Kinship and social organisation/Lecture II

by W. H. R. Rivers

 

LECTURE II

In my last lecture I began the demonstration of the dependence of the classificatory terminology of relationship upon social institutions by showing how a number of terms used in several parts of Melanesia have been determined by the cross-cousin marriage. I showed that in places where the cross-cousin marriage is practised there are not merely one or two, but large groups of, terms of relationship which are exactly such as would follow from this form of marriage. To-day I begin by considering other forms of Melanesian marriage which bring out almost as clearly and conclusively the dependence of the classificatory terminology upon social conditions.

The systems of relationship of the Banks Islands possess certain very remarkable features which were first recorded by Dr. Codrington.^[1] Put very shortly, it may be stated that cross-cousins stand to one another in the relation of parent and child, or, more exactly, cross-cousins apply to one another terms of relationship which are otherwise used between parents and children. A man applies to his mother's brother's children the term which he otherwise uses for his own children, and, conversely, a person applies to his father's sister's son a term he otherwise uses for his father. Thus, in the following diagram, C will apply to D and e the terms which are in general use for a son and daughter, while D and e will apply to C the term they otherwise use for their father.

DIAGRAM 2.

B=a

A=b

C

D

e

In most forms of the classificatory system members of different generations are denoted in wholly different ways and belong to different classes,^[2] but here we have a case in which persons of the same generation as the speaker are classed with those of an older or a younger generation.

I will first ask you to consider to what kind of psychological similarity such a practice can be due. What kind of psychological similarity can there be between one special kind of cousin and the father, and between another special kind of cousin and a son or daughter? If the puzzle as put in this form does not seem capable of a satisfactory answer, let us turn to see if the Banks Islanders practise any social custom to which this peculiar terminology can have been due. In the story of Ganviviris told to Dr. Codrington in these islands^[3] an incident occurs in which a man hands over one of his wives to his sister's son, or, in other words, in which a man marries one of the wives of his mother's brother. Inquiries showed, not only that this form of marriage was once widely current in the islands, but that it still persists though in a modified form. The Christianity of the natives does not now permit a man to have superfluous wives whom he can pass on to his sister's sons, but it is still the orthodox, and indeed I was told the popular, custom to marry the widow of the mother's brother. It seemed that in the old days a man would take the widow of his mother's brother in addition to any wife or wives he might already have. Though this is no longer allowed, the leaning towards this form of marriage is so strong that after fifty years of external influence a young man still marries the widow of his mother's brother, sometimes in preference to a girl of his own age. Indeed, there was reason to believe that there was an obligation to do so, if the deceased husband had a nephew who was not yet married. The peculiar features of the terminology of relationship in these islands are exactly such as would follow from this form of marriage. If, in Diagram 2, C marries b, the wife or widow of his mother's brother, and thereby comes to occupy the social position of his uncle A, the children of the uncle, D and e, will come to stand to him in the relation of children, while he, who had previously been the father's sister's son of D and e, will now become their father. An exceptional form of the classificatory system, in which there is a departure from the usual rule limiting a term of relationship to members of the same generation, is found to be the natural consequence of a social regulation which enjoins the marriage of persons belonging to different generations.

The next step in the process of demonstrating the social significance of the classificatory system of relationship will take us to the island of Pentecost in the northern New Hebrides. When I recorded the system of this island, I found it to have so bizarre and complex a character that I could hardly believe at first it could be other than the result of a ludicrous misunderstanding between myself and my seemingly intelligent and trustworthy informants. Nevertheless, the records obtained from two independent witnesses, and based on separate pedigrees, agreed so closely even in the details which seemed most improbable that I felt confident that the whole construction could not be so mad as it seemed. This confidence was strengthened by finding that some of its features were of the same order of peculiarity as others which I had already found in a set of Fijian systems I have yet to consider. There were certain features which brought relatives separated by two generations into one category; the mother's mother, for instance, received the same designation as the elder sister; the wife's mother the same as the daughter; the wife's brother the same as the daughter's son. The only conclusion I was then able to formulate was that these features were the result of some social institution resembling the matrimonial classes of Australia, which would have the effect of putting persons of alternate generations into one social category.

