piyu in Kawaiisu denotes only the mother who has not lost a child; in the other three Plateau dialects the term pia, pie, piya means mother, without being so limited. Luisenfo tu', mother's mother, appears to correspond to Plateau togo'o, togo, togqu, toho, which always denotes the mother's father. 1917] 367 368 University of California Publications in Am. Arch. and Ethn. [Vol. 12 Luisefno ka', father 's father or father 's mother, seems to be from the stem of Kawaiisu kagu, Ute kagu, Kaibab kahu, all of which denote the mother 's mother; while Luisenfo kwa, mother 's father, perhaps is the etymological equiv- alent of Plateau kenu 'u, kuno, k6nu, father 's father. These correspondences are not certain, and perhaps they should be interchanged; but if they hold either way, there has been a specific alteration of meaning. These two comparisons in conjunction make it clear that terms of relationship have a history quite like that of all other words. They alter in meaning, become obsolete, drop out of usage altogether, and new stems, which originally had another significance, come to take their places. If kinship terms are more conservative than most other parts of a language 's vocabulary, the difference is merely one of degree. Whether they are more conservative is a subject neither for reasoning nor for assumption, but a problem of fact to be established by purely philological comparison. In short, kinship terms are an integral part of the tongues in which they occur and are therefore subject to linguistic influences like all other words. This being so, they cannot be a perfect nor even a reliable mirror of institutions. WINTUN I secured an outline of the Southeastern Wintun system as used in the vicinity of Colusa. It is so extraordinary that I include it for comparative purposes, although Mr. Gifford subsequently obtained fuller and better verified lists from several parts of the Southeastern and Southwestern Wintun territory. It appears that I have missed one or two terms; but the skeleton of the system as here presented is substantially correct. Wintun terms are used with possessive affixes, but differ so much for the first and second persons that it is desirable to give both forms. In general, "my" is -chu, and "your" is mat-. Parent Class tan-chu, ma-tan, father, father's brother. na-ku, ma-nin, mother, mother's sister. te-chut, mat-mutle, son, daughter, man's brother's child, woman's sister's child. Brother Class laba-chu, mat-laben, older brother. otun-chu, mat-usun, older sister. tlan-chu, ma-tlan, younger brother or sister. The method of application of brother-sister terms to cousins was not learned. Kroeber: California Kin-ship Systems Grandparent and Uncle Class apa-chu, mat-apan, mother's brother, mother's father, father's father, great- grandfather. ama-ku, mat-aman, father's sister,15 mother's mother; presumably also great- grandmother. sakan-chut, mat-sakan, father 's mother.'6 tai-chu, ma-tai, woman 's brother 's childl5 or man 's sister 's child, that is, any cross nephew or niece; also, any grandchild; presumably also any great- grandchild. Father-in-Law Class tes-ba or tes-win, ma-tes, parent-in-law or child-in-law. Tes-win, which seems to contain the stem for "person" (cf. Wintun, Patwvin), is used only for the son-in-law; tes-ba denotes the daughter-in-law and either parent-in-law. In the second person the suffixes disappear and the terms are identical. Brother-in-Law Class tiran-chu, ma-tiran, sister's husband. boksen-chu, mat-boksen, brother's wife. nai-tlen, ma-tlen, spouse's brother or sister. GENERAL FEATURES The extreme condensation of this remarkable system would tend to prevent any considerable reciprocity. In fact, there is none dis- cernible, unless the two forms from the stem tes be looked upon as a single self-reciprocating term. The uncle class has been totally merged in the parent and grand- parent classes. This may be a carrying further of the Miwok principle by which the parallel uncle is called father. However, Miwok does not merge cross-uncles with grandparents, nor cross-nephews with grandchildren, whereas there is some inclination toward the classing together of nephews and grandchildren among the Pomo and Yuki who are near neighbors of the Wintun. The latter people seem there- fore to have used the simplifying tactics peculiar to the systems on both sides of themselves. The Wintun also agree with the Miwok in naming only one grand- father, but with the Pomo and Yuki in distinguishing the paternal from the maternal grandmother, if there is no error about sakan-chu. The Wintun brother-in-law terms correspond with the Miwok ones, except that there is only a single equivalent to three of the latter: 15 Mr. Gifford 's informants gave the term for older sister as denoting the father 's sister, and for younger sister as denoting a woman 's brother 's child. 16 This term was not obtained by Mr. Gifford, whose informants included the father 's with the mother 's mother under ama-ku. Sakan-chu must therefore be considered doubtful. 1917] 369 370 University of California Publications in Am. Arch. and Ethn. [Vol. 12 wokli, kolina, and apasti. Pomo, however, is still more similar to Miwok, the terms having the same signification throughout except that there is a single term instead of kolina and apacsti. It therefore seems that there are certain tendencies of terminolog- ical classification more or less common to the Miwok, Wintun, and Pomo, and several in which they differ; and that Wintun utilizes any of these methods that aid reduction of nomenclature. The result is a system even more compact that the English one, and as free from reciprocity, but constructed on utterly different principles. PM0M17 Parent Class E, father; harik, my father; address: harika. Te, mother; nik, my mother; address: nika. Ghawe-l-ip, son; any grandson; anl nephew except a man 's sister 's son. This term seems connected with hawi, boy, and mi-p, he. Address: harika, as for father. Ghawe-l-et, daughter; any granddaughter; any niece except a man's sister's daughter. Evidently connected with hawi, boy, and mi-t, she. Address: nika, as for mother. Esh, son or daughter, presumably also grandchild, nephew, or niece. A term of endearment or ceremonial usage. In address: esha. Brother Class Meh, older brother; address: meha. Deh, older sister; address: deha. Duhats, younger brother or sister; address: duhatsa. Grandfather Class Madili, father's father; address, the same. This word denotes also the father's father's father. Among the Eastern Pomo south of Clear Lake, it includes the father's older brother; but this is not so among the people on the north side of the lake. Mats, father's mother; address: matsa. Gach, mother's father; also his brother and his father. This inclusion of the great-uncle (or aunt) and great-grandparent seems to apply to all Pomo grandparent terms. In address, gacha. Ghats, mother's mother; address: ghatsa. There are no terms for grandchildren. The words chiefly employed are the "boy" or "child" derivatives used for son and daughter, it is said; but a reciprocation by the grandparents to the grandchildren is not unknown. In this case the reciprocity seems to be exact, i.e., madili denotes a man Is son's son or daughter, and so on. 17 Eastern dialect, of Clear Lake. See present series, XI, 320-346, 1911. Kroeber: California Kinship Systems Uncle Class Keh, father's brother, also stepfather; in address, keha. On the death of a married man his brother generally married the widow. His step-children, however, continued to call him keh, not harik, father. See madili, above. Weh, father's sister; address: weha. Tsets, mother 's brother; address: tsetsa. The reciprocal is dah. Tuts, mother's older sister; address: tutsa. Sheh, mother's younger sister, stepmother; address: sheha. Dah, man 's sister 's child, boy or girl; in address, daha. This is the reciprocal of tsets. It is said to be the only term of nephew-niece type in Pomo, son or daughter being used in all other cases. Father-in-Law Class Sha, father-in-law, mother-in-law. Dimot, son-in-law. This word is said to denote one who supplies or gives in return for favors, and can be used of a woman who visits her lover more or less regularly. Shomits, daughter-in-law. On account of a species of the parent-in-law taboo, these three termis are not used in address, but the plural demonstrative pronoun, hibek, "those" or "they," is substituted. Or, at greater length, a father-in-law may be addressed as butsigi hibek, "old man those"; a mother-in-law as daghara hibek, "old woman those ': a child-in-law as esh-bek, "child those." Even in reference to the relatives in question the plural hibek can be added. The brothers, fathers, uncles, etc., of the parents-in-law are also addressed in this polite way; and presumably the old people apply the form of deference to their children-in-law's brothers and sisters. If the spouse dies, the former parents and children-in-law continue to address one another as if he or she were still living. If the marriage is broken off, they revert to normal singular forms. Brother-in-Law Class God, sister's husband; also his brother, and, it seems, his sister. In address, goda. Mi, brother's wife; also her sister, and, it seems, her brother. Ha, wife's brother or sister. Reciprocal to god. Ghar, husband's brother or sister. Reciprocal to mi. Brothers-in-law and sisters-in-law are addressed directly, without pluralizing circumlocution. Husband and Wife Baili, husband; also kak, "man"; in address, butsigi, "old man." Dat, wife; also da, " woman"; in address, daghara, "old woman." Giashi is a vocative term of endearment used reciprocally by husband and wife. GENERAL FEATURES AND RELATIONS With the Pomo we encounter a reversion from the extreme re- duction of the Wintun system. There is a marked tendency to class juniors under as few designations as possible; and this suffices to prevent any great development of reciprocity. Four terms, however, 1917J 371 372 University of California Publications in Am. Arch. and Ethn. [Vol. 12 reappear for grandparents, and there are specific uncle-aunt desig- nations; so that in this point we are back at the general Californian practice common to Washo, Northern