Force and Fission in Northern Somali Lineage Structure by: I.M. lewis

7/29/2019 Force and Fission in Northern Somali Lineage Structure by: I.M. lewis

1/20

Force and Fission in Northern Somali Lineage StructureAuthor(s): I. M. LewisReviewed work(s):Source: American Anthropologist, New Series, Vol. 63, No. 1 (Feb., 1961), pp. 94-112Published by: Wiley on behalf of the American Anthropological AssociationStable URL: http://www.jstor.org/stable/667340 .Accessed: 25/01/2013 15:06

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Wiley and American Anthropological Association are collaborating with JSTOR to digitize, preserve andextend access to American Anthropologist.

http://www.jstor.org

This content downloade d on Fri, 25 Jan 2013 15:07:0 2 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=blackhttp://www.jstor.org/action/showPublisher?publisherCode=anthrohttp://www.jstor.org/stable/667340?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/667340?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=anthrohttp://www.jstor.org/action/showPublisher?publisherCode=black


2/20

Force nd Fission n Northern omaliLineage tructureO

I. M. LEWISUniversity College f Rhodesia nd Nyasaland

I

AS IS well-known, the term "segmentary lineage system" is applied to thosesocieties whose politics are cast in a genealogical idiom and whose domi-

nant political ideology is that political relations are a function of genealogicaldistance. In such communities political relations diverge or converge accord-ing to the genealogical proximity of the persons or groups concerned. Struc-tural relations thus depend upon genealogical position, and groups whichthrough descent from a common ancestor are genealogically equivalent arealso, in principle, politically equal.

As such a community, divided into a system of descent groups, expandsover time all its segments do not necessarily develop in man power or wealth atthe same pace. At every point of ramification in the lineage system naturalgrowth does not result in the production of an array of lineages exactly bal-anced in population, wealth, or in other respects. Thus, two or more segmentsdescended from a common ancestor are not

necessarily equivalentin size or re-

sources. And, in societies where such differences are of direct political signifi-cance, this means that, as lineages develop historically, discrepancies tend toarise between their genealogical position and their actual political power. Theresolution of such discrepancies by genealogical manipulation and by fictionsof clanship (in terms of the dominant political ideology) is a well-known fea-ture of segmentary lineage systems and, indeed, has been taken as a general, ifnot necessary, characteristic of this type of political system (Fortes 1953: 28).

This, however, is not true of the segmentary lineage system of the northern

Somali where inconsistencies between the genealogical positions of groups andtheir actual political power are generally not resolved by genealogical manipu-lation or fictions, but by contractual alliances which leave their genealogies un-altered. In discussing some of the morphological and functional characteristicsof northern Somali lineages in this paper, I argue that this feature is consistentwith the fact that clanship is not the sole idiom of Somali politics. For in north-ern Somali society all political units owe their solidarity to explicit contracts ofgovernment which define the specific legal and political obligations of theirmembers (see Lewis 1959a). Thus, although society as a whole is structuredagnatically, agnation itself is not the sole basis of political unity, and it isthrough contract that the implicit values of agnation are given effect in theconstitution of political groups. And while working generally within the frame-work of agnation, contract enables agnates (and sometimes non-agnates) tounite in corporate political groups sometimes irrespective of their genealogicalproximity.

94

http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/20

[LEWIS] Somali Lineage Structure 95

IIThe northern Somali are a patrilineal Hamitic people living as pastoralists

in central and northern Somalia, part of Harar Province of Ethiopia, and

British and French Somaliland. Their number in these territories is probablyabout a million and a half.2 The desert, or near desert conditions which char-acterize much of this arid region impose a nomadic response in which theSomali pastoralists move over considerable distances annually with their herdsof camels, flocks of sheep and goats, and less commonly cattle. There is littlecultivation except in the better-watered regions in the west of the British Pro-tectorate and Harar Province of Ethiopia where, within the last 60 years or so,pastoral Somali have established mixed farming communities partly engagedin the production of sorghum.

Although corporate and individual interests in wells and market centersand townships set limits to any completely nomadic pattern of movement andlead to some development of attachment to locality, the lineages into whichnorthern Somali society is divided are generally not firmly localized. Ties tolocality are weak, and every lineage does not possess a specific territory. Andin contrast to the transhumant Nuer of the Sudan (Evans-Pritchard 1940) andsemi-nomadic Beduin of Cyrenaica (Evans-Pritchard 1949), the territorial dis-persion of groups does not mirror their genealogical relationships. Moreover,while some wells are subject to specific rights of possessionn, others where wateris abundant are not, and prescriptive rights are not generally asserted to pas-ture which is regarded as a gift of God to mankind in general. Thus, men ofdifferent and often hostile lineages meet and interpenetrate in the pastures, thedistribution of grazing and water being the primary determinant of movement.

The occurrence of good grazing and adequate water varies with the seasonsand from year to year, leading to corresponding variations in grazing move-ments and patterns of settlement. In general, both these vital necessities arescarce, and competition and strife over access to them are acute and a constant

feature of Somali pastoralism. Even under modern administration such dis-putes are frequently resolved by open conflict and the ultimate sanction in theuse of pasturage and water is still the time-honored one of force. Thus, apartfrom the ecological limitations to free movement, the individual pastoralist isalways constrained by the fear that his grazing movements may take him toofar from his agnates to marshal their support when he requires it. And al-though entire lineages do not occupy determinate territories, where men settletemporarily with their stock they tend to cluster together according to theirlineage affiliation. These patterns of lineage grouping in a place of temporarysettlement are particularly marked in time of war when agnates cluster to-gether for mutual safety. Thus, although the members of any lineage are oftenfairly widely dispersed in the pastures within the general area of movement ofa clan, conflict galvanizes lineage solidarity and allows agnates to gather to-gether to defend themselves against their adversaries.

