Kinship and social organisation/Lecture III

by W. H. R. Rivers

LECTURE III

Thus far in these lectures I have been content to demonstrate the dependence of the terminology of relationship upon forms of marriage. In spending so much time upon this aspect of my subject I fear that I may have been helping to strengthen a very general misconception, for it is frequently supposed that the sole aim of those who think as I do is to explain systems of relationship by their origin in forms of marriage. Marriage is only one of the social institutions which have moulded the terminology of relationship. It is, however, so fundamental a social institution that it is difficult to get far away from it in any argument which deals with social organisation. In now passing to other examples of the dependence of the terminology of relationship upon social conditions, I begin with one in which features of this terminology have come about, not as the result of forms of marriage, but of an attitude towards social regulations connected with marriage. The instance I have now to consider is closely allied to one which Professor Kroeber has used as his pattern of the psychological causation of the terminology of relationship. Both in Polynesia and Melanesia it is not infrequent for the father-in-law to be classed with the father, the mother-in-law with the mother, the brother-in-law with the brother, and the sister-in- law with the sister. The Oceanic terminology of relationship has two features which enable us to study the exact nature of this process in more detail than is possible with our own system. Oceanic languages often distinguish carefully between different kinds of brother- and sister-in-law, and, if it be found that it is only certain kinds of brother- or sister-in-law who are classed with the brother or sister, we may thereby obtain a clue to the nature of the process whereby the classing has come about. Secondly, Oceanic terminology usually distinguishes relationships between men or between women from those between persons of different sex, and there is a feature of the terminology employed when brothers- or sisters-in-law are classed with brothers or sisters in Oceania which throws much light on the process whereby this common nomenclature has come into use.

The first point to be noticed in the Oceanic nomenclature of relationship is that not all brothers- and sisters-in-law are classed with brothers and sisters, but only those of different sex. Thus, in Merlav, in the Banks Islands, it is only the wife's sister and a man's brother's wife who are classed with the sister, and the husband's brother and a woman's sister's husband who are classed with the brother, while there are special terms for other categories of relative whom we include under the designations brother- and sister- in-law. Similar conditions are general throughout Melanesia. If, as Professor Kroeber has supposed, the classing of the brother-in-law with the brother be due to the psychological similarity of the relationships, we ought to be able to discover why this similarity should be greater between persons of different sex than between persons of the same sex.

If now we study our case from the Banks Islands more closely and compare the social conditions in Merlav with those of other islands of the group, we find definite evidence, which it will not now be possible to consider in detail, showing that sexual relations were formerly allowed between a man and his wife's sisters and his brothers' wives, and that there is a definite association between the classing of these relatives with the sister and the cessation of such sexual relations. If such people as the Melanesians wish to emphasise in the strongest manner possible the impropriety of sexual relations between a man and the sisters of his wife, there is no way in which they can do it more effectually than by classing these relatives with a sister. To a Melanesian, as to other people of rude culture, the use of a term otherwise applied to a sister carries with it such deeply-seated associations as to put sexual relations absolutely out of the question. There is a large body of evidence from southern Melanesia which suggests strongly, if not conclusively, that the common nomenclature I am now considering has arisen out of the social need for emphasising the impropriety of relations which were once habitual among the people.

The second feature of Melanesian terminology which I have mentioned helps us to understand how the common nomenclature has come about. In most of the Melanesian cases in which a wife's sister is denoted by a term otherwise used for a sister, or a husband's brother by a term otherwise used for a brother, the term employed is one which is normally used between those of the same sex. Thus, a man does not apply to his wife's sister the term which he himself uses for his sister, but one which would be used by a woman of her sister. In other words, a man uses for his wife's sister the term which is used for this relative by his wife. This shows us how the common nomenclature may have come into use. It suggests that as sexual relations with the wife's sister became no longer orthodox, a man came to apply to this woman the word with which he was already familiar as a term for this relative from the mouth of his wife. The special feature of Melanesian nomenclature according to which terms of relationship vary with the sex of the speaker here helps us to understand how the common nomenclature arose. The process is one in which psychological factors evidently play an important part, but these psychological factors are themselves the outcome of a social process, viz., the change from a dition of sexual communism to one in which sexual relations are restricted to the partners of a marriage. Such psychological factors as come into action are only intermediate links in a chain of causation in which the two ends are definitely social processes or events, or, perhaps more correctly, psychological concomitants of intermediate links which are themselves social events. We should be shutting our eyes to obvious features of these Melanesian customs if we refused to recognise that the terminology of relationship here "reflects" sociology.

