Chapter V · A CRITICISM OF THE HYPOTHESIS OF PROMISCUITY

(Continued)

We are indebted to Mr. Lewis H. Morgan for information as to the names of various degrees of kinship among no fewer than 139 different races or tribes. This collection shows that very many peoples have a nomenclature of relationships quite different from our own. Mr. Morgan divides the systems into two great classes, the descriptive and the classificatory, which he regards as radically distinct. “The first,” he says, “which is that of the Aryan, Semitic, and Uralian families, rejecting the classification of kindred, except so far as it is in accordance with the numerical system, describes collateral consanguinei, for the most part, by an augmentation or combination of the primary terms of relationship. These terms, which are those for husband and wife, father and mother, brother and sister, and son and daughter, to which must be added, in such languages as possess them, grandfather and grandmother, and grandson and granddaughter, are thus restricted to the primary sense in which they are here employed. All other terms are secondary. Each relationship is thus made independent and distinct from every other. But the second, which is that of the Turanian, American Indian, and Malayan families, rejecting descriptive phrases in every instance, and reducing consanguinei to great classes, by a series of apparently arbitrary generalizations, applies the same terms to all the members of the same class. It thus confounds relationships, which, under the descriptive system, are distinct, and enlarges the signification both of the primary and secondary terms beyond their seemingly appropriate sense.”435

The most primitive form of the classificatory group is the system of the “Malayan family,”436 which prevails among the Hawaiians, Kingsmill Islanders, Maoris, and, presumably, also among several other Polynesian and Micronesian tribes.437 According to this system, all consanguinei, near and remote, are classified into five categories. My brothers and sisters and my first, second, third, and more remote male and female cousins, are the first category. To all these without distinction I apply the same term. My father and mother, together with their brothers and sisters, and their first, second, and more remote cousins, are the second category. To all these without distinction I apply likewise the same term. The brothers, sisters, and several cousins of my grandparents I denominate as if they were my grandparents; the cousins of my sons and daughters, as if they were my sons and daughters; the grandchildren of my brothers and sisters and their several cousins, as if they were my own grandchildren. All the individuals of the same category address each other as if they were brothers and sisters. Uncleship, auntship, and cousinship being ignored, we have, as far as the nomenclature is considered, only grandchildren.438

From this system of nomenclature all the others belonging to the classificatory group have, according to Mr. Morgan, been gradually developed. The system of the Two-Mountain Iroquois differs from that of the Hawaiians essentially in two respects only, the mother’s brother being distinguished by a special term, and so also a sister’s children. The Micmac system is somewhat more advanced. Not only does a man call his sister’s son his nephew, but a woman applies the same term to her brother’s son; and not only is a mother’s brother termed an uncle, but also the father’s sister is distinguished by a special term, as an aunt. A father’s brother is called a “little father;” and a mother’s sister, a “little mother.” Still more advanced is the system of the Wyandots, which may be regarded as the typical system of the Indians.439 A mother’s brother’s son and a father’s sister’s son are no longer called by the same terms as brothers, but are recognized as cousins; and women apply to their mother’s brother’s grandsons no longer the same term as to their sons, but call them nephews.

It is needless to enter into further details. Those who shrink from the trouble of reading through Mr. Morgan’s extensive tables, will find an excellent summary of them in the fifth chapter of Sir John Lubbock’s great work on ‘The Origin of Civilization.’ It may, however, be added that the most advanced system of the classificatory group is that of the Karens and Eskimo, which differs from our own in three respects only. The children of cousins are termed nephews; the children of nephews, grandchildren; and a grandfather’s brothers and sisters, respectively, grandfathers and grandmothers. “Hence,” says Sir John Lubbock, “though the Karens and Eskimo have now a far more correct system of nomenclature than that of many other races, we find, even in this, clear traces of a time when these peoples had not advanced in this respect beyond the lowest stage.”440

From these systems of nomenclature Mr. Morgan draws very far-reaching conclusions, assuming that they are necessarily to be explained by early marriage customs. Thus, from the “Malayan system,” he infers the former prevalence of “marriage in a group” of all brothers and sisters and cousins of the same grade or generation; or, more correctly, his case is, that if we can explain the “Malayan system” on the assumption that such a general custom once existed, then we must believe that it did formerly exist. “Without this custom,” he says, “it is impossible to explain the origin of the system from the nature of descents. There is, therefore, a necessity for the prevalence of this custom amongst the remote ancestors of all the nations which now possess the classificatory system, if the system itself is to be regarded as having a natural origin.”441 The family resulting from this custom he calls, in his latest work, the “consanguine family,” and in this, consisting of a body of kinsfolk, within which there prevailed promiscuity, or “communal marriage,” between all men and women of the same generation, the family in its first stage is recognized.442 Mr. Morgan believes, however, that as a necessary condition antecedent to this form of the family, promiscuity, in a wider sense of the term, may be theoretically deduced, though, as he says, “it lies concealed in the misty antiquity of mankind beyond the reach of positive knowledge.”443

It is needless here to consider whether the last conclusion holds good. I shall endeavour to prove that Mr. Morgan’s inference of a stage of promiscuous intercourse even within the prescribed limits is altogether untenable. All depends on the point whether the “classificatory system” is a system of blood-ties, the nomenclature having been founded on blood-relationship, as near as the parentage of individuals could be known. Mr. Morgan assumes this, instead of proving it.