Paiute, Yokuts, and Tiibatu- labal. The distinction of the mother's sisters according to age is too widespread in the region to be regardable as a specific Miwok re- semblance. The brother-in-law terms equal the Miwok ones, at least in involved plan, aiad express conceptual reciprocity. The special three-step terms of the Miwok are unrepresented, terms of two-step relationship being extended to cover them, as apparently by most the tribes of California. Conceptual reciprocity is found in the one word of the nephew class; and there is an incipient or obsolescent tendency toward self-reciprocity in the grandparent group. In short, the Pomo system shares some of the individualized traits of the Wintun and Miwok plans, but in other respects is on a generic Californian basis. YUKI I failed to secure either a complete or a wholly consistent Yuki system. Dr. S. A. Barrett, while on a visit to Round Valley Reser- vation, undertook to supply the deficiencies; but his material proved insufficient for entire clearness, and showed apparent contradictions in the terms for the same classes of kindred in which I had encoun- tered difficulties, namely, uncles, aunts, nephews, nieces, and grand- children. I suspect a factor of classification to be involved here which both our inquiries failed to reach. The data on these groups of relatives must therefore be used with reserve. Parent Class K 'lun, father. K 'an, mother. K'l-(i), son, daughter. Brother Class K-'ich, older brother, older sister. La'n, younger brother. Mu 'n, younger sister. Uncle Class Kaint, father's brother, stepfather. Kikan, mother's older brother. Aint, mother's younger brother. Panchet and p'oyam were both obtained for father's sistei Naint, mother's sister, probably older. Kroeber: Calif ornia Kinship Systems Kansh, mother 's sister, probably younger; stepmother. Difficulty and confusion were experienced in securing these terms from informants. The possibility must be reckoned with that some of the terms differ radically as they are used in reference or address; or that other factors are involved. Chant-kan, man 's brother 's child, woman 's sister 's child, that is, parallel nephew or niece; or, as it might be defined, potential stepchild. Ipima or ipimich-kan, man's sister's child. Kup was obtained with the same meaning; it may be a term of address only. Omsa-kan, woman's brother's child. Some informants add woman's sister's child, and man 's sister 's daughter, but this seems unlikely. Grandparent Class Osh, father's father. Pit, mother's father. Pop, father's mother. Tit, mother's mother. Asam-ap-kan, son's child. Asam-chant-ka", am-chant-kan, daughter's child. Evidently from chant-katt, parallel nephew-niece. Informants were not wholly consistent as to the meaning of the two grand- child terms. Parent-in-Law Class O 'l-am, parent-in-law. Wit(-i), son-in-law. Kim(-a), daughter-in-law. Sut-am was obtained with the same meaning. Brother-in-Laiw Class Lanyan, wife's brother. Chat, wife's sister, brother's wife. Tanshit, husband's sister, husband's brother, sister's husband. Chat and tanshit are reciprocal; that is, any woman called chat says tarshit to the man or woman so addressing her. The reeiprocal of lanyan, however, is also tanshit. These terms were obtained identically by Dr. Barrett and myself, and may therefore be relied upon. GENERAL FEATURES AND RELATIONS In spite of the imperfection of the record, it can probably he inferred that the Yuki system is once more of the generic Central Californian type. The resemblance of grandchild and nephew terms indeed is evidence of some influence of the inclinations that have shaped the Wintun system and in part that of the Pomo. But the terminology for uncles and aunts, that for grandparents, and an apparently moderate degree of conceptually reciprocal expression- self-reciprocal terms have not been found-indicate that the Yuki system is sprung from the same basis as that which has originated the systems of the Washo, Northern Paiute, and Yokuts. 1917] 373 374 University of California Publications in Am. Arch. and Ethn. [Vol. 12 YUROK All terms were obtained with one of the suffixes ne- or n-, my, ke- or k-, your, and we- or u-, his, her. There are also suffixes, especially -osh, which are not part of the stem. Most of the terins are verified by information independently obtained by Dr. T. T. Waterman. Father Class Ne-pshets, u-pshits, father. The term of address is tot, but we-tot-osh, his father, was also obtained. The difference between the two stems is not clear. It is not one of sex of the speaker. We-ts-eko or u-kok-osh, mother. In address: kok. There are no words meaning son or daughter. N-oukshu is "my child," ne-megwahshe "my boy" or "son, " ne-weryernerksh "my girl" o'r " dauighter." Brother Class Ne-mits-osh or ne-mit-osh, older brother. Ne-pin-osh, older sister. Kits-pe'l, older brother or sister. Probably from pe'lin, large. Tseihkeni, ne-eihk-eu, younger brother or sister. Tseihkeni means ''snmall."' Tsits or chich, vocative, and ne-choch-osh, first person, were obtained by Dr. Waterman as meaning younger brother or sister. Ne-pa', brother, male cousin, or more distant male relative of a man. Ne-wveyits, sister, female cousin, or more distant female relative of a man. Ne-lai, brother, male cousin, or more distant female relative of a woman. Dr. Waterman gives an apparent contraction: let. The first five of these terms, which refer to age, and the last four, which express sex, overlap. The former have more or less exact equivalents in all the Californian languages. The latter are of a much rarer type, but similar terms recur among the neighboring Karok, so that a secondary development local to northwestern California may be involved. Grandparent Class Ne-pits-osh, grandfather, as in English, that is, both the father's father and the mother's father. Ne-kuts-osh, grandmother. Ne-k'ep-eu, grandchild. Also used for nephew and niece, in addition to the terms specifically denoting these relationships. Uncle Class Ne-ts-im-osh, father 's brother; mother 's brother; that is, "'uncle " as in English. Ne-tul-osh, mother's sister; father's sister; that is, "aunt" as in English. Ne-k-tsum, brother's or sister's son, that is, "nephew" as in English. Ner-ramets, brother's or sister's daughter, that is, "niece" as in English. All four of these terms are also used for relatives of the cousin class, which see. Ne-k'ep-eu, grandchild, is sometimes also employed for nephews and nieces. I base this statement on concrete cases within my experience. Conversely, I have had ne-pits-osh, grandfather, translated as mother's brother, and ne-tul-osh, 374 University of California Publications in Am. Arch. and Ethn. [Vol. 12 YUROK All terms were obtained with one of the suffixes ne- or n-, my, ke- or k-, your, and we- or u-, his, her. There are also suffixes, especially -osh, which are not part of the stem. Most of the terins are verified by information independently obtained by Dr. T. T. Waterman. Father Class Ne-pshets, u-pshits, father. The term of address is tot, but we-tot-osh, his father, was also obtained. The difference between the two stems is not clear. It is not one of sex of the speaker. We-ts-eko or u-kok-osh, mother. In address: kok. There are no words meaning son or daughter. N-oukshu is "my child," ne-megwahshe "my boy" or "son, " ne-weryernerksh "my girl" o'r " dauighter." Brother Class Ne-mits-osh or ne-mit-osh, older brother. Ne-pin-osh, older sister. Kits-pe'l, older brother or sister. Probably from pe'lin, large. Tseihkeni, ne-eihk-eu, younger brother or sister. Tseihkeni means ''snmall."' Tsits or chich, vocative, and ne-choch-osh, first person, were obtained by Dr. Waterman as meaning younger brother or sister. Ne-pa', brother, male cousin, or more distant male relative of a man. Ne-wveyits, sister, female cousin, or more distant female relative of a man. Ne-lai, brother, male cousin, or more distant female relative of a woman. Dr. Waterman gives an apparent contraction: let. The first five of these terms, which refer to age, and the last four, which express sex, overlap. The former have more or less exact equivalents in all the Californian languages. The latter are of a much rarer type, but similar terms recur among the neighboring Karok, so that a secondary development local to northwestern California may be involved. Grandparent Class Ne-pits-osh, grandfather, as in English, that is, both the father's father and the mother's father. Ne-kuts-osh, grandmother. Ne-k'ep-eu, grandchild. Also used for nephew and niece, in addition to the terms specifically denoting these relationships. Uncle Class Ne-ts-im-osh, father 's brother; mother 's brother; that is, "'uncle " as in English. Ne-tul-osh, mother's sister; father's sister; that is, "aunt" as in English. Ne-k-tsum, brother's or sister's son, that is, "nephew" as in English. Ner-ramets, brother's or sister's daughter, that is, "niece" as in English. All four of these terms are also used for relatives of the cousin class, which see. Ne-k'ep-eu, grandchild, is sometimes also employed for nephews and nieces. I base this statement on concrete cases within my experience. Conversely, I have had ne-pits-osh, grandfather, translated as mother's brother, and ne-tul-osh, 376 University of California Publications in Am. Arch. and Ethn. [Vol. 12 Separation of relatives in the male and female line is so frequently accompanied by a development of true reciprocal expression in Cali- fornia, in the Great Basin region, and in the Southwest that the two phenomena must be taken in connection. As might be anticipated, the Yurok evince little feeling for reciprocity, not only in the kinship classes just mentioned but in the other group which lends itself readily to reciprocal formulation, the relatives by marriage. This is the more remarkable because in the Oregon region, as instanced by the Takelma and the Chinook, systems of California-Plateau-Southwestern type seem again to prevail. It is necessary to look as far as the Coast Salish, or the tribes of the eastern United States, before terminologies of the general plan of the Yurok one are again encountered.. As the Yurok are Algonkin, the interesting problem is raised whether it is possible that they have brought the outlines of an ancient system with them from their presumable eastern source of origin, and suc- ceeded in maintaining the same for an undoubtedly long period in an entirely different cultural setting. This query can be answered only after we know the kinship systems of the tribes immediately adjacent to the Yurok: the fellow Algonkin Wiyot; the Athabascan Hupa, Tolowa, and Chilula; and the Hokan Karok. It may prove that we have to deal with a surviving and re- invigorated importation; or, on the other hand, with a new local development due to obscurer causes. The two or perhaps three classes of brother-sister designations in Yurok are very interesting, but more must be known concerning the distribution of the phenomenon, as well as of the etymology of the words in question, before a satisfactory interpretation is possible. THREE-STEP RELATIONSHIP Kindred removed by three steps of relationship, such as the great- grandfather or brother-in-law's parent,'8 can of course be designated in all languages, either by compounds, by more or less descriptive additions, or by mere extension of meaning of the terms denoting nearer kin. Some systems, however, contain specific designations for certain three-step relations-like English "cousin." Such terms average two or three in number in the Californian systems, but their frequency as well as their meanings vary greatly according to language. 