This leads me to question for a moment whether it may not be the same with that custom of our own society which Professor Kroeber has taken as his example of the psychological causation of the terminology of relationship. Is it as certain as Professor Kroeber supposes that the classing of the brother-in-law with the brother, or of the sister-in-law with the sister, among ourselves does not reflect sociology? We know that there are social factors at work among us which give to these relationships, and especially to that of wife's sister, a very great importance. If instead of stating dogmatically that this feature of our own terminology is due to the psychological similarity of the relationships, Professor Kroeber's mind had been open even to the possibility of the working of social causes, I think he might have been led to inquire more closely into the distribution and exact character of the practice in question. He might have been led to see that we have here a problem for exact inquiry. Such a custom among ourselves must certainly own a cause different from that to which I have ascribed the Melanesian practice, but is it certain that there is no social practice among ourselves which would lead to the classing of the wife's sister with the sister and the sister's husband of a woman with the brother? I will only point to the practice of marrying the deceased wife's sister, and content myself with the remark that I should be surprised if there were any general tendency to class these relatives together by a people among whom this form of marriage is the orthodox and habitual custom.

Till now I have been dealing with relatively small variations of the classificatory system. The varieties I have so far considered are such as would arise out of a common system if in one place there came into vogue the cross-cousin marriage, in another place marriage with the wife of the mother's brother, in another that with the granddaughter of the brother or with the wife of the grandfather, and in yet other places combinations of these forms of marriage. I have now to consider whether it is possible to refer the main varieties of the classificatory system to social conditions; as an example with which to begin, I choose one which is so definite that it attracted the attention of Morgan, viz., the variety of the classificatory system which Morgan called "Malayan." It is now generally recognised that this term was badly chosen. The variety so called was known to Morgan through the terminology of the Hawaiian Islands, and as the system of these islands was not only the first to be recorded, but is also that of which even now we have the most complete record, I propose to use it as the pattern and to speak of the Hawaiian system where Morgan spoke of the Malayan. If now we compare the Hawaiian system with the forms of the classificatory system found in other parts of Oceania, in Australia, India, Africa or America, we find that it is characterised by its extreme simplicity and by the fewness of its terms. Distinctions such as those between the father's brother and the mother's brother, between the father's sister and the mother's sister, and between the children of brothers or of sisters and the children of brother and sister, distinctions which are so generally present in the more usual forms of the classificatory system, are here completely absent. The problem before us is to discover whether the absence of these distinctions can be referred to any social factors. If not, we may be driven to suppose that there is something in the structure of the Polynesian mind which leads the Hawaiian and the Maori to see similarities where most other peoples of rude culture see differences.

The first point to be noted is that in Oceania the distinction between the Hawaiian and the more usual forms of the classificatory system does not correspond with the distinction between the Polynesian and Melanesian peoples. Systems are to be found in Melanesia, as in the western Solomons, which closely resemble that of Hawaii, while there are Polynesian systems, such as those of Tonga and Tikopia, which are so like those of Melanesia that, if they had occurred there, they would have attracted no special attention. The difference between the two kinds of system is not to be correlated with any difference of race.

Next, if we take Melanesian and Polynesian systems as a whole, we find that they do not fall into two sharply marked-off groups, but that there are any number of intermediate gradations between the two. It would be possible to arrange the classificatory systems of Oceania in a series in which it would not be possible to draw the line at any point between the different varieties of system which the two ends of the series seem to represent. The question arises whether it is possible to find any other series of transitions in Oceania which runs parallel with the series connecting the two varieties of system of relationship. There is no doubt but that this question can be answered in the affirmative.

Speaking broadly, there are two main varieties of social organisation in Oceania, with an infinite number of intermediate conditions. In one variety marriage is regulated by some kind of clan-exogamy, including under the term "clan" the moieties of a dual organisation; in the other variety marriage is regulated by kinship or genealogical relationship. We know of no part of Melanesia where marriage is regulated solely by clan-exogamy, but it is possible to arrange Melanesian and Polynesian societies in a series according to the different degrees in which the principles of genealogical relationship is the determining factor in the regulation of marriage. At one end of the series we should have places like the Banks Islands, the northern New Hebrides and the Santa Cruz Islands, where the clan-organisation is so obviously important that it was the only mechanism for the regulation of marriage which was recognised even by so skilful an observer as Dr. Codrington. At the other end of the series we have places such as the Hawaiian Islands and Eddystone Island in the western Solomons, where only the barest traces of a clan-organisation are to be found and where marriage is regulated solely by genealogical relationship. Between the two are numerous intermediate cases, and the series so formed runs so closely parallel to that representing the transitions between different forms of the classificatory system that it seems out of the question but that there should be a relation between the two. Of all the places where I have myself worked, the two in which I failed to find any trace of the regulation of marriage by means of a clan-organisation were the Hawaiian Islands and Eddystone Island, and the systems of both places were lacking in just those distinctions the absence of which characterised the Malayan system of Morgan. Only in one point did the Eddystone system differ from the Hawaiian. Though the mother's brother was classed in nomenclature with the father, there was a term for the sister's son, but it was so little used that in a superficial survey it would have escaped notice. Its use was so exceptional that many of the islanders were doubtful about its proper meaning. In other parts of the Solomons where the clan-organisation persists, but where the regulation of marriage by genealogical relationship is equally, if not more, important, the systems of relationship show intermediate characters. Thus, in the island of Florida the mother's brother was distinguished from the father and there was a term by means of which to distinguish cross-cousins from other kinds of cousin, but the father's sister was classed with the mother, and it was habitual to ignore the proper term for cross-cousins and to class them in nomenclature with brothers and sisters and with cousins of other kinds, as in the Hawaiian system. One influential man even applied the term for father to the mother's brother; it was evident that a change is even now in progress which would have to go very little farther to make the Florida system indistinguishable in structure from that of Hawaii.

