The Origins of the Galla and Somali Author(s): Herbert S. Lewis Source: The Journal of African History, Vol. 7, No. 1 (1966), pp. 27-46 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/179457 Accessed: 10/08/2010 08:56 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=cup. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. 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 forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to The Journal of African History. http://www.jstor.org Journal of African History, VII, I (1966), pp. 27-46 27 Printed in Great Britain THE ORIGINS OF THE GALLA AND SOMALI BY HERBERT S. LEWIS THIS study presentsa reconstructionof the originsand majormovements of the Galla and Somali of NortheastAfrica which departsfrom most of the previousliteratureon the subject. The traditionalview has been that the Gallaoccupiedmost of the Horn of Africauntil the Somali,beginning about the tenth century, swept south and south-west from the shores of the Gulf of Aden driving the Galla before them.1 The pressure of the Somali has also been consideredthe majorimpetus to the Galla invasions of Ethiopiain the sixteenthcentury. It is the thesis of this paperthat both the Gallaand the Somali originatedin southernEthiopia,that the Somali expandedto the east and north much earlierthan the Galla, and that the Galla lived only in southern Ethiopia and northern Kenya until their migrationsbegan about I530. These hypotheseswere presentedin brief form in 1962 and have since been strengthenedby Fleming'sworkon the comparativelinguisticsof the Somaliand by Haberland'srathersimilarconclusionsabout Gallahistory.2 The traditionalreconstructionhas a considerablecorpus of supporting literature,however, and thereforeI shall criticallyexaminethis literature as well as present further evidence in support of the new reconstruction. At the sametime, I hope this will serveto emphasizethe value of historical linguisticdataandmethodsforthe reconstructionof populationmovements. HISTORICAL REFERENCES TO THE SOMALI The first clear written referenceto any Galla or Somali group is found in the writings of the thirteenth-centuryArab geographer,Ibn Sa'id. Ibn Sa'id saysthat Merca,a town on the southernSomalicoastnearthe Shebeli River, was the 'capital of the Hawiye country', which consisted of more than fifty villages(or districtsor tribes).3This areais todaythe home of the HawiyeSomaliclan-family,so thereis goodreasonto assumethatthe Merca region has been occupied continuouslyby the same Somali group for the past 700 years. In fact, we can probablyextend this to 800 years, for the geographeral-Idrisi remarksthat Merca was the region of the 'Hadiye' in the twelfth century. It is quite likely that the extanttexts containan error,and that it should be 'Hawiye', as Guillain, Schleicher,and Cerulli 1 This view is presented most fully by Cerulli (I957), I, and I. M. Lewis (I959a, 1960). 2 H. S. Lewis (I962); Fleming (1964); Haberland (1963), 3-6. Murdock (I959), 319-20, 323-4, suggested that the Galla and Somali originated in the highlands of south-eastern Ethiopia but in most other respects followed the traditional reconstruction. 3 Guillain (1856), I, 238-9; Abu al-Fida (I848), II, 232; Cerulli (1957), I, 94; Schleicher (1892), ix. 28 HERBERT S. LEWIS . Tana' v KENYA Fig. i. Ethiopia and Somalia.  ORIGINS OF THE GALLA AND SOMALI 29 ",- Q s ::: 4iiiiiii MiOGOGOD0 ./ BARARE Fig. 2. Distribution of the Galla and Somali. 30 HERBERT S. LEWIS suggest.4 The earliest written evidence, therefore, indicates that some Somali were quite far south at least as early as the twelfth or thirteenth century. The next reference to the peoples we are concerned with is contained in the victory song of King Yeshaq of Ethiopia, who reigned from I4I4 to I429.5 In this song the name 'Somali' is first recorded. They are merely listed as one of the many peoples the king had fought, however; thus the reference tells us only that the Somali were within striking distance of the Ethiopians at this time, but does not locate them more definitely. A third relatively early document containing recognizable references to Somali groups is an Arabic chronicle detailing the wars between the Muslims of Adal and the Christians of Ethiopia.6 The Somali groups mentioned in the Futuh al-Habasha are generally ones which are still found in north-west Somaliland-the area which was once Adal or adjacent to it. These include the Yabarre, Bartirre, Marrehan, Geri, Habr Magadle, and various Dir groups. The evidence of this work, written between I540 and 1550, is, therefore, that the composition of the Somali population of north- west Somaliland has not changed substantially since then.7 In addition to these accounts which give the names of Somali groups there are several references to the attributes of the people living in the Horn. Yakut, a twelfth-century Arab geographer, says that the inhabitants of Mogadishu were 'Berbers, of a colour between that of the Abyssinians and the Negroes'.8 Ibn Battuta, who travelled along the coast in I33I, reports that the desert country between Zeila and Mogadishu was inhabited by dark-skinned people, many of them 'heretics', who possessed camels and sheep.9 Much earlier, a ninth-century Chinese text states that the people of Po-pa-li (presumably Berbera) 'do not eat any cereals but they eat meat; more frequently they prick a vein of one of their oxen, mix the blood with milk, and eat it raw. They have no clothes, but they wrap around their waists a sheep's skin which hangs down and covers them.'10 These tantalizing, fragmentary glimpses of the people of the Horn suggest a picture not markedly different from that of today. They are too general and partial to allow positive identification of any people, but there is certainly no indication of any major change during this period, nor any indication that these were not Somali.11 4 Guillain ix; Cerulli (I957), I, 92. (1856), I, 193, 203; Schleicher (1892), 5 Cerulli 6 Shihab ad-Din (I897). (I957), I, iii. 7 I have not been able to examine this work directly. This was the conclusion of Schleicher) 1892), xi, and it follows from the remarks of I. M.Lewis (1959a), 34; (1 960), 223. 