Preface It has been almost 30 years since I wrote my thesis on Sicite tone. After completing the thesis, I returned to the village of Kotoura in Burkina Faso, where I continued to live for several more years, living and working among the Sicite people. This allowed me refine some of my understanding of the Sicite language through further research. Two articles have since been published in the Journal of West African Languages, one on the challenging analysis of a variety of mid tones in the language, and the other on the phonological structure of the Senufo word. The latter corrects some of my misconceptions of the vowel system that I had at the time I wrote my thesis. This work led to the development of the Sicite orthography which resulted in literacy programs and publication of literature in Sicite. I am so thankful to God to be able to contribute in a small way to the development of the Sicite language and proud of those who continue to teach and write and read in Sicite. In this way, the results of my research are not relegated to a little-read thesis for a few linguists, but serve the people as a whole. Two people in particular have contributed to my further understanding of the Sicite language: Moussa Traoré, who served as my language assistant, and transcribed hundreds of pages in his language, and Zanga Lassina Traoré, who enriched his linguistics studies by researching in his own language and by directing the first literacy program in Sicite; my whole hearted thank-you to these two friends and colleagues. One person and one institution have greatly contributed to ensuring continued access to the original thesis despite the fact that it was not officially published: My father, Leonard W. Garber, a computer programmer, created special characters so that I could write my thesis on the computer way back in the 1980’s. Since then, he updated the fonts and the software, and put it online on my personal website. And now Max Planck Institute has gone one step further by ensuring a longer on-line life and incorporating it into their Language Description Heritage library. I want to thank both of these people for making it available to the linguists who want to investigate the Sicite language both now and in the future. Finally, I want to give special thanks to both of my parents, Leonard and Doris Garber, and my husband, Daniel Kompaore, my most faithful supporters and encouragers in my quest to share with others what I have learned. Anne Garber Kompaoré Ouagadougou, Burkina Faso, October 28, 2014 INFORMATION TO USERS While the most advanced technology has been used to photograph and reproduce this manuscript, the quality of the reproduction is heavily dependent upon the quality of the material submitted. For example: • Manuscript pages may have indistinct print. In such cases, the best available copy has been filmed. • Manuscripts may not always be complete. In such cases, a note will indicate that it is not possible to obtain missing pages. • Copyrighted material may have been removed from the manuscript. In such cases, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, and charts) are photographed by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each oversize page is also filmed as o n e exp o su r e and is available, for a n additional charge, as a standard 35mm slide o r as a 17"x 23" black and white photographic print. Most photographs rep r o duce accep tably on positi v e microfilm o r microfiche but lack the clarity on xerographic copies made from the microfilm. For an additional charge, 35mm slides of 6"x 9" black and white photographic prints are available for any photographs or illustrations that cannot be reproduced satisfactorily by xerography. Order Number A tonal analysis of Senufo: Sucite dialect Garber, Anne Elizabeth, Ph.D. University of J1linois at Urbana-Champaign, 1987 Copyright @)198'1 by G81'ber, Anne Ensabeth. All rights reserved.. U·MmI 300 N. Zeeb Rd. Ann Ar'Jor, MI 48106 8121638 PLEASE NOTE: In a" cases this material has been filmed in the best possible way from the available copy. P roblems encountered with this document have been identified here with a check mark_..f_. 1. Glossy photographs or pages 2. Colored illustrations, paper or print 3. Photographs with dark background 4. Illustrations are poor copy 5. Pages with black marks, not original copy 6. Print shows through as there is t�xt on both sides of page 7. Indistinct, broken or sma" print on several pages S. Print exceeds margin requirements 9. Tightly bound copy with print lost in spille __ _ a/' __ 10. Computer printout pages with indistinct print 11. Pagels) author. lacking when material received, and not available from school or 12. Page(s) seem to be missing in numbering only as text follows. 13. Two pages numbered . Text follows. 14 . Curling and wrinkled pages v:: 15. Dissertation contains pages with print at a slant, filme d as received 16. �her V ____________________________________________ University Microfilms International A TONAL ANALYSIS OF SENUFO: SUCITE DIALECT BY ANNE ELIZABETH GARBER B.A., A.M., University of Ottawa, University of Illinois, 1978 1980 THESIS Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Linguistics in the Graduate College of the University of Illinois at Urbana-Champaign, Urbana, Illinois 1987 UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN THE GRADUATE COLLEGE MAY 1987 i :1 WE HEREBY RECOMMEND THAT THE THESIS BY I ANNE ELIZABETH GARBER ENTITLED, L�AN�A LY �F�S�EN�U�F�O�: S � � IS � O � � �O �N� A� -BA�T ____ _________________________ SUCITE DIALECT BE ACCEPTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY \, u�w- k� Wee o � Final Examination", � W. J� &k.� Director of Thesis Research Head of Department Com m , Chairperson -_........ t Required 0·517 for d octor's degree but not for master's. I il !J I ® CaPYri ght by Anne El i zabeth Garber 1987 iii TONAL ANALYSIS OF SENUFO: SUCITE DIALECT Anne El izabeth Garber . PhD . DePartment o� Li n9u i stics Uni versi ty o� I l l i nois at Urbana-ChamP�i 9n . 1987 Char 1 es Ki sseberth . Advi so'" Suci te. a Senu�o lan9uage Ot the Gur language 9rouP. i s spoken i n southwestern Burki na Faso. Its tonal system o� three l e'vel tones and several contour tones exh i b i ts a consi derable number Ot compl ex al ternations. Thi s dissertat i on Provi des a descriptive anal Ysi s o� the tonal al ternati ons in Sucite. Wi th the hel P o� the autose9mental approach and Cl ements (1981) system o� tone teatures, we ProPose a doubl e ti ered apProach to tonal ana l YSis i n the attemPt to anal yse the behaviour Ot the vari ous tYPes Ot Mi d tone tound in the lan9Uage. The d i ssertati on consi sts o� an i ntroducti on and si x chaPters. In Chapter 1, we Provide a br iet descr i pt i on Ot the sound system. the m�rPhol o9Y. and syntax Ot Suci te. chapter 2 descri bes the tone and morPhol ogy of the verb and i ntroduces the concept Ot two tiers tor tonal anal ysi s. tone and morPhol ogy i n Chapter 3 The di scussion Ot noun bri ngs to l i ght the need to re-ex ami ne the accePted uni versal o� associ at i ng tones to segments from left to right. ChaPter 4 is a descri ption and anal ysi s of the tonal behaviour of verbs and verbal Particles when Preceded by �omi nal and verbal el ements at various tones. In Chapter 5. we exami ne how the nomi nal el ements atf ect each other tonal l Y within a noun Phrase. Both tonal behaviour across word boundari es and tonal al ternations wi thi n comPl ex nouns are exami ned and anal ysed with the use Ot the double-ti ered apProach . I n Parti cular . thi s chapter h i 9hl i 9hts the need tor several dit terent tYPes Ot underlYing rePresentati ons t or Mi d tone. Chapter 6 iv d i scusses the tonal behavi our o� the adverb Phrase. quest i on �ormati on . and the noun class cl i t i c , the l atter of whi ch Poses sPecial anal yti cal Probl ems. Final ly. the orderi n9 o� the tonal rul es presented i n the thesi s i s di scussed . --- -- -- - ---- -- ----- ----- --------�-- -- - v Dedi cated to mY fri ends i n Kotou�a and to the One who i s al waYs wi th us ----- - vi ACKNOWLEDGEMENTS Thi s studY woul d nat have been possi ble wi thout the cooperati ve suPPort o� numerous Peopl e. My � i rst exPressicn o� grati tude must be di recteq to mY language consul tants who 91ad l Y Provi ded the data used i n this thesi s. Pri marY consul tants have been Ouattara Nama (�armer ) . Tra�r4 Mamadou de Katien (student > . Traore Mamadou de Kat i l e (student ) . Traore Issiaka de Kati l e (Youn� farmer > . and Traore Gnodjouti en (school teacher > . I wi sh also to thank mY �riend and co-worker . Gai l Wi ebe Toevs. �or her comPani onshi p and �or shari ng wi th me her i nsi 9hts an Suc i te dur i n9 our Years t0gether i n Burkina Faso. The Commi ssion on Overseas Mi ssi ons o� the General Conterence Mennoni te Church and the A�rica Inter-Mennon i te Mi ssion are al so to be acknowl edged �or kindlY permi tting me to Pursue the research requi red �or this di ssertati on . as thei r emplOYee. and �or Provi ding a measure o� � i nancial assi stance. Thanks also 90 to the members o� mY commi ttee �cr thei r comments and SU99Psti ons. in Parti cular . to my advi sor . Char les Ki sseberth �or Patientl y gui din9 m e through the numerous revi si ons requi red �or t h e thesi s. Final ly. I wish to thank al l o� mY �riends �or the suPPort they provi ded durin9 thi s l on9 ordeal . as wel l as mY si ster . Carol Yn . who tYPed much o� the aPPendi :< and hel ped to PrePare the maPs in this thesi s. --- - - --------- ----- Thank-You ! vii TABLE OF CONTENTS INTRODUCTION • 1 • 1 4 8 Cl assi f i cati on of Senufo within the GUR l an9uage 9rouP Suci te as a SENUFO l an9uage (di al ect ? ) Senufo l an9uage Research PurPose, APProach and Presentat i on NOTES • CHAPTER 1 II. BASIC FACTS ABOUT SUCITE - INTRODUCTION 12 THE SOUND SYSTEM A. Consonants • 1. Pre-nasal i zed Consonants and Consonant Cl usters 2 . Stress and Consonants • B. Vowel s • C. Tone 12 12 14 15 18 23 I I I . MORPHOLOGY AND SYNTAX A. Nouns B. Pronouns • Personal Pronouns C. The Noun Phrase D. Verbs E. The Verb Phrase F. The Adverbi al Phrase G. The Sentence H. Ne9ati ve Formati on • I . Questi on Formati on • • • SUMMARY • 25 25 26 28 28 31 32 33 34 36 36 • 37 • 38 NOTES CHAPTER 2 - 40 THE SUCITE VERB 40 I. INTRODUCTION I I. THE COMPLET IVE VERB A. Structural Descri pt i on B. Tonal Descri pt i on • • I I I . THE I NCOMPLETIVE VERB · A. structural Descri pt i on • B. Tonal Behavi our of the IncomPl etive Verb • 1 . The Under l Yi n9 Tone of the i ncompl eti ve suff i x . ------ ---- 12 • • IV. 9 11 - -- -- --- - ----------------------------- 40 40 42 42 42 51 51 vi i i 55 2 . Root Tone Mutati ons i n the Incomplet i ve Verb IV. HISTORICAL SPECULATI ONS AND TONAL FEATURES V. (;ONCLUS ION . . . NOTES CHAPTER 3 I. - 61 . . . . . . . . . . . . . . . THE NOUN 68 70 71 • INTRODUCTION A. Noun Structure • B. Noun Cl asses and SuH i x i ng C. Noun Tone 71 72 72 73 • • • • • • • • • • • • THE INDEFINITE SUFF I X - TYPE I A. Structure • • • • • • • • • • • • • • • • • B. The Tonal Nature of I ndefi n i te Nouns - TYPe I • • 1 . The tone of the I ndef i n i te Suffi x - TYPe I • • • 2. Associ at i on conventi ons and the Indef i n i te Noun . 3 . Tone and Pre-nasal i zed Consonants • 4 . Hi gh f i nal nouns • • • • • • • • • • • • • • • • II. . • • • • • • • • . . . . . 75 75 80 82 84 87 89 91 I I I . THE INDEFINITE SUFF IX - TYPE I I ( Cl asses 2 and 6 ) . 91 A. Structure • • • • • • • • • 91 1. Cl ass 6 Indef i n i te Suff i x . 94 2. Cl ass 2 Indef i n i te Suf f i x • • • • • • 96 3 . SummarY • • • • • • • • • • • • • • • • • • • • • • • • 97 • • B. Tone of TYPe I I Indef i n i te Suf f i xes 97 1. Tonal Behavi our ot the Suff i x • • • • • • • • • • • • • • • 99 2 . Segmental Del eti on and Tonal Behavi our • • • • • 116 3. Root tone l ower i ng (High Del eti on ) • • • • • • • • • • a. Hi gh Del eti on wi th Suf f i x Reduct i on ( 122 ) ; b . Hi gh Del et i on wi th Ful l Suf f i x (124) ; c. Parti al Suff i x Del etion and Resi stance to Hi gh Del eti on (125 ) IV. THE DEFIN ITE SUFFIX A. Structure - Type I • • • • B. structure - TYPe I I nouns C. Def in i te Suf f i x Tone . V. CONCLUSION . . . . . . . . . . . . . . . . . . . . . CHAPTER 4 - TONE I N THE SUCI TE VERB PHRASE • • • • INTRODUCTION Domai n Ot Verb Tone Behavi our ---- ---- . . . . . . . . . . . . . . . . NOTES I. • • - -- - - ._ --- • • 128 129 132 134 141 142 145 145 145 ix II. NOUN OBJECT TONE AND THE VERB : DESCRI PTION AND ANALYS I S . . . . A . Descri pt i on B . Anal ysi s • 1 . Low SPread i ng Ru l e . . . . . . . . 2 . High Tone SPread i ng 3 . Mi d Tone Loweri ng . . . . . . . . . • • • • • • • • • • • . . . . . . . . . . • . · • . • I I I . THE VERBAL PARTI CLES A. Order With i n t.he Verb Phrase . B. AsPect i n the Verb Phrase 1 . The I ncomPl eti ve Aspect 2 . The Compl eti ve Aspect Marker C. Tanse, Modal , and Other Parti cl es 1 . The NA Parti cl e 2 . The NIl Part i cl e 3 . The Future Tense . . 4. The I ncompl eti ve n'7' 5 . The CA Parti cl e 6 . The Habi tual m4?a 7 . The Modal SA 8 . The BA Modal 9 . The Pre-Tense Part i cl es 10 . The Ne9ati ve Parti cl e • • • • • . . . . . . . . . . . . . . . . • · . . . . . . . . . . · . . • • • . . . • • • • • • • • • IV. CONCLUS ION NOTES . . . . . . . CHAPTER 5 - . . . . . . . . . II. • • . . . . . · . . . . . . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • ---- - --- -- ---- ------ 198 201 • TONAL BEHAVI OUR W I TH I N THE NOUN PHRASE A . H i 9h Tone Nouns B. Low Tone Nouns C. M i d Tone Nouns D. Mi d and M i d -Low Tone Nouns E. Weak M i d and Mi d-Low tone Nouns 1 . Desc:,- i Pti en ( i ) Rai s i n9 (2 1 3); ( i i ) Low tone sPread i n9 (2 1 6) 2 . Anal ysi s a. Mw and ML as Under l Yi n9l y H i gh Tone b . Mw and ML as Under l Yi ngl y HI (i) Low tone SPread i ng (229); ( i i ) High Tone SPreadi ng (231); ( i i i ) Di st i ngu i sh i ng Weak Mi d and M i d-Low Nouns (235); ( i v) Mi d -Low Tone and I ndePendent Low Tone Li n k i ng (235); ( v ) Weak M i d Tone and Feature Swi tch (239); (vi ) Rul e Order i n g (241 ) F. Low-Hi gh Nouns and RS Del i n k i ng G. Concl usi on ot Mi d tone behavi our on Nouns • 1 73 1 73 1 75 1 76 178 183 183 185 186 188 188 1 89 191 193 1 94 1 96 200 INTRODUCT ION A. The E��ect o� Noun Phrase Structure on Tonal Behavi our B. Nomi n al Tone • . · . . . . TONE I N THE SUCITE NOUN PHRASE • · 146 146 1 49 1 49 157 1 69 • • • • • • • • • • • • ----- ---- ---- ---- 20 1 202 204 205 205 206 207 210 212 213 218 218 227 243 247 :( 248 I I I . THE DEF I N I TE SUFF IX REV I S I TED • • IV. TONAL BEHAV I OUR OF THE COMPLEX NOUN • • • • • • • • • • • • • • • • • A. ComP l ex Nouns (HI + HI ) and the Ob l i 9atorY Contour Pr i nc i pl e . 1. The OCP Sol ut i on • • • • • 2. Rul e Order i n g for tile OCP • 3. H i 9h SPreaD i n g and w;P 4. Low Spreadi n9 and the OCP 5 . OCP and the Assoc' i at i on Conventi ons B . H i 9h Del i n k i n9 • • • • • • • • • • • • • • • • • • • • • • C. Lex i ca l Changes 1 . Hi9h Del eti on • • • • • a. Hi9h Del eti on I (269) ; b. Hi 9h Del eti on I I (275) 2. Low Del eti on • • • • • • • • • • • 3. Other Lex i cal Changes • • 4. Tone on Nomi nal i red Verbs • • • • • D • . Concl us i on of Study of COmPl ex Nouns • • • • • 278 292 293 288 CONCLUSI ON • • • • • A. The Doubl e t i ered APProach to Tonal Anal ysi s • B . SummarY • • • • • • • • • 29 9 289 293 • V. NOTES . . . . . . . . . . CHAPTER G - THE SUCITE SENTENCE A. B. C. D. E. F. . . . . . . . . . . I ntroducti on . . . The Adver b i al Phrase • The Noun Cl ass C l i ti c . . . . Yes-No Quest i on s • . wh Questi on Format i on and Frontshi fti n9 Rul e Order i n9 G. Concl ud i n9 Notes . · · · · · · 295 . . . . . . . . . . . . . · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 252 252 255 256 257 258 259 263 269 269 . . . . 297 . . . . . . · · · 297 297 301 313 314 3 16 3 26 BIBLIOGRAPHY 328 BIBLI OGRAPHY OF SENUFO LANGUAGES 330 APPENDIX • • • • • • 332 332 352 368 NOUN LEX ICON OF SUCI TE. VERB LEX I CON OF SUC I TE WELMER ' S QUESTIONNA IRE VITA • • • • • ----- ---- - -- - . . . . . . · . . . . . . . . . . . . . . . 399 1 INTRODUCTION The sub ject ot thi s d i ssertati on i s the studY ot the tone ot Suci te [ Suc� tel . a Senuto l anguage ot southwestern Bur k i na Faso, i nvesti gated on l ocati on by mYsel t between the Years 1992 and 1995. Thi s i ntroduct i on t i rst si tuates the Senut o l anguages as a whol e both geograPh i cal l y and geneti cal l y bet ore d i scuss i ng the rel ati on o t Suc i te with the other Senuto l anguages. I t continues wi th a revi ew ot the l i ngui sti c research i n var i ous Senuto l anguages and ends wi th a d i scussi on ot why and how the subjec t ot Suci te tone i s to be Presented i n the thes i s . CI4"ifiC4tion of e.nufo within the SUR 14ngY491 grouP It has been w i d elY acknowl edged by authors such as Lavergne de Tressan ( 1953), Greenberg ( 1966) and Westermann ( 1970) that the Senufo l anguages bel ong to the Gur grouP of the Ni ger-Congo l anguage fami l y . The Gur or Vol tai c l anguages are l ocated Pri mar i l y with i n the bas i n of the Vol ta Ri ver between the Sahara Desert to the north and the troPi cal forest to the south. In terms of pol i t i cal boundar i es , thi s area stretches IIfrom the southeast corner of Mal i . across northern I vorY Coast , throu9h a l arge Part of UPPer Vol ta [Bur k i na Fasol , northern Ghana, northern T090 , northern Dahomey [Beni nl i nto Niger i a . (Bendor-Samuel , II p. 141, 1971) Senu f o l anguages are l ocated i n the southwestern corner ot th i s area, bounded to the west by Mande l anguages , to the south by Kwa l anguages, and to the north and east by other Gur l anguages. The i r area i s spl i t among pri mari l y three countri es : I vorY Coast boasts the l argest number of sPeakers (over 700 .000 ) (Mensah , 1983), Mal i comes c l ose beh i nd (690,272) ( �, 1991) and Bur k i na Faso has an estimated 100,000 sPeakers. Northern Ghana al so has a t ew ----- 2 i sol .1f:ad grouPs. Senu�o l angua9as are tYPi cal o� Gur l anguages in that they have a noun cl ass su��ix system and verbs are marked tor asPect (comPletive and incomple�i ve) . In ganer�l p the consonant and vowel systems � Senu�o l anguages do not have anY particul arl y stri king �.aturesl . Thera i s a d i sti nct di vergence in werd order . however . Where most Sur l anguages are SVOp Senu�o langua�s are SOVAdv2• A �ew other nei ghbouri ng Gur l anguages i n Burkina Faso, such as Toussi an3, also have this word order. Tonal l y , there i s no one common Pattern throughout Gur languages. Samuel ­ Bendor ( 1971 ) nOtesp however, that a two tone downsteP system aPPears to be an areal �eature cutti ng across Gur l anguage divisions SPreading into the Kwa grouP. Thi s two-tone areal �eature stoPs short at the Senu�o languages. To both the east (southeast ) and west D� the two-tone set o� languages, we �ind entire blocks 0+ l anguages wi th SYstems Ot three level tones. The Senuto lan9Ua9&Sp which are located to·th. w.�t o� the two ton. sYstems, as wel l a. a �.w other mi scQl lanaous SUr l an9Uages bordering on the Senu�o area, Poss.ss three l evel tones. On the western side at Senu�o l .1nd, there are a numb.r o� /'Iand. l anguage. which have pri mari l y twa di screte ton... I� Mi nYianka, a Senuto l anguage, does actual l Y have onl Y two l evel ton•• , •• has b••n r.port.d4, i t maY b e due to a more Prol onged contact with Mande l anguages in the �ar north­ western corner o� the Senu�o area. In the �ar southeastern corner, on the other hand, we �ind two rePorted �our-tone Senuto l anguage., Jiminie and TakpEr6• which just haPPen to be i n the vicini ty o� several tour-tone languages o� other language grouPs - BEtE (Kru ) . Atti' <Kwa) , and Gban (Mande) <See adjoi ning maP �or location ) . Again , i t aPPears that l anguage contact maY play an important rol e in t�� devol opment o� ---- - - .-. --- ----- ------- --- ---- -- _._-_. --'--- 3 LJtN�"AGE li 110 UPS SURKII\IA fASO MALI M Ke)': �"'Q� Gt-o",� ��Q,"\t$: L�IJA3e 8o�i�: Mandl.: M Go...... : C. KwQ.� Kw Kl"'u: K... _ ___ _ 4 a l an9Uage's tonal .vstem. Further comParative studies in the tonal svstams o� Wast A�r i ca cou l d Potential l v vi el d verV i nteresting insights as to whv and how tonal svstems change and $hi �t through the centuries. Sycit- ., t SENUEO ItaiM'" (di.lect?) Unti l the l ast coupl e o� decades, Senu�o l an9Uage classi �i cation was primar i l v a matter o� specul ation . In a Personal communi cation to Bendor-Samuel ( 1968) . R. Mi l l s divided UP Senu�o l anguages i nto three di alect. grouPs and l abel l ed the.. Northern. central . and Southern Senu�o. Bandor-�iamuel ( 1 971 ) aPParentl v agreed wi th these divisions. Two recent publ i cations. Mensah ( 1983) and Mi l l s (1984) avoid these geographic l abel s. but thev grouP the Senu�o l anguages i nto roughlv the same cate90ries as R. Mi l l s ( 196 8) . apparentlv al l owing �or more �uzzines. between group boundar�... In certai n cales. i ndi vidual dialect. have not been adquatel v i nvesti gated cation �or accurate classi f i ­ Added to that, the compl i cation o� mul t i pl e names �or a number of diale:ts creates a comPl .x and. a. vet . i l l -d � i ned l i nguistic grouP o� l an9UageS"' . Before mV arr i val in Burkina Faso i n 1982. I onl Y knew o� SucHe as "Tagba". Lavergne de Tre�5an ( 1 953) (and PerhaPs before him. Tauxi ar ) , incl udes Tagba in his l ist � 30 Senu�o dialects. He had apparentl v i denti f i ed two locations �or Tagba. one i n nor�hwest IvorY Coast and the ather i n Burkina Faso. He gave no indi cati on o� what relation i t had to other Senu�o l an9Ua9as. I specul ated that Tagba mi ght be rel ated to SuPVire, the Senu�o l anguage directl v to the wast across the border i n Mal i . MY hunch Proved to be correct . as I l ater di scover­ ed. For not onl Y were Tagba and SuPVi re closel v rel ated . but the real name �or Tagba. SUci te. was a re� lection o� the close phonol ogi cal ----- ._-- - -- .- -- --- ---- ------- _. _-- .- 5 MAll SENUFO LANGUAGES SYENtftE J<EY� A 5 C 0 E Cepiri Po�Q.r� KClSlt"f. Gbot\'3.ororo TQ\"I!jG.Y"G ---- - ---- - -------- F G. H I :r Fodonon Gbo�7.o"'o FQn+a.Y""Q Pa\'\1(U"Q NaFQ\1Q (BowJou ko\.l') --- ----- - --- -- 6 correspondp.nca betNe.n the two. Bath SuPYire and Sucite bel ong to the northern grouPing o� Senu�o l anguages. According to Mi l l ' s (19 84) ma�. no disti ncti on i s made betNeen the tNO dia­ lectsJ both are under the same label . SuPYire. SUPYi re i s located in the Si kasso area o� southeastern Mal i . Sucit. is an eastNard continuation of SuPYi re i nto Burkina Fasa. According to l egend. the Suci te-sPeaking peopl. origi nate i n Mal i and the Sucite saY they sPeak the same language as the SuPYire. In �act. soma People in Burkina Faso cal l thei r language • SuPYire , rather than ·Suc i te ' . Chance encounters betNeen i ndi vi dual s o� the tNO groups has shewn that the two dial ects are qui te mutual l Y i ntel l i gi bl e. When sPeaking to an outsider . a Sucite sPeaking Person Ni l l saY he speaks Senu�o or perhaPs Bamana. as non-Senufo outsi ders Nould cal l him. Another Senu�o grouP to the South Ni l l cal l these Peopl e Tagba. The NQ;"d Tagba i s geosraPhi cal in nature. The Suci te peopl e l i ve on the Ta9auara pl ateau. Suc i te is Nhat the Peopl e cal l thei r ONn l anguage. Deri ved �rom the same root are the Nords ' si c3' (person) and · s�p{le' (P80ple) . The Suci te-sPeaking Peopl e in Burkina Faso number apProx i matel y <actual fi gure unkncwn) . 2�.000 TheY are l ocated i n the Kol oko Prefectul"e i n the Province of Kenedougnu about 1 1 0 kil ometres Nest of Bobo-Dioulasso in south­ Nester"n Burki na Faso. According to authori ties in Koloko. there are 32 vi l l ages. but thi s � i gure is not exact si nce a �eN vi l l ages are not Tagba. and at least one Tagba vi l l age i s l ocated autside the Pre�ectur• • The l i ngui stic nei ghbours o� Suci te are Nanergue. a northern Senu�o language to the ncrth . SuPYi re, to the NeSt , and a Central Senufo language tc the scuth knoNn as Senar o� Kankalaba (Prost. 1 964) . To the east and southeast are a number of smal l and di verse grouPs: Turka (Gur ) , Samagho (Mande) . Dicula ----- ------ -- - ------- ---- ---- ------ LANCUAGES OF SOUTUWESr I3URKJWA 7 FASO , • , MALI ,­ I • • • � , 8URKINA FIfSO I , Key: 1 .:."'. �"ca.\� QPpn",.lmo.+.''I SUC,'/oe: if' * Se"a.ri {KCI"kClIa.oQ: y Se.�a.ri: ---- - ._-- - .- X WOt'Cl : 10 � Na. f,'o,.o : -& Oicn... /o.: 0 SQm0,Jl1o I'n,le$ 1u.r-ka.: v NOt"�h Tou$:u'on: 2S GCU..L,'", � � $ou�h Sl!.mc,: � RQo.d.: ---------- -------- ------- .. Ta\ol�$iQ,": C 8 ( Mande) , and Si amou or S£me (Kru ) . The l atter has el uded cl assi � i cati on , though it has been suggested bv Person (1966) that Si amou i s an i sol ated Kru l anguage. A�ter visi t i ng 6 major vi l l ages in the area, mV Partner , Gai l Wiebe and I chase to l i ve i n Kotoura. a geograPhi cal lv central i zed vi l l age o� aPProx i matelv 2,000 Peopl e . A � e w c i vi l servants and a COUPl e o� Dioul a speaki ng �ami l i es l i ved on the Peri Pherv of the vi l l age but otherwi se the vi l l age was Purel v Senu�o. The Per i od o� l anguage invest i gati on began an arri val an November 25, 1992, and conti nued unti l mV departure on June 9, 198�. s.oYfA l,aiM•• Bey.rch Before the 1�0Is" tone was nat gi ven anV i mPortance i n Senufo studies­ Cheronls (192�) descri pt i on o� Mi nvanka does not even ment i on tone. Prost (1964) acknowledges the ex i stence of tones but makes no attemPt at transcri P­ tion . Seri ous research of Senu�o l anguage5 began i n the m i d 19�01S when Conser­ vati ve Bapti st mi ssi onaries settled i n the Korhogo area of northern I vorV Coast . Thei r ear l i est descript i ons o� Cabara ( Mi l l s , 19671) ref l ect thei r re�l i zati on that tone was an essential part of thi s major Senufo l anguage. To date, numerous art i c l es maV be found on a var i etv o� Senufo dialects· and three theses or d i ssertat i ons have been wri tten . one on TV.bari (Laughren . 1973). another an Nvarafol o (Baese, 1983), and a third on Fodonon (Bout i n , 1981). In addi tion , at l east ten l anguages/d i al ects are currentlv being researched . A sin9l e book has been PUbl i shed in recent Vears; Seaufg Phgngl ggy (1984) bv El i zabeth Mi l ls . Al l o� these resources , publ i shed and unpub l i shed , have Proved helpfu l i n the anal Vsi s of the basic features of Suc i te. The anal vsi s of tone, however , has nev"r been gi ven much sPace in al l of ----_ . ---- - .--- 9 these works. Mi l l s as wel l as a few others do describe the basi c tonal Patterns o� the l anguages that they have investi gated , but l i ttle attemPt has been made at a ri gourous anal Ysi s. a tonal anaisYi s an Bob Carl son i sl:.h� -F i rst . to mY knowledge, to attemPt a Sanu�o l anguage. He has wri tten an unPubl i shed descri P­ tive draft of SuPYi re tone, as wel l as the article "Downstep in SuPYi re" ( 1 983) . Hi s most recant unpubl i shad PaPer , SypYira tRn. ( 1 9� ) , i s h i s fi rst attempt at a more comPrehensi ve anayl sis. al though the scoPe o� h i s PaPer does not al l ow him to Present a thorough examinati on o� al l tonal behavi our. PUr_e. ApprO.ch .ad !tuentlt{on The PurPose o� this research i s to explore the tonal Patterns of Suci te, a dialect whi ch has not Previousl y been investigated . It i s hOPed that the Presentation o� these data wi l l �urther the knowl edge o� the tonal behaviour of Senufo l anguage. i n general . The scoPe o� thi s studY wi l l be l i mi tad to the anal Ysi s � the simple non­ complex sentence i n Suci te. Tone in comPl ex sentence structure and di scourse wi l l not be d�4Ji: wi th in detai l �or reasons of time. SP.ce, and l ack o� su�� icient data. However. prel i mi narY observati ons indi cate that tonal rules outl i ned i n the di ssertati on are not contradicted in complex sentence struc­ tures. Several theoreti cal tool s wi l l be used in the attemPt to provi de a satiS­ factorY anal Ysis. Firstly, the underlYing tones that provi de the basis for rules wi l l have to be determi ned . Research into the Possi b l e histori cal Process of tonal devel opment with the hel P o� Cl ements ' prOPosed feature system. as outl i ned in ChaPter 2, contri butes t owards this end. Secondl y , ---- ---_. --.- an autosegmental apProach to tonal analYSis wi l l be undertaken. ._------ ---- --- ------ 10 However , it will be qui ckl y seen that the autosegmental aPProach , as it has been general l y aPpl i ed wi th re�erence to tone, maY requi re some modifications here. In ChaPter 3, it wi l l be noted that the assumed Association Convention o� le�t to Right linking Poses some Probl ems. An al ternative sol ution ot Ri ght to left l i nking is proposed, di scussed , and �inal l Y adoPted . In addition , with the hel P of Cl ements' ProPosed feature system, whi ch invol ves def ining tone through the use of both a Pri marY regi ster as wel l as a subregi ster level , we borrow some basic concePts �rom the recent developments in non-l inear Phonology and ProPose a double ti ered aPProach to tonal anal Vsi s. Thi s concept is f i rst i ntroduced in ChaPter 2, and i s brieflv di scussed i n ChaPters 3 and 4, whi le a more thorough presentation o� a double ti ered aPProach �or tonal analYsi s can be �ound in Chapter 5. Final l y it wi l l be noted in Chapters 3 and 6 that rul e orderi ng is cruci al for the ProPer aPPl i cati on � tonal rul es. Certain l exical rul es wi l l have to take pl ace be�ore other rul es such as rules sPreading tones �rom one morPheme tD another'. The presentati on o� data wi l l be organized accordi ng to chaPter. The �irst chapter wi l l present the basic �acts about the Sucite language, includi ng Prel iminary information about tonal behaviour. The second chaPter wil l exami ne verb tone and morphol ogv, whi l e Chapter 3 wi l l focus on nen-complex noun tone and morphol ogy. Chapter 4 wi l l examine how the tonal envi ronment i n whi ch the verb or verbal particle is pl aced, af�ects thei r tone, whi l e Chapter 3 wi l l concentrate on tone i n COmPl ex nouns and noun Phrases. ChaPter 6 wi l l comPl ete thi s descri ption bv describing the tonal behavi our of adverb Phrases and of the sentence as a whol e. The noun cl ass cl i tic. whose tonal behaviour can al ter tone in the sentence i n unexPected wavs. wi l l be tnm �ocus of attention here. ---- - ---- - -- - 11 NOTES 1 . For A general surveY o.f Gur l angUAgeS, see Bendor-Samuel ( 1 971"> . 2. A di scussion concerning the hi stori cal impl i cations of ward order in Senufo can be .found i n "Ward Order Ch."ge and the Senufo Language", Garber ( 1980) . 3. For descriptions of Toussian , see Prost ( 1964 ) and Burdon ( 1984 ). 4. Euni ce Pi ke, uPon Prel i mi narY investigationf suggested that Mi nYi anka has twa level tones and one fal l ing tone. s. Information from Mensah ( 1983) indicates that Jimini i s a four tone l an­ guage. However , MensAh , hi msel f, questions"the rel i abi l i ty of his source, which, un.fortunateI Y, he does nat name. Wol fgang Stradner , a l i ngui st working an Jimini wandered i r Jimini might b. underlYi ngl y a twa tone language. To mY knowledge, l i ttle serious anAl Ysi s of tone has been undertaken in Jimi ni . 6. For a phonol ogical descri ption of TakPEr, a Tagbana di alect of Senufo, see Heraul t ( 1973) . 7. References .for l ingui stic maPs and classi fi cati ons o.f Senu.fo di al ects i ncl ude the fal lowing: Lavergne de Tressan ( 1933) , Westermann ( 1970) , Bendor­ Samuel ( 1 971 ) , Boutin ( 1982) , Mensah ( 1983), and Mi l l s ( 1 984 ) . 8. The reader i s asked to re.fer to the "Bibli ography of Senufo Languages' located at the end of this di ssertation for a .ful l l i st of publ i shed and unPubl ished materi al an Senu.fo l anguages. CHAPTER 1 - '11. BASIC FAcTS ABOUT SUCITE I. INTRODUCTION Chapter 1 out l i nes the sound, morPhol gi cal , and syntac tic structures of the l anguage 50 that the reader maY examine data in succe�di ng chapters wi th some degree of fami l i ari ty. Informati on on tone is l i mi ted to outl in in g the Phoneti c tones found i n Suc i te and l i st i n9 vari ous tonal comb i nati ons on three-syl l ab l e words. After desc�i bi ng basi c morphol ogi cal structures of nouns. ��ouns, and verbs. the word order of the sentence i s di scussed i n rel at i on to where tonal al ternati ons take Pl ace among i ts consti tuents. Al though examples of tonal chan9aS are 9i ven . the rul es i nvol ved are l eft for thorou9h exami nat i on in subsequent chaPters. I 1. 11£ SOUND SVSTaI A. Cgn'OD.nt. In terms of frequencY. voicel ess consonants are general l y more frequent i n Suc i te than thei r voiced counterParts. The voi ced doubl e stoP, Igbl i s also qui te common . On the other hand . the voiced consonants , Ib/. Id/ . Iv/ . and Izi are rel ati velY ra�a. The reader wi l l not. in chart ( 1 ) , the complete absence of the voiced velar stoP, l iL. Car l son analYse. the SuPYire vel ar �ri �ativ8 as being underl Yin9ly 19/ . It i s possi b l e that the velar �ri cat i ve i n Suci te, lxi, is al so under l Yi ngl y Ig/ . However , at thi s Poi nt in t i me . I have oPted to conti nue usi n9 the symbol Ixl i n the transcriptions. Thi s vel ar fri cat i ve aPPears to be ei ther voiced or voicel ess dePendi n9 on i ts Phonol 09i cal envi ron­ ment. It i s never found i n word i n i t i al Posi tion. ---- .--- - - _._------ ---- -------_ ._- - --_.. _. 13 Bel ow i s a char� o � �he consonan�s �ound i n Slldts . (1) � P plosi ve k n k [ gg) mp [bb) n� [dd) ? Nasal b d mb nd m n I Lateral Fr i caHve � gb nY � '19b I) I)m r h s v z c AHrica�e ndjjl j nj SuPYire, Y w, wY Gl i des a Senu�o d i al ec� c l osel Y rel a�ed to Suci te, has a number o� attested Pal atal i zed consonants. The chart above shows three such consonants. I�y/, /ny/ , and /Wy/l. However , i n Suci te , i t aPPears that this � eature i s i n the Process o� being l ost . In the exampl es bel ow, the palatal i zed consonant versi on o� �he word �reel Y al ternates wi th a non-pal atal i zed versi on . (2) a. l t x o b . mE x i N l Y�x o ' to get ol d , N mYaK� ' name ' i ncomPl et i ve aspect ' A more thorough anal Ysi s wi l l have to be made be�ore the status o� palatal consonants can be determi ned. 14 The sYmbol Icl represents the al veoPal atal affri cate [ t J. 1. Pr.-na.aliz.d Con.cnant. and eon.anant clu.t.r. There are numerous envi ronments where a nasal consonant can be i mmed i atel Y fol l owed by a stoP or by an affricate. In al l cases. however, these consonant clusters are found in morpheme i ni t i al Posi t i on . In some i nstances, these nasal consonants are der i ved . For axamPI •• the future tense verbal parti c l e consi sts of sf + n . When a verb fol l ows i mmed i atel Y after this verbal particle, the nasal causes Prenasal i zati on of stoPs and affri cates, and voi c i ng of fri cat i ves as shown bel ow i n ( 3) 2. ( 3) a. Pan ' I wi l l corne ' corne nd� st mPan [bban l b . k �r! go nd� sf nk.lr! [gg.lrf J ' I wi l l 90 ' c. B eat nd� sf ndi ' I wi l l eat d . ja to be abl e nd� sf nj.l ' I can ' e. gba drink nd� st mgba ' I wi l l dri n k ' f . yld get uP nd� sf njld • I wi l l 'aet uP ' g. S:J buY nd� st Z:J • I wi l l buy ' Some nouns wi th a nasal i n i t i al consonant cl uster aPPear to be der i ved from veros whi ch do not have an i n i t i al nasal . The specul at i on then . i s that Pref i x i ng a nasal to a verb in effect, nomi nal i zes i t , as is shown in the exampl e bel ow. (4 ) a. It ----- > ndl eat -----> f ood There are however , Pre-nasal i zed nouns and verbs for whi ch there i s no ---- - ---. - -- .. ---- . 15 evidence that the i n i ti al nasal i s derived. Some at these are l i sted below. ( 5 ) a . mblJKf ' oPen, reveal ' b . mp a. 'protec t , detend' c . mPill 6 ' hi l l ' d . nj i de ' tongue ' e. n k�nl a ' t ooth ' What I cal l Pre-nasal i zed voi c�l .. s consonant s are nat phoneti cal lY real i zed as such . In the consonant chart (1 ) , in ( 3a and b ) , and in ( 5b ,c , and .) above, their Phoneti c transcr i pti on i s gi ven as a d oubl e voi ced consonant s. SuPYire cognates at these words actual l Y Possess a Phoneti c pre-nasalized consonant . However, i n Suci te, these Pre-nasal i zed voicel ess consonan t s soun d l i ke a del aYed rel ease voiced stoP. The nasal not bei ng Pronounced , one has the imPressi on that i t i s SNal l awed or that the ai rst ream th rough the nasal Passages i s somehow b l ocked , the Pressure bui l d s uP in the mouth and i s rel eased onl Y atter the art i cu l at i on o-F the stoP, gi vi ng the e-Ftect at a torti s voi ced stoP or a l engthened or gemi nate voiced consonant . They contrast wi �h voi ced stoPs a5 �el! as with voicel ess and Pre-nasal i zed voiced stops. Below are two set s at mi ni mal Pai r s to il l u str ate thi s tact . (6) bi pill � 2. ' theY ' " body ' mbi ' Hour ' mPill 8 [bb il1 61 ' hi l l ' strns and Consonants Consonant s alternate i n such a waY that wh i l e the majori ty ai consonant s are al l owed in morPheme ini t i al Posi tion , these same consonants are nat al lowed to begi n sub sequent sYl l ab l e s at the same morPheme. Al l Pre-nasal consonant s, most .---- - .- 16 stops such as P, t , k. d . gb . the a��ri cates �Y, s , v , and z. c and j , and most �ri cati ves, � , ar a restri ct ed to morPheme i n i t i al Posi tion . I n (7) are some examPles o� these consonants in word i ni tial Posi t i on . The nouns in (8) are comPosed o� two noun rootsJ thus the above consonants can be �ound at the beginning o� each noun root. (7) a. ��x� ' Yam, i nde� i ni te ' b . �a-xo ' to l engthen , i ncompl eti ve asPect ' c · �..{ ' to shake, comPl .t i ve asPect ' (8) a · a-d ' Person ' b . a-ki ' goat ' c · ka-�� ' wi nd ' On the other hand. there i s a verY smal l set o� consonants that are a ll owed onl Y i n non-morpheme i n i t i al Position . These are �, �, and the glottal stoP, Z. Below are a �ew examPles. (9) a. n9u e,a-xo ' smoke ' ' l oad ' (�rom tux a ' to carrY ' ) b . i: u KJ1-r� c . k .ln :a 1, ' vi l l age ' Certain co nsonants such as nasal s. the voiced b i l ab i al , �, and the l at eral . can be �o�nd i n both Posi t i ons: (10) a. lQ70 b . a.7al a c . !l{a ' water , i nda4i n i te ' twi st, wind , wri the ' twi n ' ' owner ' �a� ' tel l a l i a' ' I ather , �oam ' The nasal vel ar , �, is most �requentlY �ound i n non-morpheme ini tial ---- ---- ---- ------ ---- -- --- --- -_. - 17 Posi t i on . It Noul d b e conceivab l e t o suggest that those consonants al l o Ned on l Y i n non-morpheme i ni t i al Position are in compl ementarY di stri but ion Ni th certai n morPheme i n i t i al consonants. For examPle. Posi t i on. alternates Ni th Li kewise. �. i. Nhi ch occurs i n morPheme i n i t i al which i s never � ound i n morpheme i n i t i al Posi t i on . k maY al ternate Ni th � or 2. One maY a l so suggest that stress pl aYs a rol e i n the d i stri buti on o� consonants. stress Pl aced on the � i rst sYl l ab l e o� the morPheme wou l d then al l a N �or the mul t i pl i c i ty o� consonants. Nhi l e consonants �ound in Neak l y stressed Posi t i on s are more l i mi ted and tend to b e the weaker versi ons � thei r stressed counterParts. Evi dence that stress i s l ocated i n word i n i t i al Posi t i on can be found i n the behavi our o� the noun cl ass cl i t i c . whi ch . as Ni l l be seen l ater in ChaPter 6 . c l i t i ci zes to i ts surroundi ng environmen t . When the n oun cl ass cl i t i c . �. i s � ound i n sentence i n i t i al Posi tion. as can be seen bel oN i n ( lla ) . i t Possesses a consonant al l oNed onl Y i n morpheme i n i t i al Posi t i on . However , when the c l i t i c i s Preceded by a verbal Parti c l e i n the sentence as seen in ( llb ) . i ts stoP consonant i s � r i cat i vi zed. Nhi ch wou l d be expected i f i t s Pos i t i on i s nON consi dered to be unstressed. ( 1 1) a. stressed : ka t�x� ' carrY i t ! ' i t carrY b . unstressed: Na� � t�x� ' he carri ed i t ' he-TA i t carrY It must be noted that the cl i ti c is the onl Y morpheme that al ternates betNeen ----_ ._-_._------- ------ --------- -.-. . _._.__. ----- 19 stressed and unstressed consonants. One mi 9ht be temP�ed �o sU99es� tha� i ns �ead of stress bei n9 a factor . the i ntervocal ic environment of stoP consonants l eads to their �ricat i vi za�i on . However. i � one �akes a second l ook at comPound nouns (Si. i t i s seen that the consonant of the second root of the compound does not undergo fricat i v izat i on even �hough i t . too. i s i n i ntervocal i c posi t i on . Thus. the hYPothesi s that stress i s a major factor of these particular consonant alternati ons is put forward . The matter of str ess wi l l be consi dered again brief l y in the discus ­ sion o� vowel s. B. ygw.l. S�c i�' has 7 vowel s. al l of whi ch are verY common to West Afri can lan9uages. (�2) i u [al a a Al l vowel s can be l engthened and nasal i zed . A l en9�hened vowel i s tran ­ scri bed wi th two vowel s: VV. Nasal i zed vowel s are wri tten as Vn . The on ' o t the nasal ized vowel . however . can be someti mes confused wi th �he on ' in InYI or i n anY Pre-nasal i zed consonant . For exampl e th e noun rlrJa ' rai ns ' i s wri tten i n our transcript i on as z�nya. Wri tten as such . i t i s i mPossib l e t o know whether 19 the ' n ' nasal i zes the Previ ous vowel , [ z�-y31 or whether the ' n ' al ong with the adjacent ' y' �orm a Palatal nasal , [zA-nY81. There4ore, to avoid an ambi guous readi ng o� ily 3 , an aPo$trophe i s i nserted to i ndicate sYl l able divi si on : zAn ' y3. In ad di t i on , when a ' n ' i s Preceded by a vowel and fol lowed b y certain stoP consonan ts, ambigu i ty ari ses in trYing to discern whether the ' n ' bel ongs to the preced i ng vowel or the �ol l owi ng consonant. The word kAkonk6 c ou l d be phonet i cal l Y ei ther [ kAkok 61 or [kAkogg61. The �ormer is correct . To disti n ­ guish Pr e-na sal i zad consonant. from nasal i zed vowel s a superscript ' :: wi l l be used �or the �ormer onl Y in the si tuat i ons where ambigu ity ari ses. Therefore, i n ( 1 3a ) no suPersci pt ' "' i . wri tten �or the nasal i zad vowel , wh ereas i n (b ) , there i s no nasal i zed vowel but rather a Pre-nasal i zed consonant . In cases where th ere i s no ambi gui ty , the ' n ' wi l l not b e i n superscri pt, as .is sho wn i n ( 1 3c) . ( 13) a. kAk gok ' [k 1kok 6] ' col a, cough ' b. ka :kur� [ ka99ur�] 'fi ve ' c · nkanl 8 ' tooth ' [9g113] The reader wi l l al so note the Presence of the schwa i n square brackets i n the vowel chart. The schwa i s found i n two sPec i f i c envi ronments. I n both i nstances, it aPPears that the schwa is in compl ementarY distr i buti on wi th some other vowel . The f i rst vowel of anY morpheme can be anY � the ei ght vowel s 9i ven i n the chart above. However, the vowel Ii i i s qui te rare i n thi s Posi tion , whi l e the schwa i s qui te frequent , leading one to the susPici on that the schwa maY be an underlYi ng Ii i whi ch i s central i zed i n certai n envi ronmen ts. Wh ile data i s not suff i ci ent to COmPl ete an i n vest i gati on on thi s issue, the i tems gi ven bel ow -----. ------. ---- 20 seem to i ndicate that a h i gh front vowel (as the f i rst vowel o� the morpheme) tends to remain �ronted when l ocated between two palatal or dental tyPe consonants ( 1 4) , whi l e i t tends to be central i zed el sewhere ( 1 5) : ( 14) a· . H..de ' breast ' b. nc ile ' bal aPhone ' c. t id. ' creePi ng vi ne ' ' goat ' hr. ' excrement ' b . s lea ' Parson ' t abe ' medicine ' c . b .a ' seed ' tak e ' tree-DEF ' ( 1 5) a. s �i However , even i � we cou l d explai n these aPParent eXcePti on s by Phonol ogi cal condition i ng by adjacent consonants, we are sti l l faced wi th the fal lowing near mi nimal Pair contrasti ng � and ( 1 6) a . f al. b . f lle •• ' aPProach ' ' beat (a f l oor ) ' Phonetical l y, the l atter i s nat a l ang vowel . Whether underl Ying vowel l ength has something to do wi th thi s contrasti ve Pai r wi l l be l eft �or anal Ysis at a l ater date. For the t i me bei ng, schwas have been retained i n transcri pt i on . The second envi ronment i n which schwas are found i s i n the second or third sYl lab l e at a morpheme where the sYl l able concerned i s nat i n word � i nal Posi tion . Consi der the examPles bel ow: ( 1 7) a. b�nl �o ' to l engthen. incomPleti ve asPect ' b. kap�?�.Id ' f i ngers ' c. feramc ' urine ' d. � aJ.'x a ' rack ' - �-- ----------. ----- 21 The examPl e i n (a> consists o� a s i mpl e verb wi th an incomPl et i ve su��i x . The �irst sYl l �b l e contains the vowel 10/ . The second vowel i s a schwa. It i s i n the second sYl l ab l e o� the morpheme and i t i s not i n word �inal Posi t i on . The third vowel is not a schwa and i t i s i n word � i nal Position . The examPl e i n ( b > has two nomi nal morPhemes along wi th a nomi nal su� � i x . Nei ther the �i rst nor the second vowel i s � schwa - both are the � i rst vowels o� thei r resPecti ve morPhemes. The thi rd vowel i s a schwa and i t i s the second vowel o� the second marphemQ �d i s not i n � i nal Posi t i on � the word. In al l the other examPles. the same pattern is repeated; there is always a schwa when the vowel is in non ­ w ard f i nal and non-morPheme initi al Position . The choice o� vowel s Precedi ng and �ol l owing the schwa does not seem to be a factor for schw� format i on 3• There are at l east two possible an �l Yses for the schwa in this Posi t i on . The fi rst i s to su9gest that i t i s underl Yingl y Ii i . as was suggested for schwas i n morPheme initial Posi t i on . However . since the structural envi ronment seems to b . such a cruci al f �ctor for schwa formation . i t maY al so be suggested that the schwa could be underlYingl y anY vowel found in thi s Posi t i on . I n order to Pursue thi s argument . i t is necessarY to take a l ook at the general distri but i on � vowel s in a simpl e. non-compl ex word. It aPPears that e�ch morPheme i s sPeci� ied far � si ngle vowel regardl ess of the number of sYl l ab l es it Possesses. If a morPheme has two sYl lables. the fol l owi ng vowel patterns emerge: • Push , ( 18) i f !l ' ' beat (a f l oor > a hIe ' approach ' e gb are ' be short 'meet tagether . gr OUP ' E P€r € ' sel l ' 'di l ute. draw out ' a b &?& ' accuse' u fulo 22 Al thaugh the abave are al l verbs4, the .ame Patterns can a l sa be faund far nauns· If a marpheme has th ree sYl l abl e., a si mi l ar, but sl i ghtly al tered Pattern &marges, t h i s t i me i nval vi ng a schwa. ( 19) i na verbs a Y�r�K� ' rai se, strai ghten e Y,,"a')t • staP, t ransi ti ve ' ' bather, heat ' u kc1ra"� a t6rax 6 ' stumble, bumP i nta ' ' send. accamPanY ' ' push raughly, jastl e ' ' 9avern . t ransl ate ' In each case af the examPles i n ( 19) , the med i al cansanant i s a schwa. Phanet­ i cal l Y t h i s parti cu l ar vawel i s bath reduced i n l ength and cent ral i zed . Because bath the vowel Preced i ng and the vawel fal l owing the schwa share same af the same features. ane maY be l ed ta the canclusian that underlYi ngl y . the schwa al sa shares these same features. Th i s underl Ying vawel then , i s reduced ta a schwa when the structural candi ti ans �ar schwa farmati an are met . In the discussian an cansanants. i t was suggested that st ress accurred an the f i rst syl l ab l e af each marPheme. It maY also be cancei vab l e ta suggest that the infarmatian concern i ng vawel features i s sPec i f i ed far the � i rst vawel af the morPheme whi l e subsequent vawels adapt these features, which then are mad i f i ed thraugh ru l es such as vawel reduct i an and central i zati an . Rul es �ar thi s behavi au r wi l l nat be farmulated. hawever, because the i ssue af vawel behaviaur i s a l i t t l e mare campl icated than i s Presented here, and a mare thoraugh anal Ysi s needs ta be Presented . Suff i ce it ta saY, hawever, that schwas wi l l be retained in the arthagraphy for the Presentatian af data in thi s thesi s. 23 c. Tgnw On the sur�ace. there are three l evel s � tone i n Suci te. Hi gh ( � ) . Mid (unmarked ) . and Law ( ' ) . Gl i des �orming anY combinat i on o� t�ese three l evel s can be �ound i n the lan9uage. However . the most common are High-Low (HL) and Mi d-Low <ML) . The i ssue o� underl Ying tones is rather compl i cated and wi l l be the �ocus o� attenti on i n the �ol l owin9 chaPters. Tonal rul es oPerate within words. and between components of a verb phrase or a noun phrase. Noun cl ass cl i t i cs. Postpositions. verbs . and verbal parti c l es are especial l y suscept i b l e to the tonal environment i mmed i atel Y precedi n9 them. The Possib l e combi nat i ons of tone withi n a three sYl l ab l e noun <disre9ardi n9 the number of morphemes therei n ) are quite numerous . as shown bel ow i n ( 20) . As wi l l be seen l ater . the d i stribut i on o f tones w i th i n i ndividual morpnemes i s l i mi ted . In add it i on , tonal rul es do b l ock Possi bi l i ti es �or certain combi nati ons whi l e en9anderin9 others. (20) <Number in an9l e brackets i ndi cates the number of tonal l Y i dentical three sYl l ab l e i nde� i n i te sin9ular nouns �ound i n a l i st of approx i matel y 362 threesYl l ab l e nouns) : 1 . L-L-L k4s�)(�'" war 2. L-L-M k45�ke war-DEF 3. L-L-HL k4Y � 1 @ compound 4. L-L-H k4YUni! comPound-DEF 5 . L-M-L �na.i Packrat <1> 6. L-M-M k.ico)(� encl osure <3 1> <12> <1 3 > ·Unl ess otherwi se marked , al l nouns 9i ven here are i n the i ndef i ni te form. Those marked DEF are def i n i te nouns. ----- . . -- - - . .- ----_ ._--- ---- ----- 24 7. L-M-H k.icok e encl o5ure-DEF 8. L-H-L mp�de)( . rib <2> 9. L-H-M k.im6l u ant: < 4> 1 0 . L-H-HL nY.ip!S l g L-H-H bundl e o t 9rass < 4) 0 <2) 1 1 . M-L-L nan.il � + i sh (sPeci es> 1 2 . M-L-M toa!,)e chi et . owner-DEF 1 3 . M-L-H zanl ��n' kidneY-DEF 1 4 . M-L-HL zan l �l e ki dney 1 5. ML-L-HL canb�!')e umbrel l a 1 6 . M-M-L 9bunna!')� wal l <3) 17. 1'1-1'1-1'1 tunkYaxl: di arrhea <9) 18. M-M-H tunkyak e di arrhea-DEF 19. M-M-HL ku ru9ba matted overhan9 M-H-L <4> 0 20 . M-H-M 9bab'ne roam of a house-DEF 21 . MH-M-M I aal a!')e butterf l y <1> M-H-H 0 22. M-H-HL 9bab'1 @ room of a house <2> 23. H-L-L PC1ny�x� c l ay <2> 24. H-L-M PC1nyt ke cl aY-DEF 25. H-L-H UnyMge st:ump of a tree-DEF 26. H-L-HL Uny�!')e st:ump at a t:ree <1> 27. H-M-L t'Pada beam < 4> 29. H-ML-M sa." t QI')E quai l -DEF 29. H-M-M i4me!')E. dew ------ .�- --- .�---- < 3> 25 30. H-M-H kamel')g, dew-DEF 31 . H-H-L n'H U t} buHocks 32. H-H-M �a.Uke rock-DEF 33. H-H-ML ma.?anja beard <1> di vi ner < 7> H-H-H 34. H-H-HL <1> 0 c'-f61 � I I I . MORPHOLOGV AND SYNTAX A . �Nqun. As i s the case wi th manY Ni ger Can90 l anguages. Senufo nouns and Pronouns are marked �or c l ass. Thi s c l ass marker is su�fixed to the noun. The cl ass marker not onl Y conveYs cl ass a-ff i l i at i on but also def i ni teness and Pl ural i ty. The ei ght noun cl asses have been grouped into -f i ve genderse. whi ch inc l ude three si ngular/pl ural Pairs and two mass/col l ecti ve noun cl asses. Each cl ass has twa types o-f suf-f i x i ng: Indefi ni te and De-f inite. Below is an examp le of each tYPe for Gender 2 (Cl asses 3/.) : (2 1 ) a. 9ba-xa ' house ' Inde.fi ni t. (Si ngu lar ) CIa•• 3 b . gba-k' ' the house ' Def i nite (Singul ar ) Class 3 c · gba-Ya ' houses ' d . gba-nY€ Indef i n i te ( Pl ural ) Cl ass 4 ' the houses ' De� i n i te (Pl ural ) Cl ass 4 Bel ow i s a noun chart showi ng the rePresentati ve indefi n i te and de.f i n i te forms .for each cl ass. Suf f i xes are underl ined . - - -- ---- ---- 26 (22 ) Gender Cl ass 1 . ' wi ' CI . 1 Engl i sh Indetini te Det i n i te gba 9ba-!lR ' ri ver ' gba-w gba-i::bi ' ri vers ' 2 . ' ki ' CI . 3 gba-u. gba-Jsj ' house ' cl . 4 gba-� 9ba-an. ' houses ' 3. ' I i ' CI . 5 gba-U. gba-A::o.a ' torehead ' gb.-:al.i. gb.-lt\:Id. ' toreheads ' 4. ' ti ' CI . 7 su-� sil-ia ' mai n d i sh ' 5. ' Pi ' CI . S sa-1iIL si-m::b& ' oi l ' CI . 2 CI . 6 Al though these nouns are rePresentat i ve at their resPecti ve cl asses, there are a substantial number at Phonol ogi cal al ternati ons tor each sutf i x , whi ch wi l l be descri bed in Chapter 3 . The complex tonal nature at nouns wi l l be the toPi c of di scussion in ChaPters 3 and 5. B. PrMPUD. Each noun c l ass has a general . emPhat i c , Part i t i ve , interrogati ve . demonstrat i ve, and rel ati ve Pronoun. The chart bel ow provides a qui c k view of the shapes of some of these Pronouns: (23) Cl ass 1 Cl i t i c wu EmPhati c Parti tive 2 3 4 5 6 7 S bi ka Yi la ki ta ba wur� Per� kar� Yir� lara kera tar� Par� w� pi k� Y� U ki U p� Demonstrati ve n9' mP! n k' nj! nd' nk{ nU mpa Interrogati ve n9a mPi n ka nj i nda nki nta mPa ---- - --- - -- - - - --- ---- - ---- ------- - -- --- 27 Bel ow are exampl es of each Pronoun type. It must be noted that Pronouns ( i n Parti cu l ar . the emphati c and the demonstrati ve > can be understood best onl Y i n context of the d iscourse i n wh i ch they are used. (24) Cl i ti c a. m (functi ons as a Possessi ve. Ref l ex i ve o r General Pronou n ) ' he i s her e ' nYe nanan b . waa !iW. w" he h i m saw c · waa � 9b�l l d. m ' he saw h i m ' h. h i mself i njured ' he i njured h i msel f ' ' hi s meat ' kaUe (2S) EmPhati c Cl i t i c < al so Referenti al ) a. � wi ' i t ' s him ' b . � ka�te ' hi s ( ref ) meat ' � Ya c . Ng� but e£:i jub� k�d"e ctn min theY-E Neg-T these-E words ' mean i ng know not ' But they d i d not understand the meaning of these words ' (26) Parti ti ve a. m�l �r")e � ' some r i ce ' b . waa � we, ' he saw some ' c . c�"e � ci nan� ' Some chi l d was here ' (27) Demonst�at i va a. nda� � S::J I thi s bought b . a.ta kaUe d. ndi Yi ' 1 bou�"t thi s ' ' th i s meat ' cen aki mo Ya mbf r f mEn I Ne9-T know what You T think Neg ' I do not know what YOU are thi nking ' (28) �nterrogati ve a. aii tun wu y� ----- - ----- - -- -- - - - ------- ' who i s i t? ' 28 There i s also a cl ass i dent i � i er , which sur�aces i n phrases, such as the �ol l awi ng: (29) a. nAA � ' he ' s a man ' b . nUI)e � ' he ' s the man ' c . n1maa bi. ' theY d . n1maab! b1 are men ' ' they ' re the men ' The tone o� Noun cl ass cl i t i cs i s qui te var iable . In i sol at i on or i n sentence i n i t i al Posi t i on , they are mi d tan •• Otherwi se, they maY be High , Mi d , or Low depending on the tonal environment . (See p . 33 �or ex amPl es o� tone variabi l i ty and Chapter 6 for a tonal anal Ysis o� the noun cl ass cl i t ic . ) The other pronouns are more stab l e tonal l y . P.,.sanal Pranoun. Bel ow i s a chart of. the Personal Pronouns. (30) wllri , w • we ' ' YoU , S9 . ' Y� ri . yl ' YOU, pl . ' ' he, she, i t ' bi ' theY ' -'de, ni. • mo, ma wu I' The thi rd Person Pronouns are the Cl asses 1 and 2 Pronouns. + Sllclt6, as wel l as other Senuf o l an9Uages, creates a l arge number of words ++ by comPoundi ng two simple nouns or by nomi nal i z i ng a Noun In �act , the l atter i s qui te productive . (31 ) a. k1-1�-x� thing + water cl . Below are examPles of each . ' shower room ' b . ta-wa-xa PI .ce + be drY c . t'-Pa-dA tree + spread across + cl . -_ ----- - _ ._._ .. _-_._---. cl . Verb eXPression . ' desert , drY pl ace ' ' beam ' 29 In these examPl es. the noun cl ass marker is tYPi cal ly at the end o� the ward . Noun + Adj ecti ve constructions are essenti al l Y the nominal i z at i on o� the Noun + Verb and oPerate in the same waY as the examPles above. There seem to be verY �ew real adjecti ves i n s�c� t'. The concepts normal l y handl ed by adjecti ves are �ound in stati ve verbs. as seen bel ow . (32) a. 9ba-kaA I E C the house is ol d ' C the old hause ' b . 9ba-l €-ke The adjecti ve c 90od ' �aY be one o� the rare true adj ecti ves. It cannot be accePted as ei ther a noun o r a verb i n isolat i on . It i s �ound onl Y i n adject­ i val or adverbi al Posi tion . (33) a . *kaa ce�e b . *cef)e wi C i t is 9oodness ' c . p�n-C�I)I\ d . waa P�r€ C i t is 9Dod ' c a 900d dog ' c�n-mi c he sold wel l ' In Noun + Number constructi ons. each consti tuent retai ns its awn noun c l ass. Numbers bel ong to the c wi ' 9ender . i . e • • Cl asses 1 /2 . As an i nde� i n i te <Class 1 ) . i t possesses no su� � i x (35a) whi l e the noun wh ich i t modi � i es Possesses the i ndef i ni te Cl ass 4 suffi x . I n ( b ) . both the noun and the number acquire the de� ini te su� � i x o� their resPecti ve noun cl asses. (34) a. gba-Yi suunn{ b . 9ba-nyt s��nn�-'" twa houses. INDEF twa houses. DEF However, i n Noun + Ordi nal Constructions . there i s onl Y one noun cl ass marker . whi ch i s l ocated in ward t i nal Posi tion . cl ass of the noun . ----- - -- - - -- - Cl ass i s determi ned b y the 30 (35) a. 9ba-suun-��-ke ' the second house ' house-two-ard-cl-DEF b. Ye-suun-�9�-n@ ' the second year ' Year-two-ord-cl -DEF In gen i t i ve construct i ons, consti tuents retain thei r awn noun cl ass markers. In the examPl e bel ow, the Possessi ve noun is i n Cl ass 1 , whi l e the second noun i s in Cl ass 3. ' the man ' s hause ' Tonal l y, a string of nominal el ements al on9 with thei r suffi xes oPerates as a uni t . Each consti tuent i s sensi ti ve t o tonal rules, wh i ch wi l l b e di scussed l ater i n greater detai l . Below are a few examPl es of tonal perturbati ons wi thin the Noun Phrase. The wards an the l eft side of the arrow shaw the tane of the i ndi vi dual wards as gi ven i n ci tati on f arm, wh i l e the wards on the r i 9ht side of the arrow show what haPPens to the tone when the wards are grouPed together i n noun Phrases. (37) a. nda kacUe --> nd� kU. te ' my meat ' b. 9baYi sMnn� ---> 9baYi suunnf ' twa houses ' c . I �Y� s��ni ---> l uY� s�unni ' two creeks ' d . ntara + fol� ---> ntaraf�l� ' l and chi ef ' e· mo foB ---> mo f61� ' Your chi ef ' ' The Noun Phrase i tseH is nat sensi t i ve to the tenal eni vi ronment surroundi ng it 401i$$ the Noun Phrase be9i ns wi th a Noun Cl ass cl i t i c . di scussed bri ef l y on p . 34 and i n detai l i n Chapter 6. ----- -- _ . _ . ._ . _--- ------ Thi s si tuat i on i s 31 D. y..b, The basi c verb i s obl i gatori l y marked �or asPect , either cOmPl eti ve (per�ective, singul ar ) or i ncomp l et i ve (continuous, i mPer�ecti ve, pl ural ) . Sandor-Samuel ( 1971 ) suggested a si ngu l ar - pl ural d i st i ncti on � or the two �orms si nce the cOmPlet i ve descr ibes one comPl ete, � i n i shed acti on whereas the incomPl et i ve descri bes a Pl ural i t y or rePet i ti on o� act i ons. In addi t i on , he al so poi nted out ( and th i s is true � or S�c� t') that the ' Plural ' � orms o� the noun and verb are Phonol ogical l Y simi l ar . There ·also aPPears t o be some transi t i vi z ing and act i vi z i r.9 suH i xes , but the semanti c l oad o� verb su� � i xes has not yet been studi ed . The comPl et i ve form of the verb aPPears t o have a zero morPheme. From the basi s o� thi s �orm, the i ncomPl et i ve asPect i s der i ved . Tonal l y, an i ncompl eti ve su� � i x vari es according to th. tone o� the root . Segmental l y , i t can have a vari ety at shaPes. The most common are c _i ' . ' -ri ' , c -x o ' , c -ni ' . and ' -I i ' . Same of these suf � i x es can have Phoneti c variants, to b. seen in Chapter 2 . There are al so a �ew verbs whi ch do nat �it i nto anY o� these cate9ori es. Selow are Slome examPl es showi ng these var i ous su� � i x es : (38) Compl eti ve 9b�x� gb�d -i gather t0gether- Cy' cYe-xO f"e�use; kci.n ka.-n1 boi l y� y(i-l f steal c&) c6-d catch, grab The tonal nature o� verbs i s a toPic reserved �or di scussi on in Chapter 2 . ---- - . -- - - ._ -- 32 The basi c verb phrase consi sts of a verbal Parti cle and a verb . The verbal part i c l e contai ns i n�ormat i on concerni ng tense, mode, and Al so asPect . As wi th other Ni ger-tongo l an9UageS. modal Particles tend to have been verbs hi stori cal l y . Bel ow are a � ew examPl es o� tense and modal mar kers. (39) a. wu A Pan --> waA Pan ' he came ' b. wu na Pan --> wu na pAn ' he has came ' c · wu 51 ba Pan -> wu s1 ba pAn ' he wi l l come ' he TA M come d . wu 51 Pan e· wu 51 he TA ' he wi l l came ' pan ka � i n i sh come � . wu Ya ma. g. tof U sf da he T ' he is comi ng ' m a. ' he wi l l be comi ng ' come-i ncomPI . A h . wu caa m he T�M ' he wi l l end uP coming ' ' he was cDllli n9 ' come-i ncompl . Seri al verbs or verbal chai ns are qu i te common i n s�clt" just as they are i n manY w�st Afri can l anguages . I n the examPl es bel ow, the two verbs are separated by a tYPe of connecti ve whi ch coal esces to the f i nal vowel o� the � i rst verb . (40) a. waa wu l �-� he-T i t take- Pan came b . waa p�ra.-a k� he-T sel l - � i n i sh ------ - - - - ' he brought i t ' -- - ------- ' he � i n i 5hed sel l ing ' 33 Li ke the noun phrase, the verb Phrase al so operates as a un i t tonal l y . I� the verbal Part i c l e and the verb are adjacent to each other . the tone � the Part i c l e a��ects the tone o� the verb . Th i s is i l l ustrated in the examPl es below. (41 ) a. wu � Pan --> wa� pan ' he came ' b . wu na Pan --> wu na PAn ' he has come ' c . wu � kar! --> waa k�r! ' he went ' d. wu ni. kid ' he has 90ne ' Unl i ke the noun Phrase. whi ch . as a uni t . i s not a��ected by the tonal environ­ ment o� the �Qnsti tuents precedi n9 or �ol l owin9 i t . the verb Phrase i s sensi ­ t i ve tonal l Y to the noun Phrase whi ch Precedes. See (4�) �or examPles o� tonal var i at i ons on the verb Phrase. as wel l as Chapter 4 �or a detai l ed anal Ysi s o� these var i at i ons. E. Ibm Adverbi•l Phr••• The adverbi al Phrase i s l ocated post verbal l Y . and consi sts � a Noun Phrase + Postposi tion or Pronoun + Postposi t i on . The tonal ru l es that operate between these two consti tuents are verY si mi l .ar to the rules that oPrah on the verb Phrase. Bel ow are a � ew examp l es : (42) a· wu nYe wu kAtgxg ' he i s behind h i m ' b . wu nYe mg k.tgxg ' he i s blhind ygy ' c · ka nYe wy Hin ' i t i s b" i dp h i m ' d . k a nYe mg t�D ' it is ----- - - -- -- - - - - ---- ----- biliii gi �gy ' 34 B. Th- Sent."C_ We have al readY �al ked abou� �he naun Phrase and �he verb phrase as sePara�e ent i ti es . In �his secti on wa wi l l exami ne �heir l ocat i on i n �he si mpl e sentence and bri e� I Y d i scuss tonal phen�Mena that oPerate across Phrasal boundari es. The basic word order i s out l i ned below: (43) Subject + Verbal Parti c l e + Obj ect + Verb Phrase + Adverb Phrase S CNP) O CNP ) V. P. Advp (NP+PP) VP Tonal l y, nomi nal + verbal el ements are grOUped �ogether as tonal uni �s . A verbal el ement cannot be attected by the tone ot a Preced i ng nomi nal el ement . howevar . The l i n k i ng i n the i l l ustrati on bel ow demonstrates tonal sensi t i vi t y whi l e the square brackets rePresent b l ocks t o the aPpl i cati on a t tonal rul es . (44) [Noun P V. P . ] [Noun P Verb P] [Noun P Postposi t i on ] Tha tone at verbal Particl es i s at�ected b y the tone at the noun subject. I n add i t i on . the tone o� the verb i s sensi t i ve t o the tonal nature o� anY word that precedes i t whether that word i s verbal or nominal . The examP l es bel ow show var i ous tonal variations o� the underl i ned words. ( 45) a. lNd� ----- ..-- - I b. c. d. -- �1.l tut w"] [mo You l ook at at] [mo *] man tut l ook at [NU S1' ] man -Fut [NU n4 kad ] man Past l eave [NAA _ . You t wA · 1 wi l l l ook at yOU ' id&] " A man wi l l l ook at yOU ' " A man wi l l l ook at some ' some l ook at " A man has · gone ' 35 e· IlA kid ] past l eave (Nd � I t I have gone ' ( I ' m l eavi ng) As ment i oned earl i er, a Noun Phrase is general l y nat sensi t i ve to the surround­ i ng tonal envi ronment. However , i� an NP begi ns wi th a noun cl ass cl i t i c . the cl i ti c is sensi ti ve to the tone at anything whi ch precedes and as wi th anY ather nou n . can also attect the tone at anythi ng wh i ch tal lows. as shawn bel ow . (46) cl i ti c V. P. cl i t i c Verb P cl i t i c Postposi tion Below are exampl es whi ch i l l ustrate how the Presence at a noun cl ass c l i t i c or the l ac k thereat i nt l uences tonal changeS on the sentence. (47) a. wu � b . wu n� c . wu � d. � � w" w" kar� wee ' he l ooked at hi m ' -> wu na � wee ' he h.s l ooked at hi m ' -> wa� kiCa w�e -> wu kar� w�e wu kar� w" e. wu � -> waa � we, ' Look at his meat ! ' kari wee -> waa � kara wee ' He l ooked at his meat ' wu -> wu na kar� we' t . wu n� kar� w" g. wu na ' he l ooked at meat ' � kar� w" -> wu na � k,{r,i w" h. wu � wu kan �-�n -> wa� wu kan � i. wu � wu l � �-�n - > waa wO l � �� l � t�ryu-� -> waa � l � t��uO j . wu � k . wu � a � l� � t he has l ooked at meat ' ' he has l ooked at h i s meat ' he gave i t to h i m ' ' he took i t tram h i m ' t he took i t tram the tather ' t��u-6 -> waa � l � � t��uo t he took i t tram h i s tather ' It �, ' meat ' i s i mmed i atel Y preceded by the verbal parti c l e , it does not undergo tonal change. Al l other nouns behave i n this waY. However , i t a noun cl ass cl i t i c is inserted , the tone ------- ----- ------ ------ at kic4 i s also attected . Any tonal changes 36 on the cl t i c can tri gger tonal pertubati ons on the noun Phrase and verob Phrase el ements whi ch fol l ow, as can be seen i n the exampl es above. These phenomena wi l l be d i scussed i n greater detai l i n Chapter 6 . The ne9ati ve sentence i s formed by i nserti ng a negati ve morpheme i mmed i atel Y before the verbal part i c l e and at the end of the sentence. (48) a . b. wu .\ Pan - wu Yi .\ Pan min he N > wa.\ Pan ---- > he came wu Yi Pan min " He d i d not come ' T c ome N c . wu ni. Pan he has come d . wu y� na Pan min ---> wi � na Pan min " he has not come ' As wi l l be seen i n Chapter 4, both Parts of the negat i ve morpheme can be affected by the tone of the Precedi ng word . The negati ve parti c l e . affect the tone of the fol l owing verbal part i c l e , whi l e mtn affects the tone at the Yes-No Questi on Marker. t. Au"tim FormUig" Yes-no questi ons are formed by addi ng " l a ' to anY sentence. ( 49) a . wa� Pan " he came ' b . wa� Pan l a " Di d he come? ' c . waa k.\d " he l eH ' d . waa k.lr! l a " Di d he l eave'? ' e· nda la " Me? ' f . mo l a " You? ' �. can also 37 9 . kalte l A · the meat? ' h . m�B l � ' ri ce? ' As the exampl es above show. the tone o� the Yes-no question marker . lao vari es i n rel at i on to the tone o� the Precedi n9 word . Frontsh i � ti ng takes P l ace �or WH questi ons. A Q marker ' Y� ' i s al so Pl aced at the end o� the sentence. (50) a. San Where he-T � b . San ' Where d i d he 90? ' waa k�r{ Ye wa� Q w6 9b�ri Yi [9b�rt � ] ' Where d i d he meet h i m? ' Where he-T him meat Q ' What did he see? ' c . NYi?a wa� nYa Ye [nY£ � ] What he-T see Q A �ront shi �ted const i tuent i s treated as an i sol ated tonal unit and theretore has no et� ect on the tone o� the const i tuent whi ch �ol l ows. Fronbh i H: i n9 al so takes pl ace qui te �re>1uentl Y in answers to WH quesb ons (51 ) a . San mu Ya see? b . caank6 l ao ndi Ya s6 ' Where are You 90i n9? ' " I ' m 9Oi n9 to the market ' market to. I am 90i n9 IV. SUl'lMARY Thi s chapter has been a sketch o� the basi c structure o� Suci te. Those acquai nted wi th Senu�o l an9uages wi l l have rec09n ized numerous Patterns and structures tYPi cal of Senu� o l anguages. The i nventorY o� Suci te consonants i ncl udes stoPs . � r i cati ves. and l ateral s. 38 as wel l as pre-nasal i zed stoPs and palatal i zed consonants . which awai t further i nvesti gAti on . Seven vowels were i sol ated , whi l e an ei ghth . the schwa. i s tentati vel v consi dered not t o hAve phonol ogi cal status. However . not enough research has vet been done to cl ar i f v thi s Poi nt . Both consonants and vowel s aPPear to Pl av a rol e in signal i ng stress on words. It was ProPosed that the f irst svl l ab l e of each word i s stressed . Thi s f i rst svl l ab l e al l ows for the greatest var i etv of consonants. and i t i s al so the Pos i t i on where a ful l v speci f i ed vowel aPPears to dictate the under l vi ng vowel qual i tv of the vowel s i n the remai n i ng unstressed svl l ables of the word . Agai n . more research wi l l have to be done on the rel ati on between segmental behaviour and stress before anV def i n i te statements can be made. It WAS al so seen that there are three l evel s of tone, as wel l as two fal l i ng tones i n Suci te . In sketching out sentence structure and word order. i t was noted that nomi nal el ements can affect the tone of fol l owing verbal s but that. verbal s cannot affect the tone of nomi nal s. The di scussion concerni ng the toPi cs of under l vi ng tone. the di stri buti on of tones in nouns and verbs. and the rules governi ng tonal changes in the sentence wi l l be deal t wi th i n succeedi ng chapters of this theSi s. NOTES 1 . Whether these consonants are under lvi ngl v Pal at ized or whether thev consi st of a combi nat i on of consonants .i s a matter that must wai t unti l further research is done. 2. See ChaPter 3 for nasal condi ti on i ng of consonants on noun cl ass suf f i xes. 3. Transcri pt i ons in this thesis do not al wavs show a schwa in this Posi t i on . Re9ard l ess of t h e transcri pt i on gi ven . however . a vowel i n morpheme med i al Posi t i on is al wavs reduced i n length and tends to be at least sl i ghtly central ­ i zed . Further i nvesti gati or. w i th a l anguage cansul tant shou l d cl ear uP these discrepancies i n the transcri pt i ons. ----- .-- - . � . - . .--- -.----- ----- 39 4. A si gni � i can� number o� d i sYl l ab i c verbs Possess a � i nal I i I i ns�ead o� harmon i z i ng wi th �he i n i � i al vowel : ct}B hanri �ori ' sPread out ' ' construct . bui l d ' ' gO ou� . aPPear ' s. There has been a quas�i on among �hose s�udYi ng Senu�o l anguages as �o how �o number and organi z e the noun cl ass system. Those worki ng an Cebara (Mi l l s et al > have grouPed si n9Ul ar and pl ural �orms �ogether . cal l i ng each 9rouPi ng a cl ass. Thus. the �i rst �hree ' cl asses ' con�ai n bo�h si ngu l ar and Pl ural �orms and the last twa contain mass and col l ect i ve nouns. Because Mi l l s set the preceden�. a number o� ather peopl e have �ol l owed �hei r lead . Carl son adoP�ed thi s sYs�em o� organi zat i on but renamed the ' cl asses ' . ' Genders ' . Recently, i n a d i scussion wi �h MarY Laughren . she sugges�ed pl ac i ng each si ngu l ar and Pl ural �orm i n i �s awn cl ass and added �hat Jean Cauvi n . who has wr i �ten an Mi nYi an ka. a Senu�o l angUage. uses the same - sYstem. Carl son argues �or hta sYs�am i n saYing tha� there i s nat �he mi x i ng and matchi ng o� noun cl asses as i s � ound in Bantu l an9Uages. In add i � i on . he saYs tha� the s i n9U l ar and Pl ural {orms o� each gender are cl osel Y l i nked and �eel s �hat �he system used shou ld re� l ect thi s i ntimate Pai r i ng � si ngu l ars and Plural s . One advantage o� assi gni ng a cl ass number � or each si ngul ar and each Pl ural woul d be in caPturi ng a cer�ai n general i z at i on . �o be seen in ChaPter 3. tha� goes across �he si ngul ar/pl ural di s�i nction . Instead o� di scussing the si ngular i nde� i ni tes and d� i n i tes seParatel y � rom the Plural s . it might be mare conveni ent to examine the si ngul ars and pl ural s an the same level when di scuss­ ing tonal and segmental behavi our . The anal ysi s woul d then nat be re�erring to si ngular and Pl ural �orms but on l Y �o i nde� i n i te and de� i ni te su� � i xat i on o� cer�ain noun cl asses. For the above reason . I have uti l i zed the Laughren/Cauvin apProach �or anal Ys i s in pre�erence �o Mi l l s ' and Carl son ' s. ChaP�er 1 . however . does re�er to Carl son ' s gender system when i ntroducing Noun Cl asses. 6 . Wi �hin �he contex� o� �h is di sserta� i on . �he term ' noun phrase. narrowl y re�ers to the s�r i ng o� nominal el ements al ong wi�h thei r su� � i xes . 7. The �erm ' verb Phrase ' wi th i n the contex � o� �hi s thesi s i s m.an� t o re�er to a s�ring o� verbal el ements onl Y and no� �o �he broader d� i n i � i on wh i ch al so i ncl udes objects and adverbi al Phrases. -- - - - - - ----- 40 CHAPTER 2 - THE SUCITE VERB I . INTRODUCTION In Chapter 1 , we saw that verbs are marked �or asPect . Segmental l y, the cOmPlet i ve farm of the verb is the basi c stem, and the i ncomPl eti ve suf f i x i s added to that stem. The f i rst Part o� thi s chaPter wi l l Provide a descri pt i ve account of the morphoPhonol ogy of the i ncomPl eti ve su�f i x , based - an a studY of about 300 common Suci te verbs. The root tone o� verbs someti mes changes when an i ncomPl eti ve suf f i x is added . Th i s studY wi l l al so l ook at these tonal changes i n an attemPt to d i scover anY general tendenci es worthy o� rul e formu l at i on . The possi b l e underl Ying tone �or the i ncomPl eti ve su�f i x wi l l also be di scussed and tentati ve rules wi l l be Posi ted , though the anal Ysi s o� a broader range of data i s needed before def i n i t i ve Proposal s can be made. I I . THE COMPLETIVE VERB A. structural Dwtcriptipn The bas i c verb i s comPosed � one, twa, or three sYl l ab l es with a l i mi t of three Tone-beari ng uni ts. The fal l owi ng sYl l ab i c structures can be found an verbs : (1) CV B ' eat ' CVV wee ' l ook ' CVCV cel € ' di vi ne ' CVVCV jc,c,l i ' SClW ' CVCVCV cal 'r')� ' heal , be heal thy ' Any a� these farms can be prenasal i zed i n stem-i n i t i al posi t i on : NCVCVCV ndax'l 6 ----- -- - - - - - _._--- ----- ' smel l ' 41 There i s al so a smal l set o� monosYl lab i c verbs that Possess a di phthon9 i n Phrase f i nal Posi t i on : (2) H il ' sprout ' Paon ' come ' .feo ' run ' ' �eed ' cie ' do ' • jie wash , ' enter ' A d i phthon9 is not as 1 0n9 as a 1 0n9 vowel . and shoul d be considered as a si ngl e Tone-beari n9 uni t . Di phthon9S i mi tate the tonal behavi our o � verbs wi th short vowel s rather than that o� 1 0n9 vowel s . As a resul t . i t seems more feasi b l e to consi d er them as havi n9 one Tone-Bear i n9 unit rather than two. In the exampl es bel ow. al l o � the verbs are Hi 9h tone i n ci tation form. However . when preceded by � ' he-Recent Past ' . the d i sYl l ab i c verb and the verb wi th the 1 0n9 vowel acquire a Low-Hi9h tone. whi l e both the d i phthon9i zed verb and the verb with a short vowel acqui re a simPl e Low tone. (3) Verb structure CVCV cv: CV cvv ptrt • sel l , waa p�rt ' he sol d ' wee ' l ook ' waa w�e ' he l ooked ' l� ' take ' waa B ' he took ' He ' enter ' waa j l � ' he entered ' When a di phthon9i zed verb i s i n Phrase medi al Posi t i on . the second vowel o� the di phthon9 i s l ost . (4) ----- . -- - . - Paon -- ----- wa4 em l a ' di d he s;gma? ' waa £i l a ' di d i t sprgyt? ' wa4 ii. 1 .. ' di d he � i n? ' 42 B. Tpn.l DetcriptipD Re9ardless o� the number o� sYl l ab l es or tone-beari ng uni ts (TBU ' s) i t maY have, a verb i n i sol ati on has a choi ce o� onl Y three di��erent tone patterns: Hi gh , Mid and Low. Underl Yingl y, there are no contour tones o� anY tyPe. The exampl es bel ow show the verbs in the i mPerati ve �orm. (5) 1 TBU 3 TBU ' s 2 TBU ' s a· H l� · take ' P€r€ b. M s:,) · buY ' tuxo • c. L U c show ' t�d cul ur')� C be heal thy ' carrY , �uxari • rummage , · vomi t ' k�r�r')� · goverr. , turn ' • sel l , It apPears then , that onl Y one tone need be sPec i � i ed �or each verb . Th i s tone woul d b e Pl aced on a t i er seParat e �rom the segment , and b e l i nked t o al l o� the TBU ' s o� the verb . Thus, as shown i n the examPl e below� there i s a si ngl e High tone at the autosegmental l evel wh i ch can be l i nked ei ther to a si ngl e TBU or to a mul ti TBU verb . C take , P€r€ V H • sel l , cul Y r')i \ i/ " be heal thY ' H Mi d tone and Low tone verbs can be descr i bed i n the same waY. The under l Y i ng nature of the M i d tone i tsel f wi l l be di scussed l ater i n thi s chaPter . I I I . THE INCOMPLETlVE A. structur.l VERB De.sriPtigD The IncomPl et i ve form of the verb i s deri ved by suf � i x i ng to the basi c ComPlet i ve �orm o� the verb . There are f i ve basi c tYPes of incomplet i ve suf f i x i ng �rom whi ch . I ProPose, other types of suf f i x i ng are der i ved . -- - - - - -- -�----- - -- - -- ------- 43 The � and most papul ar tYPe i s simpl e VOWEL SUFFI XATION. On verb stems wi th a si ngl e TBU. the h i gh vowel suff i x ( l et us suPPose that i t is under l Yi ngl y i ) assi mi l ates to the root vowel i n terms at backness. then the ti nal root vowel is rai sed to the l evel at the sut t i x vowel . root vowel is a then the sut t i x i s lowered to • &. However . i t the There are three excePtions to these descri pti ve rul es. Below are three i l l ustrations at the i mPl ementation at these rules . each tol l owed by a short l i st at verbs whi ch behave i n the same waY. The verbs i n the t i rst column are cOmPl et i ve torms at the i ncompl eti ve ver�s. whi ch are i n the second col umn . (7) t6 + { t6 + d -----> -----> t6 + 6 ' H ow ' - tdd BACKING RAI SING ' draw a bow ' a. !)mo ' -F i nish ' c . k�n ( 8) ' cut ott ' k66n N/A we + { -- we + { ----> wi ! ' l ook ' a. bYe bH ' teed. rai se ' b . tY€n tUn ' H ower ' c . kin kHn ' moan ' -> BACKING RAISING excePt ion : sutt i x l ower i ng and parti al rai si n9 d. U (9) ka + { a. ca ' shaw ' tee ---> ki." ' chew ' da ' l ook t ar . search , ------ -�-. --- ----- LOWERING 44 ., b. ca.n cu.n ' droP, destroY ' c. nYa nY.la � see ' el<cePtions: rai si ng rather than l owering d . gba 9b�u ' dr i n k ' e. ja Hi ' shoot (Pul l tri gger ) ' There i s one smal l categorY of singl e sYl lab l e verbs where vowel rai sing or l oweri ng takes pl ace, but there aPPears to be no suf � i x i ng as such , or if there was, the suff i l< was del eted after the rai si n9 or l ower i ng rul e. ( 10) Vowel Rai sing a. se si ' gi ve bi rth . be born ' b. fo fu ' emi grate ' c. jo Yu ' sPeak. saY ' d. to tu ' fal l ' B e. f. U If ' eat ' U ' weave. brai d ' ( 1 1 ) Vowel l oweri ng a· Pan ma ' come ' b . y i. y a. ' be si c k. ache ' Di sYl l ab i c verb stems wi th two TaU ' s can al so be suff i xed by the Hi 9h front vowel , ::i.. The f i nal vowel of the root i s then total l Y assimi l ated to the suH i x vowel . ( 12) a. saxe ----_ .. _-- si)x � -i ' wai t ' b. tare c . cel t - .