Among the western Papuo-Melanesians of New Guinea, again, the systems of relationship come very near to the Hawaiian type, and with this character there is associated a very high degree of importance of the regulation of marriage by genealogical relationship and a vagueness of clanorganisation. We have here so close a parallelism between two series of social phenomena as to supply as good an example as could be wished of the application of the method of concomitant variations in the domain of sociology.

The nature of these changes and their relation to the general cultures of the peoples who use the different forms of terminology show that the transitions are to be associated with a progressive change which has taken place in Oceania. In this part of the world the classificatory system has been the seat of a process of simplification starting from the almost incredible complexity of Pentecost and reaching the simplicity of such systems as those of Eddystone or Mekeo. This process has gone hand in hand with one in which the regulation of marriage by some kind of clan-exogamy has gradually been replaced by a mechanism based on relationship as traced by means of pedigrees.

If this conclusion be accepted, it will follow that the more widely distributed varieties of the classificatory system of relationship are associated with a social structure which has the exogamous social group as its essential unit. This position has only to be stated for it to become apparent how all the main features of the classificatory system are such as would follow directly from such a social structure. Wherever the classificatory system is found in association with a system of exogamous social groups, the terms of relationship do not apply merely to relatives with whom it is possible to trace genealogical relationship, but to all the members of a clan of a given generation, even if no such relationship with them can be traced. Thus, a man will not only apply the term "father" to all the brothers of his father, to all the sons' sons of his father's father, and to all the sons' sons' sons of his father's father's father, to all the husbands of his mother's sisters and of his mother's mother's granddaughters, etc., but he will also apply the term to all the members of his father's clan of the same generation as his father and to all the husbands of the women of the mother's clan of the same generation as the mother, even when it is quite impossible to show any genealogical relationship with them. All these and the other main features of the classificatory system become at once natural and intelligible if this system had its origin in a social structure in which exogamous social groups, such as the clan or moiety, were even more completely and essentially the social units than we know them to be to-day among the peoples whose social systems have been carefully studied. If you are dissatisfied with the word "classificatory" as a term for the system of relationship which is found in America, Africa, India, Australia and Oceania, you would be perfectly safe in calling it the "clan" system, and in inferring the ancient presence of a social structure based on the exogamous clan even if this structure were no longer present.

Not only is the general character of the ficatory system exactly such as would be the consequence of its origin in a social structure founded on the exogamous social group, but many details of these systems point in the same direction. Thus, the rigorous distinctions between father's brother and mother's brother, and between father's sister and mother's sister, which are characteristic of the usual forms of the classificatory system, are the obvious consequence of the principle of exogamy. If this principle be in action, these relatives must always belong to different social groups, so that it would be natural to distinguish them in nomenclature.

Further, there are certain features of the classificatory system which suggest its origin in a special form of exogamous social grouping, viz., that usually known as the dual system in which there are only two social groups or moieties. It is an almost universal feature of the classificatory system that the children of brothers are classed with the children of sisters. A man applies the same term to his mother's sister's children which he uses for his father's brother's children, and the use of this term, being the same as that used for a brother or sister, carries with it the most rigorous prohibition of marriage. Such a condition would not follow necessarily from a social state in which there were more than two social groups. If the society were patrilineal, the children of two brothers would necessarily belong to the same social group, so that the principle of exogamy would prevent marriage between them, but if the women of the group had married into different clans, there is no reason arising out of the principle of exogamy which should prevent marriage between their children or lead to the use of a term common to them and the children of brothers. Similarly, if the society were matrilineal, the children of two sisters would necessarily belong to the same social group, but this would not be the case with the children of brothers who might marry into different social groups.

