CHAPTER X.

Thus, to take a concrete example, there is not much danger that a writer who discusses the question of marriage in civilised communities will deal with one form of union of the sexes, while his readers may imagine that he is dealing with another form. For marriage is the form of sexual union recognised by the law of the land, and its legal sanction distinguishes it from all other forms of sexual union, however permanent they may be, and however short may be the period before the marriage is dissolved by an appeal to the courts of law. In fact in civilised communities the fulfilment of legal forms and ceremonies constitutes marriage, whatever might be said of a union sanctioned by legal forms but unaccompanied by the cohabitation of the parties. When, however, we are dealing with a people ruled by custom and not by law, the case is far different. The force of custom may and usually does in such cases far exceed the force of law in civilised communities. In the lower stages of culture there is far more reluctance to overstep the traditional lines of behaviour than is felt by the ordinary member of a European state, and this though there are penalties in the latter which do not necessarily exist in the former case. But law, in the sense of a rule of conduct, promulgated by a legislator and enforced by penalties inflicted by law courts and carried out by the agents of the state, does not necessarily exist, and, at most, exists only in a very inchoate state. If therefore we read of marriage among such a people, we are left in complete uncertainty whether it is a union corresponding to marriage in civilised lands, or whether it belongs to a different category. The difficulty of the case lies partly in the inability of the observer to distinguish de jure from de facto unions, partly in the fact that one may be transformed into the other, and no ceremony of any sort mark the change. An Australian may, for example, have a wife who is recognised as his by tribal custom and tradition; if she is abducted the aggrieved husband may vindicate his rights but will not necessarily be supported by even his own kin, and will certainly not find anything to correspond to the tribunal before which an Englishman would sue for the restitution of conjugal rights. If the aggrieved husband proves the weaker, he necessarily abandons his wife, and she becomes ipso facto the wife of the aggressor; divorce is in fact pronounced by the issue of an ordeal by combat. So far the matter is clear to the observer.

But if the aggrieved husband take no steps to vindicate his rights, the woman will equally pass to the aggressor, and in this case there will be no customary ceremonial to mark for the benefit of the observer the exact moment of the transition from a marriage, recognised by public opinion, or tribal custom, with the first husband A to the same kind of union with B.

Again, even where no second mate intervenes to complicate the question, the observer may be confronted with delicate problems; at what point, for example, does a mere liaison pass into something worthy of the name of marriage? What is the status of a union in which the parties are more or less permanently associated, but which confers no rights as against aggressors? If by native custom the union is not of such a nature as to confer on the male party to it any rights over the female, such as the liberty to chastise or punish without fear of the intervention of the woman's kin, are we to regard the tie as equivalent to marriage if only it is permanent? At what point does mere cohabitation pass into marriage?

All these are questions which have to be debated and decided before we are in possession of a suitable terminology for dealing with the unions of the sexes in the lower stages of culture. But they are commonly neglected in controversies as to the origin and history of human marriage.

We have seen above that in a European community we mean by marriage a union between two persons of opposite sexes, entered into with due legal formalities, and not dissoluble simply at the will of either or both the parties concerned. When we go further afield the connotation of the term is extended to embrace (1) polygyny, in which one male is associated with two or more females, (2) polyandry, in which one female is similarly associated with more than one male, and (3) the condition which I propose to term polygamy, in which both these conditions are found. In all these cases the union is properly termed marriage, in so far as it cannot be entered upon without due formalities nor be dissolved without the concurrence of the authority upon the carrying out of whose conditions in the preliminary steps the union depends for its marriage-character.

When however we come to the so-called group marriage, using the term in its original sense of limited promiscuity, we are dealing with an entirely different state of things, and it is difficult to see any justification for the use of the term marriage in this connection at all. By group marriage is meant a condition only removed from absolute promiscuity by the existence of age-classes or of two or more exogamous classes in the community; it demands no special ceremonies prior to the individual union141, it permits this union to be dissolved at will, and it consequently confers no rights on either of the parties to it, other than perhaps the right to the produce, or some of the produce, of each other's labour.

If the confusion did not extend beyond the terminology, the advance of knowledge would perhaps be but little impeded; but experience shows that confusion in terminology is apt to go hand in hand with confusion in ideas. As will be shown later, this seems to be particularly true of investigations into the history of marriage and sexual relationships. It seems desirable therefore to clear the way by classifying the ideas with which we have to deal, and by defining the terms corresponding to them.