Yet in the terms themselves there is, generally, nothing which indicates that they imply an idea of consanguinity. Professor Buschmann has given us a very interesting list of the names for father and mother in many different languages.444 The similarity of the terms is striking. “Pa,” “papa,” or “baba,” for instance, means father in several languages of the Old and New World, and “ma,” “mama,” means mother. The Tupis in Brazil have “paia” for father, and “maia” for mother;445 the Uaraguaçú, respectively, “paptko” and “mamko.”446 In other languages the terms for father are “ab,” “aba,” “apa,” “ada,” “ata,” “tata”; those for mother, “ama,” “emä,” “ana,” “ena,” &c. According to Buschmann, there are four typical forms of words for each of these ideas: for father, “pa,” “ta,” “ap,” “at”; for mother, “ma,” “na,” “am,” “an.” Sometimes, however, the meaning of the types is reversed. Thus, in Georgian,447 as well as in the Mahaga language of Ysabel,448 “mama” stands for father; whilst the Tuluvas in Southern India call the father “amme,” and the mother “appe.”449

The terms used often fall outside of the types mentioned. In the Lifu tongue, for example, one term for father is “kaka;”450 in the Duauru language of Baladea, “chicha”;451 in the Maréan tongue, “chacha” or “cheche.”452 Again, among the Chalcha Mongols and some related peoples, mother is “ekè.”453 In the Kanúri language, of Central Africa, the mother is called “ya”;454 while the Kechua in Brazil call the father “yaya.”455 Among the Bakongo, as I am informed by Mr. Ingham, “se” means father; in Finnish, “isä.” Again, by the Brazilian Bakaĭri, the mother is called “ise”;456 and, by the people of Aneiteum, New Hebrides, “risi.”457

Similar terms are often used for other relationships. The Greek, “πάππος” signifies grandfather, and “μάμμα” grandmother. In the Kanúri language, “yaya” stands for elder brother;458 and, in Lifuan, “mama” and “dhina” are terms for brother, whilst mother is “thine.”459

The origin of such terms is obvious. They are formed from the easiest sounds a child can produce. “‘Pa-pa,’ ‘ma-ma,‘ 'tata,’ and ‘apa,’ ‘ama,’ ‘ata,’” Professor Preyer says, “emerge originally spontaneously, the way of the breath being barred at the expiration, either by the lips (p, m), or by the tongue (d, t).”460 Yet the different races vary considerably with regard to the ease with which they produce certain sounds. Thus the pronunciation of the labials is very difficult to many Indians,461 on account of which their terms for father, mother, or other near kinsfolk, often differ much from the types given by Professor Buschmann.

It is evident that the terms borrowed from the children’s lips have no intrinsic meaning whatever. Hence, if a Bakaĭri child calls its father and father’s brother “tsogo,” its mother and mother’s sister “tsego”;462 if a Macúsi names his paternal uncle “papa” as well as his father, and an Efatese names his father and all the tribe brothers of his father “ava” or “tama”;463 if the Dacotahs apply the term “ahta” not only to the father, but also to the father’s brother, to the mother’s sister’s husband, to the father’s father’s brother’s son, &c., and the term “enah” not only to the mother, but also to the mother’s sister, to the mother’s mother’s sister’s daughter, &c.;464 if, among the New Caledonians, an uncle, taking the place of a father, is called “baba” like the father himself, and an aunt is called “gnagna” like a mother;465 if, as Archdeacon Hodgson of Zanzibar, writes to me, a native of Eastern Central Africa uses the words “baba” and “mama” not only for father and mother respectively, but also, very commonly, for “any near relationship or even external connection;” if, finally, the Semitic word for father, “ab” (“abu”), is not only used in a wide range of senses, but, to quote Professor Robertson Smith, “in all dialects is used in senses quite inconsistent with the idea that procreator is the radical meaning of the word,”466—we certainly must not, from these designations, infer anything as to early marriage customs.

Of course there are other terms applied to kinsfolk besides words taken from the lips of children, or words derived from these. But though considerable, their number has been somewhat exaggerated. Thus, for instance, Professor Vámbéry, in his work upon the primitive culture of the Turko-Tartars, says that the terms for mother, “ana” or “ene” have originally the meaning of woman or nurse, being derived from the roots “an” and “en.”467 Exactly the reverse seems to be the fact, the terms for mother being the primitive words. In the same way, I cannot but think that Professor Max Müller and several other philologists are in error in deriving “pitár,” “pater,” “father,” from the root “pa,” which means to protect, to nourish; and “mâtár,” “mater,” “mother,” from the root “ma,” to fashion.468 It seems, indeed, far more natural, as has been pointed out by Sir J. Lubbock and others, that the roots “pa,” to protect, and “ma,” to fashion, come from “pa,” father, and “ma,” mother, and not vice versa.469 I am the more inclined to accept this explanation, as Mr. A. J. Swann informs me, from Kavala Island, Lake Tanganyika, that among the Waguha, the words “baba,” and “tata,” which mean father, also have the meaning of protector, provider.

I do not deny that relationships—especially in the collateral and descending lines—are in some cases denoted by terms derived from roots having an independent meaning; but the number of those that imply an idea of consanguinity does not seem to be very great. Mr. Bridges writes that, among the Yahgans, “the names ‘imu’ and ‘dabi’—father and mother—have no meaning apart from their application, neither have any of their other very definite and ample list of terms for relatives, except the terms ‘macu’ and ‘macipa’ son and daughter. These terms refer to ‘magu’ which means parturition; ‘cipa’ (‘keepa’) signifies woman or female.” In Bakongo, according to Mr. Ingham, “se” and “tata” denote father; “mama,” “mbuta,” and “ngudi,” mother; “nfumu,” elder brother or sister; “mbunzi,” younger brother; and “mbusi,” younger sister. “Nfumu” means also Sir, chief; “mbuta” means “the one who bore,” from “buta,” or “wuta,” to beget; and “ngudi,” “the one we descended from.” Again, Mr. Radfield informs me that, in the language of Lifu, the term for father means root; the term for mother, foundation or vessel; the term for sister, forbidden or “not to be touched;” and the terms for eldest and younger brother, respectively, ruler and ruled. It is possible—I should even say probable—that, in these instances also, the designations for relationships are the radical words. Besides, it should be observed that, in Yahgan, “the terms for relatives are strictly reserved for such, neither are they interchanged,” and that in Bakongo, the terms “tata” and “mama” are used as signs of respect to any one, whilst the terms “mbuta” and “ngudi” seem to be applied exclusively to the mother.