If, however, there be only two social groups, the case is very different. It would make no difference whether descent were patrilineal or matrilineal. In each case the children of two brothers or of two sisters must belong to the same moiety, while the children of brother and sister must belong to different moieties. The children of two brothers would be just as ineligible as consorts as the children of two sisters. Similarly, it would be a natural consequence of the dual organisation that the mother's brother's children should be classed with the father's sister's children, but this would not be necessary if there were more than two social groups.

I should have liked, if there were time, to deal with other features of the classificatory system, but must be content with these examples. I hope to have succeeded in showing that the social causation of the terminology of relationship goes far beyond the mere dependence of features of the system on special forms of marriage, and that the character of the classificatory system as a whole has been determined by its origin in a specific form of social organisation. I propose now to leave the classificatory system for a moment and inquire whether another system of denoting and classifying relationships may not similarly be shown to be determined by social conditions. The system I shall consider is our own. Let us examine this system in its relation to the form of social organisation prevalent among ourselves.

Just as among most peoples of rude culture the clan or other exogamous group is the essential unit of social organisation, so among ourselves this social unit is the family, using this term for the group consisting of a man, his wife, and their children. If we examine our terms of relationship, we find that those applied to individual persons and those used in a narrow and well-defined sense are just those in which the family is intimately concerned. The terms father, mother, husband and wife, brother and sister, are limited to members of the family of the speaker, and the terms father-, mother-, brother-, and sister-in-law to the members of the family of the wife or husband in the same narrowly restricted sense. Similarly, the terms grandfather and grandmother are limited to the parents of the father and mother, while the terms grandson and granddaughter are only used of the families of the children in the narrow sense. The terms uncle and aunt, nephew and niece, are used in a less restricted sense, but even these terms are only used of persons who stand in a close relation to the family of the speaker. We have only one term used with anything approaching the wide connotation of classificatory terms of relationship, and this term is used for a group of relatives who have as their chief feature in common that they are altogether outside the proper circle of the family and have no social obligations or privileges. They are as eligible for marriage as any other members of the community, and only in the very special cases I considered in the first lecture are they brought into any kind of legal relation. The dependence of our own use of terms of relationship on the social institution of the family seems to me so obvious that I find it difficult to understand how anyone who has considered these terms can put forward the view that the terminology of relationship is not socially conditioned. It seems to me that we have only to have the proposition stated that the classificatory system and our own are the outcome of the social institutions of the clan and family respectively for the social causation of such terminology to become conspicuous. I find it difficult to understand why it has not long before this been universally recognised. I do not think we can have a better example of the confusion and prejudice which have been allowed to envelop the subject through the unfortunate introduction of the problem of the primitive promiscuity or monogamy of mankind. It is not necessary to have an expert knowledge of the classificatory system. It is only necessary to consider the terms we have used almost from our cradles in relation to their social setting to see how the terminology of relationship has been determined by that setting.

This brief study of our own terms of relationship leads me to speak about the name by which our system is generally known. Morgan called it the "descriptive system," and this term has been generally adopted. I believe, however, that it is wholly inappropriate. Those terms which apply to one person and to one person only may be called descriptive if you please, though even here the use does not seem very happy. When we pass beyond these, however, our terms are no whit more descriptive than those of the classificatory system. We speak of a grandfather, not of a father's father or a mother's father, only distinguishing grandfathers in this manner when it is necessary to supplement our customary terminology by more exact description. Similarly, we speak of a brother-in-law, and only in exceptional circumstances do we use forms of language which indicate whether reference is being made to the brother of the husband or wife or to the husband of a sister. Such occasional usages do not make our system descriptive, and if they be held to do so, the classificatory system is just as descriptive as our own. All those peoples who use the classificatory system are capable of such exact description of relationship as I have mentioned. Indeed, classificatory systems are often more descriptive than our own. In some forms of this system true descriptive terms are found in habitual use. Thus, in the coastal systems of Fiji the mother's brother is of ten called ngandina (ngane, sister of a man, and tina, mother), this term being used in place of the vungo already mentioned. Similar uses of descriptive terms occur in other parts of Melanesia. Thus, in Santa Cruz the father's sister is called inwerderde (inwe, sister, and derde, father). This relative is one for whom Melanesian systems of relationship not infrequently possess no special designation, and the use of a descriptive term suggests a recent process which has come into action in order to denote a relative who had previously lacked any special designation.

If "descriptive" is thus an inappropriate name for our own system, it will be necessary to find another, and I should like boldly to recognise the direct dependence of its characters on the institution of the family and to speak of it as the "family system."

