AFRICA Rivista Semestrale Di Studi E Ricerche

AFRICA Rivista semestrale di studi e ricerche

N.S. I/1, 2019

VIELLA AFRICA. Rivista semestrale di studi e ricerche © 2019 CSPE - Centro Studi per i Popoli Extraeuropei “Cesare Bonacossa” e Viella editrice N.S., I/1, 2019 ISSN 2612-3258 ISBN 978-88-3313-034-7 (carta) ISBN 978-88-3313-184-9 (e-book) Registrazione presso il Tribunale di Pavia n° 2/2019 dell’8/4/2019

La rivista è pubblicata anche grazie al sostegno dell’ISMEO - Associazione Internazionale di Studi sul Mediterraneo e l’Oriente/Il Novissimo Ramusio direttore responsabile Pierluigi Valsecchi (Università di Pavia) editor comitato di direzione Alessandro Bausi (Universität Hamburg), Alice Bellagamba (Università executive editorial board Milano-Bicocca), Gérard Chouin (William & Mary), Giacomo Macola (University of Kent), Antonio Maria Morone (Università di Pavia), Fabio Viti (Aix-Marseille Université), Massimo Zaccaria (Università di Pavia) comitato scientifico Bahru Zewde (Addis Ababa University), Giorgio Banti (Università di editorial advisory board Napoli L’Orientale), Barbara E. Barich (Università di Roma La Sapienza) Salvatore Bono (Università di Perugia), Carlo Carbone (Università della Calabria), Federico Cresti (Università di Catania), François Dumasy (Aix-Marseille Université), Anna Maria Gentili (Università di Bologna), Nora Lafi (Zentrum Moderner Orient), Thomas C. McCaskie (University of Birmingham), Jonathan Miran (Western Washington University), Mohamed Haji Mukhtar (Savannah State University), Paul Nugent (University of Edinburgh), Ian Phimister (University of the Free State), Irma Taddia (Università di Bologna) redazione Valentina Fusari, Marco Gardini, Ettore Morelli, Luca Puddu editorial staff contatti Africa c/o CSPE “Cesare Bonacossa”, Università di Pavia, contacts Corso Strada Nuova, 65 - 27100 Pavia - [email protected] website www.viella.it/riviste/testata/15

amministrazione Viella s.r.l., Via delle Alpi, 32 - 00198 Roma administration tel./fax 06 84 17 758 - 06 85 35 39 60 [email protected] [email protected] www.viella.it abbonamento annuale Italia € 70 (carta/print) € 90 (carta/print + digital) annual subscription Abroad € 85 (carta/print) € 110 (carta/print + digital) Digital (enti / institutional) € 60 Numero singolo (Italia) € 35 modalità di pagamento c/c bancario IBAN IT82B0200805120000400522614 terms of payment c/c postale IBAN IT14X0760103200000077298008 carta di credito Visa / Master Card indice / table of contents

Torna AFRICA is back 7

Ar t i c o l i / Ar t i c l e s Thomas McCaskie, The Parameters of Asante Historical Consciousness: Reflections onT Ͻprɛ and Temporality 9 Elara Bertho, L’empire de Samori Touré: pour un point de vue africain de l’histoire coloniale (Mallam Abu, Labarin Samori, 1914) 29 Alessandra Brivio, The Wade Harris Legacy in Ghana: On Ruptures and Continuities 49 Mariano Pavanello, Pawnship and Domestic Slavery in Chieftaincy Dis- putes (Nzema Area, SW Ghana) 69 Samuel Andreas Admasie, Cycles of Mobilisation, Waves of Unrest: Ethio- pian Labour Movement History 87 Luisa Revelli, Onomastica del contatto italo-eritreo 107

Rass e g na / Re v i e w a r t i c l e John Burton Kegel, Post-Genocide Rwanda 123

Re c e nsi o ni / Re v i e w s Yves Person, Historien de l’Afrique, explorateur de l’oralité (Fabio Viti) 129 Stuart Doran, Kingdom, Power and Glory (Tawanda Chambwe) 131

Au t o r i / Co n t r i b u t o r s 133

Africa, N.S., I / 1, 2019

torna AFRICA is back

Africa riprende la pubblicazione come Africa has resumed publication as a rivista semestrale, dopo un’assenza pro- biannual journal after a nine-year-long trattasi per nove anni. hiatus. Si tratta di un nuovo inizio, che tuttavia While this is a new beginning, this riannoda i fili della lunga vicenda del- historic Italian Africanist journal also la storica testata italiana di studi afri- calls upon a long legacy, whose dis- cani, che fece la prima apparizione nel tant origins date back to 1946, the year 1946 come Affrica, un mensile, per poi in which Affrica made its first appear- dare luogo nel 1957 ad Africa. Rivista ance as a monthly magazine. Renamed bimestrale di studi e documentazione Africa. Rivista bimestrale di studi e dell’Istituto Italiano per l’Africa. In se- documentazione dell’Istituto Italiano guito la periodicità divenne trimestrale per l’Africa in 1957, the journal later e l’Istituto cambiò la propria denomina- became the quarterly publication of the zione in Italo-Africano, per essere infine newly redubbed Istituto Italo-Africano, incorporato, nel 1995, nell’IsIAO (Istitu- which would eventually be incorporat- to Italiano per l’Africa e l’Oriente). ed into the IsIAO (Istituto Italiano per Tuttavia la vera data di nascita come te- l’Africa e l’Oriente) in 1995. stata scientifico-accademica può essere But Africa’s actual date of birth as a sci- considerata il 1965, quando Teobaldo entific journal can be considered to be Filesi ne assunse la direzione, indiriz- 1965, when Teobaldo Filesi took over zandola in maniera decisa verso il conte- as editor, steering it decisively towards sto degli studi, radicalmente trasformato the academic study of a continent which dalle vicende della decolonizzazione. was being radically transformed by de- L’inglese e il francese si aggiungevano colonization. It was under Filesi’s stew- all’italiano quali lingue di pubblicazione ardship that English and French were e Africa cercava di assumere un respiro added to Italian as languages of publi- internazionale in quanto sede di scambio cation, and that Africa sought to gain e confronto fra cultori di diverse discipli- an international scope as a forum for ne: “la storia, l’etnografia, la filologia, la scholarly exchanges between experts in sociologia, il diritto, l’economia, la poli- different disciplines. As the inside front tica, la letteratura e l’arte”, come recitava cover of the first issue of that year clari- la seconda di copertina del primo fasci- fied, “history, ethnography, philology, colo di quell’anno, precisando che la ri- sociology, law, economics, politics, lit-

Africa, N.S., I / 1, 2019 6 t o r na a f r i c a is b a c k vista si prefiggeva come scopo precipuo erature and art” all fell within the am- continuità rispetto alla storia iniziata nel main relevant in the present, Africa also lo sviluppo dei rapporti fra studiosi di bitious remit of Africa, whose primary 1965, seppure con cambiamenti che da wants to steer clear of the shortcomings diversa nazionalità, in particolare quelli purpose was the promotion of relation- un lato sono risposta a criticità emerse da that any dispassionate evaluation of the operanti nel Continente africano. ships between scholars of different na- una valutazione spassionata della storia journal’s past history must recognize, La testata si apriva realmente a una va- tionalities, beginning with those based passata della rivista e, per altro verso, and to reflect the scholarly needs of the riegata platea internazionale di contribu- in the African continent itself. riflettono esigenze inaggirabili poste dal current landscape of African studies. tori. Scorrendo i volumi degli anni Set- Africa truly catered to a richly diverse quadro attuale dell’africanistica. The composition of Africa’s governing tanta-Novanta colpisce l’alta percentuale audience. Browsing through its 1970s- Gli organi direttivi e il comitato scien- bodies and scientific committee attests di autori provenienti da diversi paesi di 1990s volumes, one is struck by the in- tifico danno conto nella loro composi- to the intensely international dimen- Europa e Africa, ma specialmente emer- ternational character of its pool of con- zione della dimensione intensamente in- sion of contemporary African studies. ge una linea editoriale esplicitamente tributors and by an editorial line explic- ternazionale in cui respirano gli odierni This is also borne out by the transition orientata ad accogliere e sostenere atti- itly intended to welcome and actively studi africani, della quale sono peraltro to English as the prevalent language of vamente le voci dei contesti accademici support the voices of African academic evidenze la transizione linguistica verso scientific communication and the un- africani. Tale impostazione fu sostan- milieus. This policy remained in place l’inglese come strumento prevalente di precedented extent of scholarly mobil- zialmente mantenuta dalla direzione di under the direction of Gianluigi Rossi, comunicazione scientifica e le diaspore a ity, which has transformed what were Gianluigi Rossi, che succedette a Filesi who succeeded Filesi in 1994 and led livello mondiale degli studiosi originari once more nationally rooted schools nel 1994 e firmò la rivista fino al dicem- the journal up to its 65th volume of De- di quelle che in passato erano scuole na- of scholarship. Africa embraces these bre 2010, completando il volume LXV, cember 2010, the last of the original se- zionali specifiche e connotate. Di questo changes – changes in which Italian Afri- l’ultimo della serie originaria. ries. mutamento, che interessa in maniera ma- canists themselves are deeply enmeshed La pubblicazione fu quindi sospesa e poi The publication was suspended and, croscopica anche l’africanistica italiana, – while maintaining its own tradition of cessata formalmente a causa della crisi then, formally discontinued as a result la rivista prende atto positivamente, pur operational trilingualism. e quindi dello scioglimento dell’IsIAO, of the crisis and subsequent dissolution mantenendo la propria tradizione di tri- Today, as in the past, the discourse on cui fece seguito un complicato processo of the IsIAO, which was followed by a linguismo operativo. Africa’s realities and the deep motiva- di ricollocazione dell’importante patri- complicated process of relocation of the Oggi come in passato il discorso tions of its social actors is all too often monio librario, archivistico, museale e Istituto’s important library, archives and sull’Africa, le sue vicende, le realtà e le informed by questionable assumptions editoriale dell’Istituto ad opera del Mini- museum by the Ministry of Foreign Af- motivazioni profonde dei suoi attori so- and simplistic preconceptions, which stero degli Affari Esteri e Cooperazione fairs and International Cooperation. ciali è troppo spesso sotteso da assunti militate against the prospect of genuine Internazionale (MAECI). It was in this context that, in 2017, the discutibili e preconcetti riduttivi, a tutto understanding. Fu in questo quadro che nel 2017 il CSPE (Centro Studi Popoli Extraeuro- discapito di quella comprensione della African studies are called upon to en- CSPE (Centro Studi Popoli Extraeuropei pei “Cesare Bonacossa”) of the Uni- complessità necessaria a una corretta in- gage in the important task of dispelling “Cesare Bonacossa”) dell’Università di versity of Pavia – with the support of terpretazione. stubborn ambiguities and recurrent dark Pavia – con il sostegno di alcuni studiosi some scholars who had been members Gli studi africani sono chiamati a un spots. Africa aims to contribute to this membri dei comitati direttivo e scientifi- of Africa’s editorial and advisory boards compito importante nello sciogliere effort by publishing rigorous analyses co di Africa – rivolse all’allora Vicemini- – came up with a proposal to revive the ambiguità pervicaci e dissipare ombre and empirically grounded research and stro Mario Giro una proposta per rivita- journal. Addressed to the then Deputy ricorrenti. Come proprio contributo a by providing Africanists with an inter- lizzare di propria iniziativa la rivista. La Minister, Mario Giro, the request was questo sforzo, Africa si propone di dar disciplinary venue for in-depth discus- richiesta fu accolta a seguito di una labo- accepted following a laborious pro- voce ad analisi accurate e ricerche empi- sion in the areas of the humanities and riosa procedura del cui successo va dato cedure, whose ultimate success owes ricamente fondate, offrendo una sede in- the social sciences. In keeping with the pieno merito, insieme al Viceministro e much to the resolve of both the Deputy terdisciplinare qualificata di confronto e journal’s tradition, special attention will alla sua segreteria, alla determinazione Minister and his staff and the office re- dibattito attenta ai percorsi delle scienze continue to be given to the work of Afri- dell’ufficio responsabile per la liquida- sponsible for liquidating the ex-IsIAO. umane e sociali. Una speciale attenzione ca-based scholars. zione dell’ex-IsIAO. This first issue of the new series of Af- è riservata – secondo la tradizione della The first issue of the journal’s new se- Questo primo fascicolo della nuova se- rica, which is now an independent jour- rivista – al lavoro degli studiosi residenti ries contains articles and reviews re- rie di Africa, ora testata indipendente nal published by a university center, e operanti nei paesi africani. ceived in the wake of the announcement pubblicata da un centro universitario, emphasizes both continuity and change. Il primo volume della nuova serie si of Africa’s revival in May 2018. The appare in una veste che vuole rimarcare While Filesi’s original ambitions re- compone di articoli e recensioni perve- months that followed have involved t o r na a f r i c a is b a c k 7 vista si prefiggeva come scopo precipuo erature and art” all fell within the am- continuità rispetto alla storia iniziata nel main relevant in the present, Africa also lo sviluppo dei rapporti fra studiosi di bitious remit of Africa, whose primary 1965, seppure con cambiamenti che da wants to steer clear of the shortcomings diversa nazionalità, in particolare quelli purpose was the promotion of relation- un lato sono risposta a criticità emerse da that any dispassionate evaluation of the operanti nel Continente africano. ships between scholars of different na- una valutazione spassionata della storia journal’s past history must recognize, La testata si apriva realmente a una va- tionalities, beginning with those based passata della rivista e, per altro verso, and to reflect the scholarly needs of the riegata platea internazionale di contribu- in the African continent itself. riflettono esigenze inaggirabili poste dal current landscape of African studies. tori. Scorrendo i volumi degli anni Set- Africa truly catered to a richly diverse quadro attuale dell’africanistica. The composition of Africa’s governing tanta-Novanta colpisce l’alta percentuale audience. Browsing through its 1970s- Gli organi direttivi e il comitato scien- bodies and scientific committee attests di autori provenienti da diversi paesi di 1990s volumes, one is struck by the in- tifico danno conto nella loro composi- to the intensely international dimen- Europa e Africa, ma specialmente emer- ternational character of its pool of con- zione della dimensione intensamente in- sion of contemporary African studies. ge una linea editoriale esplicitamente tributors and by an editorial line explic- ternazionale in cui respirano gli odierni This is also borne out by the transition orientata ad accogliere e sostenere atti- itly intended to welcome and actively studi africani, della quale sono peraltro to English as the prevalent language of vamente le voci dei contesti accademici support the voices of African academic evidenze la transizione linguistica verso scientific communication and the un- africani. Tale impostazione fu sostan- milieus. This policy remained in place l’inglese come strumento prevalente di precedented extent of scholarly mobil- zialmente mantenuta dalla direzione di under the direction of Gianluigi Rossi, comunicazione scientifica e le diaspore a ity, which has transformed what were Gianluigi Rossi, che succedette a Filesi who succeeded Filesi in 1994 and led livello mondiale degli studiosi originari once more nationally rooted schools nel 1994 e firmò la rivista fino al dicem- the journal up to its 65th volume of De- di quelle che in passato erano scuole na- of scholarship. Africa embraces these bre 2010, completando il volume LXV, cember 2010, the last of the original se- zionali specifiche e connotate. Di questo changes – changes in which Italian Afri- l’ultimo della serie originaria. ries. mutamento, che interessa in maniera ma- canists themselves are deeply enmeshed La pubblicazione fu quindi sospesa e poi The publication was suspended and, croscopica anche l’africanistica italiana, – while maintaining its own tradition of cessata formalmente a causa della crisi then, formally discontinued as a result la rivista prende atto positivamente, pur operational trilingualism. e quindi dello scioglimento dell’IsIAO, of the crisis and subsequent dissolution mantenendo la propria tradizione di tri- Today, as in the past, the discourse on cui fece seguito un complicato processo of the IsIAO, which was followed by a linguismo operativo. Africa’s realities and the deep motiva- di ricollocazione dell’importante patri- complicated process of relocation of the Oggi come in passato il discorso tions of its social actors is all too often monio librario, archivistico, museale e Istituto’s important library, archives and sull’Africa, le sue vicende, le realtà e le informed by questionable assumptions editoriale dell’Istituto ad opera del Mini- museum by the Ministry of Foreign Af- motivazioni profonde dei suoi attori so- and simplistic preconceptions, which stero degli Affari Esteri e Cooperazione fairs and International Cooperation. ciali è troppo spesso sotteso da assunti militate against the prospect of genuine Internazionale (MAECI). It was in this context that, in 2017, the discutibili e preconcetti riduttivi, a tutto understanding. Fu in questo quadro che nel 2017 il CSPE (Centro Studi Popoli Extraeuro- discapito di quella comprensione della African studies are called upon to en- CSPE (Centro Studi Popoli Extraeuropei pei “Cesare Bonacossa”) of the Uni- complessità necessaria a una corretta in- gage in the important task of dispelling “Cesare Bonacossa”) dell’Università di versity of Pavia – with the support of terpretazione. stubborn ambiguities and recurrent dark Pavia – con il sostegno di alcuni studiosi some scholars who had been members Gli studi africani sono chiamati a un spots. Africa aims to contribute to this membri dei comitati direttivo e scientifi- of Africa’s editorial and advisory boards compito importante nello sciogliere effort by publishing rigorous analyses co di Africa – rivolse all’allora Vicemini- – came up with a proposal to revive the ambiguità pervicaci e dissipare ombre and empirically grounded research and stro Mario Giro una proposta per rivita- journal. Addressed to the then Deputy ricorrenti. Come proprio contributo a by providing Africanists with an inter- lizzare di propria iniziativa la rivista. La Minister, Mario Giro, the request was questo sforzo, Africa si propone di dar disciplinary venue for in-depth discus- richiesta fu accolta a seguito di una labo- accepted following a laborious pro- voce ad analisi accurate e ricerche empi- sion in the areas of the humanities and riosa procedura del cui successo va dato cedure, whose ultimate success owes ricamente fondate, offrendo una sede in- the social sciences. In keeping with the pieno merito, insieme al Viceministro e much to the resolve of both the Deputy terdisciplinare qualificata di confronto e journal’s tradition, special attention will alla sua segreteria, alla determinazione Minister and his staff and the office re- dibattito attenta ai percorsi delle scienze continue to be given to the work of Afri- dell’ufficio responsabile per la liquida- sponsible for liquidating the ex-IsIAO. umane e sociali. Una speciale attenzione ca-based scholars. zione dell’ex-IsIAO. This first issue of the new series of Af- è riservata – secondo la tradizione della The first issue of the journal’s new se- Questo primo fascicolo della nuova se- rica, which is now an independent jour- rivista – al lavoro degli studiosi residenti ries contains articles and reviews re- rie di Africa, ora testata indipendente nal published by a university center, e operanti nei paesi africani. ceived in the wake of the announcement pubblicata da un centro universitario, emphasizes both continuity and change. Il primo volume della nuova serie si of Africa’s revival in May 2018. The appare in una veste che vuole rimarcare While Filesi’s original ambitions re- compone di articoli e recensioni perve- months that followed have involved

Africa, N.S., I / 1, 2019 8 t o r na a f r i c a is b a c k nuti in seguito all’annuncio di ripresa intense work on the part of authors, re- diffuso nel maggio 2018. Sono stati mesi viewers and editors, who spared no ef- di lavoro intenso per autori, revisori, e fort to ensure the resumption of publica- redattori, che non hanno risparmiato im- tion at a very short notice. Africa wishes pegno per la ripresa della pubblicazione to thank them and expresses a debt of in tempi tanto brevi. Africa li ringrazia ed gratitude to all those who actively con- esprime inoltre un debito di riconoscen- tributed to its resurrection, particularly za verso altri che hanno contribuito fat- Alberta Bonacossa, a constant source of tivamente a rendere possibile la rinascita support for the CSPE’s initiatives, and della rivista, in particolare Alberta Bona- Adriano Rossi, the president of ISMEO- cossa, partecipe sostenitrice delle attività Associazione Internazionale di Studi sul del CSPE e Adriano Rossi, presidente di Mediterraneo e l’Oriente/Il Novissimo ISMEO-Associazione Internazionale di Ramusio. Studi sul Mediterraneo e l’Oriente/Il No- vissimo Ramusio.

PV Tom McCaskie

t h e p a r a m e t e r s o f a s a n t e h i s t o r i c a l consciousness : reflections o n tϽp r ɛ a n d temporality

Abstract

This paper follows on from recent contributions I have made to the problem of the nature of history as conceived of by the Asante. Here I am concerned with the issues surrounding consciousness of the past as history, memory and/or unknowing. The core of the paper, from which broader arguments are teased out, is a detailed treatment of the richly documented but somewhat neglected ritual performance of ‘the dance of death’ (tͻprɛ). If the paper has an underpinning it is the thought that African history as lived, experienced and thought about by Africans needs to be grounded anew in an approach that eschews the paradigms imposed by western historiography and ethnography. Simply, I call for deep and sustained long-term research into specific African cultures like Asante, so as to unravel the complexities bound together in and by certainty and uncertainty. k e y w o r d s : a s a n t e , d a n c e , d e a t h , a f r i c a n h i s t o r y, o r a l t r a d i t i o n s ; tͻp r ɛ

A brief preface

The evaluation and use of oral traditions have long been a touchstone of studies of the African past, although their status continues to be a contentious matter. Led by the pioneering Vansina, historians have deployed oral traditions in combination with insights from disciplines like linguistics and archaeology to study the more remote past of African peoples1. It is the case, however, that followers of Vansina like Henige and Thornton have manifested a skeptical reluctance to take oral traditions on what we might describe here as their own indigenous intellec-

1. Canonically, J. Vansina, De la tradition orale: essai de méthode historique, Tervuren, Musée Royale de l’Afrique Centrale, 1962, and J. Vansina, Oral Tradition as History, Madison WI, University of Wisconsin Press, 1985. See too the evolving interpretations in J. Vansina, Paths in the Rainforest: Toward a History of Political Tradition in Equatorial Africa, Madison WI, University of Wisconsin Press, 1990, and in J. Vansina, How Societies Are Born: Governance in West Central Africa Before 1600, Charlottesville VA and London, University of Virginia Press, 2004.

Africa, N.S., I / 1, 2019 10 a r t i c o l i / a r t i c l e s tual terms. In pursuit of what they conceive of as a recuperable veridical truth, such historians privilege, or take refuge in, the written, the chronological, and hence the externally authored2. Anthropologists, and most notably Sahlins, will have none of this. To such commentators the dismissal or marginalization by historians of the mythological or unbelievable in oral traditions is simply a category error, for such items are inscribed in memories and lived as Habitus. Thus, in confining their concerns to documented reality, historians embrace a barren posi- tivism that fails to address the fact that history itself, to paraphrase Sahlins, is necessarily atemporal and cultural through and through3. Others have now taken this point of view forward and called for a comparative anthropology of history making, a task that remains nebulous as its proponents cannot say what the actua- lization of such a project might look like4. This present paper seeks neither to resolve nor to dissolve the contentions just described. Instead, and following on from some of my recent publications, it sets out to explore some crucial aspects of the dialogue between the experiences of an African people and their consciousness of temporality, memory and unknowing in and as history5. The setting is the three-hundred-year-old Akan Twi-speaking West African forest culture, society and polity of Asante, now located in the republic of

2. See the influential D. Henige, The Chronology of Oral Tradition: Quest for a Chimera, Oxford, Clarendon Press, 1974. For Thornton’s continuing if now slightly more qualified skepti- cism about oral traditions see J. Thornton, “Modern Oral Tradition and the Historic Kingdom of Kongo”, in P. Landau, ed., The Power of Doubt: Essays in Honor of David Henige, Madison WI, Parallel Press, 2011, 195-208; J. Thornton, “Bridging the Gap between Oral and Written Sources: The Kingdom of Kongo”, in G. Castryck, S. Strickrodt and K. Werthmann, eds., Sources and Meth- ods for African History and Culture: Essays in Honour of Adam Jones, Leipzig, Leipziger Univer- sitätsverlag, 2016, 27-44; and most recently J. Thornton, “The Origins of Kongo: A Revised Vi- sion”, in K. Bostoen and I. Brinkman, eds., The Kongo Kingdom: The Origins, Dynamics and Cosmopolitan Culture of an African Polity, Cambridge, Cambridge University Press, 2018, 17-40. 3. See most recently and pertinently M. Sahlins, “The Atemporal Dimension of History: In the Old Kongo Kingdom, For Example”, in D. Graeber and M. Sahlins, On Kings, Chicago, HAU Books, 2017, 139-221; for some insight into the development of Sahlins’s ideas in the context of his lifelong research in Polynesia see M. Sahlins, Apologies to Thucydides: Understanding History as Culture and Vice Versa, Chicago, Chicago University Press, 2004; and for Sahlins in context consult now A. Gol- ub, D. Rosenblatt and J. Kelly, eds., A Practice of Anthropology: The Thought and Influence of Mar- shall Sahlins, Montreal and Kingston, McGill-Queen’s University Press, 2016. There is shrewd comment in M. Lambek, “Marshalling Sahlins”, History and Anthropology, 28, 2, 2017, 254-263; and compare here T. Wiseman, “Roman Legend and Oral Tradition” in his Historiography and Imagina- tion: Eight Essays on Roman Culture, Exeter, University of Exeter Press, 1994, 23-36. 4. Consult C. Stewart, “Historicity and Anthropology”, Annual Review of Anthropology, 45, 2016, 79-94, and S. Palmié and C. Stewart, “For an Anthropology of History”, HAU: Journal of Ethnographic Theory, 6, 1, 2016, 207-236. 5. See T. McCaskie, “Telling the Tale of Osei Bonsu: An Essay on the Making of Asante Oral History”, Africa, 84, 3, 2014, 353-370, reprinted in my Asante, Kingdom of Gold: Essays in the History of an African Culture, Durham NC, Carolina Academic Press, 2015, 967-983; and T. Mc- Caskie, “Dreamworlds: Cultural Narrative in Asante Visionary Experience”, in T. Green and B. Rossi, eds., Landscapes, Sources and Intellectual Projects of the West African Past: Essays in Hon- our of Paulo Fernando de Moraes Farias, Leiden, Brill, 2018, 353-376. McCaskie, The Parameters of Asante Historical Consciousness 11

Ghana. The core matter to be interrogated is the context, content and meanings embedded in and expressed via a single, and quite singular, but frequently recurring historical event that is well documented empirically and also richly and vividly memorialized in Asante oral tradition. It will be seen that in their own understandings of this action, the Asante reveal much about their consciousness of what their history is or might be. At the same time, they throw light on their conceptualizations of the frontiers between a recuperable temporality and the more elusive but none the less palpable trace and impact of memory and unknowing.

The dance of death (tͻprɛ) performed

The recurring event in question was severally observed and reported on in detail by nineteenth-century European visitors to the Asante capital of Kumase. The event itself took the form of ritualized capital punishment for the most hei- nous of crimes, notably murder (awudie) and adultery (ayɛfere) with the wives of the Asante king6. Despite a plethora of evidence, this event has been little explored by modern scholars of Asante. Relative neglect is probably attributable to two conjoined factors. First, abundant eyewitness reportage is all at once disqui- eting, distasteful and repellent in its own right. Second, the external sources that recount its transaction in markedly horrified but prurient detail are a part of that racist and imperialist literature that for its own purposes stigmatized precolonial Asante as a barbaric polity addicted to irrational bloodletting7. In Asante Twi the event in question is called (a)tͻp(e)rɛ goru, and simply tͻprɛ in everyday vernacular speech8. The full name is derived from the verbal roots perɛ, meaning to struggle, to contend, to expend great effort, all with the implication of contorting the body, and goru, meaning to perform, to enact, to play, all with the implication of dancing to the accompaniment of drums and/or horns. Thus, tͻprɛ is

6. On adultery consult T. McCaskie, “State and Society, Marriage and Adultery: Some Con- siderations towards a Social History of Pre-Colonial Asante”, Journal of African History, 22, 4, 1981, 477-494, reprinted in my Asante, Kingdom of Gold, 121-140; and J. Allman, “Adultery and the State in Asante: Reflections on Gender, Class, and Power from 1800 to 1950”, in J. Hunwick and N. Lawler, eds., The Cloth of Many Colored Silks: Papers on History and Society Ghanaian and Islamic in Honor of Ivor Wilks, Evanston IL, Northwestern University Press, 1996, 17-65. 7. For pertinent comment see T. McCaskie, “Exiled from History: Africa in Hegel’s Aca- demic Practice”, History in Africa, 46, 2019, forthcoming. 8. There is a summary account of tͻprɛ in T. McCaskie, State and Society in Pre-Colonial Asante, Cambridge, Cambridge University Press, 1995, 254-7 and 314. For general discussions of the killing of human beings in Asante consult C. Williams, “Asante: Human Sacrifice or Capital Punishment? An Assessment of the Period 1807-1874”, and I. Wilks, “Asante: Human Sacrifice or Capital Punishment? A Rejoinder”, International Journal of African Historical Studies, 21, 3, 1988, 433-41 and 443-452; see more generally R. Law, “Human Sacrifice in Pre-Colonial West Africa”, African Affairs, 84, 334, 1985, 53-87. For relevant discussion about Asante see T. McCaskie, “Death and the Asantehene: A Historical Meditation”, Journal of African History, 20, 3, 1989, 53-87, reprinted in my Asante, Kingdom of Gold, 305-331.

Africa, N.S., I / 1, 2019 12 a r t i c o l i / a r t i c l e s most often rendered in spoken English by Asante people as the dance, performance or theatre of death. In briefest summary at this point, the transaction of tͻprɛ involved a day-long process of execution through ritual dismemberment, and it con- cluded with the dead victim’s disjecta membra being cast away into awerɛfiri, the material and metaphysical oblivion of erasure and forgetting. It is evident that the frequency of tͻprɛ fluctuated with the rate and number of capital crimes that fell under its rubric. In addition, all informants are agreed that the personality, will or even mood of the reigning king of Asante affected outcomes. Thus, the third Asantehene Kusi Obodom (d. 1764) was memorably reluctant to resort to capital punishment and favoured banishment and exile instead. By contrast, the ninth Asantehene Kwaku Dua (d. 1867), or ananse (the wise), was punctilious in carrying out all of the required orthodoxies of custom and law9. He was also an authoritarian who thought and said that his subjects needed to be kept in order by the constant threat of dire consequences for transgression. Many external sources furnish more or less detailed accounts of tͻprɛ between the 1810s and the 1870s, and it is perhaps not coincidental that the fullest such narratives come from horrified Christian missionaries commenting on the actions of Kwaku Dua. Let us start with the most circumstantially detailed eyewitness account of what actually happened in the act or performance of tͻprɛ. This took place on Monday 8 July 1844, by Asante reckoning kwadwo or day 34 in the adaduanan calendrical cycle of 42 named days. It followed on the recurring celebration of the adae kɛseɛ on day 33, and all the evidence confirms that kwadwo was the prescribed day for carrying out ritualized executions by tͻprɛ10. The observer on this occasion was George Chapman, a twenty-six-year-old Methodist missionary, on his own at the time and often afraid or ill in the deeply alien world of Kumase. On this occasion the man undergoing tͻprɛ was convicted of murdering five people in the course of a prolonged rampage in Kumase and the surrounding villages. He himself “belonged to the company of executioners”, and so his sentence was carried out by his peers and colleagues, or even his friends and patrilineal kin. It is apparent that a horrified but disbelieving fascination explains the lavish detail in Chapman’s account, and perhaps accounts for the fact that it was set down in a somewhat breathless narrative wholly devoid of paragraphs. Be that as it may, it is reproduced here in all of its revelatory entirety. In order that his death might be the more public, he was reserved until the “Adai Mon- day”, and at 9 a.m. commenced his tortures. A number of executioners were chosen,

9. Consult T. McCaskie, “Custom, Tradition, and Law in Precolonial Asante”, in E.A.B van Rouveroy van Nieuwaal and W. Zips, eds., Sovereignty, Legitimacy and Power in West African Societies: Perspectives from Legal Anthropology, Hamburg, Lit Verlag, 1998, 23-47, reprinted in my Asante, Kingdom of Gold, 465-482. 10. For equivalences between Asante and European calendars consult T. McCaskie, “Time and the Calendar in Nineteenth-Century Asante: An Exploratory Essay”, History in Africa, 7, 1980, 179-200, reprinted in my Asante, Kingdom of Gold, 99-120. McCaskie, The Parameters of Asante Historical Consciousness 13

who commenced their work, first by pinioning the man, and then passed two knives through his cheeks, to which another two were again added. Afterwards the cartilage of the nose was cut, so as to admit a cord being passed through, by which he was conducted through almost every street in the Town [Kumase]. Before leading him forth two more knives were thrust into his shoulders, in a line with the backbone to a depth of five or six inches, and in this condition he was conducted to the house of every Chief in the Town, who plentifully regaled his tormentors with rum or palm wine. While being thus dragged through the Town [...] his ears were cut partly off, and occasionally one of the executioners would take hold, with his teeth, of the piece hanging down, and pulling with considerable force put the poor wretch to great pain. Frequently, the knives were withdrawn from his shoulders, and the bleeding wounds filled with the juice of limes, or some other sharp liquor. Often for the diversion of his tormentors, he was compelled to dance. This they accomplished either by sharply pulling the cord in his nose, or the application of the point of the knife to different parts of the body. This cruel sport was continued until 5 p.m., when the King [Asantehene Kwaku Dua], with many of the Chiefs took their seats in public. The work of torment then commenced in ear- nest. The man being brought forward a circle was formed, around which he was led, attended by seven or eight executioners, each armed with his horrid knife. Here again he was compelled to dance for a considerable length of time, his tormentors in savage derision joining with him. The knife was now applied more freely, pieces of flesh were cut off from various parts of the body, and the wounds rubbed with [gun]powder, and the juice of the lime. The wretched man shrank beneath this operation, but his writhing with torment only furnished an opportunity to his tormentors to employ the most bitter sarcasms. “O don’t be afraid, we will nor hurt you”. “Don’t think we are killing you, nothing of the sort. You want medicine, and this we apply is good for you”. Cutting away another piece of flesh, one would exclaim, “Now we are not giving you pain. O no, nothing of the kind. We are teaching you a little, giving you a little sense, and this is good for you”. Then placing a number of knives in the ground, with the points up- wards, two or three of his tormentors compelled him to sit down upon them, at the same time addressing him with, “Now this is a fine seat for you. Do you not want a good one? O this is most excellent”. Raising him up, he was required again to dance, but growing exhausted, the poor wretch moved slowly. Lime juice and [gun]powder were again applied, and again he was led on by the executioners. Standing still for a few seconds, the skin covering the head was cut open, and pieces taken from the back of the head. These were put in his face, while his tormentors jeeringly asked, “Did you ever before see the back of your head? O you did not. Well, look, look. Now you see it, we show you what you never before saw”. Pieces were then cut from the buttocks, and the back part of the thighs. These were shown to him, accompanied by the same laughing expressions. But nature was now fast sinking. What remained to be done must be done with speed. The cap of the knees (patella) were now taken off, and the flesh from the leg. The bone being laid bare was next scraped, but the man was fast sinking. Accordingly the fingers were cut off the hands, and then the feet. Just then the miserable wretch was at his last gasp, and in an instant, the head was severed from what remained of the body11.

11. Rhodes University, Grahamstown, South Africa, Cory Library, Ms. 15 104, “George Chapman, Journal (25 January 1843 to 15 February 1857)”, entry Kumase, 9 July 1844. Chapman

Africa, N.S., I / 1, 2019 14 a r t i c o l i / a r t i c l e s

With varying nuances of circumstantial detail, Chapman’s narrative of tͻprɛ is confirmed or occasionally even amplified in the several accounts of it left by external witnesses in the nineteenth century12. Thus, on 24 March 1862, also 34 kwadwo, the Methodist missionary West witnessed a convicted murderer being done to death. His extended account of tͻprɛ parallels that of Chapman, but adds some information. So, when at last the condemned man was brought before As- antehene Kwaku Dua, the king “made a speech to the public, telling them, that if any of them committed murder he would have them tortured in the same way”. West added that the denouement that followed, when the victim was “literally cut to pieces”, took place in a Kumase “street” called “Blood never dry”, a matter that is returned to below. Salient in all external accounts of the performance of tͻprɛ were three general themes. One was the very threatening mien, outlandish appearance, and ag- gressively unpredictable physical and verbal behaviour of the executioners, the last being fuelled by the relentless consumption of alcohol. Bowdich in 1817 reported they were “men disfigured with immense caps of shaggy black skins”, and also impresarios of a “horrid barbarity” that had an unforgettable impact on all who witnessed them at work. In the 1840s Freeman opined that “no language can describe the wild ferocious appearance of these men” with their long matted hair flying all about “as they spring wildly about in frantic gesture”. West in 1862 noted that the executioners augmented “the natural blackness” of their faces by “the free use of charcoal”, with the consequence that they looked like “demons” and behaved as such by “grinning, staring, throwing themselves into all sorts of threatening attitudes and making the most hideous faces”. The second was what de Heer in 1866-7 termed the insistent beating of “the Atopieree drum”, a constant refrain throughout all the hours of the procedure, and its abrupt silence only at the climax immediately upon the condemned person’s beheading. This was a

(1817-93) was in the Gold Coast and Asante (1843-5), and then served as a missionary in the East- ern Cape in South Africa from 1849 until his retirement. 12. See inter alia W. Huydecoper, trans. G. Irwin, Huydecoper’s Diary: Journey from Elmina to Kumasi 28 April 1816 to 18 May 1817, Institute of African Studies, University of Ghana-Legon, 1962, entry for 10 July 1816; T. Bowdich, Mission from Cape Coast Castle to Ashantee, London, John Murray, 1819, 33; L. Yarak, “A Dutch Embassy to Asante in 1857: The Journal of David Mill Graves”, History in Africa, 21, 1997, 272, entry for 22 August 1857; SOAS Archives, London, (Wesleyan) Methodist Missionary Society Archive, West Africa, Gold Coast Correspondence, MMS/06, Letter W. West to General Secretaries, Cape Coast, 9 June 1862; SOAS Archives, MMS/06, T.B. Freeman, “Life and Travels: Gold Coast, Ashantee, Dahomey”, unpublished Ms., c. 1860, 95-97; Koninklijk Instituut voor Taal-, Land-, en Volkenkunde, Leiden, Ms. H 509, P. de Heer, “Aanhangsel: Extract uit het Journaal gehouden te Comassie door eenen tapoeier”, 1866- 1867, entries for 19 December 1866 and 1 February 1867 (in both of which tͻprɛ is mentioned by name as “Atopieree”); Basel Mission Archives, Basel, F. Ramseyer, “Tagebuch, 1869-74”, Ms. entry for 5 June 1871 and passim; and C-H. Perrot and A. van Dantzig, eds., Marie-Joseph Bonnat et les Ashanti: Journal (1869-1874), Paris, Mémoires de la Société des Africanistes, 1994, 600-603 and passim. All materials cited in this and the following paragraph are drawn from the references given here. McCaskie, The Parameters of Asante Historical Consciousness 15 signal to all in Kumase of what was taking place and when it was completed. The third was the attendance of fascinatedly excited crowds of Asante onlookers. Ex- ternal observers deplored this as a sign of bloodthirsty voyeurism, conveniently suppressing all recall of Europe’s own recent past of levels of public torture and death every bit as calculatedly violent as anything witnessed in Kumase. In As- ante eyes the stark finality of tͻprɛ all at once condensed and underlined the in- evitability of death, and this is reflected in well-known proverbs. Thus, tͻprɛ twene yɛmpata no (lit. “don’t leave things till too late”; impl. “past a certain point there is no coming back”), and (a)tͻprɛwuo nyɛ mpatawuo (lit. “death by torture and execution is hardly unexpected” impl. “whether you are a strong or weak person but a guilty one, you should not be surprised by having to pay for it with death”)13. There is a remarkable uniformity in reports of the procedural details of tͻprɛ as these were recorded by uncomprehending foreign witnesses. This suggests that it was prescriptively ritualized in terms of its sequential temporal and structural unfolding. It is fortunate for the historian then that this understanding was confirmed in the 1920s when the anthropologist Rattray wrote down an account of tͻprɛ recounted to him by the Kumase Adumhene Kwabena Asamoa Toto. During the last three decades of the nineteenth century this “delightful, humane, and be- nign old gentleman” served as an executioner in Kumase, and he “had taken a leading part on several occasions” in the events he decribed14. The participant testimony supplied by Kwabena Asamoa Toto to Rattray is a ritual catechism, a prescribed listing of the order and timing of successive acts to be undertaken throughout a day of tͻprɛ. His evidence can be readily tabulated as follows.

● The condemned man was seized without warning and an executioner’s knife (sɛpͻ) was driven through his cheeks to prevent him from cursing the As- antehene. ● About six in the morning he was taken off to the place in Kumase called nkram (lit. “in the midst of blood”; that is, West’s “Blood never dry”.) He was seated there and the gyabom sumaŋ was laid in his lap15.

13. Consult P. Appiah, K.A. Appiah and I. Agyeman-Duah, Bu Me Be: Akan Proverbs, Accra, Centre for Intellectual Renewal, n.d., 1086-1087, proverbs 6312-6313. 14. R.S. Rattray, Religion and Art in Ashanti, Oxford, Clarendon Press, 1927, 87-89. Much has been written about Rattray, but the most useful essays for present purposes are A.W.B. Simpson, “R.S. Rattray and Ashanti Law”, in his Legal Theory and Legal History: Essays on the Common Law, London and Ronceverte, Hambledon Press, 1987, 403-426, and G. Pescheux, “Ronald Suth- erland Rattray: un structuro-fonctionnaliste avant la lettre”, Journal des Africanistes, 75, 1, 2005, 101-113. 15. For the location of nkram and some other sites mentioned here consult the map in Mc- Caskie, State and Society, 148-50; and see too Rattray, Religion and Art, Fig. 43 between 112-113. For an illustration of the gyabom sumaŋ and some discussion see McCaskie, State and Society, 256 and 312. For the composition of the gyabom sumaŋ and its use against “evilly disposed disembod- ied human spirits” consult R.S. Rattray, Ashanti, Oxford, Clarendon Press, 1923, 99-100.

Africa, N.S., I / 1, 2019 16 a r t i c o l i / a r t i c l e s

● The nasal septum was pierced and a thorny creeper named kokora was threaded through the holes and used to lead or pull the condemned man through Kumase16. ● Four more sɛpͻ knives were thrust into various parts of the body, due care being taken not to damage a vital organ by inserting them too deeply. ● The condemned man was led or pulled by the kokora through his nose to the place in Kumase called tͻprɛ dua ase (“beneath the tͻprɛ tree”)17. ● The condemned man was led by the nose to the asafo quarter of Kumase, where the Asafohene/Akwamuhene who resided there cut off the left ear. ● The condemned man was led by the nose to bantama just outside of Ku- mase, where the royal mausoleum (baamu) was situated, and the Kontihene/Ban- tamahene who resided there severed the right ear and scraped the right shin down to the bone. ● The condemned man was led by the nose back to tͻprɛ dua ase where his itinerary through Kumase had commenced. There, shaded by the tree, he was compelled to dance for the rest of the day while keeping time to the rhythm of the tͻprɛ drum. ● After nightfall the condemned man was dragged to the place of public assembly named bogyawe that was adjacent to the royal palace. There, the Asante- hene and his office holders were already seated in state “to witness the final dis- patch of the victim”. The arms were hacked off at the elbows, and the legs below the knees. The eyelids were cut away so that the victim might more clearly “see” what was taking place. ● The condemned man was ordered to continue dancing in time to the tͻprɛ drum. His ruined body was unable to do this, and so his buttocks were sliced off and he was “set down on a little pile of gunpowder, which is set alight”. A “slab” of skin was then cut off the back, and put before his face with the pleasantry, efise wo ni wo’se ewo wo, wa hunu wo’kyiri nam? i.e. “Since your mother bore you and your father begat you, have you seen the skin off your back?”. ● The office of executioner is patrilineal, and now the “small” ones, youths and boys in training, go to their “father” the chief of the executioners and com- plain agya wa gye te’sekan (“Father, he has taken our knife”). They receive the formulaic reply ko gye wo ade (“Go and take what is yours”), following which “they are let loose on the dying man and cut pieces of flesh from various parts of

16. The kokora plant is Smilax kraussiana Meisn.; see F.R. Irvine, Woody Plants of Ghana with Special Reference to their Uses, London, Oxford University Press, 1961, 768-769, and O.B. Dokosi, Herbs of Ghana, Accra, Ghana Universities Press, 1998, 502. The slender stems of this plant are covered in prickly thorns and it was used in this form in tͻprɛ; when stripped of its thorns it was used as a strong rope or twine. 17. The species of this tree remains unidentified in the sense that there are several possibilities. Informants say that this tree was seen to weep blood from its leaves when an Asantehene died. For context see T. McCaskie, “Trees and the Domestication of Power in Asante Thought”, in R. Cline- Cole and C. Madge, eds., Contesting Forestry in West Africa, Aldershot, Ashgate Publishing, 2000, 104-123, reprinted in my Asante, Kingdom of Gold, 543-558. McCaskie, The Parameters of Asante Historical Consciousness 17 his body”. The chief executioner now reports to the Asantehene that the condemned man is almost dead, and he is given the necessary royal permission to strike off the head. ● After the beheading the pieces of the corpse were collected together “and cast away in the hollow near the spot formerly called diakomfoase”18.

Let us note that the ordeal of tͻprɛ was reserved for men. Female murderers, of whom none is known for certain from the sources, were most often arraigned on charges of witchcraft (bayi) and strangled if convicted. In a like manner, adul- terous royal wives were done away with by strangulation. If any such offenders had fallen pregnant by their paramours, then they were “kept in the stall” or locked up until they gave birth. They were then killed by means of a noose put around the neck (again kokora, but this time stripped of its thorns and made into rope), and the infant was dismembered over the gyabom sumaŋ and cast away into the bush. As performative transaction tͻprɛ lasted all day. Much of that time was taken up with parading the condemned man through Kumase. Indeed, custom ordained that this progress should pass through all of the canonical seventy-seven quarters or wards (abrͻno aduosonson) of the Asante capital. That number is not to be understood literally, but rather as a well-known symbolic marker of Asante ideas of appropriate wholeness and completeness19. As has been seen, parts of this itinerary were not only prescribed but also ritualistically foregrounded and emphasized. The condemned man was brought in turn before both the Kumase Asafo- hene/Akwamuhene and Kontihene/Bantamahene at their residences. These office holders were respectively second in command and commanding general of the Kumase army, both being charged with guardianship of the Asante capital and of order within it. These two office holders mutilated the condemned man in a precisely set out and meaningfully charged way. As was the norm in Asante, the junior preceded or went before the senior. In this case the junior amputated the ear from the subordinate left hand side (benkum) of the body, with the ear and the shin of the superior right hand side (nifa) being reserved for subsequent attention by the senior. All of this chimes with embedded and well understood Asante attitudes towards the mediation and parcellization of time and space, a roster of socio-cultural order that was affirmed and renewed in inherited and repeated ritually approved enactments20. As outlined above, tͻprɛ was all at once a procession, a progress and a spectacle that lasted through an entire day. European observers deplored the chaotic

18. “The eater of priests” (diakomfoase) was an execution site close by dwaberem, the largest Kumase market place and the main venue for public assemblies. 19. The iconically numbered “Seventy-Seven Laws of Komfo Anokye” are to be understood in the same way; consult McCaskie, “Custom, Tradition, and Law”. 20. The supreme instance was the performance of the annual Kumase odwira celebration, for which in detail see McCaskie, State and Society, 144-242.