8 Guillain (1856), I, 234. 9 Freeman-Grenville (I962), 27. Cerulli (I957), I, I63, claims that Ibn Battuta found Mogadishu to be the capital of a small Arab-Somali state. 10 Freeman-Grenville (I962), 8. 11 Wheatley (I964), I43-4, I70-I, believes that this may refer to the Galla, because the northern Somali today do not drink blood while many Galla do. This trait is so wide- spread in Northeast Africa, however, that it cannot be used to positively identify any one group. Today the southern Somali drink blood, while not all Galla do. And the Qu'ran (v, 3) forbids the drinking of blood. It is more economical to assume that the northern Somali gave up this trait within the last eleven centuries than to postulate the presence of Galla near Berbera in the ninth century on this evidence.  ORIGINS OF THE GALLA AND SOMALI 31 ANGOT1 SHOA ADAL DANMOT Harar? -ATAJAR. Fig. 3. Ethiopian provinces and kingdoms in the sixteenth century. 32 HERBERT S. LEWIS These exhaust the contemporary documents relevant to the origins of the Somali. None of them permits us to hypothesize major movements or a place of origin. They do, however, present a picture of stability in the Horn of Africa in both the way of life and the occupation of specific areas by particular Somali groups. HISTORICAL REFERENCES TO THE GALLA There is no known recognizable reference to the Galla before the middle of the sixteenth century, but their record is relatively full from then on. The first written references to the Galla are contained in several contem- porary Ethiopian and Portuguese accounts of the wars between the Galla and the peoples of southern, eastern, and central Ethiopia. The Portuguese Jesuit, F. Alvares, who was in Ethiopia about 1526, fails to mention the Galla in his book, but the chronicle of the Ethiopian king Galawdewos tells of a series of campaigns against the Galla beginning about 1545.12 A text written in the 1580's or I 59o's by an Ethiopian priest named Bahrey claims that the Galla actually began their invasions in the reign of Galawdewos's father, Lebna Dengal (I508-40).13 Manoel d'Almeida, who visited Ethiopia in the I620'S, gives the year I537 as the one in which the Galla entered the south-eastern province of Bali, supposedly the first province they captured.14 Almeida was largely dependent upon Bahrey's text for his knowledge of the Galla, however, and thus Bahrey's History of the Galla is our most valuable primary source on the early history of the Galla.15 Bahrey's home was in the region of Gamo, on the shores of Lake Abaya (Margherita) in southern Ethiopia, which was one of the earliest areas to be attacked by the Galla. He was an eyewitness of the expansion of the Galla, and his remarkable curiosity about, and knowledge of, the con- quests, customs, and political organization of the Galla demand that serious consideration be given to his text. According to Bahrey, 'The Galla came from the west and crossed the river of their country, which is called Galana, to the frontier of Bali.... ' Since the word galana means 'river' in Galla one cannot be definite about which river was meant, but two major rivers of southern Ethiopia, known today as the Galana Sagan and Galana Dulei, seem to be good candidates. Both are close to Gamo, to Lakes Abaya and Shamo, and are south-west of Bali. Even if the Dawe or Ganale rivers were meant, however, the Galla could still have moved from the west into Bali, and come from the same general area. 12 Alvares (i88I); Budge (I928), II, 341 ff. 13 iii. Bahrey (1928), ii, 603; (I954), 14 On Beccari's map, after Almeida, reprinted in Coulbeaux, n.d. 15 Almeida borrowed heavily from Bahrey, and Ludolf derived much of his information on the Galla from Tellez's version of Almeida's work. Bahrey is, therefore, the basis for most of our knowledge of the Galla in the sixteenth century. 16 Bahrey (I954), III; Budge (I928), II, 603, translates this as '.. .[and came] to the frontier of Bali. . .'.  ORIGINS OF THE GALLA AND SOMALI 33 But Bahrey gives further clues to pin-point the original area of the Galla and to place this area closer to the Sagan or Dulei than to the other rivers. He tells us that some Galla groups 'came out of their country by way of Kuera' and that they then attacked Gamo and Waj.17 Since Kuera, known today as Koyra, is located just east of Lake Shamo and south of Gamo, this argues that their origins, according to Bahrey, were south of Lake Shamo and Koyra and in the Sagan-Dulei region.18 This region fits all of Bahrey's statements. It is also the exact area which linguistic data point to as the probable Galla homeland. Bahrey indicates that there were two major Galla divisions and that each took a different route. The Baraytuma (Barentu, Bareituma, etc.) went west to Bali and then, by the 1580's or 1590's, into Dawaro, Adal, Fatajar, Gafi, Angote--all on the eastern side of the Ethiopian empire and largely Muslim-and even as far north and west as Amhara, Begemder, and Dembea. The Borana went through Koyra, Gamo, Waj and Damot. With- in fifty years the Galla were waging war against organized states and con- quering land as far as 500 miles away from their original homeland. These movements are well documented in the chronicles of the Ethiopian kings, the accounts of the Portuguese Jesuits, and the traditions of the Muslim state of Adal as well as by Bahrey.19 The historical record is far less ample regarding the presence of Galla in southern Somalia, however. There is evidence that Darod Somali groups drove Galla groups westward, out of southern Somalia, within the past two centuries.20 These movements have generally been regarded as only the last in a series of similar movements which began much earlier. The crucial question is, however, when the Galla and Somali first arrivedin those regions. We have seen that the Somali were evidently there from the twelfth or thir- teenth centuries, but there is no documentation of the presence of the Galla in southern Somalia before the seventeenth century. Father J. Lobo encountered Galla at Malindi and at 'Jubo' along the coast some days north of Pate, presumably near the Juba River, in i624.21 Father P. Velasco, writing in the I620's, claims that 'Britamos'-evidently Baray- tuma Galla-lived west of the Somali in the Shebeli-Juba region.22 And, according to Kirkman, the archaeology of Gedi and the oral traditions of the coastal Bantu argue that the Galla were attacking and moving into the 17 Bahrey (I954), 114. 