Hausa Songs in Algeria: sounds of trans-Saharan continuity and rupture

ABSTRACT What can North African music, as a system of embodied knowledge, tell us about the trans-Saharan movements of people and practices? What could it mean to listen to such a history of continuity and rupture? This essay details the existence of two bodies of Hausa songs still performed today in Algeria within the Dīwān of Sīdī Bilāl, a music and trance ritual tradition that originated, coalesced, and developed out of the trans-Saharan slave trade. These songs, categorised as the Hausawiyyn and Migzawiyyn suites, feature Hausa ritual objects, song texts, musical aesthetics, and spirit names from the Hausa bori (ceremony) pantheon among many other sub-Saharan echoes. However, dīwān communities rarely recognise these precise connections and simply consider them generically ‘sub-Saharan’ (sūdānī). Nevertheless, because the Hausa repertoires express the most purely sūdānī qualities of the entire ritual corpus, they serve as the primary example of both admirable ‘black African’ authenticity (and, therefore, ritual power) and risky alterity. Such viewpoints serve as a basis for crises about how practising Muslims should engage with the supposed ‘pre-Islamic’ origins of dīwān while preventing ancestors’ histories from disappearing. Informed by recent (2013–2016) first-hand ethnographic fieldwork with dīwān communities as well as both primary and secondary historical sources, this essay is not only the first research to document these Hausa songs in Algeria but also highlights how and why these songs matter in Algerian history and to local communities today.


Introduction
What can North African music, as a system of knowledge, tell us about the trans-Saharan movements of people and practices? What could it mean to listen to such a history of continuity and rupture? In this essay, I detail the existence of two bodies of Hausa songs still performed today in Algeria with Hausa ritual objects, song texts, and musical aesthetics. 1 As communities in Algeria debate the nature of these centuries-old songs, problematise their texts, and claim or reject them as part of their own lineages in various ritual and festival settings, the contemporary meanings of trans-Saharan movements resound from an emic point of view.
The Hausawiyyn and Migzawiyyn song suites belong to the robust musical repertoire of Algerian dīwān communities. Dīwān, or Dīwān of Sīdī Bilāl, remains one of the many ṭ uruq (paths, pl.) in Algeria grounded in Sufi epistemologies and ritual practices. What makes dīwān practice unique, however, is its origins in sub-Saharan Africa and its emergence in North Africa via trans-Saharan trade routes. Hausa songs in dīwān represent just one ethnolinguistic category of many suites in the dīwān ritual musical corpus. Although most of the songs in the corpus assert sub-Saharan origins, the largest degree of trans-Saharan connection is traceable to Hausa origins through words, phrases, ritual objects, names of spirits from the Hausa spirit pantheon, and musical aesthetics. Hence, one could think of these songs as bridges between chasms of time and space.
McDougall and Scheele (2012), as well as Lydon (2009), have been particularly influential to a shifting view of the Sahara as a continuum and dynamic 'bridge' rather than as a north-south, religious and racial barrier between 'black' and 'white' and 'Muslim' and 'non-Muslim' (Lydon 2009, 41). Such a black-white binary, very much a product of French intellectual colonialism, continues to persist discursively around and within dīwān communities in Algeria, who largely consider dīwān a sub-Saharan, animist (sometimes 'voodoo'), and pre-Islamic practice prior to its emplacement in North Africa. 2 While the Sahara in the histories of dīwān did function as a bridge to some extent, trans-Saharan passage involved great trauma, loss, suppression, and suffering. As follows, it is important to consider how Saharan crossings not only fostered connections but also how they destroyed them by reconfiguring and obliterating subjectivities.
Indeed, dīwān practice and discourse suggest uneven and fractured networks of practice and aesthetics, demonstrating both trans-Saharan connection and disconnection, remembering and forgetting, construction and destruction, and power and powerlessness. Today, dīwān musicians do not understand the non-Arabic language they sing in and most do not recognise the direct, trans-Saharan, Hausa connections in the dīwān ritual; rather, they recognise them generally as 'sub-Saharan' (sūdānī). Nevertheless, dīwān communities vehemently debate Hausa songs and texts and struggles emerge around whether these songs should still be performed and which troupes claim 'authentic', ancestral links that merit the right to perform them. In this way, Hausa songs also map dynamic moral, affective, and religious territories, particularly when the songs begin to circulate outside their original ritual context.
Resonating with James McDougall's call that trans-Saharan studies would do well to consider 'spaces in between ' (McDougall and Scheele 2012, 75), these Hausa songs, indeed, lie somewhere in between memory and forgetting, in between who ancestors were and who their descendants are, and in between the 'old ways' and the 'new.' That is to say, while Hausa songs in Algeria are valuable sonic and performative vestiges of trans-Saharan movements and histories in Algeria, they are much more than this. As follows, this essay draws on recent (2013)(2014)(2015)(2016) first-hand ethnographic fieldwork with dīwān communities as well as primary and secondary historical sources to both document these Hausa songs in Algeria while addressing how and why these songs matter today. 3

Dīwān background
Dīwān (lit. 'assembly'), a music and trance ritual tradition, originated, coalesced, and developed out of the trans-Saharan slave trade with the displacement of various sub-Saharan ethnolinguistic groups. 4 During three centuries of Ottoman rule, these communities were heavily influenced by the local, popular religious practices and socio-political organisation of Sufi lineages. Situated in the Ottoman context, dīwān developed into a stratified Afro-Maghrebi ritual and music practice predicated on many of the same structures of other musical traditions within the fold of popular Islam in North Africa: saint veneration, trance, and ritual healing. 5 Dīwān insiders, or ūlād dīwān (lit. 'children of dīwān'), for the most part, descend from these many sub-Saharan ethnolinguistic groups and they identify Bilāl, the first muezzin (caller to prayer) and a former Abyssinian slave, as their spiritual father (al-'abb al-rūḥ ī). Following the Sufi pattern of taking the name of the leader or shaykh, ūlād dīwān refer to their community as 'Bilāliyya'. 6 The dīwān ritual typically lasts six to eight hours and unfolds through dozens of song suites. Ordered by hagiography and ethnolinguistic origin, the suites contain songs for historical figures, saints, spirits, and, some say, the jnūn (supernatural entities widespread in Islam). While some of the songs (brāj, pl; borj, sing) cultivate trance for those adepts so inclined, all of the songs require distinct sensory-aesthetic protocols. Particular physical movements or ritual actions must be performed during the brāj to depict the character of the song's namesake (e.g. a saint, spirit, prophet) such as donning coloured cloaks (abāyat), dancing with special ritual knives or hide whips (bulālat), utilising incense, and distributing specific foods (e.g. boiled eggs, hard candies, rwīna) that need to be consumed during the brāj. 7 Finally, it is critical to note one very fundamental way in which 'brāj' differ from 'songs' in English. While some brāj narrate the story of a human, historical figure like a saint, and, ideally, produce the atmosphere (ḥ āl) associated with that saint's personality (Turner 2020b), brāj for non-human spirits or jnūn function at a different register. In the latter case, these brāj sonically invoke or 'call' (ʿaīt) non-human entities. Because brāj and non-human entities are made up of the same materialenergyparticular sounds resonate nonhuman entities, materialising them. In other words, non-human entities exist materially as their songs: spirit/jinn and the sound/music collapse ontologically (Turner 2021).