While I thus reject the term "descriptive" as a proper name for the terminology of relationship with which we are especially familiar, it does not follow that there may not be systems of denoting relationship which properly deserve this title. In Samoa a mode of denoting relatives is often used in which the great majority of the terms are descriptive. Thus, the only term which I could obtain for the father's brother's son was atalii o le uso o le tama, which is literally "son of the brother of the father," and there is some reason to suppose that this descriptive usage has come into vogue owing to the total inadequacy of the ancient Samoan system to express relationships in which the peoples are now interested.

The wide use of such descriptive terms is also found in many systems of Europe, as in the Celtic languages, in those of Scandinavia, in Lithuanian and Esthonian.^[1] A similar mode of denoting relationships is found in Semitic languages and among the Shilluks and Dinkas of the Anglo-Egyptian Sudan, and since it is from these peoples that I have gained my own experience of descriptive terminology, I propose to take them as my examples.

In the Arabic system of relationship used in Egypt many of the terms are descriptive; thus, the father's brother being called 'amm, the father's brother's wife is mirat 'ammi, the father's brother's son ibn 'ammi, and the father's brother's daughter bint 'ammi, and there is a similar usage for the consorts and children of the father's sister and of the brother and sister of the mother.

Similarly, many Shilluk terms suggest a descriptive character, the father's brother being wa, the wife of the father's brother is chiwa, the father's brother's son is uwa, and his daughter is nyuwa. The father's sister being waja, her son and daughter are uwaja and nyuwaja respectively. Similar descriptive terms are used by the Dinkas. The father's brother being walen, the father's brother's son is manwalen and his daughter yanwalen; the mother's brother being ninar, the mother's brother's son is manninar and his daughter yanninar.

According to the main thesis of these lectures, these descriptive usages should own some definite social cause. The descriptive terminology seems to be particularly definite in the case of cousins, and it might be suggested that they are dependent, at any rate in part and in so far as Egypt is concerned, on the prevalence of marriage with a cousin. Marriages with the daughter of a father's brother or of a mother's brother are especially orthodox and popular in Egypt, and different degrees of preference for marriage with different classes of cousin would produce just such a social need as would have led to the definite distinction of the different kinds of cousin from one another by means of descriptive terms.

It is more probable, however, that the use of descriptive terms in the languages of the Semites and of the Shilluks and Dinkas has been the outcome of a definite form of social organisation, viz., that in which the social unit is neither the family in the narrow sense, nor the clan, but that body of persons of common descent living in one house or in some other kind of close association which we call the patriarchal or extended family, the Gross-familie of the Germans. It is a feature of the Semitic and Nilotic systems, not only to distinguish the four chief categories of cousin, but also the four chief kinds of uncle or aunt, viz., the father's brother, the father's sister, the mother's brother and the mother's sister, all of whom are habitually classed together in our system, while some of them are classed with the father or mother in the classificatory system. The Semitic and Nilotic terminology is such as would follow from a form of social organisation in which the more intimate relationships of the family in the narrow sense are definitely recognised, but yet certain uncles, aunts, and cousins are of so much importance as to make it necessary for social purposes that they shall be denoted exactly. The brothers of the father and the unmarried sisters of the father would be of the same social group as the father, while the brothers and unmarried sisters of the mother would be of a different social group, which would account for their distinctive nomenclature, while within the social group it would be necessary to distinguish the father from his brothers. It would be too cumbrous to call this variety of system after the extended family, and I suggest that it should be called the "kindred" system.

Analogy with other parts of the world suggests that all those of the same generation in the social group formed by the extended family may once have been classed together under one term, and that, as later there arose social motives requiring the distinction of different relatives so classed together, descriptive terms came into use to make the necessary distinctions. You must please regard this only as a suggestion. We need far more detailed evidence concerning the social status of different relatives among the peoples who use these descriptive terms. Such knowledge as we possess seems to point to the dependence of the Semitic and Sudanese terminology upon the social institution of the extended family, just as our own system depends on the social institution of the family in the narrow sense and the classificatory system upon the clan.

If this descriptive mode of nomenclature be thus the outcome of a social organisation of which the essential element is the extended family, I need hardly point out how natural it is that we should find this kind of nomenclature so widely in Europe. The presence of this descriptive terminology in Celtic and Scandinavian languages, in Lithuanian and Esthonian, would be examples of the persistence of a form of nomenclature which had its origin in the kindred of the extended family. On this view we must believe that, in other languages of Europe, this mode of nomenclature has gradually been replaced by one dependent on the social institution of the family in the narrow sense.

