Kinship & marriage in early Arabia,

Front matter

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover.

Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the publisher to a library and finally to you.

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to prevent abuse by commercial parties, including placing technical restrictions on automated querying.

Author's Preface

The object of the present volume is to collect and discuss the available evidence as to the genesis of the system of male kinship, with the corresponding laws of marriage and tribal organisation, which prevailed in Arabia at the time of Mohammed ; the general result is that male kinship had been preceded by kinship through women only, and that all that can still be gathered as to the steps of the social evolution in which the change of kinship law is the central feature corresponds in the most striking manner with the general theory propounded, mainly on the basis of a study of modern rude societies, in the late J. F. McLennan s book on Primitive Marriage. The correspondence of the Arabian facts with this general theory is indeed so close that all the evidence might easily have been disposed under heads borrowed from his exposition ; and for those who are engaged in the comparative study of early institutions this would probably have been the most convenient arrangement. But the views of

XI xu KINSHIP AND MARRIAGE

my lamented friend are not so widely known as they deserve to be, and several of the Essays in which they are expressed are not very accessible- Moreover I wished to speak not only to general students of early society but to all who are interested in old Arabia ; for if my results are sound they have a very important bearing on the most fundamental problems of Arabian history and on the genesis of Islam itself. I have therefore thought it best to attempt to build a self-contained argument on the Arabian facts alone, following a retrogressive order from the known to the unknown past, and not calling in the aid of hypotheses derived from the comparative method until, in working backwards on the Arabian evidence, I came to a point where the facts could not be interpreted without the aid of analogies drawn from other rude societies. This mode of exposition has its disadvantages, the most serious of these being that the changes in the tribal system which went hand in hand with the change in the rule of kinship do not come into view at all till near the close of the argument. In the earlier chapters therefore I am forced to argue on the supposition that a local group was also a stock-group, as it was in the time of the prophet ; while in the two last chapters it appears that this cannot have always been the case. But I trust that the reader, if he looks back upon the earlier chapters after reaching the end of the book, will see that this result has been

AUTHOR'S PREFACE xiii

tacitly kept in view throughout, and that the substance of the argument involves nothing inconsistent with iL

The first chapters of the book do not, I think, borrow any principle from the comparative method which cannot be completely verified by Arabian evidence. These chapters are rewritten and expanded from a course of public University lectures delivered in the Easter Term of the current year, and my original idea was to confine the present volume to the ground which they cover. I found, however, that to break off the argument at this point would be very unsatisfactory both to the author and to the reader, and that, to round off my results even in a provisional way, it was absolutely necessary to say something as to the ultimate origin of the tribal system. And here it is not possible to erect a complete argument on the Arabian evidence alone. But it is, I think, possible to shew that the Arabs once had the system which McLennan has expounded under the name of totemism (chap, vii.), and if, as among other early nations, totemism and female kinship were combined with a law of exc^famy, it is also possible to construct, on the lines laid down in Primitive Marriage^ a hypothetical picture of the development of the social system, consistent with all the Arabian facts, and involving only verct causa ^ t,e., only the action of such forces as can xiv KINSHIP AND MARRIAGE

be shewn to have operated in other rude societies in the very way which the hypothesis requires (chap. viii.). I have thought it right to limit myself, in this part of the subject, to the briefest possible outline. The general principles of the hypothesis, as laid down by J. F. McLennan, are not, I believe, likely to be shaken, but it is premature to attempt more than the most provisional sketch of the way in which they operated under the special historical conditions existing in the Arabian peninsula.

The collection of the evidence on which my arguments rest has occupied me at intervals since the autumn of 1879, when I put together a certain number of facts about female kinship and totemism in a paper on '' Animal worship and animal tribes among the Arabs and in the Old Testament," which was published in the Journal of Philology^ vol. ix. At that time I had access to no good library of Arabic texts, so that I could only pick up what lay on the surface of the unsearched field ; but the results of this provisional exploration appeared so promising that it seemed desirable to publish them and to invite the cooperation of scholars better versed in the early literature of Arabia. Several orientalists of mark responded to this invitation ; in particular Prof. Th. Noldeke sent me some valuable observations, which have since been incorporated in his review of Prof. G. A.