This idea was supported by the system of relationship of the Dieri of Australia which possesses at least one feature similar to those of Pentecost, a fact I happened to remember at the time because Mr. N. W. Thomas^[4] had used it as the basis of a reductio ad absurdum argument to show that terms of relationship do not express kinship. The interest of the Pentecost system seemed at first to lie in the possibility thus opened of bringing Melanesian into relation with Australian sociology, a hope which was the more promising in that the people of Pentecost and the Dieri resemble one another in the general character of their social organisation, each being organised on the dual basis with matrilineal descent. When in Pentecost, however, I was unable to get further than this, and the details of the system remained wholly inexplicable.

The meaning of some of the peculiarities of the Pentecost system became clear when I reached the Banks Islands; they were of the same kind as those I have already considered as characteristic of these islands. When I had discovered the dependence of these features upon the marriage of a man with the wife of his mother's brother, it became evident that not only these, but certain other features of the Pentecost system, were capable of being accounted for by this kind of marriage. The peculiar features of the Pentecost system could be divided into two groups, and all the members of one group could be accounted for by the marriage with the mother's brother's wife. All these features had the character in common that persons of the generation immediately above or below that of the speaker were classed in nomenclature with relatives of the same generation.

The other group consisted of terms in which persons two generations apart were classed with relatives of the same generation. Since the first group of correspondences had been explained by a marriage between persons one generation apart, it should have been obvious that the classing together of persons two generations apart might have been the result of marriage between persons two generations apart. The idea of a society in which marriages between those having the status of grandparents and grandchildren were habitual must have seemed so unlikely that, if it entered my mind at all, it must have been at once dismissed. The clue only came later from a man named John Pantutun, a native of the Banks Islands, who had been a teacher in Pentecost. In talking to me he often mentioned in a most instructive manner resemblances and differences between the customs of his own island and those he had observed in Pentecost. One day he let fall the observation with just such a manner as that in which we so often accuse neighbouring nations of ridiculous or disgusting practices, "O! Raga!^[5] That is the place where they marry their granddaughters." I saw at once that he had given me a possible explanation of the peculiar features of the system of the island. By that time I had forgotten the details of the Pentecost system, and it occurred to me that it would be interesting, not immediately to consult my note-books, but to endeavour to construct a system of relationship which would be the result of marriage with a granddaughter, and then to see how far my theoretical construction agreed with the terminology I had recorded. The first question which arose was with which kind of granddaughter the marriage had been practised, with the son's daughter or with the daughter's daughter, and this was a question readily answered by means of a consideration arising out of the nature of the social organisation of Pentecost.

The society of this island is organised on the dual basis with matrilineal descent in which a man

must marry a woman of the opposite moiety. Diagram 3, in which A and a stand for men and

Diagram 3.

A=b

B=a

A=b

A

a

B

b

women of one moiety, and B and b for those of the other moiety, shows that a marriage between a man and his son's daughter would be out of the question, for it would be a case of A marrying a. It was evident that the marriage, the consequences of which I had to formulate, must have been one in which a man married his daughter's daughter.

It would take too long to go through the whole set of relationships, and I choose only a few examples which I illustrate by the following diagram :

Diagram 4.

A=b

D = c

e

F

f

This diagram shows that if A marries e, c, who previous to the marriage had been only the daughter of A, now becomes also his wife's mother; and D, who had previously been his daughter's husband, now becomes his wife's father. Similarly, F, who before the new marriage was the daughter's son of A, now becomes the brother of his wife, while f, his daughter's daughter, becomes his wife's sister. Lastly, if we assume that it would be the elder daughters of the daughter who would be married by their grandfathers, e, who before the marriage had been the elder sister of F and f, now comes through her marriage to occupy the position of their mother's mother.

When, after making these deductions, I examined my record of the Pentecost terms, I found that its terminology corresponded exactly with those which had been deduced. The wife's mother and the daughter were both called nitu. The daughter's husband and the wife's father were both bwaliga. The daughter's children were called mabi, and this term was also used for the brother and sister of the wife. Lastly, the mother's mother was found to be classed with the elder sister, both being called tuaga.

For the sake of simplicity of demonstration I have assumed that a man marries his own daughter's daughter, but through the classificatory principle all the features I have described would follow equally well if a man married the granddaughter of his brother, either in the narrow or the classificatory sense. There was one correspondence, according to which both the husband's brother and the mother's father were called sibi, which does not follow from the marriage with the own granddaughter, but would be the natural result of marriage with the daughter's daughter of the brother i.e., with a marriage in which e was married by A's brother.

