Monographie Journals of the Near Fast General Editor: Giorgio Buccellati Afroastatic Gnguistics Editor: Robert Hetzron, Santa Barbara Associate Editor: Russell G. Schuh, Los Angeles Advisory Boa Ariel Bloch, Berkeley John B. Callender, Los Angeles Talmy Givén, Los Angeles ‘Thomas G. Penchoen, Los Angeles Stanislay Segert, Los Angeles Volume 5 Issue 1 December 197 Chadie Classification and Reconstructions by Paul Newman @ Undena Publications Malibu 1977AFROASIATIC LINGUISTICS AAL includes contributions in linguistics within the vast domain of Afroasiatic (Hamito-Semitic) languages. Articles of general, theoretical interest using Afroasiatic material, descriptive, historical and comparative studies are included. Editor: Robert Hetzron (1346 San Rafuel, Santa Barbara, Ca. 93109, US.A) Associate Editor: Russell G. Schuh (15337 Hart St., Van Nuys, Ca. 91406, U.S.A.) Advisory Board: A. Bloch, J. B. Callender, T. Givén, T. G. Penchoen, S. Segert. MONOGRAPHIC JOURNALS OF THE NEAR EAST AMINE is a system of joumals on the Near East, with each journal devoted to a specialized study area, and each issue consisting normally of a single article. Current journals in the system are Afroasiatic Linguistics, Assur, Computer Aided Research in Ancient Near Eastern Studies and Syro-Mesopotamian Studies. General Subscription. — For a prepayment of $15.00 the subscriber selects random issues from within the entire system 2s desired, up to a total of 200 pages. The subscriber is also entitled to (1) periodical lists of abstracts from all journals in the system, and (2) reservation to any journal within the system, whereby issues of a given Joumal are sent on approval immediately upon publication (and may be retumed within two weeks). Library Subscription. — A prepayment of $15.00 for each journal in the system secures all issues of a single volume as soon as they are published. This subscription schedule does not allow the selection of random issues. Library subscriptions are available to both institutions and individual scholars. Individual issues are numbered sequentially within each volume. Each issue has its own pagination. A volume is closed when a total of about 200 pages is reached, AA title page and a table of contents listing all issues within each volume are sent to all subscribers at the close of a volume. Periodicity in the order of appearance of issues is not predetermined. A volume, however, is generally completed ‘within one year. Institutional and Professional discount of 20% on single subscriptions entered within six months of the beginning of any given volume (higher on larger orders). Payment must accompany orders from individuals, A handling fee of 804 will be charged to Libraries if order is not prepaid. Order from: UNDENA PUBLICATIONS, P.O. Box 97, Malibu, California 90265, U.S.A, ©1977 by Undena Publications. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photo-copy, recording, or any information storage and retrieval system, without permission in writing from the publisher.Monognaphic Journals of the Near East Afnoasiatic Linguistics 5/1 ‘ecenber 1977) 0. L. CHADIC CLASSIFICATION AND RECONSTRUCTIONS by Paul Newman Rijksuniversiteit te Leiden This paper contains a comprehensive classification of the Chadic family and new reconstructions of Proto-Chadic phonology and lexicon. The classification shows the position of all known Chadic languages from the level of sub-group to major branch, It is supplemented by an index in which all distinct languages as well as dialect variants and alter- native nanes are identified. ‘The phonemic system of PC is reconstructed with a rich inventory of con- sonants and a sparse inventory of vowels. For the consonants, a large number of sound laws are described leading from the PC inventory to the modern reflexes in individual languages and language groups. ‘The PC word List includes 150 itens attributable to the proto-language with a high degree of certainty. Each item includes a reconstructed for in- dicating vowels as well as consonants and supporting citations from at Teast two distinct major branches of the family. CONTENTS page IMIROOUTION 25 ese ee eee ase te 2 GHADIC CLASSIFICATION 2. eee 4 1.1. The three major branches tiie 8 112! The Masa isolated branch |) 21 > 1 Bees Geeice ease a Subclassification within the West) 22222222 L PDE iit 8 Subclassification within Biu-Mandara sf ee 9 Subelassification within the East. +). sess eee eon 9 PROTO-CHADIC PHONEMIC INVENTORY 2... ee ee eee ee gree 2.1. Consonants eee ee ee aoe 9 PAM Coo oendoc s0udddu4oGudcG00 9 2182 Pla em temerone meee ee sacubodGHUdDG 10 21S SPC ices teu jeune cc ewe teeg cece e ee 10) eae ECA ie er Ct ee ee eee 10 2.1.5. The absence of glottal stop 2.21. Be ee ee 10 ouihGs PChegeaeetts gee ere ees bdbcosd ty Delt PC Sie eee eee re soobod Mm 211.8. ‘The problem of prenasals see cece eee 2.1.9. Palatal(ized) and labialized consonants 2 22. ey 2.2, VOWS oe ee ee et ee ees ya 2:21: Nunber| of] vowolicontrasts| ) <4; )2/f ale 2: cle <6: rez AALS, 12 P. Newnan [AL 2.2.2. Vowel length oe oe cece eee eens 21213. Tone/Accent 122222 I It Se bs cul pepo pongo Geb eGanpd0G9bes04ccu0G5006 ‘Two changes affecting subbranch WST-A TTDI ttt Seven changes affecting Hausa...) 12D lt Loss of distinctive voicing in Tangale and Pero 2222 lt Intervocalic weakening in Kanakuru. 2... so ee ee Alteration of sibilants within the Bole group’ <1 2. Boacade Phonation changes in the Angas group... +... ssc s eee Booddd 4g > y/w in the Bade groups. ves ttt Development of hlaterals within’ the Zaar group 2220102 ee 3.9. Two changes affecting the M branch»... 000s secre a 310. Weakening of velar stops in the Tera groups.) ss sss t ste : 3.11. Phonation changes in the Tera group... sss sss tee cnet Gt Sy i ree pee ee 3:13. *n > £ in the Bura/iligi major group 2. Go 3.14. “n> x in the Bura group... 2. ss Bee ee eee 3.15. Devoicing in the Bura group 22 itil 3.16, Labio-alveolars and labio-palatals in'the Bura group 2222007 3:17. Loss of final consonants in Higis ss. eee ee eee 3:18. *t > £ in the Nandara group. 2 Lt ee 3.19. "> w in the Mendara group 2 birt 3.20. Loss of hiaterals within the Bata grow 2.222222 te Bla. #t >A in Bahama... ee ee etter eee 3.22, * > a and other dialect-specific changes in Kotoko Ss.) 2S. t. 3.23. Change of sibilants toh in Bulut... ss... sss s esses 5.20. Loss of hiaterals in the ESt branch 22 222222222222 0002 3125. Three changes affecting Tak... 1. +s ssc r lessee ae 3.26. Some Masa correspondences .. 0s. vst sve sss ssse neces 3:27. Common Chadic neutralization of phonation contrasts) F.2D tt 4. PROTO-CHADIC WORD LIST... es Be 4.1. ‘he reconstructions 221.2 ge ee 4.2. The citations... 21. Bens eee er 4.2.1. ‘The transcription system 2222200 t eee seen (tse (couirces sce rears eye teense emer ea ert ene a ist4if Mord}list) 6 $e) 2 a 8 eee ee Bae S. CHADIC LANGUAGE INDEX ee ee eee eee eee ee eee REFERENTS) 67 ce Me ee Nees ©. INTRODUCTION 5/L 2 13 B 13 5 15 15 15 15 16 16 16 16 16 16 7 V7 v7 V7 vv V7 18 18 18 18 18 18 19 19 19 20 20 20 a 22 35 8 ‘The first comparative Chadic word list with Proto-Chadic reconstructions was published a ‘*Thus study was partially supported by a National Science Foundation Grant SOC-7205563. Field work on Study of Nigerian Languages, Bayero University College, Kano. Kotoko and Warji was carried out under the auspices of the Centre for the T an grateful to Thilo Schadeberg for conments on an earlier draft of this paper and to Russell Schuh for in- numerable ideas on questions of Chadic language history. AL 5, 21977) Chadée Classigication and Reconstructions 3 mere ten years ago (Newman and Ma 1966 {henceforth N/Ml]).! The List included 145 etymolo- gies of which 25 were labelled as second-level confidence, The reconstructions indicated Consonants only, hyphens being used to mark the position of vowels, It was understood at the time, hovever, that the use of hyphens for vowels was simply a'notational device (in view of our then inability to reconstruct the vowels) and not an indication of their lex- ical significance or insignificance in Chadic. As the first such list of its type, the N/M set of reconstructions becane a point of refer- ence for comparative work within Chadic and between Chadic and other Afroasiatic languages. However, being a first attempt, and depending as it did on primary sources of unequal re- Liability, the N/M list inevitably contained a number of false (or questionable) etymologies, erroneous language citations, and inexact reconstructions. Neanvhile, the energence over the past decade of good quality descriptive materials on nunerous Chadic languages has nade possible the discovery of additional Proto-Chadic items— including some important, basic words—as well as the development of better ideas about the original nature and subsequent history of Chadic phonology. A work that would update the N/M study thus seened called for. One of the major purposes of this paper is to present a new Proto-Chadic {henceforth PC] word list, building on the earlier N/M list, but taking advantage of insights and knowledge acquired Over the past ten years. As in thé earlier work, all of the comparisons are intra-Chadic and no attempt has been made to relate the PC reconstructions to lexical items elsewhere in Afroasiatic (see Cohen 1947, Gouffé 1969/70, Hodge 1968, Pilszczikowa 1960). ALL of the items in this new list can be ascribed to PC with a high degree of confidence, being well attested throughout the fanily and in conformity with sound laws identified a3 having taken place. To emphasize the fact that this list is intended as a qualitative in- provenent on N/M and not a multiplication of speculative etymologies of questionable valid- ity, the number of reconstructions has purposely been limited to approximately the number in the earlier work, The addition offered in this work is not to the gross number of proto- forms reconstructed, but to the reliability of the etymologies and to the fullness of the reconstructions per’se. In this work, for example, the consonants have been specified nore explicitly and a first’ attempt has been made to incorporate vovels in the reconstructions. ‘The other major aim of this paper is to present a new classification of languages in the Chadic fanily, with greater and nore precise internal structure than provided in any pre- vious classification. This classification is accompanied by a comprehensive index that permits immediate identification of all Chadic languages that we now know of, with cross- referencing provided for altemative nanes and dialect variants. ‘The organization of the paper is as follows: (1) the classification; (2) the phonological inventory postulated for PC; (3) selected sound laws affecting various languages and lan- guage groups within the family; (4) the PC word list; (5) the language index; and (6) ref- erences. ‘an earlier paper by Pilszczikowa (1958) consisted mainly of comparisons between Hausa and other West Chadic languages. It did not contain reconstructions. The comparative study by 111i¢-Svity€ (1966), published in the sane year as N/M, was limited to etymolo- gies containing a bilabial consonant, but it did contain sone valuable reconstructions. However, having been published in Russian, it did not at the time receive the attention among Chadicists that it deserved. MAL 5, 34 P. Newnan [AL S/L 1, CHADIC CLASSIFICATION ‘The Chadic family of the Afroasiatic phylum includes over 100 languages spoken in northern Nigeria, northern Cameroon, western and central Chad, and, in the case of Hausa, Niger. Sone twenty-five years after Greenberg first advocated the unity of the Chadic family as we now know it, it is now possible to provide a fairly accurate picture of the interrela- tionship of the many languages that make up the family. The composition and internal class~ ification of the fanily that I have adopted is given in the following outline. For the Listing of language nates and for the menbership of the lower level groups and clusters, 1 have depended on recent works of Caprile and Jungraithnayr (1973), E. Wolff (1971), Schth (n.d. (a)), and, nost importantly, Hoffmann (19714). ‘The higher groupings showing the gen- eral structure’of subclassification within Chadic are my own. ‘The units of classification employed are "branch", "subbranch", "group" (which may be conbined with another group into a "najor group"), and""Ianguage".