- -.---. � di vine ' ---- 45 d. Yexe yex £ -£ " quest i on ' e. f Y�?t h@?i-i " be qui et ' f . gb�x� gb�d -i " gather together ' g. I')ndB I')m� H - £ " sl eep ' h . dd dd-£ " pl ant ' i . fulo fill l -i " Push ' j . gbc\r.\ gbc\r l -i " agree, meet ' In ChaPter 1 , i t was suggested that onl Y the f i rst vowel of a non-comPlex word i s stressed and sPec i f i ed for -Features, whi l e .,y subsequent vOrrlel s deri ved their shape �rom the f i rst vowel . In the examPles of ( 12) , i t appears that when the i ncomPl eti ve suf f i x i s added to the verb stem, the unstressed second vowel adoPts the f eatures of the suf f i x vowel instead of from the stressed root vowel . We have seen the behaviour of vowel suf f i x i ng on single TBU and doubl e TBU verbs. Tri pl e TBU verbs also have their disti nct mode of beh.vi our . Instead of a hi 9h · front suff i x vowel . however , the f i nal vowel of the verb root i s rePl aced �y the high back vowel . � Thi s cate90rY i ncl udes not on l Y verbs with three sYl l ab l es (see 14) but also anY verb with three TBU ' s , such as di syl l ab i c verbs wi th l on g vowel s ( 13) . Doubl e TBU verbs whose second sYl l ab l e begins with a nasal also bel ong t o t h i s grouP ( 15) 1 . ( 13) Long-short verbs a. c�::Jnri d�nr-u " resol ve. choose ' b. ft:Ui. -Fa-iJ -u " bal ance ' c . f U nn� fH nn-u " be c l ea.., ' d . ctl.tl.l� cLltl.l-u " be lch ' -' . - . - -' ._. ._. _--- --- - -- ------- 46 ( 1 4) Short-short-short verbs a. ce7.1t: ce7el -u ' i nsul t , l au9h at ' b . cc1l ',,� cc1U�c1 ' heal , be ,eal thy ' c . fc1xa.-! .fc1x�-c1 ' .fri ghten ' ( 15) Fi nal n.sal sYl l ab l e a. P�n� p�-u ' sPi n ' b . k�,,� k�"-u ' l ather ' The su.f .f i x � l owered to g,. ( 16) a . j ::J::Jx i never occurs a.fter the consonant , �. Rather i t aPPears to be j��x-c ' sharPen ' b . wc1r6x ! w6r'x-6 ' mi x uP, mi stake ' kal�d kiUx-6 ' sPoi l , be sPoi l ed ' c· Th i s subcate90rY of verbs provides evi dence for the f act that Suc i te verbs al l ow, at the most, onl Y three vowel s or three tone-beari ng uni ts on the verb2 • If the i ncomPleti ve suf .f i x i s added , thi s restri c t i on remai ns i n .force. and consequentl y the su.f .f i x repl aces the .f i nal vowel o.f the verb w i th the � suff i x i nstead of addi ng a f i nal vowel . In fact , thi s i s the onl Y tYPe of suf f i x al l owed on tri pl e TBU verbs. At thi s pai nt , I wi l l not di scuss whether the under l Y i n9 f orm of the i ncomPl eti ve vowel suf f i x i s The sRSQnd tYPe of suf f i x i s ( 1 7 ) a. I E ::me U -xo ' be ol d , get ol d ' b . PE n ptn-xo ' be unhappy, d i s9Usted ' CY' CYe-XO ' refuse ' H H -x6 ' be .far , 90 far awaY ' c· d• H H . - .-- -�-------- =i or �. 47 e. Pdl � pt U-xo ' be bi g , get b i g ' t . Y£ ri Ytr�-xo ' counsel (6 of 19) Note that thi s suf f i x can be added to verbs with one or two tone-bear i n9 uni ts. The � suf f i x , �, has been i dent i f i ed on single TBU roots. ( 1 S ) a. yll y6-U ' steal ' b . kll k6-U " di e ' There i s a certai n set of verbs whi ch exh i bi t a � suffi x . From the examPles gi ven bel ow, the reader wi l l note that al l of them have d i sYl l ab i c verb roots and that these roots al l end i n Ii I . ( 1 9) c�r� ------ > c�di ' sneeze ' P�n� --- > P�i ' l ose ' sf l ! ------ > sfdf ' begi n ' tt x f -----> t'd! ' pl ace, Put ' Rather than suggest that � i s a comPletel Y dit terent suf f i x , one maY specul ate that i t i s under l Yi n9ly a =LL suff i x . Two b i ts at evi dence ral l Y i n suPPort of thi s proposal . Fi rst of al l , there are no exampl es of a d i sYl l ab i c verb root acqu i r i ng a =LL suf t i x o n the surface, whi l e monosYl l ab i c roots acqu i r i ng a :d1 suff i x do not ex i st . Thus, i t i s Possi b l e that these two suttix es are i n COmPlementarY d i stri but i on . Secondl y, there is evi dence that � was deri ved trom the del et i on at the t i nal Ii I and the coalescence ot the t i nal root consonant wi th the sut f i x consonant . In ChaPter 3. i t i s seen that the coal escence of � at a noun root and - --- ---------- l ot a certain noun cl ass sutt i x , 48 results i n d.: cere + Ie ---> cede � calabash ' . Al though the mechan- i cs o� coal escence on verbal su�� i xes maY be a l i ttle d i��erent than that on nouns. si nce a l arger var i ety o� � i nal verb root consonants are i nvol ved , this examPl e o� coal escence on a noun does Provi de evi dence that 1 -> d in certain Phonological environments. The foyrth tYPe, �. i s added onl Y to singl e TBU roots. verbs whi ch use the =nl suff i x . There are 25 O� these, 1 7 Possess nasal stems Possi b l y indicati ng that the nasal i ty o� the suff i x i s cond i t i oned by the nasal i n the stem ( 20) . The other eight show no evi dence of nasal i ty i n the root (2 1 ) . (20) a. son sO-ni � WClrshiP ' b . tun t�-ni • send ( someone) c · ka.n ki-n f ' boi l ' d . no no-ni • b i te ' I ( 4 o� 17) (21 ) a. Pu p�-ni • swel l b . wu ...u-n i � Pour , SPi l l c . su si)-ni ' de�ecate ' d . t6 t6-nf • c l ose, burY I I I (4 o� 8) I� the :nL su�� i x were a resul t o� nasal cond i t i oni ng, then one wou l d have to seek out the under l Yi ng suff i x . One l i kelY candi date i s :li. A compl i cati ng factor i s the Presence � :nL su�� i xed verbs l i ke these i n (21 ) that have no trace of nasal i ty i n the root . Although it is possi ble that they were h i stori­ cal l Y nasal . there is no evi dence as yet that this was the case or that sYn­ chroni cal l Y , a 1 -> n ru l e can be moti vated3• In add i t i on , there are numerous ----- �- - .--- ---- ------- 49 cases o� co-occurrence o� a nasal i zed vowel wi th 1 i n Suci te. where the 1 is not nasal i zed . Bel ow are a �ew such ex�mPl es o� complet i ve verbs: (22) kan l a • uProot , hnh " i ncl i ne ' �aan l a " H aUer ' The � type of su� � ixat i on �or the i ncompleti ve asPect is :Bl. (23) a. k6 k6-d " draw (water ) ' b . kun ku-d " crunch C i n eat i n9) ' c . c� c6-d " catch. grab ' d . j6 j66-r{ " swal l ow ' ( 4 of 5 ) Most o� the d i sYl l ab i c verbs whi ch take the � o� ::Ki. or � su�� i x have the � inal sYl l ab l e i n the comPl eti ve �orm. In each case. the H nal sYl l ab l e o� the camPI eU va verb i s rePl aced by the suH i x . ::r:i.. as shown bel ow. (24) a· s6-d s6-d " burn ' b . mbu-d mbu-d " oPen ' c . mb�-x i mbu-d " suck ' (25) a. so-?o so-ri " cook ' b . a -?a li-d " return . 90 back ' c · wa-?a wa-ri " drY ' d . Ya-ri Ya-ri " cal l ' There i s al so a set o� verbs whose cOmPl eti ve and i ncomPlet i ve �orms are i denti cal se9mental l y . The common character i stic i s that they each end i n a ----�- --- -,-�-.----- 50 h i 9h vowel . "(26) or i.. sl l i i ' be strong ' b . c:t)B c:: t)I i ' shi ver ' c . nY�?{n nYi?{n ' sU r ' d . k�l � k�l u ' rol l ' ito si i l i There are onl Y a c:: ouPl e of suPl eti ve exampl es. ( 27 ) a. kad b. 9b6 s' ' 90 ' kjJH , ki l l ' Final ly, there are a tew verb torms where the comPl eti ve and i ncomPl et i ve torms are obvi ousl Y phonologi cal l y rel ated to eac::h other . but each Pai r stands in a c::l ass at i ts own . These have been grouPed together below. (28) a. 1 6x 6 ndjJrjJ ' hear , I i sten ' b . w�l o wjJjJ ' Pi ck uP, i:ake oH . trom ' c . tan I!. Uan ' pl ease ' d . nYel i nY� n i ' c::r Y ' e. Pan m a. ' come ' In summarY . i t aPPears that al l tYPes Ot i ncomPlei: i ve sut t i x i ng i s charac:: t eri zed b y the Presence Ot a h i gh vowel . In most c:: ases, thi s vowel i s =1. but i t was seen that tri pl e TBU verbs requi re the vowel . �. There are onl Y a tew excepti ons t o thi s general statement . Al though the structure Ot a parti cu l ar verb maY l i mi t i t t o a restri c:: ted set Ot sutt i x es . there seems to be no cl ear phonolo9i c:: al rules c:: ond i t i oni ng the c:: h oi c::e Ot i ncomPleti ve sut t i x . The choice seems to be Pri mar i l y arbi trarY. and ----- --- - -- - ---- -- - - --- --- --- ---- ----- ---- - -- 51 with a few Phonol ogi cal constrai nts. It must be noted that other Senufo l anguages d i sPl aY an equal l y confusi ng vari ety of i ncomPletive suf f i x es. ComParati ve studY maY Possib l y shed some l i ght on this Problem. It remai ns cl ear . however. that the number of tone-beari ng uni ts i n the stem has a bear i ng on what type of i ncomPl eti ve suff i x i ng is avai l ab l e to that verb . Where m�osYl l ab i c stems hava �r.cess to al l the types of suffi x i ng descri bed above. d i sYl l ab i c stems cannot use the :n1 suffi x . and al l stems wi th three tone-beari ng uni ts (trisyl l ab i c and l ong-short b i sYl labic verbs) are entirely l i mi ted to Vowel Suf f i x ation. B. TRn.1 Behayiour of the Incompletiyw Verb 1 . Th. UnderlYin9 Ton. of the incomPI .ti ve .uffix Al though there are a vari ety of segmental shapes for i ncomPlet i ve suff i x­ es. the one feature that l inks them al l together i s tone. Re9ard l ess of the bPe of segmental suH i xati on . the i ncomPl eU ve suH ix is always High tone after a h i gh tone verb. and Mid tone after Mid tone verbs. (29) COMPLETIVE INCOMPLETIVE a. Piri pid-i. " sel l ' b. ta-?a ta-ti, " Put (an f i re) ' When the i ncompl eti ve suffix i s attached to a Low tone verb. it is general l y Mid tone, as the underl i ned suf f i xes i n ( 30 ) i nd i cate. ( 30) wu Ya t�x i -L l a (he i s vomi t i ng Q ) wu Ya :<a k�-&D. moeJ (he i s i t gi ving you-to) " Is h e vomi ti ng? ' " He i s giving i t to You . ' wu Ya xa k�-&D. n��u� <he i s i t gi vi ng man-to) " He i s gi vi ng i t to the man ' ----- �-- - . --- ..------ 52 However . i f the Low tone i ncomPlet i ve verb i s fol l owed by a noun class c l i t i c (as i n (31b . c» or i s i n Phrase f i nal Posi t i on (3 1 a ) . the tone of the suf f i x i s l owered t o Low tone. ( 31 ) a. wu Ya tUx � -1 -> wu Ya tUx � t " he i s vgmitin9 ' b. wu Ya x, k�-An �6 -> wu Ya x, k�� �6 " he i s 9ivin9 i t to hi m ' c . wu Ya t�x � -1 wu l a -> wu Ya tUx � t wd l a " he i s vgmitin9 on i t ' DePendi ng on the tonAl envi ronment , then , the i ncompl eti ve suff i x can be High tone, Mi d tane, or Low tone. Before trYing to determine the under l Yi ng tone of the i ncomPleti ve suf f i x l et us observe the behavi our of i ncomPl eti ve suf f i x tone i n Cebara. a major Senufo l anguage i n the I vorY Coast . 4 In Cebara, the underl Yi ng tone for the i ncomPl eti ve sutt i x i s Hi gh after Low and Hi gh tone verbs. but Mi d after Mi d tone verbs. (32) Complet i ve IncomPl eti ve Suc i te Engl i sh Low tone kp�l t kp�l i gbel t . gbel t i wound. i njure Mi d tone nYAari nYaari nYt £ ri . nY��ru wal k . strol l Hi gh tone Pari P€r€ , P€ri i sel l (Mi I I s. 1967?) If we l ook at Cebara. we see that Hi gh tone fol l ows Low and Hi gh tone verbs. Thi s l eads us to susPect that the underl Yi ng tone ot the suf t i x is Hi gh tone. If h i stori cal l y. the i ndef i ni te suf f i x was Hi gh tone i n Senufo languages i n general . then i t maY be Possi b l e that thi s suf f i x i s al so underlYi ngl y Hi gh tone in Suc i te. However . posi ting an i ncompl eti ve Hi gh tone tor the Suc i te i ncomPl eti ve suff i x . would require an expl anati on of how this High tone was l owered to Mi d and . i n certai n cases. to Low tone. -- _ . - - - - -. . - ----- ----- - _._ -- ------ ---_ . 53 One Possible waY to account for Hi9h tone l ower i n9 after Low tone i s to tri 9ger a Low tone SPreadi ng Rul e . stated as fol l ows: (33) LOW SPREADING - When a Low tone i s fol l owed by a Hi 9h tone across a morpheme boundarY. the Low tone sPreads to the r i ght. Xl X l" ,, 1 H The fol lowin9 exampl e i l l ustrates thi s sPread: ' vomi t ' If the verb i s i n non-Phrase f i nal Posi t i on . both Hi gh and Low tone wi l l remai n l i nked to the suf f i x vowel . Since. general l y a Low-Hi 9h contour i s not Pronounced on a si n9l e tone-beari n9 uni t , thi s contour i s simpl i f i ed on the surface and becomes Mi d tone (see (30) for examPl es) . Hi 9h does del i n k after Low tone sPread onl Y when the verb i s i n Phrase f i nal Posi t i on or before noun cl ass cl i t i cse• Thi s Low SPread and Hi gh Del i nk i n9 i s i l l ustrated in the examPles bel ow. -> wu Ya t�xi -i II (35) wu Ya t�x i -{ · �' 1 ' He is vomi ti n9 ' f1 HIGH DELI NKING L SPREADING (36) wu Ya t�x i -i wU l � -> wu Ya t�x i-i � l � \t'1 L H I H LOW SPREAD v--1 I L H ' He i s vomi t i n9 on ii ' H HIGH DEL INKING ·On l Y those words whose tones are rel avant to the di scussion at hand wi l l be marked for tone at the autose9mental l evel . Thei r fi nal surface tone wi l l be i nd i ca.t:ed by the tone di acri tics above the words. _----- - --- -- -. - - - --- ---- ----- 54 Al though we cannot Proceed with a thorou9h analYsi s at thi s stage. we can alreadY see that the High del i nk i ng of ( 36) i s a resul t of a combinat i on of Low tone SPreaci ng f rom the lef t and the High tone noun cl ass cl i t i c an the r i ght . High del i nk i ng aPPears to be a waY of decontouri n9 a Law-Hi gh gl i de l i n ked to the same TBU when a Hi gh tone fal l ows or when i t is in Phrase f i nal Posi t i on . The ProPosed rul e wou l d be stated as fol l ows. ( 37) HIGH DELINKING - Del i n k a Hi gh tone whi ch i s preceded by a Law tone -. l inked to the same TBU, and fal l owed by a Hi gh tone l i nked to the fal lowing TBU or i s i n sentence f i nal Posi tion. X X �{� \ By Posi t i n9 an underlYing h i gh tone fer the i ncompl eti ve suffi x , the above anal ysi s can easi l y account for the tonal behaviour of the suf f i x after Hi 9h and Law tone verb stems. We are f orced . however . to l ook fer another expl anat i on for why thi s putati ve Hi 9h tone suf f i x i s Mi d tone af ter a Mi d tone verb stem, rather than the ex�ected Hi gh tone: ( 38) *ta -d ' be putti n 9 ' Rather : ta-ri What i s the nature of Mi d tone verbs that all ows for thi s apparent anomal y? Do Mi d tone verbs have a f i nal f l oati ng Law tone that tri ggers Low tone sPread onto the suf f i x ? Or i s the under l Yin9 tonal natur"e of the suff i :c somethi n9 other than a Pure Hi 9h tone? These questi ons cannot be ful l y answered unti l more data are Presented . Later i n Chapter 6 . we shal l see that the tonal behaviour of the noun class cl i t i c and the i ncompl et i ve suffix is verY si mi l ar . An anal ysi s of the underlYin9 tone of the cl i t i c i n Chapter 6 wi l l show that this aPParent Pecul i ar i ty i s not just restri cted to the i ncomPl et i ve suff i x . � -- - -- - - - �. - -- ------ � --- - ss Be�ore �ermi na� i n9 �hi s d i scussi on . however . �e must take a l ook at the root tone changes that take pl ace when the i ncomPl et i ve su�f i x is added . 2. Raat Ton. Mut.tions in the Incompl .ti ve Verb what has not been stated so far i s that verb root tone often changes when an i ncompl eti ve su�f i x has been added. The d ia9ram below i l lustrates the root tone changes from complet i ve to i ncomPl eti ve. The thi ckened l i nes i ndi cate the most common tonal changes. The changes rePresented by the thi n l i ne account for 12 or 13 verbs each of the 300 i n the corPus. For exampl e. onl Y 1 2 or 13 Mi d tone verbs retain their Mid tone root when an i ncompl eti ve su� f i x i s added . The broken l i nes represent onl Y 3 verbs each. (39) IncomPl eti ve root tone ComPl eti ve tone H ....-------� '-" H L------- L The tone of Hi gh tone and Low tone roots normal l y do not change when the i ncompl eti ve suff i x is added : (40) Hi gh tone ----- ----- ptrt " sel l ' cel l " di vi ne ' - --- --- - - - -------- -- - -- -- -- --- - ------ ---- - 56 Yex U ' quest i on ' cuU� ' heal , be heal thy ' .fuxru ' .fr i ghten ' .fY�?� .fy�?U ' be qui et ' gbOxO gbOx U · ' gather together ' gbarA gbarU ' a9ree, meet ' .f H n ' ni H i n ' nu ' be cl ean ' cMl o c')ill u ' belch ' ct}B ct}l i ' sh i ver ' ( 4 1 ) Low tone The .f i rst str i k i ng observati on i s that mgat Mid tone verbs are l owered �o Law tone when an i ncomPl eti ve su.f .f i x i s added . (42) a. waA U l a? I ..? b . wu Ya " c· waA fuxari l a? d . wu Ya f')xQru la? �x U ' he bgugbt? ' COMPLETIVE ' he i s byYing? ' INCOMPLETIVE ' he rymmaged? COMPLETIVE ' he i s rummaging? ' saxe $ ' wai t ' tare t�l"li ' 9ri nd ' c:l::mri d�nru ' resol ve, choose ' ce?el t ct}?t}lu ' i nsul t , l augh at ' j:l:lxi j��xo ' sharpen ' son sOni ' worsh i P ' tun t.)ni ' send (someone) ' ----- . . ---- - --- -- - --- INCOMPLETIVE ---- --------- -------- 57 In �act. those Mi d tone verbs that do not undergo root tone lowering are i n a de�ini te mi nori ty and �al l i nto disti nct phonol ogi cal cate9Dri es. as wi l l be seen shortl Y . Al though Carl sen (n . d . ) di d not i denti�Y root tone l owering a s a common resul t o� i ncomPl eti ve su�� i x i ng i n SuPYi re, he d i d note that a number o� Mi d tone verbs acqui red the behaviour � Low tone verbs i n the i ncompleti ve. In Cebara, no menti on has been made o� Mi d tone l ower i ng. In Suci te. however. Mi d tone lowering i s so re9ular that some attemPt must be made to account for the change. It i s not known whether the tonal nature a F the suf f i x has something to do with the Mi d tone Lower i ng, or whether the tonal structure of a Mi d tone verb i tsel � l ends i tsel � to tone l oweri ng whenever anY su� f i x i s added, regardl ess of i ts tonal structure. Most l i kel Y, the tonal nature of both the root and the suff i x Pl ay a role i n root tone l ower i ng. If we l ook ahead to the tonal behaviour of suff i xed nouns, we wi l l see that root tone lower i ng takes Pl ace onlY in the Presence of a parti cul ar type of suff i x as shown in ( 43b ) . ( 43) a. ii-I e b. it-xale c bal aPhon -IND ' Class 5 Suff i x c bal aPho�s-IND ' Class 6 Suf f i x I � we suggest that the tonal nature of the suf � i x cal l s for root tone l owering, we are �orced to ask why and how. At thi s poi nt, i t is qui te uncl ear that a synchron i c sol uti on can be �ound to account �or this root tone lower i ng. For now. we shal l cal l root tone l oweri ng a tonal mutation . A mutation i s a Phenomenon which occurs at the Poi nt that words are �ormed but be�ore the aPpl i cat i on of other tonal rul es. Mi d tone verbs whi ch undergo the mutation of tone l ower i ng behave exactl y l i ke Low tone verbs, once the l owering has taken Pl ace. Thi s means that the i ncomPl eti ve formati on of a Mi d tone verb usual l Y ---- . . -- .-. - -- -.. ----_ ._--- ---- ------ 58 l i rst consists ol Mid tone l oweri ng, and then the ad justing ol the tone o� the i ncompl eti ve su� � i x i n accordance with the lact that the verb root is now low tone. ( 44) s� - G I I M H -> s� -G I l I -> s�u �1 H ' be buYing ' lH Tone Mutation low SPread M->l Whi l e most verb roots mai ntai n Hi gh and low tone on verb roots when the i ncomPleti ve sul � i x is added , and l ower Mid tone in this same envi ronment , there i s a smal l set o� verbs where a low tone root i s raised to Hi gh tone i n the i ncomPlet i ve , whi l e the Mid tone verb root remains Mid except i n one or two cases, where i t i s al so rai sed to H i gh tone. Bel ow i s a l i st � the verbs which fol low these Patterns: ( 1 0 ) a. se si gi ve bi rth , be born b . fo �u emi grate c· jo Yu speak, saY d. to tu �al l B H eat f. U U weave, braid m� come y a. be sick, ache k6r! draw (water) b . kGn kGr! crunch < i n eat i ng ) c· c� c6r! catch , 9rab d . j6 j66r! swal l ow (24) a. s6-x! s6-r! burn e· ( 1 1 ) a. Pan b. ya , (23 ) a. k6 , 59 b . mb6-d mb6-d oPen c. mbil-x � mb6-d suck (25) a. so-10 so-ri cook b. la-1a li-d return, 90 back c. wa-1a wa-ri drY d . Ya-ri Ya-ri cal l HSi a. Yil y6U steal b. kil k6H die The common characteristic of al l these verbs i s that they have monosYl l ab i c , si n9le tone-beari n 9 roots. Even those verbs wi th d i sYl l ab i c comPl eti ve forms aPPear to be actual l Y monosyl l ab i c roots wi th a tYPe of comPl eti ve suff i x that is rePl aced by an i ncomPl eti ve suf f i x (see (24) and ( 25 » . Thi s set of excePtions i s aPParent l y qui te ol d i n Senufo because Cebara Produces the exact same tYPe of behavi our . In fact, the tonal correl ati on between the two l an9auges for thi s set of excePti ons i s much h i gher than for the ' normal ' tonal behavi our of i n comPl eti ve suff i xat i on . See examPles of normal i ncomPl eti ve suf f i x at i on for Cebara i n (32) above. Below are examPl es of the excePti onal set of verbs i n Cebara, alon9 wi th the Sucite equ i val ents: (45) ----- - - -- - Cebara Suc i te En9l i sh IH, IH B, U eat 1 il9il, l 6r6 ndilx � , nd6rf Plant mil9il, muru mbilx � , mb6rf suck H 9� , U d <!'Ii 1 1 s, 1967? ) t�l d , Ud descend It is i nstructi ve to note that non-l owered Mi d tone verbs are found onlY - --- -- ---- --- - --- --- ------- 60 in �hi s i rregul ar grouP of verbs. Therefore. �he Probl ems di scussed above tor �he behaviour of �he i ncomPle� i ve suff i x tone af�er Mi d tone verbs shou l d not comPl i cate �he analysi s of the under l Yi ng tone Ot the incomPl e�i ve suf f i x it we consi der the i ncampl eti ve suHi x aH on o·f non-lowered Mid tone verbs as i rregu­ l ar alon9 wi �h the rai sin9 of Low tone verbs �o Hi gh tone. Unfor�unateI Y . time and sPace do not al l ow for a more ri9Durous studY of thi s type of tonal behavi our whi ch rai ses Low tone verb roots to High �one and Preven�s the normal I O��i· i n9 of Mi d �Dne verb roots to Low tone. Wi �h the Presen� data. �here appear to be some def i n i te �endencies for tonal change bas�d on sYl l ab l e type and a Parti cul ar �YPe of i ncomPleti ve suf f i x i n9. The Precise reasons f or this deviation remains obscure . however , forcing us to remai n a� �he descri pt i ve l evel of �he analYsi s. Wi th this in mi n d . the Low Rai sin9 rule i s posi ted bel ow. (46) LOW RAISING When a si n9l e TBU verb roo� acqu i res an i ncomPl eti ve suf f i x of the shaPes. 0 . -ri o or -I i , Hi9h and Mid tone verbs general lY retai n thei r roo� tone. whi l e Low tone roots are rai sed t o Hi9h : o suf f i x L - > HI [Xl -ri -l i root In contrast. when a si n9l e TBU verb Possesses a Vowel suft i x or the :ni suft ix . the verb root tone fol l ows the normal Pattern Ot loweri n9 ot �he Mi d tone root i nstead of rai sin9 a Low tone root to Hi gh tone. ----- - ---- - - - ------ ---- ----- ---- ------- - - - 61 IV. HISTORICAL. SPECULATIONS AND TONAL FEATURES Exami nation � these root tone mutations reveal s one common tendency. In both the grouP where Mi d tone l ower i ng takes pl ace, and the grouP where Low Rai si ng occurs, the three waY tonal d i stinction �ound i n the comPl eti ve i s reduced t o two tones i n the i ncomPleti ve. The l argest 9rouP. wh i ch i nvol ves Mi d tone 1 0weri n9, di stingui shes between Hi gh and Law tone, whi l e the second . " i rregular ' 9rouP contrasts High and Mi d tone. There are no examPles ot a Hi gh or Low tone verbs mutating to Mi d tone i n the i ncomPl eti ve. As a resul t , the vast majori ty ot i n comPl eti ve verbs are ei ther Low tone or High tone. These observati ons l ead us to speculate that histori cal ly, the Suc i te tonal system maY have evol ved trom a two tone system. Other evi dence suPPorting this possi bi l i ty can be found by exami n i ng mi n i mal Pai rs. There are numerous mi n i mal Pairs that contrast Hi gh and Low tones as wel l as Hi gh and Mid tone verbs, but tonal Pai rs contrasti ng Mi d and Law tone are rel at i velY rare. ( 47) a. High vs. Mi d ( 1 5 0+ 28 mi n i mal Pai rs i n 300 verb corPus) t� ' gri l l ' ' blow, wi nnow, swel l ' t6 " .f l ow, driP ' ' emi 9rate ' b. Hi gh vs. Low ( 10 at 28 mi n i mal Pai rs) mbil:(� mbax! ' open ' man! ' assemble, add ' c . Mi d vs. Low " suck (an orange) , C l i ght, Pl aster ' (3 ot 28 mini mal Pai rs) Yiri ' cal l ' Yid " get uP ' kul i " gather ' kilH " shave ' No mi nimal tonal tri pl ets have been tound tor verbs. Bel ow is near mi n i mal tri pl et: ( 48) tad ----- - --- - · di9 ' tuxo • carrY ' tilx i " vomi t ' - -- ----- ----- ---- ---- ------ 62 The same can be found for nouns; no Mi d versus Low mi nimal Pai rs have been attested . In addi t i on there are manY more Hi gh tone verbs than Mi d or Low tone verbs ( 1 35 vs. 95 vs. 71 i n a data samPle o� about 300 verbs> . Thi s cou l d mean that i n the d i stant Past � there was no tonal contrast between the Mi d tone and Low tone. and as a resu l t . the l exi con d i d not devel op mini mal Pai rs. It has been ProPosed by numerous authors throughout the Years that proto­ Niger-Congo was a two tone l anguage. Proto-Bantu . i t has been suggested . was also two tone. Cl ements ( 1979) suggests that the aggl uti na�i ve character of manY Bantu l anguages maY have been a factor i n thei r trend toward an accent tYPe system whi l e i solating l anguages tended to retai n the tonal d i stinctions and even to al l ow for Prol iferation of tonal contrasts. In West Afr i ca. there are two. three. and t our tone languages. It was noted i n the Introduction that Senufo d i al ects have Pri mar i l y three tones. whi l e a few aPPear to have two or tour tones. The possibi l i ty that Senuto maY have evol ved � rom a twa tone sYstem seems Pl ausi b l e i n the l i ght of ProPosal s gi ven by other l i ngu i sts fami l i ar with Afri can l anguages. The waY that a three or tour tone s�stem maY have evolved t ram a twa tone sYstem has been the abject ot studY by vari ous authors. Maddi eson ( 1 974 ) d i scusses the possi bl e mot i vati ons for tonal change or ' tone-spl i tt i ng ' . as the h i stori cal Prol i fe -abon of tones i s oHen termed . These i ncl ude " the downdri ft model " i n whi ch he suggests that "new constrasti ve tone l evel s" evol ve "as a result o� the suPeri mPosi t i on of i ntonational PaHerns. such as downdri H " (p. 29) . the "sandhi model " where tone sandhi rul es can tri gger Permanent l ower i ng or rai si ng ot tones. and the " ' Phonation type ' model " .- - -- ---------- where tonal ----- 63 change was tr i 99ered by se9mental � actors. such as certai n types o� consonants. In discussi n9 the ' tone sandh i ' model . Maddi eson i nvi tes us to " i ma9i ne how a l an9uage wi th two bas i c tone l evels but with a tendency �or h i 9h tones to be somewhat l owered when a l ow tone fol l ows and for l ow tones to be somewhat rai sed when a h i 9h tone �ol l ows cou l d develop into a l an9uage with �our tone level s if the condi t i onin9 envi ronment was l ost or absorbed. Thi s is essent- i al l y the k i nd of Process that i s' sketched for Fe?Fe? by HYman ( 1 972) and for Dschan9 by Tadadjeu ( 1974 ) . A merger of the raised l ow and l owered h i gh tones i nto a mid tone could resul t i n a three-l evel system evol vin9 �rom an earl i er two level one. " (49) Maddieson : h I I -> lowered h 1 h -> rai sed I h -> ml -> mh Another tYPe of tone spl i tti n9 was su9gested by DwYer (vi a Maddieson ) : "DwYer < 1 973: 248-250) has su9gested that Southwestern Mande d i d develop a third contrasti ve tone when the compl ementarY di stri bution of mi d and l ow Phoneti c level s i n disYl l ab i c nouns was ' spoi led ' by the borrowin9 o� l ow-low nouns �rom Northern Mande. and the d i stri buti on of the three tones does sti l l l argel y reHect the ear l i er compl ementaritY. " ( p . 30i (50 ) DwYer : Southwestern Mande H L /\ M L + Low tone l oanwords From Anderson ( 1978) we � i nd that Stalhke notes that the �our l evel 1 gede tone system "seems to have devel oPed from an ori 9i nal two level system. by the spl i tt i ng of each l evel i nto two di sti nct re9i sters" (p. 171 ) . -- -_ . _._ -------- 64 (51 ) Stalhke: 1 gede H L \ / / \ Cl ements ' ( 1 9S1 ) formal i zati on of tonal f eatures ( as a revi sion of ViP ' s ( 1980) theorY) ref l ects the theorY of h i storical tone-spl i tti ng. Fi rst . he notes that i n certai n d i al ects of Ewe, the High , Mi d , and Law tones do nat stand i n symmetri cal rel at i onship to each ather . Rather the Mi d and Law t.ones "al ternate i n h i gh l y Productive Patterns whi l e nei ther tone al ternates Product i vel Y wi th the High tone. " (p . 57 ) . Thi s. Cl ements demonstrates. can be ref l ected in a system of tonal features whi ch consi sts of two Pr i marY re9i sters whi ch can be spl i t to create sYstems of three or four regi sters. " Thus. we cou l d saY. ( i n synchroni c terms) that the Ewe system i nvol ves a spl i t of the l ower register i nto a h i gher and a l ower subregi ster . normal l y desi gnated as "Mi d tone" and "Law" tone. respecti vel y. " (p. 57) Cl ements rePresents the Ewe l aw tone spl i t as gi ven bel ow: (52) Cl ements: Ewe - three waY spl i t h I / \ h 1 ( p . 58) A four tone sYstem wou l d natural l Y invol ve the spl i tt i n g of bath the h i gh and low tone registers. (53) Cl ements: four waY spl i t h 1 / / h \ \ --------�- - ---- ------ I h 1 ------ 65 studies i n other l anguages such as Lendu ( Tr i � kovi c , 1 977 ) have l ed to hYPotheses that three tone and �our tone l an guages evolved �rom two tones through the Process o� tone spl i tting. Th i s tone spl i tt i ng maY have come about through the Process � product i ve tone al ternati ng rules devel oPing i nto phonol ogi cal tonal contrasts and through the i ntroduct i on o� l oan words that hel Ped to estab l i sh new tonal l evel s. The i mpl i cat i ons of these hi stor i cal devel oPments �or synchroni c anal Ysi s can be seen in Cl ements ' ProPosed tonal .feature system. Clements i ntegrates thi s tonal .features ProPosal i nto the �ramework o.f autosegmental phonol ogy by pl acing t�e .features in rows. The toP row incl udes the tones o.f the Pri marY regi ster . whi l e the bottom row consists o.f tones assigned to the "subregister wi th i n the Pri marY regi ster det i ned by the entrY in the .f i rst row . " Hi s rePresenht i on o.f a M-H-L tonal sequence in Ewe is gi ven as .fol l ows: ( 54 ) I h h I I cv cv cv The .features above i ndi cate that Mi d tone is a resu l t o.f a spl i t at the Low tone register. The Hi gh tone regi ster , on the other hand , d i d not undergo a spl i t , and thus i s rePresented onl Y by a si ngle teature gi ven i n the toP row. In Suci te , we have suggested that Mi d and Low tones maY have had a common ori gi n . thus l eadi ng us to opt tor the not i on o.f a hi stori cal tone spl i tt i ng at the Lower tone re9i ster . The teature system ProPosed by Cl ements then Pred i cts that Mi d tone as .found on the verb wi l l be rePresented by the .features � and Low tone wi th �. For PurPoses at readi ng ease. the Pri marY regi ster tone i s - ---- ------- ----- 66 represented by a caPi tal whi l e the subre9i ster tone i s i ndi cated by a smal l l etter . In add i t i on . instead of pl aci n9 the two rows of features above the se9ment. the � eatures wi l l be Pl aced bel ow the se9ments wi th the Pri marY re9ister tone c l oser to the se9ment than the subre9i ster tone. as shown bel ow. cvcv I L I h Previousl Y we ProPosed that the under l Yin9 tone of the i ncompl eti ve suf f i x was Hi 9h tone. Si nce, at thi s poi nt . there i s no aPParent tone spl i tt i n9 of High tone. the si n91 e feature. �. i s suf f i cien t . If the Mi d tone verb. bein9 anal ysed as under l Y i n9l y Ln. is fol l owed by a H suf f i x , i t cou l d Potent i al l Y tri 9ger the Low tone sPreadi n9 rul e, si nce the Low tone sPreadin9 rul e has no tone sPec i f i ed on the subre9i ster tier of the Low tone (see (33» . Theoreti ca l l y , then , ei ther L h o r LI tones cou l d sPread onto Hi 9h tone as shown bel ow. ( 56 ) x 1. L X ... I 1 H h The possi b l e consequence of Lh tone sPreadin9 onto a Hi 9h tone i s the format i on of a Mi d-High contour tone on the suff i x . Th i s , unfortunatel Y , i s not an acceptabl e tonal shape 'for the i ncomPl eti ve suff i x . I f . however . we al l owed a Lh verb to sPread onto a Hi9h tone suff i x . but then moti vated a ru le del etin9 the subre9i ster h i 9h tone, the correct surface form wou l d be Produced . In add i ti on , thi s Process would hel P to expl ai n whY a Mi d tone verb i s Lowered to High tone before a Hi gh tone suf f i x . Bel ow ----�- - .-- .. ---- we -- - -- - - - -_. . see an ex amPl e of thi s suggested 67 Procedure. su - U (57) L 1 L ... � - > sil - u C be buYi ng ' l -- 1H H L I 6) L SPREAD H DELETION I� h i stor i cal l y . LI and Lh were one si ngl e tone. caPab l e o� sPreadi ng onto h i gh tones. then even a�ter the spl i tt i n9 o� the Low tone regi ster . i t is concei vabl e that this spreadi ng conti nued to occur . but that i t requ i red subregi ster Hi gh tone del et i on. It wi l l be seen . i n ChaPter 4. however . that Lh sPreadi ng onto Hi gh tone is not al l owed across word boundar i es . and it can be Permi tted across morpheme boundar i es o� a noun and verb � i � subregister High del et i on takes Pl ace. These l i mi tati ons wi l l SU9gest that on l Y those Low tones not speci � i ed on the subregi ster t i er �or Hi gh tone can tri gger Low tone spreading. Once evidence of this probl em i s presented . adj ustments to the Low tone sPread in9 ru l e wi l l be made. An al ternati ve anal Ysi s is to moti vate a rul e del et i ng the subre9i ster Hi gh tone of a Lh verb. Once thi s 5ubregi ster High tone is del eted . the Low tone then spreads onto the �ol l owi ng Hi gh tone as shewn bel ow. (58 ) su - U I L I I H ® H DELETION - > sil - u t/ 1 c be bUYi ng ' LOW SPREAD Thus. i � �or some l ex i cal reason . the subregister h i gh tone i s b l ocked �rom ----- - -- - - -- - -- ---- ---- --- ------ 68 deletion. Low tone sPread i n9 i s also b l ocked t rom aPpl i cat i on . The i mPortance ot orderi ng the High Del eti on rul e betore Low tone Spreadi ng wi l l be made more cl ear i n the tol l owing chapters si nce at this Point . sut t i c i ent evi dence is I ac/ d ng . The subregi ster High tone del etion rul e can b e tentat i vel Y stated as f ol l ows: (59) SUBREGI STER HIGH DELETI ON : Del ete the subregi ster H i gh tone of a Lh tone when i t is tol l owed by a H i gh tone. x LI ® X � As has al readY been noted . thi s rul e aPpl i es to the majori ty of Mi d tone verbs. However . the smal l group of Mi d tone nouns whi ch do not l ower to Low when fol l owed by an i ncompl eti ve sut t i x do not al l ow for the aPpl i cat i on ot the Hi gh del etion ru l e . V. CONClUSION In summarY. then. the posi ted tour rul es are l i sted below . ( 46) Low Rai si ng (59) Subregi ster Hi gh tone Del et i on ( 33 ) Low tone SPreading ( 37 ) Hi gh Del i nk i ng We have al readY suggested that Hi gh Del et i on must take Place before Low ----- --- _ . _ ._ .. 69 tone SPread i ng . so that those Lh verbs whi ch have undergone High del eti on can al so tr i 9ger Low tone SDreadi ng. Low Rai si ng, l i kew i se. must take plnce betore Low tone Spreadi ng but t er the oPPosi te reason : i t must bloc k Low tone sPread­ i ng trom taking pl ace. In other words, both the Hi gh Del eti on and the Low Rai sing rul es are tone teature changing rules wh i ch attect the appl icat i on ot the Low SPread rule. The t i nal rule. Hi 9h Del i n k i n g . was seen to take pl ace as a resul t of t.he Low SPread ru l e and therefore i s obl i gatori l y ordered aHer the Low SPread rule. The major questi on addressed i n thi s -chapter was how to anal yse a Phonet i c Mi d tone i n Suci te. Cl ements ' tone feature system was ProPosed al on9 wi th the su9gesti on that Mi d tone was an hi stori cal spl i t f rom Low tone. ci ting as evi dence the smal l number ot Mid versus Low tone mi n i mal Pai rs . as wel l as the rather producti ve al ternati on of tone between Mi d and Low tone. According to Cl ements ' system then , the Suc i te Mi d tone verb is underl Yi ngl y Lh tone. There are other possi b l e anal Yses t or the Mi d tone. however . that need to be consi dered here. One possi b i l i ty is that the Mi d tone is under l Yin9l y toneless. If t h i s i s the case. then Mi d tone woul d be assi 9ned to the tonel ess segments at some poi nt i n the derivat i on. In addi t i on . a rule assi gnin9 a Low tone to toneless se9ments whi ch are fol l owed by a H i 9h ton� i ncomPl etive sut t i x woul d have to take the place ot the Present H i gh Del et i on rule. Th is is a potenti al l Y Pl ausi b l e hYPothesi s t or the Suci te verb . However , i t wi l l be seen . as data is Presented i n l ater chaPters. that this hYPothesis runs into ser ious probl ems. A second al ternati ve hYPothesis i s that Mid tone verbs are under l Yi n9ly a contour tone consi sti ng at a Low and Hi9h tone. These Low-Hi9h tones wou l d then be si mpl i f i ed to Mi d tone on the surtace. The troub l e wi th this anal Ysi s i s ----- - ---- - -- " - - ---- ---- ------- " - --- - 70 that theoretical l y , a d i sYl lab i c Law-High verb shou l d have a Law tone l i nked to the � i rst Tau and the Hi gh tone l i nked to the second TBU. I n actual i ty , there are no such Law-Hi gh verbs. Mi d tone verbs are Mi d tone in comPl eti ve t orm regardl ess Ot the number at tone-bearing uni ts . Therefore the hYPothesis that the Mi d tone verbs are under l Y i ngl y Low-Hi gh tone cannot be accepted as a vi ab l e anal Ysi s . We are l et t . then wi th our ori gi nal anal ysi s that Mi d t�ne i s a spl i t from the Low tone, with i ts features gi ven as Lh . Later chapters wi l l reveal other types Ot Mi d tones tound on nouns and verbal parti cles wh i ch requ i re d i t t ering anal Yses . NOTES 1 . It i s Possi b l e that these med i al nasal s were hi storical l y tone-beari ng or sYl l ab i c , and as a resu l t sti l l retai n the tYPe at sut � i x i ng requi red t or tri pl e TBU verb torms. 2. In skimmi ng the verb l i st i n the aPPendi x , the reader wi l l note that nasal s on pre-nasal i z ed consonants occur on both disYl lab i c and tri syl l ab i c verbs . At this Poi nt , it does not aPPear that Pre-nasal s on verbs are tene bear i ng or sYl l ab i c , thus a verb with three TaU ' s can al so al l ow a Pre-nasal i z ed conso­ nant . 3. A l ->n rul e can possi b l y be moti vated el sewhere i n the l anguage. In ChaPter J, where a dist i ncti on i s made between noun roots with nasal i z ed vowel s and those wi th a ProPosed under l Yi ng t i nal nasal consonant. the l atter i s seen to cause nasal i zati on ot the i, where the t ormer does not . It is possi b l e that the verbal sut t i x , :01 i s deri ved t rom :li in the same waY . However . at t h i s poi nt , not enough research has been done to permi t such a det ini tive ProPosal . 4 . In his wri t i ngs en SuPyi re. Carl son gi ves verY l i ttle attenti on to the tone et the incomPl eti ve verbs. From h i s descri pt i ons and data. it aPPears that . according to h i m . there is no general l y no change i n tone between the comPle­ t i ve and the i ncompl eti ve t orms . 5 . It shou l d be noted here that noun cl ass cl i ti cs varY tonal l Y themsel ves . TheY can be Hi gh . Mi d . or Low tone dependi ng an the tonal envi ronment . Atter High tone verbs . cl i tics are al ways Hi9h tone. and they are also High tone at ter incomPlet i ve Law tone verbs. For more intormat i on concerni ng the tonal nature at the noun cl ass cl i ti c . the reader i s asked to rater to Chapter 6. ------ _ .---- ------- 71 CHAPTER 3 - THE NOUN I . INTRODUCTION The Precedi ng chapter deal t with the morPhol ogy and tone o� the verb . The exami nati on of the tonal behavi our o� the verb and the incomPl et i ve su� f i x Yiel ded several tentat i ve rul es . It wi l l be seen i n thi s chapter on the noun that same of the tonal behavi our i denti f i ed on the verb is also found on the noun . For ex ampl e , it was ProPosed in Chapter 2 that the Low tone of the verbal root sPread onto the Hi gh tone of the compl et i ve suff i X . Li kewi se, i t i s proposed i n thi s cnapter that certai n nomi nal su� f i xes whi ch are Hi gh i n i t i al are l owered to Mi d tone through the process of Low tone SPread . Another poi nt o� si mi l ar i ty i s the fact that nominal roots w i th certai n types of tone l ower to Low tone when fol l owed by a part i cu l ar k i nd of Hi gh i n i t i al suff i x . j ust as Mi d tone verbs l ower to Low when fol l owed by the High tone i ncomPl etive suff i x . Other i ssues i n thi s chapter concern the mel odi c nature of the noun and a compl icati on in aPpl Yi ng the Assoc iat i on Conventi on of Left to Ri ght Li n k i ng. The matter of l i n king tones to segments is further compl i cated by the struc­ tural behavi our of certai n nominal suff i xes . One nominal suf � i x wh i ch . dePend­ i ng on several factors. can undergo var i ous tYPes of segmental del et i ons. poses probl ems � or the rel i n k i n9 of tones whi ch have l ost thei r se9ments to del et i on . These and other i ssues, such as the under l Yi ng tone of the various nomi nal su��i xes. wi l l be the f ocus of di scussi on i n th i s chaPter . The di scussion of the underl Yi ng tone o� the nomi nal roots themsel ves wi l l . however . have to wai t unt i l ChaPter 5 where a �ul l er di scussi on of the noun Phrase as a whole wi l l enable one to exami ne nominal tonal behavi our i n the context of the Phrase . Thi s chapter wi l l f i rst surveY the general morphol o9ical and tonal structure of nouns as wel l the di ff2rent tYPes of nomi nal suf f i x i ng found in Suci te. It 72 wi l l then exami ne each tYPe oi su�� i x . Provi d i ng a descr i pt i ve account o� i ts •• gmental behaviour b.�ore descri b i ng and analYz i ng the tonal behaviour . A. Noun structure Non-complex noun roots are normal l Y comPosed � a noun root and a suii i x . Noun roots are ei ther monosYl l abic or d i syl l ab i c and contai n . at the most . two vowel s. The syl l abi c sequences al l owed on noun roots are noted below: (1) a. CV ji ' son l b . CVCV tol� ' owner . chiai , head at c. CVV tii ' Python I The i n 1 U al consonant at a noun root maY be Pre-nasal i zed . d . NCV ndl toad Hrom I i ' to e . NCVCV "gura-xo smoke-suf� i x (irom w� eat I ) " to b l acken ' ?) B . Noun CI ..... Ind Suffixing Al l nouns i n Suc i te are grouPed i nto noun cl as.es whi ch are most trequentl y marked by a sui�ix . There are 8 noun cl asses i n Suci te, three singular . three Pl ural . and two mass / col l ecti ve cl asses. Each cl ass has an i nd� i n i te and a def i n i te suf f i x . For the PurPoses of analYsi s. these cl asses are 9roUPed i nto two cata90ries ( l abel ed Type I and TYPe I I ) , accordi n� to the Phonol09i cal shape of the i ndei i n i te sui f i xes. Below are examPles at dei i n i te and i ndef i n i te nouns ior each noun cl ass wi th the suifi xes bei n� underl i ned . 73 (2) Class c. TYPe Indef i ni te Oef i n i te Engl i sh ' WOIItln ' 1 I c..-1Id. ce"'!la 2 II gba-alA gba-i::bi ' ri vers ' 3 I gba-u. gba-� ' hause ' 4 I gba-n gba-an ' houses ' 3 I gba-U, gba-�-aa ' -F ereh ead ' 6 II gbl-2ilA 9b�-:za:Id. ' -Foreheads ' 7 I su� si)-a ' main d i sh ' 8 I si-U si--m-ba ' ai l ' New" Ipne Sucite nouns exhibit nine d i -F-Ferent tonal pattarns. Al l excePt one can be -Found an Cl ass 1 nouns wi th a 0 su-F-F i x . The ni nth . Low-Hi gh tone. can be -Found onl y on nouns that Pos...s a cl a.s su-F-Fix. (3) Hi gh <6 >- wtrt ' maney ' Hi gh-Mi d < 16> -F Ycla ' f i sh ' Hi gh-Law <5> j� ' Pecket ' Mi d gba ' ri ver ' Weak Mi d <39) caan ' market ' Mi d-Law canB ' Younger sibl i ng ' Mid-Lcw III < 42> ji ' son ' Lew c� ' chi l d ' gbOn-a ' 9r"anarY ' < 42> <7> <77) Lcw-Hi 9h <22> Al though there are a vari ety o-F tonal .haPes an nouns. ward -F i nal tone i s - < > . number of examPles in a data samPle o-F simple non-compl ex nouns a-F al l cl asses. ----- - - - . - . - . ._-- 74 ei �her High , Mid or Low tone. It wi l l be seen l ater i n Chapter 4 that each o� �h.5. three tone. a��ects the �ol l owing verbal el ement i n d i �� erent waYs. On the sur�ac. there are also onl Y three l avals o� tone word i n i t i al ly. However , i � wi l l b. seen i n Chap�er � that the i ni ti al tone o� nouns exh i b i ts a lour waY contrast. The l abels, High , M i d , Weak Mid and Low rePresent thi s �our-waY contrast, whi ch beccmn evident when the noun i n questi on i s Preceded by another nominal in a noun phrase. There ara three types of Hi ih initial nouns , Hi gh , Hi gh-Mi d . and High-Low. Al though al l o� these nouns can trigger �onal ru l es . �hei r own �one never changes. Almost al l Hi gh and most Hi gh-Low nouns are l oan words. High-Mi d nouns seam to ba more i nd i genous to Senu�0 1 . One character i st i c of the l a�ter is �hat �heY seem �o requ i re two tone bearing uni ts. Altogether High in i t i al nouns account �or on l Y 27 nouns i n the data samPl e ol about 255 nouns. The questi on that i mmediatel Y comes to mi nd is, why such a smal l number? Th i s questi on wi l l b a deal t with i n ChaPter � when we discuss the underlYi ng tone ol nouns. A d i sti nction is made between weak Mi d nouns and Mid nouns because, al though they have the same pi tch i n c itati on �orm. they di�ler i n tonal behavi our i n other contexts. The weak Mi d nouns as wel l as the Mi d-l ew W nouns. are more suscept i bl e to certain tonal changeS than are Mid and Mid-Low S nouns. Thi s wi l l b e made cl ear i n Chapter 5. Thera are two tYPes o� Low i n i ti al nouns. Low and Lcw-Hi gh . Low tone nouns �orm the l argest grouP in the data samPle <about n nouns) . Low i n i ti al nouns are sub ject to a Hi gh SPreadi ng rul e. which wi l l be descri bed and di scussed in chapter 5. Since most nouns Possess a su�� i x of some tYPe, further di scussion of nomi nal tone can be found i n section I I . B. ---- ------ - ---- -- -- - --- -- - - --- - - - - - - - - - ---- ---- -- - - -- 75 II. TJ£ INDEFINITE SUFF I X A. - TYPE I &trudur- As w�s ment i oned above. there are two types ol i nde� i n i te 5u� � i x es. TYPe 1 ind� i n i te suff i x i ng i s the si mpl est . Its basic �orm i s -eve The consonant vari es accordi ng to c l ass afl i l i �t i on (See chart (2) above) and the vowel i s ei ther the s�me o r verY si mi l ar to the vowel ol the noun root, reg�rdl ess al class a�f i l i ati on . The onl Y vowel s not al lowed on the suff i x are the h i gh vowels. 11 1 and 11.11. It aPPears, lram l ooking at the ex�mPl es below i n (4) . that the suH i x vowel i s e.senti al l Y a coPY o� the root vowel except when the reat vowel is [ + h i 9h l , at whi ch Poi nt the suf f i x vowel is [-h i 9h l , but i t acquires the same feature f or [ + or - bac k ] as the root vowel (see (4f , g) below) . The onl Y feature that seems to be consi stent �or the TYPe I suff i x vowel , then , i s the [-h i gh ] feature. (4) a. gba-xV -> gba-xa " house ' b . cln-xV -> CE n-XE " sauce ' c · gb�-I V -> gbdn-1 53 " 9ranarY , d . fd-XV -> fd-xd " corn ' e. ki-rV -> ka-r4 " meat ' f . 50-I V -> s6-ld • H ear ' g. cl'-I V -> ci-Ie " th i gh ' Wh.ther the TYPe I cl ass su�� i x vowel i s under l Yi ngl y �eaturel ess or whether it does Possess the [ -h i gh l le�ture i s an issue that cannot be �ul I Y addressed here, wi thi n the scoPe of the thesi s . However , i nvest i gati on ol the d�ta i n the Noun Lex i con , wh i ch i s l ocated in the aPPendi x . wi l l reveal the patterns al behaviour verY bri � I Y sketched above. ---_ . . --- - .- . ---_ ._--- ---- -------'-- 76 Al l the nouns of Noun Cl asses 1 , 3, 4, 5, 7 and S Possess the TYPe I i ndef i n i te suffix . As was mentioned above, the basic shape of the suff i x i s -CV. However , not al l TYPe I nouns bear this basic shaPe. Same aPPear to have no suffix at al l . The l eftsi de column of chart (5) indi cates that three c lasses have examPles of suff i x l ess nouns. Whi l e onl Y the mi nori ty of nouns i n Cl asses 3 and e are suff i x l .. s, the major i ty of Class 1 suff i xes are i n thi s grouP. One possib l e reason f or this i s that the -CV suff i x for Cl a�s 1 nouns, -wV, was h i stor i cal l y reduced to a vowel and then f i nal l y deleted . However , at thi s stage, there is no evi dence of an ear l i er -wV suf f i x on most suff i x l ess Cl ass 1 nouns (but see d i scussion of di syl l abic noun roots, p . 79) . Because there i s no aPparent i ndef i n i te suff i x on this set of nouns, i t i s d i f f i cu l t to ascertain at this Poi nt to Nh i ch cl ass they actual l Y bel ong. However , in chart (90 ) i n section IV, on the def i n i te suff i x , cl ass af f i l i ation i s made cl ear b y the tYPe of def i n i te suf f i x used. ( S) MonosYl l ab i c Noun Roots o SuH i x -cv SuH i x CI . 1 ji " san ' (33)·· c:e-.... CI . 3 ni " Hre ' (3) b -d ' Pl ace' <33> • • ...aman - (N ) CV Suf f i x ' (1) d n-I')i • ...Ye-I')� " l eaf ' <25> "'Yen ' yi • antel ope ' < 1 > Cl . 4 0 t!: -Yi CI . S 0 ci-l� · thi9h ' (37) SE -nE • sU ng ' <5> CI . 7 0 su-r� ' main dish ' < 19> k��-n� " caHan ' < 1 > ta-be ' medi cine ' < 1 > sa-m� " mi 1 1 et beer ' ( 10) Cl . S juu • sPeech ' (2) Pl aces , l eaves ' •• Numbers in angled brackets indi cate the number of l i ke examPles in the data sample of simP l e non-comPlex nouns, rePresented by the word in the chart. For example, jl <33> means there are apProx i mately 33 monosYl l ab i c Cl ass 1 nouns, i n a l ex i con of about 255 simpl e monosYl l ab i c: and di sYl l ab i c nouns. 77 There i s another set o� nouns wh i ch have pri mari ly nasal i nde� i ni te su��ixes. Al though data on Def i n i te su� f i x i ng (see chart (90» Provi des cl earer evi dence for cl ass aff i l i at i on , the ri ght si de col umn of chart (5) provi des examPl es of nasal i z ed i ndef i n i te suf f i x es for each cl ass. What seems to be hapPeni ng i s that there ex i sts a set of noun roots whi ch Possess a � i nal under-lYing nasal consonant . When thi s consonant i s fol l owed by an i ndef i ni te suf � i x i t assi mi l ates to the i ndef i n i te suff i x , whi ch i n turn , becomes [ + na­ saI l . The two coal esce and become a singl e nasal consonant , as shown i n (6) below . ( 6 ) �uN-xV -> fu?-xV - > � u�-� - > fu�V -1-1-> fu�� ' i nside ' Si nce there i s no such thi ng as a word that ends i n a nasal consonant i n Suc i te . i t maY b e debatabl e whether one can Posi t an underl Ying nasal consonant i n these cases. An al ternat i ve anal Ysi s i s to suggest � i nal nasal i z ed vowels, which are i ndeed Plent i � u l i n the l anguage. However . the eXamPles bel ow , (7) , indicate that nasal i zed vowel s apParent l y do not tri gger the nasal i z ati on o� i nde� i ni te suf � i xes: (7) a. CEn-XE ' sauce ' *CE �E b . 9bon-l� ' f i rePl ace ' *9bonj For thi s reason , we ProPose that i t i s an under lYi ng f i nal nasal consonant which causes the nasal i zati on of i ndef i ni te su� � ix es . Further research needs to be done, however, on the nature of nasal i ty in Suci te. ----- - _ . _ ._ •.._-- 78 Disyl l ab i c roots �ol l ow the same Patterns of suf f i xat i on . The reader wi l l recal l that the types of consonants al l owed i n i t ial sYl l ab l es) are ex tremel Y l i mi ted . en unstressed sYl l ab l es ( i . e. non- As a resul t , the consonants al l owed on the second sYl l ab l e of nouns are restri cted to the ones 9i ven below in Row of chart (8) . Noun roots whi ch are terminated by a nasal are 9i ven i n Row b . •• (8) a. -1- -r- -?- -x- b. I -VN r-VN ?-VN x-VN CI . 1 a· f oB <22> • owner , b. svu < 10> · bee ' c*"Hr < 1 > ' orPhan ' CI . 3 a. 9b.I �x� <7> ' wel l ' b. ;'9ura-xo < 12> ' smok e ' n k.r�-I);;' <2> , f al l ow I and ' CI . 4 a. 9b.I � -Y� • wel l s ' b. ;'9Ura-YO k�n-ya ' smokes ' ' vi l l ages ' nk�r l -Y;;' mr n?r n-nY� ' fal l ow l ands ' • stories , CI . 5 ku-d� < 1 4) ' chai r ' a. b. CI . 7 a. cl . 8 a. no examPl es ----- ------ k�n?� <6> ' vi l l age ' mEn?E n-l)i <6> ' storY ' !i�?�-U < 1 > ' l arge basket ' s'?'-nr < 1 > ' Pal m nut ' k�?�-r� < 1 > ' dance ' b. b. fY�xu <2> ' earr i n9 ' fera-mE <2> · ur i ne ' ------ tuxu-r� <2> ' l oad ' ndt)x a-n� < 1 > • seeds , 79 A coupl e of Cl ass 1 di sYl l ab i c roots shown i n chart (9) have a f i nal h i gh b.cked vowel . It was menti oned earl i er that the ProPOsed underlYing suff i x for Class 1 i s -wV, whi ch . i n manY cases aPPe.rs to be total l Y del eted . In the exampl es gi ven .bove, however , it aPPe.rs that the -w- of the suf f i x maY have caused the rai si ng of the f i n.l root vowel before bei ng del eted . as shown bel ow: (9) sarV-wV -> s Wou-wV -> wu • bee ' In Cl ass 3 of chart (9) , the noun roots that Possess . gl ottal .toP aPPe.r to have no suffi x . I t i s possi b l e that the suf f i x coal esces wi th the gl ottal ­ i z ed sYl l abl e. resul ting i n a si n91 e sYl lable ( kAn?A - xA - > kAn?A) . When the root i s marked for a f i nal nasal . thi s coal escence does not take Pl ace: ( 10) *mEn?� mt n?t N - w V -> m'n?E �t ' stor i es ' Cl ass 5 ( I i ) has a number of nouns where the second sYl l abl e of the root i s underl Yingl Y �. However, the conson�t of thi s second sYl l ab l e coal esces wi th the suf f i x consonant , -1- . Produci ng . surf.ce -d-. 3 ( 1 1 ) kurV - I V --> kur -I V --> kud� ' chai r ' We know that there i s an underl Ying -r- by l ooki ng at the plural c lass ( Cl ass 6) of this grouP of words. The i ndef i n i te Pl ural of ' kud� ' is · k�ra-I o ' . --. --- ,�---- . 80 I, The Tpn.} NatYre of Indefinite Npyn. - Tvpe I In i solat i on , the TYPe I i nde� i ni te noun has the �ol lowing tonal Patterns : ( 12) KeY. ( ) CI ••• number . a. Hi gh <6> (L) LoanwordJ < > Number of examPl es in data 1 Syl l abl e 2 SYl lables 3 SYl l ab l es o wirt ( 1 ) <5> H.l a-xi (3) < 1 > C moneY (U ' b. Hi 9h-Mi d < 16> �Yia ( 1 ) (4) ' li sh ' c. Hi gh-Low (5) d, Mi d <42> ' Packet ' C H aar ' e. Weak Mid- (39) c..n ( 1 ) <6> ' market ' <7> C strength (U ' 9ba-xa (3) <36> C dream , ' house ' sa-xe (3) (29) .f ere-me: (8) (5) ' ur i ne ' C bush ' conB ( 1 ) 0 ' orPhan ' ' bee ' 56-1 0 (5) (3) ' ri ver ' � . Mi d-Low saru ( 1 ) (9) j! ( 1 ) ( 1 ) gba ( 1 ) <9> c rock ( U ' o ' Younger si b l i ng ' 9, Mi d-Law W*(42) ji ( 1 ) <7> ' son ' h . Low (77) c� ( 1 ) ( 1 8) ' chi ld ' i . Low-Hi 9h (22) 0 .fu-I"):5 (3) (30) ' i nsi de ' tuxu-ro (7) (5) C l oad ' Pil-l o (5) <50> C body ' gbon-l! (5) ( 1 9) C mat ' nd�r�xg ( 3 ) (3) ' YaIII ' There are manY fewer tonal mini mal Pairs on nouns than there are on varbs. • Mi d Weak nauns di f.fer .from Mi d tone noun. i n that they are mare suscept i b l e to certai n tonal changes . whi ch wi l l be di scussed i n chaPter 5. 81 Thi s i s l i kel Y partial l y due to the �act that there i s mare o� a vari ety of ton.l shaPes on nouns and that the Pl ural i ty o� noun c lasses al ter the se9mental shaPe. The most common mi n i mal Pai rs seem to i nvol ve a Low tone noun. Hi9h­ Mi d nouns are the onl Y Hi gh tone nouns i nvol ved . ( 13) a. Low vs. Mi d-Law kt)d�. k�-ne " seat. the seat ' kud�. kil-ne " road. the road ' n� " co.., ' n3 " mather ' � . sam-b. " oi l ' sami . sim-b. " sorghum beer ' b . LON vs. Hi9h-Mi d "1.l t . se1.-ne • ••1��i . s.1e-�-9f " basket . the basket. baskets. the baskets ' 5'1'"£ . s'1.-n'. s'1�i . s'1e-�-9f " pal m nut . the P. n . , pal m nuts. the p . n . ' c. Low vs. Low-Hi9h I t)l � . l �t)-ne " bi le. the b i l e ' I t)16. I t)�-n' " sh •• nut . the shea n . ' b.l�. b�-ne " seed . one o� ' b�l t. b�ne " ground nut . the 9r . pea ' d . LON VS. Weak Mi d nd�K •• nd�k. " root . the root ' e. Mi d-LoN vs. Mi d k�ra-x�. k�re-k. " i nheri tance. the i nheri tance ' brax:::h bra-k' " br i c k mol d . the br i c k mol d ' � . Weak Mid vs. Mi d saxe. sa-ke. nd� se-ke " bush. the bush . mY bush ' saKe. sa-ke, nda sa-ke " �.ather , the �eather . mY �eather ' A tinY hand�ul a� mi nimal tri pl ets have been �ound whi ch i nvolve morPhol 09i cal tone: ----- -- - . _ . _ - ---- 92 m�n@ ' a sweet ground nut ' Law High m�n' ' the sweet ground nut ' Law Mid une ' sweet ground nuts ' h . Law Mi d sU n ' peopl e ' Mi d-Law sUn ' rel ative , f ami l Y ' Hid-law Mid s1'i n ' Parents ' g. Law-High (Fa l l > Note that there are no minimal Pai rs invol ving Law tone and Hi d tone, though there i . one contrast ing Pai r of Low and Weak Mid tone nouns. Nate i n ( 1 3f > that Weak Hid tone nouns are mare suscept i b l e to tonal changes than are regu l ar mi d tone nouns. These d i fferences wi l l be di scussed in ChaPter 5 . 1 . Th. ton. of the Indef i ni t. Suff i x - TYPe I The f i rst questi on that must be asked concerns the under l Y i ng tonal nature of the i ndef i ni t e suf f i x . UPon exami nation of the examPl es in chart ( 12) , one sees that the suff i x , whi ch is separated from the rest of the ward by a hyphen , can have a Hi gh , Mi d , or Law tone. The tone an the fi nal sYl l ab l e of the root does not seam to be a factor in determi ni ng the tone of the suff i x ei ther 7 for a Low tone suff i x can b. Preceded by a Hi gh , Mi d or Low tone root . A Mi d tone suf f i x can be Preceded by a High or Mi d tone root , whi l e Lew and Hi gh tone can precede a High tone suff i x . Si nce there i s such var i ab i l i ty for the tone of the suf f i x , i t is rather d i f f i cu l t to di scern anY underl Yi ng tene for the suf f i x . We are then faced wi th the possi bi l i ty that the i ndef i ni te suf f i x maY nat be marked for anY particular tone. I f one surveYs the tonal Patterns of i ndef i n i te nouns, one di scovers that the mel odi c Patterns seem to i ncl ude the indef i n i te suf f i x . If a noun i s Low tone, one tonal feature, Low, can be used to l i nk to 83 al l the TBU ' s ( tone-beari ng u n i ts) of the noun , i n c l udin9 the i ndef i ni te suH i x . ( 14) c� � ch i l d ' I L p�-l b � body ' '( If the noun has a contour tone. such as Mi d-Low, the fol l owing Pattern emerges. On a noun with a singl e TBU, the Mi d and the Low tone are l i n ked to the same TBU: ji A � son ' ML When a noun has two TBU ' s. the Mi d tone l inks to the f i rst TBU. whi l e the Law tone l i nks to the f i nal TBU. regardl ess of whether that f i nal TBU is an i ndef i n i te suffi x or not : ' Younger sibl i ng ' conB I I M fu�j I I M L ' insi de ' L A noun wi th three TBU ' s normal l y consists of a d i syl l ab i c noun root and an i ndef i n i te suff i x . I f the tone of the suffi x were simpl y consi dered to be a tonal coPY of the f i nal tone of the noun root , one mi ght exPect a Mi d-Low-Low Pattern. Instead , a Mi d-Mi d-Low p�ttern emerges: tuxu-rb V / M L In fact there are never anY c ases of a Mi d-Mid-Lew tonal pattern to be found on non-complex nouns. Since al l nouns whi ch carrY a Mi d tone and a Low tone have a Pred i ctab l e waY of l i nki ng to the se9ments of the noun . one can suggest that Suci te nominal tone i s mel odi c i n nature . The behavi our of the other contour tones such as Hi 9h-Low. Low-Hi 9h , and ----- ---- 94 Hi 9h-mid �urther suPPorts this mel odi c hYPothesis. Al l o� them have a Pred i ctable waY o� l inking tone to segments and al l o� them l i nk tone to segments accordi n9 to the same pattern described above for Mid-Low nouns. That is. if a noun root with two TBU ' s as wel l as an i nde� i n i te suf � i x Possesses a contour tone. the second tone o� the contour l i nks onl Y to the f i nal TBU o� the word . Thi s i s shown bel ow f or al l contour tones. ( 15) tuxu-rb \I I M L ' l oad ' c'r6-lje V I H • orphan ' nd�r�-x� V I M L ' yam ' H 2. Assaci.tion conventions And the Ind.f i ni t. Noun Now that a mel odi c Pattern for these Indef i n i te nouns has been isol ated . i t i s aPPrOPri ate t o d i scuss the conventions or rul es needed to associ ate the tones to the segments in a predi ctab l e waY. Pul l eYbl ank ( 1983) proPosed the fol lowing version fer the l i nki ng ot tene to tone-beari ng uni ts on the segmental tier : ( 16) " ( 14) Assoc i ati on Conventions : Map a sequence Ot tones onto a sequence at tone-beari ng uni ts. a) �rom l e�t to ri 9ht b) i n a one-to-one relati on . " (p. 3 1 ) Any le�tover tone-beari ng uni ts. he conti nues. wi l l be assi gned tones accordi ng to language sPeci f i c rul es onl Y and anY l ef tover tones wi l l be assi gned a TBU onl Y i f sPec i � i ed by a l anguage speci f i c rule. Thi s apProach is a l i ttle di �ferent from authors such as Wi l l i ams ( 1971 ) who proPose to incl ude the l i nking o� extra tones and TBU ' s as Part o� the uni versal Association Conventi ons. At thi s poi nt, there does not seem to be anY need to posi t seParate rules fer thi s extra l i n king i n Suc ite. Thus. anY extra TBU ' s wi l l be ----- - -- - - - - - ------- --- ---- ------- ------- 85 associ ated to the adjacent tone, whi l e anY extra tones wi l l be l i nked to the .djacent TBU, as shewn bel ew. ( 1 7) 0. ' mat ' JA ' son ' i'lL j� l ' pocket ' Let us consi der the convention of Left to Right Linking. If we, aPpl y Left to Ri ght l i nking to nouns with contour tones. we i mmed i atel Y run i nto Probl ems. It was stated above th.t the f inal tone of the contour is found onl Y on the l .st TBU of the noun . Li nki ng tones from left to right across the word, however , would i ncorrectl y Predi ct that the second tone of the contour is l inked to both the second .nd third TBU of a three TBU word. From the examPle wi th three TBU ' s above ( 17) , i t is cl ear that the second tone of the contour i s agl l inked to the second TBU of the noun . Consi dering that the l .st TBU of a three TBU noun i s a tonel ess i ndef i ni te suffi x , one could devise adjustment rul es in which the second tone of the contour would also l ink to thi s f inal TBU. and then the f i rst tone of the contour would sPread onto the second TBU; f i nal ly, the second tone of the centour would be del i nked free the second TBU, produci ng an acceptab l e tonal Pattern for the three TBU noun. Thi s ProPosed sequence of rul es i s i l l ustrated by the examPle gi ven below. ( 1 8) tux�-ro -> tux�-r� -> tuxu-r� I �/ i'l L Association \tV 1'1 L SPreadi ng � 1'1 L C l oad ' Del i n k i ng Thi s seems to be a cumbersome waY to deal wi th what seems to be a rather simPl e Pattern . However , before we di scard this apProach and seek better al ternati ves, it is i nteresti ng to note that the neighbouring Senufo d i al ect, SuPYire, seems ----- ------ ------ 86 to have evi dence sUPPorti ng the rul e o� spreadi n9 o� the � i rst tone o� the contour onto the second TSU and mot i vati ng a rul e del i nk i ng the second tone o� the contour �rom the second TBU. In the SuPYire case, however , the f i rst tone o� the contour continues sPreadi ng to the � i nal TBU o� the word as wel l , and tri ggers the del inki ng o� the � i nal tone o� the contour, such that the sur�ace outPut of the noun Produces no tonal contour . In �act, there are no sur�ace contour tones on i nde� i n i te nouns i n SuPYi re. However , tonal behaviour i ndicates that certai n nouns do have underl Ying contour tones. Let us consi der one exampl e o� thi s type o� sPreadi ng in SuPYire. What are cal l ed Low-Hi gh nouns i n Suci te are referred to as Low-weak Mi d by Carl son ( 1985) in SuPYire. He moti vates a total Low SPreadi ng rul e where a Low is sPread onto the fol l owing TBU whi ch is l inked to a weak Mid tone. Thi s weak Mi d tone i s then del i nked as a resul t o� Carl son ' s Low-Mid Si mpl i f i cati on Rul e . Th. LOW-SPreadi ng and LM SimPl i � i cati on Rul es are then rePeated �or anY remai n i ng TSU ' s i n the word. unti l al l TSU ' s i n the word are l i nked to Low tone. ( 1 9 ) a. tUU-90 I V L Mw --> tUU-90 --> tuU-90 --> �V L Mw L-SPread � L Mw LM-Si mpl i f i cati on tuu-� --> tUU-90 \L / ( \ Mw L-SPread � L Mw UM-Si mP. Carl son al so shows how this Pattern o� rule appl i cati on works �or Mi d-Low One wonders, however . whether this i s a viab l e anal Ysi s �or Suci te. I� we moti vated these rul es �or sPreadi ng and del i nkin9 i n Suc i te. we wou l d have to b l ock thi s SPreadi ng and del i nkin9 �rom aPPIYi n9 to the � i nal TSU of the noun . for unl i ke SuPYire. the fi nal tone o� a tonal contour is Preserved on the f i nal - �-- -.--- ---- 87 TBU o� the noun . The secand Prabl em wi th this apProach invol ves data that wi l l b e deal t wi th i n Chapters 4 and s. Since a Mid tane sPread i ng rul e and a Low tane sPreadi ng rule can be mati vated el sewhere i n the language, i t i s canceivabl e that they cauld also be moti vated ward i nternal ly. However , a · High tone sPread i ng rul e cannot be moti vated in Suci te (thaugh i t can in SuPYi re) ; ther�are i t woul d be somewhat cumbersome to al l aw a Hi gh sPreading rul e word i nternal l y �or Hi gh-Mi d and Hi gh-Law nouns whi l e b l acking i ts aPPl i cati an el sewhere i n the language. An al ternati ve sol uti an is to �are9O the l i nk i ng o� tanes to the segments unti l the addi ng a� the i nd� inite su� � i x � and then ta l i nk the tanes �ram right to l e�t acrass the ward� as seen below. 'J tuxu-rd M I c l aad ' L '� I nd�r�-xg ' Yam ' L H This apPraach a� Ri ght to Le�t Li nking seems to be a much mare e�� i ci ent waY to deal with basi c tonal patterns a� nouns i n Suci te. It el i mi nates the need �ar a series a� rul es ward i nternal lY� and al lows the natural tonal Pattern a� nouns to be rePresented i n a simPle and uncomPl icated waY. In add i t i an , the arguments Presented here wi l l be suPParted by mare evi dence in the ensu i ng chaPters. 3. Ton• .nd Pr.-n...l iz.d Cansonants It was menti oned i n ChaPter 1 that there ex i sts a set a� nauns wi th eca: nasalized bgnsgnants. Care�ul scrutinY o� the tonal behaviour � nouns with Pre-nasal i z ed cansanants revealed that some at these cansonants bear a Low tane whi l e athers do nat seem ta b� marked �or anY parti cular tone. The di st i nctian ----�- --- ----------. - . - - - ---- ------ 88 between those that carrV a Law tone nasal and those which do nat is barel v Percepti bl e i n ci tation �or=. However , i n connected sPeech , thi s d i stinct i on i s made clearer . The l aw tone o� a nasal i s o�ten l inked to the � i nal svl l ab l e of the Precedi ng ward (see 2 1 a and c ) . whi l e those nouns that do nat carrV a Law tone pre-nasal have no aPParent e�fect an the tone of the Precedi ng ward (b and d) . (21 ) a. nda �k6nA -> nd� n k�nA ' mv waaden bawl ' b . nda m9bf n@ ' mv bamboo mat ' c . ndi va �k6n� nvAA -> ndi v� nk6nl nvAa ' I am seein9 a bawl ' d . ndi va m9bf ne nvAA ' I am seei ng a bamboo mat ' Ward i ni ti al nasal consonants are al sa sameti mes �ound to be tone beari n9. Thi s i s rePresented bv doubl i n9 the nasal in same cases. thou9h i n real i tv . l on9 nasal s are nat alwavs percepti ble. Below are l i st � examPles showi n9 that a Law tone nasal can be found precedi n9 Mid and High i n i t i al nouns. (22) a. L-Mi d S b . L-Mi d W c . L-Hi 9h-Mi d I1IPi (2) ' rabb i t ' I1Ima ( 1 ) ' 9i H ' llIbi (8) ' Hour ' Ma-ra (7) �nva ( 3 ) ' 9rass ' '')m::l-b (5) ' knite ' "nvo (3) ' mouth ' �da-xe (3) ' ear ' Ma ( 1 ) ' scorPion ' �n6-lu (5) ' 9ui nea hen ' ' ryU ( 1 ) ' twi n ' iftpd-lu ( 1 ) • sPi der , d. L-Hi9h 'z'b@ ( 1 ) ' 9Dod luck charm (Loan ) , e. L-Hi 9h-Lew "kana (3) • ' l and ' wooden bewl ' If one aPpl i ed the Ri ght to Le,t Linking Conventi on to Law initial werds 89 with prenasal ized consonants, however . an i ncorrect surface form would be Produced� as shown bel ow. (23 ) *"j�'ne \1 / I L H /II ·"U de rather VI "tide ' bat ' L M Thi s probl em cou l d b. easi l y remedi ed wi th the suggesti on that tone-beari ng nasa l s be sPecif i ed for Low tone and that thi s Low tone i s l i nked to the nasal b�ore the appl i cation of the Assoc i at i on Conventi ons. The l ow tone is l i nked to the nasal and the remai nin9 tones are l i nked from ri ght to left. 4. Hi;h final naun. The Preceding naun charts have i ncluded Hi gh f i nal nouns. However , i n ciatati on form or i n Phrase f i nal Posi t i on these Hi gh f inal i ndef ini te nouns, namel y, High and Lew-Hi gh tone nauns, exh i b i t a f i nal fal l i ng tone. ( 24) a. nd�r�x� ' yam ' ' rock ' Indef i n i te nauns which acquire a f inal Hi gh tone through rule aPpl icati on al so exhi b i t a f i nal Hi gh fal l i ng tone. b. soxo ' mortar ' ' man ' s mortar ' Poru s6x6 ' daughter ' s mortar ' Thi s fal l i ng tene di saPPears, however, when the hi9h tone i s not in Phrase H nal Posi ti on . ----�- -�- --- ----- ------ ------- ---_ . 90 (25) waa nd�r�x� wee waa nd�r�x� nY� ' he l ooked at a yam ' ' he saw a yam ' In contrast, Hi gh tone verbs and Postposi tions, and Hi gh � i nal de� i n i te nouns have no f i nal fal l i ng tone. The natural first assumPti on that one i s l i kel Y to make is that fi nal Hi gh f al l i ng nouns actual l Y have a f i nal Low tone. In order to test thi s Possi b i l i ty, one must comPare the behavi our of these nouns with those that are known to have a Low f i nal tone. Low � i nal nouns are known to tri gger Low tone sPreadi ng onto Hi gh tone verbs (see ChaPter 4 � or detai l s) : (26) a. waa m�l � wee -> waa m�l � w.e However . fl14xa and nd�r�x' ' He looked at r i ce ' do not tri9ger Low tone sPreadi ng: b. waa nd�r�x� wee -> *waa nd�r�x� w�e ' he l ooked at a yam ' Therefore. one i s l ed to concl ude that the �al l i ng tone exhibi ted on Hi gh fi nal nouns i s a Phoneti c Pecu l i arity whi ch is mani fest onl Y when a Hi gh f i nal indef i n i te noun i s i n Phrase f i nal Posi tion8• I n order to characteri ze thi s Phrase f i nal tonal adju5tment : a Low tone i nsert ion rul e maY b e f ormul ated : (27) LOW TONE I NSERTION: Insert a Low tone to the f i nal TBU of a High fi nal noun when it is i n Phrase f i nal posi t i on . Thi s Low tone Insertion woul d effecti vel Y create a fal l i ng tone on a Hi gh fi nal -- - - .- - - -- - - ---- .---- ---- ------ noun: � �� -> ��� H 91 " rack ' \ H L I I I . THE INDEFINITE SlFF I X - TYPE II (Cb.... 2 ."d 6-) A, structure We hAve just examined the tanal behaviour o� nouns whi ch use the TYPe I i nde Hni i:e suff i x . The nouns o� Cbsses 2 And 6, however , use a di Herent tyPe o� suf f i x , l abel ed here as the TyPe I I i nd� i n i te su� fi x . As plural cl asses, they have the same noun roots as Cl asses 1 and 5, resPecti vel y . Al though thei r indef i n i te su�f ix es are underlYi ngl y disyl l Ab i c , the surfAce farm is often monosYl labi c . Tonal ly, the two class su�� i xes behave i n the same waY. The ini tial tone � both suff i xes al ternates between Hi gh , Mi d and Law tane dePendi ng an the tonal envi ronment, whi l e the � i nal tone tends to be a mare stable Mi d tone. Because o� the cOmPl exi ty o� bath the se9mental and tonal Processes. I have chasen to di scuss the segmental ChAngeS f i rst be�ore deal i ng wi th tone. S . Cl ••• , Indefinite Buff i N . in SuPYire. i t i s :iil1 or I n Cebara. the Cl ass 6 su�f i K i s :SIla. whereas �. In Suci te, we hAve a vari ety o� su��i x i n9 whi ch probabl Y can be attributed to the same underl Yi n9 morpheme. -xVI V. In certai n cases. descri bed below. the vowel i s h i 9h . How vowel hei ght can be accounted �or usi n9 the Present underlYi n9 rePresentat i on . -xVlV. wi ll nat be di scussed here. (28) CLASS 5 ExamPles with the �ul l di sYl labi c suf� i K are gi ven bel ow: CLASS 6 <24> a. Pil-l� P�x al o " bodies. trunks ' b. ci-It} c�-xal e " thi9hs ' c . nci -I e nc� -xal e " bal aPhans ' 92 Someti mes the fri cati ve vel ar is gl ottal ized : (29) CLASs 5 CLASS 6 < 1 1) a. co-l � c�-7al o ' clay Pots ' b. ��-l� ��-?al � ' debts ' ManY times. however, the d i syl l ab i c su� � i x i s Part i al l y del eted or Practical l y di sappears al together . A �ew nouns l ose the � i nal sYl l able o� the sut� i x . � i n a Process that shal l be cal led SUFFI X FINAL DELETION. When thi s haPPens. the sut� i x vowel i s always high. (30 ) tLASS 6 <4) SFD a. k�r� - xvl v -> ti - xVlv -> k�r�xi ' seat. chairs ' ti t-x i ' cock ' s combs ' In certain cases. thi s shortened 5Ut � i x i s al so nasal i zed . Note that nasal i za­ tion is not tri ggered necessari l y by a nasal environment : CLASS 6 < 1 2> -> m�d-ryi ' hi l l s ' ntan - xvlv -> nt��-ryi ' courtyards ' b . mPu - xV1 V j3r3 - �vl v -> j�-r�ryi ' breast s ' It i s i nterest i ng t o note that when sut � i x t inal del et i on takes pl ace, singl e Tau roots comPensate for the l oss o� a syl l ab l e by lengtheni ng the root vowel to two TaU ' s (see (31a> , whi l e roots wi th two underl Ying TaU ' s do not change (31b > . (31 ) a. Sin91 e Tau roots mP�-d-ryi C hi l l s ' ti-t-x i c cock ' s combs ' nU-�-ryi ' courtyard ' 93 b . Double TBU roots j�r�-�i ' breasts ' s�7�-�i ' bAsket ' s'76-�i ' Pal m nut ' The � inal resu l t i s that al l o� the above �orms Possess three TBU ' s. A second l ook at the data reveal s that al l Cl ass 6 noun roots wi th twa underlvi n9 TBU ' s seem to requi re some kind of su�+ix del eti on , i n order to maintai n the l imit o� three TBU ' s lor a TVPe I I noun, whi l e sin9l e TBU roots wi th Partial l v del eted su��i xes require a l en9thened vowel to keep uP a three TBU mi nimum. SUFFI X INITIAL DELETION is another waV to reduce the size o� the su�fi x . Some di svl l ab i c noun roots wi th a second svl l ab l e of the shaPe, �, tri 9ger the del eti on of the � irst svl lable o� the su� � i x i nstead of the second. (32) CLASS 6 (8) SID kir� - xVlv -> kl-ra-le ' countries ' c�r� - xVl v -> c�-ra-le ' e99s ' There are also a tew examples where the enti re i ndef i n i te Plural suf f i x i s reduced t o a sin9l e vowel ( SUFF I X REDUCTION) : (33) CLASS 6 <4> SUFFI X RED nvi - xvl v -> nvl-i VE - xvl V -> Y�e ��n - xvl V -> f�-un ' eves ' ' Years ' ' peanuts ' Nate that thi s set o� nauna does not requ i re a f i nal outPut of three TBU ' s. as is the case for ather �orms � the Cl ass 6 suff i x . ---- �� - .