I hope these examples will be sufficient to show how a number of features which might otherwise seem so absurd as to suggest a system of relationship gone mad become natural and intelligible, even obvious, if it were once the established practice of the people to marry the daughter's daughter of the brother.

Such inquiries as I was able to make confirmed the conclusion that the Pentecost marriage was with the granddaughter of the brother rather than with the daughter of the daughter herself. After I had been put on the track of the explanation by John Pantutun I had the chance of talking to only one native of Pentecost, unfortunately not a very good informant. From his evidence it appeared that the marriage I had inferred from the system of relationship even now occurs in the island, but only with the granddaughter of the brother, and that marriage with the own granddaughter is forbidden. The evidence is not as complete as I should like, but it points to the actual existence in the island of a peculiar form of marriage from which the extraordinary features of its system of relationship directly follow.

When I returned to England I found that this marriage was not unique, but had been recorded among the Dieri of Australia,^[6] where, as I have already mentioned, it is associated with peculiar features of nomenclature resembling those of Pentecost.

I must again ask, how are you going to explain the features of the Pentecost system psychologically? What psychological resemblance is there between a grandmother and a sister, between a mother-in-law and a daughter, between a brother-in-law and a grandfather? Apart from some special form of social relationship, there can be no such resemblances. Further, if there were such psychological resemblances, why should we know of their influence on nomenclature only in Pentecost and among the Dieri? The features to be explained are definitely known to exist in only two systems of the world, and it is only among the peoples who use these two systems that we have any evidence of that extraordinary form of marriage of which they would be the natural consequence.

 

I have now tried to show the dependence of special features of the classificatory system of relationship upon special social conditions. If I have succeeded in this I shall have gone far towards the accomplishment of one of the main purposes of these lectures. They have, however, another purpose, viz., to inquire how far we are justified in inferring the existence of a social institution of which we have no direct evidence when we find features of the nomenclature of relationship which would result from such an institution. I have now to enter upon this part of my subject, and I think it will be instructive to take you at once to a case in which I believe that an extraordinary form of marriage can be established as a feature of the past history of a people, although at the present moment any direct evidence for the existence of such a marriage is wholly lacking.

When I was in the interior of Viti Levu, one of the Fijian islands, I discovered the existence of certain systems of relationship which differed fundamentally from the only Fijian systems previously known. Any features referable to the cross-cousin marriage were completely absent, but in their place were others, one of which I have already mentioned, which brought into one class relatives two generations apart. The father's father received the same designation as the elder brother, and the son's wife was called by the same term as the mother. As I have already said, my first conclusion was that these terms were the survivals of forms of social organisation resembling the matrimonial classes of Australia, but as soon as I had worked out the explanation of the Pentecost system, it became evident that the Fijian peculiarities would have to be explained on similar lines. At first I thought it probable that the difference between the Pentecost and Fijian systems was due to the difference in the mode of descent in the two places. For long I tried to work out schemes whereby a change from the matrilineal descent of Pentecost to the patrilineal condition of Fiji could have had as one of its consequences a change from a correspondence in nomenclature between the mother's mother and the elder sister to one in which the common nomenclature applied to the father's father and the elder brother. It is an interesting example of the strength of a preconceived opinion, and of some measure of the belief in the impossibility of customs not practised by ourselves, that for more than two years I failed to see an obvious alternative explanation, although I returned to the subject again and again. The clue came at last from the system of Buin, in the island of Bougainville, recorded by Dr. Thurnwald.^[7] The nomenclature of this system agreed with that of inland Fiji in having one term for the father's father and the elder brother, but since the people of Buin still practice matrilineal descent, it was evident that I had been on a false track in supposing the correspondence to have been the result of a change in the mode of descent. Once turned into a fresh path by the necessity of showing how the correspondence could have arisen out of a matrilineal condition, it was not long before I saw how it might be accounted for in a very different way. I saw that the correspondence would be the natural result of a form of social organisation in which it was the practice to marry a grandmother, viz., the wife of the father's father. Not only did this form of marriage explain the second peculiar feature of the Fijian system, viz., the classing of the son's wife with the mother, but it would also account for several features of the Buin system which would otherwise be difficult to understand.