At this point I should like to sum up briefly the position to which our argument has taken us. I have first shown the dependence of a number of special features of the classificatory system of relationship upon special forms of marriage. Then I have shown that certain broad varieties of the classificatory system are to be referred to different forms of social organisation and to the different degrees in which the regulation of marriage by means of clan-exogamy has been replaced by a mechanism dependent upon kinship or genealogical relationship. From that I was led to refer the general features of the classificatory system to the dependence of this system upon the social unit of the clan as opposed to the family which I believe to be the basis of our own terminology of relationship. I then pointed to several features of the classificatory system which suggest that it arose in that special variety of the clan-organisation in which a community consists of two exogamous moieties, forming the social structure usually known as the dual organisation. I considered more fully the dependence of our own mode of denoting relatives upon the social institution of the family, and then a study of the descriptive terminology of relationship has led me to suggest that certain modes of denoting relationship in Egypt, the Sudan and many European countries may be examples of a third main variety of system of relationship which has arisen out of the patriarchal or extended family. We should thus have three main varieties of system of relationship in place of the two which have hitherto been recognised, having their origins respectively in the clan, in the family in the narrow sense, and in the extended or patriarchal family. These three varieties may be regarded as genera within each of which are species and varieties depending upon special social conditions which have arisen within each kind of social grouping, either as the result of changes within each form of social organisation or of transitions from one form to another. We know of a far larger number of such varieties within the classificatory system than within those due to the two forms of the family, and this is probably due in some measure to the fact that the classificatory system is still by far the most widely distributed form over the earth's surface. Still more important, however, is the fact that among the peoples who use the classificatory system there is an infinitely greater variety of social institution, and especially of forms of marriage, than exist among civilised peoples whose main social unit, the family, is not one which is capable of any extended range of variation. The result of the complete survey has been to justify my use of the classificatory system as the means whereby to demonstrate the dependence of the terminology of relationship upon social conditions. It is the great variability of this mode of denoting relatives which makes it so valuable an instrument for the study of the laws which have governed the history of that department of language by which mankind has denoted those who stand in social relations to himself.

You may have been wondering whether I am going to say anything about the merits of the controversy which has till now given to systems of relationship their chief interest among students of sociology. I have so far left on one side the subjects which have been the main ground of controversy ever since the time of Morgan. You will have gathered that I regard it as a grave misfortune for the science of sociology that the topics of promiscuity and group-marriage should have been thrust by Morgan into the prominent place which they have ever since occupied in the theoretical study of relationship. Even now I should have liked to leave them on one side on the ground that the evidence is as yet insufficient to make them profitable subjects for such exact inquiry as I believe to be the proper business of sociology. Their very prominence, however, makes it impossible to leave them wholly unconsidered, but I propose to deal with them very briefly.

I begin with the question whether the classificatory system of relationship provides us with any evidence that mankind once possessed a form of social organisation, or rather such an absence of social organisation, as would accompany a condition of general promiscuity in which, if one can speak of marriage at all, marriage was practised between all and any members of the community, including brothers and sisters. I can deal with this subject very briefly because I hope to have succeeded elsewhere in knocking away the support on which the whole of Morgan's own construction rested.

Morgan deduced his stage of promiscuity from the Hawaiian system, which he supposed to be the most primitive form of classificatory nomenclature. In an article published in 1907 I showed^[2] that it rather represents a late stage in the history of the more ordinary forms of the classificatory system. My conclusion at that time was based on the scanty evidence derived from the relatively few Oceanic systems which had then been recorded, but my work since that article was written has shown the absolute correctness of my earlier opinion, which I can now support by a far larger body of evidence than was available in 1907. It remains possible, however, that the Hawaiian system may have had its source in promiscuity, even though this condition be late rather than primitive, but it would be going beyond the scope of these lectures to deal fully with this subject here. I cannot forbear, however, from mentioning that Hawaiian promiscuity, in so far as it existed, was not the condition of the whole people, but only of the chiefs who alone were allowed to contract brother and sister marriages, while I have evidence that the avoidance of brother and sister in Melanesia, which has so often been regarded as a survival of man's early promiscuity, is capable of a very different explanation.^[3] Our available knowledge, whether derived from features of the classificatory system or from other social facts, does not provide one shred of evidence in favour of such a condition as was put forward by Morgan as the earliest stage of human society, nor is there any evidence that such cuity has ever been the ruling principle of a people at any later stage of the history of mankind.