*’ Within a group, more closely related languages are set apart from less closely related languages by the use of letters and/or by the use of semi- colons as opposed to connas. The tree diagram which accompanies the classification does not show subdivisions below the level of the "group." My aim has been to provide a comprehensive inventory and classification of languages in the Chadic family. It has not always been easy, however, to decide exactly what to include and ‘what not to include since the dividing line'between Separate languages and dialects of the sane language is not discrete and the very sam language may be known by a variety of nanes depending on cultural, historical, or geographical factors. In order to make the classifi- cation easier to use for referential purposes, I have tried, wherever possible, to extract altemative spellings, dialect variants, village names, etc. fron the major listing and to relegate such infornation to the index.” In the case of well-known altemative names, how ever, both are shown in the classification, one being treated as the primary name and the other being enclosed in parentheses. Chadic Language Fanily 1, West Branch (ST) A. Subbranch WST-A 1, Hausa group Hausa, Gvandara 2/3. Bole/Angas najor group 2. Dole group a, Bole (olanci), Bele, Ngano, Maha, Kirfi, Deno, Kubi, Galanbu, Gera; Karekare b. Tangale, Pero, Kupto; Kanakuru (Dera) 3. Angas group a. Angas, Sura, Kofyar, “hip; Goemai (Ankwe), Montol, Tal, Pyapun, Koenoem b. Gerka "the groups are all referred to by the nane of a well-known language in the group. This is not an ideal system, but it seens preferable to multiplying new and wnfaniliar desig1977) Chadic Classification and Reconstructions 5 4. Ron group? a. Ron, Sha, Kulere, Karfa, Shagawu b. Fyer B, Subbranch WST-B 1/2. Bade/Warjit major group 1, Bade group Bade, Ngizim; Dwai* Warji group® Warji, Pa'a, Siri, Diri, Jimbin, Miya, Mburku, Kariya, Tsagu Zaar group* fa. Zaar (Sayanci), Barawa, Zeem, Polchi, Geji ... b. Gunntun, Ju .t. ¢. Boghom, Nanas’... IL, Biu-Mandara Branch (BM) A. Subbranch BA 1, Tera group a. Tora, Jara D. Ga'anda, Hona 2/3. Bura/tligi major group ‘Jungraithnayr (1970) uses the term "Ron" for, the group as such and refers to the indi- vidual languages as "Kon-Bokkos", "Ron-Sha", "Ron-Kulere", etc. I have departed from this practice by adopting "Ron" as the nane’for the language spoken in Bokkos and by referring to the other languages in the group simply as Sha, Kulere, etc. “Hoffman (1971a) follows Greenberg (1963) in including Mober, Auyokawa, and Shirawa in this group. None belongs. Mober is a dialect of Kanuri while the other two are now extinct, having been replaced by Hausa. °called "North Bauchi" group by Hoffmann (1971a) and Skinner (1974). The List of lan- guage nanes has been copied directly from Skinner. It is not clear how many of these hanes actually represent separate languages and how many could be interpreted as dialects of a smaller number of Languages. “called "South Bauchi" group by Hoffann (1971a). I have rejected the terms "North Bauchi" and "South Bauchi" for two main reasons. First, the terms are geographically misleading since the Chadic languages spoken closost to’ Bauchi town are Gera and Geruna in the Bole group. Second, the terms give the impression that these are subsections of a single group, which scens not to be the case. Ina recent survey, K. Shimizu recorded sone 30 different names for languages in this group. Most of these’are village nanes and cannot be taken as independent languages. "the available evidence indicates that the Jaar group consists of three discrete clusters within each of which one needs to distin- guish no more than tvo or perhaps three separate languages. To list 30 language names would give a completely false picture of the true linguistic situation. In this case, I would suggest that it is better to err on the side of listing too few nanes than too many: it is mich easier to add a language name at a later date than it is to expunge a wrong or redundant nane once it gets established in the literature, cf. footnote 4, AAL 5, 5mm. B 3. P, Newnan, Bura group A . a. Bura-Pabir,’ Chibak, Putai D. Margi, Kilba Higi group igi, Bana (Nbana) 4/5/6. Mandara/Matakan/Sukur major group Mandara group a, Mandara (Wandala); Paduko; Glavda, Guduf, Dghwede, Gvoko >. Lamang Matakam group Matakan, Mofu, Gi ‘Sukur group ‘Sukur ga, Mada, Hurza, Muktele Daba group Daba, Gawar, Hina Bata group Bata-Bachama,* Gude (Cheke), Neangi (Jeng); Gudu Subbranch BM-B 1. 2. 3. Kotoko group Kotoko, Logone; Buduma Musgu group Mus gu Gidar group Gidar East Branch (EST) Subbranch EST-A AL 1 2. 3 Somrai. gro Somrai., » Ndam; Sarwa, Gadang, Mod, Miltu Nancere group | Nancere, Lele; Gabri, Kabalai Kera group Kera, Kwang Subbranch EST-B 5 2. Dangla group ‘ fa. Dangla (Dangaléat), Migana, Mahwa, Jegu, Mogum, Bidiyo bi Mi, Masmaje, Kayakse, Birgit, Toran Nokulu group Nokulu [MAL S/1 "The terms Bura and Pabir refer to two ethnically different peoples, who, however, speak ‘@ common language. “The terms Bata and Bachana refer to two ethnically different peoples, who, however, speak a common Language. MAL S, 6197) Chadéc Classigication and Reconstructions 3. Sokoro group Sokoro, Barain, Saba IV. Isolated Branch Masa group Nasa, Zime, Mesne, Marba, Musey BRANCH SSUBBRANGH crour _ Hausa —— Bote —<—<— Sess Pon Bade pe vt daar WEST Tera Bura Higi Mandara Matakan Sukur Daba Bata Kotoko — \ cider somai A <= Nancere Kora cast awnic, Cai sTU-MANDARA _— Dangla \ B ——" Sokoro MAL 5, 7 LANGUAGE (Hausa, Oxandara) (Bole, Kanakura ...) (Angas, Goemai ...) (Ron, Kulere ...) (Bade, Ngizim ...) (arji, Pata...) (Zar, Boghom . (Tera, Ga'anda (Bura, Margi. (igi, Bana) (Mandara, Lamang atakam, Gisiga (Sukur) (Daba, Hina...) (Bata, Gude...) (Kotoko, Buduna ...) (wusgu) (Gidar) (Sonrai, Tunak ...) (ancere, Cabri...) (Kera, Kwang) (Dangla, Mubi...) (Wokulu) (Sokoro, Barain ...) (Nasa, Zime ...)8 P. Newman [AAL 5/1 1.1, The Three Major Branches In N/M, the East and West branches were combined into a,single branch called "Plateau-Sahel", which was considered to be coordinate with Biu-Mandara.* This combined Plateau-Sahel branch was subsequently adopted by most other Chadicists (e.g. Hoffmann 1971a). In Newman (in press (a)), however, I reversed my earlier opinion and broke up Plateau-Sahel into its two branches, which were taken to be equally distinct from one another and from BM, thereby set- ‘ting wp three coordinate branches for Chadic. This was done because continued’ research drawing on new and better data on EST languages still failed to turn up any features—whether lexical, grammatical, or phonological—that could be said to be exclusive innovations shared only by’ EST and WST,”and that cold therefore serve as evidence in support of the Plateau- Sahel hypothesis. Aithough there is nothing wrong a priori about the idea of EST and WST forming a single branch, the evidence points more and more to their independence. By contrast, the unity of the BM branch seems certain, in spite of the considerable internal diversity. ‘While Kotoko and Musgu (and Gidar?) occupy a place apart within IM, they are Clearly more closely related to other BM languages such as Tera or Nargi than they are to any WST or EST language. 1.2. The Masa Isolated Branch ‘The Masa group has traditionally been treated as being closely related to Musgu (Westermann and Bryan 1952) and thus by extension as belonging to BM (Hoffmann 1971a, Newnan in press (a). Linguistic justification for this classification, however, turns out to be hard to provide—at least With the inadequate materials we now have at our disposal. Rather it looks as if the cultural and geographical proximity of the Masa and the Musgu has led us to ‘Suppose a close linguistic connection that does not in fact exist. Not only does Masa not seem to bear an especially close relationship to Musgu, but it doesn't even seem to fit in the same branch of, the family. Masa may not have any Special features that would allow one to classify it as EST or WST, but it doesn't share the distinctive characteristics of the BM branch either. I have decided therefore not to include the Masa group in any of the three major branches but rather to consider it provisionally as a separate, independent category. By treating the classification of Masa as a completely open question, the problem can be approached with a fresh perspective, leading hopefully to a quick solution once fuller descriptive data becone available. 1.3. Subclassification within the Nest Within WST, the Hausa, Bole, Angas, and Ron groups constitute a subbranch (WST-A) as opposed to the Bade, Warji, aid Zaat groups (WSI-B). Within WST-A, the Bole and Angas groups bear an especially close relationship, which is indicated by referring to the two groups together as a "najor group". The relationship between Hausa and this "major group” is also very close. In the case of the NST-B subbranch, the ties between the groups making up the sub- branch are less evident. This may be because the languages are in fact less closely related to one another or it may simply be because our knowledge of the Warji and Zaar languages is Jess advanced. Nevertheless, since the Bade group shares certain distinctive features with Warji and Warji shares certain features with Zaar, the hypothesis of a WST-B subbranch co- ordinate with WST-A seems reasonable. Within WST“B, the correct grouping is almost certainly Bade-Warji as opposed to Zaar, contrary to earlier indications on the question (e.g. Green berg 1963). °At the time it,was thought that the, two branches might correspond to Lukas" distinction between "Ghadic" and "Chado -Hanitic"” (Westermann and Bryan 1952). This was wrong (see Newnan, in press (a)). The major branches of Chadic differ from iukas' earlier classi fication both in substance and in conception, AALS, 81977] Chadéc Ceassigication and Reconstructions 9 1.4. Subclassification within Biu-Mandara The, large punber of languages that in Lukas’ earlier classification made up the "Mandara" or "Chadic" family as opposed to the "Chado-Hanitic” family constitute a distinct subbranch (INA) within EM as a whole. The other, and mich smaller, subbranch (IM-B) consists of the Kotoko, Musgu, and Gidar groups, although the assignment of Gidar to BN-B is extrenely tenta- tive. ‘while the intemal structure of the BN-A subbranch has not yet been worked out, we can be sure that its eight groups could not be equally distinct from and coordinate with one another. However, constructing a neat family tree diagran turns out to be extrenely difti- cult because of overlapping innovations and criss-crossing isoglosses. For the monent, then, T have been content to start building from the botton up by conbining groups into major groups—nanely 2/3 Bura/iligi and 4/5/6 Nandara/Natakan/Sukur—and by suggesting a conmon node for these two major groups. In this latter respect, I have taken a different position fron Greenberg (1963) who put Bura, Bata, and Tera together in one large group as opposed to Mandara and Matakam, Which were assigned to two different, separate groups. 1.5. Subelassification within the East Within the EST branch, there appears to be a fairly sharp distinction betwoon the languages belonging to the Somrai, Nancere, and Kera groups (EST-A) and the languages belonging to the Dangla, Mokulu, and Sokoro groups (EST-B). By contrast, the exact relationship anong the groups ‘making up each subbranch is mich less clear; and’I have therefore refrained from in- dicating further internal structure until fuller data on these languages become available. 2. PROTO-CHADIC PHONEMIC INVENTORY 2.1. Consonants ‘The consonantal inventory of PC can be reconstructed as follows: Pp a e k we » a i 9 a ” 5 ¢ " é 4 (ah) x os - z 4 he n © y 2.1.1, The *p/*§ contrast As proposed by Greenberg (1958) and supported by I11it-Svityt (1966), *p and *4 were probably distinct phonenes in PC, In wany present-day Chadic languages, e.g. tlausa, these phonemes do not contrast, cnly one or the “ther being found. As a result, it is often difficult in the case of individual lexical items to know which consonant to reconstruct for PC. Nevertheless, the evidence taken as a whole indicates that *p and *{ were phonemically contrastive in PC MALS, 910 P, Newnan’ LAL S/L and that the common non-distinctiveness of the pair is due to mich later (often independent) mergers. 21.2. PC Ab PC probably did not distinguish "b from *v, although the question is still far from settled. In Ngizim, for example, both 6 and v occur’ in native Chadic words, e.g. bato ‘give’, vaj ‘monkey’, "but the difference does not correspond systenatically to anything elsewhere in the family that would allow one to conclude that the contrast represents the retention of a PC feature rather than being the result of a localized, internal split. In most Chadic lan guages that have a 6/v contrast—many do not—only one of the two consonants normally occurs in native Chadic words, the other being prinarily restricted to loanwords. Although I have adopted the letter b to represent this proto-phoneme, its pronunciation in PC may well have varied between bilabial stop, bilabial fricative, and labio-dental fricative. 2.1.3, PC ax The connon occurrence of velar fricatives in Chadic languages, especially in BM, suggests that *x (and its labialized and palatalized counterparts) could have been present in the PC consonantal inventory. It has not yet proved possible, however, to determine the modem re- flexes of this presumed PC phonene. It may be that /x/ in BM is a direct continuation of RC *x, hich in NST merged with some other consonants), such as ty 4 (or sh), or §. | Comare Br, 230, N. kota ‘steal’; T. xat, B. sat ‘hand’; T. xadi, Kk. fade ‘illness’. But until additional ‘evidence is provided,’ the existence of *x in PC mist be considered a likely but unverified hypothesis. 214. PC'S In addition to *6 and *d, PC had a third glottalized consonant, here, indicated *'J. ‘This consonant was probably a’glottalized palatal stop, differing fron °f prinarily by position rather than manner of articulation, and not an ejective like Hausa k' nor an approximant Like Margi 'y. | In present-day Chadic languages this proto-phonene is variably realized as ‘oy, ‘a, k', 'w, 'y, % or ¢ 2.1.5. The absence of glottal stop PC did nor have glottal stop as a phonone, either internally or at the beginning of words (ewan 1976). Contrary to the characteristic Afroasiatic word structure which requires all words to begin with a consonant (e.g. Diakonoff 1965), PC words such as am ‘water! and *aku 'fire' were truly vowel initial. 2.1.6. PC My In N/M it was observed that 4 in WST and EST languages sometimes corresponds to BM 4 and sonetines to BM fit. On the basis of the evidence then at hand, it was assumed that there was only one proto-phonene *4, which had undergone a split in EM. In Newman (in press (b)), a paper devoted to the problem of "hlaterals" (i.e. lateral fricatives), I rejected this earlier position and showed that in addition to *s, one had to reconstruct another sibilant for PC, for which T have adopted the synbol *4. What phonetically distinguished *y fron *s 4s not inom, but it seens clear that PC M4 wiS an s-1ike or, nore likely, sh-like fricative and not a hlateral as it later becane in BH.!® ‘Russell Schuh and Suzanne Platiel (independently in personal conmmication) have suggested that the sibilant *s was not merely 4h-like, but rather was probably the PC sh as such. AAL 5, 101977) Chadéc Classigéeation and Reconstructions u 21.7. PC hE ‘The comparative evidence points to the conclusion that PC had one and only one lateral phoneme (Neiman, in press (b)), Phonetically it was unusual in that it was probably pro- hounced as a voiceless hlaterai |$), rather than as a nomnal frictionless |1}, at least in initial position. Because of this "aberrant" pronunciation, I have adopted *he rather than ‘¢ as the symbol for this proto-phonene even though structufally speaking it constituted sus PC /1/. 