Before classifying the various forms of sexual relationships, it may be well to say a few words on the definition of marriage in general. Dr Westermarck has defined it from the point of view of natural history as a more or less durable connection between male and female, lasting beyond the mere act of propagation till after the birth of the offspring.

It may not be possible to propose a better definition from the point of view selected by Dr Westermarck, which is certainly the one from which anthropology must regard sexual relationships. At the same time it is not entirely free from objection. In the first place we are employing the word marriage in a sense which has but little in common with its ordinarily accepted meaning. Suppose, for example, we are dealing with marriage in Europe, it is confusing to be compelled by our definition to regard as a marriage the faux ménage, not to speak of the not uncommon fairly permanent unions in which there is no common residence. Such monogamous relationships may be, technically speaking, marriages, in Dr Westermarck's sense, but it seems desirable to make use of some other term for them and reserve marriage for the unions sanctioned by legal forms. Or take the union of two people, each of whom has prior matrimonial engagements. Such a union may, as the records of the divorce court show, be anything but impermanent; but it does not make for clearness to call such an union marriage. Let us take a third example—a New Hebridean girl purchased, or in Upa stolen, for the use of the young men, who, of course, reside in their club-house. If any of the bachelors there resident chooses to recognise her children, they are regarded as his children; if not, they are supported by the whole of the residents in the club-house. How are we to classify the position of the mother of these children? The union is obviously fairly permanent, although some of the group enter into sexual relationships of an ordinary type and join the ranks of the married men, and others enter the club-house from the ranks of those hitherto shut out from the enjoyment of the privileges of the adult unmarried male. But the relationship established with the whole body of unmarried men and indistinguishable, so far as definition goes, from polyandry, hardly seems to be a permanent union of the type which Dr Westermarck had in mind when he framed his definition, much less a marriage in any accepted sense of the term.

For Dr Westermarck's general term marriage it would be well to substitute gamé or gamic union, to express all kinds of sexual relationships other than temporary ones. As sub-heads under this we have:

(1) Marriage, a union recognised by law or custom, which imposes duties and confers rights on one, both, or all the parties to it.

(2) Free union, a relationship not recognised by the community as conferring rights, but at the same time not punished and not necessarily regarded as immoral. Temporary unions we may classify as (a) promiscuity where marriage does not exist or is temporarily in abeyance: (b) free love, the relationships of the unmarried: (c i.) temporary polyandry or polygyny of married people, where the unions are limited and recognised by custom: (c ii.) marital licence where the husband is complaisant in the face of public opinion: (c iii.) adultery where neither the husband nor public opinion permits them.

(3) Liaison, a union in which one or both parties have other ties, which renders them liable to punishment, or to some kind of atonement.

Ten various possible forms of sexual relationship actually found or assumed to have existed may now be classified.

A. PROMISCUITY.

I. Unregulated Promiscuity. (a) Primary unregulated promiscuity is the hypothetical state assumed by Morgan and others to be the primitive state of mankind. It may be noted that promiscuity de jure, which is all that is implied by Morgan's hypothesis, is not necessarily also de facto promiscuity. Unless it be assumed that jealousy was absent at this stage, it is clear that free unions must have been the rule rather than the exception. But if this be so, the only distinction between Morgan's promiscuity and the ordinary state of things in an Australian tribe is constituted, intermarrying rules apart, by the fact that the Australian husband is at liberty to reclaim his wife, if he can, without fear of blood feud if perchance he slays his successor in the affections, or perhaps rather in the possession, of his wife, whereas in Morgan's primitive stage might was right and the abductor was on an equal footing with his predecessor and successor. (b) Secondary unregulated promiscuity is distinguished from primary promiscuity by the co-existence of other forms of sexual relations. It may temporarily supersede these as in Australia; or it may take their place, as among the Nairs.

II. Regulated Promiscuity. This again falls into (a) primary regulated promiscuity, the hypothetical stage postulated for Australia before the introduction of individual marriage; and (b) secondary regulated promiscuity, which is found in certain tribes as an exceptional practice. With this custom I deal in greater detail below.

