A COM PARATIYE GRAMMAR OF THE SOUIH-AfRICAN BANIU LANGUAGES f^SlOBi-o.-Klway BOUGHT WITH THE INCO FROM THE SAGE ENDOWMENT THE GIFT OF m^nv^ M. Sage r89i ilTOTg ME FUND llllJllr..... A.yjPii !i,i.,: f 1924 077 077 612 iHQ\f :o /907 -3j 3^. no paDE^d ajtsJ nosjad ana hbih ajoin ;Cq papaan sapog ■pa^tiEM jfx'aanasqB s,j3 -iiojjoq Snunp njn;aj nam joj apem s^nam -a3uBJiB JO 'X!jBaq;i aqi 0} paujtuaj aq Pinoqs . - spouad ssaaaj Saunp/U t\ I' jT^HjiHr*''* papaan ^on sifoog ■ ^ •sjaqio jCq papaan ^on, sj 3iooq B H3IIM. 'saSaj .-TAud iBMauaj q^M*. 's>[aaM. oM% zo} samn -^6a 3Ag paMojiB aJB BJaMojjoq pajirajl •a^Bin aq pin6qs ^sanbaj lEiaads ' • , B s3[aaM OAv; pno/Caq papaan naqM '. a^qtssod _ SB igoos SB panan^aj aq pjnoqs japBJBqo x^-ia -naS B JO siBOipouad; ■' -sja^oj -joq lie o| S5[aa4i jnoj 01 paiimii ajB qajBas .31 JO nofionjisni joj asn nt ^on sqooq nv 'U3>|e^sGAiaiun|0A sm:^ u9i(Afi sm.oi|s ovep am. Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924077077612 A COMPARATIVE GRAMMAR OF THE SOUTH-AFRICAN BANTU LANGUAGES. ST-AUSTIN'S PRESS, DESCLfiE, DE BROUWER AND CO., BRUGKS. SOUTH-AFRICA REFERENCE MAP TO ACCOMPANY THE Comparative Grammar OF THE South -African Bantu Languages BT J.TORREND, S.J. N. B. The names -printed in red are those of the languages more particularly dealt with in this work .Non-Bantu Lan§ua§es; Wi Bantu intermixed with non- Bantu Lan^ua^es., aoi,ie[» S* -^u^ti.'ti A comparative" GRAMMAR OF THE SOUTH-AFRICAN BANTU LANGUAGES COMPRISING THOSE OF ■ ZANZIBAR, MOZAMBIQUE, THE ZAMBEZI, KAFIRLAND, BENGUELA, ANGOLA, THE CONGO, THE OGOWE, THE CAMEROONS, THE LAKE REGION, ETC. BY J. TORREND, S. J., OF THE ZAMBEZI MISSION, AUTHOR OF " AN OUTLINE OK A XOSA-KAFIR GRAMMAR ". LONDON KEGAN PAUL, TRENCH, TRUBJSTER & CO;, Lt-i. PATERNOSTER HOUSE, CHARING .CROSS ROAD. 1891. CORNEil UNIVETtelTY ^"^ LIBRARY^ A. H-'Soqi The rights of translation and of reproduction are reserved. TO THE MOST HONOURABLE THE MARQUIS OF BUTE, K. T., THIS WORK IS BY PERMISSION DEDICATED. PREFACE. However favourably my friends may have thought of this work when still in manuscript, I cannot flatter myself that it comes near to the 'perfection to which I should have wished to have been able to bring it. Any criticisms, corrections, additions, or suggestions, will be received with hearty thanks. There is no need to call the attention of any one to the impo-rtance of the study of Bantu. Independently of its scientific interest, it is a key for opdning one half of an immense continent to Christian civilization. ■ I will only add a word of thanks to all those tO whom I am indebted for help, whether from their published works in the same line as this, or from private advice and inform- ation. I feel particularly indebted | to the following friends : The Rev. J. T. Walford, S. J., for having very kindly looked over, and corrected, the greater part of the En- glish of my MSS. and proofsheets. Dr. R. N. Cust, for having no less kindly given me a number of modern publications on the Bantu languages, among others most of those of the S. P. C. K. His " Sketch of the Modern Languages of Africa " has also been to me an invaluable guide throughout. The Revv. Fathers Causse, Temming, Ronchi, Andrd, S. J., for information or MSS. regarding the Kafir, Chwana, Senna, and Kilimane languages, respectively. The Rev. Father Ldvesque, of the Socidtd de Notre- Dame des Missions d' Afriqiie, for having kindly sent me all the publications of his Society on the Bantu language^ VIII Preface. , The Rev. Father L. Cheikho, S. J,, for copious in- formation on South-Africa derived from the ancient Arab writers. The Custodian of the Grey Library in Capetown, and the officials of the British Museum. Above all, the Rev. Father Depelchin, S. J., the founder of the Zambezi Mission, at whose bidding I undertook these studies. When he came back to the Cape Colony in 1883 from his laborious missionary explorations in the far interior with broken health, but an undaunted spirit, I had the advantage of enjoying his company for nearly two months at St. Aidan's College, Grahamstown. All this time he was constantly saying to me : " For the love of God learn the native languages. I have come across millions of men who need but to hear Our Lord's words and deeds to become so many good and happy Christians ". These words have been ringing in my ears ever since that time, giving me courage and strength to persevere in my attempt to do so. But for them, this work probably never would have been undertaken ; certainly it would not have been brought to an end. i I pass by some other friends, who will not allow their names to appear in these pages, but whose kind help will not be forgotten. God grant that this little work be not useless to the evangelization and civilization of Africa ! St. Aloysius' College, Jersey. Whit-Sunday, May ly, i8gi. CONTENTS. INTROPUCTIQN. Page I. Division of the South-African Languages xv II. Bantu Literature. — Sources xxil III. The Origin of the Bantu xxxill Chapter I. GENERAL PRINCIPLES AND PHONETICS i I. Alphabet 2 II. Characteristic Features of the Bantu Family of Languages 6 First principle 6 Second principle 8 Third principle 9 Fourth principle 11 III. Comparative Phonetics of the Principal Bantu Languages 14 Tonga 14 Yao 15 Nyamwezi 17 Sagara, and Gogo 18 Shambala, and Boondei 18 Taita 19 Kamba 19 Swahili ... 20 Nyika, or Nika, and Pokomo. ... 21 Senna (including Tette and Nyassa) 23 Karanga (aZ/af Kalaka) 24 Ganda 26 Kafir (Xosa, Zulu, and Tebele) 27 Herfero 28 Bihe. ... i 20 Mbunda, Lojazi, Nano, and Ndonga 30 Rotse ; 31 Runda, or Limda, and Luba 33 Rua .- 33 Angola, Mbamba, and Fiote, or Lower Congo 34 Middle Congo Languages 36 Nywema 37 Kua, or Mozambique, and Chwana 38 Tshagga, and Hinzua 44 Mpongwe 44 Dualia. ... i 46 Fan 47 X Contents. I Page Ferna^dian (Fernando Po) 49 Languages of the Congo Forest 50 Semi-Bantu 51 Conclusion 51 IV. More General Phonetic Changes 53 § I. Changes caused by the Collision of two Vowels S3 § 2. Various Phonetic Changes 56 V. On Accentuation in Bantu 61 Chapter 11. ON SUBSTANTIVES 63 I. On Articles 64 II. The MU-BA Class and the Sub-classes connected with it 67 § I. Transformations of the Classifier MU- 68 § 2. Transformations of the Classifier BA- 70 § 3. The Sub-Class — 'BA 71 § 4. The Sub-Class MU-MA 72 § 5. Substantives which belong to the MU-BA Class and the Sub-Classes , connected with it. 72 § 6. Etymologies. — Varia... : - 73 III, The MU-MI Class 76 § I. Transformations of the Classifier MU- 76 § 2. Transformations of the Classifier MI- 77 § 3. Substantives which belong to, the MU-MI Class 79 § 4. Etymologies. — Varia 80 IV. The IN-ZIN Class 82 § I. Transformations of the Singular Classifier IN-., 82 § 2. Transformations of the Plural Classifier (Z)IN- 84 § 3. Substantives which belong to the IN-(Z)IN Class 85 ,§ 4. Etymologies. — Varia.. 86 V. The LI-MA Class. 88 § I. Transformations of the Classifier LI- 88 I. Polysyllabic stems which begin with a consonant 88 11. Monosyllabic stems. ... 90 III. Stems which begin with a vowel 91 § 2. Transformations of tjie Classifier MA-. 91 § 3. Substantives which belong to the LI-MA Class 91 § 4. Etymologies. — Varia ^. 93 VI. The BU-MA Class and the Sub-classes connected with It 96 § I. Forms in the Class BU-MA and the Sub-class MA 97 § 2. The Sub-classes BU without plural, and BU-{Z)IN 94 §3. Substantives which belong to the BU-MA Class and the Sub-Classes connected with it 99 § 4. Etymologies. — Varia 100 VII. The KU-jyiA Class... 102 § I. Forms \ 102 § 2. Substantives which belong to the KU-MA Class 103 § 3. Etymologies. — Varia 103 Contents. xi Page VIII, The LU-(ZIN) Glask and the Sub-classes connected with it 104 § I. Transformations of the Classifier LU- 104 § 2. Plural Classifiers. corresponding to LU- 105 § 3. Substantives belonging to cl. LU- 106 § 4i Etymologies. — Varia... ... ... ... , i 107 IX. The GI-ZI Class 109 § I. Transformations of the Classifier CI- 109 § 2. Transformations of the Classifier ZI- ill § 3. Substantives which belong to the CI-ZI Class. 112 § 4., Etymologies. — Varia 113 X. The diminutive Class KA-TU and the Sub-classes connected with it. 115 § I. Forms 115 § 2. Substantives which belong to the KA-TU Class and the Sub-Classes / connected with it 119 § 3. Etymologies. — Varia 120 XI. tocative Classifiers, and Prepositions.. 122 § I. Transformatidns of the Locative Classifier PA- 123 § 2. Transformations of the Locative Classifier KU- ... 127 § 3. Transformations of the Locative Classifier MU- 1:8 § 4. Plural Locative Classifiers ... ... 129 § 5. Effects of the Locative Classifiers on the other Prefixes of the Substantives. 129 §6. On the Use of the Locative Classifiers... ;.. ... 130 § 7. Prepositions which are not Classifiers 131 § 8. The Particles -L/, -A'/J, etc., in Locative Expressions 133 § 9. Etymologies. — Varia 134 Xir. Copulative Prefixes' before substantives 136 _ XIII. The Particles which introduce Substantives after Passive Verbs.... 138 XIV. Suffixes of Substantives 138 XV. bnomatopoetic Substantives. ... 139 XVI. Retrospect. — Varia '140 Chapter III. ON ADJKCTIVES. 142 I. Quantitative Adjectives 144 § I. Adaptation of the Quantitative Adjectives to the different Classes of Nouns ; 144 § 2. Effects of the Phonetic Laws upon the Forms of Quantitative Adjectives. 145 § 3. On the Use of Quantitative Adjectives as Epithets. ... '. 146 § 4. On the Use of Quantitative Adjectives as Predicates 148 II. Non-Quantitative Adjectives., .• 149 III. Comparatives and Superlatives 150 Chapter lY. ON PRONOUNS 152 I. Connective Pronouns 152 § I. Forms. ... " 153 § 2. Connective Pronouns prefixed to verbs as Subjects 155 XII Contents. Page § 3. 'Connective Pronouns prefixed to verbs as Objects iS7 § 4. Reflexive Pronouns 158 II. Substantive Personal Pronouns 159 § I, Forms 160 I. Enclitic Forms 160 II. Self-standing Forms 162 III. Copula-containing Forms 164 § 2. Use of the diflferent Forms 165 I. Self-standing Forn»s ' ... 165 II. Enclitic Forms , 167 III. Copula-containing Forms 169 \ 3. Varia ' 169 III. Demonstrative Pronouns 170 § I. Fundamental Forms 171 I. Formation of these pronouns 173 II. Use and place of these pronouns 174 § 2, Emphatic Forms. 177 § 3. Copula-containing Forms 178 IV. Relative Pronouns, and Relative Particles 181 § I. Forms of the Relative Particles 181 § 2. tfse of Relative Particles and Construction Of Relative Clauses in General. 184 I. Relative clauses in which the antecedent is represented by the subject of the verb. 184 11. Relative clauses in which the antecedent is represented by an object of the verb. i85 V. Pronouns in Possessive Expressions i8g § I. General Principle 189 § 2. Connective Pronouns Suppressed. ... 190 § 3. Possessive Expressions after Locatives 191 VI. Relative and Possessive Kxpressions used Substantively 194 VII. Relative and Possessive Expressions equivalent to our Adjectives... 197 YIII. Pronouns referring to Substantives understood, and Pronouns used as Conjunctions >• ... 198 Appendix on the Lunda Language... 201 IX. Numerals 203 § I. Bantu Numeration ... 203 § 2. Formation and Use of the Numbers from " One " to " Six " 206 § 3. Formation and Use of the Numbers above " Six " 207 § 4. Complex Numbers 207 § 5. Ordinal Numbers and Numerical. Adverbs 208 The negative particle before the number " one " 208 X. Interrogative Pronouns and Various Determinatives 208 § I. The Pronoun " Hovi'many?" 209 § 2. The Pronoun and Adjective " What.' What sort of... ? ".* 209 § 3. The Pronoun " Who ? ". 211 § 4. The Discriminative Pronoun " Which ? " '. 211 § 5. Interrogative Pronouns used Indefinitely 212 § 6. The Pronoun and Adjective "All, whole" 212 § 7. The Pronoun ./4 — -/?'/iv " Alone, By himself"- '. 213 5 8. The Pronouns A-a-la-kue " He also ", A-ba-la-bo " They also, " etc.... 214 ^ g. The Pronouns rendering " Self. " 215 § 10. The Pronoun -7«W " Other, Different, Foreign. " 216 § II. The Pronouns " One... another ''," Some... others. " 216 Retrospect on the Article 217 Contents. xiii Page Chapter V. ON VKRBS 219 I. Fundamental Forms of the Simple Verb 219 § I. Principal Parts of the Verbs in Bantu 219 § 2. Fundamental Forms derived frorn -^o«a ■ 221 I. Imperative form fona " see " 221 ' II. Indicative form «ifa-&«ffl " I see ". ' 222 III, Infinitive form fe-fo»ffl " to see " 225 §3. Subjunctive Form «^/-^^K^ 225 § 4. Perfect Form «rf2-i5o«zfl5? 227 § .5. The Forms »rf«-i5o«a«f« and «(/2-^o«ag'a 229 , § 6. The Negative Form C/a^«(//-i5f«t. ... 23d II. Auxiliaries 231 § I. General Principles 231 § 2. The Negative Auxiliaries • ... 232 I. Forms ... ... .^. 232 A. Absolute indicative^ clauses 233 B'. Relative clauses ^ 233 C. Subjunctive, clauses. ... ,. ... 234 X). Imperative clauses and the infinitive 234 II. Examples 235 § 3. The Auxiliary -.4 237 § 4. The Auxiliary K4 " to Go " 242 § 5. The Auxiliary £'iVZ'.i4 " to Go ", and various Continiiative Auxiliaries... 243 § 6. The Auxiliaries ^^Z^ and AW/4 " to Sit, to Remain " 247 § 7. The Auxiliary Z^ or /Z/i " to Come " ... 248 § 8. The Auxiliary A' C/5 251 § 9. The Auxiliary if^ 251 § 10. The Auxiliary /A^5/... ' 254 § II. The Auxiliaries MA and BA " to Stand, to Stop. " 255 § 12. The Auxiliaries C/, ir/, 5/, 5^, etc 256 § 13. The Auxiliary TVG:^ 258 § 14. The Auxiliary r/ " to Say. " 260 § 15. The Auxiliary .S^/F^" to Comeback. " ._. 261 § 16. The Auxiliary il/^A''^ " to Come to an End. '..' 362 § 17. Various Auxiliaries 263 in. The Verbs " To Be " and ." To Have. " 263 § I. Copula Understood 264 § 2. The verbal Forms LI, LE, ELE, IRI, etc., used as the Copula 264 § 3. The Verb KU-BA " to Become, to Come to be, " used as the Copula. ... 266 § 4. The Verbs -KALA and -NNA or -IN A " to Sit, "used as the Copula... 267 § 5. The yerb -^ATZ)^ used as the Copula. 268 § 6. Various Copulative Particles 268 § 7. The Copula in Negative clauses 269 § 8. The Verb " To Have. " ...270 § 9. The Verbs "To Be" and "To Have " in Locative Expressions 271 IV. Derivative Verbs ." 272 § I. Passive Verbs 272 § 2. Other Derivative Verbs 275 XIV Contents, I. Applicative verbs II. Causative verbs III. Intensive verbs IV. Reversive and expansive verbs V. Reciprocal verbs Cbnclusion '.. Retrospect on Adverbs, Prepositions, and Conjunctions 'Appendix I. Ethnographical Notes in Tonga. \. On the Rotse. I]. On the Karanga III. On the Tonga Appendix II. Specimens of Kaflr Folk-Lore. I. The Bird that made Milk II. MlonjaIoniani,his Sister, and a Mbulu... III. Ths Gqongqos an^ Qajana. IV. Tanga-io-mlibo Alphabetical Index. Page .. 276 ' ...277 •■ 279 .. 279 .. 280 .. 280 ... 281 ... 283 283 ... 286 289 .. 296 .. 296 ... 3°° 305 314 INTRODUCTION. I. Division of the South- African Languages. /. Whatever may be the correct division of the native races of South- Afrika, the languages of this country constitute three plainly distinct groups viz. the Hottentot-Bushman, the Masai, and the Bantu. With the first two 1 have not to deal in this work. If they are mentioned here', it is only to set them soon aside. 2. The Hottentot-Bushman group. — This comprises the languages spoken mostly by nomadic, or only half-settled people, who are found in the least accessible parts of the South-African deserts. Living in caves or in wretched huts, too lazy to cultivate the soil, eating such food as bull-frogs and lizards, wanting in what the Kafirs call marriage-laws, having no notion of political union, they are generally despised, and persecuted, or kept in subjection, by their Bantu neighbours. They are of every description with regard to colour, stature, physique, and dispositions. Some are yellow-white, others red, others reddish-black. Most of them are dwarfish in size, scarcely above four feet, but they also number fine specimens of humanity, such as the six-foot Lan^e Berg Bushmen near the Orange River. Some, have fine proportions ; others are of the very lowest type, with short foreheads, and hair on their, bodies and legs. Some are of gentle disposition, ready to do any service V^thers wage war on all living beings, and cannot be trusted with anything (')■ J. They used to be found in ancient times — as possibly, they may be found yet — even in the north-eastern deserts of Africa, and from the fact of their living in caves (Sia to TpwyT^a; iTtoSeSuxevai. (2)) were known to the Greeks under the name of Troglodytes. The most generic name they have among their Bantu neighbours is that of Ba-tua, or in Chwana pronuncia- tion Ba-roa, which now means " slaves ", and is synonymous with ba-bua, or in-ja, " dogs ". Southern Kafirs distinguish, as we do, between the pure Bushmen and the more civilized Hottentots, whom they consider to be a mixed race. These they call a Ma-lawu, which according to regular phone- tic changes seems to stand for Arabu, as if they once had had something to do with the Arabs. Probably fhe southern Bushmen are related as a race to the dwarfs who live on the north-eastern affluents of the Congo. The latter, however, seem to speak semi-Bantu languages (n. 242 of this work), I. See Anderson's Twenty-five Yean in a Waggon, London, 1887, vol. I, ppj 282, 296, etc., and vol. II, p. 74 ; also the Proceedings of the R. G. S., 1B86, p. 438. , z. Geogr. Graeci Minores, Didot, 1861, vol. II, p. 248. XVI Introduction. There is ground to believe that either these, or the Bantu proper, have preserved the original language of South-Africa the best, while the southern Bushmen, whose ancestors were, perhaps, the slaves of foreign gold and diamond diggers, have forgotten it entirely. ^. The most prominent features which distinguish the languages of the Hottentot-Bushman group from Bantu are : — i°) a great abundance of those peculiar consonants which are termed clicks (nn, 35-38), and have been compared by Herodotus to the screeching of bats, TeTptvaai xaOaTcsp a'. vuxTspiSe; (Iv, 183); --- 2°), a grammatical system built nearly exclusively on sex-denoting suffixes, while the Bantu mechanism consists mostly of pre- fixes which imply no such reference. 5. On the whole this group of languages differs perhaps more from the generality of the Bantu languages than from any other. The late Professor Bleek has remarked in it signs of affinity with some North-African lan- guages ('). He has even come to the remarkable conclusion that " all those sex-denoting languages known to us in Africa, Asia, and Europe, arc members of one large family, of which the primitive type has, in most respects, been beat preserved to us in the Hottentot language (^) ". 6. The Masai group. — The Masai are warlike tribes with pastoral and nomadic habits, which occupy a large belt of ground south of the equator from Mount Kenia to south of Mount Njaro, or Kilima-Njaro. They are said to resemble in a high degree the Somali. They are divided into Masai proper and Kwafi (3). H. H. Johnsto» has observed that Latuka, 5° north of the equator, and Bari, on the White Nile, between 4° and 6° Lat., are members of the Masai family of languages (*). Many points of contact might be shown to exist between Masai and Bantu, but, as it would require a somewhat lengthy explanation to bring them out, I have thought it better not to touch them in this work. It has certainly more in common with Galla than with Bantu. 7. The Bantu group. — The third, and more important, group of languages spoken in South-Africa, the one which I have attempted to describe in this work, may be said to comprise the idioms spoken by all the agricultural black tribes of this country. Bieek, who did more than any one else to throw light on its numerous ramifications, proposed to the scientific world to term it Bantu, because this word, which properly means " people " in most of the languages of this group (n. 322*), is principally used by the natives when speaking of themselves in contradiction to white people (s). This terra, whatever may be thought of its correctness, has been adopted on so good an authority, and is now the current name. 8. There can be no doubt that these people must be identified with the 1. Comparative Grammar of South-African Languages, I, p. viil. 2. See Reynard the Fox in South-Africa, pp. xiv-Xix. 3. See Introduction to the Vocabulary of the Enguiuh Eloikoh, by the Rev. L. Krapf, Tubingen, 185^. 4. The Kilimanjaro Expedition, London, 1886, p. 450. 5. MS. 214 of the Grey Library, Capetown, entitled " Thirty chapters of Zulu Tradition", chapter v. Introduction. xvii Zindj of the ancient Arab geographers. I grant that I find no distinct mention made by them of the western Bantu, but they distinctly include under the name of Zindj all the eastern tribes dwelling between the Juba River and Delagoa Bay ; and this says enough, as it means all the Bantu tribes known to them. g. It also seems certain that " the large country called Agi-sumba, or y4^2-jy»«(5a: ", by Ptolemy, the existence of which was known to this geo- j;rapher as far as the i6* parallel of south latitude Q, is no other than the Bantu field. The JVEasai still call the Swahili La-shumba-n, and the Kavirondo, a non-Bantu tribe dwelling north-west of Lake Victoria Nyanza, call them Wa-ki-ckumbi. A few Bantu tribes also call themselves Ma-fZimba, or in Mozambique pronunciation Ma-rimba (n. 173), which, perhaps, may be etymologically identified with these words. Then there are the Ki-rimba or Ki-zimha islands north of Mozambique. In some parts of the Congo basin the chief-town of a king is still called 7^a-j«;«5«, as formerly' that of the Monomotapa was called Zimba, or Zimba-we, or Zimba-bye, — all words in which we probably find the element sumba, or symba, of Ptolemy's Agi- sumba f"). 10. It has been repeatedly said that the Bantu have no generic national name for themselves. This is not quite correct. My native informants, those of the Zambezi as well as those of Kafraria, gave me independent evidence that all the native tribes of which they had any knowledge, the Bushmen and Hottentots excepted, were included under the generic name of Ba- nsufidu. This is the word which is variously pronounced Ba-sutu, Be-suto, Ba-suto, A-sutu. I do not know whether it may not be traced in Ba-sundi, which is the name of a large Bantu tribe on the Congo. Certainly it must be identified with the word A-suut, or A-suur, oi the. Fan tribe on the Upper Ogowe. It seems to mean " thedark-brown tribes ". This at least is the meaning which southern Kafirs assign to it. I should not be astonished if it were found to be related to the word Soudan, " Blacks ", of the Arabs. 11. Languages distinctly Bantu are heard in all the well-watered parts of South-Africa from the Keiskamma River in Cape Colony to the equator in the east, and from Walfish Bay to the Old Kalabar River on the 5* parallel of north latitude in the west. In most parts of Central Africa the Bantu field extends but little north of the equator, There are some Bantu enclaves in the Soudan, on the Niger, and further to the west. Philo- logical science has not yet determined what is the exact relation of ithe languages of the other black tribes in the north-west to Bantu. For myself, I have come to the conclusion that several of them have at least as much in common with the southern Bantu languages as certain Aryan languages between themselves, English and Greek for instance. But, except for a ftw ±. Geogr., I, 8 and lo ; iv, g. a. In several eastern Bantu languages, the word simba means " lion ", which is synonymous wi^h ' king ". Perhaps it is also related to Agi-sitmia. * , XVIII Introduction,. short digressions on this subject (nn. 245, 598, and 830), I have limited my field of study to those languages, which differ from one another lio more than English does from German. 12, Classification of the. Bantu languages. — Notwithstanding the exist- ence of a considerable amount of literature, the study of the Bantu lan- guages in general must still be said to be in its infancy, and I think that any attempt at their scientific classification must fail for some time. Bleek attempted one. It is not only inadequate, but entirely misleading from be- ginning to end to one who has comparative philology in view. He does not seem to have noticed, for instance, that Chwana has much more affinity with Kua of Mozambique than with Zulu, nor that Mpongwe differs more from most of the languages of the Congo than from those of Mozambique. When I began these comparative studies, one of the first things which struck me was the existence of a group embracing Chwana, Mozambique, and Mpongwe, and further researches have only confirmed this view. But I have found no other neatly defined group. Hence, taking all the languages that have some particular affinity with those of Mozambique to form the Kua, or Chwana-Mozambique-Mpongwe, group (169 and 246), nearly all the others may be provisionally considered as forming the main group. Those of Fernandp Po, and, probably, certain little know^n Bantu languages of the Cameroons and the Soudan, do not come well into either the main or the Kua group. They also provisionally may be considered as forming, the Fernandian group. 13. Dr. Robert Needham Cust, dealing with these languages in his " Sketch of the Modern Languages of Africa ", follows a geographical. method throughout. Hence his classification necessarily has its defects, but less than any other that I know of ; and I think it may be adopted until more is known of some languages, principally those of the Congo basin. Only it should be so modified as to pay due regard to the existence of the Chwana- Mozambique-Mpongwe group, and to certain obvious affinities between various languages: Thus, instead of a general division of these languages into a Southern, ah Eastern, and a Western branch, I should begin with their division into the main and the Kua group, with the addition of the Fernandian. Then each of the first two I should subdivide into an Eastern and a Western half-^group. The meridian of the Victoria Falls would be the approximate line of demarcation between east and west, as nearly all the tribes to the west of this limit are included by the natives under the names of Ma-mbunda or Ma-mbundu, Ma-kwango, E-xi-kongo, and Am-pongwe, all of which mean " western people " ('). The word Si-ongo, vihicla. is the native name of the Falls, seems even to mean "the separation, or beginning, of the west. " L. It appears that in the Portuguesa colony of Angola the word A-miundu, or A-mbunda, is thought to mean ' ' the invaders ' '. Tnis certainly cannot be its origihal meaning : for the simpler word mbunda means ' ba(^ ", h«ioe " west ", in several of the Mbunda, or Mbundu, languages Introduction. XIX Each of the half-groups may further be subdivided into clusters, according to the greater or lesser afifiinity of the various languages. l/j.. Hence the following might serve as a provisionaldassification of the best known among these : — I. xnain (5roup. Eastern Half/ Kafir cluster. Karanga CLUSTER. , Tonga cluster. Sagara cluster. Senna cluster. VlTI CLUSTER. Gangi cluster. Xosa or Kafir proper, spoken in Kafraria and the Transkei. Zulu, in Natal and Zululand. Mfengu, in Swaziland.' TabeleJ or Tebele, in Matabeleland. Vumbe (the Se-kalaka of the Bechwana) in Southern Matabeleland Shona, in Eastern Matabeleland. Karanga proper, by Wange's people north of the Middle Zambezi. ' Yeye, on the Zoiiga River and round Lake Ngami. Tonga proper, between the Kafuefue and the Zambezi. Lea, east of the Victoria Falls; Subia, west of the Victoria Falls. Bue, on the Zambezi, north-east of Moemba's. Kova, between the Kafuefue and the Loangwe River. Bisa, between the Loangwe and the Chambezi River. Bemba, north-west of the Chambezi River. Nyassa Tonga, east of the Loangwe River. Senna proper, at Senna. Shire, on the Shire River, ' Sofala, at Sqfala. Tette, at Tette. Zumbo, or Ntsua, at Zumbo. Nyassa, on Lake Nyassa. Gindo, from the Rufiji to the Lfndi River. Ngoni, west of Lake Nyassa. Viti proper, on the Upper Rufiji. ,, ,j-, Bunga, noith-east of Lake Nyassa. ' '\ '.- Gangi proper, or Henge, Ungu cluster. Ziraha, Kwenyi, Nkwifiya, „„ jhe Upper Rufiji N'l""'^^' f and its affluents. Bena, Sango, Kimbu, Nyaturu, "^" J- on Lake Rukua and its affluejits. Fipa i Kaguru, or Sagara proper, Itumba, Kondoa, , ' in Usagara. Kami, Khiitu, Gogo, in, Ugogo, Hehe, on the Upper Rufiji. -^W- XX Introduction. Nyamwezi CLUSTER. in Unyamwezi. Taita cluster. NiKA, or Nyika, CLUSTER. round Mombasa. INyanyembe, Sumbua, Sukuma, ) Tusi, or Ha, north-east of Lake Tanganyika. Regga, or Legga, west of Lake Mut'a nzige. Ganda cluster, -f Ganda, north of Lake Victoria Nyanza. ' I- Nyanibu, south-vrest of Lake Victoria Nyanza. Pare, near Kilima Njaro. Tambi, \ Teri, >■ on the hills between Kilima Njaro and Mombasa. Ji". ' ' Daruma,^1 Rabai, Giriama, I Digo, Pokomo, on the Tana, or Pokomo, River. Kamba, from Mount Kenia to Kilima Njaro. Lamu, in Lamii Island. Gunya, in Patta Island. Mvita, at Mombasa. Pemba, in Pemba Island. . Unguja, at Zanzibar. ' Shambala proper, on the Shambala hills. Boondei, betweeij the coast and the Shambala hills. Zeg'ula, inland from Zanzibar. ^ Nguru, west of the Zegula, f Lima, on the coast opposite Zanzibar. I Ibo, in Ibo Island (12° 20' S. lat.). Zararao, in tjzaramo, south of Zanzibar. Konde, on the Lower Rovuma. Yao, between the Upper Rovuma and the Lujenda River. SWAHILI CLUSTER. Shambala CLUSTER. Ibo CLUSTER. Western Half. HeRERO CLUSTER, Benguela . CLUSTER. Herero, in Damaraland. Ndonga,- on the Kunene River. Lojazi, near the sources of the Kwando, or Southern Kwango, River. { Bihe, on the Upper Kwanza. Nano, in the district of Benguela. Kwango, or Mbunda proper, west of the Rotse Valley. ( Rotse, on the Upper Zambezi. I Nyengo, on the Nyengp River, west of the Rotse. f (Ci-)Boko, between the Upper Kwanza and the Upper Kasai. I Yakka (?), on the Northern Kwango River. [ Angola proper } Rotse cluster. (Ci-)BoKO CLUSTER, in the district of S' Paul de Loanda. Angola cluster. IMbamba Mbangala, at Kasanje. Serta.0, at Ambaka. Lower Congo, or Fiote, at, and round, S. Salvador. Lunda, between the Upper Kasai and the Upper Lualaba. GUHA CLUSTER. I _ 'I east of the Upper Lualaba. Rungu, Bamba, Kusu, west of Nyangwe. ,, „^ „„ / Bamba, east of the Lualaba, north of the Lukuga River. NYWEMA CLUSTER. ■( ' , & • Introduction. xxi Chwana cluster. Riia, west of the Middle Lualaba. Luba, on the Lower Kasai and the Lulua River. Yansi cluster, i '^^''^' '""""^ Stanley PooL. I Yansi, spoken by the native traders above Stanley Pool. II. Eua Group. Eastern Half. ■ s Tlhaping, or Chwana proper, \ Rolong, >• in Bechwanakud, and the Transvaal Mangwato, / Suto, in Basutoland, and the Orange Free State. Kololo, on the Zambezi, above the Victoria Falls. -, c Gwamba, south-west of the Lower Limpopo. NYAMBANE I XT u . , J r u I Nyambane, at, and round, Inyambane. \ Chiloane, in, and round, the islands of Chiloane. IKilimane, on the Kwakwa River. Tugulu, in, and round, the island of Mozambique. Gunda (?), on the Lukugu River (?). See n. 97. ,, , 'I inland from Ibo Island. Medo, J ' Masasi, north of the Lower Rovuma. Tshagga CLUSTER. I ^,^ ^^^ ]• near Kilima Niaro. I Gweno J Hinzua, in Hinzua Island. Angazidja (?), in Great Comoro Island. CLUSTER. Comoro cluster. | Western Half. Buma, on the Congo, at,, and round-,' Bolobo. ' Mpongwe f Mpongwe, on the Lower Ogowe and the Gabiin ■CLUSTER. t Shekiaiii, or Bulu, on the River Gabiin. IKele, or Kali, along the Bembo River. Benga, on the islands of Corisco Bay. Dualla- round the Cameroon Mountains. Subu, or Isubu, north of the Dualla. Fan, or Pahnin, on the Upper Ogowe. III. Fernantiian Group. Fernandian (Banapa | , „ ,, ^ \ Banni } in Fernando Po Island. CLUSTER. 11 \\ . lUreka ) '■}{ /J. The length of this list of languages might lead the reader to think that it implies a great diversity between them, something like that existing between the Indo-European languages. This would be a false notion. In general the languages of the same cluster must be considered as mere dialectic varieties, Thisj for instance, is the case with Xosa, Zulu, and Tebele, in the Kafir cluster; with Tlhaping, Rolong, Suto, and Kololo, in the Chwana XXII ^Introduction. cluster, etc. There are even several clusters which might quite appropriately be said to form together a single language. For instance, the differences between the Senna, Gangi, Nika, Shambala, Sagara, and Ibo, clusters cannot be said by any means to be as great as those which may be remarked between several French patois. The greatest noticeable divergencies are found to exist between the Mpongwe cluster and the languages of the main group. These may be said to amount to something like the difference between Latin and French, or between English and German. 11. Bantu Literature. — Sources. i6. Writing is unknown to the Bantu in general. According to my Tonga informants from the Middle Zambezi; God said to the Ma-nkua (the whites) that they must learn tp write, and to the Tonga that they must learn to. speak. The only Bantu known to write are those among the coast tribes which have fallen most under foreign influence. On the west coast Roman characters alone are known. On the east coast the Arabic alphabet has probably long been in use and is still prevailing. Daniel J. Rankin, M. R. A. S., formerly Acting British Consul at Mozambique, says that, even the Makua of the coast of Mozambique, though they have so long been under Portuguese inflluence know how to write only in Arabic characters. " In most of the large villages ", he adds, " the children of the better class receive lessons in reading and writing, the universal and only lesson-book being the Koran. Beginners are taught to read and. write the alphabet and simple sentences on religious subjects by means of a board called " ubau ", formed of a hard kind of wood — answering in its use to the slate of European schools — from which ink-marks can be effaced when desired. This stage passed, a well-thumbed copy of the Koran does duty as a reading-book. The Arabic alphabet having been learned, and pronunciation of the words acquired, the education of the average native ceases. Correspondence is afterwards carried on in Swahili by those who have attained greater proficiency in their studies (') ". 77. We do not know when Bantu thus began to be written on the east coast. No Bantu literature originally writen in Arabic characters has been preserved, except two small poems in Old Swahili, published in Roman characters by Dr. Steere in his collection of Swahili tales (2), and a longer one, left in manuscript by Dr. L. Krapf, and lately published in the Zeit- schrift filr afrikanische Sprachen, 1887. 18. Still less do we possess anything of the period preceding the occupation of Eastern Africa by the Arabs. Not a few remarkable monuments of an ancient civilization have indeed been discovered in the Bantu field south of 1. Arab Tales translated from Swahili into the Tugulu dialed of the Mdkua Language, by Daniel J. Rankin, 1886. ■^. Swahili Tales, by Edwarf Steere, 2"l edition, London, 1889. Introduction. xxiii the Zambezi, but either no inscriptions have been found near them, or, if any have been noticed, there is every appearance that they are not in Bantu. ThuSi if we may rely on a paper of Farini, which was read in 1886 before the Royal Geographical Society, this traveller (?) had then discovered in the Kalahari desert about 23°^ S. lat. by 21°^ E. long, what may have been the work of ancient diamond-diggers, the right place, it seemed, to loc'k for inscriptions, but he found none. " It had evidently been ", he writes, " a huge walled inclosure, elliptical in form, and about the eighth of a mile in length. The masonry was of a Cyclopean character; here and there the gigantic square blocks still stood on each other, and in one instance the middle stone being of a softer nature was weatherworn... In the middle of the ellipse was a kind of pavement of long narrow square blocks neatly fitted together, forming a cross, in the centre of which was what seemed to have been a base for either a pedestal or monument. We unearthed a broken column, a part of which was in a fair state of preserv- ation, the four flat sides being fluted... We sought diligently for inscrip- tions, but we could find none (') ". Several descriptions have also been given by various writers of the ruins of Zimbabye, near the gold-fields of Masho- .naland, but no inscription has ever been mentioned, unless we may consider, as such certain carvings found there by thp traveller Anderson : " There are, " he writes, " several beams inserted in the walls, projecting eight feet, composed of a hard and fine-grained stone of a dark colour. Upon one of them are carvings, diamond-shaped, one within' another, separated by wavy lines... Several old. diggings are in the vicinity". The same writer, after having mentioned a large number of old ruins and forts in the vicinity of various ancient gold-diggings, speaks also of numerous rocks somewhere near the Limpopo " with carvings of animals, snakes, and figures, on them ", which may turn out to be some kind of hieroglyphics. He mentions one circular rock in particular, with " no other stones near it, fifteen feet in diameter, similar to a ball cut in the centre..., covered with carvings... representing paths with trees and fruits on each side ". " Upon one of the trees, " he adds, " is a snake crawling down with a fruit or round ball in its mouth ; near it is a figure, and a little distance off another figure with wings, almost like an iguana, flying towards a man who is running away. His left foot is similar to that of a horse, the right one has two points... the intermediate spaces have many stars. " The writer adds that, though the rock is very hard, some portions of the carvings have been rendered nearly smooth by large animals rubbing against it, from which he concludes that they must be very ancient ('). Mr. O'Neil, formerly consul at Mozambique, writes that he was told by the Capitao-mor of Gorongoza of many ancient inscriptions to be seen in the Manica gold-fields, and that, judging from the description given of them.he thought they were in cuneiform or wedge-shaped X. Proceedings of the R. G. S., i885, p. 447. 4 Anderson's Twenty-fiue Years in a Waggon, Vol. I, pp. 196, 197, Vol. 11, pp. ijo, 201, 202, etc. XXIV Introduction. characters ('). But of course, so long as our knowledge ends there, we must rest satisfied with a " perhaps ", as far as this has anything to do with Bantu. Perhaps on those rocks and ruins we have ancient inscriptions, and, if so, since they are in the Bantu field, perhaps they are couched in Bantu. Probably they are not. What is certain is that no native can give any account of their origin. Neither could the Arabs do so 400 years ago, when they were first met with by Vasco de Gaiitia near the coast of Sofala. Certain drawings were found on rocks near the Congo by Captain Tuckey in 1 8 16, and they have been compared by Mr. de Laborde to similar drawings which are mixed up with the inscriptions of Wadi Mokatteb in Arabia ('). There is even less probability of these being Bantu inscriptions than there is in the case of those mentioned by Anderson and O'Neil. It therefore seems that, waiting further discoveries, the history of Bantu literature must begin with the first Christian Missions to South- Africa. jp. Bantu Literature of the seventeenth Century. — It appears that two catechisms were written in the seventeenth century by Dominican mis- sionaries stationed at Tette on the Zambezi, but they never have been published (3). This cannot be too much regretted. To preach God's Word to the natives of Africa, then to go off without having given it to them in • writing, and yet to expect that these material people and their children will abide permanently by it, is to expect from God's grace as great a miracle as if they were to embrace the faith without anybody preaching it to them. The missionaries of Angola and the Congo did more permanent work, as is well known, and I have little doubt that the result was due in a great measure to the works they published. 20. The first Bantu work ever printed seems to have been a translation into the language of S' Salvador of Father Jorge's treatise on Christian Doctrine. It was made by the priests at the court of Congo with the aid of Fr. Matthaeus Cardozo, S. J., and published at Lisbon in 1624 ("*). 21. In 1642 there was printed at Lisbon a catechism in the language of Angola, written by Father Pacconio, S. J., and abridged by Father de Coucto, S. J. This work has passed through several editions. Father Canne- cattim, writing in 1805, finds it full of defects, such as laconicisms, redundancy and useless circumlocutions, neglect of the grammatical rules laid down at the end of its Roman edition, etc. But H^li Chatelain, author of two Angola Grammars, justly remarks that Cannecattim's criticisms are not only excess- ive, but unjust {-■). It may be added in particular that the rules laid down at the end of the Roman edition are not Father de Coucto's, but of the Ca- puchin editor,, and that the greatest defect of the work might have been its agreement with those rules, as they are more artificial than correct. Indeed 1. Proceedings of the R. G. /., 1885, p. 443. 