The subject of group-marriage is one about which I do not find it possible to speak so dogmatically. It would take me more than another lecture to deal adequately with the Melanesian evidence alone, and I must content myself with two remarks. Firstly, I think it desirable to throw aside the term group-marriage as only confusing the issue, and to speak rather of a state of organised sexual communism, in which sexual relations are recognised as orthodox between the men of one social group and the women of another. Secondly, the classificatory system has several features which would follow naturally from such a condition of sexual communism. I have evidence from Melanesia which places beyond question the former presence of such a condition, with features of culture which become readily explicable if they be the survivals of such a state of sexual communism as is suggested by the terminology of the classificatory system. This evidence comes from only one part of the world, but it is enough to convince me that we have no right to dismiss from our minds a state of organised sexual communism as a feature of the social development of mankind. The wide distribution of the classificatory system would suggest that this communism has been very general, but it need not have been universal, and even if the widespread existence of organised sexual communism be established, it would not follow that it represents the earliest stage in the evolution of human society. There are certain features even of the classificatory system itself which suggest that, if this system be founded in sexual communism, this communism was not primitive, but grew out of a condition in which only such ties of kinship were recognised as would result from the social institution of the family.

I must be content with this brief reference to the subject. The object of these lectures is to demonstrate the dependence of the terminology of relationship upon social conditions, and the dependence of the classificatory system upon a condition of sexual communism is not now capable of demonstration. The classificatory mode of denoting relationship should, however, act as a suggestion and stimulus, and as a preventative of dogmatic statement in a part of our subject which, in spite of its entrancing interest, still lies only at the edge of our slowly spreading circle of exact knowledge.

In conclusion, I should like to point out briefly some of the lessons of more general interest which may be learnt from the facts I have brought before you in these lectures. I hope that one result has been to convince you of the danger lying in the use of the reductio ad absurdum argument when dealing with cultures widely different from our own. In the literature of the subject one often meets the adjectives "absurd" and "impossible" applied in some cases to social conditions in which the actual existence of the absurdities or bilities can be demonstrated. I may take as an example the argument of Mr. N. W. Thomas, which I have already mentioned, in which the classing of the maternal grandfather with the elder brother by the Dieri is regarded as reducing to an absurdity the contention that classificatory terms express ties of kinship. If Mr. Thomas had had a more lively faith in the social meaning of terms of relationship, he might have been led to notice that the Dieri marry the granddaughter of a brother, a fact he appears, in common with many other readers of Howitt, to have missed; one result of this marriage is to bring about just such a relationship as Howitt records without a man being his own great-uncle, as is supposed to be necessary by Mr. Thomas.

Still another example may be taken from Professor Kroeber. He states that the classing together of the grandfather and the father-in-law which is found in the Dakota system, when worked out to its implications, would lead to the absurd conclusion that marriage with the mother was once customary among the Sioux. Here again, if Professor Kroeber had been less imbued with his belief in a purely linguistic and psychological chain of causation, and had been ready to entertain the idea that there might be a social meaning, he must have been led to see that the features of nomenclature in question would follow from other forms of marriage, and two of these, whatever their apparent improbability in America, cannot well be called absurd, since they are known to occur in other parts of the world. Following Riggs, Professor Kroeber does not specify which kinds of grandfather and father-in-law are classed together in Dakotan nomenclature, but in the full list given by Morgan, it is evident that one term is used for the fathers of both father and mother and for the fathers of both husband and wife. The classing of the father's father with the wife's father would be a natural result of marriage with the father's sister, while the common nomenclature for father's father and husband's father would result from marriage with the brother's daughter. It is not without significance that the features of nomenclature which would be the result of one or other, or of both these marriages, occur in a system which also bears evidence of the cross-cousin marriage, for these three forms of marriage occur in conjunction in one part of Melanesia, viz., the Torres Islands.

The foregoing instance, together with many others scattered through these lectures, will have pointed clearly to another lesson. In the present state of our knowledge a working scheme or hypothesis has largely to be judged by its utility. A way of regarding social phenomena which obstructs inquiry and leads people to overlook facts has its disadvantages, to say the least, while a scheme or hypothesis which leads people to worry out and discover things which do not lie on the surface will establish a strong claim on our consideration, even if it should ultimately turn out to be only the partial truth. I will give only one instance to illustrate how a belief in the dependence of the terminology of relationship on forms of marriage might act as a stimulus to research.