Not only has Mr. Morgan given no evidence for the truth of his assumption that the “classificatory system” is a system of blood-ties, but this assumption is not even fully consistent with the facts he has himself stated. It is conceivable that uncertainty as regards fatherhood might have led a savage to call several men his fathers, but an analogous reason could never have induced him to name several women his mothers. Hence, if a man applies the same term to his mother’s sisters as to his mother, and he himself is addressed as a son by a woman who did not give birth to him, this evidently shows that the nomenclature, at least in certain cases, cannot be explained by the nature of descent.470

There can be scarcely any doubt that the terms for relationships are, in their origin, terms of address. “The American Indians,” says Mr. Morgan, “always speak to each other, when related, by the term of relationship, and never by the personal name of the individual addressed.”471 From a psychological point of view, it would, indeed, be surprising if it could be shown that primitive men, in addressing all the different members of their family or tribe, took into consideration so complicated a matter as the degree of consanguinity. Can we really believe that a savage whose intelligence, perhaps, was so deficient that he was scarcely able to count his own fingers, applied the same term to his cousins as to his brothers, because he was not certain whether, after all, they were not his brothers and that, when he did make a distinction between them, he did so because they were begotten by different fathers? Facts show that savages generally denominate their kindred according to much simpler principles, the names being given chiefly with reference to sex and age, as also to the external or social, relationship in which the speaker stands to the person whom he addresses.

In every language there are different designations for persons of different sexes. In the rudest system of nomenclature, the Hawaiian, father and other kinsmen of the same generation are called “makua kana;” mother, mother’s sisters, father’s sisters, &c., “makua waheena,” “kana” and “waheena” being the terms for male and female. A son is called “kaikee kana,” a daughter “kaikee waheena,” whilst “kana” alone is applied to husband, husband’s brother, and sister’s husband, and “waheena” to wife, wife’s sister, brother’s wife, &c.

There are also separate terms in every language for relations belonging to different generations. Among the lower races especially, age, or, more exactly, the age of the person spoken to compared with that of the speaker, plays a very important part in the matter of denomination. According to Dr. Davy, the Veddahs appear to be without names; “a Veddah interrogated on the subject, said, ‘I am called a man: when young, I was called the little man: and when old, I shall be called the old man.’”472 The Hawaiians, as we are informed by Judge Andrews, have no definite general word for brother in common use. But “kaikuaána” signifies any one of my brothers, or male cousins, older than myself, I being a male, and any one of my sisters, or female cousins, older than myself, I being a female; whilst “kaikaina” signifies a younger brother of a brother, or a younger sister of a sister.473 Such distinguishing epithets applied to older and younger are, in fact, very frequently met with among uncivilized peoples. Thus, touching the Andamanese, Mr. Man states that “brothers and sisters speak of one another by titles that indicate relative age: that is, their words for brother and sister involve the distinction of elder or younger.” A like system is adopted by them in respect to half-brothers, half-sisters, cousins, brothers-in-law, and sisters-in-law.474 In certain languages, too, there are special terms for an uncle on the father’s side older than the father, and for an uncle younger than he;475 and in the Fulfúlde tongue, the age of the uncles is so minutely specified, that the first, second, third, fourth, and fifth uncle, on both the father’s and the mother’s side, are each called by a particular name.476

The wider meaning in which many terms for kinship are used bear witness in the same direction. The Rev. J. Sibree states that, in Hova, “ray,” father, does not take the sense the corresponding word in many Semitic languages has, of “maker” of a thing, but it is used in a wide sense as an elder or superior; and “rény,” mother, is also used in a wide sense as a respectful way of addressing an elderly woman.477 Mr. Swann writes to me that, among the Waguha, West Tanganyika, men advanced in years are termed “baba,” father, whilst, in other parts of Equatorial Africa, according to Mr. Reade, old men are addressed as “rera,” father, and old women as “ngwe,” mother.478 The Russian “batushka” and “matushka,” as also the Swedish “far” and “mor,” are often used in a similar way. Again, Mr. Cousins asserts that, among the natives of Cis-Natalian Kafirland, the terms for father, mother, brother, and sister, are not restricted to them only, but are applied equally to other persons of a similar age, whether related or otherwise. “‘Bawo,’ father,” he says, “means elder or older, ‘bawo-kulu’ means a big-father, one older than father.” Probably “bawo,” as belonging to the type “pa,” was originally used as a term of address, from which the sense of elder or older was derived; but this does not interfere with the matter in question. The Rev. E. Casalis, writing of the Basutos, states that “in addressing a person older than one’s self, one says, ‘My father, my mother;’ to an equal, ‘My brother;’ and to inferiors, ‘My children.’”479 The Finnish “isä” and the Votyak “ai,” father, the Lappish “Aja,” and the Esthonian “äi,” grandfather, are evidently related to, and probably the roots of, the Finnish “iso” and “äijä” which mean big.480 The Chukchi use, besides “atta” for father and “mámang” for mother, “empynátchyo” and “émpyngau” respectively, which obviously have the same root as “émpytchin,” elder or older.481 The Brazilian Uainumá call a father “paii,” but also “pechyry,” i.e., old.482 “Les jeunes Australiens,” says Bishop Salvado, “ont coutume d’appeler ‘mama’ ou ‘maman’ (c’est-à-dire-père) tous les vieillards, comme aussi ‘N-angan’ (ou mère) les femmes avancées en âge.”483 According to Nicolaus Damascenus, the Galactophagi denominated “all old men fathers; young men, sons; and those of equal age, brothers.”484 In German, the parents are “die Eltern,” the older (“die Aelteren”), and they are also called familiarly “die Alten;” the father, “der Alte;” and the mother, “die Alte” or “Altsche.”485 Again, among the North American Indians, old people are very commonly named grandfathers and grandmothers;486 whilst the Finnish “ämmä” does not signify grandmother only, but old woman in general.487 Among the Tsuishikari Ainos, the maternal grandfather and grandmother of a child are called both by him, and his father, “henki” and “unarabe” respectively.488

As to the collateral line, it should be observed that, in Ćagatai, an elder sister is called “egeći,” which actually means old woman “ege,” old, big; “eći,” woman, sister.489 In Hungarian, where “bátya” stands for elder brother, an uncle is “nagybátya,” i.e., a big elder brother.490 Among many Ural-Altaic peoples, the same term is applied to an elder brother as to an uncle, to an elder sister as to an aunt.491 Were we to follow Mr. Morgan’s way of reasoning, we should, from this nomenclature, come to very curious conclusions as to the early marriage customs of the peoples in question.