B. MARRIAGE.

III, Polygamy. This state is constituted by the union of several men with several women. It may be distinguished, as before, into primary and secondary polygamy. We may further distinguish (α) simple and (β) adelphic polygamy; and the latter may be (i) unilateral or (ii) bilateral, according as either the males or females, or both males and females, are brothers and sisters. A further sub-division is constituted by the relations of the groups of males or females, or both, within themselves. I distinguish these unions by the names of dissimilar (M.) and dissimilar (F.) according as one husband or one wife has a position superior to the others142.

IV, Polyandric and V. polygynic unions fall into the same divisions, save that they are naturally always unilateral. As a designation for the hypothetical stage postulated by Mr Atkinson in Primal Law, we may take "patriarchal polygyny," meaning thereby the state in which (a) in the earlier stage all the females of the horde143 are ipso facto mates of the one adult male of the horde; or (b) in the second stage all females born in the horde are equally allotted to him.

Finally we have VI, monogamy.

To the three forms of marriage we can apply the determinants "regulated" and "unregulated," "temporary144," "permanent," as in the case of promiscuity.

We have further two well-marked types of marriage and a mixed form in which (a) the husband goes to live with the wife; (b) he lives with the wife for a time and then removes to his own village or tribe; and (c) the wife removes to the husband. For the first of these Maclennan has proposed the name beena marriage; Robertson Smith has proposed to call the third type ba‘al marriage, and to include both beena and mot‘a marriages under the general name of ṣadīca. This terminology is unnecessarily obscure and has the further disadvantage of connoting the domination or subjection of the husband, a feature not necessarily bound up with residence. I therefore propose to term the three types matrilocal, removal, and patrilocal marriages. I suggest compounds of pater and mater, not as being specially appropriate, but as being parallel to matrilineal and patrilineal, denoting descent in the female and male lines respectively.

For the somewhat complicated relationships of potestas in the family I propose two main divisions, (a) patri-potestal, (b) matri-potestal; the latter may be further subdivided according as the authority is in the hands (1) of the actual mother, (2) of the maternal uncles, (3) of the mother's relatives in general, and so on.

CHAPTER XI.

IV. The Hawaian Custom of Punalua145, giving the Malayan Form of the Classificatory System146.

V. The Tribal Organisation, i.e. totemic exogamy plus promiscuity, giving the Turanian and Ganowanian System147.

VI. Monogamy.

The objections to this theory or group of theories are numerous, and it will not be necessary to consider them all here. Were it not that no one has since Morgan's day attempted to trace in detail the course of evolution from promiscuity to monogamy, it would be almost superfluous to discuss the theories of a work on primitive sociology dating back nearly thirty years.

With some points Morgan has failed to deal in a way that commends itself to us in the light of knowledge accumulated since his day; with others he has not attempted to deal, apparently from a want of perception of their importance.

First and foremost among the points with which Morgan has failed to deal is that of the constitution of the primitive group. Was it composed of parents and children only or were more than two generations represented? If the former, why were the children expelled? if the latter, how are brother and sister marriages introduced, when ex hypothesi the father of any given child was unknown and may have been any adult male? If Morgan and his supporters evade this difficulty by defining brother and sister as children of the same mother, they are met by the obvious objection that no revolution in a promiscuous group would result in the marriage of children of the same mother. Ex hypothesi there were several child-bearing women in the group, and their children, if a reform were introduced prohibiting marriage outside one's own generation, would intermarry; but the children of these women are, on the definition adopted, not brothers and sisters.

If brother and sister does not mean children of the same mother, what does it mean?

By what process are these names supposed to have come into existence in a promiscuous group? If brother in this sense is taken to imply common parentage, the name must clearly denote the relation between two males because, although a whole group of men had access to the mother, the male parent was or may have been the same person in each case, and this whether the mother was the same or not. Now, quite apart from the fact that primitive man was unlikely to have evolved a term for such an indefinite relationship, except in so far as it involved rights or duties, it is obvious that great complications would arise which would in practice make the nomenclature unworkable. For to call two boys brothers because they have the same group of men as possible fathers is only practicable in a society which has already evolved a system of age grades, and has established restrictions on intercourse between different generations, to use a somewhat indefinite term. For it is clear that in a state of promiscuity the class of adults is continually being recruited and that the boy passes at puberty, in so far as restrictions in the nature of initiation ceremonies are not imposed, from the class of sons to that of fathers. In other words, if a group consists of M1 M2 M3 M4, and they have male children of all ages N1 N2 N3 N4, as soon as N1 reaches puberty he becomes a possible father of the children O1 O2 O3 O4, who differ in age from N4 only by a few years at most and reckon as his brothers. But this means that N1 is the son of M1, for example, but at the same time the father of O1, who is likewise the son of M1; in the same way O1 is the brother of N4, who is the brother of N1; but O1 is not the brother of N1. The extraordinary complexity of the relations that would arise is at once obvious, and it seems clear that relationship terms could never come into existence under such circumstances unless they implied something beyond mere relationship and denoted rights and duties148. But if they denoted rights and duties, these must have preceded the relationship term, which consequently need not be held to apply to kinship in any proper sense of the term.

