THE VOCABULARY OF DEATH IN CHADIC AND HAMITO- Daniel BARRETEAU itr IBRISZIMOW Herrmann JtT" NGRAITHMAYR ABSTRACT In this paperare discussed the reconstructionsof the roots for the items "to die" and "corpse". In al1 the 150 , which are spoken in ,Cmeroon and , as well as in the whole Hamito-Semitic family, "to die"is reconstructed by one single root: *mw . On the opposite, there are at least four roots in Chadic for the gloss "capse": A *mwt, bg-, *bn and D kum, with some borrowings. The comparison and the analyses of thesetwo items give a sample of the complexity of the Chadic language history:on the one hand, a fundamental notion, "to die", which carries forms coming directly from the original Hamito-Semitic language craddle, on the other hand, a more "cultural" notion, "corpse", which has been considerably affected through the contactsin the new African home. This may indicate that the funeral rites have undergone more recent changes in the environmentof Basin. Keywords : Hamito-Semitic, Chadic, Nigeria, , Chad, comparison, reconstruction, die, corpse RÉSUMÉ LE VOCABULAIRE DE LA "MORT" DANSLES LANGUES TCHADIQUES ET CHAMITO-SÉMITIQUES. Dans cet article sont presentees et commendes les reconstructions pourles termes "mourir" et "cadavre". Dansles 150 langues tchadiques, quisont parlees au Nigeria, au Cameroun et au Tchad, ainsi que dans 'ensemble du phylum chmito-semitique, "mourir" est reconstruit avec une seule racine : *mwt. A l'oppos6, dans les langues tchadiques, on relBve au moins quatre racines pour le terme "cadavre": *mw t , bg-, * bn et kum, dont certains emprunts. La comparaison et l'analyse de ces deux items donne une idCesur la cornplexit6 de l'histoire des langues tchadiques: d'un côte, une notion fondamentale, "mourir", quise traduit par des formes remontant directementB l'origine des langues chamito-semitiques; de l'autre côte, une notion plus "culturelle", le "cadavre", qui a et6 considerablement affectee par des contacts dansle nouvel environnement africain. Cela pourrait indiquer que les rites funeraires ont subides changements rkcentsdans l'environnement du bassin du lac Tchad. Mots-clés : chamito-semitique, tchadique, Nigeria, Cameroun, Tchad, comparaison, reconstruction, mourir, cadavre * **

This paper is an enlarged version of a poster which was presented at the exhibitionduring the conference. A firststudy was conducted by Jungraithmayr (1990) onthe problem of reconstructing the roots for "death" and "to die" in the Chadic languages. In the present article, the data come mdnly from thecomparative work on Chadic languages by Jungraithmayr and Pbrisaimow (in press). Some unpublished data come as well Erom persona1 materials on Cameroonian Chadic languages collected by Bareteau. Here, only the roots for the items "to die" and "coqse" are discussed, both having been reeonstmcted by Jungraithmayr and Ibrisdmow.

P. lYira die" Tnere are only few glosses which are reconstructed with 8 single rsot for the whole Chadic family, such as: "to die", "to dream", "to drink", "eye", "finger","fly", "four", "head", "neck", "nose", "to ripen", "sheep", l'ta swallow". In al1 the 156 Chadic languages, which are spoken in Nigeria, Cmeroon and Chad, "to die" is a representative of this group of glosses to be recons(ructed by one single root. The proposals, up to now, were: *me t by Newman (Y977), *mwt by Jungraithmayr & Shimizu (198%)and 2~mv~t by Jungrdhnayr brisdmow (in press). At the same time, this is the most fundamental Harnito-Semitic keyword in Chadicsince it occurs in al1 families with the possibleexception of Cushitie [which may be assumed as a secondary absence; CJ Murtonen 19&9:256].So, this root, %mvt "to die", would be attested wiek a history of approximately ten milleria, in Asia minor and in the whole of North Africa dswn to Lake Chad Basin. As fa as theHamito-Semitic languages in general are concemed, one can eonsult also the study of Vycichl (1983). The importance of this word must be seen in the aspect of the genetic heritage which ha been considerably shnk due to the following facts: - the Chadic languages were one of the first to separate from the eommon Hmito-Semitic ancestor; - in their historieal development, the Chadic speaking peoples had long and nmow contacts with the peoples 0% the Sudan uea who speak "Nigritic" (Le. non-Hmito-Semitic) languages; - the 150 Chadic languages represent the most numerous and diversided linguistic unit within Harnito-Semitic as well as the widest geographical distribution if one does not csunt the more recent expansion of habic, after the VIIth centanry A.D. EGYrn Egyptian Coptic smmc maa t Ugaritic rnt Phoenician me Hebrew mlw t Arameic mwt, rnyt Arabie mw t 23 1