2.1.8. The problem of prenasals The reconstructed consonantal inventory presented above still does not include prenasalized consonants (mb, nd, nj, ng, etc.) and the problem of the origin of prenasals in Chadic re- hains unsolved. Greenberg’ (1958) had postulated their existence—mb specifically—not only for Proto-Chadic but for all of Afroasiatic. Recognizing that the original evidence was thin, Greenberg subsequently reaffirmed his position on prenasals and offered a list of nineteen etymologies as evidence "tending towards the establishnent of an original [Afto- asiatic] *nb" (1865:89). Of these nineteen etymologies, fourteen drew on evidence from Chadic languages (nainly Hausa); and of these fourteen, only one is reconstructable for PC and this not necessarily with a’prenasal (see word list no. 97 place"). Thus as far as Chadic is concerned, Greenberg still hasn't begin to prove his hypothesis. Similarly, the eight Chadic etymologies with *mb proposed by I11i®-Svityé (1966) are mch too weak to pro- vide any teal support for the idea of prenasals in PC. ‘The issue of prenasals is complicated because there are in fact two problens to solve: (i) did PC have prenasalized consonants and if so what subsequently happened to them in Chadic Linguistic history? and (ii) what is the origin of the prenasals one now finds so wide- spread in the Chadic family? These may turn out to be the same or related questions but not necessarily so. They may be entirely independent questions and the failure to recognize this may partially account for our inability to make any progress towards solving the problem. In the comparative word list presented here, there is one item (no. 15) where an mb in NST corresponds to an mb in BM and a few others where 6 in WST corresponds to m or mb in EM. What the significance of these scattered exanples is I cannot say. 2.1.9, Palatal (ized) and labialized consonants PC almost certainly had palatalized and labialized velars (7, 9", etc.). These are included in the consonantal inventory and are used in the reconstructions, ‘where their correctness in individual cases will ultimately depend on decisions regarding the PC vowel system. PC nay also have had palatalized and labialized bilabials (b", {*, etc.) such as occur, for example, in Margi and Iigi; but evidence for these is yet to be found. As far as the pala- tals (c, 'J, etc.) are concerned, we can assume that they were present in PC, either as palatalized’alveolars (i.e. ¢ = #) orasaseparate palatal series defined in’terms of a dis- Einct position of articulation, Because of the inherent phonetic instability of palatals, it has been impossible to establish regular correspondences or find other direct evidence that would prove the validity of this assumption. Nevertheless, when one takes into account the indirect evidence bearing on the problem, and considers the’ inventory of PC consonants and vowels as a phonenic system, the probability of PC having had phonenically contrastive palatals sens very high. 2.2. Vowels PC can be reconstructed as having had ar most four phonenic vowels i 2 u, and possibly only two, v and a. a AL 5, 11a2 P. Newnan [AAL 5/1 2.2.1. Number of vowel contrasts Anong present-day Chadic languages, Tangale has been described with nine vowels (Jungraith- mayr 1971), Dangla with seven (Fédry 1971), and Bole with five (Lukas 1970-72); but the characteristic Chadic pattem is six vowels a, 2, 4, u, ¢, 0 (ll, Wolff 1959). 'Of these six, € and 0 often have a secondary status, being restricted to loanvords, being tecently derived from diphthongs, or being conditioned variants of other vowels. In many languages, more~ over, even the four remaining vowels are not fully contrastive, the distinction between and 'u, 2 and é, and/or 9 and u being neutralized in specific phonological environnents. ‘The interesting’studies by Mirt (1969),'! Parsons (1970), Mohriang (1972), and Hoskison (2975) illustrate various languages in which a wide range of phonetic vowels can be reduced to two (or in sone positions, three) phonemic vowels contrasting only in vowel height. ‘The comparative evidence points in the sane direction for PC, although it would be premature to claim at this time that PC only had *9 and *a and that *<’and *u were merely non-contras- tive phonetic variants. On the whole, the choice does seem to lie between a two-vowel system anda four-vowel system with a’balanced *4, *a, *u system representing a remote pos- sibility.'? ‘The most likely explanation, however, is’that PC was characterized by the same type of distributional restrictions that’one finds in present-day Chadic languages. Thus no blanket statement that PC had this or that nunber of vowels would be correct as’such. Rather, one would have to specify how many vowels and which vowels did PC have in initial position, how many and which vowels in medial position in open syllables, etc. In the re- constructions in this paper, I make use of all four vowels in final position, two vowels (é and a) in initial position, and, with a few exceptions, two vowels (o and a) "in medial posi- tion. It should be emphasized, "however, that these vowel reconstructions are extremely tentative and in each case mist be subjected to critical examination before being accepted as correct. 2.2.2. Vowel length PC probably had phonemic vowel length, but only with a and only in word medial position. This feature is not indicated in the reconstructions. The other vowel(s) probably did not have a length contrast. 2.23. Tone/Accent PC must have had distinctive tone or pitch accent, although it is not clear which. If PC were a tone language, it would have had a simple, primitive two-tone system, In this paper, no attempt has been made to consider tone or aécent in the reconstructions, nor is tone indi- cated in the individual language citations. 1m my opinion, Nirt's modest paper represents a major breakthrough in the analysis of Chadic vowel syStens. 1the best comparative study of Chadic vowels is a yet unpublished paper by Schuh (n.d. ()). In tackling the problem of Chadic vowels, I have benefitted greatly from this paper and from recent discussions with Schuh, whose ideas on the subject have indepen- dently developed in a direction more or less'parallel to mine. AAL 5, 121977) Chadie Classification and Reconstructions 13 3. SOUND LAWS ‘The identification of true cognates, and the reconstruction of proto-forms depends on the knowledge of phonological changes (""sound laws") that have affected a family from the time of the proto-language to the present. These changes are of different sorts: varying in age and scope, applying to a whole branch or to an individual language, applying to a phylogenetic group or to a geographical area, being historically specific or being repeated independently at various places and times, being essentially exceptionless in the manner of a sound “aw” or being describable in terms of drift or tendencies, being strictly a fact of history or still being operative as a productive or semi-productive synchronic mule. For Chadic we are now in a position to describe a number of sound laws—at least as far as consonants are concened—and thus cone to a better understanding as to why certain phonetically dissimilar items are cognate (e.g. N. psa, T. wuzon 'grass'; or Tm. he, J. 4ote ‘two') and why certain phonetically similar items are ‘not (e.g. H. wutaa, G. waata "fire'). Some of these phonological changes have been described elsewhere (e.g. N/M; Newnan 1970; Schuh 1976); some are being described here for the first tine. in either case, the sound laws are generally quite straightforward, if not entirely regular, and well supported by the available evidence. The accompanying examples are thus meant merely to illustrate the sound laws in question, not to prove their validity. The sound laws de- scribed here are of course not intended to be exhaustive. What I have tried to do is to outline some of the more inportant or phonologically interesting changes plus sone pethaps less important changes when they concern better known Languages. In describing the changes, it is to be understood that the starred form always refers to the sound undergoing change at the time the change took place without necessary reference to its ultimate source in PC or in the proto-language of the relevant branch. ‘The presentation of the sound laws follows the order of the branches adopted in the classification, namely WST, IM, EST, and Nasa. Within each branch, the individual groups and languages’ discussed are identified according to the classificatory system used in the language index. 3.1. Two Changes Affecting Subbranch WST-A So far it has not been possible to document any regular changes applying between PC and the WST branch as a whole. Two changes can, however, be described at the subbranch level, both affecting WST-A. (a) *hl >2, e.g. S. fuoa 'meat', cf. N. fluted and L. hewsd. If I am correct that PC (and proto-WSr) ‘did not distinguish /il/"from /1/ (see’ section 2.1.7), then this change would have consisted in a simple phonetic shift, originally without structural consequences. (b) 4 >4, e.g. S. sum 'nane', cf. M, fiom and Mb. sami. ‘This change, also found in the EST branch and in Masa, did away with the PC distinction between *, and’ the nomal “4, a distinction preserved in different ways in WST-B and in BM, 3.2. Seven Changes Affecting Hausa Because Hausa (1.A.1) has been relatively well studied, we know more about the phonological changes it has undergone than in the case of most other Chadic languages. Seven sets of changes are described here. (a), **>y, e.g. Sooyaa 'fry', cf. B. suru; mai 'oil', cf. A. mit. This change seems to have been perfectly regular’ and’ thus present-day Hausa cannot qualify as a reflex.of a and should not as a rule be identified with in comparisons with other languages." ‘there are a few anomalies in this regard, see word list nos. 37, 64, and 74. AL 5, 15u P, Newman, IAAL 5/1 Examples of #2 > y are also found, e.g. biyu 'two', cf. Kk. belu, but this change is of a sporadic nature only and docs not’ represent a regular Sound law. (b) * >, e.g. garaa ‘termite’, cf took place after the * > y change. R, ngae; harshee 'tongue', cf. Kl. ats, This change (c)_#N > 9 in final position, e.g. zaaboo 'guinea-fowl', cf. N. zaaban. This sound law, discovered by Schuh (1976), applied both to *x and to *m. Other examples of words with lost final nasals are uku ‘three, haneti nose", and guzaa ‘monitor’ (for lost *n) and hadaa ‘crocodile’, wsux ‘nouse’,’and k'éisiéé ‘thirst’ (for lost *m). @ PC has been reconstructed with a third glottalized consonant *'J in addition to 6 and In sone of its instances, llausa k’ is a reflex of, this proto-phonene, e.g. k'ashec "pone' <*"Jagu, cf. T. 'goh€, Nore often, hovever, k' ig a reflex of *&, having doveloped by an intemal Split of #k into & and k', é.g. kad thead', cf. K. kod; j2k'a ‘become wet", cf. K, yeke. While the fact of the split seens certain, given the comparative evidence, it is not certain what was the crucial conditioning environment. In Newman (1973) 1 suggested that the najor conditioning factor for the *k >k' change was the presence of a voiced con- sonant in the preceding or following syllable. This hypothesis—which I still feel to be on the right track—avaits verification. (e) Glottal stop achieved phonemic status in Hausa only recently in conjunction with the introduction of loanwords from Arabic (Greenberg 1947). Prior to the period of Arabic loans, glottal stop already occurred in initial position as a phonetic feature of word onset, but not as a phoneme in its own right. Before the phonemicization rule |?] > /?/, words that now begin with a glottal stop, such as 7édoo'eye’ or ?ajee'put down’, would have been truly Vowel initial from a structural point of view. All of this is well-fnown—although too often forgotten. hat is not know is that initial fi!* in Hausa had essentially the same origin as”, having developed from a low-level, non-phonemic feature of word onset (Newnan 1970), e.g. harshee "tongue' < *arshee, cf. R: alis. Since h developed internally out of ¢ the only legitinate comparisons involving Hausa h are with ¢ or other consonants known 0 have derived from 9 rather than with historically old h-Like or x-Like consonants. (4) In syllable final position, velar obstruents weakened to u, alveglars to #, and bilabials to u (the latter change being restricted to the "standard" dialéct).'* These well-known hanges (KLingenheben 1927/28) gre best {Llustrated by synchronic alterations within Hausa suchas the foligang:, tefake Feomoner’, fatauedd poverty fade four, ue dude "four each"; makaagoo "blind man’, makau-niyaa 'blind woman" (g) The palatalization of alveolars before front vowels, t/d/s/z + e/j/sh/j, still operates aS an active synchronic rule in tlausa, In considering words that contain these palatals, one simply has to indo the palatalization rule, e.g. c¢ 'eat' <*¢i. For historical purposes the important point to note is that j can Tepresent either *d oF #2, e.g. jee ‘go’ <*de, cf. Ne da; jékié 'body' <*ziké, cf. B, 2iino. teference here is to h before @ and not to the phonetic/orthographic ft before o and w which is an allophone of /£/. “The resultant a, and thus the existence of the historically prior obstruent, is often hidden because of an accompanying rule changing ia to uu, e.g. juwnaa ‘each other’ < *jiuna y (followed sometines by a secondary change y > w or y > @); while in Pero, *4/z > c (with automatic intervocalic voicing as indicated above), e.g. K. yo, Tg. Jo, P. co ‘leg’, cf. Hl. sau. In Bele (1.4.2), a language more closely re- Jated to Bole than’to’these three languages, *4/#z > h, ¢.g- dhé 'fire', cf. B, (whosés héko "body', cf. B. zééwo. This universally comon change of sibilant to h'has been re- peated independently a nuiber of times in the Chadic family. 3.6. Phonation Changes in the Angas Group In proto-Angas-Goenai (I.A.3), initial voiced obstruents becane voiceless (with a few excep- tions) and voiceless obstruents became ejective (Greenberg 1958, Hoffman n.d.). Thus the original voiced/voiceless contrast was preserved although in a different phonetic form. This is the present state of affairs in the Goomai cluster. In the Angas cluster, however, the ejective consonants lost the ejective feature and merged with the corresponding voiceless consonants, e.g. Gm. pwee, A. puo ‘mouth’, cf. B. bo; Gn, p'et, A. put ‘go out", cf. Be pataa; Gn. sek, A. shék "body", cf. Ge. zwoi; Gn. 4'em, A. som 'name', cf. Ge. Suma. 3.7. § > ylww im the Bade Group ‘The PC distinction between * and *s was apparently preserved in the NST branch until the split of the tvo subbranches. In WSI-A, the consonants then merged into 4, In the Bade group (I.B.1), *} changed into a semivovel (y or u), c.g. N. yaanau ‘tooth’, cf. G. hean and Jg. san; N. wand "send", cf. Lo. han and Ms. sun.’ In the other two WST-B groups, Warji (1-B.2) and’Zaar (1.B.3), 4 now appears variably as c or sh or even as fi—the sound laws that account for these reflexes have not yet been worked out. *8In citing Tangale and Pero examples, I have folloved the original sources in indicating intervocalic voicing even though this'is not necessary on phonemic grounds. '7P includes p/§, 6, 6; T includes 2, d, d} and K includes & and g. ‘In the Shani dialect, there is now a tendency to replace intervocalic w by 6, e.g. tubd = tuwi. This applies to’all w's in that position and not just to those derived from P. AAL 5, 15)16 P. Newnan [AL 5/1 3.8. Development of Hllaterals within the Zaar Group Jn the Zaar-Barava subgroup of fhe Zaar group ({B-3), prevocalic *s apd "5 changed into the corresponding hlaterals ht and ht, e.g. 7 klya'drink', cf. B. 4a; 2. hee ‘body’, cf. B. ‘ziudo. ‘The resulting voiceless hiateral merged with the hé deriving directly from PC *he. 3.9. ‘Two Changes Affecting the BM Branch In Biu-Nandara, two changes can be described as applying to the proto-language of the branch as a whole. (a) *6 >v, e.g. T. vara 'give', cf. N. bar2 and D. bere. As indicated in section 2.1.2, PC probably did not distinguish 6 from y and thus the phoneme *b may already have had both pronunciations in the proto-language, either in complementary distribution, in free variation, and/or as dialect variants. (b) “>, e.g. M, heom "name", cf. Ge. sama, Mb. samé, and Ms. Sam-ma. In BM, this hE from *$ merged with the he coning’from PC *h, the history of the two thereafter being indis- tinguishable. The origin of the voiced hlateral fZ, now extremely common in BM, is unknown. ‘The ht/AL contrast seems to be very old (perhaps proto-BN-A if not proto-BM) but it cannot be traced back to and identified with the PC *$/*h{ contrast. Rather it looks to be a secondary split subsequent to the merger of the ht's coming from the two PC sources. 3.10. Weakening of Velar Stops in the Tera Group In the Tera group (11.4.1), initial velar stops have comnly, but not systematically, changed to yor w, eng. G, yirga 'tish', cf. Br. kélga; T. wuzon ‘grass’, cf. Kk. gozm. In Hinal position, velar stops are usually realized as x, ¢.g. T. gomox ‘roan antelope’, cf. H. guoanbid, 3.11. Phonation Changes in the Tera Group In the Ga'anda-Hona division of the Tera group, all obstruents have deyoiced, thereby con- pletely eliminating the voiced/voiceless contrast, e.g. G. {ara ‘blood’, cf.'D. baer; Hona hum 'ten", cf. K. gum, In the Tera-Jara division of this group, the tendency has been in ‘the opposite direction. With a few unexplained exceptions, all’ fricatives in initial posi- tion (including hlaterals) have become voiced, e.g. T. vat''four', cf. Be. {wat, J. Ata ‘cow’, cf. Gs. héa. ‘The result of the criss-crossing voicing rules is that a pair of examples such as T. 2a, G. sa ‘drink’ gives no hint as to the voicing of the original proto-consonant, for which oné has to depend on languages such as Hausa in WST, Bachama in BM, or Dangla in'EST, where the voicing contrasts of PC are usually preserved. 3.12. St > fin Tera Tera, in this case the individual language rather than, the group, has also undergone a fairly systematic change of intervocalic *t to d, e.g. T. moda 'die', cf. N, mata. 3.13. > £ in the Bura/Iiigi Najor Group In the Bura/Higi major group, * shifted to £ in a regular manner. In the Bura group (II.A.2), the * > £ change was an unconditioned sound law applying to *4 in all positions, e.g. Br. fa ‘dig’, cf. T. aa; Br. sat leg’, cf. T. sata; Br. hitga ‘fish’, cf. G. yirga; Br. pela ‘stone’, cf. Be. fata. “In the Higi group (II.A.3), the *a >£ change took place in all posi- tions EXCEPT in the environment CatV where the C was an alveolar obstruent, e.g. La ‘dig’, AAL 5. 161977} Chadie Classification and Reconstructions uv hotpe "fish", pate ‘stone’, but sara ‘leg’ and tore 'noon' (cf. T. ndvta). The condition- ing environment, which seems strange at first, ceases to be so when ane thinks of the Car Sequences as Ct'clusters, their more usual prommciation (Mohrlang 1972:50). 3.14. #2 >A in the Bura Group In the Bura group, final *n changed to x, e.g. Br. hia ‘tooth’, cf. T. fén; M. maar ‘three’, cf. G. naxkan, This change mist have taken place after the * >£ change. It is not possible to know whether Higi also underwent this change since it has Tost all final consonants. : 3.15. Devoicing in the Bura Group In the Bura group there was a general devoicing of obstruents, e.g. M. kum 'ten', cf. J. guom; Br. som 'eat', cf. Gs. zom. ‘The existence of this historical rule is quito evident from’ the comparative evidence even though all languages in the Bura group now exhibit a voiced/voiceless contrast with obstruents. 3.16. Labio-alveolars and Labio-palatals in the Bura Group The most striking phonological feature of the Bura group is the existence of similtaneously articulated labio-alveolar and labio-palatal consonants, bd, pe, ps, etc. (Hoffmann 1963) .!? Historically these resulted fron the loss of an intervening’ vowel ahd thé consequent. fusion of the resulting consonant cluster, i.e. ACVCV > "CCV > CLV, e.g. Br. ped ‘sun’, cf. Mg. {uti; M. bits 'forge', cf, T. Gail). "AS a gonoral nule applying to the group a a whole, the vowel reduction and consonant fusion only took place if the initial C was a labial. "In Margi, the rule was extended to initial velars, which becane bilabial in the process of the fusion, e.g. M. paar ‘grass’, cf. Br. kusans M. ptat ‘chief’, cf. G. kutina, 3.17, Loss of Final Consonants in Higi In Higi (11.A.3), all word-final consonants have been dropped and intervocalic *d’has weakened to, e.g. za 'nan', cf. Br. sa€; sé ‘beans’, cf. Br. tsar; hae 'mouse', cf. Mt. wom; furre "Four", cf. M. sade. 3.18. *1 > in the Mandara Group In the Mandara group (II.A.4), # changed systematically to £, but with a muber of restric tions. In syllable final position, the * > £ change took plice unless the "x abutted with a voiced consonant, e.g. Mn. kofpo ‘fish’, cf. Ge. keruga; Mn. voe give’, cf. T. vara; but Mi. otdza "scorpion", cf. Bd. wurjon. Mediaily the restrictions on the change were the same,as that described above for Higi’ (section 3.13) with the addition of kia) as a "cluster" protecting the *, e.g. L. patak ‘stone’, cf, Bc, fara, but L. s{a)na ‘leg’, cfs T. sata; L. k(alAana ‘crocodile’, cf. T. jéran; L. hed ‘dog’, cf. Hi. kote. 3.19. *m >w in the Mandara Group In the Mandara, group, *m changed into w, as seen in the equivalent nanes Nandala = Mandara, evg. Mn. yaws ‘water!, cf. Gs. yam; L. we ‘mouth’, cf. T. me. While a few exceptions }aLthough written as digraphs, these have to be treated as unit phonemes Like the labio- velars kp and gb So common in West Africa. AALS, 1718 P, Newnan [MAL S/1 exist, this shared rule constitutes a distinguishing feature of this group. 3.20. Loss of Hlaterals within the Bata Group In most of the Bata group (II.A.8), the former hlaterals, both voiced and voiceless, have becone frictionless £, e.g. Bc. Lin-to ‘tooth’, cf. T. Abin; Gd. ta ‘stand’, cf. M. a. As a result, Bachana, for example, no longer has any hlaterals. Gude does have hl’ in a small number of words, but this seems to be a re-introduction rather than a retention. Gudu, however, although clearly a closely related menber of the Bata group, did not undergo this change and thus preserves the historically earlier hlaterals. 3.21, *¢ >A in Bachama In Bachana (II.A.8), intervocalic *¢ changed to A, e.g. gorey 'sun', cf. Mg. fut. Inter- estingly, the tule did not apply to the suffix -t) used with feminine nouns (e.g. hoo-t2 ‘she-goat', Liki-ta 'noon') but it did apply to the third person feminine pronoun *-22, e.g. baago-no ‘her father". 3.22, 1 >a and Other Dialect-specific Changes in Kotoko In Kotoko (I1.B.1), intervocalic *n changed to 2 while final *n changed to n, thereby pro- ducing the consonantal alternation seen in many singular/plural pairs, e.g. gap ‘hut’, pl. gore; shan "tooth", pl. share, The #n > % change mst have taken place very recently’ since it was not shared by the closely related Logone language, e.g. ko. gotan, Lo. ganam ‘woman’ ; Ko. sauare, Lo. sudane ‘dream’. In the Makari dialect, from which all Kotoko examples in this paper’ are cited, hlaterals have been completely lost, *h having altered into sh, and *he! (a glottalized hlateral of undetermined origin) into’c', e.g. siu ‘meat’, c'a ‘Laugh’, cf. flu and nkt'a from the Gulfei dialect. In the Gulfei dialect, * has weakened to y under undetermined conditions, e.g. yin ‘know’, gbivim ‘mouse’, cf. Makari ay and kusum. The sibilant 4 nevertheless remains a common consonant in the Gulfei dialect. 3.23. Change of Sibilants to h in Buduna In Buduna (I1.B,1), all initial *s"s (including those derived from *z and *ht) changed into h, e.g. hue ‘drink! hum "eat", hanai "tooth", cf. Lo. se, zm, hind. Nedial *s's were also lost—present-day Buduma’ does not have any Sibilant$—but what they changed into is not Inown. 3.24, Loss of Hlaterals in the EST Branch Since hlaterals are not found in any EST languages, it is reasonable to assume that the change of PC *h€ to some other consonant(s) must have taken place very early in that branch, if not in proto-EST, The details are yet to be worked out, but the available evidence suggests that *h€ changed into ¢ in initial position and to a palatal affricate or frica- tive in medial position, e.g. Tm. Lob ‘hit’, cf. T. fitabo; Mb. icca ‘cough', cf. M. ‘ada. As far as PC *{ is concerned, the change to 4 in proto-ESI was straightforward and regular, e.g. Mb. sand *nane', cf. M. iom and KE. sum; Jg. aso ‘bone’, cf. Lo. alte and He Rashée. 3.25. Three Changes Affecting Tunak Tunak (IIT.A.1), one of the few EST languages for which we have a substantial word list (Caprile 1975), "has undergone a nunber of phonological changes, such that its words are now1977] Chadic Classification and Reconstructions 9 phonetically quite different from cognate forms in related languages. Three inportant changes are described here: (a) #6 >k in initial position, * > j elsewhere, e.g. he ‘drink', cf. Ke. saa; hin 'tooth’, cf. D. saayo; aj 'cough', cf. Sn. aso; aaj ‘laugh', cE. Nb. gamas~? “The one now finds in Tumak represents a new introduction and cannot be considered a reflex of the PC sibilant. (b) * >,9 or & [a retroflex vibrant], e.g. ba ‘blood’, cf. D. bar; he ‘two', cf, D, seers dah "noon', cf. Ne. ke-dere. One possible explanation for the two reflexes of"*a is that * vas lost in final position and changed to & in intervocalic position, followed by a sub- sequent dropping of the final vowel. (c) *£ >a in non-initial position, e.g. do-bor ‘ashes’, aru-nan ‘tree (pl.)', cf. D. butu and eto. 3.26. Some Masa Correspondences As long as the classification of the Masa group (IV) is unsettled and the extent of its intemal differentiation unknown, one cannot hope to provide an accurate historical descrip- tion of its phonological development. AL1 one can really do is to match PC phonemes with corresponding Nasa phonenes without describing the processes by which one changed into the other. Some of the regular correspondences inclule *g = 4, *% ©, and *-d- =n, e.g. Ms. aat-na "eye's cf. Br. RUM, Ge. 4a NS. hulug=fa‘Eish'y ef, Gov keru{as Me. Gno-na iy", ef. Sm. dou. 3.27. Common Chadic Neutralization of Phonation Contrasts A characteristic feature of the Chadic family is the existence of three phonation types for stops: voiced, voiceless, and glottalized. As far as I an aware, all Chadic languages that allow final stops—many do not—neutralize the voiced/voiceless contrast in that position and many neutralize all three contrasts, e.g. Kf. paat ‘five’, dapit 'nonkey', swat ‘ashes’, cf. B. baadi, bido, bute respectively. ‘Since the neutralization is a conmon, ’recurring Chadic feature, the individual manifestations can be interpreted easily and do not need to be accounted for by language-specific historical rules. 4. PROTO-CHADIC WORD LIST ‘The following list includes 150 Chadic etymologies, each with @ starred form reconstructed for PC. About half of the etymologies appeared in'N/M; their inclusion here can be taken as further confimation of their probable correctness. The other half are new. The N/M items that do not appear in this list are not necessarily errors as such. Sone are, but others have been replaced because (a) they are valid etymologies for one of the branches but not for the entire fanily; (b) there is a question whether they are cognates or early loan~ words; or (c) the evidence to justify their inclusion is relatively weak, even though they nay be perfectly good etymologies. In N/M, a distinction was made between items that were "second-level" confidence and those that were presumably first-level confidence. This dis- tinction has been discarded here since there is really a continuum, not a dichotomy, between the "sure" items and those where greater doubt exists. Also, the confidence one places in a possible etymology depends not only on the quantity and quality of the extemal evidence but also on the relative weight accorded different variables and on one's scientific view point and intellectual disposition. AAL 5, 1920 P. Newnan TAAL 5/1 4.1, The Reconstructions Each etymology is accompanied by a starred form which represents the hypothesized PC word. ‘The reconstructions make use of the inventory of consonants presented in section 2.1 with the occasional indication in parentheses of possible prenasalization, Four vowels (a, 2, 4, uw) are used in the reconstructions although, as indicated in section 2.2, the contras” tive status of the latter three is open to doubt. Vowel length and tone/accent have not been reconstructed. In initial position, only two vowels appear, @ and 4. In fin: tion, all four voxels have been used except in the case of polyconsonantai verbs ("poly- verbs"), where only @ and 9 have been reconstructed in accordance with my theory of Chadic verb classes (Newman 1975). Medially, all four vowels have also been used but the instances of 4 and u are extremely rare. Sonetines it has not been possible to reconstruct a specific vowel for a proto-form; in these cases a hyphen is inserted as a place marker for the vowel. Alternative reconstructions for whatever reason are separated by @ slant line; doubtful units are enclosed in parentheses. ‘The general appearance of the PC reconstructions is very mich in line with our usual in- pressions of Chadic languages. For example, we find diconsonantal words to be the norm, Ronoconsanantal words to be less numerous but still quite common (but mostly with verbs’ and function words), “and triconsonantal words to be rare and limited alnost exclusively to nouns. We find, moreover, that consonant clusters do not occur, that syllable-final and word-final consonants do occur but only in the case of sonorant3, and that words never con- tain more than one glottalized consonant. The only really unusual feature—unusual from an Afroasiatic point of view—is the existence of words beginning with a vowel (see section 2.1.5). While these are less common than consonant-initial words, they are far from rare in the reconstructed list. 4.2. The Citations Each PC reconstruction is illustrated by selected examples from present-day Chadic Languages. ‘The citations are grouped according to the three major branches with Masa citations kept apart: (1) = NST, (2) = BM, (3) = EST, and (4) = Isolated (Masa). The individual languages cited are identified according to the abbreviations given in the language index. I have generally limited the examples to two per branch (drawn where possible from different sub- branches or different groups) unless there was sone particular point to be made that required further citations. Many well-known cognates in liausa, for example, are not included because ‘the reconstructed proto-forns are adequately supported by equally good cognates in other NST branch Languages. 