18 Koyra is south-east of the present region of Gamo but directly south of the area Chiomio (I938), xi, xviii, believes was once also called Gamo-the area directly east of Lake Abaya. This eastern region is occupied largely by Galla today, whereas the Gamo west of the lakes is not. For these reasons, and others to be discussed below, it seems possible that Bahrey's Gamo was on the eastern side of Lake Abaya. 19 For example: Almeida, Tellez, Lobo, Ludolf, Budge, Basset and Cerulli. There is a short summary with bibliography in Guebre Sellassie (I93I), I, 580-I. 20 I. M. Lewis (I960), 225-7; Cerulli (I957), I, 67-8. 21 Lobo (I886), 24. 22 Cerulli (I957), I, 65; I. M. Lewis (I960), 225. 3 AH VII 34 HERBERT S. LEWIS coastal areas as far south as Malindi at just about the beginning of the seventeenth century.23 The sudden appearance of indications of the Galla in southern Somalia and north-eastern Kenya at the very time that other Galla were invading vast areas to the north strongly suggests that the Galla had just arrived in these regions. There is no evidence that they were being driven out at this time. It is more probable that their movement into this area was one aspect of their greater expansion than that their general expansion was due to their expulsion from southern Somalia. There is no hint in Bahrey, or any other written document, that the Galla might have lived in the Horn before the sixteenth century.24 Bahrey's testimony that the Galla moved from west to east and his failure even to mention the Somali or any people pressing upon the Galla renders most unlikely the hypothesis that the Somali were pushing the Galla from the east. In the I500's and I6oo's the Galla were an expanding people, moving north, north-east, east, and south-east, and it was undoubtedly at this period that contact between the easternmost Galla and the westernmost Somali began. The contacts have continued over the centuries across a wide front, from Ifat and Adal in the north, through the Ogaden in the centre, and as far as the Juba and Tana rivers in the south. There is no written evidence for any earlier conflicts between these peoples. GALLA TRADITIONS OF ORIGIN The oral traditions of the Galla relevant to the problem of their origins are in remarkable agreement with each other and with the written evidence. Because of the consistency of different accounts, their agreement with Bahrey and with the linguistic evidence, their unembellished nature, and the relatively short period of time involved, these traditions are worthy of careful consideration. The Galla do not claim to have lived in the Horn of Africa, nor do they speak of any homeland other than south-central Ethiopia.25 Although 23 Kirkman (I954), 74-6. See also Cerulli (I957), I, 255; Elliot (1926), I52. 24 Almeida (1954), I34, stated that '(The) lands lying between Bali and the sea, the coast of which the sailors call "the desert", is the real home of the Gallas.' This has been taken by Cerulli (1938), 30, as confirmation of the origins of the Galla in the Somalilands. However, on Almeida's map (1954) he shows the 'fatherland and birthplace of the Galla' to the south-west of Bali, exactly where Bahrey's account placed it. Almeida's knowledge of the geography of the Horn is at fault because, as the map indicates, he thought Bali was farther to the east and that the coast ran more sharply from east-north-east to west-south- west and thus that an area south-west of Bali could be between it and the sea. He did not mean that they came from the Horn. 25 Huntingford has written that the traditions of the Galla derive them from Arabia (1955, 19; Beckingham and Huntingford, I954, lxxi, III). This seems to be based on the tradition heard in Shoa by Cecchi (i886), I, 473, and Soleillet (i886), 252-3, which claims descent from Abraham, Isaac, Esau, Eliphaz, and Omar. Such a random tradition, from a Christian area, cannot be taken seriously.  ORIGINS OF THE GALLA AND SOMALI 35 there are minor differences of detail, virtually all of the recorded traditions of the Galla indicate an origin in or near the Borana region.26 Some Galla claim to have come from Borana itself.27 Other Galla say that they came from the area of Mt. Wolabo, which is located about thirty miles east of Lake Abaya, just north of Borana. Cerulli heard this tradition from both the Gombicu Galla of Shoa and the Arusi Galla of the upper Shebeli basin, while Cecchi, Ratzel, and Budge report the same tradition but give no provenance for it.28 The third place of origin recorded in Galla traditions is Bahrgamo.29 Bahrgamo can be identified as Lake Abaya and the region around it- Gamo. Bahrgamo literally means 'Gamo lake' in Amharic.30 Almeida and subsequent writers and map-makers clearly indicate a 'Bargamo' as the southernmost province of Ethiopia before the incursions of the Galla. Almeida's map shows this province south of Lake Zwai and Waj, west of Bali, just east of a province called 'Sugamo', and immediately north-west of the area he calls the 'fatherland and birthplace of the Galla '.31 This map, therefore, clearly places Bargamo in the Lake Abaya area. It also indicates two Gamos, with Bargamo as the eastern one, and strengthens the possi- bility that