The sub-Saharan musical aesthetics in dīwān pointedly reenact dīwān's trans-Saharan origins. All of the musical elements, from the instrument construction and sound to the compositional qualities of the songs, stand out as 'other' in their North African context. Dīwān aesthetics share substantially more with sub-Saharan African musics than with North African or Arabic musics. For example, the primary ritual musical instrument, the ginbrī, is a fully transpierced lute built from a long, rectangular wooden box covered in camel skin, which acts as a resonator for the three gut strings suspended over a wooden bridge tied at the distal end of the neck. The ginbrī descends from sub-Saharan lutes like the Hausa molo, Bambara ngoni, and Fulani hoddu, comparable in its construction, tuning (pentatonic melodic modes), sound quality, and technical playing style. 8 Dīwān brāj, both undetermined in terms of age and composers, diverge in their musical form: rather than employing verses or refrains with forward moving temporality, they are cyclical in nature, utilising repetition as a key musical technique. Brāj phrases return again and again, intensifying and gradually accelerating with additional layers of vocal parts or changes in instrument volume, giving brāj a spiralling, intensifying quality. The elastic feel of musical time in dīwān brāj differs from the Algerian musical milieu: metric cycles (like 'measures') are not evenly divided within beats but offset with a lilt; for example, duple groupings in time may consist of articulations with durations of 47:53 rather than 50:50. Because of this 'lilt', brāj can be simultaneously divided into groups of two or three, and, thus, can be heard in different ways depending on how one hears the recurring emphasis. 9 Second only to the moqedm who oversees the ritual proceedings, the mʿallem possesses all musical knowledge; he is the 'master' of the quintessential ginbrī. Sitting beside the mʿallem, the kuyu bungua term borrowed from Songhayenjoys the next most important role as the the main singer. 10 The kuyu bungu leads the 'call' part of the call-and-response singing while directing the chorus of six to ten response singers (rqīza or genēdīz). Each response singer also plays a pair of qrāqeb, double-headed metal clappers, resulting in a booming choral response driven by sonically piercing percussion. In all dīwān brāj (songs), the most important aspect is the main theme, the rās el-borj ('head of the borj'), responsible for identifying brāj, particularly when the name of the borj might vary from region to region, as we will see below. 11 On this note, while debates fluctuate about the song texts and meanings of songs, the melodies and themes of songs persist over long distances and periods of time. 12 In this way, musical knowledge accumulates as sonic, embodied knowledge that can sometimes be traced in ways that language cannot be.

Dīwān and its sub-Saharan roots and traces
Because it emerged via the displacement, encounters, and adaptation of sub-Saharan populations in the north, dīwān naturally embodies the epistemologies and supernatural worlds that people brought with them. It is well understood that Islam and animism or other spiritual or religious traditions often did, and still do, exist together in various arrangements. Such 'syncretic' amalgamations consistently reverberate in dīwān. 13 For example, several song texts that honour religious figures flow directly into texts about an 'animist' spirit; other songs reference both a religious historical figure, such as Moses ('Mūsa'), and a jinn by the same name. Nevertheless, despite the long and complex history of Islam in sub-Saharan Africa, local discourses both within and outside of dīwān communities in Algeria posit that dīwān's origins rest in 'animist, black Africa' but that dīwān was modified over time with the gradual Islamisation of sub-Saharan peoples upon their arrival in North Africa and the subsequent 'cleansing' of non-Islamic practices. 14 Because the Hausawiyyn and Migzawiyyn repertoires locally represent the most purely sūdānī (sub-Saharan) of all the songs, they serve as the primary vestige of both admirable 'black African' authenticity (and, therefore, ritual power) and risky alterity.
Dīwān discourse asserts that its ritual repertoire draws from seven sūdānī 'tribes' (using either tribus in French or qabāʿil, plural; qabīla, singular in Arabic), yet without consensus on which seven groups comprise it. The most commonly cited are Hausa (specifically Katsina in Algiers), Songhay, Bambara, Fulani/Peul, Bornu, Boussou, Gurma, and Zozo, in decreasing order of occurrence. The gradual coalescence of these varying groups from ethnolinguistic formations into somewhat connected 'communities' in Algeria leads us to the term ūlād dīwān, ūlād being a common referent of tribal affiliation. 15 Despite this narrative, only Hausa, Bornu, and Bambara/ Bamana ethnolinguistic groups retain possible connections through songs in dīwān. 16 On rare occasion, my interlocutors stated that the borj duo, 'Gurma-Jayba' (two brāj always played back to back), was associated with the Gurma tribe from the inner Niger River Bend, between the Mossi states and the Sokoto Caliphate (Lovejoy 2000, 156). However, Gurma-Jayba is classified by musicians under a special kind of Baḥ riyya spirit pantheon of rivers and to some groups qualifies as 'Bambara' (see also Andrews 1903, 9). Jallūl Moṭ am, the ritual leader (moqedm) of the dominant dīwān community in Saida, described Baḥ riyya as a 'tribe' (tribu, using French), throwing into question what is meant by the term exactly: assemblages of spirits, humans, or simply qualities or personages?
That said, it is important to note here that politics of ethnicity complicate matters. 17 An identification of 'tribe' as such might not necessarily indicate the origin of the brāj. It could also reference the behaviour and qualities of the spirits or personages depicted in the brāj. 18 Ethnolinguistic origin can be difficult, if not impossible, to determine because east-west, interior trade, domination, war, and slave raiding was ongoing. Lovejoy (2000, 156) details how slave trading took place across sub-Saharan Africa, so that one found slaves in the Asante Kingdom from as far away as the Sokoto Caliphate as well as Gurma and Mossi states. We know that, with so much interaction, ethnolinguistic groups regularly borrowed ideas, sounds, instruments, and approaches from others. Such integrations can also be seen at the level of religious and occult practices. For example, the Songhay spirit pantheon includes Hausa and Tuareg spirits (Rouch 1989) just as Songhay, Tuareg, North African, and Fulani spirits populate the Hausa bori spirit pantheon (Besmer 1983;. Despite these complexities of ethnicity, the Hausawiyyn and Migzawiyyn suites plausibly reveal a Hausa origin for several reasons: the songs' musical temporality and melodic phrases differ remarkably from all others in dīwān, they contain intact phrases in Hausa, call out to Hausa spirits, and use Hausa ritual objects. In these suites, numerous and still-intelligible phrases in Hausa sing of magic, medicine, and healing. The most common references are 'mai' (master) and 'magani' (medicine, magic, or charm) and 'bāwa', meaning 'slave'. Furthermore, the term 'bori' usually signifies spirit inhabitation phenomena in dīwān, which links it to the Hausa bori 'possession' ceremony of present day Northern Nigeria (notably Tremearne 1914;Besmer 1983;. 19 Keepers of the bori ceremony (see Tremearne 1914;Besmer 1983), the Maguzawa subgroup of Hausa (pronounced by my interlocutors primarily as 'Migzawiyyn'), have also regularly been described as 'pagan'that is, in direct opposition to Islamand, quite interestingly, as the 'oldest Hausa tradition progressively excluded' from other Hausa groups (Nicolas 1981, 204; see also Khiat 2014). 20 Hausa bori worlds resound throughout dīwān. At one time, it seems 'bori' was the name of one kind of healing ceremony in Mostaganem (Lecomte 2000, 7). The important and potent Dogowa spirits of the bori pantheon (see Besmer 1983;, also recognised in Songhay pantheons (Rouch 1989), regularly appear in the dīwān Migzawiyyn repertoire, most obviously in the borj 'Bani Shawara Dogowa' (locally meaning, 'Give me consult, Dogowa'). But many other invocations of Hausa spirits such as Jato, Migzu, and Rima and related terms such as 'serki' (chief) animate songs, not to mention an entire dīwān song suite dedicated to the city of the spirits, 'Jan Gari' or 'Jangare' (Tremearne 1914;Besmer 1983). This is especially important here because bori, a well-studied ritual practice of great importance to Hausaland, is typically understood as a pre-Islamic, 'animist' ritual that has struggled to survive in increasingly Muslim contexts (see particularly . The same holds true in dīwān: musicians often make a point about dīwān being 'cleansed' of its pre-Islamic roots: musicians spoke of precarious (wʿ āra) songs that were no longer performed and that they had never heard in their lifetimes because the songs had been eliminated ('supprimé') from the repertoire. Even so, the Hausa and especially Migzawiyyn repertoires still push back at this discourse (see more below).