The pastoralists are a warlike people, rifles today to a considerable extentreplacing spears in battle, and all disputes, whether they spring directly from



4/20

American Anthropologist [63, 19616

competition for sparse resources or from other causes, tend ultimately to be re-solved by force of arms. Political status depends primarily upon fighting po-tential and consequently upon male strength, the proverb, "Either be a moun-

tain or attach yourself to one," well expressing the importance of militarysupremacy in Somali society. Thus, unlike the transhumant Nuer (Evans-Pritchard 1940: 211-213), the basis of dominance does not depend upon mys-tical ties to the earth irrespective of effective strength, but upon numericalsuperiority. It is strength (b&oog)which is coveted and which every lineageaspires to achieve.

It is not to be imagined, however, that all injuries sustained lead to im-mediate retaliation in kind, or that disputes cannot be settled except by anactual trial of strength. In general, homicide and other delicts are compound-

able and settled between lineages by payment of compensation, especially whenthe parties concerned are under direct administrative pressure. But in thesecircumstances equally, strength (in terms of man power and wealth) is impor-tant, not only because the strong are in a better position to press for the settle-ment of outstanding debts than the weak, but also because the larger a groupis generally the less its individual members have to contribute in payment ofcompensation. Thus in lineage relations, size exerts a marked effect on theability to retaliate, resist attack, or to pay or exact compensation.

III

Agnation (tol) is the basic principle of Somali social relations. The com-munity as a whole is divided into a vast hierarchy of highly segmented line-

ages, all genealogically connected. The lineage corresponding to the highestlevel of agnatic relationship may conveniently be distinguished as a "clan-family." There are four clan-families in northern Somaliland: the Dir (popula-tion about 200,000); the Isaaq (about 400,000); the Daarood (about 500,000);and the Hawiye (about 400,000).3 But although individual members of these

units have a strong sense of clan-family loyalty, they are too large and un-wieldy to act as corporate units in the traditional political system.4 Theirgenealogical span varies between some 20 and 30 named generations.

Each clan-family is segmented into a vast system of lineages of which thelargest and most clearly defined unit may be distinguished as a "clan," usingthis word in a special sense. These are for all practical purposes the largest effec-tive political units and a clansman counts from 15 to 20 named generations tohis clan eponym. Although they are not strictly localized, clans are generallyvaguely associated with specific areas of movement. Some, but not all, are

traditionally led by a clan-head or "Sultan" (variously styled Suldaan Boqor,Garaad, Ugaas, etc.) and this is the only traditional political office in northernSomali society. It is not, however, a powerful one, and in the clan which rangesin population from 10,000 to 10 times that number, policy is decided by the adhoc councils which all adult men have the right to attend and speak at.

Within the highly segmented clan, the next order of segmentation which itis convenient to distinguish is that of the "primary lineage." The salient fea-



5/20

LEWIS] Somali Lineage Structure 97ture of this unit is that it is generally exogamous, marriage being most fre-quently between primary lineages in an area of agnation where agnatic ties aresupplemented by affinal bonds (although the clan is not endogamous andmarriage between clans is common). But the most clearlydefined politicalnit, corresponding to the axis about which corporate political relations mostfrequently revolve, is the so-called "dia-paying group." This title has beendopted in the British Protectorate because the individual pastoralist'siability n payment and receipt of

blood-wealth (Arabic, diiya; Somali, mag)5iesprimarily at his level of grouping. In act the dia-paying group is a lineage,r ollection of small lineages, some 4 to 8 generations in span, whose membersre ound not only by their closeagnatic ties but also by an explicit treaty orontract

(heer)6 defining the terms of their collective responsibility. And inractice his is the basicjural

nd politicalunit of northern Somali society,anging n male strength from aew hundred to a few thousand members. It sthrough membership of a dia-paying group that a person's jural andpoliticaltatus s primarily defined, and no Somali has political status except throughelonging o such a group. The group has traditionally no formal office ofleadership ssociated with it, but the Administrations have appointed leadinglders f dia-paying groups as headmen (in the British Protectorate generallytyled LocalAuthorities). Inpractice as at other levels of lineage grouping theroup's olicy is controlled by its elders, meeting in ad hoc councils, and thefficialeadman

remains essentially a figurehead representing the group in itselations ith Government and through Government with other groups.nerms of the strong military bias ofSomali society andof the system ofpaying ompensation for injuries, one important criterion of the viability of a

dia-paying roup is its ability to discharge collectively the blood-debts andtherresponsibilities ofits members. Below a male strength of between 200nd

300members, aroup cannot function satisfactorily as aiable jural andoliticalnit. Yet ize is not the sole criterion. The stability or instability of agivenia-paying group is the result of the interaction of many different fac-ors. ther than size, some of the most important of these are the structuralontextf opposition, the geographical location of component segments andmembers,the distribution and trength of leadership within the group; and thealancer lack of balance within a group may be due to the action of all or anyfhese factors and to others. Since I discuss this question more fully in aforthcoming ublication

(Lewis 1961' Chapter 7, Section 8) I do not pursue itfurtherere,heroup owes

its distinctive character to the form of political contract(heer)hich binds its membersand

ranslates intospecific terms their agnaticolidarity. or this is the most frequent

level of corporate legal and politicalsolidarity.Buthe dia-paying unit so defined is not an absolute unit' the in-ividualastoralist's range of effective loyalty extends beyond his dia-payingrouphrough aierarchy of segments to the level of the clan and clan-family.husaccording to

the context of strife, dia-paying groups combine togetherlonghe lines of agnation as larger, more inclusive, jural and political units



6/20

98 A merican Anthropologist [63, 1961

with the aid of wider political contracts at all levels of lineage division (seeLewis 1959a: 290-21).