4.2.1. ‘The transcription system Citations have been taken from approximately fifty different languages representing every group in the fanily.’’ In order to best present data fron such a large number of Languages, T have taken the liberty to make mincr adjustments in the original source materials where it seemed called for and to standardize the examples according to one transcription system, ‘The adjustments include the following: (a) substitution of ¢ and 0 for « and 9 in languages such as Pero and Sura where these vowels do not contrast; *0"Group" here refers specifically to the unit of that level employed in the classification. MAL 5, 201977) Chadie Classification and Reconstruetions a (b) substitution of 2 for Margi g (Hoffmann 1963) ; (©)_wse of » for the lower centralized vowel in Bachana, often transcribed as 6; (@) substitution of 2 for all high vowels in Mandara; (e) substitution of x for h in all languages where the h is clearly understood to represent a voiceless velar fricative; (£) deletion of initial glottal stop in all languages where the glottal stop appears not to be phonemic. ° : eG The standardized transcription system used in the examples employs the following special symbols and conventions. we, he voiceless and voiced hlaterals (lateral fricatives) [usually transcribed in sources as 2, de or 4£, zl] oi voiceless and voiced (pre-)palatal affricates ah, zh voiceless and voiced (pre-)palatal Fricatives uy woiceless and voiced velar fricatives cs rolled 4 (but only when it contrasts with another * as in Hausa) retroflex vibrant mb, nz, nj, etc. — prenasalized (""semi-nasal") obstruents bd, bz, pe, ete. coarticulated obstruents 6, 'g, etc. _—_glottalized ("“implosive") obstruents 0 glottalized senivovels ejective obstruents ° glottal stop aa, ii, ut, etc. long vowels If the meaning of a particular Language citation is different from the meaning or meanings given for the etymology a5 such, this is so indicated. A question mark after a citation in- dicates that the cognation of the particular item is doubtful even though the etymology it- self might be quite certain. Hyphens within a word indicate morpheme boundaries. These are most often used in the case of languages such as Bade and Bachama where the citation form of words contains a suffix in addition to the root. Examples cited with final hyphens Tepresent underlying roots or stems as indicated in the original sources. 4.3. Sources Language data cited in the word list and elsewiiere in the paper have been dram from a large number of published and unpublished sources. The major published sources utilized are listed in the references. Other published Sources that were consulted, but less intensively, ccan be found in my earlier Chadicibl iography (Newnan 1971). Regarding unpublished material5, MAL 5, 212 P. Newman, [AL 5/1 T have been fortunate in having access to word Lists, field notes, and draft manuscripts that were put at my disposal by the following scholars, to whon I expréss my utmost thanks: Y. Abubakar (Zaar, Barava), D. Barreteau (Nofu-Gudur) , N. Cyffer (Buduna), 2. Frajzyngier (Pero), C.H, Kraft (Natakan, Daba, Pero, Tangale), R.M. Newman (Boghom, Ga"anda), M. Sachnine (Lane); N: Schneeberg (Barava), M. Schuh (Karekaré), R.G. Schuh (Ngizin, Bade, Géra-Kirfi cluster), K. Shimizu (Barawa, Gurintun, and other 2aar group languages), M, Skinner (Bachana) , and E. WoLff (Lanang, Dghwede, Guduf, Gvoko). In addition, I have mado’ extensive use of my own field notes on Tera, Jara, llona,Gudu, Kotoko, Bole, and Warji. 4.4. WORD LIST 1. accept, answer “hava (2) K. Lawé, AL La; (2) Gd, Lowa, T. Abas (4) Mae hed. ashes *batu (2) B. buto, N. babat; (2) G. gac-, Gs. agtso; (3) D. butu, Ta. dobar; (4) Ms. buduu. 3. baobab — *kuba (2) Kk. hued, Ne uk; (2) T. Rubio, Bes kaw-ta, Kanuri kwon is undoubtedly a borrowing from Chadic, as is the case with a number of terms for native flora and fauna. 4. bear, give birth "ua (2) Gs. ye, Mg. wa; (3) Mb. waa, Ke, wna, belch *g¥aht (Q) Kk, jie, Ba. gaaht-; (2) T. jaheds (3) D. getyes (4) Ms. gihtar, 6. bird day (2) P. diyo, Bm. duaié; (2) Gs. diyew, L. diyaka; (3) D. diida, tm, deri. 7. bite/chew, ead (2) B. ngadu, N. kid; (2) T, hada, M, bda; (3) Tm, gad, 8. bite/chew, ada @) AL at, B. wodus (2) Be. ada, Mt. je; (5) Mb. aad, Sm. yida; (4) am, ite. 9. bitter GQ) H daaedé, KE. deet; (2) T. dyot, L. uy-data "gall"; (3) D. djuutu, 10. blood bax (Q) 2. vonan (2); (2) J. vara, Me. gels (3) D, baar, Tn, ba. NAM also reconstructed a *d-N (more correctly, *d-m) root for ‘blood’ based on citations Such as N, dadam and Tg. tom, This root is not, however, found outside the WST branch and thus Cannot be reconstructed for PC, notwithstanding’ the existence of this root elsewhere in Afroasiatic, The question remains open whether this *d-m root is actually a PC form that was preserved only in the WST and lost in the other branches or whether it is a more recent WST innovation. AALS, 221977] n. 12. 13. ue 1s. 16. ww 18. 1. 20. ai 22, 23. Chadie Classigécation and Reconstructions 23 blow 46 G) S. Gi, R. furs (2) Bre fi, Los Gas (3) Tm, pow body #24 GQ) B, zitwo, 2. eis (2) Gi. zu, Mg. 55 (3) DL id, Tm. hig. This word, like the word ‘mouth', commonly occurs with a ké suffix as seen in i. jékid. ABM-A innovation was the replacenent of the PC form by a root *(g-]va. bone *'Jagu () S. dyes, W. b’aasu-na; (2) Br. dyehtu, T. "gahe; (3) Jg. aso, Ne. ese; (4) 2m 50. bow (n.) — *aaga (1) AL Ad, K. maha; (2) Be. Aagey, Hi. Légyi. Assuming that this etymology is valid—there is the possibility that it is an early Jomword—it is remarkable in being the only reconstruction in this list with initial *., break, *bahto (2) B, boku, H. batee, Pa. bahtu; (2) M. Gal, ME. Bahe-. break, — *4aga (Q) Hy fasaas (2) G. gahe, M. pokes; (3) Tm. paj, Ke. pese "hatch", breaks kaha GQ) H, kawaa, Re ngyel, Ne kahtos (2) T. wulta, Ga. ngahto. breast, milk — Mvadé (1) Kis wudé, So ay (2) T, uudi, Be. undin-to; (3) Mk. udlt, So. wat. brother, friend — *mahe- (1) AL mooe, K. mots (2) T mohe-, Mt. makbi. buffalo, bush-cow — *kaban (2) K. hime, Du, kabans(2) T. ngover, GS. guvog5 (3) Nb. kébend, Tm. guns (4) MS. agun“ta (2). buy = *maso (2) H misaayaa "exchange", N, mass; (2) T. masa, Br. mas. calabash — *d-ka (2) Te. dégo, Ki. kodos (2) T. deka (3) D. ddoka. call *ua (1) P. yos (2) T. ‘ga, Be. was (3) Md. una, Tm, wag; (4) Ma. yi. Compare no. 4 'bear, give birth’, for which an identical reconstruction is provided. In Mubi, the two words constitute a mininal tone pair suggesting that this could have been the case for PC, although I doubt it. More probably the words were either hononyms or else one or the other reconstruction is faulty in detail. MAL 5, 25Fy 24, 25. 26. 27. 28. 29. 32. 35. P, Newnan [AL 5/1 carry *kano (2) H. bat, Ne kabana; (2) Be. guts, Br. hats (3) Tm gor. close/cover —*4u (1) Ge. fee, KL. gu; (2) T. mou, Le $03 (3) Mk. app-. come, *(-}89 (2) R. yes, Z. helo "go's (2) Be. shé, ME. 4-, Ko. S00 ‘enter’ (3) D. ase, Mi. as-. cone (imperative) ya (Q) H. yaa-ka, K. wo-ko, Ne yen; (2) Be. hua, Sk. yo, Lo. awa, In all of the citations this is a special suppletive imperative form and not the ordinary verb ‘to cone’. This must also have been the case in PC. cook da () P. to-fo, Bm. té; (2) Gd. da, L. d-, Mg. da; (3) Jg. temo 'cooked". cough * Jahta (2) N. aafita, W. athbi; (2) M. 'wuhta, Hi. 'gyahtas (3) Mb. écca, Tm. aj; (4) MS. The. con hla (1) K. da, N. hlas (2) T. Mea, Mg. htays (3) New sé tox" (7). My guess is that Hausa saa, fem. saaniyaa, is not a reflex of this proto-form, as is usually thought, but rather is a more recent loanword, perhaps from Tuareg (cf. Gouffé 1974). crocodile — *adam (@) B. adam, Bd. ogdom-on; (2) Br. ngotom, Nn. camo; (3) Ke. ms (4) Ms- hurum-ma. In subbranch EM-A of Biu-Mandara, PC *kodom was replaced by *koxam. This was a lexically specific change and not an instance of a general phonological change. Note that the Masa citation supports the reconstruction of the original PC form with intervocalic d3 otherwise the Masa ord would now appear with an £, its normal reflex of proto *a. The Ngizim word ka¥am is interesting in that it is a borrowing from Kanuri, which at an earlier stage must have borrowed the word from some Chadic language. cure kona () K. jané, 2. wukn ‘medicine’; (2) T. kana, Br. kuti, L. wind ‘medicine’. cut *hta (1) P. ta, 2. hea (2) 7. fita, Gd. ea, dic “mata (2) H. matu, N. mata; (2) T. mado, M. mea; (3) D. mate, So. mita; (4) Zm, mat. dislike, not want #49 (1) Re sha; (2) T. fio, L. t=; (3) Md. ese. MAL 5, 241977) Chadée Ctassisication and Reconstructions 25 36. do *ya/*ov (1) H. yé, P. yu, Pa. ayo (verbal noun); (2) M. ‘ya, Cb. ya, T. ca} (3) Je. yd, Ke. fav. Although the examples with initial y are numerically predominant, the reconstruction with %o has the advantage of making it possible to relate all of the citations by natural phonological processes that could have been repeated independently in the different branches. 37. dog *kan- (1) H. kanee, R. kyara; (2) L. kone, Hi. hake; (3) Ne. gare, Th. ga. In Hausa, y not A is the normal reflex of PC *x, Two alternative explanations can be offered for the citation here: (a) kaxee <*karne (cf. the pl. Ralinaé) in which environ- ment the * > y change was blocked; or (b) katee is a loanword and not a direct survival of the PC form. The PC reconstruction as such, however, seems well established on the basis of the citations which are both numerous and which conform to regular sound laws (e.g. * > in Higi, or x > 9 in Tumak). Nevertheless, one camot discount the pos- sibility that we are'dealing here with a widespread, early loanword and that the true PC word for dog was sonething like *ada, as exenplified by Kk, ada, Bd. j-aan, and Mu. di-na. 38, dream *samn0 (2) P. cuna, N. sauna (vb.), Suan (n.); (2) Br. suné, Lo. suuane; (3) Mb. suno, Ne. sune. Compare no. 116 sleep". 39. drink "sa () H. shaa, Pa. sa3 (2) T. za, Ko. se3 (3) D. see, Tm. hee 40. car gam (2) Hi, fame, Lo. heimé, Bu. hano; (3) Mb. sunaana, Ne. seme. N/M treated 'ear' and 'nane' (see no. 90) as a conbined entry. As suggested by Hoffmann (1970), these words were distinct, though phonologically similar, in PC. In the "ST branch, this PC form was replaced very early by a *k-m- root, e.g. H. kun-nee, Kk. humo, Bn. kun-sé. Al. eat Ate (1) He od, Ne ta (3) D. fee, Mb, t5 (4) 2m, ti. ‘This was the basic verb indicating 'to eat staple food’. In BM, this PC form was re- placed by “zuma, e.g. T. zama, Br. 4am, Bu, hun, an innovation not shared by the Masa group. 42. egg tapi (1) P. fd, We suu-nas (2) Pd. heédeya, Mg. ehbe; (4) Ms. asé-na. 43. elephant *g? oan (2) H. géiuaa, Du, géiwon; (2) T. juvan, Br. cian; (3) Mb, gauyay, Ne. june. 44, excrement *isi (2) Kk. she, Bm. yd 5 2) Mn, shewa; (3) So. dad, Ke keusd. AALS, 2526 P. Newman [AL 5/1 45. extinguish —*(m)bota (2) Gm. pit, Kk. mbutu, 2. mbats2; (2) T. mbid> "to cool’, Be. mburo. This is one of the rare PC words where a reconstruction with a prenasalized stop seems called for. 46. eye Aida (2) H. Pdoo, Ze y AT. fill -*n-(y-) (2) N. niga, W. yanau; (2) Me nya, Be. naz (3) Tm. an, 5 (2) Mn. aca, Mt. dis (3) De odo, So. ddé-m; (8) MS. dnnue 48. fire *aku/*ak"a (1) N. aka, W. kaw; (2) G. uma-ta, Mt. alia; (3) D. ako, So. obo; (4) Ms. aku-da, ‘The *w-t- form reconstructed by N/M,is a WST-A innovation and not a PC word. ‘The pho- netically similar wata found in Ga'anda, a BM language, can be analyzed as wa, < *hwe (by phonological weakening) plus a suffix ta used in the citation form of noms—thereby in- Validating the comparison with WST forms such as Hausa wutaa. The example points up the danger of basing reconstructions on isolated surface similarities. 49. fish *kangi (1) Hh, Réigdd, KE. coneps (2) T. guava, Me. LiLig; (4) NS. ulus fae In the EST branch, this proto-form has been replaced by *h-4- in EST-A and *b-s~ in EST-B, e.g. Sm. gose and'Nc, kusa, but D. bovsa and So, busi. The similarity between the EST-A forms and nearly identical itens in the Ron group (e.g. R. gushe) is nothing but mere coincidence. 50. five *bada (2) Kk. baadie, N. vaad, (3) D. beediy, Ke. wiidiw, Sl. €ly, jump *para (2) Ga. planar; (2) Cb. fala, Db. mbix, Ko. fa; (3) Md. bir, Ke. faa (2). 