Mehri mwt Soqotri mii(t) Gelez mwt Tigre mwt BERBER Chleuh mta Kabyle amma t Tuareg amma t sus mma t Zenaga ëmmwi, ëmi cusmnc Rendille a-mu t! Boni -wud- Somali m5d Due to the above mentioned facts, the present-day Chadic reflexes display sometimes an extreme "deformation" as a result of linguistic erosion, e.. *mwt > *mutu> *ntu>*tu > ru. The first radical (Rl) is for the most part preserved in its original quality: * > m in al1 branches, Western, Central and Eastern. It has changed to - in some Central Chadic languages and in Mokilko (Mokilko group) due to the loss of an interna1 vowel and therefore directly preceding a dental (partial regressive assimilation): *mwt > *m t > nt . See for example: Ngwaxi nti Cibak nti Gude ht6 Mokilko 21nd- The nasal was weakened to in some languages of the Central Branch: Higi-Baza wta or even disappeared totally: Kilba dtii Hizli t àw Wandiu t a-w Tsuvan zà Gudu 9u 232

8 to the triradicality of the reeonstmction for the Chadic lmguages - in spite of the sbvious biradicality of most of the reflexes -, the argument derivea from the fact that 8 number of the langages display reflexes with a long voevel, a possible signn of a lost (we&) comonant, e.g.:

Gerka Siri Kabalay or in some spwiific verbal forms, e.g. Mubi

%nsome laquages of the Central Branch the first and the second radicals *mw have fallen together, being hardened to mb, wxeh is a curent phoneme in these groups: 233

Mwulyen -mb6 t ii Njanyi mbfi t à Gbwata mb% t ii Bacama ùibfir6 The Ulird radical (R3) has undergone regular changes either to c, t , s, , 1, r, d and even tozero. Generally, a trend towards palatalization - * t > t y > c - can be observed (mainly) in Central Chadic, e.g. Ziziliva kan màcd Ga voko mdcà(w6) Southern-Mofu 'màc' Besleri m%c

This development led to such forms like Fali-Mucella m I kyf and even Kiriawqk'i. A more fronted development, i.e. to an alveolar (with further palatalization) or to an affricate, is found in languages of the Northern and Southern Bauchi group, as well as in Central Chadic, e.g. Tsagu miiG s èn Gurd'ung masse / mIs . Geji mI sh Buli masshi / m%shü / mIsh Tule màshl Zaar mîshî / mIs Wandala mts-a / 3mts5mts8 Palasla mats5 Buwal màts Examples of * t > y : Warji miy- Kariya mkyà Miya mi y- Pa'a mlyà / miya / miyà In the following examples it can not be definitely established whether y is a reflex of the second (*w) or the third (* t ) radical, or even represents a fusion of both: Siri miwa / miya / mîyü Mburku miy- Kwang miiyé Ndam mkyâ 234

Examples of * t > 1 : Bole mBB 6 Jixnbi Illu%- Examples Qf :kt > P : Baeama Dera HWOna Ga'anda Gabin Boka GUdU Musgu Munjuk Maj ra Sina Sibine Tobanga Examples of * t > d Tangak KWami Gera e Mmlam Malgbe .Mpado MOkilkO Examples of * t > d : THZl Mbxa A total loss of the third radical (R3) is attested in some eastern languages: Mera m6 1 m5a- Kabalay mUW5 Ede mz Tumak mi5 Cases in whieh only 83 haslbeen preserved exist but are rare (in the Central Braneh):

Gudu HU Tsuvan zii 'Ihe form wqki in Kiria is an interesting exarnple of how assimilation and metathesis have been at work as it often occurs in Bura-Margi and in Kgi groups: *mw t > *mwk i > *mk i > wr~ki . 235

CHADIC Western Branch Hausa group Hausa mEitau Gwandara mEiti3 Sura-Gerka group Sura mùu t,pl. mur ap Angas mù t Cip mut Ankwe mût Ta1 mik t Gerka mwath Ron group Fyer mot Daffo mot Bokkos mot Kulere mot / mw5th Sha mô t Bole-Tangale group Karekare meet- /metaw Bole mà16 / mot- Ngamo mat0 Maha muto Tangale mydq Kwami mùd- Kupto miit- Pero miitù Dera mur- Kirifi mutto/ mbk-ki3,VNmÙtB Gera mudu- Gerumawa mutalla Northern Bauchi group Warji ml- / mkyZug / miy- Tsagu miiïjsèn Kariya mkyà - Miya miy- Pa’a m’iyà / miya / miyà Siri miwa / mkyil / miyü Mburku miy- Ji mbi mul- Diri m5t6 / matu Dira mùs i Burma mIsè Southern Bauchi group Boghom mas Kir muse, mas! Tala mas-