Thus, despite its distinctiveness as a primary unit of political association,

the dia-paying group is only one level of corporate grouping, and like othersegmentary lineage systems a salient feature of the northern Somali system isits extreme fluidity. An individual belongs to a wide range of agnatic units ofwhich only the more clear-cut and most common levels of action have been dis-tinguished in the foregoing. In keeping with this characteristic, there are noterms in Somali to designate specific levels of grouping. The various termsqabiil, qolo, ilib, and reer, which all denote agnatic segments, are all used rela-tively to refer to lineage-groups at different levels of division. What in one con-text is a qabiil, is in relation to another order of lineage segmentation a qolo;and all are reers, groups defined by agnatic descent (tol) from a common an-cestor by reference to whom they are exactly specified. Thus, for example, theagnatic descendants of a man called Mahammad are known as the ReerMahammad and specified in terms of Mahammad's genealogy. And this man-ner of referring to a lineage by qualifying its eponymous ancestor's name withthe word reer (lineage) is applied to all levels of segmentation. In referring tospecific lineages later in this paper, however, I omit the term reer and use onlythe name and surname of the apical ancestor concerned.

IV

Somali discuss the expansion of lineages and lineage segmentation, as onewould expect, in terms of the birth of sons, for, as they say, "when a son isborn, the agnatic line extends." Each son (or daughter) is given a first nameand takes as surname the first name of the father. So that each generationbears a new name; and it is thus that the genealogies which record agnaticdescent are built up. Thus, for example, a man called Mahammad Hassanbears a son 'Ismaan Mahammad, and a daughter, Khadiija Mahammad. 'Is-

maan in turn marries and begets two sons, Yuusuf and Faarah. These are thenconnected to their great-grandfather (awow) by the agnatic genealogy Yuusufand Faarah, 'Ismaan, Mahammad, Hassan. When Khadiija marries, heragnatic line stops since her children are born to her husband's lineage. Theterm "genealogy" (abtirsiinyo), which Somali equate with the Arabic nasab(which is also employed), is a compound of ab or aabbe (father, agnatic genera-tion, tiri (to count, or reckon), and sii (to give) and means literally "reckoningof agnation." From an early age, every Somali is taught the genealogy whichlinks him to the eponym of his clan-family, and genealogies are proudlycherished and sometimes recorded in writing. To a marked degree a person'srange of genealogical knowledge varies with his age and status. Important clanelders and Sultans usually not only know their own genealogy to the clan-family founder, but also, in keeping with their wider political responsibilities,the total genealogical structure of their own clan and often much of that oftheir clan-family.



7/20

LEWIS] Somali Lineage Structure 99

Wherever an ancestor begets two or more sons, the genealogies bifurcate(kalabayd) providing points of division in subsequent lineage relations. Butnot all men have the same number of sons, nor do their descendants multiply

at the same rate. Natural increase leads to an uneven proliferation within thedescent system and over successive generations lineages do not all expand atthe same pace.

Thus the true paradigm of Somali lineage development is that shown inFig. 1 rather than that in Fig. 2, which illustrates the type of segmentation de-scribed for the Nuer (Evans-Pritchard 1940: 193):

A

| B C |D

| E I I I I I

FIG. 1

F

I|G |H

|I |J |K |L

FIG. 2

This differential process, of which Somali are acutely aware, is recorded inthe genealogies and is expressed by saying that one group has "increased"(batay, from badnan, plenty, abundance, increase) at the expense of another.Of two unequally balanced lineages, the more numerous and stronger is de-scribed asfar'amay orfarabatay (fromfar', progeny, ramification etc.; andJar,finger, and batay, increased) whereas the less numerous and consequentlyweaker group "has not multipled" (ma tarmin, from tar to multiply, and to

help). These terms which are used to distinguish between lineages according totheir numerical strengths refer specifically to growth and ramification and arealso applied to plant growth.

The genealogical equivalents for collateral lineages, one of which counts agreater number of generations to a common apical ancestor, are respectively"short branch" (laangaab) and "long branch" (laandeer). And in fact, al-



8/20

100 American Anthropologist [63, 1961

though these terms refer primarily to a disparity in the number of generationscounted to a common apical ancestor, they are used loosely to distinguish alarge lineage (laandeer) from a smaller collateral branch (laangaab). Indeed

where there is a significant difference in the number of generations counted bytwo lineages to a common eponym those of the "long branch" are usually morenumerous, and correspondingly more powerful: they possess the "decisivestrength" (koogroon).

To take a typical example. The Dulbahante clan who dwell in the east ofthe British Protectorate number some 100,000 men, women, and children, andare divided into 50 dia-paying groups. From the clan genealogy (Fig. 3) whichshows the estimated male strengths of component segments, it will be seen thatthere is a general correspondence between genealogical span and numerical

strength (the genealogy only includes the ancestors of extant lineages). Fromthe clan eponym Si'iid Harti (nicknamed "Dulbahante") descend four line-ages, the Muuse Si'iid (A1), Ahmad Si'iid (A2), Malbammad Si'iid (A3), andthe Yuunis Si'iid (A4). Of these the most important are the Muuse Si'iid whomake up the majority of the present Dulbahante clan, which is highly seg-mented into numerous lineages of every order of division. The Abmad Si'iid("Hayaag") are a primary lineage with a male strength of about 1,000, and theremaining Mahammad and Yuunis Si'iid are small, insignificant, and incapableof independent political action. Living members of these small segments tracedescent from the clan founder through only a few generations. Because it islarger, the Hayaag lineage has a longer genealogy and its members trace de-scent to "Dulbahante" through some 8 generations. In the next generation,that of Muuse Si'iid's sons, 'Abdalle's (B4) three brothers Abokor (B3),Mahammad (B2), and Barre (B1) are the ancestors of small lineages whosemembers count only a few generations to the clan eponym. Thus, for example,a man of the Mahammad Muuse lineage counts only 7 ancestors to Dulba-hante.