-- --- ------ 94 Fi nal l y , there are a �ew CAses where the enl Y indicati en e� pl urAl i ty i s the tenal change. In the examPl es bel ew , the Class S, er si ngular �erm, i s cemPared wi th Class 6: (34) CLASS S CLASS 6 <3) mane ' sweet greund nuts ' bal e " greund nuts ' We wi l l net attempt te �ermul ate rul es here te acceunt �er .u��ix reductien and P�t i al su� � i x del eti en . Seme Senu�c d i al ects maintai n the entire su� �i x , whi l e ethers reduce i t t e a min i mal uni t . There maY b e seme synchrenic Phenel D9i cal l Y cendi tiening rul es �er 5u� � i x reductien , but , �rem these eXamPles, i t i s net clear that anY such rules ex i st . Hewever. i t wi l l be seen shertl Y that the � i nal tonal shaPe e� a word wi l l dePend en the type e� su�� i x del eti en . 2. CI ... 2 Ind.finit. Su�f i x . The Cl ass 2 i nd.� i n i te ( pl ural ) suf f i x i s hi sterical l y disyl l ab i c alse. but i n Sucite. the � i nal eutput is menesYl l ab i c . In Cebara, e� IverY Ceast , the pl ural ferm i s bala • I n SuPYire, the Cl ass 2 suf f i x is ei ther -I i . er -mi l i . In Suci te, the mest cemmen �erm e� su�fi xat i en i s -IV (PerhaPS threugh SUFF I X INITIAL DELETION) : (35) s6nl u - CVI V -> s6n l u - I V -> s6nla-l� " parakeets ' Balew are ether examPl es i n compari sDn wi th their si ngul ar ceunterParts e� Cl ass 1 . ( 36) CLAsS 1 CLASS 2 <38) P4an-I a ' al l i gaters ' fy a-l a " f i sh ' --------�- ----- --------- 95 s6n l u s6n l a-lj ' parakeets ' Poru PDr'-lo ' daughters ' Nouns wi th singl e Teu roots exhi bi t a l ang vowel . The second TaU o� the l ang vowel maY be a resu l t of comPensatorY lengtheni ng of the root vowel or it is Possi b l e that i n these cases, onl Y the in itial consonant was del eted . (37) ja - Cvlv -> ja - IV -> ja-V-Vl v -> ja-a-l a SUFFI X INITIAL DELETION ar VOWEL LENGTHENING. ja -cVl v -> ja - Vl V -> ja-ala a. j� j66-lj ' packets ' b . ji jatl- l a ' sans ' CONSONANT DELETION For reasons that maY seem arbi trarY for the moment, I shal l chaa5e the rul e of Su�f i x Ini tial Del eti on where bath the consonant and the vowel are del eted . If we chase Consonant onl Y Del eti on , the noun roots wi th twa TeU ' s ( 36 ) , wou ld require an add i t i onal rul e o f vowel del etion . There are examples indicat i ng that the under l Ying ( or h i storical ) farm of the suf f i x maY have been :mili or �. These nouns have droPPed the � Part of the suf f i x < SUFFIX FINAL DELET ION) , but retai ned the f irst Part i n the farm of ::m1 • (38) CLAss 1 a· CLASS 2 < 1 0> SFD "nn.-mi ' scorPi ons ' b. ctnr')t cUn-ai ' antel oPes ' c. carar')£ car�mi c orPhans ' i\naa Agai n . as i n Cl ass 6 above , there are a few cases where the onl Y i nd i cat ion of indef i n i te i s the add i t i on of a tone-beari ng vowel (SUFF I X REDUCTION) . ----�- - --- ------- ---- -------- - - - --- ----- 96 (39) CLASS 1 CLASS 2 a. n3 nH SUFFIX RED " mothers ' " cwners, ch i e� , head c� c. nc� nctla ' sheep ' It shou l d be ncted that i n Kan9ala, a vi l l age situated 5 ki l ometres �rcm Kctcura. the speakers have nct al lcwed Suf f i x Reducti cn . Here, i nstead c� saYi ng £ia ' cllWlers ' , they ret:ai n the :Ul suf f i x : :£Bl..R. 3. au..arY. The fol l owing chart. Provides a summarY of the di�ferent tYPes cf suf f i x i ng fcr TYPe I I Indef i n i te Nouns: (40) SyffiX TVPe elas$ 6 Cl ass 2 Fu l l pc.-xa l c <24> ' bcd i es ' ------- cc.-?al o < 1 1 > ' d ay pots ' ---- Suf f i x Fi nal DEL t H -x i <4> ' ccck ' s combs ' -- --- --- ntU-l')i < 12> ' courtyard s ' d � n-mi ( 1 0) ' antel cpes ' k� ra-I e <8> ' countri es ' jatl-Ia <38> Suf f i x Reducti on nyl- i <4) ' eyes ' nl-i Tone Change On l Y m�-ne (3) ' sweet ground peas ' -------- Nasal SuH ix Ini U al ' sons ' Del et i cn ' mcthers ' Si ngl e Teu noun roots can retai n the enti re suf f i x as wel l as undergo anY tYPe of suf f i x del eticn . Nouns rccts with twc TBU ' s obl i 9atcr i l Y under90 ei ther Suf f i x Fi nal Deleti on or Suf f i x Ini tial Deleticn but never Suf f i x Reduct i cn . Other than these general 9u i del i nes , hcwever , there dc nct seem tc b e anY Phonol ogical reasons 90verni ng the choi ce of the shape cf the suffi x . .- - . � .- ----- ----- 97 B. TM. of TYpe II Indeftnit. Suffix" The two Type I I Inde� i n i te 5u�� ixes share the same tonal behavi our and thus seem to have the same under l Yi ng tane. There are several �actors that contri ­ bute t o t h e tonal changes o � the su�� i x . Fi rst o� al l . t h e tone at the noun root does attect the tone Ot the sut? i x . Secondl y . the segmental changes on the sut ?i x at?ect the waY tone is l i n ked to the word . Fi nal l y , i t wi l l be seen that the segmental nature at the sutt i x actual l Y contri butes to tonal change at the noun root. The i nterpl ay at these three tactors can create a rather compl i cated tonal si tuat i on . The �ol lowi ng i s an attempt t o d i scuss each seParatel y and al so i n con juncti on with the other tactors. 1. Ton.l SIIh.viaur oF the Suffi x Let u s f i rst exami ne how the tone at the noun root at?ects the tone at the TYPe II suf � i x . Basi cal l y , it the root is Low tone, the sut t i x tone is Mi d . It the root i s High. Mid . or Low-H i gh . then a High-Mid shaPe 9eneral l y evolves. Atter Mi d-Low roots the tonal shape at the ward i s Mid-Low-Mid-. In the chart bel ow, the si n9u l ar t orm at each noun i s also gi ven , under the label , TYPe I . The si ngul ar torm Ot the nouns gi ve a clearer i dea at the underl Ying tone ot the noun root: (41 ) Root tone ----- -- TYPe I TYPe I I Engl ish ' body, bod i es ' a· Low p�-l�, P�-xal o b . Low-Hi gh gbdn-1 8, 9bOn-x'l o ' granarY, 9ranar i es ' c · Mi d gba. 9ba-'la'"' ' ri ver , ri vers ' d . Hi gh m9bf -n@, m9b! -n'le ' bamboo mat , bamboo mats ' e· Mi d-Low j i, ja-U ae ( san. sons ' _ . _ ._ . . _-- ------ 98 In observi ng the examples in ( 4 1 ) one di scovers that the onl Y envi ronment in which a Hi gh-Mi d tonal shape i s not found i s when the f i nal tone of the root is Low tone. in whi ch case the tone of the suff i x is M id tone. If an under l Ying High-Mi d suf f i x is Posited . then a rule l oweri ng a Hi gh tone to Mi d tone must be sough t . In Chapter 2� a rule sPread i ng Low tone onto a f ol l owi ng Hi gh tone was proPosed . whi ch i n certai n cases. resulted i n a Low-Hi gh contour si mpl i fyi n9 to Mid tone. Thi s same Low tone SPread rul e (33 ) can be used here. Consider the exampl e� pb=xalg ( 41 a) . If the underl Ying suf f i x tone i s Hi 9h-Mi d � Low tone can then b e moti vated t o spread onto the High tone of the suff i x . as shown i n the deri vati on below. t j j -> Pt:xfl , -> P�-xal o ( 42) P -x l L H ili L ' bodies · H ili L SPREAD The lack of Hi gh tone in ( 4 1 e ) can al so be eXPl ained by the aPPl i cat i on of tha Low SPread rul e . The noun 9b�n-x'lg ( 4 1b) aPPears to have a Low tone on the noun root . However � a l ook at the tone of the root in i ts singu l ar form reveal s a Low-High tone. The resul ting surface pattern can be eXPlai ned if we apply the Hi gh Del inking rul e (37) ProPosed i n Chapter 2. where the Hi gh tone of a Low-Hi gh contour i s del i nked i f i t i s fol l owed by another Hi gh tone. The aPpl i cation of th i s rule i s i l l ustrated below. ------ - ----- - ----- ---- ------_ . _- _ . -- -_ . .. . - -- _ . .. _ - 99 H DELINKING Segmental del et i on � the su�� i x compl i cates the anal Ysis scmewha� . however . i n that scme of the su��ix tone i s l i nked to segments of the ncun root. Th i s maY b e con� us i ng when trYing to determine the d istinct i on between the under l Yi ng tone cf the noun root and the tone o� the suff i x . The f ol lcwing di scussion wi l l hel P to clar i fy these amb i gui t ies and wi l l al so attern�t to establ ish a �ormul aticn �or l i nk i ng tones to ncuns wi th partial l ¥ del eted inde� i n i te suH i xes. The chart below Provi des examPl es � what haPPens when the var ious tones of the noun rco�s meet uP wi th the vari ous segmental tYPes of suff i x i ng. Sinc� the the number of TBU ' s on the root is a factor a�fecting �he tonal behaviour. examples of noun roo�s wi th single TBU ' s and doubl e TBU ' s are gi ven. An as�eri sk indi cates tha� there are no accePtabl e examPles for that particu l ar categorY. whi l e a dotted l i ne indi cates that whi le i t i s concei vab le that there ex i sts an exampl e. no exampl e has been found to date. In addi tion . this chart shows that some noun roots are l owered �o Low tone in certai n envi ronments. Thi s phenomenon wi l l be di scussed l ater in the chaPter. These l owered noun roots are marked in bol d . ---- --- - - -- - ---- ------- - - -- -- 1 00 (44) Tone #TSU FULL SUFFI X SUF INI DEL SUF F IN DEL SUF RED LOW 1 TSU P�xal o co-o-I o I)m��-I)i ncA-a ' bodies ' ' nets ' ' corners ' kira-I e j �ra-I)i ' countri es ' ' breasts ' • 2TSUs LOW-HIGH lTBU gbOn-x'l o - - --- * - --- mPC)-6-l)i .f:�-un ' hi l l ' ' granM"ies ' 2TBUs ' sheep ' mM6-l a ' peanuts ' pA16-mi ' el ePh�nU asi s ' ' bucket ' MID lTSU la-x'la . U-x.l. ' Pregnanci es ' 2TBUs • gbaal a ' ri vers ' cal 6-1 a .£-i-I)i - - - - - ' sU ngs ' --- -- - -- ' Pi gs ' WEAK MID 1 TBU .f:o-? l o --- -- - - - - - ---- - - ' debts ' MID-LOW l TBU 2TSUs HIGH 2TBUS d-xal. j a-�-l a ' th i ghs ' ' sons ' * n j i ra-Ie ku-c)-x i ' tongu e ' ' roads ' m9b{n�le - * - - -- - n1-1 ' mother5 ' -- - f�. ' ch i eofs ' - - * ' bamboo mat ' HIGH-MID 2TBUs * fri.-l a car6-mi ' f i sh ' ' orPhans ' * In the d i scussion of TYPe I Indef i ni te noun tone, it was su99ested that the TYPe I i ndef i n i te suff i x be added to the noun root before the aPpl i cation of 101 the Associ at i on Conventi ons , so that the f i nal tone o f the root cou l d l i nk d i rectly to the tonel ess suf f i x . TYPe I I suf f ixat i on can al so take Pl ace before the aPpl i cati on of the Associ at i on Conventions. The crucial questi on to be asked here is. however , shou l d se9mental del eti on a lsa take Pl ace before the aPpl i cati on of the Assoc i at i on Conventions? For manY cases, thi s is a verY Practi cal approach for assi9n i n9 the correct tones to se9ments. After the deletion of the vari ous se9ments. it is a simpl e matter to l i nk the tones from ri 9ht to l eft in a one to one fash i on across the word . Hi9h . Low-Hi 9h , and Mi d tone nouns are espec i al l y amenabl e to thi s approach , as shawn i n the rul e deri vations bel ow. (45) a. Hi 9h tone root - suff i x under90es Suff i x Ini t i al Deleti on ' bamboo mats ' SUF INI DEL Ac (ASSOCIATION CONVENTIONS> b. 2 TBU Mi d tone root - suf f i x under90es Suffix Ini tial Del et i on cal a -CVlV M - > cal a - H M M I e -> H M SID c. 1 TBU ca U - la � t I AC Mi d tone root - suf f i x under9Des Suf f i x Final Del eti on ' sti n9s ' SFD V LENGTH' AC 102 d . 1 TBU Low-Hi gh tone root - suffix undergoes Suf f i x Fi nal Del eti on mptl -�VIV -> mPtl - �f -> mptl u �f -> LH H M LH H M LH mPtld�i -> mpud�i Al l HM AC V LENGTH SFD 41 1 LH H M LH H M H DELINKING Note in (45d ) that the rule of Hi gh tone Del i n king al so takes Pl ace. si nce. after the appl i cat i on of the Assoc i at i on Conventions. a Low and a High tone are l inked to the same TaU and are fol l owed by another High tone. Low tone noun roots whi ch have undergone Suf f i x Ini t i al Deleti on al so conform easi l y to l i nk i ng tone to segments after segmental del eti on of the suffix . Note in the examPl e bel ow. that Low tone sPread is tri ggered once the tones have been l i nked . e. 2 TaU Low tone root - suf f i x undergoes Suf f i x Ini tial Deleti on k � r� - CVl v -> k � r� - I e -> k� r�-le -> k � ra-l e L H M L H I I M L,1 L H I 1'1 L SPREAD Low tone root - suf f i x undergoes Suf f i x Ini tial Deletion c� - CVlv L M AC SID f . 1 TBU I L H • countri es , H M --> c� -1 0 --> c� L H M SID L 0- - 1 0 --> c�-o-lo -> cool o H M V LENGTH I I / L H M AC �1 I " nets ' LH M L SPREAD Certain Low f i nal nouns. however . have Probl ems with aPplYing the Associ ation Conventi ons after segmental del eti cn . The deri vati ons below show how such a rul e ordering produces i ncorrect resu lts. ----- - -- - -- - - - - - - -- - ----- - - - - -- --- ------- - - -- - -- 103 (46 ) a. 2 TBU Low tone root - su�� ix undergoes Su�� ix Fi nal Del et i on . ' countr i es ' L H 1'1 L Rather: jM�-"i H I'I AC SFD b . ct n - mVIV -> C E n - mi L H M L L SPREAD H M SFD L H M L L SPREAD AC ' sheeP. pl . ' H M ML HM SID Rather : cttnmi Rather : nc4a At d. ja Cvl v -> ja -la -> j a a la -> j i'l a ->*jaal a MI. HM LA M L HM SUF RED L H M LENGTHENING A\ ' antel opes ' c t m H M c . nca - CVI V -> nca- a ->*nc:aa I. 1 , j -> .ct�t -> ci-En-mi -> ML HM LENGTH Ai I A\I ML HM MLH M At L SPREAD Al l o� these counterexamPles consi st o� AC a�ter segmental del eti ons Produces an a ' sons ' Rather: j a4l a Low � i nal root � the aPpl i cati on o� i ncorrect Mid tone on the penul t i mate TBU, rather than the approPri ate Low tone. It might be suggested that the correct sur�ace �orm coul d be arr i ved at by using a High Del inkin9 rul e a�ter Low tone sPread , as shawn bel ow. (47) jM�-"i �4 I L H M The Prob lem with this proPosal , however , i s that there i s alreadY a Hi9h Del i nkin9 rul e (37) , wh i ch works i n a verY restri cted envi ronment . The rul e ---- ----- - --- - --- --. --- ------ 104 sti pul ates that a Hi9h at a Low-Hi9h contour is del inked it it is tol l owed by a Hi 9h tone ( See Chapter 2, p. 54) as shawn i n ( 4Sa) bel ow. It was al so shawn that betore Law and Mi d tones the High at a Low-Hi 9h contour was Prohibi ted tram del i nkin9 ( 4Sb ) . �66 vt V (49) a. wu Ya x, k ��n I I M I M H LH ' He i s 9ivin9 i t to h i m ' H b. wu Ya x, k�an l a \ I M I M H �1 ' Is h e 9i vi ng i t? ' I LH M There�ore, i t one tried to moti vate Hi gh Del i nk i n9 i n the setting i l l ustrated in ( 47 ) above, this attemPt wou l d contrad i ct the restrict i ons set tar the High Del i nkin9 rul e . It , however , t h e Association Conventions aPpl i ed b�ore segmental delet i on . the correct surtace torms woul d be Produced t or the examPles bel ow. (49 ) a. k��-�vl v -> kar�i I L I I I H M AC L I - SFD I I L H M I L SFD c . nC L H M L SPREAD I -> c�-t n-mi H M I L H M L HM SUF RED I LH M L SPREAD ---- ---- - -- -- --.-- --- ------ - H M LENGTHENING j - CYl ! -> nCi- )\ -> n/� L AC l--1 H M b . c�n - mVI V -> c�n - mi I > k�r�-ryi > c��nmi -> �'� 1 L # H M L SPREAD ' sheep. pl . ' c££n-mi ' antel opes ' � L HM 105 Si nce orderi ng AC before Segmental Deletions seems to work so wel l i n these cases. l et us test this ordering on other TYPe I I nouns. I n general . i f tones are l i n ked before segmental del etion . once segmental del eti on and subsequentl y . vowel l engtheni ng take Place. there are free tones and TBU ' s that wi l l requ i re l i nk i ng. Although the manner of l i nking i s def ini telY predi ctable. the rul es needed to moti vate correct l i nk i ng are not immed i atel Y cl ear and the tonal and morphologi cal si tuat i on is comPlex . One reason for the compl ex i ty of ten.l Patterns i s the morphel ogi cal nature of the suf f i xes. Neuns wh i ch have undergone SID l i nk thei r tones i n a sl i ghtl y different waY than nouns whi ch have under9Dne SFD. In addi tien . Mi d tene nouns whi ch have undergene SFD, behave in a d i fferent waY than SFD Lew tene nouns. Fi nal l y, i t i s neted that Patterns isel ated for the Previ ous greuPs ef neuns are vi el ated by certai n nouns whose reot tone is a contour tone. The anal Ysi s bel ow attempts te account for these varyi ng behavieurs. It shoul d be neted that Hi gh-Mi d and Hi gh-Low neuns have been excl uded from this Present anal Ysi s because ef thei r somehwat unPredi ctabl e behavi our . A descr i Pt i on of their behavi our wi l l be gi ven fel lowing the anal Ysi s. Let us f i rst consi der these neuns whi ch have undergone suf f i x Initial Del eti on . Bel ow are examPl es ef nouns which have al readY undergene S I D and Vowel Lengtheni ng but have net yet had tones and TBU ' s reI i n ked . The f i nal surface f erm is gi ven alongsi de each examPle. (50) a. 9ba-a-l a -> gba� a J M I H M c · ki ra- I e -> k� ral e / L ' ri vers ' I H M b . cal a- l a -> cal al a V M ' countri es ' I H M d . ce-o-l o -> c�el o , L ' Pi gs ' I H M ' Pots ' 106 e. mgbi na- Ie -> mgbf nal e / H I c bambee mats ' H M In each case_ we must deal wi th a hi gh tene wh i ch has l est the se9ment te whi ch i t was l i nked . In the eXamPl es (a) and (d) . an extra TBU has been created through Vewel Lengtheni ng. Since beth a free tene and a free TBU are feund adjacent te each ether_ the Asseciatien Cenventiens can be reaPpl i ed at this Pei nt. l i nking the twa. (51 ) gba-�l a I � I M H M The ether exampl es have ne free TBU ' s_ hewever . Judgi ng frem the f i nal eutput ef al l ef the eXamPl es_ it is cl ear that Hi gh tene dees net l i n k te the f i nal TBU. Al l f i nal TBU ' s are l i nked enl Y te a Mid tene. In C50b ) there i s cl ear evi dence that the Hi gh tene i s l i nked t e the secend TBU. The deri vati en beleN shews that the Mid tene must al sc be del i nked from the second TBU i n erder te si mpl i f y the Mid-Hi gh centeur. (52) caU-l a � I M H M At thi s Pei nt. I wi l l net di scuss the rule ef Mid tene Del inking excePt te nete the M. bei ng an the l eft s i de ef the conteur del inks frem a TBU l i n ked te Hi gh when i t al readY l i n ked to the Precedi ng TBU. (53) MID TONE DELINKIN6: x X M H � 1 07 The sur�ace �orm � the examPl e i n ( 50c > does not show a High tone on a second Tau, but rather a Mi d tone. Thi s can be expl ained i f the �ree Hi gh tone i s l i nked to the second Teu and the Low tone i s agt del i nked . as shown bel ow: (54) Recal l that in ChaPter 2 a Low-Hi gh contour whi ch was created by means of a Low tone sPreadi ng onto a High tone. si mpl i � i ed to a Mid tone on the surface. In the examPl e in (50d ) , the Hi gh tone i s l i n ked to the �ree second Tau by the uni versal Pri nci ple of tone Associat i on . The f i nal surface Mi d tone on that TBU can be eXPl ained by tri ggeri ng Low tone to sPread onto the Hi gh tone TaU. as shown bel ow: (54) c� 610 -> c� al a I I I �1 I L H M L H M LOW SPREAD Final l y. the free High tone of the examp le i n ( 50e) can al so be l i nked to the second Tau, though in this case, the l i nking i s hardl Y necessarY. since there i s al readY a Hi gh tone from the ncun root l i nked to i t . (55) mgb! na-l e l/. MI H H Al l f i ve of these examples, then, undergo a Process where the free High tone l in ks to the second TBU. Si nce thi s i s not. i n al l cases . Predictab l e by waY of the A$soc i at i on Conventi ons. a rul e must be formu l ated . speci fYi ng the l in k i ng. In attemPti ng to formulate a rule, one maY Pose a quest ion as to whY the free High tone l inks �o the l ef t rather than to the r i ght. It has been suggested in --- ------ -- -- -- ------- --- - - 109 the l i terature (abtained through Persanal communi cati on wi th c. Ki sseberth ) that tane wi l l gravi tate i n the di recti an of the tri gger af the segmental delet i an ru l e . In thi s case, the l eft sYl lable af the suf f i x was del eted . As a resu l t , the Precedi n g TBU shoul ders the resPansi bi l i t y af carrYing the TBU of the suf f i x . Taki ng i nta cansi derati on these observati ons, then , the rule can be stated as fal l aws : (56) SUFFI X H IGH LINKING: Link the free Hi gh tone af the TYPe I I suf f i x ta the f i rst TBU ta the l ef t . X \ H X I M The secand set of nauns to di scuss are these whi ch have undergane Suf f i x Fi nal Delet i an . I n these Part i cul ar examPl es <57) , i t is the Mi d tane rather than the Hi gh tane whi ch l ases i ts se9ment . Unl i ke the Hi gh tane o f the Previ ous exampl es, the Mi d tone here has no choi ce af segments to l i nk to. and thus i t l i nks ta the same segment as the Hi gh tane. a. sE -E -�i -> sE € �i ( 57 ) /\ I M ' st i ngs ' H M b . �me-e-ryi -> �m�e�i I " L ' carners ' H M c . jara-�i -> jara�i I II L ' breasts ' H M We knaw, hawever , that the f i nal outPut daes nat al l aw a HM contaur on the f i nal TBU. sa samehaw, the High tane must be del i n ked from the f i nal TBU. The surface forms af examPl es (a> and (b ) shaw that Hi gh tane of the suf f i x end� uP an the f i nal TBU af the naun roat . Thi s can be explai ned by waY of Posi ting a ---- - - - - - - •.- --- ---- - 1 09 rul e label ed Hi gh tone Shi � t , stated as �ol lows: (59) HIGH TONE SHIFT: Gi ven that ' sh i � t ' means del in kin9 a tone from one TBU and rei i nk ing i t to an adj acent TBU, shi ft a High tone of a TYPe I I suf� i x l inked to the same TBU as a Mi d tone � the su�f i x , t o the f i nal TBU o f the noun root ( i . e. the second TBU of the word ) . Xl X '\ - > H Xl X ' ... .,A H M Such a rule then , can be aPpl i ed to the noun gi ven i n (57a) above, as fol l ows: (59) a. s£-'�i -> s£€�i I M / ' st i n9s ' H M H SHIFT Before jumPi ng to the possi bl e concl usi on that thi s proPosed Hi gh Shi ft rul e and the Suf f i x Hi gh tone Link rul e ( 56) , which l inks a free Hi gh tone to the TBU on the l ef t , can be general ized i nto one singl e rul e formulat i on , the reader is asked to consi der the exampl es in (57b) and ( 57c ) . In these exampl es, it is cl ear that Hi gh tone does agi shi f t to the Precedi ng TBU. the reason for thi s? What mi ght be One thing these two examples have in common i s that they both have a Low tone noun root . Instead of al l owing Tone Shi f t , i t aPPears that these two nouns tri gger the Low tone to sPread onto the fol l owi ng Hi 9h tone su� f i x , as shown in (60) bel ow. ----- --- - -- -- -- - --- ----- --------------- 1 10 Once this haPPens, the Law and Hi9h are si mpl i i i ed ta Mid tane, with the i i nal autput remai n i n9 an acceptabl e Mi d tane. One waY ta Prevent Hi9h tane Shi i t �ram taki n9 Place an these Law tane nauns i s to order the Hi9h tane Shi i t rul e � Law tane SPread . Ii Law tane SPread accurs, as i n (60) abave, there are two reasans �or the Hi 9h tane nat ta shi � t : 1 ) the si mpl i f i ed Law-Hi gh tane l i nked t a the same Teu resu lts i n a surtace M i d tane and thus, si nce there i s n a surface cantaur tane, there i s n o reasan �ar the Hi gh tane to shi f t ; and 2) Hi gh tane ,aDDot , by virture a� the wel l farmedness candi t i an whi ch states that the Assaciat i an l i nes cannat cross, l i n k ta the Preced i ng TBU wi thout crossi n9 the Associat i on l i ne l inkin9 the Law tone to the �i nal Teu. I� the Hi 9h tane Shi ft rul e must take Place a�ter Law tone SPread and the Suif i x Hi9h tone Linki ng rule ( 56) takes place befgr, Low tane SPread , as was i l l ustrated i n ( 54 ) , then i t becomes cl ear that i t i s theareti cal l y not vi ab l e t o trY t o make ane rul e f armu l at i on for the two Processes. Nouns that under90 Suif i x Reducti on do not Produce anY additi anal campl icat i ons far anal Ysi s. When a suff i x i s reduced ta a si n9l e vawel , the on l Y tane al l owed an the raot i s Law tane. In the exampl e bel aw (61 ) . we see the Law tone sPreading onto the �ol l owing High tone, creati ng a Law-High contour tane wh i ch si mpl i t i es ta Mid tane: (61 ) ncA-a //1\ [ HM The rul es Posi ted so iar , however . sti l l do not caver al l examPl es in the data. We have purposel y l e�t the nouns wi th cantaur tanes unti l l ast . mast cammon contaur tones found on nouns are Law-Hi 9h and Mi d-Low. ---- --- - --- -- - - - - - - -- ----. - -- -- -- -- --- --- The two Low-high 111 nouns do nat Pose a probl em for the anal Ysi s, as 9i ven thus far . In the fi rst examPl e bel ow, Hi 9h Li nking takes pl ace once the i n i ti al Part of the suf f i x i s deleted, whi l e i n the second exampl e, Hi gh Shift takes Pl ace after Suf f i x Fi nal Del eti on . ( 62) mar' - CVI V - > m�r' - l a - > mar'-la L� �� L� H� L �'H� Ac SID H LINK AC SFD H SHIFT ' el ePhantiasi s ' The fal l owing Law-hi gh noun ( 63) has onl Y one TBU i n the root . Vowel Lengthening after SFD creates an add i t i onal TBU, which acqui res a Hi gh tone after Hi gh tone Shi ft . A Law-High contour fal lowed by another Hi9h then tri ggers the Hi9h Del i nk i n9 rul e , i ntroduced i n ChaPter 2. 11 ' -1\ (63) mpu-u-�i -> mp�-G�i LH HM H SHIFT A \I LH ' hi l l s ' HM H DELINKING The fol l owi n9 Mi d-Law noun (64) , causes Problems, however . Si nce i t under90es SID, one mi 9ht exPect Hi 9h tone Linking to occur, after which Law tone SPread wou l d Pl ace take Pl ace. Thi s unfortunatel Y wou l d resul t in the i n- correct surface tone of ML-M-M. --- - - -- . - -- - - ------ . -- -- (64) 1 12 �-i�l i ->* j�7�-li I'1L H M H LINK ML H ' sons ' Rather : ja�l a M L SPREAD It i s conceivab l e to suggest that si ngl e TBU noun roots wi th cent our tones are excePtions to the stated rul es . Let u s SU9gest then . that i f a contour tone ( l i nked to a Si ngl e TBU) i s fol l owed by a free TBU. the rel i nkin9 of the second tone of the contour to the fal l owi ng TBU takes precedence aver anY ather Let us cal l thi s rul e the Contour Sh i ft rul e ( C Shi f t ) l inking possi b i l i t i es. (65) CONTOUR TONE SHIFT: Gi ven that � shi f t ' means del in kin9 of a tone and l inki ng i t to an adj acent TBU. the second tone of a contour tone l i nked to the same TBU shi f ts to anY free TBU to the r i ght. Tl does not equal Tz Thi s rul e is agi part of the Association Conventi ons. It is a rul e that al l ows tar the r el i nkin9 ot contour tones for the PurPoses of si mpl i fY i n9 the contour and 9i vi n g each tone i ts TBU. Below is an exampl e of a deri vati on usi n9 the C Sh i f t rule. (66) ji � ML AC - Cvl v I I H M - > ji � - la I ML H M SID -> jat\ a la\ ML - HM V LENGTH > ja � l a -) ja � la � I M L HM C Shi f t I l,. '" HM M L L SPREAD Si nce the second tone at the contour i s now l i nked to the second TBU. one maY 1 13 wender to whi ch TBU the tree Hi gh tone ot the sutt ix maY be l i nked . It the High tone Li nk rul e i s al lowed to oPerate after the C Shi ft rule. there i s nothi ng to keep i t trom l i nking to the second T9U. too. Thi s . untortunatel Y wou ld Yi el d an i ncorrect surface torm: *jaal a. 1/\ I ML HM H LINK Therefore. Hi gh Linking must be Prevented trom l inking atter the C Shi f t rule. I f i nstead . it occurred before C Shi f t . Hi gh Li n k i ng woul d i ncorrectl y el im inate the envi ronment for the C Shift ru l e� *jaal a. :4I! j M' LHI'! i HIGH LINK We see. then . that High tone l i nking can be al l owed nei ther before or after the C Shi ft rul e . but rather . when the C Shi ft rul e aPpl i es. it excludes the � i ght for Hi gh tone l i nk i ng to aPpl y. even though the condi t i ons for ru l e aPPl i cation . out l i ned i n the formu l ati on of the High Linki ng rul e. are met . As a resu l t . i n order to Predict a correct outPut. an add i t i onal condi tion wi l l have to be i nserted i nto the formul ation of the rul e - Below i s a restatement of the rul e (56) with the new condi ti on. ( 67) SUFFI X HIGH L INKING: Li nk the free Hi gh tone ot the TYPe I I suff i x to the t i rst TBU to the l ef t . on condition that the Contour l i n k i ng rul e has not al readY l i nked a tone to that TBU. 18 X I H x I M The tree High tone of ii4l&. then i s not al l owed to l i nk to the second TBU through the High Li nking rul e. The onl Y opt i ons lef t to it are 1 ) to l i nk to the f i nal TBU� at whi ch Poi nt . the Low tone SPread rule woul d be triggered . sPreadi ng the Low tone onto the Hi gh tone. thus creat i ng a Low-Hi gh Contour 1 14 whi ch si mPl i � i es to M i d tone (see (68a» or 2) not to l i n k at al l (see ( 68b » . In ei ther case, the correct sur�ace �orm wou l d resu l t . MY Pre�erence i s to l i nk the Hi 9h tone, i� possible, si nce in general , the Hi 9h tone tends to be l i nked to at l east some TBU. ( 68) a. jaU a b . j aU a ML HM ML HM It 1\ 1/ I LOW SPREAD The other contour tone word i n thi s data samPle Poses no Probl ems for this new C Shi f t rul e . Below i s a deri vation o� mpu6�i . ( 69) mPu-CVI V -> mpu-u-�i -> mPu u �i -> mPu 6 �i -> mpu 6 �i !\ HI MI LH A " HM 1\ LH SFD � V LENGTH AC A/ HM" / /" HMI LH LH C SHIFT H SHIFT The behaviour of su�f i xed High-Mid nouns i s somewhat probl emati c . When SUFFIX FINAL DELETION occurs, two possible shapes sur�ace: HHM or HMM. The f ormer i s more �requent . ( 70 ) a . HHM vlnlu vUnmi ' cr ic kets ' caral")£ c'r'mi ' orPhans ' ' pal m nuts ' sli7lin£ b . HMM t6u r\jliline t6umi ' caterpi l l ars ' ' stone When SUFFIX INITIAL DELETION takes pl ace, a9ai n both HHM and HMM are possi b l e 1 15 tonal shapes of the noun � though i n this case� the l atter i s more f requent . (7 1 ) a. Ht1t1 f �al a ' f i sh ' k'el E ' monkey ' s6n l al� ' parakeets ' b . HHM mP(il u ' sP ider s ' k�nlS ' beads ' what i s haPpeni ng here i s that two Hi gh Mi d sequences are vying for the three avai lable TBU ' s. Possible reasons for a l ack of an adequate anal ysi s are 1 ) the tonal nature of Mid tone i n a Hi gh Mi d naun root is unknoNn . 2) l ack of accurate transcri pt i on . 3 ) arbi trarY, l ex i cal exceptions. or 4) si mpl y free vari at i on on the part of the SPeakers. The two Hi gh-Law nouns in the data samPl e behave different l y from each other. One opti onal l y mai ntai ns the Low f i nal root tone. the other does not . (72) a. jS ' pockets ' j661� s6K�lo. s6x'lo l l b . s(il � ' f l oors ' It shou l d be noted that since al l Hi gh and Hi gh-Mid tone nouns Possess two TBU ' s. they are al l ei ther subject to Suf f i x Fi nal or Suf f i x Initial Del eti on . There are n o i nstances where the enti re suf f i x has been retai ned or reduced. In summarY . then . i t has been estab l i shed that the Associ at i on Convent i on s must aPpl y before the segmental delet i on rul es . A s a resu l t . several n e w ru l es • ---- - AccuracY of tone transcript i on uncertain . - -- - - -- - ---- - - -- --- - -- ---- --- ----- 1 16 are requi red to Provi de �or the reI inking of �reed TBU ' s or tones. Accordi ng to the d i scussions abovep i t has been seen that these rul es aPpl y i n the �ol l owing order : Assoc i ation Convention s Segmental Deleti ons: SU1� i � I n i t i al Del eti on . Su�� i x Fi nal Del eti on . Suf f i x Reducti on Vowel Lengtheni ng (6S) Contour tone Shi � t (33) Low tone SPread OR ( 67) Suff i x High Li nking <Ch . 2) (53) Mid tone Del i n k i n 9 <59) Hi gh tone shi f t (37) Hi gh tone Del i nk i ng (Ch . 2 ) 3. Root ton. IONrin9 (Hi 9h o.l .Uon) One i tem that comPl i cates the analYsi s of the tonal behavi our of nouns wi th TYPe I I suf f i x es somewhat i s the f requent l oweri ng o� the tone of noun roots when the suf f i x i s added . Al l types of Mid i n i t i al nouns as wel l as Law-Hi gh nouns can al so be subject to root tone l owerin9. whi l e Hi gh i n i tial noun roots a.waa::. l ower . Below are elCamPles of nouns with l owered root tone: TYPe I TYPe I I Mi d nci - l e nd-xal e ' bal aPhon . bal aphons ' Weak Mi d f o-b .f�-?ol :::l ' debt . debts ' Mi d-Law U -d� tar-al e ' l i ana. l i anas ' Low-High n k�n-Ia nk�n-?al a t tooth. teeth ' (73) UnderlYi ng Root Tone 1 17 In order to analYze the behavi our of Root tone Loweri ng, both the tonal and segmental envi ronment must be consi dered . It wi l l be noted i n the fol lowi ng discussi on that a H i gh tone suftix seems to Pl aY a rol e in the root tone l ower i ng ot both nouns and verbs. In addi t i on � examinat i on Ot the data reveals that root tone l ower i ng i s most common on nouns with ei ther ful l or reduced TYPe I I sutf i xes , whi l e verY l i ttle root tone l ower i ng occurs on nouns wi th parti al ly deleted sut f ixes ( i . e. those which have under90ne ei ther Suttix Ini tial or Sut f i x Fi nal Del eti on . ) . Let us f i rst consi der the types of tones whi ch have a tendency to lower to Low tone and attempt to determine whether an under l Yi ng general ization can be made about them. The examPl es above in (73) i nd i cate that Mi d , weak Mi d , Mi d­ Low and Low-High nouns can al l undergo root tone l ower i ng. One possible waY to motivate the l oweri ng of the Low-Hi gh tone to Low tone is to si mpl y Posit a Hi9h del eti on rule: H -> 0 . When the Hi gh tone is del eted , the Low tone remai ns. In Chapter 2 , i t was PrOPosed that Mi d tone verbs were underlYi n9l y Lh, that i s , they Possessed a Low tone on the PrimarY t i er . whi le being sPeci ­ t i ed on the 5ubregi ster t i er tor Hi gh tone. It was sU9gested . there. that Mid tone verbs were l owered to Low tone through the Process Ot subregi ster H igh tone del eti on : h - > 0 . The appl i cati on of thi s rul e then resul ted i n a remai nin9 Low tone on the Pri marY tier . If the Mi d tones of weak Mi d . Mi d . and Mid-Low nouns cou l d al l be Posited as Possessi n9 a Low tone on the Pri marY tier and a High tone on the subregi ster tier . the rul e Ot subre9i ster Hi 9h tone del eti on cou l d a l so effecti vel Y aPPl y to each Ot these tones to Produce a Low tone. It must be noted . however . that at thi s poi nt . the underl Yi n9 tone Ot Mi d tone nouns has not been anal Yzed; theretore the hYPothesis here remains tentati ve unti l a thorough i nvestigation Ot Mi d tone nouns is made in ChaPter 1 18 5. Assumin9 that the hYPothesis is correct. however . we see that the 1 0wer i n9 of both Low-Hi 9h and Mi d i n i ti al nouns can be tri 9gered by a Hi 9h Del eti on rul e for the former t h e del eti on of the Hi 9h tone on the Pri marY tier . whi le the second set under90es the del eti on of the subre9ister Hi 9h tone. Si nce these twa Processes paral l el each other i n behavi our . l et us ProPose a sin9le rul e to cover for both. l abel ed si mPl y. the Hi9h Del eti on rule. Thi s Hi 9h Del eti on rul e wi l l not onl Y i nd i cate that the Hi 9h tone to be del eted can be on ei ther ti er . it must al so i ndi cate another fact that al l of these prOPosed under l Yi n9 tones have in common . namel Y a Low tone. In al l cases. a Low tone i s found on the Pri marY t i er ei ther ad jacent to a Pri marY tier Hi 9h tone or l i n ked by a Hi9h tone on the subre9i ster tier . Wi th this i nformat i on . the f ol l owi ng rul e i s PrOPosed : (74) HIGH DELETION: Delete a Hi 9h tone on the Pri marY or subre9i ster tier i f a Pri marY Low tone i s found to the left of a pri marY Hi 9h tone or l i nked to a subre9i ster Hi9h tone. < H. h > -> 0 I L ___ L I The examPl es 9 i ven bel ow show how thi s rul � woul d work : M : nci-dle I L J ® ML: U ral e '" L L V ® Once the root tone i s l owered to Low tone. i ts tonal behavi our i s i denti cal --- ,----- ------- ---- -------- ---, -, ---- 119 to that o� under l Y i ng Low tone roots i n that the Low tone o� the root spreads onto the Hi gh tone of the TYPe I I suf� i x . as shown in the examPl e bel ow . (76) nci - xale -> nc � - xale -> nc� - xale I I , I I I L , "' l L H M L H M L H M I I " bal aPhons , I ® h AC L SPREAD H DELETION (77) nk�n - ?al a --> n k�n - ?al a --> nk� - � a --> nk� - ?ala � � -� AC t 1� {® � � " teeth ' .... L-SPREAD HIGH-DELETION Si nce Hi gh Del eti on does Provi de the envi ronment �or Low tone SPread . i t must then b e ordered before the Low tone Spread rul e. Thi s rul e order i ng i s in contrast to the Hi gh Del i n ki ng rul e i ntroduced i n Chapter 2 . where a Low tone Spreadi ng onto a Hi gh tone TBU contri butes to the envi ronment needed f or High Del i nki ng. Recal l that when an i ncompletive suf�ix has been sub j ect to Low tone sPreadi ng. the Hi gh tone i s del i nked i f i t i s fol l owed by a Hi gh tone ncun c l ass cl i ti c . as i s shown i n the example bel ow : wu Ya xa k�-an ryda -> wu Ya xa k�-�n ryaa (78 ) I M I I M H �'1 L H LOW SPREAD V H I M I M I H V\ V L ' He is gi ving i t to hi m ' H H HIGH DELINKING The two examPl es gi ven bel ow. have the same under l Y i ng tonal shaPes. Yet one undergoes High del eti on (a) . whi l e the other . for unknown l ex i cal reasons. i s - ---- - --- -- ---- --- ---- -- --- - 120 exempt �rom Hi 9h Del etion . Instead , i t under90es Hi9h Del inkin9 < b ) . ( 79) a. n k�n - ?al a --> nk�n - � l a --> n k� - ?'l a --> nk� - ?ala A LH /\ L@ I I H M I I H M HIGH-DELETION b . 9b�n -xal o -> 9b�n -xal o 1', LH 1 / H M " LH I I L ., ' l I L " teeth ' H M L-SPREAD " 9ranarY ' H M H DELINKING These two rul es c l earl Y Yi el d d i � t erent resul ts. In the t i rst . the t i nal output i s LMM whi l e in the second . i t is LHM. It Hi9h Del eti on were ordered atter Low tone sPread. the der i vati on wou l d Yi eld the same resul ts as Hi 9h Del inkin9. i l l ustrated i n ( 79b ) above. I n this set at examples. then . the Pri marY ditterence between Hi 9h Del i nk i n9 and Hi gh Del eti on is where they are ordered in resPect to Low tone SPread i n 9 . Hi 9h Del eti on takes place at a l ex i cal l evel at the der i vati on . whereas Hi 9h Del i nk i n9 can take Place at both the lex i cal and the Phrasal l evel . The above di scussion has centered around how the tonal nature ot Mi d i n i t i al and Low-Hi9h tones can contri bute to the phenomenon o� root tone Lower i n9. now known as Hi9h tone Del et i on . There are . however , envi ronmental �actors, both tonal and se9mental . whi ch al so seem to be Present when the root tone o� the noun is l owered to Low tone. The t ol l owi n9 di scussion wi l l examine each of these factors in an attemPt to determi ne whether they shou l d be consi dered i n the formu l ation ot the Hi 9h Del eti on rule. Let us consi der f i rst the tonal envi ronment tor the Hi 9h Del eti on rul e . Hi 9h Del eti on occurs on nouns when the noun root i s tol l owed by a Hi 9h i ni tial TYPe ----_ . _--_.. - .- _.------ 121 I I su�� i x . When one recal l s that Hi 9h Del eti on on verb roots al so takes Pl ace when �ol l ow.d by the Hi 9h tone i ncomPl eti ve su��i x as i n (80) , one be9ins to susPect that PerhaPs a Hi gh tone suf � i x i s a cruci al �actor in the aPpl i cati on o� High tone del etion . ka- an - > k�-� --> k�-an ( 80 ) I L I \ I I H V\ L H L I ' gi ve - i ncompl eti ve J H ® h H-DELETION L-SPREAD The ProPosed revi sion o� the �ul e wou l d then i ncl ude the sti Pul at i on that Hi 9h Del eti on takes place when � ol l owed by a Hi gh tone su�� i x , as shown below: ( 81 ) HIGH DELETION: Delete a Hi gh tone o� the Pri marY or subre9i ster t i er i � a pri marY Low tone is �ound to the l eft o� a Pri marY High tone or l i n ked to a subregi ster Hi gh tone, and fol l owed by a Hi gh i n i t i al suf f i x . {H, h > - > 0 / L ___ L 1 [suH i x ] Note agai n that the envi ronment f or Hi gh tone deletion i s surPr i si n9l y si mi l ar to that requi red tor Hi gh del i nk i ng. In both cases, the High tone i s del eted or del inked when f ol l owed by another Hi gh tone. As menti oned above. however . Hi gh Del eti on is l i mi ted to aPPI Yin9 at the word l evel , whi l e Hi gh Del i nkin9 can take Pl ace across word boundar i es and the two are ordered d i ��erentI Y with respect to Low SPread . Examinat i on o� the data shows that not � words with Mid i n i t i al or Low­ Hi 9h tones are subject to Hi 9h del et i on when fol l owed by the Hi 9h i n i tial TYPe ---- ---- - -- -- - - -- ---- - - - -- - -- --- - - --- --- - ---- ---- 122 I I su�� i x . Ei ther the choice �or High Del eti on is pri mari l v arbi trarv or else there i s another envi ronmental �actor to take i nto consi deration . The data gi ven an the �ol l owi ng Pages shows that the segmental envi ronment seems to Pl av a rol e i n determi n i ng whi ch Mi d i n i tial and Low-High tones are sub ject to H i gh Del et i on . I t was noted i n the i ntroduc t i on to thi s sect i on that Root tone lower i n g ( now high Delet i on ) occurs al most wi thout excePti on a n nouns wi th reduced su� � i xes and on nouns with �ul l su�� i xes. whi l e tone l ower i ng is l ess l i ke l v on nouns w i t h part i al l v del eted su�� ixes. The three subsections bel ow describe (a) Tone l oweri ng wi th Reduced Su�� i xesp ( b ) Tone l ower i ng wi th tul l su� � i x and # ( c ) Parti al Su� � i x Del eti on and resi stance to tone l oweri ng. A- Hith o.l.ti an wi th Suffix �uctian When Tvpe II c l ass su� � i x under9Des Su� � i x Reduction . the root tone i s a al ways Law tone. Anv Mi d . weak Mi d . Mi d-Low. or Low-Hi gh nouns whi ch undergo Su� f i x Reducti on are also subject to Hi gh Del et i on . (82) Mi d-Low C l ass 2 Cl ass 1 a. n3 n � -i ' cow. cows ' b . P3n p�-un ' dog. dogs ' c. to B -i ' t ather. fathers ' d . ceewi ce-e ' woman . women ' e. toB te-e ' chie� . chi eh ' Mi d-LON Cl ass 6 Cl ass 5 Weak Mi d ----- ---- - ---- - ' eye. eves ' ny�-i � . nveU Cl ass 6 - ---- No Su� f i x Reduction ----- ---- --- -- - - -- ------ ----- 123 No Su.f .f i x Reduct i on Mid Low-Hi9h Cl ass 5 Cl ass 6 9. s�l� s�-o - mi l l et s9. and pl . ' h .fu-un - Peanut . peanuts ' i . baa b�-l e ' 9round Pea. 9round Peas ' j . man@ lIa-ne ' sweet ground Pea. sweet 9round peas ' • .f �n l a Bath Cebara and SuPYi re experi ence root tone l ower i ng on Mi d Low nouns whi ch have reduced su.f.f i x es. Interesti ngl y. Mi l l s ( 1 984) and Car l son ( 1 980 ) have chosen the same examPles to i l l ustrate thi s l ower i ng. Root tone l ower i ng (Hi gh Del et i on ) al so takes Pl ace i n the Sucite cognates o.f these words. as seen below: Cebara SuPYi re Suci te Engl i sh a. P:dn. P:hn:::J PPun. ppu-un P�n . pUun " dog. dogs ' b . c:::J I :::J� . cabal a cae-we. ci-e c�. c�e ' woman . women ' c . si kaa. si kaal a si ka. si kxa-a si ki. si kla " goat . 90ats ' (83) In compounds, onl Y roots i n The l atter ex ampl e i s comPound i n structure- contact with the su.f .f i x are suscept i b l e to tone l ower i ng. Carl son ( 1 980) states that in SuPYi re. most o.f the Mi d Low nouns exPeri ence Root tone l ower­ ing. Mi l l s ( 1 984 ) cal l s root tone l oweri ng " i rregul ar " . but states that manY "lll. c:lass and some i!LU. c l ass nouns" acquire a Low-Mi d pattern , wi thout i denti.fYin9 the ori 9i nal tone o.f the root . Her exampl es show Hi gh . Mid. and Mi d-Low noun roots that have exPeri enced tone l owerin9. ---- - ---- -_. -- - - - --- - --- --- -- - -_. _ - - --_._-- ------- 1 24 b. Hi gh D.I .tion with Ful l Su��ix Hi 9h Del eti on al so tends to occur on noun roots wi th the �ul l 5U�� i x . (94) Mi d-Low Cl ass 6 a. ci l t} d -xal e Weak M i d Cl ass 6 " th i gh , thighs ' b . ')m:lb ')m�-?:Jb " kni �e, kni ves ' c . �ob ��-?ob " debt , debts ' Mi d Class 6 d . nci l a nci -xsl e " bal aphon , bal aPhons ' e. l al a U-xal a , h-xttla ' PregnancY, Pregnanci es ' � . IlImal a IlImi-?al a " bud. buds ' 9. Pel E P�-xal£ " bowl , bowl s ' Low-Hi gh Cl ass 6 h. l ill 8 l �-xal o " shea nut , shea nuts ' i . nkin I a nkin-?an l a ' toot h , teeth ' except i on : j . 9b�U 9bdn-x'1 0 ' granarY, granar i es ' Al l weak Mi d nouns (except �or a COUPl e of possi b l e contradi ctorY cases ) take the � u l l i ndef i n i te 5Ut� i lC and al l ot them undergo Hi9h Delet i on . There are no weak Mi d Cl ass 2 nouns. Cebara has one example, i n i ts l i st Ot i rregu lar Plural s , of a cognate o� a Suc i te weak mid noun wh i ch l owers when a Class 2 suH i x i s added : Cebara (95) '):ln� . ')�?�n� Suci te Engl i sh ')m:l l � , ')m�?al o " kni �e. kni ves ' It is not known i � there i s a tonal categorY o� nouns i n Cebara whi ch ----- - ---- - -- - ---- ----- ----- ------ 125 corresPonds to the ' weak Mi d ' categorY of Suci te. Weak Mi d nouns do ex i st i n SuPYi re. In fact , i t was Carl son who i ntroduced the term Weak Mi d . Of the few Weak Mid examPles that Carl son Provi des. onl Y one one d i d not undergo root tone Lowering. (86) SuPYire Cl ass 1 Cl ass 2 f ya .fy�-a ' f i sh . pl . ' shi n shi -i n ' Persons ' Class 5 c. SEn-I: sil')-i i ' sti ngs ' l a-a U-h i i c Pregancies ' Partial Suffix n.l .ti an and Resi .tance to Hi gh Del eti on It has al so been observed that when a suf f i x has undergone ei ther Suffix Ini t i al Deletion or Suf f i x Fi nal Del eti on . the noun roots are l ess l i kel Y to undergo High Del eti on . In the examPl es below, the few nouns wh i ch do undergo High Del eti on are marked i n bol d . ( 87) a. Mid-Low Cl ass 2 nouns ji jaU a ' sons ' sHn sH n h ' fami lY. rel ati ves ' kal t kal j}l t ' gods ' b . Mi d-Low Cl ass 6 nouns kud� ku <x ) tlx i ' roads ' nj i dj} n j i r�l e ' tongues ' njedi nj*".le ' ! iver. ' i:id� tr.l e ' U an•• ' ---- -- - -- - -------- ---- ------- 1 26 c . Mi d Class 2 nouns gba gba-aI a H in f i { nl e . f i €nmi ' bl i nd Persons ' cal au cal 'l a ' Pi gs. pork ' Paru por'l o C daughters ' ' ri vers ' d . Mi d Cl ass 6 Nouns led. l ir.l, ' u .... ' nYen, nv""i ' horns ' tile t i < x ) 'K i C crests o� cocks ' SE nE sd,)i c st i ngs ' nti de nteral e C bats ' mpudo mPorala c horn trumpets ' • •U"i e . Law-Hi9h Cl ass 6 Nouns mgbi l @ mgb i { ')i C threshing or beating st i c ks ' mpill e mP�6')i Chills' ntUn I a ntMn,)i c baskets (f:i ght l y woven ) ' ar� nr'l a c wi nnowi ng baskets ' p.u � p,U'mi ' pai l s ' Previ ous to f:he di scussion of the segmental envi ronment . it was menti oned f:hat High Del ef:i on takes Pl ace before Low tane SPreadi ng. so that the deri ved Low tone can al so tri gger Low tone SPreading. Now we need to ex ami ne the rul e ordering of segmental del eti ons and High Del ef:i on . I n the above descri pti on i t was assumed that High Del eti on fol l owed s�gmental deletion . However . before argu i ng for thi s rule orderi n9. l et us take a lock at the vi ab i l i ty of order i ng High Del eti cn before segmental del et i ons. ----- --- - ---- - ----- --- -- -- ------ ---- 127 Most nenderi ved Lew tene reots take Part i al l y del eted suffixes. whi l e most deri ved Low tone roots take f u l l or reduced suff i xes. If Hi gh Del eti on were ordered f i rst . there wou l d be no waY to dist i ngu i sh deri ved Low tones frem Low tone roots. and thus no waY to Pred i ct that the deri ved Lew tone roots tend te ept for ful l er reduced suf f i xes. whi l e underl Y i ng l ew tene roots prefer Parti al l y del eted suf f i xes. In add i t i on . i f Hi gh Del eti en occurred f i rst . decidi ng whi ch Mi d . Mi d-Low . er Lew-Hi gh tone reot to l ewer to Low tene wou l d be Primar i l y an arbi trarY dec i s i on . On -the other hand. order i ng Hi gh Deletion after segmental del eti on wou l d bri ng some Predi ctabi l i ty t o t h e rul es. Whi l e the tYPe of segmental del eti en wou l d be Partial l y an arbitrarY dec i si on . once the del etions are made. High Del eti en wou l d be ebl i gatorY on anY Mi d i n i ti al or Low-High noun wi th a f u l l suff i x or a reduced suff i x , and rare on such neun roots wi th Part i al l y del eted suffi xes . With the segmental Del et i ons rul e ordered before High del et i on . then . one can now use segmental informati on i n the formul ation ef the High Del eti on rul e. The revised Hi gh Del etion rul e gi ven bel ow in (as) wou l d new i ncl ude the i nformat i en that High Del eti on occurs on nouns wi th ful l or reduced suff i xes. (88) HIGH DELETION: Delete a Hi gh tene on the Pri marY or subre9i ster t i er if a Pri marY Lew tene i s f eund to the left of a Pri marY Hi gh tone or l i nked to a subregister Hi gh tone. and fol l owed by a High i n i tial suff i x . If the suf f i x i s l --'\+ a nomi nal suff i x . i t must have undergene ei ther n o segmental al terati on or el se have undergone Suf f i x Reduct i on . <H. h } - > 0 I L L _I ---- - --- .- --- H (suH i x - i f N. then ful l er reduced ) --_ ._- -- -------- _._- ------ 128 Thi s rul e maY seem somewhat odd in that i t i nc l udes not on l Y tonal i nformat i on but al so se9mental and morphol ogical i n�ormati on as wel l . However . at thi s Poi nt , these are Precisel y the �actors that seem t o govern Hi gh tone del et i on . Below i s a H nal examPl e i l l ustrati ng the ordering � Hi gh Del eti on after segmental del .t i on and b�ore Low tone SPread . (89) sO - xal o --> I I " LH H M ' mi l l ef: ' siS A LH -0 f\ HM SUF RED --> sCi 6 A t\ t.® HM H-DELETION -> sCl6 --> t' l\ sCio L HM L-SPREAD Thi s compl etes a lengthy di scussion of the TYPe I I i ndef i ni te suf � i x . whi ch are found on the two Plural noun cl asses. 2 and 6. We have seen that there are several factors affecting the tone of nouns wi th TYPe I I i ndef i n i te suf f i x es. Fi rst of al l . the tone of the root can alter the tone o{ the suf f i x . SPeci fi cal l y . the Low tone of the root can sPread onto the Hi gh tone of the suff i x . Secondl y , the tYPes of segmental del et i ons that take Place not on l Y hel P to determi ne how the remai ning tones and TBU ' s can be reI i nked ; they can al so Pl ay a rol e i n tri gger i ng Hi gh tone del eti on (or Root tone Loweri ng ) of the noun root tone. IV. THE DEFINITE SUFFIX Th i s secti on di scusses the structure and tonal behavi our of the Def i n i t e suff i x . Si nce the structure of the def i n i te suf f i x of the TYPe I nouns i s somewhat d i f� erent f rom that of TYPe I I nouns. they shal l be di scussed seParatel Y . Thei r tonal behavi our . however . i s si mi lar enough that i t shal l b e d i scussed together i n one subsecti on . ---- -- - -- ------- --- .----- -- 129 A. structure TyPe I - The De� i ni te �orm � TYPe I nouns, that i s. nouns whi ch take the TYPe I i nde� i n i te su�� i x . i s deri ved by repl aci ng the i nde� i n i te with the de� i ni te suff i x . In the chart bel ow, the def i n i te �orm of the noun i s gi ven beneath the i ndef i n i te form . Those nouns wh i ch have a 0 i ndef i ni te suf f i x are gi ven i n the f i rst row, whi l e. those wi th -CV and (N) CV su��i xes are gi ven i n the succeedi ng rows . Noun Cl ass ( 90) Suf f i x 1 ( wi' > 3 ( ki ) 4 ( Yi ) 5 <l i ) 7 <H ) 0 ji <33> ni <3> na-Y.l o o ji-l')e " sCin ' ni-ke " H re ' ni-nY£ " f i res ' ce-w� < 1 > b -xt <33> te-Y� c i -a <37> su-ro< 1 9) ta-be < 1 > H-ke " Pl ace ' i:e-nY£ " Pl aces ' ci i -ne " th i gh ' su-te " mush ' ta-be " med i c i ne ' s£-n£ <5> koo-no < 1 > s£-n' ' st i ng ' koo-nde " cot.ton ' CV ce�l')e ' woman ' <N) CV c�n-"� < 1 > ----- c�n"e ' antel ope ' wYe-,,� <25> wYil")-ge " l eaf ' wYen ' y� wyinY£ ' l eaves ' juu <2> j uub' " speech ' sa-m� ( 10) sa-mbe " sorghum beer ' The TYPe I def i ni te su� f i x is comPosed of a consonant. i ndi cat. i ng noun c l ass and de� i n i t.eness and a Mi d f ront vowel . General l y . the consonants of the i ndef i n i te and def i n i te suff i x of the same noun cl ass are at. the same Poi nt. of arti culat i on . whi l e the manner of articulat i on maY varY. For examPl e. the consonants i n Cl asses 3 and 7 suff i xes are converted from [-stop) to [ +stop) (x -> k. r -> t ) . Thi s cont.rasts with the d i al ectal variat i on f ound i n some nei ghbouring vi l l ages < Kol oko> as wel l as <to the west ) across the border i n Mal i . where Cl asses 3 and 7 def i ni te suf f i x consonant.s remai n the same as those of the i ndefi n i te suff i x . --- - --- -- .- - - ---- - - - -- - - -- ------ 1 30 Dial ectal vari at i on (91 ) Kotoura Sucite gba-xa� gba-k' (3) gba-xa� gba-x' " the house ' su-rO, s\l-t:e (7) su-ro. s\l-re " the mush (main di sh ) ' - Noun stems with � i nal nasal s cause voi cing o� voi cel ess consonant suf f i x es o� Classes 3 and 7: ( 92) �uN ke (3) --> �a�-ge kOoN-te (7) --> keen-de ' i nside ' ' cotton ' Aside �rom the nasal in� l uence of some noun roots. nasal i ty seems to be a feature for the def in i te suffixes of Cl asses 1 , 4 and 5. Instead of a [ +stopl consonant rePl aci ng a [-stop] consonant� as i s the case for Classes 3 and 7, a nasal consonant i n the def i n i te suff i x is at the same poi nt of art i c u l ation as i ts i ndef i n i te counterPart. Thi s is i l l ustrated bel 0... . (93) Cl ass Inde H n i t:e C Exampl e Def i n i te C Engl ish 1 w � ce-...d. ce�-�e • woman , 4 y nY te-yt . ti-nYe · Pl aces ' 5 1 n ci-a� cU-ne " thigh ' When a Cl ass 5 noun is suff i x ed for def i niteness. the vo...el of the root is l engthened . It i s Possi ble, in thi s case. that the i nd� i ni te suf f i X was n ever total l Y del eted; with the add i t i on of the def i nite suff i x , the -1- droPPed out creat i ng a 1 0n9 vowel 12• In SuPYire, this i ndef i n i te -1- never does show uP on the surface. (94) Sucit:e SuPYi re Indef i n i te 9ben-1 8 9bu-un ' granarY ' Def i ni te gbOOn-n' gbuun-n� " the granarY ' Vowel lengtheni ng does occur sPoradical l y i n the defi n i te suf f i x at i on of other 131 noun cl asses. However. there does not seem to be anY predi ctable pattern as i s found i n the Class 5 nouns. The d i sYl l ab i c roots fol l ow the same Pattern of defini te suffixat i on . The onl Y excePti on i s i n the case of Cl ass 5 nouns whi ch have the extended -rV­ root . Unl i ke monosYl l ab i c Class 5 nouns whi ch seem to retain some sembl ance of the i ndef i n i te suf f i x when addi n9 the def i ni te suff i x . these nouns droP both the f i nal sYl l able of the noun root. -�-, and the i ndef i n i te suf f i x before addi n9 the def i ni te suff i x . (95) k�de [ kUr�-I�J -> k�-ne ' the chai r ' *k�-�-na *k�r�-ne Below is chart ( 96) with the deH n i te forms of disYl labic nouns. Pl ease refer to (8) fqr En9l i sh 910sses. 1 (wi ) (96) -1 - IND fol� (22) DEF foU-l')e -r- r-V (N) 5 (l i ) 7 (b ) 9balax� <7) 9b�la-ke �9uraxo < 12> k�d� < 14> sara-I')� �9ura-k€! k�-ne c�r�l'), < 1 > nk�ral')� <2> feramE <2> c�ra-I')€! nk�ra-1')ge fera-mb€! ?-V (N) kan?a <6> s�?�l � < 1 > kCl?ClrCl <1 > kan?a-ke s.?e-ne kCl?CI-te nYa?al')a (6) se?enE < 1 > nYa?�-1')ge se?e-ne fYQx� <2> tux urCl <2> .fyQx�-')e tux�-te x-V <N) nd�xan� < 1 > nd�xa-nde ----- ._--_. - 8 (Pi ) saru < 1 0> -?- -x- 3 ( ki ) .-- .- .- - ------ -'--'- -- -- -- ._- --- 1 32 B. structure - TYpe II npun. The Def i n i te Suf t i x es of TYPe I I nouns are si mi l ar i n structure to the def i ni te suf f i x of TYPe I nouns in that the i n i t i al consonant of the suf f i x i s a t the same Pai nt 0+ art i culat i on a s the consonant a t the i ndef i n i te sutti x . Thus tar class 2 . the det i n i te suf f i x i s b i l ab i al . as shown below i n ( 97a) and t or Class 6. i t i s vel ar . (97) Class Indef i n i te C Det i ni te C Indet . Det i n i te a. 2 m. b -> b "n�-mi. -> ?lnaa-m-bi ' sons ' b. 6 x -> k P�18l.a. -> P�-lC�-1d. ' bodi es ' In ci tation form. as gi ven i n ( 97) . i t apPears that the det i n i te suf f i x at TYPe I I nouns i s monosYl l ab i c . However . when tol l owed by a vowel i n i t i al morPheme. such as the verbal parti cle, l. whi ch coal esces to the t i nal vowel at the subject. one observes the aPPearance of an -1-. as seen below. ( 98 ) Cl ass 6 nYU -1d. ' eyes ' nYU -k11.a-.l wU y � ' the eYes hurt hi m ' " the goats have gone aut ' Thi s observati on l eads to the hYPothesi s that the underlYi ng shaPe of these suff i x es are; Cl ass 2 - bill and Cl ass 6 - k1li13 • Instead of rePl acing the i ndef i n i te suf f i x wi th the def i n i te suff i x . as was the case for TYPe I nouns. TYPe I I def i n i te nouns are f ermed by add i ng the def i ni te suf t i x to an i ndef i ni te suf f i xed noun stem. Thus. for the noun al:L ' eyes ' . the def i ni te form i s nl-l -ki . However, when the i ndef i n i te suf f i x consists of two TBU ' s . the add i t i on of the def i ni te suf t i x causes the reduction of the two TBU suf f i x to one TBU, as shown i n ( 99) . ---_ . -- - .- .- - --- ----- 1 33 (99) p� - xalo - k! -> p� - x� - k{ Ind Root " the bod i es ' De-F The e-F-Fect i � t.hat de-F i n i.t:e TYPe I I nouns consistent l Y have onl Y three TBU ' s. the -F i rst TBU bei ng reserved -For the noun root . the second tor the i nde-F i n i te su-F f i x . and the thi rd fer the def i n i te suff i x . Below are examPles o-F waYs the i ndel i n i te suf -F i x of Cl ass 6 nouns is reduced to one mora when fol l owed by a def i n i t e suffi x . ( 1 00) Cl ass 6 CL 5 IND CL 6 IND CL 6 DEF a. p�l � p�-xolo P�-x�-k! " bodY. trun k ' b . col � c�-?ol o c�-?�-k! " cl ay Pot ' c. j i de j�r�-I')i j�r�-I')-g! " breast ' d. sE nE s£i-I')i sE f-l')-g{ " sti ng ' e. t i l e tH-x i t i -x i-k! " cock ' s crest ' f . H da kira-I e H r-�-k! " countrY ' 9. kud� ku�-l< i ku-x�-k! " rule. road ' h. cedE cer�lt cer-i-k! " cal abash� gourd ' i . nYel a nYi -i nYi -i-k! " eYe ' Whatever -Iv suf f i x i n9 there was i n the i ndet i ni te torm i s del eted wi th the addi ti on ot the del i n i te sutt i x . However. the i n i t i al sYl lable of the i ndef i ­ nite sut t i x tends t o b e mai ntai ned ( examPl es a, b ) . I n the case ot the nasal i z ed i ndef i n i te suff i x . the nasal i s retained and as a resul t . i t causes voi cin9 of the Cl ass 6 defi n i te sut f i x consonant ( c, d ) . The extended -rV- roots tend to - ---- ------- ------- ----- 134 retai n thi s extensi on . causi ng the segmental del eti on o� the ent i re i nde� i ni te suH b , (� . h ) . Therc� i s , however , one instance oF the -rV- del et i ng. al l owing the � i rst sYl l ab l e o� the i nde� i n i te suf� i x to remain (g) . Cl ass 2 nouns undergo the same tYPe of segmental reduction Processes. as seen bel ow: ( 1 01 ) Cl ass 2 Cl l IND Cl 2 IND Cl 2 DEF a. j 9 j66-b j6a-b! ' Poc:ket ' b . fY;(a fyaa-I a h-aa-bf ' H sh ' c . PUn P�an-I a p��-b! ' al l i gator? ' d . Poru Por6-l o Para-bi ' daughter ' 56n l a-b s6nl a-bf ' parakeet ' f . �n'a "n""-mi "n�a-m-b{ • g. n3 n i -i n i -i -b f " mother ' a. s6n l u scorPion , c. D!finit. Syffix Iqn. The tone of the def i ni te suf f i x vari es according to the noun oF the root . For both TYPe I and TYPe I I nouns. the def i n i t e su�� i x i s Mi d tone when Preceded by Low f i nal and H i gh tone roots and High tone when Preceded by a Mi d f i nal or a Low-High noun. Examples for TYPe I def i ni te nouns are gi ven bel ow in ( 101 ) . Note that the ent i re contour of Mi d-Low nouns i s l i nked to the noun root when fol l owed by a def i n i te suff i x ( b ) , whi l e Law-hi 9h nouns l i nk onl y the Low tone of the contour to the root . Thi s contrast i n behavi our wi l l be explai ned shortlY. ._-_. - .- _ . _- _.. --- --_. _-- ------- 135 ( 102) TYPe I Nouns Root tone IndeH n i te --> DeH n i te m�l a-I')e ' ri ce ' .f:u-I')� --> .f:iln-ge ' i nside ' c . Hi gh-Low sCt-l� --> slJCt-ne < .f: l oor ' d . Hi gh .f:'l'-d -> f.l ..ke ' rack ' e· Mid gba-xa --> gba-ke ' hause ' .f. Mi d III sa-xe --> sa-ke ' bush ' g. Hi gh-Mi d .f� -> f �a-I')e ' H sh ' h . Law-Hi gh 9b�n-l g --> gbililn-ne ' granarY ' a. Low m�B b . Mi d-Low In certain cases to be d iscussed i n Chapter 5, a Law tone root becomes Mi d tone when Preceded by another Mi d tone ward . When thi s happens, the def i n i te ' Your rice. ' Recal l that TYPe I I def i n i te nouns are formed by add i ng a de.f: i ni te suf f i x to an i ndef i n i te suff i xed noun . Si nce al l TYPe I I i ndef i n i te noun stems end i n a Mid tone, the def i n i te suff i x , when added to TYPe I I nouns. i s Predi ctab l Y Hi gh tone. The examPles i n ( 1 03) . ( 104) , ( 1 05) , and ( 1 06) make thi s Poi nt clear . In each set of exampl es. the i ndef i n i te si ngular < of Classes 1 and 5) for each Type I I noun (Pl ural classes 2 and 6) is gi ven . fol lowed by the i ndef i ni te form of the Type I I noun . and f i nal l y in the r i ght hand col umn . by the def i n i te TYPe I I ncun . The examPles i n ( 103) are nouns wi th a Mid tone noun root. The Hi gh-Mid tones of the i nde.f i ni te suf f i x are bath l i nked to the second TBU of the def in i te noun. In the structural descri pt i on of the TYPe I I def i n i te noun i t was shown that the noun root . the i ndef i n i te su.f.f i x . and the def i ni te suf f i x 136 are each 9i ven a si n9l e mora. In the examPl es bel ow. i t wi l l be noted that there are no segments that cou l d be l abel ed as bel ongi n9 to the i ndef i n i te suff i x . Yet . the tone of the i ndef i ni te suf f i x sti l l shows uP on the second TBU. The same is true for the nouns wi th Low-Hi gh noun roots found i n ( 1 04) . ( 1 03) M-HM -> M-HM-H Ind 99. Ind . Pl . Def . Pl . a· 9ba 9baal a gbaibf ' r i ver ' b . ntide nter'l e nterSH ' bat ' c . cal ou cal'l a calSbf ' Pi g ' a. m9b i l @ mgb i f�i m9b) l�gf ' threshi n9 sti ck ' b . f�ra f�ribf ' wi nnowin9 basket ' ( 104) L-HM -> L-HM-H f�r'l a Hi 9h-Mi d nouns i nvar i abl y have onl Y a Mi d tone l i nked to the second TBU. ( 105) H-HM or HM-M -> H-M-H a. mgbfn@ ngbin'l e m9bf nabf ' bamboo mat ' b . jS j661 � j60bf ' pocket ' c . f�a f�al a fy'abf ' fi sh ' d . se?enE se?el')gf Low tone noun roots exh i bi t sl i 9h t l y different behavi our . Recal l that the Low tone Spreadi n9 rul e SPreads the Low tone of the root onto the High tone of the i ndef i n i te suff i x . creati n9 a surface Mi d tone. When a def i ni te suff i x i s added to such a noun stem. as shown bel ow i n ( 106) , the def i n i te suff ix i s PredictablY Hi 9h tone. However , the tone of the second TBU i s ogi Mid tone, as --- - - --- - -- -- --- ------ -- -- - 1 37 mi ght be expected, but rather Law tone. The reader is asked to refer to ChaPter 5 for a ProPosed anal Ysi s of thi s behavi our . ( 1 06 ) L-MM or LL-M -> L-L-H a. d c�ob c��b{ " net ' b . "m� U "m��"i "m��"g{ " carner, angl e ' c · k l d� k l ral e k�dk{ " countrY ' d . l lll � l �x ol o l �x�k£ " bi l e ' e. ceewt c�e c�b{ " woman ' f . njed� nj�rale nj�dk! " tongue ' 9. gbonB gb�n?:Jb 9b�n?�k{ " fi rePl ace ' h . fob f�?ob f�?�k{ " debt ' i . nci l e nd xal e nd d k1 ' balaPhon ' j . n k�nI a nk�n?an l a nk�n?�nk{ " tooth ' The und.rI Yin9 tone 01 the d� i n i t. suffi x i s not at al l strai ghtforward . Al thou9h the def i ni te suffi x tone i s ei ther High or Mi d tone on the surface, i t al so seems t o Possess a f i nal f l oati ng Low tone whi ch never l i nks t o the suffi x . Thi s is suspected because i t tri ggers the same type of tonal behavi our as other Low f i nal nouns. I t wi l l be seen i n ChaPters 4 and 5 that Low f inal nouns tri gger Law tone sPread onto adjacent High tone verbs and weak Mi d and Mi d-Law nouns. Al l def i ni te suffi xed nouns also do thi s regardl ess of whether the surface def i n i te suffix tone i s Hi gh or Mid. The examPles bel ow shaw that the def i n i te suf f i x ed noun causes a Hi gh tone verb to become Law-Hi gh . ( 1 07 ) a . wa� gbak� P�r€ /Plr€/ b . waa m�l a"e P�r€ c. /p€rl l waa gbllunn� P�r€ Ipi ri/ " he sol d the house ' ' he sol d the r i ce ' ' he sol d the granarY ' ----�- --�- ---- . ----- . ---. --- ------- 1 38 Early observati ons reveal ed that the def i n i te suf f i x tone was one uP from the f i nal root tone and one down trom a h i gh tone root . I t was c l ear . too. that the def i n i te suf f i x never had the same tone as the f i nal root tone. If we were to SUggest that the def i n i te suf f i x tone is Hi 9h CLow) . then the Presence of a Mid tone after Hi gh tone and Low tone nouns woul d then have to be expl ai ned . We have al readY ProPosed a Low tone spreadi ng rule in other contexts where a Low-Hi gh contour was simPl i f i ed to Mid tone. Migh t a Mid tone suf f i x atter Low tone noun root be a resul t of Low tone sPread onto High tone? Such a Possi bi l i ty i s i l l ustrated by the examPle below. ' the r i ce ' LOW SPREAD An observant reader maY recal l that Hi gh tone del i n king takes pl ace i n Phrase f i nal Pos,Hi on and betore Hi gh tones. The l ack of anY tYPe of High tone del i nk i n9 tak i ng pl ace here can be eXPl ained by the Pervasi ve Presence of the ti nal fl oat i ng Low tone. Low-Hi gh nouns in the det i n i te form Possess onl Y a Low tone root on the surface. They are d i st inguished from real Low tone verbs . however . by the fact that the det i n i te suf f i x is Hi gh tone and not Mi d tone: ( 109) Ind Det Low-Hi gh bC}-I@ -> bc}-ne ' ground nut ' Low bC}l � bc}-ne • seed , -> Thi s can be easi l y eXPlai ned by the tact that Low-Hi gh roots are not Low f i nal and therefore do not tri gger Low tone sPreadi ng onto the suff i x . The Low-Hi gh examPl e i n C l09) i s al so subject to Hi gh tone del i nking rule ( 37 ) i n order to ---- ----- ----- 139 simpl i f y the Low-Hi 9h contour before a Hi9h tone. as shown bel ow: ' 9round nut ' b�-n€! A I L H H Thi s Process i s t o be exPected here because both the Low and the Hi 9h are l i nked to the same TBU. However . the mot i vati on for Hi 9h del inkin9 is not quite so obvi ous on Low-Hi gh nouns wi th two TSU ' s. such as nd�r�-k'. Soth left to ri ght and ri ght to l ef t l i nking conventi ons wou l d l i n k nd�r�ke i n the same manner and Produce the form. nd�r'-k'. However . the f i nal surface form i s not nd�r6k' but rather nd�rak'. It aPPears then. that i n add i t i on to the l i nking conventi ons that some morpheme i nternal Low tone sPread ing has been takin9 Place. fol l owed by Hi 9h del inki ng: ( 1 10 ) nd�rp-k€! -> nd�r�-k€! t// / L H H L SPREAD V1: / L H H ' the yam ' H DELINKIN6 Hi gh tone nouns create a Probl em. however. l ower t o M i d after another Hi gh? Why woul d a Hi 9h tone suf f i x There are a couPl e o f cl ues. In SuPYire of Farakala (Mal i ) . there i s a verY Productive downsteP rul e. stati n9 that when two adjacent Hi9h ' s are adj acent to each other . the second automati cal l Y down stePs to mid (Carl son . 1993) . We do not have thi s same producti veness i n Suci te but there are a few hi nts of such a downsteP phenomenon i n word +ormation Processes. Ear l i er . we observed that the truly Senuto (non-l oan ) Hi gh nouns general l y adopted a HM contour. In addi t i on . Mi l l s ( 1994) states that i n Cebara. HM i s an al l otone o f Hi gh tone nouns. Secondl y . there are var i ant Pronunci at i ons in some i ndef i ni te Hi gh tone Cl ass 4 nouns. Normal l y. the tonal Pattern i s HHH but i n some cases, i t maY be HHM. Sel ow is an eXamPle of the variant tonal Pattern : ------ - -- - - - - -- - ---- - - -- - 140 mPa�-Ya � t.obaccos ' , • Porri dges , rat.her t.han t.he expect.ed s�r { -ya The wards t.hat. accePt. t.hi s var i at. i on are � orced P l ural s . t.hat. i s. Pl ural s o� wards t.hat. are not. normal l y Pl ural i z ed . What.ever t.he mot.i vat. i on maY be. t.here seems t o be a k i nd of downstePPin9 Phenomenon (or l ower i n9 of t.he second Hi 9h t.one) in the def i n i te suffi xat i on Process. A second al ternat i ve anal Ysi s is to su9gest that the def i n i te suf f i x i s underlYi n9ly Mi d t.one. I t. has al readY been ProPosed that. Mi d t.one i s a comPosi t.e of t.wo �eat.ures Placed on seParate but. l i n ked t i ers. Thi s wou l d be a rel at. i vel Y simpl e sol ut.ion for Hi9h and Low tone nouns si nce t.he M i d tone remains unchanged. ( 1 12) m� a-"e " t.he r i ce ' V/ L M (L > ( 1 13) fa l lt-ke V I H " rack ' M (L> However . aft.er Mi d t.one nouns and Low-Hi9h Nouns. this Mi d tone becomes Hi9h t.one. ( 1 14 ) 9ba-ke --> 9ba-k� nd�rlt-ke -> nd�r�-ke " the rice ' " t.he yam ' Perhaps i f Mi d t.one consi st.ed of a comPl ex � Hi 9h and Low tone f eatures t.his aPParent. rai sin9 phenomenon cou l d poss i b l y be eXPlai ned by same tYPe of Law tone del et.i on . Earl i er i n ChaPt.er 2. the Lower i n9 of Mi d t.one verbs to Low tone was ex plai ned by the Process o� Hi9h delet i on o� the comPl ex Lh tonal feat.ures. ---- ---- - --- - - - -- ---- ----- -- -- --- - 141 It is possible that some type of Low del eti on Process cou l d be moti vated for the def i ni te suf f i x . However . at thi s Poi nt. this apProach can not be ade­ quatel Y defended . Therefore . the Problem of the def i n i te suf f i x wi l l be taken uP agai n when the general tone anal Ysi s of Suc i te has been further devel oped i n ChaPter 5 . V. CONCLUSION Thi s chapter has been a d i scussi on of noun morPhol ogy and tone. of i ndefi n i te suff ixes were i dent i f i ed . Two tYPes The TYPe I Indef i n i te Suff i x was shown to Possess no underl Ying tone of i ts own whi l e the Type I I suffi x was posi ted as havi ng a Hi gh-Mid tone. Each suff i x tYPe brought wi th i t Probl ems of assoc i ati ng tone to segments. mel odi c . i t was gi ven � Having establ i shed that nomi nal tone was seParate t i er . Discussion of the Type I i ndef i n i te nouns reveal ed the need to change the di recti on of assoc i at i ng tones and TBU ' s from the conventi onal Left t o Ri ght Linking to Right to Left Li n k i ng. The vari ous tYPes of segmental del et ions of the Type I I i ndef i nite suff i x brought u P the i ssue of when tones should be associ ated t o TBU ' s. I t was establ i shed that the Associ ation Conventions be ordered before the Segmental del eti ons. As a resu l t . adjustment rules were requi red to l i nk l eftover TBU ' s and tones. The reader maY refer to p. 1 16 for a l i st of these rul es. In add i t i on to the probl ems encountered concerni ng the assoc i at i on of tones to the segments. tonal variation on both noun roots and suff i xes was observed . The l oweri ng of a Hi gh tone def i n i te suf f i x ( 1 15a) and the Hi gh-Mi d TYPe I I suf f i x ( 1 1 5b ) to Mi d tone before Low tone roots was explai ned b y ProPosi ng the Low tone SPread rul e of ChaPter 2 (33 ) . Thi s Low SPread i n9 is i l l ustrated bel ow. - --- -------- 142 ( 1 15) a. PLlLl-nJ! It "' I L " the body ' b . pa-xal o �"1 I L H " bodi es ' H M The Hi gh tone Del i nk i ng rule of ChaPter 2 ( 37) al so found its aPpl i cati on i n nouns, sPecifi cal l y i n Low-High nouns whi ch were fol l owed b y a Hi gh i n i ti al def i ni te suff i x ( 1 16a) and TYPe I I sutf i x ( 1 16b ) . ( 1 16) a. ba-ne A I LH " ground nut ' b . gban-xalo A LH H I I " granaries ' H M Fi nal l y, i t was observed that the root tone Ot some Mid i n i ti al and Low-hi gh nouns changed to Low tone when fol l owed b y the Hi gh tone TYPe I I suff i x . It was al so seen that the Mi d tone verb l owered to Low tone when fol l owed by a Hi gh tone i ncompleti ve suff i x . A tentati ve sol uti on for both nouns and verbs i nvol ved the use of a doubl e ti ered aPProach for tone and the Posi ting of a High Deletion rul e. whi ch i ni t i al l y was formul ated i n Chapter 2 as del et i ng a H i gh tone on the subre9ister ti er when l in ked to Low tone on the PrimarY t i er ( 59) , but revi sed i n ChaPter 3 to al so i nc l ude the del eti on Ot High tone on the Pri marY t i er ( SS) . It was acknowl edged . however , that thi s ProPosed rul e remai ns qui te tentati ve pending a more thorough anal Ysi s ot Mi d tones i n ChaPter 5. NOTES 1 . Mi l l s noted that Hi gh tone nouns i n Cebara maY be fol l owed by a t i nal Mi d : "Al l otones [ ' ] and [ ' -] are features of the grammati cal class Ot nouns. The f inal mi d tone fol l owing h i gh or r i si ng tones occurs onl Y on noun suf f i xes. on some noun stems, and on adject i ve stems. Thi s cou l d be extended to noun phrases , as some adjecti ves bear the h i gh-mi d al l otone. " ( 1 984, p. 1 17) 2. The word t i nal vowel tends to be l owered sl i ghtl y atter High vowel roots. cf-I V ca-I V -> ci-l � -> ci -l e -> co-Ia -> co-l � ' th i gh ' ' cl ay Pot ' Low vowel s remain as they are: ---- ----- -- --- -- --- ------- 143 c�n-rV ' ashes ' -> c�n-r� Si mi lar behaviour i s �ound i n other Senu�o l anguages such as Cebara. 3 . See a brie� desc r i pt i on o� a simi l ar tYPe o� coal escence i n ChaPter 2 . 4 . In SuPYi re. there i s a set o� Mid-Low nouns whi ch . i n the i nde� in i te �orm. are real ized on the sur�ace as Mi d tone nouns. It is onl Y when a def i n i te su��ix has been added that the Mi d-Low contour sur�aces: SuPYi re Suci te Inde� i ni te si ka. si ki. De� i n i te si ki-ryi si ki-�e ' goat . goat-DEF ' ' goat. goat-DEF ' 5. There i s a set o� compound-l i ke nouns exh i b i t i ng a Hi gh �al l i ng tone on the � i nal noun root whi ch do behave l i ke Law f i nal nouns. however . Perceptual l y there i s no tonal d i ��erence between the pi tch o� the under l Yi ng fi nal High tone and the tone o� the � inal sYl lab l e of thi s set o� nouns. However . l i ke Low �i nal nouns. they tri gger Law tone sPreadi ng onto Hi gh tone verbs. as shown below. waa n��an w" > waa n��n w.' ' he l ooked at a brick ' wa� kurugba w" -> wa� kurugb� w.' ' he looked at a shelter ' - These nouns are Pri mar i l y l oan wards. They behave l i ke comPound nouns tonal l Y and segmental l y i n that they are composed of twa stressed comPonents. each Possessi ng i ts own tonal mel odY. I f the l ast component of the ' COmPound ' Possesses onl Y one TBU and has a Hi gh-Law tonal mel odY. thi s Hi gh-Low tone bei ng l i nked to a si ngl e TBU Produces a High fal l i ng tone on the sur�ace. as shown i n the two examples . bel ow. kuru-gba V A ' shel ter (grass overhang) ' HL M UmA-tt V A ' tomato ' (French. tomate) HL L When a Hi gh-Low tonal mel odY i s assi 9ned two TBU ' s. each tone l i nks to a TBU. The second comPonent o� the comPound nouns gi ven bel ow i l l ustrate this. ' Senu�o basket ' l et-i H I II H HL �?� 1/ HL ' water python ' • ' Porr i dge ' (Dioula. baga) 144 6. Cebara fol l ows a surpri si n91 v si mi l ar pattern . Observe the examples bel ow. Hi gh Mi d Low Mi d-Low bdrd=bele t jari =gele B : gel e ko-bi : =9�l e ' bread ' (pi ) ' calabashes ' ' meal s ' ' Paths ' Mi 1 1 s ( 1 984) 7 . In SuPVire, the Hi gh of the High-Mi d suff ix shi fts to the root of a Mid tone noun , wi th the resul t that the Mi d tone of the root is compl etel v del eted . Bel ow are a COUPl e of examPles wi th corresponding words i n Suci te: Supvire Suc i te Engl ish a. c�{-l i cal'-l a ' Pigs ' b. s€ �-i i • st: i n9s • s£ € -f')i 8. SuPVi re al so has a set of underlYing Mi d-Low nouns corresPondi ng to the same group i n Suc ite, but thi s ML contour is simpl i f i ed to Mi d when add i ng i ndef i­ n i te suffixes. As a resu l t , when adding TYPe I I suff i xes, the outPut is a 1'11'1 sequence i nstead of MUM. ComPare the SuPVire and Suci te example bel ow: SuPVi ra Suc i te Engl i sh cE:En=l i i ci � n-le ' YOunger si b l i n9s ' � V M L 1'1 II M L \ 1'1 9. V LENGTH refers to a rul e not di scussed i n this haPPens that i f the suffi x i s parti al l v del eted , a wi l l l enghthen i ts vowel . Thi s Process seems to be attemPt to maintain three TBU ' s on nouns with TVPe thesi s. I t frequentl v noun root wi th a si ngl e TBU a Part of a more general I I indef i n i te suff i xes. 1 0 . Thi s rul e refers t:o the deri vational hi storY of the word i n Hs formula­ t i on . Al though certai n authors have used deri vational hi storv i n the formula­ ti on of thei r rul es, this apProach has been contested i n the l i terature (through Personal communi cat i on , C. Ki sseberth ) . Whi l e acknowl edgi ng the controversi al nature to this apPrua�h , I shal l keep thi s formul ati on of the Suffix High Li nkin9 rul e unti l such time that a mare sati sfactorv sol ution maV be -Found. 1 1 . The same i nformant gave bath forms i n the sPace of a couple of minutes. Bath were Pronounced in isolat i on . 1 2 . I n fact, i t i s possi b l e that hi stori cal l v , def i n i te suffi x es were suff ixed to indef i n i te noun stems. Whether this kind of informati on wou l d be hel pful t er svnchron ic anal vsis remai ns to be seen . 