In a system from the United Provinces recorded by Mr. E. A. H. Blunt in the Report of the last Indian Census, one term, bahu, is used for the son's wife, for the wife, and for the mother.^[4] Mr. Blunt puts on one side without hesitation the possibility that such common nomenclature can have been the result of any form of marriage, and ascribes it to the custom whereby a man and his wife live with the husband's parents, in consequence of which the son's wife, who is called bahu by her husband, is also called bahu by everyone else in the house. The causation of the common nomenclature which is thus put forward is a possible, perhaps even a probable, explanation. In such a case we should have a social chain of causation in which the son's wife is called bahu because she is one of a social group bound together by the ties of a common habitation. It can do no harm, however, to bear in mind as an alternative the possibility that the terminology may have arisen out of a form of marriage. It is evident that the use of a common term for the wife and the son's wife would follow from a form of polyandry in which a man and his son have a wife in common. A further result of this form of marriage would be that the wife of the son, being also the wife of his father, would have the status of a mother.^[5] We have no evidence for the presence of such a marriage in India, but our knowledge of the sociology of the more backward peoples of India is not so complete that we can afford to neglect any clue. The possibility suggested by the mode of using the term bahu should lead us to look for other evidence of such a form of polyandry among the ruder elements of the population of India, of whose social structure our present knowledge is so fragmentary.

Another important result of our study of the terminology of relationship is that it helps us to understand the proper place of psychological explanation in sociology. These lectures have largely been devoted to the demonstration of the failure to explain features of the terminology of relationship on psychological grounds. If this demonstration has been successful, it is not because the terminology of relationship is anything peculiar, differing from other bodies of sociological facts; it is because in relationship we have to do with definite and clean-cut facts. The terminology of relationship is only a specially favourable example by means of which to show the value of an attitude towards, and mode of treatment of, social facts which hold good, though less conspicuously, throughout the whole field of sociology. In social, as in all other kinds of human activity, psychological factors must have an essential part. I have myself in these lectures pointed to psychological considerations as elements in the problems with which the sociologist has to deal. These psychological elements are, however, only concomitants of social processes with which it is possible to deal apart from their psychological aspect. It has been the task of these lectures to refer the social facts of relationship to antecedent social conditions, and I believe that this is the proper method of sociology. Even at the present time, however, it is possible to support sociological arguments by means of considerations provided by psychological motives, and the assistance thus rendered to sociology will become far greater as the science of social psychology advances.

This is, however, a process very different from the interpolation of psychological facts as links in the chain of causation connecting social antecedents with social consequences. It is in no spirit of hostility to social psychology, but in the hope that it may help us to understand its proper place in the study of social institutions that I venture to put forward the method followed in these lectures as one proper to the science of sociology.^[6]

It may be that there will be those who will accept my main position, but will urge that these lectures have been devoted to the criticism of an extreme position, the position taken up by Professor Kroeber. They may say that they have never believed in the purely psychological causation of the terminology of relationship. In reply to such an attitude I can only express my conviction that the paper of Professor Kroeber is only the explicit and clear statement of an attitude which is implicit in the work of nearly all, if not all, the opponents of Morgan since McLennan. Whether they have themselves recognised it or not, I believe that it has been this underlying attitude towards sociological problems which has prevented them from seeing what is good in Morgan's work, from sifting out the chaff from the wheat of his argument, and from recognising how great is the importance to the science of sociology of the body of facts which Morgan was the first to collect and study. I feel that we owe a debt of gratitude to Professor Kroeber for having brought the matter into the open and for having presented, as a clear issue, a fundamental problem of the methods of sociology.

Lastly, I should like to point out how rigorous and exact has been the process of the determination of the nomenclature of relationship by social conditions which has been demonstrated in these lectures. We have here a case in which the principle of determinism applies with a rigour and definiteness equal to that of any of the exact sciences. According to my scheme, not only has the general character of systems of relationship been strictly determined by social conditions, but every detail of these systems has also been so determined. Even so small and apparently insignificant a feature as the classing of the sister-in-law with the sister has been found to lead back to a definite social condition arising out of the regulation of marriage and of sexual relations. If sociology is to become a science fit to rank with other sciences, it must, like them, be rigorously deterministic. Social phenomena do not come into being of themselves. The proposition that we class two relatives together in nomenclature because the relationships are similar is, if it stand alone, nothing more than a form of words. It is incumbent on those who believe in the importance of the psychological similarity of social phenomena to show in what the supposed similarity consists and how it has come about—in other words, how it has been determined. It has been my chief object in these lectures to show that, in so far as such similarities exist in the case of relationship, they have been determined by social conditions. Only by attention to this aim 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.

1. See Tables in Morgan's Systems..., pp. 79-127.
2. Anthropological Essays presented to E. B. Tylor, Oxford, 1907, p. 309.
3. For the full evidence on these topics see my forthcoming book The History of Melanesian Society.
4. Census of India, 1911, vol. xv., p. 234.
5. In such a case the use of the term by other members of the household, including women, would be the result of a later extension of meaning.
6. See also "Survival in Sociology" Sociological Review, 1913. vol. vi., p. 293. I hope shortly to deal more fully with the relations between sociology and social psychology.