DIAGRAM 5.

A=b

C = d

E

F

f

If, as shown in Diagram 5, E marries b, the wife or widow of his father's father, he, who had previously been the elder brother of F and f, now comes to occupy the position of their father's father, while d, the mother of E, will now come to stand to him in the relationship of son's wife.

I need only mention here one of the features of the Buin system which can be accounted for by means of this marriage. The term mamai is used, not only for the elder sister and for the elder brother's wife, but it is also applied to the father's mother; that is, the wife of the elder brother is designated by the same term as the wife of the father's father, exactly as must happen if E marries b, the wife of his father's father. A number of extraordinary features from two Melanesian islands collected by two independent workers fit into a coherent scheme if they have been the result of a marriage in which a man gives one of his wives to his son's son during his life, or in which this woman is taken to wife by her husband's grandson when she becomes a widow. If the practice were ever sufficiently habitual to become the basis of the system of relationship, we can be confident that it is the former of these two alternatives with which we have to do.

If you are still so under the domination of ideas derived from your own social surroundings that you cannot believe in such a marriage, I would remind you that there is definite evidence from the Banks Islands that men used to hand over wives to their sisters' sons. It is not taking us so much into the unknown as it might appear to suppose that they once also gave their wives to their sons' sons.

I have taken this case somewhat out of its proper place in my argument because the evidence is so closely connected with that by means of which I have shown the relation between features of systems of relationship and peculiar forms of marriage in Melanesia. I have now to return to the more sober task of considering how far we are justified in inferring the former existence of marriage institutions when we find features of systems of relationship of which they would have been the natural consequence. It is evident that, whenever we find such a feature as common nomenclature for a grandmother and a sister or for a cross-cousin and a parent, it should suggest to us the possibility of such marriage regulations as those of Pentecost and the Banks Islands. But such common designations might have arisen in some other way, and in order to establish the existence of such forms of marriage in the past history of the people, we must have criteria to guide us when we are considering whether a given feature of the terminology of relationship is or is not a survival of a marriage institution.

I will return to the cross-cousin marriage for my examples. The task before us is to inquire how far such features of relationship as exist in Fiji, Anaiteum or Guadalcanar, in conjunction with the cross-cousin marriage, will justify us in inferring the former existence of this form of marriage in places where it is not now practised.

If there be found among any people all the characteristic features of a coastal Fijian or of an Anaiteum system, I think few will be found to doubt the former existence of the cross-cousin marriage. It would seem almost inconceivable that there should ever have existed any other conditions, whether social or psychological, which could have produced this special combination of peculiar uses of terms of relationship. It is when some only of these features are present that there will arise any serious doubt whether they are to be regarded as survivals of the former existence of the cross-cousin marriage.

One consideration I must point out at once. Certain of the features which follow from the cross-cousin marriage may be the result of another marriage regulation. In some parts of the world there exists a custom of exchanging brothers and sisters, so that, when a man marries a woman, his sister marries his wife's brother. As the result of this custom the mother's brother and the father's sister's husband will come to be one and the same person, and the father's sister will become also the mother's brother's wife.

This form of marriage exists among the western people of Torres Straits,^[8] and is accompanied by features of the system of relationship which would follow from the practice. The mother's brother is classed with the father's sister's husband as wadwam, but there is an alternative term for the father's sister's husband and there was no evidence that the mother's brother's wife was classed with the father's sister. It seemed possible that the classing together of the mother's brother and the father's sister's husband was not a constant feature of the system of relationship, but only occurred in cases where the custom of exchange had made it necessary. The case, however, is sufficient to show that two of the correspondences which follow from the cross-cousin marriage may be the result of another kind of marriage. If we accept the social causation of such features and find these correspondences alone, it would still remain an open question whether they were the results of the custom of exchange or of the marriage of cross-cousins. The custom of exchange, however, is wholly incapable of accounting for the use of a common term for the mother's brother and the father-in-law, for the father's sister and the mother-in-law, or for cross-cousins and brothers- or sistersin-law. It is only when these correspondences are present that there will be any decisive reason for inferring the former existence of the cross-cousin marriage.