AUTHOR'S PREFACE xv

Wilken's book, Het Matriarchaat bij de oude Arabieren {Oester. Manatschrift f. d. Orient, 1884), and Prof. Ignaz Goldziher contributed a list of important references to the hadtth and other sources in a letter to the Academy, July 10, 1880. The hadtth (traditions of the prophet) was not used at all in my paper, but I had begun to search through it in the winter of 1879-80, when a visit to Cairo enabled me also to procure extracts from fabari's Coran commentary, of which some specimens are given in the notes to the present volume. The next contribution to th^ subject was Prof. Wilken s book, already cited, which appeared at Amsterdam in 1884. Most of the facts on which Prof. Wilken builds are simply copied from my paper and Dr. Goldziher's letter, but he adds a very useful collection of the traditional evidence about mot a marriage, for which he had the assistance of Dr. Snouck Hurgronje. On this topic I had briefly touched in a note to my Prophets of Israel (1882), p. 408 ; but Prof. Wilken was the first to bring it into connection with the rule of female kinship. Another new point to which Prof Wflken devotes considerable attention is the importance attached in ancient and modem Arabia to the relationship of maternal uncle and nephew; and what he has said on this head plays a chief part in the controversy between him and Dr. Redhouse, which has produced the two latest publications on the subject of female kinship xvi KINSHIP AND MARRIAGE

in Arabia (J. W. Redhouse, Notes on Prof. E. B. Tylor's ''Arabian Matriarchate'' [1885]; G. A. Wilken, Eenige Opmerkingen naar anleiding eenercritiek van mijn '' Matriarchaat bij de oude Arabierenl' Tfie Hague 1885). Some points in both these papers are touched on in the following pages, but I have not found occasion to go into the controversy in detail, as my interpretation of the whole evidence differs fundamentally from that of the Dutch scholar. It will be seen from this survey that by much the larger part of the evidence which I have used had to be collected without assistance from any predecessor, and I have not been able to extend my search over more than a moderate part of the vast field of early Arabic literature. On the other hand, while I have tried to give specimens of all the types of evidence that have come under my observation, I could easily have multiplied examples of many of these types.

The notes appended to the volume contain a variety, of illustrative matter, and in some cases take the shape of excursuses on topics of interest which could not have been brought into the text without breaking the flow of the argument^

In conclusion I desire to express my thanks to my friend and colleague Prof. W. Wright for

1 [For the greater convenience of the reader the majority of these notes now appear in their proper place as footnotes.]

AUTHOR'S PREFACE xvii

valuable help in all parts of the book, and to my friend Mr. D. McLennan for many important criticisms and suggestions on the first six chapters.

W. ROBERTSON SMITH.

Chapter I

The Theory of the Genealogists as to the Origin of • Arabic Tribal Groups

PACK

^ The Arabic tribal groups — Theory of the genealogists — Date of the

^ genealogies — Genealogy of Coda a — Method of the genealogists —

Bakr and Taghlib — Meaning of brotherhood and fatherhood —

— Fictitious ancestors — Tombs of ancestors — Tamtm son of Morr

^ — Basis of the genealogical S3rstem — Unity of tribal blood —

— Female eponyms — Bafn 9sAfakhidh I

CHAPTER II The Kindred Group and its Dependents or Allies The Arab kindred group — The kindred group or f^yy — Dependents

of the kindred group — Protector and protected — Adoption of

\ slaves and refugees — Rights and duties of protected allies — The

bkx>d covenant — Antique conception of kinship — Kinship and

property — The tribe and the &mily — Names for kinsmen . 40

Chapter III

The Homogeneity of the Kindred Group in Relation to THE Law of Marriage and Descent

Endogamy and exogamy — Three types of marriage — H4tim and M&wiya — Testimony of Ammianus — moia marriage — Marriage of Omm Kh&rija — Marriage by capture — Capture and contract — The husband as lord — The dowry as a price — Capture and purchase — Marriage by purchase — ** Inheriting women against their will" — Marriage with a step-mother — The law of divorce — Property rights of women — Could women inherit ? — Property of women — The condition of women — The woman and her kinsfolk