The sameprocess

ofdifferentiation occurs at every level of generation allthrough the clan genealogy. The point at which there is greatest proliferation

is at the generation descended from Shirshoorre Habarwaa (D2), where hissons, especially Faaralt and Mah. muud, are the founders of highly populatedand widely segmented lineages. Indeed Faarah (El) and Mahamuud Shir-shoorre E2) (commonly known as Faarah and Malhamuud Garaad) have givenrise to two great congeries of lineages which might be distinguished as "sub-clans." Following the segmentation down, it will be seen that the "longbranches" of the Faaral Garaad are the Hagar Aadan (HI) and the 'Ali Geri

(G4) of which only the segmentation of the former is shown in any detail. The"long branches" of the Mal.amuud Garaad are the Jaama' Siyaad (G6) (whosefurther egmentation is not shown) and the Naalleeye Alhmad K10). Membersofthese lineages count from 15 to 20 generations to the clan eponym, whereasthose of the "short branch" lineages count at every generation a lesser numberofancestors to the clan founder. Thus while a man of the Barre Muuse (B1)reaches he clan eponym after 6 or 7 generations, a man of the Faarah Hagar



9/20


(I1)counts back between

15and

18 generations to the same point. Thus it wille evident that at every generation through which descent is traced from thegroup founder ramification and segmentation are coordinate with numericalstrength. The larger agroup the more highly it is segmented,and generally, thelonger ts genealogy. This correlation is found in every Somali lineage and in

every genealogy.his pattern of genealogical expansion where size and genealogical spangenerally go together raises an immediate problem to which I wish to referonly briefly here. The continued existence of collateral lineages, whether theyre f similar strengths or not, requires reproduction through an approximatelyimilar number of generations. If two lineages, one large and the other small,race descent from a ommon apical ancestor it is reasonable to assume thatthey must have multiplied, thoughunequally, over approximately the sameumber f generations. In the one case, at each generation only one son wasborn, r in aew cases two or three, while in the other ateach generation sev-ral ons were born to become the ancestors of new subsidiary segments. Yet,s as been seen, "short branch" lineages have shorter genealogies than "longranch" ollaterals. Thus it appears that genealogies which represent merelyheontinuance of a line and do not give rise to new points of growth are fore-shortened r telescoped, and by a kind of structural amnesia, names which are

not ignificant points of segmentation drop out of the genealogical record.Somali, fcourse, strongly repudiate this suggestion, asking how a person canforget he names of his forefathers which as a child he has been taught to re-ember nd respect. Yet if elision did not occur it would be difficult to under-tand ow this correlation between strength and genealogical span could arise.ndhe interpretation proposed here is supported by the fact that there issometimes onfusion or uncertainty in the upper reaches of genealogies, par-

ticularlyn the case of small lineages.husit seems that in the upper portions of genealogies names which areot

structurally significant tend tobe forgotten(cf.Evans-Pritchard 1940:198-00),nd when one asks the names of sons of a clan founder, for example,hoseiven first are almost invariably the eponyms of important lineages.therons whose descendants are

few, or whose lines have disappeared en-tirely,re only dimly recalled by clan elders and Sultans expert in the historyfheir clan. They arenot common knowledge. In explaining their relation-shipsnd giving their

genealogies, moreover, Somali usually follow throughheain lines of descent first, concentrating upon the "long branches" at theexpensef subsidiary "short branch" descent lines. Thus, usually it is the an-estorhose descendants are most numerous, andwho is therefore most im-portanttructurally, rather than the first-born('urad) whose name is men-

tionedfirst.romhat has been said it can be concluded that northern Somali gene-logiesell a quasi-historical record in which adjustments aremade accordingoifferential rowth. At the same time, the population is increasing7 and iteemshat the genealogical systemasa whole is also expanding in step with



10/20


population growth,8 descent lines being foreshortened or telescoped in keepingwith uneven development.

V

Ihave already emphasized the primacy of force and the consequent impor-tance of strength in Somali politics. The members of a "long branch" lineagewith their superior fighting potential (at least in terms of numbers) arecounted "aristocrats" (gob) as opposed to those of weaker lineages who are de-rided n private, at least, as gun (lit. "bottom"). And Somali are acutely sensi-tive to political prestige and status, although in their acephalous society this isnot a matter of formal political offices but of the size and power of lineages.Knowledge of lineage strength is handed down from generation to generation

and revised in the light of new conditions. The strength of different groups isassessed partly from gossip and direct observation in pastoral movements andis also judged from the manner in which groups behave. Here an importantsource of knowledge is the settlement of political disputes involving compensa-tion where the size of groups is a factor affecting the proportions in which com-

pensation is paid. It may be added that this, with contractual treaties lodgedin district offices, is also a source on which government officials lean heavilyin their estimates of population. Thus there is informed knowledge amongstSomali of the size of different lineages and constant preoccupation with lineagestrength and power.

And even if today every lineage dispute does not lead to a direct test of

military supremacy, numerical strength is still the primary factor in the evalu-ation of political status. If a dispute is resolved by payment of compensationinstead of by open conflict, compensation is generally collected on a basis ofmale strength, each lineage contributing and receiving according to its male

population. Somali, in fact, assume generally (though not invariably) that

liability or concern is proportionate to male strength, at least as far as payment

or receipt of compensation is concerned; and the procedure most commonlyfollowed in settling lineage compensation is that known as "penis-counting"(qooratiris) where the financial responsibility of a group is reckoned accordingto the number of males it boasts, a tally which includes infants as well asadults. Moreover, weak groups tend to settle their outstanding debts to

stronger groups with greater alacrity than the latter do when the situation isreversed.