Just like with the Hausa term 'bawa' (slave) occurring in all categories of dīwān brāj (not just these two suites), Hausa bori traces materialise throughout the entire ritual in other ways. One very popular borj, 'Bori ya Bori Mana(n) Dabu' that is not classified as a Hausa song in dīwān references the Hausa bori directly: bori, a spirit, and magic, dabu. In this borj, dancers must lash themselves with bulālat, ritual hide whips of Hausa origin. However, despite this abundance of evidence, ūlād dīwān rarely recognise an explicit, practical connection to the Hausa ethnolinguistic group or the bori ceremony. This discontinuity can be understood in several ways.
Dīwān histories, as trans-Saharan histories, are histories of rupture and forgetting as much as they are histories about survival. Even when my interlocutors had clear family ties south of the Sahara, they typically only knew roughly from where their family lineage came and nothing more; these family histories commonly 'began' from the arrival in present-day Algeria. Such personal and social histories and their relation to dīwān communities that coalesced out of the slave trade were scantly documented and only partially passed down through the generations. Hausanot to mention other sūdānī languagesis no longer spoken or understood in dīwān communities, according to my interlocutors and my own fieldwork observation, and ūlād dīwān do not recognise ritual objects as Hausa, nor are they usually familiar with the bori ceremony. Some ūlād dīwān reject aspects of African history, notably that of the trans-Saharan slave trade. In other words, immense gaps of time, distance, and the trauma of slavery separate dīwān adepts today from their sub-Saharan ancestral lineages; thus, it is not surprising that such descriptions as 'Hausa' and 'Bambara' might not (always) be used literally but, nevertheless, do important work to signify ideas and imaginaries of who these ancestors were and what they were like. These depictions, then, also have ramifications for how ūlād dīwān relate to, perform, and choose to transmit (or not) these songs today.

The Hausawiyyn and Migzawiyyn repertoires today
All dīwān communities perform Hausa songs as part of their musical corpus, from Algiers to Oran and throughout the west from Mostaganem to Relizane and Arzew to Sidi Bel Abbess. However, Hausa songs are particularly strong in the High Plateau region of Saida and Mascara. 21 Given that the Northern Nigerian Hausa 'Maguzawa' after whom the Migzawiyyn are undoubtedly named, were cultivators (see Besmer 1983), this might help explain why we find these brāj here, in some of the richest agricultural regions of Algeria. 22 In fact, according to ūlād dīwān in Saida, up until about a decade ago, one dīwān elder, Lʿaīd Canon, understood Hausa although he refused to speak it.
Muḥ ammad Amīn Canon, one of his nephews and a respected dīwān mʿallem, repeatedly told stories about his uncle's resistance to pass on his Hausa knowledge, seeing it as expertise that must be earned. Muḥ ammad lamented that only after fifteen years of imploring his uncle did Lʿaīd slowly begin to share some of the song textsthat is, to sing them slowly and clearly enough that they could be recorded for posterity. 23 Quite important here is that ritual knowledge has always been and continues to be transmitted aurally through keen observation and repetition. The tragedies of transmission echo in every dīwān locus: 'ancestors did not share their knowledge openly' and 'questions were not welcome.' As so many musicians explained and as I witnessed in ritual, singers often mumbled 'sūdānī' words, whether intentionally or unintentionally, making it nearly impossible to understand them and, according to my interlocutors, to prevent others from 'stealing' the words' powers. Like ʿAzzeddīn Benūghef lamented, 'The elders are gone. The people now, they don't know. We do not even have the meanings of these songs. We just sing them but we do not know [the meaning]'. Sometimes my interlocutors recounted songs in our conversations that, in three years of attending rituals, I had never heard performed. Indeed, ʿAzzeddīn too noted that his ancestors spoke about songs that he had never heard himself and that have since disappeared from the corpus. Moqedm Jallūl Moṭ am of Saida shared a similar grievance, explaining, 'From Independence, dīwān began disappearing. Each time an elder died, he took his knowledge with him' ('Yeddī mʿ ā ṣ wāleḥ '). 24 Dīwān musical and ritual knowledgeand the attitudes about thempass on through family lines. This means that in both Saida and Mascara, a family lineage asserts their own categorisation and 'order' (tartib or ṭ req) of Hausa and Migzawiyyn songs. In Mascara, the Ūlād Meriem kin group sang for me twenty-nine Hausawiyyn brāj and seven Migzawiyyn brāj, whereas in Saida, my interviews and ritual observations produced roughly eleven Hausawiyyn and fourteen Migzawiyyn brāj. Even between these nearby towns, the song names, orders, texts, categorisation, and the meaning of the songs varied, sometimes quite significantly. The figures below represent approximations of song performance order, pronunciation, and categorisation for particular kin groups in Mascara, Perrigaux, and Saida. Note also that songs in green and red are discussed in depth below (Figures 1 and 2).
Discrepancies in the ritual corpus between towns and between troupes in the same town is not a new phenomenon. Even in the 1950s, Pâques noted four different dīwān groups (grāba) in the town of Saida that each specialised in particular song suites (1964, 510-517) associated with spirits from different earth elements of the ritual order: spirits of the water, fire, air, or earth (ibid., 575-586). Given the variety of ethnolinguistic origins constituting dīwān, and that transmission primarily moves between kin, we might presume that certain family lineages always specialised in certain repertoires and, unlike today, single troupes did not necessarily perform the entire musical corpus but shared the ritual between groups. Today in Saida, the dominant Canon-Farajī-Būterfās kinship group who built and maintain the impressive Zāwīya Sīdī Bilāl are locally considered the experts of the Migzawa spirit pantheon. However, according to Pâques's notes ('Busarji' 1964, 582) and my own research with the Sarji kin in Oran and Mostaganem, the Migazawa repertoire was presumably the specialisation of the Sarji family, who left Saida for Mostaganem decades ago.