The first conclusion, then, is that some of the features found in association with the cross-cousin marriage are of greater value than others in enabling us to infer the former existence of the cross-cousin marriage where it no longer exists. Next, the probability that such features as I am considering are due to the former presence of the cross-cousin marriage will be greatly heightened if this form of marriage should exist among people with allied cultures. An instance from Melanesia will bring out this point clearly.

In the island of Florida in the Solomons it is clear that the cross-cousin marriage is not now the custom, and I could discover no tradition of its existence in the past. One feature, however, of the system of relationship is just such as would follow from the cross-cousin marriage. Both the wife's mother and the wife of the mother's brother are called vungo.

Florida is not only near Guadalcanar where the cross-cousin marriage is practised, (the two islands are within sight of one another), but their cultures are very closely related. In such a case the probability that the single feature of the Florida system which follows from the cross-cousin marriage has actually had that form of marriage as its antecedent becomes very great, and this conclusion becomes still more probable when we find that in a third island, Ysabel, closely allied in culture both to Florida and Guadalcanar, there is a clear tradition of the former practice of the cross-cousin marriage although it is now only an occasional event.

Again, in one district of San Cristoval in the Solomons the term fongo is used both for the father-in-law and the father's sister's husband, and kafongo similarly denotes both the mother-in-law and the mother's brother's wife. This island differs more widely from Guadalcanar in culture than Florida or Ysabel, but the evidence for the former existence of the marriage in these islands gives us more confidence in ascribing the common designations of San Cristoval to the cross-cousin marriage than would have been the case if these common designations had been the only examples of such possible survivals in the Solomons. Speaking in more general terms, one may say that the probability that the common nomenclature for two relatives is the survival of a form of marriage becomes the greater, the more similar is the general culture in which the supposed survival is found to that of a people who practise this form of marriage. The case will be greatly strengthened if there should be intermediate links between the supposed survival and the still living institution.

When we find a feature such as that of the Florida system among a people none of whose allies in culture practise the cross-cousin marriage, the matter must be far more doubtful. In the present state of our knowledge we are only justified in making such a feature the basis of a working hypothesis to stimulate research and encourage us to look for other evidence in the neighbourhood of the place where the feature has been found. Our knowledge of the social institutions of the world is not yet so complete that we can afford to neglect any clue which may guide our steps.

I propose briefly to consider two regions, South India and North America, to show how they differ from this point of view.

The terms of relationship used in three^[9] of the chief languages spoken by the people of South India are exactly such as would follow from the cross-cousin marriage. In Tamil^[10] the mother's brother, the father's sister's husband, and the father of both husband and wife are all called mama, and this term is also used for these relatives in Telegu. In Canarese the mother's brother and the father-in-law are both called mava, but the father's sister's husband fails to fall into line and is classed with the father's brother.

Similarly, the father's sister, the mother's brother's wife and the mother of both wife and husband are called atta in Telegu and atte in Canarese, Tamil here spoiling the harmony by having one term, attai, for the father's sister and another, mami, for the mother's brother's wife and the mother-in-law. Since, however, the Tamil term for the father's sister is only another form of the Telegu and Canarese words for the combined relationships, the exception only serves to strengthen the agreement with the condition which would follow from the cross-cousin marriage.

The South Indian terms for cross-cousin and brother- and sister-in-law are complicated by the presence of distinctions dependent on the sex and relative age of those who use them, but these complications do not disguise how definitely the terminology would follow from the cross-cousin marriage. Thus, to take only two examples: a Tamil man applies the term maittuni to the daughters of his mother's brother and of his father's sister as well as to his brother's wife and his wife's sister, and a Canarese woman uses one term for the sons of her mother's brother and of her father's sister, for her husband's brother and her sister's husband.

So far as we know, the cross-cousin marriage is not now practised by the vast majority of those who use these terms of relationship. If the terminology has been the result of the cross-cousin marriage, it is only a survival of an ancient social condition in which this form of marriage was habitual. That it is such a survival, however, becomes certain when we find the cross-cousin marriage still persisting in many parts of South India, and that among one such people at least, the Todas,^[11] this form of marriage is associated with a system of relationship agreeing both in its structure and linguistic character with that of the Tamils. I have elsewhere^[12] brought together the evidence for the former prevalence of this form of marriage in India, but even if there were no evidence, the terminology of relationship is so exactly such as would follow from the cross-cousin marriage that we can be certain that this form of marriage was once the habitual custom of the people of South India.