2. Voyage de I Arab'ie PiMe, par L. de Laborde et Linant, Paris, 1830, p. 71, and Illustrations. 3. Atudes religieuses, philosophiques, historiqties et Uttiraires, 1878, p. 797. 4. Bentley's Dictionary and Grammar of the Kongo Language, p. XI, 5. Grammatica elementar do Kimbundu, p. xv. Introduction. xxv as far as I am able to judge, Father de Coucto's catechism is still now one of the best Bantu works we possess. I have made use of it constantly in writing this work. 22. In 1650 the Capuchin Father Hyacinth Busciotto de Vetralla publish- ed in Rome a vocabulary in four columns, Congo, Portuguese, Latin, and Italian. I have not seen this work. In 1659 the Propaganda at Rome published a Congo Grammar of the same author, entitled " Regulae quaedam pro difficillimi Congensium idio- ■matis faciliori captu ad Grammaticae normam redactae. " This is a good work,' and one which shows much insight into the language. It has been lately translated into English by Mr. H. Grattan Guinness, of the Livingstone Congo Mission. J J. In 1697 Father Pedro Dias, S. J., published at Lisbon an Angola Grammar entitled " ArU da lingua de Angola ". According to Hdi Chate- lain the author of this little work shows that he understood well the mechanism of the language with which he dealt ('). I have found in it several precious observations which I have noticed nowhere else. The first series of publications in and on the languages of South-Africa seems to have come to an end with this book, unless we add to it an abridged grammar of the language of Kakongo, which forms the 19* chapter 6f a History of Loango published. in 1776 if). About this time a very good French-Congo Dictionary was ready for the press. Unfortunately it is still in manuscript, waiting in the British Museum for publication (3). Its coun- terpart, the Congo-French Dictionary, has been discovered at Rome by P^re Duparquet, of the Congregation of the Holy Ghost (■*). 2^. The revival of Bantu Literature. — Bantu studies were finally re- sumed at the beginning of this century by the Capuchin Father Bernardo Maria de Cannecattim. He published at Lisbon in 1804 an Angola Dic- tionary, and in 1805 an Angola Grammar. He undoubtedly must be praised for his initiative, but his works cannot be said te be as valuable as the preceding. His Dictionary is one of those dry collections of words without a single example to establish the proper value of any one of them. His Grammar is retrograde as compared with the little work of Father Pedro Dias, which he does not seem to have known. Since then Bantu literature has been steadily increasing in the number of its volumes until such publications have become matters of frequent occurrence. 25. The most famous is Bleek's Comparative Grammar of South-African Languages. This work was ' intended to reveal to the scientific world the extent, as well as the proper, features, of the great Bantu family of languages, and at the same time to determine its relation to the Hottentot- i. Grammatica eleme?itar do Kimbundu, p. xvi. a. Histoin de Loango, par M. I'abb^ Proyart, Paris, 1776. 3. Add. Mss. 33, 779, Grenville Library. 4. Missions CathoUques, 1886, p. 400. XXVI Introduction. Bushman family, and, perhaps, to other families as well. It was to be published in four parts. The first appeared in 1862. It contains a classifi- cation of the South- African languages best known at the time, followed by a study of their phonetics. The first section of the second part was published in 1869. It is a very careful comparative study of the prefixes and suffixes of substantives both in Bantu and Hottentot. Unfortunately Bleek died before he could carry his work any further than this first section. His piremature loss will ever be a matter of regret to the scientific world. 26. The other treasures of Bantu literature down to 1883 have been des- cribed at length in Dr. Cult's classical " Sketch of the Modern Languages of Africa ". This is not the place to do the same work over again. It is simply astonishing that Dr. Cust was so successful in picking up the vast amount of information on Bantu languages and their literature which he has embo- died in his work. I cannot say I have got at all the available sources mentioned by him for the study of these languages. I think, however, I have perused nearly all those which were to my purpose. The want of the others, if want it be, is compensated, at least in part, by the information I have obtained directly fromnatives of various parts of Africa, by the possession of several works which have appeared since 1883, and by the perusal of certain MSS. of Livingstone and other travellers which are in the Grey Library in Capetown. In mentioning the materials which I have thus had at my disposal, I refer the reader for further information to Dr. Cust's work. 27. Sources for the Kafir cluster. — See Cust, pp. 301 (Xosa) and 299 (Zulu). 1. Doehne's Zulu-Kafir Dictionary, Capetown, 1857. 2. Davies's Kafir Diftionary, Xosa and Zulu, London, 1872. 3. Callaway's Nursery Tales of the Zulus, Natal, i858. 4. ,, Religious system of the Zulus, '^zX'A, 1868. 5. Appleyard's The Kafir Ldtiguage,Y^t\g^''i\!C\zxcC%ToyiVi, 1850. 6. Grout's Grammar of the Zulu Languagej^ii^X^, i86i, etc., etc. Kafir is the Bantu language I know best, having learnt it through five years' intercourse with the Xosa-Kafirs, during which purposely I never spoke to them but in their language. Most of the Kafir sentences given in this work are taken from tales which I wrote under their dictation, or which they wrote for me. One of these was published in 1886 in my " Outline of a Xosa-Kafir Crammar. " Four others are appended to this work as specimens of the traditional literature of these people. 2S. Sources for the Karanga cluster, — See Cust, pp. 307 (Kalaka), 310 (Yeye), and 307 (Shona). When I had learned Tonga from the three Zambezi boys whom I shall mention hereafter, one of them gave me the Karanga translation of most of what I had written in Tonga. He was a very intelligent native, about thirty years of age, belonging to the family of Wange, whom he made out to be the direct representative of the old house of Monomotapa, and about whom Introduction. xxvii more may be seen in the second section of the first appendix to this work. He therefore belonged to those Karanga who crossed to the north of the Zambezi, when driven by Mzilikazi out of what is now Matabeleland. His native name was Siacibi. I do not know that anything has ever been publish- ed on the dialects of the important Karanga cluster. There is a Vocabu- lary of Yeye in Livingstone's Vocabulary (MS.) to be mentioned hereafter. 2g. Sources for the Tonga cluster. — See Cust, pp. 322 i^oV&,alias Tonga), 325 (Bisa), 329 (Tonga), and 364 (Bemba).. This again is an important cluster on which nothing worth notice has yet been published. I take Tonga as the standard language throughout this work ('). I learned what I know of it in 1884 from three natives who had come down to the Cape Colony from the Interior in the company of Fathers J)epelchin and Croonenberghs, S. J. One of the three was the Karanga named Siacibi mentioned just above. He pretended to speak pure Tonga like the other two, saying that all the subjects of Wange have learned to speak this language since they crossed the Zambezi, though they all knovv Karanga also. As I told him I had heard that they had ■adopted the Kololo language, he said that this was quite false, and that not a single subject of Wange knew Kololo, adding that this language was very difficult to learn, while Tonga was easy. Whenever he gave me any inform- ation in Tonga, I got his two companions to repeat what he had said, in order to make sure of the correctness of his idiom. The second of these " boys ", as they are called in South- Africa, belonged to the Lea tribe (alias Ba-lea, Ma-leya, etc.), dwelling below the Victoria Falls. His own native language was Lea, which is a Tonga dialect, but he was quite used to speak pure Tonga, according to the standard received on the Middle Zambezi. His pronunciation was somewhat indistinct. The third of the three, whom we only knew by the name of Joe, was one of the independent Tonga who recognise Monze as their paramount rain-maker {^). His immediate chief was the well-known Sinamane, on the Zambezi River. His pronunciation was wonderfully clear and distinct. Unfortunately he was too young to give I. I believe it will readily be seen by those who will peruse this work that the Tonga language of the Middle Zambezi represents well the proper features of the larger number of the Bantu languages. As the name of Ba-tonga is common to several South-African tribes, it may be as well for me to state here what r think of them. I consider the Tonga of the Middle Zambezi, who have no other name than this, to be the purest representative of the original Bantu. They alone, it seems, have never been tributary to any empire : they say thatrthey have never had any but independent chieftains, or patriarchs, who may recognise a paramount rain-maker, but no king in the proper sense of this word. Neither slavery, nor anything like higher and lower class, is known amongst them, they all are the " children " of the chiefs. Then, well protected in their peninsula by the Kafuefue on one side and ■ the Zambezi on the other, they may easily have guarded themselves against invaders, as they do in our own days. The .other tribes known by the name of Tonga in other parts of South-Africa I should equally consider to represent the aborigines with respect to their neighbours, or to the upper classes intermixed with them. They are all peaceful agricultural tribes. Such are, for instance, the Tonga of Sofala and the Lower Limpopo, also called Ma- Gwamba or Ma-kmafa " people of the place ", and Ma-hlengwe, or, as the Xosa-Kafirs pronounce this word, a Ma-mfen^u; the Tonga of Lake Nyassa, also called Wa-kamanga; the Tonga or Tanga of the Katanga ; the Tonga or Tanga of the Gabiin, also called Naka, etc. .i. See the third section of the first Appendix to this work. XXVIII Introduction. much information, being at the time only thirteen or fourteen years of age. Some specimens of the kind of information I obtained from these natives are appended to this work (Appendix I.) Livingstone has written a great deal about the Tonga in his " Missionary Travels ". He writes their name Ba-toka according to Chwana pronunciation, instead oiBa-tonga. In January 1885 I was kindly allowed to copy in the Grey Library in Capetown a MS. of Livingstone which contains a Tonga vocabulary. It is entitled " A Comparative Vocabulary of the Languages of the Ba- khoha or Ba-yeye, Ba-shubea (=?= Ba-subia), Ba-lojazi, Ba-ponda ( = Ma- mbunda), Ba-rotse, Ba-toka (= Ba-tonga), Ba-nyenko, Be-chuana, and English. " Too many words in this MS. remind one of the Chwana scholar, but with this exception it is sufficiently reliable. I have no other source than this MS. for Subia. For Bisa and Bemba there are short collections in Last's precious Polygldtta Africana Orientalis, a work to be often referred to hereafter. Another collection of Bisa words is found in Stanley's Comparative Vocabulary at the end of " Through the Dark Continent ". With regard to the Tonga dialect of Lake Nyassa, see n.65. jd. Sources for the Senna cluster. — See Cust, pp. 307 (Zizulu = Tette) and 323 (Nyai = Tette and Zumbo), 326 (Ravi = Nyassa), 330 (Nganga = Nyassa), and 331 (Sena). In 1885 a native of Kilimane, by name Justino, whom I met in the Cape Colony, wrote out for me vocabularies, dialogues, fables, and a short history of the life and passion of Our Lord, in Senna and Portuguese. I have mostly iuade use of these MSS., all well written and perfectly consistent. My other sources are : — 1. MSS. kindly lent to me by Father Ronchi, S. J., containlng-vocabularies, fables, etc, 2. Elementos de Grammatica Tetense, pelo R. P. Victor Jose Courtois, S. J., Mo9ambique, 1889. 3. A Grammar ofthf Chinyanja Language as spoken at Lake Nyassa..., by Alexander Riddel, of the Livingstonia Mission, Edinburgh, 1880. 4. Dictionary of the Kiniassa Language, by the Rev. John Rebman, Basle, 1877. 5. Specimens of Gindo in Dr. Steere's Short specimens of three African Languages, 1869, and in Last's Polygl., pp. 90-92. 6. Bleek's Languages of Mozambique, London, 1856. 7. The Tette Language, MS. in the Grey Library, Capetown, attributed to Livingstone. 8. The Senna, Tette, ancl Maravi, Languages, MS. attributed to Rebmann, kindly lent to me by the late Father Weld, S. J. J/. Sources for the Viti cluster. — See Cust, p. 301 (Ngoni). 1. A few words here and there in Montagu Kerr's Far Interior. 2. Last's Polyglotta Afr. Or., pp. 139-141 (Bunga). 3. Stanley's Viti or Tuta Vocabulary at the end of the Dark Continent. 32. Sources for the Gangi cluster. — See Cust, pp. 343 (Henge), 362 (Bena), 363 (Sango). I. Last's Polygl. Afr. Or., pp. 93-96 (Gangi), 105-108 (Ziraha), 117-120 (Kwenyi), 109-112 (Nkwifiya), 113-116 (Ndunda), 121-123 (Bena), 124-127 and 225-226 (Sango), 231 (Kimbu), 157- 159 (Nya-turu). jj. Sources for the Ungu cluster. — Introduction. xxix 1. Stanley's Voc. in \h^ Dark Continent, (Fipa, Rungu (?)). 2. Last's Polygl. Afr. Or., p. 128-130 (Ungu). J/. Sources for the Sagara cluster. — See Cust, p. 352 (Sagara), p. 362 (Hehe), p. 36s (Gogo). 1. Stanley's Voc. in the Darh Continent, (Sagara,- Gogo). 2. Last's Grammar of the Kaguru Language, London, 1886. See note to n. 77. 3. Last's Polygl. Afr. Or., pp. 57-60, 221-222, and 233 (Kaguru), 61-64 (Ituniba), 65-68 (Kondoa), 69-72 (Kami), 73-74 (Khutu), 97-100 and 223-224 (Gogo), 101-104 and 227 (Hehe). 35. Sources for the Nyamwezi cluster. — See Cust, pp. 365 (Nyamwezi), 367 (Tusi), and 373 (Sukuma). 1. Stanley's Sukuma Voc. in the Dark Continent. 2. Dr. Steere's Collections for a Grammar of the Nyamwetii Language, London (no date). 3. Last's Polygl. Afr. Or., pp. 146-149 (Sukuma), 150-153 (Sumbwa), and 154-156 (Tusi, or Ha). j6. Source for Regga. — See Cust, p. 377 (Regga). last's Polygl., pp. 203-212. ^y. Sources for the Ganda cluster.. — See Cust, pp. 374 (Ganda), and 373 (Zongora = Nyambu). I have mostly availed myself of the excellent " Essai de Grammaire Ruganda, par un P^re de la Soci^te des Missions d'Afrique, Paris, 1885. " My other sources are : — 1. Katekis'mu Ruganda, Alger, 1887. 2. St Matthew's Gospel in G^aKi/a, British and Foreign Bible Society, 1888. 3. An Outline Grammar of the Luganda Language, by Rev. C. T. Wilson, M. A., F. R. G. S., C. M. S. Missionary to Uganda, London, 1882. 4. Stanley's Voc. in the Dark Continent (Ganda, Nyanibu). 5. Last's Pol. Afr. Or., pp. I7'3-I75 (Ganda), and"i6o-r63 (Nyambu). jS. Sources for the Taita cluster. See Cust, pp, 350 (Teita), 357 (Taveta), and 354 (Pare). 1. A pocket-Vocabulary of the Ki-swahili, Ki-nyika, Ki-taita, and Ki:kamba Languages, compiled by ^. Downes Shaw, C. M. S. Missionary in East-Africa, London, 1,885. 2. Wdrterverzeichnis aus dem Kidschagga und Pare, in the Zeitschrift fiir afrikanische Sprachen, - 1887-1888, pp. 72-76. 3. Ki-taveita Vocabulary inH. K.Johastoa's iTilimanjaro £xpediiion,'London, 1886, pp. 521-540. jp. Sources for the Nika cluster. — See Cust, p. 355 (Nyika or Nika). I. Downes Shaw's Pocket Dictionary, just mentioned. , 2. A Mika-Eiiglish Dictionary, compiled by the late Rev. Dr. L. Krapf and the late Rev. J. Rebmann, edited by the Rev.T. H. Sparshott, S. P. C. K., 1887. ^o. Sources for Pokomo. — See Cust, p. 359. 1. Zur Grammatik des Ki-pokomo, in the Zeitschrift f. a. S., 1888-89, PP' 161-189. 2. Kipokomo Wdrterverzeichnis, von Missionar Ferd, Wiirtz, Ibid. 1889-90, pp. 81-105. 4.1. Sources for Kamba. — See Cust, p. 3 5 9. 1. Last's Polygl. Afr. Or., pp. 53-56 and 217-218. 2. Last's Grammar of the Kamba Language, London, 1 885. 3. Shaw's Pocket Vocabulary already mentioned. 4. Krapf's Deutch Ki-kamba Worierbuch inOxs.Zeitschriftf. a. S,., 1887-88, pp. 81-123, 42. Sources for the Swahili cluster, — See Cust, p. 345. Swahili I have studied mostly from Dr. Steere's " Swahili Tales as told by Natives of Zanzibar, 2^ ed., London, 1889, '' and the " Arab Tales, trans- lated from ...Swahili... into the Tugulu dialect of the Makua Language, by XXX Introduction. Daniel J. Rankin, M. R. A. S., ex-Acting British Consul at Mozambique, London, 1886." My other sources are the three following remarkable works: 1. Krapf's Dictionary of the Swzhili Language (LonAon, 1882), which, with its copious examples intended to bring out the proper meaning of the words, is a good specimen of what every Bantu Dictionary should be. 2. A Handbook of the Swahili Language as spoken at Zanzibar, by the late Edward Steere, LL. p.. Missionary Bishop for Central Africa, 3* edition,... by A. C, Madan, M. A., London, 1885. 3. Graminaire Kisuakili, 'pa.i \e P4re Delaunay, de la Societe des Missionnaires de N.-D. des Missions d'Afrique, Paris^ 1885. ^j. Sources for the Shamiala cluster. — See Cust, pp. 351 (Zeguha and Nguru), and 353 (Boondei and Shambala). 1. Dr. Stsete's Collections for a Handbook of the Shambala Language, 1867. 2. Last's Pblygl. Afr. Or., pp. 41-44 (Shambala), 4952 and 213-214 (Zeguha=Zegula), 45-48 and 215-216 (Nguru), and 37-40 (Boondei). 3. Collections for a Handbook of the fioondei Language, by Rev. H. W. Woodward, of the Uni- versities' Mission to Central Africa, S. P. C. K., 1882. /f4. Sources for the Ibo cluster. — 1. Ibo Vocabulary in Rankin's Arab Tales, mentioned above, pp. 43-46. 2. Last's Polygl. Afr. Orl, pp. 33-36 (Lima). ^J. Source for Zaramo. — See Cust, p. 344. I. Dr. ?ilseie^s Short specimens of three... African Languages, London, 1869. if6. Sources for Konde. — See Cust, pp. 341 (Konde), and 343 (Donde). 1. Last's Polygl. Afr. Or., pp. 77-80. 2. Konde Vocabulary in Rankin's Arab Tales mentioned above, pp. 43-46. 4.7. Sources for Yao. — bee Cust, p. 334. I. Introductory Handbook of the Yao Language, by the Rev. Alexander Hetherwick, M. A., F. R. G. S., S. P. C. K. 1889. .2. Dr. Steere's Collections for a Handbook of the Yao Language, S. P. C. K., 187 1. 3. l.ast'&.Polygl. Afr. Or., p. 87-89. 4.8. Sources for the Herero cluster. — See Cust, pp. 309 (Herero), and 311 (Ndoriga). 1. An English- Herero Dictionary, by the Rev. F. W. Kolbe, Capetown, 1883. 2. Dr. BUnst'sSprachfiihrerfiirReisende in Damaralani, and Mdrchen der Ova-herero in the Zeitschriftf. a. S., 1887-88, pp. 252-294, 189-216, and 295-307. 3. Bleek's note on Sindonga in his Comparative Grammar (212-216). ■ ^. Lojazi Vocabulary ia Livingstone's Comparative Voc. MS. mentioned above. ^.g. Sources for the Benguela cluster. — See Cust, p. 390 (Nano). 1. Bleek's note on Nano in his Comparative Gr., pp. 216-220. 2. Pangela Vocabulary in Koelle's Polyglotta Africana, London, 1854, 3. Stover's Obsiervafions on the Grammatical structure of the Umbundu Language, Boston, 1885. 4. Sander's Vocabulary of the Umbundu Language, Boston, 1885. 50, Source for Kwango, or Mbunda proper. — See Cust, p. 390 (Ponda or Mbunda). Mbunda Vocabulary in Livingstone's Comparative Voc. MS. mentioned above. //. Sources for the Rotse cluster. — See Cust, p. 389 (Luina). 1. Barotse Lahgua«e translated into the Sichuana, MS. in the Grey Library, Capetown, attri- bute^ to Livingstone. 2. Holse Vocabulary in Livingstone's Comp. Voc. MS. mentioned above. 3. Nyengo Vocabulary in the same MS. Introduction. xxxi 52. Sources for the (Ci-)boko cluster. — See Cust, p. 397 (Kioko), and p. 399 (Yakka). Quioco Vocabulary in Capello and Ivens' From Benguella to the Territory of Yacca, London, 1882, pp. 327-330. 53. Sources for the Angola cluster. — See Gust, p. 393 (Bunda = Angola). 1. Arte da lingua de Angola, pelo P. Pedro Dias, S. J. Lisboa, 1697, supra, n. sj. 2. Father de Coucto's Catechism, 1661, supra, n. 21. 3. Helx Chatelain's Grammaiica elementar do jCintliundu, Genebra, 1888-89. Do. Die GrundzUge des Kimbundu, in the Zeitschrift f. a. S,, 1889-90. Do. Sammlung von Mbamba und Mbangala WSrlern, ibid. 1889. 4. N-biinda Vocabulary in Capello and Ivin's' From Benguela..., pp. 304-325. 5. CoUefao de Observances graiiiinaticaes sobre a lingua Bunda, por Fr. Bernardo Maria de Canne- cattira, Capuchino..., Lisboa, 1805. 6. Cannecattim's Diccionario da lingua Bunda on Angolense, Lisboa, 1804 . 7. Kasands Vocabulary in K.oelle's Polygl. Afr., London, 1854. 5f. Sources for Lower Congo. See Cust, p. 405. 1. MS. French-Congo Dictionary, 1772, British Museum. 2. Bentley's Dictionary and Grammar on the Kongo Language, Baptist Miss; Soc, 1887. ■^.Grammaire Fiote, par le Rev. P. Alexandre , Visseq, de la Congregation du Saint-Esprit, Pi^ris, 1889,, 4. Regulae quaedam... pro... Congensium idioniatis... captu, a P. Hyacintho Brusciotto a Vetralla, Concionatore Capucino, Romae, 1659, supra n. 22. 55. Sources for Lunda. — See Cust, p. 399. 1. Ruunda Vocabulary in Koelle's -Polyglotta Africana. 2. Lunda Vocabulary in Capello and Ivens From Benguela..., pp. 329-331. 3., Carvalho's Methodo pratico para fallar a lingua da Lunda, Lisboa, 1890. See n. 788'''^ 56. Sources for the Guka cluster. — See Cust, pp.371 (Guha), and 363 (Rungu). 1. Last's Polygl. Afr. Or., pp. 170-172 (Guha). 2. Stanley's Comparative Voc. in the Dark Continent (Guhha, and Rungu (?) ). 57. Sources for the Nyi&ema cluster. — See Cust, p. 372 (Nywema, and Kiisu). I. Last's Polygl. Afr. Or., pp. 183-187 and 232-233. 58. Sources for Rua.. — See Cust, p. 371. 1. Cameron's Kirua Vocabulary va. Across Africa, London, 1877. 2. Last's Polygl. Afr. Or., pp. 167-169. 5p. Sources for Luba. — See Cust, p. 400. Dr. Buttner's Zur Grammatik der Balubasprache in the Zeitschrift f. a. S., 1888-89, PP- 220-233. 60. Source for the Yansi cluster. — See Cust, pp. 409 (Teke) and 410 (Yanzi). See hn. 159-162. 1 61. Sources for the Chwotna cluster. — See Cust, p. 305 . v In 1885 I collected some materials for the study of Chwana with the help of a native of the Ba-kwena tribe from Pretoria, and a Mo-suto subject of the late Moshesh. Bat in writing this work I have not made so much use of these as of the " Notes towards a Secoana , Grammar, collected by the Rev. William Crisp, Canon and Chancellor of Bloemfontein Cathedral xxxii Introductwn. (2^ edition, London, 1886)," and of the Chwana Catechism of Father Tem- ming, S. J. My other sources are : — 1. An English and Secwana Vocabulary, by the Rev. John Brown, London, 1876. 2. The Chwana New Testament. London, i888.' 3. Hyffins in Chwana, by Father Temming, S. J., Marianhill, 1887. 62. Sources for the Nyambane cluster. — See Cust, pp. 302 (Gwamba), 303 (Hlengoe), 303 (Nyambane), and 308 (Siga = Nyambane). 1. Bleek's Languages of Moiambiqui (Lourenzo Marques , Inhambane), London, 1856. 2. Koelle's Polyglotta Africana (Nyamban =, Nyambane). 3. Lemons de Shigwamba, plar le Missionnaire P, Berthoud, Lausanne, 1883. 63.' Sources for the Mozambique cluster. — See Cust, pp.333 (Roro = Gunda (?) ), 333 (Kua). 1. Rankin's .i4roi Tales, mentioned above. 2. Chauncy Maples' Collections for a Handbook of the Makua Language as spoken at Mascui, London, 1879. 3. Elementos para un Vocabulario do diakcto falado em Quelimane, por Gustavo de Bivar Pinto Lopes, Mo9ambique, 1889. 4. English- Tshigunda Vocabulary (no title page). 5. BleeWs Languages of Mozambique (Quellimane, Mozambique). 6. Koelle's Polyglotta Africana (MetoJ Kiriman, Matataii). 7. las.'Ci Polygl. Afr. 0;-.,pp. 81-83. (Lomwe), 84-86 (Mozambique ). <5/. Sources for the Comoro cluster. — See Cust, p. 339. 1. Last's Polygl. Afr. Or., pp. 179-1S2. (Anzuani, or Hinzua). 2. Bleek's Languages of Mozambique (Anjoane). 3. Dr. Steere's Short specimens of three African Languages (Angazid}a ). 65. Sources for the Tshagg'a cluster. — See Cust, p. 357 (Chagga). 1. Worlerverzeichnis aus dem Kidshagga und Pare, in the Zeitschrift. f a. S., 1887-88, pp. 72-76. 2. H. H. Johnston's Chagga and Gwmo Vocabulary in The Kilimanjaro Expedition. 66. Sources for Buma. — See Cust, p. 409. p. H. Johnston's Voc. in The River Congo, 446-463. 6^. Sources for the Mpongwe cluster. — See Cust, pp. 417 (Pongwe), and 420 (Shekiani). 1. Dictionnaire Fran^ais-Fongoue, par les misSionnairgs de la Congregation du S'-Esprit, Paris, 1877. 2. Dictionnaire Pongoue-Fran^ais, par le R. P. Gachon, de la Congregation du S'-Esprit, Paris, 1881. 3. Grammaire de la Langue Pons^ouie, par le R. P. Le Berre, de la Congregation du S'-Esprit, Paris, 1873. 4. Mpongwe Gospels, by American Missionaries at the Gaboon, 3"* ed.. New York, 1879. 68, Sources for the Dualla cluster. — See Cust, p. 426 (Dualla), 428 (Isubu), 420 (Benga), 415 (Kele). 1. Saker's Grammatical elements of the Dualla Language (incomplete), with Vocabulary and MSB. (in the British Museum), 1863. 2. C. Meinhof's Ein Mdrchen aus Kamerun in the Zeitschrift f. a. S., 1889-90, pp. 241-246. Do, Das Zeitwort in der Duallasprache, ibid., 1888-89, pp. 1-34. Do. Benga und Dualla, ibid., pp. 190-208. Do, Das Verbum in der Isubu-Sprache, ibid., 1889-90, pp. 206-234. Do. Das Zeitwort in der Benga-Sprache, ibid., pp. 265284. 3. Bleek's Notes on Dikele, Benga, Dualla, and Isubu, in the Compar. Gr., pp. 231-240. Introduction. ' , xxxiii dp. Source for Fan. — See Cust, p. 422. Vocabulary of the Fan Language, by Senor Don Amado Osorio Zabala, S. P. C. K. 1887. "JO. Sources for the Fernanctidn cluster. —^ See Cust, p. 426 (Ediya). 1. Bleek's Note on Fernandian in the Cofnpar. Gr., pp. 248-251. 2. OsCar Baumann's Beitrdge zur KenMnis der Bube-Sprache auf Fernando Poo, and Vocabular des Banapa- (Sta Isabel) Dialektes... von Padre Don Jose MaTtinez y Sanz, S. J., in the Zeitschrift / a. .$'.,1887-88, pp. 138-ISS, It heed scarcely be said' t)iat the materials thus placed at my disposal are more' or less reliable. In this work my conclusions are generally drawn only from those which I thought could best be trusted. III. The Origin of the Bantu. 7/. Before we begin to form a comprehensive view of the various Bantu languages, and their general and proper features, it may be good to put together a certain number of data regarding the origin of the various tribes that speak them. The sciences of ethnology and philology have so many points of contact that they must, as it were, go hand in hand. In a subject like this, in particular, the conclusions to which philology seems 1;o lead may be right or wrong. It is therefore important to see what foundation history gives to them. A special reason for giving here some of the histo- ricaL and ethnographical data which I have come across regarding the Bantu is that, if we may judge from various current and upfounded theories, they scarcely seem to be known to exist. , 'J2. First origin. — The most probable account of the first origin of the Bantu seems to be the one found in Mas'oudi's " Golden Meadows ", a work written A. D. 943. Mas'oudi had crossed several times from Arabia to the east coast of Africa ('), and thus had been able to collect accurate information on the Bantu, or the Zindj, as he with the other Arab writers calls them. This is what he says : . — / " When Noah's posterity began to spread itself over the earth, the chil- dren of Kush, the son of Kanaan (Cham), followed a westerly direction and crossed the Nile. There they formed two groups. Some of them, the Nubians, the Bedjah, and the Zindj, turned to the right, between east and west ; the others, in great numbers, went westward in the direction of Zagawah, Kanem, Markah, Ghanah, and other parts of the land of the Blacks and the Dendemeh. Those who had taken the right, going between east and west, soon separated again, thus forming several tribes of the Zindj, such as the Makir {alias Mex, Meska), the Maskar {alias ^iktar, iMeshku, Mashku, Saka, Seka), the Marira, and others (="). " A little further in the same work (?), Mas'oudi adds the following details : — 1. Ma90udi, •' Les jPmiries d' Or ". Texte et traduction par Barbier de Meynard et Pavet de Cour- teille. Paris, 1861-1877, vol, I, p. 233. 2. ibid., vol. Ill, p. 2. 1 3. /bid., p. s.. , XXXIV Introduction, ■/: " As we have said above, the Zindj with other Abyssinian tribes spread themselves to the right of the Nile, down to the extremity of the sea of Abyssinia. Of all the Abyssinian trjbes the Zindj were the only ones who crossed the canal which comes out of the Upper Nile (Juba River ?). They established themselves in this country and spread themselves as far as Sofala, which is on the sea of the Zindj the furthest limit whither ships sail from Oman and Siraf. For, as the Chinese sea ends at the land of Sila (Japan or Corea), so the limits of the sea of the Zindj are near the land of Sofala and that of the Wakwak (Hottentots and Bushmen), a country which yields gold in abundarjce with other marvels. There the Zindj built their chief-town. Then they elected a king whom they called Falime (or Wa- fatime) (^). This has been at all times the name of their paramount chief. .... The Falime has in his dependency all the other Zindjan kings, and commands 300,000 mounted men. The Zindj use the ox as their beast of burden {f) ; for theiir' country has neither horses, nor mules, nor camels ; they do not even know these beasts. There are among them tribes which have very sharp teeth (3) and are cannibals (*). The territory of the Zindj begins at the canal derived from the Upper Nile, and extends to the land of Sofala and that of the Wakwak . " "jj. These are interesting assertions in the light of modern discoveries. A great empire in South-Africa with itq chief-town in the land of Sofala, — nothing could tally better with the descriptions given of South-Africa by the latest explorers. For this country is now found to possess remarkable remnants of an ancient civilization.. Let us remark here that the land formerly called Sofala by the Arabs was not limited to the coast which has retained this name, but comprised all that part of South- Africa which lies between the Limpopo and the River Rovuma. Even in the times of the Portuguese Livius, de Barros; Sofala, or Cefala, as he spells it, was a synonym for " the empire of the Monomotapa " (5). This therefore is the land where we must most expect to find the first seat of the Zindjan Empire. What was more exactly its situation? If we believe Abulfedaand Edrisi, I. I Have explained in the Grammar, nn. 365!=) and 344, that Mas'oudi's word Falime, plural of dignity VVa-falime, — which may also be read Falimo, Wa-falimo, etc. — must be identified with tlie Swahili M-falme, or M-faluine, " a king ", pi. Wa-falme, and with the Mozambique Ma-limii " a chief ", lit. " a man of learning (?) ". It seems that the original pronunciation of this word must have been mf'a limo = m-fo u-a limo, " 3. man of liMo ", whatever the exact meaning of limo may be. Certaiii it is that the Bantu stem which is pronounced -zimo in the main group, a.ad-limo, oi-dimo, or -rimo, in the Kua group, forms immediately the word ^//-w«o (Chwana /(a-/(?»o, ox Ba-dimo) "the ppirits of the kings of old, " " the departed chiefs ". See n. 365 (6). As to the word m-fo, it is often heard in Kafir, and means " an elderly man ". a. So the Kafirs only a few years ago still liad their pack-oxen. Now they have horses. 3. See the note to n. so. See also the Proceedings of the £. G. S., 188;, p. 775, and Bateman's Firsf ascent of Ihe Kasai, ■^. ifi. -. 4. The Nywema are not the only Bantu tribe accused of cannibalism. TJie Yao themselves, east of Lake Nyassa, sometimes indulge in feasts on human flesh (Proceedings of the R. G. S., 1887, p.' 468). Ancient traditions say that Senna itself was a mart for human flesh before the advent of the Portuguese. 5. Oa Asia, dec. I, lib. X, <^. I .; quoted in the Ji/utics lieligieuses, 1878, vol. I, p. 388. Introduction. xxxv in their time (before the 14* century) the chief-town of Sofala was Siyuna, which 1 think must be identified with the chief-town of Ma-shona-land, or the country of Senna ('). Is not the natural inference from this that Ma- s'oudi's seat of the first kings of the Zindj was somewhere in Mashonaland ? If the answer must be affirniative, the ruins of Zimbabye, or Zimbaze, which, discovered by Mauch a few years ago, have lately amazed the pioneers of the Chartered Company, seem to point out the exact spot for which we are looking. The descriptions given of these ruins well corroborate this conclusion. Their features which most struck the Correspondent of the Times (^) a^q : — 1° A series of circular walls within one another, the outermost of which is 4 feet high, and may be over 500 yards in diameter ; 2° One of the inner walls "from 30 ft. to 35 ft. high, 80 yards in dia- meter, about 10 ft. in thickness at the base, and tapering to about 7 ft. or 8 /ft, at the top, built of small granite blocks, about twice the Size of an or(3inary brick, beautifully hewn and dressed, laid in perfectly even courses, and put together without the use of a single atom of either mortar or cement " ; 30 On the eastern side of this enclosure, a narrow entrance, and plose to it, at a place where the wall is 30 ft. high, " a conical shaped tower, or turret, 35 ft. in height and 18 ft. in diameter at the base, built of the same granite blocks, and consisting of solid masonry " ; 4° " On the south-east front of the wall and 20 ft. froni its base a double zigzag scroll, one third of the distance round, composed of the samesized granite blocks placed in diagonal positions ". According to the newspapers, indeed, an expert sent to study these ruins is- inclined to think them to be of Phoenician origin. But, from the descrip- tion given, I rather suspect that they are, on the whole, of purely native, or Zindjan, origin. In our own days the Gcaleka Kafirs, with whom I lived some time, never, when they can afford it, build for their cattle any but round stone kraals, which, though they cannot be compared with the ruins of Zimbabye, seem to belong essentially to the same style of building ; and with many Bantu tribes zigzag-shaped drawings are the usual pattern for all kinds of attempt at anything, like artistic (Jesigns. Finally, another good reason for identifying Zimbabye with both the Siyuna of Abulfeda and the seat of the first kings of the Zindj is, that the actual-occupiers of the country round it, variously called Zindja, Ba-nyai, Ma-shona, etc., are properly part of the Karanga, who certainly have been for centuries the paramount tribe of the vast empire of the Monomotapa. 7^. If, however, it were replied that, notwithstanding these evidences, Zimbabye rnay yet be found to have been the work of foreign gold-seekers, 1. See note to p. 25 in this work. < n. The Times ot Oct.. 7, i8go. Cf. Anderson's Twenty-five Years in a Waggon, Londpn, 1887, vol, II, p. 202. ' XXXVI Introduction. and that the first chief-town of the Zindj must be sought for not in Mashor naland, but somewhere near the Victoria' Falls, I should not deny a certain probability to this opinion. It would readily explain why they are considered by the natives as being not only God's abode, but also the town of the ancient kings (munzi ua Leza, munzi ua Mizimo). See Appendix I. 75". But whatever may be thought of this question, I see no reason to doubt of Mas'oudi's trustworthiness when exposing, the traditions he had either picked up on the spot, or found in previous Hebrew, Christian, or Mohammedan writers, regarding the first origin of the Zindj. His veracity seems to be warranted by his exactitude in details of minor importance, such as the filed teeth and the cannibalism of certain tribes, the Bantu name of the king, the use of ^ack-oxen, the want of horses and camels, the gold- mines of the country, the exact extent of the Bantu field on the east coast, the location of the Hottentots south of the Zindj, etc. etc. It may be added that Mas'oudi agrees with, universal tradition, and with the most ancient Egyptian inscriptions, in considering the Blacks as children of Kush. He is mistaken only in calling Kush the son of Kanaan. Writing of the language of the Zindj, Mas'oudi says that " they express themselves with elegance, and are not wanting in orators," ('). This is a.nother evidence of his veracity. yd. There is in Mas'oudi's narrative one detail which deserves particular attention. According to him the Zindj at first occupied only the eastern parts of South-Africa between the Upper Nile and the Ocean, and further south the land of Sofala. The black tribes which originally occupied the western parts would like the Zindj have descended of Kush, but from the earliest times they would also have constituted a quite distinct group. This, I think, is a valuable clue to the study of South-Western Africa. It is mostly in the west that we find non-Bantu tribes. In the south they are met with either isolated, or mixed up with the Bantu, as far north as the upper streams of the Kwanza. Perhaps some of them may still be discover- ed living in the mysterious caves of the Katanga. Then going further to the north-west, 'we meet with them in the Congo forest, and still more to the north they occupy the country all to themselv es. Then, if we look at the physical features, of those tribes in the west which speak Bantu languages, we find that they belong to at least two distinct types, the one very similar to the most refined Bantu of the east, the other approaching more to the Bushman. Further, the ruling tribes of the greater part of the Congo basin and the Kwanza seem to have belonged until quite recently to what was called the Jinga nation. All these considerations lead me to form a view of the south-western nations of Afriqa which agrees entirely with Mas'oudi's account. The original occupants of the Damaraland, Benguela, Angola, the Congo, and in general of nearly all that part of South-Africa which is to the west of the meridian 1. Vol. Ill, p. 30. Introduction. xxxvii of the Vittoria Falls, were not Bantu. It is only in comparatively recent times, probably not before the Christian Era, that Zindj invaders from the east, called Jinga (== Zinga), overran their cbuntry, and imposed upon them both their rule and their language. yj. What was the' origin of those non - Bantu tribes ? This is a difficult question to answer owing to the want of positive documents. General tifadition, handed down to us mostly by the earliest Fathers of the Church, considers Phuth, the third son of Cham, as the father of the original occupants of Western Africa. If therefore his name meant " west '■' as the word Mbunda, or Ponda, or Puta, or Mbwudu, I should suspect that the Ma-mbiinda are children of Phuth. If we must allow with Mas'oudi that they are descended from;Kush, and this I think is the most correct opinion, it may be that, being originally Kush's children, they had to submit, even before the earliest Jinga invasions, to the yoke of people descended from Phuth, and that they bor- rowed from these first rulers the name of Mbunda, which most of them have kept to this day. ■ What is certain is, that several of the Bantu languages of South- Western Africa, or the so-called Mbunda languages, have a certain number of words in common with those of the Bushmen, as if these were the true aborigines of those parts. yS. In any case, at least one' of the above conclusions seems to be safe, and may serve as a good starting point, viz. that the , original Bantu, or Zindj, were of Kush's race. How much foreign blood has filtered into theirs, and transformed it in the course of ages, even in the land which was theirs from the earliest times, is another question, the solution of which would shed light on the history of South- Africa, its modern inhabitants, and its languages. But a thick veil of mystery hangs over it. South-Africa has long been the terra ineognita of classical writers. Sparse data may however be picked up here and there regarding the relation of its occupants to the outer world, which, if brought together, may at least shew that the land which was unknown to some civilized nations was not necessarily so to all. 7p. Relations of the, Bantu to northehi nations in Central Africa. — It appears certain that there has existed continued intercourse in ancient times between the eastern Bantu and the tribes to the north of them, but I find no evidence that such relations, generally hostile or strictly comniercial, have ever produced any mixtures of races in the "Bantu field. The manner of acting of the Masai with respect to the Bantu in our own times may perhaps be regarded as the type of what has been going on for centuries. These warlike tribes have penetrated from the north into the Bantu' field as far as the S* parallel of south latitude, forcing their way through the Kamba, the Sagara, the Rangi, and other Bantu tribes, all of which are agricultural ; but, instead of amalgamating with their enemies, they have kept their own language and customs, entirely distinct from those of their neighbours. There is nothing to show that the same hostile spirit between xx^viii Introduction. the two races has npt been going on for centuries, or that it has ever produced other effects than it does now. 80. Even the nearest approach I can find to friendly relations between the Bantu and the northern nations in ancient times was not of a nature to create a mixture of blood and languages. I read it in the " Christian Topography " of the Egyptian monk Cosmas Indicopleustes, a work written about A. D; 547. It is a typical description of the manner in which trade used to be carried on in Central Africa in his time. This is what he writes ('): — " Beyond Barbaria (also called Troglodyticq, .i. e. the actual Somali-land), there stretches the Ocean, which has there the name of Ztyyiov {Zingi, the sea of the Zindj of the Arabs, whence Zanzi-\>^r). Bordering on the same sea, there is the land called Sasos (South-EasternAfrica), which possesses abundant gold-mines, [xsxaXXa TtoXXa yt^^Mrsiw syou by my Tonga informants under the name oi Ma-punu, which cannot fail to remind one of the PuHy or Pu)i(, of the hieroglyphic inscriptions. I also think it probable that the same sort of relations between the Sabaeans and the Bantu are implied by those chapters of the third Book of Kings and the first of Paralipomena, in which the coming of the Queen of Saba to Jerusalem is coupled with the narrative of the expedition to Ophir. For, however much may have been written to the contrary, we may still be allowed to think that the first Arab traders whom the Portuguese met at Sofala with ships laden with gold were correct in saying that this was the place where Salomon's ships used to come to get the precious metal, if not the other curiosities mentioned in the Bible. Some have even long since thought that they had. shown on other evidence that the lands of Ophir, Paz, Upaz, artd Parua-im, whence the Hebrews and Tyrians used to export treasures are in the neighbourhiod of Cosmas' Sasos and Hero- dotus' Table of the Sun ('). I think that Solomon's Ophir, called Su 700. On the whole, my opinion is that the Bantu race is more mixed than it is thought to be. But its languages may rank among the most primitive ('). I. While going for the last time over the last proofsheet of this introduction, I noticed in the Times \!a«. follpwing'cablegi'am, which, perhaps, may indirectly throw further light on the origin of the Baiitu and their language : " Cape Town, Aug. lo. The Zimbabye ruins, which are being explored by Mr. Bent, are reported to be the most unique in the world. The walled enclosure, 86o yards round, containing many phallic emblems, is regarded as 'being a phallic temple. The walls in some places are i6ft. thick and 40ft. high. Two attempts have been made to open the large tower, which is solid and shows no opening at the top. There are ruins on a hill close by of the same age and style. These consist of nimierous walls and steps, archer and walled-up caves. There arc indications that three persons occupied these caves. The original builders were probably Phoenician Arabs. The natives have found a phallic altar sculptured with birds and large bowls, atid with a frieze representing a hunting scene. There are four quaggas at which a man is throwing • a dart while holding a dog in a leash. Behind are two elephants. Some blue and green Persiaii potttry and a copper Wade plated with gold have also been found, but no'inscriptions. Mr. Bent remains a few wpeks longer, hoping to discpver who built the ruins. " A COMPARATIVE GRAMMAR OF THE SOUTH -AFRICAN BAlNTU LANGUAGES. G)[)apter I. GENERAL PRINCIPLES AND PHONETICS. 