Africa, N.S., I / 1, 2019 18 a r t i c o l i / a r t i c l e s train of men, women and children who followed it from place to place with avid engagement, for in their eyes this noisy throng only magnified the clamant sav- agery and barbarism of what they had decided they were already seeing. In 1862, as noted, Asantehene Kwaku Dua delivered a harangue in which he threatened any future criminally errant subjects with the horrors of tͻprɛ that they had just witnessed. No doubt there was Schadenfreude involved in gazing upon what was taking place. A well-known proverb says tͻprɛ twene yɛnka nni w’akyi a wose ɛyɛ dɛ: that is, “if the tͻprɛ drum is not played behind you, then you say that its music is sweet and pleasing”. The drum was always played behind the person undergoing tͻprɛ, and so this apothegm is glossed as meaning that “if you are not yourself involved, you can always enjoy a spectacle, even if someone else is suffering”21. As might be imagined, indigenous eyewitness reports of tͻprɛ are scarce but do exist. Thus, between 1917-23 a school teacher and writer from Akuapem talked with a number of elderly people, principally from Abetifi in Kwawu, who had been slaves in Kumase during the reign of Asantehene Kofi Kakari until they fled when the British occupied and burned the city in 1874. These eyewitnesses described tͻprɛ as being a “torture play”, or as a dramatic performance “designed for the public to watch”. Apart from its function as a demonstration of the royal power to punish, it was “a means of entertainment” customarily transacted on a Monday (34, kwadwo) following an adae keseɛ (33, kurukwasie) when Kumase was packed with already excited people. So, “at every public place the party would halt and play music for the prisoner to mumble songs, dance and entertain the public”. From time to time the condemned man might be offered palm wine or morsels of food to revive him so that tͻprɛ was prolonged for the benefit of the spectators who jeered at and sometimes struck him22. There is nothing very surprising in this, for we know from the records of literate societies that ritualized torture, involving pain, bloodletting, death and dismemberment, enthralled audi- ences even as it made them uneasy and disturbed. In such situations spectators were simultaneously invited and forced to confront their own mortality in the violent corporeal dissolution of another being and identity that played out before them23. The performance of tͻprɛ was at bottom dialogical, a transaction between the condemned man and the executioners charged with putting him to death. Informants are all agreed that among those classified as royal executioners few emerged as being qualified to carry out ritualized killing. The successful con-

21. Appiah, Appiah and Agyeman-Duah, eds., Bu Me Be, 1086, proverb 6311. 22. Consult A.A. Anti, The Ancient Asante King, Accra, Volta Bridge Publishing, 1974, especially 38-40 and 44-48. 23. Famously, the horrendous public torture and execution of the failed regicide Robert- François Damiens in Paris in 1757 fascinated and raised issues of the sort mentioned here for Michel Foucault; a historical account that supplies the details and analyzes the eyewitness responses is in P. Friedland, Seeing Justice Done: The Age of Spectacular Capital Punishment in France, Oxford, Oxford University Press, 2012, 176-183. McCaskie, The Parameters of Asante Historical Consciousness 19 duct of tͻprɛ was supremely challenging. It involved calculated bloodletting and mutilation over many hours, but with the objective of keeping the victim alive and sentient until his ordeal ended at nightfall. Then he was brought before the Asantehene, in a state of consciousness that allowed him understanding of the final recitation of his crime(s) and of the instruction to behead him. To have brought about the untimely death of the condemned person before this ritually prescribed denouement before the Asantehene was a capital crime in itself. Thus, in the early 1870s the executioner Yaw Dapaa of Kwapra, a few miles northwest of Kumase, was done to death for killing a man undergoing tͻprɛ before the victim could be brought before Asantehene Kofi Kakari24. Inquiries over many years about the qualities making for a skilful and successful executioner generally produce as a first response the concept of innate “aptitude” (biribi a wotumi sua no ntɛm na woyɛ no yie)25. Beyond that, capacities can be ascribed to the fact that like all royal servants (nhenkwaa) executioners were trained up to specialist office and function by their fathers and paternal uncles. So, a man who was skilled in the practice of ritual killing might pass on his physical and metaphysical attributes to one or more of his sons or patrilineal nephews26. Yet others might derive their ‘aptitude’ from the favour and guidance of a shrine, such as beposo asiakwaa twaa at Sereboso. As the name itself implies, the metaphysical entity in this shrine communicated via a bundle of knives as used by executioners, but of supernatural origin. However, the fundamental relationship in play was between the executioner and his individual access to tumi (as given in the Twi phrase for ‘aptitude’ above), an otherworldly source of power that enabled specific forms of disciplined mastery over an individual selfhood and of the practices it was engaged in. It is averred that while involved in tͻprɛ the most accomplished executioners “saw the procession” (santen) of former Asante kings standing together with the living ruler, at once a mnemonic and a sanction of the past in the present27.

24. Manhyia Record Office, Kumase, CRB 20/7, Kwaprahene Kwame Dapaah (Pl.) vs. Yaw Owusu Sekyere (Def.), 1928, evidence of Defendant, Kumase, 29 August 1928. 25. I am most grateful to Asantehene Opoku Ware II (1970-1999) who assembled together a number of executioners (abrafoͻ and tͻprɛfoͻ) for my benefit in 1990, 1994 and 1996, and was an interested participant in my talks with them. 26. Compare here on the asokwafoͻ (hornblowers) and batafoͻ (carriers of the king’s trade) T. McCaskie, “‘You Are the Music While the Music Lasts’: Kwame Tua between the Asante and the British”, and T. McCaskie, “‘History Has Many Cunning Passages’: Kwasi Apea Nuama between the Asante and the British”, Africa, 88, 2, 2018, 205-221 and 222-237. 27. The concept of tumi is complex and its workings diverse. For a useful overview consult E. Akyeampong and P. Obeng, “Spirituality, Gender, and Power in Asante History”, International Journal of African Historical Studies, 28, 3, 1995, 481-508. For some of its expressions see most recently T. McCaskie, “Unspeakable Words, Unmasterable Feelings: Calamity and the Making of History in Asante”, Journal of African History, 59, 1, 2018, 2-20, and T. McCaskie, “Dream- worlds”.

Africa, N.S., I / 1, 2019 20 a r t i c o l i / a r t i c l e s

The generality of executioners (abrafoͻ) served as a constabulary that guard- ed the Asantehene, had oversight of his wives, palace and regalia, kept order in Kumase, and were charged with the task described in their name with regard to all sacrificial immolations and some criminal executions. Importantly, they also functioned as the custodians of oral poetry in their capacity as bards in the recitation of apae (royal “praise poems”). The abrafoͻ were organized into four groups, each with its own stool holding chief, and all served the king under the authority of the Kumase Adumhene. All of these groups traced their roots to the older historic Akan polities that lay to the south of Asante. Thus, the Ano group originated in Akwamu, and escorted Osei Tutu from there to Kumase when he was summoned home to succeed Obiri Yeboa and become the first Asantehene; the Twen- uroase group emigrated from Amantuom in Denkyira; the Nkram group came from Nyaadom in Adanse in the reign of the second Asantehene Opoku Ware (d. c. 1750); and the Boaman group was created at the order of the fourth Asantehene Osei Kwadwo (d. 1777)28. Many of these nhenkwaa had only a supporting role at most in the specialized ritual of execution that was tͻprɛ. For example, and as the name suggests, members of the Nkram group were selected to provide the guard around a man undergoing tͻprɛ when he was at the venue called “in the midst of blood” (nkram). The elaborated, prolonged, and very demanding form of execution called tͻprɛ was carried out by individuals selected on a case by case basis from among a small cohort of specialists quite distinct from all other executioners, and known simply if unsurprisingly as tͻprɛfoͻ. They originated in the very old settlement of Akrokyere (or Akrokerri) in Adanse, and defected to Asante during the confronta- tion between their overlord Denkyirahene Ntim Gyakari and Osei Tutu which eventually led on to the defeat and death of the former in 170129. They brought with them their drum known as tͻprɛ, the origins of which will be discussed in due course, their horn known as aworobɛn, and the music and dance associated with both. As noted, the prescribed day for tͻprɛ to take place was 34 kwadwo, and hence (a)tͻp(e)rɛ to a etwa dwoͻda (“tͻprɛ is to take place on that Monday”). The tͻprɛ drum was carried behind the condemned man and played the same refrain repeatedly: asobrakyeɛ wate?, wokͻ tͻ (“Asobrakye, have you heard? You are going to die”). Then, immediately before and after the beheading the drum beat out the short declarative finale:yɛresu ara, oo, asɛm na aba, oo (“We are all weeping, for something or other has taken place”). At the start and the finish the staccato announcement awͻmu (“piercing”) might also be played30.

28. A useful summary with excellent photographs is to be found in K. Ampene and Nana Kwadwo Nyantakyi III, Engaging Modernity: Asante in the Twenty-First Century, Ann Arbor MI, Maize Books, 2016, 166-175. 29. For context see T. McCaskie, “Denkyira in the Making of Asante, c. 1660-1720”, Journal of African History, 48, 1, 2007, 1-25, reprinted in my Asante, Kingdom of Gold, 741-765. 30. I discussed the term asobrakyeɛ with the late ͻkyeame Owusu Banahene. The word suggests a derivation from aso (an ear) and bra (a law), with the implication of calling people to listen McCaskie, The Parameters of Asante Historical Consciousness 21

Two observations can be added here. First, a single member of the tͻprɛfoͻ was chosen to dance in front of the condemned man as he processed through Ku- mase. No one else was permitted to do this – odi atͻfͻ kan (“leading the condemned”) – and would be executed if they did so. Second, and to repeat, the victim still had to be alive at the conclusion of his ordeal when he was brought before the Asantehene for final admonition and beheading. At this stage, to bring an already dead person before the king was in and of itself a capital offence, and it meant execution for the offending member of the tͻprɛfoͻ. Simply, the prescribed and sanctioned performance of the ritual demanded that the victim remain conscious of everything that was happening to him. At the same time, neither the mutilated victim nor the executioners drenched in his blood were permitted any physical contact with the Asantehene. Blood (mogya) of any sort, let alone that of a criminal transgressor, was at once ambiguous and deadly because of its intimate intertwining with the fundamentals of life and death31.

The dance of death (tͻprɛ) interrogated

In the last two decades of the twentieth century I passed a deal of time talking with Asantehene Opoku Ware II (d. 1999), either one on one or in an invited gath- ering (saadwa) where drink and talk functioned very much as in the Athenian symposium described by Plato32. One recurring topic of conversation was the Asante concept of anidaha, a word that embraces ideas of consciousness of history and of the past, as it also does of other types of awareness. At that time, I was engaged with colleagues in the preparation for annotated publication of the manuscript ‘History of Ashanti Kings’ by Asantehene Agyeman Prempe (d. 1931)33. One item was strikingly different from anything else in this text. It recounted an extraordinary tale in which two Adanse hunters hidden in the forest witnessed a troop of “Contronfis” (kontromfi) or baboons (Papio anubis chorus) – with knives, a cord, and a drum – carrying out the recognizable rituals of tͻprɛ upon one of their own number34. The burden of this tale of “Toppler” (tͻprɛ) was that the Akan

to the drum proclaiming a law being enacted; see too J.H. Nketia, Drumming in Akan Communities of Ghana, Legon-Accra and Edinburgh, Thomas Nelson, 1963, 132, 137 and 190. 31. For succinct treatment consult McCaskie, State and Society, 203 and 300. 32. For saadwa see ibid., 306. The occasions I am talking of occurred several times between 1992-8. Prominent participants other than the Asantehene included Yaw Andoh, Kwame Arhin, Buasiako Antwi, Owusu Banahene, Osei Kwadwo and Twumasi Panin. I thank each and all of them. Of course, the parallels with an Athenian symposium embrace radical differences as well as striking similarities; see, in brief, N. Spivey, The Sarpedon Krater: The Life and Afterlife of a Greek Vase, London, Head of Zeus, 2018, 80-101. 33. See A. Adu Boahen, E. Akyeampong, N. Lawler, T. McCaskie and I. Wilks, eds., ‘The History of Ashanti Kings and the Whole Country Itself’ and Other Writings by Otumfuo, Nana Agyeman Prempeh I, Oxford, Oxford University Press for the British Academy, 2003. 34. The relevant material is at ibid., 119-122, together with editorial comment.

Africa, N.S., I / 1, 2019 22 a r t i c o l i / a r t i c l e s of Adanse and then Denkyira had copied this practice from these apes, and that it was in the second of these places that it became associated with punishment for committing adultery with a royal wife and for murder. In a quest after elucidation I talked over this passage with Opoku Ware II and with others. The provenance of the specific account remained opaque, but discussion around it threw up a number of points that are sharply pertinent here. Asante tradition and custom, it was agreed, were imaginatively and sedu- lously accretive in the busy grafting of additional detail onto existing textual matter. With respect to tͻprɛ, the dense choreography of the elaborated ritual prac- tised during and reported from the nineteenth century was the creation of the workings of tumi in the thoughts and actions of successive Asantehenes, and of their political constituents. These sought to add to or to build upon received orthodoxies by incorporating items that demonstrated their willingness to stand united together (yɛ ne wͻn na yebesi afram mpreŋ; der. to mingle, to join with, to agree) in defence of the authority that underpinned existing understandings and dispositions. It is known, for instance, that the Asantehene Osei Tutu Kwame (d. 1823) added the text entitled adinkra (yɛ de brɛbrɛ be kum Adinkra; “however slowly we shall surely kill Adinkra”) to his drums including tͻprɛ so as to com- memorate victory in 1818-9 over Gyamanhene Kwadwo Adinkra35. The reverse of this process was also the case. If Asante oral histories were accretive, then they were also mutable and might even fall into redundancy. The Akan philosopher Gyekye rightly points out that custom and history were viewed through the evolving prism of temporal change. So, the proverb mmɛre di adanŋan or “times change” and similar sayings encapsulate acceptance of the fact that old and hal- lowed historical traditions die away and have to be laid to rest or rethought de novo36. Thus, tͻprɛ has now obviously fallen into desuetude, but this is a predict- ably normative outcome in a changed world, and so hardly worthy of comment, let alone the defence of a now vanished past. Discussions of accretion, mutability and redundancy circled around the matter of anidaha in the sense of a historical consciousness. The Asante conceptual- ize and name an era that was remote antiquity. This was tete, and the people who lived then were the tetefoͻ. Mnemonic traces of them survive in fragments of worked stone, or in scratchings on the walls of caves37. These are atete de, or remains of the ancient past. In the constructions placed upon and authorized by

35. There is an account in Nketia, Drumming in Akan Communities, 128-33; for historical context see McCaskie, “Telling the Tale of Osei Bonsu”. 36. See K. Gyekye, Tradition and Modernity: Philosophical Reflections on the African Expe- rience, Oxford, Oxford University Press, 1997, especially 262-263. For context see too K. Gyekye, An Essay on African Philosophical Thought: The Akan Conceptual Scheme, Cambridge, Cam- bridge University Press, 1987, and compare the views advanced in K. Wiredu, Cultural Universals and Particulars: An African Perspective, Bloomington and Indianapolis IN, Indiana University Press, 1996. 37. Consult the oral testimonies in K. Daaku, Oral Traditions of Adanse, Institute of African Studies, University of Ghana-Legon, 1969. McCaskie, The Parameters of Asante Historical Consciousness 23 anidaha the Asante and the other Akan peoples see themselves as being wholly distinct from the tetefoͻ who preceded them in occupation of the same environment. The tetefoͻ were succeeded on the land by the nananom, literally the Akan ancestors. An enormous volume of oral historical evidence records and commem- orates this supposed transition. Comment here, however, is stripped back to matters of immediate concern. First, oral historical tradition records the emergence of what are still termed the akanman pieseɛ nnum, the “five first-born Akan places”, within the auriferous basin formed by the watersheds of the Pra and Ofin rivers. In order of creation, and so of the seniority in age conferred by that, these were Adanse, Akyem, Asen, Denkyira and Asante. In the 1940s, oral traditions were collected from the important settlement of Akrokyere (Akrokerri), already mentioned, in Adanse. These fascinate in the slew of innovations attributed to the earliest inhabitants of this town. They included inter alia the inception of the matrilineage (abusua), the pioneering extraction of gold that was worked up into everything from weights to hilted swords, and during the reign of the putative seventh Akrokyerehene Asare Kumanini we are told of the institutionalization of tͻprɛ or “Otopire, a playing for the criminal persons.” The victim, so we are told, was preceded by men bearing drums, [and] his hands were pi[ni]oned behind him. A sharp thin knife was passed through his cheeks to which his lips were crossed like the figure 8 (but on its side). One ear was cut off and carried before him, the other hung to his head by a small piece of skin. There were several gashes in his back, and a knife was thrust under each shoulder blade. He was led by a [?] passed through a hole in his nose. This torture is very horrible38. The implication is that tͻprɛ and a panoply of other customs, practices, and skills were simply inherited from the tetefoͻ and, pace Asantehene Agyeman Prempe and his tale about the causal example of the kontromfi, this transfer was viewed as a seamless process requiring only assertion without explanation in the great majority of oral historical traditions. Here we are close to the core of the As- ante understanding of anidaha, of the relationship between the akanfoͻ and their predecessors the tetefoͻ, and of the borderland between what constituted history qua history and memories of and unknowing about the legacy of a pre-Akan past. In a further collection of Adanse oral historical traditions from 1963 the account of its primacy among the akanman pieseɛ nnum is repeated, but here we are given the added information that all five Akan communities were created in sequence by “the Supreme Deity of the Universe (Odomankama-Obe-Ade)”39.

38. See Manhyia Palace, Kumase, Papers of Asantehene Osei Agyeman Prempeh II (1931- 70), “History of Adanse Containing Customs and Institurions etc.” in “History of the Immigrants from Takyiman”, typescript, c. 1942, 24. This text also uses the term oman mehwe for tͻprɛ, in the sense of scraping, scratching or cutting (hwe) a person on the orders of the community. 39. See Institute of African Studies, University of Ghana-Legon, Ashanti Stool Histories, IAS/ AS 303: The Five Akan States/Adanse Traditional History, collected by J. Agyeman-Duah, 1 May 1963.

Africa, N.S., I / 1, 2019 24 a r t i c o l i / a r t i c l e s

Granted the centrality of odomankama in much Akan tradition it is at first sight surprising that so little has been written about this entity beyond confused and confusing attempts at assimilation to the Christian tradition40. Indeed, the only sustained historical attempt to try to assess or to understand the quiddity of odomankama is in a long essay by Wilks, and in that the treatment of it plays only an enabling or supporting role in the author’s ongoing attempts to rework and to salvage his ‘Big Bang’ thesis from the 1970s about the late emergence of sedentary agriculture among the Akan and its implications41. Be all that as it may, Wilks does furnish some valuable forensic insights into the understanding of odomankama among the Akan, including the Asante. He takes his cue from the Akan philosopher Wiredu who notes that in oral historical traditions odomankama is hardly ever anthropomorphized, and barely ever assumes human form. Hence, “what sort of creation are we talking about?”. Wiredu responds to his own question by arguing that “the Akan conception is one of demiurgic fashioning out of order from a pre-existing indeterminate stuff rather than creation out of nothing.” Thus, creation was not ex nihilo, but “can only have been a process of transformation”42. That is to say, change and alteration from the world of the tetefoͻ to that of the akanfoͻ and the new order of the akanman pieseɛ nnum. Interest here is in the cognitive and intellectual relevance of the foregoing in relation to Asante historical consciousness. It is commonly said that odomankama bͻͻ wuo maa wuo kum no, that is “Odomankama made death but death killed him”. This is because as the “maker” (bͻͻ adeɛ; see above “Obe-Ade”) of the Akan, odomankama was understood to have come full circle and completed his labours of invention and creation. So, “when death killed Odomankama things in the world were left to the acts of men of affairs” (odomankama kͻwule no ͻde n’akyi gyaa agyina), that is to say to the Akan communities. Strictly speaking then, the human world before the Akan has a discrete niche in consciousness. It

40. A pioneering, authoritative and hugely influential attempt at assimilation is to be found in Rev. J.G. Christaller, A Dictionary of the Asante and Fante Language Called Tshi (Chwee, Twi), Basel, Evangelical Mission Society, 1881, 89. There the Basel missionary defines odomankama as “God, the Creator”, saying that the Twi usage “as a name of God seems to mark him as the bound- less, infinite, interminable, immensely rich Being, or as the author, owner and donor of an inex- haustible abundance of things”. 41. See I. Wilks, “The Forest and the Twis”, Journal des Africanistes, 75, 1, 2005, 19-75. Wilks’s ‘Big Bang’ theory of Akan origins has been severally criticized. Consult, most recently, G. Chouin and C. DeCorse, “Prelude to the Atlantic Trade in Southern Ghana: New Perspectives on Southern Ghana’s pre-Atlantic History (800-1500)”, Journal of African History, 51, 2, 2010, 123- 145; M. Pavanello, “Reconsidering Ivor Wilks’s ‘Big Bang’ Theory of Akan History”, Ghana Stud- ies, 14, 2011, 11-52; G. Chouin, “The ‘Big Bang’ Theory Reconsidered: Framing Early Ghanaian History”, Transactions of the Historical Society of Ghana, new series, 14, 2012, 13-40; and J. Boachie-Ansah, “Excavations at an Earthenwork Site at Asaman and their Implications for the Ar- chaeology of the Forest Areas of Southern Ghana”, in J. Anquandah, B. Kankpeyeng and W. Apoh, eds., Current Perspectives on the Archaeology of Ghana, Accra, Sub-Saharan Publishers, 2014, 18-44. 42. Wiredu, Cultural Universals and Particulars, 120-121. McCaskie, The Parameters of Asante Historical Consciousness 25 cannot be resurrected in historical recall, for it lies beyond the frontier of the Akan past in the alien world of the tetefoͻ and so it can only be accessed via the speculative workings of memory (nkae; and note that, differently inflected and pronounced, this word means “the others”). Asked about some of these matters, Asantehene Opoku Ware II reflected as follows. I never thought about history until I came back to Kumase in the 1960s and took up work as a lawyer. Then I took interest in it. I asked my mother and others who knew things about our past. I learned that Asante custom was from time immemorial, the time when it was made by Domankama and then polished and added to by many, many generations of Asante people. Kings and chiefs added to it or took things away from it as need required. But there remained still things that we could not understand from the past times. And so it was said that Domankama, who made all, made these things before making Asante. So when our people learned of them they saw that Domankama made them but in many cases they could never understand why or for what purpose they were made. We say that some traditions are real but are older than we are and we must try to remember them even if we cannot really understand them. The time of Domankama was there before us. It is a living thing but it is often hard to know the meaning of it43. Part of the puzzled unknowing voiced here surely arises from the formulaic way in which the memory of odomankama has been preserved. Telling odomankama tales is the province of a drum language (ayan) in which a specialist player (kyerɛma) of the atumpan drum is trained up to replicate and to transmit canonical texts exactly. With the passing of time the meaning of elements in these texts became obscure, occluded, and at length lost to living memory44. However, as generally happens with human beings, loss of transparency generally encouraged to inventive speculation rather than to the quietus of resigned bafflement. The tales of odomankama are unhelpful in this matter, because their often archaic language and subtle ambiguities, equivocations, riddles, and asides – features all duly noted and even celebrated by the Asante – are seductively open ended. If, as the drummers always play at the opening of these pieces, “long, long ago Odomankama made a thing” (firi tete odomankama boͻ adeɛ), then the implication was that ‘a thing’ might embrace both anything and everything. For instance, the drum texts say that odomankama created the first executioner (ͻbrafo titire), ancestor by vocation if not by direct descent of all of those generations of func- tionaries who carried out tͻprɛ. Of course, no unthinkingly naïve association is in operation here. Thus, almost forty years ago, Isaac Kwadwo Agyeman, well-born, well-connected, and sometime private secretary to the Asantehene Osei Agyeman Prempeh II, talked to me about the nature of Asante history. “Kings and those

43. Asantehene Opoku Ware II, Kumase, 6 August 1996. 44. Kwame Arhin told me many times that the interpretation of ‘praise’ names (mmeraneɛ) from the remote past was often baffling; see by way of illustration K. Arhin, “The Asante Praise Poems: The Ideology of Patrimonialism”, Paideuma: Mitteilungen zur Kulturkunde, 32, 1986, 163- 197, and for comment McCaskie, “Telling the Tale of Osei Bonsu”.

Africa, N.S., I / 1, 2019 26 a r t i c o l i / a r t i c l e s with power”, he said with pointed emphasis, “can do as they will, but they cannot do as they please”. Pressed to elaborate, he went on to say in an echo of Locke that, “an Asantehene can add to, change and abolish custom, but whatever he does it must be sanctioned by the past. Reasons for anything of this sort must be sought and found in our history. Otherwise power can do nothing, and if it insists it will come to ruin”45. It will be readily understood that the quest after such “reasons” allowed for a deal of ratiocinative latitude. Thirty years ago, I published an assessment of Komfo Anokye, celebrated in oral historical tradition as adviser to the first Asantehene Osei Tutu, and as a seer whose supernatural insights are said to have given decisive shape to Asante custom and law46. At that time I construed him as being, arguably, a real person, but chose to lay stress on his status as a necessary organizing metaphor for bringing the new polity of Asante into ordered being. In light of what has been argued here, I think this interpretation can be underlined and extended. All of the dense oral traditions surrounding Komfo Anokye point to the embedded understanding that he functioned as a conduit, and as an impresario, between the world(s) of the past of odomankama and the tetefoͻ and the foundation of a newly minted Asante. It is most useful, perhaps, to imagine Komfo Anokye as a selective archivist of the past and an interested moulder of the future. However, a careful reading of all of the texts concerning him shows that his role was to propose and not to dispose. The mechanism for the ordering of items like tͻprɛ was that Komfo Anokye suggested, but Osei Tutu – and all of his successors – decided. There is nothing odd in this arrangement. When presented with a listing of authorized historical possibilities, established power, if and when it can, will always select that which supports and reinforces it.

A brief afterword

In one of the contributions cited at the beginning of this paper that is dedi- cated to the search for a comparative anthropology of history making, we find the following summary exposition. Historians and anthropologists are at cross-purposes on the question of truth. For the former, the whole point is to make true statements about the past, and this truth rests on verification in relation to evidence. Anthropologists may also operate in that mo-

45. Isaac Kwadwo Agyeman, Kumase, Adum, Kumase, 1 September 1979. For some context on this man and his life consult T. McCaskie, “Akwankwaa: Owusu Sekyere Agyeman in His Life and Times”, Ghana Studies, 1, 1999, 91-122, reprinted in my Asante, Kingdom of Gold, 505-531. Compare Agyeman’s remarks here in the light of the arguments concerning formal and informal legitimacy in Gyekye, Tradition and Modernity, 171-191. 46. T. McCaskie, “Komfo Anokye of Asante: Meaning, History, and Philosophy in an African Society”, Journal of African History, 27, 2, 1986, 315-339, reprinted in my Asante, Kingdom of Gold, 279-304. McCaskie, The Parameters of Asante Historical Consciousness 27

de, but an important goal is to capture the truth of how a particular people see the world. Historians work with a correspondence theory of truth, whereas anthropologists often work with a coherence theory that establishes the social and cultural contexts in which groups accept statements as credible. Their ethnographic accounts capture how societies understand and represent the past47. This is a partial and limited assertion, as commonplace as it is naïve. Phi- losophers have long known that the coherence theory of truth rests on “an irrational rationalism”, just as the correspondence theory is “vague or vacuous”48. Of course, this is not to say that some historians and anthropologists, and perhaps indeed the majority of both, do not operate with some implicit or largely unexam- ined adherence to versions of the correspondence or coherence theories of truth. That they do so can be economically demonstrated by the consideration, already noted, that many historians of Africa remain wedded to empirically documented verification (itself a nineteenth-century invention) and are suspicious of oral historical tradition, while numbers of their anthropological colleagues continue to think that the historical is simply grist to the mill of a comparative social scientific theorizing and abstraction (arguably, also a nineteenth-century invention). There was a time when some among us thought that the investigation of Af- rican cultures on their own terms might lead on to an intellectual decolonization, or even a termination of the applicability to such societies of the techniques of historical and anthropological inquiry as developed in the west. At one important level this hopeful expectation foundered on the changed, and still changing, conditions of research. The much-shrunken academic world of short-termism, lack of funding, and the demand for demonstrably practical relevance (to what, and for whom?) has impacted the detailed, long-term study of the African past in obviously negative ways. And yet, to endeavour to unpack or to unpick the quiddities of the African past must remain an intellectual obligation to any scholar worthy of the name and serious about the continent. In this venture, as I have tried to show here, long immersion and the understanding that the historical “truth” of any culture is not a veridical fixity or a comparative abstraction but a sequence of provisional thoughts, always subject to destabilization from further vexingly un- anticipated complications, is the beginning of sense (in all the meanings of that term). Let me offer two final reflections here. Seven years ago, Reid wrote an important survey of what he termed “the historical foreshortening” of African history, in part because of the “exaggerated significance” attached to the twentieth

47. Stewart, “Historicity and Anthropology”, 85. 48. See W.V. Quine, Quiddities: An Intermittently Philosophical Dictionary, New York, Vi- king Penguin, 1987, 212-3; a more recent discussion, far too long and complex to be summarized here, is in C. Hubert, “Truth”, in B. Cassin, ed., Dictionary of Untranslatables: A Philosophical Lexicon, Princeton NJ, Princeton University Press, 2014, 1159-79. For an Akan view of the coherence theory of truth, and of Tarski’s modifications of it, consult Wiredu, Cultural Universals and Particulars, 105-112.

Africa, N.S., I / 1, 2019 28 a r t i c o l i / a r t i c l e s century. He was kind about my work on “the culture and indeed philosophy of authority” in Asante, but entered the caveat or disclaimer that the sources for that kingdom were “dense enough to support this kind of reconstruction.” Elsewhere in this paper, however, he suggested that the sources for precolonial African history were not nearly as scant as some chose to believe49. The squaring of this circle relates to what has been argued in the previous paragraph. That is, the jury must remain out on what is possible in the reconstruction of the precolonial Afri- can past. The plain truth is that the history of few societies in sub-Saharan Africa have been subjected to the prolonged intensity and diversity of interrogation that has been applied to Asante, and it is to be feared that this state of affairs will not change any time soon. My second reflection concerns the nature of writing history itself. This is now a much-discussed topic, albeit not so much in African studies. However, a recent intervention by Hunt, a distinguished scholar of the Congo’s imbricated past and present, uses Simmel’s ideas on history and form to suggest possible new pathways into the study of the poetics and aesthetics of the African past50. Here I neither endorse nor refute what she has to say, but content myself with the observation that we need very many more innovative and challenging discussions of the sort she provides about novel ways to write about the African past.

49. R. Reid, “Past and Presentism: The ‘Precolonial’ and the Foreshortening of African His- tory”, Journal of African History, 52, 2, 2011, 135-155, especially 145 and 154. 50. N. Hunt, “History as Form, with Simmel in Tow”, History and Theory, 57, 4, 2018, 126- 144. Elara Bertho

L’e m p i re d e Sa m o r i To u r é : p o u r u n p o i n t d e v u e a f r i c a i n d e l’h i s t o i r e c o l o n i a l e (Ma l l a m Ab u , La b a r i n Sa m o r i , 1914)*

Abstract

Focusing on an unpublished ajami manuscript written in Hausa by Mallam Abu in 1914, this article analyses the story of Samori Touré with an African view. The translation from Hausa to French of this manuscript provides an insight of the arrival of the colonization in West Africa. Based on archival investigations, this article presents the historical and intellectual context of the manuscript. It also gives literary analyses on the formulaic style and repetitions of this original source, which describes Samori Touré both as guilty and heroic. k e y w o r d s : h a u s a , m a l l a m a b u , s a m o r i t o u r é , c o l o n i a l h i s t o r y

Raconter de l’intérieur l’histoire des vaincus de la colonisation1, lorsque l’on vient de vivre la guerre coloniale et que de surcroît on l’a perdue, c’est raconter l’histoire de la fin d’un monde et de l’avènement d’un nouveau où toutes les cartes du pouvoir sont rebattues. C’est la tâche à laquelle s’attèle Mallam Abu à Wa, en Gold Coast, lorsqu’en 1914 il entreprend d’écrire en haoussa, en ajami2, une chronique de l’empire de Samori Touré3 depuis son émergence dans les années 1880 jusqu’à sa chute en 1898.

* CNRS, LAM (UMR 5115). Ce texte est le résultat d’une recherche au long cours. Il a fait l’objet d’une première présentation lors du cycle de conférences de la Société des Africanistes au Musée du Quai Branly Jacques Chirac le 19 mai 2016. Un premier état de la réflexion a été publié sous le titre “Histoire de Samori” de Mallam Abu (c. 1914) : une chronique haoussa face au tournant colonial” (avec Souleymane Ali Yero), in U. Baumgardt, ed., Littératures en langues africaines. Production et diffusion, Paris, Karthala, 2017, p. 143-156. Nous présentons ici une version complè- tement remaniée et enrichie d’analyses de ce document et non plus seulement de présentation d’une traduction. 1. N. Wachtel, La vision des vaincus : les Indiens du Pérou devant la conquête espagnole, 1530-1570, Paris, Gallimard, 1971. 2. Transcription de langue africaine (en l’occurrence le haoussa) à l’aide de l’alphabet arabe. 3. Samori Touré fut empereur du Wassoulou et s’opposa à la colonisation française et britannique pendant près de vingt ans, de 1881 à sa capture en 1898. Il meurt en exil au Gabon en 1900.

Africa, N.S., I / 1, 2019 30 a r t i c o l i / a r t i c l e s

Ce manuscrit, conservé à la SOAS (School of Oriental and African Studies, Londres), a été translittéré en alphabet latin par Stanislaw Pilaszewicz dans une annexe non publiée de sa thèse. Labarin Samori (“L’histoire de Samori”, 182 pages) constitue la première partie du manuscrit : la seconde, consacrée au chef Babatu (Labarin Zabaramawa) et sensiblement plus courte (141 pages), a été traduite et publiée par Stanislaw Pilszewicz4. En collaboration avec Souleymane Ali Yero5, j’ai effectué une traduction en français de cette première partie consa- crée à Samori Touré à partir de la translittération du haoussa en alphabet latin6. Permettant d’opérer une “histoire à parts égales”7, ce texte fournit de pré- cieux renseignements sur les débuts de l’empire de Samori Touré, sur les alliances diplomatiques africaines et sur les perceptions locales de la capture de l’al- mami8. Labarin Samori appartient donc à ces “voix africaines” qui constituent assurément des matériaux précieux pour rendre l’épaisseur et la complexité des guerres coloniales, à lire en regard des sources coloniales déposées dans les fonds d’archives9. Ce qui frappe d’emblée, c’est l’apparente sécheresse de ce texte, son style formulaire, sa cadence éminemment répétitive dans la description des villes capturées par Samori Touré : précisément, c’est ce sentiment de décalage qui est l’objet de cet article. Il signale en effet que ce texte est au croisement de traditions historiographiques variées, pris entre plusieurs mondes en profondes mutations, où sources orales, sources écrites en arabe et exigences coloniales d’informations

Pour une histoire complète de son empire, voir la somme que lui a consacré Y. Person, Samori, une révolution dyula, Paris, IFAN : Centre de recherches africaines, 3 tomes, 1968-1975. 4. S. Pilaszewicz, The Zabarma Conquest of North-west Ghana and Upper Volta. A Hausa Narrative “Histories of Samory and Babatu and others” by Mallam Abu, Varsovie, Polish Scien- tific publishers, 1992 ; voir aussi une étude de S. Pilaszewicz, “On the Veracity of Oral Tradition as a Historical Source : The Case of Samori Ture”, in Unwritten testimonies of the African past : proceedings of the international symposium held in Ojrzanów n. Warsaw on 07-08 November 1989, Wydawnictwa Uniwersytetu Warszawskiego, 1991, 167-180. 5. Souleymane Ali Yero est doctorant de l’Université Abdou Moumouni de Niamey, professeur d’Histoire-géographie au lycée Olinga et au lycée français La Fontaine. 6. E. Bertho, Mémoires postcoloniales et figures de résistants africains dans la littérature et dans les arts. Nehanda, Samori, Sarraounia comme héros culturels, Thèse de doctorat, sous la direction de Xavier Garnier, Paris 3 - Sorbonne Nouvelle, Paris, 2016, 558-607. 7. R. Bertrand, L’Histoire à parts égales : récits d’une rencontre Orient-Occident, (XVIe-XVIIe siècle), Paris, Seuil, 2014. 8. Titre de chef religieux, vraisemblablement dérivé de l’expression arabe “prince des croyants”. 9. La place des “voix africaines” dans la recherche est un débat constitutif du champ des études africaines, voir ainsi J.L. Comaroff and J. Comaroff, Ethnography and the Historical Imagination, Boulder, Colo. ; Oxford, Westview Press, 1992. Pour les écrits en langues africaines sur la colonisation dans la région, voir M. Sani Umar, Islam and Colonialism. Intellectual responses of Muslims of Northern Nigeria to British colonial rule, Leiden ; Boston, Brill, 2006. Sur la nécessité de prendre en considération les sources écrites en arabes ou en langues locales pour opérer une histoire située de la colonisation, voir C. Lefebvre, “Zinder 1906, histoires d’un complot : Penser le moment de l’occupation coloniale”, Annales. Histoire, Sciences Sociales, 72, 4, 2017, 945‑981 ; C. Lefebvre, Frontières de sable, frontières de papier. Histoire de territoires et de frontières, du Jihad de Sokoto à la colonisation française du Niger, XIXe - XXe siècles, Paris, Publications de la Sorbonne, 2015. Bertho, L’empire de Samori Touré 31 cohabitent. Ce manuscrit est le reflet de ces croisements et de ces hybridations, proposant une perspective sur le monde, un regard particulièrement original. Comment se dit la conquête coloniale en haoussa, quasiment au moment- même où elle vient de s’achever? Comment dire la défaite? Quelles interpréta- tions sont données de la capture de Samori Touré et quelles orientations du discours sont ainsi révélées ? Après une analyse des conditions de production des savoirs historiques par cet “informateur colonial”10 peu banal, j’analyserai les caractéristiques de la narration africaine de l’histoire coloniale, avant de présen- ter les raisons endogènes fournies pour expliquer la défaite de Samori Touré face aux Européens.

Le scribe et l’empereur : la place de l’informateur, à la croisée des frontières coloniales

Nous n’avons malheureusement que peu d’informations sur l’auteur, Mal- lam Abu. Il prétend avoir été témoin des événements qu’il raconte, tant auprès de Babatu que de Samori, ce que réfute Stanislaw Pilaszewicz11, arguant du fait que les distances à parcourir auraient été trop longues pour un seul homme. Cet indice fragmentaire de l’identité de l’auteur disséminé dans le texte est donc contre- dit par la critique. Voici la notice fournie par la SOAS, qui présente de bien mai- gres indices : Papers of Frederic William Hugh Migeod. “Histories of Samory and Babatu and other raiders, written in Hausa about 1914 by Mallam Abu, who said he was with them, for Dr. J.F. Colson in the Northern Territories of Gold Coast”. Hausa text in Arabic script, with notes by F.W.H. Migeod in English and transliterated Hausa, c. 1914. (Given by Dr. Corson to F.W.H. Migeod in January, 1926. Ms 98017. Elle nous apprend que ce texte ajami a été une commande d’un officier de l’armée coloniale, Dr. J.F. Corson (à la graphie d’ailleurs mouvante, hésitant avec Colson), dans le Nord de l’actuel Ghana, qui aurait incité l’auteur à rassembler ses souvenirs, éventuellement en les confiant à un scribe, et ce avant la première guerre mondiale. Le manuscrit aurait ensuite changé de main, avant d’être transféré à la SOAS de Londres. Ce que confirme et développe la notice donnée par John Hu- nwick12 : ABÛ, known as Malam Abû, was active late 19th century.

10. Cette large catégorie plastique fait l’objet de nombreux travaux contemporains. Nous re- joignons la définition très large donnée par C. Van den Avenne, De la bouche même des indigènes. Échanges linguistiques en Afrique coloniale, Paris, Vendémiaire, 2017 ; C. Labrune-Badiane, É. Smith, Les Hussards noirs de la colonie : instituteurs africains et petites patries en AOF (1913- 1960), Paris, Karthala, 2018. 11. Pilaszewicz, “On the Veracity of Oral Tradition”, 170. 12. J. Hunwick, Arabic Literature of Africa, Vol. 4, The Writing of Western Africa, Leiden, Boston, Brill, 2003, 565-566.

Africa, N.S., I / 1, 2019 32 a r t i c o l i / a r t i c l e s

Malam Abû belonged to Yeri Nayiri, a Muslim section of Wa comprising “warriors” rather than “scholars”. He was directly involved in the upheavals of the late nineteenth century that resulted from the intrusion of first the Zabarima and then Samori into the Volta basin. In or about 1914 Dr J. F. Colson, Medical Officer in the Wa, encouraged Malam Abû to record his recollection of the period. He did so in the form of six hundred and sixteen tales (labarin) in hausa, which he probably dicted to a scribe. Malam Abû was also one of the principal informants of Ishâq ‘Uthmân Dabila b. Ya’qûb. Qui était Mallam Abu ? Et qui était cet officier britannique au nom incertain pour qui aurait été écrit ce texte, ou qui lui aurait été immédiatement donné ? Ni de J.F. Colson ni de J.F. Corson ne sont mentionnés dans la Colonial Office List13 des National Archives de Londres. Nulle trace non plus dans l’Imperial Kalendar ni dans la liste des Medical officers14. On retrouve pourtant une piste en Gold Coast où J.F. Corson est cité pour avoir aidé à vaincre la mouche tse-tse en juillet 192015. En 1935, il est au Tanganyika, où il est médecin de l’empire britannique, “esquire”, décoré de l’ordre de Saint Michael et Saint George16. Il aurait donné le manuscrit qui nous occupe près de dix ans auparavant, en 1926 à F.W. Migeod, soit en Gold Coast, puisque ce dernier a séjourné régulièrement en Afrique de l’Ouest de 1902 à 192817 et où il a étudié notamment le haoussa18, soit au Tanganyika, où il était man- daté par le British Museum pour effectuer des fouilles de 1924 à 193119. À son re- tour à Londres, Migeod dépose le manuscrit dans les archives de la SOAS, où il est aujourd’hui. Quant à Mallam Abu, l’auteur du texte, son identité reste encore plus mysté- rieuse que celle des deux officiers britanniques qui ont eu ensuite le texte en leur possession. S’il était plus “guerrier” que “lettré” comme l’indique John Hunwick20,

13. National Archives, Londres. Office of commonwealth relations, The Dominions Office and Colonial Office List, Londres, Waterlow & Sons, 1860-1960. 14. A. Peterkin, W. Johnston, W. R. Macfarlane Drew, Commissioned Officers in the Medical Services of the British Army, 1660-1960, Londres, Wellcome Historical Medical Library, 1968. 15. National Archives, Londres. CO323/821/76 Gold Coast, Control of tse-tse fly. Executive recommendation for annual grant. 16. National Archives, Londres. CO448/44/11 Colonial Office Honours List. 17. Né le 9 août 1876, Migeod a eu une carrière plus aisément retraçable que celle de Corson. Sur sa carrière de Colonial Administrator, voir National Archives, Londres, ADM/196/82 : dans la Royal Navy de 1889 à 1898, il devient assistant paymaster jusqu’en 1893, puis transport officer au Nigéria de 1898 à 1899, affecté aux douanes en mars 1900, puis dans les Ashanti Field Forces en 1900, où il reçoit la Ashanti Medal. Voir également des mentions à Aix-en-Provence de son nom, ainsi d’un voyage au Sénégal, attesté en 1909 : ANOM AOF/III/2 “Visite de Migeod en AOF, 1909”. 18. Il est l’auteur de nombreux ouvrages sur les langues africaines : The Mende Language, 1908, The Languages of West Africa, 1911-1913, Mende Natural History Vocabulary, 1913, Gram- mar of the Hausa Language, 1914. Voir aussi ses carnets de notes sur le haoussa, conservés à la SOAS, MS98024 Army Book. 19. Ses carnets de croquis de fossiles et de dinosaures sont disponibles au Natural History Museum de Londres et à la Royal Geographical Society. 20. Hunwick, The Writing of Western Africa, 565-566. Bertho, L’empire de Samori Touré 33 l’identité et la catégorie socio-professionnelle de Mallam Abu n’en sont pas pour autant définitivement éclaircies : était-il éduqué et accompagnait-il à ce titre la cour de Samori? Était-il simple soldat, ayant participé à de nombreuses campagnes et ayant entendu par ouï-dires les récits de celles auxquelles il n’aurait pas participé ? Il est évident en effet que Mallam Abu n’a pas pu être un sofa engagé dans toutes ces expéditions, tant les distances sont longues, et les dates, rapprochées : il n’a pas pu être sur tous les fronts. Comment avoir été à la fois à Sikasso, au Mali, et en pays Kissi en Guinée, par exemple ? Il s’agissait de deux fronts fort éloignés, que Mal- lam Abu relate pourtant à la suite (pages 115 à 120). La probabilité serait plus forte qu’il ait été de l’entourage de la cour, en ayant ainsi accès aux nouvelles de toutes les campagnes à la fois. L’autre hypothèse est qu’il aurait entendu parler de toutes ces campagnes militaires, par un ou plusieurs relais, cette dernière hypothèse ayant la préférence de Stanislaw Pilaszewicz21, qui admet cependant que Mallam Abu a certainement été très proche d’Umaru22. Il y a dans le texte un segment de phrase qui irait dans le sens de la collecte de données auprès d’autres témoins : cette courte proposition au début du page 120, à propos du siège de Sikasso, “Dugutsi a donné l’information” ; ce qui reste néanmoins très elliptique, sur l’identité d’éven- tuels autres informateurs. Déjà Hérodote distinguait dans ses “Histoires” ce qu’il avait vu distinctement et ce qu’il avait appris de relata. Quoi qu’il en soit, deux informations majeures peuvent être retenues des circonstances de production : le texte est le résultat d’une commande23 d’un représen- tant de l’administration coloniale britannique (qu’elle ait été informelle ou non) d’une part, et l’informateur était très bien renseigné d’autre part, ce qui fait de lui l’un des premiers acteurs24 de l’historiographie africaine25 du règne de Samori. Son titre de mallam, “savant, lettré, marabout, maître coranique”, lui confère une position d’autorité qui rend plausible l’hypothèse qu’il ait été un conseiller à

21. Pilaszewicz, “On the Veracity of Oral Tradition”, 170. 22. Mallam Abu nomme ainsi Sarankèn Mori, l’un des fils les plus influents de Samori Touré ; ses haut-faits sont relatés tout au long du dernier quart du texte, aux pages 159-176. 23. Jouant un rôle similaire à celui de passeur, d’“intermédiaire” : S. Dulucq, C. Zytnicki, “Présentation : ‘Informations indigènes’, érudits et lettrés en Afrique (nord et sud du Sahara)”, Outre-mers, 93, 352, 2006, 7‑14 ; B.N. Lawrance, E.L. Osborn, and R.L. Roberts, ed., Intermedia- ries, Interpreters, and Clerks : African Employees in the Making of Colonial Africa, Madison, University of Wisconsin Press, 2006. Sur cette relation à l’administration coloniale, voir J-H. Jézé- quel, “Voices of Their Own? African Participation in the Production of Colonial Knowledge in French West Africa, 1900-1950”, in Ordering Africa, Manchester, Manchester University Press, 2007, 173‑198, même si la situation est ici différente puisque Mallam Abu rédige lui-même son texte en langue haoussa ou le dicte à un scribe. 24. D. Gary-Tounkara and D. Nativel, ed., L’Afrique des savoirs au sud du Sahara, XVIe-XXIe siecle : acteurs, supports, pratiques, Paris, Karthala, 2012. 25. Voir la vaste collaboration dirigée par N. Kouame, C. Coquery-Vidrovitch, E. Meyer, et al. ed., Historiographies d’ailleurs comment ecrit-on l’histoire en dehors du monde occidental ?, Paris, Karthala, 2014. Le Projet d’Encyclopédie des historiographies non-occidentales poursuit ce travail collectif, sous la direction de Didier Nativel et de Catherine Coquery-Vidrovitch pour la partie sur l’Afrique et l’Océan Indien.

Africa, N.S., I / 1, 2019 34 a r t i c o l i / a r t i c l e s la cour, ce qui expliquerait les multiples références aux “demandes de conseil” de Samori à ses gens ou à ses marabouts : dans le texte, ces échanges ritualisés avant chaque combat entre le chef et ses gens sont mis en scène de manière stylisée et formulaire et constituent une étape incontournable avant chaque décision ma- jeure. Les sièges particulièrement longs ou les entreprises particulièrement ris- quées font de surcroît l’objet d’une demande auprès d’un ou de plusieurs marabouts – qui conseillent généralement une série de sacrifices, par lots de cent ou de mille26 – et Samori prend garde de ne jamais aller contre leur volonté. Ce motif du conseil, dont nous supposons qu’il est une trace si ce n’est allégo- rique du moins métonymique du statut de l’auteur, est mis en scène dès l’attaque de Fita, la première ville du manuscrit. Fita est d’abord assiégée par l’imam Bukari qui est tué par le souverain, Abali (fils de Alhaji Muhammadu Zuhi), puis par un es- clave de Samori du nom de Sagak’iki, enfin dans un dernier temps par Samori. Ces trois attaques successives se révèlent vaines. Ce n’est qu’après avoir consulté (ro- ko27) ses marabouts qu’il a pu prendre Fita, la première d’une longue suite de victoires : Samori, il a appelé ses marabouts, il les a consultés sur le fait qu’il allait se rendre à Fita. Les chefs de Samori sont revenus, ils ont juré sur le nom de son père Kufila. Ses marabouts ont dit : “Samori, fournis des aumônes28 en grand nombre”. Samori, Sa- mori a dit : “Moi, que vais-je faire comme aumônes ?”. Ses marabouts lui ont dit : “Pour l’aumône, [donne] cent chevaux, cent brebis, cent chèvres, cent esclaves, cent boubous29, et de l’or, et de l’argent”. Samori l’a fait, il a donné à ses marabouts, ils ont invoqué Dieu30. À ce moment-là. Samori, fils de Kufila, le chef de la guerre ici- bas, Samori le Blanc des Noirs, le chef de la ruse, Samori, le chef de la danse31, Sa- mori le chef des tambours de guerre32 (page 21).