Sometimes, the Hausawiyyn and Migzawiyyn suites fall together under the single umbrella term: 'Khelā'wiyyn' from el-khelā', the wilderness, void, or desert. 25 Some of my interlocutors emphasised that the term, el-khelā', indexes the intrigue, mystique, and precariousness of this 'old' repertoire from 'black Africa.' 26 In other areas, the two suites of songs are conceived of as separate. 27 These latter groups typically consider the appellation 'Khelā'wiyyn' to refer only to Migzawiyyn. In my circuitous efforts to try and clear up these boundaries, one female ritual leader (moqedma) from Sidi Bel Abbess explained that 'Khelā'wiyyn' indicates the musical suite whereas 'Migzawiyyn' specifies the name of the tribe (tribu) to which the suite belongs. In general, while the 'Hausawiyyn' suite is considered by ūlād dīwān to be of Hausa origin, most do not recognise Migzawiyyn as also being 'Hausa'. Given that historically, the term 'Hausa' referred quite broadly to many states, peoples, and diverse practices, it is no wonder that this produced a pixelated historical memory around 'Hausa' in Algeria. These different ideas also emerge in the regional variations between song names and classification.
However, despite their suite names (Hausawiyyn and Migzawiyyn), both ostensibly reveal a Hausa origin. The musical aesthetics of both suites significantly differ from the rest of the dīwān repertoire and are thus considered exceptional by all dīwān adepts. Whereas most songs in the dīwān corpus feature long, virtuosic, layered ginbrī phrases (four to eight bars of musical time) followed by soaring vocal calls and responses, the Hausawiyyn and Migzawiyyn songs consist of short (two-bar) musical phrases on the ginbrī, followed by brief sung responses. Whereas other songs develop over twenty minutes through various stages of phrases and refrains, both the Hausawiyyn and Migzawiyyn songs utilise remarkably simple musical material, typically lasting a couple minutes or less. In both these suites, the ginbrī plays a repeated, fairly restricted accompaniment line, sometimes called the 'Hausa groove' (sūg, to drive); two versions of the Hausa sūg from Mascara dominate audio example number two below, Ūlād Meriem performing Yurah (from 0:00-2:48; and 2:48-5:30). These specific musical aesthetics, which emphasise constrained, repetitive instrumental accompaniment to slightly more embellished singing, can arguably be traced to sub-Saharan roots, if not at least rural roots of dīwān before it developed in urban centres. 28 The further south one travels in Algeria, the more one finds such increasingly restricted musical material with emphasis on accompanied singing. 29 The two suites are mainly differentiated by their places in the dīwān pantheon and their associated ramifications. 30 As briefly mentioned above, the Migzawiyyn brāj carry a distinctive, unanimous reputation for being the most 'wild', non-Muslim, and 'pagan' personages: they were described to me as half-man and half-spirit, qualities that seem to reflect the appellation Khelā'wiyyn (again, 'wild', 'other'). On the other hand, even though the Hausawiyyn braj reference Hausa spirits like Rima, they do not necessarily depict 'non-Muslim' or 'pagan' personages and, thus, attract less suspicion. Nevertheless, ūlād dīwān consider both the Hausawiyyn and Migzawiyyn repertoires to be 'the oldest' or most authentically sūdānī songs of the musical corpus. 31 Absolutely critical to the mystique and intrigue surrounding these repertoires is the nature of their texts: both suites are mostly sung in a language that ūlād dīwān identify as 'Kuria'. Kuria is quite probably an extinct, hybrid sūdānī language that might have been spoken between the slaves of various ethnolinguistic groups and their descendants. 32 While Hausa terms populate Kuriaand this is particularly importantmany other terms appear to be Songhay or perhaps Kanuri (personal communication with Dr. Yusuf Baba Gar of Humboldt University, 4 September 2020). 33 Today, Kuria indexes a secret and powerful language that can call the jnūn. Given its name, it may connect to the Bānū Kūrī spirit pantheon of the Bornu region (see Jankowsky 2010, 84-88). While Hausa and Songhay words can be identified in the musical texts, translations of the texts are difficult if not impossible. This is important to emphasise. 34 In addition to this linguistic intrigue, Hausawiyyn and Migzawiyyn suites are also considered somewhat scarce. This has much to do with their temporal place in the dīwān ritual. While the average ritual begins at around nine in the evening, the Hausawiyyn and Migzawiyyn repertoires typically occur towards the end of the ritual, usually around three or four in the morning. 35 At this point, sometimes even the die-hard connoisseurs (muḥ ebbīn, lit. 'lovers') begin to tire and depart so that the musicians abridge or bypass large sections of the corpus. With so few people still in attendance at this point, rarely is an adept present who masters these songs in trance. Only in Saida and Mascara did I witness a ritual commitment to these songs and their specific ritual requirements: special ritual knives that must be stabbed into the trancer's thighs in the Hausa borj 'Yajrou/ Natiro' (discussed below), the Migazwiyyn incense kajiji used only in Migazwiyyn songs, or the once alcoholic ritual drink, doghnu, only served during Saida's annual wʿāda (dīwān gathering).

The Hausawiyyn suite
The first borj of the Hausawiyyn repertoire is 'Bilāl', the spiritual father of dīwān. Despite the fact that Bilāl was not Hausa, this position critically ties Bilāl to these quintessential 'black African' songs. Given that many singers pronounce it 'Bulāl' or Bulālī', it is also possible that the song once referenced the Bulala ethnolinguistic group of the Fitri region; this might explain its place in the sub-Saharan 'Hausa' repertoires. In Saida, the borj contains many Hausa words but, overall, references the spiritual father, Bilāl. In any case, this borj is absolutely mandatory. Almost always, dīwān troupes go straight from Bilāl's borj into the sub-suite for the Hausa spirit, Rima consisting of three songs: Rima, Rima (K)obana, and Jangerima. What follows, however, depends on the ritual needs and preferences of the ritual actors. Like all the other brāj suites in dīwān, the 'best of' or favourites in the Hausawiyyn grouping dominate the ritual time.
In my efforts to collect information about these elusive songs and notorious use of 'Kuria', I heard the most compelling and detailed description about the Hausa song, Yurah, that is still remarkably intelligible to Hausa speakers I later consulted. One afternoon in 2015, Bel 'Abbess Zerwālī, an elder of Perrigaux best known for his knowledge of the local ṭ req, elucidated that, 'Yurah is a poor, black, Fulani slave who travels selling (miga) leben (chanunu), a yogurt drink: 'Hadi, Yurah, ʾabd hāda, kḥ al meskīn, ydūr f dūār ybīʿa lben'. Zerwālī sang it as such (see audio example 1): There are references in this song to both the Fulani (the ethnic group that the original singer belongs to) and the Hausa. We believe the reasoning behind this is that the Fulani and Hausa often have a joking relationship. The Fulani milk seller is calling out to the Hausa people in a playful way to come and buy his product. For example, there is a line in which the singer calls to a Hausa man named Abdullah saying, 'I am calling to Abdullah. You are not coming.' Likely, he is being playful with the man (at least in the original context). 36 If this is correct, Yurah might be a striking example of a popular or folk song from one of the many ethnolinguistic groups who historically came to constitute dīwān communities and their largely-shared musical corpus. Indeed, dīwān music not only consists of pantheons but, precisely, also references sub-Saharan and trans-Saharan histories, such as in the borj, Gafla (qāfila), referencing the infamous caravans. 36 While Zerwālī sung the text quite clearly, ritual acoustics and amplification prioritise sonic force and affective impact such as in the peformance by the Ūlād Meriem group of Mascara (see audio example 2).