At the same time, strength also to some extent influences the actual amountof damages exacted for an injury, despite the fact that there is a regular tariff

in which different injuries and wrongs are covered by standard amounts ofcompensation. The following is a typical illustration. Two men of two lineages

(A and B) quarrelled over precedence in watering their camels at a well. Oneof the disputants (of lineage A) was killed and camels were seized by the

government from lineage B which happened to be the stronger of the two. Onehundred beasts were handed over to the bereaved lineage A. A few days later,however, the brother of the deceased, dissatisfied with this settlement and



11/20

LEwIS] Somali Lineage Structure 103spurred on by the desire for revenge, shot a man of lineage B. The elders of thislineage promptly went in strength to the elders of lineage A, demanding dam-ages which the latter agreed to meet. A offered B 200 camels, a rifle, and a girlho was aclose agnate of the man of their lineage whose murder

had provokedhe feud. A hundred camels with the rifle and the girl were taken by the brotherof he deceased, and the remaining 100beasts divided amongst the other mem-bers of lineage B. Lineage B, whose numerical strength is several times that ofA,was entitled to claim higher damages than the statutory amount, since thekilling was n revenge for an injury which had already been compounded. Buthegenerosity of A and the speed with which they paid compensation weregenerally ttributed to their fear of reprisal from a stronger group.But

for the presence of an impartial administration, there is nothing toake he strongfulfill

their debts in bloodto the weak and this is recognized inroverb which Somali readily quote: "The weak have found the European asprotector."

Thus, strength in terms of man power is the criterion of politicalpower ather than lineage seniority by order of birth.his ssumption which is basic in Somali political philosophy makes line-ages hich are genealogically equivalent through descent from a common an-estorpolitically disproportionate and unbalanced if they are not of similarstrengths. or, although in principle, political relations are a function of gene-alogical istance and people act towards each other according to "the numberf

enerations they count apart" (immisa awow yay kala tirsinayaan) or"where heir genealogies come together," what I propose to call the "size fac-or"upervenesto alter the political relations of groups, often in contrast toheirtrict genealogical position. Thus in Fig. 1,which represents the true pat-

tern f Somali lineage segmentation, for all practical purposes the people oflineage regard themselves asanit opposed not to the whole of B, but to itsnumerically (and therefore structurally) equivalent segment E. Similarly, themembers f lineage Eee themselves as a unit in opposition to lineage C, and

soon. Thisan be seen in the genealogy of the Dulbahante clan (Fig. 3). Withinhelan, the Hayaag lineage (A2) s genealogically equivalent to the whole ofheast Muuse Si'iid

(A1) system of lineages, but their strengths are in theatiof1o 40. In act, the Hayaag act ithin the clan not as aegment struc-turallyquivalent to the Muuse Si'iid, but as equivalent to subdivisions ofimilartrength such as the Khaalid

(D1), BahArarsame (F1), and Wa'aysadanH2). Externally they act jointly with the other segments as the Dulba-antelan. Equally, further down the clan genealogy within the Faarah.araadongeries of lineages, the Barkad(F2)

andBahArarsame (F1) do nottandpolitically opposed to all the descendants

of l:mad Faaral~ (F3), al-houghhis is what their genealogical position implies in terms of segmentaryineagerinciples.Inactual fact they are structurally equivalent to constituentsegmentsf the

Almad Faara4, such as the'Ali Geri (G4), Hagar adan (Hi)andther like primary lineages. Thus, groups of similar size are politicallyquivalentt different levels and points of segmentation in the genealogy.



12/20

104 American Anthropologist_r4l

[63, 1961

o

O ;r

r

EL3 C

I.IJ I-"


13/20


The Hayaag (A2) are equivalent to the Naalleeye Ahmad (K10), to the Khayr'Abdi (F6), and to others, and each of these to other groups of similar strengthin different genealogical positions. And while a quarrel between, for example,men of the Barkad

(F2)and 'Ali Geri

(G4)is

likelyto

developinto a feud be-

tween the two lineages this does not mean that all the Ahmad Faarah (F3) ne-cessarily unite with the 'Ali Geri against the Barkad.

Here the "size factor" intervenes and interrupts segmentation according tostrict lineal (and genealogical) principles. Structurally equivalent groups ofcomparable size and strength oppose each other to some extent independentlyof their genealogical positions. Genealogical proximity is not always the chiefdeterminant of political unity, except where collaterally defined segments areevenly balanced. Thus, again in the Dulbahante clan, to a certain extent all

the Faarah Garaad (El) stand opposed to their approximately numericallyequivalent collateral, the Malhamuud Garaad (E2), and a fight between theircomponent segments may spread to a general cleavage between these twogreat lineages. Here, generally, the more distant groups are genealogically, thegreater is the likelihood of extended strife between the parent groups of whichthey are subdivisions. Thus, hostility between one lineage of the Dulbahanteclan and one lineage of the rival Habar Tol Ja'lo clan, while initially involvingonly the two lineages directly concerned, is more likely to lead to fighting be-tween the two parent clans than is strife between one segment of the Faara.Garaad and one of the Mahamuud Garaad to a general war between theselineages.

Thus, although agnation is fundamental to the structure of Somali polit-ical relations, and although there is an explicit ideology that political solidar-ity and closeness in agnatic kinship go together, the weight which is attachedto fighting potential makes genealogically equivalent lineages politically equalonly when they are of similar strengths. To this extent actual power politicsdiverge from strict genealogical principles.

VIThis disparity between the genealogically defined position of a group and

its actual political status in terms of man power, wealth, etc., is not of coursepeculiar to the Somali and has been observed in many lineage societies. It ismost commonly adjusted or resolved by genealogical manipulations andthrough the forging of fictitious clanship links. This process has been well de-scribed among the Gusii (Mayer 1949:33), the Tiv (Bohannan 1952:309), andthe Amba (Winter 1956:211-226). It has also been indicated for the Yoruba

(Lloyd 1955:244) and has been taken as a general attribute of segmentarylineage societies (Fortes 1953:28). Amongst the Luo, on the other hand, thediscrepancy between the genealogical position of a lineage and its actual politi-cal status is not generally resolved by manipulation. Instead, the genealogicalparadigm is regarded as an ideal from which it is legitimate that customarypractice should depart (Southall 1952:32).