26. La pratique des “mille sacrifices” se retrouve dans d’autres traditions sur les Camara, les ancêtres de Samori. Voir T. Geysbeek, “A Traditional History of the Konyan (15th-16th century) : Vase Camara’s Epic of Musadu”, History in Africa, 21, 1994, épisode 19, “Wiikèlèn Sacrifice”, p. 73. Voir également pour une autre source africaine, en français cette fois, “Histoire locale” de Dji- guiba Camara, l’une des sources d’Yves Person, en cours de publication (Elara Bertho, Marie Ro- det, Jan Jansen ; Leiden, Brill). Dans ce dernier texte également, Samori a fréquemment recours à son entourage ou à ses marabouts avant d’entreprendre une campagne de guerre, mais le texte de Mallam Abu en fait un motif systématique. 27. Roko, en haoussa : “consulter un oracle, faire de la divination, adresser une demande à Dieu, prier”. Nous préférons “aumônes”, puisque Samori donnera des boubous et de l’argent. Sur les traduc- tions du haoussa, nous nous référons du dictionnaire de référence, P. Newman, R.M. Newman, A Hausa-English Dictionary, New Haven-London, Yale University Press, 2007. Pour une plus grande profondeur diachronique, nous nous référons ponctuellement à celui de G.P. Bargery, Hausa-English and English-Hausa Dictionary, version en ligne : http://maguzawa.dyndns.ws/ [1934]. 28. Saddaka, “aumônes, sacrifices”. 29. Riga, “robe ample, boubou”. 30. Roko, “adresser une demande à Dieu, prier, invoquer”. 31. Samori sariki rawa, “le chef de la danse”. Il s’agit vraisemblablement des danses exécu- tées avant les combats pour stimuler l’ardeur des guerriers. Voir l’apposition suivante. 32. Samori sariki kidi kidi. Kidi : “les tambours de guerre”. Bertho, L’empire de Samori Touré 35

Après trois échecs, ce n’est que lorsque la communauté est toute entière mo- bilisée (que ce soient “ses gens/ses sujets”, mutaninsa, ou “ses marabouts”, mal- lamaisa), et que les demandes en sacrifices ont été remplies, que la victoire peut être enfin remportée. Il a “mangé”33 la ville, il l’a conquise. De la constitution de l’empire de Samori aux conquêtes orientales menées par son fils Umaru, en passant par les rezzous menées par Babatu, Mallam Abu retrace dans ce manuscrit l’histoire d’un territoire qui s’étend de l’actuelle Gui- née au Ghana, en intégrant le Mali, la Côte d’Ivoire, et le Burkina Faso contemporains. Il serait hasardeux de statuer définitivement sur les fonctions qu’il a oc- cupées auprès de ces trois chefs, ni même sur sa présence effective ou non à chacun des sièges relatés, puisque les conjectures échafaudées seraient trop fragi- les. Que Mallam Abu ait été scribe au service de Samori comme il le prétend, ou simplement un relai secondaire de récits entendus à Wa, ce récit constitue un document exceptionnel pour l’étude de la constitution de l’empire de Samori, vue depuis l’intérieur.

Écrire l’histoire militaire africaine : la perspective d’un historiographe de Wa

La plus grande partie du texte est centrée sur les conquêtes de Samori, avant sa rencontre avec les Français, et adopte tout au long du récit un style qui peut paraître à première vue tout à fait déroutant : il n’est formé, en effet, que d’une seule et longue liste de villes, assiégées, attaquées et vaincues par Samori, jusqu’aux dernières pages, où le procédé s’inverse, et où c’est cette fois Samori qui fuit, de ville en ville, tandis que ce sont les Français qui, à leur tour, assiègent, attaquent et remportent ces villes. Le procédé n’est pas seulement répétitif : il constitue l’intégralité du texte34. Labarin Samori ne se structure que dans et par la liste de conquêtes militaires, énumérées à l’aune des centres urbains, qui bascu- lent sous une allégeance ou l’autre. Signalons qu’un texte semblable en haoussa présentant l’arrivée des Euro- péens est le Labarin Nassaru (“L’histoire des Chrétiens”) de Alhaji ‘Umar (1858-

33. Ci, en haoussa, signifie “manger” au sens littéral, et notamment “vaincre un ennemi, prendre possession de, remporter une victoire, conquérir un territoire” au sens figuré. Ci est particuliè- rement polysémique et les usages métaphoriques en sont très nombreux en haoussa. En contexte, Ya ci Fita, se traduit par : “Il a conquis Fita”. Pour une étude détaillée sur le verbe ci, voir P.J. Jaggar, M. Buba, “Metaphorical Extensions of “eat”[overcome] and “drink”[undergo] in Hausa”, in The Linguistics of Eating and Drinking, 84, Philadelphia, John Benjamins Publishers, 2009, 229‑251. C. Gouffé, ““Manger” et “boire” en haoussa”, Revue de l’École Nationale des Langues Orientales, 3, 1966, 77‑111. 34. Dans une moindre mesure, on trouve ce procédé répétitif dans des manuscrits de la biblio- thèque de Tombouctou : A. Saguer, A. Sinno, G. Bohas, Le roman d’Alexandre a Tombouctou. Histoire du Bicornu, le manuscrit interrompu, Arles; Lyon, Actes Sud ; ENS Éditions, 2012, p. 61 et suivantes (Labour, le roi des Indiens, le Maghreb, l’Andalousie, jusqu’au terme du monde).

Africa, N.S., I / 1, 2019 36 a r t i c o l i / a r t i c l e s

1934), probablement rédigé en 190635. Labarin Nassaru offre un tableau saisis- sant de la politique agressive des Européens dans la région ainsi que celui de l’incapacité des chefs locaux à s’unir pour affronter la menace. Samori Touré y est également un personnage important, présenté dès l’initiale du récit, mais il est accusé de n’avoir apporté que des divisions parmi les forces politiques africai- nes36. La dernière partie du Labarin Nassaru constitue un éloge de la colonisation, de la paix retrouvée et de la liberté de circuler ; ceci alors même que la pre- mière partie du texte déplorait la cruauté des Chrétiens dans leurs guerres de conquête. Pour Mallam Abu, la pénétration des Blancs en Afrique n’est pas une avan- cée vers l’intérieur qui aurait mené de manière téléologique à l’occupation de tout le continent, comme l’est le modèle de l’historiographie occidentale au début du XXe siècle glorifiant les “héros de l’empire”37 : c’est avant tout une invasion. Et cette invasion est traitée comme toutes les autres guerres que Mallam Abu a connu et décrit jusqu’à présent : comme un combat de chef à chef, entre Samori et Ar- chinard (“Arshanari”), ce dernier devenant dans le texte l’incarnation quasi-allé- gorique des “Européens”38. Les rencontres entre Samori et les Européens – les Français et les Britanniques – sont d’ailleurs les épisodes décrits avec le plus d’originalité. Très peu de temps après la première rencontre inaugurale, Samori doit fuir devant Archinard, qui est le seul personnage nommé parmi les Français, et qui est présenté dans les derniers paragraphes du texte comme le seul homme ayant réussi à le faire plier. Arrivé au Soudan en 1888 comme proconsul, il réussit

35. S. Pilaszewicz, “The Arrival of the Christians” : A Hausa Poem on the colonial Conquest of West Africa by Al-Haji ’Umaru”, Africana Bulletin, 22, 1975, 55‑129, présente le texte dans son intégralité en haoussa et en traduction anglaise. Voir aussi M. Al-Munir Gibrill, A Structural-func- tionnal Analysis of the Poetics of Arabic Qasidah. An Ethnolinguistic Study of three Qasidahs on Colonial Conquest of Africa by Al-Hajj ’Umar B. ’Uthman Krachi (1858-1934), Indiana Universi- ty, Department of Near Eastern Languages and Cultures, 2015, la section p. 322-338. Pour une presentation détaillée du Labarin Nassaru, voir aussi H. Weiss, Between Accomodation and Reviv- alism : Muslims, the State and Society in Ghana from the Precolonial to the Postcolonial Era, Studia Orientalia, Helsinki, Finnish Oriental Society, 2008, notamment p. 156, 162-165. 36. S. Pilaszewicz, “The Arrival of the Christians”, 65 (correspondant au folio 1 recto, verso). 37. E. Berenson, Les Héros de l’Empire : Brazza, Marchand, Lyautey, Gordon et Stanley a la conquête de l’Afrique, trad. Marie Boudewyn, Paris, Perrin, 2012 ; sur la construction de l’objet conceptuel “Afrique” comme objet de conquête à “civiliser”, voir W.B. Cohen, Français et Afri- cains les Noirs dans le regard des Blancs : 1530-1880, trad. C. Garnier, Paris, Gallimard, 1981. Pour un exemple particulièrement révélateur du paradigme de la “découverte”, voir L.G. Binger, Du Niger au Golfe de Guinée par le pays de Kong et le Mossi, Paris, Hachette, 1892 [Réédition Paris, Société des Africanistes, 1980]. 38. Louis Archinard a une relative postérité littéraire, comme incarnation de la colonisation : voir aussi bien plus tardivement B. Zaourou Zadi, Les sofas; suivi de L’œil, Paris, Pierre Jean Oswald, 1975. Pour une biographie et une description du type de l’ “officier soudanais”, voir M. Cuttier, Portrait du colonialisme triomphant : Louis Archinard, 1850-1932, Panazol, Lavauzelle, 2006. Bertho, L’empire de Samori Touré 37

à rallier à sa cause le parti colonial. La déroute de l’almami décrite ici date certainement des premiers mois de 189339 : Voilà une autre histoire de Samori Il s’est levé pour aller à Balato. Le nom de la ville est Balato. Le nom du chef de la ville est Bubu. Samori s’est levé pour partir, il s’est rendu non loin de Balato, il s’y est installé. L’Européen40 de France aussi, il a cherché Samori. Le nom de l’Euro- péen de France est Arshanari, Arshanari. C’était lorsque Samori est parti à Balato. Le nom de la ville est Balato. Le chef de Balato Bubu a accueilli Samori [par les armes]. Ils se sont rencontrés. Chien en sang, singe en sang41. Samori a tué Bubu le chef de Balato. Il est resté à Balato. Samori, il a régné, il promet la guerre. Le chef d’ici-bas, Samori. Cette histoire s’est arrêtée encore. Voilà une autre histoire de Samori Il a séjourné dans Balato. L’Européen de France Arshanari, Arshanari a entendu la nouvelle, selon laquelle Samori était à Balato. Arshanari s’est levé pour aller chez Samori, à Balato. Arshanari a chassé Samori. Samori a fui, il a fui. Arshanari l’a poursuivi dans son propre pays. Samori a fui, il a fui, Arshanari l’a suivi jusque dans la région de Nafaki. La région de Nafaki est de l’autre côté de la rivière. Arshanari l’a suivi, Samori a fui, il a fui. Arshanari, l’Européen de la France, c’est un grand stratège. Cette histoire s’est arrêtée encore. Samori. Cette histoire s’est arrêtée encore. Samori (pages 145-146). Archinard est ici un personnage fondamental puisqu’il met un terme à l’ex- pansion vers le Nord de Samori, et qu’il le pousse à déplacer son empire, en fuyant devant l’avancée des Français (“il a fui, il a fui” : ya gudu, ya gudu). Ce sera le début des conquêtes vers l’Est, menées en grande partie par Umaru (Sa- rankèn Mori), dans la version de Mallam Abu et qui correspond aux conquêtes de Korhogo, Kong, Dabakala, Bouna, entre 1895 et 1898. Aux marges orientales du second empire de Samori, Bouna était précisément l’objectif d’une colonne anglaise, menée par Henderson42, qu’Umaru met en déroute :

Voilà une autre histoire d’Umaru fils de Samori Il est resté à Guna [Bouna]. Il a entendu les nouvelles des gens de Guna d’une autre ville43. Le nom de la ville est Dorikito. Lui et l’Europen de l’Angleterre. Le nom de

39. Y. Person, Samori, une révolution dyula, tome 3, 1423 et suivantes. Après cet épisode, Samori choisira d’éviter les troupes françaises, en se repliant vers l’Est. 40. Bature, “Européen”. 41. Kare jini, biri jini : “chien en sang, singe en sang”. Expression proverbiale montrant deux animaux traditionnellement décrits comme ennemis, dans les contes, deux adversaires à bout de force, tant la bataille a été rude. 42. Sur la colonne Henderson et la victoire de Sarankèn Mori, voir Person, Samori, une révo- lution dyula, tome 3, 1818; I. Wilks, Wa and the Wala : Islam and Polity in Northwestern Ghana, Cambridge, Cambridge University Press, 2002, 128-132. 43. Forme fautive, il manque les prépositions. Le manuscrit comporte de nombreuses fautes de syntaxe et de pluriels, ce qui laisse à penser que le locuteur n’était pas haoussaphone natif.

Africa, N.S., I / 1, 2019 38 a r t i c o l i / a r t i c l e s

l’Européen est Mazzabi. D’autres l’appellent Mazzafi44. C’est cela le nom de l’Euro- péen. Mazabi a promis la protection aux gens de Guna de manière mensongère. Mais lui, de fait, Mazabi n’a pas tenu sa promesse. Alors les gens de Guna sont restés à l’intérieur de Dorikito. Le nom de la ville est Dorikito. Ils sont restés avec l’Euro- péen attendre le fils de Samori, Umaru. Cette histoire s’est arrêtée encore. Voilà une autre histoire d’Umaru fils de Samori Un jour Umaru fils de Samori est parti pour aller voir les gens de Guna. L’Européen Mazabi a promis qu’Umaru fils de Samori, lorsqu’il viendrait, il ne passerait pas la nuit dans cette ville. Eh, Mazabi, il a menti ! Il a vu Umaru fils de Samori, et il a eu très peur. Cette histoire s’est arrêtée encore. Voilà une autre histoire d’Umaru fils de Samori Il a pris conseil pour la guerre. Il s’est levé, lui et l’Européen Mazabi et les gens de Guna, ils se sont rencontrés pendant trois jours. Mazabi a eu peur, lui ainsi que les gens de Guna. Umaru fils de Samori les a chassés. Ils ont fui, ils ont fui jusqu’à Wa. Ils ont fui fui. Mazabi est arrivé à Wa. La ville de Wala. La ville de Wala, le chef Wala, lui et ses gens ont fui aussi. C’est à ce moment-là que Mazabi a fui. Les Euro- péenns anglais, et eux, et Babatu ils étaient dans une autre ville. Le nom de la ville est Yagu. C’est ce qu’Umaru fils de Samori a fait, chef d’ici-bas. Cette histoire s’est arrêtée encore (pages 174-175). Contrairement à ce que le texte de Mallam Abu pouvait laisser supposer par son apparente stylisation ou par l’absence de dates, cette source haoussa se ré- vèle être finalement fiable pour ce qui est de la chronologie, et supporte tout à fait la comparaison systématique avec les descriptions historiographiques existan- tes45. Mais ce qui fait l’originalité de ce document à notre sens – mise à part l’ex- trême précocité du témoignage écrit en langue africaine, rédigé dès 1914 –, c’est la coexistence d’épisodes qui deviendront par la suite canoniques (comme le siè- ge de Sikasso en 1888 contre Tièba du Kènèdougou), avec des épisodes moins connus (comme l’émergence de l’empire sur sa marge occidentale dès les années 1870, ou comme les relations diplomatiques nouées aux franges orientales avec Babatu46 entre 1890 et 1896), donnant ainsi un tableau complet des expansions samoriennes. Pourtant, malgré ses exploits militaires dont l’ampleur est souli- gnée par la narration, Mallam Abu nous présente le personnage de Samori sous un angle étonnant pour un chroniqueur, puisque le héros est autant dénigré pour sa traîtrise qu’il est loué pour ses prouesses. C’est que les raisons d’écrire l’his-

44. “Celui qui donne chaud”, littéralement. À noter la proximité avec mazazabi, “qui a le pa- ludisme”. 45. Ce que M. Delafosse, Essai de manuel pratique de la langue mandé ou mandingue : étude grammaticale du dialecte dyoula, vocabulaire français-dyoula, Histoire de Samori en Mandé, étu- de comparée des principaux dialectes mandé, Paris, E. Leroux, 1901, 145-146 (dès 1901, soit treize ans avant la rédaction de Labarin Samori) cherchait déjà à prouver en recueillant des traditions orales, avec une méthodologie parfois sujette à caution. 46. Voir sur ces relations J. Rouch, “Les cavaliers aux vautours. Les conquêtes zerma dans le Gurunsi (1856-1900)”, Journal des Africanistes, 60, 2, 1990, 5-36, section “Pacte avec Samory (1890-1896)”, p. 21 et suivantes. Sur Babatu, voir les analyses de S. Bornand, Le discours du griot généalogiste chez les Zarma du Niger, Paris, Karthala, 2005, 277-308. Bertho, L’empire de Samori Touré 39 toire de Samori sont nombreuses et s’entrecroisent en mêlant des influences litté- raires diverses.

Samori, un héros paradoxal. L’ère du soupçon dans la chronique historique

“L’histoire de Samori” est donc la première partie d’une commande, effec- tuée par un médecin de l’armée britannique47, à un ancien soldat, et cela a son importance dans les cadres de références convoqués pour l’écriture de l’histoire. Il s’agissait de raconter ses souvenirs afin de témoigner de l’histoire de la région au XIXe siècle. Si le texte ne correspond absolument pas aux normes du récit auquel nous aurions pu nous attendre dans une telle circonstance de production – aucune trace de la première personne du singulier, aucune marque d’implication subjective apparente48, très peu de commentaires métatextuels ou de remar- ques du narrateur en incises –, il s’apparente plutôt au genre des mémoires, où un témoin particulièrement bien placé et informé se fait le peintre de son temps, en relatant les événements majeurs auxquels il a assisté, sans néanmoins livrer d’information sur la nature de cette “place”. Et l’on constate en effet que Mallam Abu ne se consacre qu’aux “grands hommes” et à leurs haut-faits, sans témoigner de sa propre position. Seuls les chefs et les meneurs d’hommes l’intéressent : pour lui, raconter l’histoire, c’est d’abord et surtout faire le compte des victoires. Pour être juste dans le récit du temps passé, il faut peindre le pouvoir aux mains des hommes illustres. Il y a là un impératif éthique dans le récit historique, qui consiste à respecter avec minutie la place à accorder à chacun en fonction de son mérite. Et le mérite se jauge à l’aune des captures, souvent bien plus qu’à celle de la piété49. Il est donc tout naturel que le texte se présente comme un récit de Babatu et de Samori, les deux grands chefs au service desquels Mallam Abu se serait placé, et comme une liste de leurs exploits. “L’histoire de Samori”, et ses quelques deux cents labaru (pluriel de labari, “histoire”, qui correspond à chaque

47. Il ne s’agit pourtant pas du même statut que les “informateurs” qu’a étudiés C. Van den Avenne, “De la bouche même des indigènes”. Le statut de l’informateur dans les premières descriptions de langes africaines à l’époque coloniale”, Linguistiques et colonialismes, vol. 20, Glottopol. Revue de sociolinguistique en ligne, 2012, disponible sur , (consulté le 2 octobre 2018), puisque Mallam Abu est l’auteur de son texte, qu’il en choisit la langue et la graphie, sans être vectorisé uniquement par l’attente d’une réception coloniale. 48. Comme dans d’autres cas de commande coloniale, comme par exemple chez Amadou Kouroubari en 1901, dans M. Delafosse, Essai de manuel pratique de la langue mandé ou mandingue, ou Kélétigui Berté dans R. Colin, Kènèdougou au crépuscule de l’Afrique coloniale, Paris, Dakar, Présence africaine, 2004, 357-383. 49. Selon une conception ordalique de la victoire, où le vainqueur a toujours-déjà raison. Voir sur cette conception de la guerre comme ordalie F. Viti, Guerra e violenza in Africa Occidentale, Milano, FrancoAngeli, 2004. Cela rejoint également les analyses sur le “prestige du “c’est arrivé”“ de Roland Barthes commentées par M. de Certeau, L’écriture de l’Histoire, Bibliothèque des histoires, Paris, Gallimard, 1975 [2011], 67-68.

Africa, N.S., I / 1, 2019 40 a r t i c o l i / a r t i c l e s paragraphe court du récit), s’apparente aussi par bien des aspects aux genres connexes de la chronique historique orale ou écrite, du tariq écrit50, de l’épopée, du merveilleux, des récits de règnes, de la généalogie voire parfois de la louange (kirari) ou du commentaire moral. Il nous semble que c’est cette conception particulière de l’écriture de l’histoire qui éclaire le texte de Mallam Abu : le mémorialiste chargé de se faire l’historiographe de son temps se trouve à la croisée des influences littéraires (entre les genres, entres les traditions historiques, entre les pratiques scripturaires et les pratiques orales de l’histoire51). Ce faisant, il déroule avec précision plus de trente ans de règne en les déployant dans un vaste tableau guerrier fait de captures et de morts, tout en présentant une ambivalence fondamentale dans la conception du “grand homme”, à la fois fascinant car toujours victorieux, et néanmoins hautement contestable. La langue choisie a été le haoussa, qui est, au Ghana et à Wa, une grande langue véhiculaire “de prestige” pour l’écriture littéraire au début du XXe siècle52. Le haoussa est également la langue privilégiée par Lord Lugard dans l’administration et dans l’armée, comme le souligne John Edward Philips53. Dans le texte de Mallam Abu, la syntaxe en est souvent très simple, les phrases sont courtes, ce qui correspond à une langue de l’écrit noble et mesurée, exhibant son propre souci de concision, de clarté et de précision. Cette langue qui peut paraître aujourd’hui relativement sobre voir sèche54 était en réalité perçue comme em- preinte d’une retenue, gage de justesse, qui se distinguait en cela de l’oralité quotidienne. Hiskett fournit des exemples similaires de ce style haoussa de la

50. I. Wilks, “The Growth of Islamic Learning in Ghana”, Journal of the Historical Society of Nigeria, 2, 4, 1963, 409‑417, atteste de l’existence de tariq en langue haoussa dans le Nord du Ghana dans les années 1960, et analyse la tradition islamique d’écriture de l’histoire dans les centres urbains. Sur les pratiques scripturaires du Nord-Ghana, voir J. Goody, “Restricted Literacy in Northern Ghana”, in J. Goody, ed., Literacy in Traditional Societies, Cambridge, Cambridge Uni- versity Press, 1968, 199‑264. 51. Précisément, cette hybridité des références convoquées remet en cause l’idée d’un grand partage tel que développé par J. Goody, La raison graphique. La domestication de la pensée sau- vage, Paris, Minuit, 1978 : le kirari, résolument oral, est intégré dans la chronique écrite dans le texte écrit, “métamorphosant” l’oral, pourrait-on dire en reprenant G. Ciarcia and E. Jolly, Méta- morphoses de l’oralité́ entre écrit et image, Paris, Karthala, 2015, 13, en empruntant à W.J. Ong, Orality and Literacy : the Technologizing of the Word, Londres ; New York, Methuen, 1982 [2002], le concept d’ “oralité seconde”, se “recomposant à partir de l’écriture” (Paul Zumthor, Introduction a la poésie orale, Paris, Le Seuil, 1983, 36). 52. Sur les relations de prestige entre langue haoussa et arabe, G. Furniss, Poetry, Prose, and Popular Culture in Hausa, Washington, Smithsonian Institution Press, 1996, 192. 53. J.E. Philips, “Hausa in the Twentieth Century : An Overview”, Sudanic Africa, 15, 2004, 55‑84, voir notamment p. 59 sur le rôle de Lord Lugard dans l’imposition du haoussa dans l’administration et dans l’armée britannique au Nigéria. 54. Sur les difficultés de lecture, et de la critique, face à ce style formulaire, voir les analyses de B. Salvaing, “À propos d’un projet en cours d’édition de manuscrits arabes de Tombouctou et d’ailleurs” [en ligne], Afriques. Débats, méthodes et terrains d’histoire (2015), disponible sur , (consulté le 10 mai 2018). Bertho, L’empire de Samori Touré 41 chronique dérivé du modèle arabe classique, depuis le Tarikh al-khulafa (“His- toire des califes”) de Jalal al-Din al-Suyuti au XVe siècle, jusqu’à des exemples contemporains du Labarin Samori, comme les “Chroniques de Sokoto” d’Abu- bakar d’an Atiku55. Le scribe, quel qu’en ait été le statut, a porté une grande attention à souligner l’honorabilité de son récit, et de son narrateur, par voie de consé- quence, grâce à des marques ostensibles, voire ostentatoires, de scripturalité. Mais le texte emprunte également à d’autres filiations : ainsi, l’irruption du merveilleux dans le quotidien, qui engendre un pacte avec les puissances magiques et une prédiction de la destinée du héros – dont le récit n’est en fait qu’une longue mise en scène de son accomplissement –, constitue un schéma topique que l’on retrouve dans les contes ou les épopées. Ici, c’est le serpent qui apparaît au milieu du chemin et en jouant ce rôle prophétique, son discours fonctionne comme une prolepse pour le lecteur. Il est d’autant plus important qu’il donne à Samori son nom. Ce baptême initial fait entrer Samori dans l’univers de la fiction et rend possible la suite des péripéties : Samori et son ami faisaient du commerce56. Lui, Samori, il obtint des richesses57. Il distribua aux jeunes la richesse. Les jeunes jouirent de cette richesse, ils l’ont suivi, la richesse s’est tarie sans qu’ils s’en rendent compte58. Samori se remet à faire le commerce. Samori, il gagnait encore, il retournait chez lui, lui et son ami. Il distri- buait encore aux jeunes. Ils sont contents, lui et lui et son ami encore, Samori. Jusqu’à59 ce qu’un jour Samori, lui et son ami, [alors qu’] ils faisaient du commerce, ils étaient en route. Ayant entendu un appel derrière eux et s’étant arrêtés, alors ils se retournèrent et ils regardèrent derrière. Quand ils se retournèrent, ils virent un serpent. Le nom du serpent en Soninké60 serait Sai, et avec Mori, cela donnerait “maî- tre61 serpent”. Voici l’origine du nom de Samori – maître, maître serpent. Cette histoire s’arrête ici62. Et puis voilà encore une autre histoire de Samori

55. Voir de mêmes procédés stylistiques dans M. Hiskett, A History of Hausa Islamic Verse, Londres, University of London School of Oriental and African Studies, 1975, 136-137, 161-162 pour la chronique historique. 56. Futuci : fatauci. 57. Iya samu(n) : ya samu. “il est en train de s’enrichir”. Variante de l’Ouest du pays haoussa (Dogondoutchi). 58. Samari su yi raha suna bisa. Tournure étonnante *suna bisa n’est pas attesté. 59. Hali : note har, “jusqu’à ce que”. 60. La langue des Soninkés, marchands descendus au Sud-Est. Le nom mythique de Katsina est Wangara. Sur le Wangara, voir I. Wilks, “Wangara”, Encyclopedia of Islam, vol. IX, 2ème Éd., 2002, 137-38; ainsi que son ouvrage précédemment cité. Voir aussi A.W. Massing, “The Wangara, an Old Soninke Diaspora in West Africa?”, Cahiers d’études africaines, 158, 2000, 282-308. Le serpent est une divinité fondatrice chez les Wangara. Plus précisément dans les récits sur Samori, un serpent apparaît souvent en rêve au père de Samori, ce qui annoncera l’illustre destin de l’enfant qui venait de naître. 61. Mallam, “maître coranique, marabout, professeur”. 62. Wana (“cette”) est au féminin, tandis que labari (“histoire”) est masculin en haoussa. Cette faute est reconduite dans tout le manuscrit.

Africa, N.S., I / 1, 2019 42 a r t i c o l i / a r t i c l e s

Quand le serpent l’a appelé, il lui a tenu ce discours. Le serpent a dit : “Toi, tu sais, je63 te dis avec certitude, je ne suis pas un serpent. Moi, je suis un génie, mais je me suis transformé64 en serpent. Je conclus un pacte avec toi – tu comprends ?”. Samori dit qu’il a entendu. “Samori, toi et ton ami. Samori, tu auras la royauté dans la vie65. Moi, je suis ton génie. Je conclus un pacte avec toi. Prends la royauté et gouverne le monde. Samori, moi le génie, je te donne un nom aujourd’hui. Ton nom est Samori. Ton ami aussi, qu’il prenne la royauté dans la vie, et le pouvoir66 dans la vie – la vie. Quant à lui, son nom est Sagak’iki. Moi le serpent, toi Sagak’iki, toi aussi, gouverne le monde. Moi le serpent, je dis, moi, je suis un génie. Moi le serpent, je te fais ce discours, toi et ton ami, vous aurez le monde à votre portée. Avez-vous compris?”. Samori dit qu’il avait bien compris. Le génie-serpent leur avait dit de dormir. Ils se sont endormis. Le serpent. L’homme n’est pas patient. Cette histoire s’arrête encore (pages 2-3). Le génie reviendra une seconde fois dans le cours du récit sous la forme d’un jeune homme élancé (page 52), pour lui rappeler ses engagements. Mais comme dans les contes, le héros rompt le pacte qui l’unissait à son bienfaiteur, et cela entraîne sa chute. La transgression dans le texte de Mallam Abu est à la fois constituée par la mort de Sagak’iki, avec qui il avait débuté sa carrière de dyula (colporteur), et par les nombreux massacres de musulmans, qui font l’objet de la désapprobation du narrateur. Chez Mallam Abu, l’écriture induit une codification et une stylisation qui impose tout un univers de références. Pourtant ces codes côtoient également l’épique, le conte, le merveilleux, la louange; l’oral vient alors compliquer le texte d’autres motifs appartenant à d’autres univers de références parallèles. Or l’écriture de l’histoire de Samori, c’est avant tout une histoire de guerres, et ce n’est d’ailleurs que cela chez Mallam Abu. On ne trouvera en effet pas de description qui ne soit immédiatement utile au déroulement des prises de villes. Pas de superflu donc : le texte est entièrement constitué de préparatifs de guerre, d’assauts et de victoires. La ville conquise ne vaut que parce qu’elle est l’amorce et le point d’appui vers une nouvelle conquête. Samori ne cesse de “se lever” (tashi) et de “voyager” (tafi) en vue d’une nouvelle prise. Le héros est incessam- ment en marche, cheminant de conquête en conquête, toujours en mouvement. Cela correspond au quotidien des campagnes militaires de Samori, qui n’a que très peu résidé dans ses capitales que sont Bissandougou, Beyla, Sanankoro, Hé- rémakono... Cela explique également que les toponymes que nous avons pu resi- tuer sont assez souvent des villes frontières, des villes des marges, ou même des villes razziées en dehors de l’empire.

63. Ya ce, “il dit”, dans le texte. 64. Na zzama : la forme est redupliquée : “je me suis transformé”. 65. Dunyia, “la vie, le bas-monde, le monde terrestre”. Fréquemment dans le texte, ba dunyia, “donner la vie” signifie “donner le pouvoir de régner sur le monde terrestre”. À noter : l’usage du prédictif : “tu auras la royauté”. 66. Malaka : “le pouvoir coercitif, le pouvoir dictatorial”. Bertho, L’empire de Samori Touré 43

Mais comment se dit la guerre, au quotidien ? Il y a peu de descriptions de l’armée de Samori en elle-même, si ce n’est quelques considérations annexes qui jettent un éclairage “par les marges du texte” du fonctionnement des régiments. Clé de l’organisation militaire, ses soldats étaient bien armés et la mention des fusils constitue la seule description de l’armement par l’auteur : Là où Samori s’est installé ; il avait quinze mille fusils, tous européens. Ce sont tous les fusils dont Samori disposait (page 116). Samori était le Sariki mi bbaki wuta, “le chef des bouches de feu” (page 68), précisément parce qu’il importait ses fusils, qu’il les faisait venir des côtes et qu’il jalonnait les routes commerciales qui les lui fournissaient. Ils étaient “euro- péens”, ce qui montre qu’ils étaient à répétition comme le Gras et le Kropatchek (voire des Lebel67), et non pas seulement à un coup68. Cette modernisation tech- nologique de son armée lui permit de prendre le dessus sur ses ennemis, lorsqu’ils n’étaient pas retranchés dans un tata imprenable comme celui de Sikasso. La tactique était la même à chaque fois et le récit calque ce modèle répétitif : Samori envoyait des émissaires au chef ennemi, avant de lancer ses troupes à l’assaut. Il stationnait en bordure de la ville, comme pour un siège, mais il n’y avait souvent pas d’encerclement véritable, ce que confirme Yves Person pour le siège de Sikasso en 1888 : pendant près d’un an, Samori a campé sous les murs du tata, sans jamais réussir à encercler l’ensemble de la ville, si bien que Sikasso était souvent mieux ravitaillée en nourriture et en eau que les troupes de Samori elles-mêmes, qui dépérissaient et souffraient de l’insalubrité des campements. Quelle qu’en soit la durée dans le texte (sept mois, sept jours, plusieurs nuits), le siège vise surtout à se poster face à l’ennemi pour faire se rencontrer les deux armées. L’affrontement en lui-même est d’ailleurs peu décrit par Mallam Abu; ainsi de la bataille de Kumu : Samori a appelé ses gens, il leur a dit : “Demain69, quand le jour se lèvera, moi et les gens de Kumu, nous nous verrons”. Samori a vaincu les gens de Kumu en un seul jour (page 178). Ce procédé d’ellipse se retrouve également dans nombre d’épopées, qui se contentent souvent de mentionner que le choc fut terrible et que le montant des morts fut important, en donnant un chiffre d’ailleurs symbolique. De même chez Mallam Abu, il s’agit souvent de tournures euphémistiques (“ils se sont vus, ils se sont rencontrés”), de formules générales (“La rencontre des hommes, c’est le massacre des hommes”, page 17, 18), voire de proverbes courts tel que le très récurrent, kare jini, biri jini, “chien en sang, singe en sang”, dont la répétition à travers tout le document suffit à rendre compte de l’atrocité de la bataille. Le décompte des victi-

67. Person, Samori, une révolution dyula, tome 3, 909 et 993. 68. Pour un témoignage sur la question délicate de l’armement de Samori, A. Nebout, “Vingt et un jours chez Samori”, Journal des voyages, Paris, sect. 150 ; 151 ; 152, 1899, 306‑308 ; 326‑327 ; 343‑345. 69. Kobi : gobe “demain”.

Africa, N.S., I / 1, 2019 44 a r t i c o l i / a r t i c l e s mes sert ensuite, après coup, à souligner l’ampleur du conflit, ou l’impiété de Sa- mori selon les cas : Samori a égorgé les habitants de Bailo, des milliers, jusqu’à quinze mille. C’est ainsi que Samori a massacré les habitants de Bailo (page 46). Quand il a tué les gens de Kumu, il a égorgé les musulmans de Kumu par milliers et par milliers, jusqu’à soixante-dix mille. C’est ce qu’il a fait (page 178). Ces scènes de massacre se retrouvent à de nombreuses reprises, où la cruau- té70 est révélée et par le nombre de victimes, et par les circonstances du massacre, le pillage des villes saintes étant régulièrement accomplis, dans l’espace du texte, un vendredi : l’hérésie accroît alors le scandale de la mort. Entre cruauté et grandeur, il y a un vacillement axiologique manifeste dans la narration. Samori est le plus grand des conquérants de la région, il a su soumet- tre ville après ville des régions entières, mais son orgueil est démesuré et en cela il a été châtié : Samori est donc chez Mallam Abu un héros paradoxal, où l’ “ère du soupçon”, pourrait-on dire en usant de l’anachronie71, fait son apparition dans le genre pourtant très formaté et protocolaire de la chronique historique. Or, une révolution du regard apparaît dans notre texte : l’on ne sait plus très bien si Samori est un héros ou un imposteur, un guerrier sans égal ou un decei- ver72. Il y a une admiration manifeste de la part de Mallam Abu pour son personnage, qui se renouvelle à chaque paragraphe où les appositions sont majoritaire- ment laudatives : Samori est le chef d’ici-bas (duniya, “la vie, le monde terrestre, l’ici-bas”), par opposition au monde des cieux, royaume de Dieu : il est donc le plus grand parmi les hommes, littéralement sans égal, puisqu’hormis le Tout- puissant, il est le plus puissant d’entre tous (ba ka da kama ba !, “tu n’as pas ton pareil !”, “tu es sans égal !”, lui répète d’ailleurs régulièrement Mallam Abu, en l’apostrophant). Lorsqu’il prend une ville après un siège singulièrement difficile, comme c’est le cas pour Fita, la première des villes du texte, Mallam Abu com- mente l’exploit, en chantant l’une des devises de son héros : Samori, dain Kufila, sarkin yakin dunya, Samori Baturi babaku mutamni, “Samori, fils de Kufila, chef de la guerre d’ici-bas, Samori le Blanc des Noirs73” (page 21), qu’il reprend en la modulant à chaque épisode.

70. Pour une analyse de ce concept, voir F. Viti, “À la guerre comme à la guerre. De la cruau- té dans l’art du combat (Baoulé, Côte d’Ivoire, 1891-1911)”, in F. Viti and D. Casajus, eds., La terre et le pouvoir. À la mémoire de Michel Izard, Paris, CNRS éditions, 2012, 249‑270, notamment p. 251. 71. S. Piron, postface de P. Nuss, Dialectique du monstre : enquête sur Opicino de Canistris, Bruxelles; Le Kremlin-Bicetre, Zones sensibles; Les Belles lettres, 2015, 181, reprenant G. Didi-Hu- berman, distingue “l’anachronie (comme expérience de la traversée du temps, impliquée dans tout forme de connaissance historique) de l’anachronisme (erreur d’attribution à des traits caractéristiques d’une autre époque)”. 72. Purposeful deceiver, selon les catégories de Jung, du type de Renart dans le Roman de Renart. Le temps des “héros épiques” selon Bakhtine (M. Bakhtine, Esthétique et théorie du roman, Paris, Gallimard, 1978 [2011], 455) semble achevé. 73. Samori Baturi babaku mutami, “Samori le Blanc (l’Européen) des hommes noirs”. Bertho, L’empire de Samori Touré 45 Et pourtant de manière concomitante – et c’est cela qui est intrigant – Mal- lam Abu juge très sévèrement son personnage. La fin en exil qui lui est réservée sonne comme une punition divine en châtiment des massacres des musulmans qu’il a perpétrés au cours de son règne. Ainsi se clôt le texte : Voilà une autre histoire de Samori Les Français l’ont attrapé, ils sont partis avec Samori en Europe74. Le chef de la France a ri, il était content de la capture de Samori. Leur chef a élevé Arshanari en grade. Cette histoire s’est arrêtée encore. Voilà une autre histoire de Samori, et le chef de la France [a dit] : Qu’on l’amène au milieu d’une rivière, qu’on lui construise une maison ! Samori et sa femme au milieu de la rivière. Le chef de France a donc déposé Samori et sa femme à cet endroit. Jusqu’à la mort de Samori. On dit : La vie est faite d’illusions75. Eh ! Le pouvoir qui n’est pas celui de Dieu n’est pas un vrai pouvoir76. L’histoire de Samori fils de Kufila est devenue un mythe77 (page 182). Samori est mort en exil au Gabon, sur une île, deux ans seulement après sa capture78, ce que le narrateur ne glose que comme le résultat du dessein de Dieu (Labari Samori dain Kufila ya zama almara). Samori a été trompeur et parjure : il a rompu le pacte d’amitié qui le liait avec Sagak’iki, chef de Gawaguna79. Ana- lysons deux passages exemplaires de ce vacillement du jugement : Il était complètement païen80. Samori était musulman en apparence seulement81. Il ne fait rien que des tromperies. En matière de tromperie82, Samori dépasse tous les Noirs. Le chef, chef d’ici-bas, le fils de Kufiti, chef des tromperies, le musulman Samori, le chef des trompeurs83, des musulmans et des païens. Samori, tu n’es pas d’accord avec Dieu ! La tromperie de Samori est tromperie de ce bas-monde (page 7).

74. Su tafi da Samori Turi Farasi : soit c’est une incorrection pour Turawa, “eux les Français” (sont partis avec Samori), soit il manque la préposition a devant le complément de lieu “en Europe, en France”, ou bien “chez les Français”, qui pourrait tout à fait être compris comme “Saint Louis du Sénégal” dans cette seconde hypothèse locative. 75. A ci duniyya kida [gidan] kariya ne : “On dit : La vie est une maison de mensonge”. Autre- ment dit, la vie est un théâtre, et tout est vanité et illusion. 76. Ashe iko ba na Allah ba, kariya ne : “Eh! Tout pouvoir, si ce n’est pas celui de Dieu, c’est un mensonge”. La proposition entière devrait être au féminin (ce) et non pas au masculin (ne). 77. Labari Samori dain Kufila ya zama almara : almara, de al amari, dérivé de almura’at, est attesté par Bargery, qui donne dans son dictionnaire la définition suivante : “fiction funded on fact”. 78. C’est le capitaine Gouraud qui capture Samori à Guélémou, et non pas Archinard comme l’indique Mallam Abu, mais il faut noter que le personnage d’Archinard fait office de personnage unique dans la narration, incluant et englobant tous les officiers blancs. Voir, à partir des archives de Gouraud, l’étude de J. d’Andurain, La capture de Samory, 1898 : l’achèvement de la conquête de l’Afrique de l’Ouest, Outre-mer, Saint-Cloud, Soteca, 2012. 79. Gbankundo, dans Y. Person, Samori, une révolution dyula, tome 3. 80. Kafiri, dérivé de l’arabe kafir, “cafre, mécréant, incroyant, païen” : aucun des peuples du Livre, donc ni musulmans, ni juifs, ni chrétiens. 81. Il se dit musulman mais il ne pratique pas. Il n’est musulman qu’en apparence. Baka ciki : de bouche et de ventre. 82. Yawudara : “ruse”, hila : “tromperie”. 83. Rudi : “ruse, falsification”.

Africa, N.S., I / 1, 2019 46 a r t i c o l i / a r t i c l e s

Samori, le chef de la guerre d’ici-bas, Samori, le chef de la ruse, Samori, le chef des tambours de guerre (page 21). Le premier extrait révèle que Samori venait d’une famille dés-islamisée, parmi les Touré84, ce qui rend son appellation ultérieure d’almami hautement contestable, dans l’espace de la narration. Le second passage montre comment deux catégories incompossibles entre elles – “chef de la ruse”, “chef d’ici-bas” – puisqu’elles ap- partiennent à deux univers axiologiques différents, sont pourtant juxtaposées. Sa- mori appartiendrait en cela à une sorte spécifique de “héros romanesque”, pour reprendre les catégories de Mikhaël Bakhtine85 : héros spécifique, puisque non seulement le héros va à l’encontre de la société, comme un héros picaresque pourrait le faire ou un Quichotte, qui fonde le genre du roman, mais il est en outre à la fois admiré et réprouvé, ce qui produit des héros romanesques paradoxaux, voire des héros sur lesquels pèse un soupçon86. Comme si un soupçon pesait sur lui, et que les deux sentiments de fascination et de répulsion coexistaient à la fois. Comment expliquer ce paradoxe qui va jusqu’à menacer l’interprétation glo- bale du texte? Mallam Abu, après le récit de toutes ces victoires et l’énumération de toutes ces louanges, est-il finalement favorable ou non à Samori? L’a-t-il suivi de son plein gré et pourquoi raconter son histoire, si son comportement était si répréhensible? Nous pensons que cela est peut-être dû au caractère extrêmement rapide de l’ascension de Samori : en l’espace d’une génération, cet homme fils d’un talaka (“homme du peuple”), d’abord simple colporteur, se taille un empire de plusieurs milliers d’habitants, qu’il déplace ensuite intégralement vers l’Est en continuant à combattre et les Français et les Anglais sur plusieurs années encore. Les qualités de stratège d’un tel homme forcent le respect. Mais il manque une assise à ce pouvoir qui n’est appuyé ni par la religion adossée à des réformes sociales, comme l’ont pu être les jihads peuls d’Ousmane dan Fodyo ou d’El Hadj Omar Tall, ni par une ascendance illustre qui aurait légitimé pour lui son pouvoir en le référant à une dynastie toute entière. Et à ce titre, Samori ne reste qu’un simple humain, certes exceptionnel, mais néanmoins faillible. C’est ce qui expliquerait les retours répétés du narrateur sur l’origine modeste de Samori (Samori, fils de Kufila, fils de talaka, “l’homme du peuple, l’homme de rien, le pauvre”), que l’on retrouve d’ailleurs pour son fils Umaru-Sarankèn Mori alors même qu’il est loué pour la prise de Boli87 : “Un fils de colporteur reste un colporteur”, page 163. La sentence est implacable.

84. Ce que rapporte Y. Person, sur l’origine des Touré : Y. Person, “Les ancêtres de Samori”, Cahiers d’études africaines, 13, 4, 1963, 125-156. 85. Bakhtine, Esthetique et theorie du roman. 86. À l’image des personnages romanesques de l’après-guerre ; N. Sarraute, L’ère du soup- çon, Paris, Gallimard, 1956 : il ne s’agit évidemment pas de considérer que Mallam Abu intègre une critique du roman du XIXe siècle européen dans sa chronique. Il s’agit de souligner à quel point le texte fait coexister en son sein deux univers axiologiques différents, faisant du personnage-Samori un cas limite, au plan narratif, entre modèle héroïque épique et modèle romanesque. 87. Certainement Bolé au Ghana actuel, puisque e et i ne sont pas différenciés dans la graphie ajami, et que Sarankèn Mori se trouve tout près de Wa et de Bouna à ce moment-là. Bertho, L’empire de Samori Touré 47

Ni révolutionnaire jihadiste, ni dauphin d’une grande lignée, Samori est un dyoula. Il nous semble que c’est précisément ce qui fonde le narrateur à porter des jugements en apparence aussi contradictoires sur le personnage, où son ascension provoque une fascination véritable, et où sa chute, vécue dans l’espace de la même génération, ne peut être compréhensible et appréhendable que comme un châtiment divin. En 1914, à l’époque où les bouleversements politiques et so- ciaux sont immenses, où la colonisation commence à s’installer de manière pé- renne, les catégories de “pouvoir juste”, de “pouvoir légitime” qui structuraient habituellement les chroniques historiques sont sérieusement ébranlées. Cela n’empêche pas Mallam Abu d’être assez critique, voire moqueur, envers les An- glais, puisque la colonne Henderson (le personnage nommé Mazabi, dans le texte) est mise très vite en déroute par Sarankèn Mori. Mais la colonisation induit des changements de fond, plus latents, dans la structure de la représentation du pouvoir : à échelle d’homme, alors que les royaumes dont Mallam Abu parle ont disparu, et que le plus grand d’entre eux était né de la volonté d’un seul homme, il est fort difficile de présenter un jugement univoque sur les trente dernières an- nées. La seule issue, en termes de narration, est d’expliquer sa chute par la répu- tation qui l’a précédée dans le Nord de la Côte d’Ivoire et de la Cold Coast – où la conquête a été plus rapide, et où elle a été surtout menée sur de nombreuses villes saintes musulmanes (dont Kong) – : l’impiété, l’impudence, le goût de la conquête. Puisque Samori est humain et faillible, il est dès lors concevable qu’un autre homme le surpasse et ait finalement raison de lui. Dans le récit, ce sera Ars- hanari – Archinard – qui est une concaténation de l’Archinard historique, mais aussi du capitaine Gouraud qui le capture en 1898, et des autres officiers français qu’il a combattu tout au long de sa vie, Combes, Humbert, Marchand, Monteil, Braulot... C’est le sens de cette glose du narrateur, après la mention des fuites successives de Samori vers l’Est : L’Européen Arshanari il a chassé Samori, Samori fils de Kufila. Eh bien, les hommes se dépassent les uns les autres ! [Mazaji su fi juna] (page 148). “Les hommes se dépassent entre eux”, autrement dit : l’on peut toujours trouver plus fort que soi d’une part, mais également : celui que l’on croyait invin- cible a trouvé son maître. Ces renversements de situation sont possibles dans le domaine de l’ici-bas, de la vie terrestre (duniya), de l’humain (mutumin, “le mor- tel” au sens étymologique, celui qui est destiné à la mort, mutu).

***

Labarin Samori correspond d’abord à une visée de témoignage, demandée par le Dr. Corson : écrire ce que l’on a vu. Mallam Abu cherche également à at- teindre une visée historique – écrire le vrai d’un événement passé – mais elle se mêle rapidement à une visée morale et didactique : décrire pour inciter à la ver- tu.