Another popular Hausawiyyn borj, known as Natiro (pronounced nahteero) in Mascara and as Yajrou (yah-juru) in Saida and elsewhere demonstrates intelligible Hausa words (serkin, magani) alongside many other unclear or unknown words. Even while the distance between Mascara and Saida spans approximately only seventy-five kilometres, variations of pronunciation and thus, potential meaning, vary considerably. Again, such discrepancies, although highly contested between troupes, might be explained by different family lineages, Hausa dialects, and quality of transmission. I most often heard the song pronounced 'yajrou' or some close variation of these sounds. Given the difficulty of capturing song texts during ritual, the clearest representation of this version I collected was in a written text supplied to me by Muḥ ammad Amīn Canon: Yajrou albarka dami Yajrou banajrounia Yajrou matelbigui jrou However, in an extended, audio recorded conversation in 2014 with ūlād dīwān in Mascara, ʿAzzeddīn Benūghef, a beloved kuyu bungu (lead ritual singer), sang it as such (see audio example 3): Consulting with Hausa speakers on several versions of these songs, many dīwān 'terms' contradict standard Hausa orthography or are unintelligible in Hausa. 37 Familiar with ritual practice and fluent in multiple languages, Ibrahim Saani believed that 'natiro' was closer to Hausa orthography and pronunciation while 'yajrou' did not strike him as resembling any formulations of Hausa with which he is familiar. He presumed that 'natiro' was possibly from the Hausa 'na ciro' (pronounced nah cheero) or 'I nominate him', which also gestures at the 'ch' sound in the common 'yajrou' pronunciation. Translations here prove especially difficult: again, the singers do not speak Hausa and claim to simply repeat what they heard over many years in ritual, that which has been passed down from their ancestors. In such a context where pronunciation shifts over time, meaning becomes even more difficult to parse out. Moreover, the many dialects of Hausaa tonal languagevary considerably and can be influenced by neighbouring languages in the region, shifting the use of consonants. 38 Finally, adding to these complexities stands the issue mentioned previously around musicians mumbling lyrics and the possibility that many of these words indicate origins in languages other than Hausa.
With all of this in mind, based on audio recordings of multiple performances, the following renders an educated guess at the original text of 'Natiro': 39 What struck me about this possible interpretation of Natiro was that it emphasises medicine (magani) and the forest (daji), 40 two unanimous themes collected in my fieldwork about these repertoires. Interestingly, when I contacted ʿAzzeddīn in September 2020 and mentioned the other version I was consulting, 'yajrou', he replied that 'yajrou' was in fact the proper pronunciation. In any case, the point here is that texts are flexible: musicians improvise, change their minds, and texts evolve with time.

The Migzawiyyn repertoire
The Migzawiyyn song suite is the least well known of all of the ritual suites, even compared to the Hausawiyyn suite. Being the concluding suite of the night, and because the musical phrase lengths are even shorter than most Hausa songs, Migzawiyyn brāj are the easiest to cut short. They can be as brief as thirty seconds if need be. Even when these songs are not cut short, most dīwān rituals only deliver the favourites, if any are played at all. That is, some rituals entirely skip this repertoire altogether. The rarity of the full performance of this suite and the fact that it is not well known contributes much to its reputation as being particularly 'special', 'exotic', and sometimes more valuable than other suites. Indeed, like the Kuria words within, some consider the suites to contain secret knowledge.
Like other songs of Hausa origin, the forest looms again as a common theme. In fact, Migzu, the namesake of the brāj suite and the chief of the Migzawa, is declared 'Lord of the Forest', Sīd al-Ghāba. As noted above, in dīwān communities, the Migzawiyyn songs embody pagans, and, as ʿAzzeddīn Benūghef put it, '[American] Indians'. I often heard this description from my interlocutors, just as Dermenghem (1954) and Pâques (1964) did. Indeed, ʿAzzeddīn had the most to say about Migzawiyyn of all the dīwān musicians I met. He considered them as a clan of 'expert lion hunters' within the larger grouping of Khelā'wiyyn, what he called 'black Apaches', who were from either Guinea or Mali and who played some kind of ginbrī.
Given this discussion, I sought to confirm that the Migzawiyyn personages evoked in brāj were humans and not spirits. However, ʿAzzeddīn clarified that, 'they are spirit-like', referring to the fact that the Khelā'wiyyn wore black face paint. 41 ʿAzzeddīn indicated that the Migzawiyyn borj, Kirem Mdawa, announced by a blown animal horn, instigates a two-to three-day hunt after which the Migzawa hunters bring back their live prey in cages. Indeed, today during the Migzawiyyn brāj in dīwān, we see a great deal of ritual theatre having to do with hunts. Migzu, the namesake, happens to prefer wild rabbit and, in one ritual I observed in 2016, a live rabbit was turned loose and chased down by a woman in trance. Both in ritual and in stage settings, dancers 'depict' the Migzawa by carrying bows and arrows, and feign hunting prey. During this borj, according to tradition, hunters release live prey, hunt them with bow and arrow, and eat them raw without their hands. 42 Aside from these qualities, the meanings of these brāj evade documentation. Some of my interlocutors considered them as a mixture of spirits or jnūn of wild places like the Sahelsuch as the spirit Jato of Peul origin whose song classifies as Migzawa. Jato has received some attention in previous scholarship: he is a genie who, in one of the legends, eats human excrement (e.g. Rouch 1989, 64). Finally, many Migzawiyyn brāj share the same names of Bānū Kūrī songs and spirits in Tunisian sṭ ambēlī: Migzu, (Baba) Kūrī, Jamarkay (also Joumarki), and Nikiri (see Jankowsky 2010, 86). Like in sṭ ambēlī, those in the dīwān Migzawiyyn repertoire represent blackness and a definite connection with sub-Saharan Africa: Migzawiyyn brāj also correlate with the colour black. Furthermore, when these brāj resound in a dīwān ritual, the typical ritual incense (bkhūr) is removed and distinctive ritual props come out: feathered headdresses, spears, bells, and animal pelts, black cloaks (abayat), and sometimes kajiji, special Migzawiyyn incense (Figure 3). These ritual materials must always remain separate, kept in a different trunk (maḥ alla) from the ritual materials of the other songs, lest they pollute them.