1, — What makes it possible to embrace in one work the nume- rous languages which are heard from Angola to the Comoro islands and from Kafirland to Fernando Po, is that, however manifold they may seem to be in point of vocabulary, they -are none the less essentially one and the same in point of grammatical structure, and that, by elucidating certain phonetic laws, we may even bring out the identical origin in different languages of a large number of words which, at first sight, might have been thought to have nothing but their meaning in common. 2, — In this variety of languages, it was necessary, in order to avoid confusion, to select one as our standard, so as to borrow exam- ples from it for all general laws throughout the work. Our choice has fallen on Tonga, which is the predominant language of that peninsula which is formed by the Chambezi, the Zambezi and the Loangwe. The plain reason of this determination is, that, of all those languages on which a fair amount of materials has been available to us, Tonga is the one which, on the whole, best represents the peculiar features of the whole group. And, as it is also the most central, it is only natural it should be so. It might be asked whether Tonga has also the advantage of being more primitive than the better known coast languages, such as Kafir, Swahili, Herero, Angola, Mozambique, Mpongwe, etc. But this is a question we prefer to leave to the judgment of the reader. I. — Hlptiabet. 3. — Unfortunately the various scholars who have dealt with Bantu languages have adopted different alphabets, thus giving in many instances to the same letter widely different powers. Hence it. was no easy task, in a work like this, to keep uniformity without creating confusion. In this difficulty, no better plan has suggested itself than to attribute to every letter the value which is now attached to it by the larger number of Bantu scholars, without taking divergencies on the part of the others into consideration. ■4. — N. B. i.lna few cases this work so far yields to deeply rooted customs, as to follow them when they attribute to a letter or to a combination of letters, in one particular dialect, a value different from that which it has in most of the. others. Thus, in Zulu and Xosa, the letters c and x are used to represent dicks or peculiar sounds proper to these dialects, though these same letters have a different value in the other languages. Again, in certain Eastern languages, we represent by cA a sound which differs little from that of cA in church, though in the other languages the same sound is represented by the simple c. 5. — 2. In certain cases, where it was necessary to distinguish slight varieties of sound proper to particular dialects from the more common pronunciation, confusion is avoided by giving a Gothic shape to certain letters. Hence the following are the values of the letters used in this work : 6. — a = a va. father. E"x. t^.t^., my father. 7. — b — (5 in bone. Ex. hama, my mother. Kxception. — In Tonga and several other languages, b before ti and o sounds nearly like the Dutch w in wijn. Thus mu-bua, " a dog ", is pronounced nearly like mu--wua. 8. — C or ch = ^^ \n church (approximately). Ex. ci-ntu, a thing. N. B. To be more exact, this sound comes between that oich in church and that of / in tune. Exceptions. — i. In Chwana it is necessary to distinguish the two sounds cand ch. The simple c sounds nearly exactly like ch in church, while ch adds an aspiration to the same sound. 2. In Kafir (Zulu and Xosa), c represents a click-sound (of. n. 36). See also n. 25. 9. — d = «/in done. ^x. m-dezze, beard. Exception. — In Chwana rf represents a sound which stands halfway between rf and r, as in mo-sadi " a woman ". It is even written ^ by Livingstone and some other authors. Others represent the same sound by /. N. B. We represent by li a sound similar to that of ih in this, that (82). 10. — e = ai in chair. Ex. im-b&lQl^, sheep. Exception. — In Kafir, when e is followed immediately by a syllable which contains i or it, it sounds like the French ^ in bont^. Ex. ttsakleli, he lives still ; w&tu, my dear. Pronounce : usahlili, wdtu. In Chwana also, the letter e represents slightly different sounds in different positions, but the laws which regulate these differences have not yet been brought to light. 11. — f=/in fall. Ex. ku-iua, to die. Exception. — In Chwana, / sounds nearly like the Dutch v in vader. In certain dialects of this same language, it sounds more like a sort of labial h. Ex. le-fatshe " the ground " (also spelt le-iiatshe). Alphabet. 12. — %~i in gone. Ex. i-golezia, evening. N. B. We represent by g the sound o{g in bring. Ex. in-^ombe, cattle. Kxception. — In Chwana, when g is not immediately preceded by «, it sounds like the Dutch^ in goed (Arabic ghain). Ex. -a-gago, thine. 13. — h = A in Aome. We never use this letter in Tpnga proper. Ex. in Kafir : t-hode, a dove. N. B. I. Of course h has not this value in those instances in which the sound aich in church and that oi sh in shall »x^ represented by ch and sh (8,29). 2. In Chwana, the singular custom has prevailed of rendering by sh the sound oish in shall, though in this same language tsh is used to represent the sound Is followed by an aspiration. 1-4. — i = 2 both in ravine and in tin. Ex. ku-s\ka, to arrive; cx-tonga, the Tonga language. TV. B. The sound of i in tin and in the Tonga word ci-tonga is rendered in this work by i in a few instances where it was necessary to call the attention of the reader to its susceptibility of being changed to e or of being elided (Cf. n. 270). 15. — j —j in juice (approximately). Ex. i-]ulu, the sky. N. B. I. To be more exact, /is the counterpart oi c, representing a sound which holds the middle between j in juice and d in due. Exception must be made for Kafir and apparently for a few Swal^ili words, where / has almost exactly the sound ai j in juice. Ex. tt ku-iika, to turn round, (in Kafir). 2. The sound of the French y in /okt- is represented by^ (without the dot). This sound does'not exist in Tonga nor in most of the interior dialects. It is heard in AngoJa, Karanga, Chwana, etc. ^x. go-ja, to eat, (in Chwana). 16. — k = /§ in key. Ex. ku-^Lala, to sit. N. B. We represent by ft a sound similar to that of the German ch vabuch. Ex, wu-iua grass, (in Karanga). 17. — l = /in /ia!OT(f (approximately). Ex. lala, lie down. JV. B. I. To be more exact, / represents in most dialects a sound which is midway between that of / and that of r. After the vowels a, e and 0, it sounds more like /, while after the vowels i and u it sounds more like r, as if these sounds z and r as well as u a.ndr had some sort of affinity. In some cases it sounds more like d. In fact, in most Bantu languages, /, d, and r are essentially one and the same letter, the pronunciation of which varies slightly according to position. In Chwana / and d are to r proper what d is to t in the other languages. 2. In Kafir, / is pronounced entirely as in English. 18. — m = w in mine, embers. Ex. va.u-lova.be, a boy. 19. — n = % in nail, stand. Ex. ivx-jivia, lice. N. B. We represent by n a Mpongwe sound which stands halfway between n and /. Some authors render the same sound simply by n, others by nl. Ex. o-v.ome, a husband, (alias o-nome, o-nlome). 20. — = in boy. Ex. mu-oyo, the heart. Exception. — In Kafir, when o is followed immediately by a syllable which contains u, it sounds like in rope. Ex. z n-dlovu, an elephant. In Chwana also, the letter repre- sents slightly different sounds in different positions, but, here again, the laws which regu- late these differences have not yet been brought to light. South-African Bantu Languages. N. B. We represent by o a sound which is midway between a and o. Bleek renders the same sound by a. Some Mpongwe scholars render it by A, and others by o. 21. — P=/ in /aj-j. Ex. ku-'^ia, to burn. 22. — q = a click sound (cf. 37). 23. — r = r in rude. This sound, in Tonga, is merely a phonetic modification of / (n. 17). It exists as a sound plainly distinct from that of this letter in Chwana, Karanga, Mozambique, etc. Ex. go-xata, to love (in Chwana). Exception. — In Kafir, we represent by r a sound similar to that of the German ch in Nachi, though somewhat more guttural. Ex. u ku-razula, to tear ; i rati, a great man. 24. — S = .s' in see. Ex. ku-samba, to wash. TV. B. We represent by 8 a sound which stands halfway between ih in think and s in see. It is not heard in Tonga. It exists in Karanga, Kamba, Herero, etc. Ex. u-iwika, to arrive, (in Karanga). 25. — t = ^ in tin. Ex. -tatu, three. N. B. In Kafir isk fs used to render the sound of ck in church. Ex. 11 ku-isha, to burn. (Cf nn. 8 and 4.) 26. — u = « in rude. Ex. mi-buyVi, baobab-trees. 27. — v = z' in over. Ex. im-vula, raxn. N. B. We represent by b a Mpongue sound which is said to approximate to hu in the French huitre. 28. — w represents a sound not quite so full as our English w. Generally it is a remnant of a weakened labial sound. Ex. awo, there { = apo, n. 693, tables). N. B. U between a consonant and a vowel has been written iv by various authors in many cases where probably it should not be so, and vice versa. Thus the word for '' child " should probably be written mw-ana, not mu-ana in Swahili, because here the semi-vowel sound ti is more consonantal than vocal, as we see that in this language the substantives of the same class as mw-ana generally drop the vowel u of their prefix mu, as in m-itt, a person, 7n-ji, a village, etc. (= mu-iu, vni-zi, cf. 366) ; while the same word should be written mu-ana in Shambala, because in this language the « of the prefix is generally kept, as in mu-ntu, a person, mu-iue, a head, etc.. 29. — X or sh = .s'^ in shall. This sound is not heard in Tonga. It exists in Chwana, Karanga, Angola, etc. Ex. xe " the chief " (in Karanga) \go-sha " to die ", (in Chwana). Exception. — In Kafir *• represents a click-sound (cf 38). 30. — y—y in year. Ex. ku-yoya, to breathe. TV. B. I. When ji is preceded by n, don t, the two sounds are combined into one. We thus obtain the three compound sounds ny, ify, and fy, which have no exact equivalents in English. The nearest approaches to them are nt in onion, dia duty and t in tune. Of these three sounds ny alone is heard in Tonga, as in inyaii, a.buffalo. Dy and ty are used mostl)' in Kafir, as in ii ku-dyoba, to bemire, u ku-tya, food, etc.. 2. Ty in Herero sounds apparently like c in Tonga, (n. 8). :^:.^5^;v;" Alphabet. 5 31. — Z — z'm. zone. Ex. ku-zala.i to become full. N. B. We represent by \ a sound which is to z what g is to s. Ex, ti-\wara, " to beget, " (in Karanga). ADDITIONAL SOUNDS IN CHWANA. (Suto, Tlhaping, Rolong, Kololo.etc.) 32. — tl, in Chwana, approximates to tl in bottle. Ex. tlala, hunger. tlh sounds more strongly aspirated than tl. Ex. tlhapi, a fish. ADDITIONAL SOUNDS IN KAFIR. YXosa, Zulu and Tebele.) 33. — hi approximates to the Greek combination /X. Ex. u ku- hlala, to sit. This sound has also been spelt kl and si by various writers. tl, in Kafir, represents a sound similar to that oihl, but preceded by /. In fact, it is a mere modification of hi, caused by the presence of n before it. Ex. in-tlalo, a sitting. 34. — dl represents the two soft sounds corresponding to ;^/and tl. When not preceded by n, it approximates to gl in the Dutch glorie. 'Ex.u-ku-dleka, to be spent. When preceded hy n, it sounds more exactly as it is spelt. Ex. in-dleko, expenses. 35. — The above sounds are not yet what have been termed clicks. These are still less easy to describe, being produced, as they are, rather by drawing in than by expressing sound. They have some analogy to k and^. They are six in number, viz. : 36. — C, produced by drawing a hard sound as if from the front teeth inwards. Ex. u-ku-canda, to split, gc, a soft sound corresponding to c. Ex. ingca, grass. 37. — q, produced by drawing a hard sound as if from the palate downwards. Ex. i qaqa, a muir-cat (musk cat). A'. B. This click-sound is sometimes heard in Suto. gq, a soft sound corresponding to q. Ex. in-gqwelo, a wagon. 38. — X, produced by drawing a hard sound as if from the side^ , teeth inwards. Ex. u ku-xoxa, to converse. gx, a soft sound corresponding to x. Ex. in-gxoxo, a debate. II. — ctjaractemtic Features of t^e Bantu lEiiamilp of Jianguaffcs. 39. _ irst PRINCIPLE. — In these languages, concord is established by means, not of suffixes, but of prefixes, which being, as a rule, expressed first before the substantive, are then repeated, under a form sometimes identical and sometimes modified, before every expression which has to agree with it. ■40. — These prefixes are, in the best favoured dialects, eighteen in number, some of them importing generally a plural, the others a singular meaning. 41. — The same stem, by assuming different prefixes, obtains various meanings, sometimes quite opposite. Ex. i) Ulu-fonga, a Tonga. 3) M.u-samo, a tree, a medicine. 5) \-samo (or li-samo), a beam. 7) Bu-ionga, the Tonga territory. 8) Ku-tut, an ear. 9) In-samo, a whipstick. 11) Ci-samo, a stump of wood. 13) 'Ka.-samo, a stick. 15) Ija-limi, the tongue. 2) 'Ba.-^onga, Tonga people. 4) IS/li-samo, trees. 6) 'M.k-samo, beams. 6) M.a.-tui, ears. 10) In-samo (or zia-samo) whipsticks. 12) ZA-samo, stumps. 14) Tu-samo, sticks. 10) \a-dimi, tongues. 16) h-nsi (or-pa-nst), down. 17) Ku-nsi, below. 18) Mu-««, underneath. 42. Examples illustrating the general principle of concord 1. Mu-a«a u-ako u-a/ua; The-child he-yours he is dead; Your child is dead ; I have buried him. 2. Bsi-ana hSi-ako ha.-afua ; The-children they-yours they are dead; Your children are dead ; I have buried them. 3. l/l\isamo M-ako vi-afua ; The-tree it-yours it is dead; Your tree is dead ; I have cut it down. 4. IJli-savio \-ako vafua ; The-trees they-yours they are dead; Your trees are dead ; I have cut them down. 5. \A-mue sekua ( = U-sekua) Xv-ako nda-X\-jana ka-W-fuide. It-one duck it-yours I have it found when-it-dead. I have found one of your ducks dead. nda-tau-zika. I have him buried. nda-hSi-sika. I have them buried. nda-u-Uma. I have it cut down. nda-\-tema. I have them cut down. Characteristic Features. 6. yia-sekua a,-ako, ma.-mgi nda-a-Jana ka-Srfuide. The-ducks they-yours they-many I have them found when-they-dead. I have found several of your ducks dead. 7. Bu-« \i\x-ako hoonse hu-amana, fu-a-hu-h'a. The-honey it-yours all it is finished we-have it eaten. All your honey is finished, we have eaten it. , 8. "Ka-mue ]iin-tut k.Vi-ako n-k.u-6i, It-one ear it-yours is dirty, One of your ears is dirty, wash it. 9. I-mue n-gombe ( = in-gombe) \-ako It-one cow it-ycurs u-\i\X-sambe. you it wash. i-afua, it is dead, iua-i-stnza. we have it skinned. One of your cows is dead, we have skinned it. In-gombe zVako zi-ingi zi-afua, Cows they-yours they-many they are dead. Several cows of yours are dead, we have skinned them. tua-z\-stnza. we have them skinned. II. 13' 14. IS' Eci ci-ntu ci-ako n-cibi, u-ci-sambe. This thing it-yours is it dirty, you it wash. This thing of yours is dirty, wash it. .£zi zi-n/u zi-ako n-zi-bt, u-zysambe. These things they-yours are they dirty, you them wash. These things of yours are dirty, wash them. Ka-»2«iJ ysA-cece k.a.-angu ka.-afua, nda-lca-ztka. It-one baby it-mine it is dead, I have it buried A baby of Inine is dead, I have buried it. Tu-cece tu-eiu t\l-mue tu-afua, tn-mtie Babies they-ours they-some they are dead,' they-some Some of our babies are dead, others are still sick. 'Lu-sabtra The-httle-baby IvL-aftgu it-mine lu-aftM, it is dead. tu-ci-fua. they still are-sick. nJa-lu-ztka. I have it buried. My little baby is dead, I have buried it. 16. hrfuefui ( = Close : ■pa-fuefui) a Mpande, to Mpande, pa-// a mu-longa. there is with a-river. There is a river near Mpande. 1 7. YiVi-nsi (ku) -a bu-sanza \sM-a-bikua mu-lilo. Underneath (it) of the-table there was placed fire. Under the table was placed fire. 18. yin-nganda (ys\\x)ako mu-Ia-si'a : tinsi ndi-la-TOM-njira. In the-house (in) yours therein is dark : t'is not I will therein enter. It is dark in your house : I will not enter therein ('). I. A series of Zulu and Herero sentences, similar to the above, all based upon the theme: " Our handsome So-and-so appears, we love him ", may be seen in Bleek's excellent " Comparative Grammar of the South-African Languages", pp. 96-100. Unfortunately it is necessary to warn the reader that the Zulu sentences in that series are not quite correct in the sense in which they are intended. For the expressions si-«i~ia.nda, si-ba-tandu, etc., which Bleek renders by " we love him, we love them, etc. ", are never used by natives with this meaning without being determined by some other e.Kpression. (Cf. nn..844, 846, 915, etc.) 8 SotUh-African Bantu Languages. 43. — It may be noticed already here that locatives and locative •expressions, such as those in the last three sentences may serve as what are subjects from our point of view, so that even verbs, adjectives and other determinatives are made to agree with them. This is the cause of very great difficulties to the student of these languages, because it is. the source of an incredible variety of con- structions which are entirely unknown in our own languages (cf. nn. 53O-508 ; 693-704, etc.). 44. — lid, PRINCIPLE. _ Monosyllabic stems of verbs and nouns (substantives, adjectives, and pronouns) are in nearly all the Bantu lahguages subjected to special laws tending to give them prefixes or suffixes in cases where other stems have none, as if, in polite Bantu, there were, or at least had been, a general aversion to monosyllables, or, more exactly, to pronouncing an accented sound without its being accompanied by a weaker one. Thus, in those dialects which do not express in nouns the prefix li, this same prefix is found to be expressed or replaced by something else before monosyllabic stems! (cf. 413, 414). Again, in nearly all the dialects, though the imperative exhibits generally the bare stem of the word, the law is found to change when there is question of monosyllables (cf. 837-841). Cf. also nn. 283, 325, 368, 389, 472, 611,661, 765, 808, etc. This principle may be termed " the law of avoiding monosyllables or single sounds ". It may be compared with triliterality in the Semitic languages. , 45. — The chief difficulty connected with the application of this principle is to know when a stem is really monosyllabic and when it is not so, because the accent is not always sufficiently marked to exclude all doubt, but principally because, in some cases, the very same stem, apparently identical in two different languages, may however happen to be perfectly monosyllabic in the one and yet to consist really of two sounds in the other, so that in these cases analogy is often misleading to tlie inattentive. Thus the principal element of the pronoun which means " we, us, " is in Tonga sue, in two inflections of the voice, the first (su-) on a lower, the second (-^) on a higher tone, while, in Swahili, it is swi, a single voice-inflec- tion, variously written sui and si. 4;0< — ^- ^- '• Hence, when monosyllables are met with in Bantu authors, they Characteristic Features. must generally be considered as enclitics or as proclitics, or they are onomatopoetic words (n. 596). 2. The stems which begin with vowels are generally governed by principles which have much analogy with the applications of the law of avoiding monosyllables. 3. There are many instances of stems which are monosyllabic in certain languages, while in others they begin with a vowel. For instance, the Tonga stems -iza " come " and -ha " steal " have in Swahili the forms -ja and -/ia. Possibly, in such stems as -iza and -z'da, the initial vowel is not radical, but is a mere application of the law of avoiding monosyllables. 47.— Hid PRINCIPLE. — Phonetic changes being, as migh^ be expected, one of the main sources of differences between the various Bantu languages, it is to be noted : 48. — i) That, on the whole, they affect consonants more than vowels. This principle, though apparently new in philology, can be so readily verified that it needs no proof here. 49. — 2) That those among these changes which affect vowels bear mostly : a) On vowels which begin a stem, as t in -injila or -njila, enter, b) On the weaker of two vowels which are next to one another, as m (alias w) in -fua, -fwg., or -fa, die. — ■ Other instances will be mentioned in their proper place (cf 200, 213, 237, etc.). 50. — 3) That those among these changes which affect consonants may be traced, in a large proportion, to different conformations of lips and nose, with the well-known additions or absence of lip-rings, nose-rings, the various sorts of artificial gaps in the teeth, etc. ('). I. A large proportion of the Bantu tribes haVesuch marks which necessarily modify the pronunciation of certain consonants. Thus the Tonga Icnoclc out their upper incisors, when they come to the age of puberty. My informants used to say that the gap thus produced is their national marli, exactly as circumcision is the national marlc of the Kafirs. It is noticeable that tlie Lea people, who are a Tonga tribe hving near the Victoria Falls have given up this practice, since they have yielded their liberty to the Sotse. Livingstone says that " when questioned respecting the origin of the same practice, the Tonga reply that tlieir object is to be like oxen, " and that" those wh ) retain their teeth they consider to resemble zebras. " (Missionary Travels, London 1857, p, 532.) The Nyamwezi are also mentioned as knocking out their upper incisors' (Giraud, Les lacs de I' Afrique iquatoriaU, i8go, p, 303). My informants added that the tribes which inhabit the country near Jhe Loangwe, or, as they used to call them, the Mhara, have the custom of filing their front teeth to a point, this being likewise their national mark. It is well known that this custom is more general, as it is common to a large number of tribes near Mozambique and on what used to be called formerly by the Portuguese, ' ' the Senna Rivers, (rios de Senna)" — The Hehe have also filed teeth (Giraud, Lcs lacs de V Afrique iquatoHale, p. 141). Cf. W. Montagu Kerr's The Par interior, p. n6, regarding the Mashona. The KtimU, on the Kunene river, knock out the two middle incisors of the lower jaw and file the two corresponding teeth of the upper jaw to the shape of an inverted V (Missions catholiqiiei, 1888, p. 269), A similar custom has been noticed by Dr. Hahn among the Herero, (/iid.,'-p. 270). According to Johnston (T/ie River Congo, 1884, p 402), the two front teeth of the upper jaw are occa- sionnally chipped among the Congo tribe of Pallaballa, and further up the river, this custom is regular. The same writer mentions that "among the Ba-bwende of Ma-nyanga and the surrounding district large nose-rings are passed through the septum of the nose " (Ibid.). — The Up-rings of the women gn the Mozambique coast are too well-known to require description. lO South-African Bantu Languages. 51. — 4) That the nasals n and m have in many cases the beneficial effect of retaining consonants which, according to the general laws, should have been weakened or dropped altogether (nn. 93, 95, 192 compared with 172, 116, 126, 148, etc.), though in other cases those same nasals n and m have the apparently contrary effect of modifying the consonants which they precede (cf. ^i, 74, n^ 69, 99 note, etc.). — This note is very important. 52. — The explanation of this 3^^ principle alone with its various exceptions and particular applications would require a whole vo- lume. It will form the basis of the next article. Meanwhile a few of its applications may be seen in the examples given below(*). A large supply of more striking examples may be seen in the chapters on substantives and adjectives. * SPECIMENS OF PHONETIC CHANGES. to shape Steal see recover [mtr.) burn (Mr.) hear die Tonga ku-bumba -ba -bona -pona -pia -nvua -fuai Subia ku-bumba -eba -bona -ifuba -fua Yao ku-gumba -iwa -wona -pola -pia ... -uwa Sagara i ku-umba -hidja. -bona ... -fua Shambala ku-umba -uya -ona -bona -wa -ta. Boondei kuumba -bawa -ona -bona -ya ... -fa Taita ku-umba -iva -ona -bona (?) -iya -fwa Nyamwezi -iwa -wona -pia ... ■cha Kamba ku-umba -uya -ona -wona -iwa -gwa Swahili kuumba -iba -ona -pona ... ... -fa Pokomo ku-umba -iva -ona -bfona -bfia -fwa Nika ku-umba -ia -ona -vona -via ... -fua Senna ku-umba ■ba -ona -psa -bva -fa Karanga u-wumba -iba -wona -psa -wua -fa Yeye -iba -mona -pia -iva -fa Ganda ku-umba -ba ... -wona ... -fa Xosa-Kafir ku-bumba -ba -bona -pola -tsha -va -fa Zulu-Kaflr ku-bumba , -eba -bona -pola -tsha -zwa -fa Herero ku-ungura -vaka -muna -pia -zuva -ta(i28) Bihe -iva -mona -pola -pia -yeva -fa Kwengo ... -eba -mona ... ... Lojazl ... ... -sa Rotse ... -ifa -mona -bola (?) -bia -yopa -fa Nyengo -mona ... -yuba -fa Rua ... -bona (.?) ... -va ... Angola -iya -mona ... -bia -ivua -fua Mb'amba ... ... ... -hia ... ... Lower Congo wumba -yiya -mona -vula -via -wa -fua Mozambique w-upa .,"'>"^„^ -ona -vona ... -iwa -kwa Kilimane •iba{?) -ona -vola -pia(i') -iwa -ukwa „. (weak ChwanaJ3t„„g go-bopa -popa ■ucwa do. -bona -pona -fola -phola -sha do. -utlwa do. -shwa do. fweuk Mpongwe ,t,o„g goma do. -yufa -dyufa -yena ■dyena -vona -pona -via -pia -yogo -dyogo -yuwa -dyuwa Dualla ... -iba -jena ... -bwa -wo Characteristic Features. II 53. — N- B. I. For many dialects, viz. for Subia, Lojazi, Angola, etc., the scantiness of materials at our disposal is the only cause of the blanks left in the subjoined tables. With more knowledge, most of these might probably be filled up with the exact words required. 54;. — 2. In the same tables we give in every column only such wor(is as seem to have been originally identical in form' or nearly so. However, as may be readily observed, some words contain in certain languages one element more than in the others. For instance, in the Herero word -vaka " to steal, " the first element {vd) is essentially the same as the Tonga -^a in the same column, but the element -^a is superadded. Likewise in the Sham- bala word -itanga " to call ", the element nga is superadded to the Tonga -ita, etc., etc. 55. — IV* PRINCIPLE. — The preceding principle causes a great many words to appear in the very same dialect under two or even three different forms, according as they are connected or not with a nasal sound, n or m. ' SPECIMENS OF PHONETIC CHANGES. (Continued.) dawn leave arrive come dress {mir.) become full beget Tonga -cia -sia -sika -iza -zuata -zala -ziala Subia ... ... -iza -fala(?). Yao -cha ... -ika -isa -wala -gumbala Sagara -cha ... -ija -vala Shambala -xika -iza -vala ... Boondel -cha -sia ... -eza -vyala Taita -cha ... -fiica -ja -ruara -vala Nyam-fivezi ... -xika -iza -zuala -okala -wyala Kamba -cha (?) -iria (?) -vika ... -iwatoa(?) ... -chaa Swahili -cha -fika -ja -vaa -jaa -zaa Pokomo ... -yadsa ■fika -dza ... -dzaa -wyaa Nika -cha -sia -fika (.?) -dza -fuala -dzala -vyala Senna -cia ,■ -sia -fika -dza -bvara -dzara -bala Karanga ►.. -Bwika -ja -mbara -jara -jwara Yeye ... ... -ya ... ... Ganda -kia ... -tuka -ja -ambala -jula -zala Xosa-Kafir -sa -shiya -fika -za -ambata -zala -zala Zulu-Kafir -sa -shiya -fika ■za -ambata -zala -zala Herero -tya -sia ... -ya -koata Bihe ... -sia . . . -iya -wala ... K-wengo ... ... -iya Lojazi ■sa(?) ... Rotse ... -dia ... -ya ... Nyengo -iya -zala Rua ... ... -fika ... -vala Angola -xia -bixila -iza -zuata -vala -vuala Mbamba . .. . .. -zuala Lower Congo -kia -xisa -nyeka -iza -vuata -zala -uta Mozambique -hia , -pia — -wara -chara -yara Killmane -tia fia — -ambala -bala Giiwanah^ak strong Iweak Mpongwej,,^„„g -sa(S) .' -sia -fitlha -tla -apara -tlala -tsala do. do. -phitlha -tlha (?) do. -tlhala (?) do. ... ... -wia -bia -wora -bora -yana -dyana Dualla. -sa -dia ... .-ya -boto (?) ... -yaa(?) 12 South-AfriQan Bantu Languages. 56. — Thus, in Tonga, the word for " sun, " is in most cases pronounced i-zuba. Now this is a weakened form equivalent to li- zuba, which is heard only when emphasis is laid on the first syllable (41 1). And, if the copula n (cf. 582) be placed before it, the same word changes to di-zuba. Hence we may hear three different forms of the same word, viz, i-zuba, li-zuba and di-zuba, or, to be more exact, three distinct forms of the same grammatical prefix to the word, viz. ?',' li and di. 57. — Again, in Tonga, the word for " down " is in most cases pronounced a-nsi. But this apparently is a weakened form ofpsi-nsi, which reappears after n, with the effect of changing this n to m. Hence two forms for the same element, viz, a and pa. 58. — Again, if a dialect changes/ to ,^ in the generality of cases SPECIMENS OF PHONETIC CHANGES. (Continued.) cook buy go in sit, remain drink eat lie do-svn Tonga -jika .(g)ula -(i)njila -kala -nyua -lia -lala Subia -kika -gula rikara -ria -lala Yao -jinjila -ngwa -lia ... Sagara -ambika -gula -ingila -kara -nyua -dia ... Shambala -dika -gula -engila -kala -nua -ja ... Boondei ' •ambika -gula -irgila -ekala -nwa -da ... Taita -gula -ngila -kala -nwa Nyamwezi -deka -gula -ingila -ikala -ngua -lia -I'aia Kamba -wia -ua -ikia -kaa -nioa -iya -mama Sivahili -pika -ingia -kaa -nywa -la -lala Pokomo -mbika -guya -ntyia -kaa -nwa -tya -hara Nika -jita (?) -gula -ing.ira -kala -noa -ria -lala Senna , -pika -^ula -kala -mwa -dya ... Karanga -bika -nguina -gara -nua -ria -rara Yeye -wora -njena ... ... ... -ran kara Ganda -sika -gula -ingila -nyua -lia ... Xosa-Kaflr -peka -ngena -h'lala ... -tya -lala Zulu-Kafir -peka -ngena -hiala -dla -lala Herero ... -kara -nua -ria -rara Bihe ... -nua -lia -lala Kwengo -bela -kara -ria -lala Lojazi -twena -ikara Rotse -ola -ingena -ikara -nua -riya -rankana Nyengo -ola -ingena -kala ... Rua -ipika -ota(.?) -twela -shia -laia Angola -bika (.?) -kaia -nua -ria -lambarala Mbanil)a ... ... -nua -dia Lower Congo -Icala -nua -dia -lavalala Mozambique -apea -Iceia -kala -lia -thala Kilimane -gula -vira -kala -umua -oja _ , f weak -apaea do. -tsena do. ■nwa do. -ja do. -laia -tala Mponewej:;::„\ -gola -kola -yingina -dyingina ... -yonga -dyonga -nya do. -nana do. Dualla -ingea -ja -nyo da -nanga Characteristic Peatures. 13 as in Nika, then the word for " three " which is " taiu " in the larger number of the Bantu languages, will, in this particular dialect, appear generally under the form /m/m, as in [^z-hi vi-Jiahu, three things (= Tonga zi-ntu zi-tatu); but it will recover at least partly its proper form when influenced by n, expressed or even understood, as in n-Qombe tahu, three cows (= Tonga in-^ombe n-tatu). Cf. nn. 479, 608, JT,, 83, etc. 59. — The applications of this principle are chiefly remarkable in Chwana and Mpongwe. And this is the real cause- why, in the subjoined table it has been necessary to distinguish in these dialects between weaker and stronger ioxm.?,, the latter being in most cases nasal, as will be shown later (cf. 183-197 and 214). Cf. also the table of adjectives, n. 601. SPECIMENS OF PHONETIC CHANGES. (Continued.) cry hoe bite bring ■walk send call Tonga -lila -lima -luma , -leta -enda ■tuma -ita Subia -rira ... ■enda -tuma Yao -lila -lima -luma -enda -tuma -wilanga Sagara -lila -lima . -luma (.?) -genda Shambala -lila -lima -luma -I'eta -genda -tuma -itanga Boondei -lila -lirha , -luma -leta -genda -tuma -itanga Taita -lila -lima ■luma -leta -tuma -ita Nyam'wezi -lila -lima -luma -ita Kamba -iya (?) -ima -uma ' -ette. -enda -tuma -ita SwahlH -lia -lima -uma -lela -enda -tuma -ita Pokomo ... -muma -yeha -enda -liuma Nika -rira , -rima -luma -reha ■enda -huma -iha Senna -lira -lima -ruma -enda -tuma -itana Karanga -lira -lima -luma -reta -enda -tuma Yeye -rira -enda -toma Ganda -lira -lima , -ruma -leta -genda -tuma -ita Xosa-Kaflr -lila -lima -luma -tuma -biza Zulu-Kaflr ^ -lila -lima -luma -tuma -biza Herero -rira -rumata -eta -enda -tuma -isana Bihe -lila -lima -lumana -nena -enda -tuma Kwengo -lila -enda -tuma -zana(?) lioja-zi -lila -enda -tuma -zana^'j Rotse -lila -lima -moma -leta -enda -tuma ■izana Nyengo -lila ... -enda -isana Rua -jila (?) -jima (?) -suma -enda -ita Angola -rila -lumata -enda -tuma -ixana Mbamba -suma -neha Lo-wer Congo -diia -enda -tuma Mozambique -unia -lima -luma -I'eia -eta -ruma -ihana Kilimane -lila -luma -enda(?) -rruma Chwana] , I strong -lela -lema -loma -lere -eta -roma -bitsa -tela -tema -toma -tere do. -toma -pitsa (weak Mpongwe|^^,„„^ -lena -dena -noma do. -yira '•■ -dyira ^ -genda -ken da -roma ■toma -vwelia -fwelia Dualla -eya ■■• ... ... -loma -bela III. — Comparative fifjonetics of tge firincipal Bantu liang;uage0. 60. — The Bantu languages interpenetrate each other so much that the principles which find application in one of them exc/uswefy are very few Indeed. This article will therefore be a mere attempt to classify some notes of greater or less import, according to the languages in which their application seems to predominate. 61 . — ^- ^- !• Those phonetic laws which are common to the largernumber of the Bantu languages, such as the change of « to m before i5 and/, will not be mentioned here, but only in the next article. 2. Concerning our sources for the various languages which are dealt with in this article, and the locality where they are spoken, cf. Introduction. TONGA. (Spoken between the Victoria Falls and Lake Bangweolo.) 62. — Tonga, whicli is taken in this work as the standard lan- guage for the reasons given above (cf. 2 and 52 table), may be said to represent fairly well, on the whole, the generality of the Bantu languages. Its most striking feature is, perhaps, to have regularly y? and cz where a large proportion of the others have the sharper sounds £-i and ki (cf. 8 and 15), as in the word -injila or -njila, which, in most of the other languages, sounds rather like -ingila (cf. 52 table). However this latter form is heard in Tonga also, a fact which shows that the difference is not very important. With regard to ci and ki, cf n. 492. 63. — It may be added that the plain sharp sounds z and s appear to be more common in Tonga than in any of the other Bantu languages, Kafir itself not excepted. This again may be seen exem- plified in the above table of verbs, in the columns of the words -sia, leave, -sika, arrive, -iza, come, -zuata, dress, -zala, become full, -ziala, beget. 64. — It may be well also to notice that j?!" is not heard in Tonga, or is replaced by w, in some cases in which many dialects have it, unless it be after n. Thus im-pewo " wind ", is pronounced mpepo in Senna and several other languages, and ansi'' down " (pansi after m) is always pronounced pansi in Kafir and several other languages (cf 5; and 534-540). Comparative Phonetics. 15 zi-ntu, things i-ndu (H-ndu) in-dezu beard n-deu (n-de'ul) 65. — Tonga may be considered as forming one language with Subia (spoken on the Zambezi, above the Victoria Falls), Bisa (') (spoken East of Lake Bangweolo), and Bemba (=) (spoken North of Lake Bangweolo.) A particular dialect of Tonga is spoken near Lake Nyassa. It differs considerably from that which is described in this work. Judging from the scanty documents at hand (^), it looks very much like a mixture of the Tonga and Senna or Shire language. ' YAO. (Spoken on the tableland between J-ake Nyassa and the coast.) 66. — 1° Of those words which are common to Tonga and Yao many are greatly reduced in form in the latter language, mostly through the fall of 5:, v or f. Perhaps it might be more correct to say that z, V andy are then changed into a mere aspiration. Ex. Tonga Yao Tonga - Yao im-vula, rain ula fula ?) ma-futa, fat, oil ma-uta(ma-^utal) t-zuba, sun li-ua (li-ua) ma-zina, names Jtteiia ( = maina) N. B. Nz is in a few cases changed to s. Ex. sala, hunger (Tonga in-zala). 67. — 2° On the contrary, those stems which in Tonga are monosyllabic or begin with a vowel or n nasal, and a few others, are found to have richer forms in Yao. Tonga Yao -injila or -njila, (to) enter -jinjila -unn, healthy -jumi(-yumi&i\.zxn) -ingi or ^nji, numerous -jinji ( -yinji 2SX.QX n) ■zala, (to) become full -guinhala 6 8 . — N. B. I. From these examples it may be seen, among other things, that/ is in favour in Yao before the initial i and u of the other languages., Cf the use of g in Sagara, n. 'jj. 2. Some stems which begin with j/ or z in Tonga, have 7 instead in Yao. Ex. li-joka, a snake (Tonga in-zokd). li-juni, a bird (Tonga i-yuni). I. Concerning Bisa, cf. Stanley 's Vocabulary in "Tlie Dark Continent ", and Last 's collection in " Polyglotta Afrieana Orientalis ", pp. 135-138. a. Concerning Bemba, cf. Last 's " Polyglotta, " pp. 131-134. 3. In fact I have seen no other specimen of this language than a., small primer which has been kindly presented to me by M. Cust. I suspect that the Tonga of Lake Nyassa are of those who are mentioned by Livingstone as having- gone to thfe Ba-mbala, and having never returned, "because they liked tha country better than theirs. " (Liv. Miss. Travels, p. 532). If so, it must be said that they have adopted, partially at least, the Mbara language which is a dialect of th6 Senna group (cf. 98). Ex. Tonga Yao -ba, (to) steal -jiwa t-ji, an egg li-jele -ita, (to) call -wilanga -kulu, aged -chekulu i6 South-African Bantu Languages. 