1 3 . Cebara def i n i te suf f i xes for these same cl asses are b i svl l ab i c ; Cl ass 2 bel e, and Class 6 - gel e. ------- ----- ------ 145 CHAPTER • - TONE IN 11£ SUCITE VERB PHRASE I . INTRODUCTION Th. Preceding two chaPters di scu••ed the tonal behaviour of noun and verb root. wi th thei r su��i xes. Several rules �ere proposed, includi ng Low tone SPraadi ng (33 - Ch. 2) , Hi gh Del inking (37 - (88 Ch . 2) , and High Del etion - Ch. 3) , which were �ound to applv to both nouns and verbs. The Purpose o� thi s chapter is to examine the tone behaviour o� verbs across ward boundaries. The di scussi on wi l l be �ocused on anal Vzing how the tone of the verb and verbal particle are a��ected bv the tone o� the Preceding nominal and verbal el ements. It wi l l be seen that two rule. which operate across morpheme boundaries ward i nternal l v , also occur across word boundarie •• Low tone SPread and High Del i nki ng. However, wi th the i ntroduction o� new data, the �ormulation. of the.. two rules as wel l as the Association Conventi ons wi l l be be requi red to undergo mi nor revi si ons. This chaPter wi l l �irst review the domains wi thin which verbal tone maV be al tered, and then procead wi th a descripti on and anal vsi s o� how the �inal tone of noun object. can aHeet the tone of the verb. Once the ruin have bean �ormul atad and thorough l v di scussed , the tonal behaviour o� verbal parti cles wi l l be examined . reveal i ng that rules alreadv ProPOsed can also appl v to verbal Parti cles. Qpmti o of y.rb TRo. lIh.yi OUf In ChaPter 1 , we l earned that tonal chan9as take pl ace between elements a verb Phrase. It was al so stated that the noun Phrase (NP) of immediatel v preceding the verb Phrase a��ect. verbal tone. There�ore, as i l l ustrated b.low. the tone ---- --- - - - -- of the sub ject NP a�fact. the tone of the �ol l owing verbal Particl., ----- ----- - -- ---- ------- - -- - 146 whi ch i n turn . can in�l uenc. tonal chang. an the verb. ( l ) a. SUbject + Verbal Particle + Verb -- � I� a str i ng o� verbal .l ement. is spl i t UP by the i nsartian o� a direct ab ject. that di rect abject wi l l al so a��act the tone � the i mmediatelY �al l awing verbal .l ••ent . In contrast . however . the tone o� the verbal el ement cannot a��ect the tone � the naun abject . Thi s barrier to tonal change is i l lustrated by the use a� square brackets bel ow. b. [SUbject + Verbal Particle] + [Naun object + Verb] "--"'" "---" I I . NOON OBJECT TONE AND n£ VERSI DESCRIPTION AND ANALYSIS A. Qncriptt m Let us � i rst l ook at how the tone � the noun object a��8Cts the tone a� the We have al readY s.en that a verb i s .i ther Hi gh , Mi d . or Law tone. Nouns verb . have a vari ety a� tonal shaPes: however , i n exami n i ng their behavi or with v.rbs � theY can b. cata9ari zed i nto thr•• grouPs accordi ng to thei r ward � i nal tane. Hi gh � i nal , Mi d � i nal , and Law � i nal nouns and Pronouns a��ect the verb i n d i ��erant waYs . Law tone verbs undergo one tonal change. When a High Dr Law H nal naun Precedes i t , the Law tone verb undergoes no change. However . a�ter Mi d �i nal nouns a Low tone verb becomes Mi d-Hi gh . (2) a. L + --> L L L ' augi a man ' b. H + L �al file':' ibad ---- .-- - . --- ---- --> H L ' 1III&i a rac k ' 147 L -> M M PerU tt --> Peru tl --- Peru gb�rl + c. M Paru ibk1 poru k.k�1'll > Peru -> H ' � a daughter ' / kral)a ' maat a daughter ' ' � a d�ughter ' It i s i nteresti ng to note how thi s resulting Mi d-Hi gh tonal contour i s l i nked to verbs wi th v�rYi ng numbers of TBU ' s . If a verb has onl Y one TBU. both tones re l i nked to that si ngl e TBU. If there are two TaU ' s. each tone l i nks to sePar�t. TBU ' s. whi l e on � three TBU verb. the High tone of the Mi d-Hi gh contour is l i n ked onl Y to the f i n�l TBU of the verb. Mi d tone verbs rllll� l i n Mi d after M and L nouns but are l owered to Low after H f i nal nouns when the verb i s i n sentence f i nal Posi t i on . + (3) a. L + Peru '_ M � I -F�'Mi I ' ... a daughhr ' -> M f eil 'lea , Dl!& a man ' M -> Dl!& + c· H M Dl!& nAA b. M L -> M II WA II fal '� h.? l, II H L --> .faUx£ nY. '_ --> .ftl,KA tllxCl ' � a roc k ' a rock ' --> fiuKi" c.'i'.U ' i nsuIt a rock ' When the Mi d tone verb i s no l onger i n phrase-f i nal Posi t i on , this tone l ower i ng rul e is bl oc ked -From appl Ying. ·waA b • waA -FiU. ' Di d h. IRR a rock? ' nyA l a f.t'l 'xa t.t.wl Ia ' Di d he � a rock? ' ----- 148 Hi gh tone verbs remai n Hi 9h excePt when Preceded by a Law � i nal noun. Any noun wh i ch end. i n a Law tone i nc l ud i n9 Mi d-Law. High-Low and al l def i n i t e suf� i xad nouns. whose f i nal Law tone i s never l in ked to the su� � i x i tsel f (see th. 3) . sPreads i t. f i nal Law tone onto the Hi 9h tone verb lower i n9 al l tonebearing units of the verb excePt the f i nal uni t o� mul tiPle TBU verbs. A si n91 e TBU verb i s si mP l y l owered to L i n Phrase f i nal Posi ti on . + (�) a. H -> H H H f;U '__ etd. b. M + ' ull. a rock ' H --> H gbaxa etr:t c• L + •dB ' aall a house ' H -> U I# m�B . .dB eta. L L > H .dB B ' tua r i ce ' --> �B w.' ' l,ggk at ri ce ' - �B ' Wl, r i ce ' - > ndB ksfi'd -> P�ri ndB kU�d ' aegU rice ' When in nan-phrase final posi t i on . a l owered si n9l e TBU Hi9h verb i s general l y Mi d (as seen i n (6b-d ) , whi l e mu l ti -sYl l abi c verbs remain Law-High . However . examPle (6e) shows that i t remains Low if a noun class clitic �ol l ows. (6) a. waa m�B li ' He tggk. r i ce ' b . Waa m�B U l a ' Di d he take ri ce? ' c. wea m�B U nll")Ull ' He took r i ce from the man ' I d. waa ndB U fc1l'kull ' He took r i ce from the rock ' e. wea m�B li ')66 ' He took r i ce from hi m ' IncomPletive verbs whi ch are formed b y suf � i x in9 an underlYin9 Hi gh tone - - _ . _ . _ - . _-- ------ 149 morpheme to the COmPletive stem. are subject to the same tonal phenomena as camPl eti ve verbs. Recal l that on the surface, the i ncamPlati ve suffix i a Mid tone a�ter Law and Mi d tone varbs and High after Hi �h tone verbs. (7) a. ---> M tH --> gbaxa M -- > H L -> -fIlI'NI Ud + H -> L LH !!dB 11 ---> �B B ' e.ting r i ca ' m�B W --) ndB l �ci ' taking r i ca ' M gbaxa b. H L-M + + fal 'K� Uti. c. L MH ta' ' ,hgwing a housa ' ' placing a rock (an) ' The.e. then , are tha basic tonal changes on verbs. Each tonal cate�orY of v.,.b. i s affected by one tonal changes L (noun ) H ( verb) --- > L L H M (naun ) L (verb) ----> M M H H ( naun) " (verb ) ---> H L The fol l owing d i scussi on invol ves the axami n at i on and anal Ysi s of each one of these tonal changes. It aPPears that the f i rst two al ternati ons represented above i nvolve sPreading rul es. wh....as the l ast one se... to di sPl ay a type of . di ssimi l ation phenomenon , whi ch operates onl y i n Phrase final Posi tion. B. An.l Y1Ii I 1 . Law SPrHdin9 Rul e Th. f irst tonal chan� to b. d iscussed her e i s th e matter o f the High tone verb acqui ring a Low-Hi gh contour when Preceded by a Low final noun. In ----- ------- - --- - - - -- 1� ChaPters 2 and 3, a simi l ar Phenomenon was observed where a High tone suH ix became Mi d tone wh en Preceded by a law tone noun or verb raat . Thi s loweri ng to Hid tane was anal Yzed to be the resul t of the law tone o� the noun root sPreadi ng onto the Hi gh tone of the su�fi x , creating a law-Hi gh contour which, in mast cases. si mPl ified to Mi d tone. Thi s law tone SPread rul e was stated in Chapter ( 33) 2 as fol l ows. lOW SPREADING - When a law tone i s fol lowed by a Hi gh tone acro•• a morpheme boundarY, the law tone spreads to the right . Thi s rule as stated . can also appl y across word boundari es. A. a resu l t , one can propose that the law tone o� the noun sPreads onto a �01 1 owi n9 Hi 9h tone verb. Thi s i s parti cul arl y cl ear an sin9l e TBU verbs. such as the one shawn bel ow. ' take rice ' lOW SPREAD If a Hi gh tone verb which has undergone low tane SPread . has onl Y ane TBU and i s �ound i n phrase final Posi tion. the conditions for High tone Del inking are met . Raeal l that the Hi gh tone Del inkin9 rul e. as formul ated in Chapter 2. states that the Hi9h tone of a low-Hi gh contour tone l i nked to a singl e Teu wi l l be del inked i f found i n Phrase final Posi tion or i f fol l owed by another High tone. In .ffect , thi s i s what haPPens to si n9l e Teu Hi9h tone verbs, as shown i n the examPl e. bel ow. --- - - -- ---- ------ 151 b. l � II. m�l � HIGH DELINI<ING � L H High Del inkin9 al so takes Pl ace a�ter naun cl ass cl i ti cs� whi ch are� i n thi s parti cular envi ranment , Hi gh tane. When a singl e TBU verb i s subject ta Law tene SPread , the deri ved Law-Hi gh cantour i s fal l awed by anather Hi gh tane. As a resu l t , High Del inkin9 9Ces i nto e��ect del i nkin9 the Hi 9h a� the cantaur tone. c. weaa 'II� I � B ')66 - > waa ndB B V I V H H L ")dCi � "' lH HV L L SPREAD -> waa nd B B � L r')CiCi ' He tack rica V H H fram h i m ' H DELINKING Hawever, i f anything but a Naun class c l i t i c lol l ows. the resul ting Low-High contour en the High ten. verb i s si mPl if i ed an the surface to Mid tane. Thi s Mi d tone could b e explai ned by the �ai lure af the Hi gh tane to del i nk i n nanPhrase � i n.l Position . waa m�l � l� l a -> waa m�l � l� la V L ... .. \ I ' Di d h. take rice ' H M L SPREAD The abave anal Ysi s of Law tane SPreadi ng inval ved the nouns whi ch are Law tane. It has al so been suggested in bath Chapters 2 and 3 that Mid tane be anal Yzed as rai sed Low tene. A distincti on af twa tyPes o� Law tene leads ta a daubl e tiered approach ta tanal anal Ysi s where bath tYPes a� Law tane are gi ven a Low tene an the Pri marY t i er whi l e the rai sed Law tene (ar Mi d tene> i s posi ted with a Hi gh tane an the subregi ster tier and the Law tane. wi th a - ---- ------- ----- 1 52 subre�i ster Low tone, as shown below: (9) x Low Mi d I X I L L I 1 I h Therefore. i � both o� th.s. ton.. Pos •••• a Low tone on the Pri marY tier. then theoreti cal l y both .houl d be abl e to sPread onto Hi gh tone verbs. since the Low tone SPreading rul e as stated above. does not sPeci�y a tone �or the subre9i ster ti.r. The .xamPl. below i n ( 10) i l l ustrat •• this theoretical possi bi l i ty. Howev.,. . as the deri vati on an shows. a lh tone sPreadi ng onto a Hi gh tone Produces incorrect sur�ace �or •• In �act. no tonal al ternati on occurs on High ton. verbs when preceded by a Mi d tone noun • ( 1 0) waa gbaxa l � -> V L I h \ H • � Rather I wa� gbaxa l � ""L ..�H ' h. took a t a hous. ' I wa� gbaxa I h LOW SPREAD Because Lh tone must b. b l ocked �rom sPreading the Low tone SPread rule must be r�ormul ated to al l oN the sPreading o� onlY Ll tones onto High tone nouns. Thi s revi sed Low SPreadi ng rul e i s restated b.low. ( 1 1 ) LOW SPREAD When a Ll tone i s �ol lowed by a Hi gh tone across a morpheme boundary, the Low tone sPreads to the righ t . Xl X ... ... 1 L H I 1 ---- 153 Let us now consider Low tone sPreadi ng on Hi gh tone double TSU verbs. In the . �ol lowi ng deri vati on ( 12) , the Low tone spread. onto the �irst High tone l i nked TSU. Thi s re.ults in a Low-Hi gh contour on the �i rst TSU o� the verb, whi le the �inal TSU i s sti l l l i nked to the Hi gh tone, thus creating the environment needed �or the del inking o� the High tone o� the Law-Hi gh contour . ( 1 2) � V waa �l� Plrt -> waa m�l � pJ�l -> waa m�l � ptrt V V L H �H ... L L SPREAD H ' He sol d r i ce ' L H DELINKING The setti ng �or Hi gh Del inking i n ( 12) i s sl ightly di��erent �rom the si tuations described earl i er (see (37 ) , Ch. 2) . In the ear l i er i nstances, the second High tone, whi ch , in e��ect , tri ggered High Del inking, was l i nked to the �ol l owing word or morPheme. In thi s ca•• , there is no seParate Hi gh tone l i nked to a di��.rent morpheme. Rather , the same Hi gh tone i s l i nked to both TSU ' s: to the one with the Low-Hi gh contour and to the �ol lowing TSU. There�ore, the Hi gh Del i nki ng rul e can be general i zed to incl ude the del i nking o� a Law-Hi gh contour which i s �ol l owed by anY TSU which is also l i nked to a Hi gh tone regardl ess � whether that Hi gh tone belongs to another morphame or not . Sal ow i s a r�ormulation o� the Hi gh Del inking rule. ( 13) HIGH DELINKING - Del ink a Hi gh tone which i s Preceded by a Law tone l i nked to the same TSU, and 1 ) i s �ol l owed by a TSU whi ch i s al so l inked to a High tone, or 2) i s i n sentence �inal Posi tion . or X A/ L H -_._ . _ ._ ---- ------- Xl 154 Tripl e TBU verbs Pose A b i t o� a problem. As the deri vAti on bel ow shows, simPly APPl Yi n9 LON tane SPread and Hi 9h Dal i nkin9 does nat produce the correct sur�ace �orm. L-SPREAD H DELINklNG In order to Produce the cOl"'rect surfAce �orm of Wii m�l� k.lAx£ , one HOul d hAve to apply LON tone SPread and Hi�h Del i nkin9. a s.cond tl .. . A S .hown ' He sPoi led rice ' LOW SPREADING L-SPREAD H DELINKING H DELINKING However , al ieNing Law SPreAd and High Del inking to aPply a second time, in ather wards , i terativel Y, seems somewhat cumbersome. In addition , i � i terati ve SPreading were al lowed, there i s nothi ng in the rules to Prevent Law tone SPread And Hi gh Del i nking to APpl y to the l ast TBU as wel l . thus Producing a LON tane an al l three TBU ' .. In �ACt , the �inAI TBU o� a tHO 01'" three Teu verb i s � Law tone in anY WMvironmant2• Ther�ore AnY i terati ve sPreadi ng And del i nking rul es woul d have to be constrai ned with the condi tion that SPreadi ng and Del i nkin9 do nat occur an the � i nal TBU o� a mul tiPle TBU noun. ---_ . . _- _ . _ -_ . _-- ._-_._--- ---- ------ 155 An al ternAti ve sol uti on would be to l i mi t the aPpl ication of the Assoc i ation Conventions by having each tone l i nking to onl Y one TBU. Such a restrai nt would require the fal lowing reformulation of the Associ ation Conventions: ( 16) ASsOCIATION CONVENTIONS MAP a sequence of tones onto a sequence of free tone-beari ng uni ts 1 ) from ri ght to left 2) i n a ona-te-ane rel at i on until al l tones have been l i nked . Recal l i ng that in Suc i te. tone. are l i nked from ri ght to left. a three TBU Hi gh tone verb would then initial l y have onl Y the final TBU l i nked to the Hi gh tane. as shawn bel ow: ( 17) kal AKi I H Del aYing the l inking of the ather TBU ' s to the Hi gh tone would al l ow mare freedom for the SPreadi ng of the Law tone across the ward. As the deri vati on below shows. the Low tane SPreadi ng rule as stated above ( 1 1 ) . has the Law tone spreadi ng and l i nking to the final High tone l i nked i ntervening TBU ' s: TBU as wal l as al l of the ( 19) waa �l� kal .xf -> waa �l� kal axl -> waa �l � I L I H ' He sPoi led rica ' L,; �" l H r L-SPREAD ---- - - -- - - -- - ---- L -> waa m�l � H �� �x{ � I L H L DELINKING Si nce it al readY has been stated that final TBU of a mul tiPle TBU verb . ���� �I we a Hi gh tone cannot be del inked from the know that Hi gh Del i nking cannot aPpl y --- - - - -- ---- ------ ---_ . 156 here. TO . Prevent Hi gh Del i nkin9 �rom taki ng pl ace, a rul e del i nkin9 the � � of the resul ting Low-Hi gh contour must be �ormul atad. First o� al l , Low Del inkin9 i nvol ves Del i nkin9 the tone on the l e�t si de o� the contour rather than on the right sid. as i s the nature of the High tone Del i nkin9 rule. It must be noted that Low tone does not del ink �rom a Low-High l i nk.d segm.nt i� it is l i nked to onl Y one TBU ( 1 9a) or i� it has sPread to onl Y one TBU ( 19b ) . ( 19) a. 9bon-ci,,� ' 9ranarY-9oad ' 1\ .vL t� ' h. rice took? ' b. waa m�l � l � l a 1/1 L H It d.l i nk. onl Y i f the Low ton. i s al l owed to l i nk to other TBU ' s wi thin the word that has been subject to L sPreadin9, as i l l ustrated bel ow in (c) c. waa m�l � L H Ther�ore, LOW DELINKING can b. stated as fol l ows: (20) LOW DELINKING D.l i nk a Low tone from a Low-Hi 9h contour when the Low tone has alreadY l in ked to other TBU ' . in the word: [ X Xl l/\ L H There are a few environments whare both Low tone Del inki ng and Hi gh tone Del i nking can be moti vated accordi ng to the sPec i f i cations 9i ven above. For eKamPl e. i f a multipl e TBU verb whi ch has been subject to Low tone spread i s -- - - -_ . 'f 157 Also �ol l owed by the High tone noun ClASS cl i t i c , whi ch, as we hAve al readY • seen, tri 99ars Hi gh ton. Del inking o� Preceding Low-Hi gh contour, which rul e takes Precedence? Such a di lemmA i s i l l ustrAted below. 'He mad. on. �or him ' In order to make the correct Prediction, one must order Low Del inking be�ore High Del i nking. However, when the Hi gh tone verb consi sts o� onl Y one TeU. Hi gh tone deli'n king bkes PlAc •• AS ..., i n (22) b.l ow. Why? We sea thAt the conditions for Low tone del i nking cAnnot ba m.t - Low tone hAS spread to onl Y on . TBU of the word - thara4or. High tone del inking CAn take Place. (22) waa w� B �Ci V1 V L H H In conclusi on , the anAl Ysi s ProPosed hare requi res the reformulation o� the Low tone SPreadi ng rul e ( 1 1 ) , the Hi gh Del inking rule ( 13) , And the Association Conventi ons � or SucHe ( 16) , as wel l AS the Addition o� the Low Del i,lk.in9 rul e (20) . The naw Law Del i nkin9 rul e WAS s••n to be ordered a� ter the Low tone SPreAdi ng rule but ba4or. Hi9h tone Del inkin9. 2. Hi 9h Ton. SPr••ding The second tonal Process to be exami ned here i nvol ves the rai sin9 o� Law tone verbs to Mi d-Hi9h tone when Preceded by A Mid tone. Thi s process can be more easi l y anal Yzed on Low tone incomPleti ve verbs thAn on compl eti ve verbs. Recal l that Low tone i ncomPletive verbs consist o� a Law tone verb root �ol l owed by a Hi gh tona incomPletive sui�i K ' as saan bel ow i n (23) . The most 158 obvious sol uti on would be to moti vate a tYPe of Mid tone sPreading rule. whi ch woul d spread the Mi d tane onto the Low tone root . Since there i s no trace of a Low tone an the verb after thi s Mid tone SPread , ei ther the SPreadi ng would automati cal l Y have to tri gger del eti on of the Low tone. or a seParate Law Del eti on rul e would have to be formul ated . (23) a. CV CV-CV l., � 1 I @ H M The dif�erlnce between Mi d SPreadi n9 and Law SPreadi n9 i . that when a Law tane spreads onto a Hi gh tone, that Hi 9h ton. can sti l l effect tonal changes an fal lowing wards, whi l e Mi d SPreadin9 onto a Law tan. obl i terates anY Power that Law tone maY have had to tri gger tonal chang••• Thus w. saY that Mi d sPreadi ng tri99ars Law tane del .ti on , whi le Law SPreading, an the ather hand cal l s f or d.l i nkin9 of the High tone onlY when certain condi�ions are met . We run into Probl ems, however, when trYing to aPpl y the Mi d SPreadi ng rul e as stated above. to Low tone ,gmpl.tiv. verbs. I t was stated earl i er that a Low ton. verb becomes Mid-Hi gh when preceded by a Mi d tone noun. For incomPl eti ve su�fi xes, al l we had to do was to moti vate a Mi d tone SPreadi ng rul e onto Law tone. The resulting Mi d tone verb tol l owed by a High tone suffix resul ted i n a Hi d-Hi gh tone an the verb. There is, however , no Hi gh tone suffix an a comPleti ve verb . As a resul t , si mpl y sPreadi ng Mi d tone onto a Low tone comPletive verb would nat Produce the desired Mi d-Hi gh tone, as shawn b.low. (24) cv cvcv ->*cy J;�V � ,--" "! � M M © Poru gb�l -> ·poru gbara rather' Peru sbar M SPREAD It wi l l be seen that Mi d SPreadi ng onto Law tane nouns Yiel ds the Predi ct- - --- --- . ----- ,. 159 abl e result of a si mpl e Mi d tone, so why a Mi d-High contour would surface an Law tone verbs i s somewhat Puzz l ing. The fi rst attempt at a solution would be to establ i sh whether a compl etive verb has a tonal structure consi sti ng of a root tone fol lowed by a tonal morPheme. One Possib l e historical eXPlanation i s that the compl et i ve morPheme did contain a Hi gh tone at one stage. I n Chapter 2, we suggested that a Mi d tone comPl eti ve verb maY have been a raised Low tone. If we speculated that the farmer Law tone was rai sed by a Hi gh tone ComPleti ve morpheme, than we can extract a tinY bit of evi dence that thi s Hi gh tone came from a hi stori cal High tone comPl etive morpheme, whose trace can sti l l be sean i n the aPPl ication of a Mid tone sPread rul e. Whether thi s theorY holds anY val idity or not , however. does not real l y bear uPon a sati sfactorY sYnchroni c sol uti on , because in al l other environments, Low tone comPl eti ve verbs never behave as though they have a final Hi gh tone. For exampl ., they contrast with Low tone incomPl eti ve verbs <whi ch do have an underlYi ng Hi gh tone suff i x ) when fol lowed by a Noun Cl ass cl itic. In (25) below. note that both verbs are Low tone on the surface. However . the cl i ti c i s Mi d tone <underlined ) when Preceded by a comPl etive Low tone verb <a) and Hi gh tone after an incomPl.ti ve Law tone verb <b) . The eXamPle in <c) i ndicates that the cl itic i s also High tone after completive and i ncompi eti ve High tone verbs. <2�) a. Compl et i ve Low tone verb wa� wu w�l � � 11 ' he took <C) i t off of hi m ' b . Incampl etive High tone toW Ya Hn� � 11 ' he l i ed <C) an hi m ' c. ComPl eti ve and Incompl etive Hi gh tone. waa w� l � � ' he took <C> i t from him ' ' h. is taki ng < IN> i t off of hi m ' ---- 160 Low tone comPl eti ve verbs. therefore, contrast wi th Hi gh f inal verbs i n behaviour. and a s a resul t , cannot b e consi dered t o bear an underlYi ng Hi9h Hnal tone. Since Posi ting an under lYing Hi gh tone suff ix morPheme i s not feasible, same tyPe of rul e wi l l have to be moti vated to al l ow �or a Mid-Hi gh contour. One possi ble sol ution i s that , i n addi tion to Positing a Mid tone SPread and Law Deletion rul e on a Low tone verb, a High tone in.erti on rule i s al so introduced to insert and l ink a Hi gh tone to the ?i nal TBU of the verb. Thi s parti cular development would al so reClUire the rule a? Mi d tone Oel inking (53) . al readY introduced i n Chapter 3 (p. l06) , i n order to simpl i fy a Mi d-High contour on a singl e TaU. Thi s tentative series o� rules i s i l lustrated bel ow. (26) C...VC�V me k�r;}r)tl tv ;V I I-:�� - 1'1 1'1 I -> me kar.., a -> me karat) ' -> cy t;vg�� -> cy t;V�� � f _� H_, M M H SPREAD H INSERTION C guide You I M OELINKING Another Possi b l e soluti on i s to take advantage of the suggesti on alreadY made i n thi s thesi s that the underlving features a? Mi d tone are best rePresented on two tonal tiers, with Low tone Posi ted on the PrimarV t i er and High tone on the secondarv or subregi ster tier , as shown bel ow in ( 27a) . Al though Cl ements did not introduce the i dea of l i nking the two tiers nor the Possi b i l i ty of l inking a subregi ster tone to more than one tone on the primarv tier . (as shown i n (b» . nor the unl i kel Y suggesti on that tones of the subregi ster tier maV al so l in k directl y to the segmental tier (c) . the �ol lowing anal Vsis wi l l ProPose that these i deas are not onl v ?easi ble. they al so Prove to be verY practi cal for thi s particu l ar anal ytical Probl em. -- _ . _ - ---- ----- 161 (27) a.CVCV V I I L L L h h I c. CVCV b. CVCV V V h If we al low mul ti pl e l i nking of the subregister ti er . then l ogi cal l y tone sPreading rules could also be mati vated at this level . Returning to Mid tone SPreading an i ncomPl eti ve verbs. one can see how the sPreadi ng of tane on the subre9i ster tier can create the effect of Mi d tone sPreadi ng. SUBREGISTER HIGH SPREADING When the High tone of the subre9i ster tier sPreads to the fol lowing Low tane. i t tri ggers the del eti on of the subregi ster Low tone. The reason for thi s del etion is that the l inking of twa subre9ister tones t o one PrimarY regi ster . " .. tone cannot b. al l owed by vi rtue of the fact that such a construction has no concei vable surface real i zation. If. historical ly. Mi d tone was consi dered a raised Low tone. i t i s conceivable that a rai sed Low tone could rai se the tone of subsequent Low tones. Thus the l ower l evel � rePresents this ' rai si ng ' Phenomenon through , the spreading to fol lowing Low tones. Let us now consi der the mare comPl icated case of Mi d tone sPreading onto compl.U ve verbs where Mi d tone sPreading ."esu I ts in a Mi d-High contour. If the subre9ister n spreads onto PrimarY re9i ster tones. would it also be possible for the n to also i ndePendentl Y l i nk to the se9mental t i er? High contour could be moti vated . ---_. . _- -_ .. - - -.--- If so. then a Mi d- (29) r 162 c cv, L INDEPENDENT SUBREGISTER SPREADING TO SEGMENTAL TIER I '. L I V ,' ... h ' It seems that the lower register high tone not onl Y rai sed the Low tone through the Process of · sPre.ding ' , i t conti nued that rai sing effect by l i nking i ndependentlY to the segmental tier. The effect of independent l i n ki ng creates a cantour tane. The questi on i s whether thi s i s just a fancY mechanism to descri be an otherwi se i nexpli cable deri ved contour tone or whether thi s aPPraat:h can be moti vated elsewhere in the language or even in other languages. In SUPYire, a nei ghbouring dial ect in Mal i , Mid tane does not sPread anto Law tane verbs (al though a Mi d tone sPreading rule oPerate. on nouns in the exact same waY as i t does in Suci te) . Instead, Law tane verbs acqui re an Hi9hLow contour when Preceded by Hi gh final or Mid fi nal nouns. (30) u nYYE na f�un 5h6� /sh��/ he TA TA Peanuts buy u nYYE na mP� sh�� he TA TA sheeP buy · he i s bUYing Peanuts ' Suci te: wu Ya f�un sua · he i s buYing a sheeP ' Sucite: wu Ya nc� 500 Thi s sl i ghtl Y di fferent phenomenon coul d be eXPl ai ned by the inde�endent l i nking of the subregi ster Hi gh tone to the Low tone segment wi thout l i nking to the Law tane i tsel f , and thereby creati ng a Hi gh-Low contour . (31 ) f�un shdo ---- ----- cv I cv /1 I �I � L -,- L · buYi ng rocks ' · buYing peanuts ' cv I H V/ h tv -1 / L I 1 --'--_._---- '---- ------- '" 163 In his anAlYsis of SuPYire tone, CArl son refuses to Posit an underlYing tone for thi s set of verbs, because Law became. Hi gh-Lew i n so manY envi ronments. The aPProAch presentad Above can moti vate an underlYing Law tone verb which has been subject to indePendent sPreadi ng of the subre9i ster hi9h tone. Therefore, , independent subre;ister Hi gh tone sPreAdin9 does recei ve same suPPort i n SuPYire3• Hawever , this description sti l l does nat explain why i ndePendent spreAdi ng of a subre9i ster High tone to a segment shoul d take Place an verbs but nat an naunsThe quesU an also remains 'fas ta wh.ther thi s naUan of indePendent l i nking of the subre9istar tane CAn be defended wi thin the theoretical framework of autasa9mental phonol ogy. The nation of l inki ng and sPreadi ng of the subregi ster level to the Pri mArY regi ster l evel i s a rather navel idea in itsel f . It has bean Proposed that this subre9istar l evel i s an a sePArate tier f rom the tone an the Pri marY regi ster tier . However , i t was Painted aut that i n accordance wi th ' the tone feAture system as prOPosed by Clements. thi s subre9i ster tier is Si mpl y a fi ne-tuni ng of the Pri marY regi ster tone, exi sting onlY as a functi on of the Pri marY re9i ster tane. Wi thi n the Clements system, the subre9i ster tane does nat oPerate AS a seParAte and ful l tone itsel f . Therefore, wi th these l i mitati ons i n mind , it would be somewhat Probl ematic conceptual l y, to 9rant indePendent l i nking al the subregi ster tier to the segmental tier . One Possi ble aPPr oach i s to el eVAte the status of the subre9i ster tone. Cl ements based his model an the concePt of tone-spl i tting. WhAt i f the nation � tane spl i tting could b e further stretched theoretiCAl l y with the canceP� a� fusi n94 tones from two seParate but indePendent tiers? 164 ( 32) FUSION CV I � The no�ion o� h i erarchv would bv necessi �v s�i l l remain. !� no�, �hen HI and Lh would no� be di s�ingui shable wi thi n the model . In ChaPter 5. �he necessitV o� thi s distinction wi l l be demonstrated . Yet the notion o� �usi on o� two tones woul d al low Possi ble l i nking o� the lower level tone to i ndependentl v l i nk in certain l imi ted environments. One maV be tempted to specul ate that thi s svnchroni c �usi on Process was Preceded hi storical lv b v the Presence o� a � sequence tones that eventual l v �used into one. I t has al readv been stated that no preci se statements can be made at thi s time about the histori cal evol uti on o� tone in Suci te. For the PurPoses o� thi s thesi s. we shal l pl ace the second tier o� tones bel ow the � i rst and vet al l ow the second or subregi ster t i er to l i nk to the se9mental tier bv Process o� trans�ormation o� the subre9i ster tone to a pri marv regi ster tone as i t passes through �he fi rst �ier as seen in ( 33) bel ow. ( 33) x I L I h -> x " , L H ,,'" h The ��ect i s �hat o� rai sing a tone to �he extent that i t l eaves the Low �one regis�er and moves i n�o the Hi gh �one regi ster . Returni ng �o the Suci te examPl e. then . let us examine the envi ronment o� subre9ister Hi 9h tone l i nking. The subre9i ster Hi gh sPreads to the PrimarY ----- -- - - - - - - ----------- ---- --- ------ 165 tier Law tone of the verb as wel l as to the segmental tier onl Y when that subregi ster Hi gh is l exi cal l Y l inked to a Law tone i tsel f . as .hown i n (34a) . If the subre9i ster Hi gh was l ex ical l Y l i nked to a Hi gh tone then no Hi gh spreadi ng woul d take Place (34b ) . (34) a) X X I L LI h I b) *X I f' X 1\ 1):) h I V: ' Thi s i s in contrast to SuPYire where indePendent Hi gh SPreadi ng occurs on Low tane verbs from bath Hh and Lh final nouns (see (31 ) above) . At thi s stage then. a subre9i ster h i gh sPreadi ng rul e woul d be stated as fol l ows: (35) HIGH SPREADING A subre9ister high. l ex i cal l Y l i nked to a Low tone. sPreads to a fol l owi ng Low tone. thus del eting anY subre9i ster tone formerl Y l inked to that Law tone. X X L L h CD I I L. , 1 A seParate rul e must be devi sed for independent Hi gh sPreadi ng. (36) HIGH SPREADING TO TBU SPread a subre9i ster h . l exical l Y l inked to a Law tone. to the f i nal TBU of a Low tone verb . cv I cv [ \) V-.1 L ----- 166 Thi s assumes that High tone h.s alreadY sPread to the Law tone. del eting the subregister Law tone. Below i s an examPle showing both Hi gh Spreadi ng rul es. (37 ) Poru U -> Poru tl L L L L h ID I 1 1 � .. h (i) (.. ... 1 H SPREAD r\ J H SPREAD TO TSU It wou l d b. more convenient , Perhaps, to state these two rules in ane rul e. si nc:e 'l both o.f: them are tri ggered by the same envi ronment and one is a .f:ol l aw-uP o.f: the other. It wi l l be seen. however . that the .f: irst . but not the second rule aPpl ies to nouns. Ther�are. at thi s juncture. Hi gh sPread to the Pri marY tone Low tone verb. with mare than one Teu whi ch are subject to High SPreading require some addi t i onal adjustment rul es in order to l ink the tanes to al l the TBU ' s. From the d i scussion on Low tone SPreadi ng it was deci ded that the Associ ation Canven�j ons l i nk tones .f:rom Right to L�t , l inking a tane to onl Y one TBU. There.f:ore. the Law tone is l i nked to the .f: i nal Teu a.f: doubl e TBU verb . as shown i n the deri v.ti on in (38) . a Law tone The High tone SPreadi ng rule sPec i � i es that the subre9i ster High tone sPread. to the PrimarY tier Low tone a.f: the verb . Note that i .f: i t was al l owed to spread to anY se9ments precedi ng the Pri marY tier Law tone. a incorrect Hi gh-Mi d tone would evolve an mul tipl e Teu verbs Cb) . There.f:ore. the subre9i ster Hi gh tone i s restrained to sPreading initial l y to the PrimarY Low tone. Secondly. we observe the rul e o.f: s�eadi ng the High tone to the .f:inal TBU o.f: the verb and l i nking it to the Teu a.f:ter the Law tone of the verb. Thi s now resu l ts i n • Mi d-High contour tone l inked to the .f: i nal TBU and no tone l i nked to the � i rst TBU. In order to ._- _ . _ . _ -- ---_ . _ - -- ---- ------ 167 correc� �hi s Problem, some l inking adju5�men�s wi l l have �o be made. A rule cal l ed Li nking mus� ba formul a�ed (see (39) below) �o l ink �he remai ning TBU ' s of the verb �o �he verb �on• • Whe�har �hi5 i s done befQra or after Hi gh �one SPread is no� cruci al � bu� such a l inking Process mus� �ake Pl ace af�er Low tone SPread so �ha� �he Low �one can have �he freedom �o sPread �o �he ri gh�mos� TBU (see Precedi ng di scussi on on Low �one Spread ) . Af�er Linking �akes pl ace, �hen � �he Mid-high contour �one i s si mpl i f ied by del inkin9 �he Mi d �one, �hat i s, �he �ona on the l ef� si de of �he con�our. (3e) a . wa� Poru gb�r� 1/ I I Mf_ L -> II Ml.. L L - "'1 h <D H SPREAD \{Lj wa� peru gbara -) wa� Poru gbara I L I .-"""_"" r-- r, L .. J H SPREAD TO TBU II ML L Ii LINKING -> V� waa poru gbara " ML L L h MID DELINKING b . *wa� Poru gb�ra II ML I / I L L' h I L I 1 (39) LINKING - Li nk anY free TBU ' s �o �he �one of the morpheme to whi ch i t belongs. [x Xl " J T The rul e label ed Mid Dal i nking i s verY si mi lar �o �he Low Del inking rul e (20) s�ated above. I n both cases. �he l ef� si de of �he con�our tone is del inked when that �one i s al so l i nked �o another preceding TBU wi �hin �he morPheme boundarY. For �hi s reason. I proPose a more general rule to cover bo�h Mi d �one --- ---_. ---------- --- - 168 Del i nk i n9 and Low �one Del i nk i n9 - Le��side Del i nkin9 (LS Del i n kin9) : (40) LEFTSIDE DELINKING Del i n k the tone trom the l ett si de ot a contour tone when �he l ettsi de �one has al readY l i n ked to o�her Preced ing TBU ' s i n the word. In examPl e ( 39a) above, we have see how LS Del i nk i n9 aPPl i es to verbs. In Chapter 5, i� wi l l be shown that LS Del i n k i n9 i s equal l y appl i cabl e to nouns. The anal Ysi s ot Hi gh tone sPread anto Low tone verbs produced the compl i cat i on of �rYi ng to determi ne how Hi gh tone sPreadi ng cou l d Yiel d a Mid-High contaur . Two rul es were ProPosed �o cover t er thi s phenomenon , ,I:he High SPread and the Hi gh tone SPread to TBU. Althou9h the Hi gh tone SPread 'co TBU rule maY seem to Pose some theoreti cal Probl ems, and thus stand wi l l b e seen i n Chapter 5, en a shakY ground , i t that the notion at l i n k i ng a subre9i ster tone i ndependentlY �o a se9ment can be fur�her subs�antiated . Because the Associ ation Conventi ons as stated i n the di scussi on on Low tone SPreadi n9 l i mi t the l i nkin9 of tones to TBU ' s , an add i t i onal rule i s requi red to l i nk anY remai n i ng free TBU ' s to tones a�ter the appl i cati on o� the spreadin9 rul es. Thi s ru l e , stated above. has been cal l ed the Linkin9 rul e (39) . Fi nal l y , i t was consi dered necessarY to revise the Low Del inkin9 rule to incl ude the del i nk i n9 of anY tone on the l eft side contour tone gi ven the condi t i ons in the rul e ( 40 ) . In summarY, then , �he rul es are 9i ven below: --- ---- - -- -- ------ 169 (35) HIGH SPREAD (36) HIGH SPREAD TO TBU (39 ) LINKING (40) LS DELINKING 3. Mi d Ton. Loweri ni The � inal rul e we shal l discuss wi lh in thi s secti on i s a sentence final rul e invol vi ng Mi d tone verbs. A Mi d tone verb i s lowered to Low when preceded by a Hi gh final noun (4 1 ) M - > L I H ___ 01'" Pronoun and �ound to be i n sentence .f i nal position. II ev CVCV## -> CV evcv## I H V M I V Wil� �' tv -> waA � t� H H L M H L ' He carried a rock ' Why a Hi gh tone would force the lowering of a Mi d tone i s uncl ear at thi s Poi nt. At the outset i t aPPears to be a type of a di ssi mi l ation Process. A �ew Paral lel s can be drawn with other sentence final Phenomena. Fi rst of al l . recal l from ChaPter 3 that Hi gh final indef i n i te nouns exhibit a fal l i ng tone Phoneti cal lY when in sentence final Posi tion ( see (42a) bel ow) . A phrase fi nal Low tone In••rtion rule ( 27) was ProPosed. It seams that this tendency for a High tone to fal l i s transferred to the Mid tone verb, which col l apses from the weight of the fal l and becomes Low tone (42b ) . Not. that the Hi gh final noun . which would normal l y exhibit a �al l i ng tone i n sentence final Posi tion. no longer fal l s when fol lowed by a Mid tone verb or anY other el emeni: . Thai: fal l is transferred to a Mid tone verb . however . onl Y as l ong as the verb itsel f i s ----- �- - . -..- . --- --- - -- -- -- �-- ---- - 170 i n sen�ence �inal Posi �ion . As soan as same�hing else �al laws �he verb . no �al l ing phenomena �akes Pl ace. as shawn in ( 42c ) . ' rack ' b . wa� UUxa hxo -> wa� � � aK.{ bnc� c . wa� ��l '�a �uxo l a On. ' He carri ed a rack ' ' Di d he carrY a rack? ' may ask why a Mi d tone verb wou l d b. susceptible to thi s tal l but not a Hi gh tone verb. tor examPI •• It a Mid tone verb i s Anal Yzed as a rai sed Law tane. as has al readY bean suggested . thi s taI l i ng Phenomenon could be descri bed as Pushi ng the rai sed Low tone back to Low. or i n terms a� �eatures. as n (43) CV cvcv J I H L I :t­ h h Such a Hi gh Del i nking rul e maY be tormul ated as tol l aws. (44) SUBREGISTER HIGH DELINKING: Del ink a subre9i ster High tone �rom a Pri marY level Low tone when Preceded by a Hh tone and i n Phrase �inal Position . x X II H L I I h I * h The moti vation ot this Hi gh del i nking i s nat enti rel y cl ear . however. Recal l . --- - - - -.. -. ---- --- --- .- --- -- ------ ----- . 171 al so. Another Phrase � inal rule. al so labeled Hi gh Del i nking. but whi ch invol v•• the del i nkin9 a� a Pri m�Y tier Hi gh tone i n Phrase � i nal Posi tion when i t shares a TBU wi th a sPreading Low tone. as shown bel ow: (45) CV CV II l,./1 L H Al though the Lh �eature bundle is not a sequence � tones and there�ore. not a contour . it i s Possible that some �eature simPl i � i cation Process not unl i ke contour si mPl i f i cation of d.l i nking a Prim�Y Hi gh tone � a Low-Hi 9h contour . i s taki ng Pl ace an Lh verbs. triggering the del inking o� the subre9i ster Hi gh �roll the L.ow tone. The �al l i ng o� the Hi gh � i nal noun in sentence � i nal posi tion and the lowering o� Mi d tone verbs a�ter High �i nal nouns in sentence �inal Posi tion maY al so lead one to susPect that these two Processes are moti vated by the same underl Yi n9 �eature. A �al l i n9 tone an a Hi gh �inal noun can be represented through the insertion � a Law tone �eature in sentence � i nal Posi tion : However . i� one i nserted a �i nal Low tone after a Mi d tone noun . a Mi d-L.ow tone rather than Law tone i s produced. as shown below. nYi .. /"' .... L t � I� the Law tone was inserted b.tween the noun and the verb. a Low-Mi d __ant our -- - - .- .-.. ---- ----- ---- ------ 172 would resul� on the verb. In ei ther even� . High Del inking would sti l l have �o be motivated . i n order �o del i nk the subregi ster Hi gh tone. In sPi �e of the paral l el that can be found between the behaviour of �he Hi gh final noun and Mid tone verbs which are Preceded by Hi gh final nouns. there does no� seem to be a clear waY to capture thi s general ization in a si ngl e rul e. The rul e of subregi ster Hi gh Del i nking i s also verY simi l ar to the rul e of Hi gh Deletion. i ntroduced i n ChaPter 2. Both rules ei ther d.l et. or del ink the subre9i star Hi gh tane � a Lh verb and both have the same effect � l oweri ng a Mi d tone verb to Low tone. However . they occur in di fferent envi ronments and at different stages i n the deri vation . High Del etion takes Pl ace early i n the deri vation before Law tone SPread wi thin word boundaries. whi le High Del inking takes Place in a speci f i c Phrasal envi ronment . Thus High Deletion aPPears to be a l exi cal l Y based rul e whi le High Del inking takes pl ace at the Phrase level . Thi s Hi gh Del inking rule. then. i s simi l ar to other �onal Phenomena found i n the l anguage. It take. Pl ace at the same stag� o� derivation as the PrimarY ra9ister Hi gh Del inki n9 rul e and the Law tone insertion rul e. and it al so Produces the same surface form as Hi gh Del etion. Any possi bi l i ty of making a general ization covering al l of these simi l ar character i stics wi l l have to wai t for further research . - " -- ---- T 173 I I I . THE VERBAL. PARTICLES Thi s sec�ion exami nes �he �onal na�ure of �h. verbal Par�i c l es i n addi � i on �o discussing how their �on. a��ects the tone o� the verb . As the tone o� the verb i s sensi t i ve to the tone o� the Precedi ng NP, namel Y the objec� , 50 the Verbal Particle is subject to tonal rules tri ggered by the tone o� the preced­ i ng Subject NP. Verbal Par t i c l es are tYPical l y never � ound i n i solation ; thei r underlYing tone can onl Y be determi ned i n the contex t o� the verb Phrase and 5en�ence. An a��empt at anal Ysis i s turthar comP l i c ated by the tac� that even though Verbal -Parti cl es Possess onl Y one� or at the most � two� tone-bearing uni ts, theY aPPear to be tonal l Y comPl ex . Thi s studY wi l l attemPt to di scover what the underlYing tones mi ght be, and outl ine an anal Ysi s that would account tor thi s compl ex i ty . A. Order Within th. y.rb pbr... � ore going on �o d i scuss tone� i t might be hel P� u l t o examine br i ei l Y the nature at thi s Verbal Parti c l e ' sl ot ' wi thi n the verb phrase. A variety o� verbal parti cles can be i nserted i n this Posi t i on i mmedi atel Y Precedi ng the verb. A di rect object� by i ts Posi t i on in the sentence, however . �an seParate al l o� these Parti c l es from the mai n verb . No nominal el ements can be i nserted among the verbal particles . The terms, Tense, Aspect , Modal (TAM) markers, would not adequatel Y cover tor al l at these Particl es � because the negati ve marker i s also i ncl uded, and there are a COUPl e o� par t i c l es Precedi ng the TAM markers that I si mpl y reter to as Pre-tense markers� tor l ac k o� a better l abel . Precedi ng even the negat i ve marker . whi ch normal l y precedes al l Verbal Parti cles, Pre-tense markers seam to be somewhat adver b i al i n tuncti on . Sel ow i s a schema o� the al l owed order ot these parti cles . Tense and Modal markers ----- - -- --.. - -- - ---------- ---- ----_._--- --- _ ._-- - 1 74 are nHen marked for ei ther Compl etive <C> or IncomPl eti ve < I ) AsPect <A> . (47) PRE-TENSE NEGATI VE - HABITUAL-A TENSE-A MODAl-A � " saa.' gO ' <U> nH PAST <C> Yi ' agai n ' i '" m na PAST <U) ' also ' PAST <C> , saa ' gO ' < I > ca� PAST <C) caa PAST ( I > , b;aa' come ' <U) Unmarked for asPect = <U) st FUT (C) (normAl l y indi cates cOmPl eti ve> COmPl eti ve = IncomPl eti ve baa • come ' < I ) s! da FUT ( I > (C) Ya PRES < I ) = na?i PRES < I > , (I> Though thi s l i st contai ns most verbal particles. i t i s not exhausti ve. Wi thin narrati ves. a number o.f these parti c l es also carrY di scourse functi ons. combi n- ing i n numerous waYs. It i s not wi thin the scoPe of this thesis to discuss the semantics and di scourse ProPerties of these particles. The purPose i s to present most of the parti c l es and examine thei r tonal behaviour wi thin the simpl e sentence. , (48) a. mo na Bel ow are some Possible combi nati ons of parti cl es. ca� S:l'i';l ' You had cooked ' You PAST <U> PAST <C> cook <C> b. mo na caa s��d You PAST (U> PAST < I ) c . mo m�?a s! cook ( 1 ) P�r€ You HAB FUT (C> sel l <C> ---_ .- . _- _ . _ ._ - ' You had been cooki ng ' ' You wi l l sel l habi tual l Y ' T 1 75 d . rna n U saa ' You went to sel l PirU I You PAST <C) M ' gg l < l ) sel l < l ) e. me sen?' n { � ' You have came agai n l Pan You a9ai n PAST <C) came <C) � . mo m ' yl ni l min ' You d i d nat came ei ther I pan You al so NEG PAST (C) came (C) NEG Nate that the translations o� these Parti cl •• are nat Precise . For instance. I have nat i nd i cated at what Points i n the Past each Past tense marker i s used �or . Determi nat i on o� the .xact meani ngs o� these particles i s rather comPl icated and requi res a rather rigaurous studY o� the di scourse behavior . B. Awew:t in the Ver b Phr'M CalnPl .U v. asPllCt an TAM markers i s sDRl8ti mes unmarked . that i s . Possessi ng no se9menta1 or tone features. More �requent l Y . however. the comPl eti ve asPect i s characteri zed by a �i nal l ow tona someti mes associ ated wi th tha vowel - 1 . The incanPl .tiv. asPect . on the ather hand . i s consi stentl y Mi d tone and someti mes associ ated wi th the vowel :&. Sal ow are same eXAmPles � unmarked and marked completi ves i n contrast to their in comPl eti ve counterParts. (49) Inc;gmpl.tiv, COmpl eti ve <Unmarked) , " wu na s� sb?b Iso?ol cook wu n-' sa pi r€ wu I ' he had cooked? I - ---- ------ ------ sbbd IsoDl"'i l ' he had 90ne to be cookin9, se1 l i n9 1 Inc;gmpl,tiye COmpl etive (with L tone) " , sel l ' he had 90ne to cook . sel l wu nei. ca-a so?o 1 a '" na sa-iI, cook , wu n� ca-a soari l a ' he had been cooki n9? 1 176 AsPect is marked on connecti ves between verbs in a verbal chai n . as seen in the examPles bel ow. ( �0 > ComPl eti ve a. wu l ptrt l k� --> waa pir�� k� ' he � i n i shed sel l ing ' --> waa w� l �� Pan ' he brou9ht i t ' C f i n i sh he C sel l b . wu l wu (C> l Pan H C come he C i t take c · wu l twd ' He vomi ted too much ' he C vomit d . nde l wid I l Y� r � --> ndaa wir�� Y�r� ' I got uP ear l y ' C get uP C be ear l Y (51 ) I ncomPleti ve ( IN > a· wu YA wtrU DA IN 90 ( IN ) he IN be ear l y b . wu YA �yU � h. IN hurrY (52) ComPl et i ve ' he i s hurrying �u 90 ' 5' aa � m a. ' He i s hurrYing and comi ng ' IN come UN) + IncDlRPl eti ve a. wu l k6 he C .f in i sh b . wu l k6 he C .f i n i sh aa P€r { { --> wa� ko na P€r { { ' He f i ni shed sel l i n g ' IN sel l IlA soori --> wa� ko na sO�d ' He � i n i shed cook i ng ' IN cook 1 . The IncOllPl eti ve AsPect The IncomPl eti ve asPect marker consistentl y mani� est5 a Mi d tone. As a resu l t , one wou l d exPect i t to be consi dered underlYin9ly Lh and thus tri gger Hi gh tone sPread onto Low tone verbs, si nce thi s i s the behaviour � Mid � i nal 1 77 nouns. However . as the examPle i n (33a) shows . such Hi gh tone SPreadi ng Produces an i ncorrect sur�ace �orm. ( 33) a. wu � t� � i --> .wu Ya tUK i ( ' he i s vomi ting ' Rather: wu Ya t�xi i . However . the i ncomPl eti ve tone behaves i n the same manner as a Hi gh � i nal noun by l oweri ng Mi d tone verbs to Low tone in phrase �inal Position . as shown below: b. wu n SDOI"i --> wu U s�i ' He i s cooking ' c . wu sf ai4 SDDri -> wu d . wu ca-� saari --> wu c� soori sf Q1a saari ' He wi l l be cooki ng ' - He had been cooking ' Thi s observati on l eads us to susPect that the i ncomPl etive marker is underl Yi n91 y Hi gh � i nal . I� we ProPose that the underl Yi ng shaPe o� the incomPl eti ve marker i s a combi nati on o� a Low tone and a Hi gh tone. the l i nking o� both tones to one tone bear i ng uni t WDul d create a Mi d tone on the surface. Thi s Hi gh � i nal parti c l e would then tri gger High Del i nking on a fol lowi ng HI verb . as shown below . (34) wu ca-& " �ri II -> wu c a-& saar l II I LA L Ih h II I A LH II Ih I L t h H DEL INKING Thi s di scuss i on . then . shows that there are at l east two tYPes � Mid tones i n Suc i te. the Lh tone tentati vel Y posi ted for nouns and verbs. and the LH tone for the i ncomPl eti ve asPect . The Lh tone can tri gger High tone SPreadi ng whi l e --- ---- 1 78 a LH tone cannot . On the ather hand. the LH i ncompl et i ve can tri gger Hi gh tone Del i nk i ng an a �ol l owing Lh verb in Phrase f i nal Posi t i on . Whi l e the Lh tone can be subject to tonal change. through such rul es as Hi gh Del etion or Hi gh 0.1 i nki ng. the LH tone o� the i ncomPl eti ve never undergoes tonal changes. It does nat even under9a the PrimarY regi ster Hi gh Del in king rule when fal l owed by • Hi gh tone verb . Recal l that H i gh Del i n k i n9 takes Pl ace on a Law-High contaur when it i s fal l owed by another Hi gh tone. The eKamPle below shows that Hi gh Del i n k i n9 wou l d Produce an i ncorract surface farm. (55) wu Ya pir f -� -> *wu Y.\ p(rf -{ " I I A "JH HI LH H H LH Ih h h lh 11 II I I I I Rather : wu Ya P€r i i h h H IGH DELINKING Ther�ore, i t i s necessarY t o b l ack the LH tone o f the i ncompletive from Hi gh tone Del i n ki ng. Since mast Law-High contours found el sewhere in the l anguage do under9a High Del i nk i ng. it seams that the onl Y al ternati ve here is to l ex i cal l Y mark the LH incomPl eti ve to refrain from Hi gh Del i n king. 2. Th. ComPl.ti ve AsPect Mark.. Earl i er . i t was noted that the mast sal i ent �eature o� the compl eti ve asPect was that i t was Law tone. and that i t was often associ ated wi th the vowel �. Si nce i t i s vowel i n i t i al . i t coal esces with the vowel of the Precedi ng ward , be i t a verbal Parti c l e (56a ) , naun or Pronoun ( b ) , or verb (c) . (36) a. wu ca-& Pan � he had come I b . wu � !)an -> wa-& pan ' he came I c· wu-& NO l � & pan -> wa& wu l �-� pan ' h. taok i t and came l 1 79 Li ke Law � i nal nouns. the Law tone a� the comPlet i ve aspect marker al so tri ggers Law tone SPreadi ng onto the � ol l owi ng verb . Thi s i s i l l ustrated i n the examPles below. In (57b > . note that the cOmPl eti ve asPect i s l ocated in two Posi t i ons in the sentence. the � i rst i mmed i atel Y a�ter the subject. and the second as a connecti ve between the two verbs. Both tri gger Law tone sPread onto the � ol l awi ng Hi gh tone. kJrj. lr� " 1 (57) a. wu ca-a L H I h I � L I 1 LOW SPREAD 1,': ' I Ls.,; .. H L I h I I I I LOW SPREAD H I h LS DEL INKING t -> b . wa-a w€rJ -a kir L " He had gone ' -> wu ca-a kad H I h wa-� ��i-� karf �I � L H L H I I I I I h t " He went earl y ' h LS DELINKING The above deri vati ons. however . are not yet quite COmPl ete. When the comPl et i ve marker is �ol l awed by a Law tone. the Law tone an the marker i s i tsal � dal i n ked and i t assumes the tone o� the ward t o wh i ch i t has coal esced . Thi s hapPens regardl ess a� whether the Law SPread has takan place or not . There�are. i � the marker has coal esced to a ward whose sur�ace tone i s Mi d tone (58a and c > . i t becomes Mi d tone, and i t becomes Hi gh tone a�ter a Hi gh � i nal ward (58b . d. and e) . (59) a. wu " He had gone ' ---- ------ --- - -- - - --- - -- - 180 b . wa-� wir f -f k�rf -> waa wirf f k�rf ' He went earl y ' I �'V H�'V a:. H .11 l I I I I h h I I " He aweed ' d . nd. "' w�rf 4 �i ri -> nd� na wir.aa Yi r � 1I-' f H e. ibake � p.l i -> gbaka� p.l i l I " I got u P ear l y ' l " the house i s b i g ' After law tone wards, however , there i s a sl i ght devi at i on f rom the pattern gi ven above. As a connecti ve, i t remains law ton� ( 59a) , but when i n verbal particle Posi t i on , after the subject, i t acqui res a Mi d-law tone (b and c ) . (59) a. waa t�Ki l s��ri -> waa t��4 s��i b . w�� yal a na b! gbddn • He vami tad tao much I " We shou l d h i t them ' we-TA shou l d I N them hi t-IN c . n��-& yal a m� Kal e t��n wd man-TA shoul d __ y6 ' Man shoul d l ove God ' God pl ease hi m sel f-to Thi s Mid-low contour shows uP on l Y when Preceded by a Law t i nal nomi nal subject. It i s entirel y possi b l e that in thi s Posi t i on , the se9ment . 4, i s morPhol ogical l Y compl eK , contai n i ng the comPl etive marker as wel l as PerhaPs some i ndi cat i on at tense, and thus maki ng A mare tonal l Y comPl eK . The reason for this sl i ght compl i cati on , however . remai ns unknown at thi s parti cular stage at research. let us return to the d i scussi on of the compl eti ve law tone del i nk i n9 when fal l owed by a low tone. The i ssue fac i ng us here i s how to formu l ate a rule for ----- - _ . _ ._ .._-- ----- ---- 191 thi s tonal behaviour. Thi s type o� Law tone Del i n k i ng aPPears to be the same tYPe o� contour si mPl i � i cati on rule as Hi gh tone del i nk i n9 i n that both tri 9ger the del i nk i n9 o� the ri 9ht side o� the contour tone when the �01 1 0win9 tone i s the same tone as the r i ght s i de contour tane. Thi s Law tone del i n ki n9 i s compared wi th Hi gh tone del i n k i ng i n the examPles bel ow. (60 ) a. Law Del i n k i n9 waa 9b&r& �l VL I'lL b . High Del i nk i n9 waa w& l � �66 � V L H H I� one made a 9eneral i zed rul e del i nk i n9 the r i ght si de of a contour tone when the �01 1 0win9 tone is the same as the ri 9ht s i de contour tone. then the twa tonal Processes i l l ustrated above in (60) can be col l aPsed i nto one rul e . The on l Y compl i cati on that ari ses i s that Law Del i nk i n9 i nvol ves a contour tone an a l an9 vowel . whi ch in Suci te. rePresents twa tone-beari n9 uni ts. whi l e Hi gh Del i nkfn9 i nvol ves onl Y one TBU. Even wi th thi s Prob l em. however . it wou l d nat be di l � i cu l t to ProPose a si n9l e rul e lor bath Processes. Thi s rul e , label ed Ri 9htside Del i nk i n9. i s stated below: (61 ) RIGHTSIDE DELINKING: Del ete the ri ghtside at a tonal contaur ot ei ther a 1 0n9 vowel , composed at 2 TBU ' s or a si n9l e vowel . when 1) i t is t ol l owed by a surtace tone whi ch i s the same as the r i 9htsi de tone of the contour. or 2) when it is i n sentence t i nal Posi t i on . (v. W) ,,- - - -- - ----- ---- ---- ------ 1 82 Thi s ru l e wou l d work wel l �or the examples 9i ven above. Unfortunatel y. however . thi s rul e does nat appl v in al l cases· In (62a> bel ow. a Mi d-Low tone an a sin9l e TBU does nat undergo RS Del i nk i n9 when fal l owed bv a Low tonee , whi l e i n (b> , a noun with a contour tone on a 1 0n9 vowel i s does not submi t to the rul e . ( 62) a. wa� j� 9b�r� A VL i'lL b . wa� �Vaa nVa II I HM t he met a son ' t he saw a f i sh ' M Both � these counter examPl es i nvol ve nouns, and nei ther i nvol ve a Low-hi 9h tone. Therefore. i t aPPears that thi s RS Del i nk i n9 has a fai r l v l i mi ted appl i cati on - del i nk i n9 the Hi9h of a Low-Hi 9h contour (th i s has been attested on nouns as wel l as verbs> . and the del inkin9 of a ,gmpl.ti ve Low tone when i t i s fol l owed b v a Low tone. These Probl ems brin9 i nto doubt the feasibi l i ty of moti vati n9 the RS Del i nkin9 rul e. However . despi te i ts controversial posi t i on . thi s rul e shal l b e used �or the PurPose of convenience . It was menti oned i n the RS Del i n k i n9 rul e that the tone o� the f ol l owi n9 TBU must be the same on the surfaci as the Right si de tone o� the contour . Thi s means that i f a comp l et i ve Low tone i s fol l owed bv a Low-Hi9h contour wh i ch has si mpl i f i ed to Mi d tone. then no RS Del inkin9 takes Place. Thi s can be seen i n the exampl e bel ow. (63) wu � l � la -> wa� l � l a -> wa� l � l a I 1'1 l I H I i'l l 1 L .. H I 1'1 LOW SPREAD RS DELINKING N/A t Di d he take? ' 183 However , i� RS Del i nk i n9 a� the Law-Hi9h contour an the verb takes place, as .hawn i n (64) . then the comPl eti ve Law tone, wh i ch i mmed i atel Y Precedes thi s Law-Hi9h contour, al so undergoes RS Del i n ki ng. (64) a. WA� II l� ry�� . -> waa l� ?d� I H ML V H 10 V ML H H It i s cl ear from the above examPl e that the RS Del i n k i ng of the Law-Hi 9h contour must take Pl ace before the Del i n k i n9 of the comPl eti ve Law tone. Th i s maY l ead · ane t o raise the questi on that these twa tone processes might be better stated i n twa di ��erent rules. and that the one. Hi9h tone Del i n k i ng. be ordered to appl Y be-Fare the ather, Law tane del i nk i n9. However . anether waY to approach the Probl em i s te suggest that RS Del i n k i n9 appl i es �ram right to l eft acress the senience. Thus. i � H i gh iane i s del i nked from the Law-High centaur . a sur�ace Lew tane resul ts. creat in9 a sui tab l e envi ronment fer the del i nkin9 ef the compl et i ve Lew tone, situated te the l e�t e� the Law-Hi 9h centaur . c. TID••• Mpdal , and ethIC pvHc:ln 1. Th. NA Pvticl. Th. Q& Parti cl e i s a Past tense marker , but i t d i ffers i n use from the simple recent Past :i. di scussed abeve. Carl son 9i ves an aPt descr i ption of f or SilPyka, whi ch seems to aPpl y equal l y wel l te SilcHa: liThe funct i on of oa Eli i n si tuat i n9 events relat i ve to each ether maY have somet h i ng to do with i ts devel oPment i nto a Past tense marker . As a TIA marker. its CDmIIICI1l1St use i s te set the tense in the f i rst clause of a narrat i ve . II ( 1985. p. 12) H e 90es an t e / saY that Di i s nat used i n subsequent cl auses. Th i s i s al so true far Silc i ta . In " addition , he saYs "Another -Funct i on a-F ail i s emphat i c . Ii; i s used te lissert ----- -- - - --_ . ---_ 1 84 stron9l Y that somethi n9 did take Pl ace ...hen doubt has been exPressed . " Al though a& �aY be accomPani ed by a comPl eti ve La... tone marker ( 65a) . i t i s often occurs al one ...i thout anY other verbal part i cle. When i t occurs al one . aa i s consistentl y f ol l o...ed by verbs i n the COmPl eti ve asPect (b) . (65 ) a. ...u ni� k�{ ' he went ' ' he ....nt ' Its tonal behavi or i s i denti cal to the tonal behavior of Hi9h tone verbs. When preceded by a mi d or h i 9h tone subject. aa is High ton., as sho...n i n the examPles b.l o... . (66) a. ' He has 9cn. ' n4 �rf wu b . mo � k�r( ' You have 90n. ' c · Peru "" kid ' A daughter has 9one ' d . f al �� n� kar( ' a rock has 90ne ' When Preceded by a Low f i nal subject , the Low tone sPreads onto � creat i n 9 a Lo...-Hi 9h contour, as shown bel owo Th i s contour tone then simpl i f i es to Mi d tone if i t is fol l o....d by a non h i gh tone verb (67a ) . However , if fol l owed by a Hi 9h tone verb, the envi ronment for RS Del i n k i n9 i s created resu l tin9 i n a surface Low ton. -For aa <b > . (67) a. n�.l nEt 9b�;.l -> nA� n� 9b.lr.l -> n�� nil I L I I H 1,, ' 1 L L H I 'V1 L H L n� L H I k�rf -> n�� I I H na k'r{ -> n.l� n� kQrf 1 ., 1 r H I H LOW-SPREAD I NG . _-_. - -- .- . --- - - -'Jl. --- ---- ' A man a9reed ' L INKING L-SPREADING b. n�� 9bclr.l \M V L H ' A man has 90ne ' H LINKING & RS DEL INKING -- -- ------- - -- - 185 � Since i s Hi gh tone, it tri ggers the l ower i ng a� Mi d tone verbs in �i nal Position through subregi ster Hi gh Del i n k i ng ( c ) . even when i t has been subject to Law tone SPread Cd) . c. wu nS Pan -> wu n! PMt I H I ' he came ' I I H L L � t � � HIGH DELINKING d . n�� � Pan -> n�� n� Pan -> n�� na p� 1.. ... 1 L H \1/1 I I L L H L 1 h h I I I I I h h ' A man has came ' I f LINKING , HIGH DELINK I NG L-SPREADIN6 2. Th. NI l Particl. The a11 part i cl e si gnal s Past tense. The Prec i se semant i c �uncti an a� thi s parti cl e i s net known at the Present t i me. When used wi thout ather verbal Particles. i t i s al waYs accomPani ed by a comPlet i ve verb . (68 ) wu nU Pan ' he came ' � i s comPosed ai a High pl us Law tone. It i s bel i eved that thi s Law tene i nd i cates the Presence a� a Law tone COmPl eti ve marker . Thi s Law tone tri ggers Law tone sPread ante High tone verbs, and then . i s i tsel f subject to RS Del i nk i n9, as seen bel ow. (69) wu n { � �r{ -> wu n { l k�r{ -> I M / LL"'�H1 H L-SPREADI NG ----- --- - -- ----- I M I \P"1 H L H LS DELI NKING NU I M n { { k�r{ -> wu n { { k&r{ '''*JL H I ' He went ' H RS DELINKING --- ---_. _-- ------- ---_ . 186 When nit i s Preceded by a LON tone. i t is also subject to LaN-spreadi ng. The resul ti ng LH contour on D11 i s simPl i � i ed to Mi d tone� n11 n { � Pan -> n11 n i � Pan -> n11 ni � Pan (70) I / II L L H L L L I h 1 h 1 I I I J ... I � A man came ' I 1 \ H L L I I I h I h LOW SPREADING 3. Th. Futur. T.n•• The �uture ·tense mOrPheme i s characteri zed by a H tone segment . at (a ) . Because o� the � i nal nasal o� the morpheme, the consonant o� anY verbal el ement i mmediatel Y � ol l owing the �utur e marker i s nasal i zed : • I Ni 1 1 l ook ' . However . i � an abject i ntervenes. the nasal di saPPears� nd� 5£ x� we' � I wi l l l ook at You ' . When a �uture tense morpneme i s marked �or comPl eti ve asPect . a LON tone i s l i nked t o the parti c l e . creati ng a HL contour . (71 ) mo st mPan . � � YOU wi l l come ' Thi s � i nal L tri ggers Law tone sPread onto Hi gh tone verbs and i s subsequent l Y subject to RS Del i nk i ng . (72) NU st n kar{ I M � HL I H ->NU st n k1rf -> I M �':" 1 HL H L-SPREADING NU I M st nk�rf -> NU 5£ nk�f I\.P1 H L H LS DELINKING I M IV H L 1 • a man Ni 1 1 90 ' H RS DELINKING When 11 i s i t••l f subject to Law SPread by a Precedi ng Law tone. the resu l t i ng LHL tonal contour is si mPl i � ied to Mid-Law. as seen bel oN. _- - - _ ._ . ----- ----- ---- ------ 187 (73) n�� st mPan --> !� L n�� st mPan -> n�� s� mPan I t / 1\ HL L L I ' a man wi l l come ' I h h LOW SPREADING When the � uture tense marker i s used with the i ncompl et i ve asPect , i t i s �ol l owed by the i ncamPl at i ve asPect parti cl e , t4. Thi s �arti c l a i s apparent in i ncomPletive i mPerati ve �orms o� the verb : ' be 90i ng l ' ta wi ! ' be l ook i ng ! 1 When �ol l awed by the �uture tense Parti cle, the t i s nasal i zed , produc i ng the Phonetic lorm, [ ddal o� Inta/ : wu s{ nta sa ' he wi l l be goi ng " . The particle, iA is Mi d tone an the sur�ace and is never suscept i b l e to anY tonal changes. However , as with other i ncomPl eti ve Parti cl es. tA trl �. the High Oel i nkin9 o� Lh verbs and there�ore has been posi ted as bei ng under l Yi n9ly Low-H i gh . The deri vati on bel ow shows how this underlYi ng Low-High tone tri9gers H i gh Del inki ng on the Lh ver b . (7�) wu s{ nta soari ->wu s{ nta sOor� -> wu s{ nta soor� ' he wi l l be cooking 1 1 H A LH I L I I H " LH h LINKING � L f h H IGH DELINKING The �uture tense Parti cl e , Ii, i n conjunction wi th the i ncompl eti ve aSPect particle, conti nues to be sub ject to Low SPread when Preceded by a Low � i nal subject. as shown below. In thi s case. the resul t i ng Low-Hi gh contour i s si mPl i f i ed to Mi d tone. ----�- - - --- ----. ----- ---- ------ --- - -- - - - - -- 1 88 " a man wi l l be comi ng ' L SPREAD The i ncomPlet i ve Progressive part i c l e . � i s also marked far the i ncom­ Pl et i ve LH tone. The l atter i s attached to the f i nal tone-bearing uni t of creati ng a HM contour � an �. the f i nal Teu: " He i s coming ' (77) wu n H LH H � � When � i s subject to L-sPraadi n9 . the fol l owing take. pl ace l (78) n�� n"?a ma. -> nU n.l� / I� \ L L.;,;'/� H LH H L ma I -> n�.l .�? H LH H L L-SPREADING s. mi. -> nU n�?i HI:. H LS DELINKING Th. CA Particl. The mean i ng of � seems to i nd i cate acti on further i n the Past than the tocal Poi nt at the narrati ve. I t i s trequentl y combi ned with the � past tense marker . The parti c l e do ' . The verb QR � maY have been deri ved from the verb and the Part i c l e � � meaning ' to are both Mi d tone. ca is nat subject to anY tonal rul es. I t i s not known whether Lh . si nce Qi. � i s under l Yi ngl y Law-High tone or i s al wavs � ol l owed by an i ncomPlet i ve or a comPI.tive mar ker . whose tonal behavi our do not reveal anY cl ues as to the possi b l e underlYi ng When fol l owed by a camPl eti ve Law tone. the resu l t i ng comPlex part i c l e. �. --- - - - - - - - -- --- ---- - ---- ---- ------ 199 �ri ggers Low �one SPreadi ng on�o Hi gh �one verbs. as shawn i n (79a ) . When fal lowed by �he LH i ncomPlet i ve mar ker . High Del i nk i ng �akes pl ace an anY fal l owi ng Mi d �ane verb (90b ) . ( 79) ComPleti ve a. wu caA so?o ' He had cooked ' b . wu ca-A kar{ -> wu caa k1r� ' He had gone ' (L SPREADIN6 ) (90) IncamPl e� i ve a. wu ' He had been going ' caa H b . wu caa saari -> wu caa s��r� ' � had been cook i ng ' (H DEL INKING ) 6. Th. Hel tu.l a? When �he Habi �ual marker i s i n a cOmPle�i ve Phrase i t i s no� marked wi �h �he comPlet i ve Law �one. In fact . in mast envi ronments, �he d i f ference between �he ComPl et i ve and i ncomPle� i ve farms i s no� Percepti b l e. The examPles given below shaw tha� in bath comPl e� i ve and i ncomplet i ve Phrase. � �ri ggers Hi ih �one Del inki ng an bath farms of the Mi d �one verb. �hus l ower i ng a Mi d �one verb to Low tone. (91 ) COMPLETIVE wu m1?a sj?j lso?ol ' He cooked habi tual l y ' INCOMPLETIVE wu mA?a s��r� Isoori l ' He is cooking habitual lY ' However . when fal l owed by a Hi gh tone verb . the compl etive and i ncomPl eti ve Phrases exh i b i t di fferent behavi our . In the comPletive Phrase. � is com­ Pl etel Y Law tane (92a) whi l e i n (b ) i t remains Low-Mi d . (92) a . COMPLETI VE b . INCOMPLETIVE wu m�?� kar� wu m�?a s' ' He usual l Y went ' ' He was usual l Y goi ng ' ----- ---- --- ------ 1 90 I� one posi ted the underl yi ng tone a� mAZa to be underlYi ngl y Law-Hi gh , wi th an i nternal Law tone SPreadi ng rul e sPreadi ng the Law of the f i rst TBU onto Hi gh tone of the second TaU, the � i nal sYl l abl e of ma2A can then be l i nked to bath a Law and a Hi gh t ane, creati ng a f i nal Mi d tone, as shawn i n the f i rst Part a� the der i vati on i n (83) . Once the Law SPreadi ng takes pl ace, a Law-Hi gh contour �al l awed by a High tone verb is the envi ronment needed for RS Del i n k i ng. ' He usual l Y went ' LINKING & RS DElINKING LOW SPREAD However , when the LH i ncamPleti ve tone is aSSOCi ated to the U nal TBU of m42a, RS o.l i n ki n9 i s bl acked because the sequence of twa Hi gh tones requi red for the appl i cati on of the RS Del i nk i n9 rul e an Law-hi 9h contours, has been broken uP by the LH marker. Thi s is i l lustrated below. Subj ect tone does nat seem to a�fect the tane a� m424. One mi ght exPect Hi gh tone sPreadi ng to take Pl ace when the habitual marker i s Preceded by a Mid � i nal subj ect. Thi s , i n e�f ect , does nat take pl ace. (85) mu m�'i'a I I m i. �,� I H I I " I L L H LH h I h lh LOW SPREAD h *mu ma'i'i 1 L I m� � , ' Yau are usual l Y comi ng ' L H LH H I h <D h l h h L,"1 f " HIGH SPREAD The ComPl et i ve Law tone i s also i mPervious to Hi gh tone sPread . as shawn bel ow. ------- ----- ----- ------ ---_ . 19 1 (86) IOU � Pan --> moa Pan . *1000 PAn ' YOU came ' PerhAPs i t i s pecul i ar to verbal particles that they are not suscepti bl e to High tone SPread . The reason �or such an excePti on i s un known At the Present t i me. Th, Mgdals Modal s never occur al one i n a verbal Particle stri ng. A TA Particle al ways Precedes a modal . 7. The l1adal SA The l'Iodal , D is deri ved �rom the Hi gh tone verb . sa. meani ng ' to 90 ' . In a complet i ve Phrase. � i s unmarked �or asPect . Below Are two charts �aci l i tati n g a quick overview o� the e� f ect of d i �ferent combi nations o� tone wi thi n the comPl eti ve sentence on the modal . In (87) . a Hi gh tone verb is used whi l e the examPles in (SS) f eature a Mi d tone verb. A bri e� gl ance At the two charts wi l l reveal that sa al ternates between a Law and Hi gh tone when a High tone verb fal lows whi l e i t al ternates between Mi d and High tone before a Mi d tone verb. I t is al so seen in ( SS) that a Mid tone verb is al ways Law tone after i n d i cati ng that i n al l sett i ngs &i has �. an underlYi ng Hi gh f i nal ton•• EACh of the charts has th�ee raws and two columns. Each raw �eatures a d i ��erent TA marker . whi l e each col ulDn f eatures a d i Herent subject. Nc'Cr4 that the tone of the medal vari es accordi ng to the underlYi ng tone of the preceding verbAl particle. The Presence of the two columns shows that whi l e the tone of the TA marker maY varY accordi ng to the tene o� subject. the tone of the medal whi ch f al lows the TA marker. remai ns unaffected by the tene of the subject. --- _ .- ._-- ---_ . _ - -- 1 92 (87) TA Marker ' he _ aa sal l ' ' a man aa .el l ' _ i A. waa a4 Pirf b . n�a � Pir€ s1 < n ) c . wu s f � ptrt na e. wu na d pfr€ d . n�� si � Pfrf (88) TA Marker ' he _ sa cook ' c . wu 51. � ' sel l ' ' a man went to sel l ' t . n�� n� � Pfrf ' a man s� cook ' _ p€rt ' a man i s 9ai n9 to sel l ' ' A man Nent to sel l ' 5:21::1 ' cook ' .�� t . n�a e. ·wu n� do .�� n� Ii. s�� ' a man went to cook ' I t one Posi ted � a. bei ng underl Yi ngly Hi gh tone. the tonal al ternati ons gi ven above can be easi l y explained . The � i rst two TA markers i n each chart . bei ng LON f i nal . woul d tri gger LaN tone SPreadi ng onto ai. which . once havi ng acqui red the LaN-High contaur . undergoes RS Del in king i � �ol low.d by the High tone verb . but i s simPl i f i ed to Mid tane i t �ol l owed by a Mid tone verb. The deri vati ons ot (87a) and ( 88a) are gi ven bel ow to i l l ustrate how the tone o� bath the verb and the Precedi ng verbal parti c l e atfect the surface tane at (87a) wa� si ptr€ -> wa� s� Pfrt I( ML I H I 1 H �"1 ML H I H LOW SPREAD - > waa s� P€r€ 1,t/1 '� ML H H LINKING & RS DELINKING (88a) wa� s-&. so10 -> Na� � 5010 -> wa� H 5010 -> wa� sa 5010 II I H I I ML 1 h I �"1 I I ML H L ML H h 1 h h I h 1 I I I LOW SPREAD ---- ._--- - ._.- 1 1/1 L . _ . - - -.- I :f h HIGH DELINKING .. ... --- I I L --- -- RS DELINKING N/A �. 193 For the incomPl et i ve ai sa, the LH i ncompl eti ve suff i x , �, is add.d . Si nce i t tao is Hi gh fi nal , the i ncomPlet i ve farm of saa also tri ggers Hi gh Del i nk i n9 ai Mi d tone verbs. The eXamPles bel ow. shaw the i ncampl eti ve saa wi th (89a) a Hi 9h tone verb , ( b ) a Mi d tone verb, and ( c ) a Law tone verb . (89) a. ma na s1-& PirH " YOU Past 9oi n9 a9reei ng ' b . ma .n s-{-& � � lsaari I " YOU Past 9oin9 cooki ng 1 c . mo n{ sa-a gb4r� l " YOU Past 90ing agreeing 1 When the i ncampl.ti ve saa i s Preceded by a Law tone, i t i s natural l Y subject to Law ton. SPread . Hewevl!I"" , i t no l onger undergoes RS Del inkin9 when fal l owed by a Hi9h tone verb , because the i nterveni ng i ncamPl ati va suf f i x destroYs the environment far RS Del i n k i ng. The resu l t i ng surface tone far the i ncamPl et i ve saa i s M i d tone. Thi s can be seen i n the deri vati on below. (90) wa� .a-a Pirf f I L I " H LH I H -> wa� sa-a Pirf f � "1 1\ \// L H LH " he was (9oi n9) sel l ing ' li LOW SPREAD & LINKING 8. The BA l1ad.l The Modal . bA i s probabl Y deri ved from the Mi d tane verb, ean. It i s consistentl Y Mi d tone an the surface, though underl Yi ngl Y, i t i s d ear l Y Hi gh f i nal . si nce it triggers Hi gh Del i nk i ng on Mid tone vl!I""b s. Thus, ba appears to have the same underlYi ng tone as the i ncampl eti ve aspect . LH. The fact that ba aPPears to have deri ved from a Mid tone verb whi ch is anal Yzed as bei n g under l Yi ngly Lh ( that i s a rai sed Low tone> maY i ndi cate that there is same h i stor i cal connection between a Law-Hi gh contour tone and a rai sed Law tone. ---- � -- - �- �- ----- - - -- ---- ------ 194 When used i n a comPlet i ve Phrase. ba ramains unmarked for asPact. When the i ncomPl et i ve suff i x . �. i . addad . i t remains Mi d tone an the surface but i s undarlYi ngly Hi gh f i nal . Selow i s a chart showing b4 i n combi nation with vari ous verbs and verbal parti c l es . both in the comPl eti ve and the i ncomPl eti ve form. In al l cases. ba remai ns Mi d tone. Th. sentences i n (91 b ) and (g ) shaw that both the comPleti ve and i ncomPl .ti ve farms tri gger Hi gh tone Del i nk i ng of Mi d tone verbs. IncomPlet:i ve ( 91 ) ComPl .ti ve a. wu na ba Pir€ f . w u n a ba-a P€r£ ! " he ____ sel l ' b. wu n� ba s�?� 150701 g . wu ni. ba-a �d lsaar i l c he __ cook ' c. wu ne. ba gb"� h . wu n2. ba-a gb�U c he __ agree ' d . wa� ba gbM� i . wa� ba-a gbuU C he __ agree ' e. wa� ba P€r€ j . wa� ba-a p€rH " he ___ sel l ' 9. Th. Pr.-Tan•• PArticla Pre-Tense Part i c l es aPPear at the beginning of the Verbal particle stri ng. I have i denti f i ed two such parti cl es : Iin2R means c a9ai n ' i ndi cating that the act i on of the verb Phrase has bean repeated a second time. The partic l e � seems to have a c l oser connection wi th the subject. i n that i t means " al so ' . The subject. i n addi t i on to another Previ ousl y menti oned subject . commi t the same act . It aPPears that thi s i s the onl Y Posi t i on i n the sentence where ei ther Part i c l e i s al l owed . Both � and � maY be combi ned wi th numerous other tense and modal particles both i n the i ncomPl eti ve and comPlet i ve aspect . They are nat marked for asPect themsel ves. Si nce they both end in an underlYing Hi gh tone they behave as tYPical Hi gh f i nal wards in that they tri gger the Hi gh tone Del i nking ---- ----- - --- - - - --- ----- -- ---- 195 o� Mi d tone verbs. Both are susceptible to rul es tri ggered by the tone o� the subject. � Let us take a l ook at As eXPected. when Preceded by a Mid tone. � remains High · Csee (92a > , but when Preceded by a Low tane. Low tone Spread i ng goes i nto eHect and a LH contour resu lts (92b and c ) . Because mii" hAS two TBU ' s. the LH contour does not si mPl i�Y to Mi d tone. ( 92) a. ' YOU al so are cooki n g ' maa Ya s�i me ' A man i s also cook i ng ' ' a woman i s also cooki ng ' The parti c l e i8n2a ramains Low-Mi d on the sur�ace ( 93a and c ) unt i l i t i s Preceded by a Mi d tone el ement C d > . When thi s i s the CAse, the parti c l e i s MidHigh in tane. Thi s behaviour can be eXPl ai ned i� iaD2& is posi ted as havi ng An under l Yi ng Low-Hi gh tane with an internal Low tone SPreadi ng ru l e. Thus, the Low tone woul d SPread onto the second TBU o� the Particl e creAti ng a Low-Mi d tone. Thi s i s i l l ustrated i n Ca> below. When i8n2R i s �ol lowed by a Hi gh tone . . verbal part i c l e . as in ( b > , it undergoes RS del i nki ng, Producing a sur�ace Low tane on the � inal TBU 0+ the Parti cl e. Ca man is coming agai n ' ( 93) a. nU s�n?e Ya ma I L � ; 1 !\ I L H LH H LOW SPREAD j b . naa s�n?e n ' karf -> naa s�?� na karf ., L �"1 L H H LOW SPREAD ._-- - _ .- ---.---- I H VL LV1H HI 'JH LINKING & RS DELINKING C a man has 90ne again ' 196 "a rock i s comi ng agai n ' When Preceded b y a Mid tane however , i t aPPears that Hi gh tane sPreadi ng takes Pl ace be�ore Low tone sPreadi ng has a chance to aPpl y , as the deri vati on below shows. thus Produci ng a Mi d-High tone. d. mo un?e Ya moS: -> I I L L I A H LH I H I sen?' Ya ma 'h 'h Ih me L � 1 I II I h (!) h l h h L V ... I I A I H LH H I II 1 " YOU are comi ng here ' h LOW SPREADING N/A H SPREADI NG 11. Th. Ndati 'le Particl. The ne9at i ve part i c l e has been deal t wi th l ast. but this is not because i t i s parti cu l ar l Y compl i cated. I t i s a simPl e Low tone which tri ggers the Low SPread rul e i � the condi ti ons have been met . and oPti onal l y vulnerab l e to Hi gh tone SPread . Hi gh SPreadi ng onto the Negati ve parti c l e seems to be a matter o� �ree vari at i on among sPeakers � S�cit'. � comes a�ter the Pre-tense Parti c l es menti oned above. but be�ore anY other particles. The Negative sentence i s �ormed by the i nsertion of � i n tha verbal Particle stri ng and b y Pl acing the morPheme mfn at the end � the sentence. The onlY other element that is al l owed to come a�ter mtn i s the Yes-no questi on marker. lao Here. then . are some assertive sentences with their negati ve counterparts : ---- ( 94) Asserti ve Negati ve a. mu m-a3 n� p� b . mo ria � na Pan m!n. " You also d i d not come ' c . mu maa mA?a s�?� d . mu ma-a � m.l?a S::l?{ " YOU al so di d not Hab . cook e. wu n � � kAd � . wi t ni i k.lr� " he did not -- _ . _ _ . --- ------- mtn 90 ' 1 97 In ( 94) � . � coalesces wi �h �he 5Ub jec�: wu Yi -> wi i . Wi�h the recent Past particle, i and the presen� i n compl eti ve parti c l e , �, the tonal presence o� � i s l ess Percepti b l e : ( 95) Asserti ve Negati ve a. wa� Pan b . wu � Pan mto. ' he d i d not come ' c . wu Ya Sddd d. wu � sccri aWl ' he i s not cooking ' Below i s an examPle � �ree var i at ion concerni ng the use o� the Hi gh sPreadi ng rul e l ( 96) a. mu n -a P�n ' You have come ' c . mui na Pan min b . mui na Pan min I ��1 L L I H h I 1\ I L LL H I V I \ H L I ' You have nat come ' H h h h I HIGH SPREAD LOW SPREAD In ( 96b ) , Law SPreadi ng occurred because Hi gh SPread ing was nat i n e�tect. I n ( c ) , however , Hi gh SPreadi ng has Preceded Low SPreadi ng thus blocking the l atter tram appl Yi ng. The sentence t i nal negati ve morpheme, mto!,. oPerates tonal lY. much l i ke the det i n i te su�t i x . I t is Hi gh tone atter Mid tone verbs and Mid atter Hi gh and Law verbs: ( 97) wu yi Pan min ' he d i d not come ' ' he did not gO ' wu Ya k�d mE n Ya gb�r� NY wu Ya 9b�ri-{ �1 L H - .. - - . - - mE n I ' he di d not agree ' -> wu Ya gb�i-i L4 L H H min ' he i s not meeting ' I H RS DEL INKING LOW SPREAD ---_ . __.... -> wu Ya 9b�t mEn . --- --- ----... ---_ . . . .