18 I count the brother, sister, and wife as one step removed. Kroeber: Ccalifornia Kinship Systems The commonest three-step term is the self-reciprocal one denoting the parent of a child-in-law. This forms part of the Luisefio, Kawaiisu, Tiibatulabal, Yokuts, Miwok, and Northern Paiute systems. These are all found in southern and central California or Nevada. If the gaps in our knowledge were filled, the distribution of terms with this meaning would probably be found to be continuous. On the other hand, there is an area in north central California in which specific terms for the child-in-law 's parent have not been found. In this area are the Wintun, Pomo, Yuki, and perhaps Washo. For some of these information may be imperfect; but, on the other hand, the area may extend much farther northward. Mohave is also not known to possess a term of this significance. This may be due to mere over- sight in recording, or to proximity of the Mohave to the Pueblo Indians, who do without. In the northwest, however, the relationship is expressed in Yurok, though possibly the primary meaning of the term is more general. Great-grandparents and great-grandchildren are next most fre- quently denoted. Again Mohave stands out from a central and southern group, which consists of Luiseno, Kawaiisu, Tiibatulabal, Yokuts, and Northern Paiute. The terminology in most of these is obviously secondary: derivatives from words meaning brother or sister, mother's sister, and ghost occur. Sex is sometimes denoted and some- times not; the number of terms varies from one to four. In Luisefio there are terms for ancestors as far removed as the sixth generation. Miwok, the three north central systems, and Yurok lack words of this class, ancestors or descendants of the third generation being merged in those of the second. Mohave uses outright brother terms. Specific cousin terms are restricted to the extreme south. Luisefio possesses one for cross-cousins, Mohave three or four narrowly limited words for particular kinds of parallel and cross-cousins.- Luisefio is the only language known to have a term for grand- mother's brother or grandfather's sister. Several other relationships are included, but they are all three step. Miwok, finally, has specialized in developing four terms to denote kindred of relatives-in-law: haiyeme, kumatsa, moe, pinuksa. The peculiar Miwok terms may possibly be connected with a type of kin marriage that is best known among this group; though the correlation remains to be established. The cousin, great-grandparent, and child-in-law's parent terms, on the other hand, fail quite clearly to correlate in their distribution with any social practices. The last 1917] 377 378 University of California Publications in Am. Arch. and Ethn. [Vol. 12 might be thought to be associated with parent-in-law taboo; but, while still imperfectly known, the spread of this custom seems to run with- out relevance to that of the term. It is therefore difficult to avoid the conclusion that the occurrence of all these classes of terms is due to a merely conceptual attitude-a habit of mind or manner of thought which, originating among one people, was often gradually imitated by others. CLASSIFICATION OF THE CALIFORNIAN SYSTEMS The twelve systems that have been analyzed fall spontaneously into three classes. The first comprises the Mohave and Luiseino, both in southern California. The second consists, in the present state of knowledge, of Yurok alone. The third includes all the remainder, from the Yuki in the west and north to the Northern Paiute in the east and the Yokuts and Kawaiisu in the south. The geographical distribution of these three types, which have been established solely on the basis of what seems to be their inherent nature, coincides with the distribution of types of native civilization generally accepted for California; in other words, the three primary culture areas-the Southern, the Northwestern, and the Central. Within the central group of kinship systems a generic and a specialized subtype are distinguishable. The former is represented by Yokuts; bv Northern Paiute and Washo, which must be treated as a unit; by Tiibatulabal and Kawaiisu; and probably by Yuki. No two of these systems are alike, but their differences are particularities of comparatively little moment as against their similar features. All of them are peripheral in the territory in which they occur. As the center of this tract is approached marked divergences begin to appear on the one side among the Pomo and on the other with the Miwok, until, in the heart of the area, among the southern Wintun, the specializing tendencies reach their height. The characteristics of the southern Californian type of kinship are an enormous development of reciprocal expression, and a striking reduction of the terms denoting connections by marriage. Perhaps equally important intrinsically is the consistent recognition of the factor of lineage, as expressed terminologically in the distinction of cross and parallel relatives; but this is not an exclusive southern peculiarity. All of these traits seem typical also of the systems of the Kroeber: Calif ornia Kinship Systems Southwest, with which region southern California has many cultural correspondences. The central Californian type, in its generic and presumably original form, is marked by consistent reciprocity within the grand- parent and uncle classes of terms, but little at other points; by the distinction of cross and parallel relatives throughout; and by a fairly elaborate development of nomenclature for connections by marriage, parents-in-law, however, being denoted by a single term. This type of system seems to extend with but little variation across the Great Basin, whose Shoshonean inhabitants, it may be added, are culturally somewhat affiliated with the central Californians. The specialized southern Wintun subtype is characterized by an extreme merging of relationships into one another, and a consequently small number of terms. This tendency has completely wiped out two of the three traits typical of the generic central form of system: the reciprocity and the abundance of affinity designations. The third feature, the consciousness of kind of lineage as expressed in difference of terms for parallel and cross kindred, remains in vigor only in the uncle class. Perhaps the salient trait of the system is the merging of near lineal with near collateral relatives as a consequence of the general reduction in terminology. The Miwok and Pomo follow the Wintun scheme less radically, and add certain characteristics of their own which must be looked upon as local individualizations. The northwest Californian type, finally, if Yurok may be regarded as indicative of such a one and is not merely representative of its own particularity, disregards the distinction of cross and parallel relatives and reveals virtually no impulse toward reciprocal expression. The Yurok, to put it differently, come much nearer ourselves and the majority of Plains Indians than do any central or south California people in thinking in nearly every instance of the sex of the denoted relative'9 and only rarely of the sex of the intermediate one.20 There are some scattering data on several tribes not formally treated here. In general, these indicate systems of the tvpe prevailing in the region of each tribe. The available Costanoan data21 are in contradictory shape, but it 19 English, in 95 per cent of cases; Arapaho, Dakota, Pawnee, 90; Yurok, 85; other Californians, 60 to 80. 20 English, 0 per cent; the three Plains tribes, 10 to 30; Yurok, 20 (wholly among connections by marriage); Wintun, about the same; other Californians, 40 to 60. 21 Present series, xi, 437, 471, 1916. 