Again, in the Galibi language of Brazil, “tigami” signifies young brother, son, and little child indiscriminately;492 and several languages have no other words for son and daughter than those for lad and girl.493 Thus, in Hawaiian, a son is called male child, or more properly, little male; and a daughter, female child or girl.494 Mr. George Bridgman states that, among the Mackay blacks of Queensland, the word for daughter is used by a man for any young woman belonging to the class which his daughter would belong to if he had one.495 And, speaking of the South Australians, Eyre says, “In their intercourse with each other, natives of different tribes are exceedingly punctilious and polite; ... almost everything that is said is prefaced by the appellation of father, son, brother, mother, sister, or some other similar term, corresponding to that degree of relationship which would have been most in accordance with their relative ages and circumstances.”496

All those names, refer, as previously mentioned, not to the absolute, but to the relative, age of the person addressed. Often, too, there is a certain relativity in the use of words denominating sex. Mr. Dall remarks, for instance, that among the Eskimo, the form of the terms of relationship “appears to depend in some cases more on the sex of the speaker than on that of the person to whom the term refers.” In Eastern Central Africa, “if a man have a brother and a sister, he is called one thing by the brother, but quite a different thing by the sister.”497 And several other instances of the same kind are to be found in Mr. Morgan’s tables.

As for the third factor influencing the terms of address—i.e., the social relationship which exists between the addresser and the one addressed,—it is obvious that different designations are applied to enemies and friends, to strangers and members of the family-circle, nay, generally, to persons to whom one stands in an altogether different external relationship. The importance of this factor is evident from several statements. Thus, among the Hovas, according to Mr. Sibree, the words for brother and sister “are also used widely for any person whom one meets and desires to act towards in a friendly manner.”498 The Fuegians says Mr. Bridges, form certain kinds of friendships, and “speak of aunts, uncles, brothers, sisters, cousins, nieces and nephews, &c., which are only so through the friendships established.”499 Among the Waguha, strangers are called “ndugu,” brother, if of the same tribe;500 and Mr. Hartshorne tells us that the Veddahs applied to him the term “hura,” or cousin.501 We can understand, then, why the same name, as a rule, is used by the savage to denote just the persons of the same sex and of like age who belong to his own family-circle; and why, as a consequence, the nomenclature is rich or poor according as that circle is small or large. The Yahgans, for instance, who live in families rather than in tribes, have a very definite list of terms for kinsfolk. They have different appellations for nephews and nieces on the brother’s side, and nephews and nieces on the sister’s side, and their words for uncle and aunt differ according as this relationship is paternal or maternal. They have also special terms for father-, mother-, son-and daughter-, brother-and sister-in-law.502 On the other hand, the larger, the body of kinsfolk that keep closely together, and the less it is differentiated, as regards the functions of its various members, the more comprehensive are generally the terms of address. The “classificatory system of relationship” must, therefore, have emerged at a time when the separate families had already united in larger bodies.

The same principle explains how it happens that a maternal uncle is almost always distinguished from a father by a separate term, whilst this is not the case with an uncle on the father’s side, the former generally living in another community from his nephew, and, besides, very frequently standing to him in a quite peculiar relationship through the rules of succession. It may be fairly assumed, too, that a mother’s sister much oftener than a father’s sister is called a mother, because sisters, among savages, keep as a rule, far more closely together, when married, than brothers and sisters; sometimes even, especially among the North American Indians, they are the wives of the same man. If we add to this that a father’s brother’s son and a mother’s sister’s son are more commonly addressed as brothers than as father’s sister’s son and a mother’s brother’s son, it becomes obvious to how great an extent the nomenclature is influenced by external relations. But as a certain kind of external relationship is invariably connected with a certain degree, or certain degrees, of blood-relationship, the designations given with reference to the former have been taken as terms for the latter.

The basis on which Mr. Morgan has built his hypothesis must be considered, then, altogether untenable.503 It cannot be proved that, where the “classificatory system” prevails, the nomenclature was intended to express the degree of consanguity so exactly as he assumes, or that it had originally anything whatever to do with descent. On the contrary, I have endeavoured to show that the case was probably just the reverse; so that no inference regarding early marriage customs is to be drawn from the terms for relationships. Even now, in Spanish, a brother’s great-grandson is called grandson; in Bulgarian, as also in Russian, a father’s father’s brother is termed a grandfather, and a father’s father’s sister a grandmother; the Greek “ἁνεψιός” appears to have been applied to a nephew, a grandson, and a cousin; “neef,” in Dutch, still expresses these three relationships indiscriminately; in Flemish and Platt Deutsch, “nichte” is applied to a female cousin as well as to a niece; and Shakspeare, in his will, describes his granddaughter, Susannah Hall, as “my niece.”504 Surely, nobody would look upon these designations as relics of ancient times, when there really might have been some uncertainty as to kinship in the direction which the terms indicate. Mr. Morgan himself admits that, in Latin, “nepos” did not originally signify “either a nephew, grandson or cousin, but that it was used promiscuously to designate a class of persons next without the primary relationships.”505

Thirty years ago, in a work of prodigious learning,506 the Swiss jurist, Dr. Bachofen, drew attention to the remarkable fact that a system of “kinship through mothers only” prevailed among several ancient peoples. Moreover, partly from actual statements of old writers, partly from traditions and myths, he came to the conclusion that such a system everywhere preceded the rise of “kinship through males.” A few years later, though quite independently of him, Mr. McLennan set forth exactly the same hypothesis, being led to it chiefly by extensive studies in modern ethnology. While, however, Bachofen explained the phenomenon as a consequence of the supremacy of women, Mr. McLennan regarded it as due to the uncertain paternity which resulted from early promiscuity. “It is inconceivable,” he says, “that anything but the want of certainty on that point could have long prevented the acknowledgment of kinship through males; and in such cases we shall be able to conclude that such certainty has formerly been wanting—that more or less promiscuous intercourse between the sexes has formerly prevailed. The connection between these two things—uncertain paternity and kinship through females only, seems so necessary—that of cause and effect—that we may confidently infer the one where we find the other.”507