82. fly (n.) *ddue (1) Kk. déyau, N. juorab; (2) Gs. jéjwred, Mg. aduay; (3) D. déwo, Sm, dous (4) MS. 53, forget *mn(t)- (2) H, mantaa, Ke mone; (2) Be. myenta, T. mona, While the stem final ¢ seems like a suffixal clement, it has quite a wide distribution within the family. 54. four — *g%ado (1) B. pode, We gadis (2) M. fund’, Db. fie SS. fry) outa (2) B. suvu, II. sooyaa, 2. uuklias (2) T. zuro, Pd, sata, +; (3) Mb. gada, Ke. wade, 56, fry, thaw (1) Te. kow-, W, 2aw-aus (2) Bus kawes (3) D. ogive, Ke. save (2).1977) 57. 58, 59. 60. o. 62. 63. 64. 6s. 66. 67. Chadie Classégication and Reconstauctions 27 give *bara (1) Kk, baru, Ns baras (2) T. vara, Gs. vols (3) D, bere. go de/*j- () H, jee, H, daawoo"come back", N. ju, Ne dee "come", Kk. ndu, S. ji 'come's (2) ME. dow, Ko. d23 (3) Mb. nja, Jg. jasto; (4) "2m. do. ‘This word is extrenely difficult to reconstruct with precision. One reason is that it is comon in Chadic to find morphological and senantic shifting and overlapping anong various verbs of motion; 'go', ‘cone’, 'retumn', etc. (Compare Hausa, for example, where jee 'go" and zoo ‘come’ Have comp to be interpreted asa morphological pair al- though historically they were not.) Another reason. is that the words 'go' and ‘cone’ often have a number of gramatically determined variants within the sane language. (In Ngizim, 'go/cone' has three suppletive variants depending on tense/aspect_anong other factors (Schuh 1972a:17).) Finally, a problem is posed by the fact that the d/j altemation seen, for example, in the’ Hausa and Ngizim citations, camot be dis- Counted as a surface manifestation of low-level phonological rules, but rather mst be viewed as a possible shared feature of real historical significance. non ty (2) Ng. de-tu "come"; (2) T. dd, Me dt (4) Zn. "ya "leave’. grate’; (3) D. aade ‘follow’, Ke. dee: go out *pata (@) H. féta, Gm. p'et; (2) G. pod, Db. put. goat alu) ku, (1) H. rwaakéd (pl.), KE. aby Ne aakus (2) Me kus Lo. ogus (3) D. autos (4) MS. ahu-na. grass 4g"azon GQ) KK. guzon, He gaafid "type of grass’ (2); (2) T. wuzon, Me paar, Hi. gwezu. grinding-stone — *bona 1) K. bund, 2. yun; (2) T. vona, Mh. aura, Gi. buna, grow old gare (2) Ki, gaato Told", N. garas (2) T. gona, Gs. gat; (3) So. gue. guinea-fowl —*zaban (Q) H. zaaboo, N. zaabanu, 2. Agepm; (2) Gs. touvon, Lo. zawan, A distinctive EST innovation is the replacement of the PC fom by *z-p-t, e.g. D. zoputo, Tm. hibse; cf. no. 140 ‘wash oneself". hair *gasé (2) H. goashéd, W. guzhé; (2) T. yos, Gl. gufas (4) am. ngisa head “ka (2) K. kod, Gn. kaa; (2) Ko. go, Gi. ké; (3) D. baa, Ke. ed. In Subbranch BM-A, PC Ska was replaced by a *g-n root, e.g. L. yan, M. kok. In subbranch WST-B, the PC fori was replaced by *g-m, e.g. N. yam, "Z. gaan. (I'am assuming that this was a proto-NST-B innovation even though *g-m is no longer found in the Bade group.) AALS, 2728 68. 69. 70. an m2. 73. m 75. 7%. 7. 78. 1. 80. P. Newman (AAL S/L hit *htaéa (2) P. ovo; (2) T. AtaBa, Mg. heab-; (3) Tm. Lab, Ke. Laa. hole, pit *bak- () Ki. boyo, Ba. vak~an, S. gun (7)3 (2) Gs. veg-edy Mn. augas (3) D. bee. honey, bee and (2) H. ‘tanya "beehive"; (2) M. ami, Gs. aman; (3) Jg. amo, Ta. do (4) Ms. ayunema, hunger maya (2) Ny maya, He yurwaa (2); (2) Be, miya-ta, Pd. maya; (3) D. meya, Tm, may; (4) w8."nayea! hut, house ban () B. bono, 2. vin; (2) Be. vaney, Ko. fan; (3) Mb. bené "build", MK. biino, if *Imb- (1) K. bo; (2) T. ma or mba, Hi. ma; (3) Ne. ma. in-law Ag-a- () A, sin, Ny saurrak; (2) Be. shewey, Ng. suta, Kill | *d- (2) 2. £94 (25 (2) Mn Jas (3) Jz. d-, So. das (4) Ms. ci. The *d-k- root for 'beat" found in the NST, e.g. Kk. duku, N. taba, could be etymological ly related. Imee — *gagu CQ) N, kugu, He gwiiwna(?); (3) Jg. géfo, Tm, gubs (4) Ms. gégif-fa. N/M proposed two forms for "knee': *F-1-m and *e-t-m, ‘The former was probably an Angas/ Bole innovation that spread to other WST languages in the sane geographical area, The latter could be the PC word for 'to kneel", c.g. Kf. kuru, T. kwtuma, although more citations are needed outside BM to justify’such’a reconstruction. Imife — gub- (2) KL. suk, Ki, shonges (2) T. heugu, ME. heww-ed, Gi. heoho; (3) Tm hugan. know sana GQ) Hy sani, KL. syens (2) T. zona, Ko. san; (3) Mk. suwiiyo, Ne. sen, ‘The replacement of PC *sana by *b-n- in the Angas and Bole groups (see Hoffnann 1970) is one of the shared innovations that characterizes this major group. Yaugh —*ganso (1) N, gams2, 2. gyomhta; (2) Br. kumsiui, Gd. ryvas-an "Laughter"; (3) Mb. gamas-, Tm. ga. left (side) Ag-d (1) Ge. gyada, We gadis (2) T. godau, Be. tyeade, AAL 5, 281977] Chadie Classification and Reconstructions 29 ‘There is also a widespread *g-L- root for ‘left', e.g. S. kul, Mf. guta, D. geete. However, since £ in present-day languages is so Often’a Teflex of sone ther conso- nant(s); it is difficult to know how to interpret these various *g-¢- forms. 81. leg, foot asa (Q) R, Say, Te. yos (2) Lo. asa, Mg. azés (3) D. ase; (4) MS. ase-nu, In BMA, the PC form was replaced by *6-n-, e.g. G. Saxa, Br. Sat. 82, Lie down xan (@) K, anu, Bd, xana "pass the day’; (2) G. una, L. ind, Db. wan; (3) Mk. kon- 'sit'. 83. meat, animal *h€av~ CQ) KE. tua, W. hewu-na; (2) L, hewwi, Ko. shu; (4) Ms. heéw-na, 84, met — g-ma @) H. gam, We gamau; (2) G. kam, Gv. gama, 85. monkey — *badé (2) Kk. bado, N. vajés (2) M. peu, Le vafds (4) am, vin. 86. moon tana (D)_jBee ones We coraenay (2) Ts nana, Mn. f0t05 (5) Mo. LAL, Ke, bi-tiny (A) Me. 87. mouse, rat *kusam (2) Kk. casam, 2. kusans (2) Mt. fiom, Ko. kusumu; (3) Jg. kum, D, humo, Ke. konom, ‘The medial 4 of the proto-form seems to have been lost independently in BM-A and in the EST branch, 88, mouth ba (1) H. baa-kéii, B. bo, Z. viz (2) Db. ma, L. ewe, Mg. ma; (3) D. bi, So. bo-. ‘The replacenent of *ba by *ma is a distinctive BM innovation evidenced in both sub- branches. The Bade group forms with an initial nasal, e.g. Ns miua, have to be treated as special cases which do nov affect the general statenen 89, ma *tab- (a) B, ebb, N. abo; (2) Be, snbwe, Ko, ndabes (5) Tm, dubo, 90, name *yam (1) Ké, sum, 2, sums (2) Mo heam, Hi. Ali, Bu. homu; (3) Md. sami, Sm. sumo; (4) Ms? sanéma: Compare no. 40 tear". 91. neck — *wora Q) H, mya, N wura; (2) Be. wuta, Gi. wwe; (3) Jg. were, 92. night *badi (1) B. bodé, Bd. agveed=an; (2) Db. vudi, Lo. vade. AAL 5, 2930 93, 4. 95. 96. 97. 98. 99. 100. iol. 102. 103. dod, 10s. P. Newman [AAL 5/1 nose *atan (2) H, hancéé, R. atin, Pa, atan; (2) M. mein, Sk. shin; (3) D. tein), Jg. etento; G) MSL acéné nu! ‘The etymology of the Hausa word is *atin > *antin (by insertion of an epenthetic nasal) > *anéti (by the general loss of final nasals) > *hantéi (by phonenicization of a feature Of word-onset) > hanoce (by automatic palatalization)« not un Q) K wo...u, Pa. uaz (2) Go wes, .ua, Hi. wes (3) MK. ye, Tm. ay. ‘This was presumably the general PC negative marker used at the end of negative sentences. ‘The connection, if any, between this proto-form and the extremely comon ba negative, found for example in Hausa and Ngizim, is yet to be determined. oil nar () He mad, 2, méins (2) Ts mor, MEL mae. person natu (Q) Ha, mutum, Ba. md~an; (2) Be. mune, M. mda (3) Jg. méto, So. mati ‘man’. place ba () K, bod, KE. pes (2) T. ma, Lo, mba, pour po () Ge. fut, Ne pas (2) Me pa, GS. pes (3) Tn, pe. put 59 GQ) Hi saa, K. awe; (2) G. 495 (3) D. tse, New See python wads () H. (Sokoto) muudiwaa, B. midt, S. mins (2) Gd. myadan, Ko. mudd. rain *ban (2) S. gua, 2. voan; (2) Hi. va, Mg. fanz (3) Jee bons ran gam (@) Ky gam, KL, nzam; (2) T. gam, Sk. gannak; (3) Ne. gamge (pl. of 'sheep'). refuse “hur UD) K hued, Ne kuna, Bm. kad; (2) Me ngat, Ge kar. Compare no, 24 'carry'. return tna CQ) Ke mae, Dus ma-5 (2) Gs. mes (4) 2m. mbu "come". ripen ap (1) Kk. ra, Pe mid, We news (2) Be. na, Ko. nas (3) D. nee, Tm, ni. In Newnan (1975) I argued that the final vowel of Chadic verbs was lexically distinctive and not just gramatical as is commonly thought. The contrast between this verb and AAL 5, 301977] Chadéc Classification and Reconstauctions 31 the verb *na ‘see’ (no. 111) strongly supports this position, In some present-day Languages, ‘ripen’ appears with an a, sometimes becoming homonynous with 'see'; but the comparative evidence taken as a whole shows thet the two verbs mist have been distinct in the proto-language and that the difference lay in the quality of the vowel. 106, roast, bum *b-ho (1) H. babbakee, B. bokku, Ne baka; (2) Gs. gok, Mg. gung-3 (3) D. bike. 107, root *par- (2) Hh say-wia, P. coorés (2) G. kan, Mn. htatwes (3) Ne. sar-, Ke. ka-sar, 108, rope *zawwi (2) Ne cayé, We zhau-na; (2) Hi. zowe, Gs. zhewred} (3) Tn. ha, 109, say *p-do/tp-na (2) He gadti, Kk. dagu, By porus (2) Br. pot, Gi. pee. 110, scorpion |x) atda () Ba. wurj-an, Diri aridwins (2) Le anda; (4) Ms. hurduu-da. Al, see tna GQ) KK naa, W, nahaus (2) T. na, Be. ma. 412, send *gono (1) N. wana, Ws canes (2) Bes £yena, Lo. han; (4) Ms. sun. Compare no. 150 'work", where the lexical difference seems to lie only in the quality of the final vowel. 13. shave *saka (2) As sak, Ne soba, Heeaseéd (2)5 (2) Mn 42x, Be, Sawa, T. xaz9 (2). 44, sheep *tambi GQ) H, aumaabié(pl.), WN. tanakuays (2) J. ndomox, Ng. adamak; (3) Mb. tunak, Ne. dange. While the final ké can be analyzed as a suffix, its widespread occurrence in Languages throughout the fanily indicates that it may already have been fused to the root for ‘sheep’ in the proto-language. as. skin *zam (2) Gn, shams (2) Ts kozam, Br. késams (5) De zaamay Tm. hum. 116, sleep *s-nl-] (1) Pe cans (2) Gis sind, Lo, sans (3) Tm. nun (2)5 (4) 2m, shene, N/M treated 'sleep’ and ‘drean' as a conivined entry. 1 have given them separate recon- structions here even though it is very possible that these separate forns could be derivational variants of the sane root. 117, smell guna (2) He sunsunaa, We saean-3 (2) Br. sury (3) Tm, huston.32 us. 1. 120. qa. 122. 123. 12a, 125. 126, 127. 128. 129, 150, P.. Newman (AAL 5/1 ‘smoke *"Jan (1) Ba, oon, W. k'yanu-na; (2) T. ‘gan, M. "yar, spear *gas () P. kac, N. ngas; (2) J. gas, Mt. guosh "knife", The word-final obstruent suggests that we may be dealing with an early loanword rather than with @ true PC form, spew, spray *pasa GQ) Ki. géshu, Pa. pasu; (2) T. pasha, Be. giisa; (3) D. Geese (2). spit *toga/*tuga (2) H. toofaa, Ne tapka; (2) M. ntaga, Lo. tugus (3) Mb. tugga, Ke. tugés (4) 2m, tugo. stand up hee (2) Ki. faa, N. h€as (2) L, he-, Lo. hla; (3) D. utye, Ke. duu "go up's stay, stop dana (1) Kk. "yaaru, Bm, dé, It, tstayaa; (2) Mn. dat ‘climb, Ko. (Afade) 2s'arga; (3) Mb. dar. steal *xato (Q) Kk. cw, Ne teary (2) Gd. xara, Br. xat, N/M reconstruct a root *m-A- on the basis of citations such as Gera mooxa and Tera mura. My guess is that these *m-t- foms represent not another PC word but rather a reduced form of a *n-x-X- stem indicating "thief", composed of an agential prefix *m- plus the PC verb "steal", suck —*sa6o (2) Be. sao ‘drink", ME. asab-, Ko. s'aga (23 (3) Mk, aéb- "drink", Ke. sobes (4) 2m. sobo. sun, day * hati (A) KK, gti, 2. §éta; (2) Ub, pie, Ma. futis (3) D. patos (4) 2m, gute. A diagnostic feature of the Bade/Warji major group is the shared loss of the ¢ in this word, e.g. Ne asa, We 6 tail Seaton (1) Ki, biti, Bl. ubton~anz (2) Le xtini, Gi. hutra, ten gtane (2) Hh, gooma, Ne quuma; (2) J. guom, Ko. kan. termite *gahta (1) Re mgal, We guahtata-na; (2) T. gahe ‘termite hill’, M. kakla; (3) D. gudyina (7). thirst kozam (2) B. kazam, N. gajé; (2) T. xujun, Ko. stam "hunger". AAL 5,1977] 131. 132, 133, 134. 135, 136, 137. 138. 139, 140. an, Chadéc Classégécation and Reconstructions 33 this, that, the — #d GQ) K, din, 8, dis (2) 6. di, da, Mg. das (3) D. ede *here’, Tm. do. ‘The glosses are not meant to be taken literally. They merely serve as a memonic for a PC deictic particle whose exact function and meaning are still to be determined, ‘The entry acts as a reminder of the possibilities for reconstruction available in morphological domains outside the lexicon in the narrow sense. three *k(*)an (2) Ki, wm, N. fwans (2) Sk. maton, G. maxkan, L. xkana, Lo, gakora; (4) Im, hindi (2). In almost all BM languages that have this root, one finds either a ma prefix, a velar prefix, or both, In the EST branch, the PC form has been repalced by 2 root’#s-B-, e.g. D. subba, Ke, scope, tie *g-no (2) AL gyén, Bm, haan; (2) T. gona, Db, kon; (3) Mo, ewe (2)5 (4) dm, jan. tongue *ahtasé () A. Lids, KL. atush, He harshees (2) Mg. alas; (3) D. Leese. ‘the body-part prefix in the Hausa word is not ha, as suggested by Leslau (1962), but @, as seen in the Kl, and Mg. citations, ‘The source of the initial is explained in section 3,2(e) and in Newman (1976). tooth — ®gant-) (Q) Ne yaanau, 2. shins (2) Pa. heira, Ko. shany (3) Jge sano, Tm. hin. tree Ait (1) H, Picee, KL. yéemau; (2) L, udzu, Ko. sé; (3) De eto, Tm, aru-nan (pl.). two *sunl-] (1) Ba, sanan, W. ma-sun "seven"; (2) J. sane, Gi, suls (3) Md. sdén, Sm, sit. A shared innovation of the WST-A subbranch is the replacement of this PC form by a *b-L- root (see Hoffmann 1970), e.g. H. béyu, Kk. betu, R. apit. uncle, nephew *kun- (2) K. bon-ak, S. eons (2) Te nonyin-ké, Mg. ona; (3) Tm, gunman "cousin' (2). untie — *para (0) AL win, Ne gotkas (2) T. pata, Ko. fats (3) D. épite, Nev pure. wash oneself, bathe — *bana (2) B. binaa, H, wankaas (2) T. vana, Pd, para, Im the EST branch, the PC form has been replaced by *p-£-, e.g. Mk. opét Ne. petes cf. no, 65 'guinea-fowl", wash something *e-ba (1) kK. jobe, 2. top; (2) T. ciba, Hi. yabes (3) Mb. cuubé. AAL 5, 331a, 3. asa. 1s. 140. a7, as. a. 150. Pe Newnan [AAL 5/1 water tan (1) Tg. am, Ne ams (2) Ske dyam, Lo. am; (3) D. amay, Tn. nam. what? *nd/4mg (A) R, mi, Ny army (2) M, mé, Be. muna; (3) D, maa, Ne, me, mene; (4) Zm. mi. ‘The vowel contrast seen in Hausa mee (in Sokoto, mi) 'what" and wan 'who' is recon~ structable for PC, It is difficult to know what significance to ascribe to the scattered examples where this word has a second syllable containing an n, where? *ina (D) Werénaa, AL anes (2) Te fo-na, Lo. éna-ke; (3) Kes anes ‘This etymology is much less certain than the ones for 'what' and 'who'. ‘There may be a confusion here with the word for, 'which', often found with a root *n-, since a number of Chadic languages express 'where' not by’a separate word but by the phrase ‘which place?", white *p-() (DAL pye, He fardé, W. pyau-na, 2. fyalés (2) M. partu, Lo. pou; (3) Mb. genet, ‘This