1917] 379 380 University of California Publiiations in Am. Aroh. and Ethn. [Vol. 12 is clear that there was some merging of nephews and grandsons and probably of uncles and grandparents; in other words, a definite affiliation with the Wintun subtype. Salinan, from whose two dialects thirty-four terms of relationship have been preserved,22 though very variously rendered, does not show this trait. On the other hand, there is conceptual without verbal reciprocity in the grandparent and uncle classes. Indications there- fore point to Salinan belonging to the generic central type. Chumash23 is also central in character, with some leanings toward the southern type, as manifested, for instance, in distinct words used for "son" by father and mother. The primary distinction among grandparents appears to be on the basis of lineage, and among brothers and sisters on the ground of seniority, the denotation of sex being wanting or incidental. In the uncle class there are indications of four terms for seniors and four for juniors, exactly reciprocal but verbally distinct. The father-in-law and mother-in-law are denoted by one word. For the Northwest, there are scraps from three languages. Wiyot, if the translations of its half dozen known terms may be trusted,24 is of Yurok type. Hupa25 may have grandfather and grandmother terms of English and Yurok type, but the uncle-aunt nomenclature is likely to be generic Californian. Chimariko,26 finally, gives no evidence of leaning to Yurok methods. Uncertain as these meager data are, they hint that Yurok is representative of a specific California- Algonkin rather than a Northwestern Californian type. KINSHIP AND TYPE OF CULTURE A theoretical inference emerges from the distributional coinci- dence of types of kinship systems and types of culture in California. The correspondence can scarcely be accidental and meaningless. The type of culture must therefore be regarded as having helped to shape the kinship system. Now, the three Californian cultures differ but little in specific content. Nearly all the arts and ideas of one tribe recur among all the others. An inspection of a balanced museum collection from the various groups in the state invariably yields the impression of great uniformity, except as to finer detail; and reviews 22 Same, x, 169-172, 1912. 23 Same, II, 42, 1904, and a few unpublished notes from Santa Barbara. 24 Same, Ix, 407, 1911. 25 Same, III, 15, 1905. 26 Same, v, 352, 363-370, 1910. Kroeber: California, Kinship Systems of the immaterial elements of civilization have always led to the same conclusion. There are distinctive customs and practices: slavery and plank houses in the northwest, masks and moieties in the central region, sand paintings and emergence myths in the south; but relative to the totality of cultural facts such peculiarities are few. What, then, constitutes the reality and the essence of the cultural types prevailing in the three regions? Obviously, if it is not the substance of culture, it is its form; if not the discrete elements in any important degree, then their organization. An art or a custom may be practiced both in the south and in the northwest, but its emphasis or weighting be quite diverse, its associations and therefore its sig- nificance be thoroughly distinct. In short, the values of closely similar material are notably different. This is true of course of all cultural types as determined by history and ethnology and framed in culture areas or cultural periods. But in a compact and restricted territory such as California constitutes, the similarity of the civil- izational material has an opportunity to be so high as to reach sub- stantial identity; and its formal and associational individualizations become proportionally evident. These organizations or values of cultural content are in their nature general and relative as compared with the more discrete and directly given cultural material. They are also more definitely " mental, " more " psychic." When therefore we find cultures as wholes underlying kinship systems we must conclude that the latter have each been considerably influenced by the associational complex that we may denominate the "psyche" of its culture, that is, the ways of thinking and feeling characteristic of the culture. In this sense, then, we must recognize the influence, upon systems of kinship designation, of factors that, for want of another term, may be called psychological. Exactly the same conclusions are reached from an examination of the subtypes within the central Californian culture. The recog- nition within this culture area of a generalized fringe, a more definitely organized core, and a highly specialized nucleus in the region of the southern Wintun, can be established for the ceremonial aspects of religion, for instance, exactly as for kinship systems. The remoter and mountain tribes are addicted only to uncorrelated and unspecial- ized practices, which nevertheless must be accepted as representing the basis of the religion of the entire area. Inside, within the great valley, a definite ceremonial organization prevails; and this in turn appears 1917] 381 382 University of California Publications in An. Aroh. and Ethn. [Vol. 12 to reach its greatest development, and to have received most of its formative impulses, from the peoples near the center of this valley, notably the southern Wintun. In the matter of religion, the dis- tinctive achievement of the Wintun took the outward form of an elaboration; as regards kinship system, of a simplification. But in both the ritualistic elaboration and the terminological simplification there is involved a stronger adherence to an ideal scheme, more con- sequential carrying out of a consistent set of concepts, more order and organization, in short, a more developed revelation of "mental" or rather cultural activity. It would be absurd to posit the Wintun esoteric religious society and its impersonations of gods as the deter- mining cause of the abnormal Wintun system of kinship nomen- clature. But it is undeniable that they are parallel manifestations of the same manner or degree of "psychic" or civilizational operation in culture. KINSHIP AND SOCIAL INSTITUTIONS On the other hand, there are but few clear indications of an association, regional or otherwise, between types of kinship systems and types of social institutions pure and simple, that is, practices connected with marriage, descent, personal relations, and the like; and equally few instances of particular traits of kinship nomenclature according with specific institutions. Unfortunately, society is as yet perhaps the least understood aspect of the native culture of California. But we know something; and practically all the available information points in the direction of the conclusion just stated. The Mohave and Luiseino systems have been seen to be similar. Yet the Mohave are organized into clans, whereas among the Luisefio there are only halting and somewhat doubtful approaches to clans, according to the most recent information secured by Mr. E. W. Gifford. In central California a system of hereditary moieties is found among the interior Miwok, all the more northerly Yokuts, the western Mono, and probably the Salinans; and again in parts of southern California.27 It may have prevailed among a few other tribes, but its further extension can not have been very great. It is not known to have existed among the Wintun, Pomo, Yuki, Washo, Northern Paiute, or southerly Yokuts. The distribution both of types of kinship systems and of special traits of kinship designation fails to agree with 27 According to information secured by Mr. E. W. Gifford and embodied in a paper soon to appear from his pen in the present series of publications. 17roeber: California Kirnship Systems the distribution of these moieties. If there were any considerable causal correlation, the Miwok should form a unit as against the Wintun, the Pomo, the Washo, the southern Yokuts, and the other tribes of central California; whereas it appears from the previous discussion that the relations of these systems are quite otherwise. In the northwest, it is difficult to recognize any specific social factors that might be correlated with the peculiar system of this region, or at least of the Yurok. There is extant for this area Dr. Goddard's excellent monograph on the Hupa, and I have undergone repeated association with the Yurok themselves, with the opportunity of seeing much of their intimate life; yet I cannot name a single strictly "social" aspect of their culture which is not closely similar to the corresponding institutions of all the other Indians of the state, with the lone exception of the fact that northwestern marriage is a definite purchase and the wife true property. With the best endeavor I cannot, however, devise a satisfying connection between this phe- nomenon and the peculiarities of Yurok terminology for relatives. It might be said that the purchase obliterates the personality of the wife and merges her in the husband, so that the distinction of paternal and maternal relatives follows as a consequence. But I cannot wholly persuade myself that the Yurok mind works along this channel, even in its deepest unconsciousness; and there is the contrary argument that if the wife is a chattel and only the husband a person, the dis- tinction between the mother and the father, and their respective relatives, might be conceived of as being emphasized. The parent-in-law taboo is in force over considerable parts of central California: among the Yokuts, Miwok, Pomo, and presumably Southern Wintun, of the groups here treated. It is not practiced by the Yurok, Yuki, Tiibatulabal, Kawaiisu, Luisefno, or Mohave, and probably not by the Paiute and Washo. The custom might be cor- related with the Wintun subtype of kinship system; but the corre- spondence does not seem very exact. The taboo of the name of the dead, and of any allusion to them, is universal in California, and the various tribes adhere to the ob- servance with much the same scrupulousness and emotional intensity; yet devices for avoiding or altering the designations of affinities by marriage after the decease of the person connecting them seem to vary considerably. Of course such devices ensure only a formal compliance with the taboo precept; in substance they can be regarded as just as potential for emphasizing the remembrance of the death. In fact, 1917] 383 384 University of Californi,a Publications in Am. Arch. and Ethn. [Vol. 12 we cannot be sure that such is not their true subconscious function. It would seem therefore that such terminological devices may prove to be the product of several interacting and perhaps conflicting cul- tural attitudes. Where customary marriage of relatives prevails, it would seem likely to have some influence on kinship systems. This seems to be the chief reason for the undoubted correspondences of terminology and social practice in certain parts of Australia and Oceania; exogamy, descent, and marriage classes appearing to be involved, whereas a theoretically or actually prescribed marriage to certain kindred is the true shaping factor. With such marriage, definite and prescribed personal or functional relations between non-marrying relatives may be associated as a concurrent influence. In