237

Gudu 13 Fali-Mucella mIkyl Fali-Bwagira mt i Gava fit sigànà Nakatsa ms Laamang group Laamang mt- Wandala group Wandala mts-a / 3mtsiimtsii Ga lvaxdaxa mt s- Guduf mt s’( a)gànà Dghwede ficà / mtssyà Ga voko mdcà(wé) Xa di màt‘ Mabas mà t (ùkir) Sakun group Sakun -sggwus Mafa group Mafa màtsa Mefele mztsH(y) Cuvok mdtsa Southern Mofu ‘mà c . Northern Mofu mà c Southern Giziga m3c Northern Giziga m6c Mbazla muc Merey mii t Dugwor m6 ta Zalgwa mü t’ Ma lokwo mG t Mada mH t Mu yang m6 t Wuzlam ‘mii t . Mata1 ‘m6ts‘ Mbuko màts Palasla mets5 Daba group Daba mà c Besleri (= Hina) m6 c Mba dam mats Gava mfci Buwal màts Kacfa-Munjuk group Kacfa (= Gidar) mt’i Musgu mgr a Mbara micfPq Munj uk mlrf 238

.'qchrpseq' There are at least four roots in Chadlie for the gloss "corpse": A *mwt,B bg-, C *bm and D knm. They are reconstructed roots, with an asterix, or possibly reconstructed roots, without any asterix. There are also some forms which still expect explanation: these are marked in the list below with a dash. 239

"Corpse" can be circumscribed and rendered by "dead (person)". Thus the variousreflexes which are put togetherunder root A, *mwt, are undoubtedly based on the root for "to die", *mwt;cf. e.g. "to die" "corpse" Boghom mas mas dw Masa mf t mi tnà Lele mZ tümidü Al1 the reflexes for which root B, bg- , has been reconstructed are only tentatively assumed to go back to one single root, therefore it is presented without an asterisk. We are possibly dealing here with forms related to that of the root *bg reconstructed for "body". Still unexplained remains the co- existence in Southern Molù (Mafa group) of vàw "body'' and vàgày "corpse". As to Hausa (Hausa group) gd6wdd and Ngizim (Bade-Ngizim group) gùvil, both tentatively considered for this root, a metathesis is assumed. The fact that neighbouring Benue-Congo languages have -v5gb (Menemo) and u-:kWê (Kambari) for "corpse" makes the Chadic reflexes of the B root appear to be -Congo loans. On theother hand, the borrowing may have, in individual cases, also gonelater in the reverse direction, cf. i-gàawà (Ura) which is probably of Hausa origin. Roots C, *bn, and D, kum (including the similar forms maked'with D in the list below), also have obvious cognates in Benue-Congo; e=$ Kwanka pun , Birom vino, Legbo *~-bono,etc. and Chori, Kenyi kom,Yashi kum, etc. Still the reflexes of root C * bn show regular Sound correspondences. The ungraded Yedina (Kotoko group) reflex b in seems not to be related with the C root, but would come from Kanuri kawîn . The unmarkedreflex in Diri (Northern Bauchi group) aru may be related to the root * wr for "fetish". Western Braneh Roots Hausa group Hausa B? g6dwdB Gwandara B 9% Sura-Gerka group Sura D k6m Kofyar D kum Goemai D 'um Ron group Daffo C fun Bokkos C fun Bole-Tangale group Karekare B? gùbù Tangale - dQP?k Kupto - tii689 m., tilt t6sh6 . Dera A miiI pl. mil j6q Galembi C bànà 246

B c B B B B B

B

A madi W6 e

a A a

A Kotoko -

Kwang Kera %degroup Mabalay Lele Sibine group Sibine A mhdë Ndam A mHl t' Sokorogroup Sokoro B goi Dangla-Migama group Dangla mây t ?ni? (Dajo-) Migama -/A gIm6 kà mlàtêntà Mokilkogroup Mokilko - pii6 t è Mubi-Toram group Jegu A moo t Birgit A gL mààti Mubi r immè (Arabic)

CONCLUSION The comparison and the analyses of the two items "to die" and "corpse" are a sample of the Chadic language history: on the one hand, a notion which carriesforms coming directlyfrom the original Hamito-Semitic language craddle, on the other hand, a notion which has been considerably affected through the contacts in the new African home. In addition there is a great diversity of forms for other notions connected with "death", such as "grave" or "funerals". They arerendered through various innovations like semantic analogy ("hole" for "grave"), derivation, compound words and loan words. This may indicate that the funeral rites have undergone more recent changes in the environment of Lake Chad B asin. ORSTOM, UTAH, Bondy, France University of Frankfkrt-Am-Main, Allemagne

BIBLIOGRAPHY An extendedbibliography is not given in this paper: for additional informationone may consultthe sources mentioned below. Maps of Hamito-Semitic and Chadiclanguages may be found in Jungraithmayr (1981), Dieu & Renaud (1983). BARRETEAU D. & D. JUNGRAITHMAYR, 1989. "Chadic lexical roots: Nombre de racines reconstruites pour chaque item", Langues et cultures dans le bassin du lac Tchad (Barreteau Cd.), P