It is clear that the same difficulties apply when we try to work out the development on the hypothesis that a group of mothers existed. We are therefore reduced to the supposition that the term brother denoted originally a person born within a given period of time, and that this period was the same for whole sections of the community; in other words that the name brother was given to all males born between, let us say, B.C. 10,000 and B.C. 9,990. This is of course equivalent to the establishment of age grades and is in itself not unthinkable; age grades are of course perfectly well known among primitive peoples; but the establishment of age grades implies a degree of social organisation; and, what is more important, this hypothesis makes the term brother quite meaningless as a kinship term; for at the present day a common term of address for members of an age grade does not imply any degree of consanguinity, and unless it be proved that age grades are a product of the period of "group marriage" it cannot be argued that they ever did imply kinship.

It is sufficiently clear from these examples that Morgan entirely failed to work out the process by which the transition from pure to regulated promiscuity came about. But if the process is uncertain the causes are equally obscure. In Mr Morgan's view, or at any rate in one of the theories on which he accounted for the change, it was due to "movements which resulted in unconscious reformation"; these movements were, he supposes, worked out by natural selection. These words, it is true, apply primarily to the origin of the "tribal" or "gentile" organisation, as Mr Morgan terms totemism, but they probably apply to the original passage from promiscuity to "communal marriage," and I propose to examine how far such a theory has any solid basis.

Natural selection is a blessed phrase, but in the present case it is difficult to see in what way it is supposed to act. The variation postulated by Mr Morgan as a basis for the operation of natural selection is one of ideas, not physical or mental powers. Now under ordinary circumstances we mean by natural selection the weeding out of the unfit by reason of inferiorities, physical or psychical, which handicap them in the struggle for existence. But it cannot be said that the tendency to marry or practice of marrying outside one's own generation is such a handicap to the parents. How far is it injurious to the children of such unions? Or rather, how far have children who are the offspring of brothers and sisters or of cousins a better chance of surviving than the offspring of unions between relatives of different generations?

It is at the outset clear that savages are not in the habit of taking account of such matters. Even if it were otherwise, it is not clear how far they would have data as to the varying results of unions of near kin. For though on this question, so far as the genus homo is concerned, we have very few data on which to go, such data as we have hardly bear out his view. Modern statistics relate almost exclusively to the intermarriage of cousins, and apply, not to primitive tribes, such as those with which, ex hypothesi, Mr Morgan is dealing, but to more or less civilised and sophisticated peoples, among whom the struggle for existence is less keen owing to the advance of knowledge and the progress of invention, and among whom possibly the rise of humanitarian ideas not only tends to counteract the weeding out of the unfit, but even makes it relatively easy for them to propagate their species. What the result of the intermarriage of cousins is when war, famine, and infanticide are efficient weeders out of the unfit, we cannot say. Possibly or even probably the ill results would be inappreciable. It must not be forgotten that the marriage of near relatives is only harmful because or if it hands on to the children of the union an hereditary taint in a strengthened form, a result which is likely to follow in civilised life because hereditary taints are allowed to flourish unchecked by prudence and controlled by natural selection only so far as humanitarianism will permit it. These hereditary degeneracies however are probably largely if not entirely absent among savages. It is therefore open to question how far intermarriage of cousins would prove harmful under such conditions.

Statistics of the influence of cousin-marriage are not however what Mr Morgan wants. It is essential for him to prove that father-daughter marriage is more harmful than brother-sister unions.

It might be imagined that the data for estimating the effect of the union of father and daughter would be non-existent, but this is not so. Within the last few years it has been stated that such unions are common in parts of South America, and that the children, so far from being degenerates, are remarkably healthy and vigorous149. This is of interest in connection with Mr Atkinson's speculations as to the history of the family. In this connection it may be pointed out that such unions, ex hypothesi, are unlikely to result in continual in-and-in breeding, and would in all probability seldom be continued beyond the first alliance of this nature.