California, however, we hear very little of specifically determined relations between kindred; and other than the universal levirate, and its reverse, marriage with the wife's sister, the only form of marriage of kindred recorded is the cross-cousin wedlock of the Miwok and some neighboring groups. Mr. Gifford has shown very convincingly by analysis of circumstantial evidence that Miwok cross-cousin marriage, which does not accord with the Miwok designations of kindred, is a secondary result of the marriage of a man to his wife's brother's daughter. This form of marriage, and the marriage of the brother's widow or the wife 's sister, are reflected in Miwok nomenclature to the extent that a dozen kinship terms are in thorough accord, in their full range of meanings, with each of the practices. It is thus clear that certain forms of what might be described as statutory marriage have helped to shape and color kinship terms among the Miwok; and the same condition may be expected to prevail among other tribes. The marriage to the wife's brother's daughter I am disposed to regard as a local modification, under the influence of the moiety system, of the widespread Californian custom of marrying the wife's daughter. Where there are moieties, the wife's daughter must be of the same exogamous division as her stepfather and therefore in- eligible to him; the wife's cross-niece, that is, her brother's daughter, is the nearest relative available to take her place. Over most of California, accordingly, it is marriage to the wife's daughter, the wife 's sister, and the brother's widow that would have to be examined as potential influences upon the kinship system. There are a number of indications that this influence has been realized. Such, for instance, is the designation of the mother's younger sister and the stepmother Kroeber: Californita Kinship Systems by a single word. On the whole, however, my material is so much less complete than Mr. Gifford's Miwok data, especially in lacking most of the remoter meanings of the recorded terms, that any intensive examination of the degree of correlation on these points would be prematurely unsatisfactory. To return once more to general social structure, it is highly prob- able that all differences in the formal organization of society are superficial in California. Most tribes lack any such formal scheme; and where it exists, as among the Miwok and the Mohave, it rests lightly upon the whole cultural fabric. Its points of contact with the civilizational complex are few, its impressions of the lightest. This is shown by the fact that organizations like that of the Miwok remained undiscovered for many years. On a reasonably wide view, accordingly, society appears to be substantially the same in type in all parts of California; in contrast with which condition, kinship systems display a rather profound diversity. SUMMARY In fine, types of kinship classification exhibit so close a distri- butional correlation with types of culture as complex wholes, that it must be concluded that these cultural wholes have been influential in determining the fundamentals of kinship systems. The characteristics of such culture wholes consist in associations or relations rather than in content; and it is the formalizing or "psychic" impulses implied in these associations or relations that accordingly have largely shaped kinship terminology. On the other hand, specific social structure on the whole shows very little correlation with kinship classification in California. At one or two points a specific-element of culture content, especially prescribed marriage between relatives, has unquestionably affected kinship terminology at specific points, without, however, ap- pearing to affect its fundamental plan consequentially. THEORETICAL CONSIDERATIONS Some years ago I tried to substantiate a conviction that the custo- mary discrimination between "classificatory" and "descriptive" kin- ship systems was erroneous and misleading; that a truer and more useful distinction between these two kinds of consanguineal termin- ology could be found through a consideration of the differences of method employed by various nations in handling certain groups of 1917] 385 386 University of California Publications in Am. Arch. and Ethn. [Vol. 12 concepts, in short, through an analysis of psychological factors; and that in general such psychological factors were chiefly determinative of kinship designations.28 This position has been reviewed and combatted by Dr. W. H. R. Rivers in his admirable little book, Kinship and Social Organization,29 devoted to the thesis that kinship nomenclature is shaped chiefly by social institutions. Nearly every one who has subsequently discussed the matter in print has wholly or largely endorsed the view of Dr. Rivers. I must admit that my essay is characterized by some over-state- ments. I do not wish and have never wished to maintain so sweeping and unqualified a proposition as that terms of relationship reflect psychology wholly and sociology not at all. When it is the custom among a people for a man to marry his mother's brother's daughter, and also the custom for him to call his father-in-law his mother's brother, it would be dogmatic and a waste of time to argue against the very high probability of the two practices being connected. In regard to what may be construed as a retraction, I will only urge that the view which I was criticizing, and which Dr. Rivers has come to rescue, had been practically unquestioned for nearly forty years, and had attained considerable vogue even outside of specific ethnological circles. It had also been held without any real exami- nation of the validity of its involved assumptions. That in venturing into opposition I was led-in one or two of several recapitulations of my position-into an unnecessary curtness of expression, was there- fore perhaps natural. What is more to the point, I believe it to be a matter of little moment to the real issue. The uinderlying aspects of this issue are touched upon in the last paragraph of both Dr. Rivers' essay and mine. In this conclusion I deplored the inclination of modern anthropology to "seek specific causes for specific events," and maintained that "causal explanations of detached anthropological phenomena can be but rarely found in other detached phenomena." Dr. Rivers, on the contrary, affirms that kinship nomenclature presents a case "in which the principle of determinism applies with a rigor and definiteness equal to that of any of the exact sciences. " He avows as his chief object the demonstration that the forms of kinship designation have been determined by social conditions; and concludes that "only by attention to this aim [deter- 28 Journ. Roy. Anthr. Inst., xxxix, 77-84, 1909. 29 London, Constable & Co., 1914. 1roeber: California Kinship Systems ministic proofs] throughout the whole field of social phenomena can we hope to rid sociology of the reproach, so often heard, that it is not a science; only thus can we refute those who go still further and claim that it can never be a science." Dr. Rivers thus maintains and I deny that social science is a true science. If I understand him correctly, he is interested in why things are, I primarily in how they are. His steadfast motive is to explain social phenomena, whereas I deliberately limit my purpose to char- acterizing them. Without a recognition of this diversity of conception of the aim, and therefore the method of ethnology, the essential re- lation between the views held by Dr. Rivers and myself in regard to the comparatively small question of kinship designations can not be thoroughly and significantly apprehended. From the one point of view, an intrinsic interest inheres in any group of social phenomena as such. If they are analyzed, it is chiefly that they may be more fully apperceived; if they are synthesized with others, it is because the phenomena themselves become more truly known in proportion as their relations to the whole of civilization are visible and realized. To the other attitude of investigation, phenomena are only a starting point. This method seeks abstractions; it deter- mines causes and effects. However frequently it returns to actual phenomena, it perpetually uses these only as a ladder by which to mount to higher and wider generalizations. Dr. Rivers maintains that a non-deterministic ethnology is not science. I do not consider an ethnology which professes ability to explain much of culture to be ethnology. On this general distinction of purpose hinge the differences of opinion as to kinship terms. From Lewis H. Morgan to Dr. Rivers, generic stages of social development or broad principles have been sought; and kinship systems as a rule have been only pegs on which to hang theories concerning such stages. Whatever value my paper may or may not have had, it did not share this aim, and represents a genuine attempt to understand kinship systems as kinship systems. The concepts or categories with which the essay operates are not new. All of them may be found distinguished, for instance, in the work of Dr. Boas. But a systematic and comparative application of them led to the recognition that the current divisions of systems into "classi- ficatory" and "descriptive" was misleading in that it did not refer to the most essential features of our systems as contrasted with those of so-called savages. I then attempted to show that a deeper classi- 1917] 387 388 University of California Publioati'ons in Am. Arch. and Ethn. [Vol. 12 fication, and therefore interpretation, could be based on an analysis of the use to which the categories are put by various nations. Right or wrong, serviceable or not, this was at least an effort at construction, and therefore the essential part of the essay, as appears from the fact that three of the four propositions in the summary are devoted to this interpretation. It is exceedingly significant that these proposi- tions have been entirely ignored by Dr. Rivers, and by nearly every one else who has been concerned with the subject; whereas my fourth proposition, which was essentially negative in that the primacy Iwhich it awarded to psychological over social determinants assailed the cur- rent method