The position of women 73

xix

XX KINSHIP AND MARRIAGE

CHAPTER IV Paternity PACK

Fatherhood and its rights — Fatherhood in Old Arabia — ^'Amir ibn $a'9a*a — Step-father as father — Meaning of paternity — Original sense of fatherhood — Fatherhood and kinship— Tibetan polyandry — Conditions for a custom of polyandry — Rise of the custom — Infanticide . 129

CHAPTER V Paternity, Polyandry with Male Kinship, and with Kinship through Women

Evidence of Strabo — Relics of polyandry — Growth of conjugal fidelity — Conjugal fidelity and chastity — Priority of female kinship — Milk brotherhood — Blood and flesh — Ceremony of the *actca — The two systems of kinship— Conflict of the two systems — Decay of tribal feeling 156

CHAPTER VI Female Kinship and Marriage Bars

Forbidden degrees — The tent in marriage — Significance of the tent — Metronymic fiunilies — Licence of the poorer classes — Beena marriages among the Hebrews — Bdal marriage and bdal worship — Residual proUems — Totemism and heterogeneous groups 191

CHAPTER Vn Totemism

Totemism — Relics of totemism — Tribes named from animals — Evidence for Arab totemism — Lion and horsemanship — Sacred doves — ^Jinn — Tribal marks or wasm — Tattooing in religion . 217

CHAPTER Vm Conclusion

Totemism of the Northern Semites — Animal gods in Syria — Social aspect of Totemism — Totemism and heterogeneity — Illustration from Australia — Origin of the tribal system — The Yemenite migration — Disappearance of old nations — Migrations of the Semites — The Northern Semites 252

Additional Notes 283

Index 317

COMPARATIVE TABLE OF PAGINATION nr Ed. Old Ed.

New Ed. Old Ed.

New Ed. Old Ed.

New Ed. Old Ed.

42 = 37

3= 3

43 = 38

3= 3

84 = 68

4= 4

257^-

5= 5. 246

45 = 39

86 = 70

87 = 70*7.

128 = 106

8= Tiq,

48 = 41. 259

129 = 107

9= 8^.

91 = 74

51 = 43^.

13="^.

53 = 45, 260

14=13, 353

134=1"

135 = "2 t8=i6 jy.

58 = 49 ao=i7jy.

31 = 18 ^.

103 = 8557.

64=53. 262*7.

105 = 87, 269

35=33

36=33

108 = 89. 271

30=37

70=57, 266

32 = 28

72 = 58

33=29

73 = 59

34=30. 257

74=60

115 = 94. 273

155 = 130

156 = 131

37 = 32^.

77 = 62. 2671^.

157=133

158 = 133

39=34

40=35

160=135

41 = 36

81 = 651^.

X)

ci

161 = 136 xxu

183=155^- 184=1561^.

315 = 1835^.

336 = 194, 301

New Ed. Old Ed.

358 = 33357., 311

New Ed. Old Ed.

376 = 340

309 = 307

313 = 313^-

315 = 315^-

316 = 316

Chapter I

THE THEORY OF THE GENEALOGISTS AS TO THE ORIGIN OF ARABIC TRIBAL GROUPS

The Arabic tribal groups — ^Theory of the genealogists — Date of the genealogies — Genealogy of Coda'a — Method of the genealogists — Bakr and Taghlib — Meaning of brotherhood and fatherhood — Fictitious ancestors — Tombs of ancestors — Tamtm son of Morr — Basis of the genealogical system — Unity of tribal blood — Female eponyms — Batn and fakhidh.

At the time when Mohammed announced his prophetic mission, and so gave the first impulse to that great movement which in a few years changed the whole face of Arabian society, the Arabs throughout the peninsula formed a multitude of local groups, held together within themselves not by any elaborate political organisation but by a traditional sentiment of unity, which they believed or feigned to be a unity of blood, and by the recognition and exercise of certain mutual obligations and social duties and rights, which united all the members of the same group to one another as against all other groups and their members.

