1 C. Giesing and V. Vydrine, Ta:rikh Mandinka de Bijini (Guinée-Bissau), La Mémoire des Mandinka et des Sòoninkee du Kaabu (Leiden, 2007); E. L. Osborn, Our New Husbands Are Here: Households, Gender, and Politics in a West African State from the Slave Trade to Colonial Rule (Athens, OH, 2011).

2 J. C. Miller, Kings and Kinsmen: Early Mbundu States in Angola (Oxford, 1976), 7.

3 The title of my article is inspired by Oosten's magnificent study of marriage strategies among the Indo-European gods in his The War of the Gods: The Social Code in Indo-European Mythology (London, 1985). For Launay, see R. Launay, Beyond the Stream: Islam and Society in a West African Town (Berkeley, CA, 1992).

4 Ibid. 55.

5 Osborn, New Husbands, 43, mentions this principle, but analytically she takes a different road from the one here.

6 Miller argues that, around 1970, anthropologists demonstrated the political logics of the civilizing hero, who, as an outsider, can accomplish legitimacy not tainted by ‘any connection with local interest groups’ (Miller, Kings and Kinsmen, 9). For an argument that a complex of mask dances is, historically, at the origin of the Sunjata epic (and not a set of events in which human actors featured), see J. Jansen, ‘Masking Sunjata: a hermeneutical critique’, History in Africa, 27 (2000), 131–41.

7 Miller, Kings and Kinsmen, 12.

8 Ibid. 16, emphasis added.

9 Ibid. 9.

10 For a research agenda on Kaabu, see Green, Toby, ‘Architects of knowledge, builders of power: constructing the Kaabu “Empire”, 16th–17th centuries’, Mande Studies, 11 (2009), 91–112Google Scholar.

11 For details, see Giesing and Vydrine, Ta:rikh, 32 et seq.

12 Giesing and Vydrine, Ta:rikh, 77–81 for the narrative, 179 et seq. for the analysis.

13 Blacksmiths had made a web of iron and put it at the entrance to the cave, thus preventing Tenenbaa from returning inside. The authors claim that she also clearly represents Islam through her name (Ibid. 180).

14 For other examples of this code, in relation to the organization of armies for communal defence, see Jansen, J., ‘The younger brother and the stranger: in search of a status discourse for Mande’, Cahiers d'Etudes africaines, 36:144 (1996), 659–88CrossRefGoogle Scholar. Giesing and Vydrine, Ta:rikh, 182, might be a reference to this idea. The warrior groups implicitly emphasize badenya, a central term which means both ‘harmony’ and ‘descent from the same mother’; badenya thus represents centripetal forces, in contrast to fadenya, ‘children from the same father (but different mother)’. See C. S. Bird and M. B. Kendall, ‘The Mande hero: text and context’, in I. Karp and C. S. Bird (eds.), Explorations in African Systems of Thought (Bloomington, IN, 1987), 13–26.

15 Giesing and Vydrine, Ta:rikh, 176–7, 179.

16 Sarah Brett-Smith demonstrated that the porcupine metaphor for Sogolon is related to a (description of a) Komo mask, which, she argues convincingly, represents a vagina, in Brett-Smith, S., ‘The mouth of the Komo’, RES: Anthropology and Aesthetics, 31 (1997), 71–96Google Scholar, esp. 87. This analysis adds evidence to the thesis that the Sunjata epic has evolved from mask performances (see Jansen, ‘Masking Sunjata’).

17 Some variants describe the hunters’ attempt to have sexual intercourse with Sogolon very explicitly and with a great sense of humour. See the lists by S. P. Belcher, Epic Traditions of Africa (Bloomington, IN, 1999), 94–5. The list in chapter five shows that this theme is present in almost all published versions of the epic.

18 See, for instance, the genealogies in Jansen, ‘Younger brother’.

19 Matrilineality says little about female power or the status of women. In the Western Sudan, princesses from matrilineal societies do not become queens; their brothers and their sons become kings. Power and kingship stay in the hands of men, but it is the king's sister who will give birth to the king's successor.