It must be observed that the facts adduced as examples of what Mr. McLennan calls “kinship through females only” in most instances imply, chiefly, that children are named after their mothers, not after their fathers, and that property and rank succeed exclusively in the female line. If these customs were to be explained as relics of ancient promiscuity, we certainly should have to admit that such a state was formerly very prevalent in the human race. Yet we could not be sure that it prevailed universally. For, though the number of peoples among whom descent and inheritance follow the mother’s side only, is very considerable,508 the number of those among whom the male line is recognized, is scarcely less—even apart from the civilized nations of Europe and Asia. At present, when anthropologists affirm with so much assurance that a system of exclusive “kinship through females” prevailed everywhere before the tie of blood between father and child had found a place in systems of relationships, it seems appropriate to give a list of peoples among whom such a system does not prevail—a list, however, which cannot pretend to completeness.

Starting, then, with North America, which is acknowledged to be, or to have been, one of the chief centres of “mother-right,” or metrocracy, we meet there with many aboriginal nations among whom a son, as a rule takes the father’s name and becomes his heir.509 Thus Cranz states that, among the Eskimo of Greenland, “when a husband dies, his eldest son inherits his house, tent, and woman’s boat, and besides must maintain the mother and children, who share the furniture and clothes amongst themselves.”510 Among the Indians bordering on the south-east coast of the river St. Lawrence, according to Heriot, the eldest son took the name of his father with the addition of one syllable.511 The Californian tribes512 and the Dacotahs513 recognized chieftainship as hereditary in the male line; and, with reference to the latter, Mr. Prescott remarks that they cannot well forget relationships, as the names of father and mother are both recollected for three or four generations.514 Among the Ahts, the eldest son takes all the property left by his father, and the head-chiefs rank is hereditary in the male line.515 The paternal system prevails, moreover, in thirteen other tribes mentioned by Mr. Frazer in his essay on “Totemism.”516

In Mexico, Yucatan, San Salvador, Honduras, and Nicaragua, succession ran from father to son; and in Vera Paz, according to Las Casas, kinship was so exclusively recognized in the male line, that the people there thought the most remote kin in their own lineage to be more closely related than the daughter of their mother, provided she was not of the same father. On the other hand, Piedrahita tells us that, among the Chibchas, the sons of sisters, and, in default of such, the brothers of the king, were the heirs to the crown of Bogota, but that the sons had a right to the personal property of their father; whilst, according to Herrera, the property was inherited by the brothers, and if there were none living, by the sons of those who were dead.517

Among the Caribs, kinship was reckoned in the female line, but the authority of the chiefs was hereditary in the male line only, the children of sisters being excluded from the succession.518 Among the Macas Indians in Ecuador, property descends from father to son;519 among the Guaycurûs, Abipones, and Araucanians, nobility, or chieftainship, was hereditary in the male line;520 and the Brazilian aborigines, or at least some of them, laid particular stress upon kinship through fathers.521 Again, with reference to the Yahgans of Tierra del Fuego, Mr. Bridges writes, “A child belongs equally to the clan of its father and mother as regards duty of revenge, but is always reckoned a member of the father’s clan only. Children are generally named after their grandparents, paternal or maternal indifferently. They are quite as much attached to their mother’s relatives and these to them, as to their paternal relatives; the only difference is that they are integral parts of the father’s clan, not of the mother’s.” Speaking of the same people, M. Hyades remarks, “L’héritage se transmet à l’époux survivant, ou à défaut, au fils aîné.”522 In short, the paternal system, so far as we know, predominates among the aborigines of South America.

Passing to the Pacific Islands, we find that, though rank and clan are commonly inherited there through the mother, property generally goes in the male line. In Tonga, the son succeeds his father in homage and title,523 and here, as well as in Fiji, on the father’s death, his possessions descend to his children.524 Ellis tells us that, in Tahiti, the child of a chief was invested, soon after its birth, with the name and office of its father,525 and in the case of there being no children, the brother of the deceased assumed the government. In other families property always went to the eldest son.526 Among the Hawaiians, the rank of the principal and inferior chiefs, the offices of the priests, as also other situations of honour and influence, descended from father to son,527 although on the whole, the female line predominated.528 In the Hervey Islands, children belonged either to the father’s or mother’s clan, according to arrangement; usually, however, the father had the preference.529 In New Caledonia, kinship is reckoned in the male line,530 and in Lifu, as Mr. Radfield informs me, children belong to the paternal clan. In the Caroline Group, landed property succeeds mostly from father to son, children are named after their father’s father or mother’s father, and, apparently, the rank of the father influences that of the son, at least if he be a chief.531 Among the Rejangs532 and Bataks533 of Sumatra, as also in several other islands belonging to the Indian Archipelago,534 and in New Guinea,535 the male line prevails. In the Kingsmill Islands, “if a chief has several children by different wives, the son of the mother of the highest rank is the successor.”536 And, in New Zealand, nobility was inherited both in the male and female line; but on the death of a man, his eldest son took the family name which his father had held before him.537

Australian children are generally named after their mother’s clan; but this is not the case in every tribe.538 Among the Gournditch-mara, Turra, Moncalon, Torndirrup, and some other tribes, the male line prevails.539 With reference to the Narrinyeri, the Rev. G. Taplin states that a man’s children belong to his tribe (i.e. clan), and not to their mother’s; that property descends from father to son, and that, in case of a man dying without issue of his own, his possessions are always transmitted to the brother’s children.540 Again, in the Dieyerie tribe of South Australia, the sons take the father’s clan, the daughters the mother’s.541 Even where children are named after their mother, inheritance may go from father to son. Thus, among the West Australians, the hunting ground or landed property descends in the male line, though “children of either sex always take the family name of their mother.”542