Amongst the northern pastoral Somali the problem is not viewed in thislight, and although genealogical fictions do occur in special circumstances-to



14/20

American Anthropologist [63, 196106

be mentioned presently-they are generally rare. Instead extra-lineage ties areemployed to enable unequally balanced lineages to achieve political parity andgenealogical rearrangements are usually unnecessary. Thus, when describing

their lineage divisions, Somali distinguish explicitly between the genealogicalposition of a group and its structural position in relation to other lineages withwhich it acts corporately. This distinction between the genealogical (andquasi- or part historical) order and the order in which lineages are grouped intheir political relations at any point in time is usually scrupulously observed.For, inasmuch as they record the agnatic relationships of individuals andgroups, genealogies provide the framework of politics upon which actual politi-cal relations are a variation in which account is taken of the actual strengths ofgroups.

The commonest type of alliance invoked to redress an uneven balance ofpower among collateral lineages depends upon the recognition of uterine ties.This is modelled on the divisions which occur within the polygynous familywhere siblings are grouped according to uterine descent within agnation.Fortes (1953:33) refers to this process as "complementary filiation" and re-gards it as the "principal mechanism by which segmentation in the lineage isbrought about." In the Somali system, however, it generally plays a supple-mentary role to agnatic segmentation and is particularly a means wherebyweak

groupsunite with other weak

segments againststronger collaterals. The

paradigm is illustrated in Fig. 4 where B, C, and D are weak collaterals with a

joint strength approximately equivalent to A against whom they form a uterinealliance on the basis of real or putative uterine connection.

x

A |B C |D

FIG. 4

Such uterine groupings are usually denoted by the word bah (indicatinguterine generation) or habr (lit. old woman), prefixed to the name of the com-mon mother or to the name of her lineage. And sometimes when there is notrue uterine connection the prefix is added to the name of the dominant lineagein the consortium. Thus, to take a typical example from the Dulbahante: at

the level of segmentation represented by 'Igaal Faaralb (J2), 8 collateral butunequally blanced lineages descend from his 8 sons as shown:

'Igaal Faarab

I I I I l I lJaama' Mahamuud Ismaa'iil Iirsi Mabammad 'Arraale Abmad Naalleeye

Bah Hawiye Bah Ugaad



15/20

LEWIS] Somati Lineage Structure 107The first three lineages born of the same mother (of the Hawiye clan-family)orm a single dia-paying group called the Bah Hawiye with a male strength ofbout 400. The remaining 5 lineages descended from another wife of 'Igaal arenown collectively as the BahUgaad

(male populationabout 1,500) and areivided into 3separate dia-paying groups: the Hirsi Igaal (400), the Maham-ad Igaal (600), and the 'Arraale 'Igaal with the Haad Faarab (together 420).he Ahmad and Naalleeye 'Igaal together boast no more than about 80 mennd act jointly with the Bah Ugaad without being directly attached to any

subdivision within it.Similarly at a higher level of division within the Dulbahante clan, theH.assan Ugaas (E5), H.aamud Ugaas (E6), Mahamuud Ugaas E7), and Hus-eenUgaas (E8) lineages (of which the first is the strongest) are joined in aoosepolitical alliance,

putatively on a uterine basis, as the Bah Ugaas andopposed s a unit to each of their more powerful collaterals, the Faaralb ndMahamuud Garaad. nd in this they are often supported by other segmentsho tand outside the Faarah nd Mahamuud Garaad, uch as the KhayrAbdi F6), Hayaag A2), Khaalid D1), and Yahye (C1), and by the Reeriqi hinni who are a lineage of Hawiye rigin living amongst the Dulbahantebut ot genealogically identified with them. Thus, in 1949, he Hassan Ugaas,ahamuud Ugaas, I:usseen Ugaas, ndH. aamud Ugaas, nd the Khayr 'Abdi,Khaalid, ahye, Hayaag, nd Fiqi Shinni petitioned the local district com-missioner sking the government to recognize them as a political entity ofequivalent tatus to the Faarah and Mahamuud Garaad. nd they asked to begranted salaried leader of similar standing to the two Sultans (Garaads) wholeadhe Faarah and Mahamuud Garaad ineages.uchterine alliances which often unite groups against the grain of strictgenealogical roximity occur at all levels of segmentation, and are as much aartf the system as is agnatic segmentation. Ideally they are forged whereomeort of uterine relationship exists, even if it is only a common agnatic re-lationship etween the

wivesof the men from

whom the lineages concerneddescend. Butmany political unions are struck within agnation and even some-imesutside it without this justification and to aorresponding degree with-outonor. These are known generally as

gaashaanbuur, aerm which appearsliterallyo mean "pile of shields" and which expresses clearly the fundamentalimf the alliance to secure fighting potential and political power. Unliketerinelliance, this is strictly outside lineage morphology and is merely aoliticalregrouping of unequally balanced lineages. And, as in the case of trueterinenions, genealogical fictions are not normally invoked, except some-imeshen the allianceincludes agnatically unrelated strangers. Most fre-quently,owever, these pacts occur within the field of agnation and the gene-logiesre not manipulated. The

Ma.amuud Barkad dia-paying group of theDulbahante lan affords a good example. Its internal structure in 1957 shownn the genealogy (Fig. 5). Within the Mahamuud Barkad ia-payingroup,hese segments are arranged in four politicaldivisions' (1)Igaalaham ud; 2)

'Iise Mahamuud and Beddir, Subaan, and M uuse Shirwa';



16/20


(3) Faaral~, Mablammad, Yuusuf, Maah, and Diiriye 'Ali and Faaral~ Naal-leeye; (4) Amar Barkad, Wiggid Mahamuud, Kooshin Naalleeye, and Guuleedand [-irsi 'Ali. These units, which are at the moment internal segments of one

dia-paying group, were formerly separate dia-paying groups, and their presentunity represents the threat of external pressure, in particular bitterness be-tween the Barkad and 'Ali Geri. It will be seen that these lineages are groupedin approximately equivalent units whose structure does not correspond exactlyto genealogical principles. And, except in the case of the Baha 'Ali, they are allgaashaanbuur alliances where no uterine connection is claimed or postulated.