Africa, N.S., I / 1, 2019 48 a r t i c o l i / a r t i c l e s

Pour aucune de ces trois entrées, de témoignage, d’historicité, de morale, l’on ne sait si Mallam Abu exalte ou condamne Samori, puisqu’il fait les deux à la fois. La raison en est que le récit se trouve à la croisée de nombreuses influences et des intertextes charriés par tous ces univers référentiels. Alessandra Brivio

Th e Wi l l i a m Wa d é Ha r r i s Le g a c y in Gh a n a : On Ru p t u r e s a n d Continuities*

Abstract

This paper reconstructs the genealogy of William Wadé Harris’ movement in Ghana, discussing the interaction of ideas, religious traditions, political constraints and personal agency. My aim is twofold: to underline continuities between African Independent churches, Missionary churches, Pentecostal churches and traditional religion, on the one hand, and to illustrate the recurrent dialectic between rupture and continuity. k e y w o r d s : r e l i g i o u s m o v e m e n t , g h a n a , a f r i c a n independent c h u r c h e s , f e t i s h , w i l l i a m w a d é h a r r i s

Introduction

I first met Andrew, anɛsofo (priest) of the Twelve Apostles Church (TAC), in August 2012 in his home village of Ellohin, a few kilometres west of Beyin, in Ghana’s Western Region. The TAC had emerged out of prophet William Wadé Harris’ stay in the Gold Coast in 1914. Harris only spent three months in the country, but his preaching had very long-lasting effects. It was his followers, who would eventually become known as Water Carriers because of the centrality of water in their healing rituals, who had founded the TAC in the 1920s. Andrew had recently returned to Ellohin from Tarkwa, where he had lived and worked as a young man. In Tarkwa, Andrew had been a follower of the Pen- tecostal church and had begun a training to be a pastor. He, however, had economic and personal problems, and his life was not going well. One night he dreamt that a mallam, an “old Muslim man”, was suggesting him to go back to his village and take care of the old, abandoned “garden”, the open space that had belonged to his great-aunt and where the TAC congregation performed their rituals and

*This article was completed thanks to funding from MEBAO (Missione Etnologica in Benin e Africa Occidentale), and the European Research Council as part of the ERC project 313737: Shadows of Slavery in West Africa and Beyond: a historical anthropology. It is the result of field and archival research I undertook between 2012 and 2014 in Ghana.

Africa, N.S., I / 1, 2019 50 a r t i c o l i / a r t i c l e s ceremonies. He listened to the dream because he recognised the spirit as the same one who had inspired his great-aunt’s healing work, and in so doing, he told me, he found the path to righteousness. When we met, he was reorganising the garden and building a big hut to host his patients. Every Sunday morning he celebrated “mass” in the garden of the oldest and most powerful ɛsofo of the village, Kanga Soma, who by that time was confined to her bed1. During the Sunday mass, ɛsofo Andrew made good use of the pentecostal way of preaching. Rather than posing a problem, for him, his passage from the Pentecostal church to the TAC was a sign of continuity on his path to God. He, however, was aware of being a beginner and was thus determined to continue to improve his knoweldge of traditional healing and the power of ancestral “spirits”. I accompanied him on his visits to the gardens around the region and saw him consulting the old traditional priestess (kɔmenle) of Ellohin and, occasionally, following the suggestions of the mallams who sold medicines in the local markets. As he explained me, if one wished to defeat “bad ghosts” and protect people he should know how to deal with them. He was in contact with water spirits and mmotia (dwarf spirits dwelling in the forest); for this reason, every morning he preached on the beach in order to control and repulse insidious spirits. Andrew’s attitude toward religion was not an exception and needs to be understood in the historical context of the TAC, one of the main African Independent Churches (AICs). The image of the AICs was brought closer to that of traditional religions by the close association in Pentecostal studies and in the Neo-Pentecostal rethoric between breaks with the past and the embracing of global modernity2. However, recent studies claim that a focus on the ostensible differences between Neo-Pen- tecostal churches, Pentecostal churches, AICs and traditional religions obfuscates more than it reveals and does not contribute to a full understanding of the religious processes in Africa3. By the same token, the discontinuity between past and present is more easily asserted in narratives of the self than is truly enacted in everyday life; ruptures with the past always go hand-in-hand with a complemen-

1. Kanga Soma was one of the ɛsofo interviewed by Ernesta Cerulli, see E. Cerulli “I Water Carriers nove anni dopo”, in Aa. Vv., Religioni e Civiltà, vol. 1, Bari, Dedalo, 1972, 69-124. An- drew’s aunt discovered her spirit thanks to Kanga Soma. According to Cerulli, Kanga Soma had inherited her spirit, called Mallam, from her mother. We can suppose that the same spirit was believed to have passed to Andrew’s aunt and today to Andrew. 2. B. Meyer, “Christianity in Africa: From African Independent to Pentecostal Charismatic Churches”, Annual Review of Anthropology, 33, 2004, 447-474. 3. See between others: R. van Dijk, “Time and Transcultural Technologies of the Self in the Ghanaian Pentecostal Diaspora” in A. Corten and R. Marshall-Fratani, eds., Between Babel and Pentecost: Transnational Pentecostalism in Africa and Latin America, Bloomington, Indiana Uni- versity Press, 2001, 216-234; R.L. Blanes, R.Sarro, “Prophetic Diasporas: Moving Religion Across the Lusophone Atlantic”, African Diaspora, 2, 2009, 52-72; M. Engelke “Past Pentecostalism: No- tes on Rupture, Realignment, and Everyday Life in Pentecostal and African Independent Chur- ches”, Africa, 80, 2, 2010, 177-199. Brivio, The William Wadé Harris Legacy in Ghana 51 tary desire for realignments, today as at the time of Harris’ passage in the Gold Coast4. Even the global and transnational dimension – read as desire for modernity – was already a feature not only of the first steps Christianity took in Africa, but also of the traditional religions with their strong itinerant attitude and a deep interest in foreign gods5. In this article I reconstruct the genealogy of Harris’ movement, discussing the interaction of ideas, religious traditions, political constraints and personal agency, in order both to underline commonalities between AICs, missionary churches, Pentecostal churches and traditional religion, and to show the constant and paradigmatic tension between rupture and continuity. The study of prophetic Christianity has a long tradition in anthropology. In the anti-colonial context of the 1960s and 1970s, most authors tended to link prophetic religious movements and political oppression, understanding the rise of such movements primarily as a response to colonial religious institutions, whether they were mission churches or “reified” traditional religion6. Several influential works have thus emphasised the political significance of religious movements in Africa7. Other authors, such as Cornel West and Karen Fields, focused their analysis on the cultural capacity of oppressed peoples, while Eboussi Boulaga concentrated on the emancipatory role of African prophets8. Jean Comaroff sought to intersect the processes of colonial domination with the cultural signification of the independent churches9. Many of these approaches, however, have not stood the test of time, partly because they focused on the workings of colonialism as the sole explanatory framework10. Nowadays, the scholarly consensus is that African religions cannot be studied solely as “religions of the oppressed” and that their theological dimensions are no less deserving of careful analysis than their political ones11. The AIC, after all, did survive the end of colonialism, while the Christian imagination has conti-

4. See: Engelke, “Past Pentecostalism”. 5. For a discussion consult: M. Augé, Génie du paganisme, Paris, Gallimard, 1982. 6. T.O. Ranger, “Religious Movements and Politics in Sub-Saharan Africa”, African Studies Review, 29, 2, 1986, 1-69, 2. 7. See: G. Balandier, “Les mouvements d’innovation religieuse en Afrique Noire”, in H.C. Puech, eds., Histoire des Religions, vol. 3, Paris, La Pleiade, 1976, 1243-1276; V. Lanternari, The Religions of the Oppressed, New York, Alfred A. Knopf, 1963; T.O. Ranger, “Connexions Between ‘Primary Resistance’ Movements and Modem Mass Nationalism in East and Central Africa. Part I”, Journal of African History, 9, 3, 1968, 437-453; T.O. Ranger, “Connexions Between ‘Primary Re- sistance’ Movements and Modem Mass Nationalism in East and Central Africa: II”, Journal of African History, 9, 4, 1968, 631-641. 8. C. West, Prophetic Deliverance: An Afro-American Christianity, New York, Westminister Press, 1982; K. Fields, Revival and Rebellion in Colonial Central Africa, Princeton, Princeton University Press, 1985; F.E. Boulaga, Christianisme sans fetiche, Paris, Presence Africaine, 1981. 9. J. Comaroff, Body of Power, Spirit of Resistance: The Culture and History of a South Afri- can People, Chicago, University of Chicago Press, 1985. 10. T.O. Ranger, “Religious”, 2. 11. J.D.Y. Peel, Religious Encounter and the Making of the Yoruba, Bloomington and India- napolis, Indiana University Press, 2000.

Africa, N.S., I / 1, 2019 52 a r t i c o l i / a r t i c l e s nued to be a source of creativity and innovation in postcolonial times: churches do not cease to appear, grow, diversify and disappear12. The success of Harris in Ghana was not the result of a sudden political uphe- aval, or the expression of resistance to colonialism, or a dramatic conversion, but the consequence of pre-existing local dynamics. Harris’ preaching was generally ignored by colonial officials, accepted by the African elite and, to a degree, welcomed by traditional chiefs and Christian missionaries. The latter had been working in the Gold Coast long before Harris’ arrival and, despite the shortcomings and failures in their conversion process, had already prepared the terrain. The first attempt to open a Roman Catholic mission in Axim, in the Western Region, had taken place in August 1882. Local chiefs had welcomed the priests and offered them a big piece of land. However, the scarcity of Catholic religious personnel in the Gold Coast had thwarted their plans. The Western Region was abandoned until 1897, when a group of young traders from Elmina headed by John Charles Anthony Rhule funded a small society in Axim called the “Roman Catholic Union”, in order to distinguish themselves from the Methodist Club which had been in existence for 20 years. After a few weeks, the Union counted 56 members, but most of them were not baptised. A priest was sent once a year to bring them the sacraments. It was only in 1902 that Mgr. Albert and Fr. Pleger went to Axim and decided to establish the Mission. Mr. Rhule immediately put his house at their disposal, one of the rooms serving as chapel13. In Nzema, the oldest Christian church was the Methodist, linked to the Axim Circuit, whose activities came up to Sanwi in southeastern Ivory Coast. In 1914 it served about two thousand Methodist adherents14. The Methodist presence among the influential Nzema people who migrated from the Gold Coast to the Ivory Coast was crucial to Harris’ success, since he arrived in the Ivory Coast as a Christian preacher trained in the Methodist Church, and “in his preaching he spoke essentially as a Christian”15.

Wadé Harris in the Gold Coast

William Wadé Harris was a Kru man. He was born at Cape Palmas, Liberia, and educated in the American Episcopal Church, where he worked as a catechist. He was in his fifties when, in March 1914, he crossed the Tano lagoon into the

12. R. Sarro, The Politics of Religious Change on the Upper Guinea Coast: Iconoclasm Done and Undone, Edinburgh, Edinburgh University Press, 2009. 13. Archives of the Società Missioni Africane, Rome, (SMA), 3A4, Fond Heesewijk - Histoi- re d’Axim et Half Assinie, 1902-1961, 2-4. 14. I wish to thank Pierluigi Valsecchi for sharing with me this information, as well as the documents from the Archives nationales du Sénégal. 15. V.L. Grottanelli, Una società guineana: gli Nzema, vol. 2, Ordine morale e salvezza ter- rena, Torino, Boringhieri, 1977, 161. Brivio, The William Wadé Harris Legacy in Ghana 53

Gold Coast near Half-Assini. His first stop was probably Beyin, the capital of Western Appolonia16. On the 20th April, he left Atuabo, the capital of Eastern Ap- polonia, for Axim, where his headquarters would be located for at least a month. He never travelled any further east and, in July, he decided to return to the Ivory Coast: he intended the break to be temporary but he never returned17. The prophet’s message was simple. In order to be converted, people had to burn their idols and “fetishes” or cast them into the sea. They were to stop worshipping many gods in favour of the Supreme God; to do this, they had to read the Bible and join one of the mission churches, of whichever denomination. Harris confirmed the conversion with the rite of baptism. During the Wesleyan Method- ist Synod held in 1915, the Rev. Ernest Bruce gave a summary of the extraordinary work Harris had conducted in the Axim District, where at least 52 villages had embraced Christianity as a result of his predication18. Harris was able to involve the kɔmenle and the local chiefs, such as the chiefs of Axim and Beyin, who made a bonfire of their “fetishes”19. The same occurred in Atuabo, where Harris baptised the omanhene (Paramount Chief) of Eastern Nzema20. When Harris had first arrived in the Ivory Coast, he had found people ready to embrace his message; the most militant among them hailed from the Gold Coast. The men and women from Dixcove and from the Nzema area (expecially Axim, Eastern and Western in Appolonia) who had settled to trade in Grand Bassam, in the Ivory Coast, had brought Methodism with them. They had economic clout and had gained positions of power in the Ivory Coast, where the French Colonial Govern- ment appointed them as members of the Tribunal or Local Council. The Christians among them congregated for Sunday worship. John Serbah, a man of the Cape Coast Methodist Church who had a business in the Ivory Coast, had built them a two-storey house and offered a plot of land, thus providing them with a place of

16. For details on the biography of William Wadé Harris, see: J.E. Casely Hayford, William Waddy Harris, London, C.M. Phillips, 1915; SOAS Archives, London, (Wesleyan) Methodist Mis- sionary Society Archive, MMS /17/01/01/035, P. Benoit, “Story and Finding of the Prophet William Wadé Harris”, (Miss Thompson typescript traslation from French), 1926; D.S. Ching, Old Man Union Jack. The Story of Prophet Harris, London, Cargate press, 1947; G.M. Haliburton, The Pro- phet Harris, London, Longman, 1971; D. Shank, Prophet Harris, the “Black Elijah” of West Afri- ca, Leiden, Brill, 1994; S. Walker The Religious Revolution in the Ivory Coast: The Prophet Harris and His Church, Chapel Hill, University of North Carolina Press, 1983; R. Bureau, Le prophète de la lagune. Les harristes de Cote d’Ivoire, Paris, Karthala, 1996; J.P. Dozon, La cause des prophèts. Politique et religion en Afrique contemporaine, suivi de La leçon des Prophètes par M. Augé, Paris, Seuil, La Librairie du XXe siècle, 1995. 17. The Gold Coast Nation, 7 May 1914. 18. The Gold Coast Nation, 25 February 1915. 19. The reporter of The Gold Coast Leader wrote that in Axim a kɔmenle who had the sole control of the “fetishes” in town had asked all his “subordinates” to surrender their “fetishes” as “he had personally examinated Harris and found him to be not an ordinary man”. The Gold Coast Lea- der, 4 July, 1914. 20. J. van Brakel, The SMA Missionary Presence in the Gold Coast (1906-1924), vol. II, Oost- erbeek, 1994, 282.

Africa, N.S., I / 1, 2019 54 a r t i c o l i / a r t i c l e s worship, which later became the Methodist Mission House. For years, the Axim sta- tion in the Gold Coast was responsible for the supervision of that church21. J. Casely Hayford was one of the most outstanding intellectuals in the Gold Coast. An important political leader, he was an activist in the Aborigines’ Rights Protection Society and a member of the Methodist Church. For him, Harris was a black man sent by God to help Africans on their path to independence and eman- cipation from white rule. In Axim, he personally welcomed Harris as a reformer, “a dynamic force of a rare order”, who would “move this age in a way few have done”22. However, Harris’ legacy would soon prove to be problematic23.

The Harris movement: John Swatson and the Missionary churches

Harris had neither the time nor, perhaps, the wish to establish a new church. He moved quickly from place to place, and the only institutional step he took in every village he visited was to identify twelve elders, calling them the Twelve Apostles and entrusting them with the task of continuing his missionary work. But there was no structure ready to welcome the impressive numbers of converts he left behind. The Roman Catholic, the Methodist and, later, the Anglican churches all tried to exploit Harris’ work, but were unable to receive and process all of his converts24. This lack of supervision led the missionaries to fear that a new “fetishism” could replace the old one. A review of Casely Hayford’s pamphlet William Waddy Harris, published on The Gold Coast Nation, claimed that the inrush of converts under Harris’ preaching was one of the greatest problems confronting the Christian churches25. Theological problems also arose, including a controversy over the legitimacy of Harris’ baptisms and complaints about the converts’ general theological ignorance, especially in the case of women, who formed the majority of Harris’ followers26. Crucial to the diffusion of Harris’ teachings was John Swatson, a native of Beyin, born to a royal mother and a European father attached to the royal court of Beyin27. Between 1912-1913, he had served as a Methodist agent in charge of the congregation of Aboisso, in Ivory Coast. When Harris made his appearance, he

21. F.L. Bartels, The Roots of Ghana Methodism, Cambridge, Cambridge University Press, 1965, 173-174. 22. Casely Hayford, William, 6. 23. The Gold Coast Nation, February 25, 1915. 24. The Methodist Church did not have enough personnel to gather all the people who had embraced Christianity on account of Harris. They lacked the required financial and human capital and they also faced competition from the Roman Catholic church. Between 1916 and 1920, Father G. Fisher, of the Roman Catholic Fathers of Lyons, was able to welcome about four thousand converts. H.W. De- brunner, A History of Christianity in Ghana, Accra, Waterville Publishing House, 1967, 273-275. 25. The Gold Coast Nation, 9 December 1914; The Gold Coast Nation, 18 February 1915. 26. SOAS Archives, London, MMS 17/01/01/035, W.R. Griffith to Mr. Godie, 4 September 1914. 27. Haliburton, The Prophet, 217-227. Brivio, The William Wadé Harris Legacy in Ghana 55 had followed him into the Gold Coast and then, when they were recalled to Assinie, back to the Ivory Coast. Swatson was probably the first and only person ordained as a prophet by Harris28. In September 1914 the two were welcomed in Half-Assinie. The visit was reported in The Gold Coast Leader in the following terms: “Mr. Swatson, the catechist of Wesleyan Mission Church at Aboisso, Ivory Coast, arrived on a visit. He mounted the pulpit on the 5th and conducted the service both morning and evening, in our native language. The Church was crowded with members and we were pleased to hear the sermon going on in our language”29. In January 1915 the same newspaper published the news: “rumour is current that Mr. Harris is not coming back again and Mr. John Swatson who was ordained by Mr. Harris as a prophet has now taken up the work by casting out devils, idols, etc. at Kinjarboe [Krinnjabo] and many other districts”30. However, Swatson’s reputation soon deteriorated among Methodists. After Harris’ departure, he gave up his direct involvement in the Methodist Church and joined the Anglicans. In July 1918, The Gold Coast Leader criticised Swatson and the presumed “errors” committed by the Anglican bishop: If the Bishop had followed our suggestion he would have known something about this Wesleyan Catechist who had been harrowing the Society throughout all his connections with them at Half Assinie and Aboisso in the Ivory Coast, baptising persons when he had not the authority to act as much [...] enrolling himself again in the work of God during the Professor Harris’ era, covering himself with the authority of this wonderful man, procured one of his (Harris) wands, baptising for money in all the bush villages contrary to Harris’ Rules, subsequently realising what was in store for him after being caught with others of his line in the Government trap thought it expedient to attach himself to Dr. Hayford’s Baptist Church with Nakkabah [Nackabah] and others and as they seem disappointed resorted to their sacrilegious worship again, until Swatson recently communicated with the Bishop to get him attached to the SPG [Society for the Propagation of the Gospel] to enable him to carry on his tactics31. Mark Hayford, the brother of Casely Hayford, was a dynamic religious busi- nessman who in 1898 had founded the Baptist Church in Cape Coast, in connection to the Native Baptist Church of Nigeria. He had been to England and Canada to collect funds for his church, school, and teacher-training centre, and had contrib- uted to make the existence of an African Christianity known in the world32. How- ever, in the Gold Coast, he had a bad reputation. The Gold Coast Leader wrote

28. During Harris interview with Benoit, that took place in Las Palmas, at the question “Have you often appointed minor prophets?” Harris answered “One only – Rev. Swanson of Big Assinie. I appointed him as a preacher and I brought him before the governor at Bingerville. The governor gave him a permit to continue his work. He works at Assinie” SOAS Archives, London, MMS /17/01/01/035, Benoit, “Story”. 29. The Gold Coast Nation, 3 October, 1914. 30. The Gold Coast Leader, 9 January, 1915. 31. The Gold Coast Leader, 6-13 July, 1918. 32. Debrunner, A History, 237-238.

Africa, N.S., I / 1, 2019 56 a r t i c o l i / a r t i c l e s ironically in August 1918 that “we are not surprised to hear that Mr. Swatson the Apollonian ‘Bishop’ has again reverted to Dr. Mark Hayford’s Baptist Mission, what name do you give to a man of this nature and those connected with him?”33. His reputation as a greedy, cheating man also reached the Ivory Coast, where he had first travelled in 1905 to collect funds and later, in 1920, for evangelising, collecting money and selling sacred images34. After Harris’ departure, both Nackabah and Swatson began seeking their evangelical path. First, they joined Mark Hayford’s Baptist church, an encounter which did not bear the expected fruit. Swatson then decided to look to the Anglican Church, which had a minor presence in the area35. Swatson obtained important results in the Sefwi and Denkyira regions and made many converts despite opposition from local chiefs and colonial officials, who accused him of causing social and political unrest. He met the Anglican Church ministers E.D. Martinson and C.H. Elliot, who took him to Archdeacon Morrison in Kumasi. As denounced on the pages of The Gold Coast Leader, he asked the SPG to take over his congregation, and gave Morrison a list of more than a thousand converts who were then accepted as Anglicans by Bishop O’Rourke of Accra36. Over the years, the Anglican Church changed its opinion about Swatson. He was described as “a rather undisciplined and a very ignorant, self-constituted missionary”37. However, they did take advantage of Swatson’s unconventional preaching, and some years after his death the Anglican bulletin, The Golden Shore, recognised the importance of his work: He was undoubtedly used by the Holy Ghost, [...] and by the very irregularity of his methods achieved in a short time remarkable changes in primitive “fetish” country. John Swatson had been through the Denkyira district carrying a crude cross staff, denouncing “fetishism” and proclaiming the true God. In village after village the people, moved mainly by fear, burned their fetishes and submitted to Swatson’s Christianising ceremony38.

The birth of the Twelve Apostles Church

Grace Tane, John Nackabah and John Swatson met during Harris’ time in the Gold Coast. Grace Tane was a traditional priestess converted by Harris, probably

33. The Gold Coast Leader, 24 August, 1918. 34. Archives Nationales du Sénégal (ANS), 5G62, Le prophet William Harris, 1914-1920, Letter from the Administrator en Chef des colonies, Inspecteur des Affaires Administratives to Monsieur le Lieutenant-Governeur de la Cote d’Ivoire, Bingerville, 29 May 1920. 35. J. Pobee, The Anglican Story in Ghana: From Mission Beginnings to Province of Ghana, Accra, Amanza Limited, 2009, 154. 36. Haliburton, The Prophet, 218-221. According to J.S. Pobee, Bishop O’Rourke’s approach to the mission was “very audacious”. He declared Swatson’s baptisms valid and polygamists were fully accepted inside the Church. Pobee, The Anglican, 155. 37. Rev. A.H. Candler, “Earlier Days”, The Golden Shore, 9, 10, 1943, 159. 38. Rev. Candler, “Earlier Days”, 159. Brivio, The William Wadé Harris Legacy in Ghana 57 on the banks of the Ankobra river. She lived in Krisan, a few kilometres east of Atuabo, where she established her church. Today a version of her story recalls that she was such a powerful woman that it took several days for Harris to drive out Tanoε bozonle (spiritual entity) from Grace39. After her conversion, she followed Harris and became his third wife40. They separated when Harris left for the Ivory Coast. John Nackabah was converted in Axim and then went back to his village, Essuawa, where he immediately started preaching and baptising41. By founding the TAC, Nackabah and Grace Tane attempted to overcome the lack of institutional structures which had characterised Harris’ religious action. Nackabah and Grace Tane met again some years after their first encounter, when she was sent to Nackabah to be healed. On this occasion, they probably decided to work together to carry on Harris’ mission. The two are said to have founded the Church around 1918, and a huge mass of converts soon joined it. Grace Tane was regarded as the “owner” of the Church, and Nackabah, who recognised her authority, was her staunch supporter42. This hierarchical arrangement was based on her having been the first of the two to be converted and Harris’ wife. Tane and Nackabah emphasized healing over the other dimensions of their ministry. Through the act of healing they built strong bonds with their followers and those bonds were the demonstration of the successful transfer of spiritual power. Following the typical structure of affliction cults, the sick, once recovered, became part of the Church and could, if they had the “spirit”, become prophets, set up healing activities of their own and manage their own shrines, called “gardens”. Each prophet owned a “garden” and regularly attended to patients, whose numbers depended on the prophet’s skills. In order to host patients, each “garden” had facilities such as sleeping rooms, kitchen and bathrooms. The main healing exercises were held on Fridays.

39. The tufuhene of Beyin, Mr Egya Kanra, related to Grace Tane abusua, told me about her bozonle resistance to conversion. Her name actually means “one born of Tano” and signifies that her birth was attributed to the river bozonle (god) Tanoε. Interview with Mr Egya Kanra, Beyin, Ghana, 22 August 2012. 40. We have no clear information about the other wives. According to SMA documents, they all became the founders of the Water Carriers, one at Anfukum, one at Ankobra and one at Atuabo (likely Grace Tane). SMA, 3A4, Fond Heesewijk - Histoire d’Axim et Half Assinie, 1902-1961, Abstracts from the Axim Diary occasionally supplemented by Father Joseph Stauffer, 16. 41. On September 1919, during a visit to the villages along the Ankobra river, Father Stauffer stopped in Ashiem, where he founded two chapels: one for the Wesleyans and one “for the Nakaba”. SMA, 3A4, Fond Heesewijk - Histoire d’Axim et Half Assinie, 1902-1961, 25. 42. Harris’ preaching resulted in a proliferation of African Independent churches both directly linked to his activity, and imitating his example. Among them one can cite the “West African water healing society” by John Kojoe Baidoe who was the secretary of Nackabah. Kojo Monnor founded the “Wonder worker and founder of the Kajirfeh divine healing society”, which successfully spread among the Nzema in the 1940s. He claimed to be linked to Harris’ preaching and to be a Roman Catholic. See Debrunner, A History, 273-274. In that period the downfall of many Christians was caused by Monnor’s church, the success of the Esongo healing cult, and the absence of doctors, hospital and dispensary. SMA, 3A6 Fond Heesewijk- Church in Nzima 1905-1962, Fr. Kerkoffs Report, 48.

Africa, N.S., I / 1, 2019 58 a r t i c o l i / a r t i c l e s

Very soon, observers noted the proliferation of “gardens” and condemned this instance of religious non-uniformity and fragmentation, probably influenced by the pervasive fear of the re-emergence of “fetishism”. In 1930, a colonial official described the situation in the following terms: The Missions then followed Harris up with sometimes somewhat amusing results. For instance, the frontier Chiefs were so nonplussed by the varied forms of Christian worship that they divided their villages into enough section to meet the number of them and consequently each worshipped in accordance with the population of his house. There is still a Sect, the twelve Apostles who claim to be the followers of the prophet Harris. You can imagine my astonishment the other day when a messenger entered my office and solemnly announced the twelve Apostles desired to see me. Nevertheless, at the time Harris himself did much good43. Ten years later, A.Q. Kyiamah, a clerk employed by the District Commis- sioner (DC) of Sekondi, confirmed this trend: The number of water-carrying professors, mostly illiterates of female sex, is growing daily [...] with the increase in the number of the professors, change has come. Instead of being only a mysterious system of curing disease, water-carrying has also become a religious sect with the general name “Church of the Twelve Apostles”. Even with the general name, it is difficult to class all the branches as one [...] each has his or her own method to management and, undoubtedly, his or her own creed44. Colonial concerns notwithstanding, ritual practices were actually fairly simple and homogeneous and still today correspond to DC McAuley’s description: the temple was an open shed and contained a small table, a bell and a barrel full of “holy water”. Each member, male or female, was bathed by another member. The idea was that bathing in the “holy water” freed from all diseases and evils. The sick person was required to carry the “holy water” in a basin on his/her head, while he/she looked at the sun. The person walked round in circles or staggered forwards and backwards, while both Kru and Christian songs were sung, to the accompanying sound of shaken calabashes. The senior leader carried a staff with a cross at the top swathed in white cloth, which he dipped in the “holy water” and waved “as a magic wand at the others”45. All the prophets recognised the central authority of the two leaders and the headquarters remained in Krisan, Grace Tane’s village. Until the 1920s the TAC was mainly confined to the Nzema region, but the two leaders, aware of the risk of a possible departure from Harris’ teachings, travelled regularly to check the work of the new prophets and demanded to be visited regularly by them. By the

43. PRAAD (Public Records and Archives Administration Department), Ghana, ADM 11/1/1050, Summary of December Quartely Reports, Western Province. 44. PRAAD (Cape Coast), ADM 23/1/1025, Letter Kyiamah to DC’s Office, Sekondi, 6 Fe- bruary, 1940. 45. PRAAD, ADM 11/1/1050, Letter DC McAuley to Commissioner of Central Province, Cape Coast, 21 April 1941. Brivio, The William Wadé Harris Legacy in Ghana 59 early 1930s, the movement began to spread outside Nzema. Tane worked mainly in the coastal area and sent her converts eastward along the coast. Nackabah sent his disciples in the same direction but into the bush, in the Ahanta area, northwest of Sekondi and Takoradi. In the mid 1930s, a new disciple called Michael Kojo George was sent east along the coast to work with the Fanti. By 1936, he had founded a “garden” in Saltpond, and made his first contacts in Kromantse; he already had some disciples in Accra and Kumasi by the early 1940s46.

Anti-witchcraft and “fetish” activities

In the days of Harris’ preaching and later, in the 1920s and 1930s, the Gold Coast also witnessed massive anti-witchcraft movements such as Aberewa, Hwemesu, Kunde, Nana Tango, Tigari, and others. These movements encom- passed two main elements, as their oracles cast witchcraft out of the afflicted and healed him/her of its consequences. The guilty person was compelled to drink a special “medicine”, a mix of water and herbs, or sometimes just water, to extin- guish the power of witchcraft. They, of course, were also able to prescribe treatments for other illnesses that caused disturbances to the social order and threatened individual integrity, such as adultery, infringement of taboos, curses, and so on. Through the taking of the medicine, patients would be bound to the priest; if he or she refused to take the medicine or to respect the rules imposed by the movement, the result could be death. Anti-witchcraft movements were linked to the worship of new gods, which originated in the North, in the Savannah region, and then spread all over the Gold Coast47. Tom McCaskie has pointed out one important similarity between the anti- witchcraft movements and their principal competitors, the Christian churches and the prophetic movements. In the colonial period, he stated, “countless people drift- ed in and out of a succession of religio-cultic allegiances. By their own testimonies, their peregrinations of belief and adherence were prompted by the search for a

46. P. Breidenbach, “Maame Harris Grace Tani and Papa Kwesi John Nackabah: Independent Church Leaders in the Gold Coast, 1914-1958”, The International Newspaper of African Historical Studies, 12, 4, 1979, 581-614, 595-596. 47. On anti-witchcraft movements in Ghana, see among others: H.W. Debrunner, Witchcraft in Ghana: A Study On the Belief In Destructive Witches and Its Effect On the Akan Tribes, Accra, Presbyterian Book Depot, 1959; M.J. Field, “Some New Shrines of the Gold Coast and their Signif icance”, Africa, 13, 2, 1940, 138-149; J. Goody, “Anomie in Ashanti”, Africa, 27, 1957, 356-63; N. Gray, “Witches, Oracles, and Colonial Law: Evolving Anti-witchcraft Practices in Ghana, 1927- 1932”, The International Journal of African Historical Studies, 34, 2, 2001, 339-363; T.C. Mc- Caskie, “Anti-Witchcraft Cults in Asante”, History in Africa, 8, 1981, 125-154; J. Parker, “Northern Gothic: Witches, Ghosts and Werewolves in the Savanna Hinterland of the Gold Coast, 1900s- 1950s”, Africa, 76, 3, 2006, 352-380; V. Vasconi, “Witchcraft, Medicine and British Colonial Rule. Anthropological Analysis of Colonial Documents in the Gold Coast”, in M. Pavanello, ed., Per- spectives on African Witchcraft, London, Routledge, 2017, 81-104; B. Ward, “Some Observations on Religious Cults in Ashanti”, Africa, 26, 1957, 47-60.

Africa, N.S., I / 1, 2019 60 a r t i c o l i / a r t i c l e s calming certainty”48. People, in other words, joined these congregations because they were systems for controlling spiritual forces in order to secure material benefits and to cope with changes both in their lifestyle and their political and economic circumstances49. Similarly, the TAC was able to appease people’s fears of witchcraft and to offer both solace from the general incertitude of life and a solution to illness, all of which were usually administered by ancestral spirits and “traditional” gods. In Grottanelli’s view, the following were all possible causes of illness according to the Nzema: moral misconduct, grudges, infringement of taboo, oaths falsely sworn, curses, witchcraft, punishment by invisible entities and deities or ancestors’ wrath. He also described the steps normally taken by a sick person in the event of real or suspected illness. First, there was the ninsinli, or herbalist, who could prescribe a potion due to his/her personal experience or to the knowledge imparted by mmotia. In serious cases, it was generally preferable to consult an oracle. The diviner was also usually a herbalist and thus able to prescribe the treatment. If the illness was very serious, affecting the community or causing death, the best solution was to consult a kɔmenle, a priest, and organise a ceremony. In the 1970s, of course, European medicine was an alternative solution, but it was not always available. A common practice was to obtain the diagnosis from a stranger, originating from the Northern Territories, from Mali or Burkina Faso, generally called ngεnlamo (which means Muslim). According to Grottanelli, however, most of the population sought assistance from the TAC, who were also commonly called the Water Carriers50. They combined practices of divination and the constant display of crosses, candles and the Bible with the use of music, dances and water, which, of course, played a central role in the Water Carriers séances. Harris’ teachings, unlike the anti-witchcraft movements, were iconoclastic and claimed that true spiritual power resided in the prophets’ preaching and in the spirit that inhabited them, eschewing any involvement with “fetishes” and assuring their followers that prayer was sufficient to keep witchcraft at bay and to cure illness51. By 1914, the colonial government was trying to control and ban several anti- witchcraft movements that had spread throughout the Gold Coast and the Asante region. Colonial authorities on several occasions forbade the anti-witchcraft

48. McCaskie, “Anti-Witchcraft Cults in Asante”, 137. 49. J. Annorbah-Sarpei, for instance, asserted that the spread of the prophetic churches re- duced the activities and popularity of Tigare shrines, one of the last and more successful anti- witchcraft movements. J. Annorbah-Sarpei, The Rise of Independent Churches in Ghana, Accra, Asempa Publishers, 1990. 50. V.L. Grottanelli, The Python Killer. Stories of Nzema Life, Chicago, University of Chicago Press, 1988, 26-29. 51. The reporter of The Gold Coast Leader wrote that his teaching and baptism converted 4,200 “idolators” and “sorcerers” “by means of open air confession and burning fetishes”. Rumours of the insanity of two “sorcerers” who opposed Harris came from the villages. The Gold Coast Leader, 18 July, 1914. Brivio, The William Wadé Harris Legacy in Ghana 61 movements, which they regarded as new and dangerous “fetishes”, and if necessary they even destroyed the shrines. Harris’ movement, on the contrary, apart from few exceptions, did not face overt colonial hostility. In January 1918, Nackabah was accused of anti-witchcraft activities and tried. The anthropologist Paul S. Breidenbach read the case and reported the statement of a witness, a school teacher, who described Nackabah’s questionable way of healing. Accord- ing to Breidenbach, Nackabah was combining his previous ritual traditional practices with the new teachings, but actually he was very close to Harris’ way of healing. It was probably the use of water that made them suspect it was another anti-witchcraft movement52. Eventually, Nackabah’s case was dismissed for lack of evidence. In those years, the accusation of anti-witchcraft activities became a useful political stick for religious rivals to beat each other with, but the two leaders always denied any sort of involvment in such activities. The spread of the “gardens” and the success of the TAC also caught the attention of both the traditional chiefs, who feared the growth of new powers that could destabilise their own authority, and the Methodist and Roman Catholic churches. These latter a few years earlier had welcomed Harris, but now began to see the danger of a general return to fetishism. Father Joseph Stauffer was one of the missionaries in charge of the Roman Catholic Church in Axim. He was in Axim during Harris’ period and did not hide his personal hostility53. More than twenty years later, his opinion of Harris’ followers had even worsened: I have myself seen, in this district, various performances of water carrying, complete with a cross covered by white material, and the congregation were singing, rolling in the dirt and dust of the village street and many almost insane with fanaticism. It was neither original nor obscene, but merely a mockery of Christianity, coupled with superstitious fears. One clerk told me that the women performing these antics were ‘seeing the Holy Ghost’54. According to Breidenbach’s informants, during the 1930s, the traditional chiefs opposed people attempting to establish a local TAC. There is no documen- tary evidence of these disputes, which were probably conducted and resolved locally, without resorting to the colonial courts. In the 1940 “District Quarterly

52. Breidenbach, “Maame”, 588. 53. Fr. Stauffer prevented Harris from preaching in the Catholic church. Probably his hostility was due to the strong rivalry with Casely Hayford and the Wesleyans. According to Stauffer, Harris had initially done a lot of good, bringing many people to the church, but he had been misled by some Wesleyans and he had began preaching polygamy and inciting people to revolt. The Wesley- ans, apparently, intended to profit from the anger of the prophet at being dismissed by Stauffer and they asked him to curse anyone who would dare to go to the Catholic church. After the fight with Stauffer, Harris immediatly left Axim for Half-Hassini and agreed to the request. However, he changed his mind the next day, claiming that the Stauffer was right and he, as a prophet, should preach in the streets and not in the churches. SMA, 3A4, Fond Heesewijk - Histoire d’Axim et Half Assinie, 1902-1961, Abstracts from the Axim Diary occasionally supplemented by Father Joseph Stauffer, 17. 54. PRAAD, CSO 18/12/57, Provincial Commissioner’s Office, Sekondi, 31 March 1940.

Africa, N.S., I / 1, 2019 62 a r t i c o l i / a r t i c l e s

Report”, on the contrary, we find the condemnation of a general revival at Atu- abo, in Eastern Nzema, where people were allegedly obliged to attend the séanc- es of the “Water Prophetess”. According to the Western Province Commissioner, the chiefs defended the prophetesses because they were a source of revenue to them and the omanhene55. Outside the Nzema area, there was more conflict. An important denuncia- tion of “fetish worship” by the TAC was sent to the DC at Sekondi in February 1940. The author of the long report was the colonial clerk A.Q. Kyiamah, who expounded upon events he had been familiar with since the 1920s. Kyiamah regarded the Church’s blend of medical and religious practices as dangerous. In his opinion, the main causes of the Harris movement’s degeneration were the lack of medical aid and educational support in the Nzema. Modernity, which would inevitably usher in the end of “fetishism”, had yet to reach the region. Kyiamah, who had witnessed Harris’ preaching in Beyin as a child, recalled the impressive involvement of the European missionaries and the ripples it had produced among the Cape Coast intellectual elite. But after a few years intellectuals and educated people had lost interest in the movement, which reverted to primitive practices. He reported the rumours about the alleged presence of a European supervisor, but, in his opinion, this was just a strategy to reassure the Church’s followers56. He mistrusted Grace Tane, who was, he felt, using Harris’ name simply to restore past “fetish” practices and “in order to earn the daily bread”. Tane had been a “fetish priestess” before Harris’ passage and she “naturally went about in her way, charging her client for her services. She continues as a fetish priestess under the guise of a Christian evangelistic body”57. Kyiamah underlined the backward nature of the movement, which lacked a progressive vision for the future of society: The movement was conducted by illiterates and backward people, most of whom are in a fit of hallucination, and nothing really progressive is included in any of their programs, the whole affair becomes one of the blind leading the blind. As a result its effect is debasing on the race indulging in it and it divorces them from lofty ideals58. Kyiamah’s report was the result of an enquiry following Tane’s petition in favour of her Church, which in turn was a reaction to a long series of disputes with the traditional chiefs and elders of Kromantse, near Cape Coast. Following the foundation of a “garden” by Michel Kojo George, the elders of Kromantse ordered the TAC delegates to stop the worship because their way of celebrating,

55. PRAAD, ADM 11/1050, District Quartely Reports (Western Province), 31 December 1930. 56. PRAAD, ADM 23/1/1025, A.Q. Kyiamah, “Formal Report to DC’s Office at Sekondi On Activities of Water Carriers”, 6 February 1940. 57. PRAAD, ADM 23/1/1025, A.Q. Kyiamah, “Formal Report to DC’s Office”. 58. PRAAD, ADM 23/1/1025, A.Q. Kyiamah, “Formal Report to DC’s Office”. Brivio, The William Wadé Harris Legacy in Ghana 63 so different from that of the other churches in the town59, disturbed the population, and caused “anomalous deaths from smallpox”60. The leaders of the TAC were accused of anti-witchcraft activity and of having seduced and abused a woman who had joined them for treatments. Upon investigation, the police ascer- tained that there was no substance to the accusations and, hence, no reason to close the church. DC McAuley informed the elders and the omanhene that every person had the right to worship so long as no breach of the peace was caused and warned the omanhene not to interfere with the right of individuals to worship. The omanhene’s emphasis on the TAC’s alleged anti-witchcraft activities was clearly an attempt to capture the attention of the colonial authorities. In Kro- mantse there were just two churches, the Methodist and the TAC and, as emphasised by the Saltpond DC, the petition was the result of jealousy on the part of the Kromantse Methodist church61. The elders of the omanhene of Kromantse did not give up. The case reached the Commissioner of the Central Province in Cape Coast, with accusations of witch-finding activities and of professing “a false doctrine not in conformity with the principles of Christianity”62. The assistant superintendent in Cape Coast, Kwesi Oppon, wrote his own report, confirming the Church’s involvement in fetish and witch-finding activities. The description of the TAC ceremonies evoked the deprecatory language normally used about “fetish” practices: The members have no special place for worshipping than in a yard owned and spared by one Kwesi Yennyi of the same town. They gather together at this place especially on the following week days i.e. Tuesdays, Fridays and Sundays, where there is a water in a big pan, and first begin to pray and sing songs the wording of which seems

59. In a letter to the DC of Saltpond, three members of the TAC argued that the persecution by local chiefs resulted from the influence of other denominations who had found that the numbers of their followers was diminishing. PRAAD (Cape Coast), ADM 23/1/797, Letter to DC Saltpond, 10 August 1940. 60. The omanhene’s accusation was that the member of the Church “worship at night and wearing white cloth” and “the dogs seeing them used to bark at them and these sort of things resulted to unusual occasions of deaths by small pox”. PRAAD (Cape Coast), ADM 23/1/1025, DC McAuley, in the Magistrate’s Court of the Gold Coast, Central Province, Magistrate’s Court, Held at Saltpond, 15 November, 1940. 61. “There have been squabbles and jealousies between the members of this church and the members of the other denominations living in the town of Kormantyne [Kromantse]. Eventually, the Regent of Lower Kormantyne intervened no doubt under the advice of the members of the Methodist and other Churches and ordered the Apostolic Church not to function in the town on the grounds that the functions of that church were not in conformity with those of the other churches in the town”. PRAAD (Cape Coast), ADM23/1/1025, Letter DC McAuley to Commissioner of Cen- tral Province, Cape Coast, 31 December 1940. In 1944, the leaders of a recently established TAC in Abandzi had to face the same problem. In a petition to the DC of Saltpond, we read: “the Ohene and his elders through the instigation of an unknown Church, have been molesting us with the result that they have finally told us not to continue the worship”. PRAAD (Cape Coast), ADM 23/1/797, Let- ter Kwamin Koom and Kofi Dadjie to DC of Saltpond, 9 August 1944. 62. PRAAD (Cape Coast), ADM 23/1/1025, Petition to Commissioner of Central Province, Cape Coast, 18 December 1940.

Africa, N.S., I / 1, 2019 64 a r t i c o l i / a r t i c l e s

molesting those who are not members, as perishable sinners. The following women who dress their heads with white linen, then carry brass pans containing part of the water put in the yard, go about in the town screaming and dancing round while others engage themselves in the ringing of calabash which is covered with beads, in a way that almost resembling Fetish Priestesses63. The women involved, Oppon added, had in the past been charged with witchcraft. In a subsequent letter to the Commissioner of the Central Province, the elder of Kromantse demanded the intervention of the colonial authorities because they had not been able to “stop the members of the Church from carrying on their evil designs against the people who opposed this Church”64. Then they invited the police to monitor the Friday night meetings. In April, McAuley wrote another, less favourable, report, noting that, in fact, the meetings were held on Friday, “which is the general day for the fetish worship along the Gold Coast”65. He con- cluded by admitting that the Church’s practices looked like fetish worship; at the same time, he stressed that “there seems little real cause for intervention”66. Grace Tane offered her support and encouragement to Michael Kodjo George and made several trips throughout the period of the dispute67. From her point of view, the problem was ignorance on the part of the members of the Native Tribu- nals: “a section of bigoted Africans particularly Councillors of Native Tribunals, deliberately and without just cause obstruct them in the discharge of their duties, and occasionally initiate Criminal Proceedings against them before Native Tribu- nals”. In brief, the petition stressed that the Church was a legitimate Christian institution present all over the colony. She underlined its connection with Harris, quot- ing Casely Hayford’s words to remind her readers that Harris had been supported by the Methodist elite in the Gold Coast and by the Christian missionaries in the Ivory Coast. On the contrary, she stated, present-day Christian churches and the traditional authorities blindly obstructed her Church. She also tried to make sense of the persecutions suffered by her Church by contextualising them historically: “Your humble petitioner is informed and verily believes that the History of Religion is characterised by unjust and unnecessary persecutions; but the British Justice has always favoured toleration of Worship and freedom of speech”. By way of conclusion, she reiterated that her followers “do not claim to be possessed by powers to exorcise witches, nor to detect witchcraft, neither do they practice Black Magic nor indulge in mysterious practices”. On the contrary, the Church “knows no bounda-

63. PRAAD (Cape Coast), ADM 23/1/1025, Kwesi Oppong, 9 Juanuary 1941. 64. PRAAD (Cape Coast), ADM 23/1/1025, Petition to Commissioner of Central Province, Cape Coast, 20 December 1941. 65. PRAAD (Cape Coast), ADM 11/1/1050, DC McAuley to Commissioner of Central Pro- vince, Cape Coast, 21 April 1941. 66. Three years later, the leaders of the TAC in Abanze, Saltpond district, reminded critics that it was due to the assistance of the “late Mr McAuley” if the Church in Kromantse still existed, and if “those who raised objection are staunch members of the Church”, PRAAD (Cape Coast), ADM 23/1/797, Kwamin Koom and Kofi Dadjie to DC of Saltpond, 9 August 1944. 67. Breidenbach, “Maame”, 601. Brivio, The William Wadé Harris Legacy in Ghana 65 ries of race or tribe and believes in God’s Mercy, and the beneficial influences of His Holy word”68. Tane was clear about the purpose and doctrine of her Church, which, in keeping with Harris’ preaching, excluded any connections between illness and evil spirits and witchcraft, foregrounding instead the cleansing power of the Holy Spirit. Suspicions about the doings of the Church did not abate, however. In 1949, for instance, following a letter in which Michael Kodjo George had once more stressed the purely religious nature of his organisation, the colonial officer in Cape Coast suggested that the Church had “a distinct Tigare fetish flavour”69. Notwithstanding the recurrent petitions against the TAC, Tane, Nackabah and George had been brilliant religious entrepreneurs, able to interpret, innovate and organise Harris’ original movement. At the end of the 1930s, the leaders of the TAC started considering the economic benefits of receiving direct aid from European missionaries. Their first contacts with Pentecostal churches took place in 1938. James McKeown, the leader of a Welsh and British Pentecostal group in Accra and founder of the Apostolic Church, one of the first Pentecostal churches in the Gold Coast, met some of Nackabah’s young associates in Accra and proposed working together70. The three elders organised a meeting with McKeown, who proposed transforming the movement into a branch of the Apostolic Church. The encounter led to a temporary schism within the TAC, with Tane on one side and Nackabah on the other. Initially, Nackabah decided to work with McKeown, but some months later he realised that the ideas of the white man were inconsist- ent with the ritual practices and tradition of ownership established prior to his arrival. To rid himself of McKeown, Nackabah had to resort to litigation. He subsequently re-joined Tane, but after his death, the situation changed dramatically and in 1958 the Church faced a second and definitive schism71. Although disastrous, the contact with the Pentecostal church influenced the future of the TAC. After the death of Nackabah in October of 1947, John Kojoe Baidoe was elected as the Church’s new leader with the name John Nackabah II. In 1954, Nackabah II wrote to Prime Minister Kwame Nkrumah, asking the government to recognise his “Water healing society” (the society was part of the

68. PRAAD, CSO 18/12/57, Petition of Madam Harris Grace Thannie to Governor, 17 Janua- ry 1940. 69. PRAAD (Cape Coast), ADM 23/1/3817, Letter Micah Kodjo George to Colonial Secreta- ry, 21 November 1949. Tigare was an important anti-witchcraft movement widespread during the colonial period. 70. In 1937 McKeown arrived in Accra to support Apostle Anim, who upon receipt of the magazine from the Faith Tabernacle Church in 1917 had begun preaching and healing in Christ. A new movement began then. They worked together for several years and then split. Anim named his group “Christ Apostolic Church,” while McKeown’s group remained known as “The Apostolic Church”. The Apostolic Church grew faster but further schism also led to the establishment of “The Church of Pentecost” and “The Assemblies of God”. The four churches were the main Pentecostal churches in Ghana until the 1970s. 71. Breidenbach, “Maame”, 604-606.