Among the Todas, all children belong to the father’s family, and inheritance runs through males only.543 The same is the case with most of the Indian Hill Tribes: either all the sons dividing their father’s property equally, as among the Gonds, Bodo, and Dhimáls; or the eldest son getting the largest share, as among the Kandhs, Karens, and Nagas; or the youngest born male being the only heir, as among the Hos; or the favourite son succeeding without reference to age, as among the Mishmis.544 Among the Pahárias, too, sons inherit, and nephews by sisters get no share.545 The law of succession among the Singphos gives to the eldest son all the landed property of the father, to the youngest all his personal property, while the rest inherit nothing.546 Among the Santals, children belong to the father’s clan;547 and the same is the case with the offspring of intermarriages of Lepchas and Limbus and Butias.548 Touching the Karens, Dr. A. Bunker writes to me, “A child takes a name of its own, and of neither of the parents; but usually the father, being the stronger, takes the child in case of separation. It is regarded as belonging to both parents, so far as blood goes.” If we add to this that the male line prevails in Arabia,549 Tibet,550 throughout Russian Asia,551 and among the Ainos,552 it must be admitted that the system of “kinship through females only” is of very rare occurrence in Asia, being restricted, so far as I know, to a few parts of India, Ceylon, and the Malay Archipelago.553

It is much more prevalent among the African races. Yet, even among them, there are many instances where succession runs in the male line. A king or chief of the Somals554 and Ba-kwileh555 is succeeded by his son. Among the Fulah, this dignity is transmitted to the brother, while, in other instances, succession goes from father to son.556 Among the Negroes of the Gold Coast, according to Bosman, the eldest son succeeded his father in office, though kinship was reckoned through the mother all along this coast, except at Accra.557 Dr. A. Sims writes that, among the Bateke, “the child is considered as belonging to the father and mother equally,” and takes the grandfather’s or grandmother’s name. Among the Waguha, according to Mr. Swann, children are generally named after the father. In Lánda, the eldest son inherits all his father’s possessions, wives included.558 Among the Damaras, whose divisions into clans are derived from the mother, the eldest son of the chief wife, nevertheless, is the successor of his father;559 and the same rule prevails among the Bechuanas.560 The Rev. A. Eyles states that all Zulu children belong to the father’s tribe, and are called by his name or by the name of some of his ancestors.561 According to Mr. Cousins,562 this is essentially true of various Kafir tribes, the first son, however, never being named after the grandfather, but always after the father. Warner, Brownlee, and E. v. Weber assert also that, among the people, inheritance passes from father to son.563 Le Vaillant and Kolben state the same with reference to the Hottentots and Bushmans;564 and Andersson affirms that, among the Namaquas, daughters take the father’s name, sons the mother’s.565 Finally, in the part of Madagascar where Drury was, kinship does not seem to have been, in every case, reckoned through the female, though in that island children generally follow the condition of the mother.566

As for ancient peoples, Bachofen has adduced from the works of classical writers evidence for the uterine line having prevailed among several of them. But, to quote Sir Henry Maine, “the greatest races of mankind, when they first appear to us, show themselves at or near a stage of development in which relationship or kinship is reckoned exclusively through males.”567 Several writers have, it is true, endeavoured to prove that, among the primitive Aryans, descent was traced through females only;568 but the evidence does not seem to be conclusive. Much importance has been attributed to the specially close connection which, according to Tacitus, existed between a sister’s children and their mother’s brothers;569 but Dr. Schrader observes that, in spite of this prominent position of the maternal uncle in the ancient Teutonic family, the patruus distinctly came before the avunculus, the agnates before the cognates, in testamentary succession. He also suggests that, when the head of a household died, the women of his family passed under the guardianship of the eldest son, and that a woman’s children had therefore, quite naturally, a peculiarly intimate relation to their maternal uncle.570 It is safe to say with Professor Max Müller, that we can neither assert nor deny that in unknown times the Aryans ever passed through a metrocratic stage.571

Even if it could be proved—which is doubtful—that, in former times, a system of “kinship through females only,” fully developed, prevailed among all the peoples whose children take the mother’s name and are considered to belong to her clan, though succession runs in the male line, we should still have to account for the fact that a large number of peoples exhibit no traces of such a system.572 And to them belong many of the rudest races of the world—such as the aborigines of Brazil, the Fuegians, Hottentots, Bushmans, and several very low tribes in Australia and India. The inference that “kinship through females only” has everywhere preceded the rise of “kinship through males,” would, then, be warranted only on condition that the cause, or the causes, to which the maternal system is owing, could be proved to have operated universally in the past life of mankind. From Mr. McLennan’s point of view, such an inference would be inadmissible, as he cannot prove the former occurrence of a universal stage of promiscuity or polyandry, leading to uncertain paternity—the cause to which he attributes that system.

