¹ Sauvaget, J., ‘Notes préliminaires sur les épitaphes royales de Gao’, Revue des Études Islamiques (1948), 7–8.Google Scholar

² Idem, ‘Les épitaphes royales de Gao’, Al-Andalus, xiv (1949), 123–41 (same article published in Bulletin de l'IFAN, xii (1950), 418–40).Google Scholar

³ Hunwick, J. O., ‘Gao and the Almoravids: a hypothesis’, in Swartz, B. and Dumett, R. (eds.), West African Culture Dynamics (The Hague, 1980), 251–75.Google Scholar

⁴ Idem, Shorī a in Songhay: the Replies of al-Maghīlī to the Questions of Askia al-Ḥãjj Muḥammad (Oxford, 1985), 9–10.

⁵ Lange, D., ‘Les rois de Gao-Sané et les Almoravides’, J. Afr. Hist., xxxii (1991), 251–75.CrossRefGoogle Scholar

⁶ Hunwick, John, ‘Gao and the Almoravids revisited: ethnicity, political change and the limits of interpretation’, J. Afr. Hist., xxxv (1994), 251–73.CrossRefGoogle Scholar

⁷ Viré, M., ‘Notes sur trois épitaphes royales de Gao’, Bulletin de l'IFAN, xx (1958), 375 n. 2.Google Scholar

⁸ de Moraes Farias, P. F., ‘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), 516–17.Google Scholar

⁹ Eds. and transl. O. Houdas and M. Delafosse (Paris, 1913).

¹⁰ Eds. and transl. O. Houdas (Paris, 1898–1900).

¹¹ Lange, , ‘Gao-Sané’, 264Google Scholar; see also n. 85.

¹² In the TF the jesere dunka is several times quoted as an important informant (94, 155/tr. 177, 276; s. a. 11/tr. 14).

¹³ ‘Gao revisited’, 270. The reference is here to the TF (NH).

¹⁴ On the other hand I was glad to notice that Hunwick accepted several of my suggestions, although he should have acknowledged the source of his reformulation of the conquest theory instead of implying that these ideas are his own: (1) the Zãghē were part of the Sanhãja diaspora - and therefore apparently not invading foreigners (‘ Gao revisited’, 269; ‘Gao-Sané’, 266–7); (2) the Zãghē may have been absorbed into the local community (‘ Gao revisited’, 252; ‘Gao-Sané’, 265); (3) the succession of the early Zãghē kings was non-dynastic (‘Gao revisited’, 265; ‘Gao-Sané’, 263);

¹⁵ Hopkins, J. and Levtzion, N., Corpus of Early Arabic Sources for West African History (Cambridge, 1981), 113Google Scholar [hereafter, Corpus].

¹⁶ Corpus, 174.

¹⁷ Ibid. 87.

¹⁸ Ibid. 85.

¹⁹ Ibid. 97.

²⁰ Hunwick's considerations concerning Kukiya (256–7) are based on four assumptions: (1) the Kukiya-traditions refer truly to the beginning of the dynastic history of Gao; (2) Kukiya was identical with Kukiya-Bentia (Desplagnes, Delafosse); (3) the shift of the Zãs from Kukiya to Gao was linked to the beginning of regular trans-Saharan trade; (4) regular trans-Saharan trade began in the second half of the eighth century. This is all very uncertain.

²¹ Furṭū, Ibn, K. ghazawãt Barnū, ed. Lange, D. (Stuttgart, 1987), 9–21/tr. 43–57Google Scholar; see also map p. 27.

²² Denham, D. et al. , Narrative of Travels and Discoveries, (2 vols.) (London, 1926) i, 214Google Scholar; Barth, H., Travels and Discoveries, (3 vols.) (New York, 1859) iii, 30.Google Scholar

²³ I may add the relevant part of Raymond Mauny's letter of 28 August 1990 that Hunwick wished me to quote: ‘[…] votre lettre relative à la question du site archéologique de Gao-Sané (Mali) et de sa disposition sur le fleuve Niger ou sur l'une des branches du bas-Tilemsi. C'est auprès de celles-ci qu'il faut placer Gao ancien et Gao Sané, tous deux sur la rive gauche du Niger, Gao ancien au confluent même du Tilemsi. La dune rose, elle, est sur la rive droite du Niger et il n'y a aucune trace de ville médievale à cet endroit. […] C'est done le Tilemsi et non le Niger qui séparent Gao ancien de Gao Sané, jusqu’a preuve du contraire […]’. Mauny, who had a keen eye for remains of human settlements, could not be misled by historical texts to the point that he would ignore geographical realities.