The way of life of these groups was various;

some were pastoral and nomadic, others were engaged in agriculture and settled in villages or towns, and in some towns again, as in Mecca and X^if, a chief occupation of the citizens was trade. This of course implies that some communities were much more advanced in civilisation than others ; the difference between a wild Bedouin and a rich merchant of Mecca was perhaps nearly as great then as it is now. And with this there went also considerable variety of law and social custom ; thus the Traditions of the Prophet and the commentators on the Coran often refer to diversities of 'ada, that is of traditional usage having the force of law, as giving rise to discussion between the Meccans who followed Mohammed to Medina and the old inhabitants of that town. But all through the peninsula thel t3rpe of society was the same, the social and political unit was the group already spoken of.

This is not to be taken as meaning that there was no such thing as a combination of several groups into a larger whole; but such larger combinations were comparatively unstable and easily resolved again into their elements. In the greater towns, for example, several groups might live together in a sort of close alliance, but each group or clan had its own quarter, its little fortalices, its own leaders, and its particular interests. The group-bond was stronger than the bond of citizenship, and feuds between group and group often divided a town against itself. So too among

Chap. I Origin Of Arabic Tribal Groups 3

the nomadic Arabs we find that a certain number of groups might form a confederation presenting the semblance of something like a nation ; but the tendency of each group to stand by its own members in every quarrel was fatal to the permanence of such unions. This was the case not only where the confederation rested on a treaty (casdma) and was limited in scope by the nature of the contract, but also where neighbouring and allied groups regarded themselves as brothers, united by a bond of blood. In such cases, indeed, quarrels were not willingly pushed to an open rupture ; the cooler and wiser heads on both sides were willing to strain a point to keep the peace ; but if the principals in the quarrel proved intractable the outbreak of open hostilities between their respective groups was usually a mere question of time. And then all other considerations disappeared before the paramount obligation that lay on every femily to stand by its own people, that is, by its own ultimate group.

It is the constitution of these ultimate groups, out of which all larger unions were built up, and into which these constantly tended to resolve themselves again, which must form the starting-point of the present enquiry.

According to the theory of the Arab genealogists the groups were all patriarchal tribes, formed, by subdivision of an original stock, on the system of kinship through male descents. A tribe was but a larger family; the tribal name was the name or nickname of the common ancestor. In process of time it broke up into two or more tribes, each embracing the descendants of one of the great ancestor's sons and taking its name from - him. These tribes were again divided and subdivided on the same principle, and so at length that extreme state of division was reached which we find in the peninsula at the time of the prophet. Between a nation, a tribe, a sept or sub-tribe, and a family there is no difference, on this theory, except in size and distance from the common ancestor. As time rolls on the sons of a household become heads of separate families, the families grow into septs, and finally the septs become great tribes or even nations embracing several tribes.

It is proper to observe here that in the earliest times of which we have cognisance the ultimate kindred group, which in the last resort acted together against all other groups, was never a single, family or homestead (dar), and that the group-bond was, for its own purposes, stronger than the family or household bond. Thus, if a man was guilty of homicide within his own group, the act was murder and his nearest relatives did not attempt to protect him from the consequences, but the whole group usually stood by a manslayer who had killed an outsider, even though the slain was of a brother group. In such a case they might recognise that some atonement was necessary, but they interested themselves to make for their kinsman the best terms they could. This observation, it will readily

CHAP. I ORIGIN OF ARABIC TRIBAL GROUPS $

be seen, does not square well with the theory that the kindred group is only the family grown large ; at all events if we accept the theory it appears necessary to supplement it by an explanation of the reason why the blood-bond creates absolute obligations between all the families which form a single group, and only very modified obligations towards children of the common ancestor beyond this limit. On the theory one would expect to find that the family was the real social unit, beyond which the feeling of kinship obligation was never quite absolute, but grew continuously weaker as the degree of kinship was more remote ; whereas we actually find a certain group of families within which kinship obligations are absolute and independent of degrees of cousinship, while beyond this group kinship obligations suddenly become vague. But this is a point on which the genealogists have nothing to say ; they content themselves with offering a scheme of the subdivision of patriarchal tribes by which all Arabs who possess a ntsda or gentile name can trace back their genealogy to one of two ultimate stocks, the Yemenite or S. Arab stock, whose great ancestor is Cah^an, and the Ishmaelite or N. Arab stock, whose ancestor is 'Adnan, a descendant of Abraham through Ishmael. The latter stock bears also indifferently the names of Ma'add or Nizar, the former being represented as the son and the latter as the grandson of 'Adnan.^ Ma'add, indeed, has according to the