of utilizing kinship designations for social reconstructions -this negative proposition aroused sufficient interest to cause Dr. Rivers to devote considerable part of a book to it. I am confident that if the main argument of my essay had been the unfolding of a theory-a causal hypothesis-instead of an endeavor merely to realize phenomena better and facilitate their being understood still more truly in the future, it would not have been passed over in silence. The particular form which ethnological theorizing has most fre- quently taken has been the formulation of schemes of development of institutions, with a special predilection for schemes of development of those institutions that are concerned with marriage and descent. For the elaboration of such schemes, kinship terminologies are plaus- ibly promising. And, on the other hand, if kinship terminology does not consistently mirror the organization of society, an important buttress for such theoretical reconstructions falls. It was logically necessary for Dr. Rivers to write Kinship and Social Organization before writing the History of Melanesian Society. If the contention that kinship systems are determined by psychological factors is only partly correct, one of the most serviceable methods of reconstructing former stages of society is eliminated. I can and do without prejudice avow sociological determinants beside "psychological"30 ones-for that 30I regret the term "psychological," and should use another were it not that its avoidance now might seem an evasion of the issue raised by me seven years ago, and by some would certainly be construed as an admission that I had shifted the basis of my contention. I do not mean, and have never meant, that terms of relationship can be explained directly from the constitution of the human mind. They are social or cultural phenomena as thoroughly and com- pletely as institutions, beliefs, or industries are social phenomena, and I am in absolute accord with Dr. Rivers' conviction that social phenomena can be understood only through other social phenomena. In common with most anthro- pologists, I hold any attempt to derive cultural facts directly from the nature of human mentality to be illusory. Culture and all its parts are a content. They are framed and limited indeed by mentality. But the endeavor to express the nature of the content through the nature of the mechanism of mentality is as vain as it would be to explain the quality of a substance in terms of its form, 1Kroeber: California Kinship Systems matter, economic and religious ones also. Dr. Rivers cannot concede "psychological" influences beside his sociological ones, because there- with his supposed recording instrument or index becomes inaccurate. He is establishing positive determinations of causality, or at least of sequence, and cannot admit variable and undeterminable factors into his calculations. The real question regarding kinship designations therefore is not the literal one of whether the terminology is wholly of psychological or of institutional origin. Nor does it very seriously concern the relative strength of each of these influences as a general proposition. It would be as silly to quarrel about that as to argue whether there are more flat or more round objects in the world. In such matters each case must be considered separately and no principle is involved. The true immediate issue is whether kinship terminologies are deter- mined so thoroughly by institutions that they can be reliably used to construct hypothetical schemes as to institutions, or whether their determinants are so frequently non-institutional that they cannot be or to approach an understanding of the sense of written words through a study of the pen. When, therefore, I spoke and now speak of terms of relationship as conditioned by "psychological" factors, I have in mind the sort of factors to which a philologist might properly ascribe the presence of a grammatical dual in a language. These factors would obviously be comparatively vague and abstractable. In a sense they would be characterizable, like evervthing in speech, as directly expressive of a manner of thought-not of course a spon- taneous outgrowth of the pure human mind uninscribed by culture, but rather, as it were, a more general and conditioning aspect of cultural content. Dr. Rivers' views, on the other hand, I should compare-if I may without prejudice use an unflatteringly crude comparison which nevertheless I believe to be true in spirit-to the explanation of the grammatical dual in speech as due to the prevalence of dualistic philosophy, or the institution of non-pluralistic marriage, that is, monogamy. When I state that the use of identical terms for such relatives as the father-in-law and grandfather, or the brother-in-law and brother in some languages, is to be understood as " due to " the fact that these relation- ships possess several categories of kinship in common, this abstract similarity is obviously not the ultimate or whole cause, since this interpretation leaves unexplained the fact that in most languages these relationships are denoted by distinct terms. That one language employs certain categories of kinship classi- fication and slights others, and another language employs and slights different ones, is itself obviously a cultural or social phenomenon; but it is precisely these varying tendencies of languages and nations toward the use of the cate- gories that I denominate " psychological " factors. Perhaps " sociological " would have been a better word, though probably liable to misinterpretation in other ways. If Dr. Rivers or any one else can replace my " psychological" with a less elusive term, I shall be sincerely grateful to him. Meanwhile I can only continue to use the word, and trust that what is here said in regard to its significance will be sufficient to prevent confusion, and to relieve me of the suspicion of wishing to revert to the methods of mid-Victorian ethnologists.- The words "social" and "sociological" are also capable of two constructions. In the wider sense, of course, they are equivalent to " cultural " or " civil- izational." In the sense in which Dr. Rivers uses them, or I employ them in discussing his views, their significance is much narrower, and they are sub- stantially equivalent to "institutional," with prime reference to marriage, laws of descent, and personal functions. 1917] 389 390 University of California Publications in Am. Aroh. and Ethn. [Vol. 12 utilized in such endeavors. And behind this lies the larger ultimate question whether specific social phenomena of any kind can be assigned as the sole specific causes or "determining" causes of other social phenomena; or whether the nearest possible approach to "explaining" phenomena such as kinship systems lies in tracing the features of the involved "psychic" or cultural activities common to them and other phenomena. On the immediate problem, indications that influences other than social institutions enter into kinship nomenclature have already been presented in various parts of the descriptions and analvses of specific California systems that constitute the first and larger portion of the present paper. In the section devoted to a classification of these systems, further instances of the frequent dominance of "psycholog- ical" over narrowly social determinants have been adduced, as well as some evidence that the shaping influences are generic impulses rather than specific phenomena, so that the ultimate question may also be considered as answered. The case seems therefore established on the basis of concrete data which need not be recited; but it may be worth while to add some broader considerations. 1. In the first place, the obvious fact that we approach kinship systems through the terminologies in which they are expressed con- stitutes them a part of speech, and it is therefore impossible to under- stand how the serious claim can be advanced that they, should be withdrawn entirely from subjection to those psychological and lin- guistic influences which shape all language. All words necessarily classify according to certain principles which usually are not more than half conscious. There is no conceivable reason why terms of relationship should be an exception, and no evidence that they are. When we find that one nation frequently introduces the idea of the sex of the speaker into its kinship terminology and another nation fails entirely to do so, it is obvious that their classifications have been made according to a different conceptual principle; or, to put it other- wise, that the involved psychology3' is different. Now it is of course possible to meet this situation with the explanation that the psychology itself indeed differs, but that it diverges exclusively under the influence of social institutions. This attitude is certainly logically possible, but I think it will be generally granted that it is such an extreme attitude 31 The word "psychology" is to be understood in the sense discussed in note 30. Kroeber: California Kinship Systems that the probability of its universal or even general truth is slight, and that the burden of proof is clearly upon those who hold this view. We have in English the curious habit of designating an oyster or a lobster as a "shell fish." The word "fish" unquestionably calls up a concept of a smooth, elongated, free-swimming water animal with fins. The only conceivable reason why a flat and sessile mollusk with- out any of the appendages of a fish, or a legged and crawling animal of utterly different appearance, should be brought in terminology into the class of fishes is the fact that they both live in the water and are edible. Now these two qualities are only a small part of those which attach to the generic concept that the word "fish" carries in English; and yet the wide discrepancy has not prevented the inclusion of the two other animals under the term. All speech is full of just such examples, and no one dreams of explaining the multitudinous phe- nomena of this kind by reference to social institutions, former phil- osophies, or other formulated manifestations of non-linguistic life, or of reconstructing the whole of a society from a vocabulary. Such endeavors in "linguistic palaeontology" have indeed been made; but the general consensus is that while they undoubtedly contain some truth, they are on the whole of little value because the interaction of ,social and linguistic influences is too indeterminate, and each of these sets of influences too variable, to allow of any positive conclusions being attained except possibly now and then on special points. If, for instance, it were argued that English classes the oyster and lobster with fish, and that other languages, perhaps German and Chinese, class them with turtles, because the English are an insular nation that subsists on an abundance of sea food, whereas the Germans and Chinese are essentially inland peoples, the explanation would strike nearly every one as extremely far-fetched. In addition, the conflicting contention could be set up that a maritime and fish-eating people could be expected to be far more discriminating in their desig- nation of sea animals than an interior people. It seems to me that some of the explanations of kinship systems on the basis of social custom are substantially of a type with this example. 