Bahrey's Gamo was on the eastern side of Lake Abaya.32Further- more, Mt. Wolabo is located within this region. We are at least justified in associating Bahrey's Gamo, Almeida's Bargamo, and the Bahrgamo of Galla traditions with the area immediately around Lake Abaya. Thus Galla traditions of origin point to the same homeland as do Bahrey and Almeida: north-western Borana and environs. SOMALI ORAL TRADITIONS The oral traditions of the Somali have been collected and analysed more fully than those of the Galla and they have served as the primary support for the traditional hypothesis. This support is more apparent than real, however. The traditions are both highly suspect and frequently irrelevant to the problem of origins. I. M. Lewis has amply demonstrated that Somali traditions dealing with their early history and relations to Islam and Arabia serve as mythical charters for modern religious and political life.33The Muslim Somali place high value on descent from Muhammad and his family, or at least from 26 Cf. Haberland (I963), 5. 27 Abebe Ambatchew et al. (I957), 9, Montandon (1913), 67. The present author encountered this tradition among the Galla of Jimma Abba Jifar. 28 Cerulli (I933), II, 69; (I957), II, I27. Also Haberland (I963), v, 527; Cecchi (I886), II, 473; Ratzel (1904), II, 495; Budge (I928), II, 616. 29 For example, Beke (1848), II6; Budge (I928), II, 616; Haberland (I963), 4; Harris (2844), III, 45. 30 Chiomio (1938), xi, xviii; Guebre Sellassie (I93I), II, 579. 31 Almeida (I954), 9-II, map. See also the maps in Coulbeaux (n.d.) II. 32 Chiomio (I938), xvii-xviii notes that King Yeshaq's victory song also mentions two Gamos. 33 I. M. Lewis (I962), 45-7. 3-2 36 HERBERT S. LEWIS Arabs. Many Somali clan-families and clans, therefore, claim descent from immigrant Arab sheikhs or saints who married indigenous Somali women.34 These tales are, according to Ferrand, 'generally non-concordant, some- times contradictory'.35 Hagiologies are readily invented for these ancestors, and graves and shrines may be attributed to them as current sociopolitical conditions necessitate.36 There can be no doubt that these tales contain little reliable historical material; but, even if we accept I. M. Lewis's contention that they mirror actual settlement of Arabs among Somali, this would leave un- answered the question of where the indigenous Somali had come from. The traditions of the Dir, Isaq, and Darod, which are treated in the most detail in the literature, are silent on this point. Furthermore, even after the supposed immigration of these sheikhs, the Dir and Isaq movements have been-with a few exceptions-minimal. There is therefore no indication that they did not previously come from the south and no evidence of any subsequent south or south-west movement. In the south, Cerulli has collected and analysed the traditions of various Hawiye and Rahanwin groups.37 Most of this material relates only to relatively minor local movements within southern Somalia. These affirm that Tunni, Hawiye, Ajuran and Jiddu Somali have been in the south for many centuries, as the written evidence indicated. As for their contention that they came from the north-over nine centuries ago-this can be seen as a necessary accompaniment to their myths of descent from Arabian immigrants. People who know enough and care enough to fabricate genea- logies tracing their ancestry back to Muhammad's uncle are not likely to claim to have come from southern Ethiopia. Most of the Somali clan families-Dir, Isaq, Hawiye, Rahanwin- inhabit relatively restricted areas, and there is no evidence that they have made any recent major movements. The Darod clan-family, on the other hand, has such a wide distribution that a recent expansion is indicated. This Darod movement must be seen as a recent and secondary dispersal, not an indication of the general trend of Somali migration. The Somali oral traditions about their origins are suspect, because of the great time depth involved-at least I,ooo years-and because of the overriding importance of the claim of Arabian descent. But even taken at face value they do not provide much foundation for the hypothesis of a northern origin. SOMALI REFERENCES TO THE GALLA It is often claimed that the Somali are in agreement that the Galla pre- ceded them in the Horn. Close inspection of these traditions indicates 34 I. M. Lewis (1955), 14-40 passim; (I955-56); Cerulli (1957), I, 60-2; II, 252-3. 35 Ferrand (I903), 66. 36 I. M. Lewis (1956), 153; (I960), 221; (I96Ic), 7I; (I962), 44-7. 37 Cerulli (I957), I, 57-69.  ORIGINS OF THE GALLA AND SOMALI 37 a more ambiguous picture, however. Almost all of these traditions have been collected from western Somali groups-in north-west Somalia, the Ogaden, and the Shebeli and Juba valleys-who today live not far from Galla.38 In these areas Somali have been in continual contact with Galla for about 400 years, competing for water, grazing, and agricultural land. They will certainly have difficulty remembering whether a skirmish or a shift in territory took place 500 or merely 350 years ago. Pirone, in his article on the traditions of the Ogaden Somali, warns us that, 'It is difficult to establish who the first inhabitants of the Ogaden... were, and to which epoch to refer the first human settlement of the territory. The traditions- the only source at our disposal-, vague and confused, modified as they are by continued addition of new elements referable to a more recent period, don't give the anticipated indications on this problem.'39 In addition, there is a tendency for Dir Somali to be confused with the Galla in the Ogaden Somali traditions.40 The situation is unlikely to be different in the north-west,41 and in the south, most of the traditions of Galla-Somali conflict are supposed to refer to the seventeenth century and after.42 In an attempt to demonstrate the presence of Galla farther east in the Horn, Cerulli has argued that a place name in Mijurtein