- ._. ---_ . 1 98 I n the l atter examPl e , the i ncomPleti ve form of 9bara Possesses a fi nal underl Yi ng Hi gh tone i ncomP l et i ve suf f i x whi ch i s subject to Low tone SPread and f i nal l y to Ri ght Si de Del i n ki ng. IV. CONCLUSION The di scussion i n Chapter 4 h�s l ed to the reformu l at i on of several rul es as wel l �s the i ntroduction of new ones . Low tone SPread requ ired onl Y mi nor revi sion ( 1 1 ) , whi l e the anal Ysis of Low tone SPread onto multiPl e TBU verbs l ed to the the reformul �tien of the Associ �tion Cenventi ons ( 1 6) al l owing, i n the i n i ti al l i nki ng of tones t o TBU ' s. the l i n k i ng of a tene to ne mere than ene TBU. The rul e of Hi 9h Del i n k i n9 ( 37 - Ch . 2 ) w�s renamed RS (Ri9htsi da) Del i n k i n9 ( 6 1 ) to i nc l ude ether contour .tones whese ri9htsi de tone del i n ks in a spec i f i c envi ronment. In addi ti on . the rul e of M id tone Del inkin9 (53) i ntro­ duced bri ef l y i n ChaPter 3. and the del i nk i ng of the Low tene i n a Low-Hi 9h centour , were found te take Pl ace i n the same 9eneral envi ronment. Thus. the rule of LS ( L.eftsidg) Del i nk i n9 (40) �s formul ated to i nclude beth Processes. The new rul es i ntreduced were Subregister High tone Del i nkin9 (44 ) , Subre9� i ster Hi 9h tone SPread ( 35) , Subregi ster Hi 9h tone SPread to TBU (36) , and Linkin9 ( 39) . The f i rst three empl oyed the use o� the subre9i ster tier te formul ate rules dePi ctin9 resPecti vel y 1 ) the l ower i n9 of Mid tene to Low tene i n Phrase f i nal envi ronment , 2) the rai si n9 of Low tone to Mid tone after Mi d f i nal nouns. and 3 ) the creat i on of a f i nal Hi9h tone on Low verbs which had been rai sed te Mi d tane, resPecti vel y . The rul e of L i n k i n9 invol ved a l i nki n9 Procedure usual lY assumed to be Part of the Assoc i at i en Conventi ens, and that i s . the l i n ki ng of l ef tover TBU ' s of � merpheme to i ts tenes after the aPpl i ca­ ti on of other rul es such as Hi 9h tane SPread . ---. - '- ' --_ . _ -- ------- ---_ 199 These rules . whi ch were formul ated i n the di scussion on the tonal behaviour of the verb were al so tested for use on verbal Parti cl es. Wh i l e i t was observed that verbal par t i c l es cou l d tri gger Low tone SPread as wel l as be subject to the same rule. they tend to nei ther undergo nor tri gger High tone SPread onto Low tone verbs. Si nce they are never found in Phrase f i nal Posi t i on . they were al se never subject to subre9i ster Hi gh tone Del i n k i ng. Si nce the Mi d tone verb behaved sl i ghtl y differentl y from the Mid tone verbal Part i c l e , i t was PrOPosad that these two Possessed d i fferent under l Ying tones. The Mi d tone verbal Particle, whi ch tri ggered Hi gh tone Del i nking was posited as being under l Yi ng l y Low-High, wh i l e the Mid tone verb was consi dered to be a rai sed Low tone, Lh . Al though not al l questi ons cou l d be resol ved i n these di scussi ons. i t wi l l be seen i n the f i nal chapters of this thesis, that the rules ProPosed here can be used el sewhere i n the l anguage. The fol l owing i s a summarY of the rul es i n order of thei r aPpl i cati on . ( 16 ) Associ at i on Convent i ons ( 1 1 ) Low SPread (35) Hi gh SPread (36) High SPread to TBU (39) Li n k i ng (40 ) LS Del i n k i ng (44) High Del i nk i ng (61 ) RS Del i n k i ng ._--- _ ._ _ ._ --- _._---- ----- -- -- ------ 200 NOTES 1 . One maY su9ges�, as a general i z i ng fea�ure, �o i nvoke �he Obl i ba�arY Con�our Pri nc i pl e, whi ch simPl i f i es a ser i es of l i ke �ones i nto one and then to Posit a rul e such as �he one bel ow, i . e. , del i n ki n9 a High tone whi ch i s l i n ked �o �he same TBU a Low �one and is al so l i nked �o �he fol lowin9 TBU. X X IV L H It wi l l be seen i n ChaP�er 5 , however , �hat OCP can aPply on lY i n a verY res�ri c�ed envi ronmen� . AI �hough more research maY need �o be done, i t was not deemed feasi b l e �o invoke �he OCP i n thi s situat i on . 2. Spreadi ng of tha Low tone t o al l TBU ' s of a High tone verb i s actual l Y wha� haPPens i n SuPYi re (Carlson , 1983, 1 985) . Low tone spreads onto High tone verbs and del i nks Hi9h �one from al l TBU ' s. As a resu l t , �he Hi gh �one word , N{ { , acquires a surface Low tone 'when fol l owed by a Low final noun , as shown below. baaga wi { V I Mw L V H -> baaga wH V V" Mw L " i t ' s a house ' H 3. The use of the doubl e ti ered aPProach for anal Ysi s i n SuPYi re i s not used by Ca1i son , hi mseH , though he does attempt a more tradi't ional autosegmental aPProach . 4. KaYe e t al ( 1985) i n�roduce the n ot i on � fusi ng two vowel s l ocated on two separate ti ers. 5 . Thi s s�atemen� i s debatabl e. Data wi l l have to be scrutinized more cl osel Y to ascertai n whether RS Del i n k i ng does take Pl ace on Mi d-Low nouns or not . 6 . A brief note should be made of the phonol ogical behavi our of the fi nal negat i ve marker . When a verb ends i n a Hi gh vowel (no� tone ) , i nc l ud i ng a di phthon9i zed verb ending i n a High vowel , mtn remai ns intac� (see 1 . ) . However , i f �he verb ends i n a non-Hi gh vowel , �he m i s del e�ed and the remai n i ng vowel coalesces to the vowel of the verb ( 2 ) . Al most al l i ncomPl eti ve verbs end i n i or �, and therefore d o no� tri gger m del et i on . Conversel y, mos� comPl eti ve verbs end i n a non hi9h vowel , causi n9 assi mi lat i on of the ne9a�i ve morPheme �o �he verb . 1. Hi9h f i nal " he is not mee�in9 ' a. wu Ya 9b�d min (from d i phthon9 " paon ' ) " he d i d not come ' b . wu Ya Pan min 2. non High f i nal a. wu Ya gb�a mEn -> wu Ya gb�E b . wu Ya w�e mE n -> wu Ya w�ee ---- " he d i d not a9ree ' " he d i d not l ook ' -- - -_. - --- --- ----- --- 201 CHAPTER S - TONE IN THE SUCITE NOUN PHRASE I . INTRODUCTION UP u nt i l now, we have l ooked at the tone and morphol ogy of the noun <Chapter 3 ) , and have seen how the f i nal tone of the noun al t ers the tone of the verb and verbal part i c l e < Chapter 4 ) . It has been observed that Hi gh f i nal nouns tri gger the l ower i n g of Mid tone verbs to Law tone < subregi ster Hi gh Del i n k­ i ng ) , whi l e Mi d f i nal nouns rai se Law tone verbs to Mid tone <subregi ster Hi gh tone Spread i ng ) , and f i nal l y , that Law f i nal nouns tr i 9ger the sPread i ng of thei r Law tone onto Hi gh tone verbs < Law tone SPread i n9 ) . However , there has been no d i scussi on so far about whether and how nomi nal tone can b e affected by ather const i tuents in the sentence, ex cept in Chapter 1 , where it was noted that nomi nal el ements are nat sensi t i ve to the Preced i ng verbal el ements. Thi s chapter focuses an how the tone of nom i nal el ements w i t h i n the noun phrase af fect each ather . I n addi t i on , i t was noted i n Chapter 3 that the under l Y i ng tone of nouns cou l d nat be adequatel Y exami ned out of contex t of t h e noun Phrase . Therefore, the d i scussi on o f the tonal i nteract i ons w i t h i n the noun Phrase wi l l al so i nc l ude an i n dePth studY of under l Yi ng nomi nal tone. Fi nal l y , the di scussi on of the data in thi s chapter wi l l l ead to a more comprehensi ve exam i n at i on of the orderi n9 of tonal r u l es in Suc i te . The d i scussi on wi l l Proceed wi th a n i ntroductorY secti on , menti on i ng that the seman t i c cohesi veness of the noun Phrase aff ects tonal Patterns and 9i vi ng a general overvi ew of the d i f f erent tonal shaPes of the noun . The next sec t i on w i l l d i scuss the tonal Patterns that occur on a re9ul ar basi s across word boundar i es , as wel l as across morPheme boundar i es wi t h i n the comP l ex noun . Th i s sect i on wi l l i nc l ude a thorough di scuss i on of the underl Ying tonal nature of ---- -- - -- -- -- 202 the various type. a� Mid tone i n the lan�uage. a. wel l the dilemmJ of ordering the rul .. that are PraPa.ed. The anal Ysi s PraPO.ed wi l l l ead to a re-exami na­ tion of the underlYi n� tone of the Definite noun . Final l y, the tonal b.haviaur restricted to comp I .x naun structure wi l l be discussed . Since these tonal al ternati ons occur at the l ex i cal level rather than at a Phrasal l evel , a major issue to be discussad here i s what theoretical framework for rul e ordering i s suff iCiently adequate for the proposed tonal rul es i n Sucit• • A. The Effect pf New" !hr'M structure M TM.l BIIh.yi quc The studY of tone i n the naun Phra.e i . r ..tri cted to the studY of tonal interaction. within a strin� of nominal , Pronomi nal , and nominal i zed al ements. Just a. the entire the.i. i. restricted to the studY of tone in si mPl e sentence structure. l i kewise comPlex noun phrase structure is not wi thin the scoPe of thi s chaPter. A descri ption of simple noun Phra.. . tructure maY be f ound i n ChaPter 1 . A string of nominal elements and their suff i x .. oPerate a. a tonal uni t i n Suci te. Thi s means that the various consti tuent. of the Phra.e can affect each other tonal ly. Pul l eYbl ank ( 1983) and others have observed that Phonol ogi cal rule. oPerating at the l exical l evel are .ubject to mare excePtion. than rul es operating at the post lex i cal l evel . Thi s type of v.neral i zati on aPpl ies to the Suci te noun phrase i n that the mare closel Y l i nked the nomi nal comPonents are semantical l y, the �reater the VAri ety of tonal change. wi thin the Phrase. The mast ti9htl y kni t noun Phrase i s the comPl ex noun that Possesses a si n91. noun cl ass suff ix at the end of the word. Two or mare morphemes Are strung together before a suffix is added . These morphemes maY be nomi nal , verbal , or adjecti val i n ori gi n . The noun in ( 1 . ) Pas.e.sRS two morphemes: the --- - - -_ . _-- ------- 203 fi rs� one i s nominal and �he second i s verbal J the second morPheme at the noun i n ( 1b ) is adjecti val . In each case. the suff ix is underl i ned. (1) Det i ni tl! Engl ish nYIl-tb-� nYil-tb-1ll (head-cover > hat nk�-c�-!li nkil-c.-!lIl (chicken-good) good ch i c ken Indl!fi n i te Certai n type. of noun comPoundi ng are more productive than other•• Tho.e whi ch are more Productive exhi b i t a MOre .tabl e Pattern of tonal chan;e •• Exampl •• of th.se changes wi l l b. seen a. the tonal rul •• are di scussed . The next l evel of comPlex noun. consi sts of two or more comPonents where the f i rst nominal comPonent retains i t. i ndef i n i te suff i x . but def i ni ten ••• i . marked on l Y at the end of the word. The i ndRf i n i te suf f i x i . underl i ned i n the col umn of i ndl!f i ni te forms in ( 2 ) whi l e onl Y the def i n i te suff i x is under l i ned i n the second col umn . Note that the i ndef i n i te suff i x is sti l l retai ned on the f i rst morpheme even when the enti re word ends i n a def i n i te suff i x . (2) IndRf i n i te Det i n i t. a. nYIl-!lJl-fbl ' nYil-"il-fbl �� (head-chi ef ) boss, chamPion b . ca-!l4-nY.-lt ca-";)-nYt � -c.& (daY-eYe) sun c. c E n-u.-Yar6-d cr n-xa-rir6-ki (.auce-thi n;.) sauce i ngredi ents Thi . grouP aPPl!ars to exh i b i t sl i ghtly more tonal stabi l i ty. though there are sti l l a number of l .x i cal ano.al i es. Other constructi ons found in the noun Phrase are .are l ao.. ly l i n ked, bei n g composed of two or marl! separatl! nouns, each wi th thei r own noun cl ass suff i x . Tonal chan;•• wi thi n the• • Phrases are qui te re9ul ar and form patt er ns ea.i l y defi ned b y ru les . Gen i t i ve constructi ons and Noun + number construc�i on. fal l ------ ---- ---_._---- 204 wi thin thi s categorY, a5 shawn i n (3) . ( 3 ) a. n��-� iba-kA ' the man ' . hou.e DEF ' b . gba-� soonf ' twa house. INDEF ' c . gba-ax,i .CI��!lR ' the two houses DEF ' Thi . chapter wi l l f i rst deal wi th the more regu l ar tonal behaviour found across ward boundaries as wel l a. across morpheme boundaries wi th i n the comPl ex noun . Tonal behavi our whi ch i . re.tri cted to comPlex noun structure. wi l l be d i scu••ed in the fol l owi ng secti on . B. Newtn.l Tme It was observed that verbs exhi b i ted three l evels of tone, High, Hi d, and low tone. Hi d tone on verbs was analYzed , wi th the hel p of two tonal tiers, a. lh . Nouns were also .een to have three l evel . of tone. However , tonal comPosi ­ t i on of nouns aPPears t o be mare compleM than that of verbs. Fi rst of al l , contour tone. are qui te frequent. To date, these contour tones have been l abeled Hi gh-low, Hi gh-Hi d , Hi d-law, and law-Hi gh . Second l y , i t was menti oned i n Chapter 3 that there are two dif�.rent tvpes of Hid tane behavi our and two different tvpe. of Hi d-low tonal behavi our . At that poi nt, al l Hid tone. were gi ven the tentati ve rePr..entati on gf lh. It wi l l become cl ear hare. however , that different under l Vi ng rePre.entations are requi red for each tYPe of Hi d tone. A. a bri ef review, the tone Patterns al l owed an simPle naun. are gi ven below: (4) ---- ---_ . wtr t High - Hi gh-Hi d -FYia Hi gh-law j� Hi d gba -- - .. - - - --- ' monev ' ' r i ver ' --- -- -- ------ . . -- .- 20S Mi d Weak caan ' market ' Mi d-Law S canB ' Younger .i bl i ng ' Mi d-Law W ji ' san ' Law c1 ' chi l d ' Law-Hi gh gban-l � ' granarY ' I I . TaNAL BEHAVIOUR WITHIN TIE NOUN pHRASE In thi s .ecti an the behaviour of each tone is di scu.sed .eParatel y. Since Hi gh and Law tone naun, present the l .a.t comPli cation., th.y wi l l be di .cu••ed -F i r.t . A, Hi gh Ion. Npuow Hith i ni tial noun. are never su.cePti b l . to tonal chang.. regardl e.s a� the tane a� the Precedi ng naun. The examPI .. below .haw that the Hi gh tane naun� f'14-ke. ramain. H i gh tane when preceded by Law, Mi d � and H igh � i nal nouns. (3) ' he saw the rack ' a. wa1 ��l'ke nYa b. me fa1 4ke ' Your rack ' c. wu Ya w6 fa l'ke nY�� c he i s seei ng his rack ' DePending an thei r ward fi nal tone, Hi gh tone nauns can tri gger di��erent tonal changes an the fal l owing ward. Recal l that Hi gh-Mid nouns tri gger subregister Hi gh tone SPreadi ng onto Law tane verb •• Hi gh-Law nouns tri gger Law tone sPread onto High tone verb • • wh i l e simPle High tone verbs tri gger subre­ gister Hi gh tone La....ri ng. (6) Hi gh-Mid wa1 fv.aa t�xl -> wa1 fyaa tux ! C he rack saw ' Hi gh �--- ��. - - - - ' he � i .h vomi ted ' -�- - � - --- � - - --� - --- -� -- 206 8. 19w Tgn. NguOI Whi l e High i n i ti al nouns are nat susc�ptible t� tonal changes, Law tone nouns do became Mi d when Preceded bv a Mi d tone. Thi s i ncl udes anV noun that be�ins with a Law tone, �5ther Law tone or Law-High . There are no Law-Mid tone nouns. �l� -> ma m� l � ' Your rice INO ' b . ma nd�r.x� -> ma ndar.x� ' Your Yam INO ' (7) a. me c . �ba-Y. s��n! -> ;ba-VB soon! ' ' twa houses ' Thi s tonal chan�e take. Pl ace Pr i mari l y acras. ward boundar ies. There is onl y ane known i nstance where the same Process takes Place w i th i n a comPl ex noun whose � i rst consti tuent retains the i nde� i n i te su��ix : -> s.Ka-ka� ala�a (8) a. s�xa + kl��nl a-�l ' zebra, l i t . bush donkey ' In Chapter 4, we ProPosed that Mi d tone was comPosed a� the tonal feature Lh. Pl acing the Pri mArV and secondarY �.atures on twa seParate but l i nked tiers, it was PrOPosed that a Law tone verb becomes Mid tone by simpl y sPreadi ng the secandarv �eature. n � the Precedi ng Mid tone onto the fal l owi ng Low tone (High SPreading ( lS » . Th i s was fal l owed by a rul e spreadi ng the subre9istar High tone di rectlv onto the segment (High Spreading to TBU ( 36» , creating a surtace Mi d-Hi gh tone. as shown below. (9) wal me I L t� -> wal L" I L h H SPREAD me I L tl t, ' h. showed YOU ' L� V-' h H SPREAD To TBU The exampl es gi ven i n (7) and (8) show that a Law tone noun becomes Mi d tone --- ----. .- -- 207 after a Mi d f i nal el ement and not Mi d-Hi�h. as is the ca.e for verbs (sae (9» . If one moti v.ted the Hi gh SPre.d rul e for nouns, but l i mi ted the Hi�h Spraad to TBU as a rul e that aPpl i ed on l y to verbs. the spreadi n� of the subregi ster Hi gh tone to the fol l owi ng Low tone woul d effecti ve lY Produce a surface Mi d tone, as shown i n the examples i n ( 10 ) . ( 10 ) mo I � l � -> I L L I mo I L m�l � 1, ... .. ... h h ' Your rice IND ' I L ' H SPREAD s��n1 -> gbaYiI sooni I I L H 1 L I I ' two houses ' L H 1 "/ h H SPREAD mo I L I m�B-"e I L ' Your r i ce DEF ' I H h H SPREAD c. Mid IQa. Noun. Actual l Y there are at l east f i ve tYPtB of Mi d tone behaviour in Suci te. In ChaPter 4, i t was sean that the tonal behaviour of the varbal particle, bA, differed from that of the Mid tone verb . It w.s never subject to tonal change and i t tri9gered d ifferent tonal rul es than ei ther M id tone verbs or M i d f i nal nouns. Thi s Mi d tone parti c l e was ProPosed to be under l Yi nglY [Low-h i gh ] , wi th the square brackets i nd i cati ng that the comPlex tone is .1waY, si mpl i f i ed on the surface as M i d tone. 208 It has al readY baen menti on.d that th.re are two tYPe. of Mid tone nouns. Both behave the sam. i n that theY both trigger Hi gh tone spreadi ng onto Law tone verbs. However , when preceded by a nami nal element, the one tYPe normal l y d o.. not c:hange whi l e the weak Mi d nauns bec:o,.. Hi gh tane H Preceded b y a Mi d or Hi gh f i nal noun or Pronoun, and Low-Hi 9h i f Preceded by a Law f i nal word. Thi s is i l l ustrated bel ow e (11) Precedin g Tone none Mi d Low Ala ' Your ' wur� ' hi s , emPhat i c ' Mid tone noun Weak Mi d noun gba-k' cEn-k' me ' hou.e-DEF ' ' sauce-DEF ' mo cin-ke gba-k' ' Your house-DEF ' ' Your sauce-DEF ' wur� gba-k' wur� ctn-k' ' hi s-EMP hause-DEF ' ' h i s-EMP sauce-DEF ' Al thaugh there are no Mid-Hi 9h tone nouns, there are a number of Mi d-Law nouns , most of whi ch undergo tonal change., whi l e a few undergo no tonal chan�s. Those Mi d-Low nouns whi c:h undergo tonal changes behave verY Simi l ar l y to weak Mi d nouns. As a resu l t , they wi l l be di scussed al ong wi th Neak Mid nouns bel ow. Mid and Mi d-Low nouns and Neak Mi d and weak Mi d-Law nouns tri gger tonal changes an fal l owing wards i n accordance wi th thei r ward f i nal tone. Mi d­ Law nouns tri gger Law tone sPreadi ng onto Hi gh tone verbs, whi l e Mi d tone nouns SPread thei r subregi ster Hi gh tone onto Law tone verbs. Below are a coupl e of .Hampl es shawing these tonal al ternati ons: ( 12 ) Mi d-Low tri ggers L SPread onto H verb wa� con l b wee -> wa� can l b w�e ' he l ocked at a Younger brother or si ster ' --- "-" "----" - -- - "- - ----- --- � 209 Mi d tri ggers H SPread onto L verb WA� Poru 9blr� -> wa� Poru gbar� There is a ' h. met a daughter ' third type of .ur�ace Mi d tone, however . It does nat shaw uP an simPle nan-comPlex nouns� but rather. on a couPl e o� � i rst Person Pronouns and wi thi n cerbin compl ex noun structures. Thi s tyPe o� Mid tone nat on l Y tri ggers a �ol l owing weak Mid tone to becoma Hi gh tane, i t also lowars to Low tone i n th i s envi ronmen t. The under l i ned vowel s i n the examPl es below i l l ustrate thi s behaviaur . ( 1 3 ) �da clnk' -> �d i cinke ' my sauce-DEF ' gbg,-lik' -> ibj),-U ke ' ol d 9ranarY-DEF I Thi s Prel i mi narY descr i pt i on o� the varying Mid tones i n Sucite does show the necassi ty � undertaking the anal Ysi s � Mid tone �eatures. Usi ng a doub l e ti ered approach � or tone �eatures, HYman ( 1985) suggests �our waYs t o rePresent Mid tane: Tonel e•• - that i s, segMents not marked �or tone are assi gned Mid tone as a de�au l t tone: X , M Lh - where Low tone i s i n the Pri marY register and Hi gh tone i s on the secondary or subre9i ster t i er : x I L I h HI - where Hi gh tone is a Pri marY regi ster tone and Law is �aund an the subre9i ster ti er : ----�- ---- ---�---- X I H I 1 · 21 0 LH - wh.�e bath Law tone and Hi gh tone a�e �ound an the p�i marY �a9i ste� t i e� and l i nked to a si ngl e TBUI X " L H Let us, than , expl o�e thi s Mid tone behaviou� i n an attempt to determi ne whether these Possi b l a unde�l Ying �eatu�es maY be �easi ble �or Buci ta. The � ol l owi ng sub.acti ons di scus. and anal Yze seParatel y the b.haviou� o� Mi d and Mi d-Law nouns, weak Mid and weak Mi d-Law nouns. and the tYPe of Mi d tone found an o U rtlt Persan P�onoun •• In anti c i Pati an of the anal Ysi s, it has been ProPosed that thi s l atte� tyPe � tone i tl unde� l Yi nglY Low-Hi gh . D. Mid and Mid... CW 'CIW Nmos The fi �st set � Mi d tone noun. to b. di scussed a�e tho•• whi ch are simpl y l abel ed as hav i ng a Mid tan.. These are the nauns whi ch trigger High tone s�.Ad onto LON tone v.�bs and nouns but which themsel ve. are .el dam i n� l uenced by the tonal .nvi �onmant. I� A Mid tone noun does alte�nAte in tane, i t sh i �ts to Law tane. but neve� to Hi gh tone3• Li kewise Mi d-Low nauns in thi s categorY a�e nat subject to tonal �ul •• • Selow a�. soma exampl es o� Mi d tone nouns showing the l Ac k � tanal change re9ardl ess o� whAt p�eced"l ( 14) '00: _ Mid-Law nouns Mi d tone nouns ' Your hause ' mo conla-�a ' Your Younger si bl i ng ' ' th i s EMPH hausa ' wur� conl �-�e ' h is EMPH y. sibl i ng ' bObd conl &-� ' a dea� mute ' s y. sib l i ng ' I n Chapter 4 . i t was tentat i vel y P�opo.ed that al l Mid tone nouns were underl Yi n91 y Lh. The �ecedi n9 di scussion has shown that it woul d not be wi se ----- ----- - .- -.-- - --- ----- 21 1 to posi t the sama underl Yi ng tone for al l tYPes of Mid tone nouns. The question faci ng ua hera i s . what i a the under l Yi ng natura of thi s type of Mid tone? In Chaptar 3, an exami nati on of nominal mi n i mal Pairs ravealed that no mi n i mal Pai r. contrasti ng Mi d and Low tone nouns had been found ( see pp. 81-2) . I n Chaptar 2 . the fact that a verY smal l proPorti on of minimal Pai rs contrast ing Mid and Low tona verbs were found . was used as evidence in sUPPorting the hypoth.sis that Mi d tone � 1rbs and Low tone verbs were hi storical l y of the same tone (see pp. 61ff ) . It ha. al so b ..n seen that Mi d tone nouns and verb. both al ternate with Low tone i n certai n environmants. They are both caPabl a of l owering to Low tone when f ol l owed by • Hi gh tone suf f i x (by means of the Hi gh Dal etion rul e (88) ( s.. Ch . 3. pp. 1 17-127 ) , where•• nei thar alternate wi th Hi gh tone. The hYPothesi s that M i d and Low ton.. maY hava had the same hi stor i cal ori gin l ed to the devel oPment of tonal f eatures that rePresent.d a hi stori cal tonal spl i t of Low tone. where Ll was th� lowered Low tone and Lh . the rai sed Low tone. Although i t i s d i f f i cul t to di scus. tha tonal f aaturas of M id tone wi thout discussing i t i n rel ati on to the other tvPes of Mid tone, we shal l tentativelv posi t the M i d tone noun ( i . e . the tvpe of Mid tone that rarel v changeS tone) as undarl Yingly lh . Thi s Lh nou n . then , can verY easi l v trigger Hi gh tone Spreadi ng onto Low tone nouns and verbs. The exampl e bel ow shows tha subregister Hi gh tone of noun spreudi n9 onto the Low tone of a verb as wal l as onto the f i nal Teu of the verb. ( 1 S ) wal Poru tt -> wal poru t, -> wal Poru tl I ,, l Ml I h I l II ML I I. l L 1, ... ' h HIGH SPREAD ---- ----- - -- - -- II Ml I r\ ' He showed a daughter ' l� � ... l h H SPREAD TO TSU ------ - ---- ---- --- - -- - - --- a lh ---- - 212 The Mi d-Low noun, accordi ng to this analysis wou l d be LhL, underlYi n9l y , as shown bel ow. ( 16) conB , L I ' Younger sibl ing ' , L h One maV suggest that H i gh SPreadi ng should al so take Place ward i nternal l y, creat i ng a surface M i d tone noun . Thi s , however , do•• not happen . Hi gh SPread must be l i .itad then , to spreadi ng across morPheme boundarie •• As LON ? i nal nouns these Mi d-Low nouns do tri gger Low tone SPread onto High tone verb., as i l l ustrated by the der i vati on below: ( 1 7) wa� con l � w" II I I ML L L I I h I H -> wal con l � ��, -> wa� conl � 1. 1 I J,�!.1 ML 1 L L h I I I H LOW SPREAD �, JI I � ML L L H I I h ' he saw a Younger si b l i ng ' 1 LSI DELINKING E. we.k Mid .nd Mid-LAW tAn. Noun. W.ak Mid ( Mw ) and weak Mid-Lcw (ML) nouns can have a vari etv of tonal shapes dePendi ng on the tonal envi ronment. Because of the comPl eK 1 ty o� thei r behavi our, thi s secti on i s somewhat lengthy, i nc l udi ng not on l Y an anal ysi s of underl ving tone, but al so the formu lation O? rul es and a di scussi on o� rul e order i ng. ---- - .- . .- -- -.- --- - -- 213 1 . DllSc:ri ptian weak Mi d and weak Mi d-Low (hence�orth re�erred to as Mid-Low-) are di scussed together here because o� the simi l ar i tv in thei r tonal behavi our . In �act , the onl Y t i me thev d i ll er in tonal structure is in i n i t i al Fasi t i on o.f a noun Phrase as wel l as i n ci tation �orm. Thei r di ��er i ng tonal structure in these Posi tions al so l eads them to tri gger d i ��erent ru l es . In tha examPl es balow. not. that the Mi d-Low noun , �, tri ggers Low tone SPreading onto High tone verbs ( 18a) , whi l e the weak Mid noun , � aPPears to tri gger subre9i ster Hi gh tone sPreadi ng onto Low tone verbs ( 1 8b ) , as do al l Mi d � i nal nouns. ( 1 8 ) a. Mi d-Low karl ' m.at ' ' he chewed meat ' ' he i s chewing m.at ' b . W.ak M i d wa� c£nx£ c£ nx£ tl ' Huce ' It�1 ' she showed sauce ' ' she is showi ng sauce ' However , when subject to tonal changes, ....ak Mi d and Mid-Law nouns undergo the same changes in i dent i cal tonal environments. Thes. changes, tentativel v l abel .d Rai sing and Low tone SPreadi ng, are descri bed bel ow. A f u l l anal Vsi s of these al ternati ons wi l l fol l ow the descript i on i n the di scussion i nvol vi ng the anal Vsi s o.f the under l Ying features for w.ak Mi d and Mid-LON ton.s. ( 1 ) Raising: The f i rst tonal change to be descri bed here is what is tenta- t i vel v cal l ed Rai si ng. When Preceded bv weak Mi d . Mi d , or Hi gh f inal nouns, ----- • Weak Mid-Low noun. bai ng much mare numerous than reiu l ar Mi d-Low nouns, d i scus.ed in the precedi ng secti on , wi l l be referred to as si mPl y Mid-Law from now on , whi l e the re9U l ar Mi d-LON nouns wi l l not be re�erred to again in thi s thesis. ----- - --- -- -- - -- -- --- - - - - - -- - - --- - --- - --- - -- - - - - - 214 both weAk Mi d And Mi d-Low nouns are Hi gh tone, as can be observed i n the exalllPl es bel ow: ' A dAughter ' s mortar IND ' ( 19)A. Peru SG-xo -> Poru s6xEl M M + Mw H b . Peru kA-r" -> Poru kara M + c . taxil mut Mw + H M tiL ' A daughter ' s meat I ND -> taxa iii xt Mw ML ' A tru ' s name ' H Mw + MW - No examPI .. d . 9b�n l � + .d -> gbbnH mid' L H + L H ML ' granarY ' s name ' H (hoe + smal l tree> LH ' hoe handl e ' H The same Rai sing behaviour can b e found i n ComPlex nouns . The examPles below shew that i f the f i rst roet ef a comPlex neun is weak M i d , Hi gh er Mi d tene, . i t wi l l rai se anY i mmedi Atel y foi l ONi n9 Neak Mid or Mi d-Low noun root ta Hi gh tane. ( 20) a . �mE I , + juu --> Mw + voice Mw mejd-I S M H sAVl ng, words b . gba-x a -b.l � --> gbab a 1 8 M .. --- • sPeech , --- ML M . ----- H ' ream ef a heuse ' 21� -> ntaraf61 � c . ntara + foB I'1w ' l and ch i ef ' ML d . ern-xI + Yare-x a -> crnxayir'xa Mw ' sauce i ngred i ents ' Mw sauce thi ngs Thi s derived Hi gh tone affects the tone of fol lowi ng morphemes in the .ame waY as anY other High fi nal noun . Thus anY rule or dering would have to ref l ect the fact that this ' Raisi ng ' takes Pl acm on a morpheme be-Eore the tone of that morpheme tri ggers anY rul e. on fol l owi ng morpheme •• Several exampl es of such a sequencing of tonal behaviour are shown b.l ow. In the fi rst .et of examPl e• • (21 ) , the def i n i te suffi x b ecomes M i d tone aft er the d er i ved High tone (a and b ) , just l i ke def i ni te the suf f i x becomes Mi d tone after .nY d er i ved High t one k> • (21 ) a. mo so-k' -> ML H M -> b . mo kaU6 H M mo k�te H ' Your meat-DEF ' M -> faI a ke c . fal a-k6 H ' Your mortar-DEF ' M 1'1 ML H M mo stlke ' rock-DEF ' - H After a deri ved Hi gh tone� other weak Mi d and Mi d-Low morphemes and nouns wi l l al so become Hi gh ton e . Thus, as shown bel ow, Rai sing can occur i terat i vel Y across the Phrase. ( 22) mo nYi m-b.l �-k' M Mw ML -> mo ny! m-bal a-ke -> M H Rai sing ML rna M nY! m-bal'-ke H Rai si ng H ' Your n i ght ' 21 6 Mi d tone ve�b. are l owa�ed to Low tone i n Ph�ase f i nal Posi tion th�ough Hi gh Del i nki ng, if p�ecedad by a de�i ved Hi gh tone ( 23a) , as they do fol l owi ng anY Hi gh f i nal noun (23b ) : ( 23) a. wa� eo karA nYa -> wa� mo ka �' nYa -> Rai sing b . wa� l'l 'xa nYa' wa� mo kar a nyA " he saw Your Hi gh Del i nking (�4) " meat ' -> waA lal 6x a nY� " He saw a rock ' Hi gh Del i nkin9 (44) (i t ) Law tane "-"ln91 L ow tane spreads onto H i 9h tona verbs. Therlt'fare, one wou l d exPect that Low tane Naul d al so sPread onto H i gh tone nouns. Thi s however . i s not the case, as shown bel ow: ( 24) wur� fal6ka -> .wur� f�l �ke ' hi s REF roc k ' rath.r : wur� fal 6ke Hi gh tone on nouns i s one of the most stabl e tone. i n Suc i t• • I t i s never suscept i b l e to tonal chAnge. Low tone Spre.di ng occurs i nstead an NeAk Mi d and Mld-Law nouns. When a weak M i d or a Mi d-Low noun i s Preceded by a Law � i nal noun o� Pronoun, the noun acqui �es a Law-Hi gh ca,tau r , as i l l ustrated by the examPles bel ow: (25) a . nAA � b. nU -> n.� � ' . man ' s mo�tar ' k.&::I::.l -> nU kki c . n�� k�rax� -> n1� k�r.x' d . 9b�r�tila-xe -> 9b.� �t� l �-x' ' a man ' s meat ' " . man ' . i nheri tance ' ' toP of a door ' The examPles i n (26) shaw that Low SP�ead can al so occur ac�oss mo�pheme boundaries wo�d i nternal l y . Wi thin COmPl ex noun structures. Low tone SPreadi ng can occu� on Mid-Low nouns but no examples of Low SPreadi ng have been found on ---_ .. _--- - _ .. ------ ------- ------------ 21 7 weAk Mid nouns· (26) I. + Mw non. .f:ound I. + Ml.. a. nll + nval . (mAn + .ve) -> b . s�s1n + kud� ( b l ood + road ) -> c . k1n?l + .f:al � (vi l l Age + chi e.f: ) t1l. + Mw n4nv�l t ' .f:r i end ' sas1nk�de -> ' vei n ' kan?l��I � ' vi l l age chi e� ' nane laund Ml.. + ML -> a. 'CAI)1 + nVeU (dav + eVe> b . nve?� + lal� (lAC. + chi � > c. c� + t6 (dav + cavar (vb . > -> - cA""v.1 8 ' sun ' nve? ��l ' ' al der brother ' > cint�?8 ' umbrel l A ' When A naun ha. been subject t o I.aw tone spreAd , t h e resu l t i ng � i nAI High tane (ol the I.aw-Hi gh contour) tri ggers the same rul es as anY Hi gh � i nAI noun · Agai n , as wi th Rai s i ng, rul es must b. ordered such that I.aw tone SPreadi ng must occur on the ftIOf"Ph ... i nvol ved balere the tane al thAt morpheme can tri gger tonAl chAngeS an anY � al l awi ng morPhemes. For examPl e, once Mi d-Law or weak Mid nauns acqu i re a I.aw-Hi gh contour . the � i nal Hi gh tone can the� l ower the Mi d tone verb to I.aw tone (Hi gh D.l i n k i ng) as do anY ather Hi 9h � i nal nouns. The examPles bel ow shaw bath a weak Mid ( b > and a Mi d-Law < a > noun , havi ng acqui red a l.aw-Hi 9h contour through I.aw SPread . l ower i ng a Mi d tone verb to Law tone i n Phra.e l i na l Pos i ti on . ' ( 27) A. nl-nvel i - > nlnv�18 wAa n1nv�l l nva b. n11 n11 - saxa -> n11 s�x6 nVa -> > waa nlnv� l l nvl ' he saw a fri end ' , hi s-REF /DOI"tar ' s�xe n11 lriend ' ' s�x6 nv1 ----- ---- · S.. h i s REF mortar ! ' 218 In summarv, then � weak Mi d and Mi d-Low nouns tri gger High tone sPread and Low tone sPread resPecti vel v when not a��ected bv tonal changes. However , when both become Hi gh tone through what has tentativelv been CAl l ed Rai si ng, or Low­ h i gh bv meAns � Low tone SPreadi ng, thev behave as Hi gh � i nal nouns, l oweri ng Mi d tone verbs ( Hi gh Del i nk i ng ) to Low tone in phrase f i nal Posi t i on , and rai sing �ol l owi ng weak Mi d and Mi d-Low nouns to Hi gh tone. The anal vsi s gi ven bel ow wi l l exami ne the rePercussi ons o� these tonal ch�nges more c l osel v and attemPt to PrDPose a vi ab l e sol uti on . 2. Analvsi s There are at l east two possi b l e waVs to analvze the behavi our o� weak Mi d and Mi d-Low nouns. One i s to consi der them underl vi ngl v Hi gh tone si nce thei r tonal behaviour i s verV s i mi l ar to that o� High tone verbs. A second Possi b i l ­ i tv i s to Posi t an under l vi ng Mi d tone of some tvpe that wou l d be subject to Rai s i ng and LON tane sPreadi ng rules. 2&. .... .rid /II.. .s �d.... lvin9lv HiG1 Tan. The f i rst suggesti on i s that weak Mi d and Mi d-Low nouns are under l vinglv High tane. but become Mi d or Mi d-Low tone i n certain envi ron..nts. Thi s i dea was promPted bv the fact that their tonal behaviour i s verV si mi l ar to that o� High tone verbs. Thev are both subj ect to Law tane SPread i ng and are Hi gh tone i� al most anvth i ng Precedes them wi thi n the noun Phrase . The onl v t i me thev are Mi d or Mi d-Law tone i s when thev are in i n i t i al Posi t i on � a noun Phrase. Thi s l oweri ng � Hi gh tone maV, i n effect , be a nomi nal i z i ng �eature. Evidence � thi s surfacas when Hi gh tone verbs are nomi nal i z ed . No Hi gh tane verbs become a stab l e Hi gh tone when nomi nal i z ed . Rather , in (28) , thev are ----�- - -�- -,--- ------ -- -- - - - --_ .. _. - ---_ . 219 shawn to be l owered either to Hi d , weak Hi d or Hi d-Law tone: (28) a. Hi d b . Hw "mS l !S ' sl eep ' -> 1 6x6 ' hear ' -> "dil-x e ' ear ' gt,6 ' ki l l ' -> gbuun ' funeral ' '" c. HL ' ''m:ll a-x:l ' n i ghtmare ' ya ' tt; be i l l ' -> Ya-m� ' i l l ness ' k�?i1n ' be UrG ' -> kan?a-r.l ' fati gue ' If we suPPose that weak Hi d and Hi d-Law nouns wara under !Yingly Hi gh tone however , we imllHldi atal Y run i nto an. potenti al hurd l e . There AlreAdy .x i st. A set of Hi gh tone nouns i n Suc i t. wi th the SAme pi tch l evel as Hi9h tone verbs, which are never suscePti bl e to anY tonal rul es. As shawn in (29) , Hi gh i n i t i al noun. never l ower to Hi d Dr' Hi d-Law, they do nat tri 9ger the l awerin9 of the tone of � ' mY ' ( to be s..n shartl y as a matter of Right Si de Del inking of a Law-High contou r ) nor are theY subject to the Law tone SpreAdi ng Rul e. al l i n contrAst t o the behavi our (29) of weak Hi d an d Hi d-Law tone nouns I Hi lib Noyn' ML ,nd Mw Ngyn, kAUe ' the rack ' ' the meat ' ' the sauce ' &iil feU 6ke • c.r:I1 kiate mY rock ' adA wur� fii16kp , cln ke ' my meat ' ' IIY sauce ' hi 5 aFH rac k ' wur� � ' h i s EHPH sauce ' In effect , there seems to be a tonal barri er Preventi ng anY tonal rule from APpl Yi ng to Hi gh tone nouns. It must be noted that these High i n i t i al nouns make uP a verY smal l number of the data samPl e and thAt mare than hAl f CAn be c l ailrl Y i dent i t i ad as Loan wards. Al l Purel y Hi gh tone nouns, wi th one possi b l e ----_ . _- _ . _ . _ . --- .--. . . . _. ..,.. 220 axcept i on . are Loan words. One Possi b l a h i stori cal devel opment i s that the ori gi nal sat o� H i gh tone nouns l owered to Mi d tone i n certain environments. High tone nouns i ntroduced to Suc i te, however , were not mada sub ject to thi s tone l owering Phenomenon and as a resul t . a d i sctincti ve tonal contrast between Hi gh and Lo�ad Hi 9h tona nouns devel oped . Shoul d than , thi s set � l owered Hi gh tone nouns be rePres.nted as Hi gh i � there i s al readY a set a� Hi gh tone nouns? I f these sur�ace High tan. nouns can be sPeCi al l y marked so as not to undergo anY tonal rul es, weak Mi d and Mid Low nouns .can then be posHad as underlYingl y High tone and as a r·esu l t , wou l d be subJlCt to the Low tone sPreading ru l e i n just the same waY as Hi 9h tone verbs are. Bel ow are examPl es � Low SPread an both a verb and a noun : (30) -> LtdB Plrl L H I I �B Pirl ��--1 L H -> �B Pj rJ � lOW SPREAD I I wur� Jcara -> I M L H wur. k ". I 1-:':..... 1 i'l L -> H LOW SPREAD L H LS DElINKING wur. klr � , � M L C ull rice ' -> wur. klra H LS DELINKING ' V t\ M L ' hi s REF 1IlUt.' HL PHRASE F I NAL L INSERTION After Mi d and H i gh f i nal nouns and Pronouns, thi s underlYing Hi gh tone wou l d remai n Hi gh tone. just a s h i gh tone verbs ramai n Hi gh i n t h e same environment• • thus el i m i nati ng the need for a Rai si ng Rul e. as shown b.l ow. (31) mo I lara -> L I n \J lot me I L , h kara V. ' Your meat ' H L PHRASE FINAl.. L INSERTION ---- - - -- - . - - -- 221 How then woul d one account for the loweri ng of thi s High tone to Mi d tone? Recal l that thi s • l ower i ng ' Process takes Pl ace when an underl Ying Hi gh tone noun is i n Phrase i n i t i al Posi t i on . One coul d suggest that there is a spl i t i n the High tana regi ster at thi s Poi nt. Whi l e High tone verbs rema in Hi gh tone. the nouns in Phrase i n i t i al Posi t i on are desi gnated to a l ower l evel of the Hi gh tone re9ister . The spl i tt i ng of tha Hi gh tene register i s. accordi ng to thi s aPProach. moti vated by th e fact that Hi gh tone nouns are i n phrase i n i ti al Posi tion. Therefore a rul e woul d have to be devi sed to l ower Hi gh tane at thi s Pai nt. Thi s Proce.s can be represented 'by i nserting a Law tone i n the subregi star tier . to be stated as fal l ows. (32) PHRAsE INITIAL LOW TONE INSERTION A Hi9h tone noun i n Phra.e i ni t i al position i s marked for Law tone an the subra9ister t i er : H - > HI / a. I cenXE -> canKE I H I I • m Phrase ini t i al Posi tion sauce ' H I I I t has been assumed unti l now that weak Mi d and Mid-Law nouns have the same underlYing canf i 9urati on . The questi on remai ns. however . as to how to d ist i n9u ish between the two. so that when I -insertion occurs. the correct surface farm can resu l t . One waY to deal with thi s d i l emma is to al l ow the subre9i ster I ON tone to l in k to the se9mental t i er . thus creati ng a Mi d-Law tone. The subra9i ster Low tone l i n ked to Hi9h tone creates a Mi d tone whi l e thi s sarna Law tone l inked i nd�ndent l Y to the segment.l t i er creates a surface Low tane. resulting in a Mi d-Law contour . ----- --- - -- - --- - - - - -- ------ - - -- -- ----- - 222 ( 33) I CV ji t, H' 1', H , I I I, 1./ A cartai n set � nauns. than. are l ex i cal l v marked for thi s l i n k i ng of the subregi ster Law tone to the Se9mental tie� . whi l & are another set <the weak Mi d nauns) are b l acked fram thi s l i nk i ng �acess. The Probl em wi th this procedure i s th.t whi l e thi s subregister Law tene l i nk i ng te a TBU is l ex i cal l v governed. the envi ronment wh ich Provi des tar thi s Law l i nki ng (namelv. l i nk i ng . subregi ster Law tone to • Pri marV t i er Hi9h tone) i s created bv waV of the Phrase l evel rul e. Law tone Inserti on . Thus the phrase l evel r u l e must take Pl ace betare the l ex i cal lv marked Law tene l i n k i ng can occur, as shewn bel ew: ( 34) wa1 � nVa -> wa� ja nVa -> wa1 ji nVa II ML I H I L I h II ML I I H L I h I I LOW INSERTI ON II ML f\, I H, L I h " ' he saw a san ' 1 LOW LINKING TO TBU Essanti.l l Y . there needs to be ·i nfarmati an i n the l ex i con that wi l l d i stingui sh between those whese Phrase i n i t i al tarm is Mi d and those wh i ch are Mi dLaw. However. i t does nat seem possi b l e that the l ex icon can Predi ct how a subre9i ster Law tone can behave i f i t i s nat even spec i f i ed i n the l ex i con. One Possib l e waV to sal ve thi s Probl em i s to Posi t a sl i ghtl v different underlving farm tar Mi d-Law nouns: HL. Phrase l evel I-i nserti on woul d subsequentlv viel d the correct outPUt f or beth weak Mid and Mi d-l ow nouns. A subregi ster Lew tone l i n ked to a High tone wou l d vi eld a Mi d tene whi l e a . subre9i ster Law l i nked to a HL woul d vi el d Mid-Law. as shawn below . --- - -- - -- - --- ----- ---- ------- 223 cv k Mi d Lawl cv I H HL V I I I That Mi d"'Law nouns are und.,.l Yi ngl Y Hi gh-Low i s not enti rel Y i ncancei vab l e when observi ng the cognates o� Mi d-Low nouns i n certain ather Senu�o l anguages. Althou9M numeraus other l an9Uages such as Cabar-a also exhi b i t thi s Mi d-Lew centeur. Na�ara a� Bandau keu . whi ch aPPears te have an l Y a twe tene l evel di st i ncti en . Previ des a Hi gh-Lew contour fer these same cD9nates. (36) Suci te Cebara Nahra af Bondaukou n j i d. nYi ni nUl ' tongue ' kud k::JB kit ' hand. arm ' f.,.� ": i r� H r� ' excrement ' (Mensah , 1983) I t i s Pcssi ble, ther"':ere, that Suci t e Mi d-Lew neuns are under l Yi ngl y Hi gh-Lew tene. These underl Ying Hi gh-Law neuns weuld b e diHerent from the marked Hi 9hLaw ncuns whi ch never submi t te tonal rules. Thi s Hi gh-Law repres.ntati an does create Prebl ems �ar rul e �ermu lat i on , he....v.r . The f i nal l aw tane weul d have t a b e del eted when.ver the naun i s Preceded by anather nemi nal ef anY tone sa that these HL (Mi d-Law) neuns can bacema COmPletel Y High tane. The ex amPl e, (37a) shews that thi s Lew Del eti en i s required after Mi d tane neuns, whi l e ( 37b > Prevides an i l l ustrat i en ef Law Del .ti on at the eutPut ef Low tene Spread . (37) a . ma kSrA ,. " L H L ---- . . --- -> me teara -> t L I I h " H CI) h LOW DELETION _ ._ ._---_.- me I L I lear;{ V H h LINKING --- ---- ' Your m.at ' 224 nk�-�r� -> Lor" L I H L LOW SPREAD nku-kar a -> I L l H (D I LOW DELETION n k�-k��a VV L ' chi cken meat · H LINKING AND RS DELINKING In �act , the on l � Place where the Low o� the suPPosed Hi gh-Law contour is �ound i s when tha noun i s in Phrase i n i t i al Posi t i on , the verY same envi ronment i n whi ch Phrase I n i t i al Law tone i nserti on ( 32 ) i s moti vated . Any time that the Law tone i nsertion rul e is nat in OPerat i on , then , the Law tone o� the Hi gh-Law Contour i s al so del eted . The obvious connecti on between these twa rul es however , is nat made expl i c i t in the deri vati on o� the rul es. I� the subregi st.r Law tone and the Law tone of the Hi gh Law contour are Present or absent i n the exact same envi ronments , Perhaps one mi ght specul ate that theY both maY have the same under l ying source. A second Problem for the under l Yi ng Hi gh tone hYPothesi s is the fact that the.e High tones l ower to Law tone i n certai n l exi cal envi ronments. Recal l that weak Mi d , Mi d-Law. Mi d , and Low-Hi gh noun roots were l owered to Law tone when �ol l owed by a High i n i t i al TYPe I I su�f i x . ' th i gh. ' nci - xal e -> nc�-xale ' balaPhons ' ' shea nuts ' It was suggested in Chapter 3 that thi s l ower i ng cou l d be accounted tar if al l four tones Possessed bath a Hi gh tone and a Law tone i n the ir underlYi ng con� i gurati on . The simple Process o� High del eti on wou l d then produce the same outPUt far al l of these tones, as i l l ustrated below for Mi d and Law-Hi gh tones. Hi gh Del et i on woul d l eave onl Y a Law tone l i n ked to the TBU. ----- ---- - -- ------ ---- ---- 22S (39) Mid CV Law-Hi gh I CV � L I ® If weak Mi d w�s Posi ted �s under l Yi ngl y High tone and Mi d-Low �s Hi gh-Low� the rul e of Hi gh del et i on woul d Yi el d � Law tone for HL nouns. However , Hi gh Del .ti on woul d Y i el d a toneless TBU for we�k Mi d nouns i f they are anal Yzed �s underl Yi ngl y Hi gh tone at the l ex i cal l evel . (40) cv Mi d-Law ...ak Mi d 4 CV � + Thi s Probl em coul d Potenti �I I Y be overcame by i ntroduci ng a di ssimi latorY rul e. wherebY a High tone on anY t i er is converted to Law tone when fal l owed by a <H , h > - > < L . I > I High tone suff i x : ___ H. Th i s rul e wou l d effecti vel Y produce correct outputs for Mi d (Lh ) , Law-Hi gh , �nd weak Mi d (H) nouns (see (41 » . Mi d-Low nouns. however . wou l d be bl acked f rom this l oweri ng bec�use thei r Hi gh tones are not �dj acent to the High tone suff i x (41d ) . (41 ) a. Mi d nci -x�l e -> nc� -xal e I L I h I I I , I H L h -> I I L H L I h b . Low-High , H I ' i H L -> I 1\ \ \ LL H L h , h ' she� nuts ' f o-?�l o -> f�-?olo I " LH ' b�l aPhons ' c . Weak Mi d l �-x�l o -> l �xalo I I I I H L L H L h I h I � debts ' I I d . Mi d-Low c1-x�l e -> *c�-xal e A HL , I H L , h ' thighs ' � I I HL H L I I I h L INSERTION 226 This entire approach conti nues to assume, of course, that the changel ess Hi gh tone nouns are barred �rom the tonal rul e Processes and are there�ore not susceptible to Hi gh tone Loweri ng. Verb. also cOmPl i cate thi s di ssimi l ation rule. Recal l that Mi d tone verbs are l owered to low tone be�ore the Hi gh tone i ncompl eti ve suff i x . The sol ution ProPosed in ChaPter 2 i nvol ved the del .tion o� a subregi ster Hi gh tone of the Lh verb .uf� ixed by a Hi gh t�n. morPheme. If the H->L rul e was i ntroduced i nstead, the correct outPut woul d be Produced �or Mi d tone verbs (s.. (42a» . However , this rule would al so i ncorrectl y aPpl y to Hi gh tone verbs, loweri ng Hi gh tone verbs to Low tone before i ncomPleti ve su��ixes (42b ) . This, in ��ect, does not haPPen . High tone verbs, as a general rule, remain Hi gh tone when an incomPl eti ve su�� i x i s added . (42) a. M verbs CY su-cJ -> s�-cJ I , I I L H L H I L I h -> 1 b . H v8l"'bs CY , H -> *L I h ' be buYing ' I 1 Pir! -f -> .Piri-f I I I I H H L H ' be sel l i ng ' Posi ting weak Mi d and Mi d-Low tones as underlYi ngl y Hi gh tone does create some Problems for anal Ysi s . First of al l . we have to separate the changel ess Hi gh tone nouns and mark them i nel i gible �or tonal change. Secondl y, the Posi ting � a underlYi ng Hi gh tone Pos•• Probl ems in how to d i st i nguish between weak Mi d and Mi d-Low tones i n the underl Ying structure and sti l l al low for the formul ation of sat i sfactorY rules. Fi nal l y, the formul at i on of a H-> L rul e for l oweri ng of weak Mi d and Mi d-Low noun roots to Low tone i s found to be lacking ----- - ---- - -- ------ - --- --- ---- ------ ----- 227 a �ul l general izat i on : High tone verbs must be excl uded �rom the rule aPPl i cation .nd Mi d-Low nouns whi ch should al so be l owered Pose dif�i cul ties �or the formulation of the H->L rul e. Although these Problems could perhaps be accommo­ dated wi thin the underl ving High tone hYPothesi s. an al ternati ve soluti on i s sought. 2b. !'IN ad "- as lkld.,1 Yin91 Y HI An al ternati ve to Posi ting an underl ving Hi gh tone i n the lex icon and motivati ng a Phrase l evel rul e inserti ng a subregi ster Low tone is to mark the subre9ister Low tone in the l ex i con. i tself . Mi d-l ow and weak Mi d nouns wou l d now b e considered underl Yi ngl y � ( or a lowered Hi gh tone) rather than si mpl y underlYingl v Hi gh tone. Where the Preceding anal vsis posi ted both the unchangeabl e Hi gh tones and the weak Mi d nouns as undarl Yi ngl v Hi gh tone. distingui shing the two onl Y bv blocking the unch.ngeable Hi gh tone nouns from undergoi n9 anv rul es. thi s particular anal vsi s di sti ngui shes the two tones bv Posi ting different underlvi ng features for each . Thus the unchangeabl e High tone nouns bear the feature. � that i s. High tone is posi ted on the Pri marv tier. as wel l as on the subregi ster tier . The weak Mid noun . then . i s consi dered the l owered High tone. � with a Low tone on the subregister tier . Both of these desi gnations are i l l ustrated below. (43) Hi gh tone nouns weak Mi d and Mid-Low x I X H H h 1 I I I Thi s devel OPment now s... both the High tone and the Low tone register spl i t 228 into twa tonal l evels (see (44) below) . Recal l �rom Chapter 2. that Cl ements ' proPosed tane �eature system allows far thi s possi bi l i ty. HaNever. these sPl i ts in Suci te do nat rePresent �our Phoneti c level s o� tone. as mi ght be suspect.d . Rather. i t has been ProPosed that the regular Mid tane i s underlYingl y �. whi l e weak Mid. whi ch i s at the same Pi tch level as the regular Mi d tone. i s Posi ted as being Hl. HY.an ( 1985) . ProPosed bath H I and Lh as Possi ble underlY- ing rePresentations �or Mi d tone (se. p. 209-10> . Cl ements ' system (as shawn below) does nat exclude thi s possibi l i ty of the twa underl Ying tones havi ng the same phoneti c pi tch . (44) < Cl ements ' Tone Feature sYstem H h- Ifl I - U. h L HI / � l - Ll - Furthermore. it i s entirel y possi b l e that the rai sed Law tone and the lowered High tone were at one time in the historY of the l anguage. distinct i n Pi tch, but that they l ast the Pi tch disti nction a s the tonal system of Suci te devel oPed" . Thi s ProPosed sol uti on wi l l requi re, however, the adjustment of tonal rules. In addition to modifYi ng Law Spread so that i t wi l l sPread onto HI nouns as wel l as Hi gh tone verbs, a rul e wi l l have have to be devi sed that wi l l rai se HI tone to Hh tone when preceded by Hi gh or Mi d tone nouns. The issue of how to disti nguish weak Mi d and Mi d-Law nouns usi ng the same underlYi ng representation 229 o� HI wi l l also be tested and discussed . ( i ) Law tan. !Prodi nSi Let us �irst i nv••tiSl.t. how this new devel oPment al ters the �ormul ation o� the Low SPreading rule. Previ ousl y, i t w.s stated that Low tone sPre.ds onto High tone verbs (a) . Here we � i nd that Low tone sPreads onto HI nouns (c) but not High tone nouns (b) , as shown below. (4' ) a. ka( L' ' to chew ' ,'\ , b. *��ljx", I, ', .... ,� ' H L' I h ' rock ' H , h c . c,"x€ ," I , L' '1 ' sauce ' H , I Since Hi gh tone nouns do not .l low Low tone SPr••ding. theY must be b.rrad from anY appl i cati on � the Low tone SPre.d rule. Once thi s constr.int i s Pl aced on the rule, the e�l i er formulation o� the rule gi ven in Chapter 4 would sti l l aPpl y here. si nce, as shown below, the rul e does not sti PU l ate what subregister tone must be l i nked to the Pri marY level Hi gh tone. ( 1 1 ) LOW SPREAD: When a Ll tone i s �ol lowed by a Hi gh tone .cross a morpheme boundarY, the Low tone spreads to the right . Xl X \.,'\ L H , I As a resu l t . both the HI and Hh tone words c� be subject to the Low tone SPreading rule. However , whi le Hh verbs Present no problems �or the anal YSi s as gi ven so far, the Low tone Spreading rule, as stated above does not . i n i tsel � , ----- - - - _ ._ ---------- 230 Yi eld a correct output for HI nouns. Law tone spread i ng onto a HI tone Produces a Law-mi d centeur tone. However. there are no Law-Mi d contours i n Suc ite whi ch are a resul t of a Law SPread rul e. Recal l from examPl es i n (25) and (26) that Law tone spreadi ng onto weak Mi d and Mi d-l ew nouns Yi el ds a Lew-Hi gh tone, the verY same tonal contour whi ch resul ts when a Law tone is sPread onto a Hi gh tene verb. A verY si mpl e waY of acquir ing a Law-Hi gh centour after sPreadi ng Law tene onto a HI noun i s to tri gger a rul e del eti ng the subregi ster Law tone. Thi s rul e wou l d be stated as fel l ows: (46) SUBREGISTER LOW DELETION: Delete a subregi ster Law tone l i nked to a Hi gh tone if a Lew tone from the l ef t i s l i nked to the same segm.nt as the HI tone. x X L H v-1 t I 1 <D Once the subre9i ster Law tone i s del eted. the Hi gh tone is l eft wi th no speci fi cati on of tone an the subre9i ster tier . Any tone whi ch is unSPec i f i ed at the subre9ister l evel wi l l by defau l t . acqui re the same tone as found at the Pri marY regi ster l evel . Thi s shal l be cal l ed subre9i ster speci ficat i on <SS ) . (47) SUBREGI STER SPECIFICATI ON d H -> OCHh In the der i vati on bel ow (48) , as wel l as i n succeed i ng deri vati ons of rul es , Subregister SPec i f i cat i on wi l l aPply automat i cal l Y whenever a rul e l eaves a tene an the Pri marY t i er wi thout a subregi ster tone. Si nce the aPpl i cati on i s ---- � ---� - �-- �- �--�� � - ----- - - -- --- ----- - ----- 231 automati c , i t wi l l not be necessari l y shown as a seParate step in the deri vation. Therefore, a HI noun whi ch i s subject to the Law tone SPread rule would have the fal l owi ng derivat ion : ( 48 ) wur. ClnXE -> WUI-� CE nXI -> �ur� CEnXI -> wur� dnxt I L.�'''''''1 I �?4 M L H M L I I 1 1 1 H I I LOW SPREAD I V , � M L H I I (!) 1 M L H L I h I I h LS DELINKING L DELETION � L H L I 1 I I I L SPREAD I H I <D � L H V L HL I h I hI I I LS DEL INKING L DELETION I ' h i s EMPH sauce ' I I L INSERTION nk�-�ra -> nk�-k�r� -> nk�-k�r4 -> nk�-k�r a ��-1 1\ 1\ ' chi cken meat ' II L INSERTION The moti vation for Low Del etion is not total l Y clear at this Poi nt . Possi bly i t is some tYPe of dissi mi l atorY Process, forcing a Pure Hi gh tone when i t i s l i nked to the same se9ment as a Low tone. At anY rate, subregi ster low deletion must take Pl ace i mmed i atel Y fol lowing a Law tone sPreading rul e. The second tonal al ternation that must be deal t wi th here i s what was referred to earlier as Rai sing. Thi s is the case where weak Mid and Mid-Law nouns became Hi gh tone when Preceded by Mi d or Hi gh tone nouns or Pronouns. If weak Mid and Mi d-Low nouns are now consi dered underl Yi ngl y HI , motivati ng the del etion of the subre9i ster Low tone woul d Produce, i n a verY si mpl e way, a High tone. However , i n thi s ca.e. there is no Law Spreading rul e to tri gger Law -- - . _ ._ ---- ,----_ .. _-_ ._-- ------ 232 del etion . Therefore one i s compel l ed to determi ne the general i z �tion that �l l ows for L del .tion a�ter both Mi d and High tones. In order to do thi s. l et us l ook at the under l Yi ng tone o� Mi d and Hi gh tone nouns. It has �l ready been seen that Hi gh tone nouns are underl Yingl Y Hh , whi l e Mi d tone nouns whi ch are not suscept i b l e to tonal changes (�s are HI nouns) are consi dered to be Lh . In exami ning both underl Ying ton•• • it is seen that both are spec i � i ed for High tone an the subregi ster t i er . (49) ' rock ' gbaxa t ' hause ' L I h I� pl ac i ng ei ther before a HI noun triggers Law del eti on , i t would nat be di f f i cu l t to Postul at. that this Low tone del eti on was moti vated by waY of the subre9istar Hi gh tone of a Hh or Lh noun spreading onto the HI tone. Such a Process i s i l l ustrated b.l ow i n (50) . Si nce i t i s not feasi b l e to al low twa subra9i ster tones to be l i n ked to a singl e tone an the Pri marY tier, the sPreadi ng � the subre9i ster Hi gh tone wou l d automati cal l Y cal l for the del eti on o� the subregister Low ton •• The rul e mi ght be stated as �ol l ows: (S0 ) HIGH SPREADING: SPread a suhregi ster Hi 9h tone onto a fol l owing Pri marY tier High tone and del ete the subregi ster tone l i n ked to that Hi gh tone. x I X I H T .. ... ... ... , h t Tit = anY tone Thi s parti cu l ar Hi gh sPreadi ng rul e would have the fol l owing aPpl i cation : ----- ----------- 233 (51 ) a. mo clnxl -> I L .. .. J H � ' Your sauce ' clnxt ' V'. L H L '/ I h h H SPREAD me 1 L INSERTION LINK I NG � Yam ' s sauce ' HIGH SPREAD LINKING L INSERTION Recal l , however, �ha� �he rul e ol Hi gh �one sPread has already been i n�roduced. In �he envi ronm.n� where Low tone becomes Mi d tone a�ter Mi d tone nouns. i t has been ProPosed ( earl i er i n this chapter as wel l as i n ChaPter 4 (35» that the subregister Hi gh of the Lh tone sPreads to the � ol lowin9 Low tone. as shawn below: (52) mo m�l � -> mo mol � ' Your ri ce ' \ L L ... h \ ... . L I n al l cases o� High sPreadi n9 described here. i t does seem that there is a tendency for a rai sed tonal re9i s�er to rai se the re9ister of the tones that �ol l ow. For exampl e . a rai sed Low (or Lh ) rai ses a l owered Hi9h tone as wel l as Low tones. and a rai sed Hi gh �one (Hh ) rai ses fol l owi n9 l owered high �ones. Thus it maY be P�5si h l e to pO$i � a general ized rul e, al l owi ng anY subre9i ster Hi gh tone to sPread �o the fal l owi ng Pri marY tier tone re9ardl ess of what that �one maY be. ---_ . _-- - . - - ---- - --- ---- .----- ------ 234 Seei ng the val ue o� i mPl ementi ng a hi9h spread rul e i n the examples above, i t i s necassary, then , to examine the sCOPe � the spreadi ng o� the subre9i ster h igh tone onto primarV register tones. Selow i s a l ist o� poss i b l e environments where the subre9i ster High tone cou ld technical l v sPread . (53) a·1ID dn-k. ' You,. ....c. · b . 1IID f 'l 6-ke • Your reck I c • • dn-k. '"y AIIc. ' d . "de U l a-ke " my rock ' L H L H H H H h 1 h h h 1 H1 h h L ... ... 1 e· 1IO II:Jl� ' You,. ri c. ' L L. ... '1 f . 1IID gba-k' " Your house I L 1.. - ... ..., h I L L h h L • .-. ...' g . ild• .cl� l .. ..... 1 ' .y • ,.ic. ' 1.. ... .-. - h . "de gba-k' " my house I H L H L h 1 h h l .. .. ....1 L .. -"l Above are al l the concei vab l e examPl es where a subregi ster h i gh tone cou l d Possib l v spread on t o a fol l owi ng tone. High tone sPreadi n g marked in bol d wou l d e�f ect i vel Y change the tonal structure � the fol l ow i ng word . Those not marked in bol d are redundant h i gh sPreadin9 Processes where no tonal change wou l d take Place as a resul t of the sPread. The example i n (53g) shows one envi ronment whare h i 9h tone sPread Produces an i ncorrect resPonse, however . A subregi ster h i 9h tone must not be al l owed to sPread onto a Low tone i t that subre9i ster h i gh i s not al readv l i nked underlYi ngl y to a Low tone. In other words, a subregi ster h i 9h o� a Lh noun can sPread onto a Low tone (see (53e» but the same of a Hh noun cannot . In order to al l ow for this ex cept i on , the hi9h tone sPread rule must be revi sed as fol l ows: 235 ( 54) HIGH TONE SPREAD: SPread anY subre9i ster Hi 9h tone onto the �ol l owin9 Pri marY re9ister tone, T2 and del ete anY subre9i ster tone Previ ousl Y l i nked to TIt = anY tane Condi t i on : I� T2 is Law tane, then T1 must also be Law tone. It has been assumed unt i l now that weak Mi d and Mid-Law nouns have the same under l Yi n9 tonal �eatures because o� the same behavi our they exhi b i t when subject to both the Law and Hi 9h Spreadi ng rules. As a resu l t both have been considered to be under l vin91 Y HI . However . we al r eadY knew that when nat a��ected by tonal rul es themsel ves (that i s , when in r : .rase in i t i al Posi t i on ) weak Mi d nauns tri gger di ��er.nt rules than Mi d-Law nouns. Somehow thi s di ��erence must b e accounted �or i n the underl Yi ng representati ons o� the twa tones. ( t v ) Mid-Law Tane and Ind�d.nt LON Tane Linkin9 It was Proposed earl i er that Mi d-Low nouns were under l Yi n9 Hi gh-Low. I � one l ex i cal l Y l i nks a subre9ister Low tone to both , the � i nal tonal output wi l l be Mi d-Low. (55) CV "- HL V I The probl em wi th thi s ProPosed under l Yi ng rePresentati on i s that everY time --- --------- 236 the subregi ster I i s del eted a�ter both law tone sPread and high tone Spread � the Pri marY ra9i .tar l must al so be del eted, as shown bel ow i n (S6) . Th is makes one wander whether both the Law tone o � th e Pri marY regi ster and th e Low � on the 5ubre9i ster t i er are actual l Y one and the same tone. La. tan. 'Pr�: __ Hi 9h tone sPread CD cv cv � G)� h wur� n3-�e -> wur� n�-�e ' hi s EMPH mather ' ' Your mother ' Ona sU99asti on for the rePresentati on of Mi d-Low tone i s �o Posit a si mpl e HI tone and then to ProPose a rul e l i n k i ng the subregister Law tone di rectl y to the segmental t i er whenever the HI tone i s i n Phrase i n i t i al Posi tion . Thi s � i nal indePendent l i nking ef�ect i vel Y creates a Mi d-Law contour tone� as shown bel ow in (57a) . Weak Mi d nouns, on the ather han d , would al so Possess an under l Yi ng HI tone, but they wou l d not be subject to thi s i ndependent l i nking o� the subregister Low tone (57b) . Thus, Mi d-Low nouns wou l d be marked i n the l ex i con � or thi s ex tra l i nking in Phrase i n i t i al Position , whi l e weak Mi d nouns wou l d not . In thi s anal Ysi s. we shal l rePresent thi s marking by underl i n i ng the subregister Low tone. Hl, then , re�ers to the Mi d-Low tone whi l e HI si 9n i t i e5 the weak Mi d tone. ( 57) a. Mi d-low CV 1\ H • 1,, ' 1 b. weak Mi d CV , H I I Thi s notion o� i ndePendent l i nking o� the subre9i ster ti er to the se9mental ---_ . 237 U er has al readv been introduced earlier and brieH v di sscussed i n ChaPter 4. Recal l that subr.9ister Hi gh tone SPreadi ng to the segment was i ntroduced to create a Mi d-High tonal contour on under lying Low tone verbs < ••• pp. 161�� �or di scussi on ) . The reason for ProPosing this tone � eature rePresentation here i s t o b e abl e to somehow caPture the general i zati on that ex i sts between weak M id and Mi d-Low nouns and vet sti l l account for their di fferi ng tonal shape i n Phrase .i n i t i al Posi t i on . If both can be rePresented as H I < weak Mid as H I and Mi d-Low as Hi) , then both can b. subject to Low sPreadi ng and high sPreadi ng as wal l as undergo the subragi ster l ow tone del eti on i n the appropri ate anvironments. If the underl Yi ng rePresentat i on for Mid-Low tone, Hi, can be l ex i cal l v spec i � i ed t o l i nk the subregister low tone i ndePendently to the segmental t i er , del ati on of the subre9i star Low tone wou l d automat i cal l Y del ate anY trace of Low tone o� the Mi d-Low contour because i ts source, the subregi ster Low tone wou l d hava been del eted . (58) Thi s i s i l l ustrated bel ow: CV � It was menti oned ear l i er i n ChaPter 4, that the Process of l inking the subregi ster Low to the segmental t i ar i ncluded Passing through the PrimarY register tiar . Thi s � Passing through ' wou l d ��ectivel v transform the Low tone i nto a pri marv regi ster Low tone: I -> L. (�9 ) CV 1\. H L V' I __ .•. - ..- ..---- 238 Let us ProPose the �ol l owing rul e : ( 60 ) INDEPENDENT SUBREGISTER LOW TONE LINKING. Li nk the subregi ster Low o � a marked Hl tone to the � i nal TBU of the morpheme when i t is i n phrase i n i tial Posi t i on . X I -> X '# ' -- r., H H V I 1 L 1 The appl i cati on o� the IndePendent Low Li nki ng rule on a sin91 e TBU i s a �air l Y simPl e Procedure: ' son ' ja -> ji 1 1\ H HL L" I 1 1 On a word with two or more TBU ' s . �urther l inking and del inki n9 rules are requi red , as shown below: kara , H -> kari -> karl 1\. HL I '" H L V 1 Il' 1 1 L L INKING LINKING Wi th Ri 9ht to �t l i nk i n9, HI ' meat ' LS DELINKING i s f i rst l i nked to the f i nal TBU, then the subre9ister 1 i s i ndependentlY l i nked to that same TBU. HI then l i nks to the fi rst TBU and � i nal l Y the contour simPl i �icati on Process of LS Del in kin9 takes Place and Low tone. HI del i nks from the f i nal TBU to leave it to the i ndependentl Y l i nked 239 Once thi s Independent Law Li n k i ng ru l e takes pl ace. then . Hl nouns can tri gger the Low tone spread i ng ru l e. as shewn bel ow: (61 ) wd kartl � -> watl kar� �� , � I L.t" 1 H L I H H L , 1 ' He saw meat ' H 1 LOW SPREAD LS DEL INKING Il Low Li nki ng d i d nat take Pl ace beler. Lew SPread. ne Law SPreadi ng waul d be al l ewed after Hl neuns. As a resu l t . Law Li n k i ng m it take Pl ace H rst belore Law tone Spreadi ng. The reader i s asked to reter to the subsection (vi ) on rule ordering lar lurther d iscussion. (v) We.k Mi d Tan. and F• •tur. Swi tch An observant reader wi l l have noted that whi l e Lh neuns tri ggered subregi ster High tene SPread i ng. i t has been suggested that weak Mi d (or HI ) tone i s al so responsib l e lor the behavi our that changes a lol l owin9 Lew tone t o Mi d tone. or a weak Mid tone to Hi 9h tone; yet there is no Hi9h tone on i ts subre9i ster tier to moti vate a Hi 9h tone Spreadin9 rul e . Thi s i s i l l ustrated bel ow in ( 62) . where the l i rst HI component at a complex noun in ci tati on lorm or in phrase i n i t i al Posi tion seems to be responsi b l e lor the rai si n9 ol the lol l owi n9 HI component to Hi 9h tone. (62) weak Mid CE nXE + Yaraxa -> c£nxE-�r 'x 3 I I H H I I I ' sauce i n9redients ' t One possi b l e but enti rel y unconventional sol ution i s to swi tch the t eatures 240 o� the � i rst component o� the complex noun , demoti ng H to h and bri nging I to the Pri marY t i er where i t gai ns the status o� L. Thi s i s not a total l Y crazY i dea �or a l anguage where HI and Lh tones are at the same pi tch l evel and therefore coul d easi l y be con�used by the speaker as � u l f i l l i ng some o� the same functions. Conceptual l y, the HI woul d become Lh when in phrase i ni ti al Posi t i on and then the Hi gh tone spreadi ng rul e wou l d gO i nto effect. (63) CEnl(f: -Yarsxa -> CE nXE -yar'x a.. / r I ( �, \J l 1 FEATURE SWITCH I , L H h I "" L-- ' sauce i ngredients ' H SPREADING Hi gh tone SPread onto Low tone nouns or verbs can al so take Pl ace i f the weak Mid noun undergoes the Feature Switch ru l e, as shown below . (64) Hi gh sPread CEnx£ / , �, 1 FEATURE SW I TCH t:l SPREAD &c LINKING AND LS DELINKING H SPREAD TO TBU It shoul d be noted that the same type of Probl em woul d surtace i f l owered Hi gh tones were speci f i ed as underlYingl y High tone as i n the earl i er anal Ysi se It an underl Ying High tone were in phrase i n i t i al Posi t i on and i t was tol l owed by a Low tone noun or verb , as seen i n (65) , it woul d be f i rst subject to Low Inserti on by virtue of the f act that i t i s i n Phrase i ni t i al Pos i t i on . Thi s Low Insertion woul d not i n i tself provi de the envi ronment f or n sPread i ng . However , it Feature Swi tch took pl ace, the envi ronment for High tone sPreading would be 241 created . (65) cinxi I H B I -> crnlCE B / I I I L H L I I I I L-INSERTION -> CEnlCE B / , ;, I ? - > crnxr 1 \ L I I FEATURE SW r', / h L ' L l. - ;-r ' CD H SPREAD. H SPREAD TO TBU Thar�are. re9ard l ess a� whether a w.ak Mid tone i s analYzed as under l Yi n9l y HI or High tone. the unconvention al �eature swi tchi ng rul e i s sti l l requi red for bath hYPotheses. The �eature swi tch rul e then maY be stated as fol l ows: (66) FEATURE SWITCH: Swi tch the �eatures of a HI tone to Lh when the morPheme to ,,;hi ch it: i s l in ked i s i n Phrase i n i t i al Posi t i on . HI - > Lh I*' _ (vi ) Rul. DrdlH"in9 It i s cl ear � rom the deri vati ons in subsecti ons ( i v) and (v) that bath IndePendent Low Li n k i ng (60) (�or Mi d-Low nouns) and Feature Swi tch (66) (for weak Mi d nouns) must occur b�are weak Mi d and Mi d-Low nouns can tri gger spreadi ng rules . Let us consider IndePendent Low Li nking f i rst . If Low L in king was ordered after Low tone SPread . as in (67a ) , � wou l d not be abl e to tri gger Low tone Spread . Once Low Spread is ascertai ned to be not aPpl i cab l e, Low Li n k i ng wou l d take pl ace creat i ng a Mi d-Low tone on �. Since Low SPread has al readY been tested for appl i cabi l i ty. the deri ved Low tone cannot tri ggRr Low tone Spreading onto the fol l owin g High tone verb . producing • . as a resu l t . an i ncorrect surface � orm. If . on the other hand . Low Li nking takas Pl ace f i rst (see 67b ) . a --- ---- -- --------- -- ��- 242 deri ved Low tone i s produced , whi ch. i n turn , Provides the environment �or the Low tane SPread rule. (67) a. wa� kara we' I r I '.I', ,� H H L V 1 1 LOW SPREAD N/A b. wa� karl "'" 1\ H L �,. 1 I H IND LOW LINK ING " he l ocked at meat ' ->*waa kar� wee H LOW L INKING. LINKING � LS DELINKING -> waa karl � -> N..1 H L V 1 wal karl wee ¥1H " he l ocked at meat ' H L H � 1 LINKING � LS DELINKING LOW SPREAD The same probl em woul d arise i � Feature Swi tch were erdered a�ter High tene sPread . A weak Mi d neun weul d net be abl e te tri gger Hi gh tone sPread wi thout Feature Swi tch . I� Feature Swi tch tack Pl ace a�terwards, then it weu l d be tee l ate fer the swi tched �eaturas to tri gger High tone Spreadi ng, as seen i n (68a) . The carrect sur�ace �orm can be deri ved rather . b y erder ing Feature Swi tch be�ore High tene SPread as seen in �b) . (68) a. wd clnxr t�d -> wal crnKI t�x � ->*wa� c£nx£ t�d I H I I I L 1 1 -H SPREAD N/A I rI ,� I I I I L , I h ' FEATURE SWITCH L I h �, / L I I LINKING FEATURE SW ITCH b. wal crnxl t�x i -) wa� c£ nxl t�K � ( H ,\ � 1/ I / -> I.. � �l! (!) L L\ H SPREAD. H SPREAD TO TBU LINK ING AND LS DELINKING 243 Thereiore. both Feature Switch and Law tone Li nking, bei ng Phrase i n i t i al rul e•• must take Pl ace before the aPpl i cati on of ei ther the Hi gh tone SPread rul e or the Law tone SPread rul e across the Phrase. There i s a questi on as to whether Law tone SPread and Hi gh tone SPread need to be ordered wi th respect to one another . Thi s i ssue as wel l as ather aspects of rul e order i n g wi l l be di scussed in Chapter 6 when mare data has been Presented . E. LQW=Higb Nouns Md as Dlllinkini The f inal set of nouns and pronouns which Produce a surface Mi d tone i n certain envi ronments are Law-Hi gh nouns. We have al readY seen that a Low-hi gh tone i s created as a resu l t oi a Law tone spreadi n9 onto a Hi9h tone TBU. si mPl i f i ed to a surface Mi d tone if the High tone was nat del i n ked . Thi s i s true Nhen the Low tone of a verb sPreads onto a Hi 9h tone i ncomPlet i ve suf f i x (69a) , whan a Low tone of a noun sPreads onto a si n9l e TBU Hi 9h tone verb (69b ) , and when the Low tone of a noun sPreads onto the High tone of the TYPe I I suH 1:< (69c ) . (69) a. wa4 xa t� e l a I It LH b. L c. M waa w4 l � l a L/1 pC-l<al o l/1 I .L H ' Di d he show i t? ' I ' Di d h e take some? ' M ' bodi es ' H M I t was al so d emonstrated that once Low tone Spread takes pl ace, the Hi gh tone of the resu l t i n9 Law-Hi gh contour i s del i n ked when iol lowed by a noun ---- . --- - --- - --- - --- ---- ---- ------ 244 cl ass cl i ti c , wh i ch i s Hi �h tone i n thi s particular envi ronment . Thi s tonal behaviour was l abel ed Right Si de Del i nking. The eXamP l es bel ow show that this i s true both for i ncomPl eti ve suf f i x es as wel l as for Hi 9h tone s ingl e TBU verbs. (70) wa� xa t.-� ��� C He showed i t f or him ' V'$ V L H waa w.l B 1/1 L H H ' He took same f or hi m ' ��� I H It has alreadY been mentioned that the f i rst Person Pronouns Possess a f i nal Mi d tone, (See subsecti on I I . C. ) and that thi s Mid tone not on l Y tri ggers subre9i ster High tone SPreading onto HI nouns, i t i s i tsel f l owered to Low tone in thi s environment (71a, b ) . The ex amPl es gi ven bel ow show that this i s the onl Y environment in wh i ch this Mid tone is l owered to Low tone. When Wia i s fal l owed by a Hi gh i n i t i al noun , however, i t i s not l owered t o Low tone ( f , g, h ) . Thi s l ac k of tonal change confor.s to the observat ion made ear l i er that Hi gh i n i t i al nouns are nat subject to tonal change and i n thi s case, i t aPPears they also do not parti c i Pate i n the change i nvol vi ng t h e l ower i ng of the Precedi ng Mi d tone. (71 ) a. HI i\de soxo -> i\d� s6x8 ' mY mortar ' b . Hl. i\d. k.r� -> Ad. k rS ' my meat ' c. L i\da m�B ' my rice ' d. LH "d. gbm, I B ' my granarY ' e. Lh i\da gbaxa ' my house ' f. H i\da faUxi ' my rock ' g. H-Lh i\da f yaa ' my fish ' h . I-L ' my H oor ' "da s�l � ----- -,- - --- ------- 245 Thi s l owerin9 of the Mid tone can be easi l y eKPlai ned i f ada. as wel l as � are Posi ted as under l Yi n9ly Low-Hi 9h . with an in ternal Law SPreadi ng rul e whi ch sPreads Law tone onto High tone resu l t i ng i n a Law-Mid surface tone. (72) �da ' my ' VI Wr'a " our ' V1 L H L H When fal l owed by a HI naun , the cond i t i ons for H igh tone SPread and subsequently for RS Del i n ki ng are met as shawn bel owe• (73 ) "da ka1-te --> �1 I L H I I 1h L H 1 I Ii I L L SPREAD " my meat ' I V1 H I kat-te--> "da H � QJ .. .... LINKING & RS DELINKING H SPREAD b . "da cln-ke-> �da c£n-ke-> �da cE n-ke-> �d� c€n-ke k'1 VI 1 I L H H L H H l h 1 l h @ I I I r LOW SPREAD 1.. .. 1 H IGH SPREAD va l VI I L H I I L H H I I I h LINK ING " my sauce ' H l h PI/A RS DELINKING An addi ti onal b i t of evi dence that the Mid tone o� "da is Low-Hi gh rather than Lh. for example, i s that "da does nat tri gger Hi gh tone SPread onto Law tone nouns or verbs as do Lh nouns and Pronouns. In the exampl e bel ow, the Lh Pronoun , �. can tri gger High SPread i ng ' onto the fal l owi ng Law tone noun whi le moti vati ng such a rul e for "da woul d Yield an i ncorrect surface farm. (74 ) mo 111:1 1:1-1)6 I L L. .... ... I I L H h ,- ' -- ---- ----- 246 ·i\da m::Jb-I')' VI L H , L. ... .... .... 1 I I " my rice ' L H h Recal l that the Hi gh SPreading rul e doe. not al low the sPreadi ng of the subregister Hi gh tone onto a Low tone i f that Hi gh tone is not al readY l i nked As to a Low tene. a resul t . the subre9i ster Hi gh cf a Lh noun can spread onto a Low tone. but the same of a Hh noun cannot . If the underlYi ng tone of � i s consi dered as Possessi ng a fi nal Hh . than its underlYi ng representation correctl y Predicts that i t cannot tri gg� Hi gh tone spreading anto LON tone nouns and verbs. SimPl e Law-High nouns in the i ndef i nite farm al ways consist of at l east 2 TBU ' s and as a result. the Low and the Hi gh tone are each l i nked to separate TBU ' . (75a> . However. when a Low-High noun root i . p�rt o� � COmPlex noun . i t maY have onl Y one TBU for both tone• • If fal l owed by a Low or Lh tone noun root. the Lew-High tone i s simPl i f i ed to Mid tone. as shewn bel ow i n (�a, b > . However , i f fol lowed by a HI tone. i t tri ggers Hi gh tone spread and i s subsequentl y subject to RS Del inking (c ) . / (�) a. gb�n-IS t " granarY-IND , L H b . gban-no?o-lo -> gbonno?c-I o It LH � " di rty granarY-IND ' L I h c . gbon-c�n-I')� A LH ------ --- " " geed granarY-I ND ' L - - -- ---- -- 247 ' an ol d granarY ' H SPREAD B. Cgnc1uwipn LINKING. of Mid tm. bwb.yi A'r L INSERTION & RS DELINKING po NQunw In subsecti on I I . e • • HYman ' s �our ProPosed underlYi ng rePresentations �or Mid tone. two o� whi ch use the double ti ered approach �or tone �eatures, were Presented. In the succeedi ng di 5Cu�si on , al l but one o� these rePresentations were ProPosed �or use i n Suci te. Mid tone as a de�au l t tone, to be l i nked to ton�less TBU ' s . has so �ar not been considered as a possi b l e waY to rePresent Mid tone i n Suci te. I t wou l d be d i � f i cu lt to consi der anY of the Mid tones di scussed above as si mPl y tonel ess. to be assigned a defau l t Mid tone at a particul ar stage of the der i vat i on . MY Presumpti on i s that tonel ess segments are not disPosed to tri9gering tunal rul .s, since they themsel ves do nat Possess anY tone. Yet, i t has been seen that those nouns wi th Lh or Law-hi gh tone do ��ecti vel Y tri gger the Hi gh tone Spreadi ng rule . I � ei ther one were consi dered tonel ess. i t wou l d have to acquire the Mi d tone be�ore the appl i cation � the rul esr even then , thi s d�ault Mi d tone wou l d not possess the Properties requi red �or trigger i ng High tone SPreading6• I� , on the other hand. one attempted to suggest that HI nouns were underlYi ngl y tonel ess. there wou l d b e no waY t o expl ai n how thi s tonel ess noun becomes Low-Hi gh when Pr�ceded by Law f i nal nouns. Thus it seems that each of the Mid tones di scussed here needs to be �ul l Y sPec i H ed for tone so that tonal rules can be moti vated to tri gger 248 predi ctab l e behaviour . I I I . THE DEFINITE SUFF I X REVISITED A concern that must be deal t wi th here i s the tone o� the De� i ni te Su�� i x . Recal l that the def i nite su� � i x i s M id tone a�ter Law f i nal and High tone noun roots and Hi gh tone a�ter Law-Hi gh and M id tone noun roots: ( 76) Noun root tone Hi gh fal '-ka ' the rock ' Low m� l a� ' the rice ' Mi d-Law kal-1R ' the meat ' Law-Hi gh nd�r.