While South India thus provides a good example of a case in which we can confidently infer the former existence of the cross-cousin marriage from the terminology of relationship, the evidence from North America is of a kind which gives to such an inference only a certain degree of probability. In this case it is necessary to suspend judgment and await further evidence before coming to a positive conclusion.

I will begin with a very doubtful feature which comes from an Athapascan tribe, the Red Knives^[13] (probably that now called Yellow Knife). These people use a common term, set-so, for the father's sister, the mother's brother's wife, the wife's mother and the husband's mother, a usage which would be the necessary result of the cross-cousin marriage. Against this, however, is to be put the fact that there are three different terms for the corresponding male relatives, the two kinds of father-in-law being called seth-a, the mother's brother ser-a, and the father's sister's husband selthe-ne. Further, the term set-so, the common use of which for the aunt and mother-in-law seems to indicate the cross-cousin marriage, is also applied by a man to his brother's wife and his wife's sister, features which cannot possibly be the result of this form of marriage. These features show, either that the terminology has arisen in some other way, or that there has been some additional social factor in operation which has greatly modified a nomenclature derived from the cross-cousin marriage.

A stronger case is presented by the terminology of three branches of the Cree tribe, also recorded by Morgan. In all three systems, one term, ne-sis or nee-sis, is used for the mother's brother, the father's sister's husband, the wife's father and the husband's father; while the term nis-si-goos applies to the father's sister, the mother's brother's wife and the two kinds of mother-in-law. These usages are exactly such as would follow from the cross-cousin marriage. The terms for the sister's son of a man and the brother's son of a woman, however, differ from those used for the son-in-law, and there is also no correspondence between the terms for cross-cousin and any kind of brother- or sister-in-law. The case points more definitely to the cross-cousin marriage than in the case of the Red Knives, but yet lacks the completeness which would allow us to make the inference with confidence.

The Assiniboin have a common term, me-toh-we, used for the father's sister, the mother's brother's wife and the two kinds of mother-in-law, and also a common term, me-nake-she, for the mother's brother and the father's sister's husband, but the latter differs from the word, me-to-ga-she, used for the father of husband or wife. The case here is decidedly stronger than among the Red Knives, but is less complete than among the Crees.

Among a number of branches of the Dakotas the evidence is of a different kind, being derived from similar nomenclature for the cross-cousin and certain kinds of brother- and sister-in-law. Morgan^[14] has recorded eight systems, all of which show the features in question, but I will consider here only that of the Isauntie or Santee Dakotas, which was collected for him by the Rev. S. R. Riggs. Riggs^[15]and Dorsey^[16] have given independent accounts of this system which are far less complete than that given by Morgan, but agree with it in all essentials.

In this system a man calls the son of his mother's brother or of his father's sister ta-hang-she or tang-hang-she, while his wife's brother and his sister's husband are ta-hang or tang-hang. Similarly, a woman calls her cross-cousin she-chay-she, while her husband's brother and her sister's husband are called she-chay. The terms for brothers-in-law are thus the same as those for cross-cousins with the omission of the suffix she. One of these resemblances, that when a woman is speaking, has been cited by Professor Kroeber^[17] as an example of the psychological causation of such features of relationship as I am considering in these lectures. He rejects its dependence on the cross-cousin marriage and refers the resemblance to the psychological similarity between a woman's cousin and her brother-in-law in that both are collateral relatives alike in sex, of the same generation as the speaker, but different from her in sex.

As we have seen, however, the Dakota correspondence is not an isolated occurrence, but fits in with a number of other features of the systems of cognate peoples to form a body of evidence pointing to the former prevalence of the cross-cousin marriage.

There is also indirect evidence leading in the same direction. In Melanesia there is reason to believe that the cross-cousin marriage stands in a definite relation to another form of marriage, that with the wife of the mother's brother. If there should be evidence for the former existence of this marriage in North America, it would increase the probability in favour of the cross-cousin marriage.