Africa, N.S., I / 1, 2019 66 a r t i c o l i / a r t i c l e s

TAC) and to open hospitals and clinics to the prophets belonging to his society. The rise of Nackabah II coincided with the adoption of a new stance on witchcraft. Nackabah II now openly listed anti-witchcraft activities amongst the skills offered by the Church. Talking about himself, Nackabah II wrote: “he was able also to heal with water all manner persons possessing evil-spirits, witchcrafts, and also able by Divine relieve women who would be pregnant and unable to deliver during child-birth; also could be able to clear any insanity with the performance of the prayer of Water Healing, in all such cases having proved successful”72. According to the Government agent it was not a good idea for the Government “to admit the existence of “evil-spirits and witchcraft” by conferring official recognition on this sect, because, he explained, “the object of (the) petition (which is nowhere expressed) was to appeal to Government for funds to build a hospital in which lepers, witches, the insane and the demoniacally possessed could be treated free of charge by faith and immersion”, whereas Govern- ment funds were “required to finance its own medical system”73. Nackabah II re-wrote the history of the movement. The role of Harris was, of course, acknowledged, but the emphasis was very much on the founder’s “won- ders”. According to Nackabah II, the latter had been able to “cast out persons possessed of evil spirits to go out from them, and those practising poisonous medicines to injure men’s lives, and witchcrafts”74. Tane was no longer mentioned and Harris’ legacy belonged exclusively to Nackabah75. Times had changed and the dialogue with the Pentecostal churches – that put deliverance from bad spirits, “fetishes”, and witchcraft at the centre of their discourses – had clearly influenced the society. The new leaders no longer felt compelled to defend themselves from allegations of anti-witchcraft activities. The Church had become an established institution: its 1950 Annual General Convention, held at Lagoon Park, Saltpond, attracted over 7,000 members from the Colony, Ashanti and Togoland.

Conclusion

Discourses on evil spirits, anti-witchcraft and alleged fetishism took on crucial, ambiguous roles in successive debates about the legitimacy of the TAC. For Harris, who remained close to the Methodist Church throughout his life, the fight against witchcraft had never been a central concern. Witchcraft did exist, but faith in the Supreme God would erase the problem with no need for direct intervention. The baptismal water was sufficient to cleanse the converts once and for all. How-

72. PRAAD (Cape Coast), ADM 23/1/3817, Petition John Nackabah II to Prime Minister Kwami Nkrumah, 22 July 1954.. 73. PRAAD (Cape Coast), ADM 23/1/3817, Letter Government Agent’s Office to Western Regional Office, 25th August, 1954. 74. PRAAD (Cape Coast), ADM 23/1/3817, Petition 1954. 75. PRAAD (Cape Coast), ADM 23/1/3817, Petition 1954. Brivio, The William Wadé Harris Legacy in Ghana 67 ever, his followers and the first leader of the TAC were very soon forced to re- spond to the accusation of practising anti-witchcraft activities. In his report, Kyiamah made the accusation that the popular appeal of the movement was linked to its immediate relevance to concerns such as witchcraft and childbearing. The anthropologist Ernesta Cerulli expressed a similar point of view in the 1970s, as- serting that one of the reasons for Harris’ success was his capacity to “deliver the population from witchcraft, an atavistic terror that dominated their existence”76. While speaking to people’s fears, however, Grace Tane was also able to keep Harris’ converts within the Christian fold and take useful steps towards integra- tion within the colonial system. She maintained a theological position close to Harris teachings, and even if both Tane and Nackabah emphasised the healing dimension of their ministry, they firmly denied that this could be confused with anti-witchcraft activities. The institutionalisation of the TAC was probably encouraged by the contacts with the Pentecostal churches but this was a trend also adopted by some of the anti-witchcraft movements77. From a spiritual point of view, the Pentecostal discourse had introduced a new notion of evil that permitted open talk about “bad spirits” and witchcraft to resurface once again. The early Pentecostal churches, such as the Apostolic Church, challenged mainline Christianity on the issue of the perception of evil and the way to combat witchcraft. Evil spirits, including witchcraft and a new image of the devil, ceased to be synonymous only with traditional and backward cults, and played an important role in the process of conversion to Christianity78. Following Tane and Nackabah’s tenure, second-generation leaders adopted that same spiritual stand. Over the years, Harris’ movement had been marginalised by the Roman Catholic, Methodist and Anglican churches, all of which maintained rigid barriers against contamination by “occult forces”; in the 1950s, the TAC changed attitude, openly vindicating anti-witchcraft practices. In more recent years, the increasing presence of African Neo-Pentecostal churches seems to have totally overtaken AICs. Ghanaian Neo-Pentecostalism advocates a total break with the ancestral spirits, accused of being expressions of Satan, but it also teaches that deliverance from evil entities – the central focus of the liturgy – can only be achieved by remembering the past. Thus, total amnesia about the traditional past is avoided79. The current marginalisation of the TAC is basically the effect of a general shortage of financial resources and limited gen-

76. Cerulli, “I Water Carriers”, 116, traslation by the author. 77. See: J. Allman and J. Parker, Tongnaab: The History of a West African God, Bloomington, Indiana University Press, 2005. 78. See: B. Meyer, Translating the Devil: Religion and Modernity Among the Ewe in Ghana, Edinburgh, Edinburgh University Press, 1999. 79. D. Maxwell, “Delivered From the Spirit of Poverty? Pentecostalism, Prosperity, and Mo- dernity in Zimbabwe”, Newspaper of Religion in Africa, 28, 3, 1998, 350-73; R. van Dijk, “Pente- costalism, Cultural Memory, and the State: Contested Representations of Time in Postcolonial Ma- lawi”, in R. Werbner, ed., Memory and the Postcolony: African Anthropology and the Critique of Power, London, Zed Books, 1998, 155-181.

Africa, N.S., I / 1, 2019 68 a r t i c o l i / a r t i c l e s erational turnover. However, for the moment, the TAC is far from disappearing from the Ghanaian religious scene80. Both the Pentecostal churches and the TAC claimed that illness, lack of success and misfortune all depend on “evils”, believed to operate in the guise of traditional spirits. They both trust that preaching the Holy Spirit can defeat them, but, as Andrew stated, to destroy one’s enemy, it is necessary to know it. An- drew’s return home, his move from the Pentecostal to the TAC and his interest in “traditional” spirits and traditional medicines all indicate his changeable nature and propensity towards religious innovation but they also suggest that every rupture with the past demands a subsequent realignment. In the summer of 2018 I met Esther in Axim. She was hosted by the local TAC Church, where the ɛsofo was treating her. The “spirits” who bothered her, taking her to the forest and inducing insane behaviours, were the cause of her illness; she was a member of the Christ Church but decided to be healed by the TAC, because they were able to deal with “ghosts”. How each rupture is concep- tualised and understood hinges not only on discourse but also on the decisions people make in everyday life. Andrew decided for reasons of economic opportunity and spiritual interest to go back to the past, while Esther – aware of the existence of invisible spirits dwelling in nature – chose the more suitable solution to ease her suffering. Beneath the surface, the boundaries between different religious expressions are not so sharp, and both Harris’ converts and the Neo-Pente- costal followers, although they remained within Christianity, still negotiated among several religious alternatives to find the best solutions to their personal problems and to those of the society at large.

80. In the Nzema region, all the TAC followers recognised the leadership of Grace Tane and of her descendants, Grace Tane II and Grace Tane III. They all remembered the name of the first prophet, William Wadé Harris, “the man who married Grace Tane”, as well as the name of Nackabah and of Michael George. One old prophetess I met in Ellohin also vaguely remembered that a white man had joined the Church as a prophet (Interview with Eti Dabla, Ellohin, Ghana, 5 august 2012). In the Central Region, the spiritual leader was Michael George, and later his successors Michael George I and Michael George II. Here the “gardens” tended to assume a more independent stand, each taking over the role as a Church. The new tendency is to adopt a professional Pastor to support the prophetess. Mariano Pavanello

Pa w n s h i p a n d Do m e s t i c Sl a v e r y in Ch i e f t a i n c y Di s p u t e s (Nz e m a Ar e a , SW Gh a n a )

Abstract

Stool and land disputes are important issues in the life of Ghanaian chieftain- cies, and the written records of the arbitrations carried out by the customary institutions offer rich material about Traditional Areas and the local ruling families and their internal structure of ‘royal’ and ‘plebeian’ matrilines. Litigation between families competing for a stool have been going on for generations, and the records show that the correct knowledge of the stool’s history is a crucial mark of legitimacy. Through the analysis of a case concerning the stool of Bonyere, We- stern Nzema Traditional Area (Jomoro District, Western Region), the author aims to show how the internal structure of families may be revealed during litigation through the narration of historical tradition, and what this source can reveal about broader social and political dynamics. k e y w o r d s : a k a n , n z e m a , chieftaincy, p a w n s h i p , s l a v e r y

Preface

In this paper, I want to analyze the Nzema1 traditional discourse on pawnship and slavery as it emerges from the proceedings of the stool disputes. Inter alia, these documents illuminate the relationship between would-be legitimate and illegitimate descent lines of the matrilineages entitled to the ownership of stools2.

1. The Nzema, a matrilineal Akan population, live in the extreme southwestern coast of Ghana, between the rivers Tano in the West and Ankobra in the East. The territory is divided in two Traditional Areas, Eastern and Western, corresponding to the Elemgbenle and Jomoro districts of the Western Re- gion, respectively. Some Nzema communities are historically settled in Ivory Coast (Canton Aduvlé; Grand Bassam; Abidjan). A basic bibliography on the Nzema would include V. L. Grottanelli, Una società guineana: gli Nzema, 2 vols, Torino, Boringhieri, 1977-78; M. Pavanello, Il formicaleone e la rana, Napoli, Liguori, 2000; Id., La papaye empoisonnée: essais sur la société Akan des Nzema, Saarbrücken, Éditions Universitaires Européennes, 2017; P. Valsecchi, Power and State Formation in West Africa. Appolonia from the Sixteenth to the Eighteenth Century, Basingstoke, N. Y., Palgrave Macmillan, 2011. The term Akan refers to a group of societies sharing a stock of cultural features and a common linguistic phylum settled between the rivers Comoë in Ivory Coast and Volta in Ghana. 2. The stool (Nzema: ebia, pl. mbia; Twi: akonua) is the most important symbol of Akan power. Stools are hierarchically organized according to precolonial tradition and to the modern

Africa, N.S., I / 1, 2019 70 a r t i c o l i / a r t i c l e s

These families consist of several matrilines3. Some of these do not belong to the original stock of the lineage and are collectively known as koahweabane (plebeians), a collective plural mainly referring to matrilines of slave origin (ahanra, pl. of kanra, ‘slave’, purchased person in the northern markets)4 as well as to descent lines of pawn origin (awoba)5. This is the normal structure of the Akan matriclans or matrilineages6. The Nzema area was a carrefour of the slave trade in the eighteenth and early nineteenth century, and Nzema slave merchants were well known on the Western Gold Coast7. The legitimate8 matrilines of a lineage consider themselves as the suakunlu abusua9, whereas the category of asalo abusua10 denotes the external lines of

Ghanaian constitutional arrangement (see chapter 22 “Chieftaincy” of the 1992 Constitution of the Republic of Ghana; and the Chieftaincy Act 2008, n. 759). The highest level is the paramount stool (Nzema: ebia kpolɛ, big stool), which is vested with the primary allodial title on the land of a Tra- ditional Area (in most cases a former precolonial state or chiefdom). Under the paramount chiefs, subordinate stools hold derivative allodial titles to the land of towns or villages. Out of the vast literature on Ghanaian chieftaincy, see especially K. Arhin, Traditional Rule in Ghana, Past and Present, Accra, Sedco, 1985; Id., Transformations in Traditional Rule in Ghana, 1951-1966, Accra, Sedco, 2002; K. Boafo-Arthur, “Chieftaincy in Ghana: Challenges and Prospects in the 21st Cen- tury”, African and Asian Studies, 2, 2, 2003, 125-153; A. K. P. Kludze, Chieftaincy in Ghana, Lanham, Austin and Winfield, 2000; I. K. Odotei, and A. K. Awedoba eds., Chieftaincy in Ghana. Culture, Governance and Development, Accra, Sub-Saharan, 2006; I. Owusu-Mensah, “Politics, Chieftaincy and Customary Law in Ghana’s Fourth Republic”, The Journal of Pan African Studies, 6, 7, 2014, 261-278. For the land tenure system in Ghana, see K. Bentsi-Enchill, Ghana Land Law. An Exposition, Analysis and Critique, London, Sweet and Maxwell, 1964; N. A. Ollennu, Princi- ples of Customary Land Law in Ghana, London, Sweet and Maxwell, 1962. 3. I translate as ‘matriline’ the local concept of bɛnla akunlu, literally “their grandmother’s womb”. 4. Anyi kanga; Twi ɔdɔnkɔ (pl. nnɔnkɔ or nnɔnkɔfoɔ). War prisoners (alesedeɛ) were a further category of dependent people, clearly distinguished from the ahanra. A Twi term of general use is gyaasefoɔ, the people of the gyaase (literally, the hearth), the house of a landlord, including his children, clients, pawns, servants and slaves. Cf. e.g. R. S. Rattray, Ashanti Law and Constitution, Oxford, Clarendon Press, 1929, 47-55; C.-H. Perrot, Les Anyi - Ndenye et le pouvoir aux 18e et 19e siècles, Paris, Publications de la Sorbonne, 1982, 157ff. For a discussion of some analogies and differences between kanra and awoba, see the Conclusion of the present paper. 5. Anyi aoba, Twi awowa. 6. See F. J. Amon d’Aby, Croyances religieuses et coutumes juridiques des Agni de la Côte d’Ivoire, Paris, Larose, 1960, 129; A. A. Perbi, A History of Indigenous Slavery in Ghana from the 15th to the 19th Century, Accra, Sub-Saharan, 2004; Rattray, Ashanti Law, 33-55. 7. Cf. Perrot, Les Anyi - Ndenye, 96, 160 n. 5. 8. In Nzema: dehelɛ, pl. alehelɛ (in Twi ɔdehyeɛ, pl. adehyeɛ), a term translated into English as “royal”. 9. The family (abusua) of the king’s palace (suakunlu, lit. the ‘womb of the town’). Used throughout the Akan area, the term Abusua (pl. mmusua) denotes the “matrilineal extended family”, both as matriclan and matrilineage. Every Nzema lineage belongs to one of seven clans. For a synthe- sis of the old debate on to the etymology and exact meaning of the term abusua, particularly focusing on the anthropological distinction between clan and lineage, see M. Pavanello, “Clan, lignage et mariage en pays Nzema: une reconsidération”, Journal des Africanistes, 75, 1, 2005, 209-232. 10. The family of the external hall (asalo) of the house. The inclusion or adoption of a subordinate line into a lineage was a frequent practice through the marriage or concubinage of a free man Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 71 servile or pawned origin. The asalo matrilines are normally excluded from inheritance but they often interact with the legitimate ones in succession disputes for the stools. A Nzema proverb says, “the ant-lion and the frog do not celebrate a common festival”11, which means that the plebeian matrilines do not share a common inheritance (agya) with the legitimate ones. However, even in the recent past, the matrilateral cross-cousin marriage12 with a mother’s brother’s kanra daughter (i. e. a daughter of a maternal uncle and a woman of slave or pawn origin) was very frequent, showing the continuous intermarriage between legitimate and illegitimate matrilines of the same lineage13. In such a case, a legitimate matriline may maintain a position as ‘father’ vis-à-vis a subordinate matriline, re- newing and strengthening a ‘father-child’ relation at any generation. The oath sealing the secret, which binds within one family ‘royals’ and their ‘plebeians’, is in principle never broken, as attested by another proverb: “nobody knows the secret between the frog and the ant-lion”14. Nevertheless, the parties in conflict often disclose their secrets.

The records of the chieftaincy disputes

This work is based on a type of written source generated by verbal events: the handwritten records in English of the proceedings of chieftaincy disputes, arbitrated by Ghana’s customary institutions15. Although anthropologists and historians have seldom paid attention to these primary sources, they constitute a rich heritage with a purchased or pawned woman, and the inclusion of the offspring in their father’s matrilineage. See my interview with Kwame Nyamekɛ, abusua kpanyinli of the Bawia stool in Western Nzema, 12 August 2002: “A man went to the North and bought a woman. That woman was his daughter (ra) and was brought home and was called kanra. The man could marry that woman as suanu agyalɛ, and the children were asalo abusua” (M. Pavanello, “Reconsidering Ivor Wilks’s ‘Big Bang’ The- ory of Akan History”, Ghana Studies, 14, 2011, 21). 11. Futukɔnyima nee kɛlɛne bɛngulu nu bɛnli ɛvoyia. See G. B. Kwesi, Nzema Mrɛlɛbulɛ. Nzema Proverbs, edited and with an introduction by M. Pavanello, Legon, Institute of African Stud- ies, University of Ghana, 2007. 12. This is the case of a male Ego’s marriage with his mother’s brother’s daughter (MBD), who is considered Ego’s daughter in the Crow kinship terminology system of the Nzema and the Fanti. In other Akan societies, like the Asante, the MBD is considered as a sister. In all cases, the MBD is marriageable. 13. Cf. Pavanello, “Reconsidering Ivor Wilks”. 14. Awie ɛnze kɛlɛne nee futukɔnyima bɛ avinli. 15. For the chieftaincy’s judicial system and its transformations from colonial to postcolonial times, see S. A. Brobbey, The Law of Chieftaincy in Ghana, Accra, Wrenco Ltd., 2008; A.S. Anam- zoya, “Neither Fish nor Fowl: An Analysis of Status Ambiguity of the Houses of Chiefs in Ghana”, The Journal of Legal Pluralism and Unofficial Law, 46, 2, 2014, 218-234; R. Gocking, “British Justice and the Native Tribunals of the Southern Gold Coast Colony”, The Journal of African His- tory, 34, 1, 1993, 93-113; A. K. P. Kludze, “Chieftaincy Jurisdiction and the Muddle of Constitu- tional Interpretation in Ghana”, Journal of African Law, 42, 2000, 37-63; R. Rathbone, “Native Courts, Local Courts, Chieftaincy and the CPP in Ghana in the 1950s”, Journal of African Cultural Studies, 13, 1, 2000, 125-139.

Africa, N.S., I / 1, 2019 72 a r t i c o l i / a r t i c l e s and are of great interest for the history of Ghanaian Traditional Areas and the study of local political contexts16. Any attempt to write “African history from an African perspective”17 should not neglect documents of this kind. Moreover, these records provide researchers with a vivid picture of living social relations, behaviors and discourses. The gradual bureaucratization of chieftaincy in the colonial Gold Coast and post-colonial Ghana18, while probably increasing internal conflicts, nonetheless provided the legal instruments for regulating and managing disputes, as well as for the recording of their proceedings. According to the Chieftaincy Act 759 of 2008, section 29 (1), Traditional Councils are the first level of the Ghanaian customary judicial system for the arbitration of chieftaincy disputes. The Traditional Councils’ recorders are local people with a bilingual training, and the English records under discussion are the product of a rapid process of simultaneous translation and abridgement from the local language – in our case Nzema and, occasionally, Fanti Twi. Therefore, making sense of the gap between the actual verbal events and their records is a formidable challenge for researchers, even when they are acquainted with the local language. From a linguistic point of view, the records often betray Nzema syntax underlying a flux of English words. The need for concision sometimes creates a fuzzy text, due to the difficulty of translating the peculiarities of local discourse. In addition, the recorders frequently translate concepts pertaining to the spheres of social structure, magic, witchcraft and religion with English or pidgin idioms, or colonial stereot- ypes, thereby hiding the complexity of local categories, or even misrepresenting original meanings. For example, the local concept of amonle is translated as juju, a widespread and generic pidgin term for ‘sorcery’. Yet amonle is a complex notion, ranging from ‘ritual curse’ or ‘oath to the gods’ to ‘magic object’. Moreover, the recorders constantly use the term ‘tribe’ to denote the abusua, clan or lineage, and the expression ‘tribal war’ for the Nzema civil war (1868-1871), showing the cultural influence of British colonial discourse19.

16. An important exception concerns the records of the Asante Court (Manhyia Archives, see M. Adams, “A Guide to the Historical Records of the Asante people of Ghana at the Manhyia Ar- chives”, African Research and Documentation, 101, 2006, 9-25). A specific contribution on the management of chieftaincy records from the point of view of information sciences is M. Adams, “The Management of Chieftaincy Records in Ghana: An Overview”, African Journal of Librarian, Archival and Information Sciences, 15, 1, 2005, 67-73. Specific reference to the “records of some of the ‘Native Tribunals’ of the main coastal towns”, contained «in the Cape Coast branch of the National Archives», is made by R. Gocking in C. Dickerman et al., “Court Records in Africana Research”, History in Africa, 17, 1990, 305-318. 17. See the ample historical review article on this issue by J. K. Adjaye “Perspectives on Fifty Years of Ghanaian Historiography”, History in Africa, 35, 2008, 1-24, which does not even mention the chieftaincy records. 18. The Native Jurisdiction Ordinance of 1878, and its subsequent amendments, established the statutory duties of the former State Councils (see Arhin, Traditional Rule, 90 ff.; see also R. Addo-Fening, “The Native Jurisdiction Ordinance, Indirect Rule and the Subject’s Well-Being: the Abuakwa Experience, c. 1899-1912”, Research Review, n. s., 6, 2, 1990, 29-43). 19. In many colonial documents and early ethnographic or historical works written by colo- Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 73

The Bonyere stool dispute

The records examined in this article consist of cases arbitrated by the West- ern Nzema Traditional Council (formerly State Council), at Beyin, and pertaining to the Bonyere stool dispute, an extended piece of litigation encompassing various cases brought before the chieftaincy’s judicial institutions and the appeal Commissions or superior magistrate Courts from 1940 to the present20. I selected this case for its historical significance, and because the records, despite some la- cunas, provide deep insights into the internal relations and structures of the two lineages – belonging to the Adahonle and Mafolɛ clan, respectively – involved in the litigation. The dispute has its roots in the civil war (1868-1871),21 when the town was destroyed because of its betrayal in favor of the Eastern Nzema chief, Avo. The war originated from the split of the old Nzema kingdom in the mid- nineteenth century, following the capture and death of King Kaku Aka22. Two former captains of the deposed king started ruling the two chiefdoms, the Eastern, whose capital was the town of Atuabo, and the Western, centred on Beyin. In 1867, an agreement between the British and the Dutch gave the latter control over the former British possessions west of the Sweet River (between Cape Coast and Elmina)23. Initially, the two Nzema chiefdoms accepted the Dutch flag, but due to old disagreements, hostilities came to the fore again, and the Eastern Nzema gave up the Dutch flag. A war eventually broke out (ca. 1868) between Amakyi24, the

nial officials the large kinship groupings, generally labeled as clans in later scientific literature, are often referred to as ‘tribes’. In Nzema, war is konle, and the civil war is generally referred to as konle kpolɛ, big war, or Amakyi Avo konle (as in J. A. Essuah, Mekakye bie, 3 vols, Cape Coast, Catholic Mission Press 1958, vol. 3, 169). Therefore, ‘tribal war’ is a colonial expression nega- tively referring to African or ‘primitive’ wars. 20. I base my analysis only on the cases dating back to more than fifty years ago (Book of Records of WNTC, Beyin, 1948-49 case, pp. 49-125; 1951-52 case, pp. 127-152; 1954-55 case, pp. 262-273; 1955 case, pp. 324-326; 1959 case, pp. 433-437; 1960-61 case, pp. 464-468; 1963-64 case, pp. 574-629). 21. The Nzema civil war was part of the general regional turmoil that culminated in the Anglo- Asante war of 1871-74. The tales recorded in the proceedings mention the alliance of the Eastern Nzema with the Wassa Amanfi, and the support given to the Western Nzema by the Sannvin and eventually by the Asante. 22. King Kaku Aka (who reigned ca. 1832-1849) was captured by the British in 1849 and died in Cape Coast Castle in 1851; see P. Valsecchi, “The Fall of Kaku Aka: Social and Political Change in the Mid-nineteenth Century Western Gold Coast”, Journal of West African History, 2, 1, 2016, 1-25. 23. “In 1867 came the fatuous arrangement by which the Dutch took over all the forts west of the Sweet River”. (R. W. Sanderson, “The History of Nzima up to 1874”, Gold Coast Review, vol. 1, Jun.-Dec. 1925, 103; J. Y. Ackah, Kaku Ackah and the Split of Nzema, MA Thesis, Legon, Uni- versity of Ghana, Institute of African Studies, 1965, 160, 163). 24. Amakyi, the ruler of Western Nzema after the split, reigned ca. 1849-1874. Ezoa Kpa- nyinli and Nyanzu Aka II succeeded him. See Ackah, Kaku Ackah; Crowther’s Report on Native Affairs of Appolonia, rif. ADM/11/1767, Ghana National Archives, Accra.

Africa, N.S., I / 1, 2019 74 a r t i c o l i / a r t i c l e s western ruler, and Avo, the eastern chief25. Nevertheless, in neither case did all of the towns accept to join their respective leaders. Thus, some western towns (Adusuazo, Awiane and Bonyere) supported Atuabo, while some eastern ones (Asanta, Nkroful and Sanzule) were in favor of Beyin26. From the proceedings of the cases it is possible to ascertain that the Ada- honle were the original owners of the Bonyere stool, and that the person responsible for the treachery was a caretaker, Annor Eza or Azan, belonging to an Ada- honle matriline of pawn origin. At the end of the civil war, King Amakyi wanted to rebuild the town, which had been burnt down. Since the legitimate members of the Adahonle family failed to claim their property, a new chief (or caretaker), Azan Ndoli, belonging to the Mafolɛ family, was appointed by the king. This man was the one who reported the treachery allowing King Amakyi to seek refuge at Krinjabo. Azan Ndoli was a Mafolɛ ‘son’ of the Adahonle lineage, but it is not possible to ascertain whether he was a son of a legitimate or an illegitimate matriline. The two families share a long history of intermarriage of the cross-cousin types in the past generations, and both adopted slave or pawn matrilines, which makes it very difficult to disentangle the knots of the genealogical successions.

Bonyere: oral traditions and (western) history

Bonyere is a populous coastal settlement about five miles west of Cape Apol- lonia, and twenty miles east of the mouth of the Bia-Tano lagoon system. One oral tradition holds that it was founded at the time of the Aowin king Anɔ Asɛma, at the end of the seventeenth century27; another has it that King Dehelɛ Kwesi founded Bonyere at the end of the eighteenth century28. A third oral tradition,

25. Avo succeeded Anɔ (Anaw), his supposed senior brother, who had succeeded Ɛbanyenle, the first ruler at Atuabo. It is not possible here to summarize the debate about the origins and clan affiliation of the successors of Kaku Aka. Apart from Ɛbanyenle, whose origins and clanship are still controversial, both present dynasties owning the paramount stools of Atuabo and Beyin claim descent from previous rulers belonging to the Ndweafoɔ clan (the dog clan) (see references in previous note). 26. “In Nzema today, all the towns which fought for Beyin are called kanga (the native word for Dutch), while those that supported Atuabo are nrinlinza (English). Thus Asanta, Nkroful and Sanzule in Eastern Nzema are known as kanga because they sided with Beyin, while Bonyere in Western Nzema is nrinlinza because it allied with Atuabo” (Ackah, Kaku Ackah, 160). 27. Anɔ Asɛma is described by Valsecchi, Power and State Formation, 2011, 124, 197, as the «Aowin chief and military leader» who conducted an expansionist campaign towards the coast of Quaqua in the last decade of seventeenth century. Moreover, he includes an Anɔ Asɛma in the Nvavile-Anɔna-Ɔyoko genealogy bound in a chain of marriage alliances with a supposed line of Ndweafoɔ originating from Kɛma Kpanyinli. Anɔ Asɛma is also a quasi-mythical figure of the oral tradition of the Anyi Ndenye (see C.-H. Perrot, “Ano Asemã: Mythe et Histoire”, The Journal of African History, 15, 2, 1974, 199-222; Id., Les Anyi – Ndenye). 28. Ackah, Kaku Ackah, Appendix 5, p. 4 and 6, p. 9, reports two lists of Nzema kings in which a Dikileh Kwesi or Dehele Koasi appears as a successor to Amihere Kpanyinli (who was Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 75 accepted by both families, and often recalled in the course of the dispute, has it that Bonyere was founded during the reign of King Nyanzu Aka I (1816-1832)29. This tale, however, refers most probably to the village of Akasuazo, named after Aka [Aka kpanyinli], Nyanzu Aka’s son. This village was the place of the king’s farm laborers (slaves) who supplied the Bonyere court. The catholic bishop, J. A. Essuah, in his Mekakye bie, reports another widespread tradition: [During the war between king Kaku Aka and the Sannvin, ca. 1840] one man called Gyebili set up a war trap from the beach to the forest. The side of the trap on the beach was very close to the River Kɛdɛ that meets the Domunli River. The village community settled on the beach in order to guard the trap was named Gyebili Ɛhanezo, meaning the community of Gyebili’s trap. There was a newly established community [sua fofolɛ], near Ezinlibo, named Akasuazo. The first man to live there was Aka kpanyinli. Tradition has it that one of the nephews of King Kaku Aka went to Akasuazo and asked Aka Kpanyinli how his village was doing. He answered “ɔle bɔ nyεlε” meaning the community was progressing [because] soon his village and that of Gyebili would become one big town, renamed as Bɔnyɛlɛ.30 It is likely that the above traditions of Akasuazo and Gyebili Ɛhanezo derive from family histories concerning the foundation of different small communities on the coastline between Ezinlibo (East of Bonyere) and Kabenlasuazo (Domun- li lagoon, West of Bonyere) that subsequently merged with the main centre. We can infer that the tales narrated in the judicial hearings concerning the origin of the two families refer to events relating to the period from the foundation of Aka- suazo under the reign of Nyanzu Aka I (1816-1832) to the Sannvin war in the 1840s under the reign of Kaku Aka (1832-1849). A further element confirming this hypothesis comes from a statement by the Mafolɛ plaintiff in the 1948-49 case. Among the founders of Bonyere – the Mafolɛ plaintiff asserted – was a grandfather of his who had built the war trap at Gyebili Ɛhanezo 31. On the other hand, it is likely that the conflicting traditions of Anɔ Asɛma and Dehelɛ Kwesi refer to a seventeenth- and eighteenth-century historical framework, when the polities to the West of the Cape Three Points underwent a thorough readjustment. Yet, Dutch documents and maps show that Bonyere is much more ancient than the oral tales imply. The town is mentioned as Bongere in documents concerning probably ruling from the 1750s to the 1770s; see R. Baesjou, “Trade Conflicts in Eighteen-century Western Gold Coast, and the Formation of the Nzema State”, in M. Pavanello ed., Prospettive di studi Akan. Saggi in memoria di Vinigi L. Grottanelli, Roma, Cisu, 1998, 23-54). Ackah, Kaku Ackah, 78, gives a genealogical table of the kings of Nzema from Amihere Kpanyinli to Yan Sack- ah (Yenzu Ackah [Nyanzu Aka]): Amihere Kpanyinli (died in the 1770s); Kwesi (probably the same as Dehelɛ Kwesi, died in 1801); Ezoa Kyi (1801-1816); Nyanzu Aka (1816-1832). 29. See Ackah, Kaku Ackah, 30. 30. Essuah, Mekakye bie, 1958, vol. 3, 168. 31. In the case of the Bonyere stool, 31 December 1948 – From cross-examination by defendant Awonyi Nwia to plaintiff Amihere Agyilihah: Question by Awonyi Nwia (Adahonle), “What is the name of your grandfather among people founded old Bonyere?” Answer by Amihere Agyilihah (Mafolɛ), “His name is Erhu, the one built Gyabiri hanza, the ancient war trap at Bonyere”.

Africa, N.S., I / 1, 2019 76 a r t i c o l i / a r t i c l e s the vassalage treaties that the towns of Jumore, the political area west of Cape Apollonia, signed with the Dutch WIC (West India Company) officer, J. Valcken- burgh, during his visit in 1652 and later in 165732. Bonyere, described as “the principal town of Jumore”33, was marked as Iumore in a WIC map of 165034 and as Jumore in a map by Joannes van Keulen (1704-1755)35; it was eventually marked as Bonnaree in a Dutch map drawn before 181236.

Historical tradition as legal evidence and its transformations

The records include narrations of stool or family histories given as evidence by the authorized representatives of the contending families or their witnesses. The opening words of most historical tales are tete edwɛkɛ kile kɛ, ‘the ancient history shows that’. Tete is iterative of te, to hear, which also means ‘to understand’; therefore, tete edwɛkɛ literally means ‘the heard and heard tale’. The object of the hearing and understanding is the discourse, edwɛkɛ, of the ancestors, the real owners of the history, which, by definition, is the truth37. Therefore, the ancestors’ memory can only be disclosed with their permission. The ceremony of ‘telling the history’, ka edwɛkɛ, must be preceded by a libation aimed at summo- ning the ancestors’ spirits to the meeting – meeting which is therefore regarded as a sacred moment. This custom is losing much of its earlier importance, and, because of multiple transmission, tales may change, and different versions may contain contradictory statements. In spite of this, people continue to hear and transmit the tales from one generation to another, and the variants increase conti- nuously. This is one of the reasons why people often do not care about contradic- tions and inconsistencies, which frequently characterize the traditional discourse. Historical tales – ‘oral tradition as history’38 – do not provide researchers with rigorous chronological data, nor with exact reconstructions of events39. Nonethe- less, like riddles, they offer enough elements to permit cross-checking with other

32. Baesjou, “Trade Conflicts”, 25; Valsecchi, Power and State Formation, 96. 33. Baesjou, “Trade Conflicts”, 42 n. 6. 34. AMH-7661 NA (WIC 1650). 35. AMH-7600 NA. 36. Port. 182, N. 29, “Goudkust van Guinea”, early 19th c., probably by J. Koning, Leiden University Library. A later English version was published by H. Meredith, An Account of the Gold Coast of Africa with a Brief History of the African Company, London, Longman, etc., 1812 (reprinted by F. Cass, 1967). 37. See C.-H. Perrot, “L’histoire dans les royaumes Agni de l’est de la Côte-d’Ivoire”, Anna- les ESC, 6, 1970, 1659-1677; M. Pavanello, “L’événement et la parole. La conception de l’histoire et du temps historique dans les traditions orales africaines: le cas des Nzema”, Cahiers d’Études Africaines, 171, XLIII, 3, 2003, 461-481. 38. J. Vansina, Oral Tradition as History, Madison, The University of Wisconsin Press, 1985. 39. See D. Henige, Chronology of Oral Tradition, Quest for a Chimera, Oxford, Basil Black- well, 1974; D. Henige, Oral Historiography, London, Longman, 1982. Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 77 data and interpretation. It must also be stressed that such tales often refer to well- known historical events, and may thus be of help in reconstructing chronologies. They may also refer to customs and beliefs still or no longer existing. Moreover, the need to emphasize family interests, or conceal negative memories, encoura- ges the parties in conflict to give rise to variants of the same histories or merge different stories pertaining to the various matrilines composing the lineage. A further aspect of the process of change in oral tradition is determined by the judicial procedure, and the ever-growing use of written documents. The interaction between oral and written sources is twofold. First, the use of public documents (Gazettes, lists of chiefs, archival material and other official or judicial documentation), as well as books40 and newspapers, is a well-known issue. Second, the legal procedure itself is a notable source of internal transformation of the nature and contents of the orally transmitted memory. The modern use of oral historical traditions as legal evidence watered down their sacredness and brought about the habit of narrating the tales without any traditional ritual precaution, often only swearing on the Bible, which – until the recent past – had little meaning for illi- terate elderly men. Nevertheless, the possession of historical knowledge is considered to be a mark of legitimacy. ‘Telling the history’ is a crucial performance of authorized elders of the legitimate stock of the family. Knowing or not knowing the stool history distinguishes putative royals from plebeians. The dispute between the two main traditions (Anɔ Asɛma vs. Dehelɛ Kwesi) had a bearing on the evaluation of the correct historical knowledge among the Mafolɛ family. In 1940, Annor Kpole stood in the State Council for his Mafolɛ lineage against the Adahonle Kaku Ako- soku (installed on the stool in 1939, with the name of Kabenla Kumisa II), and gave evidence recounting the tradition of Anɔ Asɛma. His family had not appointed him for that duty, but he wanted to represent the Mafolɛ because the then head of the family and original plaintiff, Annor Arizi, was in hospital. The Council rejected Annor Kpole’s tale as incorrect41, and Amihere Agyilihah, the Mafolɛ plaintiff in the 1948-49 case, later accused him of not knowing the history:

40. The main books used by chiefs and heads of families in the manipulation of oral tradition are: Annor Adjaye, Nzima Land, London, Headley, 1931, a text compiled by Annor Adjaye I, who sat on the Western Nzema paramount stool from 1920 to 1938, and printed in London at his own expense with the aim of celebrating himself, and containing administrative and political issues, as well as historical information; Essuah, Mekakye bie, a 3-vol. collection of popular legends and traditions without any critical framework, written in Nzema language by the catholic bishop of Takoradi. It is possible to find in the houses of some highly educated chiefs copies of books on the history of Africa and Ghana, such as, for example, F. K. Buah, A History of Ghana, London, Mac- millan, 1980, and C. C. Reindorf, History of the Gold Coast and Asante, Based on Traditions and Historical Facts, Comprising a Period of more than Three Centuries from about 1500 to 1860, Basel, Basel Mission, 1895, reprinted 1950. 41. The then paramount chief, Annor Adjaye II, who was chairing the arbitration panel in the 1940 case, objected that Bonyere was founded by Dehelɛ Kwesi and not by Anɔ Asɛma.

Africa, N.S., I / 1, 2019 78 a r t i c o l i / a r t i c l e s

[…] when we were ready to perform the funeral of ohene Nyamekɛ Annor of Bonyere [in the year 1939] Adahonle family came and sworn oath protesting that they own the stool of Bonyere. We responded to the oath and took action against Kabenla Kumisa defendant family. My family deputed Annor Arizi to appear [in] court on their behalf and litigated with Kabenla Kumisa. Annor Arizi went to hospital at Sekondi for medical treatment. The case was called. Annor Kpole was not authorized by my family to stand and litigate for them; but he appeared and took the said litigation upon himself. As he does not know the ancient history of the stool, judgement was entered against him. As he did not know the history, started to give evidence during the time of Andoh Assima’s reign42. The accusation is serious, because it corresponds to an open declaration of illegitimacy. Awonyi Nwia, the Adahonle defendant in the 1948-49 case, immediately seized the opportunity for inquiring about the secret structure of the Mafolɛ and asked his antagonist to clarify which of the two Mafolɛ plaintiffs of 1940 was his legitimate relative: Awonyi Nwia (Ad). Could you remember when I was not at Bonyere, Annor Arizi and Annor Kpole took action against Kabenla Kumisa ohene of Bonyere claiming the ownership of the stool of Bonyere? Amihere Agyilihah (Ma). Yes, I do remember Awonyi Nwia (Ad). After the result of the case, who was found guilty? Amihere Agyilihah (Ma). Annor Kpole was found guilty because he does not know the ancient history of the stool. […] Awonyi Nwia (Ad). You and Annor Arizi are they succeeded each other? Amihere Agyilihah (Ma). Yes. Awonyi Nwia (Ad). Is Annor Kpole also succeeded you? Amihere Agyilihah (Ma). Yes. Awonyi Nwia (Ad). Annor Kpole does not know the history of the stool and if a member of your family dies, Annor Kpole could know the rightful successor? Amihere Agyilihah (Ma). Yes43. Not knowing the family or stool history means belonging to a subordinate (pawn or slave) matriline. The Adahonle defendant was clearly aware of the lie of his Mafolɛ antagonist. Yet Amihere Agyilihah did not want to break the secret, and insisted that Annor Kpole was a legitimate member of his own family who shared the same inheritance, even though he did not know the history.

Slavery and pawnship in traditional discourse The frequent mentions in the records that both families, Adahonle and Mafolɛ, “met” king Nyanzu Aka would suggest that they came to Bonyere after

42. 30 December 1948 – From evidence by plaintiff Amihere Agyilihah (Mafolɛ). 43. 31 December 1948 – From the cross-examination by defendant Awonyi Nwia (Adahonle) to plaintiff Amihere Agyilihah (Mafolɛ). Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 79

1816. I advance the hypothesis that the king was probably residing in Bonyere and that both Adahonle and Mafolɛ were subordinate offsprings of the king’s sons. The clues supporting this interpretation are manifold. First, they were all employed as laborers in Akasuazo, the food-producing village for the Bonyere court. Suffice it to quote the following statements made in the 1964 case by the Mafolɛ representative: According to tradition, my ancestors […] generated from Abradi [Wassaw]44 reached Kabaku45. They met Awulae Nyanzu Aka I, whose village is now called Bonye- re. He permitted them to farm on part of Bonyere land. The product of the said farm was sent to the Awulae by the above named persons. While they were in the village, Kabenla Kumisa and others also came and met them at the village. They all lived together there.46 Kabenla Kumisa is the direct ancestor of the treacherous Adahonle caretaker. The traditional discourse reveals three critical issues. First, in most cases, oral traditions talking about people ‘meeting’ a king are dealing either with immigrants who pay allegiance by pawning themselves, or with offsprings whom the king (or his sons) begot from purchased or pawned women. Second, it is indispu- table that farming on the king’s land and presenting the king with produce were the tasks of dependents, which confirms the entire interpretation. Third, ‘living together’ means intermarrying and sharing the same food. Therefore, when Amihere Agyilihah says that “they all” (his and Kabenla Kumisa’s people) were “living together”, it means they were intermarrying, as was customary among subordinate matrilines stemming from the same royal stock. Indirect evidence of the pawn origin of the Kabenla Kumisa’s family is to be found in the following statement by a witness in the 1963-64 case: I know that Awulae Nyanzu Aka founded a village and many settlers went and settled with him. Kobina Inkumsa from the Eastern Nzema also went and settled there as salt boiler47. ‘Salt boiler’ denotes the activity of domestic servants or pawns employed as salt producers through the technique of marine water boiling48. This interpretative

44. In the previous 1948-49 case, when cross-examined by defendant Awonyi Nwia (Ada- honle), the Mafolɛ plaintiff, the aforementioned Amihere Agyilihah, declared on 31 December 1948 that his ancestors “came from Ashanti Mampong”. On the same occasion, he wanted to add that “Wassa Fiasi State Benso there our principal family stool lives”. This is one of the numerous surprising inconsistencies shown by oral family histories. 45. Ekebaku, a village East of Beyin, on the border between Western and Eastern Nzema. 46. 23 October 1964 – From evidence by defendant Amihere Agyilihah (Bonyere case 1964). 47. 27 October 1964 – From evidence by Agyili Aka first witness for defendant Amihere Agyilihah (Mafolɛ). 48. For salt producers on the Ivory Coast, see H. Memel-Fotê, L’esclavage dans les sociétés lignagères de l’Afrique noire: exemple de la Côte d’Ivoire précoloniale, 1700-1920. Lille 3, ANRT, 1989; Id., “L’esclavage sur la côte ivoirienne du XVe au XVIIe siècle. Faits et problèmes”, Africa.

Africa, N.S., I / 1, 2019 80 a r t i c o l i / a r t i c l e s framework is confirmed by a further clue emerging from the evidence given by both parties in the 1948-49 case. The Adahonle claimants to the Bonyere stool belong to a would-be legitimate matriline whose ancestor, Kwesi Bendobenli, is said to come from Shama (East of Cape Three Points). They affirm that the traitor was an Annor Eza, a mere caretaker, member of a pawned line, which traced its descent from Kabenla Kumisa, a man whom Kwesi Bendobenli had pawned with all his relatives at an earlier time: I remember my great grandfather by name Kwesi Bendobenli immigrated from Shama. When he was coming to Nzema land, he was accompanied by his mothers, brothers and sisters. My grandfather Kwesi Bendobenli stayed there and named the village Bonyere. That time King Yanzu Ackah I was King of Beyin. At that time, the place was thick forest. My grandfather stayed at Bonyere for number of years. One day he saw a man called Awua Kpanyinli of Kikam, Kabina [Kabenla Kumisa], Ay- iba and their sister also called Agu Manzah, came from Kikam and told my grandfather they had got a prodigious debt to be paid, and they desired my grandfather to go with them and paid the said debt for them, and they would stay with him. My grandfather agreed and paid their debts and they stayed or lived with him as his pawn in one compound. […] After all, many people of different tribes also came and stayed with him, there the town extended. […] One Effua with his sister Ebella Kyi also came from Ebonua in the French Ivory Coast49; came and stayed with my grandfather Kwesi Bendobenli. My grandfather asked Ebella Kyi to be married, and Effua gave her to my grandfather for marriage. My grandfather Kwesi Bendobenli beget […] Azan Ndoli. [...] After this, my grandfather’s relatives were dying, the whole of them dying daily. […] His pawn relatives were the people killing his relatives because they possessed plenty witchcraft50. […] Kabenla Kumisa’s people used to kill people at night with witchcraft. […] In time of Annor Eza, a relative to Kabenla Kumisa, then King Amakyi was reigning at Beyin51. Awonyi Nwia’s tale offers three interesting features. First, it plainly disclo- ses the pawning of an important matriline of his Adahonle family: the Kabenla Kumisa’s line, who occupied the Bonyere’s stool from 1939 to 1945, when Kabenla Kumisa II (alias Kaku Akosoku) was enstooled. Second, the tale de-

Rivista trimestrale di studi e documentazione dell’Istituto italiano per l’Africa e l’Oriente, Roma, LIV, 1, 1999, 1-49. For salt production on the Slave Coast, see F. A. Iroko, “Le sel marin de la côte des esclaves durant la période précoloniale”, Africa. Rivista trimestrale di studi e documentazione dell’Istituto Italiano per l’Africa e l’Oriente, Roma, XXXXVI, 4, 1991, 520-540. Oral tradition I collected during my fieldwork shows that several stool-owning Nzema families originated as domestic servants and salt producers. 49. It is surprising that the Mafolɛ elder did not object: he had previously asserted that his ancestors came from Ashanti or Wassa, whereas now his antagonist declares that they hailed from Ebonua (Ivory Coast). This is further proof of the inconsistencies in the family histories that too often are collations of stories pertaining to different matrilines composing the lineage. 50. In the Nzema language, witchcraft, ayɛne, is both a nature (ɔle ayɛne, he is a witch) and an object (ɔlɛ ayɛne, he has witchcraft). 51. 6 January 1949 – From evidence by defendant Awonyi Nwia (Adahonle). Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 81 scribes what was probably the first suanu agyalɛ52 in the history of the two families: Kwesi Bendobenli, the Adahonle ancestor, married Effua’s sister, Ebela Ekyi, the first ancestress of the Mafolɛ53. Third, Awonyi Nwia draws a very negative sketch of his ancient pawns, presenting them as powerful witches who killed almost all the members of the legitimate matriline. His aims are three- fold: ascribing witchcraft powers to his asalo people, charging them with ancient treason, and justifying the absence from the stool of the rightful Adahonle suakunlu abusua. It is quite surprising that the Mafolɛ plaintiff asked his antagonist to declare that the would-be legitimate Mafolɛ themselves were Adahonle’s pawns: Amihere Agyilihah (Ma). Can you tell the court to know that my grandfather Effua and Ebela Kyi also are of your grandfather Kwesi Bendobenli pawns or money relatives? Awonyi Nwia (Ad). Yes, were his pawns and he married Ebela54. However, after a few days, cross-examination led the Mafolɛ plaintiff to advance the suspicion that even the would-be legitimate Adahonle matriline was probably not as royal as they were claiming. In the Nzema language, Bendobenli means “they do not buy their mother”55, which sounds like a rhetorical denial of slave origin.

Amihere Agyilihah (Ma). Did you not remember Kwesi Bendobenli’s mother is named Ndofula? Awonyi Nwia (Ad). No. His mother is called Ahu. Amihere Agyilihah (Ma). Do you know Akasui of Ndoku? Awonyi Nwia (Ad). No. I do not know him. Amihere Agyilihah (Ma). Could you not remember Akasui bought Kwesi Bendo- benli’s mother as his slave56?

The arbitration panel dismissed the last question because a public accusation of slavery is strictly forbidden. Therefore, Awonyi Nwia did not answer. The questions asked by Amihere Agyilihah during the subsequent cross-examination obliged Awonyi Nwia to make some embarrassing admissions that we-

52. Suanu agyalɛ, ‘marriage in the house’, denotes cross-cousin marriage as well as marriage with a purchased or pawned woman (see Pavanello, “Reconsidering Ivor Wilks”). 53. Ɛbela ekyi is the stool-name of the Bonyere’s Mafolɛ queenmother. 54. 17 January 1949 – From cross-examination by the plaintiff (Amihere Agyilihah, Mafolɛ) to defendant (Awonyi Nwia, Adahonle). 55. The phrase is bɛ ndɔ bɛ nli (they, bɛ, do not buy, ndɔ [to buy is tɔ, dɔ], their mother, bɛ nli). A proverb says abɔnkye se ɔ nli a bɛdɔle ye a na bɛandɔ yemɔ (the goat says they buy her mother, but they do not buy herself), ironically emphasizing the attitude of those people who cannot deny the slave origin of their mother, but stress that they are not themselves slaves (because they are children of a free man). 56. 17 January 1949 – From cross-examination by the plaintiff (Amihere Agyilihah, Mafolɛ) to defendant (Awonyi Nwia, Adahonle).