²⁴ Levtzion, , Ancient Ghana, 53–60.Google Scholar

²⁵ Lange, D., ‘Das alte Mali und Ghana: Der Beitrag der Oraltraditionen zur Kritik einer historiographischen Fiktion’, Historische Zeitschrift, 255 (1992), 610–23.CrossRefGoogle Scholar

²⁶ Levtzion noted this similarity (Ancient Ghana, 26). See now also Raimbault, M. and Sanogo, K., Recherches archéologigues au Mali (Paris, 1991), 63–80, 174–84, 249–58.Google Scholar

²⁷ Raimbault, and Sanogo, , Recherches, 520–2.Google Scholar

²⁸ Corpus, 77–85. It should be noted that in one instance al-Bakrī suggests that Ghana was on the Niger (Corpus, 82). Elsewhere he locates it at a distance of three days from the river (ibid. 85).

²⁹ Lange, , ‘Das alte Mali’, 615–16.Google Scholar

³⁰ M. Delafosse considers that Ghãna is derived from the Malinké term gãna, kãna, kãnda meaning ‘chef de guerre’ (La langue Mandingue [Paris, 1955] ii, 241).Google Scholar In Borgu the name gando is given to royal slaves (Lombard, J., Structures de type féodal [Paris, 1965], 36–7Google Scholar), while in Hausaland gandu designates a large or perhaps more properly a royal farm (Mischlich, A., Wörterbuch der Hausasprache [Berlin, 1906], 155Google Scholar). The form kanta is attested in Songhay as designating foreign slaves, who supposedly were ‘ royal slaves’ (de Sardan, J.-P. Olivier, Concepts et conceptions songhay-zarma [Paris, 1982], 418).Google Scholar

³¹ Lange, , ‘Das alte Mali’, 618.Google Scholar It would be unwise to follow here the a priori judgement of Hunwick (p. 261) and others that al-Idrīsī systematically places the major towns of the Sahel on the West African Nile - this opinion is disproven by the location of Gao and Ghana.

³² In correcting ‘Nabo Kantabo’ to ‘Nabo! Kontabo!’ Hunwick insinuates that I misquoted Depuis-Yacouba (p. 258 and n. 24), while actually I quoted Jean Rouch (who has the spelling I gave) and mentioned Depuis-Yacouba only as a further reference (‘ GaoSané’, 270, n. 113). He further claims (p. 258) that I consider ‘Nabo Kantabo’ a ‘praisename’ while actually I wrote that it is a ‘devise’ (motto). He is right however to criticize my translation of Sãmaqandã (p. 258) - it should rather be translated as ‘land of Sãma’ (Barth has ‘Marka-kanda’ as ‘land of the Soninke’ [Travels, iii, 702]) - although it does not falsify the opinion that also in this case Qandã stands for Kanta since even the Songhay termganda - ‘earth, land’, which Hunwick reads into Qandã (pp. 257–8), would appear to be cognate with Kanta (Gado, B., Le Zarmatarey [Niamey, 1980], 21–2Google Scholar, see also 134, 145).

³³ For the former see Doumbia, P., ‘Étude du clan des forgerons’, Bulletin du Comité d'Etudes Scientifiques et Historiques de l'AOF, xix (1936), 334–9Google Scholar and Lange, , ‘Das alte Mali’, 605–6Google Scholar, and for the latter Urvoy, Y., Histoire des populations du Soudan Central (Paris, 1936), 56–9Google Scholar; Rouch, J., Contribution à l'histoire Songhay (Dakar, 1953), 2077–8Google Scholar; de Sardan, O., Concepts, 406–12.Google Scholar

³⁴ Urvoy, , Histoire, 56–9Google Scholar; Gado, , Zarmatarey, 21, 129–34Google Scholar; de Sardan, O., Concepts, 410–11.Google Scholar

³⁵ Gado, , Zarmatarey, 126–8, 145–6Google Scholar; de Sardan, O., Concepts, 225–30, 330–1.Google Scholar

³⁶ de Sardan, O, Concepts, 287.Google Scholar

³⁷ Rouch, , Religion, 12Google Scholar; de Sardan, O., Concepts, 287.Google Scholar

³⁸ Concepts, 287.