Among a number of peoples, some of whom form part of the Sioux, including the Minnitarees, Crows, Choctas, Creeks, Cherokees and Pawnees, cross-cousins are classed with parents and children exactly as in the Banks Islands, and exactly as in those islands, it is the son of the father's sister who is classed with the father, and the children of the mother's brother who are classed with sons or daughters. Further, among the Pawnees the wife of the mother's brother is classed with the wife, a feature also associated with the peculiar nomenclature for cross-cousins in the Banks Islands. The agreement is so close as to make it highly probable that the American features of relationship have been derived from a social institution of the same kind as that to which the Melanesian features are due, and that it was once the custom of these American peoples to marry the wife of the mother's brother. Here, as in the case of the cross-cousin marriage itself, the case rests entirely upon the terminology of relationship, but we cannot ignore the association in neighbouring parts of North America of features of relationship which would be the natural consequence of two forms of marriage which are known to be associated together elsewhere.

I am indebted to Miss Freire-Marreco for the information that the Tewa of Hano, a Pueblo tribe, call the father's sister's son tada, a term otherwise used for the father, thus suggesting that they also may once have practised marriage with the wife of the mother's brother. The use of this term, however, is only one example of a pracwhereby all the males of the father's clan are called tada, irrespective of age and generation. The common nomenclature for the father and the father's sister's son among the Tewa thus differs in character from the apparently similar nomenclature of the Banks Islands and cannot have been determined directly, perhaps not even remotely, by marriage with the wife of the mother's brother. This raises the question whether the nomenclature of the Sioux has not arisen out of a practice similar to that of the Tewa. The terms for other relarecorded by Morgan show some evidence of the widely generalised use of the Tewa, but such a use cannot account for the classing of the wife of the mother's brother with the wife which occurs among the Pawnees. Nevertheless, the Tewa practice should keethe Sioux nomenclature may depend on some social condition different from that which has been effective in the Banks Islands in spite of the close resemblance between the two.

The case for the former existence of the cross-cousin marriage should occur elesewhere in North America. So far as I am aware, the only people among whom it has been recorded are the Haidahs of Queen Charlotte Island.^[18] It is a far cry from this outpost of North American culture to Dakota, but it may be noted that it is among the Crees who formerly lived in the intermediate region of Manitoba and Assiniboia that the traces of the cross-cousin marriage are most definite. This mode of distribution of the peoples whose terminology of relationship bears evidence of the cross-cousin marriage suggests that other intermediate links may yet be found. Though the existing evidence is inconclusive, it should be sufficient to stimulate a search for other evidence which may make possible to decide whether or no the cross-cousin marriage was once a widespread practice in North America.

I can only consider one other kind of marriage here. The discovery of so remarkable a union as that with the daughter's daughter in Pentecost and the evidence pointing to a still more remarkable marriage between those having the status of grandparent and grandchild in Fiji and Buin have naturally led me to look for similar evidence elsewhere in Melanesia. Though there is nothing conclusive, conditions are to be found here and there which suggest the former existence of such marriages.

When I was in the Solomons I met a native of the Trobriand Islands, who told me that among his people the term tabu was applied both to grandparents and to the father's sister's child. I went into the whole subject as fully as was possible with only one witness, but in spite of his obvious intelligence and good faith, I remained doubtful whether the information was correct. The feature in question, however, occurs in the list of Trobriand terms drawn up for Dr. Seligmann^[19] by Mr. Bellamy, and with this double warrant it must be accepted. It is a feature which would follow from marriage with the daughter's daughter, for by this marriage one who was previously a father's sister's daughter becomes the wife of a grandfather and thereby attains the status of a grandparent. The feature exists alone, and, further, it is combined with other applications of the term which deprive it of some of its significance; nevertheless, the fact that a peculiar and exceptional feature of a Melanesian system of relationship is such as would follow naturally from a form of marriage which is practised in another part of Melanesia cannot be passed over. Standing alone, it would be wholly insufficient to justify the conclusion that the marriage with the daughter's daughter was ever prevalent among the Massim, but in place of expressing a dogmatic denial, let us look for other features of Massim sociology which may have been the results of such a marriage.

In Wagawaga^[20] there is a peculiar term, warihi, which is used by men for other men of their own generation and social group, but the term is also applied by an old man or woman to one of a younger generation. Again, in Tubetube^[21] the term for a husband, taubara, is also a term for an old man, and the term for the wife is also applied to an old woman. These usages may be nothing more than indications of respect for a husband or wife, or of some mechanism which brought those differing widely in age into one social category, but with the clue provided by the Trobriand term of relationship it becomes possible, though even now only possible, that the Wagawaga and Tubetube customs may have arisen out of a social condition in which it was customary to have great disparity of age between husbands and wives, and social relations between old and young following from such disparity in the age of consorts.