Africa, N.S., I / 1, 2019 82 a r t i c o l i / a r t i c l e s akened his position. He had to acknowledge that members of the pawn line had occupied the stool in the past as caretakers, and that the same line was presently occupying the stool:

Amihere Agyilihah (Ma). How many of your pawns also occupied or appointed caretaker on Bonyere stool? Awonyi Nwia (Ad). They were two, Annor Eza and Kabenla Kumisa. Amihere Agyilihah (Ma). Whose family now in possession of your stool? Awonyi Nwia (Ad). The stool is now in possession of Kaku Akosoku. Amihere Agyilihah (Ma). So, your pawns have occupied the stool of Bonyere? Awonyi Nwia (Ad). Kaku Akosoku litigated with Mafolɛ family, and having obtained judgement, he also sat on it57. Awonyi Nwia was probably referring to both Kabenla Kumisa, the ancient pawn of Kwesi Bendobenli, and to Kabenla Kumisa II, alias Kaku Akosoku, installed as chief at Bonyere in 1939. We may conjecture on the basis of this passage that the first Kabenla Kumisa was at the same time a pawn and a ‘son’ of the previous rulers, thus making it possible for him to be appointed as caretaker.

A contested human sacrifice

When Bonyere was burnt during the civil war, an amonle was cast. This took the form of a ritual consisting in burying a mortar (duba) and a pestle (dwɔma) in the soil along with some herbal medicines. This amonle is also known as ‘bɛ ya aserɛne’. After the civil war, Bonyere was rebuilt and its stool – according to an accepted tradition – was given by the then Western Nzema king, Amakyi, to a member of the Mafolɛ family, Azan Ndoli, because the original owners were not available. Before the town could be rebuilt, the amonle was to be removed through a purification ritual (amonle ɛsuosualɛ). The Mafolɛ tradition58 holds that the ritual included medicines and drinks for libation, the sacrifice of monkeys and fowls, as well as the human sacrifice of a virgin girl, described as a granddaughter of Azan Ndoli. The girl was a kanra daughter of the house of Azan Ndoli, as it was usual to consider the kanra children as nlɔnra (plural of alɔnra, with the twofold meaning of grandchild and uterine grandnephew/grandniece)59. Slaves and war prisoners were commonly used as victims for the funeral ceremonies of kings or important

57. 20 January 1949 – From cross-examination by plaintiff Amihere Agyilihah (Mafolɛ) to defendant Awonyi Nwia (Adahonle). 58. See evidence given by Amihere Agyilihah, 30 December 1948. 59. The slave woman was considered as a daughter (ra) or a uterine niece (awozoa). “The master of a female servant could give her in marriage to any marriageable male relative. The children were considered as children or grandchildren as the case might be”. (My interview with Nana Otu Perbi IV, ɔhene, in the ahenfie of Mampong Akuapem, Eastern Region, Ghana, 29 July 2004, see Pavanello, “Reconsidering Ivor Wilks”, 22). Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 83 chiefs, or for purification rituals60. According to the Mafolɛ, Azan Ndoli provided all the money and the other items requested for the ceremony, including the sacrificial maiden. The Adahonle challenged the Mafolɛ’s version, affirming their tradition that Azan Ndoli was not the protagonist of the purification of the site of Bony- ere, because the money was provided by their family. They also pointed out that there could not have been human sacrifice due to the colonial prohibition. Their refusal to acknowledge the historical evidence of the sacrifice called into question the strong symbolic link of the Mafolɛ with their land. A cross-examination offers a crucial clue for a chronology of the events:

Awonyi Nwia (Ad). Since the mortar pestle put or was placed at Bonyere and taken it out and the girl was killed, how many years now? Amihere Agyilihah (Ma). About 79 years ago. Awonyi Nwia (Ad). Since the English law forbidden human sacrifices how many years now? Amihere Agyilihah (Ma). I cannot tell the years61. Thus, according to the Mafolɛ elder, the purification ceremony and the human sacrifice were performed in 1869-70. Yet, a witness, summoned bythe Mafolɛ plaintiff in 1949, stated the following during a cross-examination by Awonyi Nwia:

Awonyi Nwia (Ad). From the time to this day, Bonyere was burnt how many years now? Agyili Aka. It may be about 75 now. Awonyi Nwia (Ad). How many years human sacrifices were prohibited by the Bri- tish Government? Agyili Aka. I do not know. Awonyi Nwia (Ad). Then it was about 75 years Azan Ndoli provided his granddaughter to make sacrifice for the town Bonyere to be killed? Agyili Aka. Yes. […] Awonyi Nwia (Ad). What is the name of the woman or girl Azan Ndoli offered to be killed and offered sacrifices to Bonyere soil? Agyili Aka. She is called Eduku62.

60. The issue of human sacrifice in Africa and in the Akan world is quite problematic. See R. Law, “Human Sacrifice in Pre-colonial West Africa”,African Affairs, 84, 334, 1985, 53-87; E. Ter- ray, “Le pouvoir, le sang et la mort dans le royaume asante au XIXe siècle”, Cahiers d’Études Africaines, 34, 136, 1994, 549-561; C. Williams, “Asante: Human Sacrifice or Capital Punishment? An Assessment of the Period 1807-1874”, The International Journal of African Historical Studies, 21, 3, 1988, 433-441; I. Wilks, “Asante: Human Sacrifice or Capital Punishment? A Rejoinder”, The International Journal of African Historical Studies, 21, 3, 1988, 443-452. 61. 31 December 1948 – From cross-examination by defendant Awonyi Nwia (Adahonle) to plaintiff Amihere Agyilihah (Mafolɛ). 62. 4 January 1949 – From cross-examination by the defendant (Awonyi Nwia, Adahonle) to plaintiff’s 1st witness (Agyili Aka).

Africa, I / 1, 2019 84 a r t i c o l i / a r t i c l e s

Agyili Aka affirms that Bonyere was burnt and rebuilt in 1874, a statement that appears based on uncertain memory. The year coincides with the end of the Anglo-Asante war and although it may not seem inappropriate for the epoch of the reconstruction of Bonyere, nonetheless it is not a proof that the colonial prohibition of human sacrifice was in force. A first formal interference by the British Authorities in the internal affairs of precolonial coastal states had been the Bond of 1844, “which was really a treaty with the Fanti chiefs, the Bond dealt only with criminal cases and barbarous customs”63. However, effective British control over the Nzema area was only completed with the constitution of the Gold Coast Col- ony, sometime after the end of the Anglo-Asante war. In light of this, the objection that human sacrifices were forbidden by the British at the end of the Nzema civil war is most probably unfounded. Nonetheless, the refusal to acknowledge the historicity of the human sacrifice performed for the purification of the town is common among Adahonle supporters in Bonyere64.

Conclusion: the heritage of pawnship and slavery

The families or matrilines contending the inheritance of a stool often ex- change reciprocal accusations of illegitimacy under the labels of pawnship or ‘money relationship’. Indeed, many records of stool disputes show that the liti- gating families are both of subordinate origin stemming from pawns or servants of ancient ruling groups. Moreover, they share a history of frequent intermarriage of the cross-cousin types also involving newly acquired subordinate matrilines: a practice which characterizes emerging lineages of former low status and which shows that the custom of suanu agyalɛ involving a free man and his kanra daughter/wife can work as a mechanism of social mobility. Pawnship and slavery refer to different legal conditions, although in practice both types of dependents were obliged to work to the benefit of their masters65. Yet, in local perception, the two institutions may overlap, because people given in pawn were generally of slave origin. However, the language of the proceedings often uses ‘pawn’ and ‘pawnship’ as more politically correct expressions than ‘slave’ and ‘slavery’, which are considered a serious abuse. Colonial discourse too appears to adopt a similar language, as Claridge witnesses that British government abolished ‘internal domestic slavery’ with the proclamation of the Gold Coast Colony in 1874, but the custom “in existence was that known as pawning”66. Another aspect of this kind of political correctness concerns public discourse on

63. W. W. Claridge, A History of the Gold Coast and Ashanti: From the Earliest Times to the Commencement of the Twentieth Century, 2 vols, London: John Murray, 1915, vol. 2, 175, 452. 64. Personal evidence. 65. See Perrot, Les Anyi Ndenye, 157ff.; P. E. Lovejoy and T. Falola eds., Pawnship, Slavery and Colonialism, Trenton NJ, Africa World Press, 2003. 66. Claridge, A History of the Gold Coast and Ashanti, vol. 2, 177-78, 443ff. Pavanello, Pawnship and Domestic Slavery in Chieftaincy Disputes 85 human sacrifice. This ancient custom is confined to a remote ‘precolonial’ time and its memory is banned, as is the memory of slavery. The hidden link between slavery and human sacrifice recalls the hidden link between slavery and witchcraft that is often raised by the records of the stool disputes67. A further aspect emerging from the proceedings is the relationship between fathers and sons. As the literature on the subject unanimously acknowledges68, the ‘father-son’ relation is a political pillar in the Akan world. The local concept of paternity is both individual and collective; it refers to someone’s real father, as well as to his father’s matrilineal family (the ‘fathers’), so that any male member of that family may be considered as ‘father’. Patrifiliation has a further implication in the case of marriages with enslaved or pawned women, as the children were normally included within their father’s abusua; thus, matrilineal kinship overlapped with and concealed paternity. A clue of hierarchical structure is the ‘father-son’ relation between matrilines, especially when it is repeated from one generation to another. An external observer can distinguish pawnship from slavery when the subordinate (‘children’) matrilines do not belong to the same clan as their ‘fathers’, as in the frequent case of free people “held in debt bondage as col- lateral for loans”69, which was probably the case of the first Mafolɛ ancestors. In such an event, pawns never assumed the clanic identity of their creditors, unless they had previously been members of the same clan, which was often the reason for their request for help. A clue for distinguishing ancient pawns from purchased people, when they share the same clanic identity as their old masters, is when the family memory can recollect the place of origin of the pawns, as in the case of the Adahonle Kwesi Bendobenli’s and Kabenla Kumisa’s matrilines, whose origin – Shama (Ahanta) and Kikam (Eastern Nzema), respectively – are universally acknowledged. By contrast, the place of origin of former slaves is never known. Crucial to the interpretation of the traditional discourse on slavery and pawnship is the gender issue: plaintiffs and defendants in the disputes are almost always prominent men of the contending families. This peculiarity has some implications when genealogical memory – a distinct feature of historical knowledge

67. It is not possible here to deal with these important issues, which I propose to develop in a future work. 68. For the Nzema, see M. Pavanello, “Clan, lignage et mariage”; P. Valsecchi, Power and State Formation; P. Valsecchi, “My Dearest Child Is My Slave’s Child: Personal Status and the Politics of Succession in South-West Ghana (Nineteenth and Twentieth Centuries)”, in A. Bellag- amba, S. E. Greene and M. A. Klein eds., African Slaves, African Masters: Politics, Memories, Social Life, Trenton NJ, Africa World Press, 2018, 41-60. For the Asante, see Arhin, Traditional Rule, 29, 40ff.; Perbi, A History of Indigenous Slavery in Ghana; R. S. Rattray, Ashanti, Oxford, Clarendon Press, 1923, 45ff.; R. S. Rattray, Ashanti Law, 8ff.; for the Asante ntoro patridivisions, see M. Fortes, Kinship and the Social Order: the Legacy of Lewis Henry Morgan, London, Rout- ledge and Kegan Paul, 1969, 198-202; I. Wilks, Asante in the Nineteenth Century: The Structure and Evolution of a Political Order, Cambridge, Cambridge University Press, 1975, 329. 69. T. Falola and P. E. Lovejoy “Pawnship in Historical Perspective”, in Lovejoy and Falola eds., Pawnship, 1.

Africa, N.S., I / 1, 2019 – is concerned. Queenmothers (ɔbahema, pl. abahema) are the official experts on genealogies, whereas elderly men control the male political networks of their families. This is the reason why genealogies are hardly available from the judicial records, and the frequent mentions of the ancestors of both legitimate and illegitimate matrilines are almost never supported by correct genealogical data. As the Bonyere dispute demonstrates, it was, and still is, customary that a ‘son’ may be appointed as neazo70, that is, as caretaker of the stool and the land of his ‘fathers’ when none of them is at hand or able to assume the office. A neazo may often be selected among the ‘sons’ originating from pawned or enslaved women. In such a case, the caretaker sitting on the stool often belongs formally to the same family as the legitimate owners, though he is only a temporary custodian, a truth that is not necessarily disclosed. Many disputes on stools arose from the conflict between would-be legitimate matrilines and their illegitimate relatives acting as caretakers. Indeed, the records show that pawned or slave matrilines very frequently compete for the stools and take over the succession to them. Thus, conflicts often break out within the same family, which may irrevocably split up when the internal struggle proves irreconcilable.

70. Nzema term denoting the political position of the caretaker of a stool. From nea, to watch, and zo, upon, thus meaning ‘watchman’ or ‘custodian’ (M. Pavanello, “The Work of the Ancestors and the Profit of the Living. Some Nzema Economic Ideas”, Africa Journal of the International African Institute, 65, 1, 1995, 36-57). Samuel Andreas Admasie

Cy c l e s o f Mobilisation, Wa v e s o f Un r e s t : Et h i o p i a n La b o u r Mo v e m e n t Hi s t o r y

Abstract

The history of the Ethiopian labour movement appears in the shape of repeated cyclical movements in levels of mobilisation. In contrast with the static depiction dominating the literature on the movement, this article examines how Ethiopian wage-workers have, over the past six decades, acquired, relinquished and redeveloped the collective coherence and organisational capacities to engage in sustained collective action, the strategic orientation and willingness to do so, and the resilience to fend off sustained counter-mobilization. The article discuss- es how the history of the Ethiopian labour movement is deeply entangled with, and mutually constitutive of, historical shifts in wage labour relations, and the struggle over these relations. In doing so it suggests a reciprocal relationship between Ethiopian workers’ achievements and strategic orientation. At the heart of the article is the notion of historical agency, which Ethiopian workers acting collectively have exhibited to a degree that is underappreciated in the literature. k e y w o r d s : w o r k e r s ; l a b o u r m o v e m e n t s ; t r a d e u n i o n s ; s t r i k e s ; l a b o u r r e l a t i o n s

Introduction

The historiography of the movement of Ethiopian wage-workers is severely underdeveloped. Yet, when approaching this “field” – the quotation marks are warranted, because the body of literature remains too impoverished to thoroughly warrant the term – one encounters some curious general features. One such curiosity is the tendency in the literature to deny organised wage- workers any degree of autonomy in establishing their strategic orientation and engaging in collective action. What workers have done collectively has been, through direct control or severe manipulation, viewed as being determined by exogenous political agents and events, including institutions of the state. The macro-political conjuncture has tended to become the sole lens through which workers’ collective actions have been interpreted and explained. The conditions

Africa, N.S., I / 1, 2019 88 a r t i c o l i / a r t i c l e s of wage-workers, or the internal dynamics of the workers’ movement, has not featured as factors propelling neither shifts in the strategic orientation of organised workers nor upswings in industrial contention. Another particular feature, somewhat related to the above, pertains to state- centrism in the determination of labour relations, and the concomitant denial of agency to Ethiopian workers. Workers have been viewed as having been control- led, commanded, and deployed within given relations, at the inexplicable whims of rulers. Labour relations have seemingly been entirely derived from state policy, and insulated from pressures from below. Workers have thus been portrayed as deprived of both autonomy and agency, and the historiography of the labour movement has tended to collapse into an analysis of the political conjuncture and the content of state power. As a corollary, the labour movement has been described as co-opted, dormant and inconsequential. This article draws on research in the archives of, and interviews with participants in, the Ethiopian and international trade union movement to examine the history of the Ethiopian labour movement. Far from state manipulation alone, this entails the history of Ethiopian workers having developed and relinquished the collective solidarities, coherence and capacities required for meaningful industrial contention within that contention itself. Ethiopian labour movement history is, moreover, deeply entangled and mutually constituted with the history of shifting labour relations in Ethiopia’s waged sector, and the struggle over these relations.

Theoretical considerations

The literature offers a number of general and theoretical notions relevant to the study of the history of the Ethiopian labour movement, the shifts in its strategic orientation, and its impact. A suitable starting point is the discussion on the dual nature of trade unions. Trade unions are central components of labour movements – although the term labour movement broadens the scope. In addition to trade unions, it is here taken to include also other constellations organising wage-workers in the struggle for better conditions, as well as collective practices among wage-workers, aiming to improve those conditions. Following Richard Hyman’s definition, a trade union can be conceived as “an agency and a medium of power” which permits workers to exercise a level of collective control over labour conditions which they would not have able to do acting individually1. Unions, Hyman explains, engages in a two-way relationship with external agencies, where the former seeks to influence and pressure the latter and vice versa. But to be able to act coherently and make sure that established agreements and strategies are adhered to, unions must also

1. R. Hyman, Industrial Relations: A Marxist Introduction, London, Macmillan, 1975. Admasie, Cycles of Mobilisation, Waves of Unrest 89 exercise internal control over workers. This is what constitutes the dual nature of trade unions. There is both tension and complementarity between the institutional aspect of labour movements – manifested in, for example, trade unions – as organised collectives, and labour movements as a set of practices, the principal example of which is the strike. With regards to this relationship, van der Linden has observed that while “unions cannot exist without (the ultimate threat of) the strike weapon, the converse is not true”2. However, even if strikes do not require a union, some form of organised constellations is generally desirable as a vehicle that can articulate demands, coordinate efforts, negotiate agreements, and safeguard them. In the absence of any organisation it is unlikely that durable concessions can be extracted. The broad general aims of labour and trade union movements are given by the definition: to improve or defend workers’ conditions. The precise nature of the goals and the means of achieving them, however, vary widely in reality. This is where the strategic orientation – which determines both the roles and potential impacts of movements – becomes important. Strategic orientation is generally established in contention, and subject to continuous challenge and successive shifts, whenever the balance of forces demands. As Gallas has aptly put it, “unions are not just organisations of struggle, but also fields of struggle between competing forces of labour with different strategies… and they can change over the life course of a single union”3. Union officialdom – because of corporate interests, and functional, social and economic separation from the rank-and-file – tend to press for a moderation of aims and practices, and uphold industrial peace and legality. The preservation of tranquillity at all costs is usually grounded in the logics of “institutional needs” – to preserve the organisation, strengthen it, or defend its achievements – and the defence of the organisation itself is often conflated with and placed above the defence of the interests of the membership4. The growing clout of a separate layer of officialdom is known as bureaucratisation, and its effects are enhanced by the fact that subversive external pressures towards moderation generally focus on the leadership level of organisations5. While this interaction between internal and external pressures has been viewed as a key source of moderation of strategic orientation, militancy has been

2. M. van der Linden, Workers of the World. Essays Towards a Global Labor History, Leiden, Brill, 2008. 3. A. Gallas, “Class Power and Union Capacities. A Research Note on the Power Resources Approach”, Global Labour, 9, 3, 2018, 351. 4. R. Brenner, “The Political Economy of the Rank-and-File Rebellion”, in A. Brenner, R. Brenner and C. Winslow, eds., Rebel Rank and File: Labor Militancy and Revolt From Below Dur- ing the Long 1970s, London, Verso, 2010, 37-76. 5. R. Darlington and M. Upchurch, “A Reappraisal of the Rank-and-File Versus Bureaucracy Debate”, Capital & Class, 36, 2012, 77-95.

Africa, N.S., I / 1, 2019 90 a r t i c o l i / a r t i c l e s viewed as originating in the pressures from the rank-and-file6. According to Dar- lington and Upchurch “it is the exploitative social relations at the heart of capitalist society to which the mass of rank-and-file union members are subject that provides the material basis for collective workers struggles which distinguish them from [full-time officials]”7. But even so, certain factors more directly conditions – either in a conducive or restraining manner – the prevalence of militant practices. As Kelly has pointed out, “the conflict of interest that lies at the heart of the capitalist employment relationship does not necessarily give rise to conflict behaviour”8. Following Charles Tilly, Kelly argues that for collective action to be viable, a number of conditions need to be satisfied. These encompass the social construction and definition of collective interests; collective mobilisation which transforms individual workers into collective actors; collective organisation; and opportunity. The pivotal condition of opportunity includes considerations regard- ing the relative balance of power, the costs of repressions, and contingent opportunities available. Employers and the state generally undertake counter-mobilisation, which is “likely to target all of the dimensions of collectivism”9. In other words, attempts are made to restrict opportunities and impose higher costs on collective action, overturn goals and leaders, rearticulate interests, undermine organisations, and demobilise workers. Generally, there appears to be a clear correlation between the militant energy of a labour movement, on the one hand, and its potential impact on labour relations on the other. Silver has provided a global historical account of how vibrant labour movements engaging in sustained strike activity have generally “succeeded in raising wages, improving working conditions, and strengthening workers’ rights”10. A number of scholarly accounts11 largely corroborate these findings, although other accounts point to the contingent nature of this relationship, and the impact of broader political-economic and historical forces12. If militancy tends to result in greater potential impact of labour movements, there are also indications that such impact has reciprocal effects on movements’ strategic orientation. Screpanti has pointed to such a relationship between achievements and militancy. The notion that “there is a positive cumulative effect of past

6. Brenner, “The Political Economy”; S. Cohen Ramparts of Resistance. Why Workers Lost Their Power, and How to Get It Back, London, Pluto, 2006. 7. Darlington and Upchurch “A Reappraisal of”, 88. 8. J. Kelly, Rethinking Industrial Relations. Mobilization, Collectivism and Long Waves, Lon- don, Routledge, 2002, 25. 9. Kelly, Rethinking Industrial Relations, 128. 10. B. Silver, Forces of Labor. Workers’ Movements and Globalization Since 1870, New York, Cambridge University Press, 2003, 168. 11. See, for example, R. Franzosi, The Puzzle of Strikes. Class and State Strategies in Postwar Italy, Cambridge, Cambridge University Press, 2006, 344; S. Cohn, When Strikes Make Sense ‒ And Why. Lessons from Third Republic French Coal Miners, New York and London, Plenum, 1993, 219. 12. See for example Brenner, Brenner and Winslow, eds., Rebel Rank and File on differing outcomes of successive strike waves in North America and Europe. Admasie, Cycles of Mobilisation, Waves of Unrest 91 experience in relation to actual claims”13 leads to the possibility that a virtuous, and somewhat path-dependent, cycle of increasing militancy and achievements could be created. When sufficiently developed, this could result in the type of explosive surges of industrial unrest which Screpanti termed proletarian insurgencies. However, because “claims will change in time in the same direction as achievements”14, Screpanti also acknowledged that an equally vicious cycle of defeat and despondency can be generated.

The Ethiopian labour movement in the literature

The above notwithstanding, the manner in which the Ethiopian labour movement and trade unions have been treated in the scant literature does not exhibit much appreciation for neither the potentials, practice, nor the historical achievements of organised workers. The movement has tended to be treated as persistently co-opted, dormant, weak, and inconsequential. At times when the movement has exhibited militancy and autonomy, this has been attributed to external agents. When it has been characterised by energy and activity, this has been attributed to particularities in the political conjuncture. Where progress and victo- ries have been recorded, this has almost unfailingly been attributed to shifts in the content or force of state power. The tendency is to disparage and attribute static qualities to the movement. In Killion’s pioneering account15 of organised labour in Ethiopia and Eritrea, which remains unsurpassed in historical detail and analytical sophistication, this tendency constitutes the one objectionable attribute. Working-class formation in pre-revolution Ethiopia, according to Killion, was “incipient”, “incomplete”, and characterised by “false consciousness”16. “In contrast to workers in the European colonies”, Killion claimed, “Ethiopian workers’ organisations developed a nationalist political consciousness which frequently confused the interests of the state with their own economic interests”17. Partly as a result, Killion characterised the labour movement as persistently subordinated to the state. When workers staged successful strikes on the Djibouti-Ethiopian railway and in other foreign- owned enterprises, he interpreted these strikes as originating “in the contradic- tions between the interests of a non-capitalist ruling class and foreign capital”, rather than in the contradiction between labour and capital. Bluntly put, Ethiopian workers were being manipulated. Despite Killion acknowledging how, prior 1974

13. E. Screpanti, “Long Economic Cycles and Recurring Proletarian Insurgencies”, Review, 7, 3, 1984, 530. 14. Screpanti, “Long Economic Cycle”, 531. 15. T. Killion, “Workers, Capital and the State in the Ethiopian Region”, PhD thesis, Stanford University, 1985 16. Killion, “Workers, Capital and the State”, 2-12. 17. Killion, “Workers, Capital and the State”, 12.

Africa, N.S., I / 1, 2019 92 a r t i c o l i / a r t i c l e s

Ethiopian revolution, the labour movement had gained a degree of autonomy and had produced a growing militant strand, which permitted workers to have a “significant impact” on the unfolding process18, he insisted that subordination remained the salient feature. Killion is far from alone in making such interpretations. In Markakis’ semi- nal work Anatomy of a Traditional Polity, the labour movement appears only as a peripheral phenomenon, whose weakness is identified as one of the reasons for the depressed level of wages19. Other scholars have tended to agree. “The unions were far less significant than the local neighbourhood associations”, Clapham20 opined, while Keller argued that the labour movement “was never able to present a serious militant challenge to the status quo in the industrial sector”21. According to Lefort, the labour movement was “apolitical, heterogeneous and atomised”22, while Bahru has written that prior to March 1974, the central trade union confederation “had been notable for its lethargy rather than its militancy”23. Gebru, meanwhile, has dismissed the emerging working class in toto as “small, factional, and inconsequential, both qualitatively and quantitatively”24. Several other works have discussed the role of the labour movement in the context of the 1974 revolution. What these accounts have in common is that they subsume labour militancy under the broader revolutionary process and attribute the former to the latter. Gebru, for example, has claimed that the trade union movement was “relatively placid and dormant throughout the 1960s” and that its “torpor was shattered only amid the dramatic circumstances of the 1970s”25. Ma- rina Ottaway, meanwhile, was dismissive of the movement’s role, playing down its militancy as mere pretension. The labour movement, she claimed, functioned as a vehicle for petty-bourgeois and white-collar interests26, and opposed the incoming military government from a conservative position dressed in radical rhet- oric. This is a view that she and David Ottaway would reassert, criticising the trade union confederation – by that time in open conflict with the military regime – for being unable to compromise, and labelling its leadership incompetent, con-

18. Killion, “Workers, Capital and the State”, 12. 19. J. Markakis, Ethiopia, Anatomy of a Traditional Polity, Oxford, Clarendon Press, 1974, 174-177. 20. C. Clapham, Transformation and Continuity in Revolutionary Ethiopia, Cambridge, Cam- bridge University Press, 1990, 35. 21. E.J. Keller, Revolutionary Ethiopia. From Empire to People’s Republic, Bloomington, Indiana University Press, 1988, 149. 22. R. Lefort, Ethiopia. An Heretical Revolution?, London, Zed Press, 1983, 27. 23. Bahru Zewde, A History of Modern Ethiopia: 1855-1991, Addis Ababa, Addis Ababa University Press, 2008, 231. 24. Gebru Tareke, The Ethiopian Revolution. War in the Horn of Africa, New Haven and London, Yale University Press, 2009, 23. 25. Gebru Tareke, The Ethiopian Revolution, 23. 26. M. Ottaway, “Social Classes and Corporate Interests in the Ethiopian Revolution”, Jour- nal of Modern African Studies, 14, 3, 1976, 469-486. Admasie, Cycles of Mobilisation, Waves of Unrest 93 fused and opportunist27. They were not alone in treating the labour opposition as inauthentic and illegitimate. The process of radicalisation of the labour movement and the Confederation of Ethiopian Labour Unions’ (CELU) opposition to the incoming military government has often been characterised as the outcome of “infiltration” – denoting a mechanical and external form of control – by oppositional activists28. In more recent times, the key features of the depiction have persisted. A new round of conflict between the authorities and the trade union movement in the 1990s has been described by Praeg, who insisted that “one could not have expected a different fate” than the demise of what temporary autonomy the latter enjoyed29. Melakou, meanwhile, has suggested that no independent unionism whatsoever has existed on a national level in the past four decades30. The scholarly tendency to view the movement in disparaging terms has probably increased over time. Dessalegn Rahmato, writing in 2002, claimed that the trade union movement “has been kept as a docile instrument of state policy by successive governments since the 1960s”. He described successive governments’ persistent desire to exercise control over the movement as an irrational- ity which has given “the trade union movement a significance and power far beyond its actual potential”31. Ethiopian trade union history, he most tellingly summed up, has been “the result of false perceptions and an exaggerated sense of worth”32. But as will be demonstrated, this literature only tells one part of the story. To be sure, there have been periods where the Ethiopian labour movement has indeed been co-opted, dormant and impotent – at least on a macro-level. But equally, there have been several periods in Ethiopian history where a relatively autonomous, militant and impactful movement has emerged.

Cyclical movements

If the historiography of the Ethiopian labour movement is characterised by a curious tendency to deny workers autonomy and agency and to describe the movement in disparaging and static terms, the actual history of the movement is

27. M. and D. Ottaway, Ethiopia. Empire in Revolution, New York, Africana, 1978, 101-107. 28. See for example Lefort, Ethiopia, 30; Andargachew Tiruneh, The Ethiopian revolution, 1974-1987. A Transformation from an Aristocratic to a Totalitarian Autocracy, Cambridge, Cam- bridge University Press, 1993, 141, 175. 29. B. Praeg, Ethiopia and Political Renaissance in Africa, New York, Nova Science, 2006. 30. Melakou Tegegn, State and Civil Society. Ethiopia’s Developmental Challenges, Los An- geles, Tsehai, 2013. 31. Dessalegn Rahmato, “Civil Society Organizations in Ethiopia”, in Bahru Zewde and S. Pausewang, eds., Ethiopia. The Challenge of Democracy from Below, Uppsala, Nordic Africa Insti- tute, 2002, 114. 32. Dessalegn Rahmato, “ Civil Society Organizations in Ethiopia”, 114.

Africa, N.S., I / 1, 2019 94 a r t i c o l i / a r t i c l e s characterised by dynamism and displays a very different peculiarity: the appearance of cyclical movements of mobilisation, within which ebbs and flows in in workers’ autonomous activity and industrial-political contention, shifts in the movement’s strategic orientation, and advances and retreats in the position of labour have occurred. At certain points of this cyclical movement, wave-like mo- tions of labour unrest have emerged and crested, resulting in the type of proletarian insurgencies that Screpanti described, before having tailed out in a renewed period of dormancy and industrial tranquillity. Three such cycles can be identified: one lasting from the beginning of the 1960s to the mid-1980s; a second lasting from the late 1980s to the 2000s; and an emergent contemporary cycle beginning in the 2010s. The first, and hitherto most substantive and protracted, cycle began as the Ethiopian labour movement emerged as a country-wide phenomenon. At the beginning of the 1960s, unions were not legally permitted, but workers’ associations modelled in the shape of collective self-help organisations had appeared and turned into unions in all but name. Some of them were quite successful in press- ing for improved conditions, and this propelled an intensified process where associations “sprouted up all over the empire”33. As the associations multiplied, strikes became frequent. Meanwhile, a cooperative forum for workers associations was formed, and although not formally legal, it was cautiously tolerated by the government as the lesser of evils. The heightened activity among workers and “the growing number of industrial disputes”34 led to the passing of the 1962 labour law, which permitted the registration of unions. As a result CELU was formed the year after. But no sooner had CELU been founded than it came into conflict with the government – a conflict that would demarcate the tolerated limits to its autonomy. By forcing the original president of the confederation to resign and having him replaced with a more deferentially inclined successor, the government had re-set the organisation on a very pragmatic footing. There would be no ultimatums or offensive demands from the confederation in the decade to come, neither any forceful protests when the government engaged with workers and union representative in a heavy-handed manner, as it frequently did. But the pragmatism and lack of militant energy in the leadership of the confederation was not representative of the full movement. Industrial contention and unrest on workplace and basic union level took place unabated throughout the 1960s and early 1970s. In most cases activities were of an unsanctioned nature or organised by local unions, but in some cases central leaders supported strikes underhandedly35. Yet, there were problems in the

33. Syoum Gebregziabher, The Rise of CELU and the EFE Under Haile Selassie’s Regime, Mimeo, 1969, 27. 34. A.M. Zack, “Trade Unionism Develops in Ethiopia”, in J. Butler and A.A. Castagno, eds., Transition in African Politics, New York, Praeger, 1967. 35. Interview with Beyene Solomon, Addis Ababa, 20 Feb, 2016. Admasie, Cycles of Mobilisation, Waves of Unrest 95 relationship between the central leadership, local workers’ representatives, and the rank-and-file. Dissatisfaction with the central leadership’s bureaucratic inclinations and the strategic orientation it pursued was widespread36. It was accused of subverting internal democracy and showcasing excessive deferentialism, overturning Gen- eral Assembly decisions to call general strikes and demonstrations on two occa- sions37. The leadership itself had begun to voice concern about “being unable to cope with the demands of the workers”, by the mid-1960s, which, it considered “often quite unrealistic”38. It reported that there had been “much disagreements between CELU and its affiliates” – in fact “perpetual strife was admittedly taking place” – with “one of the major criticism [being] that [CELU] was not doing enough for its membership”39. A representative of the international trade union movement based in Addis Ababa reported home that CELU’s leadership, in turn, was “divided in different groups”40. But such divisions and frictions, resulting from contention over strategic orientation, in fact prevailed throughout the structures of the movement. By the early 1970s, an organised opposition had emerged within the labour movement. Spearheaded by 30 basic unions, it briefly attempted to disaffiliate from CELU to set up a parallel confederation, but was barred from doing so by the government41. The opposition instead embarked on a campaign to move the central confederation in a more militant direction by canvassing the rank-and-file and challenging the central leadership within CELU’s institutions42. In addition to unions that had attempted to disaffiliate, it had a growing number of supporters within the rank-and-file and among representatives in CELU’s ruling bodies. The resolution of this division would only come about as the result of an outbreak of a major wave of labour militancy in the period preceding the 1974 revolution. In 1973 striking workers of the Diabaco Cotton factory and the Central Print- ing Press had been denied support from the central confederation43. But when the

36. Emmanuel Fayessa Negassa, “The Confederation of Ethiopian Labour Unions and its First General Strike (March 7-11, 1974): Causes and Impact”, MA thesis, Cornell University, 1977, 72. 37. Seleshi Sisaye, “Labor in Contemporary Ethiopian Politics. The Case of the Confedera- tion of Ethiopian Labour Unions and Its First General Strike”, paper presented at the 71st Annual Meeting of the American Sociological Association, New York, 1976. 38. CELU, “Progress Report and Future Projected Activities of CELU”, Oct. 22, 1965. ICFTU Archives, Amsterdam, IISH. 39. CELU “Progress Report and Future Projected Activities of CELU”. 40. Lennart Kindström [Letter] to P.H. de Jonge, Feb. 2, 1966, ICFTU Archives, Amsterdam, IISH. 41. Beyene Solomon, Fighter for Democracy. The Saga of an Ethiopian Labour Leader, Bal- timore, Publish America, 2010, 167. 42. Fayessa Negassa, “The Confederation of Ethiopian Labour Unions and its First General Strike”. 43. Seleshi Sisaye, “Urban Migration and the Labor Movement in Ethiopia”, in Proceedings of the Fifth International Conference of Ethiopian Studies, University of Illinois, 1979, 698.

Africa, N.S., I / 1, 2019 96 a r t i c o l i / a r t i c l e s large union of the state-owned Commercial Bank of Ethiopia called another strike that same year, the opposition piled pressure on the central leadership to intervene. CELU’s General Council vowed to take solidarity measures44, and it was reinforced by an enthusiastic response from basic unions, several dozens of which sent letters to the CELU leadership expressing their determination and readiness to intervene45. The labour movement had entered into a head-on conflict with the state, and it came out victorious: the government retreated and conceded to the bank workers’ demands46. This achievement put the deferential strategy the leadership had pursued in negative perspective, as it made apparent that workers were capable of extracting significant concessions. In doing so, it led to a sharp increase in militant self-confidence and in March 1974 the General Council called a country-wide general strike. Again, the strike forced the government to concede to CELU’s demands. Over the coming couple of years, workers from hundreds of workplaces would strike. Reports initially told of new strikes “almost daily”47. Militancy also expressed itself in other forms of industrial contention. In the compound of the Labour Relations Board crowds of workers appeared daily, blocking the entrance, beating up employers’ representatives, and even attempting to set the buildings on fire48. General strikes were again attempted – with differing degrees of success – in the September months of 1974, 1975 and 1976. Primarily because of the severe repression – in response to the September 1975 strike attempt alone, it was reported that more than 100 people were shot and around 1600 people were ar- rested49 – none of them could muster the same strength as the March 1974 strike, but each attempt saw dozens of workplaces emptied of workers. After the imperial Ethiopian government was removed and replaced, industrial contention was reinforced by political strife. The deferential leadership of the labour movement was soon ousted, and in 1975 the radical oppositional labour movement took over control of the confederation. The confederation was outlawed shortly after, but that did not put an immediate end to strike activity and industrial contention. It was only with the coming of massive repression against

44. CELU, ስለ ኢትዮጵያ ንግድ ባንክ የሠራተኞች ማኅበር የተሰጠ ውሳኔ [Resolution] 1973, CETU Archi- ves, Addis Ababa. 45. See for example, letters from the unions of Ethiopian Petroleum Workers; Indo-Ethiopian Textile Factory; Michell Cotts Union; HVA Union; Ethiopian Airlines; Addis Ababa Bank; United Wood Workers; and Addis Ababa Hilton, in CETU archives. 46. Andu-alem Aragie, “A Short History of the Employees of the Commerical Bank of Ethio- pia from 1964 to 1997”, BA thesis, Addis Ababa University, 1999, 48. 47. P.D Wyman, Apr. 17, 1974, PLUSD, https://wikileaks.org/plusd/cables/1974ADDIS04184 _b.html (accessed 5 June, 2017). 48. P.D. Wyman, “Labour Situation”, Cable 1974ADDIS12489_b from the US Embassy in Addis Ababa, Oct. 18, 1974, PLUSD, https://wikileaks.org/plusd/cables/1974ADDIS12489_b.html (accessed 6 June, 2017). 49. F.L. Luyimbazi, “Report on the Main Activities of ICFTU Research Office and General Labour Situation in Ethiopia, From October 1974 to date, January 9”, 1976, ICFTU Archives, Am- sterdam, IISH. Admasie, Cycles of Mobilisation, Waves of Unrest 97 the labour movement that strikes began to peter out. By late 1978, hundreds, probably thousands, of workers and workers’ representatives had been killed or imprisoned50. Industrial tranquillity had been re-established, together with a new, deferential, central confederation – the All-Ethiopian Trade Union (AETU). The first cyclical movement had tailed out and come to an end. The larger part of the 1980s was a period of severely curtailed autonomy, low levels of mobilisation and little activity. The 1983 strike at the Barottolo Textile Factory in Asmara constitutes the single notable exception – which tellingly resulted in harsh repression51. By the end of the 1980s, however, a number of events signalled the re-emergence of some degree of autonomy and militancy, and the onset of a new cyclical movement. In 1989, for example, 400 textile workers in Addis Ababa walked off the job protesting workplace transfers52. According to CELU’s former president, the following years saw the beginning of a workers’ “uprising” that spread from workplace to workplace in the industrial corridor stretching southwards from Addis Ababa53. The uprising was allegedly set off by the arrest of workers and union representatives who had protested against poor pay and conditions. Eventually, it came to include workers from diverse sectors, such as transportation, construction, trading, beverage and textile manufacturing, as well as fibre and sugar plantations. Evidence from the archives of the Ministry of Labour and Social Affairs in Addis Ababa also testifies to a re-emerging vola- tility in industrial relations during this period. In April 1991, for example, fuel workers at Total Mer Rouge declared their resolution to strike should their demands for a better collective agreement not be met, and threatened management with “bloodshed” should their rights not be respected54. Workers also directed their ire against the trade union bureaucracy. In the industrial district of Akaki, demonstrating workers attacked the residence of the president of the central confederation – now re-baptised the Ethiopian Trade Union (ETU) – in early 199155.

50. See J. Wiebel & S.A. Admasie “Rethinking the Ethiopian ‘Red Terror’: Approaches to Political Violence in Revolutionary Ethiopia”, [forthcoming article], for a detailed discussion on this. 51. Goitom, “Important dates in the history of Eritrean workers”. [Translation from ፍዳሜታት 180, 15 May 1985: 3-7], 1985, Tom Killion Papers, Hoover Institution, Stanford University; Com- mittee of the High Authority in the Region of Eritrea, በአሥመራ ጨርቃ ጨርቅ ፋብሪካ በ22.6.75 ተደርጉ የነ በረው የሥራ ማቆም አድማ በማስመልክት በኤ.ክ.ሀገር ከፍተኛ ባለሥልታኞች ያሉበት ኮሚቴ መርምሮ በ26.6.75 የፋብሪካው ጠቅላላ ሠራተኞች በተገኙበት የተሰጠ ውሳኔ, 1983, CETU Archives, Addis Ababa. 52. ETU, በቤሔራዊ ጨርቃጨርቅ ኮርፖረሽን የአደይ አበባ ድርና ማግ ፋብሪካ የሰብሰባ ቃለ ጉባኤ. [Minutes, Sept. 21, 1989], CETU Archives, Addis Ababa. 53. Beyene Solomon, የኢትዮጵያ ሠራተኞች አንድነት ማህበር: ውልደት ዕድገት ውድቀት, Addis Ababa, S.n, 2012. 54. P. Graf [Letter] to Ministry of Labour and Social Affairs, Apr. 23, 1991, Ministry of La- bour and Social Affairs Archives, Addis Ababa; Tadesse Tamrat [Letter]to Total Mer Rouge, Mar. 29, 1991, MoLSA Archives, Addis Ababa; Kedeme Teshome [Letter] to Total Mer Rouge, May 7, 1991, MoLSA Archives, Addis Ababa. 55. Interview with Hailu Ourgessa, Geneva, May 28, 2017.

Africa, N.S., I / 1, 2019 98 a r t i c o l i / a r t i c l e s

As the manifestations of re-emergent workers’ militancy grew in numbers, the macro-political setting again changed when the Ethiopian People’s Revolu- tionary Democratic Front (EPRDF) marched into Addis Ababa in May 1991 and established a new government. ETU was abolished by fiat, and local unions were ordered to elect new leaderships. This did not put an end to unrest. In the coming months, several strikes occurred, including in mines, printing presses, factories, and freight corporations56. This was followed by two large strikes from workers in the strategically important Commercial Bank of Ethiopia and Ethiopian Airlines. Meanwhile protests occurred among factory, dock, plan- tation, construction, trade and care workers, among others, while “numerous actions” were reported from entrenched workers57. The early years of EPRDF rule saw liberalisation measures that included retrenchments of several thousand workers58, deterioration of employment conditions, and sharply increasing living costs. In tandem with this, unions that refused to fall in line with the new dispensation faced renewed repression59 and workers’ representatives across the country faced another wave of dismissals – in the two years 1991 and 1992 more workers’ representatives were dismissed from their jobs than the prior 9 years combined60. In response, workers attempted to con- struct defensive barriers against the pressures they faced. This had two aspects. The first consisted of strikes, petitions and protests at a workplace level. The second consisted of an attempt to re-establish a central confederation to coordinate workers’ efforts. To this end, workers networked in impromptu constellations and pressure groups, eventually convincing the government of the necessity to permit the reestablishment of a central confederation61. However, the government remained intent on steering the process in a direction that would see it retain influence over such a confederation. In November 1993, the Confederation of Ethiopian Trade Unions (CETU) was established. At the founding congress, workers’ representatives had managed to counter the government’s efforts to create a wholly co-opted organisation62, but this came with its own hazards, as it created suspicion of the motives of the new

56. Ethiopian Herald Aug. 2, Oct. 16, Nov. 23, Dec. 5, 1991 57. Ethiopian Herald Oct. 16, 1991; US Department of State, “Country Report on Human Rights Practices for 1993”, Washington DC, 1994. 58. Hishe Hailu, “An Assessment of the Process of Privatization in Ethiopia”, MA thesis, Ad- dis Ababa University, 2005. 59. The unions of the Akaki Textile Factor and the Ethiopian Electric Light and Power Author- ity are two examples. 60. Ministry of Labour and Social Affairs, ማህበር መሪዎችን ከሥራ እገዳ ወይም የሥራ ዝውውር መመሪያ, 1982-1993. 3 Archival boxes, Addis Ababa. MoLSA Archives. 61. Interview with Hailu Ourgessa; CETU, “The Legal Aspect of the Dispute Between the Two Factions of the Confederation of Ethiopian Trade Unions (CETU) and the Position of the Transitional Government of Ethiopia in this Respect”, [Memorandum], 1995, MoLSA, Addis Aba- ba, MoLSA. 62. CETU, “Trade Union Situation in Ethiopia: Factual Documentation”, [Memorandum], 1995 MoLSA, Addis Ababa; R.H. Sikazwe, “Report on the Inaugural Congress of the Restructured Admasie, Cycles of Mobilisation, Waves of Unrest 99 confederation. Within nine months, when CETU felt compelled to voice its opposition to the increasingly harsh effects of the government’s structural adjust- ment programmes63, the stage was set for a showdown. Despite strong signals coming from the basic unions and workers’ representatives that they were ready to defend the confederation through a general strike, this alternative was not seri- ously entertained by CETU’s central leadership64. Instead of a strategy based on mobilising the rank-and-file, it decided to pursue a strategy based on defending the confederation in courts. This proved to be deleterious. Within three months the courts had frozen the confederation’s assets and operations, and banned its president from entering his offices. Two years of legal wrangling later, CETU was altogether dissolved by law and a new confederation with the same name was established – this time under the control of deferential government loyalists who would openly refute the notion that strikes could ever resolve industrial disputes65. In the meantime, control over basic unions and industrial federation had progressively shifted to the same group of government-aligned moderates. Two decades of inertia and despondency followed. In 2017, however, industrial tranquillity came to an abrupt end, when a wave of strikes tore through the economy66. Over the Ethiopian year that ended in September 2018, at least a dozen major strikes occurred, involving more striking workers than in any other year since the 1970s. It also involved CETU re-emerging as a combative defender of workers’ interest, threatening to call a general strike should a harsh labour bill proposed not be amended. While this wave is still ongoing at the time of writing, it does appear that a third cyclical movement is in its initial stages. Judging by the force of energies released, it holds the potential of becoming a most substantial movement, surpassing, at the very least, the second cycle. Summing up the discussion on the cyclical movements, it is noteworthy how aggregate levels of industrial contention on the micro-level has, broadly speaking, concurred with macro-level combativity of central institutions of the labour movement to form three cycles. A periodisation of relative levels of labour unrest, based on a measurement of the number of recorded strikes and expressions of unrest in Ethiopia over the past 50 years, reveals a cyclical pattern of mobilisation similar to that described above. The shifting general level of unrest is illustrated in the schematic periodisation below.

Ethiopian Trade Union Movement, Addis Ababa, 26-29 November, 1993”, ICFTU Archives, Am- sterdam, IISH. 63. CETU, “የኢሠማኮ ሥራ አስፈጻሚ ኮሚቴ የአቋም መግለጫ አወጣ”, የሠራተኛው ድምፅ 1,3, 1994. 64. Interview with Dawi Ibrahim, Amsterdam, May 20, 2015; Interview with Hailu Ourges- sa. 65. John W., “An Opponent of Strikes in Charge of a Labour Union”, Reporter, 5, 241, 18 Apr. 2001, 11. 66. See S.A. Admasie, “Amidst Political Recalibrations. Strike Wave Hits Ethiopia”, Labor and Society, 21, 2018, 431-435.

Africa, N.S., I / 1, 2019 100 a r t i c o l i / a r t i c l e s

Figure 1: Schematic periodisation of general levels of unrest.