-ki ' th. yam ' Mid gba-ki ' the house ' Weak M i d cc n-ki ' the sauce ' The possible nature of the underl Ying form of the def i n i te suff i x was brief l y di scussed i n ChaPter 3. but wi thout the hel P of the subsequentl y acqui red knowledge of the d i f� erent possi b l e �eature conf i gurations of Mi d tone. The questi on was Posed at that ti me as to whether the de� i n i te suf f i x was underl Yi ngl Y Hi gh tone or same tYPe of Mi d tone, essenti al l Y the same prob l em posed for weak Mid and Mi d-Low nouns i n this chaPter . We know for sure that the def i ni te su� f i x is nat the unchangeab l e High tone belongi ng to a verY restri cted set of nouns. because i t does al ternate between M id and High tone. We also know that it has a f i nal f l oating Low tone which is never l i n ked to the suf f i x i tsel � but can tri gger Low tone spreadi ng. SuPPose. fi rst o� al l . that the ae� i ni te suff i x was underl Yingl y HI . Bath Law tone sPread and High tone SPread wou l d produce sati sfactorY resul ts i n the ------- . -- - ---- ------ ------- ------- ------ ------ 249 �ol l owing derivat i ons: (77) �. m�l.-�e -> m�l �-�. 1.. -'1 L H (L ) I I I I LOW SPREAD L./I L I I H (L) b . nd�r�-ke -> nd�r�-k' V1 L H I I I h LOW DELETI ON I L I (j) c . gba-ke -> 9ba-k' I ' the r i ce ' I H (L) I <D HIGH SPREAD va L H I I I ' the YAm ' H (L ) l h RS DELINKING C the house ' H (L ) ��� H IGH SPREAD After Hi 9h tone noun roots. the def i n i te su� � i x i s Mi d tone. A Hi gh tone SPread rul e, however. wou l d i ncorrectly Produce a High tone on the suf� i x , as shown below: (78) ��l '-ke -> *fa l a-k' I H I H (L) ,., ...., h CD HIGH SPREAD If an under l Yi ng HI tone i s adopted for the def i n i te suf f i x . then somehow the Hi gh SPread rul e must be bl ocked from aPpl Ying i n the setti ng as descri bed i n (78) . An al tern�ti ve sol ution i s to consi der the def i n ite suf � i x as underl Yingl y Hi gh tone. The same noun tYPes that Posed no probl ems �or deri vati on above i n (77) al so Pose n o Probl em here, as can b e i l l ustrated bel ow. Not• • however , ----- - -- _ . _ ._ - . _---_ . _ -- ---- ------ 250 that i � the su� � i x i s consi dered under l Y i ngl y High tone. there i s no l onger a need � or a High SPread rule. Thus. the noun i n ( c ) requires no rul es. RS DELINKING LOW SPREAD c. gba-ke -> gba-k' I I L H (L) h h I ' the house ' I Th e d� i n i te sut � i x as underl Yi ngl Y Hi gh tone. however . sti l l does not produce a sati s�actorY resul t �or Hi gh tone nouns. In the examPle bel ow. the de� i n i te su�t i x remains an i ncorrect Hi gh tone. (80) .t�l�-k' � I H H (L ) h h I I Al though posit i ng HI for the de� i ni te sut � i x woul d necessitate the b l oc k i ng at the Hi gh Spread rul e �rom High tone roots to the det i ni te suf f i x . wh i l e sti l l Permi tti ng i t with Low-High nouns. Posi t i ng an underlYing Hi gh tone wou l d require t h e t ormu l at i on o� a new rul e i n whi ch t h e Hi gh tone suf f i x i s l owered a�t.r Hi gh tone noun roots but not a�ter Low-Hi gh roots. Whatever the sol ution maY be. i t does apPear that we are deal i n g wi th an excePt ion that does not con�Drm n i cel y to the rul es al readY ProPosed thus far . For that reason . there does not seem to be anY strong argument f or choosing one sol uti on over another. 251 I� ane daes chaase ta Pasit an under l Yi ng Hi gh tane. hawever . there i s a waY of ordering the rules such that the � armu l atian a� a rul e ta lawer the Hi gh tane su� � i x ta Mi d tane a�ter a High tane naun wi l l nat appl y ta Law-Hi gh nouns. Let us � i rst �armul ate thi s Hi gh tane l awering rule to i nval ve the i nserti an a� a Law tone an the subregi ster tier a� the Hi gh tone su� � i x . Thi s rule. to b. cal l ed High tane DownsteP. i s stated bel ow. �e 1 ) HIGH TONE DOWNSTEP: Li nk a .subre9ister Low ,tane ta the High o� tene def i ni te suf f i x i f i t i s Preceded by a Hi gh tone l in ked ta a TBU. x I H - X I I h Thus the noun -> X - I X I H (Ll H H (L > [DEFl h I £ala-ke I I wau l d have the fal l owing deri vation : A Low-Hi gh noun. accordi ng the rul e � ormu l ation above. wou l d al so be subject to the Down.tep rul e . However . i t is alreadY known that the def i nite suf fix does not l ower a�ter Low-Hi gh nouns (see ( 79» . A sol uti on to thi s c.ontradi ction can be faund if RS Del i n king i s ardered to take Pl ace before DownsteP and i� a del i nked Hi gh tone i s b l acked from tri ggeri ng Downstep. Such a deri vation i s i l l ustra�ed bel ow. 2 52 ( 83) a. nd�r�-k� -> nd�r�-ke � I L H V L H H RS DEL I NK I NG I ' Yam ' H DOWNSTEP N/A Thus. Pos i ti n9 an under l Y i ng High tone for the def i n i te suff i x does effecti vel Y Provi de for a vi abl e anal Ysi s . If one posi ted an underl Yin9 HI tone , i t wou l d also b e possi b l e to Produce a correct surface f orm for Hi9h tones by stati ng si mpl y that there is an excePti on to the High tone SPreadi n9 ru l e when a Hi 9h tone noun ( and not Low-Hi 9h > i s fol l owed by a d ef i ni te suf f i x . For PerhaPs arbi trarY reasons. I have chosen the anal ysi s where the def i ni te suf f i x i s consi dered t o b e under l Yi n9l y Hi gh tone. IV- TONAL BEHAVIOUR OF THE CCJt'IPLEX NOUN The Previ ous secti on exami ned tonal i nteracti ons wi th i n the Noun Phrase. It was al so seen that the same tonal rul es that occur across word boundar i es al so occur across morPheme boundari es wi th i n the comPlex nou n . Th i s secti on focuses on the comPlex noun . Two tYPes of tonal al ternat i ons wi l l b� ex ami ned : f i rst . those wh i ch change as a resu l t of the tonal envi ronment . and second l Y. those whose tonal changes seem to be governed Pri mar i l y by semanti c and structural consi derati ons . The l atter wi l l be di scussed under the head i ng . ' Lex i cal Changes ' (see I V . C. > . A. Complex Nouns (HI + HI ) .nd the ObliiitorY contour Principle When HI and Hi no�n roots are strung together i n a comPl ex noun. i nteresti n9 tonal al ternati ons take Pl ace. Fi rst of al l , the rul es wh i ch occur across word boundar i es al so take Pl ace across morpheme boundar i es w i th i n a comP l ex noun . 253 • )f . for l?l(amPl�: the f i rst mOI"'Phl!II'Ie 0';: a comp l ex noun i s HI .. . .. < weak Mi d ) and the second morpheme is HI or Hl < weak Mid or Mi d-Low ) , Feature Swi tch takes Pl ace on the f i rst HI morPheme. The resu l t i n9 Lh tone then can sPread i ts subre9i s- ter Hi 9h tone onto the fol l owi n9 HI ton e , as shown i n the ex ampl es i n ( 84 ) . (S4 ) a . cE n-X E + Yara-xa -> cE nxaYa'rtix 8 I I H , �, I I (sauce thi n9s ) H .. ... , CD h I ' sauce i n9redi ents ' .... J ....,� L (" H I GH SPREAD & L I NK I NG FEATURE SW I TCH b . ntara+.j: ol :l -> ntara-fo l � -> ntara-f 6 1 5 -> ntara-.f6I :J' I H I I I I , 1 I I ( H ,\ \a I' / H H H h 1 L. ... .... 1 1' meju-Ie "" L " l and ch i ef ' " H � h H SPREAD FEATURE SW I TCH c . mmE + j u u --> t L .... � L I NK ING ' speec h ' ( vo i c e + SaYin9. words) Low tone SPread i n9 can al so take Pl ace across morpheme boundari es wi th i n a complex noun . I f the f i rst n ou n root i s a marked Hl noun root (or Mi d-Low ) , Low Li n k i n9 can take pl ace, fol l owed by Low tone SPread . as shown bel ow : (S5 ) a . nye?e + f ob /. H L V 1 I H I 1 I ND LOW L I NK I NG � :t�� \'�' I -> nYe?efol � -> nYe e N 1 ..... H L H 1 I V H L , I H ' ol der brother ' ( face + chi ef ) I L I NK I NG. & LS DEL INK I NG LOW SPREAD & LOW DELETION b . ca�a + nYe l t c . carya + t6 ----- --- - --- ---- < daY + eYe ) -> <daY + cover ( vb . » - ca��nyel � ' sun ' > canto�e ' umbrel l a ' -------- ---- ---- 254 There i s a restri cted set ot HI and Hi complex nouns . however . that behave a l i tt l e d i t t erentl y . I n stead ot the t i rst HI tri 9ger i n9 a Hi 9h or Low tone Spread i n9 rule onto the t o l l ow i n 9 HI noun root . the ent i re noun takes on a Mi dLow contour . Th i s means that the l ast TBU Possesses a Low tone wh i l e al l Preced i n9 TBU ' s are Mi d tone . Ei ther noun root maY be HI ( weak Mi d > or Hi (Mid- Low ) . Yet the resu l ti ng tonal structure is the same. Observe the ex amPl es bel ow. < S6 ) a. ceewil + PoB Hi Hi + -- > c i PoB <woman + husban d ) ' 9room ' - > nyambal axe <shade + un i t > ' n i 9ht ' Yataxe < th i n9 + tree > ' nec k ' Hi HI n o examPl e� b . nYi lll£ + bal e + x e Hi HI c . Yarax a + taxa HI -> HI d. nYamE + bal e + wa + 10 -> nyabal awol d <shade + un i t + b l ac k ) < dar kness ' HI Hl. Let us consi der H rst (S6 a ) . The noun c:ipgB i s c:omF'osed ot two Hi nouns . S i n c:e these are mar ked tor I ndependent Low Li n k i n9. i t mi 9ht have been ex pected that t i rst no�n root wou l d under90 Low l i n k i ng wh i l e the sec:ond wou l d be sub jec:t to Low tone sPr ead . as we saw i n ( 85 ) above. However . in th i s case . onl Y the sec:ond noun root undergoes I ndependent Low Li n k i ng . wh i l e the f i rst remains HI : (86) a . c:i -PoB I '.1\ i 1/ l. H I H L 255 APParent l y some as yet unsPeci f i ed ru l e or cond i t i on must take Pl ace to b l oc k the l i n k i n9 o f the f i rst subre9i ster Low tone t o t h e se9mental t i er . I n (86c ) . we see a comb i nat i on of two HI noun roots, nei ther of wh i ch are mar ked f or IndePendent Low Li n k i ng. and Yet the f i nal noun root i s sub j ect to I ndependent Low Li n k i n 9 . (96 ) c . Ya - tax� I H I I \�\ H V I L A9ai n . i t seems aPParent that - some rul e Precedi n9 Feature Swi t c h . wh i ch wou l d normal l y appl y i n th i s set t i n9. must tri 9ger the Low Li n k i ng on certai n compl ex nouns consi st i n9 ot two HI nouns . As w e be9i n t h i s anal Ysi s. i t must b e emPhasi zed t h at n o t al l HI comPl ex nouns behave i n th i s waY and those wh i ch do seem to d o so f or unknown reasons· Al though there aPPenr to be no Phonol 09i cal or syntact i c reasons f or th i s behavi our . the f cl l owi n9 i s an attempt to provi de a anal yt i cal descr i pt i on for thi s pecul i ar t YPe of tonal behavi our. 1 . The OCP Sol ution One possi b l e sol ut i on i s tn i nt.roduce the Ob ! i 9a t�rY Contour Pr i nc i p l e as a l an9uage sPec i f i c rule that wou l d aPpl y on l Y to th i s desi gnated set of HI comPl ex nouns· The OCP wou l d col l aPse a seri es at HI tones i nto one . and as a resu l t of that col l aPsi ng Process , the suc�agi ster Low tone wou l d be marked t ur IndePendent Low tone Li n k i n9 . Th i s i s i l l ustrated i n the examPl e gi ven bel ow . 256 ( 87 ) Ya I + ta:<a -> Yata:<� - > yataxe V I H H I 1 I I H � H L l. l. , OCP ( th i ng + tree ) ' neck ' 1/ IND LOW L I NK. L I NK I NG . AND LS DELI NK ING As a ru l e i n Suc i te. the OCP maY be stated as fol l ows: HI ( 8S) OBLI GATORY CONTOUR PRINCIPLE: A seri es o� def i ned un i t are col l apsed i nto one H I I -- - n > HI tones with i n a semant i cal l Y tone. H I 1 Th i s OCP ru l e wou l d not work on those comP l ex nouns whose t i rst comPonents tri gger rul es onto fol l owi ng comPonents as seen above i n ( 84) and <8S) . If these nouns submi Hed to the C:CP, i ncorrect T orms woul d resul t . as shown bel ow . (89) nYe?� + foB ( face + ch i ef ) ->*nye?E f o l � / H / H l. l. I I I/i H ,/ V l. L ' ol der brother ' Rather : nYe?etol :9' < See (85b ) for correct f orm) Theretore . the OCP must not be al l owed to appl y everYwhere . Each noun t o 2 . Rul e Ordering �or the OCP Those noun comPounds wh i ch do undergo OCP d o not al l ow Feature Swi tch or Low Li n k i ng on the f i rst components of the comPoun d . Rather OCP goes i nto eff ect and then i ndependent Low tone Li n k i ng aPpl i es on l Y at the end of the word • .- -- -----------_. _------ ( 90) I nYi m-bal a-wo l o I I H H H I 1 1 , -> nYi m-bal a-wol o -> nYi m-bal a-wo l o -> nY � H , I I � H L I 1." 1 '" I -"!. .. H -1, .- ... 1 / j - L L I NK I NG � LS DEL I NK I ND L L INK ING OCP � 257 There�ore. the OCP must aPpl y before the aPpl i cat i on of the Feature Swi tch or the I n dependent Low Li n k i n9 Ru l e . If e i t her appl i ed betore the OCP . they woul d adjust the tone features i n suc h a waY that OCP woul d no l onger b e Poss i b l e . The OCP. o n t h e other han d . must aPpl y obl i 9ator i l y i n thi s set t i n 9 . and theretore must not b e b l ocked under anY cond i t i on by Feature Swi tch or Independent Low tone Li n k i n9. In add i t i o n , i t i s seen i n (90 ) that the ru l e at Independent Low Li n k i n 9 i s requ i red atter the appl i cat i on of the OCP. 3. Hi 9h SPr••d i ng �d OCP When Hi 9h tone SPread i n 9 takes Pl ace on HI compl e:< nouns. they al l end uP wi th the same resu l ts re9ard l ess at whether they were subject to OCP or not . Hi 9h Spreadi n 9 on ComPl ex nouns wh i ch do not under90 the OCP i s i l l ustrated bel ow . In the e:<ampl e bel ow . the subre9i ster Hi 9h tone sPreads onto the tol l owi n9 HI tone. The same Hi 9h tone then sPreads to the nex t HI tone. ( 91 ) wu Ya wu CE n:<, -Yaraxa suu I I I H H H CD I L - h -1 H I GH SPREAD I -> " .J '� wu Ya wu c€ nx€ -��ax� suil? I I� - - - -I h CD H H H HIGH SPREAD & L I NK I NG Nouns whi ch have under90ne OCP are al so subj ect to Hi 9h tone SPreadi n9 and produce the same resu l t as those nouns wh i ch d i d not undergo the OCP. as shown bel ow : ---- . . -- � _ . _ _. _-- ------ -- ---- ------- ------- 258 ( 92) mo Yai:ake -> mo yatake -> mo yatake I "'-J I , L H H L H H L H H H &.. ... - .., I I ' I I V I I I I I I h I 1 1 h GI l h H I GH SPREAD & DOWNSTEP OCP 4. LaM SPr.�i ng ."d the OCP Non OCP compl ex nouns al so behave i n the same waY as OCP comPl ex nouns when subjec t to the Low tone SPreadi ng rul e . F i rst . l et us l ook at nouns wh i ch do not undergo the OCP. I n the examP l e bel ow . the � i rst comPonent of the comPl ex noun wi l l . i n i sol at i on . undergo Feature Swi tch . and then tri 9ger High tone spread onto the f o l l owi n9 HI comPonent . ( 93 ) a . me-juune-> me-juune-> me-jCi6ne I I �J'I rI I H , L H h I I 'J I H H L 1/"1 I I I 1. ... .. 1 H 1 I h L I NKING & DOWNSTEP H SPREAD FEATURE SW ' speech ' When preceded by a Low f i nal word ( as i n ( 94) . however . the f i rst comPonent i s n o l onger i n Phrase f i nal Pos i t i on and thus cannot under90 Feature Swi tch . Rather i t i s sub j ect to Low tone SPread . A Low-Hi gh contour resul t s . whi ch . after b e i ng spec i f i ed on the subre9i ster t i er . then tri 9gers High tone sPread- i n9 onto the fol l owi ng comPonent . Once the SPreadi ng ru l e takes pl ace . the contour si mpl i f i cat i on Process of RS Del i nk i n9. del i n ks the Hi gh tone of the Low-Hi gh contour . wura mej uune->wura m�jcicine-> wura m�jciUne I ... 1 V$ '� L H H H L H H H L H H H I 1,.......1 L,../ , h h ( 94 ) 1.. I I I t...--1 <D (1) 1 L SPREAD & L DEL ---_ .. __ .- - '- '--- I H SPREAD ----- I CD 1 L I NK I NG . RS DEL I NK I NG. &c DOWNSTEP 259 AI�hoU9h comPl ex nouns wh i ch under90 the DCP Y i el d the same resu l �s when subject to Low tone sPread i n9 , �he manner i n wh i ch those resu l ts are arr i ved at i s a l i t t l e d i f f eren t . Dnce the DCP has col l apsed al l the HI �ones i n�o one (as shown i n ( 95 » . �he Low t on e sPreads onto the f i rst Hi 9h l i n ked TBU and Low tone dei et i on takes p l ace. S i nce t h i s subre9i ster Low tone is l i n ked to the H i 9h tone wh i ch i s in turn l i n ked to al l �he comPonents of the comPl ex nou n , al l the component s are aff ected b y i ts del eti on and , a s a resu l t . become Hi 9h tone. Th i s i s u n l i ke the case tor non DCP comPl ex nouns where Hi 9h t one sPread mus� be mot i vated to �ri 9ger Low tone del et i on on the rema i n i ng comPonents Ot the comPl ex noun ( see 94) . ( 95) k�n-Y� ka-nkur� -) kanY� ka-"k urc - ) kcinY� k�-"kCirEl -> kanY� I I L , H I I H I L "'./ H I 1 I I I I I I " , -� H I L I L SPREAD & L DEL DCP • H ve vi 1 1 ages ' (vi l l ages + f i ve ) s�m� sak a - > s�m� s�ka I L I I H H I I I L V H I DCP " f at 90at ' -> L� I I I h L I NK I NG & RS DELI NK I NG (; ' V s�m� s�ka L I I I cb kci-"kurEl H I I L SPREAD & L DEL L I NK I NG & RS DELI NK I NG ( o i l + 90at ) s. OCP .nd the Associ .tion Conventi ons In Chapter 3 . i t was assumed that the Assoc i at i on Convent i on s i nvol ved the l i nk i ng of al l tones and TBU ' s . In Chapter 4 . the search f or a sat i stactorY anal Ysi s tor Low tone SPread brought to l i 9ht the need to mod i f y the 260 Assoc i at i on Conven t i ons such that 1 ) tones assoc i ated w i th TBU ' s i n a Ri ght to Lef� fasn i on across �he word and 2 ) �he i ni t i al l i n ki ng did not i nvol ve the l i n k i ng ot more than one TBU to a tone. Wi th the i ntroducti on of OCP . one sees agai n the need to PostPone the l i n k i ng ot tone to al l TaU ' s unt i l Low tone sPread i ng takes Pl ace. Thi s becomes c l ear when comPar i ng Low tone SPread on a s i mpl e three TBU noun w i th Low Spread on a complex noun wi th two or three noun roots . O n words wi th onl Y one n o u n root . the Low tone spread$ to t h e f i nal TBU of that root . nU k� �x5 1/..:-- - 1 L H , I On words wi th more than one n oun root . the Low tone sPreads to the t i nal TBU ot the t i rst morpheme on l Y . n�� Ya-Ux@ j.. � �",,/ " man ' s n ec k ' H , 1 L The waY the ru l es have been set uP thus t ar al l ow tor th i s correct Predi ct i on i n behavi our . I t . t or examp l e . al l TaU ' s were l i n ked to tones bet ore the appl i cat i on of the tonal ru l es . the appl i cat i on of the OCP woul d i mmed i atel Y create an amb i 9u i ty between t h e word w i th a si n9l e noun root versus one wi th two or more noun roots. as shown i n (96) . to al l TBU ' s . f i rst TBU. In b oth cases . tone woul d be l i n ked When Low SPread i n9 takes pl ace. Low wou l d SPread on l Y to the There i s noth i n9 i n the l i n k i n9 Pattern to i ndi cate that Low must a l $o sPread to the second TBU i f the noun concerned is a si mpl e three TBU noun . but must not spread a second t i me i f the second TBU i s another noun root . Thi s 26 1 amb i ';u i ty created by such a l i n k i n'; Procedure makes i t i mPossi b l e for the Low tone SPread i ng ru l e to d i st i n9ui sh t h ese two tYPes and thus treat them d i ff eren t l y . (96) mu l t i Pl e noun roots n.\.\ '11 L ��",!x� si ngl e noun root ' man ' s nec k ' 'V n�.\ k��:<:, \J, '" � H Pul l eYb l an k ' s I 1 H L I 1 ( 1 983 ) proposal to l i mi t the Assoc i at i on Conven t i ons to l i n k a tone to on l Y one TBU serves a usefu l PurPose h ere. The eHed of l i n k i ng one tone t o one TBU s i 9nal s . ex cept i n the cases of under l Yi ng contour tones . the Presence of a si n9l e noun root . I f . after the Assoc i at i on Conven t i on s . the aPpl i cati on of the OCP resu l ts i n a si n9l e tone bei n9 l i n ked to several TBU ' s . we are thus i nf ormed that there are as manY noun root s as l i n kages to the tone. The Low t one sPreadi n9 rul e . then . is l i mi ted t o sPread i ng on lY to the f i rst HI tone l i n ked TBU. On a si n9l e noun root wi th three TBU ' s. thi s means that the Low tone wi l l sPread to the f i nal TBU, si nce t h at is the f i rst ( and on l y ) TBU that i s l i nked by the HI tone . On the noun wi t h two noun roots. the Low tone wi l l sPread on l Y to the f i rst TBU that is l i n ked . In (97 ) bel ow. is an i l l us- trati on of how the tonal rul es i nteract to Produce correct f orms f or ( a ) a three TBU Singl e root noun and ( b ) a comPound noun . (97) a. s i ngl e root ' hi s EMPH i nheri tance ' CVCVCV f H L I 1 A C -> CVCVCV / 1 �,. ... '" L H f 1 L SPREAD -> CVCVCV � L -> I d) H 1 DELET I ON � L H LS DEL I NK I NG 262 b . mul t i Pl e root wura n Yi m-bal a-wo l o -> wura nY� mbal aw61 6 cv-cv-cv -} cv-cv-cv I I I '-..V L H H H H L I I I I A c I I I 1 -> cv-cv-cv �' � L OCP I 1 H ' h i s EMPH darkness ' -> cv-cv-cv L L SPREAD � -> H I J� L H @ I -DELET ION RS DEL INK I NG The � i nal quest i on that remains i s whether the aPpl i cati on o� OCP i s real l y needed t o al l ow a correct tonal outpu t . I n terms Ot SPread i n9 r u l es . i t appears that the aPpl i cat i on o� the oCP i s not essen t i al �or produci ng correct resu l ts . Whether a noun has undergone oCP o r not . both H i 9h and Low SPread i ng ont o a comPound noun produces the same resu l ts . However . we recal l that oCP was i n troduced i n an ef f ort to deal w i t h a sel ect set of c omPl ex nouns wh i ch behaved l i ke Si mpl e non-compl ex nouns i n phrase i n i t i a l Posi t i o n . The nonaPpl i cat i on of the OCP i n thi s situat i on wou l d have requ i red at l east some other ki n d of rul e or cond i t i on to a rul e to b l oc k the appl i cati on of Feature Swi tch and I ndePendent Low Li n k i ng in al l Posi t i ons ex cept in word t i nal Posi ti on . The b l oc k i ng of these two ru l es word i nternal l y wou l d have to be restri cted to t h i s speci al set ot nouns. however . f or we have al readY seen how both rul es can be appl i ed word i nternal l y on other compl ex noun s . Such a condi t i on wou l d b e feasi b l e i f i t were sU9gested that Feature Swi tch and Independent Low tone Li n k i ng were sensi t i ve to the cohesi veness of a par t i cu l ar comPl ex noun · I f a compl ex noun functi oned as a semant i c uni t . then FS and Low Li n k i ng maY be b l ocked �rom aPpl Ying. whi l e they sti l l aPpl y to the end o� each morPheme w i th i n compl e:< nouns wh i ch were l ess cohesi ve i n nature. We see thi s l ac k of FS and Low Li n k i n g in another setti n9 where the OCP ----- ----- --- ----------- 263 cannot be moti vated . but where a con d i t i on b l oc k i n9 the two ru l es cou l d be aPPl i ed . When a H I noun root is f ol l owed b y a Lh noun root wi t h i n certai n compl ex nouns . nei ther FS nor Low Li n k i n9 are al l owed to appl y. (98) a. son:d + na�a -> sonna�a / / H , L I 1 h - > tasal')e b . tt: :( � + son I H I 1 ' horse ' s tai l ' I ( Pl ace + sacr i f i ce ) ' al tar f or sacr i f i ce ' L I h c · saka + nYe7e -> sakanye?ere < sheeP + be al ot ) ' al ot of sheeP ' V H , 1 I L I h Therefore, i t maY b e feasi b l e to attach a con d i ti on to FS and Low Li n k i n'; . b l oc k i n9 thei r appl i cat i on i n certai n word i nternal envi ronments. i nstead of ProPosi ng the OCP rul e , wh i ch does not cover f or the exampl es 9i ven above i n (98) . Unf ortunatel Y , t i me and sPace d o not al l ow f or a thorou9h i nvesti 9at i on of � h i s Possi b i l i �Y . B. Hi9h De1iokio9 Earl i er . i t was Poi nted out that al thou9h HI (weak Mi d ) usual l Y behaves d i f f erentl y than Lh . there are si tuat i ons where i t tri 9gers the same tonal r u l es as Lh nouns. I n order to accommodate this behavi our . a rul e cal l ed Feature Swi tch was i mPl emented to change HI to Lh . Th i s Process d i d not chan ge the surface tone; i t s i mpl y changed the under l Yi n'; tone so that i t cou l d ----- -- _ .. _- --- ----- 264 tri ';';er the same rul e as Lh nouns. It was suggested that t h i s conf usi on of i dent i ty f or these d i f f erent tYPes of M i d tones was due to the f act that thei r i dent i cal pi tch val ue contri buted to the over l apP i n9 of tonal b ehavi our . The fol l ow i n g is a d i scussi on of a tYPe of Mi d tone that apPears to be Lh i n certai n setti ngs , whi l e in other envi ronments, i t behaves more l i ke a HI or a Low-Hi gh ton e . Th i s tYPe o f Mi d tone i s n ever f ound i n word i n i t i al Posi t i on and i t i s al ways preceded by a Low ton e . tone ( Lh ) . i t i s Mi d tone. When f o l l owed by no tone o r by a Low o r Mi d When fol l owed b� a HI tone or H tone. it i s Low ton e . The under l i ned vowel s i n the eKamPl es bel ow are l i n ked to these Part i cul ar tYPes of Mi d tone . ( 100) k�-l a-xa ' sorghum- I ND ' k�-j ti.-K e k�-n tLL-)( O " wood- I ND ' " bac k - I ND ' k�-l A-p6-l � " bund l es of sorghum- IND ' k�-l a-P�?al o " bundl es of sor9hum-DEF ' k�-j l.-p6-1 � " bund l es of wood-I ND ' k�-ji..-pi:}-?al o ' bund l es of wood-DEF ' k�-nt i.l.-Yi\-n3 ' back-ache- IND ' Th i s tyPe of behavi our Paral l el s verY cl osel Y the tonal behavi our of LowHi gh contour tones whi ch are l i n ked to a s i n9l e TBU ( see pp. 243ff ) . If these par t i cu l ar Mi d tones are consi dered under l Y i n9l v Low-Hi 9h . then one can mot i vat e the Ri 9hts i d e Del i n k i ng ru l e . wh i ch del i n k s a H i gh tone of a Low-Hi 9h c.ontour when i t i s f ol l owed by another Hi gh tone. Such a der i vat i on i s i l l us-· trated bel ow. ( 101 ) k� 1 a p6l S -> k� l a p6l � / L I 1 /1 / H I L.- "'1 LH 1h 1 H I GH SPREAD I A L LH I H I I� I lh RS DEL I NKI NG 265 However , i f the segment =La i s under l Yi ngl y Low-Hi gh . the two tones shou ld l i n k to separate TBU ' s when f o l l owed by an i ndef i n i te suf f i x . as shown bel ow . Unfortunatel y, t h i s produces the wrong surface t orm. ( 1 02 ) *ka-U-x-' I I I L L H I I I 1 ' sorghum ' Rather : ka-I a-xa I h In addi t i on , kAlaxa tri ggers Hi gh Spreadi n g onto Low tone nouns and verbs, someth i ng a Hh f i nal noun cannot do. I f one Pos i ts an under l Y Lh tone for =la, then i ts subre9i ster H i 9h tone can spread onto a f o l l ow i n9 Low tone, as shown bel ow. ( l 03) wu Ya ka-Iaxa tea I / A L L LH 1 h Oh I ' He i s showing sorghum ' I.. , 1 , H I GH SPREAD If we consi der that the under l Ying tone i s Lh . however , some eXPl anati on must be found f or the l ower i ng of th i s Lh tone of =la to Low i n kalap61$. Another ru l e wi l l have to be f ormul ated wherebY the Lh tone becomes Low tone when preceded by a Low tone and fol l owed by a Hi gh tone. A si mpl e waY of perf ormi n9 the mechan i cs of th i s tone l c�er i ng i s to dsl i n k tha subre9i ster Hi gh tone i n t h i s envi r on ment : ( 1 04 ) kal a p61 S I I / L L H 1 h h I -$ I Such a r u l e wou l d be stated as fol l ows = 266 ( 105) H I GH DEL I NKING: Del i n k a subre9i ster Hi 9h tone l i n ked to a Low tone i f i t i s Preceded b y a L l tone and tol l owed by a Hh tone. wi th i n the doma i n of the word . x I L I I X 1 L X h h � I I H The reader maY recal l that an ear l i er H i 9h Del i n k i n9 rul e ( 44) was i ntroduced in Chapter 4 to mot i vate the l ower i ng of a Mi d tone verb after a Hi 9h f i nal noun and i n a Phrase f i nal Pos i t i on . Because these two Hi gh Del i n k i n9 ru l es take Pl ace i n d i f t erent envi ronments . they shal l be consi dered separate ru l es at t h i s Poi n t . Th i s Process i s qui te remi ni scent o f the R S Del i n k i n9 of a Low-Hi 9h contour . ex cept i n t h i s case. the Lh i s not a contour but rather a comPl ex bund l e of f eatures . PerhaPs. thou9h . the mot i vati on f or this tYPe of rul e i s si mi l ar to that of RS Del i n k i n9 . RS Del i n k i n9 was concerned wi th the s i mpl i f i cat i on of contour tones . Th i s H i gh tone Del i nk i n9 seems to be moti vated by some k i nd of desi re to s i mpl i f y the compl ex Lh f eature i n an envi ronment where the Lh tone i s caught between a Low and a Hi gh ton e . Li ke RS Del i n k i n9. the Hi 9h Del i n k i n9 ru l e al so t.�kes pl ace atter the appl i cati on of the spread i n9 r u l es. as shown below. ( 1 06) a. k�l a + pol 3 1 I H L L 1 .1 I h I I I -> kal a - pob I \ L 1 I L .... H � "" "l h (!) H SPREAD (sor9hum + t i e . n . ) I -> kal a-p61 � ' bund l e of sor9hum head s ' I I L L ' -J ,� I H h LINK I NG & H DEL I NKING 267 b . v�nd i �e + k E E x � (shi rt + arm) -) v�n d i - k € € x t ' sl eeve ' A si mi l ar Phenomenon occurs on nouns wi th TYPe I I suf f i x es . When a noun ends i n a TYPe I I i ndef i n i te suff i x . the f i nal tone i s Lh , and i t can tri 9ger Hi gh tone sPread onto t o l l owi ng words, as shown i n the deri vati on b el ow. ( 107) nk�n-?al a ti -> n k�n-?al a h -) n k�n-?al a tl � L......-1 H L L L H L L L L H L I , l !r I h h I I h h I h h (.. ... 1 I I I I I I LOW SPREAD I V-1 I I I .... ... , (D H I GH SPREAD ' show teeth ' h� HISH SPREAD TO TBU However , when a noun w i th a TYPe I I suf f i x i s compl eted by a Def i ni te Sut t i l( , the TYPe I I suf f i x b ecomes Low tone: nk�n-7a-k! -) n k�n-?�-k ! ' the teeth ' , It wou l d aPPear that Hi 9h Del i n k i n9 is al so taki n9 P l ace here . However . i n observi ng the examPl e bel ow i n ( 1 08) , the under l Y i n9 tonal structure ot the TYPe I I suf f i x Presents a sl i 9h t l y more comPl i cated situat i on . Atter the Low tone �f the iiOUii root sPreads oiito the Hi gh tone of the TYPe I I suff i x , a. talal Ot three tones are l i n ked to the singl e TBU , =la. ( 1 08 ) n k�n-7a- k ! 1.. - � I L I 1 ' the teeth ' HL H " I hh h LOW SPREAD In order to moti vate the l ower i n 9 of al l these tones to Low tone, i t seems that both RS Del i n k i n� and Hi gh Del i n k i n9 wou l d have to take P l ac e . However , Hh i s bl ocked tram Del i n k i n9 because i t i s not f ol l owed by a Hi 9h tone but rather by a Lh tone. and the subre9i ster Hi gh tone of the Lh tone cannot be del i n ked ----- . .- � . - .- ..--- 268 because i t i s not Preceded by a Low tone. One possi b l e sol uti on i s to somehow col l apse thi s comPl ex of tones. LI -Hh-Lh i nto a si mPl e . more manageabl e combi nat i on such as Lh . Unfortunatel Y t i me and sPace do not al l ow a di scussi on of the theoret i cal i mPl i cat i ons of such a move. However . such a Process wou l d Permi t t h e Hi 9h Del i n k i n9 r u l e to take Pl ace. resu l t i n9 i n a correct surface rePresentat i on . as shown bel ow . ( 1 09) n kan-?a-ki -> n kan-?a-k! -> nkan-?a-k ! L. ... ...� L HL I I t..-1 L H L hh h 1 h h I I LOW SPREAD I I I Tone Col l aPse I l--1 L II I 1 I H L H h h * I H I GH DEL I NKING One m i gh t t h i n k the Lh tone of kal a- wou l d al so under90 H Del i nk i n9 when fol l owed by the Hi 9h tone DeH n i te Suf f i x . However. as t h e -F i rst e:< ampl e i n ( 1 10) i n d i cates . no Hi gh Del i n ki n9 occurs. I t seems then . that t h i s Hi gh Del i nk i n9 occurs before def i n i te suf f i x es on l Y on Type I I suf f i x es . ComPare ( a ) wi th ( b ) bel ow . ( 1 10) TYPe I I SuH i x TYPe I Suf f i x a. k a l a-ke / L I L I I " th e sorghum ' H vs . b . n ka?-a-k ! -> n k a?-a-k! I L h I L I I " the teeth ' H h The Pri marY d i ff erence b etween these two nouns i s that the noun wh i ch does not undergo H i gh Del i n k i n9 i s a TYPe I noun. whi l e the second i s a TYPe I I noun . Other examPl es conf i rm that def i n i te suf f i x ed TYPe I nouns d o not al l ow Hi 9h Del i n k i n9 i mmed i atel Y Preced i ng the def i n i te suf f i x . i n contrast to TYPe ----, - _ . - - -- -.--- ----- 269 I I noun s . The ru l e of H i 9h Del i n k i n9 ( 1 05) . then i s f ormul ated i n t h e attemPt t o account f or t h e 1 0wer i n9 o f a Lh t one si tuated b etween a Low tone and a Hi 9h tone i n a compl ex nou n . The d i scussi on above ex Poses some compl i cati ons i nvol ved i n i ts aPpl i cati on . At th i s Poi nt . however . a more thorou9h anal Ysi s wi l l have to wai t unt i l more research i s don e . C. LRxiC.l Cbani•• The Previ ous t wo tonal a l t ernat i ons. the DC? and H i 9h Del i n k i n9 . took P l ace in a par t i cu l ar tonal envi ronmen t . However . i t was n oted that somet i mes their aPpl i cati on was b l oc ked f or reasons not al ways c l ear . Th i s sec t i on deal s wi th tonal al t ernati ons wh i ch I cal l l ex i cal changes. These tonal changes oPerate at the l ex i cal l evel ( i . e. wi th i n word bo��dar i es ) and are. f or the most Par t . not 90verned by the t one of ad j acent morPhemes. but take Pl ace after noun roots are comb i ned to f orm comPl ex nouns. The f ol l owi n 9 d i scussi on wi l l deal wi th each of these changes and ProPose the waY i n wh i ch these changes maY be formu l ated . 1 . High Del eti on Hi gh Del et i on takes Pl ace i n two tYPes of envi ronmen t s . The f i rst takes Pl ace i n an envi r onment that i s both t onal and morPhol ogi cal . wh i l e the second seems to be a functi on of the seman t i c and morphol 09i c �1 structure of the noun . For the sake of conveni ence . they shal l be l abel ed Hi gh Del eti on I and H i gh Del eti on I I . a. High Del .ti on I In Chapter 2 . i t was men t i oned that Mi d tone verbs l owered to Low tone when an i ncomP l et i ve suff i x was added . The ProPosed anal Ysi s posi ted the Mi d tone of ------,-,- - ---- ------- ------- 270 a Mi d tone verb as under l Y i n9 l y L h . A Hi 9h Del et i on ru l e was then �ormu l ated to tri 9ger del et i on o� the subre9i ster Hi 9h tone of the Lh verb when � o l l owed by a Hi 9h tone i ncom�l eti ve su� � i x ( See (59» . In Chapter 3 . i t was al so �ound that certai n tonal sets o� nouns al so l owered to Low tone when the Hi 9h i n i t i al TYPe I I i nde� i ni te su�f i x was added . Al thou9h . at that Poi nt . the under l Yi n9 features o� the vari ous types of Mi d tones had not been di scusse d , i t was sU9gested that Hi 9h Del eti on cou l d aPply here , too . The Hi 9h Del et i on ru l e as ProPosed i n ChaPter 3 i s restated bel ow: (SS) HIGH DELET I ON: Del ete a Hi 9h tone on the Pri marY or subre9i ster t i er it a Pri marY Low tone i s � ound to the l e� t o� a Pr i marY Hi 9h tone or l i n ked to a subre9i ster Hi9h ton e . and f ol l owed by a Hi 9h i n i t ial su�f i x . I� the suf � i x i s a nom i nal suf f i x . i t must have under90ne ei ther n o se9mental al t erat i on o r el se have under90ne Sut� i x Reduct i on . {H. h > -> [sut t i x - it N . then f u l l or reduced ] Thi s r u l e assumes that al l Mi d t ones were under l Y i n9 l y Lh or Low-Hi gh . However . now that HI tone has been posi ted for weak Mi d and Mi d-Low t ones , Hi 9h Del et i on . as �ormu l at ed above , wou l d not del E�te the Hi 9h tone . wh i ch i s now on the �ri marY t i er . One therefore wi l l h ave to mod i f y the rul e t o al l ow f or the del eti on anY t i me that Hi 9h t one i s combi ned wi th and l i nked to the same TBU as a Low ton e : ----- ,- - "-- ---- 27 1 ( 1 1 1 ) H I GH DELETI ON : Del et e a Hi9h tone on the Pri marY or subre9i ster t i er i f a Low tone on the Pri marY or subre9ister t i er i s l i nked to the same TBU as the Hi 9h ton e . I f the suf f i x is a nomi nal suff i x . i t must have under90ne ei ther no se9mental alterat i on or e l se have under90ne Suff i x Reduc t i on . [suf f i x - i f N . then f u l l or reduced ] Th i s wou l d mean t h at when the Hi 9h of a HI tone i s del eted . the remai n i n9 Low tone on the subre9i ster t i er wou l d l i n k to the se9mental t i er . Th i s i s i l l usIn ( 1 12b > . the Low-Hi 9h tone i s al so del eted through trated b e l ow in ( 1 12a> . the same Hi gh Del et i on ru l e . ( 1 12) a . n i -i I � � - > n � -i �� -> H L L HL 1 h h 1 hI , , H DELETION n i -i I II L SPREAD ' mothers ' b . nkan-?a l a -) n k �n-?ala ' teeth ' A � lh , I I H L L I I I h h HIGH DELET I ON 1 I I H L I I h h LOW SPREAD In these examp l e s . once Hi gh Del et i on takes Pl ace. t h e remai n i n 9 Low tone then sPreads onto the H i gh tone suf f i x . creati ng a Low-Hi gh contour . wh i ch is then si mpl i f i ed to Mi d ton e . If a HI morPheme marked for Hi 9h del et i on i s Preceded b y a L I tone o r a subre9i ster h i gh f i nal word . one mi ght expect Low or H i 9h tone sPread i ng to bl ock Hi 9h del eti on . However . ( 1 1 3a ) shows a der i vat i on where Hi 9h Del et i on takes P l ace before Low tone SPread and Hi gh SPread . thus Produc i n 9 a correct 272 surtace torm. I t Hi gh SPread were al l owed to take Pl ace and sPread onto a HI noun before Hi 9h Del et i on . th i s H i gh SPread wou l d e f f ect i ve l Y b l ock the envi ronment f or Hi gh Del eti on . as shown in ( b ) . I I ( 1 1 3 ) a . �da n i -i -b { -> �da n f - i -b { -> rtda n � - � -b ! -> �da n � -� -b i I I I I l I � I H @ HL H I I I I L H H HL H L 1 h I I h t I I I I 1 h h I h h H I GH DELET I ON b . �da n i - i -b { ->*�da n { -l-b{ I � \ VI L H H H L H " ....1 h I <D h ' I ' '' ' L H V L H I I 1 h 1.-1\ I V\ V\ , L H L L H I I I � I L HL H I L I I I h h h I h ' mY mothers ' h T COLLAPSE � H DEL I NK I NGe LOW SPREAD ' mY mothers ' H H L H I h H IGH SPREAD �1 h H I GH DELET I ON N/A Bel ow ( 1 14 ) i s an ex amPl e of a comPl ex noun wi th several noun root s . The second noun root . wh i ch i s Lh . sPreads i ts subregi ster Hi 9h tone onto the Th i s i s f o l l owed by a H i gh Del i n k i n9 ru l e . fol l ow i ng HI morPheme. ( 1 1 4 ) a . k�-j i -Po-l j - > k�- j � -p6-1 :S I L I I I / H L " .... 1 h CD H IGH SPREAD I L 1 L / ' bund l e of wood ' H � r1D L I N K I NG � H I GH DEL I NK I NG However . when th i s same comPl ex noun i s fol l owed by a TYPe I I suff i x . the t i nal noun root . bei n9 a d j acent to the Hi 9h i n i t i al suf f i x undergoes Hi 9h del eti on . as shown i n ( 1 1 5 ) . One maY exPec t . then . that the subre9i ster Hi 9h ot the Preced i n9 Lh morpheme wou l d SPread onto the tol l owi n9 Low ton e . but as ( l ISa ) shows, t h i s Produces ' an i ncorrect resul t . --------�- - �-- ----------- 273 ' bundl es ot wood ' ( l l S) a . ka- j i -po-?al o ->*ka- j i -po-?al o I I L L ® H L I I I I I I I 1 . ., I I I I I L I Ii I I h h H I GH DELETI ON L L' ·1 I ... I I H L I I h h H I GH SPREAD I nstead . the deri ved Low tone sPreads onto the Hi gh tone of the suff i x . as shown i n C l l Sb > . b . ka-ji -Po-?al o -> ka-j i -po-?al o I I I L L 6iJ H L I I T I I I 1 I h 1 I H I GH DELETI ON ' bundl es of wood ' � ... 1 I L L L H L 1 h 1 h h I h h I I I I ' LOW SPREAD I t aPPears. from the examPl e above. that the Hi gh sPread r u l e must b e b l ocked f rom sPread i ng onto anY Low tone that is a resu l t of a Hi gh Del et i on rul e . If this is the case. Hi gh tone Spread shoul d al so be prevented f rom occurr i ng i n the exampl e bel ow . where a M i d tone verb wh i ch has undergone Hi gh Del et i on i s preceded by t h e L h Possessi ve Pronou n . �. However . as seen i n t h e deri vati on . i n order to Produce the correct surface f orm. the subregi ster Hi gh tone of the Lh Pronoun must sPread onto the deri ved Low tone. " ( 1 1 6 ) ndi Ya mo nYa- -a - > ndi Ya mo nya-a - > ndi Ya mo nYa- I i Ii I L H LH L I I II I 1 h Ih h I I I I L H h h , I " , L H LH L I I II I 1 h lh h ' I I I L H " L H LH <E) h 1 h lh I H I GH DELETION I I I II I a I I " r see YOU ' L H L l... .... I h h H I GH SPREAD The ProPosed general i z ati on that no sPreadi ng i s al l owed ont o deri ved Low tones. then . meets uP w i t h a contrad i ct i on . I t i s possi b l e that ther e i s a 274 d i st i nc t i on i n tne apPl i cabi l i ty of tne SPread i n9 r u l es b etween sPreadi n9 across word boundar i es and sPreadi n9 across morPnemes wi tn i n word boundari es. Tn i s matter needs more research before i t can be resol ved . The order i n9 of H i 9h Del eti on wi th the OCP i s al so cruc i al . If the OCP aPpl i ed before Hi 9n Del eti on . as shown i n ( 1 1 7 ) . Hi 9h tone wou l d b e l ost al l across the wor d . Produci n9 an i ncorrect surface f orm. I I 1\ H H H L I I I ( 1 1 8) saka-a 1 1 - > <fDV sa ka-a -> saka-a - > * s�k�-a V � H H L I 1 h I " H 1 h OCP ' 90ats ' 1tC- 1' r L I 1 h H I GH DEL H L I h LOW SPREAD However . i f H i 9h Del et i on takes Pl ace before OCP. the envi ror �ent for OCP wou l d then be l ost and the correct surface f orm wou l d be Produced . , ( 1 1 9 ) sa-ka-a - > sa-ka-a -> sa-k�-a I I ®H L , T I , , , H H L , I , H I 1 H 1 I h h H I GH DEL " H I � '1'.. L H L I I h LOW SPREAD Note that the OCP occurs on th i s word when i t i s not fol l owed by a TYPe I I i ndef i n i te Suf f i x . as snown bel ow. ( 1 20 ) sa-ka -> saka -> sak;1 I I \1 1 1 1 I H H I H , OCP 1/: H L L/ 1 I ND L LINKING & LS DELINK I NG Hi gh Del et i on then i s a ru l e that takes pl ace after Assoc i at i on Conventi ons but ---- ------ ---- --- 275 before anY other rul es at the Lex i cal l evel . Th i s second tYPe of Hi gh Del eti on i nvol ves the same Process of del et i ng a H i gh tone wh i ch i s i n comb i nat i on wi th a Low tone. However . the envi ronment for t h i s Hi gh Del e t i on I I i s d i fferent . I t takes P l ace Pr i mar i l y when a HI noun root is fol l owed by some other noun root i n a noun comPound . The nature of the t one of the root does not seem to be cruci al f or the aPpl i cat i on of the rul e . Bel ow are some exampl es o f High Del eti on I I . I n t h e fol l owi ng examPl e . the noun , � ' dog ' is Hl. and ex h i b i ts a Mi d-Low tone as a si mpl e noun ( 1 2 1 a ) . When i t i s fol l owed by adj ecti val mod i f i ers. however . i t undergoes Hi gh Del et i on . In ( 1 21 b ) Hi gh Del et i on takes Place. and then the remai n i ng Low tone tri ggers Low tone SPread i n9 onto the fol l owi n9 HI noun root . When fol l owed by Lh or Low tone, the remai ni n9 Low tone tri 9gerS no tonal rul es , as seen i n ( c ) and ( d ) . ' a dog ' ( 1 21 ) a . P3n �, H L ' ,, 1 I I b . pjn-h x ( � 1 -> H I 1 H DELET I ON .9 pjn � x � � �1 L ' an ol d dog ' H I CD LOW SPREAD & L D!::I_!::r r ON A N i) LS il '=. ,-_ I N II. • •\l C- " a good dog ' d . pjn-kal a:< a ' a spoi l ed dog ' In other cases. a part i cu l ar word maY undergo Hi gh del et i on i n certai n ------ .- �- ---- -�---- --- ---- ------ 276 semant i c env i ronments. wh i l e in other seman t i c envi ronments. i t does not . The noun , � � road ' i s such an examPl e . I n ( 122a ) . i t does n o t under90 Hi 9h del eti on i n contrast to examPles ( b ) and ( c ) . Therefore. i n some cases. aPpl i cat i on ot Hi 9h del et i on maY be semanti cal l Y det i ned . ( 122 ) a. kun-boB / I I 1 H -> kun-bol � '\./ H I qv 1 -> ku-se� -> k 1,� 1 H I I 1 L I H � � travel l er ' H L I h LOW SPREAD ( l i t . road 90�r ) , L I 1 (0 H DEL � bi 9 road ' IND L L I NK I NG . L I NK I NG . LS DELINK I NG OCP I �l � � H L 1.. '" 1 kun- I 1 I 1 H b . ku-see -> I LS DEL I NK I NG. & L I NSERT I ON & L DELET I ON c . ku-sel t ' tr i P ' ( l i t . road 90i n9) Some LH n ouns al so under90 Hi gh Del et i on when in word i n i t i al Posi t i on . The word £Qali • Peanu t , i s one scch examPl e. ' the Peanut sel l er ' b . tcn-caxa � �I -> 1 H DELETI ON tcn-cax� -> t cn-c�x � L.�"' 7 L H I � L H CD LOW SPREAD & L DELET I ON LS DEL I NK I NG ' peanut stal ks ' 277 c. SymmarY H i gh Del eti on I and I I b oth i nvol ve the del et i on ot a Hi gh tone i n comb i na­ t i on wi t h a Low ton e . though H i gh Del et i on I I seems to i nvol ve pri mar i l y HI noun roots - Both take Pl ace i n l ex i cal l Y marked envi ronments. i n that not al l High tones i n comb i nat i on w i th a Low tone undergo High Del eti on . It was seen that Hi gh Del eti on I accounts t or the del eti on ot High tone at certa i n l ex i c­ al l Y marked noun roots when these noun roots Precede a Hi gh i n i t i al TYPe I I sut t i x or an i ncomPl eti ve sutt i x - The morPhol ogi cal structure at that sut t i x al so has a bear i ng o n 'whether Hi gh 'Del'eti on t akes Pl ace . Hi gh Del eti on I I . on the other hand . occurs on l ex i ca l l Y marked HI noun roots wh i ch t i nd themsel ves i n i n i t i al Pos i t i on ot certai n comp l ex nouns. They both occur at the same stage ot r u l e der i vat i on . Recal l trom the d i scuss i on on p . 274 that Hi gh Del et i on I must take Pl ace betore OCP as wel l as bet ore the sPreadi ng rul es . Al though ru l e order i ng was not di scussed tor High Del et i on I I . the examPles i n ( 1 2 1 ) i nd i cates that High Del et i on needs to take pl ace before Low tone Spread . An ex am i nat i on ot ( 1 22) wi l l reveal that OCP must al so take Pl ace at ter Hi gh Del et i on I I . It OCP t ook pl ace betore Hi gh Del et i on . the ent i re complex noun wou l d u ndergo Hi gh Del et i on . whereas i n real i t y, on l Y the t i rst noun root i s subj ected t o t h e del et i on ot i ts Hi gh tone ( see ( 1 22b ) . Theret ore . whi l e the env i r onments Ot these two rul es are d i H erent trom each other , t h eY both take p l ace at the same stage ot rul e der i vati on . and they both i nvol ve the del eti on 0+ a H i gh t one in comb i n at i on wi th a Low t on e . For these reasons , it maY be more ett i ci ent to comb i ne these two ru l es i nt o one ru l e . as stated bel ow . 278 ( 1 24) H I GH DELET I ON: On words l ex i c:al l Y marked f or Hi ';h Del eb on . del ete a Hi 9h tone on the Pri marY or subre9i ster t i er i f a Low tone on the Pri marY or subre9i ster t i er i s l i n ked to the same TBU as the H i 9h tone and 1 ) fol l owed by a Hi 9h i n i ti al TYPe I I or i ncompl et i ve suff i x < If the suf f i x is a nomi nal suf f i x . it must h ave under90ne ei ther no se9mental al terat i on or el se have under90ne Suff i x Reduct i on . ) OR 2) f o l l owed by a noun root i n a comPl ex noun . - i f N. then f u l l or reduced ] 2. La... Del .t i on Whi l e some HI noun roots i n a compl ex noun under90 H i ';h Del et i on and thus acqu i re a deri ved Low tone. there is another verY smal l set of HI nouns wh i ch are rai sed to h i 9h tone i n the same morphol 09i cal envi ronmen t . Bel ow are a couPl e of e:< ampl es i nvol vi n 9 the HI noun ( 1 25) a. ta-c�l')� b . ta-kal a:<a �. ' a good tree ' " a rui ned tree ' Wi th i n the doub l e t i ered approach . th i s " rai si n9 ' can be eff ected by means of del et i n 9 the subre9i ster Low tone of the HI noun root . Such a ru l e maY be stated as f ol l ows. 279 ( 1 26) LOW DELET I ON : Del ete t h e subre9i ster Low t one of a HI noun root , l e:< i c- al l Y marked to do so. when f o l l owed by another noun root wi t h i n a comPl e:< noun . x - X I H I I I t must be s t i Pu l ated that t h e HI noun root has to b e l e:< i cal l Y marked f or Low Del eti on . because on l Y a sel ecti ve set of HI noun roots under90 thi s ru l e . An ex amPl e of Low Del et i on i s i l l ustrated i n the exampl e bel ow . ( 1 27) ta-soonI')9uk� " the second tree ' I � L H I I I h The reader maY recal l another Low del et i on ru l e ear l i er i n t h i s chaPter . L i ke Low Del et i on ( 126) (46) wh i ch was ProPosed above . i t al so del eted a subre9i ster Low tone of a HI morpheme. However , these tWO rul es d i ff ered both i n the tonal environment and i n the stage of the rul e deri vat i on in wh i ch theY occurred . Low Del eti on ( 1 26 ) as a l ex i cal ru l e . occurs at the same stage of the deri vat i on as the Hi 9h Del et i on ru l e ( i . e . before OCP and the sPread i n9 rul es ( se e ( 1 24 ) . whi l e Low Del eti on ( 46) takes Pl ace on a HI n oun root after Low tone spread . Because of these d i f ferences. then . i t seems w i se to keep these two as seParate rul es . There i s at l east o n e case where L Del et i on ( 128) ( 126) seems to b e oPt i onal ( see bel ow ) . In the e�: ampl es bel ow . the f i rst noun root i s oPt i onal l y HH lh tone ( a > . hav i ng undergone Lew Del et i on ( b > , or Mi d tone . that i s. both the H and the subre9i ster Low tone of the .--- ------ HI noun root rema i n i n Pl ace. Note that • 280 regardl ess of whether L Del eti on takes Pl ace or not. Hi gh tone SPread occu�s because i n each case. the�e i s a subregi ster High tone to sPread onto the fol l owi ng HI noun root . ( 1 28 ) ta-I E x( - ta-I E E k� > I I I H - > I I l, ... ...1 I H H H I I (D H H V I H H V I I ' the old tree ' I LI Ni\.INC. 'I> h H SPREAD L DELETION ta-I llke H Q) h l -} ta-I llk� DOWNSTEP �aU ike -> .tal U ke I I I L H H ", ..., h <D FS SPREAD H . I \� I H H H V h I I LINKING & DOWNSTEP In a few cases shewn bel ow. Low Del eti on does not take pl ace at al l on � when i t i s Part of a compound . ( 129) a. ta - Pan l � -) ta�nl � (tree - cleared > · cl eari ng. new f i eld ' ' the f i rst tree ' However , even i f the l ex i cal rul e of Low Del et i on does not take Place. the i n i t i al HI noun reot can be subject to High tone SPread when preceded by a subregi ster Hi gh f i nal word, as shown bel own . The resul t of thi s High SPread i s a surface form that i s i denti cal to nouns that have undergone Low Del eti on . ( ComPare ( 1 28 ) and ( 1 30 » . ( 130) a. Ada tal ll ke t..1 I ! L H H H I � .. 1 I l h l l H SPREAD -> Ad� tal t t ke I $ 1 I L H H H l h I I V H SPREAD ' my oI d tree ' I � RS DELINKING 281 ' mY H rst tree ' The examPles above show that Hi 9h sPread occurred a�ter Low Del eti on . Th i s can b e ef�ected b y assumi n9 that a de�au l t subre9i ster Hi 9h tone i s pl aced on the subre9i ster ti er a�ter the del eti on of the subr'e9i ster Low tone. However , + there are some examples where n o Hi 9h tone SPread occurs after Low Del eti on . Thi s i s i l l ustrated ( 1 31 ) ta - Pada -> bel ow. ta-Pada I H H H H 1 h 1 <b I L DELETION ( tree I I I I I to cross) ' beam ' H SPREAD N/A The reason �or thi s l ac k o� Hi 9h tone SPread i s not known at th i s ti me. I n turther examPles bel ow. the ori 9i nal tone o t the f i rst noun root i s not known . They do not behave as Hi gh f i nal nouns . however . i n that they do not tr i 9ger Hi 9h tone SPread . I t mi ght be suggested that thes� parti cul ar Hi gh tone noun roots are f ol l owed by a f l oati n9 Low tone. However . i t th i s was the case. then the Mi d-Low (Hi) noun roots. which fol l ow these pecul i ar High tones. shou l d be subject to Low tone SPread . wh i ch does not happen . Bel ow are some examPles of comPl ex nouns wi th an i n i tial Hi gh tone noun root foll owed by Mi d-Low noun rooi:s. ( 1 32 ) ka - Para-x a (thi n9 + to cross) ' f l oor , as i n second f l oor ' k a -ton�j (th i ng + i ron ) ? ' rust ' kaa -taxe ' br i dge ' ' beard ' ' tYPe of snake ' ' quai l ' ....... � - � -- -.--- .--- 282 The Hi '3h tone Demonstrati ve Pronoun exh i b i ts simi l ar behavi our i n that i t too does not tri gger Hi gh tone SPread onto HI or Hi nouns. as shown i n (a) and ( b ) below" . ( 1 33) a . nte mE E ts " th i s cord ' b . nde kilne ' th i s road ' Si nce demonstrat i ve. have not undergone Low Del eti on and are not Part of a comPlex noen , i t i s cl ear that thi s l ack of Hi gh tone Spread i n certai n cases, i s 'not a resul t oof the Low DeleHon rule. Whether or not thi s Probl em can be resol ved wi l l have to wai t unti l more research can be done. 3. oth.,. Lex:. c.l Ch..,gn Thlrl irl other ton.l changes , found i n complex noun format i on . that do not currently have anY clear expl anation . For exampl e. the Low tone noun . aAA " man ' . usual l Y remains Low tone when another noun root i s attached to i t : ( 1 34 ) naa + nyel E (man naa + pel � + eYe) + I i (man + -> big + age ) -> nanYel f napele@ " fr i end ' " a verY old man ' However . thi s Low tone i s raised to Mid i n certai n cases (or i n double ti ered termi nol 09Y , i s sub ject to High tone i nsertion on the subre9i si:er t i er ) : ( 1 35) •• nafol j (man + chi ef ) ' man ' s i n-l aws ' b . nandaa " bachel or ' c . naYeril " fr i end ' In al l three of these i nstances the second root i s a ML tone. I t i s not known why thi s ML root f orced the rai s i ng of Low tone i nstead of al l owing the Low tone to sPread onto i t . 283 There are a few examPl es where Low tone is rai sed to HI or Hi gh tone. In the f i rst exampl e b el ow, the second noun root , normal l y Low tone. i s Hi gh tone when Preceded by the noun root �, whi l e , i n the second ex amPl e . both Low tone noun roots are rai sed to Hi gh ton e . ( 1 36 ) sax a + nj ( bush + cow) -> sanS l ox � + ( water + Python ) fii ' buffal o ' -> l uf ! i There are other except i ons f ound i n the data. ' water python ' However . not enough data h as been col l ected to determi n e whether there are Patterns of behavi our among these excePt i ons to mer i t consi derat i on for r u l e f ormul at i on . 4. Tone on Nomi nal i zed Verbs The tone of verbs in a ncmi nal set t i ng deserves men t i on her e . verbs are usual l Y Low when nomi nal i z ed . ' danc e ' -> ko?oro ' di e ' -} ko ' dance ' " death ' Lh verbs ei ther remai n Lh or b ecome H I . ( 1 38 ) a . -> Lh b . -) HI se ' gi ve b i rth ' -> -': z i jool i ' sew ' -> jol a-x:J ' sewi ng ' wa?a ' drY ' -> wa-xa ' drou9ht ' jo ' speak ' -> juu ' speech ' ' ur i nat e ' -> tera-mE ' ur i ne ' (Mw ) heer i c . -) Hi ' b i rth ' koori ' Peel . c l ear ' - > kooro " peel i ngs ' tuxo • ' l oad ' carrY -> tux u-ro (MU Hi 9h tone verbs al so ei ther become Lh or HI . ----- --- � - -- ---- ------ - - -- ------ t.ow tone 284 I')m:5 l :5 ' sl eep ' -> 1 6:< 6 ' hear ' -> nda-:< e ' ear ' 9b6 ' ki1 1 ' -> 9buun ' .f uneral ' ya ' to b e i l l ' -> Ya-m.a ' i l l n ess ' ka n?c3n ' be t i red ' -> kan?a-rc! " f at i 9ue ' ( 1 39 ) a . -) Lh b . -> HI (Mw ) ,- c . -> Hi <MU ' I')m:11 a-IeJ ' n i ghtmare ' Wi t h i n comPounds , v.rb.l el ements .re most common l Y i n non-i ni t i a l Pos i t i on . Once verbal nomi nal s have under90ne t h e tone .f eature changes descr i bed above. thei r deri ved tone is then sub j ect t o the apPropri ate tonal rul es al readY descri bed earl i er . In al l o.f the ex ampl es bel ow < 1 40 , 1 4 1 , and 1 42 ) , the nomi nal i z ed verb is the second el ement Ot the compl ex noun . When a n omi nal i z ed Low tone verb i s Preceded by a Hi or a Low-Hi 9h noun . as shown bel ow . no t onal changes take Pl ace on the verbal el ement • . as mi ':lht be expected . ( 1 40 ) Low verb a. Hi + L na + nY� b . LH + L s�l e + nd�x � <mi l l et + to sow ) < f i re + to l i 9ht ) -> na-nY� I')� -> ' .f l ame ' soond�x� ' mi l l et sowers ' Note that the LH i n < b ) si mPl i f i es to Mi d tone before a Low tone. When a Hi ':lh tone verb or a Lh verb becomes Lh i n nomi nal Pos i t i on i t behaves as anY other Lh nQun in that it is not af f ected by the tone of the preced i ng nomi nal el ement . ( 1 4 1 ) Hi.9h verb or Mid verb a. L + Lh -) Lh nk� + C:� (ch i c ken + 9ather ) -> n kucJ-x� ' cage ' l ux � + Y'r� ( r i ver + stand > -> l UYera-be ' water hol e ' -> tawa-xa ------ -- .- --- --- ' drY Pl ace ' 285 c . HI + Lh Yaraxa + 9ba < t h i n 9 + dr i n k ) -) Ya9ba-xa ' party ' Yaraxa + 9bex al e ( t h i n 9 + create) -) ya9bel ax e ' creat i on ' d . Lh + Lh 9baxa + wa7a <house + drY) -> 9bawaxa 9baxa + nY� �e <house + cool ) -> 9banYi �e e . LH + Lh tililx8 + koori 501 8 + (hoe + cl ear ) -> tU6.kora-x o (mi l l et + uProot ) 9bE ri - > soo9bE ro ' drY house ' ' cool house ' ' hoe for weed i n'3 ' ' mi l l et uProoter ' When a' H i 9h tone or a Mi d tone verb b ecome HI when nomi nal i z ed . i t can be sub j ect to Low tone SPread if the Previ ous noun root is Low f i nal (a) , and al so to h i 9h tone Spread if the Previ ous noun root Possesses a subre9ister H i 9h tone (b) . ( 142 ) Hi 9h verb or Mi d verb -> HI a. L + HI -> LOW SPREAD L - LH nA-pt!- lS� L L H / I L SPREAD & L DELET I ON VAarya b. Hi I L � L HI I LS DEL I NK I NG -) ML - LH /' H HL i 1 I , L L I NK I NG _. _ . - - -- --- V -} v.1.1ntor,CJ ' sheet . b l an ket ' LOW SPREAD �'.A.Y' t HL � H + to <cl oth + to cover ) + ' verY ol d man ' I ta + I E -x E -} tu-H -:{ � t, ( man + be b i 9 + be ol d ) < f ather + be ol d ) ' 9randf ather ' H I <D L SPREAD . L DELET I ON & LS DEL INKING ------- 286 �a + ca t6 (day + cover ) c · Ln + H I - > Lh-H H I GH SPREAD I NG 9ba + I E - > 9ba-l i x t I I -> cantd-�e ' umbrel l a ' c ol d house ' <house + ol d ) l,.. - 'l h Q) L H H SPREAD 9ba + 56n < heuse + spend n i9h t ) -> 9bs6nx� ' bedr oom house ' d . LH + HI -> L-H H I GH SPREAD AND RS DEL I NK I NG 9bon + I E -I E -> 9bdn-I t l t L H H I L H I h CD 1 h 1\ � '� H ' V , l - - -1 + < 9ranarY be ol d ) ' ol d 9ranarY ' L I NK I NG & RS DEL I N K I NG H SPREAD Nomi nal i zed HI verb roots can al so under90 the OCP when Preceded by ei ther Hl or HI noun roots. as shown in the examPl es bel ow. -} e · Hi + HI nYe + PE n-I E , H , 1 /' H I I -} nYi -PE n H � (eYe + be d i sPl eased ) ' j eal ousY ' H L I I r L L I NK ING. L I NK I NG . LS DEL I NK I NG OCP ceewa + j � ------ .- - .---- --- M-ML ( woman + enter ) ----- -> c i j e re - ' adul terY ' 287 f . HI + HI -) Ya + gbon-:d M-ML < t h i ng + h i t ) - > Ya9bonx� " ax e ' The nomi nal i z i n9 of these verbs. and thus the tone f eature changes that accompanY these nomi nal i z at i ons take Pl ace b ef ore the SPread i ng ru l es as wel l as b ef ore the aPpl i cat i on of the OCP. Th i s i s at the same stage of the deri va­ ti on of ru l es as the l ex i cal rul es of High Del et i on and Low Del e t i on . However , the tonal changes that accomPanY nomi nal i z at i ons do not f ol l ow a cl ear enough PaHern to f ormu l at e anY ru l es . We have al readY seen that Hi '3h tone verbs. when nomi nal i z ed i n second Posi t i on of a comPl ex noun become ei ther Lh or HI . There are some c l ues that i nd i cate that the seman t i c make-up of the noun maY Pl ay a rol e in the determi nat i on of nomi nal tone of ver b s . e:<amPl e . when preceded by i ta::. acqu i res a Mi d-Low contour < H ) : < f rom i:e: x , ( Pl ace ' . the verb tapEr�ke ei..d For al wal"� ( the sel l i ng Pl ace ' . Al l other h i gh verbs acqu i re th i s same Mid-Low contour when Preceded by thi s same noun root . However . when Preceded by other noun roots and Pl aced i n Cl ass 1 . eLti becomes Lh . Th i s behavi our i s i l l ustrated by the exampl es bel ow . ( 1 4 3 ) a . c � l � + Pt r t -> c�$P£ r£ -�e b . mj l j c· + .. the bean sel l er ' Pt rE -> mdPE rE -�e .. the r i ce sel l er ' kar� + Pt rt -> kaPE r£ -�e ' the meat sel l er ' On the other han d . when verbs are at the begi n n i ng of a comP l ex nomi nal . they seem to retai n thei r ori gi nal t one. For examPl e . the Hi gh tone verbs remai n Hi 9h . The e:<ampl es bel ow consi st of sentences contai n i n9 a H i '3h verb in i ti al comP 1 e:< noun ( under ! i ned ) • p£rt I@na ( 144) a. Nd i Ya 'I can mE do not kno... ho ... to sel l ' I VP-Ne9. sel l manner kno... not b . Ndaa pt rt kana ' I know ho ... to sel l ' c6n I -VP sel l manner kno... c. p£ rty6l al"le d . Ndi Ya ' sel l er ' ptrty61af)e c.n m€ 'I do not kno ... ...ho sol d ' I VP-Ne9 sel l -chi ef-DEF kno... not Al thou9h these nomi nal s do acqu i re a noun cl ass suf f i x , they are more verbal i n functi on than the other compl ex nouns ...e have been exami n i n9 uP to no... • Other gerund-l i ke nomi nal s i n c l ude nouns wh i ch b e9i n wi t h the f o l l o... i n9 Pref i xes, � ( f rom i: E :<t · pl ace ' ) . Ili=.. and somet i mes ka=,. ( 1 45 ) a . ta- Yazu bi nYa b i kH takgrpxi mPE n Jesus them see them chas i n 9-PP CLAUSE. i t-VP d i s9ust hi m-PP ' When Jesus saw them chas i n9 them. i t di s9usted hi m. ' b. ni- 8i nize�b( l a p Yt;!zuu �m� I S thei r 90i n 9s on . Jesus-VP sl eeP c · ka- ...u kacexekf puna�e nyi kact n�i . h i s doi ngs al l are ' As t h eY ... ere 90i n9, Jesus s l ep t . ' ' Al l that he does i s 900d ' good . D. Cgn�lY'ign gf study gf Cgmplex Ngyn, Th i s sketch i s on l Y the b egi nn i n9 of the studY of comp l ex nouns. I t i s an attemPt to out l i ne the tonal Patterns observed on comPl ex nouns and ProPose tentat i ve r u l es such as Hi 9h Del i nk i n9 ( 1 05) . OCP (88) . Hi 9h Del eti on ( 1 24 ) , and Low Del et i on ( 1 26 ) . Mi l l s, who has been studYi n9 Cebara for about ---. --- ----_ . 30 Years, has come uP wi th a 289 descri p� i on Ot �he var i ous tonal �YPes Ot comPl ex nouns i n �ha� l anguage. Her boun t i t u l da�a has a l l owed her � o c l ass comPl ex nouns i n� o three major categ­ ori es . accord i ng to morphol ogi cal and tonal i nt ormat i on . wi thout attempti ng to expl ain these categor i es . I shal l s i mpl y l i st them wi th the examPles she gi ves to tyPi. t y each categorY: " (235) 1 . margin-c;gre 2 . c;gre-mar9in 3 . c;ore-c;gre s i -sj � : n =rj su-Pe : =re s! -5j j : =d [ 5u=rO + sj� : n l [ su-ro + P': ] [su=rO + sjj : nr i ] .f ood + bad+sut t ood + suck in + sut ' bad f ood ' " t l i es ' " tood +staY ( overn i gh t ) +sut • l et tovers ' < 1 984. p. 1 8 1 > The Mi d-Low roo� i n i ni � i al Posi t i on �akes on �hree d i f t eren� �YPes ot tone dependi ng on �he �ype of compound i ng. Mi l l s gi ves an e:< tensi ve descri Pt i on of the tonal Pat�erns f ound i n each cate90rY. ci � i n9 numerous ex amples for each ca�e90rY. She does no� . however . a��emP� �o f ormul a�e anY ru l es to account for these Patterns . At f i rst 9l ance . tonal Patterns in Suc i te seem to cross these cate90ri es . A much more ex tensi ve studY wi l l need to be done o n comPlex nouns i n Suc i te to determi n e whether the type Ot comPound i n9 wi l l have an ef f ect on the tonal behavi our w i t h i n the comP l ex noun . V. CONCLUSION A. Th- Dgubl e tiered ApprgAch tg Tgn.l An.lY,i, Al l o.f the di scussi ons i n t h i s chaPter assumed the use of the doubl e t i ered apProach to tonal anal ysi s . As men t i oned in earl i er chaPters. th i s not i on of Posi t i ng tone on two t i ers or i gi nated with the concept of tone-spl i tt i ng of tonal f eatures . i ntroduced by YiP and Cl ements. Evi dence has b een brou9ht f orth 290 to argue for the f easi b i l i ty of th i s somewhat novel apProach . Fi r st of al l . i t was seen that there was more than one tYPe o f Mi d tone i n Suc i t e . In a n ef f ort to d i st i ngui sh b etween the two . i t was suggested that one . weak M i d . was a spl i t off f rom the Hi gh tone regi ster. si nce i ts tonal b ehavi our was al most i denti cal to that of H i gh tone ver b s . whi l e the o�her . Mi d ton e . was s i tuated in the uPPer part of the Low tone r e9i ster . Both Possessed the ex act same pi tch . The rePresentat i on of the these tones, HI ( l owered Hi gh t one) and Lh <rai sed Low ton e ) l ed to t h e Pos i t i n9 of these doubl e f eatured tones on two seParate 'ti ers. the second b e i ng subsi d i arY to the f i rst or Pr i marY regi ster . Secondly. i t was seen that i f tone on the subre9i ster t i er was al l owed to l i n k by waY of rul e to the segmental t i er . then some eXPl anat i on cou l d b e gi ven for the under l Yi ng HI tone whose surface rePresentat i on i n i sol ati on i s Mi d­ Low. and f or the Mi d-H i gh tone that resul ts on Low tone verbs af t er the aPpl i cat i on of Hi gh tone sPread . Th i rd l y. th i s doub l e t i ered aPProach great l y enhances the oPPortun i ty to caPture general i z at i on s that a si ngl e t i ered approach wou l d not be ab l e t o do. Thi s is esPec i al l y true for the Hi gh tone sPreadi ng ru l e . I n a si mi l ar d i al ect. SuPYi r e . Car l son ( 1985) ProPosed three ( or tour ) ton al ru l es that cou l d easi l y have b een covered by the h i gh sPreadi ng rul e . In the doub l e t i ered aPProac h . the subre9i ster High t one Spread too k care of Low tone words bei n g rai sed t o Mi d tone after M i d tone nouns. and weak Mi d words rai si ng t o Hi gh tone af ter both Mi d and Hi gh tone nouns. by the s i mpl e act of sPread i ng onto the f ol l owi ng Pri marY regi ster ton e . Fourt h l y . i t was seen that the doub l e t i ered apProach faci l i tated the e:< p l anat i on of tone f eature changes at the l ex i cal l evel . Instead of i ntroduc­ i n g ru l es 0+ tone l oweri ng and rai si ng. ru l es concer n i ng High or Low Del et i on 291 were i n troduced . The not i on of Hi 9h del et i on . was espec i al l y hel pf ul f or i nstance. when the H i 9h tone coul d b e f ound on anY t i er i n con j unct i on wi th a Low tone. The eff ect at t h i s s i mP l e rul e was the ab i l i t Y to l ower a var i ety at tones to Low ton e : Low-Hi gh . HI < weak Mi d ) , Lh < Mi d ) . and Hl<Mi d-Low) . Th i s general i z ati on wou l d have b een al most i mpossi b l e t o capture i f the doubl e t i ered aPProach had not been i mpl emen t ed . Fi nal l Y . i t must be acki'lc�d edged that not al l Prob l ems are sol ved throu9h the use of the doubl e t i ered aPProach . For examPl e . the i dent i ty of the two tYPes of Mi d tones i s somet i mes confused i n the appl i cat i on of tonal ru l es . The Feature Swi tch Ru l e was i ntroduced to al l ow the weak Mi d noun to i nduce Hi gh tone sPread i ng al ong wi th i ts Mi d tone counterpart . wh i l e the tentati ve Hi 9h del i nk i ng rul e was i nserted to take care of the Lower i n9 of a Mi d ton e i n an envi ronment where i t seemed that Low tone sPreadi n9 was tak i n9 Pl ace. HOPet ul l y f urther research wi l l uncover some answers for the anal Ys i s of t h i s phenomenon . I t i s not at al l certai n . however . that these Prob l ems cou l d be deal t wi th anY more eas i l y throu9h the use of a s i n9l e t i ered aPProach to tonal anal Ysi s. In summarY . then . the doubl e t i ered aPProach t o tonal anal Ysi s states that t one can be rePresented on two t i ers . especi al l y in cases where the Pri marY tone re9i sters have sut t ered tonal spl i ts. On the pr i marY t i er i s f ound the tone wh i ch rePresents the pr i marY tone re9i ster . Th i s tone i s ei ther H i 9h or Low ton e . The tone on the secondarY or subre9ister t i er t i ne tunes or spec i f i es wher e . wi thi n the �r i marY tone reg i ster . th i s part i cu l ar tone i s f ound . When the tone on the Pr i marY t i er sPreads. the tone on the subre9i ster t i er i s carri ed al ong by virture 0+ t h e f act that i t i s l i n ked to the Pr i marY re9i ster tone: - ----- _.----------. ------- ------- 292 ( 146) CV CV � T , t On the other han d . when the tone on the subre9i ster t i er sPreads i t does not carrY the Pri marY re9i ster tone al on9 wi th i t . It sPreads on i ts own . usual l Y to the �ol l ow i n 9 tone on the Pr i marY t i er ; but i t can a l so l i n k t o the se9mental t i er by � i rst pass i n 9 throu9h the Pr i marY t i er : ( 147) CVCV. CVCV �� V t t I� a Pri marY tone i s not sPec i � i ed � or tone on the subre9i ster t i er . a de�au l t tone i s i nserted a t the aPProPri ate stage i n t onal der i vat i on . Th i s de�au l t tone i s a coPY o� t h e Pr i marY re9i ster tone i nserted on the sub re9i ster t i er : ( 148 ) CV -> I Tl CV I Tl I tl The r u l e i nser t i n9 t h i s d e� au l t tone has been cal l ed Subre9i ster Spec i � i cati on . Usi n9 thi s doub l e t i ered aPProach . then we have i dent i f i ed f i ve d i f f erent under l Yi n9 tones : LI . Lh . Hh , HI . and Hi. The three under l Yi n 9 contour tones that have been i sol ated are Hh-Lh (Hi 9h-Mi d > . Hh-LI . and LI -H (Low-Hi 9h ) . ------- -- -- -- -- ------ ------ ------- ------- ---- 293 B. SummarY Thi s chapter has di scussed the tonal behavi our wi t h i n the structure ot the noun Phrase . I t was seen that r u l es that oPerated on the verb Phrase were al so found on nouns. However . the more comPl i cated structure ot nouns r equi red a thorou9h exami nat i on ot the und ;,.-1 Y i n9 tones and their i mPI i cat i ons tor ru l e f ormu l at i on . The tones that caused the most d i f f i cu l t i es were those i n i ti al l y dubbed as weak Mi d and M i d -Lowe Si nce both reacted to tonal rul es i n the same waY . i t seemed conven i ent t o 9i ve them s i mi l ar . i f not i dent i cal under l vi n9 tones. Two Proposal s concern i n9 the i r under l vi n9 tone were put torth. one sU9gest i n9 that these tones were under l vi n9l v Hi 9h tone. the other that i t was a l owered Hi 9h tone. HI . Af ter d i scuss i n9 the mer i t s and drawbacks tor both aPProaches . the l owered Hi 9h tone aPProach was chosen . since the l atter seemed to i nvol ve tewer prob l ems f or anal vsi s. The rul es ot Low SPread ( 1 1 ) and Hi 9h SPread (54) were then re-ex ami ned and retormul ated i n order to account t or the tonal changes on the HI noun . Once the i ssue of the under l Vi n9 tone tor these weak Mi d and Mi d-Low nouns was set t l ed . the matter ot d i st i n9u i sh i n9 between weak Mi d and Mi d-Low tone were dea l t wi th . I t apPeared that both were mar ked l ex i cal l v tor the app l i ca­ t i on ot certai n ru l es to take pl ace at the Phrase l evel . Weak Mi d (HI ) nouns underwent the Feature Swi tch rul e (66 ) , whi l e Mi d-Low ( Hi) nouns tri 9gered the i ndependent Li n k i ng ot the Subre9ister Low tone to the se9mental t i er ( 60 ) . Both ot these rul es. i t was seen . were more easi l v stated i t the subre9ister Low tone was Part ot the l ex i cal make-uP ot weak Mi d and Mi d-Low tones rather than i nserted at phrase l evel . In the l i ght ot these n ew developments. the under l Yi ng tone ot the det i n i te --------�- - ---- ------. ------ ----- . 294 suf f i :< was re-exami ned . reveal i ng that al though e i ther an underl Y i ng Hi gh tone or a HI tone coul d b e Posi ted f or the def i n i te suf f i x . there seemed to be a verY sl i ght advantage to choosi ng a H i gh tone as the under l Yi ng r ePresentat i on . The f i nal sec t i on of thi s chapter di scussed tonal behavi our wi thi n comPl ex nouns . I n manY cases i t was f ound that the Hi gh and Low tone Spread i n'; r u l es wh i ch were attested to oPerate across word boundar i es. al so took Pl ace across morPheme boundar i es wi th i n the comPl ex noun. In add i t i on to these r u l es . however . tonal behavi our was f ound that was pecul i ar to t h e comPl ex noun . Al l of them seemed to b e somewhat l i mi ted i n thei r appl i cat i on . Fi rst of al l . the val u e of appl Y i n g OCP (38) to a verY sel ect set of comPl ex nouns was d i scussed . It was deci ded that al though one cou l d Perhaps have found a waY to el i mi nate the need f or the OCP. it more easi l y Provided a correct "i output on complex nouns that requi red IndePendent Low tone Li nki ng. Secondl y , the Prob l em concern i ng the l ower i ng o f a Mi d tone between a Low t on e and a H i gh tone was verY tentati vel Y resol ved through the appl i cat i on of the Hi gh Del i n ki n 9 ru l e ( 105) . Fi nal l y . tonal behavi our that seemed to be governed by Pri mari l Y l ex i cal consi derat i ons was descr i bed . al though on l Y two rul e s . Hi gh Del et i on ( 124 ) and Low Del et i on ( 126 ) , were proPosed . At var i ous Poi nts i n th i s chapter . order i ng of the ProPosed ru l es was d i scussed and argued . It was seen that the Le:< i cal ru l es of High Del eti on and Low Del et i on must take Pl ace f i rst . fol l owed by the Ocp, Feature Swi tch and Low Li n k i ng, Low tone SPread i ng and Hi gh tone Spread i ng . Li n k i ng. LS Del i n ki n9. RS Del i n k i ng . and f i nal l y Downstep. However , orderi ng was not thoroughl y di scussed f or al l rul es . A more thorough d i scussi on i nvol vi ng the i nteract i on al l ru l es ProPosed i n thi s thes i s wi l l be Presented i n Chapter 6. -- - . _._- ---- ------ 295 NOTES 1 . The numbers oj"kin ' one ' , � ' two ' , and � " th ree ' are comPosed o� two comPonents. the � i rst comPonent i s l in ked �ow tone and the second une to Hi gh tone. Normal l y, there is an i nternal Low SPreadi ng rule. sPread i ng the Low tone of the f i rst comPonent to the Hi gh l i n ked component . Thi s creates a Low-High contour on the f i nal Tau. which becomes Mid tone i n non Phrase t i nal envi ronments. as i n (a) . In Phrase t i nal Posi t i on . however . this Hi gh tone i s del i nked <b) . a· waa sc}�� ta !t ' J L la ' he taak two? ' H b . ndcriya s � II L H When these Low i n i tial nouns are subject to M i d tone sPread however. Low tone sPread i s b l ocked from aPpl Ying to the Hi gh tone comPonent. The resu l t i ng tone i s Mi d-High as can be seen i n (7c) . 2 . Recal l f rom Chapter 3 that the underl Ying tone of the Def i nite Suff i x was not undi sPutel Y under l Yi ngl y Hi gh tone. However . at thi s Poi nt. we shal l assume that it is under l Y i n91y Hi 9h tone unti l f urther anal Ysi s can be undertaken . See the subsec t i on I v. a. , t i t l ed C Hi 9h Del i nk i ng ' i n the sectian on ComPlex nouns �or a discussi on o� an envi ronment tri ggeri ng the l owering o� Mid tone nouns to Low tone. Recal l , al so in Chapter 3 that Mi d tone noun roots l owered to Low tone betore certai n tYPes of TYPe I I suf f i x es. Thi s l atter Phenomenon wi l l be di scussed agai n i n the subsecti on on l ex i cal changes i n thi s chaPter . 3. 4 . It was mentioned i n the i ntroductorY secti on of the thesi s that most Senu�o l anguages have three l evel tones. whi l e a few in the southern area are rePorted to have tour l evel tones. If this rePort hol ds to be true , it coul d prov i de evidence to the hYPothesi s that Senufo l anguages were hi stori cal l Y of tour l evel tones. Among the three l evel tone l anguages. SuPYire also Possesses two different tYPes of Mi d tones, whi ch behave in a verY simi l ar fashi on to the Mid tones of Suci te. Cebara, on the other hand . seems to Possess onl Y one tYPe of Mi d tone. Insuf f i c i ent data on other Senu�o l an9uages i nh i bi ts turther i nvesti ­ gation at th i s t i me . 5. Certai n comPlex nouns HI (weak Mi d ) and Hi ( Mi d-Low) n ouns are opti onal l y subject to High tone SPreadi ng . If Hi gh SPreadi n9 does not take pl ace. then consequently there i s no envi ronment for RS tone Del i nkin9. The examples bel ow Possesss two acceptabl e t orms. one (a) where nei ther Hi9h Spread nor RS Del i n ki n9 take pl ace and the other ( b ) where both take Pl ace . Ada + j� + ka-n kudo-ngu-I")e - > a i\ da jAkA"k"dQr')glll")e -QRD-DEF son f i ve mY • mY H Hh son ' OR . :296 Jaka�k6d6�9u�e I I I V1: I I b . � da L H H H I 1./1 I l h Q) l H I 1 L H 1 The i �era�i ve Hi 9h SPreadi n9 whi ch takes P l ace here wi l l b e d i scussed i n the secti on on Compl ex nouns. 6. See pp. 289ff at �he end of �h i s chaP�er for a ful l er ar9ument in suPPor� of �he doub l e t i er ed approach f or tone f eatures and ru l e aPPl i ca� i on . 7 . On l Y a par�ial t onal der i va� i on of these san�ences are gi ven i n order t o s i mpl i f y the di scussi on a t h and . 8 . See sect i on IV. B . on Hi 9h Del i n k i n9 f or eXPl anati on of Hi 9h Del i n k i n9 and T Coll aPse . 9. There are a couP l e of examPl es. however . �hat seem to i ndi cate �ha� . i n certai n i nstances . t h e Demonstrati ve Possesses a f i nal f l oat i ng Low �on e . I n the examPl es bel ow. t h i s supPosed f l oat i n9 Low �one spreads onto �he fol l owi n9 verb (a) and noun ( b ) : a· aka k60 na ' �m� l i ! -> oka that cut wi �h k n i f e-wi th b. aka l �r� nYe j ecr i YQ -> that Pr i ce i s how-manY Q keo na ' �m�l i ! ' cut that wi th a k n i f e ' oka l�r� nYe j�cri YQ ' How much i s tha�? ' 297 CHAPTER 6 - THE SUCITE SENTENCE A. Introduc;tion The Previous chapters �ocused on tonal behavi our o� nouns and verbs i ndi vi dual l Y and then anal Yzed tonal behavi our across word boundari es- Thi s has l ed to an ex tEnsi ve d i scussi on about the rePresentat i ons o� var i ous surtace tones - in Parti cu l ar . those whi ch are Mid tone. Thi s chapter compl etes thi s studY of tonal behaviour in Suci te by anal Yz i ng the tonal behavi our of a few more aspects ot the Suci te sentence and then by reviewi ng al l of the rul es Presented in the thesi s and exam i n i n g how these rul es shoul d be ordered with respect to one another . B. Th. Ady.rb!.! Phr••• The adverbi al Phrase maY consi st ot a si mpl e adverb or a phrase consi sti ng ot an NP and a postPosi ti onal part i c l e. ( 1 ) �nja • Bel ow are a number ot common adverbs: yesterday , waa Pan tanja · he came yesterday ' n!nja · today ' nY� "'pal')a • tommorrow , b(nj � • l ast year ' n-&n?an ' here ' wal Pan ncin?an ' he came here ' wa " there ' waa kad w3 " he went there ' Most adverb Phrases . however . consist ot a noun plus a Postposi tion . These adverbs and adverb phrases are l ocated i n postverbal posi t i on i n the sentence- ---- ---- .