³⁹ Ibid. 227.

⁴⁰ Gado, , Zarmatarey, 124, 126, 129Google Scholar; de Sardan, O., Concepts, 227, 406.Google Scholar

⁴¹ Lange, , ‘Gao-Sané’, 270–2Google Scholar; similarly Gado, , Zarmatarey, 13, 123–80.Google Scholar

⁴² ‘Gao-Sané’, 271–2.

⁴³ Corpus, 82; Hunwick, et al. , ‘Le géographie du Soudan d'après al-Bakrī’, Mélanges en hommage à Raymond Mauny (Paris, 1981), 420Google Scholar; he is right however to read ‘ Naghmãrata’ as ‘ Wangarãta’ (ibid. 421).

⁴⁴ Lange, , ‘Altes Mali’, 615–16.Google Scholar My disagreement with Hunwick is expressed in Ibid. n. 125.

⁴⁵ TF, 38/tr. 65.

⁴⁶ Prost, A., La langue sonay et ses dialectes (Dakar, 1956), 554Google Scholar; de Sardan, O., Concepts, 383–6.Google Scholar In the Kebbi town of Gurma there exists a local community of Wangara peasants. In Borgu the name Wangara refers to Muslim quarters of local towns (Lombard, Structures, 36).

⁴⁷ Barth, , Travels, i, 479Google Scholar; iii, 111 202, 645, 646.

⁴⁸ MacAdam, M., The Temples of Kawa, i (London, 1949), 101.Google Scholar

⁴⁹ ‘Gao revisited’, 252–7, Delafosse, , Haut-Senegal, i, 239–42Google Scholar; Rouch, , Contribution, 165–9.Google Scholar

⁵⁰ de Fabregues, B. Payre, Lexique des plates du Niger, (2nd ed., Niamey, 1979), 45.Google Scholar

⁵¹ I do not think that any of the Songhay meanings which Rouch proposes for various Zã names are acceptable. Names like ‘le jeune homme jujube’ (for Kosso-Dãrē) or ‘la main sait un peu’ (for Biyay-Kayna-Kinba) do not seem to suit rulers of a kingdom (Contribution, 174, n. 13).

⁵² Corpus, 17.

⁵³ Cuoq, J. M., Recueil des sources arabes concernant l'Afrique occidentale du VIIIe au XVIe siècle (Paris, 1975), 69.Google Scholar

⁵⁴ Lange, , ‘Gao-Sané’, 267.Google Scholar

⁵⁵ Corpus, 302, 333.

⁵⁶ Ajwibat al-Maghīlī ‘an ’as’ilat al-Amīn al-ḥãjj Muḥammad Askiyã (hereafter Ajwiba), Arabic text and English translation in Hunwick, Sharīa, 14/tr. 70.

⁵⁷ Corpus, 333.

⁵⁸ Ibid. 322.

⁵⁹ Ibid. 302.

⁶⁰ It could also be argued that ‘Zaghãy’ is closer to ‘Songhay’ than to ‘Zai’.

⁶¹ Corpus, 320, 321, 354. Lange, D., ‘Un texte de Maqrīzīsur les “races des Sūdan”’, Annates Islamologiques, XV (1979), 207 n. 3.Google Scholar

⁶² Prost, , La langue soηay, 44–50.Google Scholar

⁶³ Hutchison, J. P., A Reference Grammar of the Kanuri Language (Madison, 1981), 43Google Scholar; Kraft, C. H. and Kirk-Greene, A., Hausa (London, 1973), 124.Google Scholar

⁶⁴ TS, 2–3/tr. 4–6; TF (NH), 332–6 (the second name has been amended).