Source:author’scompilation67 Source: author’s compilation.67 Factorsatwork FactorsWithinthe atoutlines workofthecyclesdescribed,anumberoffactorsanddynamicsthathavegenerated theshiftingtrendscanbeidentified.Tobeginwith,andonaquitebasiclevel,thepropellingforce behindmilitantpracticesandpositionaladvanceshasgenerallybeenthepressureemanatingfrom therankWithinͲandͲ file.theTh outlinesiscanbeillustrated of the withcyclesanumber described,ofexamples, a numberbeginning ofin thefactorsfoundation andof dy - namicslegaltrade thatunionism have ingeneratedEthiopia.The thelegalisation shiftingof trendstradeunionism can bewas identified.not–unlike whatTo beginhasbeen with, andconveyed on a inquitetheliterature basic 68level,–so muchthe thepropellingoutcomeof forcethework behindofliberals militantinthe governmentpracticesor and positionalinternational advancespressureas hasofthe generallydisruptioncreated been andthe acutepressurethreats emanatingposedbyworkers fromengaged the rank-in and-file.wildcatunrest. ThisThere canis evidencebe illustratedofsenior governmentwith a numberofficials ofacknowledging examples,as beginningmuchtovisiting in the foundationrepresentatives ofof legaltheinternational trade unionismtradeunion in Ethiopia.movement inThe1962. legalisationOneofficialexpr ofessly tradeworried union - ismabout was ‘growing not – agitations unlike whatamong hasthe workersbeen conveyed who had staged in the several literature downͲtools68 – and so walkmuchͲout the outcomeactionsrecently’, of thenoting workthat ofa liberalsgoodnumber in theofworkers' governmentassociations or internationalhadbeguntocreate pressuretrouble as ofwith theemployers. disruptionAnother createdofficial andstated acutethat threats‘theexisting posedworkers by organisationsworkers engagedwerenow ingrowing wildcat into restless movements’, and that because the government ³was worried about the possible unrest.consequences Thereof illegal is evidenceand“unwarranted” of senioractions government´,itwouldtable officialsabillregulating acknowledgingtheformation as muchand operation to visiting of trade representatives unions, hoping ofthat the this international would stem the trade growing union restiveness movement69. The in 1962.legalisation One of official trade unionism expressly in Ethiopia wasworried thus caused about by urgent “growing pressures agitations asserted by the among the workersemergent wholabour movement.had staged several down-tools and walk-out actions recently”, notingWorkers that have a furthermoregood number been offound workers’ having perceivedassociations great injusticeshad begun in their to createrelations trouble with withemployers, employers. driving theAnother development official of industrial stated thatmilitancy “the and existing the type workersof collective organisations strategic wereorientation now requiredgrowing to intoengage restless in contestation. movements”, Morehous’ and research that onbecause workers' the attitudes government in the “was1960s, worriedforexample, aboutfound thethat possible‘unwillingness consequencestogiveawage ofincrease illegalisviewed and ‘unwarranted’onlyasanattempt ac - tions”,by the employerit would to table squeeze a billmore regulating of an already the outrageous formation profit andfrom operationthe workers' of toil’ trade70. This un - ions,resulted hopinginrecurrent that thiswildcat wouldaction. stemThroughout the growingitsearlyhistory, restivenessCELU’sPresident69. The atlegalisationthetimehas of tradeaffirmed, unionismrankͲand Ͳfilein Ethiopiapressureoften wascompelled thus precipitatedunionleadersto bycall urgentstrikeswhether pressuresthey foundassertedit by the emergent labour movement. %DVHGRQXQUHVWGDWDRQQDWLRQDODQGZR UNSODFH OHYHOV ZKLFKZDVFRP SLOHG DQG SUHVHQWHGLQ6$$GPDVLH³'\QDPLFVRI$VVHUWLYH/DERXU0RYHPHQWLVPLQ(WKLRSLD2UJDQLVHG /DERXU 8QUHVW DQG :DJHVLQD6R FLR +LVWRULFDO3HUVSHFWLY H´ 3K' GLVVHUWDWLRQ 8QLYHUVLW\ RI 3DYLDDQG8QLYHUVLW\RI%DVHOZ67. Based on unrest data onLWKUHFRUGVIRUWKH\HDUVDGGHG national and workplace levels which was compiled and presented in S.A.6HHIRUH[DPSOH7.LOOLRQ³:RUNHUV&DSLWDODQGWKH6WDWH´«FLW Admasie “Dynamics of Assertive Labour Movementism in Ethiopia. Organised Labour, Unrest and 6&ODYHULH³5HSRUWRQP\ Wages in a Socio Historical 9LVLW WR (WKLRSLD´ Perspective”, ,&)78$UFKLYHV$PVWHUGDP PhD dissertation, University of Pavia and ,,6+ University /*0RUHK of Basel, RXV 2018, ³(WKLRSLDQ with records /DERXU for 5HODWLRQV the years $WWLWXGHV 2010-2018 3UDFWLFHDQG added. /DZ´ 1RUWKZHVWHUQ8QLYHUVLW\68. See, for example, T. Killion, “Workers, Capital and the State”. 69. S. Claverie, “Report on my Visit to Ethiopia”, 1962, ICFTU Archives, Amsterdam, IISH. Admasie, Cycles of Mobilisation, Waves of Unrest 101

Workers have furthermore been found having perceived great injustices in their relations with employers, driving the development of industrial militancy and the type of collective strategic orientation required to engage in contestation. Morehous’ research on workers’ attitudes in the 1960s, for example, found that “unwillingness to give a wage increase is viewed only as an attempt by the em- ployer to squeeze more of an already outrageous profit from the workers’ toil”70. This resulted in recurrent wildcat action. Throughout its early history, CELU’s President at the time has affirmed, rank-and-file pressure often compelled union leaders to call strikes whether they found it prudent or not71. Union leaders, More- hous discovered, accepted frequent strikes as a concession to “the militancy of the mass of union members”72. “CELU has always been a restraining force, because usually the workers would want to use force”, former CELU top official Fissehatsion Tekie has stated corroborating this view73. Rank-and-file pressure in favour of a more militant strategic orientation was not, however, confined to the first cycle. The second cyclical movement also emerged from localised, illegal wildcat action unsanctioned by any central institution. For the larger part of the cycle, there was in fact no central coordination or leadership, and once such a leadership had been founded it did not encourage militant practices. Yet, that leadership remained subject to pressures, and its drift into conflict with the government was partly the result of such pressure74. More- over, once the conflict had begun, it was basic union representatives that called on the reluctant leadership to organise a general strike75. With regards to the emergent third cycle, what can be stated at this stage is that it has hitherto consisted, almost exclusively, of the form of localised wildcat activity that also marked the opening stages of the first two cycles. Some factors have intermittently checked and reversed workers’ mobilisation. One such factor has been the internal control mechanisms associated with the institutionalisation of trade unionism and the dynamics of bureaucratisation. The dual role of trade unionism has often been on vivid display in Ethiopia, as exemplified by a bakery union leader speaking to CELU’s newspaper Voice of Labour in 1970: Your union has tied our hands and feet. Before you organised us we used to get substantial concessions from employers by merely threatening a walkout. The employers knew that we did not kid when we threatened – the work, the flour even the equipment was not safe. Yet, now you have advised us that through the union we have first to di-

70. L.G. Morehous, “Ethiopian Labour Relations: Attitudes, Practice and Law”, Northwestern University, 1969, 70-71. 71. Beyene Solomon, Fighter for Democracy, 109. 72. Morehous, “Ethiopian Labour Relations”, 71. 73. Interview with Fissehatsion Tekie [recording], 1983, Tom Killion Papers, Hoover Institu- tion, Stanford University. 74. Interview with Hailu Ourgessa. 75. Interview with Hailu Ourgessa.

Africa, N.S., I / 1, 2019 102 a r t i c o l i / a r t i c l e s

scuss with our employers, then if we are not satisfied we have to submit our grievance to the [labour department] and that if we are not reconciled by the section we have to lodge our dispute with the Labour Relations Board. The case does not stop there, either employers or employees can appeal to the Supreme Imperial Court [or even] be peti- tioned to His Majesty’s [court]. Yes, your union has weakened our unity. We neither have the money, the time nor the manpower to follow up our disputes through these channels. Your union has only tied us to the benefit of our employers76. Through much of the first cycle, the confederation leadership and parts of the basic union officialdom worked hard to contain and subvert militant pressure. In the second cycle too, wildcat action – advertently or inadvertently – was short-circuited once a confederation was established. While rank-and-file pressures have repeatedly moved the central leadership into confrontations that were not of its choosing, the institutionalisation of formal trade unionism has tended to result in the diversion of pressures, and the containment and diminishment of industrial conflict. While its importance has frequently been exaggerated, another factor that has checked the movement’s momentum has been that of external subversion and repression. Employers and state institutions attempting to exercise close control over the labour movement has hardly, as Dessalegn has suggested, been irrational. On the contrary: a militant and assertive labour movement has repeatedly shown that it is capable of serious disruption. At times, subversion and repression has been highly effective, at other times, however, it has not, as can be demonstrated by the appearance and reappearance of labour militancy. In other words, external agencies have not mechanically determined the boundaries of the cycles of mobilisation. This fact also relates to the idea of “infiltration” as a factor driving the development of militancy. The political conjuncture has neither generated nor unleashed cycles. As argued in the above, the conflict of interest between employers and wage labourers is the fundamental cause of collective industrial action. Moreover, as suggested by Kelly, a number of conditions needs to be considered prior to the opportunity factor in explaining collective action – including the social construction and definition of collective interests, the emergence of collective coherence, and organisation. That being said, contextual factor, such as the political conjuncture, have indeed been important by varyingly enhancing or limiting the opportunity factor open to workers. When, for example, the state has been embroiled in broader socio-political crises, that has curtailed its power to displace and subvert militant pressure from workers, and to repress its expressions. However, the relationship between opportunity factors and outcomes is neither causative nor mechanical, but complementary. The notion of a mechanical relation between external factors and the strategic orientation of the labour movement can also be refuted on an empirical level in the Ethiopian case. To begin with, the sequence does not add up. In the cycles described, labour militancy has manifested itself in longer arches, reaching deeper into the past then

76. Voice of Labour, “The Ethiopian Labour Movement: The Aftermath of Birth”, የሠራተኛው ድ ምፅ, Special issue, 1970, 11. Admasie, Cycles of Mobilisation, Waves of Unrest 103

neither the subsumption of strategic orientation to the political conjuncture nor the idea of infiltration permit. The fundamental dynamics that have propelled the cycles must be found elsewhere. Growing labour militancy has tended to have a ratchet effect in Ethiopia, and strike waves, when unleashed, have created their own internal logic arching forward. But what have the dynamics of this process been? At this stage, the shifting conditions of Ethiopian labour require mention. These are not unrelated to workers’ mobilisation, since, to begin with, the purpose of the latter is indeed to affect the former, but also because the literature has indicated how achievements can affect strategic orientation. To what degree workers have been successful in affecting change, however, requires closer examination. Taking average real wages – that is, nominal wages deflated with the rate of the rise of the general price index – as the principal determinant of the position of labour, the following figure charts their development since the early 1960s.

Figure 2: Deflated Real Wage Index 1963-2010.

Source: Compiled by author from CSA’s annual Statistical Abstract 1963 to 2012/13.

The period between 1963 and 1975 saw both sustained labour mobilisation and rapid improvements in wages. This, by itself, does not conclusively demon- strate any causal relationship between workers’ mobilisation, on the one hand, and wage outcomes (achievements) on the other. To establish this, it must be shown that successful contention led to better conditions in concrete cases, generating the aggregate movement. As it turns out, there are plenty of examples that do precisely this. Strike resolutions and collective bargain agreement at workplaces such as the Dire Dawa Textile Factory in 1963; the HVA sugar factory in Wonji between 1954 and 1964; the St George Brewery, Ethio-Synthetic Textiles

Africa, N.S., I / 1, 2019

77. Killion, “Workers, Capital and the State”, 505-506; Bahru Zewde, “Environment and Capital: Notes for a History of the Wonji-Shoa Sugar Estate”, in Society, State and History. Selected Essays, Addis Ababa, Addis Ababa University Press, 134-136; CELU 3¼[ Õ )8 p Áp Õ · ćH õ^Â, Addis Ababa: CELU, 1972.. 78. ILO, “Employment and Unemployment in Ethiopia”, Geneva, ILO, 1972, 53-54. 104 a r t i c o l i / a r t i c l e s and Chandris Meat Canning Factory in 1972 are among the many examples where the outcome of contention had been high wage increments77. It is telling that an ILO report written in 1972 contended that union pressures had resulted in a situation where near-excessive wages had come to prevail, warning that this situation needed “to be monitored carefully”, to avert the danger that “collectively bar- gained wage spills over into the non-unionised sectors”78. The relationship between strategic orientation and achievement is not necessarily unidirectional. The possibility that a reciprocal relationship can prevail has been noted in the literature. In the history of the Ethiopian labour movement, readily identifiable achievements and militancy have played into and fuelled one another. In the course of the late 1950s, the 1960s, and the early 1970s, a string of major achievements were recorded as outcomes of workers’ struggles. This began with the establishment of the first workers’ associations and the organisation of the first successful strikes; proceeded with the legal recognition of trade unionism and the establishment of a confederation; the emergence of collective bargaining; a progressive shift in workplace balance of forces; and the emergence of the labour movement as a social force capable of extracting heavy concessions from capital and the state. Through each step of this process, achievements only appear to have further reenergised the movement and ratcheted up its force and demands. The best and most direct example is probably the manner in which, first, the successful bank strike of 1973 demonstrated conflict feasibility and released energies that resulted in the 1974 general strike, and, second, the successful general strike triggered a strike wave through the spring, and generated a proletarian insurgency that lasted through the following couple of years But it is not only in hindsight that the relationship between the strategic orientation of workers and achievements becomes evident. This was also, in fact, how representatives of Ethiopian employers of the time understood the process they were facing. In 1966, the Federation of Ethiopian Employers (FEE) Presi- dent Prakken complained over a high number of strikes, and speculated that this was “a consequence of previous illegal labour actions having remained unpunished”. “It is regrettable”, Prakken wrote, “that apparently no action was taken against the offending strikers concerned who infringed the law, as this seems to have encouraged others, considering numerous strikes and slow-downs that erupted in various industries during the past few weeks”79. Seven years later, however, the situation had not significantly improved, and new achievements were feeding into new unrest, and vice versa. The FEE president of the time,

77. Killion, “Workers, Capital and the State”, 505-506; Bahru Zewde “Environment and Cap- ital: Notes for a History of the Wonji-Shoa Sugar Estate”, in Society, State and History. Selected Essays, Addis Ababa, Addis Ababa University Press, 134-136; CELU 3ኛው በየሦስት ዓመት የሚከናወን ጠ ቅላላ ጉባዔ, Addis Ababa: CELU, 1972.. 78. ILO, “Employment and Unemployment in Ethiopia”, Geneva, ILO, 1972, 53-54. 79. J.M.J. Prakken, “Memorandum on Labour Developments”, ICFTU Archives, Amsterdam, IISH. Admasie, Cycles of Mobilisation, Waves of Unrest 105

Bekele Beshah, echoed his predecessor in lamenting how unpunished unrest led to growing boldness, and increasing wages. “Industrial relations developments in Ethiopia seem to have taken a turn for the worse”, he wrote, continuing to outline his view of the causal relationship: The most recent spate of illegal strikes… are in our view a cause for concern… It is obvious that there has been a growing misuse of union power which, unless checked by law … will jeopardise the future of collective bargaining… Very important issues here, are… the vulnerability they show of business in general and of the government and public at large to industrial action in key sectors... If no immediate legal reme- dies are implemented the chain of events very likely to be created could without any doubt jeopardise economic stability80. But as important as it is to point out the reciprocity of cause and effect in having driven gains and further militancy, the reverse has evidently also been on display. Following the repression of the labour movement in the 1970s, demobilisation followed. The movement, consequently, lost all clout and leverage. As a result, wages collapsed and the position of labour slumped, generating a vicious cycle of passivity and defeat. In the tail end of the cycle, there was only resigna- tion and docility. It would take a decade before new stirrings appeared. During the resultant second cyclical movement, collective efforts were generally limited to rear-guards attempt to establish defensive barriers, and few of those barriers held when facing stress. The strikes organised, as important as they were, tended to end in few concessions from employers. Moreover, there was only limited initial progress in acquiring organisational autonomy, which was subsequently reversed relatively expediently. But most important, perhaps, there were no readily identifiable achievements – no significant improvements for the rank-and-file, nor any major concessions extracted from employers or the state. In contrast, Ethiopian wage-workers saw their conditions and position progressively deterio- rate. Resistance was dispersed, and would eventually be defeated in separation. In other words, the momentum could only just begin to build before it was halted and reversed. Instead, and only a year after the establishment of the confederation, the labour movement was on a path of successive defeats, further curtail- ments, demobilisation, and declining levels of activity – all within a context of worsening conditions, mass retrenchments, and hollowing of rights. With regards to an emerging third cycle, there are already indications that there have been achievements recorded as the outcome of specific strikes81. Whether this will translate into aggregate achievements and reciprocal momentum remains an open question.

80. Bekele Beshah, 1973, “Speech by Ato Bekele Beshah at the Annual General Meeting of the Federation of Employers of Ethiopia – September 13, 1973”, CETU Archives, Addis Ababa. 81. S.A. Admasie, “Amidst Political Recalibrations”.

Africa, N.S., I / 1, 2019 106 a r t i c o l i / a r t i c l e s

Conclusions

The history of the Ethiopian labour movement, far from the static impression conveyed in the literature, is actually one of dynamism. It is the history of a movement that has repeatedly developed, relinquished and redeveloped the collective coherence and capacities required for meaningful industrial contention, and the militant willingness to do so. It is also the history of a movement that has, at times, had significant impact on labour relations, revealing the historical agency of Ethiopian workers, which has been forged in contention with subversive and antagonistic counteragents. The strategic orientation of Ethiopian workers acting collectively has developed within and been shaped by that contention itself, affirming Screpanti’s notion of the “positive cumulative effect of past experience in relation to actual claims”. Virtuous cycles of achievements and mobilisation and vicious cycles of despondency and defeat have intermittently prevailed. There have been ebbs and flows, waves and troughs in levels of mobilisation, militancy and industrial contention. What the literature has tended to describe, however, has at most been confined to the periods in between the crests of the cyclical movement – to the troughs and the ebbs. Scholars have tended to assume that the ebb is the default state, and the flow is the extraordinary phenomena – attributing the latter to external sources in general, and the political conjuncture in particular. But this ignores the fact that it is the same dialectic that has driven the full cyclical movement: crests as well as troughs. This dialectic has pitted the corporate needs of moderate labour officialdom against the material needs of the rank-and-file, and the requirements of state and employers for cheap and malleable labour against the trade union logic of pushing in the other direction. Both aspects have crystallised in the struggle over the strategic orientation of the labour movement – producing great shifts, cyclical movements, and repeated waves, rather than static tranquillity. Luisa Revelli

On o m a s t i c a d e l c o n t a t t o i t a l o -e r i t re o

Abstract

The linguistic-cultural contact between the Italian language and the different endogenous codes of Eritrea can be described as prolonged and bilateral, despite being at the same time characterized by heterogeneity, asymmetry and top-down dynamics. The reflections of this contact on the repertoire of proper names, which can be present or absent depending on the historical or cultural setting, prove the complexity of the socio-cultural representations naturally developed by the speakers as well as the motivations connected to the language planning choices made by the political élite. Founded on field research, and also on written and oral sources, this contribution aims to propose some onomastic paradigms useful to interpret, both in a diachronic and synchronic perspective, the articulate and multifaceted effects of the contact between Eritrean and Italian languages mainly in the area of Asmara, but also in the whole Eritrean territory, as well as the retroactive impacts on Ital- ian onomastics. k e y w o r d s : i t a l i a n ; c o n t a c t ; l i n g u i s t i c l a n d s c a p e ; o n o m a s t i c s .

Il contatto linguistico: modelli descrittivi

Le situazioni di contatto fra lingue e culture sollecitano a vari livelli nelle prime e nelle seconde cambiamenti che possono distribuirsi in modo omogeneo e simmetrico o invece coinvolgere prevalentemente o soltanto alcuni aspetti di una delle unità coinvolte1. Da un punto di vista strettamente linguistico, i passaggi monodirezionali, prevalentemente costituiti da prestiti e calchi, non richiedono necessariamente la compresenza della lingua modello e della lingua replica su un medesimo territorio né un’effettiva conoscenza della prima da parte dei parlanti. Gli esiti osservabili in situazioni di concreto bilinguismo sono invece, di norma e

1. S. Dal Negro, F. Guerini, Contatto. Dinamiche ed esiti del plurilinguismo, Roma, Aracne, 2007.

Africa, N.S., I / 1, 2019 108 a r t i c o l i / a r t i c l e s seppure con proporzioni molto variabili, bidirezionali, e vengono principalmente esplorati attraverso l’osservazione di comportamenti linguistici come l’alternan- za e commutazione di codice o l’analisi dei risultati di interferenze e ibridazioni2. In una prospettiva ecolinguistica, le dinamiche dei rapporti tra codici in contatto sono piuttosto esaminate tramite indicatori finalizzati a determinare la vitalità delle lingue di minoranza o a fotografare situazioni e tendenze di dominanza e subordinazione all’interno di repertori plurilingui. Fra tali indicatori – in genere particolareggiatamente declinati in categorie riferite a fattori demografici, econo- mici, culturali, affettivo-ideologici oltre che propriamente linguistici3 – la categoria dell’onomastica tende ad essere ignorata. Eppure, sebbene dotati di confini grammaticali in parte ambigui e statuti se- mantici non sempre circoscrivibili, i nomi propri sono prodotti idiolinguistici che, per le loro caratteristiche di rappresentatività sociolinguistica e peculiarità identitaria, simboleggiano l’atto della “nominazione” in tutta la sua complessità culturale4, potendo quindi assumere utilmente il ruolo di termometri delle configurazioni del contatto sia come indici della continuità storica o discontinuità evo- lutiva, sia come icone dell’immaginario linguistico dei parlanti. Il settore onomastico su cui gli effetti del contatto si riflettono con maggiore evidenza quantitativa e significatività qualitativa è probabilmente quello relativo ai nomi di persona: se, in effetti, l’impatto di lingue esogene può incidere sui repertori dei primi nomi anche in caso di relazioni solo virtuali o a distanza – come le ricorsive mode d’importazione di esotismi evidenziano per pressoché tutte le epoche e aree geografiche – più profonde e emblematiche possono rivelarsi tanto per la loro presenza quanto per la loro assenza le contaminazioni collocate in punti di effettiva convivenza di lingue, comunità e culture. In tali contesti possono peraltro emergere non soltanto fenomeni di migrazione di forme prenominali e cognominali, ma anche – in caso di rilevante distanza iniziale della sintassi delle formule appellative – esempi di rifunzionalizzazione delle componenti o vera e propria ristrutturazione dei sistemi di denominazione5. Le dinamiche del contatto affiorano d’altra parte anche in ambiti onomastici diversi da quelli dell’antroponimia: i toponimi consentono, ad esempio, oltre che un’osservazione diacronica delle sedimentazioni di relazioni interlinguistiche concluse, di interpretare l’affermarsi di forme endonimiche e varianti esonimiche rivelatrici dello status assunto dai codici compresenti su un medesimo repertorio;

2. G. Berruto, “Confini tra sistemi, fenomenologia del contatto linguistico e modelli del code switching”, in G. Iannaccaro, V. Matera, eds., La lingua come cultura, Torino, UTET, 2009, 3-34. 3. Cfr. B. Moretti, E.M. Pandolfi, and M. Casoni, eds., Vitalità di una lingua minoritaria. Aspetti e proposte metodologiche, Bellinzona, Osservatorio linguistico della Svizzera italiana, 2011. 4. A proposito della salienza della nominazione nell’ambito della cultura umana cfr. G. Rai- mondi, “Onomastica”, in M. Salviati, L. Sciolla, eds., L’Italia e le sue regioni (1945-2011), III: Pratiche, memorie e varietà linguistica, Roma, Istituto dell’Enciclopedia Italiana, 2015, 607-662. 5. Si tratta di quanto avvenuto, ad esempio, a seguito del contatto fra i diversi popoli dell’Ita- lia antica. Revelli, Onomastica del contatto italo-eritreo 109 gli odonimi, nel delineare la memoria simbolica incarnata dalla scelta di intitolazioni dedicatorie, encomiastiche o ideologiche, riflettono interventi di politica linguistica e politica tout court emblematici degli equilibri interlinguistici e della veicolazione delle loro rappresentazioni; i crematonimi6 riflettono le rappresentazioni di prestigio percepite dai parlanti e le sensibilità linguistiche a loro attribu- ite in quanto fruitori; gli etnonimi, specie nella forma dei blasoni popolari asse- gnati fra comunità in contatto o a loro specifiche sotto-categorie di componenti, possono infine evidenziare dinamiche sociali, tensioni interculturali ed asimme- trie a livello esplicito invece criptate o tabuizzate. I dati onomastici rilevabili nei contesti di contatto plurilingue possono assumere, insomma, se non funzioni predittive degli scenari repertoriali delle aree di contatto, almeno il ruolo di indicatori utili per l’interpretazione di configurazioni e rappresentazioni sociolinguistiche e culturali passate e presenti. A partire da indagini sul campo e testimonianze tratte da fonti scritte e orali, storiche e contemporanee7, il contributo si prefigge allora la messa a fuoco di alcuni paradigmi di prospettiva onomastica ritenuti utili a interpretare gli articolati e multi- formi effetti del contatto italo-eritreo, assunto dal punto di vista della lingua italiana nel contesto urbano di Asmara e prevalentemente in relazione alla realtà tigrinofona, ma con indispensabili intrecci ed incursioni nella dimensione delle lingue diffuse sull’intero territorio eritreo e delle loro presenze riflesse nello spazio italiano.

Caratteristiche ed esiti del contatto italo-eritreo

Quello fra la lingua italiana e i diversi codici presenti sul territorio eritreo8 può essere considerato un esempio di contatto prolungato, tuttavia caratterizzato

6. Per evitare appesantimenti terminologici si accoglie qui l’uso generico di crematonimo per “nome di oggetto prodotto dall’uomo, in particolare in ambito commerciale” come definito in E. Caffarelli, C. Gagliardi, “Terminologia onomastica”, AVSI. Archivio per il Vocabolario Storico Italiano, 1, 2018, 11-55. 7. Le ricerche cui si fa riferimento, inscritte nell’ambito del Progetto “Plurilinguismi e linguistica migrazionale: profili diacronici e varietistici dell’italiano d’Eritrea” condotto da chi scrive presso il Dipartimento di Scienze Umane e Sociali dell’Università della Valle d’Aosta, si basano su fonti scritte non convenzionali (scritture diaristiche, epistolari, scolastiche, archivistiche minori) e sull’osservazione naturalistica, condotta attraverso l’integrazione di protocolli diaristici con pratiche di osservazione strutturata. La documentazione è integrata da un corpus di 32 autobiografie sociolinguistiche ed etnotesti elicitati attraverso l’utilizzo veicolare della lingua italiana, dai parlanti consultati spontaneamente acquisita in ambito familiare o lavorativo oppure appresa con buoni livelli di padro- nanza in contesti d’istruzione formale e comunque collocata all’interno di repertori fortemente plurilingui, in cui a una tendenziale dominanza tigrinofona si può per ragioni biografiche o pragmatiche sostituire la prevalenza di codici come l’arabo, il tigré, l’amarico, l’inglese e l’italiano stesso. 8. Oltre ai nove idiomi locali, riconducibili a tre differenti sistemi di scrittura (ge’ez, arabo e latino), vanno naturalmente considerate le parlate miscidate descritte in letteratura come pidgin (R. Siebetcheu, “La varietà semplificata di italiano nel Corno d’Africa in epoca coloniale: un indigenous talk?”, in C. Carotenuto, E. Cognigni, M. Meschini, eds., Pluriverso italiano: incroci lingui-

Africa, N.S., I / 1, 2019 110 a r t i c o l i / a r t i c l e s da disomogeneità, verticalità e asimmetria. Da un lato, i pur intensivi rapporti che hanno contraddistinto l’epoca coloniale non si sono distribuiti sull’intera popola- zione, avendo riguardato – sul versante italiano – soltanto i singoli individui e nuclei familiari, peraltro spesso dialettofoni, coinvolti nei processi migratori e avendo d’altra parte escluso – sul territorio eritreo – un’ampia porzione di popo- lazione rimasta estranea ai circuiti sociali e lavorativi collegati alla comunità italofona. D’altro lato, i contesti di contatto non hanno dato come risultato la sosti- tuzione di lingue: gli italiani migrati con progetti transitori o definitivi hanno conservato la loro lingua d’origine che i nativi hanno eventualmente spontaneamente acquisito o formalmente appreso in contesti educativi9, in ogni caso in aggiunta rispetto al proprio personale repertorio. Con la fine dell’epoca coloniale, il ruolo dell’italiano in Eritrea, progressivamente indebolito, è rimasto quello di codice veicolare dei rapporti economico-commerciali con la Penisola, di lingua di cultura e di prestigio circoscritta a una ristretta cerchia di parlanti e di parlata etnica all’interno della comunità dei migranti italiani, comunità in cui peraltro i fenomeni di trasmissione intergenerazionale si sono andati gradualmente ridu- cendo. Nel contesto spiccatamente variegato e pluriculturale del territorio eritreo, i ruoli svolti dalla lingua coloniale sono stati pian piano assorbiti dall’inglese10, che ridisegnando le gerarchie diglottiche non soltanto ha oggi assunto il ruolo di lingua veicolare della comunicazione interetnica, ma – anche in seguito al suo inserimento in ambito scolastico – si va generalizzando come codice della quoti- dianità anche nei contesti intrafamiliari. Gli spazi dell’italiano possono, per questa ragione, essere oggi utilmente esplorati, piuttosto od oltre che negli usi dei parlanti, attraverso le rappresentazioni che questi simbolicamente hanno matura- to nel passato e ancora oggi esprimono attraverso l’onomastica.

Appellativi personali

Le scelte di nominazione di un nuovo nato vedono l’intrecciarsi di ragioni eufoniche e fonosimboliche a più profonde motivazioni di ordine culturale, sociale, psicologico: a queste si possono affiancare e anche sovrapporre spinte linguistico-culturali e percorsi migratori in lingua italiana, Macerata, Eum, 2018, 174-189) e le diverse varietà semplificate e d’apprendimento delle lingue d’importazione, italiano e inglese in primis. 9. Per il ruolo storicamente svolto dalla scuola italiana si rimanda a G.P. Carini, R. La Corda- ra, Storia della scuola italiana in Eritrea, Ravenna, Giorgio Pozzi Ed., 2014. L’attuale sistema di funzionamento dell’istituzione scolastica statale di Asmara – oggi frequentata da circa 1.200 allievi per la quasi totalità di nazionalità eritrea che studiano l’italiano come lingua straniera – è stata descritta da A. Pagliarulo, “La diffusione della lingua italiana in Eritrea: situazione attuale e prospettive future”, Itals, 2, 6, 2004, 51-84 e più recentemente da M. Longo, “L’insegnamento dell’italiano presso la scuola statale di Asmara: efficacia e criticità”, Education et Sociétés Plurilingues, 45, 2018 (in corso di stampa). 10. L. Revelli, “Reti dell’italiano nel continente africano (e intrecci dell’area eritrea)”, in V. Noli, ed., L’italofonia in rete. Annuario della Società Dante Alighieri, Roma, Società Dante Ali- ghieri (in stampa). Revelli, Onomastica del contatto italo-eritreo 111 stico-identitarie soprattutto nei contesti in cui diversi codici si trovino a svolgere ruoli concorrenti, ciò che spesso accade nelle situazioni di contatto, specie se scaturite da esperienze migratorie11. Nello specifico del contatto italo-eritreo, le dinamiche d’imposizione dei prenomi risentono poi necessariamente di aspetti legati alla complessa e non sempre armonica storia delle relazioni fra i due Paesi: non è per questa ragione sufficiente distinguere schematicamente – secondo la tradizione degli studi dedicati all’onomastica di confine – tra fase di contatto e fase di post-contatto, perché il contatto in oggetto è mutato sulla base di movimenti migratori non privi di criti- cità, di correnti che nel corso del tempo si sono modificate fino a capovolgere gli scenari di destinazione, di flussi che hanno compreso e comprendono scenari estremamente diversificati, potendo includere il trasferimento di donne e uomini soli così come di famiglie o comunità intere; l’inserimento o l’isolamento in gruppi preesistenti; la presenza ufficiale o clandestina sul nuovo territorio; la creazione di nuclei etnicamente omogenei o invece misti e insomma una quantità di variabili capaci di promuovere o al contrario compromettere le possibilità di me- scolanza dei repertori antroponimici in contatto. In dimensione storica, un interessante esempio di mancata osmosi intercultu- rale è costituito dall’elenco dei prenomi imposti ai bambini cattolici censiti nei registri battesimali della Chiesa della “Beata Vergine del Rosario” di Asmara (la cosiddetta Cattedrale)12: tutte le forme documentate per i 246 nati nel periodo 1891-1904 si inscrivono, infatti, nella trafila italofona non soltanto in presenza di genitori di origine italiana, ma anche quando a portare a battesimo il nuovo nato siano coppie interetniche, coppie in cui peraltro la componente autoctona è sempre quella femminile. La possibilità di individuare soluzioni di sintesi tra le due differenti tradizioni linguistiche, per esempio attraverso l’imposizione al figlio di più prenomi di differente trafila, è stata in quel periodo evidentemente esclusa o preclusa per un criterio di patrilinearità onomastica che, pur comportando per le madri eritree la rinuncia al repertorio prenominale della propria tradizione, vali- dava un principio di discendenza paterna condiviso da entrambe le culture di provenienza13 e doveva comunque probabilmente essere avvertito come promet- tente per le prospettive di vita dei nuovi nati14.

11. Cfr. ad es. la testimonianza di Kouakou Kouassi, “Nomination et identité dans la migration”, Le Coq-héron, 175, 2003, 54-61. 12. L. Revelli, “Mimetismi asmarini, tra Otto- e Novecento”, RIOn. Rivista italiana di onomastica, 24, 2, 2018, 823. 13. Come evidenzia Giulia Barrera, proprio il fatto che anche la cultura tradizionale tigrina si basasse sulla discendenza paterna ha determinato dinamiche complesse: le donne eritree “per sostenere i principi tigrini in materia di discendenza e identità, educarono i figli ad identificarsi con la cultura dei colonizzatori: dovremmo chiamarla resistenza o complicità?”, G. Barrera, “Patrilineari- tà, razza e identità: l’educazione degli italo-eritrei durante il colonialismo italiano (1885-1934)”, Quaderni storici, 109, 27(1), 2002, 46. 14. L’ampio potenziale di criticità identitaria derivante dalla nominazione dei nati da couples mixtes è messo in rilievo per l’epoca contemporanea ma in diversi contesti geografici da C. Desout-

Africa, N.S., I / 1, 2019 112 a r t i c o l i / a r t i c l e s

Non è, d’altra parte, privo di significato il fatto che il repertorio dei prenomi imposti sul territorio italiano nel medesimo periodo, pur includendo soluzioni rap- presentative del contatto, non tragga mai ispirazione dall’inventario dei primi nomi eritrei: gli italiani che in epoca coloniale sceglievano di celebrare le imprese d’oltremare attraverso la nominazione di un figlio preferivano ispirarsi alle denominazioni dei personaggi italiani che le simboleggiavano, eventualmente anche attraverso neologismi ricavati dalle loro forme cognominali (ad es. Baldissera, Galliano)15 o attingere al repertorio degli urbonimi eritrei (Asmara/o, Ghinda, Macallè, Mas- saua, ecc.) piuttosto di orientarsi su prenomi evidentemente avvertiti come eccessi- vamente esotici, distanti dalla tradizione italiana o scomodamente connotati. Toponimi eritrei sono d’altra parte stati all’epoca impiegati sul territorio italiano anche come cognomi fittizi da attribuire ai bambini senza famiglia: come evidenziato da Lenci16, l’estro onomaturgico dedicato all’infanzia abbandonata ha infatti dato origine, a partire dal 1885, a una serie di nomi di famiglia di matrice geografica, spesso adattati alla funzione cognominale tramite introduzione di terminazione in -i (ad es. Agordati, Coatiti, Dancali, Massauini, Monculli). L’identificabilità del ruolo del nome di famiglia costituisce d’altra parte un tratto che differenzia significativamente il sistema onomastico italiano da quello eritreo: prenomi e cognomi appartengono a serbatoi onomastici distinti cui corrispondono statuti grammaticali e strutturali differenti, individuale e sensibile all’accordo di genere il primo, invariabile ed ereditario per linea paterna il secondo. Se, quindi, nel sistema italiano bimembre l’inversione dell’ordine dei due antroponimi non compromette la riconoscibilità del loro rispettivo statuto, ben diverso è il sistema appellativo di matrice eritrea, che presenta una sintassi rigida, caratterizzata dalla rigorosa collocazione del personale in prima posizione, segui- to dal personale del padre17, da quello del nonno e da eventuali ulteriori sequenze

ter, “Deux langues, deux cultures. Quels prénoms pour les enfants italo-français ?”, in R. Delamot- te-Legrand, ed., Mixités conjugales aujourd’hui, Presses Université Rouen, CID-FMSH-Diffusion, 2018, 213-236; C. Leguy, “Noms de personne et expression des ambitions matrimoniales chez les Bwa du Mali”, Journal des Africanistes, 75, 2, 2005, 107-128; M.-L. Moreau, “Le marquage des identités ethniques dans le choix des prénoms en Casamance (Sénégal)”, Cahiers d’études africaines, 41, 3-4(163-164), 2001, 541-556; J. Streiff-Fenart, “La nomination de l’enfant dans les fa- milles franco-maghrébines”, Sociétés contemporaines, 4, 1990, 5-18. 15. Per le caratteristiche dell’antroponimia coloniale tra Otto- e Novecento si rimanda a E. De Felice, Nomi e cultura. Riflessi della cultura italiana dell’Ottocento e del Novecento nei nomi personali, Venezia-Roma, Marsilio Editori – Sarin, 1987 e S. Pivato, Il nome e la storia. Onomastica e religioni politiche nell’Italia contemporanea, Bologna, Il Mulino, 1999. Una panoramica dei riflessi coloniali sul repertorio prenominale del XX secolo è stata recentemente riproposta da E. Papa in “Eredità coloniali nell’onomastica italiana del Novecento”, in G. Brincat, ed., Onomastica belli- ca. Da Torino a Malta. Atti delle Giornate di studio del Dottorato di ricerca in Lessico e onomastica, Malta, Malta University Publishing, 2015, 99-129. 16. M. Lenci, “Cognomi italiani di origine coloniale”, RIOn. Rivista italiana di onomastica, 13, 1, 2007, 37-50. 17. Come evidenziano le testimonianze raccolte da Barrera (“Patrilinearità, razza e identità”, 35), l’impossibilità di disporre di un nome maschile da indicare in seconda posizione ha rappresen- Revelli, Onomastica del contatto italo-eritreo 113 patronimiche cronologiche, secondo la tradizione estendibili alla menzione dei bisavoli fino alla settima generazione precedente18. Soltanto il valore posizionale, in assenza di marche pragmatiche o grammaticali che lo rendano riconoscibile in quanto tale, attribuisce quindi al primo nome eritreo il suo ruolo funzionale di identificativo personale. Il contatto storico fra due sistemi onomastici così differenti ha generato e genera equivoci: nella prospettiva occidentale, la citazione di personaggi con ruoli apicali attraverso il primo nome viene letta come inusuale segnale di confiden- za19; il patronimico eritreo in seconda posizione può, d’altra parte, essere errone- amente descritto come gentilizio tribale20 o venir trattato alla stregua di cognome, quindi considerato ereditario anche per la discendenza. I contatti con il contesto occidentale sembrano in effetti oggi orientare gli stessi parlanti eritrei verso una rifunzionalizzazione della stringa per esigenze di semplificazione e adeguamento al canone bimembre: come evidenzia il seguente estratto da un’intervista rivolta a un parlante asmarino di 28 anni21, la denominazione in seconda posizione tende a essere descritta come cognome, e quella in terza posizione a essere considerata accessoria:

I.: come ti chiami? S.: il mio cognome è S. ma tutti mi chiamano B., che vuol dire ‘marinaio’ I.: quindi B. è il tuo nome, quello più usato S.: sì, è il mio nome: qui si usa più il nome del cognome I.: ma che cosa intendi per “cognome”?

tato nel periodo coloniale motivo di forte stigma per gli italo-eritrei non riconosciuti dal padre. 18. Mussie Tesfagiorgis (Eritrea, Santa Barbara, ABC-CLIO, 2010, 236) descrive il sistema come trimembre senza citare la possibilità di riferimenti a personali successivi a quello del nonno. Le testimonianze raccolte sul campo da chi scrive evidenziano effettivamente che il ricorso alla stringa estesa non è più, se non eccezionalmente, praticato; molti giovani – pur essendo a conoscenza della tradizione – dicono d’altra parte di non essere in grado di citare patronimici oltre la terza o quarta generazione. 19. Il disorientamento indotto dalla differente pragmatica degli usi del primo nome è emerso ad esempio in occasione dell’incontro tra il presidente eritreo e il premier italiano avvenuto ad Asmara il 12 ottobre 2018. In quell’occasione, pochi quotidiani italiani hanno ritenuto opportuno citare i due protagonisti come Isaias e Conte, sulla base delle convenzioni comunicative dei rispet- tivi Paesi: nella gran parte dei titoli i due presidenti sono quindi stati denominati per esteso, secondo il principio bimembre, come Isaias Afewerki e Giuseppe Conte. 20. Così veniva presentato il sistema onomastico dell’area del Corno nella Guida dell’Africa Orientale Italiana pubblicata nel 1938 dalla Consociazione Turistica Italiana (CTI, Milano, Tipo- grafia Colombi & C., 1938, 84): “La famiglia è, come ovunque, la base naturale della società; ma la sua individualità è assorbita in una collettività maggiore che comprende più famiglie discendenti da un capostipite comune. Mentre non esiste neanche un vocabolo equivalente a quello di «famiglia», né questa ha un cognome (i figli si designano col nome proprio seguito da quello del padre) […] ognuno di questi gruppi viene designato col nome del capostipite comune che funge perciò da nome gentilizio”. 21. Intervista registrata ad Asmara il 26 marzo 2018. Qui e nelle successive trascrizioni gli antroponimi sono sostituiti dalle iniziali per tutela della privacy.

Africa, N.S., I / 1, 2019 114 a r t i c o l i / a r t i c l e s

S.: è il cognome, è il nome di mio padre. È come per gli italiani: hai un nome e poi un cognome, che è il nome di tuo padre. I.: Però il “nome” di tuo padre è diverso da quello di tuo nonno… S.: Mio padre si chiamava S. e dopo anche M., che era mio nonno. Io in pratica mi chiamo B. S. M. e anche altri nomi dei nonni dei nonni, ma dico solo B. S. perchè tutto è troppo lungo.

Si tratta di un cambiamento strutturale che, se confermato, potrebbe in futuro determinare conseguenze sulla preminenza posizionale della sintassi onomastica eritrea e anche moltiplicare i casi di omonimia, data l’attuale disponibilità di un unico serbatoio per le tipologie appellative in prima e seconda posizione. Poten- ziali effetti di tale cambiamento potrebbero peraltro riguardare anche la rappre- sentatività semantica del personale22: come emerge nella prima battuta dell’estrat- to d’intervista, proprio per la sua funzione simbolico-identificativa questo conser- va ancora attualmente significati pieni e trasparenti che i parlanti condividono volentieri offrendo dispositivi di traduzione e calco. Tale accordo di primato del senso rispetto alla forma onomastica implica anche un’ampia tolleranza nei con- fronti delle varianti appellative, tolleranza e varianti certamente sollecitate dal contesto plurilingue e dalle necessità pragmatiche da questo indotte: la possibili- tà di denominare gli individui attraverso trasposizioni – traduzioni e sinonimie – o per mezzo di appellativi aggiuntivi è in Eritrea una consuetudine molto diffusa anche al di fuori degli ambiti affettivi o confidenziali in cui invece resta di norma circoscritta nel contesto italiano. Ne porta un esempio Aman Abraha23 descrivendo il mondo del ciclismo nella seconda metà del Novecento: piuttosto che attraverso il loro nome locale, i primi campioni eritrei erano infatti noti con pseudonimi italiani (Caramello, Pe- peroncino, Pitone) che venivano adottati dai parlanti locali e dai ciclisti stessi. Operazioni di sinonimia o reduplicazione onomastica sono del resto normal- mente contemplati anche nel contesto cattolico africano, come mette in evidenza il racconto di una donna nata ad Asmara negli anni Cinquanta24:

T.: Io mi chiamo T.Y. però anche in battesimo mi chiamo O. con nome italiano per- ché sono cresciuta con le suore, ho fatto il battesimo. I.: Quindi il nome O. …

22. “African names are typically distinguished from European ones on the basis of name meaningfulness, i.e., African names carry semantic import”, R.K. Herbert, “The Dynamics of Per- sonal Names and Naming Practices in Africa”, in E. Eichler, G. Hilty, H. Löffler, H. Steger, and L. Zgusta, eds., Name Studies. An International Handbook of Onomastics, Berlin – New York, De Gruyter, 1996, II, 1222. A proposito dei significati delle denominazioni tigrine cfr. L. Ricci, “Nomi personali fra genti a lingua tigrina”, Rassegna di Studi Etiopici, 21, 1965, 111-161. 23. Aman Abraha, Oro per l’Eritrea, ‘Diario di un eritreo’, 2013, in: www.eritrealive.com/ oro-per-leritrea (ultimo accesso il 28/3/2019). 24. Intervista registrata ad Asmara il 29 marzo 2018. Revelli, Onomastica del contatto italo-eritreo 115

T.: …è il secondo nome. Mia mamma mi chiamava T. e Y., però le suore mi chiama- vano O. e anche perché mi ha presa in battesimo una signora italiana che mi chiamava così.

Cionondimeno, secondo le consuetudini comunicative eritree il personale in prima posizione costituisce il principale riferimento anagrafico e allocutivo a prescindere dalla formalità delle situazioni e dalla rigidità delle gerarchie sociali, diversamente da quanto accade in italiano – dove il nome di famiglia rappresenta l’ufficialità – e diversamente anche da quanto accadeva in Eritrea quando il titolo appellativo di goitana/guitana (padrone) costituiva il formato asimmetricamente e obbligatoriamente previsto nell’interlocuzione dei locali con gli italiani25. Residui di convenzioni impostate sulla verticalità di provenienza etnica e rango affiorano ancora oggi nei formati interazionali che, quand’anche ammet- tendo – anche per ragioni di semplificazione linguistica – il “tu” reciproco, preve- dono la consueta allocuzione diretta attraverso il prenome, richiedendo però nell’altra direzione l’unilaterale ricorso ad appellativi professionali o al titolo generico ma marcato di signore/signora. Si tratta, anche in questo caso, di un evi- dente esito dei rapporti con la comunità italofona sul territorio africano, non essendo previsto tale formato – invece usuale nella Penisola – nella comunicazione intraetnica. Ulteriori effetti del contatto potrebbero con buona probabilità essere riscon- trati in altre circostanze appellative: in questo senso, anche il modificarsi di convenzioni o di rituali legati a norme documentate per alcune zone dell’Africa sub- sahariana – come il divieto per i figli di chiamare per nome i genitori o per le mogli di pronunciare il nome del marito in luogo pubblico26 – meriterebbero uno specifico studio evolutivo di prospettiva etno-antropologica oltre che onomastica e linguistica.

La denominazione di spazi e luoghi

L’attribuzione ad Asmara dell’appellativo di “piccola Roma”27 e la più recente definizione di “Pompei africana” assegnata ad Adulis ben rappresentano l’inclinazione degli italiani a battezzare gli spazi del territorio eritreo attraverso il ricorso a

25. Riferisce una donna eritrea nata nei primi anni Venti del Novecento: “Gli italiani li chia- mavamo goitana [padrone] ‘buon giorno goitana, buon giorno’: neanche lo conoscevo e dovevo salutare così”, G.Barrera, “Asmara: la città degli italiani e la città degli eritrei”, in A. Triulzi, G. Tzeggai e G. Barrera, eds., Asmara: architettura e pianificazione urbana nei fondi dell’Istituto italiano per l’Africa e l’Oriente, Roma, IsIAO, 2008, 12. 26. Cfr. J. Fédry, “Le nom, c’est l’homme”, L’Homme, 191, 2009, 77-106. 27. Malgrado l’anacronistico stampo che la caratterizza, la denominazione è ancora oggi in voga, come ha dimostrato il suo utilizzo sistematico da parte della stampa italiana quando Asmara, nel luglio 2017, è stata dichiarata dall’Unesco “Patrimonio dell’Umanità” e inserita nella lista World Heritage.

Africa, N.S., I / 1, 2019 116 a r t i c o l i / a r t i c l e s referenti culturali a quel territorio esterni, inclinazione che ha raggiunto il proprio apice autoreferenziale nell’odonomastica del periodo coloniale: allo straordinario impulso urbanistico che rese allora disponibili nuovi centri, quartieri e strade corri- spose infatti una vera e propria proliferazione di intitolazioni italofone e italofile28. I processi di ricorsiva ridenominazione successivamente subiti dagli spazi urbani eritrei riflettono i travagliati avvicendamenti governativi del Paese, evidenziando emblematicamente le politiche di (ri)appropriazione simbolica che, attraverso strut- turati e mirati interventi dall’alto, gli odonimi possono incarnare e al contempo promuovere. A titolo di esempio con riferimento alla città di Asmara, il Viale Benito Mussolini si è tramutato nel periodo post-fascista in Corso Italia per trasformarsi in Hailè Sellassiè I Avenue a conclusione dell’occupazione inglese, quindi in Naziona- le Avenue nel periodo Derg e infine in Harenet Avenue dal 199129. Non soltanto lo stradario, ma anche le intitolazioni delle infrastrutture sono state soggette a processi di ridenominazione (l’Ospedale Regina Elena di Asmara è, ad esempio, stato rinominato Itegue Menen) che evidenziano la transitorietà cui gli odonimi possono essere soggetti e la loro stretta relazione non soltanto con motivazioni linguistiche, ma anche con la temperie politica, ideologica, culturale e sociale delle circostanze storiche in cui si collocano. Guardando specularmente al contesto italiano, d’altra parte, è possibile osservare che il processo di “africa- nizzazione” della toponomastica stradale30, dopo aver portato odonimi encomia- stici e ideologici nella maggior parte delle città italiane31, non soltanto ha subito un arresto dalla metà del XX secolo32, ma è stato successivamente esposto a pro- gressivi processi di erosione: alcune intitolazioni avvertite come politicamente connotate sono state sin dal dopoguerra messe in discussione da parte di vari consigli comunali33 e in alcuni casi anche sostituite attraverso dediche ideologica- mente antitetiche34.