and the names of two famous northern Somali raids refer to Galla.43 However, the root gal (pl. galo) in these names (Galo-od; Gal-eri; Galka'ayo) does not mean the people known to the Somali as Galla or Gallawi (pl.) but means 'pagan, infidel, or non-Muslim in general. In contemporary Somali these are separate and distinct terms.'44 The assumption that these must once have been one term has no other justification than the hypothesis it is meant to prove. The same linguistic confusion lies behind the claim that the Somali attribute various stone graves in their country to the 'Galla'. According to I. M. Lewis, the Somali actually say that galo made them, or-much the same thing-dadki hore, 'people who were before'.45 This is one of the commonest ways for people to explain away an unknown object. When I. M. Lewis excavated three of these graves he discovered that they were not more than 250 years old (on the basis of a radio-carbon date). For this and other reasons Lewis concludes that these graves 'are comparatively recent and contain Somali remains'.46 This should be a warning against lightly accepting Somali traditions which purport to deal with events of 500 or more years ago. 38 Pirone Cerulli (I957), I. M. Lewis (1954), II9; I, 57-9; (I959a). 39 Pirone (1954), 119. 40 Pirone, loc. cit. 41 Shihab ad-Din, who knew the north-west in the first half of the sixteenth century, does not mention the Galla in his long chronicle (Schleicher, I892, p. xii). The chronicle of the history of Adal-largely a compilation of oral traditions-likewise fails to mention the Galla until they attack the Harar area in the reign of Emir Nur, about 1565 (Cerulli, 193I). 42 Cerulli(1957), I, 57-8. 43 Cerulli (1957), I, 72-3, o1I-2, 163. 44 I. M. Lewis 22. (1958), 28o-I; (I959a), 45 I. M. Lewis 46 I. M. Lewis (196I a). (I961 a), I03. 38 HERBERT S. LEWIS The foregoing discussion has indicated the following preliminary con- clusions: (i) The extant historical documents and oral traditions relevant to the origins of the Galla indicate that they began their expansion about 1540 from south-central Ethiopia, probably south of Lake Abaya. (2) There is no evidence to suggest that the Galla previously occupied the Somali- lands. (3) Genuine historical documents offer virtually no evidence re- garding the origins of the Somali, and the oral traditions of the Somali are questionable and often irrelevant. In the following section I hope to demon- strate that linguistic evidence offers overwhelming confirmation of our Galla hypothesis and compelling evidence for the new hypothesis as to the origins of the Somali. Historical linguistics has put a powerful tool at the disposal of the culture historian, but until recently workers in Northeast Africa have failed to take advantage of it. Through the use of the principles of historical linguistics, and especially of migration and dispersal area theory, as developed and used by Sapir, Greenberg, Kroeber, Dyen, and Diebold, it is possible to derive, from synchronic linguistic data, hypotheses as to genetic relations among languages, centres of origin and dispersal, and directions of population movements, which can be agreed upon and replicated by independent workers.47Through comparative linguistics it is possible to classify geneti- cally related languages into distinct groups, each of which represents a previously unified speech community which, in the course of time, diverged into a number of different languages. When the component languages of genetic groups are geographically dispersed, 'since related languages are divergent forms of one language, it follows that all of the forms can be traced to movement from a single continuous area'.48 In order for related languages to have become dispersed beyond that area of origin, population movements-expansion, migration, or the intrusion of unrelated languages -must have occurred. A knowledge of the modern distribution of related languages is sufficient to produce highly probable hypotheses as to such origins and movements. Only recently have these principles been applied to historical problems in Northeast Africa, but the result has already been a new series of hypotheses which offer economical explanations for popu- lation movements. Using historical linguistics, Murdock, H. S. Lewis, Greenberg, and Fleming have all come to generally similar and mutually reinforcing conclusions: the homeland of the Eastern-Cushitic-speaking peoples must have been in southern Ethiopia-northern Kenya.49 The recognition that Somali, Galla, Afar and Saho are related languages within the larger category of Cushitic is about Ioo years old and has never been seriously disputed. On the other hand, the subgrouping of Cushitic (and more particularly of Eastern Cushitic) had to await the accumulation 47 Sapir (I916), 75-83; Greenberg (1955), 40; Kroeber (I955); Dyen (I956); Diebold (I960). Also Voegelin (1958). 48 Dyen (I956), 614. 49 Murdock (I959), 319-20, 323-4; H.S. Lewis (I962); Greenberg (I963), 43; Fleming (1964).  ORIGINS OF THE GALLA AND SOMALI 39 of fuller evidence regarding the languages of southern Ethiopia and northern Kenya. In I940 M. M. Moreno suggested a genetic classification of Cushitic and its several subgroups which, though modified as a result of the addition of new data on almost fifty languages, has remained basically unchanged in spite of independent replications by Greenberg and Fleming.50 The genetic classification and subgrouping of the Eastern Cushitic languages indicate that we are dealing with at least twenty-four languages which can be subdivided into four co-ordinate branches: (i) Somali, Rendille, Baiso; (2) Galla, Konso, Gidole, Gato, Arbore, Magogodo, Warazi, Gawata, Tsamai, Geleb; (3) Afar Saho; (4) Sidamo, Kambata, T'ambaro, Hadya, Alaba, K'abena, Marak'o, Darasa, Burji.51 It will be seen from this genetic classification that neither Somali nor Galla holds any special linguistic position within Eastern