2. It is extremely important to guard against subjective selection of interpretation in a field of such delicate refinement as kinship nomenclature. When among ourselves a minister of religion or a socialist orator addresses his audience as "brothers" we say that the speakers are indulging in metaphor. When we refer to our brother-in- law as "brother" we are merely slovenly familiar or intimately in- 1917] 391 392 University of California Publications in Am. Aroh. and Ethn. [Vol. 12 correct. On the other hand, when a so-called savage names his father 's brother " father," we immediately tend to have recourse to the levirate as explanation; when he designates his cousin as "brother," we think of this as a survival of group marriage; and when he calls the members of his clan "brothers," we are inclined to assert that in his nation the family of blood kindred is entirely merged in clan organization. We forget too often that uncivilized people are as likely as we to indulge in figures of speech and in short-cuts of expression. They would be very inhuman if they did not. But, of course, the more we can reduce them to the level of machines, automatically operating according to a few simple principles, the more convenient do they become as an instrument with which to unravel theoretical speculations. 3. An influence that is wholly terminological, and therefore at once "psychological" and linguistic, is the impulse toward reciprocal de- notation or form of kinship terms. It is evident on wholly abstract grounds that this must materially affect the systems into which it enters. The moment a term implies sex and has an exact reciprocal, it is clear that the reciprocal must express the sex of the speaker, and not that of the relative, so that a variant principle is introduced; or else both terms must denote both categories, which means that the number of distinct terms is duplicated or quadruplicated. The latter is a result that most languages evidently shrink from, and the former course is usually followed. In either case, however, there is a distinct shaping of the system as a result of the reciprocating tendency. When a Papago, whose system is thoroughly pervaded by reci- procities, has words meaning "older brother or sister" and "younger brother or sister," which are reciprocal, instead of our non-reciprocal "brother" and "sister," it is as clear that this part of his nomen- clature reflects the "psychological" tendency toward reciprocity, as that terming the father-in-law "mother's brother" reflects a social institution when it is customary to marry a cross-cousin. The use of descriptive phrases instead of radical words to denote connections by marriage is again a "psychological" trait. In French, the son-in-law and daughter-in-law are denoted by distinctive stems, the parents-in-law by circumlocutory ones analogous to those of Eng- lish. German follows the English plan, except for retaining some obsolescent radicals. The same tendency has become operative in all three languages, but with varying degrees of completeness. This is simply a philological phenomenon entirely parallel to the fact that the plural of " ox " has remained " oxen " instead of becoming " oxes. " Kroeber: California Kinship Systems No one would dream of arguing that French, English, or German marital customs must be different because the kinship terms in question are formed on a different plan. And so when Luiseiio and Northern Paiute and Papago use circumlocutory expressions for many connec- tions by marriage, and Mohave and Yokuts and Tiibatulabal and Miwok do not, there is also a distinctive difference of system without any reason for an assumption of a corresponding difference in social organization. This influence of reciprocity is particularly clear when circumlo- cutiOns and reciprocal expression are combined. A Papago woman call her son-in-law moih-ok, that is, ok or father of her mos, a woman 's dautlhter's child. The son-in-law calls her the same. There is no form of marriage or social institution that will explain why an old woman should be called the father of anybody's granddaughter, why the man referring to her should speak of himself as a female, and why he should designate her the parent of his daughter's child when it is hefl daughter that has the child and he himself is without grand- children. It is clear that there is nothing at the bottom of this usage bilt a strong tendency to call a connection what the connection calls or.tOelf, operating upon a stock of descriptive terms. There is no indonvenience or confusion in speaking of one 's mother-in-law as some- body's father; for in doing so a man uses a woman's term; which combination, by exclusion, exactly specifies the lady referred to. From oiLf literal point of view, the Papago is absurdly illogical in this matter. Blit he is practical, since his procedure not only isolates the person in question as thoroughly as ours but allows him to employ the reci- ptltdity to which he is accustomed and which satisfies a habitual psychic ne6d. 4. In most discussions of kinship systems the innermost kernel is ha1dly touched upon. Relatives only one step removed are neglected fom those two and three steps distant. We hear much of the fact that cotilins are called brothers, but little of the entirely different methods of distinguishing brothers. A great deal is made of the circumstance that the father's brother is often merged in the father, terminologi- call3l but very little attention is paid to whether parents are desig- nated by four terms or by two or by one. The mother's brother's daughter is far more important in most kinship discussions than the sister, In short, the most fundamental and primary relationships are disregit-ded because the remote ones lend themselves to readier cor- relatioUi with social institutions. Some nations have one word for 1917] 393 394 University of California Publications in Am. Arch. and Ethn. [Vol. 12 older brother-sister and one for younger brother-sister; others one word for brother-sister of the same sex as the speaker and one for brother-sister of different sex; and still others one word for male and one for female brother-sister; in short, various peoples express re- spectively only age, relative sex, and absolute sex in this class of kin- ship; while still others express them in different degrees and combi- nations. Surely if there is anything of consequence in kinship it is these nearest of relationships, and diversity of terminological classi- fication is as extreme for them as for any other group of kindred. Yet because they do not lend themselves to theoretical reconstructions of marriage systems, they have been passed over in almost complete silence. Dr. Rivers is therefore more sanguine than accurate when he states in the conclusion of his book that not only the general char- acter but every detail of systems of relationship has been demon- strated as determined by social conditions. The parts of systems that correlate with social conditions have indeed been correlated by him; but those parts that do not correlate have for the most part not even been considered. 5. Finally, it is not only theoretically conceivable but an actual fact that terminology has at times influenced marriage institutions. This is as it should be, for in the wider sense of the word terminology is as much a social phenomenon as marriage, and an a priori denial that any class of social phenomena is capable of affecting any other class is certainly unjustifiable. In Roman Catholic nations, as Andrew Lang has pointed out, the god-father does not marry the god-daughter. Here there is no kinship at all; but the mere name has resulted in a taboo of wedlock. If civilized European people can take their meta- phors so seriously as this, it is likely that rude heathens represented as living in a world of symbolism have sometimes done so. It may be suspected, for instance, that the Chinese prohibition of marriage between persons of the same family name is due to a similar secondary scrupulousness, instead of being a survival of an ancient clan system, as it is customary to state. Of course what is wanted in a situation like this one is not a conviction that this or that inter- pretation is true, but a substantiated case made out by a sinologue who commands knowledge of his subject as well as critical faculty. Yet the instance is not without suggestiveness as it stands. It is perfectly true that every one should expect customs to shape names more frequently than names shape customs. Those who are ready to recognize a variety of factors as entering into terminology Kroeber: California Kinship Systems can admit this disproportion cheerfully. But those who are bound to schemes of rigorous and exclusive explanation through social insti- tutions can not permit the introduction of even the rarer instances of priority of terminology without fatally dulling the edge of their working tool. If the issue were primarily the narrower one of the preeminence of so-called psychological and so-called social influences on kinship systems, I