Yet it is far from being so inconceivable as Mr. McLennan assumes, that “anything but the want of certainty on that point could have long prevented the acknowledgment of kinship through males.”573 Paternity, as Sir Henry Maine remarks, is “matter of inference, as opposed to maternity, which is matter of observation.”574 Hence it is almost beyond doubt that the father’s participation in parentage was not recognized as soon as the mother’s.575 Now, however, there does not seem to be a single people which has not made the discovery of fatherhood. In reply to my question whether the Fuegians consider a child to descend exclusively or predominantly from either of the parents, Mr. Bridges certainly writes that, according to his idea, they “consider the maternal tie much more important than the paternal, and the duties connected with it of mutual help, defence, and vengeance are held very sacred.” But it is doubtful whether this refers to the mere physiological connection between the child and its parents. Dr. Sims informs me that, among the Bateke, the function of both parents in generation is held alike important, and the Waguha of West Tanganyika, as Mr. Swann states, also recognize the part taken by both. The same is asserted by Archdeacon Hodgson concerning certain other tribes of Eastern Central Africa, though, among them, children take the name of the mother’s tribe. Again, the Naudowessies, according to Carver, had the very curious idea that their offspring were indebted to their father for their souls, the invisible part of their essence, and to the mother for their corporeal and visible part; hence they considered it “more rational that they should be distinguished by the name of the latter, from whom they indubitably derive their being, than by that of the father, to which a doubt might sometimes arise whether they are justly entitled.”576 Moreover, it seems as if the father’s share in parentage, once discovered, was often exaggerated. Thus, referring to some tribes of New South Wales, Mr. Cameron tells us that, although the father has nothing to do with the disposal of his daughter, as she belongs to the clan of her mother’s brother, they “believe that the daughter emanates from her father solely, being only nutured by her mother.”577 Indeed, Mr. Howitt has found in every Australian tribe, without exception, with which he has acquaintance, the idea that the child is derived from the male parent only. As a black fellow once put it to him, “The man gives the child to a woman to take care of for him, and he can do whatever he likes with his own child.”578 Again, Mr. Cousins writes that, according to Kaffir ideas, a child descends chiefly, though not exclusively, from the father; and the ancient Greeks, as well as the Egyptians579 and Hindus,580 maintained a similar view. Nay, Euripides states distinctly that, in his day, the universally accepted physiological doctrine recognized only the share taken by the father in procreation, and Hippocrates, in combating this opinion, and contending that the child descended from both parents, seems to admit that it was a prevalent heresy.581 Finally, it seems probable that the custom known under the name of “La Couvade”—that is, the odd rule, prevalent among several peoples in different parts of the world, requiring that the father, at the birth of his child, shall retire to bed for some time, and fast or abstain from certain kinds of food—implies some idea of relationship between the two.582

Admitting, however, that there was a time when fatherhood, in the physiological sense of the term, was not discovered, I do not think that the preference given to the female line is due to this fact. If the denomination of children and the rules of succession really were in the first place dependent on ideas of consanguinity, it might be expected that a change with reference to the latter would be followed by a change in the former respect also. But the ties of blood have exercised a far less direct influence on the matter in question than is generally supposed, the system of “kinship through females only” being, properly speaking, quite different from what the words imply.

There may be several reasons for naming children after the mother rather than after the father, apart from any consideration of relationship. Especially among savages, the tie between a mother and child is much stronger than that which binds a child to the father.583 Not only has she given birth to it, but she has also for years been seen carrying it about at her breast. Moreover, in cases of separation, occurring frequently at lower stages of civilization, the infant children always follow the mother, and so, very often, do the children more advanced in years. Is it not natural, then, that they should keep the name of the mother rather than that of a father whom they scarcely know? Mr. Belt tells us that the men and women even of the christianised lower classes of Nicaragua often change their mates, and the children, in such cases remaining with the mother, take their surname from her.584 According to Swann, the Creeks conferred the honour of a chief on the issue of the female line, because it was impossible to trace the right by the male issue, women only exceptionally having more than two children by the same father.585 And touching the Khasias, one of the few tribes in India among whom the female line prevails, Dr. Hooker states that they have a very lax idea of marriage, divorce and exchange of wives being common and attended with no disgrace; “the son therefore often forgets his father’s name and person before he grows up, but becomes strongly attached to his mother.”586

Speaking of certain negro tribes, Winterbottom suggested long ago that the prevalence of the female line was to be explained by the practice of polygyny,587 and Dr. Starcke has recently called attention to the same point.588 The Rev. D. Macdonald likewise remarks, in his account of the Efatese of the New Hebrides, that the idea that children are more closely related to the mother than to the father is an idea perfectly natural among a polygynous people.589 It is a customary arrangement in polygynous families that each wife has a hut for herself, where she lives with her children; but even where this is not the case, mother and children naturally keep together as a little sub-family. No wonder, then, if a child takes its name after the mother rather than after the father. This is the simplest way of pointing out the distinction between the issue of different wives, a distinction which is of special importance where it is accompanied by different privileges as to succession. It is worth noticing that, among the Negroes, who are probably the most polygynous race in the world, the female line is extremely prevalent; whereas, among the Hill Tribes of India, who are on the whole, monogamists, children, with few exceptions, take the name of the father. With reference to the Basutos, a Bechuana tribe, Mr. Casalis observes that the authority of the eldest maternal uncle preponderates to excess, especially in polygynous families, where the children have no strong affection for their father.590

Further, among several peoples a man, on marrying, has to quit his home, and go to live with his wife in the house of her father, of whose family he becomes a member. This is a common practice among several of the North American tribes,591 and prevailed, in the southern part of the New World, among the Caribs.592 In some parts of Eastern Central Africa, also, a man who marries a full grown girl “immediately leaves his own village and proceeds to build a house in the village of his wife.”593 Among the Sengirese, according to Dr. Hickson, the man always goes to his wife’s house, unless he be the son of a rajah, in which case he may do as he pleases.594 Dr. Hooker tells us that, among the Khasias, “the husband does not take his wife home, but enters her father’s household, and is entertained there.”595 And in Sumatra, in the mode of marriage called “ambel anak,” the father of a virgin makes choice of some young man for her husband, who is taken into his house to live there in a state between that of a son and that of a debtor.596

According to Dr. Starcke, this custom is due to the great cohesive power of the several families, which causes them to refuse to part with any of their members. “Since men are more independent,” he says, “they are also less stationary; they can no longer attract the women to themselves, and are therefore attracted by them.597 Under such circumstances, there is nothing astonishing in the fact that children are named after the mother’s tribe or clan, which is the case in all the instances just given of peoples among whom the husband has to settle down with his father-in-law. Indeed, Dr. Tylor has found that, whilst the number of coincidences between peoples among whom the husband lives with the wife’s family and peoples among whom the maternal system prevails, is proportionally large, the full maternal system never appears among peoples whose exclusive custom is for the husband to take his wife to his own home.598 And it is a remarkable fact that where both customs—the woman receiving her husband in her own hut, and the man taking his wife to his—occur side by side among the same people, descent in the former cases is traced through the mother, in the latter through the father.599 In Japan, should there be only daughters in the family, a husband is procured for the eldest, who enters his wife’s family, and, at the same time, takes its name.600