is one of those cases where the cognation of the various citations seens highly probable but the correct form of the PC reconstruction is difficult to ascertain, who? *ua GDR. ua, Ne tary; (2) M, ua, Bes wond; (3) D. waa, Ne. wei. Compare no. 143 'what?", wing phe (1) H, fékangikié, We pako, Bm, pakap; (2) T. Ropar, G. papax ‘feather’. with/andy do GQ) He da, KL. tu, Bd, do; (2) T. nda, Gi. dés (3) MK. £2, Kes da, with/andz — *ea/*ga (Q) B. ga, Ke ba; 2) Be. ka, Ko. 905 (3) Jae kas work — gana () N. wig, W. cuna; (2) T. hema, Gs. kana; (3) So. ussan; (4) 2m sin. Compare no, 112 'send', MAL 5, 341977) Chadée Ceassisication and Reconstructions 35 5. CHADIC LANGUAGE INDEX ‘The following index includes approximately 250 nanes culled from the literature on Chadic Linguistics. ‘The symbol = is used to indicate altemative (sonetines more precise, some- ‘times “incorrect") names for the sane language, The phrase dial. of is used to indicate dialect variants or village nanes. The nanes taken as the prinary designations for the approximately 125 different Chadic Languages are identified according to the classifica- tion presented in section 1. Each language is identified as to its branch (indicated by a Roman numeral), its subbranch (indicated by a capital letter), and its group (indicated by an Arabic numeral). ‘The notation IT.A.4, for example, would indicate a language belong- ing to the Biu-Nandara branch, subbranch B“A, and group’4, i.e. a language belonging to the Mandara group. Lower level subgroups and’ clusters are’not indicated. Finally, the index also serves as a key to the abbreviation of those languages from which examples have been cited in the word list and elsewhere in the paper. Afa = Pa'a Breme dial. of Hurza Afade dial. of Kotoko Buduma (uy) 11.B.1 Alstaghwa dial. of Lanang Buli dial. of Polchi Angas ® LAS Bure (Br) TA. Auyokawa (extinct) 1.3.1 Burrun = Boghom Babur = Pabir Butura dial. of Daffo Bachana (ic) TL.A.8 — Bwol dial. of Kofyar Bade (Bd) 1.8.1 Ghakfem-Mushere dial. of Sura Balda dial. of Miktele Chala = Ron Bana TLA.3 — Cheke = Gule Banana = Masa Chena dial. of Guduf Barain 11.8.3. Chibak (Cb) 11.4.2 Barava 1.B.3 Ghikide dial. of Guduf Barke = Mburku chip TAS Baron = Ron Chire dial. of Gabri Bata ILA.8 — Chonge dial. of Pero Baza dial. of Higi Daba (0b) 11.7 Bedanga = Sokoro Daffo dial. | Bedde = Bade Dangaléat Bele I.A.2 Dangla () 1B. Bidiyo TI.B.1 Dari dial. of 2ime Birgit TIT.B.1 Daza dial. of Bole? Boga dial. of Ca'anda Deno TAZ Boghom (in) 1.3.3 Dera = Kanakuru Bokkos dial. of Ron Dghwede TLAA Bolanci = Bole Dimmk dial. of Kofyar Bole () 1A2 pint 1.8.2 Boto dial. of Zaar AAL 5, 3536 Dirya = Diri Doga dial. of Migana Doka dial. of Chip Dorm dial. ef Gabri Duwai (bu) 1.8.1 Pali of Jilv dial, of Gude Fali of Kiria dial. of Higi Fali Fali of Fali of Mucella dial. of Gude Fyer LAA Ga'anda (@) TLAL Gabin dial. of Ga'anda Gablai = Kabalai Cabri, TIL.A.2 Gadang TILA Gala dial, of Warji Galanbu LA2 Gamargu dial. of Mandara Gasi dial. of Kanakuru Gawar ILAT Goji T.B.3 Gemjek dial. of Zulgo Gera (Ge) LA.2 Gerka LAS Goruma LA Gidar (Gi) 11.8.3 Gisiga (Gs) TL.A.S Glavda (G1) TA Goemai (Ga) LAS Guie (Ga) ILA8 Gud TLA.8 Cuduf ILA Gudur dial. of Mofu Gulei = Ndan Gulfei dial. of Kotoko Guruntun 1.3 MAL 5, P. Newnan Jonkor of Gera [AAL 5/1 (Gv) TL.A.4 LAA (LAL (Hi) 11.a.3 TL.A.7 IAL TLA.S G) TAL (Jg) T11.B.1 1.B.2 1.B.3 TI.A.2 UL.B.L WAZ (i) LA.2, LAS 1.B.2 (Ke) TIT.A.3 ILA.2 (Ki) LA? LAS1977) Kofyar Kola dial. of Daba Kotoko Rubi Kulere Kulong = Marba Kupto Kuri dial. of Buduma Kuseri dial. of Kotoko Kushi = Chonge Kwalla dial. of Kofyar Kwan of Bole Kwang Kyibaku = chibak ‘Lamang Lane dial. of Zine Lele Logone Mada Mafa = Matakan Kotoko Bata Mangar dial. of Ron? Mangas Marba Margi Margi West = Putai Marva dil Masa Masana = Nasa Masmaje Natakam Mawer dial. of Tumak Mbana = Bana Moku dial. of turza of Gisiga Chadéc Ceassigication and Reconstructions (ke) LAS (Ko) I1.B.1 LA2 ) Lad LA2 TIL.A,3 @) ILA TLA.2 (Lo) 11.8.1 TLAS LA2 TIL.B.1 Qh) TEA 1.B.3 wv @) 1LA.2 (Ws) 1V TIT.B.1 (it) ITA. MAL Mburku Nemyang = Mirrian Nesne Midah dial. of Logone? Migana Mijivin dial. of Gisiga Mileu Miriam Miya Mod, Modgel = tang Mofu_ Nogun Nokulu Nonogey dial. of Marba? Nontol Nora dial, of tharza Meer dial. of Kotoko sbi Muktele Muleng dial. of Bata Mulwi = Musgu Mmndat dial. of Sha Musey Musgoi dial. of Daba Musgu Misgun = Musgu Mutunwa dial. of Gisiga Kofyar oi) oe) oy) O) (uy (We) Oc) a 37 1.8.2 Vv 11.B.1 TILA 1.B.2 LAL TAS MI1.B.1 11.8.2 LAS TII.B.1 ILAS 11.8.2 TIT.A.2 MLA TAZ Bal38 P. Newnan TAAL 5/1 Njai = Neangi Tobanga dial. of Gabri Njeny = Neangi Toran 11.B.1 Nkafa dial. of ligt ‘Truade = Dghwede Nyimatli = Tera Tsagu 1.8.2 Neangi ILA8 Tunak (Tm) TITALL Ouldéné = Udlam Turu dial. of Lanang Ourza = Hurza Uilan dial. of Nada Pata (Pa) 1.3.2 Vame dial. of Hurza Pabir ILA2 Vengo dial. of Lanang Paduiko (Pa) 1A4 — Vizik dial. of Lamang Pero ©) LA2 Volum dial. of Musgu Peve dial. of Zine Wandala = Mandara Pia dial. of Pero Varji wm 1B2 Pidlimdi dial. of Tera Kaya dial. of Lanang Podok(w)o = Paduko Wula dial. of Higi Polchi 1.8.3 Wurkn = Pia Putai TLA.2 — Yaghwatadaxa = Cudut Pyapm LA3 Yanaltu = Tera on @) Ad Yedina = Buduna Saba 111.B,3 Yivon = Gerka Sarva MAL Zar (2) 1.3.3 Sayanci = Zaar Taladva dial. of Lamang Sha TAS Zany = Neangi Shagawa TA 2eem 1.3.3 Ship = chip Zelgwa = 2ulgo Shirava (extinct) 1B. Zime (am) wv Shoe dial. of Kotoko Zulgo dial. of Nofu Siri 1.B.2 dum dial. of Bata Sokoro (So) 111.8.3 Somrai (Sa) TILA Sukur (Sk) TL.A.6 Sura (©) LAS Tal LAS Tanbas dial. of Fyer Tangale (Tg) TA2 Teel = Montol Tera @ maa AAL 5, 38197] Chadée Ceassigication and Reconstructions 39 REFERENCES Abbreviations: AM Ageécana Marburgensia Auli Agtéka und libersee JAL Journal of African Languages SAL Studies in Agrécan Linguistics ‘Anon. 1962. A Short EngLish-Bura Voeabukary for Ready Reference. [Privately printed by Church Brethren Mission, Garkida, Nigeria.] Bargery, G.P. 1934. A Hausa-English Dictionary. London: Oxford University Press. Barreteau, D. In press. "Le mofu-gudur, lamgue tchadique du Nord-Cameroun."" All. Burquest, DA. 1971. A Paetiminary Study of Angas Phonology. Studies in Nigerian Languages, no. 1.” Zaria and Kano. Caprile, J.-P. 1971. "Les langues du Tchad." In Attas pratique du Tehad, pp. 36-37. Paris. + 1975, Lexique tunak-(angais, Berlin: Dietrich Reimer. Caprile, J.-P., and H. Jungraithnayr, 1973. "Inventaire provisoire des langues 'tchadiques' pariées’ sur le territoire de ia République du Tchad." All 6(2):31-48. Camochan, J. 1970. "Categories of the verbal piece in Bachama." African Language Studies 181-112. =. 1975. "Bachana and Chadic." In Hanito-Senitica, J. and T. Bynon (eds.), Pp. 459-68, The Hague: Mouton. Cohen, M. 1947. Essac companatif sux Le vocabutaine ot La phondtique du chanito-sénitique. Paris: Champion. Diakonoff, I.M. 1965. Seméto-Hamitéc Languages. Moscow: Nauka. Ebert, K-H. 1976. Sprache und Tradition der Kera (Tachad), Teil 2: Lexikon/Lexique. Berlin: Dietrich Reimer. Fédry, J. 1971. Dictionnaire dangatéat (Tehad). Lyon: Afrique et Langage. Frajzyngier, Z. 1976. "Rule inversion in Chadic: an explanation." SAL 7:195-210. Gouffé, C. 1969/70. "Compléments et précisigns concernant le haoussa dans le cadre de L'Essad comparatig de M, Marcel Cohen." Comptes rendus du groupe Linguistique d'études chaméto-sémétiques 14:27-43. . 1974. "Contacts de vocabulaire entre le haoussa et le touareg." In Actes “dar Pitomien conan’s international de Linguistique sénitique et chamito-sémitique, A, Caquot and D. Cohen (eds.), pp. 357-80. The Hague: Mouton. Greenberg, J.H. 1947, “Arabic loan-words in Hausa." ond 3:85-97. 1958. "The labial consonants of proto-Afro-Asiatic." Word 14:295-302. 1963. The Languages of Agtica. Bloomington: Indiana University. . 1965. "the evidence for */™b/ as a proto-Afroasiatic phoneme." In Synbotae TaGTisticae’ in honorem Georgii Kurykowicz, A. Hernz, M. Karas, et al. (eds), pp. 88-92. Krakow. Hodge, C 1968. "Some Afroasiatic etymologies.” Anthropological Linguistics 10(3) :19-29. AAL 5, 39)40 P, Newman [AL 5/1 Hoffmann, C. 1963. A Grammar of the Margi Language. London: Oxford University Press. 1970. “Ancient Benue-Congo loans in Chadic." as 3(2):3-23, + 197la, "Provisional check List of Chadic languages." Chadie Newstetter [Narburg}, special issue 1971b. "On the classification of Nancere." Jounna of west African Languages teat =o 12 + 197lc. "On the position of Paduko." In Actes du huctidme eongtés de fa —HUCELE Linguistique de L'Agrique cecidentate, pp. 221-33. Abidjan. n.d. "Towards @ comparative phonology of the languages of the Angas-Goenai group." Unpublished ms. Hoskison, J.T. 1975. "Notes on the phonology of Gude.” M.A. thesis, Ohio State University. TILiR-Svity’, V.M, 1960. "Iz istorii Eadskogo konsonantizma: labial “nye smyénye." In Jazyki Ariki, B.A. Uspenskii (ed), pp. 9-34. Moscow: Nauka, Jungraithmayr, 1. 1961/62. "Beobachtungen gur tschadohanitische Sprache der Jegu (und Jonkor) von Abd Telfan (République du Tehad)."” Aull 45:95-125, 1963/64. “Die Sprache der Sura (Maghavul) in Nordnigerien." Aull 47:8-89, ——DEn. a = 1964/65. "Naterialen zur Kenntnis des Chip, Montol, Gerka und Burrum —(Sikiplateau, Nordnigericn)." Auli 48:161-82. = - 1906/67. “Specimens of Pa'a (*Afa') and Warja languages with notes on the ~GFIBES of Ningi’ chiefdom (Bauchi Province, Northern Nigeria)." Aull $0:194-205. = 1970, Die Ron-Spachen: tschadohamitische Studion én Nordnigerien. Glikkstadt: J.J. Augustin. _ - 1971. "The Tangale vowel harmony system reconsidered." In Specéat Chadéc —Tasag, P. Newman (ed.), pp. 28-33. JAL 10 (1). Klingenheben, A. 1927/28. "Die Silbenauslautgesetze des Hausa." Zoitschrige git Eonge- borencn-Sprachen 18:272-97. Leslau, W. 1962. "A prefix ft in Egyptian, modern South Arabian and Hausa." Agréea 32: 65-68. Lukas, J. 1936. Dée Logone-Sprache im zentraten Sudan, Abhandlungen fiir die Kunde des Norgentandes, 21°(6). Leipzig. 1937. Zentratsudanische Studion. Hanburg: Friederichsen, de Gruyter and Co. + 1904/05, "Das HHitkalanci, eine Sprache un Gxoze (Nordostnigerien)." Audi ase 14. . _. 1960. "Tschadohamitische Sprachproben aus Nordnigerien (Karekare- und ———PoTaRci-Texte)." In Neue Aftikanéstische Studéon, J. Lukas (ed.), pp. 173-207. Hanburg: Deutsches Institut fiir Afrika-Forschung. __. 1970. Studéen zur Sprache dex Gisiga (Nordeamemn). Gliickstadt: J.J. ‘tin. . . 1970-72. "Die Personalia und das prinire Verb im Bolanci (Nordnigerien)." Ha -$4: 237-86 5 55:114-39, 1974/75. "Bin Text in der Sprache der Djonkor des Gera-Nassivs (République Hi Tehaa). Aull’ 58:12-26. Lukas, R. 1967/68. "Das Nomen im Bade (Nordnigerien)."" Auli 51:91-116, 198-224. AAL 5, 401977] Chadée Classification and Reconstructions a Neyer-Bahiburg, H. 1972, Studéen zur Morphotogée und Syntax des Musgu. Hanburg: Helmut Buske. 1972/73, “Texte im Masgu von Girvidik (Nordkamerun)." Aull $6:61-71. Mirt, Il, 1969. "Einige Beerkungen zum Vokalsystem des Mandara." Zeétachnégt der Deutschen Morgentindéschen GeseLeschast Supplement 1, no. 3, pp. 1096-1103. _ + 1970/71. "Zur Morphologie des Verbalcomplexes im Mandara."” Aull $4:1-76. Mohrlang, R. 1972, Higd Phono€ogy. Studies in Nigorian Languages, no. 2, Zaria and Kano. Mouchet, J. 1950. "Vocabulaires conparatifs de quinze parlers du Nord-Caneroun."” Bubtetin de ta Société d'Etudes Camerounaises 29/30:5-74. 1983. "Vocabulaires comparatifs de sept parlers du Nord-Caneroun." BaeRetin de ka Société d'Etudes Camerounaises 41/42:136-206. Netting, R. 1967. "A word list of Kofyar." Research Notes {Ibadan} 1(2):1-36. Newnan, P. 1970. "Historical sound laws in Hausa and in Dera (Kanakuru)." Journal. of west African Languages 7:39-S1. - 1971. "A Ghadic language bibliography (excluding Hausa)." In Speciat Tadic Tasue, P. Newman (ed.), pp. 101-09. “JAL 10 (1). . 1973. "the development of /f/ as a Hausa phoneme." Paper presented at the Foi anual Conference on African Linguistics, New York. = 1974. The Kanakuru Language. West African Language Monograph Series, no. 9. TREE! Institute of Nodem English Language Studies. ___+ 1975. "Proto-Chadic verb classes." Folia Orientatia 16:65-84. ___; 1976. "the origin of Hausa /h/."" SAL, Supplement 6, pp. 165-75. In press (a). "Chado-tlanitic ‘adieu’: new thoughts on Chadic language —“elasi fication.” In’ Proceedings of the Second International Conghess on Hamito- Semitic Linguistics. Florence. In press (b). “Lateral fricatives (‘hlaterals') in Chadic." In Papers in THEE. Linguistics, P. Newman and RLM. Newnan (eds.). Leiden: Afrika-Stuliecentrun, Newnan, P., and R. Ma. 1966. "Comparative Chadic: phonology and lexicon." JAL 5:218-51. Parsons, F.N. 1970, "Is Hausa really a Chadic language? Sone problems of comparative Phonology." “African Language Studies 11:272-88, Pilszezikowa, N. 1958. "Contribution & 1'étude des rapports entre le haoussa ct les autres Langues du groupe nigéro-tchadien.” Rocznik OnentaListyezny 22:75-99. + 1960. "Le haoussa et le chamito-sémitique & 1a lumiére de 1'Essaé comanatis Ee Maire Cohen." Rocznék Onéentatistyczny 24:97-130. Rapp, E.L., and B. Benzing. 1968. Dictionary of the Glavéd Language. T: Glavdé-Engtish. Frankfurt: Bible Society. Rapp, E.L., and C. Mihle. 1969. Dictionary of the Gtavdd Language. IT: English-Glavdé. Frankfurt: Bible Society. Schneeberg, N. 1974. "Sayanci. phonology." Schuh, R.G. 1972a. "Aspects of Ngizim syntax." Ph.D, dissertation, University of California at Los Angeles. 1972. "Notes to Bade dialect map." AM 5(2) :50-56. 1976. "The history of Hausa nasals." SAL, Supplement 6, pp. 221-32. " Ph.D. dissertation, Indiana University. AALS, 4142 P. Newman — Chadie Classigication and Reconstnuctions [AL 5/1 1977] n.d. (a). "The Bole-Tangale group of languages." Unpublished ms. n.d. (b). "Towards a typology of Chadic vowel and tone systems." Unpublished Shimizu, K. 1975. Boghom and 2aax: Vocabutary and Notes. Kano. Skinner, A.N. 1974, "A note on the North Bauchi language group (with map)." AM 7(1):47-50. . 1977. "N, Bauchi Chadic Languages: common roots." Agnoasiatic Lin uistics 4/1:1-49. Wente-Lukas, R, 1973/74. "Zur sprachlichen Stellung des Bana (Mandara-Gebirge, Nordwest- kanerun)." Aull §7:1-15. Westermann, D., and M.A. Bryan. 