Barkad (F2)

Mabamuud Amar 'Ali(30)

'Igaal 'Ali 'Iise Wiggid(400) | (100) (50)

Naalleeye

I I IShirwa' Kooshin Faarab

(60) (20)

'Ali Beddir Subaan Muuse(130) (80) (80)

Faarab Mahammad Yuusuf Maah Diiriye Guuleed Hirsi\ o/S / (600) (40)

Baha 'Ali(400)

FIG. 5

Such alliances, as has been said, are sometimes struck outside agnation andare the means by which small scattered minorities living as foreigners amongststronger lineages achieve political viability. Such protected minorities are re-garded as clients (generally called magan) and form dia-paying pacts with theirstronger hosts. But even here there is usually no question of their genealogicalincorporation in their hosts' genealogies. Generally they remain foreign accre-tions allied with their numerically, and therefore politically, dominant hosts.The nature of the client-protector relationship varies with the particular cir-cumstances and a wide range of different degrees of symbiosis occur. But it isgenerally only in the case of minority communities whose agnates no longer

enjoy an independent political existence in any part of northern Somalilandthat the process of alliance is taken to the point of full genealogical absorption.This occurs also in the case of a few small lineages who have so often alliedwith clans to which they are not agnatically related that they have lost theirown genealogical traditions. The members of such groups claim the lineageaffiliation of those to whom they are currently attached and are generallyscorned for their lack of lineage pride. They are known as "pretenders"



17/20


(sheegad), a status which is only assumed in extremities and betrays weaknessand insufficiency. Cases of this kind do occur, but in general, a striking featureof the northern Somali system-at least in the British Protectorate-is the

number of small segments of strong lineages which live as foreign minoritycommunities among stronger hosts to whom they may be politically allied butwith whom they are not genealogically identified. Every lineage, howeversmall, seeks as far as possible to maintain its independent identity and fulladoption is the last resource of the defenseless.

VII

These alliances which the importance of fighting potential necessitates aremade possible by the way in which Somali employ contractual unity (}eer) as

a principle of grouping distinct from and supplementary to agnation (tol).Political contract is the basis of all political unity, whether between groupsacting in accordance with their genealogical positions or in spite of them. Com-pared with other segmentary lineage systems where clanship is the all-per-vasive principle of political union, the use of uterine and purely contractualpacts within and also outside agnation in the northern Somali system appearsto leave little need for genealogical fictions. For agnation is not the sole prin-ciple of grouping in Somali politics. And the northern Somali system is char-acterized by the interaction of contract and agnation. This is seen most clearlyin the structure of the dia-paying group which is first and foremost an agnaticgroup, but whose political charter is expressed in terms of an explicit contrac-tual agreement. And more extensive degrees of agnatic relationship are mobi-lized through wider contracts which define new and more embracing spheres ofpolitical unity (cf. Lewis 1959a: 290-291).

It has already been observed that northern Somali lineages are essentiallynot land-based, and are not sharply defined territorial units. In this respectagain the northern Somali organization differs from those other segmentary

lineage societies which have been described where ecological circumstancesdictate a less fluid and more clearly defined pattern of land relationships thanthat prevailing in northern Somaliland. In contrast to these it might be arguedthat in Somaliland where local contiguity is weak, contract replaces it as apolitical principle of fundamental importance. For political unity is defined bycontract. Agnation alone is insufficient as a basis for corporate political action.

Again in relation to other segmentary lineage societies, it will be evidentthat to regard Somali genealogies as mere "charters" (cf. Bohannan 1952)would not be entirely appropriate. For the genealogies do not mirror exactly

the actual pattern of political relations at any point in time. Moreover, thelogic of Somali clanship is not so much that people act together and must there-fore be agnates, but rather that because they are agnatically related they co-operate in defined activities. And although, as I have suggested, their genealo-gies appear to be foreshortened, and in this sense convey a distorted record,they nevertheless preserve an important historical content (cf. the situationamong the Pathans, Barth 1959: 7). And the agnatic generation recorded in the



18/20

American Anthropoiogist10

genealogies which every child learns by heart and which ideally prescribe thelimits of his political affiliation, is, as it were, the main theme of Somali politicsupon which actual political relations are a variation.

Yet there is a sense in which Somali genealogies may legitimately be re-garded as a charter, and this takes us outside politics into the sphere of re-

ligion. As devout Muslims, all Somali clans and clan-families ultimately tracedescent from the lineage of the Prophet Mahammad and his "Companions."Such claims, although supported by many circumstantial local traditions and

hagiologies (some of which have been published in Arabic), have not beensubstantiated by any external evidence and appear to be generally fictitious.But whatever their historical content, their importance lies in the fact that

they validate the whole Muslim basis of Somali society.9

NOTES

1This paper s based on research arried ut in Somaliland, mainly n the British Protectorate,between 1955 and 1957, and financed by the Colonial Social Science Research Council, London,whose generosity acknowledge with gratitude. am grateful o J. C. Mitchell and J. Van Velsenfor comments n earlier drafts of this article.

2The total Somali population ncluding hat of southern Somalia and the Northern Provinceof Kenya is between wo and a half and three million. The southern Somali, nhabiting he rela-tively fertile region between he Juba and Shebelle rivers where here is a considerable evelop-ment of cultivation, have a considerably different ocial structure rom their northern pastoralkinsmen. See Lewis 1957:37-41, or a tentative description.

3 The remaining lan-families re the Digil and Rahanwiin f southern Somalia who numbertogether some 400,000 souls. And there are also considerable umbers of Daarood, Hawiye, andDir in southern Somalia, especially n the region ying south of the Juba River. The six clan-families nto which he Somali nation s thus divided are gealogically onnected nd can be placedin a national genealogy n which the primary division s that between he Digil and Rahanwiinon the one hand and the other clan-families-the Somali proper on the other. The first aremainly cultivators who have been nfluenced y contact with the Galla and pre-Hamitic Negroidinhabitants of southern Somalia, and the second who regard hemselves as a conquering ristoc-racy are predominantly astorial nomads.