1 A convenient view of the whole system, printed in the shape of genealogists a brother 'Akk; but the 'Akk are regarded as having married into Yemen and become Yemenite (B. Hish. p. 6), or even the Yemenites and the 'Akk in Khorasan gave themselves a different descent, deriving their father 'Adnan or 'Odthan from the Asd {ib. 1. i8). Ma'add, again, has four sons (B. Hish. p. 6 ult.), of these the first is Nizar, the second Coda'a (really the first-born, but on him see Additional Note A), the third Conos, who is regarded as extinct (p. 7, 1. 9), and the fourth lyad, who, however, is reckoned also as son of Nizar (so Wustenfeld's tables). Thus the three, 'Adnan, Ma'add, and Nizar are practically identical.

The elaboration of this genealogical scheme falls mainly within the first century of the Flight — though it was hardly completed so early — and is probably connected (as Sprenger has pointed out in his Life of Mohammed) with the system of registers introduced by the Caliph 'Omar I. for the control of the pensions and pay distributed among believers from the spoil of the infidel. The pension system, as Sprenger^ has explained at length, afforded a direct stimulus to genealogical research,

a series of genealogical tables, is to be found in F. Wiistenfeld, Genealogische Tabellen der Arabischen Stdmme und Familien (Gpttingen, 1852). The tables are accompanied by an index volume. Register zu den Tabellen^ which contains a very useful accumulation of traditional material, put together without criticism of the sources, so that a good deal of sifting is necessary.

^ [Das Leben u, Lehre d, Moluzmmad^ iii. pp. cxx sqq, (Berlin, 1 86s).]

CHAP. I ORIGIN OF ARABIC TRIBAL GROUPS 7

► and also, it must be added, to genealogical fiction ;

► while the vast roisters connected with it afforded , the genealogists an opportunity, which certainly

never existed before, to embrace in one scheme the relations of a great circle of Arab kindreds. At the same time, in consequence of the victories of Islam many tribes, or at least large sections of them, migrated to distant lands, where they received estates or were settled in military colonies and frontier stations. The military oi^anisation closely followed the old tribal grouping ; the feuds of the desert were transplanted to Syria and 'Irac, to Spain and Khorasan, and in all the numerous factions and civil wars that rent the old Arab empire tribal alliances and kinship played a conspicuous part. Every ambitious chief therefore was anxious to include as wide a kinship as possible among his dependents and allies, while a weak group found it advantageous to discover some bond of connection with a stronger neighbour. As the old groups were, in the various provinces, shuffled through each other in very various combinations, it plainly became an object of interest to reduce to system the relationships of all the Arab tribes. From time immemorial the population of Arabia had been divided into two great races — the same which the genealogists refer to Cahjan and 'Adnan respectively. In all parts of the empire these two races maintained their ancestral traditions of bitter and persistent feud, and this race-antagonism was a dominating feature in the whole stormy politics of the Omajryad dynasty. In such circumstances the task of the genealogists, who undertook to trace out and reduce to system all the links of kindred connecting the tribes of *Adnan and Cah^an respectively, had a very practical interest ; the questions involved were not mere matters of archaeological curiosity, but had a direct bearing on the political combinations of the time. Scientific impartiality therefore was not to be looked for; even if the genealogist himself was an incorruptible judge — and hardly any Oriental is so — he was certain to have much spurious evidence laid before him.

An example will make this clear, and at the same time shew how uncertain is even the |main structure of the genealogical tree. In the form of the genealogies which ultimately prevailed, *Adnan, NizSr, Ma'add, Ishmaelite Arabs are identical terms and embrace one great nation. All other Arabs are Yemenites or sons of Cahtan, and these again, if we neglect the remote tribes of Hadramaut, may be taken as forming two main groups : (a) the tribes of Kahlan, and (6) the tribes known under the common name of Coda'a, which are traced to Cattt^tn through Himyar, the eponym of the race whom the Greeks and Latins call Homerites. At first sight all this seems to be quite correct and to correspond with the historical fact that under the Omayyads there was a great and enduring hatred between the Caisites, a branch of Nizar or Ma'add, and the Kalbites, a branch of Coda a ; the feud of Caisites and Kalbites seems to be simply a local form of the

CHAP. I ORIGIN OF ARABIC TRIBAL GROUPS 9

feud of Yemen and Ma'add. But when we turn to the Aghanly 777 sq. we find that "the genealogists

r are at variance as to Coda'a, some maintaining him to be a son of Ma'add and brother of Nizar, while

, others make him to be Himyarite." The evidence on each side consists of verses in which Coija'a is referred to Ma'add or to Himyar respectively.