20 Paul Lovejoy wrote in 1989: ‘No one has argued as much, but it may be that matrilineality and the export trade were interrelated. They certainly reinforced each other.’ See Lovejoy, P. E., ‘The impact of the Atlantic slave trade on Africa: a review of the literature’, The Journal of African History, 30:3 (1989), 365–94CrossRefGoogle Scholar, esp. 388.

21 I refer here to work by Wyatt MacGaffey on Congo; I would like to thank Toby Green for sharing with me his knowledge and insights on descent systems in West Africa. One notes that for Kaabu the king's succession seems to have been possible through double descent, as Hecquard noted in the mid-nineteenth century that a legitimate successor could only be the son of a ñànco (a non-Muslim warrior), either of his own wife or of his sister (for details, see below).

22 N. Levtzion, ‘Islam in the Bilad al-Sudan to 1800’, in N. Levtzion and R. L. Pouwels (eds.), The History of Islam in Africa (Oxford, 2000), 63–91.

23 Ibid. 80, emphasis added.

24 Ibid. 85.

25 Ibid. 79.

26 Ibid.

27 Ibid. For Levtzion the label ‘Muslim’ refers to the religious side of this identity, while I postulate that a political-economic-military status component matters a great deal.

28 A representative collection of such narratives on jihads is M. Schaffer, Djinns, Stars and Warriors: Mandinka Legends from Pakao, Senegal (Leiden, 2003).

29 For a study with a similar multidimensional analysis, see J. H. Hanson, Migration, Jihad, and Muslim Authority in West Africa: The Futanke Colonies in Karta (Bloomington, IN, 1996).

30 When analyzing Kaabu's political system, the authors initially appear to subscribe to the idea that the genealogical accounts describe Kaabu's political system, since they describe it as a ‘discours établi sur la pluralité et oscillation des cultures, langues et identités qui est souvent introduit par le thème du mariage entre une mère autochtone et un père immigré’ (Giesing and Vydrine, Ta:rikh, 286). However, they then continue their analysis with the ‘migrant father’ as an object of study, thus bypassing the marriage obligations and implicitly suggesting that the account refers to a historical foundation.

31 Giesling and Vydrine here refer to L. H. Hecquard, Voyage sur la côté et dans l'intérieur de l'Afrique occidentale (Paris, 1853).

32 Giesing and Vydrine, Ta:rikh, 286.

33 The term sòoninkee is central to the authors’ analysis. According to Giesing and Vydrine, this label refers to the practice of performing libations/sacrifices and drinking alcohol for ritual reasons. Sòoninkee are related to ownership of the earth and have the capacity to speak with the jinns of the earth. Though in many Mande languages the term means ‘pagan’, linguistic analysis has shown that its original meaning was ‘noble warrior’ (Giesing and Vydrine, Ta:rikh, 382). See also Levtzion, ‘Islam’, 79.

34 Quotes from Giesing and Vydrine, Ta:rikh, 284–7, published with the permission of Brill, the publisher.

35 Kankan is not mentioned in pre-nineteenth-century European written sources, such as Mungo Park's Travels in the Interior Districts of Africa (London, 1799) (although Park travelled close to Kankan) and texts produced at European trading settlements along the West African Atlantic Coast. See also Osborn, New Husbands, 212–33, fns. 58 and 59.

36 Suwarian Islam advocates, among other things, an absence of proselytizing activities in relations with unbelievers, a rejection of jihad as a means of conversion, and an acceptance of non-Muslim rulers. See I. Wilks, ‘The Juula and the expansion of Islam into the forest’, in N. Levtzion and R. L. Pouwels (eds.), The History of Islam in Africa, 93–115, esp. 96–8.

37 Osborn, New Husbands, 26, sees this as a sign of tolerance. To me, both marriages by Kaba Laye represent ‘politically appropriate’ marriages, although the second one is with a woman more highly placed than the first. This suggests a positive judgement of hypergamy. But one also notes that neither of the marriages in which the Muslim man marries the warrior woman leads directly to political success, since his son professes a Muslim identity, while a warrior identity is a prerequisite for political success (see Hecquard supra on Kaabu).