69. ^ 4° Some peculiar changes are caused by the sound n when it combines with other consonants, viz. : NP is changed to MB. Ex. mbachile, I have painted, , ku-?n-bona, to see me, , a-tnbweni, he has seen itie, , ndawile, I hare bound, , n-ganik, I have refused, , n-yigele, I have carried, , n-japile, I have washed, , nindani, wait for me. This law of Yao explains why we have the following correspond- ences of words between Yao and Tonga, which is not subject to such' changes. NW { MB or MBW NT ND NK NG NJ NY NCH NJ NL N for n-pachile „ ku-n-wona „ a-n-weni „ n-taii)ile „ n-kanile „ n-jigele „ n-chapiU ,, n-lindani Ex. : Tonga Yao Tonga Yao im-pewo, wind m-bepo in-yati^ a buffalo n-jati mu-ntu, a person mu-ndu in-kani, a story n-gani etc., etc.. 70. — 5° It will be seen further on (86 1) that in all the Bantu .languages, verbs undergo certain phonetic changes of consonants in the perfect form. But Yao distinguishes itself among them all in this respect. Ex. ku-tama^ to sit Perfect, a-teme, he is seated. ku-kola, to get a-kwete he possesses. ku-kwela, to cliinb a-kwesile, he has climbed. ku-iaga, to put a-tasile, he has put, etc.. 71. — The Rev. Alexander Hetherwick, in his " Handbook of the Yao Language, "p. XIII, observes rightly that from this cause combined with the phonetic changes previously described, " words in Yao are so modified in the process of grammatical construction as to be almost unrecognizable by a beginner. Thus from ku-leka " to leave, " we have n-desile " I have left "... And from ku-pa, " to give, " we have a-m-bele " he gave me ", where not a single letter of the original has been retained. " 72. — ^- ^- A large number of common Yao words differ radically from those of similar meaning in the other languages, or at least seem to do so. Ex. Tonga and other languages -^/, bad (cf. adjectives, 6oi*) mu-oya, breath (cf class MU-MI, 377) ... . mu-alwni, husband (cf. class, MU-B.'\, 322*). ^«-/fl:, bow (cf. class BU-MA, 453) kutui, ear (cf. class KU-M.A., 462*) in-zila, road (cf. class IN-ZIN, 385*) ... . «'-^»«, stone (cf. LI-MA, 410*) Yao -chiniwa bu-musi asono iikunje li-pilikanyilo li-tala li-ganga Comparative Phonetics. 17 2. Interesting comparisons might be established between Yao and Chwana. It may even be said that most of the peculiar features of Yao have their counterpart in the languages of the Chvvana-Mozambique-Mpongwe group (169). NYA-MWEZI. 73. — The two Nya-mwezi dialect? on which we have most in- formation, viz. Nya-nyembe and Sukuma, differ from the generahty of the Bantu languages : — 1° By a pequliar tendency to weaken certain consonants after nasals. In this they go even further than Yao. Ex.: Tonga Nyanyemee Sukuma NY changed to N : inyama, meat inatna nania NT J) NH: mu-ntu, person mu-nhu inutihu NK )J NH: m-kuni, wood n-hui ? -it )J NG: in-huku, hen n-goko n-goko ND )) NH: -endela, go to fetch ^en/tela ? 11 » N: ndi, I ni ni MP JJ MB: im-pewo, cold m-biho m-beho )» »J MH; im-pande, pieces m-hande ? , MV >» MB: im-vula, rain m-hula m-bula 74!. — N- -5. What renders particularly interesting this tendency inNyamwezi, Yao, and, as we shall see further on, in Sagara and Gogo, to weaken consonants after nasals, is that, in many of the othet Bantu lariguages, the same nasals produce the very opposite effect, and that consequently , those stems which in grammar are subject to changes of form, such as -pia '' new ", -bi " bad ", etc., are found to be used in their weaker form in Nyaimwezi, Yao, etc., precisely in those instances in which they have their stronger form in Ganda, Chwana, Herero, Nika, Mozambique, etc., and riccvei-sa.. Ex. Nyamwezi Ganda Tonga ■wu-su wu-pia, new flour bu-ta bu-gia. btt-su bu-pia n-goma m-hia, a new drum e fi-goma e m-pia in-goina m-pia 75. — 2° By having often ^ where Tonga has/". Nvanyembe : ma-guta Sukujma Ex. Tonga : ma-futa, oil, fat „ i-fua, a bone „ -fuefui, short i-guha -nifii ? ? ■suhi 76. — 3° By eliding, in many cases, certain vowels which, in the other languages, are at most contracted or assimilated with those which follow them. Ex. Tonga : ;««w, eyes (for ?;2a-/j^) Nyanyembe : ?«'«(? Sukuma : w'wi'o ,, meno, teeth (for ma-ino) ,, m'l/io „ m'ino „ in-goma ziesu, our drums „ ngoinazisu ,, ? (for zi-a-isu). 2 i8 , South-African Bantu Languages. SAGARA AND GOGO. 77. — The documents available for these languages are so un- reliable (') that I cannot make Out any of their characteristic features with certainty. Apparentlif they are nearer to Tonga than the Nyamwezi language. The nasal seems to weaken the following consonant in some cases, as in Nyamwezi, and to be itself dropped in others, principally before s. H replaces the Tonga/, though not after m. G is apparently a favourite, at least in Kaguru, as it is found replacing not only the Tonga /^ but also v, j, and even/. Ex. Tonga Kaguru GOGO im-pewo, cold m-behu beho (?) a-nd, down ha-si ha-si ku-pona, to heal (intr.) ku-ona or ku-hona ? i-fua, a bone i-guha ? ku-vuna, to gather (harvest) ku-gola ? ij-anza, a hand i-ganza i-ganya li-nso, an eye. i-giso ziso SHAMBALA and BOONDEI. (Spoken inland facing the Pemba island.) 78. — These languages differ but little from one another, and both of them are closely allied to Sagara. Their most remarkable phonetic features are the following : — 1° The consonants which follow n nasal are firmer in Shambala and Boondei than in Sagara. 2° N nasal falls before a larger number of consonants in Sham- bala and Boondei than in Sagara, and generally it strengthens those before which it falls. 3° In Shambala and Boondei there is no preference for g as in Sagara, 4° 5" of the other languages is sounded x (english 5^)in Shambala, though not in Boondei, so that this seems to be the most palpable difference between these two languages. I. The Kaguru grammar published under Last's name is full of evident misprints. Thus, for example, he word for " man " is spelt inu-ntu (p. 14, etc), mu-nhu (p. 21, etc.), mu-nku (p. 124, etc.), mu-nbu (p, i3), mu-n/ta (p. 129). Comparative Phonetics. 19 79, — Ex. : Tonga tnu-niu, a person in-soni, shame im-pewo, cold im-vula, rain li-iiso, an eye t-fua, a bone N. B. On the whole, Shambala, Sagara, and Gogo, look more like Tonga than most of the other East African languages: '"■/ jHAMBAL.V BOONDEI Sagara mu-nfu mii-ntu mu-nhu (77 note.) xoni soni soni Pho ;;;:, .-.peho m-behu fula : L fula m-vula zi-xo ^zi-so gi-so mu-vuha mu-vuha ; i-guha -.' TAITA. (Spoken on the hills between Mombasa and Kilima-njaro.) 80. — Taltahas a great number of words which are 'not hea;r.d in the more Southern Bantu languages. As to those words which it has in common with them, when putting them together, we find no very regular transitions of sounds. We may observe however a tendency to weaken hard consonants after nasals, e. g. in n-gano " a story " (Tonga in-kani), n-guni " firewood " (Tonga ««-^««z), ki-ndu " a thing " (Tonga ci-ntu), etc. Possibly also it is a general law of Taita to change in certain cases into chu, and in others into ' vu, the sound su or fu of the generality of the Bantu languages, as in ma-vuta " fat " (Tonga ma-futa), i-chumu " a spear " (Tonga i- sumo, Swahili fumo), ku-chuila " to spit " (Tonga ku-suitd), etc. r. r -^-t- KAMBA. (Spoken west frOm Mombasa to Mount Kenia.) 81, — 1° Not only d, but also /, s, and/, are generally dropped in Kamba or weakened, this, with other contraction's, causing many stems to be reduced to very" short forms. Tonga Kamba Tonga ku-gula, to buy kuua kii-ulu, a foot ku-u murbili, the body mu-i kujayu, to kill ku-aa mu-ezi, the moon mu-e ku-boko, an arm k-oko 82. — 2° The Tonga s is sounded t,in Kamba. Last says in his Polyglotta, p. 3, that this sound Is similar to that of th in this, that. Ex. Tonga bu-sio, face bu-su, flour Kamba u'^io mu-tiu Tonga Kamba in-soni, shame n-'ionni kiiseka, to \2i\x^. ku-tieka 20 South-African Bantu Languages. 83. — 3° Among the changes produced by the nasal n on a following consonant, there is not only that of / to d as in Tonga and most of the other dialects, and that of w and v to b\ but also that of t to z. Ex. n-lembua, guttapercha, pi. n-dembua u-wau, a side » in-bau . u-iuka, night » «-^«/4fl(cf. Tonga to-«/J?< " night ", and n. 51) ' • SWAHILI. ,, 84. — This is said to be the most arabized of all the Bantu lan- guages. However this assertion, though probably correct on the whole, might lead to false conceptions. For, arabized as it is, Swahili remains without some Semitic features which are noticeable in several Bantu languages. Thus it has no article, and it has many words beginning with vowels. Again, Swahili proper, when not spoken by a man who knows Arabic, rejects hiatus less than several other Bantu languages. Those Arabic guttural sounds which are heard in a limited number of these same languages have not pene- trated into Swahili proper, etc., etc.. 85. — But Swahili is arabized in this sense that Arabic words often intrude bluntly into it, without even putting on a Bantu dress. Thus, in a single tale of 9 lines, the first of Steere's and Rankin's Swahili and Makua tales, I niotice no less than 7 words wshich have 110 Bantu color at all, viz. ilmu, doctrine; hcisira, anger; hatta, until; sababu,z^.\x=,Q: ; killa, each ; -rudi, to return; shekh, a chief. 86. — And again. Arabic influence must probably be seen in some of the following pecularities : — ', ■ 1° The classifying elements of those words which are in most frequent use (cf 42) are much weakened, by elisions and contractions, some of them being reduced to mere nasals, others being dropped altogether. Examples may be seen further throughout the whole of the fchaptcr on substantives.. Here are some others : — Ex. Tonga Swahili VOM-nzi u-« xa\X-ame, the king's residence, m-// w-a Ta-falme. i-zina li-« mu etke, the name of a slave, jina la m-fui/iua. hX-limi hxe iayati, the tongue of a buffalo, \x-liini vra n-yati. Comparative Phonetics, 21 87. — 2° Though Swahili has many words beginning with vowels, it prefixes /? to many others as if purposely to avoid be- ginning with them, or, more probably, to change them into perfect dissyllables. Thus the Tonga demonstrative pronouns oyu, eli, eci, etc., are in Swahili hiyu, hili,hicM, eXc. 88. — 3° Swahili drops the Tonga /, though not so often as Kamba(cf. 8i). Ex. Tonga ■ SwAHlll Tonga Swahil ku-ziala, to beget kurzqa in-zila,. a. road n-jia ^ in-zala, hunger -vjaa ku-Hla, to cry ku-lia ku-ulu, a foot m-guu cf. ku-lala, lo lie down ku-Iala 89. — 4° Thepe are some other i-emarkable phonetic differences between Swahili and the generality of the Bantu languages, . but general laws cannot be laid down. Ex. Tonga -onse, all li-nso, an eye -jika, to cook Swahili -ote (812) . jircho ox ji-to (413) -pika (52 examples) (Cf. jiko = fire-place) Tonga - Swahili mu-kazi, {eraa\e,m-ke (cf. 81) muse, earth ti-M or ii-ii ma-nzi, water via-ji (440 ex.) ■sano, five , -tdno (792) 90. — N.B. I. Some of these examples show at least traces of permutation between s aiid t or ch (cf. 63). Such double fornis iLSJicho ox ji/o, nchi or nti, properly belong to different dialects, viz. —jicKo and itchi belong to the dialect of Zaiizibar, while /zVo and nii belong to that of Mombasa., 91 . — 2. Likewise it may be added that z is less in favour in Swahili than in Torga as the former replaces in many worjis the z of the latter by v or j. The same may be said of many other languages (of. 63); NYIKA AND POKOMO. 92. — ^- B. Nyika is spoken inland from Mombasa, and Pokomo on the banks of the Pokomo river. Unfortunately, nearly all that we know on these languages has come to us through Germans who seem to have mistaken in many instances hard for soft con- sonants, and vice versa, for instance, /for v, v for/, s for z, z for s, etc.. 93. — I" These two languages, though differing considerably frpm each other, have this remarkable feature in common that they have generally- the consonant k where the main group of the Bantu languages has a t. However this letter reappears regularly, accord- ing ton. 51, under the influence of?/, expressed or dropped. 22 South-African Bantu Languages. :. Tonga ku-tuma, to send inyati. a buffalo ma-fuia, fat hu-aio, a canoe ku-leta. to bring up, mu-niu, a person -iatu. three NiKA ku-huma nyahi ma-fuha w-aho ku-reha mu-iu POKOMO kurhuma nyahi ma-fuha wahv hu-yeha piu-niu \ ■hahu (without ri) -hahu (without n) -tahu (with «) iahu (with n) 94. — 2° They have also this in common that, hke Senna (99) they have the compound sound dz (alias ds) where Tonga has the simple sourid z. Pokomo has also in common with Senna the compound sound bv (aHas hf) and probably// In Nika the sound vu (alipsyii;), and perhaps in Pokomo the sound bvu, hecotne/>/u under the influence of nasals. Under the same influence. theNika sound m seems to become .only pi. :. Tonga Nika Pokomo Senna kit-za. to come ku-dza ku-dza ku-dza ku-zala, to become full ku-dzala ku-dzaa ku-dzala i-zuba, il^e sun dzua dzua dzua -fiiide, dead -vti (without n) ? . -fu (with n) ^ ? yci.pfubafihan^ ) {oTpfupa.) ■pia. new f -via (without nasal) -bfia \-bvia \ -pia (with' nasal) -bfia [-pfa (?)] 95. 3° As many other languages, Nika and Pokomo drop out or weaken the consonant b, when it is not preceded by m[d. n. 52 examples), but this letter reappears regularly under the influence of nasals. Tonga Nika Pokomo -/(without nasal) -wi -mbi (with nasal) -mbi (?) Ex ■bi, bad 96. — 4° Pokqmo differs from Nika principally in this, that, like Kamba and Swahili, it weakens the / or r of the other languages in many cases, and drops it in others. Pokomo Kamba Swahili n-dzaa n-sa (n-dzaa?) n-jaa ku-wyaa ku-dzia(kudziaa1 ) ku-zaa mu-i mu-i mw-ili ku-ye ... ku-le mu-yango . . . vrw-ango ' muyume tn-ume , m-ume mu-tyiu ... m-kia Ex. Tonga Nika in-zala, hunger n-dzala kuziala, to beget kuiiala mti-bili, the body mu-iri ku-le or kii- re far ku re mu-liango, doorway mu-riango mu-ahtme, a husband mu-lume mu-cila. a tail mii-chira Comparative Phonetics, 23 97. — N. B. The Gunda language, which is a mixture of that of Senna and of that of Kilimane, is, as it were, the connecting hnk between Senna and Pokomo. For, like the latter language, it drops / in many words, and changes it to j* in several others, as in be, a woman's breast (Tonga i-bele), n-sia, a path (Tonga n-zilci), n-tayd, hunger (Tonga n-zala), ku-iaya, to be (Tonga ku-kald), etc. SENNA (including TETTE and NYASSA). 98. — N. B. This language, though known to the Portuguese as the " Kafreal de Senna, " is not so well spoken at Senna itself as at Tette and in the neighbourhood of the Nyassa Lake, this being probably a result of the greater contact of the natives with Eu- ropeans at Senna than in those other places. It is considered by the natives of the Lower Zambezi as being much more primitive than the language of Kilimane and far superior to it. Rebman also speaks of its Nyassa dialect in the following enthusiastic terms : " My study of the Ki-niassa was to me a continual feast. ... No sooner had I got ah insight" into it, than the dialects with which I had previously made myself more or less acquainted, appeared to me rather as so many rays of one and the same light ('). " However, lest Reb- man's enthusiasm should convey a false notion to the reader, it should be remembered that his terms of comparison were principally coast languages, viz. Swahili, Kamba and the various Nika dialects, all of which have certainly undergone more foreign influence than Senna. 99. — The most prominent phonetic feature of this language as compared with the others is that, where most of these have a sharp zor V orf, it has, in many instances, compound sounds, some entirely labial, others entirely dental, others partly labial and partly dental, variously pronounced in the various dialects. Most of these com- pound sounds are the result of a suppressed z or a suppressed nasal. Ex. Tonga Tette Senna Nyassa Nyassa (Blantyre spelling) (Rebman's spelling) ' i-fua, a bone mu-nvui, an arrow im-vwuu, a hippopotamus ku-zuata, to put on dress ku-zala, to be filled i-zina, a name zi-kowe, eyelids 100. — As in Nika, Taita, Swahili, and several other languages, the Tonga b is generally weakened or dropped altogether in Senna, as in ku-ona, to see {^ov.^2, ku-bona), dzua, the sun (Tonga i-zuba), ku-ziwa, , to know (Tonga ku-zibd), a-ntu, people (Tonga ba-ntu). 101. — In Senna the classifier MU of the classes MU-BA and MU-MIis reduced to N, though not before monosyllables nor before labial sounds (cf. 323 and 367). I. Dictionary of the Kiniassa Language, 1877, p. vii. pfupa (?) pfupa fupa pfuba mu-bvi viu-bvi mubvi mu-pfi mbvu m-bu ? in-pfu ku-bvara(?) ku-bvala ku-bvara ku-pfara ku-dzala ku-dzala * ku-dzala ku-dsara dzina dzina dzina dsina bzi-kope b}i-koJ>e psi-kope 24 South-Africmi Bantu Lafignages. , N. B. However, it must be noted that Rebman in his Ki-nyassa dictionary reduces it only to M. But it may be that in this he is no more reliable than in spelling the above examples mu-pfi, m-fifu, ktt-pfara, etc., whereas the correct spelling is probably mu-bvi, m-bvti, ku-bvara, etc.. ;>•,, ' 102. — Tetle and Nyassa are not the only varieties of the lan- guage of Senna, Others are that of Zumbo, the Mbara language of tlie Loangwe, and even the dialect of Sofala which is described in Bleek's " Languages of Mozambique ". 103. — We may probably add to these the Gindo language, very little of which is known. Dr. Steere, who supplies a short vocabulary of it, says that " the Gindos are a tribe lying between the coast Swahili from near the north of Monha to Kilwa ". Thus it may be seen that the Senna language is one of the most extensively spoken in South Africa. KARANGA {a/ms Kalaka). .104. — This, the language of the famous Monomotapa empire, is, on the whole, closely related to Senna. In fact, the three principal features of Senna, which have just been mentioned, are also features of Karanga, though the applications are somewhat different. How- ever, on the other hand, Karanga has several remarkable features which distinguish it plainly from Senna, so that it deserves to be treated as a separate language. Hence : — 105. — 1° Double consonants of a peculiar kind are met with in Karanga as in Sertna, but with some variety of pronunciation. Hepce they are written S'^, ■^w,ps, (cf. Alphabet). To these may be added/ where Senna, has dz. Ex. Tonga Karanga ku-sika, to arrive u-%wika zi-niu, things ^wi-ntu ku-ziala. to beget u-^7ciara ■pia new -pm ku-za, to come ii-ja (cf. Senna ku-dzd) 106. — 2° We hear in Karanga the sounds/ (French/) and -r (English sh), unknown in most Bantu languages. We hear also two peculiar guttural sounds, viz. g and k (cf. 12 and 16). Ex. Tonga • Karanga Tonga ku-samba. to wash u-xamba ci-fua. i-saku, a devil xaku mu-btia, Leza, God Reja , bu-tzu, a - -like. mu-sozi, a tear nii'XOjt Comparative Phonetics. 25 Karanga the breast fom a dog im-buna grass wu-kua itself -^8« 107 -3° Not only is the classifier MU of the classes MU-B A and MU-MI generally reduced to N as in Senna, (orto UN, when the word is not isolated), hut also that classifier which in the other languages is N or IN, is in Karanga reduced to I, as in Mozam- bique (385), and the classifier which in the other languages has the form KU, is in Karanga reduced to U, as also happens in Mozam- bique (cf. 175). w Tr^^Tr^/ Karanga. ^''- V isolated form : mi-kaji {?>tr^r,^ un-kazi) mu-kazi, a woman | connected form : n-haji' (Senna n-kazi) im-mla, rain i-vura (Mozambique i-fuld^ ^ ■ ku-iata to seize ri-pata{yioz?>xM(\ae. u-vara) 108. — 4° Karanga is full of contractions and elisions which are unknown in Tonga, and such that it cannot be termed an agglutin- ative language. This renders its study far more difficult than that of Tonga which is, on the whole, much more purely agglutinative. . Ex. Tonga Karanga Leza u-a-ka-tuma m-vula, God sent rain. I Reja-ka-lume-vura ( = Reja u-a-ka-tuma i-vura). U-a-lia mu-tuma mu mu-lilo u-ta-mani, U-a-kb-n-tuma viu motb-si-nppera ( = U- he sent him to the fire without end. I a-kamu-tuma mu mu-oto u-si-nakupera). 109. — In Livingstone's Mss. Coniparative Vocabulary pre- viously mentioned, there is a vocabulary of Yeye, or the language of Lake Ngami and the River Zouga. (Cf. Livingstones Miss. Trav., pp. 63-72). There can be no doubt that it is a variety of the Karanga language. 110. — The language of Mashonaland is also a dialec^ of Karanga. Perhaps it is one step nearer to S.enna than Karanga proper ('). ' I. The word Mashonaland, which has come so often before the public during these last months, is rather interesting. Shona is nothing else than the Karanga pronunciation of the word Senna fSyona or Si-yuna of the ancient Arab geographers). Hence the word Ma-shona is properly the name of the ancient " Senna Rivers " (30, note), which included large tracts of country South of the Lower Zambezi. As a great portion of this country is called Ma-nica, and the Senna language closely ressembles Nika (99-100), should we not identify the ancient Manica gold-diggers with the Manica of Mombasa, and both tiXia.'iht Amalika of ancient Arabia ? Concerning the Syona or Siyuna of the Arabs, cf " Giographie d'Abulfeia", traduiteparM. Reinaud, tome Il.i'epartie, Paris,i848, p. ?o8,and " Giographie d'Edrisi'\ traduitc par P. AmiJd^e Jaubert, Paris, 1836, tome I, p. 6§. 26 South-African Bantu Languages. GANDA. (Spoken on the shores of the Victoria Nyanza.) 111. — If Ganda be compared to the languages which have been reviewed before this, the first thing which will strike us is the repeated use of the vowel article a, e, or o, before substantives, and of the conjunction na "and ", before those verbs which are in a historic tense. The use of these particles in Ganda points perhaps to Semitic infliience. In any case, it is more^onetic than gram- matical. For such particles seem to be heard exclusively after a pause, long or short, as if to introduce verbs and substantives more gently. '.' i;l"2. — Ex. Tonga Ganda O Bu-ganda buno nm-bu-lie. (Grammaire Ruganda, p. 83.) Na.-ci-sal>aJa, lit. afidhe entered the boat, na.-a-wunguka, and he crossed, ndi-a-iuka mu kialo,andhQ came to town. (Mat. 9, I.) ' ^'■■ 113. — Among other features of Ganda we may notice : — 1° A phonetic insertion oi g, sometimes b, between vowels, as if to avoid a hiatus. Tonga . • Ganda Bu-ganda obu mu-bulie, lit. this Ganda realm, eat it. U-a-njila mu i52)a-kaii, in the middle ba-kaci (0 mu-tinia, heart, in Herero) mo-cima kutta, to call loud ko-izana -tatu, three -atu in-kuku, a hen moku inkulu, an old man n^ulu imi-ntu, a person, mo-nu suiiiue, a hyaena zoTido (?) m-ganga,a. doctor 11^^1:1,0^ im-pongo, a goat mboii^o 14-1 . — : N- B. r. The sound which we spell «j is spelt by Livingstone variously : n, fik, fig. Sometimes the very same word occurs in Livingstone's manuscripts with all three different modes of spelling. 2. We cannot warrant the correctness of the vowels in all the examples given in this work for Rotse. Livingstone who is our only authority and who was principally a Chwana scholar, does not seem to have cared much for the differences between o and «, e and i (cf n. 200). Likewise we cannot certify that y in some of preceding words is pronounced like y myear. Possibly Livingstone meant to express by it the sound of the French/(our^, 15). 3. At the end of Capello and Ivens' " From Benguellato the ienilory of Yacca ", there is a short collection of words which are said to represent the Ca-luiana language. As the Rotse call themselves Ba-loi or Ba-lteiait't, we should expect these to be Rotse words, but they are not so, or, if they are, we must say that they are considerably metamorphosed. The authors, say that probably they belong to the Kololo language. Certainly they belong to nothing of the kind. But what approaches to Kololo are some twenty words given in the same work under the heading of " Njenji "- Concerning Kololo, see n. 169. 142. — Next to Rotse is the Nyengo language, which is described in Livingstone's Comparative Vocabulary MSS. It is spoken on the River Nyengo, which is an affluent of the Upper Zambezi, Ex. 1°) mo-kathi, a jpp. 220-233. Ex. Tonga Lunda KOELLK Carvalho Capello and Ivens Luba kti-tui, an ear di-dsh di-tui di-to di-chu i-fua, a bone di-fup di-fupa mu-kazi, a woman mu-Uaje mu-kaje mu-kax mu-zike, a slave mo-ror mu-roro i-bele, woman's breast ..di-yel di-yele chi-adi im-vula, rain um-fal lu-nvula lu-limi,. the tongue ar-dim lu-dimi lu-dimi ma-nzt, water menyi meme meii kabua, a little dog ka-b ka-bua ka-bo li-no, (Kafir i-zi-nyo), a tooth di-zeu di-zeu dino li-mo, an eye di-z dice 144, — 2° There is also every appearance that the Tonga zi is sounded y? or ci in Lunda. Ex. Tonga Lunda mu-lozi, a wizard viulaji (?) or u-roichi mu-kazi, a woman mu-kaje N. B. Possibly also the Tonga sound tu- before a vowel is, in Lunda, changed to ish or dsh, as in di-dsh (K), an ear, ffonga ku-tui); umo-dsh (K), a head, (Tonga di-dsh), etc.. Cf, parallel changes in Chwana and Kafir for the sounds bu, pu, mu, etc., before vowels, nn. 122 and 202-207. RUA. (Spoken on the Lualaba, South of Nyangwe). 145. — If we may rely upon Mr. Last's collections in his " Poly- gloita Africana ", the most remarkable phonetic feature of Rua is the transition from LI to J I. There is however no trace of this in Cameron's Rua vocabulary at the end of his "Across Africa ". 3 34 South-African Bantu, Languages. Ex. Tonga Rua Last Cameron -bill, two -biji -wilt lu limi, tongue lu-jimi lu-vimi li-no, tooth ji-no li-no N. B. Guha, which is spoken West of Lake Tanganyika, is closely allied to Rua. How- ever it shows no trace of the transition from LI to JL ANGOLA, MBAMBA, and FIOTE or LOWER CONGO. 146. — In many respects these languages differ considerably from each other, but they practically agree in most of the points in which they differ from Tonga. The only regular permutations of consonants which are worth notice in them are the following : — 147- — I" The Tonga z before zand g = generally/ (French /j in the three of them. The Tonga s before i and e = generally ;ir_(EngHsh sk). Ex. Tonga Angola Mbamba CONGC mu-nzi, village mu-iji, family mu-iji, family -zima, put out a in^zila, road light ■jima n-jila ? ? -Jtma jila (l)izina, name bu-si, smoke rt-jtna ri-xi ? mu-ixi t ejtna mw-ixi ■muse, earth 'oxi ? n-xi Mbamba Congo -hia ■ via ki-f-Ua ? (o)va-nxi -ha vana 148. — 2° The Tonga p (both expressed and suppressed or weakened) = Angola b = Mbamba k (or d suppressed) = Congo v (sometimes m). Ex. Tonga Angola -/w, a) burn; b) new -bia ci-fu(w)a, bone, chest ki-fuba, bone (p)a-nsi, down b'o xi •pa, give -ba 149. — ' N. B. In Congo, those stems which have generally v where Tonga has p recover this consonant after nasals. Ex. m-pemo '' wind " (Tonga im-pewo). 150. — 3° Tongas = Angola g (?) or >^ = Congo k (cf. 258). Ex. Tonga Angola Congo /e;/-(r/V?, dawn of day ku-^a kukia mucila, tail inukila n-kila 151. — Though agreeing thus in many points, the language of Angola and that of Lower Congo seem to differ considerably on some others. Thus : — Comparative Phonetics. 35 I" In Angola, n or vi is dropped before J, x,p. '^oX. so in Congo. On this point Mbamba is apparently like Angola. The same phe- nomenon takes place in Swahili, Sagara, etc., cf. 282, 283. Ex. Tonga Angola Mbamba Congo in-soni, shame sonye ? n-soni ku-nsi, below koxi ? . ku-nxi in-ziba (?), swallow piapia, (of class IN) pieha (of class IN?) venga-m-punza lo2. — ■A'; B. I find there are in Angola a few words in which the Tonga n is replaced by i before a consonant. Ex. ku-ivua " to hear " (Tonga ku-nvua), mii-iji " a family " (Tonga mu-nzi), etc.. 153. — 2° In Congo, the classifier MU of the classes MU-BA and MU-MI is generally reduced to N (M before labials), as in Karanga and Senna (107, loi). Ex. Tonga Congo Karanga mu-kulu, elder n-kulu (u)n-kuru mu-samo, a tree n-ti (u)n-ti mu-cila, a tail n-kila (u)n-cira 154. — 3° Congo weakens also, or drops, the classifiers BU and KU of the classes BU-MA and KU-MA. Cf. 465 and 447-450. N. B. In this again, Congo reminds one of Karanga. Are these merely accidental con- nexions between the principal language of the ancient Congo kingdom and that of ancient Monomotapa? 155. — 4° In general, in the classification of nouns, Congo recedes further from Tonga and from the generality of the Bantu languages than Angola does, as will appear from the chapter on substantives. Mbamba seems to be nearer to Tonga than either Angola or Congo. 156. — A'; B. I. The Congo dialects which are described in the old Grammar of the Capuchin Father Brusciotto a Vetralla and in the Mss. French-Congo Dictionary in the British Museum, were more perfect than the modern San-Salvador dialect described in Rev. W. Holman Bentley's" Vt'c/ionary and Grammar of the Congo Language " (Lon- don, 1887). ■ 157. — 2. The Bangala language, of which Mr. H^li Chatelain has given us speci- mens in the " Zeitschrift filr afrikanische Sprachen ", 1888- 1889, pp. 136-146, is probably the same as that which is called Kasands or KasdndsM in Koelle's Polyglotta. It differs but little from Mbamba. 158. — 3- The old Angola dialect, which has been preserved to us in the Grammar of Father Pedro Diaz, S. J., and in the catechism of Father de Coucto, S. J., had fewer contractions and was consequently nearer to Tonga than the modern dialect. 36 South-African Bantu Languages: MIDDLE CONGO LANGUAGES. 159. — H. H. Johnston, in his "Journey up the River Congo ", gives us precious, though short, vocabularies of three languages of Middle Congo, viz. Teke, Buma and Yansi. They are sufficient to show that these languages differ considerably from one another, comparatively speaking, and yet perhaps more from any other known Bantu language. But they are neither sufficiently accurate, nor complete enough, to allow us to bring out any of their phonetic features with certainty. 160. — N. B. A few words in Buma and Teke have the consonant r where Tonga has /. This, as we shall see further on, is characteristic of the. Mozambique-Chwana- Mpongwe group of languages. The Buma language in particular has certainly a great deal in common with Mozambique. 161. — Here are, for the sake of comparison, a few of the words in which these languages agree best with Tonga, and consequently with the main Bantu group : — Ex. Tonga ma-bele, woman's breast hu-ato, a canoe mu-ana, a child muse, country kufua, death mu-bua, a dog in-goma, a drum kuiui, an ear ma-tui, the ears mu-tue, the head i-ji, an egg li-nso, an eye in-sui, a fish im-vuvu, a hippopotamus tn-vubu in-ganga, a doctor ma-fua, a skeleton -bi, bad ■tatu, three 162. — Here are also a few words in which, as far as we may rely on these small vocabularies, these languages differ widely from the main Bantu group. Ex. Tonga Teke mulilo (Senna m-oio), fire m-ba or vt-baa mpongo (Swahili m-buzi), a goat n-taba in-gubo, cloth zo-iho in-zoka, a snake n-tare i-zuba, the sun ma-iere Teke Buma Yansi ma-biela ma-biela ma-biela bw-atu bw-aro bw-engo mw-ana mw-ana mw-ana n-chi ki-se n-chi a-Xthedead?) saa-fwa (?) a-guii^e. dead.?) vi-bwa m-bwa m-bwa n-goma ?i-goma n-goma i-iui i-tui ma-chui {li^^ mu-chwi viu-tu mu-tu i-ke i-ke i-shu di-u li-shuu n-chwi n-tu n-cluti m-vubu n-gubu n-ga. n-ga. ma-fwa U-pfuba (99) ■M -bi -bi -taiu -saru -satu Buma Yansi m bo me-a or vie-ya n-taba n-taba ki-piu bi-lamba 7n-pili mu-shwema i-tere n-dembe Comparative Phonetics. 37 NYWEMA. (Spoken North of the River Lukuga). 163. — The materials available regarding the language of the cannibal Nywema are not yet sufficient to allow us to pass a judg- ment on the features proper to it. However it may already be said that it has much in common with the language of the Bihe, while, in some respects, it reminds one more of Mpongwe (cf. 213 and sqq.). 164. — 1° The classifiers of the class MU-MI (366), are reduced in Nywema \.o ox u in the singular (Mpongwe 0, Bihe 11), and to e or i in the plural (Mpongwe z, Bihe vi). Ex. Tonga Nvwema Mpongwe Bihe Bamba Kusu DIALECT DIALECT mu-lomo, the mouth 0-lomo u-lomo, o-lumbu mi-lomo, mouths e-lomo e-lo?no i-lumbu mu-tue, the head o-twe o-twe u-twi mi-tue, heads e-twe e-twe ... vi-twi \iiiuti (Senna), a tree] o-ti o-ti o-iindi u-ti 165. — 2° The Tonga 2 before z is replaced in Nywema, at least in some words, by /, (Mpongwe / or n, Bihe / expressed or sup- pressed, 131). Ex. Tonga NVWEMi W-eli Mpongwe mu-ezi, the moon o-gweli mu-kazi, a female w-ali im-buzi, a kind of goat m-buli m-buni Bihe u-kai 166. — 3° The sound which in Nywema is v when not influenced by a nasal, changes to p when influenced by one, as in Mpongwe, Congo, and several other languages. Ex. lu-vita, a finger, pi. pita (= m-pita, of class LU-IN). 167. — N-B. I. Nywema differs from Mpongwe, among other things, by not having r where Mpongwe has it for the Toriga t (cf. 214). Otherwise the word for " head " should not be in Nywema o-lwe, but o-rue; that for " belly " should not be o-tima, but o-rima, etc. 168. — 2. All these conclusions concerning Nywema are drawn from Last's pre- cious collections in his " PolyglottaAfricana ", pp. 183-187 and 232-233. Mr. Stanley's col- lections in his " Dark continent " would lead to different conclusions. Probably they represent different dialects from those which have been studied by Last. % 38 South-African Bantu Languages. KUA OR MOZAMBIQUE, and CHWANA. 169. — The association of Chwana with the language of Mozam- bique may appear astonishing on account of its novelty. The fact is that we are passing to a class of languages which differ on im- portant points from those reviewed until now, and that, precisely where such differences occur, these languages happen to have similar features. This part of our study is particularly interesting, because after having passed from Mozambique and the Comoro islands to Basutoland and the Kalahari, thus touching the very southernmost parts of Africa, we find ourselves obliged to retrace our steps towards Kilimanjaro, then to pass over to the Ogowe under the equator, across the whole African continent. N. B. Nearly everything that will be said on Chwana in this article is true not only of Chwana proper, but also of its numeroiis dialects (Suto, Tlhaping, Kololo, etc.). 170. — To understand the language of Mozambique and Chwana, it is necessary to distinguish with a very peculiar attention between those consonantal sounds which include a nasal and those which contain none. Hence : — 171. — 1° Considering those sounds which contain no nasal, we have to notice a set of permutations which difTers considerably from most of what we have seen until now. The correspondence of?' and t is particularly remarkable. The general tendency is to guttural sounds. < o iz; O H JLi < < u Ex. Tonga Mozambique Chwana Masasi Kilimane r,rr = r -tatu, three -raru -raro -raru )* a " n ku-jata, to hold u-vara u-varra gochwara ?» )» >' >) ku-tuma, to send u-ruma -u-rruma go-roma s,r (l)i-tama, a cheek n-rama tn-rra?na le-sama pi. ma-rama X i r, rr = d,r mu kazi, a wife mw-ari mu-arriiJ)mo-iadi (or (isl. of Moz.) mo-sari,c{.g) )J )* 1) ma-nguzu, strength i-kuru It a J» )) inu-ezi, moon mw-eri mu-erre kguiedi (or kgiceri) J) r,/ „ )) indezu, beard t'^reru e-rrelo tedu ' (isl, of Moz.) Comparative Phonetics. 39 < o o zt ' o ch, z y, b < U Ex. Tonga Mozambique Chwana Masasi Kilimane „ dropped (l)i-zina, a name nchina ,, // ku zala, to become full u-chara ,, ts ' ku ziala, to beget u-yara zu cw // ku-samba, to wash mu sisi, hair ku-sia, to leave behind „ „ „ „ „ bu-siku, night „ ,, X „ s ku-busia, to rouse „ „ ihi, ti „ dropped ku-simba, to dig se „ the, te „ tshe kuseka, to laugh me-to ?n-tava u-hapa n-hia u-hiu ii-wuxa u-tfiipa u-thea ni-zina^) le-ina go-tlala u-bala gotsala (Suto tsuala) gO'Cwa ma- to via- it I ho {KaRr a 7/ieh/o J mu-tana mo-thlana (Kafir u m-hlana) uhaba go-tlhapa (Kafir u-ku-hlamba \ 175.— K =dropped= g "" )j )) )) )> ,. ,, k ,, sh -ako, thine in-zoka, a snake ku-buka, to awake i-kutni, ten .' „ „ ., „ „ ku-kuwa, to shout 176.— C = dropped= .s- ci-ntu, a thing „ „ "„ „ ,.. mucila, tail „ „ „ „ „ cia-ku Ha, iooA 177.— F = /4 ^ sh ku-fua, to die „ „ „ „ /, h ma-futa, oil, fat 17§.— LI =: li,j^= J ku-lia, to eat „ „ / „ „ mu-Uango, door-way 179. — B ==dropped= b ku-bona, to see . „ „ ... „ c,b hu- boko, axi arm ■ao i-noa w-uwa 7ii-kuini u-kuwa i-tu mw-ila y-olia (= i-a- u-kwa ma-kura u-lia- m-lako iv-ona. n-tite II- tic u-timba u-tea ■au noa u-uwa kume u-kuwa e-lo- mu-ila u-lid) go-sta bo-sigo go-cosa go-epa go-tshega -ago noga go-coga le-shome goshua se-lo se-jo u-kua gO'Shwa ma-kurra ma-fura(pxma- hurazi. ii.) u-oja bi ku-buka, to awake (intr.) u-uwa ■■biala, to sow w-ala go-ja (20s) mo-jako go-bona le-cogo, pi. ma- bogo (414). go-coga . gojala (202) 40 South-African Bantu Languages. 180. < o ;? o H P = N O < s u Ex. Tonga Mozambique MaSASI KiLlMANE =/(i i) mu-{ni{=mu-ptni),h2inA\t ni'ini (pi. mi-vini) ,1 I) . (p)a-kati, between vari v-ari ,, ... i-fua(^i-fupa),hoiaQ ni-kuva » u (p)a-nsi, down ^a-/,^? ?'«-// Chwana liw-finii fa-gare lefatshe, earth 181. iV. B. I. This last permutation, viz. ^ = i/ ==/, should be compared with what has been noticed in Congo (148, 149), Rotse (137), etc. 182. — 2. The fact of 3 being suppressed, as in Mozambique, though mentioned more particularly in this place, is common to many other Bantu languages, as may have been remarked throughout the whole of this article. Cf, class BU-MA, in the chapter on substantives. 183. — 2° Considering those consonantal sounds which contain a nasal, we meet here with an entirely new application of the gene- ral principles mentioned innn. 55-59, viz. — the nasal is apparently suppressed, except before monosyllabic stems, and then, in Chwana, the consonant which remains is either hardened or strengthened^ or, if possible, dentalized, while in most of the dialects of Mozambique there is a marked tendency to the same effect. Hence : — 184. — o o H Ex. Tonga o MP= p = 185.— p,d = MB= 186.- la < K ' U pA im-pewo, wind p kubumba, to form ,, imbuzi, goais ,, im-bezu, seeds Mozambique Chwana NVorMV=/=/ „ „iw„ tl 187.— in-vula, rain mi-nvui, arrows ku-nvua, to hear Masasi i-pio, w-upa e-puri m-beyu (?) i-pula w-iwa KlLIMANE pevo u-uba bure beu NF= 188.— NK= k,,= 189.— NG= «o = 190.— NG= k = kh kg in-zila n-fuefui, a short . . . [road tn-kuku, a hen (z)in-kuni, firewood i-kuni -ku phefo go-bopa podi peu pula me-ctit go-utlwa tsela e khutsham kgogu kg in-^ombe, a cow i-n^ope di-kgon^ kzomo in-gonuf, drum i-koma mu-liatigo, the doorway m-lako t-gomai}) koma{—yfa.T song) mo-jako Comparative Phonetics. 41 < o Is o H o < a u Ex. Tonga Mozambique Chwana k, is ku-njila, to enter 191. NJ= /6 192.— NT= t ==^ th mu-ntu, a person ,, ,, „ „ „ m-iibt, a shield 193.— ND= A th Masasi u-kena tn-tu KiLlMANE 5> U J> )1 >) 5> 3» >» )» )J mu-lindi, a pit ku-enda, to go ku-linda, to watch over ii-lita, n-liti, w-etha. 194. NS= /, M = /M -(?««, all „ „ „ „ „ linso, an eye X h -othe ni-lho i-xoni 195.— NZ= th, d = X n 196.— NY= n = „ in-soni, shame „ in-sangu, a hoof, a shoe . . . „ mu-nsi, within m-hi-na tl (p)a-nze, outside vathe „ in-zala, hunger i-thala is in-zila, a road „ ma-nzt, water ina-xi n • in-zoka, a snake i-noa i-nyama, meat i-nama i-nyati, a buffalo inari I ... go-kena or gotsena mu-io mo-ihu thepe ino-lete u-endaQ) ^(7-i?/a(=to travel) go-leta -ote-ne -otlhe ni-to k-itlho di-tlhonti, (Kafir /'« tloni) tlhaku (Kafir i n-tlangu) moti-n mo-teni va-nje (?) kwa-ntle data tlala dila tsela ma-ije metse noa noga nyama nama narre nare 197. — ^- B. It is evident from this last permutation (ny^n) that the Mozam- ' bique word rioa, snake, and the Chwana noga interchange immediately, not with the Tonga form in-zoka, but with the Kafir form i nyoka. And in general it may be said of many words both in Chwana and in Mozambique, that they are in more immediate connexion with their Kafir than with their Tonga equivalents. 198. — This influence — open or concealed — of nasals upon other consonants, in Chwana and Mozambique, causes a great many words to have in these languages two forms each, these forms being sometimes widely different (cf. 52-59). Ex. Tonga Mozambique Chwana ba-ntu ba-iaXw, three persons a-tu a-raru ba-tha ba ba-ra.ro in-^ombe (zjin-tatu, three cows i-aope (di)taru di-kgomo ise (di)tdSVL 42 South-African Bantu Languages. N. B. I have not however sufficient evidence to trace with certainty to the influence of nasals the fact that verbs in Chwana adopt a stronger form after the reflexive pronoun / (655), as if « were suppressed. Possibly this fact might also be explained by saying that the vOwel z produces in given cases the same effect as n, as if z and w'were two cognate sounds in Bantu (cf. 152, 285, 412, 414). But this explanation does not seem to hold so well as the first in presence of the fact that the classifier DI of the class SE-DI (Tonga CI-ZI) does not cause the stems which follow it to adopt strong forms. (Compare n. 496 with n. 395.) 199. — Though Chwana and Mozambique agree very nearly in the remarkable features just mentioned, they can in no wise be considered as mere dialects of one and the same language. For they diverge in many other respects, principally in this, that, through contractions, elisions, and probably owing to European intercourse, the grammatical system of proclitics, enclitics, prefixes, and suffixes, is in Mozambique reduced to a mere skeleton, while its richness is extraordinary in Chwana. 200. — Again, Chwana, contrary to Mozambique, often changes to o the u of the other languages, and their i to e, as may be seen in the above examples. Likewise the syllable m at the end of Bantu words is regularly changed to ng in Chwana, though there is no evidence that the same is done in Mozambique (194). 