�---- 298 Bel ow are some examPl es showi ng the l ocati on o� the adverb phrase, whi ch i s underl i ned . (2) a. katake nY� Hunger i s ad. lA me an ' I am hungrY. ' Subject Vb . Pro PP b. ndi Ya �Y�1d nAk�I Abf nyA?6 lA 1 '1 am afraid o� thi eves ' am afrai d thi eves ' face on SUb. VP Vb. Noun Noun PP c. waa he some weave ma-for ' He wove some �or me ' Sub-VP Obj . Vb . Pro-PP The tone of the naun Phrase wi thi n the adverbial Phrase a�fects the tone of the PostPosi tion i n much the same waY as the tone of nouns affects the tone of �ol lawi ng verbs. High tone Postposi ti cns are Hi gh after Mi d or Hi gh f i nal nouns, but are subject to the Low tone sPread rule when preceded by a Low f i nal noun. The major Processes a��ecti n 9 Postposi ti ons after nouns are i l l ustrated i n the data in (3) . The examPles i n (3a) and (b) show two dif ferent Hi gh tone Postposi ti ons Preceded by a Lh nou n . whi l e (e) shows how a Law-Hi gh Pronoun does not af�ect the tone of the tol l owi ng Hi gh tone Postposi tion , but i s i tsel f subject to RS Del i nk i n9. In <d) and <e) are examPles � a Low f inal noun SPreadi ng i ts Low tone onto the tol l owing Hi gh tone Postposi tion . In ( d ) , the Postposi tion is onl Y one TBU and i s l ocated i n Phrase f i nal Posi t i on . Once Low SPread takes place, RS Del i nk i ng del i nks the Hi gh tone of the resu l t i ng Low­ Hi ih contour when i t i s i n Phrase � i nal Posi tion . ----�- �- .-- --- ------- 299 ( 3 ) a. waa w� kan mo-6 ' he gave some to YOU ' I I L H I h b. wu nYe mo ��n " he i s besi de 'J I L yOU ' H I h c . 'waa w� kan ;,da-ci 1/1 1 • he gave somit to me ' -> waa w� kan ;'dil-6 L H H RS DELINKING d . waa w� kan n.-�e- 6 -> waa w� kan n�-�u- L H (L) H tian -> "-,,1 ,,'.... �I HL H (L) U 1 " he 9ave some to the man ' L H (l ) H RS DEL INK INS lOW SPREAD e. wu nYe na-ke il L---1 ......-1: t.. ... , .... "I H lOW SPREAD wu nYe naka t�n A/I P'1 ' he is besi de the � i re ' HL H (L) H V 1 LS DELINKING Mi d (Lh ) tone Postposi t i ons remai n Mi d tone after Mi d tone and low tone nouns. but are subject to Hi9h Del inkin9 ( 44 ) when preceded by an underl Yi ng Hi 9h � i nal noun . ---_ .. _ .. - .- --- ----- ---- ---- - -_.- .-.... ... _ .. . 300 ( 4 ) a. kat�ke nYe mo l a � You are hungrY ' katike nYe �da l a -> kat eke nYe �da l � '-'t l I ' 1 h I L H Vl l I I 1 h � L H L h " I am hungrY ' L h HIGH DELINKING (44) b. t ��a K a la -> ��i xa l a (touch it on ) " touch i t ' There i·s no data showi ng low tone postposi H ons. thus there i s no waY of knowing whether Low tone PostPosti ons wou l d be subj ect to Hi gh tone SPread i � Preceded b y a Mi d tone noun . as i s the case f or Low tone verbs. an The i n i ti .l el ement of Thus • Precedi ng verb. • adverb phrase is not a�fected by the tone of the Hi gh tone verb does not tri gger High tone SPreadi ng onto HI nouns of an adverb phrase. Rather . as exampl e ( 5 ) i l l ustrates. the Hi noun at the begi nning o� the adverb phrase undergoes Low Li nking (60) . a rul e. the reader maY recal l . whi ch l i nks a subregister Low tone of a Hi noun to the segment when that Hi noun is in Phrase i n i t i al Posi t i on . (5) waa 1a p€r€ n9-�u - I H I h 1\ I � -> I HL H <U H I �' I 1 h h waa 1a I f\,1 P€r€ n9-�u H I h - .... � 1: e he sol d i t to the mother ' HL H (U H V I 1 h , h LOW SPREAD & RS DELINKING LOW LINKING Thi s lac k of tonal i nterac t i on between the verb and the adverb phrase i s i n keePing wi th a statement made i n Chapter 1 that there i s a b�rri ar for tonal change b etween a verb and a fol l owing nomi nal el ement . ----_ . _- _ . _ - --- ----- 301 C. Th- Noun Cl ••, Cli tic The noun cl ass c l i t i c has been menti oned in ear l i er chaPters. However . i �s comPl ex �onal behavi our has requi red del ay of i ts anal Ysi s unti l now. Each noun cl ass has i ts own cl i ti c . However , al l noun cl ass cl i ti cs. regardl ess of cl ass af f i l i at i on Possess the same tone. The noun cl ass cl i ti c can be found i n subject and obj ect Posi t i on . or a s Part o f an adverb phrase. It can al so functi on as a Possessi ve Pronoun. Bel ow are a few examPl es showi ng the cl i t i c i n (6a) subject Posi t i o n . (b ) obj ect Posi t i on . and ( c ) a s a Possessi ve Pronoun. (6) a. It& b. nda� c· It& Ya ma � nY� " he i s cOllli ng ' " r saw h i m. C Hi s meat is here ' ka� nYe nanran I t has been referred to by other authors. such as Mi l l s ( 1 984) . as a general cl ass Pronoun . The reason that i t is l abel ed a cl i ti c i s because of i ts tonal behaviour wi thi n the sentence . It was men t i oned ear l i er that tonal boundari es exi sted between verbal el ements and f ol l owi ng nomi nal el ements. Noun cl ass c l i t i cs. however , break down that tonal barri er i t Pl aced i n i n i t i al Posi tion of the noun phrase, and are subject to tonal rul es from the Preceding verbal el ement . The al tered tone of the c l i t i c can then affect the tone of succeed i ng elements of �he noun phrase. In addi t i on . when the noun cl ass cl i t i c functions as a possessi ve pronoun . i t cl i t i ci zes to the fol l owi ng noun . The fol lowing di scussi on wi l l descri be the behaviour of the noun cl ass cl i ti c wi thi n various tonal envi ronments. First . we shal l l ook at how the cl i t i c affects the tone of fol l ow i ng consti tuents. and secondly, how �he tone of the noun c l ass cl i t i c i s affected b y the tone of �he Precedi ng verbal el ement . The compl i cati ons of the cl i ti c ' s tonal behaviour wi l l be examined in the l i ght of the Present anal Ysi S ---- . ---. - - - _ ._-- ------ 302 i n an attemPt to seek out a vi abl e sol uti on . I n sentence i n i t i Al posi t i on� the noun class c l i t i c exh i bi ts a Mi d tone as seen i n (7) bel ow. The questi on i mmed i atel Y comes to mind as to what tYPe Ot Mi d tone i t i s. It one exami nes i ts behavi our i n subject Posi tion . there i s an i ndi cati on that the c l i t i c tone maY be Lh . In (7a) below. the subject � " YOU ' has a Lh tone. When �� a pre-tense marker . tol l ows. it i s subject to Hi gh tone sPread f rom the Precedi ng Lh subject. resul t i ng i n a Mi d-High tone. When s�n?e i s Preceded oY the noun cl ass cl i ti c . i t acqui res the same Mi d-Hi 9h tone, as can be seen i n ( b ) . (7) a. mo s�n?� Ya ma -) mo sen?� Ya ma " You a�e comi ng agai n ' I I , L L H ..... ... h HIGH SPREAD b . w sen?� Ya ma I I I L L H - } wu sen7� Ya m � " he i s comi ng again ' "' .... h HIGH SPREAD On the other hand, the tone of the cl i t i c does not tri gger Hi gh sPreadi ng onto the Negati ve marker as i s oPti onal l y the case tor regu l ar Lh nouns 1 • However , when put i n object Posi tion or i n an adverbi al phrase, the storY changes. Fi rst . i t must be noted that noun cl ass cl i t i cs i n object Posi ti on � tri gger Hi gh tone sPreadi n g onto Low Tone verbs. Note bel ow. that when a Low tone verb fol l ows a cl i t i c . i t does not acquire a Mi d-High tone (aa) . as i t does when Preceded b y the Lh Pronoun cab ) • ---_ . . _- _ . . _ ._ --- ----- • 303 (6) wu gbara ' meet h i m ! ' *wu 9bara I L ... t, L \) l.. -:: - - h HIGH SPREAD, HIGH SPREAD TO TBU mo gbara I L j.... .. I -> L mo gbara �, � ... I L h meet you ! L, h H SPREAD H SPREAD TO TBU, LINKING, LS DELINKING The cl i t i c , then , cannot be al l owed to tr i gger high tone SPreadi n9 onto fol l owi ng verbs Particularl y. when i t functi ons as an object . As a resu l t . the cl i t i c does not seem to be underlYi n9l y Lh after al l . A possi b l e eXPl anation for �hi s di screpancY i n tonal behavi our i s �hat the underlYi n9 tone ot the subject c l i t i c i s different from �hat of a cl i t i c in object Posi tion . Si nce i t aPPears that cl i ti cs are not under l Yi n9l y Lh tone i n object Posi t i on , thei r behaviour needs to be further exami ned . When Preceded by a cl i ti c . a Mi d tone verb or Postposi tion i s l owered to Low tone (9a and b ) . Thi s same 1 0werin9 Phenomenon i s observed when a Mi d tone verb i s Preceded b y a high t i nal noun (9c ) . (9) a. b. wu tux o -> wu wu la -> c . fal axa l a tux o wu 1 4 • carrY i t: ! ' ' on h i m ' -> t�laK� 1 4 ' on a roc k ' Thi s 1 0wer i n9 was anal Yzed i n Chapter 4 as Hi gh tone Del i nk i n9 (44) . where the subregi ster Hi gh tone of a Lh verb is del i n ked when Preceded by a Hi gh t i nal noun and i n Phrase f i nal Posi tion (see ( 10» . ---_ . _ -- ---- ------ 304 ( 10) �fU axa tuxo I > � a l ax a t�xO H L H h h h - I I " carrY a rock ' I I I L $ I h HIGH DELINKING It the PostPosi t i on is ( 44 ) al so anal Yzed as underlYi nglY Lh . the same High Del i nking rul e can be moti vated when the Postposi tion i s Preceded by a Hi gh � i nal noun such as fala&i. �4lalCa. l a -> �� Uxa (11) I I I I a l I H L H L h h I h � h HIGH DELINKING ( 44 ) Al though the cl i ti c i s not Hi gh tone on the surtace. i t could al so moti vate a Hi gh Del i nk i ng rul e i f the el i t i c i s anal Yzed as Low-Hi gh . A Low-Hi gh tone l i n ked to the same TBU woul d create a surtace Mi d tone whi l e at the same t i me tri ggering High Del i n ki n9 o� the f ol l owing Lh verb or Postpos i t i on . Such a der i vat i on i s i l l ustrated bel ow. ( 1 2 ) a. wu tuXO I " LH II wu t�l(O L " LH h Ih I Ih -> \ ' carrY i t ! ' L � II -Ii HIGH DELINKING b. wu l a " II l LH L lh I h -> wu 14 Ih h " I LH L II � " on h i m ' HIGH DELINKINI3 -- - -- -- - -- -- -- --- --- -- - 305 Un�ortunatelY, the sol uti on �or the tonal behaviour o� the noun cl ass c l i t i c i s not that si mpl e. It i t were anal Yzed a s undarlYi n9l Y Low-Hi 9h there wou l d be no waY to moti vate what aPPears to be Low tone SPreadi n9 on t ol l owi n9 Hi 9h tone verbs. In ( 13a and c) below n ote that when a Hi 9h tone verb or Postposi t i on is preceded by a c l i t i c . it acqui res a Low-Hi 9h tone in exactl y the same waY as it does when Preceded by a Low � i nal noun (see ( 13b and c » . ( 1 3 ) a. wu Plrt -> wu P�r€ ' Sel l i t ! ' ' sel l r i ce ! I ' besi de him ' • besi de rice ' Thi s tonal change was anal Yzed as the Low tone ot the noun sPreadi n9 onto the verb . as shown bel ow . (14) ' sel l ri ce ' LOW SPREAD LS DaINKING As a resul t . thi s behaviour seems to i ndi cate that the cl i t ic Possesses a Low f i nal tone. thou9h thi s Low tone does not show uP on the surtace. However . an under l Yi n9 � l oati n9 Low tone shou l d not Permi t the rule of Hi 9h Del i nk i n9 on Mid tone verbs, si nce Hi gh Del i n k i n9 requi res a Hi gh t i nal noun or Pronoun . not a Low t i nal one. The noun cl ass c l i ti c . theretore. i s gi ving contrad i ctorY si9nal s concerning i ts underl Yi n9 tone rePresentation . As Part of a noun phrase. i . e • • as a Possessi ve Pronoun. the c l i ti c exh i bits yet another type ot tonal behavi our . When tol l owed b y Low i n i t i al . Mi d or Mi d- Low nouns . i t tri 9gers no tonal changes. as shown i n ( l Sa , b . c ) . However . when ----- ---- - " "- " 306 fol l owed by a weak Mid (HI ) noun . that weak Mi d noun becomes Mi d-Law. as shown ( 15) a. wu m�l a-I")e " hi s r i ce ' b . wu 9ba-ke " hi s hause ' c . wu ja-I")e " hi s sen ' d . wu ta-ke -> wu ti-ke " hi s tree ' e. wu so-ke -> wu sc-ke ' hi s mortar ' Thi s i s exactl y what hapPens when certai n comPlex nouns are composed at twa or more HI or Hl noun roots. As shewn i n the examPle ( 16 ) below. the t i rst noun root remai ns Mi d tone. whi l e the t i nal noun root acqu i res a Mi d-Low tonal contour. Recal l tram Chapter 3 . that this behavi our was accounted tor through the use at the OCP rul e (88) , whi ch col l aPsed two HI tones i nto one Hl ton e . whi ch subsequent l Y underwent Independent Low Linking ( 60 ) i n Phrase i ni ti al PosH i on . ( 16) a. Ya - ta-ke I I I -> Ya-�a-ke -> Ya-ii-ke V H H H (L ) H I I 1 I I I OCP I H ( L) V, H L V I L LINKING I -> H (L) Ya-ti-ke I 1... '" ' neck ' H L H (L) V I L SPREAD If one posi ted the c l i ti c as bei n9 underlYi n9l y HI . i t cou l d al so be moti vated to under90 the OCP when f ol l wed by a HI noun . as shown i n (b) bel ow. IndePendent Low Linking wou l d then take Pl ace an the f i nal TBU creating a Mi dLow contour. ----- . ---- - -- -.--. ---. . ---- ----- ------- 307 b . wu I �a-ke + I / -> H H H (L ) 1 1 wu �a-ke '-J I I I -> wu ta-ke -> H H (L) I V\ H L H (U 1 � I DCP I / wu ta-ke " hi s tree, I ND ' ��(U , '/ 1 L LINKING L SPREAD In summarY. then . the noun cl ass cl i ti c conti nues to gi ve conf l icting si gnal s concerni ng i ts underlYi ng nature. As a subject, it seems to be Lh . As a noun object i n sentence initial Position . i t seems to be both Hi gh f i nal and Low f i nal dependi ng on the tone of the verb whi ch fol l ows. When foll owed by a HI noun . it behaves as a HI comPonent of that noun and ParticiPates i n the DCP. The precedi ng di scussion examined how the cl i ti c affects the tone of fol l owi ng nouns. verbs and verbal parti cl es. The noun cl ass cl i t i c i s al so affected by Precedi ng verbal parti cles and verbs. The reader wi l l recal l that normal l y a noun in Phrase i n i t i al Posi t i on is n ot affected by the tone of a Preced i ng verbal (see Chapter 1 ) . Howaver . when a cl i t i c i s i n i n i t i al Pos i t i on of a noun Phrase. anY Precedi ng verbal can al ter the tone of the cl i t i c . whi ch can i n turn. al ter the tone of fol l owing nomi nal s and verbs. For examPl e. when a c l i t i c i s i n sentence i n i t i al Posi t i on functi on i ng as an object . i t can tri gger Low tone sPreadi ng onto Hi gh tone verbs. as seen i n ( 17a) bel ew. However , when Preceded by a High f i nal verbal particle, as i n (b) . the c l i t i c i s i tsel f High tone. as wel l as the Hi gh tone verb . Whatever floati ng Low tone there maY have been underl Yi ngl y i s del eted in thi s envi ronment. When Preceded by a Low tone Par�i cle. as i n (c) , however , the cl i � i c i s Low tone, whi le �he fol lowi ng Hi gh tone verb i� Hi gh . ( 1 7) a. wu pt rt -> b . mo na c . moo wu _ .... - . - -. wu ptrt -> ptrt _.. wu P�rt -> mo n a WO Plr€ ' You sol d moo wu Plr€ ' YOU sol d i t ' . . _ ---- -- - _. _ .. . . __ ._ H' . . . _ --- 308 I t wou l d aPPear that , i n thi s case, the c l i ti c behaves as a Hi noun and i s subject tc both Hi gh tone and Low tone SPread , as shown i n the deri vations bel ow. , ( l S ) a. mo net wu P€r€ -> mo na WO Pir€ , I I I I I I I L H H L H H H H I I I h h I I h h I I (., '1 h � YOu sol d i t ' <D h HIGH SPREAD b . mo-o wu P€r€ I l, / I L L I l h I H ,l W L SPREAD & I -> H 1 h L DEL moo wi} P€r€ I '-'1 LL H t " hI '� H I h h LINKING � you sol d i t ' � RS DELINKING A cl i ti c whi ch functi ons as a possessi ve Pronoun can al so be subject to Low tone or Hi gh tone sPreadi ng, which i n turn can affect the tone of fol l owi n9 consti tuents i n the Phrase. ComPare the derivati ons of ( 19a ) and ( 1 9b ) bel ow. In (a) , the possessi ve Pronoun and the fol lowi n9 HI noun under90 OCP. Since · they are in sentence i n i t i al Posi tion . Low Li nkin9 takes pl ace, which then results in Low tone sPreadi n9 onto the fol lowi n9 Hi gh tone verb. In (b) , � kaca i s sti l l i n Phrase i ni tial Posi tion . so one mi ght exPect Low Li nking to take Place. However , if this haPPened . an i ncorrect surface t Qrm wou l d be Produced . as shown i n (c) . Instead the cl itic is subject to Low tone Spreadi n9 from the Precedi n9 Low tone verbal particle. Once the subre9i ster Low tone of wu kari i s del eted , Li n k i ng and RS Del inkin9 take place, thus produci n9 a surface Low tone on the cl i ti c and a Hi gh tone on the noun. .. ---_ ._-_ _-- ------- 309 ( 19) a. wu kara w" I I H H H 1 1 h I , wu kara w" -> I "-../ H wu kari wee '/'1 H " 1 I -> I H L H 1 h h wu kari w�e -> "v1,-.....: 1 H L H 1 h I .1 I,. " -> u� w I I I /I waa wu kara wee -> HL H H H HL 11 1 h 11 I ' .11 1 '-./ " II I I -> waa wu �ra wee I t.. .... 'v' , H L I L INKI NG waa wu �ri wee I V'I,J/ \J HL H H HL H H 1 h 11 Q) h 11 h h I " , , L SPREAD & I L DEL II LINKING I I & RS DEL INKING *waa wu kara wee � �'I -> H H Ii" 1 h H ' he l ooked at h i s meat ' "vi -> 1 H OCP c . waa wu kara wee I H LS DELI NKING & b . waa wu kara wee H L I I L SPREAD L LINKING OCP ' l ook at his meat ' I '/ L H r LOW LINKING LOW SPREAD & LS DELINKING When a cl i ti c whi ch has undergone Low SPread is fol l owed by a Low tone verb . the resulting Low-Hi gh contour remai ns l i nked to the c l i tic. creati ng a Mi d tone, as shown below : (20) a . waa I � I L H J I 1 1 wul o I L I 1 -> waa � wul o L H L ..- 1 I 1 I � wul o L L H ' -4 1 I 1 h L 1 1 , CD L SPREAD wac! I & -> I� I ' he took i t off ' L DEL LINKING When a cl i ti c whi ch has undergone Low SPread is fol l owed by a Mi d tone verb. the verb undergoes High Del i n k i ng. as shown bel ow: ---- ----- - ._ . _----- 310 b . waa � nYa -> waa � I I I l,.. ... l l H l l I 1 I 1 nYa -> waa H � nYa l H 1 1 l .:t:' I h h 1 '.......-1 l I ' I ' h l eD h " he saw h i m ' t H I GH DEL INKING lOW SPREAD & l DElETION " ,. Naun c l ass c l i ti c5 are al sa subject ta �anal rul es when i n i ni t i al Posi t i on at an adverb phrase. I n the examPles belaw, the H i gh tane verb , as wel l as the '. High �ane i ncampl et i ve suH i x an law tane verbs . tri gger H i gh Spread anta the c l i t i c (21a, b . c ) , wh i l e the low tane verb can tri gger Law tane Spread (21 d ) . (21 ) a. waa l a P€r€ � u -> waa l a Plrt I I , H H H !lAd " he saI d i t ta h i m. ' L. ' I h Q) H IGH SPREAD b . wu Ya H nu � l a -> wu Ya H nu IrliC&. l a .. ...., , L H H I I 1 h I 1 I J...-1 l l H 1 H 1 - > wu Ya H nu IrliC&. l a -> 1/1 1 1 L H l • " ....1 I 1 h CD h I h LOW SPREAD H I GH SPREAD H I V 1 h l * h H I GH DEL I NK I NG ->wu Ya U na � a L-4 L H V I H h I L t'I RS DELINK ING he VP l i e- I NC h i m an c· wu Ya IrliC&. klan � � he i s l Y i n g an h i m ' -> wu Ya IrliC&. kaan � " he i s g i v i ng i t ta h i m ' 31 1 d . waa �muuna c� I � l L H I I 1 1 nYU�ge mPa�a i -> waa �uuna cil \ H I H (L) I I 1 h / , tIlL I H L L h 1 " I nYu�ge mpa�a-i -> '" I I I I H H ( L) H 1 h I " I L I h OCP L H I h I L SPREAD � L DELETION � L INKING , H (L> H L I I I I LS DELINK ING. RS DELINKING h h � DOWNSTEP 'he-VP kni�e 9rab h i s head above-PP " he 9rabbed the kni�e above h i s head ' I t must be noted . however . that Mi d tone verbs cannot . as their structure shou l d Permi t . tri gger Hi gh tone sPreadi ng onto noun c l ass c l i t i cs. The e��ect i s that cl i ti cs do not become Hi 9h tone after Mi d tone verbs. as i l l ustrated bel ow. Rather . i n (22a) the cl i ti c tri ggerS Hi gh Del i n ki n9. whi l e i n ( b ) i t under90es Low Linki n9 and subsequentlY sPreads its Low tone onto the Hi 9h tone PostPosi ti on · (22 ) a . katike nYe I " hunger i s on h i m ' la � I I L H L I I -$ h h I HIGH DELINKING ( "'4 ) b . waa wu kan � - pd -> waa I L HL h 1 I wu �,,/"'<I V kan �ud ' he 9ave it to h i m ' H LOW LINKING. LOW SPREAD. LS DELINKING Recal l that verbal i ncomPl eti ve suffi xes. anal Yzed as under l Yi n91y Hi 9h tone. were al so Mid tone after Mi d tone verb roots. as shown i n (23a) . However . th i s aPParent underlYi ng Hi gh tone does not tri gger Hi gh tone SPreadi n9 onto a noun ---- - - - _ . - .- -- --- - - _ .. __ . 312 c l ass cl i t i c ( b ) . (23 ) a. ta - r i ' be cooki ng ' b . wu Ya ta-ri wu I A -> *wu Ya ta-ri WO I A ' he i s cooking on i t ' At thi s t i me. there does not seem to be a cl ear answer �or thi s l ac k of Hi gh tone SPreadi ng. It i s Possible that hi stori cal l Y Mi d tone verbs had a di�ferent tonal ori gi n than M i d tone nouns. PerhaPs �urther devel opment ot the anal Ysi s i n the �uture can Yield more concrete answers. How to expl ai n the variety of behavi our on the noun cl ass c l i t i c eludes us at this Point in t i me. Rul es whi ch have al readY been di scussed are used i n connecti on wi th the noun c l ass cl i ti c . However . the variety o� rul es used bri ngs about a con�usi ng arraY o� hYPotheses concerni ng i ts underlYi ng tone. As a subject . i t someti mes appears to be Lh . In object Position . i t aPPears to be Hi9h f i nal i� fol l owed by a Mi d tone verb . and yet i n the same Posi tion . i t tri ggers Low tone sPread onto �ol l owi ng High tone verbs. We can explain the Low tone sPread . as wel l as i ts bei ng subjected to Low sPread and High sPread i f we Posi t the underlYing �orm as Hi. However . a Hi word normal l y does not tri 9ger Hi 9h Del i n ki n9 ( 44 ) unless i t has i tsel f been subject to subregi ster Low tone del eti on . Thi s seems to be a case where a l owered Hi gh tone. Hi, whi l e normal l y functi on i n9 a s a Hi tone. al so functi ons i n certai n i nstances as a Hh tone and tri 9gers Hi 9h Del i nk i ng on fol l owi ng Mi d tone verbs . even when there is no indi cati on that the subregi ster Low tone has been del eted . I t maY be possi b l e to come u P wi th some k i nd of solution t o expl ai n these idi osYncraci es . At thi s time. however . I shal l l et the matter rest unti l further research can b e done. ---- -----� 313 D. ye.-No Que.tign. Yes-No Questi ons are formed by addi n9 la to the end of the decl arat i ve shtement . ( 24) wa� Paon ' he came ' wa� Pan l a? ' d i d he come? ' waa k�d ' he went ' waa k�r{ l a? ' d i d he 90? waa 9b�r� ' he a9reed ' waa 9b�r� l a? ' di d he a9ree? ' The Quest i on marker i s consi stentl y Mi d tone after verbs re9ardl ess of the tone of the verb . However . alter nouns. var i at i ons can be found. as shown by the examples below : 13 me? b . wild 1 3 us? c. You , pl . ? (25) a. "d� y�d l a d . mo 1 3 You . S9. ? e. wud l a h i m? f . m�B U ri ce? 9. ceew� la a woman? h . 9baxa l a a house? i � r.djr:;(� j . m�I �l')e k. 1a a yam? 1 a the rice? 9bake l � the house? l . nd�rake l � the Yam? These el(amples 9i ve four possi b l e surface tonal shaPes for the questi on marker . Mi d . Hi 9h-fal l i n9. Mi d-fal l i n9. and Low tone . When it fol l ows a verb . one i s temPted to saY that it i s under l Yi n9l y Lh . whi l e after nouns. i t cou l d be consi dered underlYi n9lY Hi9h tone. o r PerhaPs even Low tone. At thi s POi nt. -----���- ---. ---- ----- ---- 31 4 i t i s un known what the tonal anal Ysi s �or the questi on marker mi 9ht be. It is onl Y cl ear that i ts tone shape is i ndeed a��ected by the tone of the Precedi n9 word . esPecial l y i � that word is a noun. E. Wh Que.tion EgrmAtign An d Frgnt.biftin9 Wh questi ons are �ormed by �rontsh i �tin9 the questi on word to i n i t i al Posi t i on o� the sentence. The sentence then termi nates wi th �. Below are a �ew exampl es: (26) a. san waa k.ad y� where he-VP 9c ' Where di d he 90? Q b . San waa wu nYa y� ' Where d i d he see h i m? ' y� c . Di i naa m�x€ nYe how man ' s name i s ' What i s a man ' s name? ' Q d . DH ceewu m�x€ nYe y.a ' What i s a woman ' s name? ' e. NYa?a mu Ya kun y� ' What are You doi n9? ' what: You TA do Q The tonal behavi our of the sentence � i nal questi on marker � i s not unl i ke that of the comPl eti ve aSPect marker i d i scussed i n Chapter 4. It i s general l y Low tone. However . l i ke the complet i ve asPect marker . when i t i s Preceded b y a Low tone. a Mi d-Low tone material i zes . as seen below: (27) a. san waa wu 9bara where he-TA him meet y�-> San waa wu QbirEt ' Where d i d he meet h i m? ' Q ComPare wi th y6 man-TA should __ ' Man shou l d l ove God ' God Pl ease h i m sel �-to 315 Whethe� both ot these wo�ds possess an unde�IYi ng Mi d tone o� whether theY both submi t to some tYPe o� tone i nsertion rul e (ei ther Mi d or High tone) i s a questi on that shal l be le�t unanswered at thi s time. The �ront-shi � ted Questi on Phrase does not i nteract tonal l Y with the fol l owi ng Subject noun Phrase. It the Hi 9h Spreadi ng rul e were al l owed between the two Phrases. an i ncorrect surtace torm woul d be Produced . as shown bel ow. ceewO nYe Y� ->* san c�ewu nYe� (29) I I H 1 1 L , h l L ' Where i s a woman? ' , 1 H I- h H - - --1 (!) HIGH SPREAD b . NY47a ceewO Ya kun Y� ->*NY47a c" wu Ya kun y� · What i s a woman doi n9? ' l/t vI I L H H L H h I h 1 1 t... - I H -" I <D HIGH SPREAD Thi s restri ction i s true t or anY object or adverbi al Phrase whi ch has been �rontshi tted . In (29) , � has been tront shi tted . I� i t al l owed i ts subre9i ster Hi 9h tone to sPread onto the subject cl i t i c . the cl i ti c wou l d acquire an i ncorrect Hi gh tone. (29) 9baxa wu Ya nY44 I I L H h 1 I I -> *9baxa wu Va nY44 I I ' A house he sees ' L H " '1 h O) HIGH SPREAD Tonal rules wh i ch oPerate across the sentence aPpl y � the trontshi fting. In (30a> the obj ect . �, tri ggers Hi 9h tone sPreadi ng onto the Law tone verb . whi l e in ( b ) , the trontsh i t ted object can no l onger tri gger Hi gh tone - ---- ------�--- 316 sPreadi ng because i t i s no l onger adjacent to the verb . (30) a. ndi Ya 9baxa b . 9baxa ndi Ya � -> ndi Ya 9baxa � ' I a house aaa ' � -> 9baxa ndi Ya � ' a house I aaa ' If the tonal rul es must wai t unti l after frontsh i f t i n9, then , the resulting adjacent noun Phrases at the be9i nni n9 of the sentence must have tonal boundarY divi d i ng them so that the tone of the f i rst noun Phrase wi l l be blocked from affecti n9 the tone of the fol l owi n9 noun Phrase. E. But. Ordwrin9 Rul e order i ng has al readY been di scussed somewhat bri ef l y i n ear l i er chapters of thi s d i ssertati on . It was ar9ued i n Chapter 2 (pp. 67-68) and further conf i rmed i n ChaPter 5 ( pp . 271ff ) that the l ex i cal rule of High Del eti on ( 124) must take Pl ace before the SPreadi n9 rul es . I n Chapter 3, w e saw that Associ ation Conventi ons were best ordered before Se9mental Del eti on rul es ( pp . 100f f ) whi l e adjustments made in rel i n k i n9 tones after the Se9mental Del eti ons wel�e made both before and af ter Low tone Spread . (see p. 1 16 for l i st of rul es) . Chapter 4 i ntroduced new rul es but few environments were found to test the ordering of rul es . It was noted that LS Del i nkin9 must take pl ace before RS Del i n ki n9 (p. 1 57) and that both occur at the output of Low tone SPread . In addition , it was seen that the Associ ation Conventi ons i nvol ved the l i nking of tones to TBU ' s i n a one-to-one rel ation and that the Li nki ng of anY l ef tover TBU ' s took pl ace after the aPpl i cati on of the Low tone SPread . The rul e to account for thi s add i ti onal l i nki n9 was si mpl y l abel ed Li nki ng. ChaPter 5 i ntroduced another set of rules and data wh i ch brought more --�--�- .�.- - ---- ------ 317 oPPortun i t i es to observe the i nteractions of the rul es. In Secti on IV. C • • i t was shown that the fol l owi ng rul es must take Pl ace i n the fol l owi ng order . ( 124 ) Hi gh Del eti on , ( 126) Low Del et i on ( SS ) OCP (66) Feature Swi tch , (60) Low Li n king After these rul es, whi ch take Pl ace wi th word boundari es. are the two sPreadi ng rul es: High tone SPread and Low tone Spread. Unti l now, however , two i ssues concerni ng these sPreadi ng rules have not been di scussed : 1) thei r orderin9 wi th resPect to one another and 2) � whether they take Pl ace on two d i fferent l evels, that i s, the l ex i cal and the postlex i cal l evel , usi ng the terms of l ex i cal Phonol ogy. Let us deal wi th the l atter i ssue f i rst . Do the SPreadi ng rul es take Pl ace wi thin word boundari es before they occur across the syntacti c Phrase? There are a few cases where Low Spread does occur at the l ex i cal l evel f i rst . Doubl e TBU Low-Hi gh nouns undergo what aPPears to be a Low SPread rul e . The High tone of the resul tin9 Low-Hi gh contour i s then del i nked through the process ot RS Del i nkin9, as shown bel ow: (31 ) nd3r�-ke -> ndtr�-ke -> ndjra-ke �1 I V1 I I I I L H H L H H L H H L SPREAD ' the yam ' RS DELI NKINS The resul t i ng surface tone for the noun root , then , i s comPl etel Y Low tone. If the noun ndjr.-k� were preceded by the Lh possessi ve Pronoun , �, the 5ubregi ster Hi gh tone woul d SPread onto thi s Low tone , Produci n9 a Mi d tone on both TBU ' s of the noun root , as seen bel ow: 318 ' Your Yam ' ( 32) mo nd�r�-k� -> mo ndjra-k� I V" L I ' L H H l L f.,.. ..... h h L/ I L H H I 1 HIGH SPREAD If the 5ubre9i ster Hi 9h tone were al l owed to sPread betore the i nternal Low SPreadi n9 rule. the fol l ow i n9 i ncorrect surtace torm woul d be Produced: (33) mo nd�ra-k� ->*mo ndjra-ke I I I L I I L H H 1- ... ... V H SPREAD DOWNSTEP h I I ! L L H H I 1 h Another set of examples whi ch seems t o confirm the necessity Ot applYin9 Low SPread at two different l evel s are Low tone nouns whi ch Possess both a TYPe I I suft i x and a def i n i te suff i x . I n Section IV.B. of Chapter 5. the tentati ve rule Ot Hi 9h Del i nkin9 ( 105) was i ntroduced . Wi thin this secti on . TYPe II def i n i te nouns were di scussed . As the examPl e ( 1 09) from that secti on shows. i t aPPears that a series Ot rul es are requi red i n order to produce the correct surtace form. One of these rules i s Low tone SPread . ( 1 09) nk�n-?a-k! -> nk�n-?a-k! -> nk�n-?a-k! � ... ..." L I. 1 I 1.,.;-'1 HL H L L hh h 1 h II I LOW SPREAD 1 I , t,.....-4 I H L h 1 I I L H j:. I h h Tone Col l aPse HIGH DELINKING When th is noun i s preceded by �, i t i s cl ear that the above rules have al readY taken Pl ace before the subre9ister Hi 9h tone Ot � is sPread onto the 319 noun ; when Hi 9h SPread takes place, both the Low tone ot the root and the deri ved Low tone ot the TYPe I I sutf i x acqui re a surface Mid tone (34a > . If Low Spread . Tone Col l aPse and subsequentl Y Hi 9h Del i n k i n9 had not taken pl ace before High SPread . onl Y the f i rst TBU wou l d become Mid tone whi l e the suf f i :< tone wou l d i ncorrectl y remai n High-Mi d , as shown i n (b) . ( 34) a. mo nk�n-?�-kf > mo nkan-?a-k! I I I I I I I , L L L H L L -L H - I h h L- h � .. h HIGH SPP.�AC b . mo nk�n-?i-kf -> *ma nkan-?i-kf I It I I I � I f L L HL H L HL H L "" .... ., " I , I II I hh h 1 h h <D h h h HIGH SPREAD There i s yet another examPl e where there seems to be word i nternal sPreadi ng . before sPreadi ng across word' boundaries . The word kao4 i s a H i noun meani ng ' manner ' . I t i s used as the f i nal const i tuent of a noun Phrase tar exPressi ons that wou l d be transl ated i nto Engl i sh as " how to ' . A tew examples are gi ven bel ow: (35) a. nda� so?o-kan� cYon " I know how to cook ' I-TA cook-manner know b . nda� piri- l<in a c:Y6n "r know how to sel l ' I-TA sel l -manner know c · nda� xa cel i -k�n� cY6n I-TA i t drY-manner know " I know how to drY i t ' 320 d . nda� � i n i cel i -kana cY6n ' I know how to drY �onio (a grai n ) I I-TA �on i o drY-manner know In the examPles above. note the tonal variat i on of kan4. In (a) thi s Hi noun i s Mi d-Low when Preceded b y a Mid tone nomi nal i zed verb . whi l e i t i s Hi gh tone a�ter a High tone verb and Low-Hi 9h after a Low tone verb. Thi s behavi our can be easi l y eXPl ai ned i � High tone SPreadi ng accounted for the High tone � i n ( b ) and Low tone S�readi ng for the Low-Hi 9h k�na i n ( c ) . A deri vati on of ( c ) i s 9i ven bel ow: (35 ) c . nda� x a c:�B -kana cY6n -> nda� x a cel � -k�na cY6n ��"1 H H L 1 CD h 1 I I � L SPREAD � H I I I L I L DELETION h I H I h LS DELINKINI3 When the Phrase i s preceded by a Mi d tone noun. i t aPPears that i nternal Low SPread i ng has al readY taken pl ace when Hi 9h tone SPreadi ng occurs as shown i n ( 36a) . for i f Hi gh SPread occurred before i nternal Low SPread. the former woul d block the appl i cati on of the l atter . a s seen i n C36b ) . produci ng an (36 ) a. ndaa f i n i cel i -kana cY6n -> ndaa f i n i cel i-kan � cY6n I L V I I L H H r -- (b � � H II3H SPREAD b . ndaa � i ni c:�l � -kana cY6n ->*ndaa f i n i cel i - �n� cY6n I I I , I � - Cb 1 H L h h L HII3H SPREAD L H I •J L '" I H H CD h �- - -1 I HII3H SPREAD. LOW SPREAD N/A ---- - - --- -- - -- -- --- ----- --- ------ i ncorrect 321 Agai n . i t aPPears that word i nternal sPreadi ng must t ake Pl ace betore postlex i cal sPreadi ng. For other cases i n Suci te. however . havi ng the sPreadi ng rul es oPerate first at the lex i cal and then at the postlex i cal l evel creates compl i cati ons. For examPl e. i f a Low tone TYPe I def i n i te noun i s Preceded by �, one wou l d exPect the Low tone of the root to sPread onto the sutft i x f i rst and then to have the subregi ster Hi gh tone sPread onto the noun . Unfortunatel Y , thi s gi ves an i ncorrect surtace form, as shown i n (37a) . In order to Produce the correct surface form, Hi gh tone must sPread f i rst . whi ch resul ts i n el i minating the envi ronment for Low tone SPread ( 37b ) . (37) a. mo mjl a-� -> mo mjl a-�e ->*mo m�la-�: I I L L I H (L) I 1 h I " I h L h t... 1 ...... L H (L) L h h I I I L SPREAD ... t,......1 , I, ' Your rice ' L H (L) " .. I h H SPREAD b . mo mjl a-�e -> mo m�l a-�e I I L L. h I L H , I (l) h , " HIGH SPREAD L SPREAD N/A A couPle of other examPles show that postlex i cal High SPread must take Pl ace before l ex i cal l evel Low SPread . (38) a. ----- ------ - mo s�n7e Ya ma -> mo sen7e Ya m a I L I L I ' You too are goi ng ' H L.. 1. I h Q) h H SPREAD -- -- L SPREAD N/A - - - - ------ -------- ------ 322 b. ndi Ya mo ny�- a - > n d i Y a mo nYaa I ' I see yOU ' I I L L H h CD h 1, , 1 I H SPREAD L SPREAD N/A On the one hand. there does seem to be some These exampl es Pose a d i l emma. need to PosH a Low Spreadi n9 r u l e at the l e:< i cal l evel . other cases. however . require a Postl ex i cal aPpl i cat i on . even wi thi n word boundaries . It is enti rel y Possi bl e that a sol uti on can be tound t or thi s Probl em. However . time and i nsut t i c i ent data do not al l ow tor a tul l i nvesti 9ation . The second i ssue to deal wi th here i s how t o order Low tone and Hi gh tone SPreadi ng with respect to one another . It we order Low tone Spreadi n9 t i rst . the deri vati on i n (39a) wi l l produce a correct outPut. whi l e the one i n < b ) wi l l b e i ncorrect . (39) a . wa� wu kara wee I 11. , .. .... HL H II 1 1 1 Q) L SPREAD H H I h 1 & I I L DEL -> wa� wu kara wee 1 1/1 HL H II I 11 h .. I ... H I I h I h L I LOW SPREAD L I H I h L 1 \.....-10 II L H <D h L-1 J H � .� 11 h LINKING '� ... HL H ::l I UJ h b . wa� mo m�l j waa ->*wa� mo m�l � wee L,.."- 1 waa wi} ka ra: wea H H SPREAD I -> & H I Ii RS DEL INKING ' he saw Your ri ce-I ND ' I H IGH SPREAD. LINKING. <See ( 40b ) for correct surface form > & LS DELINKING On the other hand . i t Hi9h SPreadi n9 was ordered t i rst . (40a> wou l d be i ncorrect whi l e ( b ) wou l d produce the correct surface form. - ---- ------.--- 323 (40) a . waa wu kara wee -> waa wu kara wee ->*waa wu kara wee II \ 1 L. .... 1 \ 1 I 1 V1 I HL H H H HL H H H HL H 11 1 1 h 1 1 (D 1 h 11 I II I I I It H SPREAD NIA L SPREAD I & L \... h I I \ L H L CD h h - '\ I L V""" I HIGH SPREAD I " .J '1 I H 1 H 1 h (see ( 39a) tor correct surtace h t orm> LINKING L DEL b . waa mo mjl j wee -> waa mo m� l � wee I II , '01 ' he saw Your ri ce- IND ' H I h L SPREAD NIA 8c L INKI NG Theretore. i t aPPears that: stri ct: orderin9 tor these two rul es Produces i ncorrect resu l ts. If . however . Hi gh and Low SPreadi ng operated i n some type ot cyc l i c tashi on across the sentence. where ei ther Hi 9h or Low SPread would occur f i rst. dependi n9 on whi ch cou l d apply t i rst i n scanni n9 the sentence trom l ett to right. a correct surface torm wou l d be Produced i n each case· Note that i n ( 3 9a) above. Low tone SPread was the f i rst aPpl i cabl e sPreadi n9 rul e encountered . whi l e i n (40b > . Hi 9h SPread occurred t i rst: . Bel ow are a tew more e:<ampl es to i l l ustrate thi s i nteracti on ot rul es. ( 41 ) a. wu Ya wu k�n �u-u I It I II I I , \ " I I H LH H LH 1 1 Ih wu Ya wu kaan �uu -) wu Ya wu kaan �uu 1 A , Q) I h Ih Ih Ih 1 " HIGH SPREAD II LH HH 1 1,.. ... 1 1 I A I 14 H LH H I IV 1 Ih HIGH SPREAD ---- - --- - -- --- -- Ih & I HH 1 -> wu Ya wu kaan �UU I A II H LH H LH \ I I, '" lh II �1 HH H LH H h I, LH H H \ \ Ih -> 1\ I IV 1 Ih I I 1 \ LOW SPREAD ' He i s 9i vi ng i t to h i m ' lh RS DEL INKING --- ---- ---- ---- -> 324 b . wu Ya mo kaan I 1\ I II H L.H L LH I 1 II In I n II In wu Ya mo k aan I)Uu HH H LH L LH HH H LH L In 1 In 1 11 1 I)Uu II -> 1 1\ I II II I 1 I 1.. ..... , , h (»11 Ij wu Ya mo kaan -> ! /\ II LH I I I I/ i,... It HIGH SPREAD I 1 11 n h I)uu It HH -r 1 � L SPREAD N/A. HIGH SPREAD " He i s g i v i ng You to h i m ' After the sPreadi ng rul es . then . Li nki ng rul e ( 39 ) takes pl ace. From Chapter 4 we know that the L i n ki ng rul e was devi sed as a seParate Process trom the Associ at i on Conventi ons so that extra TeU ' s woul d not be l i nked to tones bet ore the aPpl i cati on ot Low tone SPread (see pp. 1 SS) . Atter Li nki ng. LS Del i nk i n9 takes Pl ace. An examPl e t rom ChaPter 5 i l l ustrates tile need to order Li nkin9 betore LS Del i nk i ng. In ( 42 ) bel ow. IndePendent Low Linkin9 takes Pl ace . resu l t i ng i n botn tne Mi d and Low tone l i nked t o the same Teu. Normal l y . thi s i s an accepted contour tone i n Sucite. However . i t the t i rst tone of the contour i s l i n ked other TeU ' s of tne morpheme. then the l et t si de tone can be del i n ked . ( 42 ) Thus. Li n k i ng must take pl ace f i rst . kara I H I 1 -> kari /l H L t-. ... - 1 L LINKING } kari 1/1 -> kara � H L H L 1 1 V LINKING " meat ' V LS DELINKING Nex t . RS Del i nkin9 takes Pl ace . The Hi9n Del i n k i n9 rule ( 44 ) wou l d al so take Pl ace at thi s stage ot the derivation . It i s cruc i al l Y ordered i n reterence to Hi gh tone SPread and Low tone Spread . As the e:<ample bel ow shows. Hi 9h SPread Provi des the envi ronment tor the Hi9h Del i n k i n9 0+ the Mid tone verb . ----- - - --_. - .- ------- ------- ----- ------ 325 (43) waa mo kara nYa -> waa mo I I I L H I L L h 1 h h I I I �ri nYa -> waa mo k� ra nYa " he saw Your meat ' l.. ... I I I H L L '� I L H � 1 I Q) h ... * h h LINK ING HIGH SPREAD � HIGH DELINKING The Low tone Inserti on rule (27) ( i ntroduced i n Chapter 3) whi ch Produces a f al l i n 9 tone on Hi 9h f i nal nouns i n sentence f i nal Posi t i on wou l d al so be ordered anytime after Hi9h Spread . Fi nal l y , as menti oned before. the Downstep rul e takes Pl ace. The deri·vati ons bel ow are ex amPl es of this orderin9 of rul es. (44) a . waa �muu-n' ca wu nYU�-9' mpa�a-i -> waa �muu-na ca wu nYU�9' mPa�a-i II ! / � .. 1 h 1 <D 1 HL H H (L ) L H 11 1 I " I I LOW SPREAD ! I ! H H (L ) I I II I I I 1 H H (L ) L H 1 11 1 II h l £> € l Eft O '" ... 1/1 HL ! I I h I / H L ' I I H H (L ) 1.. / 1 I , 1 h CD h h I H L 1 I h HIGH SPREAD -> waa �muun' cO wu nYU�9' mp.a��-i -> waa �muuna c� wO nYU�ge mpa�! i I , V I � , / �! " / / l/1 I I ,:" .,o j I HL H H (L) L H 11 1 II H H (L) I t.--'" ' I 1 h h h H L HL H H (L ) L H 1 11 1 1 I I I I h LOW SPREAD I ' I 1 (he-TA knife hel d h i s head above-PP) b . b i � Ada saka-a-bf 9b6 -> I I I � I HL L H H H HL H (L ) H 11 1 h ! HL L 1'1 11 1 I 1 hI h II !I I hh h J 1 h " I I II l ! h 9b6 -> bi i nda saka-a-bi 9b6 I I t\ f I H 6i) HL H <L) H I I T il I 1 1 hh h HIGH DELETI ON - -- --- ------ DOWNSTEP " He held th� knife above h i s head ' bi i Ad' saka-i-bi II I 1 1 HL I LINKING. LS DEL INKI NG. RS DELI NKING. � 11 ' 1 H H (L) V h II v1 I 1/1\ f I I h HL L H H L HL H <U H II , I 11 1 h 1 1 1\ 1 1 hh h L SPREAD < Lex i cal ) I 326 -> I t vt 1 &..-1 1 - I > bi i �da sakaab! j l V'\ l lA l 9b6 I -> bi i �da sakaabi 9bo I I V1 I 1'1 I .... '" HL L H H L L H (L) H HL L H H LL H (L ) H HL L H H LL H (L ) H l1 l h l l h h h l1 l h CD l l h 11 1 h II ! I ' I � Tone Col l aPse -> 9b6 bi � �da saka-a-b! & I H DELINK bi i �da sakaab! 9bo �� � ' VI '� HL L H H LL H (L ) H I, I , I, I I l1 l h ll h II I I a.. .. 1 I ' I I h HIGH SPREAD II I V II I 11 h I h LOW SPREAD (theY mY 90ats ki l l ed ) ' They k i l led mY goats ' � RS DEL INKING G. Coocludioi Not•• Thi s enti re di ssertati on has consi dered i n detai l the tonal behavi our wi thi n a simpl e sentence o� Suci te. a Senu�o dialect . It was seen that some Probl ems cou l d be easi l y sol ved whi l e others de� i ed a cl ear anal Ysi s . The theoreti cal apProach used here extends the autosegmental aPProach i nto the i mpl ementati on o� two tonal t i ers. borrowi n9 �rom the recent developments i n non-l i near Phonolo9Y. I t was seen that thi s apProach great ly enhanced the abi l i ty to adequatel Y describe the tonal Phenomena o� Suci te whi l e at other Poi �ts. some na99i n9 questi ons sti l l remai n unanswered . Further research wi l l b e requi red t o Pursue these questi ons whi ch . i � answered . cou l d contri bute to mod i � i cati ons o� the above anal Ysi s. It is mY hoPe. however . that the Present descri ption and anal Ysi s o� Suci te wi l l serve to Provi de a sPri n9board �or �uture anal yzes o� tone in Senu�o l anguages as a whol e. The doubl e ti ered aPPProach i tsel � needs to be tested � or i ts vi ab i l i ty i n other l an9uages. However . whether the double ti ered approach hol ds uP i n other languages or not . i t does remai n cl ear that a trad i t i onal autose9mental aPProach to anal Yz i ng the ------- ---- - 327 tone of Suci te fal l s terri bly short of adequatel Y descri b i ng i ts comPl i cated tonal Processes. Again it i s hoped that further i nvesti gat i on and the presentat i on of more data wi l l resolve the questi ons that remai n unanswered at thi s Poi nt . NOTES 1 . The reader maY also recal l that Hi gh tone spread as a rul e i s not as Producti ve between subjects and verbal parti c l es as el sewhere in the sentence < See ChaPter 4 ) . J, . ------ .-�-- --- ._--_ . _ - -- - :328 BIBLIOGRAPHY Anderson . StePhen R. ( 1 978) . "Tone Features . " Li n9u i sti c SurveY. PP. 133-175. V. Fromki n <ed . ) . Bal l a Dial l o ( 1 981 ) . Atlas du Mal i , Les Atl as de Jeune A�r i que. Rel l ure Tours. 2�me Tri m , Paris. Ione: A ImPri mer i e Bendor-Samuel , John T. ( 1 971 ) . "Ni ger-Congo. Gur . " Thomas A . Sebeok (ed . ) . Curr@nt Ir@nds i n Lin9ui stics, Vol . 7. Mouton . Par i s . Clements. G. N. ( 198 1 ) . liThe Hi erarchi cal RePresantaion ot Tone Features, " i n G. Clements. ed • • Harvard Studies i n pbonolggy. vol . 2. I ndi ana Universi tY Li ngu i stics Club, Bl oomi ngton . Cl emants. G. N. and J . Gol dsmith (eds. ) ( 1984) . �. Fori s Publ i cat i ons. Dordrecht . Aytgs@9m@ntal Stydi es i n Bantu Ewen . Col i n and John M . Anderson . ads. ( 1985) . Cambri dge Un i versi ty Press, London . Pbgnglggy Y@arbgok Fromki n . Vi ctor i a (ed . ) ( 1978) . London . Greenberg. Joseph H. ( 1966 ) . Bl oomi ngton . Ign@: A Ibl l i n9ui ?tic aurv@Y. Lan9uag,S gf Africa. 2. Academi c Press. Indi ana Uni versi tv. Herau l t . George. d i r . ( 1 983 ) . Atlas d@a l angU@? Kwa d@ cat@ d ' Ivgir@ Mgnggrapbi @s . Inst i tut d e Li ngu i st i que Appl i que. Abi djan . Tom! 1: Herau l t . Georges ( 1 983) . Atlas diS l angu@s Kwa d@ cat@ d ' I voi re Iom@ 2: List@s l@xica1@s @t etudes cgmparati v@s. Insti tut de Li ngu i st i que Appl i que, Abi dj an . HYman . LarrY ( 1 985) . "Word Domai ns and DownsteP i n Bami l eke- Dscbang. " i n C. Ewen and Anderson, eds • • phongl ggy Yearbook 2 . Pp . 47-85. HVman . L. M. and R. G. Schuh ( 1974) . "Uni versal s oi tone rules: West Air i ca . " Lingui �tic InqUirY 5: 81- 1 1 5 . eVi dence from KaYe. J . and H. KooPman ( 1982) . "Les tons du sYsUme verbal en Presented at 1 3th annual con . Air . Li ng. , Montreal . baU <gbadi ) . " KaVe. J . . Lowenstamm. and Ver9naud ( 1985) . "The i nternal structure o� Phonologi cal el ements: a theo�Y oi charm and government. " i n C. Ewen and Anderson <eds . ) phgnglggy Y@arbggk 2. Cambri dge Uni versi ty Press. Cambri dge. KiParskv, Paul ( 1982) . " From CVc l i cal Phonol ogy to Lex i cal Phono l ogy. i n van der Hul st ( 1982) . Pp. 131-176 . - ---- ---------. ------ 329 Lavergne de Tressan ( 1953) . Inyantaira l in9ui sti9ue de l ' Afri9ue gbcidentale francaise It du Tggg . M'moires de l ' IFAN. 00 . 30 . IFAN, Dakar . Le Saout , Joseph ( 1 976) . Etude descriPtive du Sban (Cete d ' Iygire) pbgneti9ue et Pbgoglggigua. SELAF, Par i s . Maddi eson , Ian ( 1974) . "A Possi b l e new c:ause at tone-sPl i Hing evi dence t ram Cama, Yoruba. and other l anguages. " The Tgne Tgme, ucLA Wgrking papers in phgneti cs , 27 : 28-46. ManessY, G. ( 1975 ) . Les Langues oti -Vglta. SELAF, Pari s . Mensah , E . N. A . and Z . Tc:hagbale ( 1 983 ) . Atlas des l anguea d ' ivgire. I nsti tut de Li ngui stisue Appl i 9U�, Abi djan . Sur de cate Odden , David ( 1986 ) . "On the rol e of the Obl i 9atorY Contour Pri ncipl e i n PhclOol o9i c:al TheorY, " i n , Langua.ge 62/2: 353-383. Person . y. ( 1 966) . "Des Kru en Haute-Vol ta. " nos. 1-2: 485-492 . Bul l. de I ' IFAN. T 28. ser e B, Pul l eYbl an k , Dougl as G . ( 1983) . Tgne in LeXical phgnglggy. Phd . Di ssertat i on . M . I . T • • Cambri dge. Massachusetts. TrHkovi c: , Mi r jana ( 1977) . " Tone Spl i tting: Lendu . " Li ngui stics, SuPpl ement 7. pp. 223-234. studiea in African van der Hu lst , HarrY and Norval Smi th (eds. ) ( 1992) . The structure gf pbonglggical representatigns (part I ) . For i s Publ i c:ati ons, Dordrec:bt . westermann . Di edri ch and M . A . BrYan ( 1970 ) . Languagea gf Weat Africa. Part I I . Internati onal Atri c:an Insti tute, London . 330 BIBLIOGRAPHY DE SENUEO LANGUAGES Anonymous ( n . d . ) . Lan9uage. ms. A Grammat i cal Sketch o� the Kati ola Di alect o� the Ta9bana AnonYmous ( 1961 ) . Wgrd Li st : TYgbara-En9I i sb . Mi ssion Bapti ste, Korh090. Boese, Li nnea ( 1993) . Nyarafglg Narra tiygs and prgcedyrals: anal v$i s. M. A thes i s , Uni versi ty Mi cro� i lms , Ann Arbor . Boese, Li nnea ( 1 993) . A discgyrse Descrpti on o� Basic NYara�ol o SYntax . ms. Bouti n . Pi erre ( 1991 ) . El 6ment$ pgyr D . E . A . , Uni versi t' de Paris I I I . yng $Y$temati 9ye Th�se du dy E�nd�nd�. Bouti n , Pierre ( 1992) . "Rel ati ons de d'i:ermi nati on en �odonon (Parl er s'nou�o de la r'9ion de Di kodoU90u , Cate d ' Ivoi re) . " Afri 9yg gt Lan9a9g 1 9 . 2 : 5-36. Burdon , Stephen and Claud i a ( 1994 ) . ms . Tbg Syllable, Wgrd and Tgng i n Tgu$si an . Carl son , Robert ( 1983 ) . 14/1 : 35-46. "Downstep i n SuPYi re. " Carlson . Robert ( 1980) . A descri pt i on of SuPYi re tone. Carlson , Robert ( 1983) . Narrati vee ms. The �unct i on o� the Con junctions k4 and Stydie$ i n Afri can Li n9yi $H cs draf t . Carlson , Robert ( 1996 ) . Grammati cal i zed Verbs i n SuPYire. 16th Conference o� Afri can Li n9u i stics, 1986. Carlson . Robert (n . d ) . Lan9uage Learni n9 Lessons (SuPYi re) . Carlson , Robert ( 1981 ) . Sketch o� SuPYire Noun Cl asses . ms. Car l son , Robert ( 1985) . Supvire Tgne. Cheron , G. ( 1 925) . Le Di al ecte Paul Geuthni er , Paris. ma i n SuPYi re Presented at the ms. ms. sengyfg dy Minianka. Li brai r i e Or i ental i ste Cl amens, G. ( 1953) . " Des Noms Vernacul ai res du TVPHLOPS dans quel ques d i al edes Senoufo. " Ngh$ Africaine' 58, 59. Cl amens, G. < 1951 > . lEAN 52 : 120-122. "AnthroPonYmie nYara�olo. " Ngte$ afric;aines, Cl amens, R.P. ( 1 952) . "Essai de 9rammai re senu�0-ta9wana. " l ' IEAN XlY, no. 4 : 1402- 1465. Cl amens, G. ( 1950) . By ll eti n de. "Des noms de Petrsonnes en dial ecta t.a9wana. " � 331 Cl amens . G. ( 1951 ) . "Lan9ue!l secr�i:es du Poro. " Ngtea Africaines. IFAN 52: 93-4. Cremer . J . ( 1 919) . "Essai sur l a langue Mi nianka. " SCient. de l ' A . Q . F 4: 560-61 6 . Garber . Anne ( 1980) . Bull . Cpm. Et- hia• vt "Word order change and the Senu-Fo l an9uages. " Studiea in the Li n9uj ,tic Sci encea. 10/ 1 : 45-57 . Herau l t . G • • et J . Ml anhoro ( 1973 ) . "Le TakPE r (Ta9bana de Ni akaramandou90u ) EsClui sse Phorlol 09i Clue et corPus lelCi cill . " UniVeraitR d ' Abid ian. Annalea. Serie H. 6/ 1 : 1 33-178 . Holas. B. ( 1957) . Lv, SenyfPa (Y cgmpris l ea Minianka) . P. U. F. Par i s . Jordan . Dean L ( ? ) . NYaratol o Tense-AsPect i n the Fol k Tal e. i n Discgyrse, PP. 84-90. paPera gn Lau9hren . MarY ( 1973 ) . AnalYSt p16remati9ue dy hebari (yo dialede a9nufp) . Th�se du Doctorat du 3e cycle. Ni ce. Lau9hran . MarY ( 1 976) . " Serial Verbs. " n o . 4 : 872-889. Byl l. de l ' IFAN T . 38 , ser . B. Mi l l s. El i zabeth ( 1 978) . The Senu-Fo Noun and Pronoun. ms. Mi l l s. El izabeth ( 1 979) . Theme in Senu-Fo Narrati ve Di scourse. ms. Mi l l s. Dick and Betty ( 1967?) . mi ssi on . Korh090. ManesseY. G. ( 1966) . TYEbara Grammar . The Conservati ve BaPti st "Recherches sur l a morphol 09i e du verbe senu-Fo. " &WJ..:.. �. 28 (9) 3/4: 690-722 . Prost . R. P. A. ( 1 964 ) . M. I . F. A . N. no. 70. Contri but i on IFAN. Dakar . � l ' etuda des l an9ues vol tai Clues . RaPP. E. ( 1933 ) . "Di e Na-Fana-sPrache au-F der EHenbei nkuste und au-F der Gol dkusi:e. " M . S . Q. S. 36, 3. 1 933: 66-69. Wel mers. W. E. ( 1 950 ) . "Notes on two l an9uages i n the senou-Fo 9rouP. I Senadi . I I SUp l i de i n Lan9uage. " Jpyrnal gf the Lin9yiaHC; Spdeh pf Amer i Ca . vol . 26. no 1 : 126-146. and no 4 : 494-532 . Wi chser . Madl en ( ! 983) . Deac;riptign de Karabgrp. ms .-- -------- • 332 APPEND IX . The d i acri t i cs f or each tone are 9i ven bel ow : Low ... Mi d unmarked vowel Hi 9h .. Low-Hi 9h R i se Low-Mi d Ri se Mi d-Hi 9h Ri se v ... � Hi 9h-Low Fal l Mi d-Low Fal l Hi 9h-Mi d Fal l NOUN LEX ICON OF SUCITE Each noun has u P to f i ve entr i es each separated by a semi -col on . W i t h i n each entrY two possi b l e forms maY be 9i ven . seParated by a comma . These i nd i cate a var i at i on i n Pronunci at i on of a parti cu l ar word. The symbol ?T i nd i cates that the tone mar k i ng i s uncertai n . The f i ve entri es. i n order of aPPearance . are : Si ngul ar Indef i n i te � S i n gu l ar Def i n i te � Pl ural Indef i n i te; " mY <nda> - Si ngu l ar Def i n i te ' . Pl ural Def i n i te ; Some nouns have no Pl ural . The l ast entrY i s gi ven onl Y f or a restri ct ed number or nouns· b. b�?�; b�?�ke� b�?aya; b�?anYE : Poi son b�l � ; b�l a�a; b�l a l a . b�l aYa; b � l ib ! . b � l inY€ : bal l . bul l et? bando?�; b�ndo?o�a; r�ndo?oYS; bando?onY€ : sesame b�rant�n ; b�ranta �e; b�rantanYa; b�rant anY€ : banana b6 � x �; bar�ke: strength bal e ; ban e ; b � Ya; b � Yaki : seed bal € ; bana; b al e ; bal aki : 9round Peas balou. bel u?T ; bal arye ; bal a l e . bel a l e?T; bal ab i ; nda bal arye : sl ave 333 b�rax � � b�r� ke; b�raY� ; b�ranY€ ; ????: ba9. sack bobQ; b ob u�e; bobol � J bobou b i : mute bOroxo ; buruke; buroyo; burunY€ : bread bux uro� buxute: addi t i on to a vi l l ag e C. c�; c�rye; n��koor� ; nikoote: c h i l d caan J caanke; caan ' Ya; caannY€ ; nd� c atnke: mar ket c�l � ; cal ary� ; cal i ya; cil � nY€ : beans cal ou ; cal a�e; cal �l a ; cal Sb f ; nda c al a�e: Pi 9 canc6l �; cAnc66ne ; canc6761 � ; canc676kf : Pi Pe ( t or smoki ng ) cantory8; cantOrygeJ cantony8: cintonY€ ; nda cintoryge: umbrel l a cary�; ciryge J cannYi; cinnY£ ; nd� canbane : day, sun . . . cary�nY�I € ; ca��nYi i ne > ca�en Y- ; c a��nYi i ; ca� �nYi � k ! ; ?? sun, I i t . day cedE ; cene; ?cer�I E Nc�ral e ; ?cerSk f Nc�r�k ! ; nda cene: cal abash . gourd ceew�Nceewu; ce�rye; c�e; c��b i ; nda c ��e : woman cef 6 1 9 ; cet61 ��e; cef�e; cef��b f : d i vi ner ce9burox o ; ce9buroke; ce9bor i Yo ; ce9boranv€ : P i ec e of wood -? cere; ceteJ nd� cete: bodY c�n�� ; c� n�e; ct � nmi ; c � � n mb ! : antel ope c�nrax � ; ctnr�ke; ctnr�Y� ; ct nranYe: l i on CE nXt ; ct nke; CEn ' YE ; ct nnY€ ; nd� c€ n k e : sauce cEnxa�r�x � ; cE nxaY03 rake: sauce i n9redi ents c i j er�; c i jete; ??: adul terY < of man ) ci k�n ; c i k�n�e; c i kuu n ; c i kuunb f ; ??: b r i d e . l i t . exci sed woman c i k�nr � ; ci k3nte; nd� ci k �n t e : weddi ng c i l � ; c i i n e ; c i x al e ; c i x �k ! ; nd� c ! ! ne : t h i 9h ca9boraxo ; c�9borake; ca9boraYo ; ca9boranY€ : c l ean i ng . cal abash ? c i l uu; c i l a�e; ci l �l e ; c i l �b ! ; nda c i l ��e: Potter ci nnde@; ci nndee�e; c i nndeel E ; crnndeeb i ; ??: el der woman c i P� I � ; c i P� I ��e ; c i Pee; c i P�eb i ; ??: groom c�r��E ; cera�e; c�rami ; cerambi ; nda cera�e: orphan c i wu l 8 ; c! woone ; c ! wuu . cuwuu?; c ! wuuk ! : week col � ; coOne; co?ol � ; co?ok i ; nd� c66ne : c l aY Pot con l � ; con l a�e; ci i nl E ; c i i nb ! ; nda conl a�e : Younger brother . si ster ----_ .- . _- _ . _ . _ ---- eYe of the 334 c� ; c�rye; c�ol j ; c��bi : net c�nr� ; c��nte; c�n ' y� ; c�nnY€ ; nd� c55nte = ashes � d�9bAnry�; d�9b�ryge; d�9b�nnY�; d�9b�n YE : gi nger d�?�sal u ; d�?�sal arye; dA?Asal al a; d�?�sil ab f : t i sh < sPeci es > dA?�s6nryo; d�?�s6nryge; d�?�s6n ' yo ; d�?as6nnYt : toad danaa; danaa�e; danaaya; danaanYE : red PePPer danb�r�ryS; dAnb�r�ryge: di rt das�l � ; das�l ��e; das�l o; dasul �b f : e l ePhant d690n l aryo , -nk-? ; d690nl aryge; d690n l anYo; d690n l anYt : duck domp�ry� ; domp�ryge ; domp�nYi ; domp�nYE : vul ture d�x�nca; d�:(�"carye; d�x�ncaa; d�x�ncaab f : sheep £ taac i i � t aac i i rye ; f aaci i ; f aaci i b i ; nda t ac i i rye : cu l t i vator tal a ; fal a�e; ?????: agr i cu l ture. cu l t i vati ng fa l axa; f� l ake; ta l f ya; t a l f nYE : rock tan7an f �l 9 ; fan?anf�l orye; f an?an f ee ; fan?anfeebi : k i ng t an l a; f an l ��e; f an l al a; f an l �b i ; nda f an l �rye: crutches tan�a; f an�ge; fan n Ya; fannYt ; nda faryge: cemeterY, grave teramE ; f erambe; nda ferambe : urine t E nl E ; f i i ne; f � n?E nl E ; f�n?� n k f ; n d � f i f ne : Pen i s ?? f � � ; f � i rye ; f � i l E ; f i � b i : python f i i n ; t i i nrye; f i f n l e, t i t nmi ; f i �n b f . f i €nmb i , ?t f i nb i nd a f i i nrye: t i nE ; t i na�e ; nda f i na�e: f on i o . grai n from crabgrass t ami l y f a l �rt; fal �rarye ; t al ar i ye ; f al ar�nY€ : f l ower ( Fren c h ) f� l� x e ; t �l �ke; fal � Ye; f al anYE : mat f an E tar � ; tanE tararye; fanE taral e ; t anE tarab i : wi ndow ( French ) f �ra; f ararye; f ar.a l a; f arab f : wi nnow i ng basket fare; f ate; f anY� ; f anYE ; nda f at e : excrement f okanzaxe; f okanzake; fokan z i ye; f okanzinYE ; nd� t6kanzake: wi n9 f 6 l 6k i de; f6l 6ki daki : l egend f ol � ; f ol a�e ; fee ; f eeb f ; nd� t6l a�e : owner . c h i ef f ol � ; f oone; to?ol j ; torok ! ; nda f66ne : debt fon l a ; f ��ne; t�u n ; tu�nkf : Peanut fonry�n ; fOnryge: Poverty b l i nd Person 335 tOX O ; toke; t UYO; tunYt : corn tU; tu�e; tuu l o ; f uubi : 9reet i n 9 tuced� ; t uc�n e ; tuceral t ; t uceraki ; ? ? : stomach (gener i c ) tUtt n?E n ; tutt n?t n k'; t utt n?tnYt ; t utt n ?t nnY€ : pal m l eat tU9bel � ; tU9b" ne; tU9b€x€ l t ; tU9b€ x ak ! : Senuto basket tujokul o ; tujokuun'; tujokoxal o ; f uj okoxak f : cuP tujox � ; fujoke; tujOY� ; tujonYE : water Pot f u kanx A ; fu kan ke : f ukan ' yA ; tukannYt ; nda f a�n ke : shou l der tUnkYax E ; tunkYaxak'; t un kYaxaYt ; tunkYaxanY€ ; nda tunkYak': d i arrhea f unm�n ; f unmb e : sweat ' fu�� ; tU�ge; tunY� ; tunYt ; nda fa�ge : i nsi des tutax e; tutike; f u t i Ye ; futinYE ; nda fatak e : trai ner (tor Yam pl ants) t v.aa ; t y,ia�' ; t yaal a ; fyaab i : f i sh f Y�x u ; f Yexa�e; t yex a l e ; f Yexamb! : earr i n9 tYE nx� ; fy?n k e ; f Yen ' y� ; t YenY� ; nda ty€ n k e : worm ill 9ba ; 9ba�e; 9baal a; 9baab i ; nda 9ba�': r i ver 9babal @; 9babane; 9bab i ya ; 9bab i yakf ; nda 9babane: room (ot a house ) 9ba l A ; 9baAne; 9bA?al a; 9bA?akf ; nda 9b aane : t orehead 9bAnmbal a�� ; 9bAnmbal u�ge ; 9bAnmbal anY� ; 9bAnmbal i nYE : bamboo 9bAram9bAbal @; 9bAram9bAban'; 9bAram9bAb i Ya; 9bAr�m9b�b i Yak i : l i ce 9bArati l ax@; 9bArat i l ake; 9barat t l i Y@ ; 9barat i l i nY€ : toP of the door 9bArax A ; 9bArake; 9bar i yA ; 9bAri nYt : door 9bas6nx9; 9bas6n k e ; 9bas6n ' Y9; 9bas6nnYE ; nda 9bas6n ke : bedroom 9bax a; 9bake; 9baya ; 9banY€ � nda 9bak': house 9bes�; 9bese�e; 9bes€€ I E . 9bes{Ye; 9bes€ t b f . 9bes�nY€ : machete 9bi i n ; 9b i � n b e : l i 9ht 9bal ax e ; 9bal ake; 9baB Ye; 9baB nYE : wel l 9bo; '�bur'/e ; kul o ; k u l ob :i ; nda 9bu�e: corPse 9bon l � ; 9buunne; 9b�n?j l j ; 9bon?ok { ; nda 9bounne: t i repl ace 9boxo; 9boke; 9buyo ; 9bunY€ ; nda 9bo k e : tam-tam 9b�n�� ; 9b��ge; 9b�nnY� ; 9b�nYE : 90ri l l a 9bun . on 1 . 00n1; 9buunbe; 1??111? : tuneral 9bunna�o; 9bunna�ge ; 9bunnanyo; 9bunnanYE ; nda 9bunna�ge: wal l 9busAn ; 9busan�e ; 9busaan l a ; 9busAan b i : trog 336 i ja; ja�e; ja�l a ; ja�b� ; nd� j a �e : son jat�x�; jat�ke; j atoY� ; jatonY€ ; nda jat�ke: roof jaxaci i . j�l<�9ban ; jaxaci i �e; jaxacH l e; ja:<a c i 'ib� : trader jexe; jeke; j i Ye; nd� jeke: soaP j i de; j i ne ; j �r��i ; j �r��9i ; nda j i neN?nd� j � ne: breast j � funam@; j � f un�mbe : potassiu� j � �zi ; j � �za�e ; j i i z i Ye. j i i z i � l e; j � � z i nY€ . j i � zii b� : thread (Di oula) j � l �xe; j � l �ke; j � l i ye; j i l i nYE : nut jaramE ; jarambe; ???????: mi l k jokurax8; jokuroke ?T; jokur ! y8; jokuri nYt : mumPs jol � ; j oone ; j o?ol � ; jo?okf ; nda j66ne: testi c l e j S ; j��e; j661 � ; j6ob ! ; n d a j��e: pocket j�l ax�; j�lake; j ol i Y� ; jol i nY€ ; nda j�lake: sewi n9 juu ; juube; nda j 66be: speech k kaafoxe; kaatoke: outsi de vi l l age l i mi ts kaa taxe� kaa t ake; kaatiye; k�atsnYE : bri dge kacer@; k�cete ; k�c i Y@; k�cinY€ : stal k <of a pl ant ) kacEnnE ?; kacEn�e; ????????: the 900d kacE nl E ; kac i i nne; kacE €n�i ; kacEfn�9f ; nda kac i i nne: fet i sh kac i l axe ; kac i l �ke; kaci l i Ye ; kaci l ?nYE ; nd� ka c! l�ke: bone kac i l axe; kac i l �ke; kaci i re ; kaci � te; nd� kacf l �ke: skel eton kac�x� ; kac�ke; kacoY� ; kaconY€ : Pen or encl c5ur� kacu; kacu�e; k�cul o; kacul ub� : mouse kad�rt; kad 'ra�e; kadar�l e; kadarabi ; nda kadar��e: door trame katanl a��; kat�nl ��ge; k�t�n l �nY�� k�t�nl anYE : donkey katol S; kafol ��e; katee; kateebi : l ord kat6r8; kaf6ra�e; kaf6r l YO ; kat6ranY€ ; nda kaf6ra�e: e99Pl ant katulo; kafuune; katol<ol o ; kato:(ob! ; ??????: surPrise katuxo; kat uke; katuYo; kafunYE ; nda katuke: vaPor . steam . hot season katYenanji de; kat Yen�nj i ne ; katYenanj � ral e; katYena"j � r�kl ; nda katYena"j i ne: whi r l wind harmaHan wi nd s . cold season ----- ----------- ------- 337 ka9bol j ; ka9boone; ka9bo701 � . -9box aI E9 ; ka9bo7ok ! ; nda ka9b66ne: mi rac l e k�j i i x e ; kaj i i ke ; kaj i i Ye ; kaj i i nY€ : f i rewood kaj i k�rax � ; kaj i k�rake; kaj i k� r i Y� ; kaj i k�ranY€ : bark kaj i nn� ; kaj�n n e ; kaj i nn e ; kaj 1 nnak! ; nda kaj ! nne: bracel et kaj 1 p� I � ; kaj1 P66ne; kaj i Po7ol � ; kaj i Po7okf : bund l e ot f i rewood kakonx � ; kak� n k e ; kakon ' y� ; kakonny€ : col d . cOU9h kal ap61 � ; kal aP66ne ; kal aPo7ol � ; kal apo7ok { ; nda kal ap66ne : sheat Ot sor9hum kal al(a; kal ake; kal aYa; kal anY£ : sor9hum kal el � ; kal e�n e ; kal e��i ; kal e��9! ; ???????: customs kal ox j ; kal oke; kal �Ya; kal �nY£ : shower room kama� rya; kama7aryge; kamS � nYa; kama ?anY€ : pai l l e en corne** kamene; kamene ; kamE € �i ; kam£ nY€ ; l ett hand : kamene ; kamene; kamee�i ; kamee�9f : sParrow hawk k a merye ; ka meryge; ka meY£ ; kamenY€ : dew kam9b H �; kam9b H ne ; kam9b U I)O kam9b i 'i1')9f : man ' s cane kam61 u ; kam6 l arye; kam66mi ; kam6omb ! : ant kam6nzax e ; kamSz i ke ; kam6nz i ye ; kamSz i nY € ; nda kamSz i ke: ant wi th wi n9s kam6onyel u ; kam6onYi i n e ; kam6onyere ; kam6onyeete : ant ( sPeci es ) kam6%or�; kam6z orarye ; kam6zoral o ; kam6zorab f : ant ( spec i es ) kam� "tdxada; kam�"t6x ane; kam�"tdxar6; kam�ntdx ate : anteater kampel � ; kampe�ne ; kamp�7e I E ; ka"'p�7�ki ; n d a k a"'peene ; t i nger ka"'pena�93; kampena�9��e; ka"'penary9�!Sb. -9aYS ; ka"'penary9jSb i ; Pouce: thumb ka"'panna�� ; kampannaryge; ka PanannY� ; ?ka Pan 1 nYe ; n d a kapana�ge: st i ck . stake ka"'p� rary� ; ka"'p( raryge ; kamp� ranY� ; kamp� ranYe ; nde kap�r( ryge: cri c ket ka,npul o ; ka"'puilne; ka"'poxel o ; ka"'poKak i ; nda k�"'puune: musc l e ka"ci nry( ; ka"c?n�ge; ka"c i nnyt ; ka"c?nnYE ; nda k a"cl ryge: nai l ( t i nger ) kandax e ; kandake; kan d i ye ; kan d i ny€ : r i 9ht hand kan?a ; kan?ak e ; kan ' YA; kan n Yt : v i l l age. town kan?ant o l S ; kan?antol orye ; kan?ant�e; kan?anteebf : v i l l age chi et kan?S nrya ; kan?a�geNkan?anrygeNY ; kan?3nnya; kAn?anYe ; Pi ege: trap ka"tal a ; ka"taane ; ka"ta?al a ; ka"ta7ak f ; nda ka "t�ne: hand ka"hwi l e; ka"tawH n e ; ka"tawex al e; ka"t:aw�l( akf ; n d a ka"i:awH n e : pal m ka"t:uxo; ka"tuke; ka"i:uyo; ka"tunY€ : bac k ka"t�Yana; ka"i:ilYane ; ka"tuYanY�; ka.nyakCinyS ?; kani:�Ya n Ye : back ache kanyakCin e ; kanYakun�: naPe ( Ot nec k ) 338 k8parax�; �Par�ke; ��par i Ya; kapar� nYE : � l oor (as i n second � l oor ) kapel i ne; kaPel i n�; kapel i ! �i ; kapel i ��9i ; nd� ka pel i ne: r i n9 kape�E ; kape�ge; kapenYE ; k ape n nyl : b r oom kaPE nl � ; kaptnne; kaPEnx� l E ; k aPE nx�ki ; ???? : sadness kapi ci i ; kaPici i �e; kapi c! i l e ; kaPic{ i b{ ; ????? ; evi l doer : kaPi l �; kaPi � ne; kaP�x�l e ; kaP�x�kf ; nd� kapf i n e : cu rse. bad kaP�n�can l a ; kaPan�caanne; kap�nacanxal a; kapan�caanxak ! ; nd� � panac aan ne : sauce sti rr i n9 sti c k kapan ; kaPon�l ; kapoun ; kaponYte: PraY i n9 mant i s kara; kaate; nda ka�te: meat �ramjl � ; kara mjla�e; karamj l al j ; k.8ramj l ab i ; nd a ki ra mj l a�e: snake ( spec i es ) kar�l oxo : armpi t kas�koQn ; kasakaan��; kasakaOn l j ; kasakaunbf : sol d i er kasana�m� l t ; kcsana�m� � ne ; kasana�m���i ; kasana�mee�9! : el bow kas�x e ; kasake; kas � Y� ; kaS� nYE : war katant�x @; katantake; katant ! y@; katant1 nYt : l adder k ate l � l � ; katel �ne; k ate l al E ; kat�e�9i ; ????????: habi t katexe ; kateke; kati Y�; kat�nYE ; nda ka teke: hunger rust katon�j ; ka tan�e; katonnyj ; katonnYE ; nda k�tan�e: kaY� l @ ; kaY� � ne; kaY� f �i . kaY�xal e ; kaY� ��9f . kaYexak f : comPound kaz�nxt ; kaztnke; kaz t n ' y� ; kaz t n YE ; nd� k�z t n ke: k�l u . ee?; k�l a��; k�el E ; kerex@; k �re�� : su�ter i n9 keeb ! : squ i rr el m on key kez e?� ; kez�?e�e; k�ze?el E ; k�ze?�b1 : chance. l ot k E E X� . kee; kE ake. ktke; keyt ; kenYE ; nda k��ke : hand kE raY� ; kE ranYE ; ????? : � i el d k ! d�; k� ne; ker�x i ; k�raki : Peop l e. c ou n trY kal t ; k al a�e ; kal el E ; kal �b i ; ???????? : 90d . God ko?oro; ko?ote: da n c e kolotjl j ; kolotol c�e: vi l l age chi e � k E rE x t ; k E rake; konl j ; kaanne; konxal j ; kjnxak i : tomb koolj ; koone; nda koone: c ju9h koono; koonde: cotton i:: o ri nY€ ; nd� k66te: Peel i n9s k66ru : kora�a; k6ral o; k6rab i : button kooro; koota; kor i Yo ; ----- ---. - ..- . - - --- .---- 339 k�n l � ; k�nla�e; k�nlal�. k�nl ay� ; k�n l ab i : beads (Di ou l a ) kjrad�nry� ; kjrad�ryge; k�rad�nYt � k�rad�nY£ : 9l ue kjrax � ; kjr�ke; k�ri Y� : k�r�nY£ ; nd� k�rake : heri tage k�raxj ; kjrake; k�r i Yj ; kjri nY€ ; nda kjrake : brick mol d kjru ; kjra�e; kjra l j ; kjrabf ; nda kjra�e: boat (French ) ku; kil�e: death kudd; kune; koral o ; kor�ki : seat kudo ; k�ne; kuux i ; kuxdki ; nd� kune: rule, road kul u9baxa ; kul u9bake; k & lu9baYa; kill u9banY€ : ancestor house kun�9�nry�, -nkan-?; kun�9�n�ge; kunA9�nnYa; kuna9annY£ : centi Pede ku;,boa ; hmbocne; kunbol(ol o ; kunboxcki : l arge or main road kilnn��c; kunn��ge; kilnn�nyc . -i Yo?; kilnn�nY£ : turtle kunna�j ; kuna�e; kun{Yj ; kilnanY€ : mushroom kunncn ; k�nne; kilnnon : kUn�ki ; ?nda kunneNnd� kune: navel kunt�ra�a; kilnt�ra�ge; kunt�r�nY� ; kunt�r� nY£ : ant kunYu�S; kunYu�ge; kunYunYS; kunYUnYE : owl kuru9ba-a l a. kUrU9baYa; kuru9baabf . kuru'ilbanYi ; han9ar : shel ter (9rass overhang) kus�@; kusa��e; kus�eI E ; kus�ebf : travel l er kus�l t; kus��ne ; kUs�xaI E ; kils�xaki : tri P kils�r€; kus�te: triP kiltunno. nk-?; kutunna��; kiltuxalo; kiltu:(�bi : monkeY kuz�n ; kuz��e; kuzuul o ; kilzuumb f : for ked stick 1 l aa l arye; l aal a�ge; l � l anYa; l aal anY€ ; nda l a�l a�ge: butteri ly l al a ; l aane; l axala. l axal a; l a:<ak! , l axak f ; nda hane: PregnancY l-a m;1: � ma�e; lamaI a; l amb i : b l ade <French > l amana; l�mana�e; l amanaa l a ; l amanaabi : nda l �mana�e: sl ing l anmb3; l anmb&�e; l anmbaila. l anmbaYa; l�nmb�ab l , l �nmbanY€ : lamP l axa; l ake; l aya; l anY€ ; nd� like: stomach l ede; l ene; l�ral £ ; l �rak f ; nda l ene: t i me. l e�£ ; l e�ge; l enYE ; l enYi ; nda l e�ge: i ntesti nes l £ rasl i l @; l £ rasi i l ane; ??: be9i nnin9 <of t i me > l O?j ; l o?oke; l o?oYj ; l o?onY€ ; nd� 1 676ke: water l o7okutara�e; l o?okutararyge ; l o?okutaranYe. -i Ye ; l o7okutaranY€ ; nda l o?okutara�ge : :340 hi PPoPotamous l �r� ; l�te; l onyo; l onY€ ; nda l �te: vegetab l e l�r� ; l �te; I �Y� ; l �nYt ; nda l �te ?T: Pri ce l u� i � ; l u� i � rye ; l U�l � I E ; l u� i � b f ; nda l u� ! � rye (?) : water Python luIS; I Mne; 1 0:< al o ; l oxaki : shea nut l u1 e ; 1 Ullne; 1 0x01 0 ; I Cl< c k l : b i l e l unti de; l unti ne; l � nteral e; l u"terak! : ankle l Ul<anv6; l ukanvoke; l uvenyov6; l unYenYonv€ : ri versi de. or mouth of the ri ver? l uxo; l uke. l ox oke; l�vo; l unvE : creek . stream l uverabe ; l uverabe; l uVeri Ye ; l �Yeranv€ : water hol e (stati onarY water ?) l uzu; l uz urye; l uz�l c; l uzulabi : hunter l uzufol6; : chie� hunter . exPert hunter? II ma?anjo; ma?anjorye; ma?anjoOl � . m�?anj ov� ; ma?anjoOb i ; barbe: beard mambel �; mambel arye; mambel al E ; mambel l abf : car m�ra; mararye; m�rel a; mar�b i : el ePhanti asi s marafa; m�rafarye; maraf.aa l a . m�ratava; marafaabf . marafanv€ : 9un mas�n�; maSEnarye; m�s�naI E ; mas�n�b i : needl e mejul 6; mejuune; mej6xol o; mej6xokf ; nda mejuune: voi ce . crY . sPeech mE raryE ; mE raryge; mE ranY� ; mE ranYE ; nda m€raryge: vi Per mE t�nx�; mEtanke; nda mE tanke: 900d name mE trt; mE trarye; mE taral e ; mE tarabi ; nda mE trarye ???: teacher (French > mE n?E nryE ? ; mE n?Enryge; mE n?E nnY£ ; mEn?tnnYE ; nda mtn?€ nryge : storY mExE ?nT; mE ke; meYE ; menY€ ; nda mi ke: name m9b � 1@; m9b � � ne ; m9b � i ryi ; m9b� ��9i : beat i n9 or threshi n9 sti ck m9bin@; m9bi narye; m9bi nal e; m9b f nabi : bamboo mat ma"c3; ma" corye; ma"caa; mancaabf ; nda manccrye: el der si ster man@; mane; mane; manaki : sweet 9round Pea ?? mma; mmarye; mma3l a, mmaYa; mmaab f . mmanyt ; nda mmarye: 9i f t mmal a ?nT; mmaane; mma?al a; mmaakf ; nda mmaane: bud mmE l E ; mmi ine; mm£?E nI E ; mmE?Enki ; nda mm! l ne: voi ce mmE I E ; mmi i ne; mE?E I E ; mmE?E ki ; nda mml f ne: roPe mond�r�; mond�rarye; mond�ral � ; mOnd�rabi : ti mePi ece (French ) moPi i n ; moPi i nrye; moPi f nl E ; moPi�nbi : enemY mCPi i nf61 �; : enemY 34 1 mot8; mot6rye; mot61�, mot6y� ; mot60b f . mot60nY€ ; nda mot6rye: motorcycle m�l � ; m�l arye; m�l aY� ; m�l anYE : r i ce munala: munaan�; muna?anl a; muna?ankf : nose munawi l @; munawHne; munawe:< al e; mcJnawhaki : nostri l mu"kudB; mu"kun£ ; mu"kuralo; mu"kur�ki : cheek muz6n��n , muz��n ; muz6n��; muz66nmi , muz��nl � ; muz6nmbi : bui l der muz�rt; muz�ra�e; mtiz�ral � ; muz�rabi : scarf ( French ) me. mPa. mPa?T; mPary�, mparye?: above mp�?a; mPa?arye; m�?aYa; mpa?anY€ : waterY Porri dge mPanla; mpaanne, mPaanne; mp.aanra; mP nte: Pi geon mPe; mpi ��; mPi i l e; mPi �b i : rabbi t mPad�x�; mPadake �Na?; mpadaY�; mPadanYE : ri b mpoxof6lS; mpoxof6I a�e; mPoxofee; mPoxofeebi ; nda mpox ot61 6rye: cl own Rlpudo; RlPun�; ",Poral o; ",poraki ; nda RlPun�: horn trumF'et mpu18; mPuun�; mpuuryi ; mpuury9i : hi l l mpul u ; mpul ury�; mPul al o ?T; mpul ab f : sPi der mPu�� ; mpuryg�; mPunY� ; mpunY€ ; nda mpUry9�: cei l i n9 mba:<e; mbak�; nda mbak�: Powder Q na; nake; naYa; nanYE ; nda �ke: t i re naa; naarye; namaa; namaab i : man An�a; Anaa��; An�mi ; Anlambi : scorPion naawaxa; naawak�; naawaYa; naawanY€ ; nda naawake: boi l naaxa; naake; naaYa; naanYE ; nda �ke: sore. i ntected wound nafol � ; natol arye; natee; nateebf ; nda natol arye: parents-i n-l aw of a husband naful o; natul arye; nattil al o , natul ayo; natulabi , naftil anYE : ri ches nambu:<u ; nambuxul'''je ; nambuxumi ; nambuxumbi ; nda n mbux u�e : + i ref l Y namem� ; namembe; nda n�membe: soot . ashes namponry� ; namponrye; nampuun ; nampuunbi ; nda nampory�: vi sitor , stranger nanal u; nanal arye; nanal al a; nanal ab f ; nda nanal arye: fresh water t i sh <spec i es) nandaa; nandaarye; nandaal a; nandaabi ; nda nandaarye: bachel or na"j ide ; nanj i n�; na"j i ral e; na"j � raki : si l houette. 9host ----- - - - _ . _ .- ---- ---_. . _-- --- ------- 342 nan j � l u; nan j � l ��e; nan j � l al e ; nanj � l �bf ; voun9 man : nankAlu; nan kal ��e ; nan kal al a; nankal�b i : thi ef nankoov�r�; nankoo� te: red measl es nanvel � ; nan ve l ��e; nanvel �le; nanyel Sb! : friend nanYl �t ; nanYl �ge; nanyl yt ; nan v i nY£ ; nd� ",nYi �ge: fl ame rtna�a ?nT; rtna�g�; rtnanYa; rtnanY€ ; nda rtna�ge: tai l na�9g; n a�9��e ; na�9��1 � ; na�9��b! ; nda na�95�e: 9arden napel e@; napel ee��; n apel eel e; napel e�b! : a verY o l d man napunnaP'l�; napunnapa1&�e; naPunnap�l i Ya; napunna�l i nY€ : mel on seed _ <tYPe ) nas�muu ; n as amu u�e ; nas�muul o ; nas�muumbf ; nda nasamuu��: spark naYeru; naYer��e; naYeral t ; naYer�bi ; nd� n'Y�ra�e: friend na%od� ; nazon�; nazcral � ; nazcr�k! : ladle. di pper nejcb� ; nejcb e ; nejcY� ; nejcnYt : cow manure nena1acn ; nene?an�e; nena1anmi . nena?al a; nena?anmbi : cattl e herder ntr�; : bad l uc k . mi sfortune n£ r�dajan?a; n£ r�dajan?a��: father ' s maternal uncl es nt rafcl g ; : someone who bri n9s mi sfortune n£ r�fcl �; : rel ati ve through a daughter of Pat . fami l y ntru; nt ra�e; ntr�l £ ; n£r�b ! ; nda nE r��e: maternal uncl e n�xasQ; n�xasu�e; n�xasul o; ntxasul abi : b i cycl e n�f�n ; n�f.t�e; n�f.a n ' Ya ; n�f�nnY€ : brick naf i ! le; naf f i ne; n�fexale; nafex�k! : b u ttocks na9un�c. -nk-?; n�9un �ge . na?; n�9UnnYC; na9UnnY£ : knee namp�c!n��; namp�c! n�ge; namp�ci nnY€; n�mpeci nnY£ : toenai l n�ncin?€nn€; n�nc:€n?tnnde; n�nctn?lnnvt; n�nc€n?ln n Yt : sal i va n�nc! l ; n�nc ! f �e; n�nci i ; n�nci i b ! : nei 9hbour . partner . co-wi fe nandee; nandee�e; nandeel E ; nandeeb ! ; nda nandee�e : mi ddl e-aged el der n�njaramE : cow ' s mi l k nan�cl l e . -dul c ; n�nt�c i � ne. -duun£ . 7; n�nt�c � i �i . -duu�i ; nant�c i l �9i . -duu�9i : heel n��e; n��ge: root? rta�ak ci l ax@; rta�ak ci l ake; rta�ak ci l iV@; rta�ak ci l i nvE ; nda rta�ak ci l ake: sPi ne n�Panla; n�paanne; n�Paanra; n�Paante: mosquito nap�l�; nap��ne; nape?el E ; nape?�ki : toe naporoxo; naPorak�: nudi ty ------ ---- - -- - ----- ---- .---- 343 naPu�� ; naPu�ge; nda napu�ge: the whol e nasi l ax@; nasi l ake; nasi l i Y@; nasf l i nYE : f oundati on nol �x�; nol �ke; nol�Yt; nol �Y�; nda nol � ke: grandmother nox�n mb�l@; nox�n mb�ne; nox�n mb�Ya; nOx�n mb� Yaki : boy n� . ne�t ; n��ge; n�nY� ; nenYE : cow n3; n3�el nl i ; ni lb{ ; nda n��e: mother �n�l u ; �n� l a�e; �n��mi ; �n��mb! : gui nea hen n�n?�n? n�n?� Ny; n�n?�nke; ???????: underneath DC. nca; nc��e; nc4a; nC44bf : sheeP nc4n4?4on ; ncAn4?4�e; nc4n4?ala. nc4n4?4mi ; ncana?4nmb{ ; berger: shePherd nc�nkonl�; nc4nkonl ��e; nc4nkon l al � ; nc4"koISbi : wi l d cat ncasj; ncas �e; ncas l a; ncas ab! : encampment . smal l haml et . settl ement nci l e . j-?; nCi i ne; ncl xal e; nc�x�k! ; nda nci i ne: bal aPhone ncananYa. j-?; ncananYa� ; nda ncananYa�e: world (Di ou l a ) �c� . j-? �-?; �c��e. -00; �c�� l � . -06-; �c�Sbf . -06-; n d s Ac��e. -0-: toP ad. nd4; nda�e: bel i ef < Di ou l a> nd� ; nd� �e: food nd�r�; nd�te: i ndi genous med i c i ne <cf . tabe> Adaxe; �dake; �di Ye; �di nYl ; nd� �dake: ear nd�xa; nd�ke: root nd�r�x�; nd�r�ke; nd�r�YS; nd�r�nY€ : Yam nd�x�no; nd�x�nde: seeds nd�x�vire; nd�x�vite; nd�x�viYe; nd�x�vi nYf : seedl i ng nd�xo; nduke; nduyO; ndunYt : odour � �guraxo; �9urake; �guruYo; �guranY€ ; nd� �garake: smoke D.i njedt ; njene; njeral e; njer�k i ; nd� njene: l i ver �jeene; �jeene; �jee�i ; �jee�g! : stone n j i de ; nj�ne; n j i r�le; n j i r�kf ; nd� n j ine: tongue njarazan l a; njarazaant ; njarazanxal a; njarazanxaki ; nda njarazaanE : ePi l epsY nj�rald ; nj�r�ke; nj�d y� ; nj�dnYe: ; mud : 344 � nkanaan ; nkanaan�e; nkanaanl a; nkanaanbi : Packrat nkanj�na; nkanj'nE ; nkanj�i �i ; nkanj� a�9i ; toothache nkanl a; n kaann� ; nkan7an l a ; nkan7anki ; tooth nkanma; nkanmbe: 9ravel nkede; nkene: a Part. side nkeenm9bada; nkeenm9bane; nkeenm9baral a ; nkeenm9baraki : neckl ace n kEEnmgb6ro ; nkE E n mgb6ra�E ; n kEEnmgb6ralo; nkE�nm9b6rabi : chamel eon nkEEn�jolS; nkE( n�jQQnl ; nkEEn�jQ?ol� ; nkEE n�jQ?ok{ : throat nkE�( ; nkE�ge; nkEnYE ; nkEnYE : branch . chi cken ' s stomach nkora�� ; nkora�ge; n kor� Y� ; nkQr� nYE ; nde nkora�ge: fal l ow l and n kucox� ; nkucoke; nkucoY� ; nkuconY� : chi cken cage nkul exe; nkul eke; nkul i ye; n kul i nY� ; nda nkul E ke : cockroach nkul o; , nku � a�e; n kuu ; nkuub{ : chi c ken �kuna; � kunake; � kunaYa; AkunanYE : wooden bowl nkunton ; nkunton�e; nku"to6n l � ; nkunto6nbi : l i zard n kunYel ebE ; nkunYel ebe ; nkunyel i Ye; nkunYel anY� : rooster ' s crow nkupol S ; nkuPol a�e; nkupee; nkuPeebi : rooster nkutun� ; n kutuna�e; n kutuxol o ; nkutux ab{ : red monkeY? � ntaanla; ntaanne; nta�n�i ; ntaan�9i : basket ( t i ghtl y woven ) nt'm9b �a; n ta m9b �ge; nt�mgb nya; nt'm9b nYt : dust ntanla; ntaanne; ntaan�i ; ntaan�9i : courtYard ntana sp?: da Pl ant . the l eaves and t l owers are used i n cooki ng ntara; ntate; nd� rttate: l an d �taraf6l S; Ataraf6l 6�e; ntarafee; Atarateeb{ ; ???: l and chi ef Abazl�x�; rttaz€lke; A taz�lY� ; At3z€€nYE ; nde ntiz€€ke: d i rt . soi l nti de; nti ne; nteral e; nter5kf ; nda nti ne: bat nta ?r; (; to�l ; nde nto�e: bur i al nton 1 9 ; ntonl a�e; ntonl i Y� ; ntonl lnY€ : termite nt��x�; nt�jke; ntooY� ; ntoOnYE : Pestl e ntu�� ; ntu�ge; ntunY� ; ntunYE : chest • nnYa; nnyake; nnYaYa; nnYanyl ; nda nnyake: grass. straw 345 nYa�a; nYa�ge; nYaYa; nYanYf : nda nYa�ge: mountai n nYaP�lg; nYap66ne; n Yapo?ol � ; nYapo?ok! ; nda nYap66ne: bund l e of 9rass nYe?efol g ; nYe?efol a�e; nYe?efee; nYe?efeebi ; nda nYe?efol a�e: ol der brother nYe?i ; nye?eke; nYeYt ; nYenYE ; nda nYe?eke: face nYel � ; nYi � ne; nYt i ; nYi i k! ; nda ny! ! ne: eYe nYenE ; nYene; n yes�i ; nyee�9! ; nda nYenf : horn nY�x� ; nY� ke: morni n9 nYi bal awol o; nYi bal awoone; nYi balaw�xab ; nYi bal aw�:<ak! ; nda nYi b�1 �w66ne: darkness nY� 9banxa. -p-?; ny� gbanke; nyi 9ban ' ya ; nY1 9bannYE : drY season nY{ mb'd� ?T; nYi mbene r?; nda nYi mbene: darkness nYi mbal axe ; nYi mbal ake; ??: ni 9ht nYi mi ?; nY�mbe; nda nYlmbe: shadow nnYam£ ?ml ; nnyambe; nda nY�mbe: shadow nYamado�o. -nt-?; nYamado�ge; nYamadonYon ; nYamadOnnYE ; nda nYamado�ge: earthworm nY� mgb�?g; nY�m9bo?Dke; nY� mgbo?Dyg; nY� m9bo?DnYe: rai nY season nY� �i ; ny� �ge; nYt Ys ; nY� nYE : 9round nYiPE n l � ; nYiPi i ne; nda ny!p! l ne: jeal ousy , l i t . unPl easant face nYi samt ; nYi sambe; nda nYi s�mbe: tear nYo; nYoke; nYOY� ; nYOnYf ; nda nY6ke: mouth , beak ny6?6�� ; ny6?o�ge; ny6?6y� ; ny6?onY€ ; nda ny6?O�ge: camel nYon?on"tanl a; nYon7on "taanne ; nYon?on"taan�i ; nYon?ontaan�9i ; nda nYon?on"taanne: chin nYon�mE n7E n�E ; nYOn�mEn?E n�ge ; nYOn�mE n?E nnYE ; nYOn�mE n?E nnY€ ; nda nYon�mE n7£ n�ge : jaw nYosc E x i ; nYosE � ke; nYoseeY� ; nYoseenY£ ; nda MY6s€€ke: l i Ps nnYo"tan l a ; nnYo"taanne; nnyo"tan?al a; nnYo"tan?aki ; ??: chi n nYoto�� ; nYota�ge; MYotonY� ; nYotanYE ; nda ny6t6�ge � l i d , cover nnY�fan l anxa; nnY�fanl anke; ???: Promi se nnY�juU; nnY�juube; nda nny�juube: sPeech, l i t . mouth words nYudunn� , -nt-?; nYudunn��€ ; nYudunn51 � ; nYudunnSbf ; nda nYutuna�e : deaf nyugod�, -nk-?; nYU9Cne; nYugoral � ; nyugorakf : sku l l nYumpinx�; nyumptnke; nYUmpen ' Yt� nyumpennY€ : bad luck nYunjoor6; nyun joote; nYunjuy6; nyunjunY€ : hai r nYu"tanxa; nyutan ke; nYu"tan ' Ya; nyutanY£ : 900d l uck ----- -- _ . _ - _._------ 346 nYu�j���; nYU�ja�ge; nYU�j�nY�; nYu�j�nY€ : headache nYu�on ; nYU�ge; nYunYon ; nYUnYE : head nYu�ufoI 9; nYu�afol o�e; nyu�afee; nYu�afeebf : chamPi on . boss . suPervi sor nyupar�m�; nyuparambe: brai n nyut i l e; nYut i i ne ; n yuti xale; nyuti x akf : peak underside nyuton8; nyutone; nYuto6�i ; nYuto6�91 : hat nYUVedE ; nYuvene; nyuveral e; nYuver ake: headband (worn by certai n old women ) nYtlzaaxa ; nYuzaake: nyuzaaYa: nYuzaany€ : comb !l '�aa; ' �aa�; ' ryaanm i ; ' �aambi ; nda '�aa�e: twi n �9b�7���; �9b�?��ge;· f)9b�?an nya; �9b�7�nY€ : bamboo bed �m� l t ; f)mi i ne; f)mee� i ; �mee�9i : corner . an9le f)mon ka��; �monka�ge ; �mo"kannY�; f)mo"kannYE ; nda f)mo "kaf)ge: breath f)�� l � ; �muune; �m�?jlj; �m�?