Cushitic, within Cushitic as a whole, or with respect to the other, in spite of their historical, cultural, and geopolitical importance. The immediate and overwhelming implication of these linguistic facts is that the origins of the Galla and Somali cannot be considered without regard to those of the other twenty- odd languages with which, at some time in the past, they were one. Any reconstructions of their origins and movements must be consistent with those of the others (and, at still higher levels, with those of the Cushitic and Afro-Asiatic languages as a whole). Having classified the subgroupings of Eastern Cushitic, it is now our task to determine where their ancestors must have lived at that time in the past when they represented just one speech community in one continuous area. Such a determination can be made for each subgroup or for all of Eastern Cushitic. One must be consistent with the other, but each will tell us about a different period and will be of value for somewhat different problems. The cardinal principle involved in making this determination has been expressed by Dyen as follows: 'The probabilities of different reconstructed migrations are in an inverse relation to the numberof reconstructedlanguage movements that each requires.'52The principle of economy of explanation requires that we postulate a centre of dispersal which requires the fewest population movements to account for the current distribution of these languages.53 In the case at hand the answer appears obvious. Since today twenty-one of the twenty-four languages of the Eastern Cushitic group are spoken in southern Ethiopia-northern Kenya (from 7? N., 40? E. to o?, 37? E.), if the Galla and Somali and their relatives had originated in northern Somalia, we would have to account for twenty-one population movements. If, on the other hand, we assume that the original homeland was in the area of 50 Moreno (1940), 315-2I; Greenberg (1963), 43; Fleming (1964), 82-3. 51 This classification and subgrouping follows that of Greenberg (1963, p. 43) almost completely. Fleming (I964), 82-3, is in close agreement with Greenberg. 52 Dyen (1956), 6I3. Emphasis Dyen's. 53 The reader is referred to Sapir, Dyen, Kroeber and Diebold for fuller discussion of the postulates and methods involved. 40 HERBERT S. LEWIS *ADDISABABA 0 Miles I I,50 H,ADYA- , ?::?I L.Zwar: ?,I?z. ?"r;- " HADYA ?5;?:? RO ALABT SI DAMO BUP?\ L.Rudolf Fig. 4. Eastern Cushitic languages of southern Ethiopia.  ORIGINS OF THE GALLA AND SOMALI 4I southern Ethiopia-northern Kenya, we need postulate only three move- ments to the north-east-those of the Afar, Saho, and Somali, thus greatly increasing the probability of our hypothesis. If we consider each subgroup separately, we find the evidence for the new hypothesis just as persuasive. Two of the branches, the Sidamo and the Galla, have all their languages represented in this southern area. Somali is the only one of three languages in its group to be found outside of this area. And Afar and Saho are so closely related that we are justified in assuming that they represent just one independent movement, with the split into two languages occurring after they were resident in the north-east. Nineteen of the component languages of Eastern Cushitic are found relatively close together in the area between Lake Zwai on the north and Lake Stefanie and Rudolf on the south. It is only the Mogogodo and Rendille, who evidently moved south, and the Somali, Afar and Saho, who moved east and north, who have left that area.54 Fleming has discussed the origins of the Somali-Rendille-Baiso group at some length, and, since I am basically in agreement with him, I shall only cite his conclusion that: 'In general. .. the watershed of the Juba River, as far south as the present Somalia border, is the homeland of proto-Macro- Somali. 55 At the moment we have no written evidence nor oral traditions to support this view, but neither, I submit, have we any evidence seriously to question it. I believe it must stand as the only reasonable hypothesis and that it is consonant with the rest of our knowledge of the western Cushitic peoples. For the origins of the Galla we fortunately have oral traditions and written history which confirm the obvious conclusion to be drawn from linguistic distributions. The Galla originated in the area between and around Lakes Shamo and Stephanie, in the area of the Galana Sagan and Galana Dulei, just south of Bahrgamo and Mt. Wolabo, in north-west Borana. It is here that the closest linguistic relatives of the Galla-Konso, Gato, Gidole, Arbore, Gawata, Warazi, Tsamai, Geleb-are found. The Galla language was once one with these, and Galla speakers even today live in this area. This is striking confirmation of Bahrey, of Almeida's map, and of the oral traditions of the Galla themselves. Conversely, the other sources confirm the value of the linguistic methods. Finally, comparative linguistic data also indicate that the Afar and Saho 54 According to Sapir (I9I6), 76-8, geographical areas which are centres of linguistic differentiation, with high degrees of diversity among the members of single language groups, tend to be areas of older occupation than those with relatively little differentiation. The probabilities are against numerous migrations of related speech communities into the same region, because migration leads to the dispersal of closely related languages rather than their concentration. Time is required for the differentiation of related languages. Where clusters of related languages are found in one region, therefore, we assume some time depth for their settlement there. Dyen concludes that 'determinable positive migra- tions are from complex areas to uniform areas' (1956), 625. Our findings in this case are consistent with Sapir's and Dyen's postulates. 55 Fleming (I964), 87. 