The meaning and mattering of Hausa songs in ritual today
Today, one could argue that the bulk of dīwān songs sung in Arabic praise the Prophet Muḥ ammad, Muslim saints such as ʿAbd el-Qādr Jilānī, and the 'Companions of the Prophet'('Soḥ āba') such as Abū Bakar and ʿAlī ibn Ābī Ṭ ālib. Many ūlād dīwān make this precise point: that dīwān has been cleaned up, even 'whitened'. However, even these songs for prophets and saints (personages held in high regard) pose problems for some ūlād dīwān because their texts sometimes contain unknown (Kuria) words. Furthermore, spirits may take Muslim names like ʿAlī (also the name of a Hausa spirit), making it simple to 'mask' and confuse who a song means to praise. 43 At other times, songs that seem to be about a human personage such as a prophet (e.g. Moses or Mūsa) slip into a song for a sub-Saharan spirit or jinn of the same name. In an attempt to sort this out, some of my interlocutors emphatically stated that dīwān is an assembly of the saints and holy 'people' (sometimes referred to as awlīya' other times as en-ness eṣ -ṣ aliḥ īn) and not a dīwān (assembly) of the jnūn. Still others speak decisively about two worlds (ʿālamīn) within the dīwān repertoire: some songs praise the awliya'personages of the human worldand other songs call the jnūn, or other personages of the Unseen (al-ghayb). Even if one accepts a supernatural dimension to the ritualas this, too, is contestedthese different worlds are meant to be kept separate musically, textually, temporally, and materially.
The sonic performance of 'Kuria' is considered to invoke particular powers that cannot be mustered with song texts in Arabic. As mentioned, the Migzawiyyn songs in particular are seen as especially 'other' and exotic, because sound and summon animist or 'pagan' people or spirits. It was precisely due to this debate that one ritual elder insisted on the importance of dīwān musicians being extremely careful when singing in Kuriaor perhaps not singing at allbecause they may inadvertently call forth the supernatural world through sound. Occasionally this argument leads ūlād dīwān of all levels of status to insist that no one should sing in anything but Arabic, even if that means the end of Hausawiyyn and Migzawiyyn repertoires altogether. * * * In March of 2015, I interviewed Mʿallem Tūfīq ʿAbd Es-Selam in Algiers. Tūfīq, a well-known ritual musician, was particularly interested in dīwān texts that supposedly contain Kuria words. Like so many others told me previously, Tūfīq particularly emphasised that certain words contain inherent agency to invoke worlds and call forth the jnūn, with or without the knowledge or intention of those singing the words. The idea that words manifest magical and/or healing effects, whether or not they carry a direct relationship with a supernatural world, prevails in West and North Africa, but has also been explored in numerous other anthropological sites and studies (e.g. Malinowski 1935;Ong 1967;Tambiah 1968). However, Stoller's (1984Stoller's ( , 1996 examination of the power of sound for the Songhay most closely echoes this thesis. 44 'I'm against all the brāj in which there is not Arabic [the language]', Mʿallem Tūfīq began. 'I don't trust them.' He went on to explain how it was problematic that certain people in dīwān continue to sing in a language that they do not understand or speak: 'If you know Kuria well, you understand it! But later when you ask [those who sing it], "What did you say?" they respond [inadequately], "From the old times, they sang it like that" [bekrī kānū gūlūha]. And that's serious [dangerous]! (et ç'est grave!).' Here, Tūfīq means that even people who appear to sing well in Kuria do not know the words' meanings and fall back on the explanation of 'that's how I learned it' or 'this is what people have always sung'. This anxiety abounds among kuyu bungu-s (ritual lead singers), moqedmīn (ritual leaders), and muḥ ebbīn (connoisseurs); I encountered such worry in nearly every conversation with ritual experts, especially when we discussed specific 'sūdānī words' that turned up in otherwise Arabic texts about saints and holy people.
For example, the text for the borj Sidi ʿAlī, the nephew of the Prophet, contains the word 'dodo'. While none of my interlocutors knew the word's meaning or history, according to Besmer's work on the Maguzawa bori spirit pantheon, it means 'evil spirit' in Hausa (1983, 6, 158). Tūfīq used the namesake Migzawiyyn song, Migzu, to demonstrate his point. He sang the main theme (rās el-borj) in what appeared to be Hausa: 'Migzu, nama, ay nama bani: Migzu! Meat! Hey, give me meat! Referencing those words, Tūfīq continued. 'That [first] line, you cannot deviate from it (ma truḥ sh) because it's the key (la clé) of the song'. That is, the line has to be sung in order for the borj to be the borj. It is, after all, the identifying theme. 'But later', Tūfīq added, 'you have to cultivate something [the words] that the public knows', meaning that, after the main theme, the text should switch to Arabic. Since no one really knows the text's meaning, other than the words 'Migzu' and 'meat', translating 'the rest of the text' would be impossible. But to argue these details would be to miss Tūfīq's point.
To illustrate his reasoning Tūfīq went on to tell a story about a dīwān. During the borj, Bū Derbālaanother name for ʿAbd el-Qādr Jīlānī, arguably the most beloved saint in Islama well-known elder from small, western hinterland city began singing in Kuria. Tūfīq complained that, first, the singing was not in rhythm with the ginbrī and qrāqeb. He demonstrated the singing, how it stumbled in time because there were too many sūdānī words to fit into the ginbrī's melodic cycle which singers must follow. In this case, Tūfīq was bothered by the performance because one should sing about such a noble, Muslim saint in Arabic, not in Kuria. In other words Arabic, as the language of Islam, represents morality while Kuria represents sūdānī precariousness.
Furthermore, part of Tūfīq's argument was that the significations of words possess power to affect circumstances whether or not those singing them realise this. If someone sings in Kuria or Hausa or a language he does not understand, he could inadvertently attract spirits or energies called by those words. Tūfīq's voice became pinched and nervous, almost seeming to scold his audience, as he explained, You can sing to [the jnūn] with certain terms! Why did I tell you that you have to know [what you are singing]? There's something like fifty percent of the terms [like this in dīwān]. There are terms that call them [the jnūn] without your awareness! You're calling them! Because [the words] are theirs, theirs! It is their domain! And they come and hang out with you. There are brāj that are theirs. 45 The lack of mastery of textsor disregard, or otherwise 'mixing' of themposes a particular problem when words that reference the world of the jnūn interweave or collide with words that reference the Prophet Muḥ ammad. During one of the several brāj for the Prophet Muḥ ammad within the musical suite Sema'wiyyn ('of the sky'), and just after singing, 'Ṣ alou Nabī, Nabīna Muḥ ammad' ('prayers upon our prophet Muḥ ammad'), singers go immediately into a new phrase, singing, 'Yay, Janari Nyam-Nyam' on the same melody without pausing (see audio example 4).
According to dīwān musicians, Janari Nyam-Nyam calls a non-Muslim spirit. Although Janari arrives earlier in the ritual within a different song suite, some claim it belongs to the Migzawiyyn. In this borj, one of the jedebbīn might eat raw meat, tearing it with his teeth 'like a wild animal'. The important point here is that to go from singing about the Prophet Muḥ ammad to singing about eating raw meat (forbidden in Islam) without skipping a beat poses a serious problem for not just Tūfīq, but for the majority of ritual experts I spoke with. In accordance with dīwān discourse, some Hausa dictionaries list 'nyamnyam' as implying meat or a wild animal. Yet even more striking is that, as mentioned in Bovill's critical volume, Caravans of the Old Sahara, nyam-nyam was a 'common name for cannibals generally in the Western Sudan' (1933, 62, fn). 46 Related to Bovill's footnote, he described a 'primitive forest tribe' 'armed with bows and arrows', known to the Arabs as Lemlem or Demdem (ibid): just as ūlād dīwān also describe and enact the Migzawa today. Thus, the song Janari Nyam-Nyam demonstrates a particularly intact example of embodied knowledge without a conscious connection to history.