4Although he clan-families o not act as corporate olitical units in the traditional pastoralpolitical ystem, hey do act corporately n the structure f modern arty politics see Lewis 1958).

5The standard ate of compensation s 100 camels (each valued at 67 shillings 50 cents EastAfrican) or a man's ife, and half that amount n the case of a woman. There is also a statutorytariff of damages or different ategories of injury based on the Shafi'ite School of Muslim Lawwhich Somali generally ollow. To some extent these standard valuations are modified by localagreements nd by the social context n which njury s suffered, nd by the political tatus of theparties concerned. The collective solidarity of the members of a dia-paying group s usually re-flected n a lower scale of damages applicable o internal homicides.

6These dia-paying reaties which ay down the precise details of collective responsibility ndenumerate what actions are regarded s wrongs within the group, etting forth a scale of damages,are today recorded n writing n Arabic or English and lodged n district offices. Since the ad-ministrations egard hese contracts as defining he legal responsibilities f the individual, heybecome a source of law in the modern administrative ituation. For examples of beer treaties seeLewis 1959a: 86-289.

7 That the population s increasing t the present time is indicated by a Net ReproductionRate of 1.3 calculated rom igures iven by A. M. Morgantini 1954) or the municipal opulationsof Somalia. This result was obtained by applying he formula NRR=-0.3536+0.001996 FR), tocorrelate he Fertility Ratio with the Net Reproduction Rate. This numerical esult can, however,

[63, 1061



19/20

LEWIS] Somali Lineage Structure Illonly serve as a general ndication hat

the population s increasing ince the population tatisticson which the calculation s based refer only to municipalities which are probably not fully repre-entative of the total (rural and urban) population. I am indebted to Professor J. C. Mitchellfor this analysis.lthough umerical data are lacking, he wholehistory

ofthe Somali, which can be recon-tructed as far back as the 10th entury at least in general outline, s one of an ever-increasingopulation, xpanding n waves from he northern hores of what is today the British Protectorate

and northern Somalia nto the center and south.There s no detailed genealogical vidence o enable comparisons of genealogical span to bemade atdifferent periods n Somali history, but a 16th century chronicle Shihab ad-Din 1897-1909) escribes Somali ineages hen n amanner which suggests hat this interpretation s correct.For t is clear rom Shihab ad-Din's account hat specific roups which are today congeries f clanswverehen clans or smaller units.Cf. Lewis 1955/56. tmay be remarked ere that contrary o Dr. L. ohannan's uggestion,(Bohannan 952:314) the recording of genealogies n writing does not prevent genealogical

anipulation r genealogical ictions. And the act that northern Somali genealogies re some-times reserved n Arabic has nothing o do with the small part fictitious clanship plays amongsthe astoralists. Genealogies re just as frequently written n Arabic by the southern ultivatingomalimongst whom genealogical ictions are extremely ommon; nd the spurious enealogicalraditions hich flourish in semi-literate Muslim Arab society are well known and by nomeansunique.

REFERENCES ITEDBACON, E. E

1958 Obok: a study of social structure n Eurasia. New York, iking Fund ublicationsinAnthropology 25.ARTH, .

1959 Segmentary pposition and the theory of games: a study of Pathan organisation.Journalf the Royal Anthropological nstitute 89:5-21.

BOHANNAN, .1952

A genealogical harter. Africa22:301-315.

EVAS-PRITC ARD,. E.940 heNuer.

Oxford University Press.1949 he

Sanusiof

Cyrenaica.Oxford University Press.

1FORDE,.938Fission an

dccretion

n thepatrilinea llans f a sem i-Bantu ommunity nouthernNigeria.Journal

f the Royal Anthropological nstitute 68:311-338.FORTES,EYER

1945he dynamics of

clanship mong he Tallensi. Oxford University Press.1953he structure of unilineal escent groups. American Anthropologist 55:17-41.

FORTES,MEYER and E. E. EVANS-PRITCHARD940frican political systems. Oxford University Press.

FRIED,MORTON1957

The classification f corporate nilineal descent groups. ournal f the Royal Anthro-pologicalnstitute 87:1-31.

LEws,I. M.1955-56 SufismnSomaliland: study in tribal Islam. Bulletin of the School of Orientalandfrican tudies 27:581-602; 28:146-160.

1957

The Somali lineage system and the total genealogy. London, Hargeisa (BritishSomaliland)nd Crown Agents for he Colonies.

1958Modern political movements n Somaliland. nternational African nstitute Memo-

randum0, Oxford University Press.1959a

Clanship nd contract n northern Somaliland. Africa 29'274-293.



20/20

112 American Anthropologist [63, 1961]

1959b The classification f African political systems. Human Problems n British CentralAfrica 25:59-69.

1961 A pastoral democracy. A study in pastoralism and politics among the northernSomali of the Horn of Africa. (forthcoming)

LLOYD, . C.1955 The Yoruba ineage. Africa 25:235-252.

MAYER, .1949 The lineage principle n Gusii society. International African nstitute Memorandum

24, Oxford University Press.MIDDLETON,OHN nd D. TAIT

1958 Tribes without rulers. London, Routledge & Paul.MORGANTINI, . M.

1954 Quelques esultats preliminaires e l'enquete ur la population de la Somalie meneeen 1953. In Proceedings f the World Population Conference, 954, vol. 6. New

York.SHIHAB D-DIN1897-1909 Futub al-Habasha. R. Basset, ed. and trans. Paris.

SMITH, . G.1956 On segmentary ineage systems. Journal of the Royal Anthropological nstitute

86:39-81.SOUTHALL, A.

1952 Lineage ormation among the Luo. International African nstitute Memorandum26, Oxford University Press.

WINTER, E. H.1956 Bwamba: a structural-functional nalysis of a patrilineal society. Cambridge,

W. Heffer.

Documents

Force and Fission in Northern Somali Lineage Structure by: I.M. lewis