Again, as to the rules of succession, Dr. Starcke has set forth the hypothesis that they are dependent on local connections, those persons being each other’s heirs who dwell together in one place. Among the Iroquois, for instance, at the death of a man, his property is divided among his brothers, sisters, and mother’s brothers, whilst the property of a woman is transmitted to her children and sisters, but not to her brothers. “Owing to the faculty of memory,” Dr. Starcke says, “childhood and youth involve a young man in such a web of associations that he afterwards finds it hard to detach himself from them. The man who, when married, has lived as a stranger in the house of another, clings to the impressions of his former home, and his earlier household companions become his heirs. But the brother who has wandered elsewhere stands in a more remote relation to his sister than do the sisters and the children living with her in the parental home, and he is therefore excluded from the inheritance.”601

Though agreeing, in the main, with Dr. Starcke’s hypothesis, I do not think it affords a complete explanation of the matter. It certainly accounts for the fact that, under the maternal system, it is just the nearest relatives on the mother’s side who are a man’s heirs, to the exclusion of other members of the clan. But, if succession really depended upon local relations only, or upon the remembrance of such relations in the past, it would be the most natural arrangement, where father and children lived together till the latter were grown up, for the father to be succeeded by his son. It seems probable that the causes which make children take their mother’s name, have also directly exercised some influence upon the rules of succession; but I am inclined to believe that the power of the name itself has been of the highest importance in that respect.

By means of family names former connections are kept up, and the past is associated with the present. Even we ourselves are generally more disposed to count kin with distant relatives having our own surname than with those having another. And upon man in a savage state language exercises, in this matter, a much greater influence than upon us. With reference to the aborigines of Western Australia, Sir George Grey observes, “Obligations of family names are much stronger than those of blood;” and a “Saurian,” or a “Serpent,” from the East considers himself related to a “Saurian,” or a “Serpent,” from the West, though no such relationship may exist.602 Among the Ossetes, according to Baron von Haxthausen, a man is considered more nearly related to a cousin a hundred times removed, who bears his name, than to his mother’s brother; and he is bound to take blood revenge for the former, while the latter is in fact not regarded as a relative at all.603 Speaking of certain Bantu tribes, Mr. McCall Theal remarks that their aversion to incestuous marriages is so strong, that a man will not marry a girl who belongs to another tribe, if she has the same family name as himself, although the relationship cannot be traced.604 Is it not a justifiable presumption that a similar association of ideas has influenced the rules of succession also,—all the more so, where community of name implies community of worship as well? It should be observed that in every case—at least so far as I know—where rank and property are inherited through females only, children are named after the mother,—but not vice versa, thanks to the direct influence of local and other connections. In China, a man is even strictly forbidden to nominate as his heir an individual of a different surname.605

It is a difficult, sometimes even a hopeless, task to try to find out the origin of savage laws and customs, and I do not pretend to have given an exhaustive explanation of those in question. But it seems to be sufficiently clear, from what has been said, that we have no right to ascribe them to uncertain paternity; nay, that such an assumption is not even probably true. No one has yet exhibited any general coincidence of what we consider moral and immoral habits with the prevalence of the male and female line among existing savages. Among the Barea, for instance, as among the Negroes of Loango, inheritance goes through mothers only, though adultery is said to be extremely rare;606 whilst, on the other hand, among the wanton natives of Tahiti, possessions always descend to the eldest son. With the Todas and Tibetans, among whom paternity is often actually uncertain on account of their polyandrous marriage customs, succession runs through the male line only. “If one or more women,” Mr. Marshall says with reference to the former, “are in common to several men, each husband considers all the children as his—though each woman is mother only to her own—and each male child is an heir to the property of all of the fathers.”607 Among the Reddies, a son—although it often happens that he does not know his real father—is the heir of his mother’s husband.608 And, in India and Ceylon, female kinship is associated with polyandry of the beena type—where the husbands come to live with the wife in or near the house of her birth; and male kinship with that of the deega type—where the wife goes to live in the house and village of her husband.609

Lastly, as Mr. Spencer remarks, avowed recognition of kinship in the female line only, shows by no means an unconsciousness of male kinship. As a proof of this may be adduced the converse custom which the early Romans had of recognizing no legal relationship between children of the same mother and of different fathers. For, if it cannot be supposed that an actual unconsciousness of motherhood was associated with this system, neither is there any adequate warrant for the supposition that actual unconsciousness of fatherhood was associated with the system of “kinship through females only” among savages.610

The prevalence of the female line would not presuppose general promiscuity even if, in some cases, it were dependent on uncertainty as to fathers.611 The separation of husband and wife, adultery on the woman’s side, and the practice of lending wives to visitors occurring very frequently among many savage nations, the proverb which says, “It is a wise child that knows his own father,” holds true for a large number of them. According to Mr. Ingham, the Bakongo, who trace their descent through the mother only, assert as a reason for this custom uncertain paternity; but nevertheless, as we have already seen, they would be horrified at the idea of promiscuous intercourse.

Having now examined all the groups of social phenomena adduced as evidence for the hypothesis of promiscuity, we have found that, in point of fact, they are no evidence. Not one of the customs alleged as relics of an ancient state of indiscriminate cohabitation of the sexes, or “communal marriage,” presupposes the former existence of that state. The numerous facts put forward in support of the hypothesis do not entitle us to assume that promiscuity has ever been the prevailing form of sexual relations among a single people, far less that it has constituted a general stage in the social development of man, and, least of all, that such a stage formed the starting-point of all human history.

It may seem to the reader that this question has received more attention than it deserves. But I have discussed it so fully not only because of the importance of the subject, but because of the insight the customs mentioned give us into sexual and family relations very different from our own, and because the unscientific character of the conclusions we have tested shows most clearly that sociology is still a science in its infancy.

Even now my criticism is not finished. Having shown that the hypothesis of promiscuity has no foundation in fact, I shall endeavour, in the next chapter, to demonstrate that it is opposed to all the correct ideas we are able to form with regard to the early condition of man.