About this borj and some of these same issues, Shaykh ʿAbbess Zerwālī of Perrigaux sang for me his version of the text in question (see audio example 5).
Ya janari yo yay kuri Ya janari kuri ya Bornu Yay janari nyam-nyam ya bawa Yay janari yo ay kuri Janari kayba ya kuri Janari kuri, koye Ay janari baba ya kuri Yah kuri m Bornu ya kuri Yay kuri madi, kuri No translation was given: again, the text was conveyed as the now-extinct and thus unintelligible language of 'Kuria'. We can spot certain terms, however. The Arabic bābā, for 'father', the Hausa term bawa for 'slave', Bornu the region, and 'Kuri' a sub-Saharan spirit, often Hausa, that appears in various pantheons (Monfouga- Nicolas and Bastide 1972;Besmer 1983;Jankowsky 2010;Khiat 2014). But regardless of some 'original meaning' even if we could arrive at one, the debates over Kuria texts between dīwān experts points to a more pressing social mattering.
Lineages, ownership, and territories of knowledge: some concluding thoughts Perceived in recent memory as the most 'sub-Saharan' of the ritual corpus, the least well known and understood, these song suites have accumulated a certain social capital and mystique. The Hausa and, in particular, Migzawa song suites sometimes function as proof of a family's or troupe's authentic connections to sub-Saharan ancestors from which ritual power and potency come, despite anxieties around non-Arabic texts and non-Muslim or pagan origins (as we saw with Tūfīq above) that might threaten one's reputation as a 'good Muslim'. While I have so far focused on ritual contexts, it is important to mention that the Migazawa repertoire is also performed in semi-private and, increasingly, in openly public contexts; that is, spaces particularly ripe for the claiming or rejecting of mythical musical territory. These two additional contexts are the semi-private, annual Saida wʿāda (large multiday gatherings of dīwān troupes from across the country) and the annual, state-sponsored Dīwān Festival of Béchar.
Every year, dīwān fans and family groups come from all over Algeria for the five-day wʿāda in Saida, hosted by the Canon-Farajī-Būterfās kinship group. An entire afternoon is dedicated to the full performance of the Migzawiyyn repertoire. Priding itself on a deep knowledge of the repertoire, the kin group utilises two sizes of barrel drums (ṭ bel), the qrāqeb (metal clappers), and the usual call-and-response singing rather than the ginbrī, which only animates night rituals. Furthermore, the wʿāda cultivates a carnival-like atmosphere where attendees dress up in costumes, including face masks, and parade in a circle through the courtyard. At the end of the dancing line, the troupe's shawsh (ritual helper), holds a bottle of doghnu, the legendary Migzawiyyn fermented drink that was formerly alcoholic. I was told that, in the absence of journalists, we may see an enactment of a rabbit hunt, the quintessential prey of Migzu. 47 Although many other troupes in Algeria claim to bear Migzawa knowledge and even family ancestry, no other family group or zāwīya (ritual space) in Algeria gives a full-day performance of the repertoire, particularly with percussion only (see video example 1).
The second, most recent and most crucial public site for Migzawa performance and rivalry is the annual Béchar festival. Both a coveted and heavily critiqued format for the staged performances of any category of dīwān brāj, the festival hosts troupes from all over Algeria. Troupes need not be from a dīwān family lineage or possess any ritual experience. In fact, the many young contestants who acquire the songs through amateur, circulated audio recordings of rituals sustain bitter criticism. In combination with the aspect of competition, such controversies about transmission give the festival a particularly contentious quality. In order to show their knowledge of the repertoire and increase their chances of winning, dīwān troupes choose what they consider to be the most impressionable, even shocking, songs. Indeed, since 2013, more and more troupes deliver the 'rare' and 'exotic' Migzawiyyn brāj on the festival stage. Playing off of the 'authentic' Migzawiyyn mystique, this strategy is also fuelled by a growing trend amongst urban, Algerian youth (but also ūlād dīwān) to 'look south' for inspiration. In other words, a steadily growing tendency to champion 'African' (read: 'black') culture and embrace such Africanness as valuable parts of Algerian worlds eclipses the former tendencies to 'look north' to France and Europe.
At the Béchar festival, these motivations result in what one might consider caricatures of the Migzawiyyn, designed to appeal to a popular imagination of 'black Africa': different troupes have donned mock animal-hide pelts or brightly coloured costumes with geometric shapes, supposedly to depict 'African' textiles. Most of these troupes, even the Sarji kin group, the historical guardians of this repertoire, enacted hunts on stage by dancing with spears, jangling bells, and walking circles around the stage, pretending to stalk prey (see video example 2).
On the one hand, such public depictions smack of exoticisation of 'black Africa' and one's own ancestorslet's remember, most performers are black Algerians in a highly racialised 'black-white' context. And yet, on the other hand, when speaking to ūlād dīwān about these choices, such enactments mean to embody the fractured myths and stories of these Migzawiyyn hunter personages. This does not mean, however, that ūlād dīwān, as primarily black Algerians, always consent to representations of dīwān's origins in 'black Africa'. In 2015, in an effort to depict Migzu, a troupe from Béchar brought a live rabbit on stage and feigned hunting and killing it, provoking an uproar among the audience and journalists present and condemnations from dīwān elders (see Turner 2017).
Just as musicians revise their texts, as ʿAzzeddīn did regarding Natiro/ Yajrou, I want to posit that the meanings and matterings of these songs continue to develop and change. While these songs embody dynamics of centuries-old movements of peoples across the Sahara, still containing words, objects, ideas, and aesthetics of dīwān ancestors, they are not static, historical objects. As dīwān musicians and dancers en-sound and embody these Hausa songs decade after decade, they also map the forgetting and rupture of their histories, imagining and living these sonic legacies with today's concerns in mind.
On a final note, a tribute is due here. On 17 September 2019, the beloved octogenarian Meriem Bel ʿArabī of Mascara passed away. Meriem was one of the last of her generation of dīwān families in Mascara, and the only adept I encountered who had been afflicted by Migzu for most of her life. Meriem's family explained that, because she had spent a great deal of time in the wilderness when she was a child, she would often fall quite ill (mrīḍ a) so that only a dīwān could comfort her. During these spells, I was told, she spoke with the deep, rattling voice of Migzu. Although I never witnessed this phenomenon, I spent several afternoons with Meriem over the course of my fieldwork between 2013 and 2016, benefiting profoundly from her inexhaustible knowledge and tenderness. When I think of Meriem now in relation to the questions posed by this essay, I am reminded that however ūlād dīwān might remember or forget, claim, revise or reject this embodied heritage of trans-Saharan continuity and rupture, the raison d'être of dīwān ritual obliges the recognition of one additional non-human agent: Migzu himself, inhabiting human bodies across time, the atemporal, spectral transmission of these transient and more-than-human worlds.