⁶⁵ Johnson, S., History of the Yorubas (London, 1921), 34, 149–52Google Scholar (Sango); Lebeuf, A.M., Principautés kotoko (Paris, 1969), 71 (Dongo).Google Scholar

⁶⁶ Patterson, J. R., Kanuri Songs (Lagos, 1926), 1, 17, 27Google Scholar (Tsongorina); Nachtigal, G., Sahara und Sudan (Leipzig, 1889) iii, 343Google Scholar (Donga).

⁶⁷ Pollet, E. and Winter, G., La société soninké (Bruxelles, 1971), 38Google Scholar(tuηka); Corpus, 79 (Tunka).

⁶⁸ Lange, , ‘Gao-Sané’, 270.Google Scholar

⁶⁹ de Sardan, O, Concepts, 122, 225–230Google Scholar; TF, 11, 94, 155/tr. 14, 177, 276.

⁷⁰ Johnson, , History, 34, 149Google Scholar; Frobenius, L., Und Afrika sprach (Berlin, 1912), 206–229.Google Scholar

⁷¹ Rouch, J., La religion et la magie songhay (Paris, 1960), 50–1Google Scholar; de Sardan, O., Concepts, 360–98.Google Scholar

⁷² Rouch, , Religion, 11Google Scholar; de Sardan, O., Concepts, 366.Google Scholar

⁷³ Wente-Lukas, R., Handbook of Ethnic Units in Nigeria (Stuttgart, 1985), 150–6Google Scholar; Chapelle, J., Nomades noirs du Sahara (Paris, 1957), 341–4Google Scholar; Alojaly, Gh., Lexique TouaregFrançais (Copenhague, 1980), 211.Google Scholar

⁷⁴ Wente-Lukas, , Handbook, 150.Google Scholar

⁷⁵ I am grateful to Philip Shea for his valuable comments on Hausa history and on the Yoruba Sango.

⁷⁶ History, 3–4.

⁷⁷ Field notes, Kareshin, April 1993. The Achifawa live south-east of Kebbi and east Of Yauri.

⁷⁸ Soden, W. v., Akkadisches Handwörterbuch (2nd ed., Wiesbaden, 1985), 1163.Google Scholar

⁷⁹ Seux, M.-J., ‘Konigtum’, Reallexikon der Assyriologie, vi, 169.Google Scholar

⁸⁰ Harris, P. G., Sokoto Provincial Gazetteer, typescript (Sokoto, 1938), 234Google Scholar (the translation of the Kebbi chronicle is here slightly truncated).

⁸¹ Arnett, E. J., The Rise of the Sokoto Fulani (Kano, 1922), 13.Google Scholar

⁸² Lange, D., ‘Das fruhe Kebbi und Mali’, Z. der Deut. Morgenl. Gesellsch., cxli (1991), 155–6, 159.Google Scholar

⁸³ Dan Ayi, Argungu, April 1990.

⁸⁴ Local informants, Argungu, Oct. 1989.

⁸⁵ Barth, , Travels, iii, 643.Google Scholar The Mossi call the present Songhay ‘Marense’, sing. ‘Marenga’ (ibid.; O. de Sardan, Concepts, 290–1).

⁸⁶ Barth, , Travels, iii, 154Google Scholar and below.

⁸⁷ Nicolaï, R., Les dialectes du songhay (Paris, 1981), 14, 52.Google Scholar Hunwick refers to the north-eastern Songhay dialects but for reasons not stated denies that this is an indication of an earlier more eastern concentration of Songhay-Zarma (' Gao revisited', 255 n. 10).

⁸⁸ Harris, , Sokoto, 231.Google Scholar

⁸⁹ Prost, , Langue, 242Google Scholar; de Sardan, O., Concepts, 222.Google Scholar

⁹⁰ Dan Ayi, Argungu, April 1990.

⁹¹ Local informants, Argungu, Birnin Kebbi, Oct. 1989.

⁹² The names Sarkawa and Kabawa were earlier synonyms (Harris, P. G., ‘Notes on Yauri (Sokoto Province), Nigeria’, J. Roy. Anthrop. Inst. G. Brit, and Ireland, lx (1930), 291Google Scholar, and own data).