201. — Again, a remarkable feature of Chwana, apparently not shared in by Mozambique, is a series of combinations of conson- ants and vowels which occur before such suffixes as begin with a vowel. They are for the most part similar to those which have been described in the note on Kafir (122-123) as affecting the con- sonants m. b, and/. A few others are new, affecting the consonants /, r, and ts. They are well described in Rev. William Crisp's " Secoana Gr. ", pp. 103-104, from which the following examples are drawn: — 202. — 1° Be- (vowel) =j- Ex. thebe, a shield; diminutive thejana, a small shield. Bo- (vowel) '=jw- or;- Ex. -tlhaba, pierce; passive voice -thlaju-a, be pierced. 2QQ bo-gobe]a,7iieio='{bogobe\io-aLtneho-o),t\\\s\smyh\ez.d. 2° Po- (vowel) =f«'- Ex. mo-lapo, a river; dimin. vio-lacwana. Fhe-{yOfie\)=chui- Ex. tshephe, a springbok; dimin. tshechwana. 204.— 5° Mo- (vowel):= imv- Ex, kgomo, nn ox; dimin. kgon■ ? (with n) naru ngombe (?) homo ngtvana (204) mtv-ana ba-thu (1^2) a-tu (di-tlhare)(\'j2) mi-iri me-tse (195) ma-xi •faru (172) -raru ) tharu (192) taru > kgomo (189) i-n%ope mw-ana unyana a-ntu a ba-nitt mi-ti i mi-ti ma-nzi ama-nzi -tatu ■tatu n%ombe i n-komo TSHAGGA AND HINZUA. 211. — N. B. Tshagga is one of the languages sppken near Kilima-njaro, Hinzua is one of those of the Comoro islands. The short specimens we have of these two languages are evi- dently insufificient to judge of their proper features. However they show plainly that both of them have some of the featufes of Mo- zambique, principally with respect to dental and liquid sounds. Ex. Tonga Tshagga Hinzua Mozambique T- - -tatu three -raru -taru [with «(?)] -raru {-taru with nasal mu-ti, a tree (Senna) mu-ri mw-iri mw-iri [influence) ma-tui, ears maru ma-ki-yo ma-ru mu-tue, the head mu-rue xi-tswa (cf. 206) mu-ru vm-tuniua, a servant m-nima(m-rumia 1) ka-rumia m-oto, fire (Senna) m-oro m-oro m-oro Z-^ mu-kazi, a female ma-nzi, water muali m-she ma-zdi (?) mw-ari ma-xi i-zuba, the sun e-ruva i-dzua n-chuwa K- --ako, thine -akue, his -aho -dhe -ao -awe P— ano (=pano), here vanu vano -ngai?(= -ngapi?),ho^ many?... -ngavi? -chanil or -ngavi ^ 212. — N. B. I. The Gweno language, of which Johnston gives us short specimens in his " Kilima-njaro Expedition ", is closely allied to Tshagga. 2. The short specimens of Angazidja which were published by Steere in 1869 represent a language of the Comoro Islands which seems to differ considerably from Hinzua. If these specimens may be relied upon, Angazidja is a mixture of Hinzua and Swahili. MPONGWE. (Spoken on the Lower Ogowe.) 213. — Strange as it may appear, it is none the less true that Mpongwe is more closely allied to Chwana and Mozambique than to the languages of the Lake region. For : — I" Here again the most noticeable permutations are from i to r, and from z to / (Chwana /, d, or r). Comparative Phonetics. 45 2° The influence of the nasal on consonants combined with it is in many respects similar to what has been noticed in Kua and Chwana, though it is to be noted that in Mpongwe, contrary to. what occurs in Chwana and in some Mozambique dialects, the nasal is retained before consonants in given cases with the effect of changing k, s, and / to iy, z to dy, etc. 3° In many words the vowels i and u are changed respectively to e and o, as in Chwana (200). 214. — Ex. Tonga Mpongwe Chwana Mozambique T -tatu, three -raro -raro -raru NT -ntatu, do. (class IN) -ntyaro -tharo -taru Z mu-ezi, the moon o-gweli khwedi NZ in-zovu, an elephant n-dyogu tlou ... S /-sue, we «-zue ro-na hi-jw NS in-soni, shame n-tyoni di-ilhon^ i-xoni B -bi, bad -we -shwe MP -mbi, do. (class IN) -mbe 91 j^j^ \ tn-kuku, a hen n-dyogoni kgogu / mu-altimi, a husband ( -lanfo, long onome (cf. mo-nona) (cf. m-amna) -la -lek ND ndanfo, do. (class IN) -nda -tekle 215. — It may be added that in Mpongwe, as in Mozambique, V and p correspond to each other as weak and strong letters, e. g. owaro oMolu, " a large canoe, " nyare ni-'^olu, " a large ox. " On this particular point, Mpongwe resembles the language of Lower Congo (cf 149). - - 216. — A remarkable feature of Mpongwe, in the same line as those just described, and noticeable principally in verbs, is that these have double forms such as -ioma and -roma, " send " (Tonga -tuina, Chwana -toma and -roma), -dyonga and -yonga, " drink " (Tonga nyua), etc. Probably the more dental, or stronger, of such forms is due to an occult influence of the nasal. 217. -- A phonetic feature proper to Mpongwe Ts the use of the consonant nY'^^/zai' nl) in many instances in which most other lan- guages have /. Ex. o-Uome, "a husband" {^ong?. mu-alumi), -baXit, " two " (Herero -6a7'-z, Tonga -dz/z), i-weXle, " a woman's breast " (Tonga i-bele). 46 South-African Bantu Languages. 218. — N. B. The principal feature of Mpongwe, as compared with the other Bantu languages, is the partial obliteration and disappearance from it both of the classi- fying elements of nouns, and of the connective elements of other words, viz. those elements which refer verbs, adjectives, and pronouns to their proper noun (cf. 42). However it must be added that the richness of Mpongwe is saved by the introduction of a great many constructions apparently foreign to Bantu. The practical consequence of such a fact with respect to this work is that less will be said on Mpongwe than on the other great Bantu languages, because our aim is not so much to dwell on the features proper to particular languages as to bring out those that are propei; to the main group. DUALLA. 219. — Dualla, the principal language of the Cameroons, has a great deal in common with Mpongwe, or scarcely differing from it. Thus : — 99n j^ (The Tonga t not preceded by n (=Mpongwe ;')=Dualla /. (The Tonga z before i (= Mpongwe / or n) = » d. Ex. :ToNGA Dualla Mpongwe Tonga Dualla Mpongwe Im-aio, c&noQ bolo ow-aro ina-futa, fat vi-ula (Chwana: ma-furd) -tatu, three -lalu -ratu ■tuma, send -loma -roma i-zina, name di?ia ina //«-tes/,a kind of goat vi-bodi m-boni /(?(73/, straight, good -lodi mi-sozi, tears misodi an-tyoni '2L'2L\. — 2° Dualla has, like Mpongwe, verbs with two forms, the one stronger, the other weaker (216). Ex.: Tonga Dualla Mpongwe Tonga Dualla Mpongwe ■buena(pxy -bona A iXtowgex -jene -dyena \ . f stronger -^/V/^i?a -dyingina (cf. 264) (weaker -ene -yena j ' 1 weaker : -/n^iSfl -yingina 222, — N. B. I. The change of the Tonga -buena ox -bona into the Dualla -jetu and the Mpongwe -dyena should be particularly noted, as it reveals another link which connects these languages with Chwana and Mozambique. Cf 202 where bo- before a vowel is found to change regularly to^ in most Chwana dialects. 2. Vowels are weakened in Dualla as in Mpongwe (213). 223. — Strange to say, if we consider Dualla from another point of view, we shall find that on the whole the Bantu grammatical elements are better preserved in it than in Mpongwe. Thus in par- ticular the classifiers of the classes MU-BA and MU-MI are not reduced to 0-A and O-I, as in Mpongwe, but they keep their consonants m and b. Ex. : Tonga Duali.a Mpongwe tnu-ntu, a person, pi. ba- tno-tu, pi. ba- o-ma mu-alumi, a husband, pi. ba^ m-nmt, pi. b'- o-nome, pi. a- mu-lomo, mouth, beak, pi. ;///- mo-lumbu, pi. ;;//- o-lumbu, pi. i- (Herero : mu-tima) heart mo-lema, pi. mi- o-rema, pi. /- Co}uparatil)e Phonetics. 47 224. — And, if we place ourselves in a third point of view, we may notice in Dualla a feature which reminds one of Swahili and Kamba of the East Coast, viz. /is often dropped (8i, 88), £x. : Tonga Dualla Swahili Kamba im-vula, rain *in-bua m-vua m-bua in-zila, a road n-gia n-jia nsia -lila, cry -eya -lia -iya 225. — Other consonants are dropped in some cases in Dualla, but apparently the laws cannot be generalized. Ex. : Tonga Dualla K i-kumi (= li-kumi), ten d'-uvi F ma-futa, fat m^-ula Tonga Dualla Z before a -iza, come -ya or -wa N nasal tnn ntu, a man ino-tu 226. — N. B, Saker says in his Dualla grammar, " tjiat the present Dualla are a very mixed people, greatly the result of the slave trade " Their language is indoubtedly quite as mixed, and consequently cannot be said to be a good repfesentative of pure Bantu. 227. — The same must be said of Benga, Isubu, and Kele, all three of which are languages closely allied to Dualla. Bengals spoken on the islands of Corisco Bay, Isubu north of the Dualla, and Kele principally along the Bembo River. 228. — The most remarkable phonetic difference between Benga and Dualla is the transition from s to h. Thus the Dualla w^ords sango " father ", duso " an eye ", bo-so " the face", esadu" small " are respectively pronounced in Benga hango, diho, boho, ehalt (Zeitschri/t, 1888-89, p. 195). 22Q, — Between Isubu and Dualla the most remarkable phonetic differences are the transition from p to /, and the use of k in many instances in which it is dropped in Dualla (Saker's Grammar, pp. 12 and 18). Thus the Dualla words mo-lopo " the head", m-boa '* 3l town ", ma-iya " blood ", mo-utu " a child ", etc., are respectively in Isubu mo- Iqfo, m-boka, ma-kia, mo-kutu, etc. 230. — Kele differs more from Dualla, Benga, and Isubu, than these latter differ from one another. Its most characteristic feature seems to be to weaken vowels more than any of the languages we have hitherto reviewed. Ex. Tonga Kele Dualla Mpongwe mu-nttt, a person mu-ty\ mo-tu ma-boko, arms ma- bo ... a-go li-ftio, an eye dish\ d-iso i-niyo mi-mie, the fingers mi-na mi-ne i-meno bu-ato, a canoe bi-ali b-olo ow-aro bu-sio, the face bo-she bo-so o-jo FAN. (Spoken on the upper stream of thp River Gabiin.) 231. ^ Judging from Don Amado Osorio Zabala 's Fan Voca- bulary lately published by Mr. Cust {1887), there can be no doubt that this is a Bantu language. It is closely allied to Mpongwe 4^ South-African Banhi- Languages. perhaps more closely related to Kele, and again forcibly reminding one of Chwana, and even more of Mozambique. This is plain from the following permutations, several of which may be considered as being regular. 232. - Ex. : Tonga Fan Tonga „ =n after n Z= / „ = n final Jj dropped before and after Oy U „ ^ 6 before t, e, a 233.- dropped n-sel -tuimgue, sent ■tatu, three ku-tue, ear pi. ina-tue inu-tue, head indezu, beard mu-zimo, soul, spirit a-lind, mu-ezi, moon gon ma-zuba, days melu lu-boko, arm u-d o go ^a-«/a, people boru'or ... [b-ur '^(3'-fo«,aflame(Guha)') /^fl-M p]. ^u-bia, tire i p\. do-a ... Fan Mpongwe -lomigue -lomio -laa -raw aid, o-rot, pi. mold pi. a-roi n-nu e-lelu i-nina o-gweli MozAMB. Chwana -rotHelie{f)-romilwe '• -raru -raro nya-ru, pi. ma-ru ..." mu-ru e-reru tedii, mu-rimoQ) mo-dimo mtveri kgwedi ma-chuwa ... /e-{:ogo(cii'jg) a-iu ba-iho i?=^, or drop- ped y=gork,etc. K dropped ma-bele, breasts -bili, two in-gulube, a pig ku-ulu, a foot mu-lomo, the beak lino, a tooth i?ia-pile(jf) inabele -Hi -bedi ikuhiwe kolobe ma-bl am-bene -be -bant 7t-gul n-gowa e-ko o-golo en-soon o-junibu ... iHo-lomo as -on i-no ti-iiio le-ino u-fua, a dying man e-gu{e-kuQ))-yuwa,x.od\&-kiva,todA&-shwa,to6i& or e-u ku-fuma, to be rich kutna, riches ... ... -fuma in-zovu, an elephant en-sok n-dyogu . . . tlou im-vuvu, a hippopo- n-sogo-usui n-guu . . . kubu [tamus / nyoka, a snake iio ... i-noa noga in-kuku, a fowl ku n-dyogoni . . . koku 234. — Evidently this is not a complete list of the phonetic permutations of consonants between Fan and other Bantu languages. I exclude particularly all reference to the influence of « nasal, because I cannot trace its law in Fan as we have traced it in Chwana and Mpongwe. However, the extent of this influence may be conjectured from the fact mentioned by several travellers that " the nasalization of the language is very marked " (Gust's " Languages of Africa ", vol. II, p. 422). Comparative Phonetics. 49 235. — A very remarkable feature of Fan is the negligence with whiqh the vowels are pronounced (230). For not only do'we find here many words dropping their final vowel, principally after n, such as engan, " a doctor" (Tonga in-gangd); ason, "a tooth" (Tonga li-no); n-bom, " a boa " (Tonga im-booma); n-suur or n-suut, " a black man " (Tonga mu-sundu, Kafir on-tsundu, Chwana mo-sutu, etc.) ; but also several accented vowels themselves have an uncertain pronunciation, as is evidenced from the fact that the author of the Fan Vocabulary writes the same words with different vowels in different places, e. g. enom or enam or enom, "husband" (Mpongwe oWome); em-borre and -vora = ow^ (Senna -bozi), etc. 236. — This furnishes probably the correct explanation of another remarkable feature of Fan, viz. that in many Fan words the vowel a = Tonga or u; likewise Fan e = sbmetimes the Tonga or a^ and the Fan = often the Tonga u, etc. Ex. : Tonga Fan Tonga Fan meso, the eyes mise mu-yuni, a bird un-Sn -kulu, ancient -kod -a-palua, filed (teeth) e-bol, etc. cf. examples above. 237. — ^- ^' These, with the phenomena described in nn. 230, 213, 200, and 122, seem to be the, most important exceptions to the general principle of the relative stability of the vowels in Bantu (48). FERN AND! AN (Fernando Po). 238. — Strange to say, Fernandian differs from Mpongwe and Dualla by using the t in the same cases as, Tonga, Kafir, etc., instead of the ;' of /which we have just seen used in several other languages. Ex. Tonga Fernandian Mpongwe Banapa dialec Banni dialect Ureka dialect bu-ato, a canoe buato b-ato bato ow-aro * ku-tue, an ear ba-tiu b-ato ■ balo o-roi mu-tue, the head e-tue e-chue e-chue (Kua vtti-ru) ■tatu, three -tta -ta -ta -raro 239. — Fernandian seems even to be fond of ^'s as it replaces often by. ^ the Tonga s, as in n-tele, " a road." (Tonga n-zila), n-tohi, " the sun " (Tonga i-zuba), etc. However, in other words we find the t of the other Bantu lan- guages replaced by s in Fernandian. Ex. bw-aiso, " a woman " (cf. Kamba mw-aii'o, " a mother "), b-osso, " fire " (in^oio in Senna, Swahili, etc.). ' 2-40. — Another remarkable feature of Fernandian, at least of 50 South-African Bantu-Languages. its principal dialects, is the one noticed by Bleek, p. 248, viz. the frequent use of b where the other languages have m. Ex. Tonga Banni dialect mu-ntu, person bo-cho mu-alume, husband b-ube {D\iaX\d^ m-omi) 24il, — N. B. As for the rest, the documents at hand are insufficient to allow of any important conclusions being drawn safely from them. However I may say that in reading these same documents I am strongly reminded of the languages of the Lower Congo (nn. 146-158), and of Bihe (131). LANGUAGES of the CONGO FOREST. 242. — We are indebted to Stanley forgiving us in his " Darkest. Africa " words belonging to the languages of the dwarfs that inhabit the great Congo forest. Unfortunately no one can tell us whether these words belong to the original language of those tribes, or whether they have been borrowed by them from the agricultural tribes in whose neighbourhood they live. I take this latter view to be the correct one, principally because we know that the more southern dwarf tribes of the Kalahari desert readily adopt the lan- guages of their neighbours. (Cf Introduction). But, whatever view we take, the fact is that a large number of the words given by Mr. Stanley as belonging to the languages of his dwarfs are unmistakably Bantu in origin. Such are not only the numbers. -(5a;'7 " two ", -sa7'o and -karo " three, " -nna " four ", -tano " five ", but also a certain number of substantives, e. g. : — 243.— Cf. in Bantu : ba-iua, (Tonga, Kafir, etc.) mu-ntu (Tonga, etc., n. 322*) mu-kazi (^ongSi, Ganda, etc. n. 322*) mu-bua (Tonga, etq.) mu-lindi (Senna, Gindo, etc.) ma-bongo, brains (Tonga, etc. n. 440*) in-du (several Tonga dialects, n. 385*) i-zuba (Tonga, etc. n. 410*) m-bua (Kamba, etc. 385*) iivVan (Angola, etc. n. 410*) ma-nyo (Ganda, etc. n. 410*) ku4ui, an ear (Tonga, etc. n. 462*) e-raka (Herero, etc. n. 133). KU-MBUTTI Ba-kiokwa (Ba-kwa forest} (Ba-Kumu forest) ba-kwa, dwarfs mo-ku, a person vio-go kali, woman kali i-bu, a dog i-bu i-tindi, a foot i-tindi ... ma-bongo, head in-du, a house ... kupa, the sun ... m-bua, rain m-bu itari, a stone mi-nyo, teeth ?ni-nyo ki-tu, the ears ki-toi i-dakka, the tongue i-dakka etc., etc. Comparative Phonetics. 51 244. — Of course, the materials furnished by Stanley are not sufficient for fixing any of the laws which regulate the transitions of consonants in these languages. There are however at least three examples which tend to' show that the Tonga t is more or less regularly sounded k o'c ghy the, dwarfs of the Congo forest. These examples are -karo "three", (Tonga -tatu, Chwana -tharo and -raro), ba-kwa " dwarfs " (Tonga ba-tua, Chwana ba-rwa or ba-roa) and mo- /eu "a. person " (Tonga mu-ntu, Chwana mo-tho). On the whole, these languages seem to have more in common with the Chwana-Mozam- bique-Mpongwe than with the main group of the Bantu languages. SEMI-BANTU. 245. — We leave it to others to compare with the Bantu lan- guages which we study in this work several of those of the Soudan, Lower Niger, Liberia, Sierra Leone, Senegambia, and other parts of Western Africa. I believe that interesting affinities might be brought to light by such a comparison. Koelle's " Polyglotta Afri- cana" andChristaller's collections in the " Zeitschrift fur afrikanische Sprachen' will be found invaluable in this connexion. Most of these so-called negro languages are in fact semi- Bantu, and I do not think that a thorough investigation of their proper features can be made without some knowledge of the more primitive and less contracted Southern Bantu languages. Cf. nn. 598 and 830. CONCLUSION. 246. — This cursory glance at the most striking phonetic differ- ences between the best known Bantu languages, while forcing upon our minds many unexpected conclusions, naturally gives rise to a number of highly interesting problems. We see that this family of languages, if it be confined to the limits we have assigned to it after the example of other scholars, has been very improperly compared by certain philologists to the Aryan fa- mily. So far from finding any such distance between the most remote members of the Bantu family as between English and Sanscrit, we perceive that the greatest discrepancies between those members of the group which are furthest apart can scarcely be said to be equal on the whole to the difference between French and Italian. This being so, what is simply amazing is that untold millions of so-called savages, inhabiting a country nluch larger than Europe, 52 South- African Bantu Languages. cind devoid of political connexions, even in these days probably so remote from the time of their original separation, should still be. found to have languages so closely related together. Again, we see that in this Bantu family a whole group is sepa- rated from the rest by a peculiar set of phonetic features, such as the transition from t to r, z to /, and k lo g ox h, when it is not dropped entirely, together with changes due to an extraordinary in- fluence of half-suppressed nasals. And then, if we look at a map of Africa, we are struck by a sight no less amazing than the first. For the tribes which speak the languages of this group live by no means in the neighbourhood of one another, but they are rather at the opposite extremities of the Bantu field. They are the Bechwana and the Ba-suto near the southern end of Africa, with the most eastern tribes of Mozambique and the Comoro islands, the Tshagga nation of Kilima-njaro, and the north-western tribes of the Ogowe, Cape Lopez, and the Gabiin River. We understand that the ancient Oriental race which South-African natives call Kua (Ma-nkua or Ma-kua or Ba-koa, whence the diminutives Ma-kuana, Ba-hiana, Wa-ngwana, and Be-chwana), after having occupied the Comoro islands and Mozambique, may have gone down along the coast of Sofala, then ascended the Limpopo and its tributaries in quest of gold. We may even understand that the same race may have gone to seek precious stones in the direction of Kilima-njaro to those mysterious caves at Elgon which have been described by Thompson in his " Through Masai-land ", pp. 300-302. But we should not have expected to find the same race settled at Cape Lopez, and we fail to see which w;ay they followed in those emigrations of a past deeply veiled in mystery ('). 1. Since this went to the press it has struck me that the word X'ua, pronounced Goa or Gva at Kili- mane, is no other than the name of Goa in India, and that the Oriental race called Ma-nkua are no others than tne Moors, Parsis, Banyans, Battias, etc. , indiscriminately Included by the natives of several parts of Eastern Africa under' the name of Goanese, probably because most of them come from Goa, and the Portuguese colonies of the same parts have long been a dependency of Goa. Now, as the harbours of Mozambique have been for the last three centuries the most noted places for shipping slaves, I much suspect that the linguistic and etimologioal affinities existing between the tribes of Mozambique and those of the Gabiin are the result of nothing else than an interchange of slaves. I also notice that for the Tonga the word Ma-nkua is a synonym oi ba-kuala " people who can write"- IV. — ffiore General fi&onetic Ci)ange0. 247. — The phonetic changes which have been described in the preceding article are for the most part so pecuHar to this or that language as to form one of its prominent features. Here we shall turn our attention to a few other changes which are more generally- met with. They occur mostly in the combination of the different elements of the words. § I. We may include them under two heads, viz. i° Changes of sounds caused by the collision of two vowels. 2° Changes caused by the concurrence of certain consonants with other sounds. § I. Changes caused by the Collision of two Vowels. 248. — The general principle of these changes may be laid down as follows, with all reserve regarding its particular applica- tions, as these are somewhat different in the different languages : — 249. — 1° A, when occurring before another vowel, is scarcely ever elided, except in Nyamwezi (cf 76), but generally either there is a sort of assimilation of both vowels, each of them changing its sound into one which is intermediary between them, so that a-i and a-e become e-e, while a-u and a-o become 0-0; or a contraction proper takes place, viz. a-i and a-e become e, a-u and a-o become 0; a-a becomes a. In some languages, e. g. in Tonga, assimilation is the rule, contraction proper is the exception. In others, e. g. in Kafir, contraction proper is the rule. When through assimilation the same vowel should.be repeated three times, two of the vowels are contracted into one. Ex. : Tonga (assimilation) Kafir (contraction) A-I =ee^e m&eso or meso(=»ia,-isoJ, the eyes, a mt,hlo ('= a ma.-ihlo) A-E = «« = « ba-ntubetza(=ba.-iza=bs.Si-lza ), a ba-ntu bt.za(^bz.-&3a,bi.-diAza) the people came. A-U=flc=(7 u-zoonvua(^u-za,-\i-nvua=--u-za- u-o-va(=u-a,-VirVa = u-ya-ku-va) ku-nvua), he will hear (cf. 948). K-0=oo=o ma-tanga oonse (=a,-Onse), aW the a 7na-tangaonke(=a.-Otike) pumpkins. 250. — ^- ^- !• I have heard in Tonga both ba-ntte bb-onse and ba-ntu be-ense, all the people, as if a-o could change not only to ao, but also to ee/ unless the form beense may be explained by saying that the Tonga stem ;onse (= all) has also the form -ense, just as we have in Kafir -odwa or edwa, alone ^815). 2. In Tonga as it is spoken, the initial? of the'verb-z«^a" to go", which is very frequently used, assimilates to itself entirely the final a of preceding words. Thus we may hear iu-a-ki 54 South-African Bantu Languages. tnka " we went " for iu-a-ka inka, uli inka " he is going off" for ula tnka, etc. This may be explained by saying that the syllable ka being particularly accented in the -v&rh-inka causes the preceding syllable to prefer the weaker sound z to the stronger sound e. No account is taken of this phenomenon in the written language (253). On the contrary the verbs -injila " to go in '', -invua " to hear ", etc., and in many cases the substantives which begin with in lose their initial i after a. 251, — 2° The weak / or i, when occurring before another vowel, is generally assimilated with it, as in c^-elo (= ci-elo), " a ceelo " ('), and in ca-a mu-luma ( = ci-a mu-luma), " it has bitten him. " x^o2. — N. B. I. In Chwana, when a week /(^^ Tonga i) is immediately followed by a vowel, it is generally entirely assimilated with it. Ex. : o-no o-reka (= o-ne o-reka), " you were buying " ; o-na a-reka = o-ne a-reka, " he was buying,'' etc. (Cf. Crisp's Gr., P- 3I-) 2o3. — 2. The principles of assimilation and contraction thus laid down both for the vowel a and for the vowel i (or a weak e) are applied principally when prefixes or suf- fixes are joined to other elements of the same grammatical word. In this case it is better that the spelling should agree with the pronunciation, as in the above examples. But the same principles have other applications in the rapid pronunciation of such words as are immediately joined to one another. It will be sufificient to warn the reader of these once for all, without confusing the written language with them : otherwise we should have two different spellings of the same clauses, the one for slow, the other for rapid pronunciation. Ex. Slow pronunciation and written language : ndabona izuba, " I saw the sun " ma- kumi a-ta-balui " a large number ", lit. " tens which are not counted, " = ndabone izuba, ma-kuma a-ta-balui, in rapid pronunciation. When the first of the two words which meet in this way is a mere particle, such as the preposition -a " of", its sound in Tonga and the like languages is always modified before a vowel, even in writing ; in Kafir, Ganda, Herero, etc., a contraction proper takes place. Ex. Tonga : ma-futa e in-zovu {=^ ;.. a in zovu) " fat of elephant " (Kafir ma-futa e ndlovu = ... ai ndlovii). 254;, ^- The- impossibility of writing certain expressions as they are usually pro- nounced is particularly felt in Karanga, which, having a special horror of hiatus, always contracts or elides in ordinary pronunciation whatever vowels happen to succeed each other. Thus the Karanga would pronounce as a single word the whole sentence : " They saw a small house, " bakabonemumbececana. Which evidently must be spelt so as to separate the different words, bakabona iimanba icecana (cf 108). 255, — Z° I proper, when occurring before another vowel, keeps very nearly its proper sound in Tonga and apparently in the greater number of the Bantu languages, such as Yao, Shambala, etc., though a beginning of assimilation is sometimes noticeable. In Kafir, Herero, etc.,,? before another vowel becomes entirely consonantal, and is consequently spelt jj/ when it is not immediately preceded by a consonant ; but it is dropped when immediately preceded by a consonant. I. A sort of evil spirit which is supposed to fly about like a bird, and to bite people's heart, thus causing their death. General Phonetic Changes. 55 In Swahili, Senna, etc., the law is the same as in Kafir, except for the plural classifier of the class CI-ZI. This keeps the i or changes it tojj/. Ex. Tonga Kafir Swahili In-ganda i-angu, my house i ndlu yam nymnba yangu i-zina li-ako, thy name i gama lako jina lako in-^ombe zi-esu, our cattle inkonio zetu naombe zeiu zi-bula zi-enu, your chairs izitulo zenu vUi vyenii (alias vienu) 256. — ^- B. I before a vowel is elided in Congo after a, and in Angola after J, e. g. in Congo : nzo zanene (== zi-a7iene), " large houses " (Vetralla). in Angola -.jinzoja tnundele (= ji-a mundele), " houses of a white man " (Hdli Chatelain, p. 14.) ', In the other cases z before a vowel keeps its proper sound in these languages, as in Tonga. 257. — Exceptions. — In some cases, z before a vowel combines into one sound with the consonant before it. Examples of this in Chwana have already been noticed in words in which the phonetic permutation is double, viz. first i is replaced by £ according to n. 200, and then be^, le-, re-, etc. are changed respectively before vowels to j-, isk., etc. (cf. 202-206). Likewise in Tonga h- before a vowel (= Chwana le) changes in some cases to/, t.g. jatisuj^haxid." = li-anza (plural ma-anza). This very natural phenomenon is common to many languages. 258. — Again, in Swahili and several other Eastern languages ki- before a vowel changes to c or ch (8). Ex. in Swahili : ki-devu ■ ch-ako = " thy chin " ( — ki-devu ki-ako). 259. — In Senna the same phenomenon takes place not only before i, but also before e. Ex. u-fumu bu-anu bu-fice, "thy (lit. your) kingdom come " ('= u-fumu bu-anu bu-Ji^e). 260. — N. B. It, is interesting to notice that the Swahili sound ki, even before a consonant, is equivalent to the Tonga ci, Herero tyi, Kafir si, Chwana se, etc. (cf. class CI-ZI, 491.) 261. — 4° U, when occurring before another vowel, keeps its proper sound in the larger number of cases, and causes no change. In Kafir and several other languages it becomes more consonantal than in the others, and is consequently written w. Ex. Tonga Kafir Swahili Lower Congo lu-limi iM-ako, thy longue. u Ivf-imi Iw-ako u-ltmi -vr-ako lu-bini Ivr-aku kufua k\x-ake, his death u-ku-fa kwake ku-fa hn-ake fwa kw-andi mu-nzi yx-enu, your town u m-zi vT-enu m-ji w-enu 262. — A-'. B. Kjand wa often sound almost like oa, by a partial assimilation of u or w with a. 263. — Exceptions. — I. Z7 before o is changed to o, or these two vowels coalesce to 0, according as the languages prefer simple assimilation or contraction. Ex. Tonga : bu-siku bo-onse, the whole night = Kafir u bu-suku bonke. 56 South-African Bantu Languages. ao4;, — 2. U before a vowel is dropped in a few cases to be mentioned further on. (Cf. 656 * passim, etc.) The most important case is in Kafir ailtr the labial consonants b and/ Ex. u bu-so bako, " thy face, " (= u bu-so b\x-ako). 2oO. — 3' Ue ox we and o are convertible in some cases. Ex. Tonga : -buena or -bona, " see. " Kafir : i ngwenyama or i ngonyama, " lion. " N. B. Hence it is that in Kafir and Chwana stems of nouns ending with o are treated in composition with suffixes as if they ended with -we (cf. 202, 203, etc.). 266. — 4- Examples may also be found in some languages in which we or ue is convertible with «, as in mw-eli or tinili " the moon ", in Mozambique. Hence the word Na-muli, which is the name of certain remarkable peaks East of Lake Shirua, is etymo- logically nothing else than a Mozambique transformation of Nya-mwezi, and consequently means as well as this word " Mountains of the Moon ". 267. — 5- C^before /sometimes causes this latter vowel to be suppressed, e.g. ^a-2a, " to come " = kti-iza, as if in such cases u were a more important vowel than i. 268. — 6. £/ before a vowel coalesces sometimes with its consonant, at least in se- veral Bantu languages, viz. Chwana, Kafir, Senna, etc. (cf. 122 and 202-204). Ex. In Senna ; nyaku-sasamba {== mn-a kii-sasamba), " a merchant. " 269. — ^- ^- It should be remembered that in grammatical elements (classifiers and collective pronouns, 637) and in some other instances, the Chwana 0=^ u of the other Bantu languages. Before a vowel the same Chwana o is generally written w, when it does ' not coalesce with the preceding consonant (202-204). 270. — I" the other Bantu languages o as well as e, not being found in any gram- matical element, occur before vowels only at the end of words. Then o is sometimes decom- posed into ue or we, according to n. 265, while in the other cases no change takes place at least in writing, according to the principle which has been laid down in nn. 253 and 254. § 2. Various Phonetic Changes. 271. — 1° In Tonga and several other languages we find a letter which, though sounded e when accented and in the middle of a word, becomes t when not accented at the end of a word. This is the sound which we represent in some instances by i to remind the reader of this very principle. Cf. 14. Ex. a-fue = near. Derivative : a-fue-fui, very near -mui = one • ,, jnue-mui, (evr i-kumi = ten ,, ma-kume-kmni, hundred mu-s'e = earth „ «-;?«, on the ground; /««-««, in the ground, etc. 272. — ^'- B- I- Probably it is due to some phenomenon of the same kind that authors often hesitate between / and e at the end of a word. Thus Livingstone in his Tette vocabulary writes madze " water " and panse " down " in one place, while in another he spells the same words madzi and pansi. 273. — 2' Th^ penult often drags the last vowel of a word to its own sound. Thus we may hear Zttlu or Zula (proper name), Ba-lunda oxBa-lundu, "the Lunda people, '' etc. 274. — 2° A changes to e before certain sounds, though only in given cases, principally before «-, ^jj/?-,//-, or similar sounds, and in Kafir before certain verbal stems, etc., as if in such cases there were General Phonetic Changes. 57 a contraction of a with an obscured i- sound, or, more probably, a peculiar phenomenon of assimilation. Ex. Tonga : Se-dseke, the people of Sesheke [sing. Mu-cisehe = Muiciseke (?) n. 266] ; Be-cikudu, the people of Cikudu (sing. Mu-dkudu); Me-ja, horns (sing, t-jaj. Chwana : 'B&-cwana, Be-sufo, the Chwana people, the Suto people (sing. Mo-cwana - Mo-sutd). Congo : 'E.-sikongo or 'E.-xikongo, the Congo people (sing. Mu-sikongo), Kafir : "Wtnyuka, he went up ; 'Wesuka, he went off; V^&gqita, etc., he passed by, (where we should expect regularly wa-nyuka, wa-suka, wa-gqita, etc.). 275. — 3° ^ when occurring before y changes to i, if preceded by a dental consonant, provided there be no danger of a double meaning. Ex. U-yuni (= tu-yuni), little birds. 276. — N- B. I. £/and i seem to be interchanged easily in pronunciation when the change partially assimilates two consecutive syllables. Thus among Kafirs the common people will generally say nd\i.-ku-fwnene for ndi-ku-fumene " I have found thee '', u mfundisxx wait for u mfiindisi wait... " the master saS.A. .." i ko/u for i'kqfi " coffee", etc. 277. — 2. Through some assimilation of the same sort, the auxiliary formsj'^ and ze change in Kafir ioyo and zo before ku. Ex. : Hamba u-yo ku-ndi-kelela 'e mlanjeni, ■' go to fetch water for me in the river " (916, 948). 278. — 3. In the X\. (!i)in-gombe. ,, ,, Class (L)I-MA. ,, (l)i-zuba, sun, pi. ma-zuba, ,, ,, Class BU-MA. ,, bu-aio, csinoe, pi. ma-aio. ,, ,, Class KU-MA. ,, ku-tui, ear, pi. ma-tui. ,, ,, Class CI-ZI. ,, ci-bula, Ai&Sx, ^\. zi-bula. ,, ,, Class KA-TU. ,, ^a-««, baby, pi. ;«-««. ,, ,, ClassLU-(Z)IN.,, lu-limi, tongue, pi. in-dimi. . (p)a or Class (P)A. „ (p)a-nsi, down, (no plural). , ku or Class KU. ,, /sk-««, below, (no plural). . tnu or Class MU. „ ot«-»«, underneath (no plural). 315. — Some substantives are found to depart from the general rule in the choice of their plural prefix. We shall treat them as forming sub-classes. Thus — with cl. MU-BA we connect a sub'class MU-MA. Ex. Mu-karanga, a Karanga, pi. Ma-karanga „ KA-TU ,, „ KA-BU. ,, ka-niabua,Re3.,pl.bu-ntabua „ LU-ZIN ' ,, „ LU-TU. ,, lu-sabtla,h3.hy,p\.iu-sabila 01 m-sabila etc. etc. i. In my " Outline of a Xosa-Kafir Grammar Grahamstown, 1887 ", I term these classifying elements " characteristic prefixes", or simply "characteristics". I now think that the term " classifier ", proposed by the Rev. F. W. Kolbe, ought to be preferred. 2° ,, ,, mu- f ) ■ mi- 3° „ „~ in- j» (zjin 4° .. „ (Oi jj ma- S" ., „ bu- ») ma- 6° „ „ ku- ji ma- 7° ,, ,, ci- ,5 zi- 8° „ „ ka- }> tu- 9° ,. „ lu- ', (z)in 10° Locative class with prefix ... 11° ,, jj 12" 9, ,, 64 South-Afvican Bantu Languages. 316. — In Angola, Yao, Mozambique, and Senna, we find sub- stantives which have two classifiers in the singular number, both of which change regularly in the plural. Ex. in Angola : k.Si-mu-xz " a shrub ", plur. tvL-m'i-xi, k.a-n-iart " a small stone ", plur. tu-ma- iarz, etc. In point of the concord required all such nouns are prac- tically considered as having, their first classifier only. Hence, for instance, k.a,-mu-xz, plur. tu-mi-;i;?, belongs to the class KA-TU. I. — On Htticles. 317. — Before we begin to study each class separately, it is necessary to fore-warn the reader against a mistake which has often been made, viz. that of confusing with the classifiers a different kind of prefix, or rather a proclitic, which is usually met with before nouns (substantive and adjective) in several Bantu languages, corresponding in some of them both to our definite and to our in- definite article, and in others to the definite article only. In those languages which have some sort of such article before nouns its ordinary form is a mere vowel. Thus in Kafir the article, both definite and indefinite, is u, i, or a, according as the classifier following it, expressed or understood, somehow or other contains u, i, or a. Ex. u mti " a tree " or " the tree ", t li-so " an eye " or " the eye", a bantu, " people "or " the people". In Herero the article, also definite and indefinite, is always 0, except before nouns of the class li-ma in the singular, where it is e. Ex. ma-yuru, " the nos- trils ", e yuru, " a nostril " or " the nostril ". In Kafir and Herero, the article, being both definite and indefinite, is generally expressed before substantives when they are pronounced or written by them- selves. In Angola the article, only definite, is always 0. In Fiote or Lower Congo, where likewise it is probably definite only, its form is 0, e, or a, according as the classifier, expressed or understood, which follows it, contains u, i, or a. As an exception, the article is 0, or e, not a, before the classifiers MA and VA [ = Tonga (P)A]. In Ganda its form is also 0, e, or a, according as the following classifier contains u, i, or a. But, as far as we may judge from available materials, it seems to be both definite and indefinite. Probably it is heard only after a pause or breath, and even then not always (m). On Articles. 65 As a rule, no article is used in vocatives, nor after negative par- ticles. In Kafir it is omitted also after demonstrative pronouns, and in a few other cases. On this subject of the use and omission of the article there are between the different languages considerable divergencies which we shall not dilate upon in this work. A''. B. In Kafir proper names themselves take an article in the same cases as other substantives. On the contrary in Herero proper names, and some other substantives which are equivalent to proper names, such as mama " my mother ", ina " his mother ", tate " my father ", the "his father ", Ka-tyiungu " Mr. Wolf " (cf. o m-bungu " a wolf "), Kaha- Vandye " Reynard " (cf. o m-bandye " a fox "), etc., are oftener used without the article than with it. Ex. : — 318.— Kafir: With article : Nditanda a ma-hashe, I am fond of horses. Aye nga pina a ma-hashe'^ In which direction have the horses gone? Ndabona u Langa-li-bakle, I saw Langa-li-bal^l^ (a Zulu chief). Without article : Yopula, ma (not u ma). Mother, take the meat out of the pot. La ma-hashe... (not la a ma-hashe), these horses... A ndi na nto (not... na i nto), I have nothing. Ufuna n-to nina (not i n-to) ? lit. What thing do you want ? 319. — Herero: Without iiidcle: Vdna/yevandye, he ndyi-pahere... (not o vanatye...), My children, get for me... {"■ Zeitschrift", 1887-1888, p. 191). Muatye uandye, ue ndyi-esa (not o mu-atye) ? My child, dost thou forsake me ? (do. p. 202). TV. B.'We however find in the same work, p. 199, the following sen- tence : O mu-ndu, zondu zepi .? Man, where are the sheep ? Kahavandye atya...,'R£ynaid said... (do. p. 200). Iheuazepere... (not o the), his father slaughtered... With article : M'o u-tuku..., o vanatye drire tyi ve-kutitra o n-dyatu, n'ariretyi va- isa mo muatye. At night the children loosened the bag, and took the child out of it. (do. p. 192). 320.— Ganda: Daura n'azala bana (not a band)..., n'agamba bana-be (not a bana-be) ...: " Bana bange (not a.-bana), O Bu-ganda buno mu-bu-lie... " Bana ne bagamba (why not a bana T) : " Kitafe, lerofe a bana bato, fe tuna Ha Bu-ganda (not Bu-ganda) ? " Daura begot children, and he said to his children . "My children, this Ganda kingdom eat it you. " And the children said : " Our father, we little children, to-day shall we eat the Ganda kingdom ? " (" French Ganda Grammar, " p. 83). m South-African Bantu Languages. Mod£rn Angola : Taf etu, uala hu maulu axile o rijinct rie, kize ko tuala o kifuxikie, . . .tubangele mu kiaiiba. (Heli Chatelain's '■' Kim- bundu Grammar'' , -p. XX). Our Father, who art in Heaven, hallowed be Thy Name, Thy kingdom come, ...deliver us from evil. 321.-— Old Angola : Taf etu, uekala ko maulu dkondeke o rijina riae, kjizf ko f uekala o kifucikiae, ...tubangele bo mu kiaiba. (Father de Coucto's " Cate- fAMw",i66t,p.i.The spelling is adapted to our alphabet). Congo : , With article : Ke Iwalu olu-kata, there is the box (Father Visseq's Gr. Tekiaki e ki-kila, there is the papaw (do.). E dnmla di-andi diabizU, his house is beautiful (do.). E mi-nsenga mi-etu miavia,. our sugar-canes are ripe (do.). P-9)- [P- 49). Without art. : Ki-nkutu ovene Npetelo (not e ki-nkulu), he has given a book to Peter (do. N. B. I. Though Father Alexandre Visseq seems to have on the whole understood the Congo.article better than the Rev. W. Holman Bentley, it is necessary to warn the reader that he has mistaken the classifier DI {= Tonga LI) for the article corresponding to it, and vice versa. What has given occasion to this mistake is that in Congo the classifier DI is generally reduced to E when there is no article before it. If we had to judge of the value of the article in Congo from the remarkably sparse sentences which we find in Rev. W. Holman Bentley's Grammar, we could no more say whether it is definite or indefinite than when it is and when it is not used. Ex. N-ti watt wambote (why not o n-ti), " this tree is good" (Bentley's Gr., p. 556). Cf. o matadi mama... i mau mama twamwene ezono, these stones are those which we saw yesterday (do. p. 526). 