¹ See my ‘Gao and the Almoravids: a hypothesis’, in Swartz, B. and Dumett, R. (eds.), West African Cultural Dynamics: Archaeological and Historical Perspectives (The Hague, 1979), 413–30.Google Scholar

² xxxii (1991), 251–76.

³ There is some ambiguity in his presentation of the role of force in effecting a change of rule. On p. 262 he speaks of ‘un pouvoir accaparé par la force’, whereas on p. 274 he discounts a military conquest but suggests ‘un mouvement de révoke religieuse instigué par des fervents musulmans’

⁴ The initial consonant being closer to that in the French ‘je’ rather than the English ‘jet’.

⁵ This is an anonymous and untitled work for which I have proposed the title ‘ Notice historique’ taken from the exordium of the work in its French translation. See my Sharī‘a in Songhay (Oxford, 1985), 4 n. 1.Google Scholar The translation appears as the second appendix to the French translation of the Ta'rīh al-fattãsh, trans. Houdas, O. and Delafosse, M. (Paris, 1913–1914).Google Scholar The work incorporates a great deal of material from the Ta'rīkh al-fattãsh but is probably not organically related to it.

⁶ See de Sardan, Jean-Pierre Olivier, Concepts et conceptions Songhay-Zarma (Paris, 1983)Google Scholar, art. ‘Sonay’; Prost, A., La langue sonay et ses dialectes (Dakar, 1956), 524.Google Scholar

⁷ l'Africain, Jean-Léon, Description de I'Afrique (2 vols.), nouvelle édition traduite de l'italien par A. Épaulard (Paris, 1956), 1, 16Google Scholar: ‘Dans la Terre des Noirs on parle diverses langues. L'une d'elles est nominée par les Noirs songhai. Elle est employée dans plusieurs régions: Gualata [Walãta], Tombutto [Timbuktu], Ghinea [Jenne], Melli [Mali (sic)] et Gago [Gao]’.

⁸ Last, D. M., ‘Beyond Kano, before Katsina: friend and foe on the western frontier’, in Barkindo, B. M. (ed.), Kano and Some of Her Neighbours (n.p. [Zaria], 1989), 129.Google Scholar According to him it still survives in the kirari: ‘Zaye ta Korau’ - ‘Zaye of Korau (first Muslim ruler)’. I do not find his further speculations on the identification of Zaghai with Zagha/Diakha/Jahanke/Wangara altogether convincing. He is certainly incorrect in identifying the ‘ Buda’ of Abraham Cresques’ map of 1375 with what he claims is another

⁹ It would also provide a path southwards for the copper which Mauny has suggested came from Tegidda and was used to make the bronze alloy for early Ife heads. See Mauny, R., ‘A possible source of copper for the oldest brass heads of Ife’, J. Hist. Soc. Nigeria, II (1962), 393–5.Google Scholar

¹⁰ The existence of pockets of speakers of Songhay dialects still living in the In Gall region south-west of Air and, until recently, of Songhay speech in Agades is not evidence to the contrary.

¹¹ See Hopkins, J. F. P. and Levtzion, N. (eds. and trans.), Corpus of Early Arabic Sources for West African History (Cambridge, 1981)Google Scholar [hereafter, Corpus], 320.

¹² Ibid. 332.

¹³ Earlier (p. 322) Ibn Khaldūn placed (a) Takrūr east of Kawkaw and apparently west of Kanem.

¹⁴ Al-Suyūt īwrote a letter to ‘ the kings and sultans of al-Takrūr in general’ and to the rulers of Agades and Katsina in particular. For a partial translation see Hodgkin, Thomas L., Nigerian Perspectives (2nd ed.Oxford, 1975), 118–19.Google Scholar

¹⁵ Eventually ‘Takrūr’ was applied in a very general sense to the Muslim areas of West Africa (e.g. in Muhammad Bello's Infãq al-maysūr fī ta'rīkh bilãd al-Takrūr [from Dãr Fūr to Futa Toro] or al-Bartilī's Fath al-Shakūr fī ma rifat a'yãn 'ulatmã al-Takrūr [from the R. Senegal to the Middle Niger]). See also al-Naqar, 'Umar, ‘Takrūr: the history of a name’, J. Afr. Hist., xx (1969), 365–74.CrossRefGoogle Scholar

¹⁶ Hopkins, and Levtzion, (eds.), Corpus, 320.Google Scholar

¹⁷ Ibid. 321, where the variant Zaghãwa is also noted. From a passage in the ‘History’ (Corpus, 327) it seems that Ibn Khaldūn thought the Zaghãwa were part of the Sanhãja.

¹⁸ Corpus, 335.