We are practically in complete darkness as to the results of brother and sister marriage in the human species. We have of course various cases of ruling families who perpetuated themselves in this way, but the data from such peoples refer to an advanced stage of culture and to a favoured class. They are not therefore applicable to similar unions among savages where they formed, as Mr Morgan suggests, the invariable practice. It is however possible to deduce from very simple considerations the probabilities as to the respective effects of adelphic and father-daughter unions. In the first place, as has been already pointed out, the father-daughter union implies only one family of in-and-in-bred children; in the case of brother and sister marriage, on the other hand, this state of things may go on indefinitely. If this is not enough to turn the scale against adelphic unions there is the further fact that, taking the descendants of the first pair of intermarrying descendants of common parents, whose tendency to disease or deformity is we will suppose x1 on both sides, and assuming that this tendency increases in a simple ratio, the offspring have the same tendencies to the second power of x. If their children marry each other the measure of degeneracy in the third generation is x4. Suppose now a father and mother with index of degeneracy each x1; a daughter of this union will have as her index x2; if the daughter bears children to the father, their index will be not x4, but x3, if the simple law which I have assumed for the purposes of argument holds good.

It is therefore clear that the offspring of adelphic unions, so far from being at an advantage compared with the offspring of father-daughter unions, are at a disadvantage in the proportion of 4 to 3. In the third place, in father-daughter unions the male is physically as well as sexually mature. In adelphic unions both parties are probably immature. Consequently from this point of view also the advantage is with the supposed injurious type of union. Now if the father-daughter union was less harmful than the brother-sister union, a fortiori are uncle-niece and similar unions less harmful. Yet Morgan supposes them to have been prohibited in favour of brother and sister unions.

Mr Morgan's reformation therefore turns out to have been no reformation at all, but a retrograde step. Assuming however that the facts were as he supposed them to be, and that the reformation was a real one, it is by no means clear how he supposes it to have been brought about. It was, as we have seen, an unconscious150 reformation; it is not supposed therefore that the primeval savage detected more pronounced signs of degeneracy in the offspring of one class of union and by the force of public opinion caused such unions to fall into disrepute and ultimately into desuetude. So far as can be seen the method which Mr Morgan had in his mind was this: certain unions resulted in offspring less able to maintain the struggle for existence, and these families consequently tended to die out. Other unions—those of sisters and brothers—on the other hand produced more vigorous children, and tended to perpetuate themselves. Whereas originally there was no tendency either one way or the other, some families developed from unknown causes, which, whatever they were, were neither moral nor utilitarian, the practice of brother and sister marriage. This diathesis followed the ordinary laws of descent, and eventually those families which were fortunate enough to be affected in that way exterminated their rivals.

Now, as will be shown immediately, this course of events seems to be in contradiction with the facts of savage society at the present day and with all probability. Apart from that however, how does Mr Morgan suppose his eugenic diathesis to be transmitted? It can hardly be maintained that this was the result of the different social conditions of the families in which brothers and sisters intermarried. Obviously there would be nothing to prevent the male in one of these unions from reverting to the other type of marriage. This would indeed be highly probable for reasons to be developed in the next paragraph. But if social conditions were not the determining factor, we are left with the somewhat grotesque theory of innate ideas. It is hardly necessary to refute this origin of social evolution.

Perhaps the strongest objection, however, to Mr Morgan's theory is the fact that in the most primitive communities the female tends to be younger, often much younger, than her mate. It is a readily ascertainable fact, though it seems to have been neglected by Mr Morgan, that the age of puberty does not coincide with the greatest development of the physical powers, but precedes it in the human subject by many years. The result of this is that the younger males are, as a rule, in the case of many mammals, held in subjection by the patriarch of the herd, the result being what I have termed above patriarchal polygyny, as long as the old male retains his powers. We have, it is true, no evidence of any such conditions among the anthropoids; but it must not be forgotten that we have no evidence of the consanguine family either among anthropoids, other mammals or human beings.

It tells against the hypothesis of patriarchal polygyny that both among horses and among camels there is evidence of the existence of actual sexual aversion between both sire and filly and dam and colt in the first case; and, as Aristotle tells us, at least between dam and colt in the case of camels; but we can hardly argue from Ungulata to Primates.