^ The later singers of CodS'a maintained the Himyarite genealogy and made a number of verses to support it ;^ but this, says Moarrij (a noted scholar who died a.h. 195), dates only from the last days of the Omayyads, and all older poets before and

r after Islam refer Coda'a to Ma'add.* And ac-

, cordingly the Aghanl shews that the famous Codaite poet Jamil, of the tribe of Sa'd Hodhaim (died a.h. 82), repeatedly speaks of his race as Maaddite.

It appears then that in this case the genealogy

. that ultimately prevailed was based on a deliberate falsification of old tradition. The motive is explained by the noted genealogist Aba JaTar Mohammed ibn liabib (died a.h. 245), quoted in the Tajy 6461 : " Coda'a was always known as Maaddite till the feud between Kalb and Cais- 'Ailan arose in Syria in the days of Merwan ibn Al-Hakam; then the Kalbites inclined to the Yemenites and claimed kin with Himyar to get their help the more readily against Cais." In point of fact, at the battle of Marj Rahit (a.h. 64)

1 So B. Hish. p. 7, 1. 7. 2 [Cp. Jarir's DiwOn (Cairo), i. 107ia— A. A. B.]

lo KINSHIP AND MARRIAGE chap, i

Merwan's party included besides the Kalbites the Kahlanite tribes of Ghassan, SalcQn, and SakasiL^

What was done on a large scale in the case of ^j Coda'a was doubtless done on a smaller scale in ' other cases. Indeed Hamdani tells us that he found it to be the r^ular practice of obscure desert groups to claim kinship with more famous tribes of the same name {Jeztrat, p. 90). But for j our purpose the point to be noticed is that it still was possible in the later days of the Omayyads to make a radical change in the pedigree of great tribes like the Kalb and other Coija'a. For this shews that the whole system of pedigrees was still in a state of flux, at least as r^arded its remoter members and the connections between distant tribes. The Northern Arabs called themselves Maaddites even before the time of the prophet; but if this term had then conveyed the definite genealogical conceptions that went with it in later times, it would not have been possible to transfer I a series of great tribes from Ma'add to Himyar. Undoubtedly the genealogists found in oral tradition and official registers a large mass of sound information as to the old affinities and subdivisions of tribes, but this material was not sufficient for their task ; it was fragmentary in character and its range was limited by the notorious shortness of the historical memory of the Arabs, To make a complete system out of such materials it was necessary to have constant recourse to conjecture, to force

1 See Additional Note A

CHAP. I ORIGIN OF ARABIC TRIBAL GROUPS ii

a genealogical interpretation on data of the most various kinds, and above all to treat modem IK>litical combinations as the expression of ancient bonds of kinship. The backbone of the system was the pedigree of the prophet — itself one of the most obviously untrustworthy parts of the whole scheme — and round this all the other Northern Arabs were grouped on the principle that every connection, real or imaginary, between two tribes was to be explained by deriving them from a common ancestor, who in turn was brought into the prophet's stemma as brother or cousin of some ascendant of Mohammed. To link all known tribal and gentile names together in this way, and at the same time make the lines connecting historical contemporaries with the common father tolerably equal in length, it was necessary to insert a number of "dummy" ancestors. These were got by doubling known names or using personal names of no tribal significance. The places in which the imaginary names should come in were of course largely arbitrary; it was known what were the actual sub-tribes and septs included in any greater tribe, and all these had to appear by their names as descendants of the father of the tribe, but it was comparatively indifferent whether they should be sons or grandsons, though in a general way it was desirable that the eponyms of the more populous groups should stand nearer the common ancestor. Accordingly when one compares different authorities one finds continual variations in matters of this sort ;