¹⁹ See Rouch, J., ‘Les Sorkawa, pêcheurs itinérants du Moyen Niger’, Africa, xx (1950), 5–21CrossRefGoogle Scholar; Harris, P. G., ‘Notes on Yauri (Sokoto province, Nigeria)’, J. Royal Afr. Soc., lx (1930), 283–334.Google Scholar

²⁰ Sharī'a in Songhay, 5.

²¹ These arguments are elaborated in my Sharī'a in Songhay, 4–7, where I also argue that the raison d'être for the development of right-bank Gao and the eventual relocation there of the Zã dynasty (of Kukiya), or representatives of it, was partly because of the need which North African merchants had to provision the new trading centre of Tãdmakkat and partly because of a new trade route which was opened up from Ancient Ghana to Egypt passing through this area.

²² There is an obvious need for extensive archaeological work in the Bentiya (Kukiya) area before any hypotheses can be verified in regard to its origins, extent and political and economic significance. See Farias, Paulo F. de Moraes, ‘The oldest extant writing of West Africa: medieval epigraphs from Essuk, Saney, and Egef-n-Tawaqqast (Mali)’, Journal de la Société des Africanistes, lx (1990), 65–113CrossRefGoogle Scholar, where a burial inscription dated to 1201 (possibly 1182) is discussed.

²³ Another false ethnonym. Malinke means nothing more than ‘belonging to MaliMalinke’ and as such was applied to warrior groups who were the power brokers of the Mali state.

²⁴ The velar occlusive ‘g’ is also represented by Arabic kãf, often modified by three dots above it (as used to distinguish the letter thã’). See Hunwick, J. O., ‘African language material in Arabic sources: the case of Songhay (Sonrai)’, African Language Review, ix (1973), 67–8.Google Scholar

²⁵ See ‘ Les légendes de Farang roi de Gao’, recueillies par M Dupuis-Yakouba in Desplagnes, Louis, Le plateau central nigérien (Paris, 1907), 421, 445, 450.Google Scholar

²⁶ Rouch, Jean, ‘Contribution à l'histoire des Songhay’, Mémoires de l'IFAN, xxix (1953), 174 n 13.Google Scholar He uses the spelling of the name given in the translation of TS: Kosoi.

²⁷ No Arabic text is available for the ‘Notice historique’, so we do not know precisely how this name is spelt or if the name is one or two words.

²⁸ Ultimately we may have in the two sources two expressions, Zã/Jã Kūkuray and Zã/Jã Tãkūray, which ‘conceal’ a single name.

²⁹ ‘Contribution à l'histoire des Songhay’, 170 n. 9, 174 n. 13, 175 n. 15.

³⁰ Dalziel, J. M., The Useful Plants of Tropical West Africa (London, 1937), 387.Google Scholar

³¹ Barth, H., Travels and Discoveries in North and Central Africa (3 vols.) (New York, 1857–1859, reprinted 1965), iii, 480.Google Scholar

³² Corpus, 87.

³³ See, for example, Amselle, Jean Loup, Logiques métisses (Paris, 1989).Google Scholar

³⁴ See above, pp. 255–6.

³⁵ See de Mézières, A. Bonnel, ‘Découverte de l'emplacement de Tirekka’, Bulletin du Comité des travaux historiques et scientifiques: section de géographie, xxix (1914), 132–5.Google Scholar

³⁶ We do not know to what extent there was awareness of the existence of Lake Chad among Arab writers of the Mediterranean world at this time, but there had been trade between Tripolitania and the state of Kanem to the north of L. Chad since the early ninth century.

³⁷ I have suggested elsewhere that al-Bakrī may have at one point confused Ghana with Ghiyaru which was on the ‘Nile’ (R. Senegal), and it is possible that al-Idrīsī was heir to this confusion. See Hunwick, J. O., Meillassoux, Claude and Triaud, Jean-Louis, ‘La geographie du Soudan d'après al-Bakrī: trois lectures’, in 2000 ans d'histoire africaine: le sol, la parole et l'écrit, mélanges en honneur à Raymond Mauny (Paris, 1981)Google Scholar, ‘ Troisième lecture’, 420.

³⁸ See McIntosh, Susan Keech, ‘A reconsideration of Wangara/Palolus, island of gold’, J. Afr. Hist., xxii (1981), 145–58.CrossRefGoogle Scholar

³⁹ Delafosse, M., Haut-Sénégal-Niger (2 vols.) (Paris, 1912), i, 55.Google Scholar See also his La langue mandingue et ses dialectes. ii Dictionnaire mandingue-français (Paris, 1955), 260Google Scholar, where he lists: ‘gbã-garã ou gã-garã (litt. “trou du terrain”), puits à or, région aurifère’.