However this may be, the objections to Morgan's theories do not lose their strength. Enough has perhaps been said of them from the point of view of theory. We may look at them in the light of the known facts of social evolution among races of low stages of culture.

If we now turn for a moment to see what light Australian facts throw on the first two stages postulated by Mr Morgan, we find that the theoretical objections are amply supported by the course of evolution which can be traced in Australian social regulations. It will be recollected that in his view father-daughter marriage disappeared first, then brother and sister marriage. Totemism apart, there are in Australia, as we have seen, two kinds of organisation for the regulation of marriage—phratries, the dichotomous division of the southern tribes, and classes, the four-fold or eight-fold division of the other areas as to which we have any knowledge. Of these the phratry is demonstrably older than the class. But the result of the division of a tribe into two phratries is to prevent brother and sister marriage, while, so far as phratry rules are concerned, father and daughter are still free to marry in those tribes where the descent is matrilineal. The result (though not necessarily the original object) of the class-system, on the other hand, is to prevent the marriage of fathers and daughters and generally of the older generation with the younger, so far as the classes actually represent generations. In actual practice the class into which a man may marry includes females of all ages, so that he is only debarred from marrying young females if they are his own daughters. But if we may assume that the original object of the classes was to prevent the intermarriage of different generations, it is at once obvious that in Australia the evolution postulated by Mr Morgan, if it took place at all, took place in reverse order, the brother and sister marriage being the first to be brought under the ban.

The objections to which attention has been called seem to make it difficult if not impossible to accept Morgan's explanations either of the processes or of the causes which led to the passage from promiscuity to communal marriage.

CHAPTER XII.

In actual practice we find that the name which a mother applies to her own child is applied by her equally to the children of the women whom her husband might have married. This state of things may obviously arise from one of three causes, (a) In the first place the name may have been originally that which a mother applied to her own son, and it may have been extended to those who were her nephews in a state of monogamy, or stepsons (= sons of other women by the same father) in a state of polygyny either with or without polyandry. (b) The theory that a name was applied originally to own and collateral relatives has already been discussed, so far as it refers to the "undivided commune." The case of regulated promiscuity is different and must be considered here. (c) On the other hand the name which she uses may have been expressive of tribal status or group status, and may have had nothing to do with descent.

It is unnecessary to say much about the first of these possibilities. First, there is no evidence to show that such a thing has taken place; secondly, we can see no reason why such a thing should take place; thirdly, if such a change of meaning did take place, it is quite clear that we have no grounds for regarding the philological evidence for group marriage as having the slightest significance.

In connection with the second hypothesis—that the names actually represent the relations formerly existing, it may be well to preface the discussion by a few remarks on the regulation of marriage in Australia. The rules by which the Australian native is bound, when he sets out to choose a wife, make the area of choice as a rule dependent on his status, that is to say, he must, in order to find a wife, go to another phratry, class, totem-kin, or combination of two of these, membership of which depends on descent, direct or indirect; on the other hand he may be limited by regulations dependent on locality, that is to say he may have to take a wife from a group resident in a certain area. There is reason to suppose that the latter regulations are the outcome of earlier status regulations which have fallen into desuetude. However this may be, all that we are here concerned with is the fact that regulations in this case also are virtually dependent on descent, inasmuch as a man is not in practice free to reside where he likes, but remains in his own group, though occasionally he joins that of his wife (this does not apparently affect the exogamic rule). The groups are therefore to all intents and purposes totem-kins with male descent.

Taking the Kurnai as our example of the non-class-organised groups, we find that the fraternal relationship once started goes on for ever; the result of this is that with few exceptions the whole of the intermarrying groups, so far as they are of the same generation, are brothers and sisters. Dr Howitt, whose authority on matters of Australian ethnology is final, recognises that on the principles on which group marriage is deduced from terms of relationship, this fact should point to the Kurnai being yet in the stage of the undivided commune (why, it is difficult to see, when they are definitely exogamous), but regards the argument from terms of relationship as untrustworthy in this instance. If it is not reliable in one case it may well be unreliable in all; we are entitled to ask supporters of the hypothesis of group marriage what differentiates this case from those in which they have no doubt of the validity of the philological argument.