38 Based on quotes from Osborn, New Husbands, 27–8.

39 Summarized from Ibid. 34–5.

40 Could an African political leader and family chief ever remain unmarried in practice? I doubt it, although this is irrelevant for my argument. I must admit, though, that reading an oral tradition about a political leader who refused to marry was a revelation for me: it was the moment when the first seeds of this argument were sown. Ideologically, the marriage is needed to transfer rule to the next generation; producing offspring through a slave woman is inappropriate as these children cannot be accepted as legitimate inheritors.

41 Referring to the work of Lansiné Kaba (himself originally from Kankan), Osborn notes (New Husbands, 56) that Alfa Kabiné ‘glossed differences of ethnicity and origin that may have differentiated them from Baté's lambila, or host families’. I note that in her description of the nabaya system Osborn takes an ahistorical and too simplistic view of issues of land entitlements in Manding and beyond when she writes on the same page about the nabaya system that ‘This practice … put into reverse marital practices that are characteristic of Mande and many other groups in West Africa.’

42 Osborn, New Husbands, 56.

43 I see a similarity with Miller, Kings and Kinsmen, 14, who notes for his material ‘no developmental chain of related events set on a time-based continuum’.

44 Toby Green has perceptively remarked that ‘what is foreshadowed is precisely Islam, not Islam as status in a matrimonial system’ (personal communication, 14 May 2014). Kankan's case is indeed, to me, a conceptual/narrational link between the warriors’ states and the Islamic reformist movements of the nineteenth century.

45 Osborn, New Husbands, 57.

46 According to Osborn (New Husbands, 57), it was mainly merchants and travellers who settled in Kankan. Her evidence consists of interviews she had with people in Kankan. Note that the link between warfare and trade was always close in pre-1850 West Africa.

47 I am referring here not only to the large jihadi armies, but also to smaller ‘migrating’ armies. For a polity where Muslim strangers introduced fire weapons and imposed rule but did not interfere with village politics, see Narena, for which oral traditions have been published in S. Camara and J. Jansen (eds.), La Geste de Nankoman: Textes sur la Fondation de Naréna (Mali) (Leiden, 1997). For a detailed study on how Kankan's northern neighbours, the rulers of Kangaba, incorporated these ‘migrating’ armies into ranks, see Jansen, J., ‘In defense of Mali's gold: the political and military organization of the Northern Upper Niger, c. 1650 – c. 1850’, The Journal of West African History, 1:1 (2015), 1–36CrossRefGoogle Scholar.

48 Osborn suggests that Kankan was a pacifist and ‘inward-looking state’, but I believe that such a state would never have survived in that era. In my reading her rich and suggestive material supports an alternative conclusion. See Osborn, New Husbands, 44.

49 I would like to express my thanks to Peter Mark for mentioning these two different types of arms as well as the time-depth of this process.

50 Jansen, ‘Younger brother’, demonstrates that the genealogies of Sunjata's male descendants – which feature as praise lines in the Sunjata epic – should be read as claims to leadership in an army organization in eighteenth-century society. The present analysis demonstrates, in addition to this, that the matrimonial relationships of Sunjata's father also refer to post-medieval political issues. This explains why the genealogies in the epic are so different in both structure and content from those we know from medieval sources. For an analysis of medieval genealogies of Mali's rulers see Austen, R. A. and Jansen, J., ‘History, oral transmission and structure in Ibn Khaldun's chronology of Mali rulers’, History in Africa, 23 (1996), 17–28CrossRefGoogle Scholar.

51 For the data on the Wolof kings, see Levtzion, ‘Islam’, 78, para. 4; 79, para. 2.

52 V. Kamara, De Kroniek van de Kamara – Een Verhaal uit Guinée (Rijswijk, 1998). This is a Dutch translation by Marloes Janson of an unpublished English transcription and translation, by Timothy Geysbeek and Jobba Kamara, of a recording made by Vase Kamara (on 21 Dec. 1985) at the request of Timothy Geysbeek (information from Kamara, De Kroniek, 11).