should still lay more stress on the former influence, because, after all, kinship systems are terminologies, terminologies are classi- fications, and classifications are reflections of " psychological " pro- cesses-just as I should expect religious phenomena to be influenced chiefly by other religious phenomena and only in a lesser degree by social, economic, or technical factors. I also construe the evidence as actually bearing out this interpretation. Yet I am ready to concede freely that "social" influences-and religious and economic ones- have entered in some measure into kinship systems, at times to a considerable degree even. But back of this aspect of the problem lies the basic issue: whether kinship terminology is determined rigidly by specific social phenomena of only one kind, and can therefore be utilized for constructive causal explanations of societies; or whether all classes of social phenomena can and do interact on such termin- ology, and the infinitely variable play of the variable factors forbids any true determinations of causality of a sweeping character. Two irreconcilable methods of prosecuting ethnology and history here con- front each other. It is the magnitude of this conflict of ideals that gives some dignity and perhaps consequence to the question of kinship terminology, which otherwise would be but a technical if not a trivial problem. I am aware that the causal and deterministic method has in its favor the appearance of far greater productivity, and that it often tempts with immediate profit. It can give the public the hard and fast formulations and the definitely final reasons for which the public hungers. It is also assured of a warmer recognition from scientists- natural scientists-who, unable to follow each historical situation in detail, tend nevertheless to see in this method a welcome extension of their tried methods to new fields. But I am convinced as I am of few things that this method as it has been and is practiced in ethnology is vain; that its results are illusory in proportion as they are plausible; and if ever cultural phenomena are subject to causal and deterministic analysis, it will 1917] 395 396E University of California Publications in Am. Arch. and Ethn. [Vol. 12 be in ways and with results utterly different from the methods and conclusions in vogue today. It is from this conception that I have approached the problem; and ungratefully negative as the conclusions may seem, I believe that the evidence bears them out. Dr. Rivers has rendered service to ethnology paralleled by few men. He has made valuable contributions to the critical methods of recording material. He has amassed noteworthy data, and has boldly and imaginatively attacked them without recourse to interpretation by physical and organic factors, and steeled himself no less against the more insidious temptation to explain culture in immediate terms of spontaneous psychology. There are those who wish that he mnight return to the path so brilliantly blazoned in the earlier Todas rather than continue in that pursued in the History of Melanesian Society. But all students of ethnology, those who differ as well as they who agree with his arguments, must be grateful to him for the consistency of his presentation, his courage, the directness with which he has met problems, and the precision with which he has defined them. In the present question the ultimate verdict must be left to others: I shall be satisfied if I have helped to clear the issue to the same degree on one side as Dr. Rivers has cleared it on the other. Transmitted October 11, 1916. .VN1VWRSITY OF CA tFTtIA PkrGATlQ tS on lD ) VoL '7.. 1> . he. Xe1/UlW4t >' U1e. "Pp4 S1406, plte 2, wit'''h . = -2.- Beetiest-Igons, -bd ptheQueson. Qf' tex*-nce of - VOL' .7-:.eon i the AO u s re oao, by WiUx ;e i; : on ud c -3...1iazi Baker., 0by: 0. A.. bart P. 3,$3406,p1te 1540w ............................ }~~~ln?~4 of the< '1 -TtCIC 8a R io- Dy N> C* Nel Q-,a 'son 0~~p $Q<~8 plte 3234 4. ; Peu?teFr ziXber,;a B~O t. .5i- Nen40, Vol.8 1fi. AW-i(n ecrdQ the;1 -Clior . C.1ns frQ a- MaitucRp In the ;~~~~n~f Z4brary, by A. L. --eb4Pb -2Z Ma7, 19w08 .25 88, plates i458 XT ly49 -- es eb 3. The Reigo >s > *~the Lusfo njleaoDwin 'f. ,So,'h4rli Call'.~~0- - ~ ~ ~ frI,~ :-stnc G6dar -2;ubuoio. :PZp> 6948. .1t 18-19.-0+- 4 r , 10* . -a-.4.V. pm r p A.-0 - . - ..18. --tSJ~~~~~ VW, M->g X -Or 1i n;0*ja- 5. r~o, Q- 0OoOI 3. D lcs ofs 4Soter 0agri b A. L ff. K.io...X s-t : ~ ~ hr Pp ,3~i . ..t. .b_ r -1909 -f .80'5 ne, x- -: --- pp. X$39469 i,, ,,,, L ., ,. - , ;,,, , ; - -'-. Vo.-1 1.-lb 7; s by R":dwa dgef.rt0g.ter th s~area Mths coliecte--'--b . .o . : . p : -2. :. a, 19r 0. ....*.5- 2.' ' ss-' The Cbniaslian Gost'oa' Lszgages 'y A. x rober Y. 8 S.The tangagao h oa i(iaothf a jnlo,b -~~~A L < h-;"- c-+.i Zrobet. .' ~345 and a.- Apxll, 124 1.50 VtoL. ;10. 1. Ph>uet1< 0od#tlt.nts g th#%l~ti~e Lang~zates Of iorz by A. :- IL Kroeber. Pp 1-2P,a' 91. .... . .. ...,. 1 W4ate . -p , -I s -. LN ?etber-,.. 9.1. . 8. PhonetIct t1erns of" te |M ha Lnag, bs A Z -Krer PP- ow r S ; - t - O od Atk I -at Z-- - &.;s Tb#t - 2V VoL iL 1 . leent.o the -.tatEem . o,.f . L* L ;; } e iags,-,> ^. ;;.yRal ~ddr.I. 1 ~~~Z,2~rigo.P 1 t8. Api,A1 .., .. 0 .-. . -A ll-9-;'_-t_ - 10 .> asi --F8.T ^ts .-6 by Pitny ............. GodarWd-. Pp. 1;-Z7 ,er y 19,15 ..................... 1.00^ U.Dhooilou Socia 0raaiEo nsut rrta&Cioni,b z~~~wr ;;epwifod -p~ 29.-( W 1916 ...-L. ...... -*05-**- -- z -7. Tb leutz -llet of\f Qoatnua $a:ed onn.Mf th Yau fD a - ~i ' -;-- -n~a b. -. Alde , : rn. Pp 3v9i4 Matlz 1 F$ ., ..-. .70 -s 7.~ California ?Ltwe Narn.t 'f jg#* -g, r, A L. r 6ebe_ -r. Pp. - * ' $b; Wler Db 1 Dt ,Ds ~~~i .. :.tf.-- n. b-A e p. 71488. Jn, ; , .70 ,,4.,Miw-lr- Mo,eti , by Z> # Sd4a% {-WI*so Glif'e2t '+-rd Pp 139-9. J;;eX-z'un : tLz . - 19153 ...... ? 1ia 1ds -t d . .. -d . .---- : w~-24 s 3 . %l f**- - ' 12 ow 5. '-Oi ltigteOV t he A.-b 'O D :e. 8P. 0 , . 0 . _b, t e ''b65%- $ - > Q Y *f -,*-W 1to - O'''I * -- A* * - ri u t - * * , - t - A s -- w ~ ' - f m , - + - t ' ' ' , ,z, ; C,? ,,,,, _ m~f7tl -d "P 26_ UNIVCRt4TY OF CALIFtONI,A PUBLICATIONS - (4ONTI UKEb)-. 6'., Tibatulabalx a ,na iawail Kinship ?erms, 1y Ddwar Winslow Gif-; ford. P ~, ~>,''-',Bp. 219-248.- PE!ebruary1917 , , .30..................... so ^ V BadeLier';s COWtrib`Atioi to e Stdy Aen Mexan Social Orgadiiaton, by T.T. Wat a. Pp. 29-28 Eebruar, 1917..- .35 a8. `iwolDk:yth, , y , dwar inlo Gifford. - Pp. 283-338 plte 6. Mat, W17 .-. -~~~~~~~~~~~............ -. .55 9. -'Califo h Syte AP.L. Zreb9er. p 39-96. May, 1917. .6 Volumsnow compte:- V.olme 1.. .190904. 4378 page a- 30 plates g '_.'^ - Volu.mt 2.- 190497, 39 and 21 3plates ........ . 3.50 Volume# . 1954 he M holog of the tupa age. 3844 pages ..... 3.50 Vom ^4; -s106-1$7. 374pages, with 5htablee p tu-, and map - 3.59 -Volme 5. . 97-191A 384 tg wth 25 pe- ..- .. .. 3,50 Volme7 1907-1910, 4 p 4 50 plts . -. -. .............. 3.50 Voum 8,;I1 *'1908-1'10. 869 'pags ad-& pXlates-'Z~* P@ .z .ww- . .-.....wPe* 3.50 V-olume 9. 1910-191, 439O * wS pag ...... .. . --. V --~--- Vo e 10.- - 1911914. ags and 41 pltds - b..........-. .50 rV t1 .14*-. l .11*907 1- 4-P&e ad-4 -............. ;......... ...... '_. .<._ -- Voume 1.1908-14916, 479-pagend: 45 pate .....3.50 - t.hTe U'niersty of VallforniaPbicto area .ofere in; exchange fo the publi- cation Df xlearned woiete and rinstitutikons, uierite an iraries:.a Coplt lisl~ of al1l -h vabll$>1catlon of the UierWt wiwll be sent upon requeSt. For sampl copies, 'lst of.f publications or other' inormator1, ad4es the Manageol' the tYnivarsity;;lPress,BerkeleZ- California, U. .S. A . ll- RltIterset sin e,-hig s:xhoude - addressd -t;o :Th Bzcphage Peartmn, Ui4ers3ity Library, BerSle, CaliforniaX .'. A. ;- V.t C..iCS.E. B.- b,l 3 . W. ilmore -andi. . Lima. , Editors. Price per1vue 43.50 -olu es sad .. .in .r..res. -ASTRO>NOMY?W.- W,* Capellditor.-. (Lick Qbservatory Mt, Hamilton, C.al): -Publiatln f thie Lck Cbs;y:r.-Vol :;sI-I complete. BOTANY-W. A. Sechl, Xditor. Price pervoum $3.5. Vo-dlun~e. 1 (pp. 418), tI- (pp.: - 360, -1III (pp.. 400), andIV (pp 397> cmlted ;-Volumes , jVI and;. progess CLSSIC:AL PHIdLOG.-.Edward B._-Olpp, Willia A.; M*rlll, Hebert C. Xttliztg,r Bditor . PLce 7 pet9vue d.. - u .. . ( .pp.270), .a..d II -(pp. 312), ce--. ; - . Vo.lume III in pro%ress . - < - X f ; d ''DUC.ATI:ON,-diFed'by *the Depatt:efltof IEduca , Pr"icelper vou $2f50. ', r_ t Gh aG-so:t uder tedir nof the eng epartens. his series wi'7- -'iill contai contibtlons fron theohleges .of Mehncs,. Dklg, 4.n Oiyl Eu5l pueb*,ins Voltme In it p ' or '-s ;' QO)GRAPH.-~lf .Io y Idio. Vol.unie I in pro0r a. 0; -,X GEO.LOG 5F.-BuU le*iof the Depetleut- of Geiolg. AndrYw C. ~Lawson and. 3*01m . Merriam, ditrs Price perU volumhe ;for tvolumep i;:VI, $:5.50; fiot volne eVII ud r'following, It$5. Volmes I- (pp. 435), II xpp. 457), ( 42) IV' (d 46g), V -(p0p, 48),:t VI (pp 454),- V-II pp. 50), VIII pp. 583, and IX (jPp. 545> completed. Volte 3 In progr.ess.--. ;-Lb MODER] PEI1OOG.-olnes- I'(pp. 400) II -(pp.: 373), III (-pp. 510S), an IVW (pp. 498) complete. - - -:-- ^:;-t - - :AToot,-YF-Fidedck P Gay, Ediitor. Prie pe;-rvoma $=~.50. Vlum- Jt- (pp. 347); > ;-complete< V-o iu-;>-?fl-a I- In prge ss-,- PUIIIOOPH..-, H. Eieber,; Ed6r Volue I- (pp. 20) -comipleted.~ *~ Vlume 11 in P rogr uss .r Ier vol0~n 2 PHY8IQLO(Y..th-S. S.Maellb )ditor- Pri,pr Dolm 00.lumes I ( ' 7 TN Ap 215), I (p. 197), a IV (pp. 228) ocmpleted. e1un(e V n pogress. PHYCUOLOT-.-eorge-. ; Srttn dtor.Vlne and IId' in pros.fgress. -.', '. OGLQGYF.W. I tteanid- .. Kfoi itors.< Price pervou e forviui tx, $3 :450; mr vl aume ZIi an oipweg, $5.00. _olues I (pAp 317), I (pp., 3 I (pop. 3) oVppeG)V (p 440),-VI(-.?P 475>,VI(p. 446) VII (pp.7), I - -(pp.39>. 3WX (;pp. 417), Xl (PP. 538), aini XI (Ppp 58>cOmpeted Volumes XII : :to RfXVI nlusie in pr;ogress.* ;3 tIVERSITY -OF CAI(I } CliaNICLI.A -officIal -record o0-f Un : llt life, -','',isud- quartely edilted-< ya,-commite of the Fclty. ,Price,- $1.0 per,-year.. C; ur - rent rv solie 0o. Ae.. -Addressaml orders or re dts foriormation concerningthe -above pubHicatios to The Unversityb of. CaUforn Prees ~erly, 0alfoni.S.-. - _, '',,',,,,',,','".,',,','.' ,