⁴⁰ When Ibn Baṭṭūṭa was on the Niger near, I believe, Sansanding in 1352 he obtained information about these two towns and was told that the inhabitants of Zãgha (read: Diakha) were Muslims of long standing. See Corpus, 287. In fifteenth-century Timbuktu there seems to have been tradition of learned Kãbara scholars migrating there, again indicative of an ancient tradition of Islam; see Tar'īkh al-sūdãn, 47–8.

⁴¹ See ‘La Géographie du Soudan d'après al-Bakrī’ 422.

⁴² Or better, flood-channel. The word used is khalij- ‘gulf, inlet’.

⁴³ Corpus, 113.

⁴⁴ ‘Notice historique’, 329.

⁴⁵ Lange refers to a letter and an oral communication from Professor Raymond Mauny on this point but does not tell us exactly what he said. None of Mauny's published contributions makes any mention of the search for right-bank Gao. The difficulties of locating such a site would, indeed, be formidable given that building materials were probably adobe and arboreal materials, that the site may have since been washed over by the Niger or covered by shifting sands.

⁴⁶ See Mauny, R., ‘Notes d'archéologie au sujet de Gao’, Bulletin de l'IFAN, xiii (1951). 837–52.Google Scholar

⁴⁷ Mauny suggests that Sané was established precisely at the limit reached by the Niger flood waters and was eventually abandoned when the floods failed to reach that far up the wadi. See his Tableau geographique de l'ouest africain au moyen âge (Dakar, 1961), 498.Google Scholar

⁴⁸ Corpus, 174.

⁴⁹ Ibid. 85.

⁵⁰ Ibid. 86.

⁵¹ Ibid. 87.

⁵² ‘Du nouveau sur les stèles de Gao: les épitaphes du prince Yãmã Kūrī et du roi F.n.dã (XIIIe siècle)’, Bulletin de l'IFAN, xxvi (1974), 511–24.Google Scholar

⁵³ The title amīr al-mu'minīn was taken by those who claimed headship of the Islamic community (the umma) after Muhammad, whereas the Almoravid rulers in the Maghrib used amīr al-muslimīn to indicate that they were not challenging the ‘Abbãsid caliphs of Baghdad.

⁵⁴ Each of the three ‘kingly’ inscriptions calls the deceased al-malik al-nāṣir li-dīn Allãh, recalling titles used by the contemporary Seljuks and others in the east. On the use of the title malik in the Islamic world, see Encyclopedia of Islam (new edition), vi, 261–2, art. ‘Malik’ by A. Ayalon.

⁵⁵ Pp. 275–6. This is an imaginative reconstruction, but it is no more than that and has no basis that I know of in historical evidence.

⁵⁶ His rendering of the putative reading Zãghī, which he connects with the Berber root m.z.gh - ‘free man’. He claims that the name Zãghē was later ‘transformed’ into Z.wã, which appears on some later tombstones and was also probably used at Gao earlier (cf. my remarks on Zã/Z.wã Jã’, Juwa etc., in Sharī'a in Songhay, 10 and n. 5.)

⁵⁷ See Sharī' a in Songhay, 4. Kavvkaw was already known to the geographer alKhwarazmī who wrote in Baghdad between 813 and 842.

⁵⁸ E.g. the Yemeni origins of the Sayfuwa dynasty of Bornu, the Baghdad origins of Bayajidda ultimate progenitor of the rulers of the seven Hausa states, the Bilãlian ancestry of the Manding, etc.

⁵⁹ ‘Notice historique’, 332. TS, 5, however, hints that he may have been: ‘Because he killed the fish, people said he was a Muslim and that his descendants apostatized after his death’. Indicating his scepticism as to the historicity of the tale, however, he continues: ‘we do not know which of them first did so, nor do we know the date of Zã Alayaman's quitting the Yemen, nor when he reached them, nor what his true name was’.

⁶⁰ The elements of this argument appear on pp. 266–7 of Lange's article.

⁶¹ For long Songhay was unclassified. Greenberg finally assigned it to the NiloSaharan group. More recently Robert Nicolai has produced hypotheses first linking the language to Mande and even more recently suggesting that it is a creolized form of Tamasheq. See his Préliminaires à une étude sur l'origine du songhay (Berlin, 1984)Google Scholar and Parentés linguistiques (à propos du songhay) (Paris, 1990).Google Scholar He is careful to state, however, that his present conclusions form no more than a working hypothesis.

⁶² Farias has discovered a Sané inscription which antedates the first royal inscription by some 12 years (Makkiyya (?) bt. Ḥasan al-ḥãjj, d. 481/1088) and has also shown that an even older inscriptional tradition existed at Essuk (Tãdmakkat). See his ‘The oldest extant writing of West Africa’, n. 21 above.