⁹³ Lange, , ‘Frühes Kebbi’, 155–9Google Scholar;, ‘Hausa-Traditionen’, 62–4.

⁹⁴ The traveller Heinrich Barth had already considered the possibility that Songhay was the speech of Kebbi before Hausa (Reisen und Entdeckungen [Gotha, 1858] iv, 215Google Scholar [not included in Travels, iii, 153]); see also Sutton, J., ‘Towards a less orthodox history of Hausaland’, Journal of African History, xx, 1 (1979), 190–1.Google Scholar Barth heard that in Argungu there was a holy tree called ‘tunka’ (Travels, iii, 638).

⁹⁵ TS, 5–6/tr. 9–12.

⁹⁶ TF (NH), 334.

⁹⁷ Ibn Khaldūn, K. al-‘ibar, tr. in Corpus, 323, 334.

⁹⁸ TF (NH), 335.

⁹⁹ Therefore Ch. Monteil thought that Songhay was subdued twice by Mali: at first by Mansa Walī (c. 1255–70) and next, some time after the revolt of ‘Alī Kolon, by Mansa Sãkūra (c. 1285–1300) (‘Les empires du Mali’, Bull. Com. Ét. Hist. Scient. de l'AOF, xii [1929], 365–72). He was followed by Rouch, , Contribution, 175–6Google Scholar, and Levtzion, , Ancient Ghana, 75–6.Google Scholar

¹⁰⁰ TF (NH), 335; Ibn Hajar al-‘Asqalãnī, al-Durar al-kãmina, part, transl. in Cuoq, , Recueil, 394.Google Scholar

¹⁰¹ In fact, it is often overlooked that the chroniclers depict Sunni ‘Alī as a bloodthirsty tyrant - and hence also as a bad Muslim - rather than as an unbeliever (TS, 64–71/tr. 103–116; TF, 43–53/tr. 81–100).

¹⁰² Al-Maghīlī, , Ajwiba, 13–14/tr. 69–70.Google Scholar

¹⁰³ See Éaulard, A., Description de l'Afrique (Paris, 1956) ii, 462.Google Scholar Marmol calls Sunni ‘Aī a Lemtuna (apud Delafosse, , Haut-Sénégal, ii, 61).Google Scholar

¹⁰⁴ Sharīa, 69 n. 1.

¹⁰⁵ According to al-Maghīlī Sunni ‘Alī succeeded his father whereas the kinglist has his father as either his fifth (TF) or his eighth predecessor (TS). In fact, on the basis of the information provided by the Ta'rīkhs it would appear to be more likely that Sunni ‘Alī was the successor of Sulaymãn Dendi.

¹⁰⁶ TF (NH), 334.

¹⁰⁷ The author of TF (NH), 327–8, makes it clear that his writing was commissioned by Askiya Dawud b. Hãrūn (1657–69).

¹⁰⁸ I.e. 400 km downstream from Kukiya-Bentia.

¹⁰⁹ Dupuis-Yacouba, , ‘Légendes de Farang’, in Desplagnes, L., Le plateau central nigérien (Paris, 1907), 383–450Google Scholar; Harris, , ‘Notes on Yauri’, 288–9Google Scholar; Rouch, , Contribution, 165–72.Google Scholar

¹¹⁰ Delafosse, , Haut-Sénégal, i, 241–4Google Scholar; Rouch, , Contribution, 165–8.Google Scholar

¹¹¹ de Sardan, O, Concepts, 365.Google Scholar

¹¹² This would mean that the Gao list includes not only Zã and Sunni kings but also rulers who do not belong to the two old royal clans of Gao. As noted above this seems also to be the case with Mãkara Komsū. It would also mean that the 28 or 33 Sunni rulers have to be squeezed into the period extending from c. 1400 to 1464, or else Ibrãhīm Kubē was purposely placed behind ‘Alī Kolon and Silman Nãri.