42 HERBERT S. LEWIS must have moved out of the south-west first, followed by the Somali, with the Galla expanding still more recently. Afar and Saho have evidently been in their area long enough to have diverged into two closely related but distinctly different languages. Among the Somali, however, the process of differentiation has not had time to go so far, for today Somali speak several distinct dialects which, though perhaps approaching the status of distinct languages, can still be considered as dialects of one language.56The Galla, it is clear, cannot have begun separating very long ago, since their 'language is so essentially constant that the women and children of the Gurri tribe, who inhabit the El Wak oasis and the surrounding districts... talk the same dialect as those of the Walega'.57 CONCLUSIONS AND IMPLICATIONS The preponderant trend of Eastern Cushitic migration and expansion has been from the south and west to the north and east rather than the reverse. The Afar and Saho were evidently the first to move north-east, while the Somali moved into the Horn some time later. It was not until the sixteenth century, however, that the Galla began their great expansion in all directions but west. The original homeland of the Galla may be placed in or around north-western Borana; that of the Somali in the south-eastern highlands of Ethiopia. For four centuries the Galla and Somali have competed with each other over a long front from Harar in the north to the lower Juba and Tana rivers in the south, and during this time there must have been numerous shifts of fortune and transfers of territory. A Galla group found within Somali country, no longer in continuous distribution with other Galla, may be considered a remnant group relative to the most recent Somali movements, but was undoubtedly once in the vanguard of Galla movement into Somali territory. In their nomadic rounds, which may cover hundreds of miles in a single year, pastoral Somali and Galla groups must have penetrated each other's territories many times since I550. The Galla, Somali, Afar and Saho once belonged to a single speech community located in southern Ethiopia and presumably shared more common cultural elements than they do now, many centuries after their separation. In view of this it would be better to view these groups as having diverged from a generally similar parent culture rather than, as in the past, a prioriascribing shared traits to more recent borrowing. Thus, for example, the age-grade organization among the Afar and some Somali may represent an inheritance from their earlier residence in southern Ethiopia. More generally, the northern Somali bias towards pastoralism and their distinc- tive lineage system might be viewed as an adaptation to the extreme poverty of their environment, while the Rahanwin and Digil Somali, who practise 56 Tucker and Bryan (1956), 122-6; Fleming (1964), 79-83. 57 Foot (19I3), vi; Fleming (I964), 85.  ORIGINS OF THE GALLA AND SOMALI 43 mixed agriculture and do not stress lineage relations for political purposes, may be closer to the original Somali pattern. Lineage does not serve as the basis for sociopolitical life among most of the Eastern Cushitic peoples of southern Ethiopia and northern Kenya. It may be the northern Somali, therefore, who are divergent and require a special explanation, rather than the Rahanwin and Digil.58 Finally, the new reconstructions presented here have implications for the methodology of African culture history in general. Reconstruction of the origins of ethnic groups should be in agreement with the comparative linguistic evidence. In I916 Sapir wrote, 'If, as may sometimes happen, the linguistic evidence seems to run counter to other evidence or to a pre- vailing theory, it should not be lightly discarded as irrelevant to historical problems. While it may be forced to yield in the face of powerful testimony pointing to contrary conclusions, its claims always deserve serious consider- ation.'59 This case, I believe, amply bears out Sapir's position, for the new hypothesis was formed initially on the basis of the linguistic data alone, although it contradicted most of the literature on the subject. I hope it has been demonstrated that there are not, in fact, data to support the older hypothesis, and that the application of migration and dispersal area theory to the problem has yielded a highly probable hypothesis capable of overturn- ing the existing reconstruction and strong enough to serve as the basis for a new one. BIBLIOGRAPHY ABEBE AMBATCHEW et al. Field trip to Nakamte, Bulletin no. 6, Ethnological Society, University College of Addis Ababa (I957). ABU AL-FIDA. Gdographie d'Aboulfeda, tr. M. Reinaud, 3 vols (Paris, 1848-83). ALMEIDA, M. DE. The History of High Ethiopia or Abassia. Extracts published in C. F. Beckingham and G. W. B. Huntingford, Some Records of Ethiopia, 1593-1646 (London, 1954). AL-OMARI,IBN FADLALLAH. Masalik el Absar, tr. and annotated by Gaudefroy- Demombynes (Paris, 1927). ALVARES,F. Narrative of the Portuguese Embassy to Abyssinia, during the Years 1520-1527 (London, i88i). AZAIS, R. P. and CHAMBARD, R. Cinq annees de recherches archeologiques en Ethiopie (Paris, 1931). BAHREY. History of the Galla (a) in E. A. W. Budge, A History of Ethiopia, 2 vols. (London, 1928); (b) in C. F. Beckingham and G. W. B. Huntingford, Some Records of Ethiopia, 1593-1646 (London, I954). BASSET, R. Etudes sur l'histoire d'Ethiopie (Paris, i882). BECKINGHAM, C. F. and HUNTINGFORD, G. W. B. Some Records of Ethiopia, 1593-1646 (London, I954). BEKE, C. T. 'On the Origin of the Gallas', Report to the 1847 meeting of the British Association for the Advancement of Science (1848), pp. 113-I8. 58 Cf. I. M. Lewis (1960), 227-8. 59 Sapir (I9I6), 82. 44 HERBERT S. LEWIS BUDGE, E. A. w. A History of Ethiopia, 2 vols. (London, I928). CECCHI, A. Da Zeila allafrontiere del Ca.ffa, 3 vols. (Rome, I885). CERULLI, E. 