In Tubetube there is yet another piece of evidence. Mr. Field^[22] has recorded the existence in this island of three named categories of persons, two of which comprise relatives with whom marriage is prohibited, while the third groups together those with whom marriage is allowed. The grandparents and grandchildren are included in one of the two prohibited classes, so that we can be confident that marriage between these relatives does not now occur. The point to which I call your attention is that the class of relative with whom marriage is allowed is called kasoriegogoli. Li is the third person pronominal suffix, and we do not know the meaning of kasorie, but goga is the term used in Wagawaga and Wedau for the grandparents, its place being taken by the usual Melanesian term tubu in Tubetube. The term kasoriegogoli applied to marriageable relatives thus contains as one of its constituent elements a word which is probably the ancient term for grandparent in the island, since it is still used in this sense in the closely allied societies of the mainland.

We have thus a number of independent facts among the Massim, all of which would be the natural outcome of marriage between persons of alternate generations. To no one of them standing alone could much importance be attached, but taken in conjunction, they ought at least to suggest the possibility of such a marriage, a possibility which becomes the more probable when we consider that the Massim show clear evidence of the dual organisation of society with matrilineal descent which is associated with the granddaughter marriage of Pentecost and the Dieri of Australia. It adds to the weight of the evidence that indications of this peculiar form of marriage should be found among a people whose social organisation so closely resembles that in which the marriages between persons of alternate generations elsewhere occur.

I have no time for other examples. I hope to have shown that there are cases in which it is possible to infer with certainty the ancient existence of forms of marriage from the survival of their results in the terminology of relationship. In other cases, differences of culture or the absence of intermediate links make it unjustifiable to infer the ancient existence of the forms of marriage from which features of terminology might be derived. Other cases lie between the two, the confidence with which a form of marriage can be inferred varying with the degree of likeness of culture, the distance in space, and the presence or absence of other features of culture which may be related to the form of marriage in question. Even in the cases, however, where the inference is most doubtful, we have no right dogmatically to deny the origin of the terminology of relationship in social conditions, but should keep each example before an open mind, to guide and stimulate inquiry in a region where ethnologists have till now only scratched the surface covering a rich mine of knowledge.

1. The Melanesians, p. 38.
2. I leave out of account here those cases in which members of different generations are denoted by a reciprocal term.
3. Op. cit., p. 384.
4. Kinship Organisations and Group Marriage in Australia, Cambridge, 1906, p. 123.
5. This is the Mota name for Pentecost Island.
6. Howitt, Native Tribes of South-East Australia, pp. 164, 177.
7. Zeitsch. f. vergleich. Rechtsvaiss., 1910, xxiii., 330.
8. Rep. Cambridge Expedition to Torres Straits, vol. v., pp. 135 and 241.
9. I know of no complete record of the terminology of the fourth chief language of South India, Malayalam.
10. I take my data from the lists compiled for Morgan by the Rev. E. C. Scudder and the Rev. B. Rice, Morgan's Systems . . ., pp. 537-566. These lists are not complete, giving in some cases only the terms used in address. They agree in general with some lists compiled during the recent Indian Census which Mr. E. A. Gait has kindly sent to me.
11. Rivers, The Todas, 1906, pp. 487, 512.
12. Journal Royal Asiatic Society, 1907, p. 611.
13. See Morgan, Systems . . ., Table II.
14. Loc. cit.
15. Dakota Grammar, Texts, and Ethnography: Contributions to North American Ethnology, Washington, vol. ix.
16. Preface to above.
17. Op. cit., p. 82.
18. Swanton, Contributions to the Ethnology of the Haidahs, Jesup North Pacific Expedition, 1905, vol. v., pt. i., p. 62. Miss Freire-Marreco tells me that the cross-cousin marriage occurs among some of the Hopi Indians.
19. See The Melanesians of British New Guinea, Cambridge, 1910, p. 707.
20. Ibid., pp. 482 and 436.
21. The Melanesians of British New Guinea, Cambridge, 1910, p. 482.
22. Rep. Austral. Ass., 1900, viii., 301.