53 Ibid. 43.

54 Ibid. 58.

55 For this classical observation, see Radcliffe-Brown, A. R., ‘On joking relationships’, Africa: Journal of the International African Institute, 13:3 (1940), 195–210CrossRefGoogle Scholar.

56 The succession to the throne of the Mali empire is a topic that has received much attention. Sources suggest a dynamic hybrid encompassing components of patrilineality, matrilineality, double descent, collateral succession, and succession from father to son. Personally, I am fascinated by the fact that the author of the Tarikh el-Fettach describes the ruler of Djenné as a humble servant of the ruler of Mali, because the Djenné delegation had to pay taxes to the wife of Mali's ruler and never met the ruler himself. This is indeed almost an insult for the Djenné delegation, who must have been patrilineal Muslims. However, if the Mali empire at that time was characterized by matrilineal succession, the transfer of money might have been made by the Mali ruler in a dignified and diplomatically respectful way. The interpretation of this anecdote is of importance for our judgment of the power of the Mali empire in the sixteenth century. For sources, and some discussion, see M. Ly-Tall, Contributions à l'Histoire de l'Empire du Mali (XIIIe–XVIe siècles) (Dakar, 1977), 52, for the Tarikh el-Fettach reference to Djenné.

57 Moussadougou is called Musadu in Kamara, De Kroniek, which gives variations on some of the story themes discussed in Djiguiba Camara's typescript below. I would like to thank Elara Bertho for giving me permission to use this typescript, which covers the origins of the Kamara up to Samori's wars. The source was written in 1955 by Djiguiba Camara, a ‘Chef Supérieur du Canton’ and one of Person's key informants, who claims to be, in this text, a direct descendant of Farin Kamara. Elara Bertho and Marie Rodet are currently preparing a critical text edition of this typescript.

58 For discussions on the origins of the Sunjata epic with an emphasis on its text editions, see R. A. Austen (ed.), In Search of Sunjata – The Mande Oral Epic as History, Literature, and Performance (Bloomington, IN, 1996). But see fn. 6 and fn. 16 for alternative ideas.

59 R. L. Roberts, Warriors, Merchants, and Slaves: The State and the Economy in the Middle Niger Valley, 1700–1914 (Stanford, CA, 1987), 19, emphasis added.

60 Roberts does not describe in detail how this economy was ‘embedded’ in culture and society, though he describes his model of the warrior state as an embedded model, thus linking it to Karl Polanyi's substantivist approach. See Roberts, R. L., ‘Production and reproduction of warrior states: Segu Bambara and Segu Tokolor, c. 1712–1890’, International Journal of African Historical Studies, 13:3 (1980), 389–419CrossRefGoogle Scholar. In retrospect I read this model as an alternative to the formal modes of economic production models that dominated the study of African history in the 1970s, the glory days of neo-Marxism. These deterministic models emphasized, Roberts argues, technological innovations (such as ‘the use of horses’ or ‘the introduction of firearms’), ecological circumstances (such as ‘drought’ or ‘geographical barriers’), and ecological specializations (such as particular forms of trade). However, the human element was absent in these models.

61 Cf. the polity organized by the rulers of Kangaba (Upper Niger), in Jansen, ‘In defense of Mali's gold’.

62 One anonymous reviewer made the intriguing suggestion that the reluctance to marry a Muslim, in many narratives, might be read as expressing an attitude directed more to the Atlantic slave trade and less to trans-Saharan trade. This is a challenging idea that needs further research, including Toby Green, The Rise of the Trans-Atlantic Slave Trade in Western Africa, 1300–1589 (Cambridge, 2012). My current analysis, however, hasn't the strength to cover this thesis, since it analyzes (not) marrying a Muslim as an expression of political agency, not as evidence of a factual marriage preference.