'Documenti arabi per la storia dell'Etiopia', Memorie della Reale Accademia Nazionale dei Lincei, Ser. VI, IV (1931), 39-10I. CERULLI, E. Etiopia occidentale, 2 vols. (Rome, 1932-33). CERULLI, E. Studi etiopici, vol. II (Rome, I938). CERULLI,E. Storia della letteratura etiopica (Milan, 1956). CERULLI, E. Somalia, scritti vari editi ed inediti, 2 vols. (Rome, 1957). CHIOMIO, P. G. Lingua Wollamo (Turin, I938). COULBEAUX, J.-B. Histoire politique et religieuse d'Abyssinie, 2 vols. (Paris, n.d.). DIEBOLD, A.R. 'Determining the centers of dispersal of language groups', Int. J. Am. Ling. xxvI (I960), i-Io. DYEN, I. 'Language distribution and migration theory', Language, xxxII (1956), 6iI-26. ELLIOT, J. A. G. 'A visit to the Bajun Islands', J. Afr. Soc. xxv (1926), 147-63. FERRAND, G. Les Qomalis (Paris, I903). FLEMING, H. C. 'Baiso and Rendille: Somali outliers', Rassegna di Studi Etiopici, xx (I964), 35-96. FOOT, E. C. A Galla-English Dictionary (Cambridge, I9I3). FREEMAN-GRENVILLE, G. S. P. The East African Coast (London, 1962). GREENBERG, J. H. Studies in African Linguistic Classification (New Haven, 1955). GREENBERG, J. H. Essays in Linguistics (Chicago, 1957). GREENBERG, J. H. 'The Mogogodo, a forgotten Cushitic people', J. Afr. Lang. 11 (1963), 29-43- GUEBRE SELLESSIA. Chronique du regne de Menelik II, 2 vols., ed. M. de Coppet (Paris, I931). GUILLAIN, c. Documents sur l'histoire, la geographie, et le commerce de l'Afrique Orientale, 3 vols. (Paris, 1856). HABERLAND, E. Galla Siid-Athiopiens (Stuttgart, 1963). HARRIS, W. C. The Highlands of Aethiopia, 3 vols. (London, i844). HUNTINGFORD, G. W. B. The Galla of Ethiopia; the Kingdoms of Kafa and Janjero, Ethnographic Survey of Africa (London, I955). HUNTINGFORD, G. W. B. 'The peopling of the interior of East Africa by its modern inhabitants.' In R. Oliver and G. Mathew (eds.), History of East Africa, I (London, 1963). KIRKMAN, J. s. The Arab City of Gedi (London, I954). KROEBER, A. L. 'Linguistic time depth results so far and their meaning', Int. J. Am. Ling. xxI (1955), 91-I04. LEWIS, H. S. 'Historical problems in Ethiopia and the Horn of Africa', Ann. N.Y. Acad. Sci. xcvi (1962), 504-II. LEWIS, H. S. 'Ethnology and African Culture History.' In W. C. Gabel (ed.) (title of volume on African history not yet known) (Boston, I965). LEWIS, I. M. Peoples of the Horn of Africa: Somali, Afar, and Saho. Ethno- graphic Survey of Africa (London, I955). LEWIS, I. M. 'Sufism in Somaliland: a study in tribal Islam', Bull. Sch. Oriental Afr. Studies, XVII (955), 58I-602; XVIII (956), 146-60. LEWIS, I. M. 'The Gadabuursi Somali Script', Bull. Sch. Oriental Afr. Studies, xxI (958), 134-56.  ORIGINS OF THE GALLA AND SOMALI 45 LEWIS,I. M. Review of Cerulli, Somalia, scritti vari editi ed inediti. Africa, XXvIII (1958), 280-i. LEWIS,I. M. 'The Galla in Northern Somaliland', Rassegnadi Studi Etiopici, xv (i959a), 21-38. LEWIS,I. M. 'Clanship and contract in Northern Somaliland', Africa, xxix (I959b), 274-93. LEWIS,I. M. 'The Somali conquest of the Horn of Africa', J. Afr. Hist. I (I960), 213-30. LEWIS,I. M. 'The so-called Galla graves of Somaliland, Man (I961 a), 103-6. I. M. 'The advance of the Somali and the withdrawal of the Galla in LEWIS, northernSomaliland,c. I Ioo-1 600 A.D.',Third Conferenceon AfricanHistory and Archaeology,School of Oriental and African Studies; mimeo (I961b). I. M. 'Notes on the social organizationof the 'Ise Somali', Rassegnadi LEWIS, Studi Etiopici, xvII (i 96 c), 69-82. LEWIS,I. M. 'Force and fission in northern Somali lineage structure', Am. Anthrop. LXIII (I96 d), 94-I12. LEWIS, I. M. A Pastoral Democracy (London, I961 e). LEWIS, I. M. 'Historical aspects of genealogies in northern Somali social struc- ture', J. Afr. Hist. in (I962), 35-48. LEWIS,I. M. 'Recent progressin Somali studies', J. Semitic Studies, IX (I964), 122-34. LEWIS,I. M. Review of Haberland, Galla Siid-Athiopiens. Man (I964), I89-90. LEWIS,I. M. 'The northern pastoral Somali of the Horn'. In J. L. Gibbs (ed.), Peoplesof Africa (New York, I965). LOBO,J. A Voyage to Abyssinia (New York, i886). LUDOLF,J. A New History of Ethiopia (London, i682). MATTHEW,G. 'The East African coast until the coming of the Portuguese'. In R. Oliver and G. Mathew (eds.), History of East Africa, I (London, I963). MONTANDON,G. Au Pays Ghimirra (Neuchatel, 19I3). MORENO,M. M. Manuale di Sidamo (Milan, I940). MURDOCK,G. P. Africa: its Peoples and their Culture History (New York, I959). PAULITSCHKE, P. Harar (Leipzig, i888). PAULITSCHKE,P. Ethnographie Nordost-Afrikas, 2 vols. (Berlin, 1893-96). PIRONE, M. 'Leggende e tradizioni storiche dei Somali Ogaden', Arch. Anthrop. Etnol., LXXXIV(1954), 119-43. PRINS, A. H. J. 'On Swahili historiography', J. East Afr. Swahili Committee, XXVIII (I958), 26-40. PRINS, A. H. J. The Coastal Tribes of the North-Eastern Bantu: Pokomo, Nyika, and Teita. Ethnographic Survey of Africa (London, I952). RATZEL,F. The History of Mankind, 3 vols. (London, I904). SAPIR, E. Time Perspective in Aboriginal American Culture. Canada Department of Mines, Memoir, no. 90 (I916). SCHLEICHER, Die Somali-Sprache (Berlin, I892). A. W. SHIHAB AD-DIN, Futuh al-Habasha, ed. and tr. R. Basset (Paris, I897). SOLEILLET,P., Voyages en Ethiopie (Rouen, i886). SWAYNE,H. G. C. Seventeen Trips Through Somaliland and a Visit to Abyssinia (London, 1903). TELLEZ,B. The Travels of the Jesuits in Ethiopia (London, I710). TRIMINGHAM, J. S. Islam in Ethiopia (London, 1952). 46 HERBERT S. LEWIS TUCKER, A. N. and BRYAN, M. A. The Non-Bantu Languages of North-Eastern Africa, Handbook of African Languages (London, I956). VOEGELIN, C. F. 'The dispersal factor in migrations and immigrations of American Indians'. In Migrations in New World Culture History, University of Arizona Social Science Bulletin no. 27 (I958). WERNER, A. 'The Galla of the East African Protectorate', J. Afr. Soc. xIII (1914), I21-42, 262-87. WHEATLEY, P. 'The land of Zanj: exegetical notes on Chinese knowledge of East Africa prior to A.D. 1500'. In R. W. Steel and R. M. Prothero, Geo- graphers and the Tropics: Liverpool Essays (London, I964).