¹¹³ Boulnois, J. and Hama, B., Empire de Gao, histoire, coutumes et magie (Paris, 1954), 104Google Scholar; Rouch, , Religion, 11.Google Scholar

¹¹⁴ Harris, P. G. has the Sarkawa come from Songhay (‘The Kebbi fishermen (Sokoto province, Nigeria)’, J. Roy. Anthrop. Inst. G. Brit. and Ireland, lxxii (1942), 23)Google Scholar, but this is not the tradition of the Dankanawa themselves (Dan Ayi and others, Argungu, April 1990).

¹¹⁵ For Mali cf. Ibn Baṭṭūṭa in Corpus, 284, 290; for Songhay: Bãghana-farma, Bingafarma, Dendi-fãrī, Kabara-farma etc.

¹¹⁶ Harris, , ‘Notes on Yauri’, 289.Google Scholar

¹¹⁷ Perhaps from Arabic nã’ib - ‘substitute’. The Egyptian nb-‘master’ should also be considered.

¹¹⁸ In Crone, G. R., The Voyages of Cadamosto (London, 1937), 93–5.Google Scholar

¹¹⁹ Lange, , ‘Frühes Kebbi’, 142–8.Google Scholar

¹²⁰ In: de La Roncière, Ch., La découverte de l'Afrique (Cairo, 1924) i, 156.Google Scholar

¹²¹ Person, Y., ‘Le Moyen Niger au XVe siècle d'après des documents européens’, Notes Africaines, lxxvii (1958), 46.Google Scholar

¹²² TF, 45/tr. 85. Kukiya corresponds to Songhay ‘gungu’ - island.

¹²³ T. al-Fattãsh, 46/tr. 89.

¹²⁴ Perhaps the etymology provided by the TF according to which the title shī (=soni/sunni) meant ‘vicar’ (43/tr. 82) corresponds to an African interpretation of the Arabic sunnī Indeed, the soni title having the meaning of the TF can be found in various kingdoms between the Nile and the Niger.

¹²⁵ The rather neutral and non-prestigious title of malik of the Gao-Sané epitaphs which attracted the attention of Hunwick (p. 18) was probably meant to indicate the inferior status of the Zãghē in relation to the distant, but highly respected Almoravids.

¹²⁶ K. al-‘ibar, see Corpus, 336, 338–9. Levtzion following H. Lhote interprets Ibn Khaldūn's reference to Takedda as a mistake for Tadmekka (Ancient Ghana, 77–8).

¹²⁷ Corpus, 322.

¹²⁸ It may be noted that the Abbasid caliphs after having lost their worldly power in 945 survived for even a longer period in the shadow of powerful secular rulers and that their history was also punctuated by occasional reversals of fortune until finally they were deposed in 1517 (Lewis, B., “Abbãsides’, Encyclopédie de l'Islam [new edition], i, 20–2).Google Scholar

¹²⁹ TS, 76/tr. 125; Hunwick, , ‘Gao revisited’, 273.Google Scholar

¹³⁰ Hunwick, , ‘Gao revisited’, 256–7.Google Scholar

¹³¹ TS 4/tr. 6.

¹³² TS, 3–5/tr. 5–8; TF (NH), 326.

¹³³ According to Harris the name Gunga applied to an ethnic group living near Yauri on the banks of the Niger is derived from the name of the old Kebbi capital (‘ Notes on Yauri’, 291).

¹³⁴ Rouch cogently pointed out the similarities between the state tradition of Zã al-Ayaman and the clan tradition of Bote, Faran Maka (Contribution, 170–1).Google Scholar He did not realize however that the state tradition (recorded in the seventeenth century) was derived from the clan tradition (recorded in the twentieth century).

¹³⁵ It should be noted that many Sorko are ritualists, not fishermen (Boulnois, and Hama, , Empire, 65–6Google Scholar; de Sardan, O., Concepts, 343).Google Scholar Anthropologists assume that those who live far from the river are descendants of fishermen, but this is not necessarily so. The name ‘Sorko’ itself, perhaps derived from sarki (Hausa and pre-Hausa: king), seems to indicate a close connection of this group to the king. ‘Koy’ and ‘Ki’, the royal titles of Songhay and Borgu, could be cognate to the second syllable of sar-ki.