�kf ; nda ��uune: knife f)m�ljxj m?; f)m�l �ke; f)mjl i Yj ; �mjl i nY£ ; nda �m�l jke: ni 9htmare e. Paan ; paan�e; paan l a ; paanbi : a carn i vorous rept i l e. 2-3 m. 1 0n9 pal 3; pal sf)e; p�l ami ; pal amb f . pal tnY€ : bucket P§ndaI �; Pcfnd.ll sf)e; pand�l al a; Pa·nd�l ab l : Pants paYast; P�Yasf f)e; P4Yasf l e; p�Yasl i b l . p�Yasl inY€ : mattress Pel E ; Peene ; pexal E ; pexakl ; nda Peene: bowl PE nxt ; PEnke; PE n ' y� ; P£nnYE ; nda p€nke: comPound PE XE ; PE ke; PeYt ; PenY€ ; nda PE ke: l arge clay Pot PSc�; pac��e; Pacaa; pacaabf : 9irl PSjenx€; PSjenke; pajeny� ; pajennY£ : t i c k PsmpannY�x�; PSmpanYake; PamPannYtY�; PSmpanYenY€ : babY PSnf)t ; panf)ge; pannY� ; pannYE : wooden drum Pl f)E ; pi f)ge; pi nYE ; P� nYE ; h i ve: Parandi i ; Parandl i f)e ; parandl f l e; Parandf i bf ; nda parandf i f)e : apPrentice (French ) Pare; pate ; Pi Ye; Pi nY€ ; nds pate: bamboo l eaf Pol � ; pol a�e; pee: peebl ; nda p61 af)e: mal e. husband Po! u; pol af)e; pol sl j ; pol abi ; ????: catf i sh Pondl i ; pondf i f)e; Pondl 1 mi ; pondi i mb i ; nda Pondl i f)e : nai l s ( French ) Poro ; poote; nda poote: mud . banco Poru ; poraf)e; Poral o ; porabl ; nda pora�e: daughter - �-- ------.--- ------ 347 pjn ; p3n�e; P�un ; P��nbi ; nda p5n�e: d09 p�l o ; PllUne ; POl< ol o; POl<oki : b odY pCinY�:<� ; pCinY� ke; PunY�Y� ; PUnY�nYE : c l aY Pusert; pus�ra�e; PUSeriYE ; PuseranYt ; nda pusera�e: dust ( French > 5. sacEnl � ; saceenne. i � n?� sac�nx E l E ; sac�nx�n k ! ; 111?1??: b i rd sfj6n?S; saj6n76�e; saj6n7i Y� ; saj6n?anY€ : sand sal a; sal ake: l az i ness sal a�ol S; : l az Y Person Sa"cEn�E ; sa"cE n�ge; sa"cE nnY£ ; sa"cE nnY€ : cat san�a; san�ge; sannYa; sannvt ; nda san�ge: the rest �"tu; sa"tu�£ ; sa "tul o; sa"tul abi : quai l sa"tu�� ; Sa"tu�ge; sa"tunv� ; sat�nYE : hYena saPi i n ; saPi i n�e; saPi ! nl E ; saPi�nb! : ennemV sara; sara�e; sar ! Ya; saranY€ : sal arY ( Di ou l a) sara ; sara�e; sar l Ya; sari nY€ : tobacco sa�u; sara�e; saral a; sarab! : bee sedE ; sedaki : h i p seeP�l<�; seeP�ke; seeF'�Y� ; seeP�nYt ; n da seep�ke: bel t se?eH ; se7ene; se?e�i ; se7er)9i : basket n arge. 1 easel Y woven ) se?enE ; se7ene; se?e�i ; se?e�9! : pal m nut sen�9a?a�a; sen�9a7a�ge; senI')9aranY.i}; sen�9a?anYE : wa:< senzenexe; senzennake; senzen l Ye; senz eni nY€ ; nda senzenake: hePat i t i s sePE nx� ; sePEnke; nda septn ke: boredom serenver@ ; serenYera�e; serenveral E ; serenYerabi ; nda serenYera�e: door l atch SE E X� ; sE � ke ; seeY� ; seenYE ; nda s€€ke: l eather , skin SE nE ; SEne; sE €�i Ns���i ; sE f�91 ; nda SEne: sti nger sEnrE ; sente; sE n ' YE ; sEnnY€ ; nda s€nte: honey si ; si �e, si i �e: l i �e saca; sacal')e; sapal s; sapal ab! : Person saca?ana; saca?ane ; saca7a�i ; saca7a�9i : shaker . rattl e sacere; sacete; nda sacete T?: madness sacera�6l S; sacera�6l a�E ; saceratee; sacerateeb l : crazY Person sacire; sac�te; nda saclte: crowd , the people ssci r@; sacite: ta9ba lan9uage 348 s�dI')J . s�c�nJ . ; s�dl')·3 e. s�conde: fool i sh sa9�nla. -nk-?; sa9�nl al')e; sa9�nl al a; sa9�n l abi . sa9�nl �nYe ; Po@l e: frYi n9 Pan si � n ; si � nl')e; si ! nl E ; si � nbf ; nda si 1 nl')e: f ami l y s� �r�. s�n�; s� � te. sl l nde; S�y�; sl nYE : nest saj � nl')�; saj � nI')ge; saj � nnY€; saj � nnY€ ; nda saj l nl')ge: carP sak�; sak�l')e; sak�a; sak��bi ; ?: 90at sakol�; sakoone; sak06l')i . sakJxal � ; sak061')9! . s.k�Kaki : woman ' s cane sama; sam�l')e; samaYa; sam�l')e: harvest (Di oul a) sambuxo; s�mbuke; s�mbuYo; samounY€ : swarm of bees sam�; sambe : oi l sam�; sambe; nd� sambe: sor9hum beer sampoxo ; sampdke; sampUY�; sampunYe ; nd� samp6ke: va9i na sane; san6l')e; san�l � ; san60bf ; nda san6l')e: buff al o. bush cow sa"taaxa; sa"taake; sa"taaYa; sa"taanYE ; nd� sataake: bow si nY�l�; si nYQl al')e; si nY� i ; si nyl l b f ; ??: sibl i n9 of same mother � f ather sal')e; sal')ge; sanYe; s i nY€ ; nda sal')ge: worshiP, sacr i f i ce sare; sate; saYe; sanY€ ; nda sate: skin sarawf l �; saraw! ! ne; sarawi xal e T?; sarawi xakf ; ?: Pore s�san ; sasanke: b l ood sasankud8; sasankune; sasankuuxi ; sasan kuuki : vei n s�s6nl')� ; saS6nl')ge ; s�s6nnyj ; s�s6nnY€ : f l y sata7A; sata7ake; sata7aya; sata7anye : arrow saxe; sake; s i Ye ; sinY€ ; nda sake: f i el ds . bush sax e; sake; si Ye ; si nY€ ; nda sake: feather saxacal u ; sa:<acal al"j e; saxacal al a; saxacaUib 1 ; nda saxaca1al')e: wi l d Pi9 < bush Pi9) saxakafan1 al')a; sal(akafanl aI')ge; saxakafanl anya, -i ya?; saxakafanl anY€ : z ebra sol S; soone; soo; sook! : mi l l et s6n l u ; s6nl al')e; s6n l al j ; s6n l ab i : parakeet sonnal')a; sonnal')ge ; sonnaYa; sonnanY€ ; nda sOnal')ge : horse ' s tai l sonxJ ; sdnke; son ' YJ ; SdnnYE ; nda s6nke: horse s� ; sol')e; sool � ; soob ! ; sta9. hart : s�I')J ; SJI')ge ; sjnyj ; S�nYE : mound f or sweet potatoes. Yams s61 0; s66ne; s6x�l o-s6xalo; s6xak! : f l oor sure; sute; nda sute: stapl e d i sh suso709baxa ; sus0709bake; suso709baya; suso709banYt ; nda suso?o9bake: k i tchen ----- ---. _ . _ . _-- 349 5uum�; suumb e ; nda suumbe: sal t 5U:<O; suk€!; suYo; sunyt ; nda suke: mortar t. taanbal@; taanban€!; taan b i Ya ; taan b i ya k i T?: arrow ta�axa; tat ake; tat aYa; tatanYE ; nda tata ke: vi l l age centre tagb�l ax � ; tagb�l ake; tagb�l i Y� ; tagb�l i nY€ ; nda tagb�lake: wound . i n jurY, cut tal i nt; tal i n€! ; tal i ! �i ; tal i ��9! : proverb tamat ! ; t amati �e : tomato tan?an�a; tan?an�g€!; tan?anYa; tan?anY€ ; nda tan?an�ge : shoe. sandal taPu�� ; taPu�ge; nda t�Pu�ge: vi r9i n l an d tasa�e; tasa�ge; ????: al ter f o r sacr i f i ce t�ta�a; tata�ge; tatanya; tatany€ : Senufo Pancake tawax a; tawak€! ; tawaya; tawanY€ ; nda � wak€!: drY pl ace t�mt; t�m��€!; t�m€ l E ; tEmE b ! : s i t ter . si eve tE Xt ; t?ke; teYt ; t�nYt ; nda t € k e : Pl ace. l ocat i on tabe; tab€!; t i Ye ; t i nY€ ; nda tabe: med i c i ne tabalara; tab a l ate: resPect . honour t i d e . tade NY ; t�ne; taral e ; tarak ! ; ????????: l i ana. troPi cal creePer t agaanf61 9 ; tagaant61 a�e. -n k-?; ta9aantee; tagaanteeb i ; sorc i er : sorcerer t i l e ; ti i n€!; t i x ax i ; t i x§k! ; ????????: crest Ot the cock H mbuxal axo; H mbu:<al ak€!; H mb u x al i Yo ; H mbuxal anYt : cl i mb i ng P l an t tanmt ; tanmb e ; t i n ' y� ; t i nnYE : n o i s e tanYu�8; tanYu�g€!; tanYuny8; tanYunY€ : stumP O t a tree ta�e ; ta�ge; tanye; tanYE ; nda ta�ge: 1 09 tapada; tapane; taparal a ; taparak ! : beam . wooden cross ta�nl�; taRSn l a�e ; taPinl { ya; ta�n l i nYt ; nda taP3 1 a�e : new t i el d . c l ear i n9 UPi ne; tal='i n€!; Upi ! �H UPi ��9! : f l ut e tap� l � ; taPoon€!; tap� x5 1 � ; taP�x §k! : tYi ng wood tarapan@; tarapane ; taraPane T?; taraPanaki ; nda tarapane : gr i nd i ng stone taraxa; tarake; taraYa ; taranY€ ; nda ta�ke: gr i nd i n9 stone tax e ; take; t i Ye ; ti nY€!; nda take: tree to; tu�e; ti i ; t i i b ! ; nda tu�e : tather tan l j ; tanl a�e; t�n l a l � . ton l ayj ; ton l ab i . tan l anYt : Prot i t tontonx � ; tanton ke; tonton ' Yj ; tontonny€ : l un9 tax a; toke: l eProsY .--- ------ 350 tClxOf Cl I S ; tOl< ofClI Cll')e; toxOfee; tClxO.feeb f : l ePer t� ; tOlje; to61 � . toY� ; to6b f . tonY€ ; nda tolje: soci ety. group. c l ub t�� x � ; t�� ke; toOY� ; toonY£ ; nd� t�5ke: Paw. t oot . l eg tuc i c �d � ; tuci ci � ne : Younger brother ot father . l i t . l i tt l e fa. tucul um� ; tucu l umbe; nda tucul dmbe: heal th tU9udo. -nk-?; tU9une; tU90ral o ; tU9ur�kf : Pi mpl e tukoraxo . uu? ; tukorake; tukor i Yo ; tukor i nyl ; nda tukorake: hoe for weed i n9 . scratch i n9 earth tu l t x � ; t u l t ke ; t u l eyt ; t u l eny€ ; nda tul t ke : 9randf ather t ul ux o ; tul oke: ethn i c 9rouP tutul')O; tutulje; tutuun l o ; tutuGn b � : messenger tuturo; tutut e : commi ssi on . errand tutuu; tutuulje ; tutuulo; tutuub{ : n i 9h t b l i ndness totuxo; tutuke; t utUYO ; tutunY£ : bran or mi l l et . fon i o, corn tut u x o ; tutu ke ; tutUYO ; tutunY€ : bean cake tuu ; tuul')e; tuumi ; tuub i : caterPi l l ar tuutan@; t uut ane ; tuutane; tuutanakf : hoe handl e tuuttln ; tuuttllje; tutun � ; tutun ab f ; n d a tuuttllje: b l ac ksmi th tuux 6 ; tuuke; tuuy6; tuunY€ : hoe tuxuro. t ox oro; tuxute. toxote; tUXUyo . toXOYO; t ux unYt . toxonYt ; nda tuxute . t 6x 6t e : l oad � vaan b � t ; vaanb � � b e : cl oth vaan"tOIjS ; vaan "toIjge; vaan"tony6; vaan"tonY£ : cover vaanl')a; vaanljge; vaannYa; vaannYt : wraP-around cl oth vaan ti i ; vaant i i lje; v4ant i i l e ; vaan t i i b f T?: weaver van d i k l i :< f' ; vandi kU ke; van d i keyt; van d i kenYt : sl eeve vannd i lje ; v4nnd i ljge ; vannd i n Yt ; van n d i n Yl : shi r t ,. bl ouse v€ n l u ; v€ n l alje ; v€€ nmi ; vt t nmb ! : cri c ket vUljj ; VUIjge; vunyj ; vunY£ : shel ter vYex u ; vYexalje: wel l l adder ( f oothol d s i n wal l ) I!t waras3 ; warasolje ; waraso61 � . warasoy� ; waras6cb ! . warasonY€ : wax a; si ckle. reaPi ng hook wak e ; waya; wanyl ; nda wake: drou9ht w€ rt; wtralje; wtr f Yt ; wtranYt : moneY - ---- -------- ---- ----- :35 1 w£ ramis�n ; w£ rami s€ n�e ; w£ rami s€nY£ ; w£ rami s�Y€ ; nda wE rami s€n�e : change ( D i ou l a ) wod�; wane; ?w�r� l �Nw�ral � ; ?w�r�ki Nw�r�k1 ; n d � w6ne: star w� ; wd�e; wdol � ; w��b i : snake wu l ac� � n ; wul ac � i n �e; wul ac i i n l E ; wu l ac i i nb i ; nda wu l ac i ! n �e : crocod i l e, caYman wYe�t ; Wye�ge ; wYen ' Yt ; wYenY£ ; nda WYenge: l eat � Yaara; yaate; Yar i Ya; Y ari nY€ ; nd� YLate: 9arbage. th i n 9 Yaced� ; Yacene; Yactral £ ; Yac,rak i ; nda yac i n e : Pre9nancY Yat al a� e ; yatal ake ; Yatal i Ye; nd� yat al ake : rePti l e Ya9ba:< a ; ya9bake; Ya9baYa; Ya9banY€ ; nda ya9bake : Party Ya9be� al ax e ; nd� yagbexal ake: creat i on Yagb�ntan�; Yagb�ntane ; Ya9b�ntane; Ya9b�ntanak{ ; nda ya9b� ntane : hand l e ot axe Ya9b:Jn:< � ; ya9b3n k e ; Ya9b�n ' y� ; Ya9b3nY£ ; nd� Yc{9bSnke: ax e Ya9buu; Ya9buu�e; ya9buUl o ; Ya9buub i ; ?????: he who ki l l s t or sacr i T i ces yakukaara; Yakuk��te; yakukAya ; Yakuk�Y£ ; n d i yakukaate: debr i s Y�l o�:J ; Y�1 0�ge ; Y�l onY:J ; Y�l onY€ : okra Yam�; yambe ; nda Y�mbe : si c kn ess Y�r�t£ n ; Y�r�t£ �e : b l ac k berrY Yarax a ; yarake; Yari Ya ; Yar i n Y€ ; nda Y� rake; thi n9s : Yasi nne� yasi nde; Yasi n nYe ; Yasi nY€ ; nda yas i nde: t ru i t yatarax e ; yataraxake; YataraxaYe; n d a Y�tarax ake : creature Yatax �; Yatike; Yat i Y� ; YatinY£ ; nda yatake: neck Yel £ ; Yeene ; Y�e; Y�e k 1 ; ???????? : Year Yi �a ; Yi�ge; Yi nY�; yinY£ ; ????? : moon . month Y l r { faxe; Y1 rl f i k e ; Yi r i f i Ye; Y l r i f i n Y€ : al b i nos Yaru ; Yara�e; yaral e ; Yarab f ; nda Yara�e: Porcupi ne Yd�r�; Y��te ; YuY�; YUnYE : i ns i de ot tree Yox o ; Yoke; YuYo ; YunY€ ; ????????: quarrel z. z an?an ; z�n?�n e ; zan ' ya; z �nnY€ ; nda z �n?�ke: rai n ' z an l u1 8 ; ' z an l uune; ' z an l dx a l o; ' z an l dKaki ; nda ' z an l uune : ki dney z ��m5 l ax S ; z��mS l ake; z ��mSl i YS; z ��m5 1 i nY£ : rai nbow z � nK � ; z � n ke; z t nt i Y� ; z , nt i nY€ : baobab tree 'zi; ' z i rye ; ndi z i �e : b i rth z l Pu ; z � pu�e; z l Puu l o ; z i puub f ; ndi z l Pu�e: s k i rt . ----- --- - ._ . ---_ ._--- ---- ( French ) �I,,", 3",,.. zon; zon�e; z��n l � . zoonmi J %oonmb ! ; nda zo�e: heart zonjan&. -�-?; zonjane; zonjanY&; z onjanY€ . zon ja��9� : heart sickness VERB LEXICON OF SUClTE Each verb Possesses two f orms. the ComPl eti ve and the IncomPl et i ve . Fol l owin9 the En9l i sh trans l at i on of the verb is an examPl e sentence wi th transl ati on . b. accuse· bi ba7a wu l a naMkaI �:(� l a : they are accusi ng h i m ot h i m of thi evsrY. be exhausted . be b red out . ndaa ba?a: I am e:(hausted . thank ( l oan word f rom Arabi c > . thanki n9 You . b ! J bH : rai se (an i mal ) , f eed . WU Ya m6 bar�x6 : theY are nk�u ndi Ya b � i : I rai se chi ckens . 9ather , harvest ground Peas. the kernel s 0+ corn . waa f �ca�e b�l e: he 9athered uP resPect , obey . honour . be thi c k . resPected thei r mother . l ook for. ,. , .- wu Ya saka caa: he i s l ooki ng for a 90at . can ; caan : l ay an e99 . droP, be Possessed . destroY. can : the chi cken l ai d e99s. ceo ci ; kun : do. ce?e; ce?! � : l augh . ce7el E ; d�7�1 u : i nsu l t . nank�ta Ya ra Ye c�?�l u: the each other . cel e; cel H : di vine. nk�l ��aa c' ral a wu sl za tu"turo eYe: he wi l l do some errands. ce?e f an?an la: l au9h loudl Y. ch i l dren are i nsul t i ng wu ci wi sa cel i . T?-sa: have h i m 90 to di vine. spread out . paco�aa vannYa cel i : cl othes (to drY) . tremble. shi ver . shudder . h i m shi ver . we�ta Ya the gi rl l ai d out her wu c�l � : t�p col d made 353 ci n ; d n ! : know , be convi nced . Persuaded . and I . we know each other . C E r aX E ; d raxo: d i s l ocat e . d i sPersed ' nda n� �nnaa wu Ye ce n : Anne nf}nYa.1 d i sPerse. C E ra:< E : the caH l e h ave refuse. Ya�a Ya ya1 ata CYf}x o : the sick Person ref u sed f ood . sneez e . k.1konxa f 6 1 a m.1?a cad i : he who has a col d , sneezes . car i ; ddi : peel ( eggs ) . shel l shel l the peanuts? carax e ; c arax6 : f ai n t . waa na?�n kuu w u ma?a caraxe T? : i f t o d i e. he fai nts. conri , j; condi : embarrass . worrY. ca m.1?a conri n.1 mpone ti:< ! i la: the c h i l d embarrasses the stranger f uun k i cari l a : d i d he someone i s goi n g • .. cod ; cud i : waa ( Peanuts ) . stran91 � . someone. mE E ta m.1?a nca w.1 cor i : the cord can strangl e 9ather . n.1n j a l �b i Ya ma?�n9bora�a cuu: the boys are in the Process of 9ather i n 9 man90es . c�n ; cuu n : tear . n d a v.1nndi ��.1 c�n : mY sh i r t i s torn . sort . choose, coonri I set t l e ( an af f ai r ) . nda si am 90i n 9 to set t l e an aff ai r . za kaana 1 .1 dd ; dd i : Pl an t . COl ul")5n ; col a�u.: h eaT. recover . be i n 900d heal th . People are recover i ng . cud ; c ild i : pl unge. s i n k , swamP . foot i n the mud . cuol o ; cuul u : bel ch ( or burP?) . mo6 l i ta mo m.1?a cuul o : when Yo� eat wel l . You bel ch ( or burP? ) . be deep. cYil ; cy6d : catch . wuu yal a na taYi c�ri i : we shou l d pl ant some trees . yambi Ya col o�u : the s i c k ndaa na t55ka curi n j jraki i : I stuck mY bal akaa cux i : the wel l i s deep . b i i n kul � CYU: they caught a ch i c ken . d. pu l l ( a cord ) . wu cu rna dal i T? rna: grab i t and Pu l l . 354 £. en l arge. mazor/a Ya ';bax a f.{a:l u : the mason i s i n of en l ar9i n9 a house. bal ance. swi n 9 . the roPe. t l atter . t emPi: . her . t "H u ; t.:tl u : the Process s� z� fa� u me: e: t a l a : l ei: ' s g o swi n 9 on wu wu Ya wu f�n l u: we are temPti ng . f l at t er i n9 construct . bui l d . bui l i: the house. n9a wa� gbak6 f aanre= i:hi s i s i:he one who exchange. repl ace. bi n� w€ra f�l � : theY made change . wu na wu ( t h e vOi ce > . l i mP. t i 1 1 , cuI b vate. H el d k€rax 6 f aa c:i : h e cul ti vated h i s • .fa:n l� ; tan I H : t o l ose voi ce. run . dr i ve . nda� yal a n a mamb�l a t � u : +tnH ; HnI H : i nc l i ne . d i P . be at an an9l e s i n k i ng. t 1:: nr i ; H nd i : Pi nch . put in a corner . b 9h t spot? H ; .f l u : germi nate. sprout . H i nni ; t H nnu: be c l ean . .f H nn� ; t H nnu: cancel . "' z an kaa cuI t i vab n'; . I l ost my I must dri ve a car · Ya fEnl i : the ma m� ya f 1:: nr i nk�: are sun i s You in a sddk � l aa f i u : the mi l l et has germi nated . tal a f H ni 1 i : the rai n • cancel ed the f al i wd�a l a : apProach t h e snake. ntane b i Ya f ! l ! i : i t is the courtyard that Pound ( a f l oor > . theY are Poun d i ng. tal < a > l e ; f al < a > l u : f ane: ; f anu : ca�9a Ya fanl � : cana� f i i nn i : the c l aY Pot i s c l ean aF'proach . H J. a ; H I l i : nda m� � ni crawl . i:el l a l i e. l uzu�aa f a l < a > l e : the huni:er crawl ed . wu Ya f anu : h e is i:el l i n9 a l i e. make �.h i te . be wh i te . make c l ean . vanl arJa c l oth is get i: i n 9 wh i te < becomi n 9 c l ean > . Pi erce. t i re > • n E x asu�aa f ar i : Ya f ant x6: the the b i cycl e i s Pi erced <has a f l at 355 to; f u = emi 9r at e . t or i ; fild i : f 6x a1 6 ; f6xal u: t al ab i l aa f a : the t u l an i emi grated . fuu: f l ow . l o?oke Ya l ose . w�raa f �n l otara�i : we l ost at the l otterY. 90 out . aPPear . Pound 1 . the water i s f l owi n9 Yf�9aa fori : the moon came out . bi n�a mbi fuxal u. b l ow . wi nnow, swel l . b l ewin9 the t i re . f :) ; t uu: roast i n hen . nga waa n�ke f � : th i s i s the one who i s wu Ya nn6l u f uu : h e i s roast i n9 a gui nea t h e f i re . B n ; Bni : m i s s . f ai l . f u l o ; fill H : push . f iln ; f ilnn i : sweat . kaf�ka Ya b i f iln 1 : filx 1 ; tcid : 9 1 ean . wu Ya filun fux f ; ex t i ngu i s h . na e:< t i n9u i shi n9· furf : f u:< ari ; fil:<�ru : T?-n£a : they pounded the f l ou r . waa sa f � n ann 1 i : h e mi ssed Anne. wu Ya f� a:<a f ill H : he pushed a stone. f ud : bi the heat makes them sweat he i s 9 l ean i n'3 Peafiuts . · T" . f urf rummage , ransac k , search . the house. it's . a .f i re wilri Ya 9bake f ilx �r il : u r i nat e . N€ n ; f H n : f l ower . sukara waa f Y£ n : wh i P , beat . f y€ nx € ; fy€nx6: Ann e i s afrai d . he uri nated . wu Yari taka Ya f i f n : we searched hurrY uP. waa +Yeer i : be qu i et . bi� staYed qu i et . suc k . �nn1 Ya fY�:< i:l : ta f yal il : hurrY uP. are the hYena to f r i ghten . tremb l e. f r i 9htened them. be afrai d , f ear . they wi i fY�?� : theY cal l ed h i m , but he i t was su9ar that he suc ked . the tree i s f l ower i n9. w�ra Ya wi:l�a fY� n r i:l : sob . cU 9be:<el aa nyel i wu has real l y cried, he sobs . we are beat i n9 the snake ma7an fyt nx 6 T?: • when a chi l d 356 9ba; 9bilu : dri n k . ndaa b i E ra 9ba l a: did I dri n k beer? a9ree. meet . wel come. waa 9bara si da sf na moi : a9reement t o 90 w i t h You . wound . i n jure. 9bere; 9bE}rE}xo : f al l asl eep. t h e b i cyc l e i n j ured h i m . ncx asu�aa wu 9bel � : b e short . shorten . wu kune < l et ' s take a short cut ) ca�aa 9bE}:< � : l et ' s shorten 9bere= uProot < mi l l et ) . 9b€rax € ; 9b€ra:< 6 : d i spute . '=Iuarrel . '=Iuarrel l i n 9 . b i sooki 9bE r i : ceeb i Ya bi gb6; k6! i : ki l l . gbox i!l; 9bi!lx H : meet t0gether . Put i n Pi l es . i n 9bcxc: the men met together . gb�n; 9b66n : h i t , beat . d i g 9round Peas. beat i ng each other . Posona�a Ya nk6l i : get t at . the road the c h i l d tel l asl eep. 9baxal a ; 9b�x al 6 : arrange , make. torm. des i 9n . rePai r . drew a scorPi on . 9bE r i ; 9bedi : he i s i n wu Ya n� 9bax al u : he he uProoted the mi l l et . ya 9b€rax 6 : the women are the poi son k i l l s . na�a Ya 9b61 ux 6 : grouPs . bi Ya b i y� namaabi l aa b� ya they are gb66n : t h e man i s get t i ng t a t . i shoot . brea k . b e ab l e l eave. wash . waa nn6l u j a : to. succeed enter . jiilu jUlu: cross (a r i ver ) . j o ; Yu : sPeak . saY, tal k i n 9 . j 6 ; j 66r:i : swal l ow . t orage. j ool U j OOl u : sew . ------ he at . wu Ya " r}m:Jb j H : j i ; cy€ n : he shot a gu i nea hen . he washed a kni f e . bi Ya nCYE n 9baki i : they entered the house. b i i kuna j i i l i : they crossed the road . the recount . nkill a�a Ya j 66r i : wu va�9aa jool i : _ . _._ -_ . _----- was ab l e to Peof:' 1 e are the c h i cken i s toragi ng. h i s c l oth i s sewn . 357 sharPen . ndi Ya 'l')muune j�j: d : I am sharPeni n'3 the kni f e . k chew . k.i'i1 : annl Ya burc1:(o Anne i s chewi ng bread . cri t i ci ze . order (somethi ng ) . they ordered a hoe and then l et t . sutt er . Yal')a Ya nkaal a : wu na taax6 ka�jo ml kar � : the sick Person i s sutt ering. d isapprove, deny . Pardon . waa w€ r�l')a Ya mi i stole moneY and he den i es i t . PaY attenti on . suPervi se, consider . take care. min : he does not consi der anybody . k.a7,k� ; k.a7,ku : scratch . sPi t . frY. wu Ya nY1 l')9a ka7ara: wu Ya ssmsY60ra k� f i : be SPoi l ed . erase. destroY. Yam i s spoi l ed . teach . read . swear . kan ; k,3an : give. l end. boi l . be sad . Cebaa kara kan �naa: uProot ( ground Peas. harvested Peanuts . l ather . foam. ndjdk.ia': kaU:d : go, l eave . Peanuts) . he read the l etter . Ti eba 9ave meat to Anne . ndi Ya kanrl sakf i : moo t aun ndi Ya kedo ereda7e: I kin ; k i f n : groan . Yal')a Ya nki l n : You l i ve at Orodara. Anne has gone to Canada. 90vern (a country ) . translate (a l an9uage) . turn. Pour the mi l l et i nto another contai ner . ca bi kanla: the soaP i s l ather i ng. annaa karl Kanadii : tease. mi streat . mi streat i ng . the the water i s boi l i ng. j ek� Ya n kanl')a : inhab i t . l i ve at . wu Ya wa kaacaa she i s f rYi n9 dou9hnuts. be ti red . t i re. Punish . to make sutter . I t i re i n the f i el ds . kanla; k�nB i : he he scratches the soi l . waa seb�l')a kal � : 1 07ak� Ya n'kan i : x� �1�1 c1: keraxu : it ' s seekl karal')a: a ch i l d the s i c k Person i s groan i ng. they are 358 kal axe; kal axo: be i ntel l i gent . k6; k6nU:<6: tear trom. k6; k6d : draw (water ) . k�7o; ko7� i : dance. plaY . k676; k6?H : hope. ant i � i Pate. kob ; kol H : cou9h . kon l i ; kon l axo: sl i t a throat . ki l l . koori ; kooru : work the soi l . cl ean oH the l and . are wor k i ng the f i e l d . k6d ; kodf : nai l . k5; kuu: f i nish, termi nate, end . running out . k5n ; kudn : cut , d i 9 ( a wel l ) . ci rcumci se. they are d i g9i ng. k5d ; k6dl : chase, punt . d i e. shave. gather . kul al 6; kul al u: crY out . kill il; kill u : rol l . hid : crunch . it's a bi t ilunkl kul i : they that theY are nai l i n9 . Ya n kdo: the moneY i s ba1 S:<a b i kudn : i t ' s a wel l i t ' s a 90at theY are chasi n9. I shaved the b eat" d . he i s gather i n9 uP the Peanuts . wu Ya n�xasur"Ja kill il: wu Ya wora ktirl : You " must endure tati 9ue. y@ : t aci � �a Ya k6l alu: the culti vator is crYi n9 out . he i s rol l i n9 the bi cyc l e . I am crunch i ng the kol a . fol d the mat . fal aka kuri : kura1 6; ktiral ti: tol d . wee�ga Ya nk6ral u: the leaf i s fol ding. waa wu Ya kuro�5 : wu Ya ')m6H na nkuraru: ----�- - " -- --_._---- wu Ya kE raka kodru: the chi ckens are dYi n9. kan?anta kil ma wal k on al l fours. al l fours. they k i l l ed a steer . door w€ra�a tol d . snore. • the woman i s cou9hi n9 . sakaa bi k6di : stumb l e . bumP i nto. he danced and then he went bi non konl i : kud ; kud i : kurar'o; kuraru : go and draw some water . ndaa m.{?,:(jcr"Ja kill i : kul i ; kildi : the wi nd pu l l ed uP a tree. waa ko70 mi i sa: n kililb i Ya kul i : kill � ; kun ; sa 1 070 k6: 9baraxa bi kudi : to endure. He i s i ntel l i gent . kafE� kaa taxa k6: cee�a Ya n kol i i : kCi; kol ! : kildi : waa k�l axi : he stumb l ed . he snores whi l e sl eePi n9. ca�a Ya ku'l:< o: the chi l d is wal ki n'; on 359 kYex i ; kY�x o : brea k . d i g and take broken . nda k€€ kaa kYex i : out Yams . mY arm is 1 laal a ; LUl u : l ick. IAiwul c; laawuu: make fun of . r i d i cule. ot h i m . return . pa�aa tasa�a l aal a: bi Ya wu 1aawau: l a?a Pan nan?'n : Peel . ku mande:ra wu Ya Peel i ng. 1 ��H ; l ��l u : l ower . bend over . l ex€lle; l ex€ll u: t i c k l e . Pri ckle. Ie: ; Hxo: 9row ol d . be ol d . the dog i s l i ck i ng the Pl ate. come bac k here. it ' s a Potato that he is 1 a'?al u: I am bendin9 over · ndi Ya l �el u : y � a y� y� they are mak i ng fun l ex al u: 90 t i c k l e Youself . mY father i s o l d . nda tu�a.a I e: : a thi ef hi de. camouf l age. (habi tual ) h i des • crack (by heat ) . B; H: eat . down . nYl �9a.a l e: rax i : m! Sa san€ : maa l � the 9round i s cracked . when You have eaten . 90 and l i e if YOU be far awaY. maa I i I ! wu sa kari : wi 1 1 1 eave You . bi Ya fct6l a 1 66: 90 far awaY . we I S ; 1 au: take . h i re. 1 6x6; nduru: understand . hear . l i sten . mo Ya nduro na kaci in&a fori : hear that the f et i sh has come out . lux ! ; l ad : cl i mb . 90 uP. wu Ya l ur! tak! i : he takes Pi ctures. You He i s cl i mbi ng the tree. m m,ar.a; m.arh : glue. conserve. sti ck together . PaPers are stuck t0gether . man ]: ; mbadi : l i ght . pl aster . mani ; manH : assembl e . add . col l ect . col l ecti n9 moneY . - --- ---. l amb��a man � : ---- - ----- bi papi nYa mara y� Ye l a : the l i ght the l amP. Ya w€r€ man! ! : they are 360 inn� sa mon kanadai : mo . mon ; moni : stay a whi l e. l ast . wh i l e i n Canada . mb Ure; mbHru: think about. Ot You . mbux i ; mbur:t : suck . mbu: d ; mbud : oPen . reveal . cee�a� wu na�a non70n reveal ed the secret Ot her husband . Anne wi l l ndi Ya mbf f ru mo l a : medi tate. c��e Ya Aeta mbur f : sanci i na Ya mpeel u: wu ja�e Ya mpannu: the woman mbux f : he rol l ed uP the Protect . defend. munca�a Ya mpdi col a�a l a: protects the Younger sibl in9. stutter . I am thi nkin9 the chi l d sucked the nere bean . mbuxal 6; mbuxal u: rol l u�. waa nci i z i mbuxal6 wu k€€ ka l a: strin9 around h i s arm . 9l i de (hover?) . stay a the b i 9 si ster the b i rd i s 9l i din9. his son stutters. D. twi st . wind. wri the. ne; neni : put . wear (clothin9) . wear a shi rt . kunaa na7ala: nda s i za vandi�ane: bri n9 i n . sak� bi n€�u 9baku: bri n9i n9 i nto the house. no. non ; ndni = bi te. arri ve. achi eve. n6n76; n6nd : di rty, be d i rty. the road is twi sted . i t ' s a 90at that wo waa wu non : ndaa n�n7Sn : I am di rty. retreat . reject . rei mburse. are wi thdrawi n9 them. ndunxa1 6; ndunxal u: smel l . sni H . pan�a Ya na mPon�a ndun:< al u : sni tt i n9 the stranger . - -- --- sow. or Pl ant. waa toxa ndil:< i : they are a snake bit hi m. ndunru�Sn ; ndunru�un : ndu:<u; ndad : I am 90i n9 to b i Ya b i ndunru�un: theY the d09 i s he sowed . or Pl anted . corn . 361 ndu:<u; ndux6: dunk ( i n a sauce) . the sauce. bel i eve. create. (generi c ) • see. Ya �iin � l oat . kal aa saca nta nYaa mEn : swi m. 1 0?akH : T? saca: God I am swi mming. di l ute. reduce ( swel l i ng) . su kara�aa nYaun the sugar is di ssol ved i n the water . nYAl a�a; nY�l a�u: riPen . be red . ma?agboraa I E wu ma?a nY�l a�: i s mature. i t i s red . nYel i ; nY� n i : crY. waa kaata nY£ n7( n : he tasted the meat . wal k around . wu S l sa sax a nYE E r i : l et ' s go wal k the bush ) . nyUd ; nyUr6: kal t wu nYEE r6: ask �or. PraY. ny€n7f n ; nY€n?l n : sti r . move. cE nka ny£n?l n : conaa nY! : nyi ; nY� n i : fill . nYi ; nY1 un : shine. l i 9ht. ny! ; nyi nl ; wake uP. i t i s Sod he i s Praying. sti r the sauce. sa wu nY1 Y1 da se: the sun shi nes . wake ( sg . ) him uP and 90 (pl . ) . be wet . be col d . cocl . cool i ng . Fjmon ; �mL1un : . --- l et ' s go hunti ng ( l i t . the clay Pot i s ful l . ca�9i Ya nYi u: b e 900d . b e Pretty. when a mango the ch i l d i s crYi ng. ca�a Ya nYi n � : ny�n7en ; nYe7� i n : taste. created man a b l i nd Person does not see. ndi Ya nYanf : d i ssol ve. he dunked i t i n wu n� r� ndux6 cE nkf i : va�9aa nYo: the broth i s the cl oth is pretty. draw a bow or sl i n9shot . waa mp�na :)mon na: �E rambH : shot at the turtl e-dove wi th a sl i ngshot • ------- he 362 I)m6n ; I')m6n i : rest . breathe. rested . wu net k..:ln7.fn mi '"Jm6: mo6�mSI S bi si suta l � : sl eep. the mush . c��a Ya I)m�d� : nurse. he was ti red and he i f You sl eeP. theY wi l l eat the chi l d i s nursi ng. push roughly. jostle. jostl i ng each other . bi Ya bi Ye I)m�ra�un : theY are e p�l a ; p�I H : .. Pan ; ma: surPri se. come. arri ve. be f at . Peo. Pe; peni : wu Yamba� nda p�l �: napo1�a tak�J pel � : shel l < l ocust bean ) . bean . Pan : his sickness surPri sed me • the stranger has arri ved . the tree i s f at . ba n€t� Pea: di sPl ease. di s9ust . not content. ndi � : what You did di sgusted me . PEri ; P€rH : sel l . mol a wu P€ ri i : l ose. w€ r�l')aa Pan� : SPi n <cotton) . f i ght . sweeP. bu bi Ye per� : nt�nna wu ci l a� Pin pan u : she is SPinning cotton . they fought each other . kaati Ya p� ni : the meat is getting i t ' s wood that he i s tYi ng. puu : do better . be haPPY. doi n9 better. i t ' s the courtyard she i s swe�Pi ng. the sick Person i s swel l . wu '3be7el)9.1a Pu : h i s cheek i s swol len. '3reet . ndi Ya m6 s.Uru: I am greeti ng You .---- ------- I)ge the moneY is l ost . wu Ya kona kanci Ya waa P� : Ada moo i t ' s rice she i s sel l i ng. be r i Pe. be wel l cooked . wel l done. tie. come and shel l the l ocust • 363 sharpen . comb . carve. hew. that I am sharPeni n9. san7an ; san7an : kara nda san7en : chew. kari nY8n nda saJ�6: i t ' s meat that unti e . reach out (the hand ) . him. bi n� i t ' s a Penci l am chewin·; . I. wu san7an : theY unti ed stretch out . waa wu Ye san7anl l fal aka la: stretched out on the mat . .. san � ; sani : vacci nate. PaY. dcxat5raa wu san� : fol a ndaa sara: Pri ck . rend. sna9. wri te. nkaanaa nda sar i : the stick poked me. am wri t i n9 . I si ; si : be b orn . 9i ve b i rth . si E n ; sHni : Produce . si i l i ; sH l i : be stron9. sHnH ; sH nlu= ai m ( t o shoot somethi n9> . that I a m aimi n9 at . si l i ; sid! : be9i n . nda c6�aa see: tak�a s i E n : mY wi fe 9ave bi rth . the tree Produced . f aci'i' si i I i : waa 9baxa s! l ! : a cuI ti vator i s stron'; ' we n d i Ya si ! n l u : he be9an a house . bi Ya sinu: mo ndi Ya sax � i : they are l Yi n9 down i t ' s You that waa s070 mi i son ; sdnni : worsh i P . adore. are ador i n9. 56n ; suun : spend the ni9ht. nda si za s6n k�n ka7i : the n i 9ht at Kan9al a. 56:d ; s6d : burn . kacE n l a bi i son : 9baxa :(a s6d : buy, be saved . escaPe. am wai t i n9 for . she Prepared PrePare. cook . then went . defecate. sa: I 5070; sori : 5U ; silni : i t ' s a snake timid. be ashamed . l i e down . wai t . the nurse vacci nated us. i t ' s a debt that I Pai d . ndi Ya sab! ! : be shy. timid . he i s <food ) and i t ' s a fet i sh that theY I am 90i n9 to SPend a house i s burni n9. nkul Ul')aa s� : ndi Ya si za su: I the chi cken escaped . am '30i n9 to defecate. 364 su; sul i : pound . sCi; suxal Ci: poke. jab . ���nary4d nda su : silild ; sililru: exagerate. much be too i t ' s corn they are Poundi ng. fcxa bi sol i : much. the need l e Poked me· wu Ya I i ma sililri : he ate too • suxalo; silxal u : crouch . su:<ari ; silxaru : si�t. rya�aa su:<al 6 : the man i s crouched . si �t the t l our . mb i i be suxari : shake. the car i s shaki n'; us. mambel arya Ya wcra su:<aru: recei ve. f i nd . commi ssi on . I taal a; Ual u : carress. Pat. hanla; taan l u : al i 9n . sacks. t�.inla; tadnlu: measure. comPare. measurin9. I i 9n ndi uP. Ya ni p6nrya taal u : recei ved Yow'" am Patting mY dog. bil b�ranYa taanla: they l i 9ned uP the i t ' s mi l l et that seo nda taanlu: I am l i ke. pl ease. be c�ntent. mo � ba nda �ri ka n� ta�n ndi i : You came to 9reet me . that Pl eased me . h?a; tar i : put on the � i re . wu Ya col a tari : he Put a pot on the t i re. it 's share. d i vi de amon9. d i stri bute. peanuts that I am di stri buting. tanba; Unbi i :, l earn . teach . a mechani c . stomp. tan?Jnl� ; ban?anl Ci: ndi Ya mekanaz E n Ya tanbi : waa ncia tan?an : stomp . he stomped me. wu Ya f i na tan?.£n l Ci: be 900d ( to eat ) . 900d to eat . he i s stompin'; ton i o . mak�rana�aa tanla tal ax i : touch . wu Ya n�ryu t Yaabi l a l o?ak i i : � i sh in the water . sti ck . gl ue. b i nd . on the l etter . b e used to. I ' m l earn i n9 to be they are touchi ng the tE mburCi tard l E tira�a la: kan?ataa nda tel i : the macaroni i s stick I am used to �ati 9ue. a stamp 365 h; tee: show. show Your husband (to me) . ma naa�a t� : tiin. ti i n ; tlan : si t down . wu Ya tian anYbodY ' s house ( f i 9 . ) . wa Y�ri mE n : he doesn ' t sit down at tinl �mi ; tlnl ama: bar9ai n for . ha99l e over , barter for . i t ' s a shi rt that I am barteri n9 for . vandi �e n d a t€nl amu: Hrt ; HrH : sl i de. sl i P . waa t�r( : he sl i PPed . Pl ace. put . set ( a dat�) . hel P. and hel P me i n the work. ba n� t€xi baara�a la: u; U: brai d . weave. tan ; tan i : make a noi se. rumbl e . crash (thunder ) . thunder i s crashin9. 9ri nd . crush . waa vanI a ti : coma he is weavin9 a cloth. the kal ��i Ya tani : he is 9ri nd i n9 mi l l et . waa soo tare: l odge at . unl oad . stayed at mY Place. t�un . t i n ; tan! : be ful l . waa I i taun : descend. nda si tax i l uki i : fal l down . ndaa to: gbu he ate unti l he was ful l . wi l l 90 down to the creek. I I fel l down . t6; t6n! : burY. close. burYi n9. tonI :! ; tonUxo: be 1 0n9. make 1 0n9. ndaa tonI:! : t6r6; t6d i : Pass . Pass on and -;0 . t6rax 6; t6r6xa: accomPanY. have sen t . send. wi l l send You a l etter . bn75n ; bnd : dil ute. draw out . the H our . t::5d ; t6di : count. tun ; tunn i : send somebol Y on an errand . an errand . t6r6 maa se: ----_ . -- - - - vomi t. it ' s a I nda ndaa ilibi i bn75n : ndi Ya sakaa t6di : clean . wi Pe. rub . hands. bl:d ; til=< H : wu Ya t6ni : I - -- - - ---- .---- we are am tal l . sf z a seba t6r6x6 m06: I I am mi x i n9 water wi th am count i n9 90ats. bi Y� nda tunni : wu Ya wa kenYa tuuxo : ca�a Ya tux i i : corPse that he i s the chi l d i s vomi ttin9. he sent me on rubb i ng h i s 366 t(tx i ; hid : tu:<o; tux i i : di9. deePen . carrY transPort . m�l a ·nda� tuxo: drY. be courageous. be f l our i s drY. l ook . vi s i t . i t ' s r i ce that I carri ed . wu na wa wu yaranYi : throw. mi spl ace. h i s objects. wa7a; war i : I ' m o i 99i n9 for a mouse . kacu ndi Ya turf : hard . d i f f i cul t . ma� nta na wi ! mE n : be raPi d, be in a hurrY. what you ' re doi n9. be b l ack, forget . b l ackened > w£ri mbi ba� mo rida w£ra�£ : rna k�rak i l i : nda fu�aa w� : wa?a: the don ' t l ook at me. heat . bother. be a pest . annoY. vex . me. wtd ; wtrax6: he mi sPl aced some of You bU9 hurrY uP illl th I for90t ( l i t . mY i ns i des • wu Ya nafAn w�nl : Pour . make bricks. wul i ; wCidf : wash . ndi Ya wudl : he is Pouri n9 bri cks. I am wash i n9 mYsel f . take out . cl ean uP/cl ear out a wel l . the wel l d i 9ger cl eared o�t the wel l . wurax f ; wCira:<6: mix uP (th i n9s > . be mi staken . ya. ; Y4: be si ck. wu Ya Yet : , Yawn . Ya?a; Yari : yal a; Y.1l axo: waa Ya�l �: abandon . Permi t . be suf f i c i ent . Ya?a waa se : the meat i s suf f i c i ent. Yere wu taun : wu Ya wu Yera�Ci : - ---- ----.--- wu he i s shaki n9 the tree. ka�ta� yal a: stoP ( i ntr . > , stand . --------� I 90t mi xed uP. he yawned . wu Ya tak� Y�?�ru: stoP ( tr . ) . ndaa wurax i : he i s si c k . l eave. reserve. 90 . shake. bal aku�a balaka wul � : stoP beside h i m. he stOPPed h i m . l et h i m :367 sa wu Yex e: questi on . ask about someone. Ytr€ ; YtrH : wear out , sPoi L damage. wear i ng out . ncia vanYa Ya YtrH : counsel , advi se. ndaa wu Y& rE h i m about the wor k . get u P . f l y. come f rom. Yari ; Yari : cal l . i nvite. come and eat . jump. waa Yari : ndi Ya mo Yari ma ba spl i t . bi Ya Yi ni : I advi sed he got uP. Ii: I am i nvi t i ng You to ndaa kampena Yaraxe : I they are j umpi ng. kanci xe ndi Ya Yuu: i t ' s wood that water . na�g�rya nd i Ya Yuu: water i ng. YO; yCil ! : mY cl othes are baara�a woke l a : rai se , ��ke uP (tr . > . straighten uP. rai sed the f i nger . Yi u . YU Yi ni : go and ask h i m . steal . tri ck. riP-ot t . a m spl itting. i t ' s the garden wara Ya yCil i : squeeze (orange> . mi l k (cow) . i s wri ngi ng out the cl othes . I wr i ng. that I am him. he steal s. waa vannYa Yux i : she 368 WELMER ' S QUESTIONNAIRE Transcri Ption and French transl ati on by Traore Gnoudj oti en . Tone Anal Ysi s and En9l i sh translat i on b y Anne Garber . The fol l owi n9 questi onnai re was desi gned by Wel mers for the PurPose of doi n9 a qui ck surveY ol the basi c structures of an Alri can l an9uage. Speci l i c vocabul arY and syntacti c structures were sel ected i n order t o faci l i tate a basic anal Ysi s and a comparati ve studY between d i al ects and l an9uages . Thi s Suci te questi onnai ;'·e i nc l udes a French and Engl i sh transl at i on lor each i tem and Engl i sh 9l osses. At this Poi nt, 9l osses are tentat i ve. pendi ng a more thorough consul tati on wi th a Suc i te speaker . The transcri pt i on . though on the whol e accurate. does have some i nconsi s�enci es that cannot be corrected unti l I get i t checked out wi th a Suc i te sPeaker . UnderlYi ng tone i s marked above the tone-beari ng uni ts esPec i al l y i n cases where the surface tone i s di flerent than the underlYing tone. Bel ow i s an eXPl anation ol the symbol s used : L Low tone H Hi gh tone M Mi d tone (Lh ) Mw Weak Mi d tone ( HI ) ML Mi d-Low tone (Hi) AbbreyiaU ons Q Fut . VP Neg· REF 1. I NC IncomPl eti ve marker Quesb on marker Future marker CLAUSE M Cl ause Mar ker PP PostPosi bon Verbal Parb d e Con . Connector Negab ve Marker EXCLAM Excl amati on marker Referent i al Pronoun ? transl ati on or transcri pt i on unced:.>rin M L M nankin son une Personne Person one one Person . - ----- 2. LM LM sl i n sUOn Persons two deux Personnes two Peopl e 3. LM M ML sl: i n kankorll Persons f i ve c i nq Personnes f i ve People kan7an vi l l age un vi l l age one vi l l age 4. 369 5. L M M ML kanY� sililnni vi l l ages two 6. kanY� ka"kurO vi l l ages f i ve 7. c i nq vi l l ages f i ve vi 1 1 ages L M se?�l � n�n k i n one basket ,., o. un Pani er one basket M L M M M ML se?�l)i suunn! baskets two 9. deux vi l l ages two vi l l ages se,?�l)i k�nkurO baskets f i ve 10. M LM deux Pani ers two baskets c i nq Pani ers f i ve baskets sakaa suun 90ats two deux chevres two 90ai:s s�Pal a Punon Peopl e al l tous l es hommes al l the men kanY� Punon vi 1 1 ages al l tous l es vi l l ages al l the vi l l ages 11. , .... . 1 ... 13. L M M LM nd� sakaabf Punt mY 90ats al l toutes me� chevres al l mY 90ats 14. SaCi nYE ?are: men manY 15. M ML M beaucouP d ' hommes - manY men SakanYE '?are: 90ats manY beaucoup de chevres manY 90c!i:s 1 ilnYE ?a:<E water manY beaucouP d ' eau much water wa someone quel qu ' un someone 16. 17. ----- -- - - - - - --- ----- - --- ---- ------- 370 19. des hommes some men pH some 1 9. M pH b H P�n ba some they come M 20 . 21 . 22 . 23. j ' ai achete du r i z I have bou9ht some rice ci nk�na l� tun sE n?i n d o manner an other a9ai n Y a-t-i l un autre moYen de faire cel a? nYe nd€l l a l aa i s that on Q i s there another waY ef doi n9 that? SE nYe?ara Peopl e manY yaci VP-not H ML Ycil M a9ai n? 28 . L M ML l aa kcro M xa jo ii: d ' autres hommes ont d i t qu ' i l s ne 1 ' avai ent Pas vu some other men have sai d that they did not see him nkuu da kanna na nYe ndud chi ckens l i tt l e onl Y be mi ne 26 . 27. . nya€n wd h i m see-Ne9. m�l a c�r� kanna rice l i tt l e onl Y Peu d ' hommes l ' ont vu few men have seen him nya€n wu h i m see-Ne9. P� l a� na ju some others saY that L 25 . des hemmes sent venus te voi r some men have come to see You m�na ndaA S::l rice I buy Peri n.i theY-REF VP 24 . mo wee You vi si t je n ' ai qUE Peu de poul ets I have but few chi c kens il ne reste que Peu de ri z there remai ns but l i ttle rice . di s-I e encore saY it a9ain saY M il l ' a fai t encore he did it a9ai n yaUci l a c� he a9ai n? ii: do lOla ML H nda yaci nda ci nkanna do m€ I not that do �anner know Ne9. �- � ---�---- je ne sai s Pas fai re ca I don ' t know how to do that 37 1 29. 30. 31 . 32 . 33. 34. 35 . 36. 37. 38 . 39. 40. ----- nya€n nda Ya� wu I not him see-Neg. je ne l ' ai Pas vu I d i d not see h i m wu Ya mi nanj�� he come today i l va ven i r aujourd ' hui he wi l l come today bi � Pan t��j�� they come YesterdaY i l s sont venus h i er they came yesterdaY bi Ya met nY�"'Panl')a theY come tomorrow i l s vent veni r demain they wi l l come towmerrow l a7a Pan ny�mPan�a return come tomorrow revi ens demain return tomorrow Mw Pan na 1 07i f namb�er�m€ come water r i ght nOI'4 aPPorte de l ' eau tout de sui te bring some water ri ght awaY wu nYe kan7an l a h e b e vi l l age i n i l est au vi l l age en ce moment he i s at the vi l l age at this moment nd� na mPan I come H mo sH na You but Vp nambed? r i ght now n�71n ta�jaa. here yesterday nci �7€£n be here-Neg. je sui s venu ici h i er . mai s tu n ' etai s Pas a I came here yesterday , but You were not there H L M waa t� f n n�7;n canncaa suunnf he staY here daYs two i l est reste la deux jours he stayed there two days yal ari n� nci wa€ food be there-Neg. i l n ' Y avai t Pas de nourri ture there wasn ' t anY food there wu gbak� nYe waa �€ [nYeaan�t ] his house be there sa mai son est la-bas h i s house i s over there L M wcr� sa ba n�ri wad �i ny�mpan�a we Fut . go there tomorrow nous i rons l A demai n we wi l l go there tomorrow -- -. . _ - _ .._------ la 372 41 . Mw L M nd� nYe nd� '�m66nnE thi s be mY kni fe ceci est mon couteau th i s i s mY kni f e 4 2. Mw nd� nYe '�m�ntafesd? thi s be kni fe-pretty ceci est un j ol i couteau thi s i s a Pretty kni fe 43. L M T? nk� nYe nY£?E�n thi s be what qu ' est-ce que ceci? what i s this? nk� 9bakaa psI � thi s house l arge cette mai sen est 9rande thi s house is l arge nta ka�taa P�n?�n:<� thi s meat tou9h cette vi and e est dure thi s meat i = tou9h 46. Mw nd� nYe '�m�ncEnnE that be kni fe-900d cel a est un bon couteau that i s a 900d kni fe 47. M L M Mw nd� nYe nd� '�m66nnE that be mY kni f e cel a est men couteau that i s mY knife L M T? nk� nYe nYE?€sn that be what '=!u ' est-ce que cel a? what i s that? nk� 9bak" kaa psI � dE that house l arge cette mai son-l� est 9rande that house is l arge M L �9atun waa k�r� j Os who he that saY qui a di t cel a? who sai d that? wa� kaf?ka fo wu kanU?Aki l i he wi nd b l ow h i s hands-i n il a souff l e dans ses mai ns he b l ew in h i s hands kaf?ka Ya fau f�n?�n l a wi nd i s b l owi n9 f orce on Ie vent souf f l e fort the wind b l ows stron9l y wu Ya �m�nn� l � h e i s breathi n9 Q resPi re-t-i l ? i s he breath i n9? 44. 45 . 48. 49 . 50. 51 . 52. 373 54 . 55. 56 . H nk� k�"c i � nYi�€ . ka it wet this wood Ya?s nta nyf mE wi l l -Neg. INC burn Ne9. ce boi s est humi de. i l ne brOl eia Pas thi s wood i s damp. i t wi l l not burn L M Mw H naka nda Yar�nYa Pun� sorax6 f i re mY thi n9s al l burn I e .feu a brOl e tous mes objets the f i re burned al l mY thi ngs �9atun wu Ya rna w€� who he come there-Q '=lui arri ve? who ' s comi ng? w� wa na nkol al o she throw and crY el l e s l est mi se � crier she started to crY wa kal (a) la� mon be-I on9 she crY el l e a eri e l on 9temPs she cri ed a l on9 � i me sap�l abf Ya nko?i k�n?�nka Peopl e are dancing vi l l age ML .fan�H i nsi de-in l es gens sont en trai n danser au vi l l age 57. 58 . 59. 60. 61 . the People are danci ng at the vi l l age H b i i kj7A� nY�ka mbax i they dance morni n9 oPei! i l s ant danse toute l a nui t they danced the whol e ni 9ht L M ML nda tor"ja ka mY father d i ed mon P�re est mort mY father is dead waa k� tfnje� he di e l ast Year i l est mort l ' an dern i er he died last Year nkatunna�a� tu nYal"}9a l a fal l ground on monkeY I e si nge est tombe sur I e sol the monkey has f al l en on the ground 62. 63. 64 . ka� tu i t fal l l o?okH the water-i n c l est tombe dans l ' eau it fel l i n the water wa� tu� he fal l kd ka the arm i l est tombe et s l est casse I e bras he fel l and broke h i s arm 65 . kE :d break - ---- ------- 374 66. i:ak.:{a t u il tree fal l Pari kiina t.he road cross l ' arbre est. t.omb� a t.raverse I e chemi n t.he t.ree has fal l en across t.he road 67. ma?an nta t Y�xa m€ You-Neg. I NC afrai d Neg. n ' ai e Pas Peur don " c be afrai d waa fYail he fear i l a eu Peur he was af raid wu na nda conl a�a b� y� 9b6n he and mY brot.her t.heY h i t / f i 9ht lui et. mon t rere se sont. bat.t.us he and mY brot.her t ought. each other waa xa � l o?ok i i 1 he i t. throw wat.er-i n h�n x� �{ na nY�ni H oat. i t. i l a j et.� cel a dans I ' eau . et. cel a a t l oH� he threw t.hat. in t.he wat.er . and i t f l oated 68. 69. 70. 71 . M L l ilka Ya kan?an ka Pad (?) � ri ver i s vi l l age crossi n9 l a riviere cou l e a travers I e vi l l age the ri ver f l ows through t.he vi l l age 72. nka nci � n�9a l a l ilka t.hi s di rect.i on on r i ver l a ri vi �re coul e dans cet.t.e di recti on s� l d?�ka Ya t uu na wat.er i s and goi ng the ri ver f l ows i n th i s di rection sancHnna yad fly b i rd l ' oi seau s ' est envoI � the bird i s in f l i ght la it. i l ne sai t. Pas vol er he doesn ' t know how t.o f l y 73 . 74 . 75. Ya?a n j a �jara mt Fut. . Ne9. be abl e f l y Ne9. H waa d?€ he l au9h 76 . i l a ri he l aughed H waa sanE he l i e down ----- ---------- i l s ' est couch€! he went t.o bed ------� -- �--,�- - " :375 77 . H waa HnM.1 nYa :( a k.1nndilka la he si t see i ts vi l l age-same? on i l a hab i U dans I e m@me vi l l a';e he l i ved i n the same vi l l age wu s£ n?in nYe si l a l aa he sh l l be l i f e wi th Q est-i l encore vi vant? i s he l i vi n9 a9ai n ? 73. 79. H bi i m�n?�nki cel .1.1 nY� ka mbilx ! they son9 sin9 mornin9 oPen 30 . 81 . i l s ont chante toute l a nui t theY san9 the whol e ni9ht H waa tUn he si t i l s ' est assis he sat H L M waa ')m�nl � cannc.1a hand he sl eep days three 32. ;I wa.1 j u na. he sPeak wi th i l a dormi trois j ours he sl ept three daYs H dax�taxa')i i chi ef-wi th i l a Par l e au chef he spoke to the chi ef 33 . wa.1 j u fan?an · l a he sPeak force wi t h i l a Parl er fort he spoke l oud lY waa ka?axa can nYa,)9a l a he sPi t drop 9round on i l a crache Par terre he sPat on the ground waa yad he get uP i l s ' est l eve he 90t uP 34. 85. 36 . H waa yere he stoP i l s ' est arr�te he stoPPed Ya?a waa j uuba he sPeech l eave. stoP i l a cesse de Parler he stoPPed b I ki ng 87 . 88. takil nYi r"]l mi ntaun tree be-wet and swel l I e boi s est devenu humide et i 1 90ntle the wood has become damp and i s swol l en wu kekaa tailn h i s arm swel l sen bras s ' est gen t l e his arm has swol l en H 89. a 376 90. 91 . 92. 93. 94. 95 . 96. 97. 98. 99. 1 00 . 101 . nda Yaa 1 070 ctn mE I not water know Ne9. je ne sai s Pas nager I don ' t know how to swi m H ndaa 1 070ka nYan tor6 r i ver swi m cross I j ' ai traverse 1 a r i vi ere a 1 a nage I swam across the ri ver L M nda Ya suun na nd� sa 1 � thi n k that I fut. i t I dn .ncee know do H j e Pense que j e sai s fai re ca I thi nk I know how to do that ndaa xa son�� ca�9a k� i t thi n k day end I j ' ai Pense a ca toute 1a j ourn�e I thought of that al l daY nda caa suun na wu sa mPan th ink that he fut. come I j ' ai Pens� qu ' i 1 a1 1 ai i: veni r I thought that h e was 90i ng t o come H waa a74 he return i 1 s ' est retourn� he returned waa bra�a kar��a he canoe return i l a rei:ourne l a Pi r09ue he returned the canoe • H waa kara we� sa ndn 1')9a l ook to k��w who he turn '" wu ma wa'" 1')ge he come there CLAUSE M L M 9baya hand t6r6 houses three Pass kanda:<e ri9ht mi i l a tourne 1 a i:@i:e Pour voi r qu i venai i: he turned h i s head to see who was comin9 wa. kara turn depasse troi s mai sons Pui s i:ourne a droi t 90 Past three houses then turn r H lht wu Yaa nblx � vomi t he i 1 vomi t he throws uP wClraa Yera ss:<aa mo we YOU wai t b e-1 0n9 nous vous avons attendu l on9temPs we wai ted tor yOU a 1 0n9 t i me L M 377 wu Ya nca na nY� �ra mt he Neg. be-ab l e wal k Neg· 1 02 . 1 03. 1 04 . 1 05. 1 06 . 1 07. 1 08 . 1 09 . 1 10 . 111. 1 12. 1 13. i l ne sai t Pas marcher he doesn ' t know how to wal k H waa l ux ! i:aki i he cl i mb tree-i n i l a gri mP' dans l ' arbre he cl i mbed the tree H waa l ux ! t 6 nYan�ga ny� nti n f i h e c l i mb unt i l h i l l toP-PP i l a mont' j usqu ' au sommet de l a coU i ne he cl i mbed to the toP ot the h i l l I u:<! na7�n mt c:l i mb here montes ici cl i mb uP here waa Ux � nYan�ga l a he descend h i l l PP i l descendu de la col l i ne he Came down from the h i l l H ML waa c��a l urax6 t�l aka nYu�! i he chi l d l it t stone head-PP i l a h i ss' l ' entant sur l a Pi erre he l i t ted the chi l d uP onto the stone H wa� keka nce he arm wash i l a l av' I e bras he washed h i s arm waa c��a taraxa tak i -i he ch i l d descend tree-PP i l a descendu l ' entant de l ' arbre he 90t the ch i l d down trom the tree wa� tor i he go out gbak! -i house-PP waa yad he l eave k�n?�nk i -i vi I l age-PP wa� �m� l a he kni te wul 6 take out i l .est sorti de l a maison he went out of the house i l a qui tt' I e vi l l age he let t the vi l l age L 9bugbQn�gi -i box-the-PP i l a pri s un couteau hors de l a boi te he took a kni fe out of the box wdra� nci gbaki -i we enter house-PP sommes entr's dans l a mai son l et ' s 90 i n the house tYanY£7€n l a� ji enter t i sh manY beaucouP de Poi ssons sont entr's dans la nasse sa�n�i net-PP - ---- ----.--- ------ 378 manY f i sh have entered the net 1 1 4. 1 15. 1 16 . 1 17. 1 18 . 1 19 . 120. 121 - 122. 123 . 124 . waa mjl �l)a ne ( ? ) beh k!-i he rice put ba9-i n i l a mi s I e riz dans I e sac he put the rice i n the ba9 l o?o ne tonna cddn!-i water Put metal Pot-pp mets de l ' eau dans l a marmite put some water i n the Pot H waa wi l � kdn m ! sana s! n he hol e cut and Post Plant il un he in waa kaYi (?) d� ne iW:JkH he meat-pi ece Put. mouth-i n i l a mi s u n morceau d e vi ande dans la bouche he Put a Pi ece of meat in h i s mouth wdra y�d kan?an l a we l eave vi l l age H L M Mw kad wdra t.ai:a 90 our Pl antati on ma to nous sommes al l �s du vi l l age a notre Pl antat i on la pp we went from the vi l l age to our Pl antat ion waa y�d kan?an l a t<1l")ja he l eave vi l l age PP yest.erday a creuse un trou et i l y a Plant� Pi quet dug a hol e and he Pl anted a Post it i l a qui tte I e vi l l age h i er he l ett t.he vi l l age Yesterday H waa kad ny� ka l a he l eave morn i ng PP i l est Par ci Ie matin he l eft i n the morni n9 wdraa non n�?an tarija we arri ve here yesterday nous sommes arri v�s ici h i er we arri ved here yesterday L H wdraa non na?an Yakdl')9a l a we arri ve here eveni ng PP nous sommes arri v�s i ci Ie soi r we arri ved here i n the even i ng waa Yi u kilna l a he jump road PP i l a saut� a traver I e chemi n he jumped across the road wa� Yi u Wa takH he j ump tree-PP i l a saut� de l ' arbre he jumped from the tree 379 1 25 . waa kunnaa Yi u �9a wu Yaa non mi i l a saut� mai s n ' a Pas Pu attei ndre he road j ump but he not arr i ve Neg . he j umped but cou l d not make i t 1 26 . mon bH juu they tal k be-l ong i l s ont par l � 1 0n9temPs they tal ked t or a 1 0n9 t i me 127. bi wil they he k� nco tal k about ( der09atorY ) i l s ont Par l � de l u i theY tal ked about h i m 128. 1 29 . 1 30 . wu Ya Yu fan (? ) an l a sPeak he torce PP i 1 ParI e f ort he speaks l oud l Y ML waa Y i u tanYu�ga l a 1 09 PP h e jumP i l a saut� par-dessus l a bucne he j umped over the l og LH n d a Y a s�c�ra Yil senoufo speak I j e par l e du senoufo I ' m tal k i n9 about senouto waa t�mpal a wa he noi se-b i g? throw i I a pouss� un cri he 9ave a crY H ndaa kill ( � ) 1 6 crY out I l< � l 6de i t hear-Neg . j ' ai cr i � mai s . i l s n e m ' ont Pas entendu I c r i ed out but theY d i d not hear me 131 . 1 32 . 1 33 . panryaa wu dog h i s 134. 1 35 . 1 36 . �ga b i Ya but they not Mw t55ka non l eg b i te MwML waa l< � k6n kaYfna takak i h e i t cut meat-Pi ece Pl ace-chew-PP Ie ch i en l u i a mordu l a j ambe the dog b i t h i m on the l e9 i l a arracne en mordant un morceau de vi ande he b i t otf /t ore ot f a Pi ece of meat waa m�U S::J?::J she r i ce cook eI I e a t ai t cui re du r i z she cooked some r i ce waa n kill a S::J?:l she ch i cken cook el l e a f ai t cui re un pou l et she cooked a ch i c ken 380 y 1 37 . H waa nkilLlbf Bri she chickens couni: 1 38 . 1 39. el l e a compi:e l es poul ets she couni:ed i:he chickens H wa.i bni:a kon nal')9i i he l i ana cui: mi ddle-i n? i l a i:ranche l a l i ane he cui: the l i ana Wa... v�.nmp� ! k6n i l a coupe Ie i:issu he cui: the c l oth waa k£ka konl i he hand cui: i l s ' esi: couPe a l a mai n he cut h i s hand he c l oth cui: 1 40 . 141 . L M ML waa k.inc� s6r6xaya Pan la he cui: � i rewood i l a couPe du boi s .i brQl er he cui: some f i rewood wu ya.i nka nk6n nk�nna ctn mE he not that cut manner know Ne9. i l ne sai i: Pas couPer ca he doesn ' t know how to cui: that 1 42 . 143. LH wu Ya ndjraYe i:Ciri Yams d i g uP he i l est en t rai n de deterrer l es ignames he i s di ggi ng uP Yams 1 44 . waa he 1 45 9ba dri nk l unYE ?E nxa water-much i l a bu beaucouP d ' eau he drank a l ot of water • ...'draa mjl a B we r i ce eai: 1 46 . nous avons mange du r i z we ate some r i ce H wdra.i kara k.i we meai: chew-eat 1 47. nous avons mange de l a vi ande we ate some meai: LM-H wa.i w€ra kan ndilCi he moneY gi ve me-PP i l m ' a donne de l ' argent he 9ave me some moneY w€ra kan nuLl moneY 9i ve me-PP donne moi de I ' argeni: ';i ve me some moneY 1 48 . 1 49 • . - �- ---. ---- 381 1 50 . ndaa x� l 6x 6 I i t hear j ' ai entendu cel a I heard that L M moo r.d� 9b6n ( :::: ? ) You me hit tu m ' as frappe You h i t me waa b i 9b6n he theY h i t N i 9h t i l l es a baHu he f ought them waa Pdnl')a hun he dog k i c k i l a donne un couP de P i ed au ch i en he k i c ked the d09 M L wd. Pdn l')a nY�?� n k6n he d09 f ace cut i l a rel:enu I e ch i en he restrai ned the do9 15! . 1 52. 1 53 . 1 54 . H sakal')e 9bo they 90at k i l l i l s ont tue une ch�vre they k i l l ed a 90at bH 155. H " ndolii wu caa 1')ge ndaa d n that he want CL know I je sai s ce qu ' i l veut I know what he wants ndaa wil d n I he know je I e connai s I know h i m 1 56 . 1 57 . H mi i na nt�l e -nt? or d? roPe pul l b re l a corde pu l l the roPe tal')ge !:al a n �?an mE 1 09 Push there over Pousse l a bQche Par l a push the 1 09 over there 1 58 . 1 59 . ML waa wil kE € k a tUUl< 0 he h i s arm rub 1 60 . H waA Yacsna tOn l �x i n a kanc�t i � he bel l Y scratch wi th nai l s-PP 161 . i l a f rotte son bras he rubbed h i s arm ci bi i ma bi they do/tel l they come i l s ' est 9ratte I e ventre avec l es ongl es he scratched h i s stomach wi th h i s nai l s d i s l eur de ven i r tel l them to come - --- ------ ------ 382 1 62. waa ma; bi ci tU,/a thei r f ather do/tel l he come d i s a l eur P�re de ven i r tel l thei r father t o come nYa?an waa jo� what he saY-C! qu I a-t-i l dH? what d i d he saY? ndaa saxasa"cuun nYa I bush cat see j ' ai ndaa vanl i\l")a j ool i I sew c l oth j ' ai cousu I e ti ssu I sewed the cloth 1 63. 1 64 . vu un l eopard I saw a l eoPard 165. 166. LM nda Yaa 'I')guraxo nduxa Ua I smoke odour smel l je sens la fumee I smel l the smoke waa nkan l aka Y::I he stick spl i i: i l a f endu I e b�ton h e spl i t the stick 1 67 . 1 68. ML waa wll sil ka fi\d he h i m skin Pi nch 1 69 . L M i l s l est Pi nce l a Peau he Pi nched h i msel f ML waa ndi\ k€€ k a f E ri he mY arm Pi nch i 1 a Pi nce mon br'as he Pi nched mY arm waa naaka f YE n he wound drai n i l a suce l a pl ai e he drai ned the wound 1 70 . 171 . 1 72 . L M pi\mp!nnY� ka na ncaa '/m�di was nursing babY 1e waa ka"kanl axa wet; throw he sti c k i l a j ete un b�ton he threw a stick waa mE ta klld he roPe b e i l a noue l a corde he t i ed the roPe bebe tl!t:ai t the babY was nursing 1 73. - ---- ------- 383 1 74 . waa sak;iiI'"Ja Po t:axa 1 03 he 90at t i e tree to i l a attache 1 03 chQvre a un arbre he tied the 90at to a tree waa bl!h ka kild he ba9 t i e i l noue I e sac he ti ed the ba9 175. 1 76. L H LH waa vanmp� ( � ?) ba ncQ he wash cl oth i l a l ave I e ti ssu he washed the cloth waa k�nYa ncQ he hand wash i l s ' est l ave l es mai ns he washed h i s hands waa 1 070 wuI ! he water wash i l a Pri s un bai n he teok a bath waa I)muuna tuuxo he kni f e w i Pe i l a essuYe I e couteau he wi Ped the kni fe bH 9ba:<a f aanri they house bui l d i l s ont construi t une mai son they bui l t a house 1 77 . 1 78 . 1 79 . 1 80 . 181 . L M bi Yaa 9baxa faanru they house bui l d i l s construi sent une mai son they are bui l di n9 a house ndaa sama s;:] I oi l buy j ' ai achete de 1 ' hui l e I bou9ht some oi l 1 82. 183. H Aka fa:l ake B this stone Pi ck uP 1 84. ramasse cette Pi erre pick uP thi s stone H waa kaYarakf Puna bal e he Pi eces al l 9ather i l a ramasse tous l es morceaux he 9athered al l the Pi eces nUsill a nYe yawi 9bol ;:, el ePhant be ani mal -l arge un el ephant est un 9rand ani mal an el ePhant i s a l arge animal 185. 1 86 . L M ML nda n;:,:·:an�bal @na nYe nda j3 thi s I i H I e boY be mY son - -- _._----- ce peti t 9arcon est mon f i l s thi s l i ttl e boY i s mY son 384 1 87. LM-H LM n� nkoo sudn nYe ndad chi l dren two be mi ne-PP 1 88 . j ' ai deux enfants I have two chi l dren LM-H M L gb�a nYe ndad nkau chickens six be mine j ' ai six poul ets f have si x chi ckens ML 1 89 . �guk� wa sii ka suu fari thorn his skin Pi erce l ' ePi ne l ui a traverse I e Pi ed the thorn went through h i s foot saru waa nda non na k££ ka l a bee he me b i te mY hand PP une abei l l e m ' a Pi que a l a mai n a bee stung mY hand 1 90 . 191. Mw WQ waa nda non na tjjka l a snake he me b i te mY foot PP 1 92 . L M i l m ' a hi t mal i t hurt me kaa nda y"d i t me hurt 1 93 . un serPent m ' a mordu I e P i ed a snake b i t mY feat L M tjjk& Ya nda yad me hurt foot j ' ai mal au Pi ed mY foot hurts wa� lo?o kaun he water boi l i l a fai t boui l l i r de I ' eau he boi l ed the water 1 94. 1 95 . H i l a fai t frire une vi ande he fried some meat waa kara kall he meat frY 1 96 . H waa barant3n kaL( he bananas roast i l a fai t r8ti r des bananes he roasted some bananas f� waa faun he Peanuts roast i l a fai t '3ri l l er des arachi des he roasted some peanuts l o?oka Ya nk","n ! water boi l l ' eau est en trai n de boui l l i r the water i s boi l i ng b i i m::n-nY£ ?un wtri they rice much Pl ant i l s ant pl ante beaucoup de ri z theY Pl anted a l ot of r i ce 1 97 . 1 98. 1 99 . . - -- ------- . :385 200. LH bi i nd�r�Yi nd�K i they Yams Pl ant i l s ont Plante des i 9names they Pl anted some Yams Yaa m3nl ��a f�� bi they rice wi nnow i l s sont en trai n de vanner I e riz the · are wi nnowi n9 the ri ce ndaa m�3n-nYE ?axa Pir€ I r i ce much sel l j ' ai vendu beaucoup de r i z I sol d a lot of r i ce 201 . 202. 203. L ML waa k��9b i n' kE E x i he sti c k break 2(04 . L ML kaF)9b ! naa kE E l<i sti ck break 205. i l a casse I e b�ton he broke the sti ck I e b�ton est casse the sti ck i s broken H waa mina ken he roPe cut 206 . i l a casse l a corde he broke the roPe H mi n," ken roPe cut l a corde est casse the roPe is broken kon ' yakon ' YS wa� Pena nca he bowl shaHer Pi eces ? i l � casse Ie Pot en morceaux he broke the bowl i n Pi eces wu Ya?a. �ja l<a kE x a mt he Ne9. fui: . be-ab l e i t break Ne9. i 1 ne Peut Pas I e briser he cannot break i t 207. 208. 209. H waa i:or6 n aran mt he Pass here 210. H wa� nda 9baka ter6 he mY house Pass 21 1 . i l a depasse ma mai son he Passed by mY house H nd' kUna waa 1 3 thi s road he take 212. i l a Passe Par i ci he Passed by here L M ML wor� ka nHan Ya qur manner VP-Ne9. Pl ease mP! sxnbi l i mt these People-PP Ne9. i l a Pri s ce chemi n he took th is road ces gens-l a nous detestent those People hate us - ---- ------- ------ 386 213. L M t1L ki:f wor� our 2 1 4. H si n b U i tMn mP:L ces gens-l � nous ai ment manner Pl ease these Peopl e-PP those Peopl e l i ke us H H ML wu ma� na"'k::J:Jra � Uan wu-i our also chi l dren Pl ease us-PP 215. chacun ai me ses enfant.s everYone loves thei r chi l dren H-L waa ju war� sa wu cu. he saY he-REF fut . h i m catch i l a essayt! de 1 ' attraper mai s i l n ' a Pas Pu H wu s� Ya nca mE he but Ne9-VP be-ab l e Ne9. 216. 217. LM nda Ya suun s� j u na mo-i n je veux te Parler I VP desire fut sPeak wi th You-PP I want. to t.al k to ML? YOU M 1 070 k�l a nYe nda l a water manner ? i s m e on 218. L M 219. he tri ed t.o trap h i m but he could not je veux de l ' eau I want. some water ML nd� l� nYe s� ni:a I desi re be f ut . Con . Inc . t�ba7about� tanbi french l earn je veux aPPrendre I e francai s I want to l earn french waa Y�ri na nc�7:L he get. uP and l au9h i l s ' est mi s a rire he st.arted t.o lau9h waa koClna Ya7a he cou9h stoP i l a cesst! de tousser he stoPPed cough i ng 220. 221 . H wa� nk� 5070 n kanna ctn he that. cook manner know i l sait Prsparer cel a he knows how to make that wu .1 aa nci wi nc(:il oF)Sn . he desire be he be-healed waa yal a ma taba 9ba med i ci ne dri n k he shou l d s ' i l veut 9ut!r i r . i l doi t. Prendre du remede i t he wants to be cured . he must t.ake some medicine LH nda Yaa nd�r�Y:L sol i Pound I VP Yams je suis en trai n de Pi l er des i 9names I am pounding some Yams 222 . 223 . - ---- ------. ----- 387 224 . bi wll n i:a1aa. Yi I a:<e 9baki -i they he p� l l house-PP i l l ' oni: i:rarne hors d e l a mai son they dra9ged h i m out of the house ndaa wu tun ntax atilx al")iln he send I c:h i ef-PP je l ' ai envoye c:hez Ie c:hef I sent h i m to the c:h i ef ndaa. wira t6r6x6 1")0un moneY send h i m I je l ui ai envoYe de 1 ' argent I sent h i m some moneY 225. 226 . 227. L s.. l:ata punaa ra ye 9b6xo peopl e al l t hemsel ves assemb l e c:aank! i market-i n tous l es gens se sont rassembl es au marche everYone ';ets together at the market bi � sacata 'Obo:<[) they People assemb l e i l s ont rassembl e l es gens they assembl ed the peopl e kafonI")9i 1 a.1 non kannya Puna l a news arri ve vi l l ages al l i n l es nouvel l es se repandent dans tous l es vi l l ages the news sPreads i n al l the vi l l ages tal aka c:an nYal"}9a l a mat. sPread 9round etends 1 a natte Par terre sPread the mat over the 9round sadtaa CEraXE Peopl e di sperse l es gens se sent di sperses the Peopl e have di spersed 228. 229. 230 . 231 - 232 . H v.1nnya Y�ral")€ n !7an m€ l aundrY han9 there 233 . H mets a. Pendre Ie l i nge l a han9 u P the l aundrY there H 1 M kan c:i 'lna PanIc! mac:hete take c:ut rod gbes@ 234 . H � n ka kdn na " I")m::ll i -i that. cut with kn i fe-PP Prends une mac:hette et. c:ouPe l a ba9ueHe i c: i take a mac:hete and c:ut t h e rod here eouF'e ea avec: un c:outeau cut that wi th a kni fe - ---- --------- ------ 388 235. waa '�muuna Ya?a l aa tu he knife l et i t fal l nY��9a l a 9round on i l a l ai ss� tomber I e couteau Par ter"re he l et the knife fal l to the 9round '�muunc1a tu nY��9a l a kni h tal l 9round on I e couteau est tomb� par terre the kn i fe has t al l en to the 9round waa suxa 9bexale he mortar make i l a f abri que un mortier he made a mortar waa tu nt�nna �m� l a 9bexale he i ron knife make i l a f abr i que un coteau for9� he made a torged knife 236. 237. 238. 239. 240. M waa Pdnndal a jool i sew he Pants il a contectionne un pantal on he made a Pai r of pants nYa?an mo Ya n kun y� what You VP do Q qu ' est-ce que tu es en trai n de fai re? what are YOU doi ng? 24 1 . H moo waa n ka 9be:(�I� l a You thi s make/do Q 242. as-tu f ai t ceci ? did You do thi s? H maa ma co�a gbex�l a l a You Your net PrePare Q as-tu Prepare ton f i l et? have You Prepared Your net? waa kan?an�9a 9bex�l e l a PrePare Q he traP a-t-i l PrePar� I e pi ege? has he set the trap? 243. 244 . M te na a ka ci n kana it do manner show me on 245. M kuna te na a road show me on 246. montre-moi comment taire ca show me how to do that montre-moi I e chemi n show me the road M '�muclnna te na a ma kn if e show me on Your montre-moi ton couteau show me Your kni te 389 247. 248. 249. ma nda Yara l aa You me cal l Q est-ce que tu m ' as appe l e? d i d You cal l me? M waa wdra y�d he we cal l i l nous a appel es he cal l ed us b i i mamkl il Yar i they Mamadou cal l i l s ont aPPel e Mamadou they cal l ed Mamadou waa kob he cou9h i l a tousse he cou9hed waa c�d he sneez e i l a eternue he sneezed waa cilill d he bel ch .. i l a eructe ? waa k t n he 9roan i 1 a gemi he 9roaned waa sawe?ar� j o ? he saY i l a c huchote he wh i spered 250. 251 - 252. 253. 254 . 255. L M ML nd� ny(i"J9a Yaa nda t ilB ? mY head VP me 256. L M 257. 258. 259. j ' ai mal a l a t@te I have a headache Mw L M nda 1 5' ka Yaa nd� w al ax c:i ? mY b i l e? me j ' ai mal a 1 ' estomac I h ave a stomachache M suna n Y� nda 1.1 di arrhea b e me on j ' ai l a d i arrhee I h ave d i arrhea M nYe nda 1.1 c i t ura bodY-heat be me on j ' ai l a t i erre de la temperature I have a tever M kat�ka nY� nda U hunger be me on j ' ai hi m I ' m hUn9rY 260 . wuu l et us s! c E n k a l a 90 market to al l ons au march� l et ' s 90 to the market 26 1 . 262. nda� baara c i nan j � ( � ) I work do today j ' ai travai l l e aujourd ' hu i I worked today M L n k a nYe 9ba9b� l S that b e house-l arge cel a est une 9rande mai son that is a l arge house bi 9bakUl p�l u the i r house l arge l eur mai son est 9rande the i r house i s l arge 263. 264 . l ilnYi '"Ja col d water 265. kan nlllln 9 i ve me-PP H l o?okaa nY� '"J€ water cool donne moi de l ' eau troi de 9i ve me some col d water l ' eau est trai che the water is cool 266. Pan ni s�nci waYi ! wu come wi th drY wood-pp so-we ba ne nak i i M Put t i re-i n aPPorte du boi s sec Pour I e feu bri n9 some drY wood t or the t i re s�nci nya Ya� wa?E t not drY woed Ie boi s n ' est Pas sec the wood i s not drY 267 . 268 . ' r;r"l::lPE nX� Jl:n. ; e-b l unt H L I')ja Ya?cJ, Ne9 . t u t . be-ab l e un couteau emousse ne couPera Pas de vi ande H 269. 270. kaata k()£ n meat cut a b l unt k n i f e wi l l not cut meat nda ' '1muunn aa PE n th i s kn i t e VP b l unt ce couteau est emousse th i s kn i fe i s b l unt H Y�ra k�n7aan a H l a Your vi l l age far Q est-ce que votre vi l l age est l oi n? i s Your vi l l age t ar awaY ? ------ - ---- ------- 39 1 271 . 272. H waa k.id taB H:d -i he ';0 Pl ace-hr-PP i l est aI l e l oi n he has gone far awaY Mw ML-H , wa.1 Pan na. sama sak.£-! he come wi th fat 90at-PP i l a aPPorte une chevre 9rasse he brou9ht a fat goat 1')9€! sakil')a pun€! nYe sam� th i s 90at al l be fat cette chevre est 9rasse thi s goat i s fat 273. 274 . , wd. Pan na se?etafeed� ! he come with basket-Pretty-pp i l a aPPorte un jol i Pani er he brought a Pretty basket Ad€! se?ena.1 nYon that basket be-Pretty ce Pan i er-l a est jol i that basket is pretty 1')9€! n.1l')a zi:lnl')a.1 nYon that man heard be-900d : ' est un homme bon tha.t i s a 900d man M kaPi I a c i wdra l a theY bad do us on i l s nous on fai t du mal theY did us an i njurY 275 . 276. 277. bH 278. ju wa.1 kapa:<ala he bad <news) tel l 279. LM-H wi:lrCi us-PP i l nous a di t de mauvai se nouvel l es he tol d us some bad news ML ML 1')9a nal')a nYa :l: dnmP� H61S that man be heart-evi l -chi ef c ' est mechant homme that i s a dangerous man nda nYe ' metonl axu-u I be roPe-l on9-PP j e veu:< ur.e 1 0n9ue corde I want a 1 0n9 roPe 280 . 281 . nda that 282. m9b � � n€-€ l a.1 ton l a d€ ce b&ton 1 .1 est 1 0n9 sti c k-there it be-lon9 EX CLAM that st ick is 1 0n9 H v.inmb� baa drt smal l d oth this ce morceau de ti ssu est etroi t thi s Pi ece of cloth is narrow LM-H i:onnacdf6nna nYe ndu-d i ron new Pot be me-PP j ' ai une nouvel l e marmi i:e I have a new Pot mpa 283. 392 2 84. naVCnl")9i new-i t c ' est neuf it i s new 285 . L ML-H wa� Pan n� sd7dl � x ! he come wi th ol d basket-PP i l a uPPorte un vi eux Pani er he brou9ht an ol d basket nd.i sd?dnaa I e: thi s basket ol d ce Pan i er est vi eux thi s basket i s ol d 286. 287. L L ML L M namp�l �d�a w�� P�n ba Yer� wee ol d man one come M You vi si t un viel homme est venu vous voi r an ol d man has come t o see YOU waa I e: he ol d i 1 est vi eux he i s old vanmbi nY� i nYe ��� cl oth-red be h i s il a du ti ssu rouge he has some red cloth 288. 289. 290. H f�l anYe: e: Yaa nYal ��� �j ! those rocks-there VP red l es cai l l oux-l a sont rouges those rocks are red 29 1 . s�e truth 292. 293. w� it i s c ' est vrai it ' 5 true H nt.i ka�taa f�n?6n th i s meat be-rotten cette viande-ci est pourri e thi s meat i s rotten H nd.i '�muunnaa tcinl� kn ife be-sharP thi s ce couteau est c!"' lpant thi s kn i fe i s sharp nta me: e: caci gbere thi s roPe short cette corde est courte thi s roPe i s short H waa 9bexala� 9bere make smal l he i l est vrai ment Petit (de tai l l e> he i s real l y smal l H waa 9be:<ala� ton b make tal l he i l est vrai ment 9rand (de tai l l e > he i s real l y tal l 294. 295 . 296. ------ . ._-- ---- -----_ . 393 297. L ta mof na. c:ome wi th M ML LM-H se?ed d i d H smal l basket-PP 298. H nka f�l akaa 9bex�laa wol axo thi s rock make smooth 299 . LH LH 302. ce b�ton-ci es� droi t thi s sti ck i s strai ght H ce b�ton-ci es� mi nce this sti c k i s thin nda m9b i i na c�r€ thi s stick smal l 301 . ce cai l l ou-ci est tres l i sse thi s rock i s real l y smooth H nda m9b i i naa t�l e thi s sti c k strai 9ht 300 . aPPorte un Pet i t Pan i er bri n9 a smal l basket H waa drE it smal l il est mi nce it i s thi n -H s�m� nYe wi -i fat be he-PP i l est 9ras he i s fat l o?okaa nam�neun hot water 1 ' eau est chaude the water is hot 30.3 . 304. H nka ka"ci kaA nY�I"JE wet this wood 305. ce boi s-c i est moui l l e thi s wood i s wet LH nda m9bH na,g mp�l l: th i s sti ck be-thi c k ce b�ton-ci est epais thi s sti ck i s thick vaanvi i � nYe ��� cloth-white be hi s il a du ti ssu b l anc he has some wh i te cloth 306. 307. ML vaan�906 nYe ��� cl oth-black be his 308. ML H nka vMnnkonk�� 9be:<�laa: pel i thi s cl oth-cut large make 309. --- ce morceau de ti ssu est troP l arge thi s Piece of c l oth is qu i te l arge H nta tux�taa 1 �x6 th i s l oad be-heavy ----- i l a du ti ssu noi r he has some b l ac k cloth - .- . ce fardeau-ci est l ourd thi s l oad is heavY 394 310. 1')9a baaral')a� wa?a thi s work be-hard ce travai l -ci est d i + + i ci l e thi s work i s d i f f i cult nda nYe kac�nn� thi s be thi n9-90od ceci est i mportant th i s i s i mPortant 31 1 . 312. H b�arant�nl')aa nY�nl al')a be-ri Pe bananas l es bananes sont mures the bananas are r i Pe ba�rant�l')a� l � non k6n mP� bananas age arri ve cut l es b ananes sent mure a cuei l l i r the bananas are ready to cut 313. 314. H wu Yaa col OI')£E he Ne9. wel l -Ne9. il est mal ade he is sick l o?o�J namaneun water hot l ' eau est chaude the water is hot 315. 316. H 1')9a ba�rant�nl')aa t�anla these bananas soH ces bananes-ci sont deuces these bananas are soft woral')aa soro kol a bi Her l es nei x de kel a sent ameres the kol a nuts are b i tter 317. 318. L ML n�.fill af ol a r i ch Person 319. L Wl: he is ML waa ci nMoU+ oB he do r i ch Person 320 . 322. i l est devenu r i che he has become ri ch LH z�n7anka Yaa ntO rain VP t al l 321 . i l est ri che he is r i ch ML i l Pl eut i t ' s rai n i n9 ML canl"'}anyHna Yaa ntO sun VP tal l I e sol ei l se couche the sun is 90i n9 down w� nYe ( n > ka f � aka n�n?-i n snake be this stone under-PP i l Y a un serPent sous cette Pi erre there is a snake under thi s stone ------- ----- ------- 395 323 . nika ne c:ona n�n?-i n B rs Put Pot under mets Ie feu sous Ia marmi i:: e pui:: i:: h e f i re under i:: h e Poi:: c:ona nYe nika l a Pot be t i re on I a marmi i:: e est sur I e feu the Pot i s on the t i re 324 . 325 . H ka tt x f nY�I'"J9a l a i t Put 9round on mei:: s c:a Par i:: erre put that on the 9round nda� wu nYa kilna l a I h i m see road on je kilna Yaa sf l aka n ksna l a road VP 90 ri ver si de on I e c:hemin longe l a rivi ere 326. I 1 ' ai vu sur I e chem i n saw him on the road 327 . ? 328. nda tac:6r�rya nYa wu wua�a mY Pl antat' on be h i s one <VL? ) ma pl antati on esi:: Pres de son Pl antation H 329. i:: aaun besi de-PP mY pl ani:: ati on i s near h i s Pl antati on ba yere � nk�d� l a M stand mY side on ti ens-toi debout � c8te de moi stand beside me ba Yere wa so?ol i stand our betwl �n ti ens-toi debout entre nous stand between us 330. 331 . wu his na and 332 . L M ML tac:6r�rya nYe nd� wd�a pl antati on be mY one k�n7�nka so70l i vi l l age between H -M Mw waa Y�re sac:ai:: a i:: o iWi -i he stand uP peopl e mi ddle i l s ' est mis debout au mi l eu des gens et s ' esi:: mi s � Par l er m:Li h e stood uP i n the midd l e of the c:rowd and began to sPeak M ya and sPeak 333. sa pl antati on est entre ma Pl antati on ei:: I e vi l l age h i s i='l antaH on i s between mY Pl antation and the vi l l age ML wu 9bak� nYe k�n?�nka n�ryH h i s house be vi l l age ceni:: r e sa mai son est au mi l eu du vi l l age h i s house i s in the mi ddl e ot the vi l l age :396 334. H waa yi!re wcJ nt�:< �ta:<arya he stand h i s c h i et ML nY.1?An 1.1 bet ore PP 335. 336. 337 . M H waa yi!re 9baka nYo 1 .1 h e stand house mouth PP ML Mw? waA ' rymuuna c� w� n YcJry9a mpAnryi i n,; k n i t e grab i l,:S head above-PP nt � i mo� ry9a ndi rya SJ?jj joe how YOU th i s tood cook saY-Q i 1 s ' est t enu debout devant son ch i et he stood betore h i s chi et i 1 s ' est tenu debout devant 1 a mai son he stood i n front ot the house i 1 a tenu I e couteau au dessus de sa t@te he 9rabbed the k n i t e above i ts head ( by the han d l e? ) comment as-tu PrePare cette nour r i t ure? how d i d You PrePare t h i s tood? 338 . nYA7an waA kan mo6 y� what he 9i ve You-PP Q qu ' est-ce qu ' i 1 ta donne? what d i d he 9i ve You? nYA7anxa nYe si!?i!n-i Y£ be basket-PP Q what qu ' y a-t-i 1 dans I e pan i er? what i s there i n the basket? nYA7an tad i i waA Pan y� moment he come Q what quand est-i l venu? when d i d he arr i ve? Yaa see san Yi where You (pI ) VP 90-Q o� al l ez-vous? where are You 90i n9? ML Yi m06 ry9utUn waa k�ra j u who-other he i tRF.F saY You-PP Q qu i t ' a d i t ce1 a? who t o l d You that? 339. 340 . 34 1 . 342. 343 . 344. H nYA7an 1 a Yeraa k.1d w�mpa� what tor YoU (pI ) 90 over there Mw Yi n k a I �r a nYe joor i that Pr i ce be how manY Q Pourquoi @tes vous al l es U-bas? why d i d You 90 over there? comb i en ce1 a coQte-t-i 1 ? how much does that cost? 397 345 . nd.a sheep L M b i nYe I')ilun Yt j oor ! be h i s Q how manY eombi en d e moutons a-t-i l ? how manY sheep does h e have? 346. 347 . sa'n mo b:il')a nYE t n where Your f ather be-Q oil est ton p�re? where i s Your f ather? L M ML Mw? n ya?an H l ax a kaari Yt what tYPe meat Q quel l e sorte de vi ande est-ce? what k i nd Ot meat i s that? 349. H waa ndlt ci wu y06 h e that do h i mseH 349. 350 . 351 . 352. il a t ai t ca l u i -m@me he d i d that h i msel f H ndaa nda ci na Yd6 I that do mYsel f j ' ai fai t e a moi -m@me I d i d that mYsel f M bii bi YS nYa they themsel ves see i l s se sont vus they saw themsel ves H M wdraa wll Ye nYa we oursel ves see nous nous sommes vus we saw oursel ves -H ta rna na I C come wi th ndll')i -i tood-PP 353 . aPPorte I e rePas b r i n g the meal H cuI Il1"'J:) be-healed Y�1"'Jaa s i c k Person 1 e mal ade est 9u€!ri the si c k Person i s cured 354 . c��b i Ya b i Y S k6d ! women TA themsel ves chase- IC 355. H Ya9bak ai Yd7dri t h i n 'ii-dr i n k 356. 357. l es t emmes se pourehassent the women are chasi ng each ct�er la f @te est i n teressante the teast is i nterest i ng H Yal at� kOl UI')5 thi n9-eat l ac k l a nourr i i:ure est i nsut t i sante the t oad is not enough H ncek .fa: kj soaP .f i n i sh - .-- ----�--- I e savon est t i n i the soaP has run out 398 358. 359. 360. 36 1 . 362. b�l �kaa wel l cil: d be-deep , silmbaa wil can mi l l et beer h i m destroY Le boi sson l ' a detru i t Dr i n k r u i ned h i m wu Yaa y6?6ru he VP gossi P i 1 bavarde he goss i Ps . tal ks al ot H l')a f::J f on i o wi nnow vannes I e f on i o winnow the f on i o H 9baka pj house sweeP �----- Ie pu i ts est profond the wel l i s deeP bal ai e l a mai son sweeP the house --- - �-- -.----------. ----- 399 VITA Anne El izabeth Garber was born i n Toronto, Canada on FebruarY 8, 1 955 . She attended Oakri dge SecondarY School i n London , Ontari o and the Col l ege General et Pro�essionel i n Jol i ette, Quebec , 9raduati ng �rom the Un i versi ty o� ottawa wi th a B. A. in Li ngu i st i cs i n 1 978 . In June 1 978. she began graduate studY i n Li ngui sti cs at the Uni versi ty o� I l l i noi s and recei ved her Years at the Uni versity at a M . A . i n 1 980. Duri ng I l l i noi s. she held research assi stantshi Ps i n Li n�lui $t i cs and Art Hi storY. as wel l as a teaching assi stantshi p i n ESL. Duri ,g the academei c Year 1 981-82 she attended the Associ ated Mennon i te Bi b l i c,tl Semi nari es i n El khart , Indi ana, a�ter whi ch she sPent the Years 19821 985 i n i<otoura, Burkina Faso doi ng l i ngui st i c research with Afr i ca Inter­ Mennoni te Mi ssi on . In June 1 987 . she i ntends to return to Burkina to conti nue working �or A�rica Inter-Mennoni te Mission .