2. Articles are found in a few languages which have not been mentioned above, such as Bihe, Nario, and other dialects of Benguella, as also in Nyambu (119), etc. But from available materials it is impossible to make out after what laws they are used. 3. If Mpongwebe compared with the language of the Bihe, it looks very probable that several of the Mpongwe classifiers were originally articles. The classifiers proper having been dropped through contractions in many cases, the articles have remained instead, and their original notion has' probably been lost. 4. Strange to say, articles used often to make their appearance in Tonga, when with the help of my informants I would try to render English sentences into this language, but I do not find a single article in the stories and sentences which I wrote under their immediate dictation (Cf. Appendix I).. In these the nearest approach to articles are substantive pronouns occasionnally placed before nouns where we should use definite articles in English. Ex. Ue miiana uangu wafua " my child is dead ", lit. " Ae, child of me, is dead ". Hence, until further researches on this point, I consider Tonga as having no article. At the same time I conclude from these facts that probably the articles of the other languages were originally contracted substantive pronouns (830). n. — jpfje Mu-BA class anil t^e Sub^tlasses connecteD toit^ it. 322. — The substantives which belong to the MU-BA class, including the sub-classes connected with it, are those which require in the singular number the same sort of concord as the word mu- niu " a person ", plur. bSi-niu*. These sub-classes connected with the class MU-BA are: — 1° the sub-class — BA, or those substantives which, though requiring in the plural the classifier BA, have none in the singular, as faia " my father ", plur. ba-iaia; — 2° the sub-class MU-MA, * EXAMPLES ■ 'i a person a man (vir), husband a child, son Siur. Plur. Stftg. Plur. Sing. Plur. Tonga mu-ntu, ba. mu-alume, ba- mu-ana, ba- Bisa mu-ntu, wa- mu-analume, wa- mu-ana. wa- Gogo mu-nhu, wa- ih-lume. wa- mw-ana, wa- Sagara mu-nhu (? , wa- iji-lume. wa- mw-ana, wa- Shambala mu-ntu. wa- m-goxi. wa- mw-ana- wa- Boondei mu-ntu, wa- m-gcsi, wa- mw-ana. ,wa- Talta mu-ndu. >.* in-lume, mw-ana. ... Nyanyembe mu-nhu, wa- m-goxi. wa- mw-ana, wa- Sukuma mu-nhu, war m-goxi, wi- mw-ana, wa- Kaxaba mu-du. a- m-j]me. a- mw-ana, a- Swahili m-tu. wa- m-ume, wa- mw-ana. wa- Pokomo mu-ntu, wa- mu-yume, wa- m-ana. wa- Nika mu-tu, a- mu- ume. a- mw-ana. ana Senna mu-ntu. (w)a- m-amuna. wa- mw-ang. wa- Karanga (u>n-tu, ba-nu norume, ba- nona. ba- Ganda mu-ntu, ba- m-saja. ba- mw-ana. ba- Zulu~Kaflr u mu-ntu, aba- ... u nyana. nyana Xosa-Kaflr u m-ntu. aba- ... ... u nyanaj nyana Herero mu-ndu, va- mu-rumendu, va- mu-ma, vanatye Bihe mu-nu. ma- u-lume, a- o mona, Mbunda mo-no, ' ba- ... ... ngw-aneke. ba- Rotse mo-nu. a- ... ... mu-ana, a- Guha mu-ntu, ba- . ■■ mu-ana. ba- Rua mu-ntu, ba- mu-lume. ba- mu-ana, ba- Angola ^ mu-tu. a- mu;lume, a- mona, ana Lower Congo mu-ntu. a- n-kaza, a-kaji mw-ana, ana Nywema o-ntu. a- ume (o-ume ?) .»• ona. ana Yao mu-ndu, wa- a-sono, a-cKa.- mw-ana, mw-ana, a-cM w- Kilimane mu-to, a- m-amna, ■ .. ana Mozambique m-tu, a- mw-amna, a- mw-ana mwane , ana-ane Ch-wanaproper mo-thu. ba- mo-nona. ba- njw-ana. bana Suto mb-tho, ba- mo-nna, ba- ngw-ana, ban a Mpongwe o-ma. a-ttaga onome. a- on w- ana. aw-.. Fan e-f^m, ba- e-n6m. ... m6n. Dualla mo-tu, ba- m-omi, h- muna,' bana Fernandian bo-cho, be- b-ube, ba- bo-Iai, ba- (Banni dialect) . . (Banapa dialect) 68 South-African Bantu Languages. or those substantives which, though requiring in the singular the classifier MU-, have in the plural the classifier MA-, as ybl-nkua " a white man ", pi. 'Ma.-nkua^ § I. Transformations of the Classifier MU. 323. — This particle may be said to have in the different Bantu languages all the intermediate sounds between mu and n, as well as between mo and o. Even in those languages in which it is most reduced traces are preserved either of its labial nasal element, or of its z^-sound. Hence more particularly the following forms: — EXAMPLES. (Continued.) a woman, wife a chief a servant God Stn^. Plur. Sing. , . Plur. Sing. Plur. Tonga mu-anakazi, ba- mu-ame. ba- mu-zike. ba- Leza Bisa mu-anakazi, wa- ... ... mu-sia, wa- Lesa Gogo m-chekulu, wa- J ... mu-lelwa, wa- Mu-Iungu Sagara m-ke, wa- m-ridewa, wa- m-fugwa. wa- Mu-lungu Shambala m-kaza, wa- ... m-xumba, wa- Mu-lungu Boondei tn-kaza, wa- ... ... m-lugoja, wa- Mo-lungu Tatta mu-ke, ... m-gosi ... m-tiimu ... Mu-lungu Nyanyembe m-kema, ' wa- m-temi. wa- m-deki, wa- Mu-lungu Sukuma mu-kima (?), wa- ... m-sese, wa- Mu-lungu Kamba mu-ndu mu-ka, a ndu a-ka ... ... mu-dedia. a- Mu-lungu Swahili mw-ana m-ke, wa-ana a-ke m-falme, wa- m-tumwa. wa- Mu-ungu Pokomo mu-ke, wa- ... ... Mu-ungu Nika mu-che. a- mu-vieri, a- mu-humiki. a- Mu-lungu Senna (u)n-kazi. a- (u)-mbuya, a- mu-lece, a- Mu-lungu Karanga tiokaji, ba- xe, ba- (u)n-ja(?), ba- Reja Ganda m-kazi. ba- kabaka. ba- mu-ddu, ba- Katonda Zulu-Kafir u mu-fazi, aba- ... u mu-ntu. aba- u Tixo Xosa-Kafir u m-fazi, aba- ... u m-ntu. aba- uTixo Herero omu-kajendu, ova- mu-hona, c va- o mu-karere va- Mu-kuru Bihe u-kai, ' a- ... ... Suku Mbunda mo-nokazi. ba- mw-ene mu-hikana, ba- Redza Rotse mo-kati, a- mo-yoande, a- mo-bika. a- Nyambi Guha m-kazi, ba- ... • >■ m-jia. ba- Kabeja(.?) Rua mu-kazi, ba- m-lohhe • *• mu-hika, ba- Virie Angola mu-kaji. a- ... mu-bika, a- Nzambi Lower Congo n-kaza. a-kaji ... *.. n-leke. a- Nzambi Nywema o-azeni, a- o-lowe, a- o-hombo, a- Kixi Yao ^-sbno, a-cA'a- m-c/42-m\v-ene wai- kapolo. a- Mu-lungu Killmane mu-yana, a- mu-enye ■ •> .... Mu-lugo Mozambique mw-ari. ari mw-ene, ma-mwene karumia. a- M-luku Chwana proper mo-sadi, ba- mo-rena, ba- mo-tlhanka. ba- Mo-dimo Suto mo-sali. ba- mo-rena. ba- mo-tlhanka. ba- Mo-limo Itipongwe onw-anto, anto o-ga. a- o-xaka, a- Anyambe Fan >>. ... ... en-saga mo-kum. Aname Dualla mu-'tu, b'-itu mo-anedi, ba- ba- Loba Fernandlan bo-adi, ba- bo-tukwe, ba- bo-taki \ ... Kadupe (Banni dialect), (Clarence dialect) (Banapa dial ■ ) Hie MU-BA Class. 69 324. — MU- generally, in Tonga, Bisa, Mbunda, Herero, Angola, Nika, etc. M- with an affection to the vowel «, in Swahili, Mozambique, Shambala, Kamba, etc 325. — N. B. I. In most of these languages, if not in all, the law is evidently to pronounce the vowel u- distinctly, when otherwise the word would be sounded like a monosyllable. Hence in Ganda mu-nttt, " a person '', not m-ntu; mu-ddu, '' a slave ", not m-ddu. Do. in Kamba, Nyamwezi, Shambala, etc. It is somewhat strange that Swahili and Mozambique should prove an exception to this law (cf. 44). 326. — 2. In these same languages the »-sound of this classifier is partly preserved before such stems as begin with a vowel. Hence vrvt-ana, *' a child ", etc 327. — ■ N- with an affection to u, in Senna, Karanga, and Lower Congo. N. B. In Senna and Karanga the u is heard distinctly when the word begins the sentence, but then it precedes the nasal instead of following it, as if the sole reason of its EXAMPLES. (Continued.) Cnames of nations) my father my mother Sittg: Plur. SifS; Plur. Tonga Mu-tonga, a Tonga, Ba- tata, ba- (ma), ba-ma Bisa Mu-bisa, a Bisa, Wa- tata, wa- ma (?) ,.. Gogo ... ... tata yaya Sagara M-sagara, a Sagara, Wa. baba mau Shambala M-xambala, a Shambala, Wa- baba mlala Boondei M-boondei, a Boondei, Wa- tate mlale Taita ... ... aba mawe(?) ... Nyanyembe M-nyamwezi, a Nyamwezi, Wa- tata, wa- mayu, wa- Sukuma M-sukuma, a Sukuma, Wa- baba, wa- mayu, wa- Kamba M-kamba, a Kamba, A- — ^,a-chakwa mw-aito, a- SwahUi M-jomba, a man of the Zanzibar coast ,Wa- baba mamangu ... Pokomo ... ... • •■ baba ... Nlka Mu-nyika a man of the desert. A- baba niayo(wangu)... Senna Mu-zungu, a Christian, a lord. Wa. ^ a-tatu — ,amaianga ma Karanga (u)N-karanga, a Karanga, Ma- tate, ma- Ganda Mu-Ganda, a Ganda, Ba- kitangi nyabu Zulu-Kaflr u Mu-tshaka, a Zulu, %Ma- u baba, 0- umame Xosa-Kaflr u M-xosa, a frontier Kafir, %Ma- tata(bawo), 0- u ma ,Herero Mu-herero, a Herero, Va- tate, tate mama Bihe ... ... tate . ... mai Mbunda Mu-mbunda, a Mbunda, Ma- n-tate Rotse Mu-loi, a Rotse, Ma- xangoe (?)... 'me ".'.. Guha ... • ■. tata, ba- maju Rua Mu-rua, a Rua,, Ba- tata lolo Angola Mu-mbundu, a black, A- Lower Congo Mu-sikongo, a man of the Congo, e- (273) tata mama Nywema ... yoni mboni Yao M-yao, a Yao, Wa- a.-tnt\,a-ck'a- a ma wo Kilimane Mu-goa, an Indian Portuguese, Ma- baba n-ma Mozambique M-kua, do. (= Tonga mu-nkua), Ma- — , a-thithi mama Chwanaproper Mo-chwana, a Chwana, Be- (273) rara mme Suto Mo-sotho, a Suto, Ba- n-tate 'me Mpongwe Fan N-suut or N-suur, a black, ... rere ngi yami ... naa (') Cualla ••• ••• tite Fernandian (Banni dialect) ... obu-lieo (?) berim (.?) 70 South-African Bantu Languages. pronunciation were to support the nasal. In such cases, as also before monosyllables, some people pronounce /««-rather than un-. 328. — MO- in Chwana and Dualla. Bo- in some Fernandian dialects (246). 329. — U- (seldom MV-) in Bihe. 330. — O- in Mpongwe and Nywema, with traces of the nasal in some nouns. 331. — O-t or -£-(?), in Fan, also with traces of the nasal in some words. A'i^i^ — N. B. \. As may be seen in the subjoined examples, the word mu-ana a child ", changes variously to mona or m\yna (zl. 265), nptana (204), nyana, (122), nona (265 and 328), etc. , 333. — 2. There is no trace of this classifier being naturally long (mu) in any Bantu language. If so pronounced in some words, it, is owing to some sort of con- traction or to position before a nasal. Bleek mentions that it is marked long in Thlaping, a dialect of Chwana. It would be more correct to say that in Thlaping, though it is written mo- as in Sato, yet properly its sound is an intermediate one between mu- and mo-. § 2. Transformations of the Classifier BA-, 334. — This classifier has its consonant more or less weakened in the different languages, probably according to the shape of people's lips. Hence the various forms : — 335. — ^z^- in Tonga, Kafir, Ganda,Guha, Chwana, Karanga, Dualla, Fan, etc. N. B. Properly speaking, in Tonga Ba- has a sound intermediate between Ba- and Wa-. 336. — WA- in Swahili, Shambala, Nyamwezi, Yao, etc. 337. — A-'m. Mozambique, Senna, Angola, Congo, Mpongwe, Kamba, Nika. N. B. In Senna a slight labial aspiration is still perceptible in this classifier. Hence in some cases it is even spelt iva-. 338. — VA- in Herero and Nano. VA- or MA- in Bihe (Cf. " Observations upon... Umbundu ", by the Rev. Wesley M. Stover, Boston 1885, pp. 13, 16 and 17). 339. — N. B. I. Be- replaces BA- before « and in some other cases, according to n. 274, as \ibe- were then a contraction for ba-i. The presence of be-, as if for ba-, is parti- cularly remarkable in the Kafir wprd a "Belungu " white people " (sing, u M.-lungu " a white man, a lord "). This phenomenon probably is due to the fact that this word is of foreign importation. (Cf. the Phenician and Hebrew word tnelekh, or molokh in the possessive expression a-molokh). It may be observed by the way in the preceding table of examples that the Bantu word Mu-lungu or Mu-luku " God " is probably no other than the Phenician Moloch. 34:0. — 2. In Kele (Di-kele) the plural botyi, " people ", is probably for ba-utyi, just as in Fernandian buchu " people " is for ba-uchu, and in Isubu bomi " men " for ba-umi. 341. ' — 3- Other phonetic changes produced by the concurrence of ba-, wa-, a- with vowels, are easily explained according to nn. 249 sqq. The MU-BA Class. 71 § 3. The Sub-class BA. 342. — There is a large proportion of those substantives which require the same sort of concord as the word mu-ntu, " a person ", though they have no classifier in the singular. Such are 1°) the words, in nearly all these languages, for " father " and " mother ", viz. (in Tonga) /«/«, " my father ", uso, " thy father ", mse, "hisfather", usokulu, " thy grandfather", etc. (Cf. 748). 34;3. — -V. B. I. In Tonga the vvords for " mother " are through politeness used in the plural instead of the singular. Hence da-ma, " my mother ", bd-nyoko, " thy mother'', ba-nyena, " his mother ", etc. (cf. 748). 344: . — 2.' In some other languages a similar law is extended to names of parents in general. It appears that in Yao it is even extended to some other substantives, as we, find that the substantives " husband, master, brother, friend ", etc. are respectively ren- dered by the plural forms n-sono, ambuje, a.-kulu, a.-mwene, etc. (cf. 354). The Yao word a.-chi-mwene " a chief", which is sometimes used tor m-chi-mwene, is likewise a plural of dignity or respect which contains the classifier cki- (502) besides the classifier a-. The fact that in this word chi- Is in the singular number, while a- is a plural of dignity, shows that the Yao themselves, must have practically lost this notion that a- is in the plural number. 345. — 3- In Senna, many substantives of this sub-class are formfed with the prefix nya- (= mu-a^ 122). Ex. nysL-ku-fula " a smith ", pi. a-n-^Orku-fula. Substantives of the same sort have in Mozambique the prefix ka. Ex. k.a-r«cw«/a " an apostle ", pi, a-Y.eL-rumia. Cf. S17. . , 346. — Such are 2°) all proper names of persons, as Monze, " the chief Monze ". 347. — J^- B. Many proper names of persons begin with a prefix which means " Father " or " Father of... ", " Mother " or " Mother of "- Hence in Tonga Si-meja, lit. " Father Tusks", Sia-pi, lit. " Father of where? ",Na-simbi, " Mother of iron ", etc. Hence also in Kafir Sa-Rili, lit. " Father Kreli ", So-ndawo, lit. " Father of the place ", No-nio " Mother of a thing ", "etc. Hence also in various languages those names of God which begin with Ka, as Ka-zova (in Nyambu), Ka-tonga (in Ganda), etc. 348. — Such are 3° several names of animals, e, g. su-ntue, " a hyena", se-kale, "a muircat ", etc. 349. — N. B. I. Like proper names of persons, many such substantives may be decomposed into two parts, the first of which is a prefix which seems to mean " father ", or " mother ", or " son ". Such are in Tonga the words just mentioned, and in Kafir u no- tnadudwane, " a scorpion ", lit. " a mother of little dances ", u no-meva, " a wasp ", lit. " a mother of stings ", etc. Such are in Senna s-ulo, " a hare " (Tonga s-ulue), nya-rugue, " a tiger ", lit. " son of a tiger " (= Tonga si-lugue,\\t. " Father tiger "), etc. etc. 350. — 2. In the'language of Mozambique some names of inanimate things, prin- cipally of fruits, belong to this sub-class. They have na- or /ia- as a prefix. Ex. »a-kuo, " a cob of maize; " pi. a-naiuo; ka-raha, " a sweet potato, " pi. a-karaka. 72 South-African Bantu Languages. 351. — The plural of all such nouns is formed in the generality of the Bantu languages by prefixing the classifier BA- to the form of the singular number. Ex. ba-suntue, " hyaenas " (sing, su-ntue), ba-sokue, " baboons " (sing, so-kue). 352. — N. B. I. In Kafir such substantives take " EXAMPLES. (Continued.) ■■ ' 7 - '' ' . T 1 Ithe back the heart a tree a baobab- tree ■ Sing. Plur. Sing. Plur. Sing. Plur. Sing. Plur. Tonga mu-sana, mi- mu-oyo. mi- mu-samo. mi- mu-buyu. mi- Bisa mu-s'ana, mi- ... ... niu-ti. mi- t .. ... Gogo m-gongo, mi- ... ... Sagara m-gongo, mi- m-oyo. mi- mu-ti, mi- m-pera. mi- Shambala mu-gongo, mi- m-oyo. mi- mu-ti. mi- m-uyu. mi- Boon del mu-gongo, mi- m-oyo. mi- mu-ti. mi- m-buyu. mi- Taita mu-gongo, mi- ... mw-iti, mi- ... ... Nyanyeaibe m-gongo, mi- m-oyo, mi- mu-ti. mi- m-pela. mi- Sukuma m-gongo. mi- m-oyo, mi- mu-ti. mi- m-pera. mi- Kamba m-mongo. mi- ... m-ti. mi- mw-amba, mi- SwahUi m-gongo, mi- m-oyo. mi- m-ti, mi- m-buyu. mi- Pokomo m-ongo, mi- m-otyo. mi- mu-hi, mi- . .. ... Nika m-ongo, mi- m-oyo, mi- mu-hi, mi- m-uyu, mi- Senna (u)n-sana, mi- m-oyo. mi- (u)n-tengo, mi^ (u)m-buyu, mi- Karanga (u)n-xana, mi- m-oyo. mi- (u)n-ti, mi- u m-buyu. mi- Ganda mu-bega, mi- m-oyo. mi- mu-ti. mi- ... Zulu-Kafir u m-hlana, i mi- ... u mu-ti, imi- Xosa-Kaflr u m-hlana. i mi- ... . .. u m-ti, imi- Herero ... ... mu-tima, o mi- mu-ti. mi- Bihe u-tima. vi- u-ti, vi- Mbunda m-ongo ... Rotse in-ongo=end of spine mii-jima, mi- mu-sito, mi- Guha m-gongo, mi- ... mu-ti, mi- ... Ru'a mw-ongo, mi- mu-ula (?) ... mu-ti (?) ... Angola ... ... mu-xima. mi- mu-xi. mi- m-bondo (?] ... Lower Congo ... m-oyo, mi- n-ti, mi- n-kondo. mi- Nywema o-vuna. e- o-tima = belly o-ti, i- ... Yap m-gongo. mi- m-tima, ■ mi- m-tela. nii- m-lonji, mi- Kilimane ... ... ... ... mu-rre, mi- m-laba, mi- Mozambique m-thana. mi- m-rima. mi- m-tali. mi- m-lapa, mo-wana, mi- Chwana proper mo-tlana, me- ... ... me- Suto mo-tlana, me- ... ... mo-wana. me- Mpongwe o-kongo. i- o-rertia, i- . . Fan ... ... e4i' Dualla ... ... mo-lema. mi- ••* Fernandian ... bu-ila, bi- ba-ti(?) bo-ti ... ... .„ 78 South- African Bantu Languages. has the contracted form U-. It is pronounced ME- in Chwana ac- cording to n. 200, and BI- in some Fernandian dialects according to n. 240. In Mpongwe and Nywema its form is I- or E-. In most of the other Bantu languages its proper form is MI-, 370. — ^- ^- "• I" Tonga I often thought I heard it pronounced like mu in the French mur. This inclines me to think that its original form was MUI. 2. These two classifiers M0- and MI- correspond to one another as singular and plural in all the Bantu languages. Bleek has it that MI- corresponds regularly as plural in Nika to the classifier U- (== Tonga BU-), and he gives as an example the word u-miro, " voice ", to which he ascribes mi-miro as plural. But it is now plain from Rebmann's "Nika Dictionary " that the whole idea is incorrect, for properly speaking the word in Nika for •' voice ", or more exactly for " word ", " speech ", is m-oro, pi. mi-oro, and certainly m-oro is regularly of cl. MU-MI, as in the Nika proverb : TA-oro ra^x-d^o ka-\x-lavia dzua, " a good word does not bring out (?) the sun. " (Rebmann's " Nika Diet., " word mora). EXAMPLES ■ (c ontinuec I.) fire a river (mu ddy) a moon, month a year Sing. Sing. Plur. Sing. Plur. Sing. Plur. Tonga mu-lilo mu-Ionga, mi- mu-ezi, mi- mu-aka. mi- Bisa mu-lilo ... ... mu-ezi, mi- mu-aka. mi- Gogo m-6to m-ongo, mi- m-lenge. mi- mw-aka, mi- Sagara m-oto m-korongo, mi- m-lenge, mi- mw-aka, mi- Shambala mu-oto mu-to, mi- mu-ezi. mi- mu-aka. mi- Boondel mu-oto m-to, mi- mw-ezi. mi- mw-aka, _mi- Taita m-oto mw-ita. mi- mw-ezi, mi- m-aka (?), mi-(?) Nyanyembe mu-lilo m-ongo,' mi- mw-ezi. mi- mw-aka, mi- Sukuma m-oto m-ongo, mi- mw-ezi. mi- mw-aka, mi- Kamba mw-aki ... ... mw-ei. mi- mw-aka, mi- Swahill m-oto m-to, mi- mw-ezi, mi- mw-aka. mi- Pokotno m-oho ... mw-esi. mi- mw-aka. mi- Nika m-oho mu-ho. mi- mu-ezi, mi- mu-aka, mi- Senna m-oto (u)n-tsinje, mi- mw-ezi. mi- ... ... Karanga m-oto ... ■■■ mw-eji, mi- mw-aka, mi- Ganda mu-lilo mu-gga. mi- mwiezi, mi- mw-aka, mi- Zulu-Kaflr u mu-lilo u mu-lambo , mi- ... ... unyaka, imi -nyaka Xosa-Kaflr u m-lilo u m-lambo, imi- ... ... unyaka,imi -nyaka Hereto mu-riro mu-ramba— torrent mu-eje, mi- ... Blhe ,. .. ... ... ... u-nyamo. vi- Mbunda (o)n-diro (o)n-donga,mi-l.. (o)n-gond B(?) mw-akwari. mi- Rotse mu-lilo mu-lonka, mi- mu-eti, mi- mu-aka. mi- Guha ... mu-fito, mi- mw-ezi, mi- • •• ... Rua mu-jilo, mi- ... ... ... ... • • ■ ... Angola mu-lengu= flame ... ... ... mu-vu, mi- Lower Congo n-laku aflame n-koko, mi- — , mi-eji = moonlight m-vu, mi- Nywema ... ... ... 0-eli, ... ... ... Yao m-oto m-lusulo, mi- mw-esi. mi- >•■ ... Killmane m-oto (?) ... ... mw-erre, mi- ... ... Mozambique m-oro m-oloko, mi- mw-eri. mi- mw-aka, mi- Chwana proper mo-lelo ... ... ... ... ngwaga ... Suto mo-lelo ... ... ... ... ngwaga ... Mpongwe o-goni o-lovi, i- o-gweli, i- o-mpuma. i- Fan ... ... ... J gon = moon \ — ; mi-el == moonlight ... ... Diialla ... mo-opi, ... ... m-bu, mi- Fernandian bo-sso ••, ... ... ... ... ... The MU-Mi Ciass. ' 79 § 3. Substantives which belong to the MU-MI Class. 371^ _ In Tonga, and, as it seems, in the generality of the Bantu languages, the substantives which belong to this class are principally : — 1° The names of such complete trees and plants as stand up without support, as in Tonga mu-samo, " a tree " in general ; mu-nga, " a mimosa-tree"; mu-konka, " a cocoa-tree "; mu-buyu, " a baobab- tree ". N. B. We shall see further on that the names for the fruits of such trees are generally of class LI -MA. 372. — 2° The names of such tools or artificial objects as remind one of the form of a tree by having branches or bushy parts, as mu-ini (alias ■ntu-pmi), " a handle, " mu-iaezio (alias mu-piaezio), " a hroom " , mu-nvui " an arrow " (bearded), mu-zuete, "clothes", niu-pdnda, " a cross ", etc. 373. — 3° The human and animal body, mu-bili, as also such of its parts as brancb off in some manner, growing out into acces- sory parts, or move up and down, as mu-oyo, " the heart ", mu-nue, " a finger ", mu-limba, " a feather ", etc. The same may be said of the similar parts of trees, as mu-yanda, " a root ", etc. 374. — 4° All beneficent elements and producers of animal or vegetable life, such as mu-ezi, " the moon", which in Africa is thought to be the great source of rain, while rain is thought to be the greatest benefit which men can receive from God (cf. the specimens of Tonga at the end of this work) ; mu-longa, " a river " ; mu-ezi, " a pool of water " ; mti-tulu, " a fertile plain " ; mu-nda,, " a garden " ; mu-se, "the soil"; mu-lilo, " the fire ", which naturally reminds these people of the food it cooks, and of the warmth in which it keeps the body during cold nights ; mu-nzi, " a living-place ". • 375. — 5° The soul, a shadow, and several objects noticeable either for their instability or their variety of design, as mu-zimo, " the soul ", the plifral mi-zimo being used principally with reference to the departed souls (Kafir i mi-nyanga or i mi-nyanyd) ; mu- zimuemue, " a shadow " ; mu-mpini-ciongue, " the rain-bow, " mu- bala, " a variety of colours, " etc. 376. — ^- ^- I" ^ ^^'^ languages, e. g. in Kafir, three or four personal substan- tives x>r tribal names belong to this class MU-MI. This seems to be due to their including some reference to the word for "spirit ", mw-zimo. 8o South-African Bantu Languages. 2>n, -^ 6° The breath, the air, and empty spaces, as mu-oya, " the breath, air, breeze " ; mu-lindi, " a pit in the ground " ; mu- liango, " the door- way ", etc. 378. — 7° Medicines, unfermented beverages, and some other products with beneficent or marvellous effects, as mu-samo, " a medi- cine", viz. anything belonging to that which to a primitive mind forms the genus " physics ", such as even secret sciences ; mu-ade, a certain supposed judicious poison, which kills sorcerers, while it exculpates the innocent (cf. appendix I) ; mu-bonobono, " castor oil " ; mu-sili, "powder " ; mu-sinzai " soup " ; mu-kande, " very light Kafir beer ", opposed to bu-kande, properly " fermented beer " (cf. 440*) etc. 379. — 8° A few names of immaterial things which occupy a fixed time, or come round at regular times, as mu-aka, " a year " ; milia, " feasts with sacrifice " (a word apparently not used in the singular) ; mu-sebemo, " a work ", etc. 380. — ^- B. In Senna the nearly total loss of the classifiers LI- and LU- has caused many words to be brought into this class MU-MI, which in the other Bantu lan- guages do not belong to it. Ex. mw-ala, " a column, a stone " (= Tonga lu-ala, a column, i-6ue, a stone). This remark extends partially to several other languages. § 4. Etymologies. — Varia. \ 381. — Judging from the sort of substantives thus admitted into the MU-MI class, it seems pretty evident that the predominant notion in this class is that of " objects which are light, move, change, grow, produce, or, in general, which contain some principle of life and production, a notion intimately connected with that of "power of growing up " like a tree. Hence I should think that the classifier MU- of this class is, like MU of the class MU-BA, radically iden- tical with the adjective -umi, alive, from the verb -ma or ^ima, " to stand ". Bleek connects it with the preposition mu which means " in ". Perhaps the correct thing is to unite both opinions by saying that the classifier MU- is directly connected in some words with the preposition mu, and in others with -umi. It may also be that in a few words its immediate connexion is with the verb -nytta (Karanga -mua) " to drink, " h. e. " to take light food " (Cf. 430). 382. — As to the classifier MI-, we should see in it the fun- damental element of the verb -mila or -mena " to grow " (cf. 280(2) ), exactly as we connect BA- with -bala (363). The MU-MI Class. 8i 383. — N. B. The verb -mila or -mem " to grow " is the applicative form of -ma or -ima" to stand " (1065). This may be another reason to say that thp singular classifier MU- is related to the latter verb. I 384. -^ The examples given under n. 366 probably must be explained etymologi- cally as follows : '' ■ 1. Mu-bili " the body " = the upright thing which has its parts two by two. From -6i/i "Fwo, double " (792). , - 2. Mu-cila " the tail " = the hanging thing, or sort of branch, which sits upon (the body). From ka, notioh of" sitting ", which changes to <: before i (cf. 257-259), and -ila, notion of " stretching along, or upon, something " (1065). 3. For mu-tue " the head ", arid mu-buyu " a baobab (tree) ", we have only doubtful etymologies. 4. Mu-lomo " the lips " ^^ that which is drawn inwards. From /o, notion of " being drawn" (cf. -/a/a "to lie down", z-/o "a bed", -j/ala "to stretch", etc.), and mo "inside" (530, 656 Tables). ^.Mu-sana "the spine, the back", lit. "that upright member lyhich sends its own shoots through the body " From sa, notion of " thrusting something^ through a body " (cf. -yasa. " to thrust a spear, to shoot "), and na or ana, notion of " close union ''{d. 363(3)). > ■_. 6. Mu-oyo " the heart ", lit. " the pfirt of the body which beat»s, going up and down ". Cf mu-oya," the air, the wind ", ku-yoya " to breathe ", etc. 7. ATu-samo "a tree", lit. "the standing thing which thrusts roots within (the ground) ". From JO, notion of "thrusting something through a body" (jK/ra 384(5)), and »?o" within, inside " (supra, 384(4)). Many languages replace mu-samo hf mu-ti, which means lit. " a thing standing in the ground ", from ti, notion of" ground " (Swahili n-ti" ground "). In Chwana the usual wprd for " tree " is setlhare (cl. CI-ZI), iri which //,4a = Tonga jo (i74) and re = ti of mu-ti {172,20c)). Hence se-ilkare means also lit. " the thing which thrusts roots through the ground ", but, as it is of cl. CI-ZI, it does not include the notion of something standing, like mu-samo. 8. Mu-lilo " fire, flattie ", means lit. " the thing which goes up eating its own bed " From //, notion of "eating" (cf. -h'a " to eat "), and lo, notion of" something drawn out " or of " a bed " (supra, 384(4)). , - 9. Mu-longa " a river ", lit. " the thing moving down, being drawn through gaps "- From lo, notion of " bed " (supra 384(8)) and «§"«, notion of " going through a gap. " 10. Mu-ezi " the moon ", lit. " the mother of water and fertility ". Mw-ezi ===■ mti-a-izi, and -izi is the same element which appears in lu-izi " a river ", mu-nzi " dwelling-place ", lit. " birth-place ",-ma-nzi " water ", etc. (cf 284). This element -izi or -nzi conveys the notion of production, fecundity. The moon is considered by nearly all the Bantu tribes as the great fertilizing power in the world. 11. Mu-aka " a year ", lit. " one station ". Connected with ku-yaka " to build "- The Bantu are in the habit of renewing the thatch of their huts every year. IV. — mtt iN-(z)iN Class. 385. — The IN-(Z)IN class includes the substantives which admit the same sort of concord as in-sz/a " a path ", pi. {z)in-zzh*. N. B. In Kafir there is a sub-class IN-MA. Ex. 1 n-doda " a man, a husband ", pi. a •axA.-doda, § I. Transformations of the Singular Classifier IN-. 386. — This classifier stands- in nearly the same relations fo the letters iV and /as the classifier MU- to the letters M and U. Hence the following forms : — 387. — NI- or NY' before vowels in several languages, viz. in Tonga, Ganda, Kafir, etc. , 388. — (I)N- before consonants \IM- before b, p, v, f, (n. 280)] in Tonga, Bisa, and Bemba, with a sound often approaching that of en. When this classifier is * EXAMPLES. a native doctor ■ < the beard flesh, meat a head of cattle Sing. Plur. Sinff. Plur. Sing. Sing. Plur. Tonga in-ganga, (zi)n- in-dezli, (zi)n- iny-ama in-gombe, (zi)n- Bisa • •')." in-ama n-gombe, n- Gogo '■^i^'K ny-ama n-gombe, n- .Sagara n-gangar.?) ... ... ny-ama n-gombe, n- Shambala n-ganga(?) ... n-de'zu, n- ny-ama n-gombe, n- Boondei ... n-dezu. n- ny-ama n-gombe, n- Taita ... gafa (.?) ... ny-ama n-gombe, n- Nyanyembe ... ... n-ama n-gombe, n- Sukuma ... n-ama n-gombe, n- Kamba jeu ny-ama n-gombe, n- Swahill ... n-defu. n- ny-ama n-gombe, n- Pokomo ... ... ... Nika *.> ... n-defu = hair ny-ama n-gombe, n- Senna ' n-ganga, (zi)n- n-debzu, (zi)ri- ny-ama n-gombe, n- Karanga i-ganga, i- i-devu, i-nyama i-ngombe (?), i- Ganda n-ganga=a sacred bird ny-ama n-te, n- Zulu-Kafir i ny-anga, i ziny- i n-devu. i (zi)n- i ny-ama i n-komo, i(zi)n- Xosa-Kafir ... in-devii. i (zi)n- i ny-ama i n-komo, i(zi)n- Herero n-ganga, zon- ... ny-ama n-gombe,ozon- Bihe n-ganga, lon- (0 n-jele), lon- situ n-gombe,olon- Mbunda n-ganga, n- n-jezu (?), situ n-gombe, n- Rotse n-ganga, n- ... ny-ama n-gombe,(ti)n-(?} Guha ... ... ... ny-ama n-gombe, n- Rua n-ganga (?) ... «... n-gombe, n- Angola n-ganga, (ji)n- xitu n-gombe, (]i)n- Lower Congo n-ganga, (zi)n-. n-zevo, m-biji n-gombe, (zi)n- Nywema ... ... Yao n-deu, (si)n- ny-ama n-gombe, (si)n- Kilimane n-ganga, n- e-rrelo, e- ny-ama (?) gombe, di- Mozambique ... i-reru, i- i-nama i-ngope, di- kgomo, di- Chwanaproper ngaka, di- tedu. di- nama Suto ngaka, 11- telu, li- nama kgomo, li- Mpongwe ... ... ... ny-are, (si)ny- Fan en-gan n-sel ... en-ds6m Oualla n-sedu (?) ... nyama=ahimal ny-akka Fernandlan ... e-sedu - n-kelapi ''■S°P°',B^Ip^?£? The iN-(z)iN Class. 83 very intimately connected with a preceding word, no trace at all of its vowel / or e is perc^tible,* so that we may hear, for instance, tu-a-kovtba m-vula, " we have asked for rain " next to tm-lapela Leza ini-vula " we have prayed God for rain. " The presence of the / in this form is particularly felt in possessive expressions, where it pTpduces, together \vith the possessive particle a, the sound ee, which we write e i (249, 253), as in mii-tue ue-ngomie, " the head of a cow" (= •■• uain-gombe), N. B. Before monosyllabic stems the classifier IN- Sounds almost like een {eem before i, p, etc.). Ex. eem-pi'e " an ostrich "- 389- - — N- before consonants (M- befpre b^p, v,f) regularly in most of the othfer languages, if we may trust to.our authorities. But several of these languages, viz. Swahili, Angola, Herero, Yao, Shambala, Mpongwe, etc., regularly drop this n before the hard consonants s, f, x, h, k, p, t, according to n. 283, as also before m and n. Here again however the tendency to avoid monosyllables comes in to prevent the n from being dropped before monosyllabic stems (nn. 283, 44, 325, 368, etc.). JJXAMPLES. (Continued.) a goat a fowl a snake an elephant Sing. Plur. . Sing. Plur. Sing. Plur. Sing. Plur. Tonga im-pongo, (zi)m- in-kuku, . (zi)n- in-zoka (zi)n- in-zovu (zi)n- Bisa m-buzi, m- nkuku, n- ... 'Gogo m-peni, m- n-khukhu, n- n-zofu. il- Sagara m-bxizi,^ m- n-khukhu, n- n-joka, n- n-tembo, n- Shambala m-buzi, m- n-guku, ri- ny-oka, ny- tembo Boondei m-buzi, m- n-guku, n- ny-oka, ny- n-tembd, n- Ta-ita m-buzi, m- n-guku, n- ny-oka, ny- n-jovu. n- Nyanyembe m-buli, m- n-goko, n- n-zoka, n- n-zovu,' n- Sukuma m-buli, m- n-goko, n- Kamba m-bui, m- n-guku, n- n-soka, n- n-zou. n- SwahlU m-buzi, m- kuku, n- ny-oka, ny- n-dovu, n- Pokomo ,-. . . , . . paa {?)' ... n-dzofu. n- Nika m-buzi, . m- kuku ny-oka, ny- n-dzovu, n- Senna m-buzi^ (zi)m- n-kuku, (zi)n- ny-oka, (zi)ny- n-jou, (zi)n- Karanga ... i-uko, i- i-nyoka, i- i-joo. i- Ganda m-buzi, m- n-koko, n- n-joka, n- n-jovu. n- Zulu-Kaflr ... i n-kuku, i (zi)n-. i ny-oka, i(zi)ny- i n-dlovu. i(zi)n- Xosa- Kafir i-bokue, i(zi)- i n-kuku, i (zijn- i ny-oka, i(zi)ny- i n-dlovu. i(zi)n- Herero on-gombo,ozon- on-dyUhua,pzon- ny-oka, zony- n-dyou. zon- Bihe hombb, lo- sanje, ■ lo- ny-oha, lo- n-jamba, lon- Mbunda m-pembe, m- ... ... ... yamba Rotse m-pongo,,(timr(?) n-goku, (ti)n- (?) ny-oka, (ti)ny- n-dopo (?) ... Guha m-busi, m- n-kuku n- .;. Rua m-buzi, ' m- n-zolo n- ny-oka, ny- holo Angola hombo, 0')- sanji, (ji)- ni-oka, (ji)ni- n-zamba. (ji)n- (zi)n- Lower Congo n-kombo, (zi)n- n-susu (zijn- ni-oka, (zi)ni- n-zamba. TJywema m-buli ... ... ... Yao m-busi, (si)m- n-guku (si)n- n-dembo, (sijn- Kilimane buze(?) ku, di-ku noa, di- doo. ^ di- Mozambique i-puri, i- i-laku i- i-noa, i- i-tepo, tlou, 1- Chwanaproper pudi, di- kgogo (i") di- noga, di- di- Suto puli, ' li- khogo di- noga, oli- tlpu. li- ■Mpong-wa Fan m-boni, (si)m- n-dyogoni, (si)n- ku m-peiie, (s)im- n-dyogu, en-sok (s)in- Dualla m-bodi, m- ... • ,,. m-bamba, m- n-dsou Fernandian m-pori n-ko, or in-ko mapa .... "i :/'''^' \ ' South-African Bantu Languages. Hence in Swahili the words n-so, " loins " ; n-si, " a gnat " ; n-xi, " the eye-brow " ; «-// or ncAi, " land " ; n-ta or n-cAa, " a point " ; n-_fi, " the sting of a bee, " etc. 390. — /- or E- in Mozambique, with strengthening of the initial consonant of the stem, according to n. 183, sqq. /- in Katanga, though without any such strengthening of the initial consonant of the stem. Dropped in Chwana^ but with strengthening of the initial consonant of the stem, according to n. 183 sqq. Here again the tendency to avoid monosyllables preserves the n before them, e. g. in n-tlit or en-Hu, " a house " ; n-ku or en-ku, " a sheep '' ; n-tlha, " a point " ; n-tsi, " a fly " ; n-ca, " a dog " ; n-che, " an ostrich " ; etc. § 2. Transformations of the Plural Classifier ZIN-, 391. — Though the substantives of this class require a different EXAMPLES. (Continued.) .shame rain a house a path , fc Sing-. Plur. Si<^. Sing. Plur. Sing. Plur. Tonga in-soni im-vula in-ganda, (zi)n- in-zila. (zi)n- Blsa in-fula in-ganda. in- in-zira. in- Gogo ... m-vula n-ganda. n- n-jira. n- Sagara soni hi-vula nrumba. n- n-gila. n- Shambala spni fula : ny-umba, ny- sila Boondei soni fula ■:•' ny-umba. ny- sila Taita ... m-vula ny-umba. ny- n-gila, n- Nyanyeinbe n-soni m-bula n-umba, n- n-zila, n- Sukuma m-bula n-umba, n- n-zira. n- Katnba n-ltoni (?) m-bua n-umba, n- n-fia. n- Swahili soni = abu^p m-vua ny-umba, ny- n-jia. n- Pokbmo • .• ... ny-umba. ny- ... Nika ... m-fula ny-umba. ny- n-jira. n- Senna ... m-vula ny-umba, (zi)ny- n-jira. (zi)n- Karanga i xoni i vura i-mumba. i- i zira, Ganda n-sonzi n-kuba (114) ny-umba, ny- Zulu-Kafir i n-tloni i m-vula i n-dlu, i zin-d u i n-dlela, i (zi)n- Xosa-Kaflr i n-tlohi i m-vula i n-dlu, i zin^dlu i n-dlela, i (zi)n- , zon- Herero honi m-bura n-dyuo, zon- n-dyira Blhe m-bela n-fo. olon- n-ji a, olon- Mbunda n-fera n-jolo, n- n-gela. n- Rotse n-fula n-do, tin-(.?) n-dela, (ti)n-(?) Guha mvula n-sese, n- n-jila. n- Rua ... m-vula Angola m-vula in-zo. (ji)n- n-jila. (ji)n- Lower Congo n-soni m-vula n-zo, zm- n-jila (zi)n- Nywema vula (m-vula ?) m-vulu. m- Yao soni ula ('ula ?, 66) ny-umba. (si)ny- Killmane ny-umba, diny- dila,' di- Mozambique i-xoni i-pula i-nupa i- i-piro Chwanaproper (tlhong), di- pula (e)n-tlu, ma- tsela. di- Suto (tlhong), li- pula (e)n-thlo. tsela, di- Mpongwe n-tyoni, ... n-ago, (s)iA- m-pono. fs)im. Fan en-san=offence ... en-da en-kon-el6f?')... Dualla ... m-bua n-dabo n-gia R-teI)6 Fernancllan *i* n-kola n-chibo,orn-iobo (Utekadi ' / •'" The IN-(Z)IN Class. ?55 concord when used in the singular and when used in the plural, yet practically they themselves generally have the same form in both numbers, viz. in-, n- or i-, as above. The following forms are there- fore the exception rather than the rule : — 392. — : ZIN- in Tonga, and probably, in Bisa, Nyamwezi, etc., when special attention is called to the plurality of the thing spoken of. 393. — ZIN- in Kafir in thte same case, and besides — a) in vocatives, as in zVOrkosi! " My chiefs ! ", — b) regularly before monosyllabic stems, as in i zin-d/u " houses ", — c) regularjy after the locative particle e, as in e zin-d/eleni " in the roads " 394;. — //N- (JI- before hard consonants, 389) in Angola, when attention is called to plurality, and regularly before monosyllabic stems, as in jim-iua " dogs " (Chatelain's Gram., p. 140), perhaps also regularly when substantives are preceded by the article o,as'inoyim-6ong0 "riches". Ci.ngulu "pigs" {^Ibid., p. 133)) n dende, " palm-muts " (Jl>/d. pp. 142, 143). JV. B. Probably similar principles are applied in several other languages. 395. — -DI- (alias LI-) regularly in Chwana, with a hardening of the following consonant ; DIN- before monosyllabic stems (390). 39o. — SIN- (SI- before hard consonants) regularly in Mpongwe, INAw given cases (Cf. Ms"^ Le Berre's Gram., pp. 4, 5). 397. — r ZON- regularly in Herero. A very extraordinary form, when compared with the others, on account of the vowel which it contains (cf. 230). N. B. I suspect that its true origin is to be sought for in some kind of imitation of the Portuguese article os. 398. — LON- in Bihe. A regular modification of the Herero ZON- (131). § 3, Substantives which belong to the IN- (Z)IN Class. 399. In the generality of the Bantu languages, we find in this class apparently all sorts of substantives, more particularly : 400. — 1° A few names of persons, as in-ganga " a native doctor ", etc. 401. — 2° A great many names of animals, principally of the milder type as im-bizi " a zebra, a horse ", im-belele " a sheep ", im- booma" a boa ", eem-pie (388 Note) " an ostrich ", in-jina " lice ", etc. Many of these substantives are often treated as belonging to cl. MU-BA(36o). 402. — The flesh and a few parts of the body, as iny-ama " flesh, meat ", in-dezu " beard ", in-kumu " the forehead " (including the nose), im-pemo " the nose ", zn-go " an ankle ", in-singa " a vein ", etc. 86 South-African Bantu Languages. 403. — 4° A few objects and phenomena in nature, as inyenyezi " a star ", in-simbi " metal ", more especially " iron ", im-vula " rain ", i^nyika " a place ", more especially " an empty place, a desert. " 404. — 5° A great many artificial objects, principally, as it seems, such as are curved, or yield to pressure, or are produced by smelting, as in-samo " a flexible rod ", in-celua " a pipe ", in-juzio "a key", in-kaba "a die", in-goma " a musical instrument ", more particularly " a drum ", in-kando " a hammer ", in-gubo '* a piece of cloth, a blanket ", im-pete " a ring ", in-sangu " a shoe ", in-tibi " a shield/', in-tiba "a' knife", in-tobolo "a gun", in-sima "por- ridge ", etc. 405. — 6° Uncomfortable sensations, as im-peho " cold ", more particularly " cold wind, winter ", in-soni " shame " ; inyaezia "danger", etc. 406. — N. B. In Kafir nearly all foreign names of things are brought into this claiss, as i kofu, " coffee ", unless they begin with s, for these are generally brought under class SI-ZI (= Tonga CI-ZI). '' § 4. Etymologies. — Varia. 407- — In this gfreat variety of substantives which are brought under the IN-ZIN class, it appears very probable that this is the proper class for all the substantives which there is no special reason for bringing under any of the others. The classifier I N or N may originally have been flo other than the indefinite adjective -mue (Kafir -nye) " one, another, sqme " (792, 828). Cf. 122, 204, 327, .517. 559. etc. 408. — As to the classifier ZIN-, it seems to be connected with the verb -ziala " to bring forth young ", so that' it would signify primarily " the progeny of beasts ", according to what has been said in n. 363. This further brings it into connection with the element nst or tst " notion of fecundity ", which we have already met with in mu-esi " the moon ", ma-nzi " water ", mu-nzi " village ", lit. " birth-place ", etc. (384(10)), and which probably furnishes the adjective -nji " many " (Kafir nt-nzi or «?-nji, etc. (601, Ex.)). 400. — The examples given under n. 385 probably must be explained etymologi- cally as follows : I. In-ganga " a doctor ", lit. " one who sees through and through "- From «?a, the notion of " going through a gap " (384(9)). The IN-(2:)IN Class. 87 2. In-dezu (= in-lezu) " the beard ", lit. " what comes out long ". From -/?, notion of " length " and zu, notion of " coming out ". The proper meaning of the elements /e and zu in Bantu is perfectly plain. We find le in ku-le " far 'f (533, Ex.), and in nearly all the transformations of the adjective which means " long " (601, Ex.). The element zu gives us the verb ku-ztta " to come out "- It may even be remarked that the last element of the -fioxA in-dezu varies in the different languages exactly as the verb -zua. Thus Ganda, Karanga, Kafir, etc. which replace zua by vwa or va (whence, in Kafir, the applicative verb -vela = -vwela, cf. 1069)), replace also in-dezu by in-derv\x, ki-le-vu, i-dewa, etc. In like manner Chwana which replaces zua by cwa (= dwa or Iwa, whence the perfect du-le oxlu-le, cf. 205), replaces also in-dezu by tedu (193, 173), etc. 3. For the words inyama "meat", in-gombe *' cattle ", im-pongo and m-buzi " a goat ", in-zoka " a snake ", in-zovu " an elephant ", we have only doubtful etymologies. 