28. La topografia di Massawa si snodava, ad esempio, fra piazza Principe di Piemonte, piazza Garibaldi, via Torino, via Roma, Lungomare Umberto I, Banchina Regina Elena, ecc. 29. Belula Tecle-Misghina, Asmara: An Urban History, Roma, Nuova Cultura, 2015, 71. 30. N. Labanca, “L’Africa italiana”, in M. Isnenghi, ed., I luoghi della memoria. Simboli e miti dell’Italia unita, Roma-Bari, Laterza, 1996, 255-290. 31. M. Lenci, S. Baccelli, “Riflessi coloniali sulla toponomastica urbana italiana. Un primo sondaggio”, I sentieri della ricerca, 7-8, 2008, 161-182. 32. Il censimento effettuato da Laura Ricci per la città di Roma (La lingua dell’impero. Co- municazione, letteratura e propaganda nell’età del colonialismo italiano, Roma, Carocci, 2005, 189 e sgg.) evidenzia come nella capitale le denominazioni celebrative delle imprese coloniali si concentrino nel cosiddetto “quartiere africano” e siano frutto di provvedimenti amministrativi ap- provati nel periodo 1920-1937: in tale intervallo temporale, l’Eritrea è commemorata attraverso una dozzina di denominazioni geografiche (via Asmara, via Assab, via Cheren, via Dancalia, viale Eritrea, via Massaua) in alcuni casi evocative di specifiche imprese belliche via( Adigrat, via Agor- dat, piazza Amba Alagi, via Coatit, via Senafè). 33. Si veda, ad esempio, il caso di Bologna illustrato da V. Perilli, “Da Dogali a Gramsci. Toponomastica e memoria coloniale a Bologna”, Zapruder, 23, 2010, 136-143. 34. Cfr. L. Serianni, “A proposito di odonimia”, RIOn. Rivista Italiana di Onomastica, 1, 1 1995, 41-49. Revelli, Onomastica del contatto italo-eritreo 117

L’immagine dell’Eritrea che emerge oggi dall’odonomastica del territorio italiano risulta di fatto, nella migliore delle ipotesi, staticamente cristallizzata al primo quarantennio del Novecento, e al contempo ancora condizionata dallo sguardo con cui le vicende coloniali hanno, anche linguisticamente, guardato oltremare. L’uni- lateralità della prospettiva trapela infatti anche dalle rese formali con cui i toponimi eritrei sono ancora oggi rappresentati nel contesto italiano, con il ricorso a calchi e varianti che riflettono più o meno significativi e non sempre motivati adattamenti fonetici35 (ad es. Decameré per Dekemhare; Debaroa per Dbarwa36; Godofelassi per Kudo-Felasi) e rimaste fedeli alle prescrizioni del regime fascista in relazione all’evitamento dei grafemi estranei all’alfabeto italiano (Addi Qualà per Addi Kua- la, Cheren per Keren, Massaua per Massawa o Medzawà). Il fatto che tali modelli onomastici non abbiano lasciato tracce al di fuori della Penisola, né nelle scelte della comunità eritrea né negli esotoponimi diffusi- si a livello internazionale, sembra rappresentare – dal punto di vista dell’italofonia – il più sfavorevole dei possibili esiti ottenibili, in quanto implicito segnale di scarso prestigio o deliberata rimozione della lingua coloniale.

La designazione dei prodotti

Gli esiti del contatto italo-eritreo emersi in letteratura attraverso il censimen- to dei prestiti scambiati nelle due direzioni, dalle lingue locali all’italiano e vice- versa, appaiono tutto sommato quantitativamente poco significativi. Se osservato dal punto di vista lessicografico, il primo dei due fenomeni risulta anzi decisa- mente marginale: Nichil37 segnala che il lemmario del GDIU38 comprende 186

35. Lo stratificarsi di interferenze dovute alla distanza fra i diversi codici grafo-ortografici e sistemi linguistici in contatto ha determinato adattamenti anche nelle rese delle forme di matrice italiana da parte degli eritrei: a tale proposito Uoldelul Chelati Dirar (“From Warriors to Urban Dwellers: Ascari and the Military Factor in the Urban Development of Colonial Eritrea”, Cahiers d’Études Africaines, 44, 3/175, 2004, 533-574) cita ad esempio il caso della denominazione del quartiere asmarino di Kommishtato, “tigrinya corruption for the Italian Campo cintato” (554) e quello del toponimo Addi Telleria, “which is the result of the combination of the Tigrinya word Addi, which means village, and Telleria, the Tigrinya corruption for the Italian artiglieria” (ibid.). 36. Corruzioni del toponimo dovevano essere ben radicate e diffuse in Europa se già M. Bru- zen de la Martinière, geografo al servizio del re di Spagna Filippo V, nel suo Grand dictionnaire géographique et critique, II (B. et C.), A la Haye, Chez P. Gosse, & P. de Hondt, 1730, sente la necessità di glossare la voce Ba r o a precisando: “Ba r o a ou De b a r o a , ou Ba r u a , ou De b a r u a , ou plus mal encore Ba r n a ; manières vicieuses d’écrire le nom de Do b a r w a , lieu où reside le Bahrnagash, en Abissinie au Royaume de Tigré” (123). 37. R.L. Nichil, “Voci dall’Africa. Un contributo per la ridefinizione dell’elemento coloniale nel lessico italiano”, in R. Coluccia, J. Brincat, and F. Möhren, eds., Actes du XXVIIe Congrès international de linguistique et de philologie romanes. Nancy, 15-20 juillet 2013; Section 5: Lexico- logie, phraséologie, lexicographie, Nancy-Strasbourg, Société de Linguistique Romane, 2016, 415. 38. T. De Mauro, Grande dizionario italiano dell’uso, Torino, UTET, 2007, I-VI e due supple- menti.

Africa, N.S., I / 1, 2019 118 a r t i c o l i / a r t i c l e s africanismi, che “al netto di lievi oscillazioni grafiche e fonetiche, si riducono ad appena 142 unità, corrispondenti a circa lo 0,043% del lessico presente nel repertorio”. In specifico riferimento all’area del Corno d’Africa, che pure – escludendo gli apporti di area araba – risulta quella dotata del ruolo maggiormente attivo, l’apporto è limitato a 34 prestiti dall’amarico, 13 dal somalo, 9 dal tigrino, 4 dal ge’ez e 1 dal caffino39. Un po’ più ampio ma probabilmente comunque sottostimato nelle indagini lessicografiche appare il serbatoio degli italianismi penetrati nelle parlate del territorio eritreo. Ai nostri fini, va osservato che secondo la consuetudine lessicogra- fica i nomi propri sono esclusi dai lemmari dei dizionari, e quindi la componente onomastica può essere osservata esclusivamente in relazione ad alcune particola- ri categorie di lessemi caratterizzati dal contemporaneo statuto di nomi propri e nomi comuni o interessati dal passaggio dall’una all’altra categoria. È il caso dei singenionimi disponibili per l’allocuzione diretta: in quest’ottica, possiamo osservare che Yaqob Beyene40 registra nel tigrino la disponibilità degli esiti bābā per “papà” e emāmā/māmā per “mamma” e la penetrazione dell’adatta- mento baadari/baaderi impiegato per “padre” come appellativo reverenziale d’ambito religioso. Ulteriori prestiti interpretabili in chiave onomastica sono rappresentati da voci interessate da processi transonimici, come beretta per “pistola automa- tica”, dal cognome di Pietro Beretta, fondatore dell’omonima fabbrica d’armi, e chianti per “vino rosso”, dal nome della regione toscana di produzione. In relazione al saho, dei 307 lemmi censiti da Banti e Vergari41 – alcuni dei quali veicolati da arabo e tigrino – due corrispondono a deonomastici: il primo te- stimonia il passaggio del marchionimo B/balilla, impiegato per la celebre utilitaria FIAT, al significato comune di “auto di piccole dimensioni”; il secondo, consiste nell’impiego del tipo S/sambarsaano “sammarzano” – dal nome del comune di San Marzano sul Sarno – per indicare genericamente i pomodori di forma allungata. È però probabile che questi così come altri deonimici documentati da Hoff- mann42 – ad es. acquaragia, biro, diesel, eternit, galvanizzare – siano penetrati in diversi idiomi del Corno d’Africa già nella loro veste di nomi comuni, in effetti acquisita precocemente nell’italiano peninsulare. Più distintivi del contatto italo-eritreo risultano esiti che, pur assenti nei repertori lessicografici, risultano ben vivi nelle testimonianze dei parlanti: è il caso del marchionimo N/necchi, passato a rappresentare il significato di “macchina da cucire” in relazione all’attività dell’italo-eritreo Giovanni Mazzola, sarto attivo e

39. Poche altre voci, 11 in totale, sono etichettate con generica indicazione di provenienza geografica come etiopiche, abissine ed eritree (Nichil, “Voci dall’Africa”, 423). 40. Yaqob Beyene, “I prestiti italiani in amarico e tigrino”, Rassegna di Studi Etiopici, N.S., 3, 2011, 97-140. 41. G. Banti, M. Vergari, “Italianismi lessicali in saho”, Ethnorêma, 4, 2008, 67-93. 42. S. Hoffmann, Il lascito linguistico italiano in Dodecaneso, Libia e Corno d’Africa: L2, pidgin e prestiti, Tesi di laurea in Lingue e Culture Moderne, Università degli Studi di Pavia, a.a. 2012/2013, www.academia.edu/9926333 (ultimo accesso il 28/3/2019). Revelli, Onomastica del contatto italo-eritreo 119 celebre in entrambe le sue patrie, o dell’uso generalizzato di M/melotti come an- tonomastico di “birra” grazie al successo della produzione avviata nel 1939 ad Asmara dal migrante italiano Luigi Melotti43. Mancano d’altra parte nelle raccolte dei prestiti, pur facendo invece parte del patrimonio della comunità locale, denominazioni italofone – usate a prescindere dalla lingua materna dei parlanti – per punti di riferimento che hanno conservato una centralità, reale o simbolica, soprattutto nello spazio urbano di Asmara: è il caso dell’edificio modernista denominato Fiat Tagliero, inaugurato come stazione di servizio nel 1938 e dedicato a Giovanni Tagliero, direttore della sede FIAT locale; del Caravanserraglio, trasformato nell’officina-mercato di Medeber; del Cimitero degli italiani, collocato all’interno del cimitero cristiano, e anche dei molti teatronimi di stampo fascista (Cinema Impero, Cinema Roma, ecc.). Per le strade della capitale – così come in quelle degli altri principali centri urbani eritrei – le scritture esposte, private e commerciali, si collocano general- mente in un paesaggio linguistico d’intensa stratificazione plurilingue (Bar Vitto- ria pastry ice cream; Gianni e Gina Beauty Parlour) in cui la componente italofona è ben testimoniata. È tuttavia indispensabile notare che gli apoteconimi (far- macia, ottica, pizzeria, ristorante, ferramenta) e le denominazioni di attività commerciali (Albergo Italia, Bar Capri, Caffè Rosina, Pasticceria Giardino, Pen- sione Pisa) espressi in italiano compaiono soltanto nelle insegne datate di diversi decenni, mentre la cartellonistica più recente adotta sistematicamente come lingua internazionale l’inglese (ad es. Multi Sports Club & Fast Food), in alternativa a o a fianco del tigrino. Mero residuo novecentesco, il paesaggio linguistico di matrice italiana risale insomma a un passato che ha veicolato prestiti spesso temporanei, non necessariamente penetrati negli usi effettivi dei parlanti o per lo meno non in modo generalizzato: già in origine, in effetti, il pubblico cui le scritture italofone esposte si rivolgevano afferiva ai circuiti della comunità dei migranti giunti dalla Penisola o dei loro discendenti. Si tratta, a ben vedere, di un fenomeno inverso e speculare a quello oggi emergente nelle città italiane, dove nuovi flussi migratori favoriscono la compar- sa di inediti paesaggi urbani, con denominazioni e insegne commerciali espresse attraverso alfabeti e idiomi tarati su pubblici etnicamente selezionati. Anche in questo caso, la valenza simbolica attribuita ai naturali dispositivi pragmatici del contatto mostra una portata che travalica di molto la sola dimensione onomastico- linguistica, come evidenzia per l’Italia la recente presentazione di una proposta di legge che vorrebbe la messa al bando delle insegne multietniche che contengono

43. Con il cambio di proprietà del birrificio, oggi condiviso dallo Stato eritreo e alcuni inve- stitori privati, la birra è stata ribattezzata Asmara, ma la denominazione di Melotti continua a essere comunemente impiegata come sinonimo di “birra eritrea”.

Africa, N.S., I / 1, 2019 120 a r t i c o l i / a r t i c l e s denominazioni espresse in una lingua diversa dall’italiano, a meno che si tratti di un dialetto della Penisola o di lingua europea44.

Le etichettature

Al di là delle macrocategorizzazioni stereotipizzanti che hanno suddiviso la collettività in negri e ferengi, gli effetti del contatto tra abitudini, stili di vita, universi culturali e lingue differenti si sono manifestati nella storia italo-eritrea anche attraverso la coniazione di appellativi creati ad hoc per denominare fenomeni o referenti venutisi a costituire come risultati del contatto stesso. In questa direzione l’esperienza coloniale ha dato luogo alla diffusione di denominazioni etnico-professionali impiegate per l’identificazione di categorie esclusive di alcuni gruppi o sottogruppi delle etnie locali: a partire da basi arabe, sono stati adottati sia dall’italiano coloniale che dalle lingue endogene appellativi come ascari (soldato), per identificare la specifica figura dei combattenti indigeni arruolati nelle truppe coloniali; dubat (turbante bianco), per indicare i componenti di truppe irregolari di arditi neri; meschin (mendicante), per alludere ai que- stuanti e da loro stessi adottato per chiedere aiuto e carità. Un’ipertrofica necessità di differenziare, categorizzare, marcare l’appartenenza o prendere le distanze, sempre comunque implicitamente esprimendo un giudizio a proposito degli effetti della diversità emersa attraverso il contatto, ha condotto ad attribuire alle donne eritree a servizio delle famiglie italiane lo specifico titolo etnico-professionale di letté, mentre il marchio eufemistico-gergale di madame45 è stato riservato a quelle, fra loro, che vivevano in contesti di concubi- nato e l’epiteto grevemente spregiativo di sciarmutte a quelle dedite alla prostitu- zione. La vena onomaturgica, aperta agli apporti di matrici linguistiche diverse – ora palesemente ingiuriosa, ora paternalisticamente mitigata – non risparmiava il mondo dell’infanzia: il titolo di “diavoletti” – per reinterpretazione paretimolo- gica dell’arabo ja ulèd (ei, ragazzo)46 – identificava i giovani locali al servizio, anche sessuale, degli italiani; quello di “cioccolatini” era invece attribuito ai nati

44. Proposta di legge presentata il 24 marzo 2018 dalla deputata leghista Silvana Comaroli e altri: “Art. 9-bis. 1. Le regioni, nell’esercizio della potestà normativa in materia di disciplina delle attività economiche, possono stabilire che la posa delle insegne esterne a un esercizio di sommini- strazione al pubblico di alimenti e di bevande sia condizionata all’uso di una delle lingue ufficiali dei Paesi appartenenti all’Unione europea ovvero del dialetto locale”. 45. Come osserva C. Volpato (“La violenza contro le donne nelle colonie italiane. Prospettive psicosociali di analisi”, DEP. Rivista di studi sulla memoria femminile, 10, 2009, 122), la creazione del termine madamato risponde in effetti a un duplice obiettivo: “definire le relazioni tra uomini italiani e donne africane con un termine diverso da quelli impiegati per le relazioni tra donne e uomini italiani e usare un eufemismo per indicare lo sfruttamento sessuale e domestico delle africane”. 46. Cfr. Ricci, La lingua dell’impero, 221. Revelli, Onomastica del contatto italo-eritreo 121 da unioni miste, altrimenti etichettati dagli italofoni come mulatti o meticci47 e apostrofati dai parlanti locali come dqala o wedi sebeyti48. Nella stessa cerchia dei coloni italiani era d’altra parte possibile riscontrare forme soprannominali che evidenziavano malcelate ostilità di categoria, attribuibili a differenze di provenienza geografica, rango socio-economico o scelte negli stili di vita. Così, i migranti italiani di estrazione proletaria o provenienza meridionale e quelli che si integravano nel contesto d’arrivo con modalità diverse da quelle attese o convenzionali venivano etichettati, sulla base delle diversità loro imputate, come indigenati49, insabbiati50, incatramati51 o anche, attraverso sprezzanti espressioni mutuate da blasoni popolari della cultura locale, gurage e ‘agāmè52. Sebbene certamente incompleta, questa rapida rassegna sembra sufficiente a evidenziare esemplarmente che le relazioni scaturite dai contatti tra comunità migranti possono dar forma a un universo onomastico che riflette una percezione esasperata dell’alterità inter- e intra-etnica, alterità che le soluzioni onomastiche stesse codificando convalidano e enfatizzano: come osserva Comberiati, in effetti, “il nome storpiato, l’epiteto affettuoso o offensivo, il tono neutro o razzista con cui si viene chiamati contribuiscono a formare […] l’identità dello straniero”53.

Conclusioni

Se le occasioni di contatto vissute dalle realtà eritrea ed italiana nell’ultimo secolo e mezzo hanno innescato dispositivi di accorciamento della lontananza –

47. Cfr. F. Faloppa, Parole contro. La rappresentazione del ‘diverso’ nella lingua italiana e nei dialetti, Milano, Garzanti, 2002. 48. Si tratta di epiteti grosso modo equivalenti a “bastardo” (Barrera, “Patrilinearità, razza e identità”, 34). 49. “l’indigenizzazione dell’italiano venne rappresentata come un lasciarsi andare, un abban- dono che sanciva la vittoria della ‘terra vergine’ sulla volontà di chi si era lasciato sedurre senza opporre resistenza. Una minaccia, dunque”, N. Poidimani, Difendere la razza. Identità razziale e politiche sessuali nel progetto imperiale di Mussolini, Roma, Sensibili alle Foglie, 2009, 121. 50. Allo studio delle caratteristiche dei cosiddetti “insabbiati” è dedicata un’approfondita ricerca condotta sul campo negli anni Ottanta dall’antropologa F. Le Houerou, L’épopée des soldats de Mussolini en Abyssinie, 1936-1938: Les Ensablés, Paris, L’Harmattan, 1994. 51. Cfr. G. Dore, “La vita nelle Colonie 1923-1941”, in M. Isnenghi, G. Albanese, eds., Gli italiani in guerra, III, Torino, Utet, 2008, 651-658. Agli incatramati ha dedicato diverse pagine il giornalista Tommaso Besozzi, autore di reportage dall’Africa raccolti nel volume Il sogno del set- timo viaggio, Roma, Fazi, 1999. 52. Come segnalato da G. Dore (Scritture di colonia. Lettere di Pia Maria Pezzoli a Bologna dall’Africa Orientale (1936-43), Bologna, Patron, 2004, 27) ‘agāmè era l’epiteto spregiativo “con cui nelle rappresentazioni inter-regionali eritree si indicavano gli abitanti della povera regione etio- pica confinante con l’Akkälä Guzāy” mentre gurage era l’appellativo usato per indicare sprezzan- temente gli abitanti dall’omonima regione, considerati “rozzi, campagnoli e disposti ai mestieri più umili” (124). 53. D. Comberiati, Scrivere nella lingua dell’altro. La letteratura degli immigrati in Italia (1989-2007), Bruxelles, Peter Lang, 2010, 40.

Africa, N.S., I / 1, 2019 122 a r t i c o l i / a r t i c l e s geografica, linguistica, culturale – creando saldi legami tra i due Paesi, la distanza rimane e appare consistente anche da un punto di vista onomastico. Si tratta di una distanza in alcuni ambiti – come quello dei prenomi – sorvegliatamente pre- servata; in altri – come quello relativo alla sintassi antroponimica – in potenziale evoluzione; in altri ancora – come quello odonimico – studiatamente ripristinata; di una distanza, in ogni caso, che una sterile onomaturgia tassonomica ha nel passato alimentato e che oggi ancora rischia di sostenere attraverso la classifica- zione di persone, comportamenti, ruoli sociali con etichette onomastiche descrit- tive di una diversità codificata in termini di alterità, differenza, non appartenenza. E se, come qui ci si è concisamente proposti di documentare, gli usi onomastici possono validare – specialmente in un quadro di iniziale marcata distanza e nelle situazioni di incontro conseguenti a migrazioni – implicite rappresentazioni del contatto, appare chiaro che le scelte onomastiche individualmente o collettiva- mente seppur non sempre consapevolmente attivate dai parlanti incidono in modo determinante sulla natura e la direzionalità dei contatti, e non soltanto di quelli onomastici. John Burton Kegel

Po s t -Ge n o c i d e Rw a n d a

Timothy Longman, Memory and Justice in Post-Genocide Rwanda. Cambridge University Press: Cambridge, 2017 Susan Thomson, Rwanda: From Genocide to Precarious Peace. Yale University Press: New Haven and London, 2018

Since the 1994 Genocide, the Rwandan Patriotic Front (RPF) has been the dominant political force in Rwanda. One of its most significant challenges has been the social, economic and political reconstruction of the country under less than ideal circumstances. Despite leaps forward, Rwanda remains a poor country in a deeply unstable part of the world. Both books under review here analyse the nature and outcomes of these strategies of reconstruction. As such, they are more than purely historical works: they actively engage with current politics in Rwan- da. In a broader sense, both Timothy Longman’s Memory and Justice in Post- Genocide Rwanda and Susan Thomson’s Rwanda: From Genocide to Precarious Peace fit into a well-established narrative, one which is highly critical of RPF’s governance.1 Thomson has written on a similar subject in her previous book, Whispering Truth to Power, and her various altercations with the RPF are well known.2 Longman’s relations with the RPF have been more nuanced, and he can boast of an exceptional history of engagement with, and study of, Rwanda. First, this review will outline the contents of both works; second, it will raise a number of specific criticisms about some of their claims and findings.

I .The central thesis of Longman’s book is that the reconciliation process initiated by the Rwandan government has been a failure, because two key ingredients have not been properly handled by the RPF. The first obstacle on the route to reconciliation, so argues Longman in the opening part of his book, is that the RPF has been too heavy-handed in imposing its view of history on the country. One of the key objectives of the RPF after taking power in 1994 was to re-educate the

1. For instance, F. Reyntjens, Political Governance in Post-Genocide Rwanda. Cambridge, Cambridge University Press, 2013. S. Thomson, Whispering Truth to Power: Everyday Resistance to Reconciliation in Postgenocide Rwanda. Madison, University of Wisconsin Press, 2013. 2. S. Thomson, “Re-education for Reconciliation: Participant Observations on Ingando”, in S. Straus, L. Waldorf, eds., Remaking Rwanda: State Building and Human Rights after Mass Violence, Madison, University of Wisconsin Press, 2011, 311-339.

Africa, N.S., I / 1, 2019 124 r a s s e g n a / r e v i e w a r t i c l e

Rwandan population, away from the divisive ethnic teachings of the Habyarimana years and towards a narrative “in which Tutsi were not foreign invaders but sons and daughters of the Rwandan soil.”3 In Longman’s view, the new official history is built around a set of unquestionable tenets. First, the essential unity of the Rwandan people during the pre-colonial period, when conflicts along ethnic lines were unknown. Second, the divisive role played by colonialism: it was colonialism which laid the foundation for the ethnic hatred which would prove so deadly in post-colonial Rwanda. Third, the period of rule by Kayibanda and Habyarimana exacerbated the ethnic tensions introduced by the colonizers, “inevitably building toward the 1994 genocide.”4 Fourth, the centrality of the 1994 Genocide to Rwanda’s history. And finally, the role played by the RPF in stopping the Genocide in 1994. In Longman’s view, the emphasis on this last assumption seeks to obliterate the public memory of RPF abuses during and after the 1990-1994 war. As heroic saviours of the country, the RPF cannot also be villains. The idea that the RPF bears any responsibility for the genocide itself, for having attacked the country without regard for the consequences for Tutsi still in Rwanda, is categorically rejected.5 Longman concludes that, in imposing its vision of history on the country, the RPF has been guilty of ignoring the lived experiences of many Rwandans. The fact that neither Hutu history nor Hutu sufferings are acknowledged has left many with a bitter distaste of the current government and acts as a barrier to reconciliation. This leads to Longman’s second explanatory factor in accounting for the failure of the reconciliation process, namely, that justice has not been done in Rwanda. Justice, Longman argues, has not been used to help find the truth or punish the people who committed crimes between 1990 and 1994; rather, its primary objective has been the subjugation of the Hutu population of Rwanda. In this reading, mass arrests following the Genocide and, again, during the years when the Gacaca courts were active were mainly geared to ensure “RPF political, social and economic domination.”6 The proof of this is that RPF members accused of war, or other, crimes committed after the Genocide have not been tried. Had the RPF been truly committed to justice and reconciliation, they would have also prosecuted their own. To substantiate his arguments Longman draws on three case studies: Buyoga, Mabanza and Ngoma. This is by far the most interesting and compelling part of the

3. T. Longman, Memory and Justice in Post-Genocide Rwanda. Cambridge, Cambridge Uni- versity Press, 2017 p.46. 4. Longman, Memory and Justice, 51. 5. Longman, Memory and Justice, 57. Let it be clear that the reviewer also categorically re- jects this idea. The very fact that a regime is prepared to use its own citizens as hostages is a sign that the regime must be overthrown. 6. Longman, Memory and Justice, 23. Kegel, Post-Genocide Rwanda 125 book, where the depth and breadth of the research conducted by Longman and his team really shines through. There is no doubt that the material presented in this section will prove useful to historians for many years to come. Using interviews with individuals, focus groups and his own substantial experience of the country, Long- man addresses the multi-faceted ways in which the Rwandan population has responded to attempts at transitional justice. Not the least commendable trait of this section of the book is that, here, Longman allows Rwandans to speak for themselves. This openness to local voices contributes to nuance his overarching arguments and shows the multitude of opinions held throughout the country. Thomson’s book is more strident than Longman’s in presenting Rwanda as a full blown dictatorship – one in which the RPF controls every facet of daily life – and in arguing that the apparent progress in many socio-economic domains amounts to a deception that serves the interests of the RPF, enabling it to curry favour from the international community and to continue accessing international aid and funds. The book opens with a derivative account of the 1994 Genocide and its causes. The politics surrounding independence and the Social Revolution (1959-1964) are briefly discussed, as are the presidencies of Kayibanda and Hab- yarimana. Next, the reader is introduced to the background of the RPF and how it consisted mostly of Rwandan refugees who grew up abroad in refugee camps. Thomson does concede that it was the military campaign of the RPF which ended the Genocide. In the following chapters, however, she argues that, following its victory, the RPF assumed total power in Rwanda by manipulating the “genocide guilt” of the international community and by ruthlessly using force to repress any dissidence within the country. Eventually, she also contends, the RPF moved away from the use of blatant force to impose its will, in favour of less open, but just as coercive, elements of state power – especially rules and laws – to control the social realm. In Thomson’s Manichean view, this total control is necessary for two reasons: first, to develop the Rwandan economy as quickly as possible, and secondly, to line the pockets, and secure the power position, of the RPF refugee elite from Uganda, the Democratic Republic of Congo and Burundi. Thomson’s polemic also posits a continuum between the current dispensation and the Hab- yarimana and Kayibanda periods. In her view, “Rwanda’s postcolonial presidents have long used the language of democracy and inclusivity to mask their dictatorial tendencies.”7

II. Despite their differences in tone and empirical strength, Longman’s and Thomson’s works share one fallacy, namely, the failure to place their respective narratives in the context of Rwandan history. This is not entirely surprising, considering that several key aspects of the country’s recent past have yet to be properly studied. Most obvious is the failure of both authors to adequately delve

7. Thomson, Rwanda, 200.

Africa, N.S., I / 1, 2019 126 r a s s e g n a / r e v i e w a r t i c l e into the history of the RPF. The books discuss the post-Genocide rule of the RPF, but do not examine in sufficient depth the history of the movement. This is clearly problematic, since every organisation draws on lessons from the past to inform its present and future actions. In the case of the RPF, the main formative experiences of its most important members were their years in exile, the Struggle for Liberation – as the RPF calls it – and the Campaign against Genocide. In the aftermath of the Genocide, these were followed by the wars in Zaire/DRC and the insurgency which raged in North-West Rwanda until 1998. Out of these critical conjunctures, only the Genocide period is addressed in detail by the two authors. Conversely, the experience of growing up in Obote II’s Uganda remains by and large unexplored, as do the transformations that the RPF underwent during the 1990-1994 Struggle for Liberation (or Civil War, as many in the West, including Longman and Thomson, call it).8 Much the same can be said of the insurgency in North-West Rwanda, which the RPF regarded as posing a fundamental threat to their precarious post-Genocide stabilisation efforts.9 Failing as they do to place the RPF in its rightful historical context, both Thomson and Longman advance a number of dubious claims. They, for instance, assume that decision making within the RPF is monolithic, with the President’s will being invariably the way, with little room for internal dissent or moderation. In all probability, this contention needs to be problematized. The old guard of the organisation, after all, is made up of Rwandans who grew up all over the Great Lakes Region. There are those who grew up in Uganda, and are mostly English speaking, and those who grew up in French-speaking Burundi. Other constituen- cies include the members of the Rwandan military who were formerly part of the Forces armées rwandaises (FAR), and those who fought against the RPF during the Congo Wars and the Insurgency in the North West, like Paul Rwarakabije.10

8. More satisfying in this respect are the relevant chapters in C. Clapham eds., African Guer- rillas, Oxford, James Curry, 1998 and G. Prunier, The Rwanda Crisis, London, Hurst & co, 1997. Also, W.C. Reed, “Exile, Reform and the Rise of the Rwandan Patriotic Front”, Journal of Modern African Studies, 34, 3, 1996, 479-501. See also: E.D. Mushemeza, “Politics and the Refugee Experience: The Case of the Banyar- wanda Refugees in Uganda (1959-1994)” PhD Thesis, Makarere University, 2003. This was also published as a book. E.D. Mushemeza, The Politics and Empowerment of Banyarwanda Refugees in Uganda 1959–2001, Kampala, Fountain Publishers, 2007. 9. On it, see P. Roessler and H. Verhoeven, When Comrades Go to War, London, Hurst & co, 2016. and R. Orth, “Rwanda’s Hutu Extremist Insurgency: An eyewitness perspective”, Working Paper, MacMillan Centre for Genocide Studies, 2001. There is also a detailed report by African Rights “Rwanda: The Insurgency in the Northwest”, 1998. but the impartiality of the organisation has come under recent scrutiny, see: R. Debelle et al., ‘Rebuttal to: “NGO Justice: African Rights as Pseudo-Prosecutor of the Rwandan Genocide,” by Luc Reydams’, Human Rights Quarterly, 40, 2, 2018, 447-465. and L. Reydams, “Protesting Too Much: A Response to Linda Melvern et al.”, Human Rights Quarterly, 40, 2, 2018, 466-473. 10. The best informed expose on the composition of the current RDF is: M. Jowell, “Cohesion through socialization: liberation, tradition and modernity in the forging of the Rwanda Defence Force (RDF)”, Journal of Eastern African Studies, 8, 2, 2014, 278-293. Kegel, Post-Genocide Rwanda 127

Last but not least are the young cadres of the party who were born and grew up in Rwanda. It would be odd if the need to take into account the interests of all these groups did not affect RPF governance. The truth is that very little is known about the decision making process, and thus the motives, of the RPF.11 This is why the view should not be taken at face value that the RPF’s policies are solely, or primarily, motivated by the desire for self-enrichment and material gains. This leads to another criticism, one which applies with special force to Thom- son’s work. The book’s originality and even its ultimate raison d’être remains open to doubt. As the anti-RPF genre has been well and truly alive for several years, Rwanda – its polemical ambitions notwithstanding – fails to open up any genuinely new avenues for debate. Nor, as already intimated above, does it present its reader with any new primary or even secondary material. Thomson herself writes that “assessing whether RPF rule could result in future episodes of mass political violence is the subject of this book.”12 Her unhelpful conclusion is that: Rwanda’s past points to waves of mass violence, occurring every forty years or so, when the ruling class fractures and ordinary people become targets of physical, eth- nically motivated violence. The ambitious, talented and heavy-handed RPF show few signs of bucking this trend. As the Yale University historian Dan Magaziner poignantly notes, ‘bodies and human suffering are the cursed currency of history, as Juvenal Habyarimana and Gregoire Kayibanda have taught and Paul Kagame regret- tably continues to teach.’ Rwandans, regardless of political affinity, socio-economic class or ethnic identity, know this all too well.”13 What Thomson does not mention is that violence in Rwanda, though pro- voked by elites, has historically been largely carried out by “ordinary” Rwandans. The suggestion that there will be violence in Rwanda’s future robs these Rwan- dans of their agency and capacity to shape their future for the better, apparently damning them to a perpetual cycle of violence and penury. This deeply cynical approach also fails to suggest any real alternatives to the path chosen by the RPF. To date there have been few, if any, examples of truly successful reconstruction in very poor, post-genocide/conflict states in fragile geo-political areas. Has Truth and Reconciliation succeeded in South Africa? Has the Balkan model worked better, with international peacekeepers still on the ground under UNMIK? And how about the Democratic Republic of Congo or Burundi? Afghanistan or Iraq? Where is the shining example the RPF is supposed to follow to a stable, prosperous, democratic state – all in less than 25 years? Criticism is easy, but finding solutions is infinitely harder, and both Longman and Thomson fail to present any.

11. What we do know often comes from sources who have permanently broken with the RPF, and whose testimony is by definition somewhat biased. 12. Thomson, Rwanda, 4. 13. Thomson, Rwanda, 254-255, Magaziner quote “#MindYourOwnBussinesss.” September 21, 2016, Africa is a country, http://africasacountry.com/2016/09/mindyourownbusiness/ (last accessed October 9, 2016)

Africa, N.S., I / 1, 2019

recensioni / reviews

Yves Person, Historien de l’Afrique, explorateur de l’oralité, Edition mise au point, présentée et enrichie de notes et de cartes par François-Xavier Fauvelle et Claude-Hélène Perrot, Paris, Publications de la Sorbonne, 2018, 220 p.

Le nom de Yves Person (1925-1982) est étroitement lié à son œuvre majeur, la somme en trois volumes qu’il consacra à Samori Touré (Samori. Une révolution dyula, Dakar, Ifan, 1965-1967-1970). Et pourtant, Person est l’auteur également d’un nombre remarquable d’articles, d’études et d’essais dont une moindre partie est reproposée dans ce recueil, consacré principalement aux sources orales de l’histoire africaine, ce qui renvoie nécessairement au rapport avec l’anthropologie. C’est sur ce sujet que s’ouvre ce recueil, avec un texte de 1971, qui ne rend que partiellement compte de l’état des rapports entre les deux disciplines phares de l’africanisme de cette époque. L’article est daté, certes, et il ne peut nullement être considéré aujourd’hui comme une contribution utile au débat. Il contient notamment quelques jugements à l’emporte-pièce et un certain flottement qui mène l’auteur, après des critiques sévères à l’encontre de l’anthropologie, à une conclusion « pacifiée », évoquant un rapprochement tout à fait convenu entre les deux disciplines. Mais avant d’en arriver là, Person maltraite les anthropologues, sans toujours les identifier précisément, jusqu’à écrire des phrases du moins surprenantes par leur généralisation hâtive, telles que celle-ci : « C’est précisément pour fuir l’histoire que les anthropologues se sont consacrés aux sociétés sans État, dont les structures donnaient une fausse impression de simplicité. Un examen attentif leur apprit que c’était là une illusion, ce qui fut important pour ramener les anthropologues à l’histoire » (p. 51). Ce propos n’a pas de sens ; c’est comme si on disait qu’inversement les historiens ont évité les sociétés non étatiques pour se cantonner à l’étude de structures sociales « faciles », car familières, comme l’État. De plus, le propos de l’historien est amplifié dans l’Introduction au recueil (non signée) en ces termes : « L’anthropologie est une discipline qui n’a pas trouvé son objet » (p. 16), ce qui radicalise jusqu’à la caricature l’affirmation déjà injustifiée de Person. Ce n’est pas fini. Person écrit aussi que « sous prétexte d’aider l’Afrique, l’anthropologue prétend lui dicter ses propres priorités. L’ère coloniale ayant boule- versé les anciennes sociétés, il pense que, pour faire oublier son éventuelle compli- cité, il doit empêcher leur étude » (p. 55). On croirait rêver : un ancien administrateur colonial qui parle de l’« éventuelle complicité » de l’anthropologue [au singulier] avec la colonisation ! Un exercice de réflexivité de la part de l’historien

Africa, N.S., I / 1, 2019 130 r e c e n s i o n i / r e v i e w s aurait été d’un tout autre intérêt, mais Person, dans sa carrière scientifique et acadé- mique n’a jamais souhaité revenir sur son parcours précédent d’administrateur des colonies, casquette sous laquelle il a tout de même conduit la plupart de ses recherches sur Samori, notamment en Guinée (1954-1958) et en Côte d’Ivoire (1958- 1960). La plupart des autres articles publiés dans ce recueil sont – heureusement – consacrés aux sources orales, à la chronologie et à la généalogie, sujets sur lesquels la contribution de Person revêt un intérêt certain. Toutefois, le choix de ces sujets, s’il respecte d’un côté l’engagement de Person et constitue un dossier ho- mogène, risque de l’autre côté de donner une image quelque peu biaisée du travail de l’auteur. Yves Person a été, en effet, un historien qui a su, avant d’autres et mieux que d’autres, conjuguer toutes sortes de sources : orales, écrites, maté- rielles, cartographiques. Or, le choix, par ailleurs pleinement justifié, de reprendre les textes de l’historien sur les sources orales a tout de même une conséquence fâcheuse, celle de réintroduire un « grand partage » entre oralité et écriture que Person avait largement dépassé, pragmatiquement sinon théoriquement. Il suffit de rappeler que dans son ouvrage sur Samori les traditions orales collectées auprès de 861 informateurs (précisément listés) côtoient un nombre tout aussi impres- sionnant de sources écrites coloniales. En d’autres termes, l’apport pionnier et la leçon durable de Person concernent surtout l’utilisation conjointe de toute sorte de source et par là l’harmonisation, au-delà de quelques rudesses polémiques, de l’histoire et de l’anthropologie des sociétés africaines. Non, Person n’a pas été qu’un historien de l’oralité ce qui serait réducteur ; de plus, il a aussi intégré à son travail des sources écrites africaines, contribution vraiment originale que l’historiographie plus récente a cherché à amplifier et qui n’est évoquée par lui-même que trop rapidement (p. 56). C’est finalement la contribution de Yves Person en faveur de l’élargissement du champ des sources pour servir à l’histoire africaine qui reste peu visible dans ce recueil. Les écrits de Person – qui concernent aussi le monde manding, l’empire du Mali et l’histoire des religions africaines – sont enrichis de nombreuses et ponctuel- les notes des éditeurs scientifiques, qui accompagnent et actualisent très opportuné- ment les textes. Dommage seulement qu’ils n’aient pas pris connaissance d’un article récent d’Elara Bertho (« Filmer la résistance à la colonisation. Stratégies postcoloniales de mémoire et d’oubli. À propos du scénario Samori de Sembène Ousmane », Cahiers d’Etudes africaines, 66, 4, 224, 2016 : 875-890) qui a retrouvé et présenté le scénario inédit que le réalisateur sénégalais avait préparé pour un film, qui n’a jamais vu le jour. Ce scénario était directement inspiré par l’œuvre de Per- son, preuve certaine de l’influence de cet historien trop vite disparu et dont le travail aurait pu, fait rare, devenir à son tour la source d’une fiction cinématographique centrée sur le « personnage » (le terme est de Person) de Samori.

Fabio Viti r e c e n s i o n i / r e v i e w s 131

Stuart Doran, Kingdom, Power and Glory: Mugabe, ZANU, and the Quest for Supremacy, 1960 – 1987. Midrand: Sithatha Media (pb R489 - 9780620 752930). 2017, 842pp.

The political and economic crises that Zimbabwe experienced since the turn of the 21st Century have generated a considerable literature by scholars keen to understand the formative years of African nationalism and the post-colonial state. Stuart Doran`s Kingdom, Power, Glory joins this growing list. Most studies whether in biographies, memoirs, or academic historical accounts are divided between the nationalist/patriotic history of the 1990s and more recent works critical of the post-colonial state. Doran`s work falls within the latter group and offers a detailed analysis of this period from the vantage point of Pretoria, London, Syd- ney, Ottawa, and the Front-Line States. Critical of African nationalist motives during the liberation war and the early years of independence, Doran examines the rise of Robert Mugabe and the pervasiveness of violence in Zimbabwe`s political landscape, not least the Gukurahundi massacres. For Doran, African politicians like Mugabe, the central figure in the book, were driven by personal ambition and the pursuit of power at all costs. Kingdom, Power and Glory describes in detail the continuous scheming that pitted rival African nationalists, political parties, and their military wings against one another. Cunning and ruthless, Mugabe emerges in the end, as the victor in this tussle for power within the nationalist movement and the post-colonial state. Doran relies on diplomatic cable and espionage reports shared between and among embassies, governments, and their respective intelligence agencies. This material is gleaned from Australian declassified documents, diplomatic telexes between Ottawa and Harare and the National Archives in both Britain and South Africa. What emerges is a fascinating outside and top down look into the nationalist movement and the war, the politics of the Lancaster House negotiations, the February 1980 elections, the dissident question, Gukurahundi and Zimbabwe African National Union’s (ZANU) efforts to consolidate power by establishing a one-party state. The book underscores the fierce antagonism that existed between and among individual politicians, ZANU, Zimbabwe African People`s Union (ZAPU), United Alliance National Council (UANC), and their military wings. These political disputes were resolved through violence and by courting the support of the Frontline States and by taking advantage of the Cold War animosi- ty between the West and the East. This was a game whose art Mugabe had perfec- ted. By the end of the 1980s, Mugabe had managed to absorb ZAPU into ZANU Patriotic Front (PF) and had eliminated the political rivalry of the UANC, ZANU (Sithole) and the remnants of the Rhodesian Front. In addition, Mugabe ensured the dominance of the Zimbabwe National Army by Zimbabwe African National Liberation Army forces at the expense of Zimbabwe People`s Revolutionary Ar- my or the erstwhile Rhodesian army. Doran argues that ZANU and Mugabe`s dominance of the nationalist and postcolonial political space came at the expense

Africa, N.S., I / 1, 2019 132 r e c e n s i o n i / r e v i e w s of the general populace who were subjected to systematic violence and intimida- tion during the war, at the time of the 1980 general election and in the years following the attainment of independence. In that regard, his study corroborates the conclusions in the memoirs by David Coltart, Masipula Sithole, Joshua Nkomo and Edgar Tekere as well as the work of Catholic Commission for Peace and Ju- stice on the Gukurahundi massacres. More than that, Doran`s work falls within recent literature that questions the role of violence in Zimbabwe`s political past and challenges the patriotic history of the 1980s that glossed over the extent and implications of the Gukurahundi carnage. Other aspects of Kingdom, Power and Glory are problematic, however. For one, the argument that violence was emblematic to ZANU PF during its formative years has been thoroughly explored by many studies. Indeed, Doran paints an undifferentiated picture of a dystopian post-colonial state plagued by violence, personal ambition, and regional ethnic hostilities. In his attempt to portray Muga- be and ZANU PF as the ultimate villains, he denies agency to important players during this period. Even more, he downplays apartheid South Africa`s role in destabilisation after 1980, reimagining the dissident question as a crisis blown out of proportion to serve Mugabe`s quest for complete political control. In addition, Mugabe`s political nemeses such as Joshua Nkomo and Abel Muzorewa or those within his ranks like Josiah Tongogara are depicted as passive victims of Mugabe`s unbridled ambition. Part of the problem has to do with the nature of the material that Doran relies on. The literature on the early years of 1980 has shown how the white settler populace and the wider international community was apprehensive of a ZANU PF led government. Not surprisingly, diplomatic cable and intelligence from London, Pretoria, Ottawa, Sydney, and former Rhodesian intelligence`s reportage of the 1980s decade reflects these concerns. This is a difficult period to write on, not only because of the sensitivity of the subject matter but because Zimbab- we government documents from the 1980s are not yet available to researchers. Even so, external sources could have been enhanced through oral testimony collected from key contemporary players. That said, Kingdom, Power, Glory is a welcome contribution to the understanding of the contentious late colonial and early post-colonial period in Zimbabwe.

Tawanda Chambwe Au t o r i / Co n t r i b u t o r s

To m McCas k i e was successively Professor of Asante History, Birmingham University, and Professor of African History, SOAS, London University, until his retirement in 2011. He is author of State and Society in Pre-Colonial Asante (Cambridge, Cambridge University Press, 1995) and Asante Identities (Edinburgh, Edinburgh University Press, 2000). In 2015, fifty of his selected papers were reprinted asAsante, Kingdom of Gold by Carolina Academic Press. This paper is the tenth he has published on Asante since 2017. He lives in France.

El a r a Be r t h o is researcher at the French National Centre for Scientific Research, in Bordeaux, Les Afriques dans le Monde (LAM UMR 5115). Sorcières, tyrans, héros. Mémoires postcoloniales de résistants africains à la colonisation (Paris, Honoré Champion, 2019).

Al e ssan d r a Br i v i o teaches Cultural Anthropology and Anthropology of Religions at the University of Milano-Bicocca. She has a great deal of fieldwork experience in Benin, Togo and Ghana. Her research interests are Africa religions and memory of slavery in Africa. She authored articles and books including Il vodu in Africa. Metamorfosi di un culto (Roma, Viella, 2012) and Donne, emancipazione e marginalità. Antropologia della schiavitù e della dipendenza (Sesto San Giovanni, Meltemi, 2019).

Ma r ian o Pa v an e l l o , retired professor of Anthropology and former Head of the Dept. of History, Cultures, Religions (University of Rome “La Sapienza”), did his main fieldwork among the Nzema of Ghana (1989-2013). His more recent works include several papers published in international journals and the books: Perspectives on African Witchcraft (London, Routledge, 2017); La papaye empoisonnée. Essais sur la société Akan des Nzema (Riga, Éditions Universitaires Européennes, 2017); Research Materials on Traditional Medicine in the Nzema Area (Ghana) (Accra, University of Ghana Press, 2011).

Sa m u e l An d r e as Ad m asi e is regional representative of the International Institute of Social History in Africa and a senior lecturer at the Institute of Peace and Conflict Studies of the University of Hargeisa. He received his doctorate from the University of Pavia and the University of Basel, researching the history of the Ethiopian labour movement.

Lu isa Re v e l l i is researcher in Italian Linguistics at the Department of Human and Social Sciences of the University of Valle d’Aosta. Her recent volume Langues faibles (Paris, L’Harmattan, 2017, with A. Tabouret-Keller and G. Varro) and the research project “Multilingualisms and migratory linguistics: diachronic and varietal profiles of the Italian of Eritrea” (in progress) focus on study of the relationships between change of language, behaviors and speakers’ representations, of the relations between alleged linguistic identities perceived and practiced in plurilingual contexts and of the secondary repertorial dynamics to migratory phenomena.

AFRICA. Rivista semestrale di studi e ricerche pubblicata dal CSPE - Centro Studi per i Popoli Extraeuropei “Cesare Bonacossa” con Viella editrice, e sostenuta dall’ISMEO-Associazione Internazionale di Studi sul Mediterraneo e l’Oriente/Il Novissimo Ramusio Copyright © 2019 - CSPE / Viella

Africa è una rivista semestrale di studi africani sottoposta a una procedura di doppia revisione anonima, che accoglie articoli in italiano, inglese e francese. Si incoraggia la presentazione di saggi afferenti a tutte le aree delle scienze umane e sociali, purché siano risultato di analisi rigorosa e ricerca empirica saldamente fondata. Africa is a double-blind peer-reviewed half-yearly journal of African Studies, hosting contributions in Italian, English and French. Submissions in all areas of the humanities and social sciences are encouraged. Published articles will invariably be the result of rigorous analysis and empirically grounded research.

Linee guida per gli autori https://www.viella.it/riviste/testata/15/autori Instructions for authors

Questo fascicolo è stato pubblicato anche con il contributo del Progetto MIUR “Studi e ricerche sulle culture dell’Asia e dell’Africa: tradizione e continuità, rivitalizzazione e divulgazione”. Finito di stampare nel mese di aprile 2019 da The Factory srl Roma