The change of terms after the death of a connecting or other relative is another feature of considerable interest. Thus, the Kawaiisu of California address the father as muwuni, but by the quite distinct term kuguni after the loss of a child.[58-v] Again, the Kootenai have one word for the father-in-law before and another after the wife’s or husband’s death. This peculiarity appears also among Californian tribes, the Chinook, Quileute, and several Salish tribes. This distribution again demonstrates diffusion from a common center. On the other hand, the probably even higher development of post-mortem nomenclature among the Timucua of Florida[59-v] cannot be ascribed, in the present state of our knowledge, to anything but independent origin, though we are not in a position to state what common cause, lacking in the intervening area, produced the common effect in the southeastern United States and in the remote regions of the Far West.
I will only call attention to one other kinship usage of more general interest, that embraced in the term ‘teknonymy’, the custom of denoting an individual in terms of his relationship to a child, viz., ‘father of Mary’, ‘grandmother of John’. This practice exists in South Africa and India,[60-v] in Melanesia,[61-v] and in the Pueblo area and on the Northwest coast of North America.[62-v] Tylor connected it with the custom of the husband’s residence with his wife’s kin, of the father’s assertion of his paternity and his ultimate recognition as more than a stranger by the wife’s family with whom a condition of ceremonial avoidance obtains. However, it should be noted that among the Zuñi and the Hopi, though the husband lives with his wife’s people, there is no parent-in-law taboo, and the wife is as often referred to teknonymously as the husband. Thus, my Hopi interpreter always spoke to me of his wife as ‘Herman’s mother’. Tylor’s explanation is accordingly inadequate and would seem to require at least amplification. But whatever result a systematic survey of the subject may lead to, it is certain that the effect of diffusion will have to be taken into account. It is inconceivable, e. g., that the practice originated independently among tribes so geographically situated and so intimately related in culture as the Zuñi and Hopi.
Special Forms of Marriage and Social Customs. There can be little doubt that a well-established marriage rule often finds expression in nomenclature. Even the exogamous principle can be brought under this head since it expresses the potential matrimonial status of members of the community. In a dual organization my ‘father’ is one who potentially, if not actually, is a mate of women of my mother’s group, while a ‘mother’s brother’ is one who can under no condition occupy that status.
Of the specific forms of marriage the levirate has already been considered and the cross-cousin marriage briefly mentioned. Dr. Rivers has demonstrated the close dependence of nomenclature on the latter practice in Melanesia. Here the custom itself is found in full swing, and it would be unreasonable to deny that the terminology had its origin in this usage even in parts of Melanesia where it cannot be observed. This does not mean that cross-cousin marriage necessarily obtained throughout the range of distribution of the corresponding terminology but that the terminology spread from a center where it reflected the social institution. Thus, in Guadalcanar the cross-cousin marriage still persists and we find cross-cousins, brothers-in-law and sisters-in-law comprised under a single appellation. In Anaiteum, cross-cousins of opposite sex address one another by the terms used for husband and wife.[63-v] It seems to me methodologically quite justifiable to interpret similar features in neighboring islands as having their ultimate origin in cross-cousin marriage. But the argument fails where similar connotations of terms occur without evidence of the marriage rule unless it can be demonstrated that no other cause could have produced the result. Thus, I must consider unsuccessful Dr. Rivers’ attempt to deduce, though with qualifications, the former existence of the institution in question from the system of the Dakota Indians.[64-v] The classification of brothers-in-law with cross-cousins might be simply a reflection of the dual organization, by which these relatives would fall within the same group; or, to put it differently, if the term cross-cousin is given the wide significance with which we are familiar in primitive systems, so as to include members of the opposite moiety and one’s own generation, a man’s brothers-in-law are necessarily members of the cross-cousin class. The superiority of the moiety hypothesis in this instance lies in the fact that the dual organization occurs among several contiguous and related tribes while the cross-cousin marriage is extremely rare in North America and its highest development occurs among remote peoples of the Pacific region. Regarding special forms of marriage, it is rather important to ascertain whether the terms used by our authorities are to be interpreted in our own or in the more inclusive primitive sense. For example, Tylor reduced the institution of cross-cousin marriage to the principle of exogamous moieties by assuming the wider significance.[65-v] As Dr. Rivers points out,[66-v] the two rules are not identical if marriage is prescribed with the own daughter of the own mother’s brother. In that case, the moiety rule is only a larger framework with which the specific institution is not incompatible but which does not determine cross-cousin marriage. Looking at the matter chronologically, I can even conceive the development of larger social groups from such specific marriage regulations. If in the absence of an own cross-cousin, a more remote cousin comes to be regularly substituted, we should have a whole class of possible mates, of whom the nearest cross-cousin would be only primus inter pares.
It must be understood that while special marriage regulations, like exogamy, tend to be mirrored in nomenclature, there is no absolute necessity for this occurrence. As the New Mexican Tewa have exogamous groups without the Dakota principle, so the Miwok of California have the cross-cousin marriage with little or no indication of it in terminology.[67-v] One factor that must always be considered in this connection is the time element. A recently acquired custom may not yet have developed an appropriate nomenclature, while, as Morgan supposed, the nomenclature may survive after the custom has become obsolete. That the frequency of marriage according to a certain rule, and the coexistence of other rules, possibly antagonistic in their effects, must have an influence, is obvious. As regards the latter point, Mr. Gifford shows that while marriage with the cross-cousin is not suggested in Miwok nomenclature, marriage with the wife’s brother’s daughter is reflected by twelve terms.