Now if Dr Howitt's doubts as to the interpretation to be put upon the Kurnai terms of relationship are correct, we may reasonably, in the absence of proof that they originated in a different way from the Malayan terms, ask ourselves upon what basis the case for promiscuity rests. Beyond a few customs, and it will be shown below that it is unnecessary to regard them as survivals of a period when marriage was unknown, the proof is purely philological, and on examination the philological proof is found to be wanting.

Dr Howitt, in his recent book, rests the case for the undivided commune (i.e. promiscuity) on the Australian terms of relationship which he discusses, viz. those of the Dieri and the Kurnai. He will not admit that the Kurnai terms point to the undivided commune; we are therefore left with the Dieri terms. But the Dieri organisation, so far from being that of an undivided commune, is the two-phratry arrangement by which a man is by no means free to marry any woman in his tribe, but is limited to one-half of the women; further, tribal customs limit his choice still further and compel him to marry his mother's mother's brother's daughter's daughter (these terms do not refer to blood but so-called "tribal" relationship, i.e. it is a woman with a certain tribal status whom he has to marry). Where then does Dr Howitt find his proof of promiscuity?

We have, it is true, a certain number of tribal legends, according to which the phratry organisation was instituted to prevent the marriage of too near kin. But, quite apart from the fact that tribal legends are not evidence, the legends merely point to a period when marriage was unregulated, when a man was free to marry any woman, not when he was de facto or de jure the husband of every woman. Even if it be proved beyond question that marriage was once unregulated, it does not follow that promiscuity prevailed.

The existence of the undivided commune is a proof of promiscuity only for those who discover proofs of group marriage in the divided commune, in other words in the terms of relationship and the customs of the ordinary two-phratry tribe of the present day. We may therefore let the decision of the question of the validity of terms of relationship as a proof of extensive connubial activities rest upon the discussion of the evidence to be drawn from the tribes selected by Dr Howitt and Messrs Spencer and Gillen, viz. the Dieri and the Urabunna.

It may however be pointed out that neither of these writers has dealt with the passage from promiscuity to "group marriage," nor shown how under the former system terms of relationship could come into existence at all. With the difficulties we have dealt above.

We must now revert to the question of the origin of the so-called "terms of relationship." Are they expressive of kinship or only of status and duties? Neither Lewis Morgan nor the authorities on Australian marriage customs—Dr Howitt and Messrs Spencer and Gillen—discuss the question at length, but seem to regard it as an axiom (although they warn us that all European ideas of relationship must be dismissed when we deal with the classificatory system) that all these terms may be interpreted on the hypothesis that the European relationships to which they most nearly correspond actually existed in former times, not, as in Europe, between individuals, but between groups. The case on which Spencer and Gillen rely is that of the unawa relationship. They argue that a man is unawa to a whole group of women, one of whom is his individual wife; for this individual wife no special name exists, she is just unawa (= noa) like all the other women he might have married. Consequently the marital relation must have existed formerly between the man in question and the whole group of unawa women. The reasoning does not seem absolutely conclusive, and our doubts as to the validity of the argument are strengthened when we apply it to another case and find the results inconsistent with facts which are known to the lowest savage. Not only has a man only one name for the women he might have married, and for the woman he actually did marry, but a mother has only one name for the son she actually bore, and for the sons of the women who, if they had become her husband's wives, would have borne him sons in her stead. From this fact by parity of reasoning we must draw the obvious conclusion that during the period when group marriage was the rule, individual mothers were unknown. If we are entitled to conclude from the fact that a man's wife bears the same name for him as all the other women whom he might have married, that he at one time was the husband of them all, then we are obviously equally entitled to conclude, from the fact that a woman's son is known to her by the same name as the sons of other women, either that during the period of group marriage she actually bore the sons of the other women or that the whole group of women produced their sons by their joint efforts. Finding that the term which is translated "son" is equally applied by the remainder of the group of women to the son of the individual woman, whose case we have been considering, we may discard the former hypothesis and come to the conclusion that if there was a period of group marriage there was also one of group motherhood. This interesting fact may be commended to the attention of zoologists.

It is perhaps unnecessary to pursue the argument any further. The single point on which Spencer and Gillen rely is sufficiently refuted by a single reductio ad absurdum. If more proof is needed it may be found in Dr Howitt's work151. We learn from him that a man is the younger brother of his maternal grandmother, and consequently the maternal grandfather of his second cousin. Surely it is not possible in this case to contend that the "terms of relationship" are expressive of anything but duties and status. It seems unreasonable to maintain in the interests of an hypothesis that a man can be his own great uncle and the son of more than one mother.