4. In-kuku " a fowl ". An onomatopoetic word, derived from the cry of this bird. 5. In-soni " shame ". This word includes unmistakably a reference to the eyes, li-nso, plur. meso. But I do not see exactly what notion is conveyed by the element ni unless it be the locative sufifix described in nn. 553-555. 6. Im-vula " rain ", lit. " what opens out (the earth) ". Related to -jula (Kafir -vula) " to open ", from zua or va " to come out ". 7. In-ganda " a house ", lit. " a protection "- Related to -yanda " to love, to protect. " 8. In-zila " a path ", lit. " what goes to a definite place ". From za " to come " and t7a, applicative suflfix (1065). v.—mbt LI-MA dags. 410. — The class LI-MA includes the substantives which require the same sort of concord as {\)i-bue "a. stone", pi. tasL-due*. § I. Transformations of the Classifier LI-. Here it becomes particularly important to distinguish the sub- stantives which have monosyllabic stems from the generality of the others. Then we must also set aside such as have stems beginning with a vowel. Hence : — li Polysyllabic stems vvhich begin with a consonant. 411. — Before the polysyllabic stems which begin with a con- EXAMPLES. the devil, the sun, a day a duck an eye s. pernicious spirit Sin^. Plur. Sixg. Plur. Sing. Plur. Sit^. Plur. Tonga li-saku, ma- (l)i-zuba, ma- (Di-stkua, ma- li-nso. meso Bisa ... i-dyoni. ma- 1-inso, menso Gogo ... ... i-zuwa, ma- nyaniwala ma- z-iso, meso Sagara i-zimu, ma- i-jua, ma- i-wata, ma- d-iso. meso Shambala ... zua, ma- wata. ma- z-ixo, mexo Boondei loho (?), ... zua, ma- wata, ma- z-iso, meso Taita vppoQ), i-jua, ma- bata. ma- iz-izo Nyanyembe li-gunhu, ma- li-uwa. ma- i-mbata, ma- 1-iso, m-iso Sukuma i-beho, in a- le-enii ... li-mbata (? , ma- d-iso Kamba ... i-jua, ma- i-kuanyungu,ma- ito. mento Swahlli zimui, ma- jua, ma- bata, ma- ji-cho. ma- Pokomo ... , dsua. ma- kaza. ma- dsi-tso, ma- Nika pepo dzua, ma- bata. ma- dzi-tso, ma- Senna saku {?), ma- dzua. ma- di-so. ma- Karanga - xaku, ma- juba, ma- ... ... -ixo, mexo Ganda mandwa (?] bata. ma- i-so. ma- Zulu-Kafir i zimo«cannibal» i langa, ma- i dada, ama- i llso, a mehlo Xosa-Kafir i zim«, cannibal » i langa. ma- i dada. a ma- i llso, a mehlo Herero ... e yuva. ma- eho, o meho Bihe e ii-abu e kumbi. o va- i-so. o va- Mbunda ...^ li-tangwa ra a- ... 1-ixo, mexo Rotse ... ' li-yoba, ma- 1-io (.?), meo Guha juwa, ma- ... 1-iso, meso Rua juva. ma- . .. j-iso, ^ meso Angola ri-abu, ma- ... r-isu, mesu Lower Congo e tombola, ma- . .. d-isu, meso Nywema ... yani li-uta (?), a- i-so, wa- Yao li-soka, ma- li-ua. ma- li-wata, ma- 1-iso, . meso Kilimane ... ... n-zua, ma- ni-bata, ma- ni-to, ma- Mozanibique n-xoka, ma- n-chuwa, ma- n-rata, ma- n-itho. metho Gh-wana proper ... le-tsatsi,ma-latsi ... le-itlho ma-t!ho Suto . .. le-tsatsi, ma- le-itlo, ma-tlo Mpong'we i-nini 0} ... izage, a- i-ntyo, a- Fan ... ... . y6 ... d-iso. mise Dualla i sangu = idol i-ve ... d-iso, miso Fernandlan ... ... i-tohi e-mipoto{?) ...■ j-oko (Banapi dial.) (Urekadial.) The LI-MA Class. 89 sonant, the classifier of the singular number in this class is : — /- generally, in Tonga, Bisa, Sagara, Kamba, Mpongwe, Dualla, etc. LI- in Tonga, only when emphasis calls for it. DI- in Tonga, after the copula n, according to nn. 286, 291 and 583. LI- generally in Yao. RI- in Angola and DI- in Congo. It is omitted in these languages when emphasis does not require it (321 (i)). LE- generally in Chwana. NI- generally in Mozambique. The vowel / is apparently very weak and, in some cases, omitted., altogether. ^- in Herero. Apparently this vowel contains the article together with the classifier (ai7. 319)- Omitted generally in Kafir, Swahili, Ganda, Shambala, Nika, Senna, etc. EXAMPLES. (Continued.) a tooth a spear a bone a pumpkin Sing. Plur. Sing. : Flur. Sing. Plur. , Sing. Plur. Tonga 1-ino, meno (l)i-sumo. ma- (l)ifua, ma- (l)i-tanga, ma- Bisa 1-ino, meno i-fumo, ma- ... Gogo idz-ino, meno Sagara ■ gego. ma- ... i-guha, ma- lengi. ma- Shambala z-mo, meno guha, ma- vuha (?), ma- tango (?), ma- Boondei z-ino, meno guha, ma- vuha (?), ma- koko, ma- Talta i-jego, ma- i-chumu, ma- ... ... Nyanyembe 1-ino, m-ino i-cimu. ma- i-guha, ma- li-ungu. m- Sukuma 1-ino kimo ... Kam/ba i-yeo,. ma- i-tumo (?) ma- i-windi, ma- i-beki, ma- Swahili Pokomo j-ino, meno fumo, ma- fupa«large bone» boga, ma- Nika dz-ino, meno fumo. ma- range, ma- Senna dzi-no. ma- dipa, ma- fupa, ma- tanga. ma- Karanga j-ino, meno fumo, ma- fupa, ma- puji. ma- Ganda li-nyo, ma- fumo. ma- gumba, ma- boga. ma- Zulu-Kafir i zinyo, a menyo ... i tambo, a ma- i tanga, a ma- Xosa-Kafir Herero i zinyo. a menyo i tambo, a ma- i tanga, a ma- eyo, ma-yo enga. ma- e tupa, ma- . .. ... Bihe eyu, va- e kepa,' va- Mbunda li-onga, ma- li-mputo, ma- Rotse li-yeo, ma- pinjei , ma-(?) ... ... ... Guha 1-ino, meno fumu, ma- Rua j-ino. meno ... i-kupa (?), ma- ... Angola ri-)u, ma- Lo'wer Congo d-inu. meno e di-onga ,madi- e-lenge. ma- Nywema li-nyu. wa- li-konga ... Yao 1-ino, meno li-panga. ma- li-upa, ma- li-ungu, ma- KlUmane !-ino. meno Mozambique n-ino. meno ni-vaka. ma- ni-kuva, ma- n-chuchu, ma- Chwana proper le-ino. meno le-rumo. ma- le-sapo, ma-rapo le-phutse. ma- Suto le-ino. meno le-rumo, ma- le-sapo, ma- le-phutse, ma- Mpongwe i-no, a- i-gonga, a- i-loge. a- Fan a-son. meson a-kon, ' ma- Dualla i-sunga , ma- Fernandian ?, belo ... ... ... 90 South-African Bantu Languages. 4tl2. — I. In Kafir the article z, and in Congo the article e, before substantives of this class must not be mistaken for the classifier. 2. In Nika we find the word Aomo " a large lip", of cl. LI-MA, derived from mu-lomo " a lip ". The dental d in this word points to the influence of a suppressed n before it. Several links connecting Nika with Mozambique may have been observed in the previous chapter. This is another. Likewise in Senna some substantives of this class LI-MA begin in the singular by double consonantal sounds which are simplified in the plural, as if the presence of these sounds in the singular were the result of a suppressed n. Ex. Xsamba " a leaf '', pi. ma-samba (Father Courtois' " Grammaiica Telense, " p. 28). Cf. 99, N. B. 11. Monosyllabic stems. 413. — In the words which have monosyllabic stems the law of avoiding single sounds (44) causes all sorts of irregularities, as may be noticed in the subjoined tables of examples under the words ^j)/^, tooth, and stone. EXAMPLES. (Continued.) a stone the sky a hoe ■Tonga Sing-. (l)i-bue,' Plur. ma- Bisa i ri-bue, ma- Gogo i-Jjue, ma- Sagara i-bue, ma- Shambala iwe, ma-iwe Boondei i-we, ma- Taita i-we, ma- Nyanyemba i-we, ma- Sukuma i-we, ma- Kamba i-bia, ma- S\irahill ji-we, ma- Pokomo dzi-we. ma- Nika i-we, ma- Senna Karanga ji-bwe, ma- Ganda j-inja. ma- Zulu-Kafir i li-tye. a ma- Xosa-Kafir i li-tye, a ma- Herero eoe. ma- Bihe e-we. ova- Mbunda le-manya, ma-nki (?) Rotse li'yoe, ma- Guha di-bue, ma- Rua ji-ve (.?), ma- Angola ri-tari. ma- Lower Congo e-tadi, ma- Nywema ... Yao li-ganga, ma- Killmane ... Mozambique n-luku. ma- Chwana proper le-ncwe. ma-je Suto le-ncue, ma-joe Mpongwe i-do. a- Fan a-kogk (? . Dualla i-dali. ma- Fernandlan i-te. ba- (l)ij-ulu, i-vimbi, vunde (.'), i-lunde, i-lunde. Pblr. ma- ma- ma- gulu, i.zulu, i Zulu, e yuru, li-elo(?) li-uilo, i-ulu rilu, e-zulu, li-unde, ma- ma- ama- a ma- oma- maulu ma-' ni-hute, le-godimb, le-golimo, d-oba(?) ma- ma- ma- Sing. Plur. (l)ij-amba, ma- i-sile, i-sire, gem be, gembe, i-gembe, i-gembe, i-zembe, jembe, jembe, paze, i kuba, i kuba, ma- ma- ma- ma- ma- ma- ma- ma- ma- ma- ama- ama- e-temo, o va- li-tema, ma- le-kao, ma- ri-temu, ma- li-jela, ma- n-hipa (?) di-bau, ma- Sing. (l)i-zina, i-sipa, i-tagwa, i-sina(?), zina, zina, i-zina, i-gina, 1-ina dz-itwa dzina,- zari, dzina, dzina, zina, li-nya i gama, i gama, e na, sina (?) i-sina (?) ri-jina, e-jina, 1-ina, n i-zina (?), n-cbina, le-ina, le-bitso, i-ni, dyg dina, i-la, Phir, ma- ma- ma- ma- ma- ma- ma- ma- ma- ma- ma- ma- ma- a ma- a ma- o ma- ma- ma- mena ma- ma- ma- ma- a- ma- ba-(?) The LI-MA Class. 91 ■414". — ' N. B. I. In Chwana the word le-n-cwe " a stone", pi. ma-jwe^ when compared with le-cogo " an arm ", plur. ma-bogo, le-satna " a cheek '', pi. ma-rama, etc., leads me to suspect that Chwana has undergone here the influence of a language like Mozambique in which the regular form of the classifier LI is NI or N. I see no other way of explaining the presence of « in le-n-cwe " a stone " (= Tonga i-bue, cf 185-I-203). . I ,have little doubt that we must have recourse to the same influence to explain the changes oibtoXheAtntaXcmle-cogo, andrto the more dentaljin le-sama {Tonga. i-/ama). 2. The variety noticeable in the formation of the words which have monosyllabic stems may be attributed in part to that sort of affinity between t and n which causes them to interchange in certain cases (285). This, coupled with transposition of letters, would explain the presence of « in the Tonga word h'-nso " an eye " (Subia li-nso, Kamba me-nto, Nyatflbu me-nso, etc.). For in these words the regular form of the stem is probably -sio, which we hnd retained in the Tonga bu-sio " the face ", lit. " the place of the eyes "- III. Stems w^hich begin vrith a vow^el. 415. — In the words which have stems beginning with a vowel- sound either the classifier LI- is somewhat transformed, as in the Tonga word (i)j-anza (256) " a hand ", (plur. ma-anzd), or a euphonic consonant, generally^, is inserted between the classifier and the stem, as in the Kaguru word i-^anja " a hand " ; or again in a few languages the classifier LI- is used without any change, as in the Nyamwezi word /i-ungu "a pumpkin" (Steere). JV. B. In the Herero e-oe. " a stone ", and the like, the vowel o must be considered as having a semi-vowel or consonantal value, or as being preceded by a sort of labial aspi- ration which replaces the Tonga * of z-i5«^. Otherwise the classifier e would probably undergo a change. § 2. Transformations of the classifier MA-. 416. — The regular form of this classifier is : — MA- in almost all the Bantu languages. N. B. The exceptional fcjirm ME- is easily explained according to the laws of contract- ion (249). A- regularly, AM- before vowels, in Mpongwe. A- or WA- in Nywema. VA- in Nano and Bihe (131). § 3. Substantives which belong to the Class LI-MA. 417. — The substantives which are brought under this class are principally such as refer to the following : — 418. — 1° Such persons or animals as are unproductive, barren or only productive of harm, and such as have a naked body, or a sleek, rigid, and relatively flat appearance, as i-saku " the devil ", i-buto " a naked slave", (such as those which are employed by the 92 South-African Bantu Ldnguages. u . Rotse to row), i-panda " a water tortoise ", i-sekua " a duck ", etp. Hence also in several languages the young of animals, as, in Kafir, i-tole " a calf, a young of animal ", i-takane " a kid ", i-tokazi " a" . heifer, a female Iamb or kid ", i-tshontsho " a nestling ", etc. 419. — 2° Fruits and those parts of bodies which are relatively hard, or bare, or flat, as i-ji " an egg ", i-buyu, " the fruit of the baobab ", i-konka "a cocoanut", i-ja " a horn, a tusk of elephant ", i-/ua " a bone ", ij-anza " a hand ", tz-nso " an eye ", tz-Ho " a tooth ", i-tama " a cheek ", i-kanda " the skin ", t-dete " a woman's breast ", which in Bantu proverbs is compared to a stone (cf. Heli Chate- lajn's Kivibundu Gr., p. 145), i-baba " a wing ", etc. 420. — 3° Such things in nature as are hard or unproductive, as i-bue " a stone ", i-zulii " the sky ", which the ancients thought to be hard as brass (Job, XXXVII, 18), i-foba " a cloud" {/bid., V. 21), i-saka " a sandy unproductive land ", t-dose " a drop of water ", i-suko "dust ", Hue " ashes ", etc. 421. — 4° The " sun ", or " day ", i-zuba, and those relations of time and place which the Bantu associate with" the day, or with the various positions of the sun, as z-/tlo " yesterday ", i-junza " to- morrow ", i-golezia " evening ", li-no " now " ; i-tale " the side of a river, or of other things ", etc. 422. — 5° Those tools and artificial objects which are hard, or flat and smooth, as ij^amba " a hoe ", i-jegeso " a saw ", i-hola " a kind of knife ", i-kuati " a table ", i-sumo " a spear ", etc. 423. — 6° Words and distinct sounds, as i-zina " a name ", i-jui " a loud sound ", i-ko " coughing ", i-zumo " a thunderclap ", etc. 424. — 7° A few actions, as z-j'ayo " a murder ", z-^ wound ", etc. 425. — To these must be added in several Bantu languages, e. g. in Tonga, Shambala, Nika, Swahili, Karanga, etc., augmentative nouns, or names of such persons or things as are remarkably tall, or high, or long, or large, as i-lundu " a high mountain ", i-yuni " a large bird ", i-samo " a high tree ", or " a large piece of timber "; etc. 426. — ^- B. I. In 5uch augmentative nouns the classifier of the singular number ' seems to be used regularly with its full form /»-, e. g. li-tui, "a long ear", li-bizi, "a large horse, " etc. The LI-MA Class. 93 4; 2 7, — 2. Some augmentative nouns have two forms, one which keeps the usual classifier together with the augmentative particle, another which drops the same classifier. Ex. jen-zoka (= li-en-zoka) or li-zoka (from the usual noun in-zoH^ " a snake '') " a large snake ". 428. — 3- Augmentative nouns are comparatively little used in Tonga, as if they were somewhat foreign to that language. § 4. . Etymologies, — Varia. 429. — The Rev. F. W. Kolbe in his ' 'Language Study based on Bantu", p. 52, considers the particle /«' to be 1°) the proper prefix for names of dead things, and 2°) to signify " in ", this, he thinks, being the reason why it is applied to " the dead teeth in the mouth, the bones in the body, the stones and metals in the earth ". The first part of this opinion may be correct enough, but th"e second part is more than probably the very reverse. And, if any classifiers signify " in ", the'se are rather the particles MU- and IN- which, as we have seen, are principally applied to such things as are covered with hair, or vegetation, or something similar, thus recalling to mind the fact that in ancient Egyptian a hieroglyph representing " a skin " is often affixed to the names of quadrupeds. , 430. — More probably the classifier LI- or RI-, in the generality of the substantives of this LI-MA class, is rather ^the naked form of the v&ch -Ha or -via "to eat", the same exactly as that which we have in the following expressions' taken from Kolbe's Herero Dic- . tionary : matu ri, '' we ate " ; ze 'sa ze tokere aze ri, " let them feed till sunset ". Hence the notion of strength which this particle con- tains. Hence likewise its augmentative power, because to a Kafir mind the notion of " king " and " lord " is convertible with that of " well fed ". Hence also its adaptation to the teeth, and to whatever has a crushing power, as stones ; and again to such things as are hard or resistant, as also to fruits, eggs, bones, breasts, or other parts of bodies which draw to themselves the best substance of these. Hence again its adaptation to the sun, which according to the manner of speaking of these people, eats all that the moon j^yx-ezi) labours to bring out of the earth (384 (10)), thus filling people both with reverence and terror by its power to cause the fearful droughts. Hence finally, on the one side its application to the eye, which is to the body what the sun is to the world, and .on the other side to sterile beings, as also to such as are the terror of weak and superstitious people. 94 South-African Bantu Languages. 4:31. — ' N- B. I. This conclusion is of some importance, as I notice that several missionaries honour the devil vv'ilh the classifier MU-, calling him mu-diaboU, or some similarly formed word, which evidently is calculated to convey to the minds the very^ opposite of the notion it is intended to express. More logical than ourselves, the natives of Angola have changed the Portuguese diabo, not into mu-diabo or mu-diaboli, but into ri-abu or di-abii. ^■SS. — 2. Both my Tonga and my Kafir informants used to say that the particle U-, sometimes replaced by izi- before monosyllables, forms " bad names ". And my Tonga informants added that this was the reason why a certain white man, whom they had heard of in the interior, and^f whom the less said the better, had not been called mu-nkua, as other white men, but izi-kua. 433. — Probably in some substantives of this class LI-MA, as in i-lo " a bed ", i/'-utu " the sky ", etc., the particle LI- is etymologi- cally connected, not with the verb -/ta, but with the element -te (Kafir -de) " long, high, far ", which seems to be itself essentially a form of the reduplicative verb -ta-ta " to lie down, to be stretched ". 434. — And in some others, as in i-j'm " a sound ", i-ko " sneezing, coughing ", i-zina " a name ", etc. the particle -tz- rather reminds of the verb -/ita " to produce a sound ". 435. — Another question is whether the three verbs -tza " to eat ", -tata " to be stretched ", -tz/a " to cry ", are themselves formed from one and the same root. But this is not the place to discuss it. 436. — With regard to the plural classifier M A-, there appears no serious reason to say with the Rev. J. Rath (Bleek's Comp. Gr,, p. 20o), that it is mainly used when speaking of things which con- stantly go in pairs. More probably the classifier MA- expresses properly " the end of natural production or multiplication ", being radically identical with the verb -mala or mana (280, 1065) " to end, to cease to produce ", exactly as the plural classifiers BA-, MI-, ZIN- are radically connected with the verbs bala, mila, ziala, all of which express production or plurality. 437. — This opinion is corroborated by the fact that the classi- fiers MA-, BA-, MI-, ZIN- are always long and accented, which is not the case with the singular classifiers MU-, IN-, LI-. And further it well agrees with the fact that the classifier MA- is precisely the plural for fruits " the end of the production of trees ", for the young of animals, for extremities of the body, for stones, bones, and other such things apparently no more subject to transformations. The LI-MA Class. 95 4;38. -^ N. B. All this of course leaves more or less room for exceptions in the different languages, according as they have been more or less modified by foreign influence or other causes. Then it should always be remembered that the same things may be Viewed in different lights, and brought accordingly under different classes. Hence from such words as raw-ala " a stone " (in Senna), e n-juba, " the sun " (in Ganda), z sl- qamo, " a fruit "(in Kafir), etc., nothing can be inferred against the above conclusions. ■4i3Q. — As to the sul)stantives which may be found under n. 410 : — I. We have only doubtful etymologies to give for i-sakii " a pernicious spirit ", li-no " a tooth ", i-tanga " a pumpkin '', ij-amba " a hoe '', and i-zina " a name ". 2. 1-zuba " the sun ", lit. " that which comes out with light "• From -zua " to come out ", and -uba " notion of light ". 3. 1-sekua " a duck " is an onomatopoetic v^ord derived from theory of this bird. When I asked my Tonga informants what they meant by an i-sekua, their first answer was " the bird which makes kua-kua... ". Bata " a duck ", of Swahili, Ganda, etc., reminds one of the Old Egyptian word api " a goose ". 4. Li-nso " the eye ", lit " that part of the body which can be veiled" is connected with ku-sia " to be veiled, dark " (285). 5. 1-sutno '' a spear ", lit. " that which disappears within (the body) ". From su " notion of disappearing " which we find in ku-snana " to disappear within one another ", and mo " inside " (384 (4 and 7)). 6. 1-/ua " a bone ", lit. " a dead member ". -From ku-fiia " to die ". The element /« (Herero /u) may be said to convey the meaning of " death " almost in every Bantu word in which it is found. 7. /-bue " a stone ", lit. " that which falls, is heavy " (in all probability). From ku-ua (Angola ku-bua) " to fall " (cf. 462*). 8. /j:ulu " the sky " lit. " that which is stretched out "- Related to the passive form -u/ua of the element -ula which forms expansive verbs (1080). VI. — jpbe Bu-MA Class and tlje Sub-Classes connecteD toitl) it. 440. — ,The class BU-MA contains the substantives which require the same sort of concord as hw-aio " a canoe ", plur. ma- «^. We connect with it the sub-class MA without singular (Ex. taSi-nzi " water "), as well as the sub-class BU without plural (Ex. hu-su " flour "), and the sub-class BU-ZIN (Ex. in Nyamwezi ■w-a(o " a canoe ", plur. ny-ato *). 441. — The reason for connecting the sub-class MA withouti singular with the clas^ BU-MA is that the same words which * EXAMPLES. the face grass fermented drink beer, wine • the brains flour Sin%. Plur. ' Tonga bu-sio, ma- bu-izu bu-kande (.'') bu-ongo bu-su Bisa ... bu-nga Gogo u-su ... ... ' . . u-sagi Sagara ... ma-nyari u-gimbi w-ongo (?) u-sagi Shambala ... uw-ongo ' Boondei m-aiii uw-ongo u-nga Taita u-xu ma-nyasi ... w-ongo u-nga Nyanyembe w-ixu ma-swa- bw-alvya w-ongo wu-su Sukuma ... w-ongo (i") u-su Kamba u-iio, n-zio ... ... ... Swahili u-so, nyu- ma-jani u-ji « gruel » w-ongo u-nga Pokomo u-so ...' (w-idzi «green») ... u-nga Nika ^u-so ... V u-ji ongo u-nga Senna ma-so ma-u-dzu bu-adua w-ongo (?) u-fa "Karanga ... ... wu-ftua wu-kube ^i") wu-rubi .. Ganda ma -so bu-so m-alua bu-ongo bu-ta Zulu-Kafir U bu-so u ty-ani u tyw-ala ', •I" ... Xosa-Kafir ubu-so u ty-ani u tyw-ala ... Hereto '/'V *" ... u-ruvi ... Bihe w-ongu u-tepa w-ongo Mbunda ... bo-ambo bo-ala ... ... Rotse mopo (=nia-upo) ma-lupo o-loi o-nga Guha . .. ... ... ... u-xie (?) Rua ... i^a-lovu ... ... Angola ... (u-isu« green ») u-alua ... ... Lower Congo ... ... ... ... Yao ... ... ma-nyasi u-tulua u-tutu u-tandi KUimane ... ma-ane u-alua ... u-to Mozambique w-ito ... ma-nyaxj ... u-koko ... Chwana proper ma-i tlho bo-jang bo-jalwa bo-koko bu-pi Suto ma- tlo bo-jang bo-jalwa bo-koko bu-pi Mpongwe o-ju, a- am-ani ... ... ... Dualla b-oso, mi- bi-ulu (?) liia-u ... ... Fan ... ... b-ut Fernandlan bu-so ... ba-u ... ... The BU-MA Class. 97 require the classifier BU- in Tonga are found to require the class- ifier MA- in a certain number of the other languages. § I. Forms in the Class BU-MA and the Sub-class MA-. 442. — The classifier MA-, both in those substantives which havp no singular, and in those which require BU- in the singular number, is essentially identical with the MA- of class LI-MA. Hence the same variations of its forms, viz. ma-, a-, me-, am-, va-, etc. 443. — But in some languages we meet with this peculiar phenomenon, that between MA- and the stem of the substantive BU- is retained under one form or another. Hence the following plurals, apparently irregular : in Angola moia " boWs " ( = ma-u-ia, EXAMPLES. (Cpntinued.) niglit a bpat (names of countries) Tonga Bisa Gogo Sagara Shambala Boondei Taita Nyanyenabe Sukuma Kamba SwahUi Pokoino Nika Senna Karanga Ganda Zulu-Kafir Xosa-Kafir . Herero Bihe Mbunda Rotse Guha Rua Angola Lower Congo Yao Kllimane Mozambique Chwana proper Suto Mpong-we Dualla Fan Fernandian iSiitg. bU'Siku bu-siku wu-zjku u-ziku u-tuka, I U'^siku u-siku u-siku, wu-siku u bu-suku u bu-suku o u-tuku u-teke bo-rike o-siko u-fuku u-sikua u-suku fuku, ma-tio u-hiu bo-sigo bo^igo o-gwera, b-ulu jDO-chio (?) Sing. bu-ato, bu-ato, Plur, ma- ma- w-ato, ny- w-aho wu-ato o w-ato borato w-ato w-ato u-kula u-lungu (?), lungu, ma- w-ato, ma- b-ote(?),ma-b-ote ow-aro, b'-olo, bi-al b-ato, am- bi- (?) Bu-ionga, Tongaland U-gogo, Gogoland U-sagara, Sagaraland U-xambala, Sambaraland U'-nyamwezi'ythe Nyamwezi country U-kamba, Kambaland U-ngujai Zanzibar U-zungii, the Portuguese territory Wu-karanga, Karangaland. Bu-ganda, the Ganda Empire 0-lumbu (?), the Rotse Empire Uzrua, the Rotse country Bo-rwajihe countryof the Bushmen 98 South-African Bantu Languages. sing, u-ta, in Herero o ma-u-ta, in Nyamwezi ma-wu-t'a, etc.), in Senna ma-u-dzu " straw ", in Nyamwezi ma-wu-ziku {}) " nights " (sing, wu-ziku), etc. , 4^4:4!. — N. B. Were it not for this last example, in which the stem has two sylla- bles, I should see a new application of the laws relative to monosyllables (44, 413, etc.) in this fact of the retention of the classifier BU- after MA-. But then it should .be said that the classifier MA- (and the same might be said of the element ZI- in the classifier ZIN-) is not so intimately united with the stem which follows it as to have a single accent {44) and to form rigorously a single word with it. MA- should therefore rather be considered as a sort of adjective preceding its substantive. 445. — The classifier which has the form BU- in Tonga has the same form in Bisa, Bemba, Subia, Ganda, Kafir, Lojazi, etc. In Kafir this classifier changes to ty- or tyw- before vowels, according to n. 122, as in a ty-ani " grass " {= a bu-ani), and apparently to EXAMPLES. (Continued.) ■wool clay life. water fat Tonga bu-oya bu-longo bu-umi ma-nzi ma-futa Blsa -1 ... mensi ... Gogo ... ... ma-renga ... Sagara ... u-longo u-gima meji ma-futa Sharabala ... u-longo ... ma-zi ma-vuta Boondei ... u-longo u-gima ma-zi ma-vuta Taita ... ... u-zima meji ma-vuta Nyanyembe w-oya wu-lolo wu-panga m-inzi ma-guta Sul^uma ... tn-inzi ma-guta Kamba u-wea u-ima ma-nzi ma-uta Swahili u-dongo u-zima ma-ji ma-futa Pokomo ... ... ma-dzi ma-fuha Nlka ... u-longo u-zima ma-dzi ma-fuha Senna u-bwea ma-dzi ma-futa Karanga ... wu-penyo ... ma-futa Ganda bu-iza (?) bu-mba bu-lamu ma-dzi ma-savui Zulu-Kafir u b-oya u bu-longo u b-omi a ma-nzi a ma-fiita Xosa-Kafir u b-oya u bu-longo u b-omi a ma-nzi a ma-futa Herero ma-inya ... omeva ma-Je Bihe ... o-tuma Q) ... va-va vete Mbunda ... mema ma-ze Rotse oia o-toko (?) me-i ma-fe Guha ... ... ma-ji ... Rua ... ... u-umi (?) mema ma-ni Angola ma-vunzu menia ma-ji Lowef Congo w-ika ' ... ma-za ma-ji Yao u-mbo (?) u-tope u-umi mesi ma-uta Kilinaane ma-taka ... ma-ije . ma-kurra Mozambique ... u-kumi ma-xi ma-kura Cliwana proper bo-boea bo-raga bo-tshelo metse raa-fura Suto vbo-ea / bo-raga bo-tshelo metse ma-fura Mpongwe o-mwa o-mbona ... a-ningo a-gali Dualla ... ma-diba m-ula Fan b-oka ma-chi ... Fernandian bi-riba(?) •■ ba-isopa (i") — . bo-opi bi-ta The BU-MA Class. ' 99 /- in the plural form of the same word, in the Swahili ma-j-ani ( = ma-wu-ani), etc. 446, — Other forms are : — BO- in Chwana, which changes to bo-j- (= bo-bo-, 202) before vowels, e. g. in bo-jang " grass " (Mpongwe am-ani), bo-j-alwa " beer " (Mbunda bo-ala, Nyamwezi, bw-alwa, etc.), as if, the origin of the_; for bo- (n. 202) in such words having been forgotten by the Chwana, they had restored bo- before it, either for the sake of uni- formity, or to prevent all doubt as to the proper class of the same words. For the same reasons such Kafirs as h^ve only a half-knowledge of their language say some- times u bu-tyvr-ala "beer ", instead oi u-tyw-ala. , AiAt7, — WU- regularly in Karanga, Nyamwezi, etc. 44'8. — U- {w- or uui- before vowels), with a sort of spiritus asper, in Herero, Swahili, Nika, Senna, Angola, etc. . 44Q. — C- (tfK'- before vowels) in Mpongwe, 41so in Rotse and Nyengo, unless Livingstone's spelling was influenced by Chwana, when he wrote his notes on these languages. 450. — It is dropped in Congo, where however it is retained under the form w- before vowels. 451 . — N- B. I. Bleak's remark {Comp. Cr., p. 273) that this particle is sometimes elided in Kafir and ,Chwana does not seem to be correct. The error comes from not noticing the change oibu or bo- to iy,j, etc., before vowels. 3. The proper form of this classifier in Dualla, Fan, and Fernandian, is not evident from ' the documents I have come across. § 2. The Sub-classes BU without Plural and BU-ZIN. 452. — Of the words which have the classifier BU- by far the larger number have no plural form, because they express pro- perly a sort of collective or abstract notion. They form the sub- class BU without plural. 453. — The sub-claiss BU-ZIN exists only in Swahili and in a few other languages. N. B. The origin of the class BU-ZIN in Swahih comes from the confusion of the class BU-MA with the class LU-ZIN, through the fall of ^ and /in LU- and BU- (86 and loo). Hence the word u-ta " a bow" is mentioned by Krapf as having two plurals, viz. : ma-ta and ny-u-ta (za...) ; u-so " face " is said to have no other plural form than ny-u-so, etc. However u-siku " night " has only its regular plural ma-siku. This reminds one that the expression " at night " is rendered in Tonga by ma-n-siku, and in Senna by ma-siku. Cf. 556. § 3. Substantives which belong to the Class BU-MA and THE Sub-classes connected with it. 454. — The substantives which have no other classifier than MA- are principally those of fluids or quasi-fluids, or again of things too South-African Bantu Languages. which melt naturally, as ma-nzi " water ", ma-lidi " sour milk ", ma-tanana " snow ", etc. N. B. I. In several languages, principally in Chwana, MA- is often used to express great namber. Ex. ma-biise " many horses " (cf. di-pitse (390, 39S) " horses "). 2. In Senna the classifier MA- is regularly used to form names of actions from ' applicative verbs. Ex. ma-li'miro " agriculture ", from ku-lima " to cultivate the ground" (applicative : ku-limira (1065)), ma-fambiro " a journey ", from ku-famba " to go ", (ap- plicative : ku-fambira), tic. 455. — The nouns which have the classifier BU- are principally those of : — 1° Things which ferment, or generate bubbles, as bu-kande and bu-koko " beer, wine ", bie>-su " flour ", bu-longo " wet cow-dung." and " pot-clay ", bii-ongo " the brains ", bu-loa " the blood ", bu-ci "honey ", which Kafirs make into a fermented beverage, bu-tale " iron ore ", bu-la " the bowels ", bust " smoke ", bu-e/e " small pox ", etc. 456. ' — Things which come into being or grow to light collec- tively and by gentle heat, as bu-ana buenkuku, " chickens ", bu-izu " grass ", etc. 457. -^ 3° " The night, " bu-siku; " the face, " bu-sio, and those feelings of the sbul which transfigure the face, as bu-botu " a good face ", i. e. " happiness " and " kindness ", bu-bi " an ugly face ", etc. 458. — 4° Authority or empire, bu-ame ; whence the sphere itself of authority, as Bu-ganda " the Ganda Empire ", etc. 459. — 5" " A canoe, " bu-ato ; " a bow, bu-ta, " etc. § 4. Etymologies. — Varia. •400. — The etymology of the classifier MA- has been suffi- ciently studied in the preceding article. With regard to the classifier BU-, it is pretty certain that it implies generally readiness to reaction and imnsformation, whether by fermentation, as in " beer ", " flour ", etc. ; or by hatching, as in " a brood " ; or by smelting, as in " iron ore " ; or by a fresh start, as in " the grass " so easily refreshed, and- in " night ", which to a Kafir mind is nothing else than the universal silent renewing of nature after " the fall of the sun " ; or by transfiguration, as in " the face ", the mirror of the soul, and in "authority"; or by plasticity, as in "clay", and in "cow-dung", which is generally used by Kafirs for plastering their huts ; or again by elasticity, as in "a bow"; or even by readiness for a change of The BU-MA Class. loi position (?), as in " a canoe ", etc. All this supplies plenty of sug- gestive materials for the study of the association of ideas. This classifier BU- in many words is unmistakably related to the verb -bumba {— bubua, cf 285) "to work clay, to shape" (cf, 52*). And' it probably is to -ua {-bUa or -gua) " to fall " (462*) what the classifier LI- is io-lia " to eat " (430). Hence it seems to mean primarily " that which falls, which cannot stand upright or firm ". ■4>61« — Coming back to the examples under n. 440, we may rtiake the following statements : — 1. In dti-st'o " the face " the element -sio is urimistakably the same as that which gives us li-fiso "the eye ", pi. meso (439 (4)). Hence the proper meaning oibu-sio must be "the place of the eyes ", or more exactly " that which falls over the eyes ". In Senna, Chwana, etc., the word for "face "is no other than that which means " eyes ". 2. Bu-izu " grass ", lit. "that which comes out (of the ground) ". Cf. 409 (2). 3. Bu-alua " fermented liquor ", lit. " a thing for bewitching ". From -lua or -loa " to bewitch ". In ma-/uvu {■perhaps m-aluvu), which is the jivor.d for " palm-wine " in several dialects pf Angola and the neighbouring countries, the element ■•?'« seems to add to bu- \ alua the notion of " vomiting ", or that of " foam " produced by fermenting liquors. 4. In bu-ongo " the brains ", and bu-longo " clay, cow-dung ", etc., the element ngo con- veys very probably the notion of " something, soft " ; but I do not see exactly what notion is conveyed by the elements oand lo. In bu-longo, however, the element 7(7 means probably " that whichis sprea,d" (Cf. 384 {9, 8, and 4)). "\ 5. In bu-siku " the night " the element si is related to the verb ku-sia " to be veiled, to be dark ", and the element ku to the adjective -kulu " great ". Hence this word means lit. " great darkness ". The first element si is replaced in some languages by the element su, notion of" disappearing", in others by/» (Herero tu), notion of" death ". Hence the words bu-suku, bu-fuku, and u-tuku, 6. Bu-su " flour ", lit. " that which is pulverized and rendered almost invisible. "From the element sti, notion of " disappearing ". 7. B-u-dto " a canoe ", lit. " a thing for ferrying across ". From a " of .", and to, notion of " carrying ". Cf. ku-td'la " to carry "- 8. .S«-^o«fa " Tongaland " (cf. 36s (5)). g. Bu-umi " life ". From the adjective -umi " alive "- 10. Ma-nzi " water ", lit. " that which fecundates " (384 (10) ). Karanga renders " water " by i-vura, which properly means " rain ". Thus it is perhaps of all the Bantu languages the only one in which the word for " water " is not of cl. MA. This peculiar exception is probably due to the custom, common to several South-African tribes, of not pronouncing the names of revered peiisons nor any of their principal parts. For I notice that the Chwana word for " water", metsi, enters into the composition of a quasi-sacred national name of the Karanga, viz. Ma-iafia-rao\si, lit. " the Water-elephants (sea-cows) ", whence the' well- known word Monomaiapa (= Mu-ene wa Mafapa(meisi), lit. " the Lord of the Water- elephants", which was the title of their King or so-tailed Emperor. In connexion with this it may be mentioned that the hippopotamus is a sacred animal with the Karanga even to this day, and that their reverence for it has passed to their conquerors the Ma-tebele, or Ma-tabele. See Kerr's " Far Interior ", p. 20. 11. Ma-futa "fat "lit. " that which melts and is sticky ". From the element /», notion of" dying " and ia, notion of" adhering to\.., sticking to... (?) ". VII. ^^ TF&e Ku-MA ciasis. 462. — This class includes the infinitives of verbs used as sub- stantives, as y^«;/«a; "to die, death", and also in many of these languages a fe\y other substantives which require the same sort of concord a.s 'ku-iui " an ear ", plur. txia^iui*. N. B. In the materials at my disposal there is no evident trace of the classifier KU-, not even before infinitives, in Dualla, Fernandian, and Nywema. § I. Forms, 463. — In the words of this class the classifier MA- is identical with MA- of the preceding classes LI-MA andBU-MA. 464. — JV. £. A. few words in Herero^ and a larger number in Ndonga, are men- tioned by Bleek (Compar. Gr., p. 207) as keeping the particle KU- in the plural together * EXAMPLES. an ear an arm a foot to fall, a fall ^ing. Plm. Sine- , Plur. Sing. Plur. Tonga ku-tiii, ma- Iju-bpko, ma- ku-ulu. ma- ku-ua, or ku-gua Bisa ku-tui, ma- ku-boko, ma- ku-ulu. ma- ... Gogo . . ■ ... ... ... ku-kagwa Sagara (467) ... ... • .• ku-gw£t 3haiiibala ... ... ku-gwa Boondei ""(467) ku-lume « right hand » ... ... ku-gwa Taita ... ... ... ... ku-gwa Nyanyembe ku-tui, ma- ku-kono. ma- ku-gulu, ma- ku-gwa Sukuma ku-tui, . ma- ku-Kono, ma- ku-gulu. ma- ku-gwa Kamba ku-tu, ma- ku-boko, ma- ku-u. ma- kii-waluka Swahlli • •■ ■>. ^ ... ... kw-anguka Pokomo •■> •■• ku-guu, ma- ... Nika • •> . .. ... ... ku-bwa Senna (467) ku-gwa Kaplanga ku-tabeso. ma- ... Ganda ku-tu, ma- ... ... ku-gwa Zulu-Kaflr ... ... u ku-wa Xosa-Kafir ... ... u ku-wa Herero ku-tui, ma- ku-oko. ma- ... ... ku-ua Blhe kw-oko, va- ku-ulu « leg» ku-wa Mbunda ku-tui, , ma- ku-boko, ma- ... Rotse ku-toe, / ma- k-oko ... ... ... ku-koa (.?) Guha ku-tue, ma-(?) ku-boko. ma- ku-gulu, ma- ... Rua ku-twe, ma- ku-woko. ma- ku-ulu, ma- ku-fiona (i") Angola (467) (lu-ku-aku) maku ... ... ku-bua Lower Congo ku-tu, ma- k-oko. m- ku-lu, ma- bwa Yao ... ... ... ... ... ku-gwa Kilfmane (ny-arro),m-arro ... . .. u-ogua Mozambique (ny-aru),m-aru ... ... . .. u-lua Chwana proper ... ... ... go-wa Suto ... ... ... . . . go-wa Mpongwe o-roi, a- o-go, a- o-golo a- poxwa DuaUa ... ' ... ... . .. ... ko Fan a-lo, molo a-bo e-ko (?) Fernandian — , ba-to 1 ba-kole ... ... The KU-MA Class. 103 with MA, e. g. o ku-ti, " field ", pi. o ma-Ku-d. Here again I notice that their stems either are monosyllabic, or begin with a vowel (cf. 44, 413, etc.) 465. — The forms of the classifier which is KU- in Tonga, Bisa, etc., are : — KU- in Karanga before ordinary substantives only, U- before infinitives. Xf/- in Congo before ordinary substantives only, dropped before such infinitives as begin with a consonant, though retained before the others under the form kw-. GO- in Chwana, where it is found only in infinitives. 0- in Mpongwe before ordinary substa:ntives, In this language infinitives are appa- rently not used as substantives. U- ( W- before vowels) in Mozambique and Kilimane, where it is used only before infinitives. It is replaced by hya (= ni-a-) in nya-ru, or nya-rro, " an ear ". I 2. Substantives which belong to the KU-MA Class, 466. — In the larger number of the Bantu languages the words" which fall under this class are exclusively : a) Infinitives (used as substantives) ; 6) the few parts of the body mentioned in the prer ceding examples (462*) ; c) the names of such rjvers as are con- sidered as being " the arms " or "shoulders " of others, or of the sea, as the rivers Ku-bango, Ku-a or Ku-ba, Ku-anza, Ku-nene, etc. 467. — ^- B. Those languages which have lost the classifier KU- in ordinary substantives have however retained traces of it, at least most of them. Thus in Senna the ,word for " ear " is kutu of class LI -MA, pi. ma-kuiu, where the syllable kti- is evidently the primitive classifier. Cf. in Angola lu-K\L-aku " an arm ", plur, mdku, not ma-)sM-aku, and in Kaguru ghutwe " an ear ", pi. ma-ghutwe, in Boondei gutwi, pi. ma-gutwi, etc. § 3. Etymologies. — Varia. 468. — The etyniology and, exact power of the classifier KU- offers no difficulty. It is originally identical with the locative class- ifier KU- (542, 563), and essentially connected with the verb kula " to grow out ", as also with the corresponding adjective ^ulu (Kamba hJ) " full grown ". Hence its adaptation to those parts of the body which grow out of the main trunk, as : — 1 . Ku-tui " an ear ", lit. " a thing protruding from the head ". From mu-ine " the head ". The change of the final e tozis caused by the transposition of the accent (»«■/«/, H-tu), and this transposition is itself due to the fact that the classifier MU- is naturally short (mie), while the classifier KU- is, naturally \on^ (kii). Cf. 271. 2. Ku-boko " an arm, a shoulder ", lit. " a thing protruding downwards at the side ". From the elements bu " notion of falling " and ko " notion of side ". 3. Ku-ulu " a foot", lit. " a thing protruding flatwise". From the element ulu " notion of something stretched out ". Cf. 439 (8). In the infinitives of verbs KU- properly refers to the notion 6T time or place (563). Hence ku-ua " to fall " means lit. " when (or) where one falls ". VIII. — ypfje Lu-(z)iN Glass anlr tje Buti^classes connecteD Wtt it, 469. — In these we classify together all that refers to the va- rious categories of substantives which have in the singular number the classifier L1J-. There is comparatively little agreement between the various Bantu languages in the use of this classifier. Some use it as a diijninutive, others as an augmentative, others both as a diminutive and as an augmentative, etc. All this causes a great diversity in the formation of the plural *. § I. Transformations of the Classifier LU-. 470. — The classifier which is pronounced LU- in Tonga, Bisa, Gaiida, etc., is pronounced : — 'i*- Dt/- after n in the same langua,ges and in Karanga (z86). HI/- regularly in Karanga and Herero. * EXAMPLES. the tongue A rope or string Stn^. Plur. Sing. Plur. Sing. Plur. Tonga lu-limi, {r.;is'„i Lu-izi, the Middle Zambezi... lu-ozi (?) in-gozi (.