This research was funded by a Saharan Crossroads grant from the American
Institute for Maghreb Studies (AIMS) and the West African Research Association (WARA). This is the first research to attend to these songs as well as the dīwān musical corpus in general. Much of this data is contextualised in the comprehensive study of dīwān in Turner (2017). Previous anthropological work for consultation: Andrews (1903); Dermenghem (1954); Pâques (1964); Lapassade (1982); Khiat (2014 would require a lengthy discussion. In brief, because Bilāl did not descend from the Prophet and did not produce a lineage (silsila), the Bilāliyya are not considered an official Sufi ṭ ariqa (path, order) although their ritual epistemology, vocabulary, and practice parallels that of other ṭ uruq. 6. Dīwān can be considered a sister tradition to the Moroccan gnawa and Tunisian sṭ ambēlī. Unlike these other two traditions, however, research on dīwān is scant and what little documentation exists (Andrews 1903;Dermenghem 1954;Pâques 1964;Lapassade 1982;Khiat 2014) does not go into depth about musical worlds and their meanings. 7. Both women and men can and do engage in trance of all types. While the rituals are gender separated, they are mixed except in areas such as the M'zab where women hold their own rituals. There is a vast anthropological and ethnomusicological literature on music and trance phenomena that is simply not possible to cover here. However, see Turner (2017) and (2020a). 8. E.g. plucking and strumming the three strings while sometimes also hitting the skin face as percussion. 9. This phenomenon has been widely studied in ethnomusicological scholarship of sub-Saharan musical aesthetics and practices in North Africa, particularly concerning the gnawa in Morocco (e.g. Fuson 2009) but also in other similar traditions (Jankowsky 2010;Polak 2010 (Last 1971) and make comparisons of garaya playing with ginbrī playing. While approaches to playing the lute instrument are similar (accompanied singing aesthetics), and while there is still an idea of the Maguzawa being 'pagan' or nominally Muslim, I was not able to make any further connections in terms of musical repertoire that would clearly tie the two. Professor Last reported that the term Maguzawa comes originally from 'magus', the singular of 'magi' for magic. This could potentially be a connection in that this repertoire in Algeria is understood as the most 'wild', 'dangerous', and ominous. 29. For example, I documented songs in Biskra, Adrar, Ouargla, and the Mzab where musical aesthetics changed noticeably. See Turner (2017). 30. For an approximate list of these repertoires in Oran, Saida, Mascara, and Perrigaux, see Turner (2017), appendix 3, 'Treq Comparisons'. The repertoires are rarely performed in their entirety and, as noted, typically these repertoires are cut short. Musicians were reluctant to verbally list their pantheon and despite years of attending rituals, I observed no definite pattern: song names and orders varied and shifted. In other words, it is more accurate to mention which songs/spirits are typical and what 'common' orderings can be heard. 31. The terms 'Hausawiyyn' and 'Migzawiyyn' are adjectives as much as labels. To be clear, I will call these two repertoires by these namesas my interlocutors dideven though I argue that they are both Hausa in origin. 32. My attempts to clarify if Kuria might have been a form of Hausa were unsuccessful. However, since Hausa was a common trade language and because there are still intelligible phrases in Hausa in the Hausawiyyn brāj, such a connection is likely. In Morocco with the sister tradition, the gnawa, a mini-language or ghos developed between slaves of various sub-Saharan origins. Many of the words here also appear to be Hausa although the language was viewed as a kind of pidgin and private language. Personal communication, Tim Abdellah Fuson, September 2020. 33. After hearing my recordings and consulting the texts of the songs Natiro and Yurah, he suggested that there may also be Kanuri words or other languages spoken in Niger. 34. Dīwān is not the only case in Algeria where we find that 'sūdānī' words in otherwise Arabic texts contain mysterious abilities to summon entities. Champault's brilliant monograph on the small community of Tabelbala in the Algerian southwest features discussion on a unique dialect, kora n-die, that primarily utilises Songhay vocabulary, Berber forms, and Arabic roots. It explains that neighbouring groups consider the Songhay language to be 'of the spirits'( jnūn) because of its unintelligibility (1969,43). 35. This is the typical song order of rituals in western Algeria. In Algiers and in the east, Hausa songs might also occur earlier. 36. 'Gafla' is the Western Algerian pronunciation of qāfila, or caravan, convoy. 37. I primarily consulted with Ibrahim Saani and Emily Williamson but also had brief exchanges with Dr. Yusuf Baba Gar at Humboldt University also regarding these two songs. Dr. Baba Gar was also able to give the most information on Yurah whereas the word(s) 'natiro' did not strike him as Hausa. He suggested that perhaps there are Kanuri or other words mixed into the texts. 38. For example, Emily Williamson gave the example that Hausa speakers in Ghana use 'f' in some words elsewhere pronounced 'h' so that the word 'hira' becomes 'fira'. Personal communication, Emily Williamson, 26 August 2020. 39. I thank Hausa speakers Emily Williamson and Ibrahim Saani for this interpretation and their hard work at making educated guesses on these words and meanings. 40. In the sung version, ʿAzzeddīn actually pronounces the word 'dagani' with a hard 'g' as in goat whereas the Hausa would be pronounced as 'j'. This change was part of Mr. Saani's educated guess on the interpretation. 41. It was not entirely clear why black face paint would make these hunters more spirit-like. However, it brings to mind the popular notion that sub-Saharan 'spirits' are typically associated with the colour black, just as they are in dīwān as well as in Moroccan gnāwa and Tunisian sṭ ambēlī (see Jankowsky 2006Jankowsky , 2010. 42. However, he adds that now people are required to cook their meat and that there are new words in Migzawa brāj texts to reflect this. He adds that they danced on all fours, might have been cannibals, and might have once lived in trees. 43. I frame 'people' in quotes because the distinction between human, jinn, or other spirit is often blurred. 44. It must be pointed out that, in Islam, sound is an exceptionally valent medium.
Of course, it is the preferred medium through which divine knowledge is conveyed--such as the Qur'ān which must be recited and heard. The idea of healing through sound is also germane here. It is upon this idea that ruqqīa works, the method of reciting holy words over a person afflicted by harmful forces, such as a jinn or black magic. In Algeria, this is taken for granted. Sound vibration does things, it is not only meaningful because of the associations it is given by humans but rather, words are constituted by energetic forces. We can see this in daily practices where one must recite certain words in order to fend off bad luck or misfortune: bismillāh, el-hammdoulillah, and masha' Allah.
45. 'Kayn des terms, ntā taʿītelhūm, ble ma ʿala belek. Tetaʿītelhūm! Parce que entʿāhūm, entʿāhūm, had el-domain entʿāhūm! Ījī yrīḥ ū mʿak ū aʿītelhūm ū gʿā. Kaynīn brāj entʿāhūm.' 46. Indeed, ʿAzzeddīn Benūghef precisely described this song as being about cannibals, but, as a kuyu bungu, he referred to it by its key sung phrase, 'yop yop ina maniyama'. 47. In former times ('bekri'), before the Islamisation of dīwān, individuals inhabited by Migzu would capture the rabbit and kill it by biting its neck. Eating raw meat and ingesting blood is forbidden in Islam, therefore such ritual acts are said to also be forbidden although they are not completely unheard of even today.

Disclosure statement
No potential conflict of interest was reported by the author(s).