1952. The Languages of West Africa. Handbook of African Languages, part 2. London: Intemational African Institute. Wolff, E. 1971, "Die sprachliche Situation im Guoza-Distrikt (Nordostnigeria)." In Special Chadic Issue, P. Newnan (ed.), pp. 61-74. JAL 10 (1). 1974/75. "Neue Linguistische Forschungen in Nordostnigeria." Aull 58:7-27. WoLff, H._ 1959. "Subsystem typologies and area Linguistics." Anthnopotogient Linguistics 1(7):1-88. AAL 5, 42,new from Undena @ The Economic Role of the Crown. by N. Yotfee 168. $18.00 (cloth), $13.50 (paper). The Economie Role of the Crown in the Old Babylonian Period stulies economic and administra tive documents under the assumption that economic relationships and their changing nature through time merit study on their own terms. Bibliotheca Mesopotamica, Vol. 5. Pp. vi Seals and Sealing edited by M. Gibson and 8. D. Higgs Bibliotheca Mesupotamica, Vol. 6. Pp. ii-160, 1 mf. $23.00 (cloth), $18.50 (paper). Seals and Sealing in the Ancient Near East is concemed with thr functioning of seals in their con: texts, the significance of sealing on cuneiform tablets of particular periods, but there are also treatments of seal use on a specific category of texts through time Mountains and Lowlands ed. by L. p. Levine and T. Cuyler Young, Jr. Bibliotheca Mesopotamica, Vol. 7. Pp. x-405, 26 pls. $21.00 (cloth), $16.50 (paper). Mountains and Lowlands: Essav's in the Archaeology of Cireater Mesopotamia is composed of essays written by those who worked with Robert H. Dyson, Jr. at Hasanlu, Atchacological, at historic and epigraphic matters are represented. The Royal Inscriptions byw. de Assur, Vol. 1, Issue 7. 47 pp. $3.95 (paper) The Royal Inscriptions of Aiur Nasir Api I 883-859 B.C.): 4 Study of the Chronology of the CCalah Inscriptions Together with an Edition of Two of These Text Using close textual analysis, itis contended that two different palaces ae referred to in the inscriptions and that the composition of most of the Clah texts was in fat well spaced out over Asm.’ reign. lish-Egyptian Index by D. Shennum ‘EYP ARTANES, Vol. 1. Pp. iv-178. $6.00 (paper). English-Egyptian Index of Faulkner’s Concise Dietionary of Middle Egyptian is intended as a useful tool for locating textual material for comparative research in Egyptian semantics, history, archaeolo- ay and religion. Space is provided for adding one’s own material ippi Immortal E; edited by D. Schmandt-Besserat Invited Lectures on the Middle East at the University of Texas at Austin. Pp. viii-62, 47 plates. $16.00 (cloth), $9.50 (paper). Immortal Egypt represents some of the latest ideas and insights in various aspects ofthe field of Egyptology by some of the most authoritative American scholars in the field: G, R, Hughes, K. Butzer, W. K. Simpson, H. A. Liebowitz, T. A. Wertime, L. M. Gallery and C. Aldred. Professional and institutional discount af 20%on single copies and new subsctiptions until June 30, 1978 (higher ‘on larger orders), All prices are postpaid. Descriptive flyers and information on desk copies availabe on request UNDENA PUBLICATIONS, Inc., Dept J, P.O. Box 97, Malibu, Cali. 90265, USA.AFROASIATIC DIALECTS Editor: Thomas G. Penchoea (University of California, Los Angeles): Berber Advisory Bourd: Giorgio Buccelat (University of California, Los Angeles): Akkadian Russell G. Schuh (University of California, Los Angeles): Chadic Stanislav Sepert (University of California, Los Angeles): Northwest Semitic Aoasiatie Dislets (1AD) seeks to provide concise descriptions of individual languages which belong to the Afroaiatic Tanguage family. Its primarily directed toward an audience consisting. on the one hand, of students of one or several Afro site languages, and, on the other, of students of linguistics. In these volumes, both these groups should find sucinct lrcatises sich ae to provide familty with the basic strctare of the language ia question in a comparative perspective. Esch description will be comprehensive im scope and sufficiently detailed in exemplification, But atthe same time the ais will be 1 cut through fo the essential and to avoid specialized argumentation. The Rol then ls nether to publish a corpus of {exhaustive reference grammars nor fo provides platform forthe analytes! defense of theoretical questions, In thie sense the Series is properly datworiented. Though the authors will necesarly be of avarety of theoretical persuasions and each will hve his wn set of preferences for presentation, not the least important goa! wil be to achieve as high a degree 3s possble fof uniformity in structure, ad in the conventional sigs and terminology used. This belag accomplished, the reader should have no difficulty in finding points of resemblance and divergence amongst the languages which concern him with regard to some point of inquiry. The term dilects’ inthe vere tlle refers not only to modern spoken vernacular bu to historically ‘efinabe stages of any language of the various branches, Publcition of studies of as many Such dialects a posible would Provide, we feel, both an encouragement to comparative work and a sound documentary bass on which alone this work may fruitfuliy progres. AAD - Berber: TAMAZIGHT OF THE AYT NDIIR by ThomasG. Penchocn. 1973,1V-124 pp., $850. The Ayt Ndhirdislect which is described belongs to one of the major Berber languages, Tamatight, spoken in the Middle Alls Mountains of central Morocco, The description is based in the main on research Undertaken with native speakers of the Ayt Ndhir teitory surrounding El Hajeb. ~ While directed to the non-specialist «quinber of point in the description proper will be of interest 0 the specialist as well: the presentation of noun and verb morphology points up a number of Fegulaiies which ore often than not hate been obscured in previous descriptions. AKO, phonological rules are sven which account forthe major share of morphophonemic complexities. The reader wil find in the appendices and ‘optional’ sections onugation tables of typical verbs including detailed observations onthe placement of shwva in verbs a cart showing the ‘main morphological patterns involved in verb derivation, a description of the phonological rules applying in. complex Sequences of morphemes of the ver group, the “asic” vocabulary contained in several wellknown lexiostatistie word ist, Sndachart ofthe Tfinz alphabet used by the Tuarea. AAD 2 ~ Ancicat Egyptian: MIDDLE EGYPTIAN by John Callender. 1975, 150 pp., $10. ‘This grammar deals withthe iterary language used in Exypt from ca. 2000 to 1200 B.C. and considered in eve later times to be the catial written form of Eayptian. The book i diected toseard the general linguist a8 well a8 the Fayptologtt: fxamples are plossod and writen in transcription and there isan index of gractmatical terms and Egyptian morpherne. A Comprehensive set of paradigns of both verbal und non-verbal predicate types s included as an appendix, together with an appendix on negation and one on the historical origin of certain constructions. — The grammar contains three main pats phonology, morphology, and syntax, of which the last receives most emphasis. The section on phonology sketees the laws St sound change to the extent they can be scovere. The section on morphology strewes the paradigmatic character of ‘ea tenses and their derivations. A distinction is made between truly paradigmatic fenses and tenses borrowed from Old Egyptian for quotation or special effect. Following Polotsky, the "emphatic forms” are treated as nominalizations under the rubric “manner nominalzations.” Unlike previous grammars of Egyptian, ths grammar discusses syntax according 10 Itansformational categories, ‘The process of "clefting interrelaes emphatic forms, the “participa statement™ and onstructons with pws relaives. The proces character of negation is emphasized, and the implications of so considering it late developed in a special appendix. A sample text is also included, accompanied by a vocabulaty anda translation, AAD 3 - Se ie: DAMASCUS ARABIC by Arne Ambros Based on both previous works and the author's own observations, the ect spoken in Damatcus, Whe strictly synchronic and writen without presupposing knowledge ofclasial Arabic, it follows traditional srangement and terminology 48 closely as posible without failing however todo justice tothe indiv= ‘Guat traits of the diseet. Appendices deal with 1) the ogula reflexes of Cassia! Arabic phonemes in Damascus Arabic, find rules governing the reduction of vowels, and 2)a discussion of momphotogicl substitutions which cannot be inter- preted as deseribing the historia! development from Classeal Arabic to Damascus Arabic [Al pres ae posal. Payment mus company ores from ines A handag fo of 80¥ wil be charge to Urars {€ otder snot prepaid, Dicount of 20% on all orders rected within one year of pubiation dite Onder fi” UNDENA PUBLICATIONS, #0. Hox 97, Malibu, California 90268, US.AAAL STYLE SHEET UNDENA PUBLICATIONS The primary goal of Undena Publications is to publish at lowest cost possible while keeping the highest possible standards. In order to do this we need your cooperation when sub- mitting manuscripts. Please read the information below. General Procedures MANUSCRIPT: Keep a duplicate copy of your submitted manuscript since this will not be included when proofs are sent to you. PROOFS: Each author will receive proofs for corrections. CORRECTIONS AND CHANGES: Corrections and changes must be kept to an absolute minimum. Major changes, i.c., changes that affect more than a couple of lines, or a great many minor changes, will be at the author's expense. You will be presented with a bill based on the amount of work involved. All changes as submitted by the author on the proofs are suggestions only and may be disregarded at the discretion of the editor. Changes that affect entire pages will not be accepted. OFFPRINTS AND COPIES: Authors will receive 10 free copies of the work. Additional copies will be available at discounts of 30% for a single order of 5 copies or more, and 40% for a single order of 20 copies or more (a single order is to be billed and shipped to the same address). Manuscript Preparation MANUSCRIPT: All material must be typed, double-spaced throughout on on-erasable and non onionskin bond: photocopies or xeroxcopies are accepted. Isolated corrections may be entered by hand, but should be printed. PUNCTUATION, ETC. Material in foreign languages should have single underline; emphasized material should be doubly underlined. Use single quotes for glosses, double quotes for everything. else. FOOTNOTES: All notes are to be typed, double-spaced, on separate pages with running numeration. Footnotes should be restricted to substantive comments, not used for references (see below). ABSTRACT: All manuscripts should be accompanied by an abstract of approximately 100-125 words. TITLES: Authors are strongly encouraged to divide their manuscripts into sections, subsections, ete. numbered and titled. There should be a table of contents (following the abstract) referring to these sections on the following model: 1. (Major heading) 1.1 (Sub-heading) 1.2. (Sub-heading) 1.2.1 (Second level sub-heading) 1.2.2 (Second level sub-heading) 2. (Major heading) REFERENCES AND BIBLIOGRAPHY: References in the text should not be made in footnotes but rather by the last name of the author followed by date of publication of the reference, e.g. Diakonoff (1965). In the bibliography for journal articles, list Author's surname, first name. Date, “Title.” Journal Volume: Pages. For books, list Author's surname, first name. Date. Title. Place of publication: Publisher. If more than one publication bearing the same date is mentioned, use smalll letters, e.g. 1965a for the first publication by a single author with that date, 1965b for the second, etc. In referencing articles in edited volumes, give name(s) of editor(s).AFROASIATIC LINGUISTICS AAL includes contributions in linguistics within the vast domain of Afroasiatic (Hamito-Semitic) languages. Articles of general, theoretical interest using Afroasiatic material, descriptive, historical and comparative studies are included. Euitor: Robert Hetzron (1346 San Rafael, Santa Barbara, Ca. 93109, U.S.A.) Associate Editor: Russell G. Schuh (15337 Hart St., Van Nuys, Ca, 91406, U.S.A.) Advisory Board: A. Bloch, J, B, Callender, T. Givén, T. G. Penchoen, S. Segert. Volume One Articles by P, Newman, R. G. Schuh, J. L, Malone, R, Hetzron, T. Givon, T. M. Johnstone, B. W. Andrzejewski, and H. Minkoff. Volume Two Articles by D. R, Cohen, C, D. Johnson, A. Bamea, R. Nir, C. T. Hodge, G. Janssens, S. Segert, J. B. Callender, J. L. Malone, T. Givén, and A. D. Corré. Volume Three 1. RG, Schuh, The Chattc Verbal System and Its Afroasiatic Nature, 14 pp.,$1.20. 2G. Buccellati, On the Akkadian “Ateributive” Genitive, 9 pp. 1d, The Case Against the Alleged Akkadian Plural Morpheme -int, 3 pp., $1.00. 3. R.Hetzron, The Agow Languages, 46 pp. $3.85. 4. J. Saib, Schwa Insertion in Berber: Un Probléme de Qhoix, 14 pp. 1.20, 5. R. Steiner, On the Origin of the héder ~ hadkr Alternation in Hebrew, 18 pp. D. Boyarin, The Loss of Final Consonants in Babylonian Jewish Aramaic (BIA), 5 pp. 1. Avinery, The Position of the Declined K\ in Syriac, 3 pp., $2.20. 6. A. Zaborski, The Semitic External Plural in an Afroasiatic Perspective, 10 pp. Id., Consonant Apophony and Consonant Alteration in Bilin Plurals, 12 pp., $1.88. E, Rubinstein, On the Mechanism of Semantic Shift: Causation of Symmetric Locativity, 10 pp. $0.85. Bibliographical Bulletin, 24 pp., $2.00. P, Abboud, On Ablaut in Cairo Arabic, 22 pp. $1.85 Volume Four 1, Neil Skinner, North Baucht Chadic Languages: Common Roots, 49 pp. 1, ‘Fly’ (Noun) and ‘Mouth’ in Afroasiatic, 12 pp., $5.15. 2 Lewis Glinert, Number Switch: A Singular Feature-change Rule in Moder Hebrew, 38 pp.. $3.20. 3. Russell G. Schuh, Bade/Wgizim Determiner System, 74 pp., $6.25. 4. Joshua Blau, The Beginnings of the Arabic Diglossin A Study of the Oriins of Neoarabic, 28 pp., $2.35 Volume Five 1. Paul Newman, Chadic Classification and Reconstructions, 4 pp., $3.55. 2. D. L. Appleyard, A Comparative Approach to the Amharic Lexicon, 67 pp., $5.65. Subscription to one volume of about 200 pp.: $15.00 Individual issues available separately Descriptive flyers, with complete lst of abstracts and prices for individual issues, are free on request Write to: UNDENA PUBLICATIONS, P.0.B. 97, Malibu, California 90265, US.A.