Among the Thonga of South Africa several interesting forms of preferential matrimonial union occur. As among the Miwok, marriage with the wife’s, younger sisters and wife’s brother’s daughter is considered peculiarly appropriate, and these affinities are subsumed under a common caption. Levirate extends only to the elder brother’s, not to the younger brother’s, wife, and quite consistently these affinities are distinguished by distinct words. A man may inherit his maternal uncle’s wife and therefore classes her with the wife. On the other hand, logic does not hold sway undisputedly. A man calls cross-cousins by the same term as parallel cousins and brothers, yet it is possible for a man to inherit his parallel cousin’s, but not his cross-cousin’s (father’s sister’s son’s), wife. The explanation given by Junod seems quite satisfactory from a comparative point of view. My cross-cousin cannot belong to my gens, my parallel cousin must belong to it.[68-v] Since the Thonga usually distinguish marriage potentialities with considerable nicety, we may reasonably infer that the present terminology for cousins is a recent innovation, which conclusion once more indicates the relatively late development of Hawaiian features.
A systematic comparison of the effect of definite forms of marriage on nomenclature, in different parts of the world is highly desirable. When we shall have examined how such an institution as the inheritance of a maternal uncle’s wife affects the systems of the Tlingit of northwestern America, of the Banks Islands in Melanesia, and the Thonga of South Africa, and know the action of whatever coexisting institutions may occur, we shall have gained considerably more insight into a very suggestive problem. It is fairly clear that a form of marriage does not determine nomenclature univocally, as the facts relating to the levirate indicate. To ascertain in how far parallelism actually occurs, is a matter of great moment.
Conclusion. The question with which this chapter opens has now received an answer. Terms of relationship form a proper topic of investigation for the ethnologist, first because they are often directly correlated with cultural phenomena, such as social usages regulating marriage; secondly, because the features of kinship nomenclature are an index of tribal relationship. Any particular system is not a unified logical whole but a complex product of internal development and foreign connections. Accordingly, its features cannot be understood by themselves any more than other cultural phenomena, but only in association with concomitant traits of the native culture and in the light of a comparative survey of like features among neighboring tribes and ultimately throughout the world. By utilizing our ethnographical knowledge in applying the method of variation it is possible to ascertain, at least to a considerable extent, the causes, whether primary or secondary, that have shaped a given system.
When, for example, we endeavor to explain the system of the Hopi, we can start with the fact that their speech constitutes them a member of the Shoshonean family, i.e., we can begin by comparing Hopi nomenclature with that of the Paiute, Paviotso, Ute and Shoshone. One fact that strikes us here is the great difference in the actual vocables employed by the Hopi from those of their congeners, an observation which by no means extends to all of their language. Morgan held the view that kinship words were the most persistent elements of speech, but however this rule may work in other stocks, such as the Athabaskan, it certainly does not obtain among the Shoshoneans, nor, I may add, within the Siouan family, where even such closely related languages as Crow and Hidatsa reveal far greater differences in the lexicon of relationships than might be expected from other vocables. It is, however, in the classification of kin that the distinctiveness of the Hopi seems most remarkable. Their system is not characterized by the prominent features of the Plateau Shoshonean terminologies, such as reciprocity and the separation of paternal from maternal grandparents. On the other hand, they employ the Dakota principle with the Hidatsa variation. That variant occurs, so far as we now know, only among peoples historically quite unrelated to the Hopi so that neither genetic connection nor dissemination accounts for the similarity. On the other hand, all the tribes having this feature share exogamous groups with maternal descent. Such clans are characteristic of the Hopi also, but are lacking among the other Shoshoneans. We infer from this that the Hidatsa variant among the Hopi is functionally connected with their clan system. If the neighboring Zuñi do not share this characteristic, a possible explanation may be found in the relative weakness of the Zuñi clan concept, as recently expounded by Professor Kroeber, when contrasted with its dominance in the social life of the Hopi. In other features the intimate cultural contact between the Zuñi and Hopi is emphatically apparent. Probably for no other tribes is there evidence for such exaggerated reliance on teknonymy, while a certain looseness in the use of terms common to both seems to be a general Southwestern trait. The Hopi system thus reflects both the social fabric of the tribe and its historical relations,—the ancient ones reduced to a few lexical resemblances, while the more complex tribal organization and recent cultural affiliations with the Southwest, and particularly with the Zuñi, stand out in bold relief.
A strictly similar inquiry might be made into the system of the Crow. Here the almost complete coincidence of certain very unusual features with Hidatsa ones bears eloquent testimony to the exceptionally close genetic relationship of the two tribes. Thus, a wife who has been married before is distinguished by a specific word, and spouses generally refer to each other not by a specific term, which seems restricted to non-vocative usage, but by a demonstrative expression. Not only is there a confusion of generations according to the Hidatsa variant, but the mother’s brother is classed with the elder brother and so is the mother’s mother’s brother. The last-mentioned features are partly found among the Mandan. All three tribes differ from the other Siouans, and indeed from all other Plains Indians in having matrilineal descent. Since this is likewise the rule among genetically unconnected peoples sharing the Hidatsa variant, we regard the latter as functionally connected with the clan organization. But there are other traits in which the terminology of the Crow differs from that of their nearest congeners, and here we must systematically consider the possible effect of all such peoples as the Oglala, or Blackfoot, with whom they have come into contact. Such divergence may be merely the effect of internal readjustment. Thus, the Crow classification of the father’s sister’s husband with the father admits of a plausible interpretation as the result of another peculiarity—the classing of the father’s sister with the mother in direct address. Instead of having two deviations from the Hidatsa norm, we should thus have at bottom only one.