From the foregoing discussion it will be clear that there are very grave, if not insurmountable, difficulties in the way of regarding the "terms of relationship" as being in reality such. In reply to those who regard them as status terms it is urged that if they are not terms of relationship, then the savages have no terms of any sort to express relationships which we regard as obvious, the implication being that this is unthinkable.

Now in the first place it may be pointed out that the converse is certainly true. Civilised man has a large number of terms of relationship, but he has none for such ideas as noa; a boy has no term for all men who might have been his father; a woman has no name for the children of all women who might have married her husband, if she had not anticipated them. To the savage this is just as unthinkable as the converse seems to be to some civilised men.

In the second place it is perfectly obvious that the savage has, as a matter of fact, no names for the quite unmistakeable relationship of mother and child. The name which an Australian mother applies to her son, she applies equally to the sons of all other women of her own status; the name which a son applies to his mother, he applies equally to all the women of her status, whether married or unmarried, in old age, middle life, youth, or infancy. If there is no term for this relation we can hardly argue that the absence of terms for other relations is unthinkable.

Morgan attempted to meet this objection by urging that in a state of promiscuity a woman would apply the same name to the children of other women as to her own, because they were or might be by the same father. But in the first place this assumes that the relationship to the father was considered rather than the relationship to the mother, and this is against all analogy. In the second place, even granting Morgan's postulate, the relation of a mother to her son is not that of a wife to the children of other wives of a polygynous husband. Poverty of language is therefore established in this case, and may be taken for granted where the obvious relationships are concerned.

It has been pointed out more than once that there are grave difficulties in the way of any hypothesis which assumes that terms of relationship, properly so called, were evolved in a state of pure promiscuity. It has now been shown that no intelligible account of the meaning of such terms can be given, even if we dismiss the difficulties just mentioned and assume that terms were somehow or other evolved, and a transition effected to a state of regulated promiscuity. If on the other hand we regard the "terms of relationship" as originally indicative of tribal status and suppose they have been transformed in the course of ages into "descriptive" terms such as we use in everyday life, the difficulties vanish.

For one proof of this hypothesis we need look no further than the terms of relationship applied by a mother to her own (and other) children, an illustration which has already done duty more than once. It is abundantly clear that what this term expresses is not relationship but status, the relation of one generation to the next in the Malayan system, of the half of a generation to the next generation in the same moiety of the tribe among the Dieri, and so on.

It is admitted even by believers in group marriage that the terms of relationship do not correspond to anything actually existing; beyond the "survivals" which we shall consider below, they can produce no shadow of proof that the terms ever did correspond to actual relationships, as they understand them. They can give no proof whatever that they did not express status.

It is therefore a fair hypothesis that unawa (noa) and similar terms express status and not relationship. From the example of mother and son we see that the Australian does not select for distinction by a special term that bond which is most obvious both to him and us. It is therefore by no means surprising that by unawa he should mean, not the existence of marital relations, but their possibility, from a 'legal' point of view. Just as he is struck, not by the genetic relation between mother and son, but by the fact that they belong to different generations, so in the case of husband and wife the existence of marital relations between them is neglected, and the point selected for emphasis is the legality of such marital relations, whether existent or not.

It is singular that anyone should regard this savage view of life as anything but natural. For the Australian the due observance of the marriage regulations is a tribal matter; their breach, whether the connection be by marriage or free love, is a matter of more than private concern. The relations of a man with his legal wife however concern other members of the tribe but little. Public opinion among the Dieri, it is true, condemns the unfaithful wife, but her punishment is left to the husband; among the Kamilaroi the tribe indeed takes the matter up but only on the complaint of the husband; and generally speaking it is the husband who, possibly with his totemic brethren, pursues the abductor. We have therefore in this insistence on the legal status of the couple and the comparative indifference to the husband's rights a sufficiently exact parallel to the insistence on status and not marital relations in the use of the term unawa.

The course of evolution has been, not, as group-marriagers contend, from group to individual terms of relationship but from terms descriptive of status to terms descriptive of relationship.

It is, in fact, on any hypothesis, impossible to deny this. Whatever terms of relationship may have meant in the past, no believer in group marriage contends that they represent anything actually existing. But this is equivalent to admitting that they express status and not relationship, and no proof has ever been given that they were ever anything else.