•■) Bisa lu-limi, in-dimi ... ... ... lu-sisi ... Gogo lu-limi, ma-limi lu-enga « a river », ... ... Sagara lu-limi, lu-ko ongo « a river », ma- 1-uzi, s-uzi Shambala lu-limi, n-dimi ... ... lu-gole. gole (?) Boondei u-limi, , n-dimi lu-tolongo « a river » ... lu-zigi. zigi(?) Taita lu-mi ^. , ... ... Nyanyombe lu-limi, n-dimi ... ... lu-goye. n-goye Siikuma lu-limi, ma-limi ... ... lu-goye. n-goye Kamba .•■ ... u-tsi « a river » - ,,, SwahlU u-limi, n-dimi ... ... ... ■ ■• • •• Nika lu-rimi, n-dimi ... ... lu-goe. n-goe Karanga ru-rimi, in-dimi ru-izi « a river », nj-izi ... ... Ganda lu-limi, n-dimi ... ... ... lu-goi • •• Zulu-Kafir u lw-imi,ilw-imi u Iw-andle « the sea > ... u-tambo «a snare>,in- Xosa-Kaflr , u Iw-imi, i Iw-imi u Iw-andle « the sea » u-tambo ^ anet;^, in- Herero ■ ■• ■ *■ ru-rondo « a rivulet J , tu- ru-sep£i « a thread » o tu- Bihe • •* • •< lu-wi « a river », o on-dwi lu-ndovi lo- Mbunda lo-lime (?) ... • ■• 1-ozi • ■■ Rotse lo-leme (?) ... ... ■ *> 1-osi (?) ... Guha lu-limi, n-dimi ... ... ... ..• ... Rua lu-jimi, n-jimi lu-wi « a river » 4.> ... Lower Congo lu-bini, tu-bini ... ... ... lu-kamba ... Yao lu-liiT^i, n-dimi lu-sulo « a river », n-sulo lu-goji. n-eoji Chwana proper lo-leme, dl-teme lo-tsitsi, « a watercourse >, di- lo-tlwa « a net >, di- Suto lo-leme, di-teme ... ... •*■ ... Mpongwe o-iieme, i-neme ol-obi « a river », il- o-goli. i- Bualla i-yeme, lo-yeme ... ... ... ... Fan ••• ... u-dsui « a river » ' ... ... Fernandian lo-belo ... ... ... The LU-(z)iN Class. 105 LO- in Chwana (n. 200), as also in Rotse, Mbunda, Nyengo, and Lojazi, if here again Livingstone's spelling has not been influenced by Chwana. In Fernandian we find both LO- and LU-. / O- in Mpongwe. U- ( W- before vowel) in Kamba and Swahili, according to nn. 81 and 88 ; and likewise in Nywema. 471 • — N- B. I. According to Bleek (C(?»«/. Gn, p. 237) the form of this classifier . is LA- or 2;A- in Kele (Di-kele). The examples given are la-ngoko " head ", pi. ma-ngoko ; \a-pa^a " hoof", pi. m.a-pa\aj la-nyui " honey-bee ", pi. nyui; la-nya\a " a flee " ; la- ndmgn " the end ", pi. tna-HdmgO. ,, , Bleek adds (p. 271)^ that "in Timneh (a semi-Baritu language spoken near Sierra Leone) rope-like or creeping plants have commonly the prefix ra- in the singular. '' 4:72. — 2. This classifier is dropped commonly in Ka!fir ; for, in such words as u Sana " a baby '', u-siba " a feather ", etc., u is not the classifier, but the article. In the same language it is retained under the form LW- before such stems as begin with a vowel, and under its proper forrn LU- before monosyllabic stems (cf. nn. 44 and 325, 368, 389,. 413). 473. — 3- This classifier LU- has almost entirely disappeared from Angola, Senna, Lojazi, etc., and apparently altogether from Mozambique and Kiliraane. iln Dualla it is regularly replaced by DL or LL of cl. LI-MA. § 2. Plural Classifiers corresponding to LU-. 474. — In the formation of the plural we meet with more variety in this class than in any other. 475. — Bleek thinks that the plural classifier which corresponds properly as plural to LIJ- is the classifier TU- (of class KA-TU). But this opinion is unwarranted, as we find such correspondence only in Herero, Ndonga, and Congo, to which may be added the Dualla group, that is, precisely in those languages which, having practically given up the classifier K A- as the regular diminutive Qlassifier (cf, 509, 5,22), replace it in many cases by LU-, and which, separate themselves on many other points from the generality of the Bantu languages. 476. — N. n. In Dualla, TU- changes regularly to L0-, according to n. 220. Ex. lo-yeme " tongues " (Congo tn-bini). 477. — In the other languages, the classifier corresponding as plural to LU- is commonly (ZI)N-, as in Tonga, Bisa, Nywema, Karanga, Ganda, Swahili^ Nika, Kafir, Chwana, etc., with the variety of forms which has been described in nn. 393-398, and with those various effects upon the initial letter of the stems of substantives which are regularly produced by n nasal expressed or suppressed. (Cf. 51 and 72,, 77, 83, 140, 93, 95, 184-196, 389, 395, 396, etc., etc.). ! io6 , "South-African Bantu Lmtguages. •4; 78. — yV. B. I. In some cases the particle I'u- is kept partially or totally in the plural, and combined with the classifier (ZI)N-. This causes some remarkable phenomena. Thus, in Kafir we have i Iw-imi, " tongues ", t Iw-andle " seas ", etc., whicb require the same concord as if we had i(zi)n-lw-andle, i (zi)n-l'w-imi, etc. Ex. i hti-imi e zixa-bini " two -tongues ". Likewise, in Kaguru we find the following plurals su-gka " plots of ground " (sing, lu-gha), su-ti " shafts " (sing, lu-it), s-uzi or ny-uzi " strings " (sing, l-uzi), s-umo or ny-umo " razors " (sing, lu-mo), etc., all of which require the same concord as if they contained the classifier Z1N-. (Cf. Last's Kaguru Crdmmar, pp. ii and 15, 17, etc.). It may be further remarked that in all the preceding examples the stem of the substantive, either is monosyllabic or begins with a vowel. Hence these phenomena seem to be due to an extension of the general laws concerning monosyllables (cf. nn. 44 and 325, 368, 389, 413, 464, etc.). Cf. Pfere Delaunay's Grammaire Kisuahili, p. 11. 4!79. — 2. The effects of n nasal, expressed or suppressed, upon the initial consonants of the stems of the words are more easily studied in this class LU-(ZI)N than in the class IN-(ZI)N, because here we have no longer the nasal both in the singular and in the plural number, but only in the plural. Thus we see plainly how under the infiuence of « nasal expressed or suppressed — a) in Chwana / changes to / Ex. &-&»z« " a tongue ", pi. rf/-fe»z« S » 11 ^S 01' ^ Ex. Ip-gonz " a piece of wood ", pi. rf/-fto«g sh „ „ ch Ex. lo-chu " death ", pi. din-chu b n ,> P Ex. lo-badi " a scar ", pi. di-paro sa gagwe " his clothes " ( = se-atla, se-aparo). EJ- before vowels, E- before consonants, in Mpongwe. /- or E- in Oualla, Benga, etc. 4:93. — N- -5- I. The proper form of this classifier in Fan is still doubtful. It seems to be ECH- before vowels, e. g. ech-um, " a young man ". Perhaps it is E- before con- sonants, as in Mpongwe. 4Q4. — 2. Bleek mentions also the form VI- in Kele and Benga. But this seems EXAMPLES. (Continued.) the chin the chest, (a bone) (a cob, a bunch, etc.) Sing. Plur. Sing. Pli, zi- Bisa ... ... ... ... ... ..'. Gogo ... ... ... ... Sagara ki-levula, vi- ki-fa, . vi- ... Shambala ki-dezu, vi- ki-fua, vi- ... Boondei ki-ev'u, vi- ki-fua, vi- .-• Taita ... . .. ... ki-konzi « a cob of com » ... Nyanyembe ■ ,;..,<.. ... ki-kuwa, fi- ... Sukuma * ... ... ... ... Kamba ki-nyezwa, i- ki-sivi. i- ki-tsakwa « a cob of com », i- wahili ki-devu. vi- ki-fua. vi- ki-tawi « a bunch >, vi- Pokomo ki-yefu, vi- • •• ... ... N-lka ki-refu, vi- ki-fua (?), vi- ki-guta « a cob of com >, vi- Senna ci-debzu, bzi- ci-fua, bzi- ci-konje « a bunch >, bzi- Karanga ci-revo, }wi- ... ... Ganda ki-levu, bi- ki-fuba, ' bV- ... Zulu-Kafir i si-levu, zi- i si-fuba, i zi- i si-kwebu < a cob >, i zi- Xosa-Kafir i si-levu, i zi- i si-fuba, i zi- i si-kwebu « a cob "i, i zi- Herero tyi-hehemeno ovi- ty-ari. vi- . Bihe ci-yeli ..." - •'■■ Mbunda ... ... ci-tea (.?) ... Rotse ... se-foba (.?) ... Guha ... ... ... ... ... Bua ... ... ci-kupa (?), vi- ... Angola Lower Congo kexu, (?) ki-fuba " bone ", i- ki-lende < a bunch J>, i- bobo, pi. Dobo vixi " bone". pl. vixi kangi « a bunch >, pl.kangi Nywema Yao chi-mbundi i- ... chi-sondei- Kllimane e-kua«the chest»(.''),vi- Mozambique ... ■ *• ... i-konyo, « a bunch », i- Chwanaproper se-ledu, di- se-fuba, di- se-gwere «a cob of maize»,di- Suto se-lelu, di- se-fuba (?), li- ... Mpongwe ... ... e-pa « bone », pl. pa e-goro < a cob », goro Fan ... ... n-cuk(?), ? ... Dualla ... e-isi « bone », be- e-sambu « beard of corn»,be- Fernandlan ... ... e-aka, bi- (?) ... The ci-zi Class. Ill to be an error, because properly speaking the Kele class VI-LA and the Benga class VI-L' correspond to the Tonga class KA-TU, not to CI-ZI. (Cf. 522). 4!90- — 3- In Mozambique this class of nouns seems to have melted into the same with class IN-(ZI)N. Hence in this language the form I- (Y- before vowels) in both npmbers, as in i-iu " a thing, things ", y-o-lia (= i-a-u-lia == Senna ci-a-ku-lia) " food " lit. " thing for eating ". This is a result of the phonetic laws(i76). § 2. Transformations of the Classifier ZI-. , 496. — The principal forms of the plural classifier of this same class are : — -'Z.WI- in Karanga, and Yeye of Lake Ngami (cf. 109). BZI- in the Tette- dialect of the Senna group (cf. 99). DZI- or B'2^1- in the Shire dialect of the same language and in Senna proper (cf.99). ZI- in Tonga, Kafir, Mbunda, etc. EXAMPLES. (Continued.) a stump a dried hide a detached hill ' or mountain an ant-hill a light-hole Tonga Bisa Gogo. Sagara Shambala Boondei Taita Uyanyembe Sukuma Kamba SiivahUi Pokomo Wika Senna 1 Karanga Ganda Zulu-Kafir Xosa-Kafir H erero Bihe Mbunda Rotse Guha Rua Angola Lower Congo Ny-wema Yao Kllimane Mozambique Gh-wana proper Suto -Mpong-SB-e Fan Dualla Fernandian SzKg: Plur. ci-samo, zi- ki-zibi, vi- •ki-siki, • vi- ci-banda,bzi- ki-kolo, bi- o tyi-pute, o VI- ki-xinji, i- xinza, pi. xinza cbe-singa, i- i-kokolo, i- se,-sipi, di- se-sipi, li- e-tenge(?)be- si-udi, bi- Sing. Plur. ci-kanda, zi- ki-kanda, ... ki-ngo, vi- ki-kanda « a bag tt ki-chingo, vi- ci-kuruo, bzi- i si-kumba,i zi- i si-kumbo,i zi- ci-kanda, se-tumba, ki-sewa, ki-seva, ki-ba. chi-kopa, i- e-banda,banda Sing. Plur. ci-lundu, zi- ki-gongo, ki-rima, ki-lima, ki-lima, ki-fumvu ki-gongo, ki-gongo, ki-ima, ki-lima, ki-rima, vi- ci-dunda,bzi- o tyi-hungo, o vi- ci-lundu (.?) kundubulu chi-tundulima, se-tlhaba, di- se-tlhala, li- Sing. ci-olu, Plur. zi- ki-bumbuswa, fi- ki-umbi, i- ki-suguli, vi- ki-so, ci-uru. VI- bz- ki-wo, bi- i si-duli, i zi- i si-duli,- i zi- o tyi-tundu, o vi- o ci-mu, o vi- se-bukomolo(?) ki-nsama, chi-kula, se-olo, di- se-thlaga, li- ej-imba, y-(.?) Sing. Plur. ci-bonebone, zi- ki-tonia, i- za, sa (.?), ki-tuli, bi- i si-roba, i zi- i siFroba,i zi- otyi-tuo(.')... se-iponi, di- se-iponi, li- 1 1 2 South- African Bantu Languages. Dl- in Chwana, spelt Ll-'va. Suto and in some other Chwana dialects (cf. 9 and 175); VI- in Swahili, Shambala, Nika, Herero, Guha, etc. N. B. In Yao it is also VI- according to Last, but Hetherwick spells it I-, while Steere spells it FI-. It is also spelt FI- in Nyamwezi. If this form be correct, it may be noted as being so far the only plural classifier which contains a hard consonant. BI- in Ganda and Nyambu,. BI- or BE- in Duallaand the neighbouring languages. /- in Angola, Mbamba, Kamba, etc. Y- before vowels, suppressed before consonants, in Mpongwe. § 3. Substantives which belong to the Class Cl-Zl. 497. — The substantives which fall under this class in Tonga, and in the generality of the Bant^j languages, are principally : — ' 1° The names of languages, as Ci-tonga " Tonga". (Cf, 484.) 498. — 2° The word for " a thing " ci-ntu, and some substan- tives in which this word is understood, as ci-tede " such and such a thing", ci-fiila-mabue "a hailstorm ", lit. " that which forges stones, " ci-indi " the past ", lit, " that which is remote ", etc. 499. — 3° The words for any sort of limited break, or cut, on land or wat^r, or on a body, as ci^knle " a national mark or cut " (such as circumcision for Kafirs, filing between the front teeth for the Herero, etc., cf. 50), ci-bongo " a small lake ", ci-sua " an island in a river ", ci-to " a ford in a river ", ci-vukumba " an opening in a rock, a cave ", Ci-ongo or Ci-ongue (in Chwana pronunciation Si- ongd) " the great Zambezi falls ", ci-limo " summer " lit. " the break in the work ", from -ttma " to till the ground " (cf, 52*) ci-liba " a well " ,-ci-bonebone " alight-hole ", etc. 500. — 4° Whatever is what the Tonga call " short ", i. e. relatively thick in one part and small in another, or halved, or protruding with a thick basis and to a comparatively small height, etc., as ci-kulukulu " a man stooping by agq ", ci-embele " an old person oi" animal ", ci-yuni " a bird with short legs ", ci-binda " a land tortoise ", ci-pembele " a rhinoceros " (short legs), ci-tapile " a potato " (from the Dutch aard-appel), ci lezu " the chin " ci-zui, " a knee ", cyi-ni " the liver ", ci-popue " a cob of maize ", ci-lala " a young palm- leaf ", ci-lundu " a hill ", ci-pami" a half ", ci-kalo " a saddle ", ci-bula " a seat ", ci-kanda " a hide, a shield ", ci-longo " a wide earthen pot '', ci-tungu " a low-hut ", ci-zumbo " a nest ", ci-sama " a low table ", ci-tale " a candlestick ", ci-lapo or ci-ldo "apaddl?", etc., etc. The ci-zi Class. 113 501. — N. B. In Congo the class KM (= Tonga CI-ZI) is the regular diminutive class. (Cf. 521.) § 4. Etymologies. — Varia. 502. — The Tonga and the Karanga still bear in mind very distinctly the proper meaning of the classifier CI-. They render it invariably by the English word " short ", or by the Dutch " kort ", and say it is identical with the adjective -ce " short ". But when they explain their mind, it can be easily made out that they attach to it in some cases a negative or privative, and in the others an intensive power, and that in many words it might be rendered by the -adjective " thick ", rather than by the adjective " short ". Thus, while it has a negative power in ci-ntu " a thing ", lit. " that which is no person " (cf. mu-ntu, " a person "), and a privative power in ci-panzi " a half ", ci-tungu " a low hut ", ci-sanza " a low table ", etc., it may be said to have rathe;r an intensive or enlarging power, at least from our point of view, in such words as ci-pempere " a rhino- ceros ", ci-romdo [Senna word) " a lord, a wild beast ", etc. This intensive power is further associated with a productive or causative notion, as in ci-lezu " the chin " lit. " that protruding part of the body which produces beard" (cf. indezu = in/ezu" beard "). Father Pedro Diaz, S. J., has noticed the privative and the intensive meaning of this classifier in his Angola Grammar, p. 32 (Lisboa, 1697), and explained' their connexion by saying that CI- (KE-, KI-) is essentially negative, but that negative expressions may convey both privative and superlative notions, as " no-man ", for instance, may signify both " less than man " and " more than man ". Cf. 634. More probably the classifier CI- has two dififerent etymologies, and this is the true explanation of its different powers. The first CI- may originally have been identical with the word which means " ground " in nearly all the Bantu languages (Tonga ■mu-se, whence n-si in pa-nsi, ku-nsi, and mu-nsi, Swahili n-chi or n-ti, Angola^, Congo «-xe or n-ci, Herero erhi, etc., cf. 533*), and it is from this meaning of " ground " that is has derived that of " something low, short, on the ground ", as also, on the one hand, that of privation and negation, and, on the other, that of production. The second CI- may originally have been identical with the Karanga word ;i;e " chief, lord ", whence its augmentative power, 114 South-African Bantu Languages. principally in Karanga, Senna, and Yao, as in ci-rombo " a wild beast ", lit. " a wild lord " ('). Analogies and phonetic laws seem to point to a relation of the classifier CI- to the verb -da " to dawn " (52*), but it seems hard to associate the notion of " something short " with that of " dawn ", unless it be said that a thing short is only a beginning or remnant of something, exactly as the dawn of day is a beginning of day and the end of night. Cf. 994. The classifier ZI- (Karanga ^ze/z-, Swahili m-) is probably related to -vula (= -vut/a ?) " to multiply ", which is itself derived from -sua, -vua or -va " to come out " (409(2)). Hence it conveys the notion of number without including that of the manner in which multiplication is obtained. Cf. 408. Possibly the elements vu and izi are closely related to one another in Bantu, as they both convey the notion of fecundity or development. Bleek thought that the original form of this classifier was PI- ('"). But this opinion cannot stand with the fact that its modern forms contain no such hard letter as P. Cf. 496. 503, — In the examples under n. 491 : — 1. Ci-ntu " a thing " seems to mean lit. " that which is no person". Cf. mu-ntu " a person ". 2. Ci-tonga " the Tonga language "- It might be asked how we can find in such names of languages the notion of " ground " which we consider to have been conveyed originally by the classifier CI- (502). We answer that in such words the classifier CI- takes from the idea of " ground " only the notion of something which is the basis of all the rest, which always remains, which is characteristic, so that, for instance, Ci-tonga means lit. " that which is characteristic of the Tonga ". A less probable explanation of such words would be that which would refer them to ci-kule " a national mark ''. 3. Ci-lezu " the chin ", lit. " the ground of the beard ". Cf. 409(2). 4. Ci-bula " a seat, a stool ", conveys the notion of something bent over itself. Cf ku- bola (Kafir u ku-buya) " to return ". Ki-ti, in Swahili, means lit. " a stump of wood ". Cf m-ti " a tree ". 5. Ci-fua " the chest ", or " a thick bone ", lit. " a ground of bones ", in opposition to the more fleshy and muscular parts of the body. 6. Ci-samo " a stump "- Cf mu-samo " a tree ". 7. Ci-kanda " a dried hide ", in opposition to i-kanda and lu-kanda " the skin ". The element -anda conveys the notion of " covering, protecting ". Cf. 409(7). 8. Ci-lundu " a hill ". Cf i-lundu " a mountain ''. The element lu, here reduplicated, conveys the notion of " something stretched out ". Cf 439(8), 468(3), 489. 9. Ci-olu " an anthill ". Here again the element lu conveys the notion of something raised, but I do not see what notion is conveyed by the o before it. 10. Ci-bonebone " a light-hole, a window ", lit. " a hole for seeing ". From ku-bona " to see ". Cf. 52*. 1. The natives of Senna consider wild beasts as the embodiments of their deceased chiefs, and consider themselves bound to feed them. 2, Comparative Grammar, p. 264. X. — JPlje tiiminutiiae Class KA-xtj anti tfie Sub^Classes connecteD toitf) it. 504. — Though the privative class CI-ZI may in some respects be considered as diminutive, yet, properly speaking, in the larger proportion of the Bantu languages such things as are small in every respect are found to take in the singular number the classifier K A- and in the plural the classifier TU-, as ka-/5«« " a small dog ", plur. \Xi-bua *. Those languages which do not agree with Tonga on this point, do not agree any better among themselves, some of them having the classifier FI- or VI- in the singular, others on the contrary using VI- with a plural meaning, others forming their diminutives by suffixing or prefixing the word for " son ", muana, etc. § I. Forms. 505. — A single glance at the subjoined tables will show that more information of a reliable kind is still wanted, However, here .A EXAMPLES. a baby (a youth) a stick, a branch the opening of the mouth Sing. Pl-ur. Sing. Plur. Sing. Tonga ka-cece, tu- ka-sanjo, tu- ka-nua Bisa ka-ana, tu- ... ... ka-nua Shambala ka-zana (youth) vi- ... ... Boondei ka-zatia (youth), vi- ka-nua Nyanyetribe ... ka-tambi. tu- ka-nwa Sukuma ka-gosia. tu- ' ... ... Kamba ka-ana. tu- ka-munsa ... ka-nyoa Swahm ... • >■ ^ (kanwa, cl. IN) Nika ka-dzana, vi- ka-nwa (?) Senna ka-wzOT-ana, tu-K/ana VsL-m,u-\\, tu-z«/-ti ... Karanga ka-ana, .tw- ... ... ... Ganda ka-ana, bu- ka-ti. bu- ka-mwa jNyambu ka-ana, j tw- • .. ... Herero ka-natye, u- ka-ti. o -u ■ ■• Bihe ... ... ... ... Mbunda ... ... ... ka-nwa Rotse ka-uzi (?) ... ... ka-nwa Lojazi ... ... • .. ka-nwa Guha ^ ... ... ka-nya, tu- Rua ... ... - >■. ka-nwa Angola ka-»zona. tu-ana (?) ka-*?«-xi, \\x-mi-yX LoTver Congo ... ... ... ^ ... Nywema ... ... ... ... Yao ka-anache. tu- ka-pichi. tu- ka-mwa Kele ..• Fan ^ ■ \' ' ... ... ■ •• Fernandlan si-neneheh, to-(?) s-aka, tw- ii6 South-African Bantu Languages. are a few conclusions which can be drawn pretty safely from the documents at hand, viz. : -^ 506. — 1° The regular diminutive classifiers are KA- in the singular, TU- in the plural, in the larger number of the Bantu languages, viz. in Tonga and all the dialects which may be grouped with it (Bisa, Subia, Bemba, Lea, etc., n. 65), in all the dialects of Nyamwezi (Nyanyembe, Sumbua, Sukuma, etc., cf. 73), in Yao, Kamba, Karanga, Guha, Regga (near the Mut'a-nzige), Luba, Lojazi, Angola, etc. 507. — 2° A few Tonga words, instead of taking in the plural the classifier TU-, require, or at least admit, another collective classifier. Ex. ka-ntabua " flees ", pi. bu- ; ka-ana ke inkuku " a little chicken ", pi. tu-ana tue inkuku or bu-ana bue inkuku; ka-bue " a pebble ", pi, tu-bue or lu-buebue, etc. 508. — N- B. 1° The use of BU- as plural to KA- seems to be the rule in Ganda {Grammaire Ruganda^'^.it). However it may be noticed in l^aai's" Polyglotta" (p. 160) that in Nyambu, which is a language closely akin to Ganda, the classifier used as plural to KA- is not BU- but TU-. Ex. ka-lumbu " sister ", pi. tu-; ka-atia " child ", tw-. EXAMPLES. (Continued.) the middle, ■ the centre (a match, embers. a little fire) an axe Sing. Sing. P}ur. Sing. Plur. Tonga ka-ti ka-lilo " a match ", tu- ka-ngone, tu- Bisa Shambala ... ,,, ... Boondei ... ... ... Nyanyembe (ga-ti?) ka-lilo"" embers " tu- ka-virunana, tu- Sukuma ... ... ... Kamba ka-ti ka-foka, tu- Swahill ka-ti \ka..) ... ... ... Nika ka-hi (.?) ka-dzoho " a little fire ", vi- ka-dzoka, vi- Senna >.. ... ... ... Karanga ka-ti ... ... ... Ganda ka-ti ... ... ka-badzi, bu- Nyambu ■ ... ... ... Herero (ka-ti (?)) ka-parua " a match ' ou- ... Bihe ka-ti Mbunda ka-ti (?) • •> > •• tu-ya"fire» ... Rotse ka-ci tu-via"fire" Lojazi ... tu-ya " fire » Guha ;.; karhia " a little fire ", pi. tu-hia ... Rua ... . .. ka-solo, tu- (?) Angola (xaxi ka ...) ... .. tu-bia " fire " ... Lower Congo ... ti-ya " fire " ... Nywema ... ... ... Yao ka-ti ka-wago, tu- Kele v-eya " firewood ", 1-eya vi-ondshi, 1- Fan : ■.. ka-ba " a flame ", do-a (do-ba T) Fernandian ... si-so (?) " a flame " ... ... ... The KA-TU Class^. 117 509. — 2° Again, in Herero the classifier U- (= Tonga BU-) is considered as the regular plural of KA-. But here two points are to be notice4 : a) KA- is by no means in Herero the regular diminutive classifier. Any one who will peruse Dr. Buttner's " Mahrchen der Ova-Herero " in the " Zeitschrifi filr afrikanische Sprachen " (1887-1888) will rather find that far more diminutives are formed in Herero with the classifier RU- than with KA-. i) Even such substantives as admit the classifier KA- are found to be treated as if they had another, names of things being treated as if they had RU-, and names of persons as if they had MU-. Ex. 'k.A-tio iM-horoti " a long stick " (" Zeitschrift ", p. 1 89), ]ii.a.-kurukase ■xxapendukire " the little old woman got up ", etc. 510. — 3° In Nika the classifier TU- is replaced by VI- of class CI-ZI, probably because according to Nika phonetics the plural classifier TU- should be pronounced HU- (cf. 93), which might create confusion with the singular classifier U- {== Tonga BU-). In Shambala also we find VI- instead of TU-. But more information is required on this I'anguage, as it seems that even in the singular number the Shambala classifier KA- is practically identified with KI- (= Tonga Cl-). Ex. ka-zana ka 'ki-£oxi " a son ", (Last's 511. — 4° In the language of the Gabiin River and the like, what we pronounce TU- in Tonga is regularly pronounced LO- or eo- (cf 220-230). 512. — 5° In Senna and Angola the classifiers KA-TU have kept the regular form, but in most words they allow classifiers between themselves and the stems of their nouns. Ex. in Senna : ka-m-beni \= ka-mu-beni) " knife " pi. tu-tni-beni; in Angola, ka-m-bika (= ka-mu-bika cf. 279), pi. tu-a-bika, etc. In one case in the Shire dialect of the Senna EXAMPLES. (Continued.) a little bird a small dog a pebble a second time Sing. Plur. Sing. Plur. Sing. Plur. Sing. Tonga ka-yuni, tu- orbu- ka-bua. tu- ka-bue, tu- ka-bili Bisa ■ ... ka-wili (?) Shambala ka-ndege. vi- ... ka-iwe, vi- ka-ili (?) Boondei ... ... ka-kuli, vi- ... ka-idi Nyanyembe ka-noni, tu- ka-bwa, tu- ka-we, tu- ka-wili Sukuma ka-noni. tu- ... ka-wili (?) Kamba ka-nyuni ka-iwia k-ele Swahili .. . • <• (kawe, cl. IN) ... Nika ka-dzuni, vi- ka-dya ka-dziwe, vi- Senna ... >.. • .• ... ka-wiri Karanga ka-nyuni, tu- ka-ja (?) ... ka-bwe, tu- ka-biri Ganda ka-bwa. bu- ka-inja, bu- ... ka-bili (.?) Nyambu ka-nyuiii, ..., ka-wili (?) Herero ka-fera, u- ka-ua. u- Blhe ... ka-»zbwa(?) ,otu- ka-we, tu- Mbunda ka-tari ... ka-bari (?) Rotse • •■ ... ... ■ •• ka-yeri (?) Lojazi ka-|ela. tu- ka-tari ■ •• ... Guha ... . •• ...,tu-wiri(?) Rua ka-yuni (?) tu- (?) ... . ka-biji (?) Angola ka-»zbua. tu- ka-?-«-tari,tu-?«a-tari ka-yari Lo-vsrer Congo ... ... ... ... Nywema fi-ulu. tu- ... ... Yao ka-juni, tu- ka-wa, tu- ... ka-wili Kele vi-noni, lo- ... • •■ ... ... Fan ... . .. ... Fernandian si-nOdi, to- ... ... ... ... 1 1 8 South- African Bantu Languages. group I find TU- changed to TI-, viz. ti-ana " children " (sing, ka-mu-ana) (Nyanja New Testament, Mat. II, i6). Cf. 517. 0I3. — 6° In Swahili I can find no evident traces of the plural classifier TU-, but I find traces of KA- used as a classifier. Ex. Ka-jua ni \i.a.-pi? " Where is the little sun?" (Krapf s Diet., p. 125); ka-ndia 'k.a.-dogo " a small path " (Krapf,'p. 128). However, it seems that, when the particle KA- forms diminutives in Swahili, it is oftener used as a mere prefix than as a f/awyf^*-, as in ka-we. Kx.iaweyA... " a little stone o(..,",notka-we'is.a... I give in the preceding tables the Swahili word ka-nwa " the mouth ", but I have no knowledge of its ever being used as a word of cl. KA. The same applies to the same word in Boondei and Nika. 514!. — In Mozambique the prefix KA- is not a classifier. It forms substantives of the sub-class BA (346, 350 and 527, 517). 515. — 3° In Tonga many diminutives, principally names of animals, are formed with the compound expressions ifl-«/a... pi. tu-nga..., lit. " little son of.., little sons of... ", in which the syllable nga is either a contraction for mu-ana (cf. 332), or a particular form of its stem -ana and then the noun following ka-nga or tu-nga keeps its regular prefixes. Ex. ka-nga sekale " a little musk cat, " lit. " a little son of musk cat ", pi. tu-nga ba-sekale ; ka-nga sulue " a little hare ", pi. tu-nga ba-sulue, fete. 516. — ^- B. I. Somewhat similar expressions are met with in Herero, with this difference, however, that ka in such Herero expressions acts as a mere prefix, not as a classifier (cf. 347, 509). Ex. ka-ha-vandye " a jackal ". 517. — 2' In Senna many diminutives are also formed by using as a sort of prefix either the word mu-ana " son " in its full form, or the particle nga- (alias nya-) which seems to be a contraction for it. Ex. mu-ana-mbua " a little dog ", pi. ana-mbua; mu-ana- mpuru " a calf ", pi. ana-mpuru ; mu-ana-mpeyq " a little stone for grinding " {mpeyo, alias pheyo = a grinding-stone). Such words as take the prefix n^a or nya seem to be rather diminutives of politeness than real diminutives. Ex. nya-rugue " a tiger ", etc. (cf. 349). This manner of forming diminutives and their particular use without any real diminutive meaning is common to several other languages, and is to all appearances borrowed from the Oriental languages, in which we continually meet with such expressions as " son of death, son of error, son of the house, son of Babel, son of a hundred years ", etc. Cf. in Mozambique the prefixes KA, NA (344, 349). 518. — 3- In Chwana and Kafir, as also in Rotse, diminutives are also formed by using the word for " son " under the various forms -ana, -nyana, etc., but here, instead of being used as prefixes, these forms are on the contrary used as suffixes. Ex. in Kafir : u mf-ana " a young man " (from um-/o"a man "), u m-?itw-ana or u m-nt-ana " a child " (from u m-ntu " a person ") ; in Chwana : ntlw-ana " a little house" (from n-tlo "a house") etc. Further, in the adaptation of such suffixes to the stems of the nouns we meet with all the various phenomena which have been previously described (nn. 202-206 and 127.) 510. — 4- In Herero and Yao the suffix -tye or -che (= Tonga -ce, cf 593) is appended to some diminutives, or even forms them by itself. Ex. in Herero : o mu-a-tye, pi. o va-natye, or o ka-na-tye, pi, tt-na-tye " child, children " ; in Yao ka-ana-the, pi. tu-ana-che " child, children ". 520. — 4° In Nywema we find the Tonga classifier KA- replaced by FI-, which evidently is radically identical with the The iKA-TU Class. ug Tonga adjective -fui " short ", cf. 6oi. Ex.Ji-u/u " a little bird ", pi. tu-fulu (Last's Poly^-^., p. i86).. 521. — ^- B. I. This classifier FI- is also found in Lower Congo, but apparently without a plural (Bentley, p. 536); In Congo the regular diminutive class is KI-I (= Tonga CI-ZI). 522. — 2. It is evidently the same classifier wich is found in Kele under the form VI-, in Dualla and Behga under the form VI- before vowels only, I- before consonants. Ex. in Kele vi-noni " a bird ", pi. /o-noni (cf 494}, vi-ondshi " a hatchet " pi. l-ondshi. < 523. — 3- In Fernandian the same classifier'has the form SI-, thus being identical with the singular classifier of the preceding class CI-ZI. Ex. si-iuki " a fly ", pi. to-iuki; si-nodi " a bird " 'pX. tu-nddi, etc. § 2. StJBSXANTIVES WHICH BELONG TO THE KA-TU ClASS AND THE Sub-classes connected with it. 524. — Unmistakably only such substantives fall under this class as express true diminutives from a Bantu point of view. Such are : — 1° Points of separation of various things, as ka-ti " the very centre or middle of a thing '', ka-kokola " the joint of the arm ", &a- ango " the centre of the breast ", etc. 525. — 2° Things which are not only low or short, but compara- tively small in every dimension, as ka-nyamankdla " a little animal ", ka-pamba " a little baby ", ka-samo " a branch, a stick, a quite young tree ", ka-nvua "a thorn ", ka-nyenyezi " a little star ", ka- sua " a small island ", (cf. ci-sua, " an ordinary island "), ka-ciocio " an ear-ring", ka-langulango " an ear-ring ", ka-lilo " a rriatch ", ka-longo " a cup ", (cf. i-longo " a high earthen pot ", ci-longo " a low earthen pot ", bu-longo " pot-clay ", mu-longa or mu-longo "■ a muddy river "), ka-ngone " a small' axe,", ka-sako " a small poisoned arrow ", ka-simbi " a nail", etc, 526. — .We must also consider as belonging to the class KA- such words as ka-mue " once ", or " the first time ", ka-bili " a second time ", ka-tatu ". a third time ", etc For though, from a European point of view, we might consider them as adverbs, they are never- theless true substantives from the Bantu point of view. In Ka!fir and a few other languages the classifier KA- has been retained exclusively for such words, and in these languages they may be said to have become adverbs proper. 130 South-African Bantu Languages. § 3. Etymologies. — Varia. 527. — The diminutive classifier KA- is probably the element from which is formed the verb -inka " to start " (Kafir mkd). There is no need to explain how this notion of " mere determination or departure " is very naturally applied to the starting point of a thing, and to things that are in their first stage of formation. This etymology throws light on another fact, viz. the peculiar use of the prefix KA- in Mozambique, NKA- in Kafir and Senna, before several substantives of the class MU-BA or of the class IN-ZIN. For it may be noticed that such substantives, when they are not diminutives, are principally either those of animals remarkable for their " rapid starts ", or the like. Ex. in Mozambique ; ka-lamu " a lion ", pi. a-kalamu (in Senna nka-lamu, pi. (zi)n-kalamu), ka-pwiti " a gun ", pi. a-kapwiti, ka-rumia " a messenger ", pi. a-karumia, ka-mruxo " sensation ", etc. In some words the diminutive classifier KA- reminds rather of the verb -kala " to sit, not to move " than of the verb -inka. 528. — The plural classifier TU- is probably derived from the verb -iuia or -iota " to take, to carry " (Kafir -twala), exactly as the other //«ra/ classifiers are respectively derived from the verbs -bala, -ziala, -mala or -mana, and -vula (-zuilaf). Hence it is that we find it used almost exclusively for such things as are taken up, and, as it were, pluralised by the hand, such as tu-samo " branches ", tu-simbi " nails ", etc. This may even be the reason why the word ka-ntabua " a flea ", pi. bu-ntabua, and the like, borrow another classifier than TU- in order to form their plural. It may be noticed that this is of all the plural classifiers the only one which has a hard letter in the generality of the Bantu languages. 529. — The examples given under n. 504 may be explained etymologically as follows : — 1. Ka-cece " a baby "- The reduplicated element ce means " short, small ". It is essen- tially identical with the classifier CI- (502). 2. Ka-samo " a branch, a stick ". Cf. mu-samo " a tree '', 384(7). 3. Ka-nua " the opening of the mouth ". I have never heard this word myself in Tonga, I take it from Livingstone's Mss. It seems to be related to the verb -nyua (Senna -mwa) " to drink ". Possibly it is related to li-tio " a tooth ", pi. meno. It may therefore be that it means lit. " the opening through the teeth " or " the opening for drinking ". 4. ICa-ti " the centre, the middle ", lit. " a point in the very ground (of a thing) " Related to muse " the ground " (Swahili n-ii or n-cAi). Cf. 384(7), and 502. 5. ICa-it/o " a match ", lit. " a small fire ". Cf. mu-liio " firtf ", 384(8). The Guha word The KA-TU Class. 121 ka-hia " a flame '', plur. tu-hia " fire " (Angola iu-bia, Rotse iu-via, etc.) is derived from -^ia " to burn " (52*). 6. Ka-ngone " an axe ". This again is a word Which I take from Livingstone's Mss. It must be related to in-kuni " wood ", and therefore signify lit. " that which goes through wood. " 7. Ka-yuni " a small bird ". The stem -yuni probably means " in the air ', from -ni " in " (553-555), and -yu, which is related to the stem -oya of mu-oya " the air ". 8. Ka-bua " a small dog ". The stem bua is onomatopoetic, being derived from the barking of the dog. 9. Ka-bue " a pebble ". Cf. i-hue " a stone ", 439(7). 10. Ka-bili " a second time ". Frorn -bili " two ", 792. XI. — Jiocatiioe Classifiers anlr ;^ Prepositions 530. — This is a subject which we must consider apart from European views concerning the cases of substantives in general and locatives in particular, because they would be an obstacle to a cor- rect perception of the Bantu mind. To explain myself, when we say, for instance, " it is dark in the house ", " he lives above me ", " he lives below me ", etc., we are accustomed to consider the expression " in the house " as a locative which has no influence at all on the verb "it is dark"; and likewise the words "above, below" are not substantives, but prepositions : otherwise we should say " above of me, below of me ", etc. On the contrary in the larger number of the Bantu languages such expressions as " in the house ", " above ", " below ", etc., are substantives of the same type as those we have examined in the preceding articles, and require after them the same constructions as if we had " the-inside-of-the-house ", " the-place- above ", " the-place-below ", etc. Thus we have in Tonga : — Ulu-ngaitda mu-ld-si'a, lit. " the-inside-of-the-house //-isdark ", i. e. " it is dark in the house ". CZ-kede k\i-tala kuangu, lit. " he lives the-place^above that-of-me, " i. e. " he lives above me ". l/-kede ku-«K k\l-angu, lit. " he lives the-place-down Ma/-of-me ", i. e. " he lives below me ". In all such sentences it may be seen how the locative elements ,MU- and KU- act as ordinary classifiers, requiring the expressions governed by them (mu-h-sm, ^in-angu) to be also determined by prefixes like themselves (MU- and KU-). 531. — It will, however, be seen further on that in some lan- guages these locative elements deviate partially from the nature of classifiers. Thus in Kafir we shall find pe-su/u kw-am " over me ", instead oipe-zulu pa-am, etc. 532. — In the generality of the Bantu languages the locative classifiers are three in number. In Tonga their forms are MU-, KU-, (P)A-. In several of the Eastern languages the classifier MU-, instead of being prefixed, is on the contrary sutfixed, and changed to -«g[ or -m or -ini (cf. 553). 533. — A good number of stems are susceptible of receiving Locative Classifiers and Prepositions. 123 the three different locative classifiers ; but then the change of clas- sifier produces a change of meaning, which seems not to have heeh sufficiently attended to in some translations of the New Testament. Thus in Tonga, for instance, three locatives are derived from the noun muse " earth ", viz. (p)a-«-i'?, ku-«-J'« and mu-«-«; but the meaning of the three is different, viz. (p)a-«J« = " on the ground, at the surface..." ; ]s.yx-nsi— " below", withanptioii of comparison; mu-«« = " fnside " (of some solid substance, such as the earth) *. § I. Transformations of the Locative Classifier PA-. 534. — The principal forms so far known of the first of these locative classifiers exhibit all the intermediary labial sounds between A- with a slight labial aspiration and PA-, viz. : — * EXAMPLES. doiHrn below within (beneath) upon above in the air T6nga (p)a-nsi ku-nsi mu-nsi (p)e-julu ko-julu mo-julu Bisa pa-nsi ... pe-iilu ku-e-ulu Gogo ha-si ... ... ku-chanya . •• Sagara ha-si su-nda-«j ... ... ... Shambala ha-xi ... ... ... . .. Boondei ha-si i-si (?) nda-» Nyanyembe ha-si ... kw-igulia Sukuma ha-nsi ... ... ... ... Kamba wa-si (?) ndi-«j ulu (it/a...) ku-ulu ... SwahlU ... lchi-»i,ti««i 1 nda-«i ... juu m-bingu-«2 Nika ... dzi-«« ... ... ... Senna pa-nsi > ku-nsi mu-nsi ku-zulu(?) Karanga pa-si feu-si mu-si pe-juru ku-denga Ganda wa-nsi mu-nda wa-gulu gulu Zulu "pa-ntsi e za-ntsi ... pe-zulu e-zulu-i«? Xosa pa-ntsi e za-ntsi ... pe-zulu ... e-zulu-2«/ Herero p-e hi kehi m-o ukoto Blhe ... ... k-iiu Mbunda ka-zi (?) ko-elo Lojazl ko-ilo Rotse ba-nje ku-inje ... ko-ilo Guha ha-nsi ... he-gulu , Rua ha-nsi ku-nsi he-ulu (?) ku-uiu (?) • .. Angola 1 b-oxi k-o xi m-o xi ... Congo va-nxi ku-nanxi mu-nxi Wywema ha-xi... la-xi . .. lu-ulu Yao pa-si ku-si ... pe-nani ... Kilimane (v)a-ti mo-ti« va-dulo Mozambique (v)a-thi mu-hi«3 va-chulu,va-zulu m-chulu Ch-wana proper te«B fa-tla-se mo-te«g go-dimo Suto te«g (ka tla-se) mo-te«j go-limo - Mpong-we gontye ... ... gw-igonu ... Fan e;dsi • •• e-yu(we-yu(?)) ... Dualla wa-si ... ... Fernandian ua-tshe(.?) /o-she (?) ... bo-ko(?) 124 South-African Bantu Languages. 535. ■ A- commonly, PA- after m nasal, in Tonga. A- commonly in Taita. Ex. ^.-ndu " a place ", a.-vuhi " near ". HA- in Subia, Nyamwezi, Mbamba, Nywema, etc. Possibly this is pronounced as A- in Tonga. 536. — WA- in Ganda, and in a few words in Kamba and Swahili. In a few other words in Swahili it has kept the form FA. Ex. ^^-hali ^-ote " in every place ", etc. In some other words both in Kamba and Swahili, as also in Nika and perhaps in Congo, this classifier is simply omitted. 537. — HA- in some Chwana dialects, the ZT being pronounced as a sort of hard labial aspirate. EA- in the other Chwana dialects (cf. ii). BFA- in Pokomo, according to the " Zeitschrift fur afrikanische Sprachen", 1888-1889, p. 164. The only example given for it is EXAMPLES . (Continued.) near (on the same level) near (on different levels) far, very far outside outride Tonga (p)a-fu(p)i ku-fu(p)i ku-le,kulekule (p)a-nze ku-nze Bisa ... ku-tali ... ... Gogo • •■ ku-tali ... Sagaira b-eiii ... ku-tali (?) ... ku-nje Shambala h-ehi ... ..■' ... Boondel h-ehi k-ehi (.?) ha-le Nyanyembe b-ihi (?) ... ku-le ha-nze" place" ku-nze Sukuma • .. ... ku-le ' ... Kamba wa-guwe ku-^ha o^zofku-atsa e-nsa (?) nsa SwahiU ka-ribu ... m-bali .;. wa-zi (?) nje Nika v-evi ku-re ... ndze Senna pa-fupi ku-fupi ku-tali ' ... ... ku-nje Karanga pa-fupi ku-fupi ku-re ... Ganda wa-mpi ku-mpi wa-la w-eru ku-sa Zulu ku-fupi ku-de pa-ndle ^-ndle Xosa ... ku-fupi ku-de,ku-de le pa-ndle e-ndle Herero ... ... ku-re (?) ... p-e ndye k-o si Blhe ... ... ... Mbunda ba-moheje(?) ko-lajalaja ku-ese Lojazi a-moyeye(?) ko-laja ... kua-lebu Rotse b-ebe . .. ko-re,korekore ba-nde Guha ha-buiyi . . . ku-le ... ... Rua h-epi ku-lele ... Angola ... ku-mbambu ... bu-kanga ... Congo va-na ndambu cu-na ndambu va-la J ku-na mbaji Nywema h-eni (?) . .. ■ *. lan-za Yao pa-ngulugulu ... ku-talika ...' pa-sa ku-sa KUlmane ... ... va-nje (?) ... Mozambique va-tama^ u-tai ( — u-tali) va-the ... Chwanaproper ga-ufe ... ... fa-ntle ... Suto ga-ufi . .. go-le fa-ntle (ka ntle) Mpongwe ba-raba ... gw-evungu (?) gw-igala Fan ... ... e-valg ... ... Dualla ... ... i.. ... ... Fernandian bi-ho ko-pie bu-sualo (?) ••• Locative Classifiers and Prepositions. 125 bfa-ntu " a place " (= Yaa pa-ndu, Sagara ha-ntu, Kamba va-ndu or wa-ndu, Taita «-«^«, Nika va-tu^ C)xvfB.na,felo=fa-elo, etc.) VA- in Mozambique, Nika, and Congo. N. B. I. In Congo the preposition NA (cf. 579) is generally appended to VA-. Hence the compound classifier VA-NA. 2. Concerning the suffix -ni or -«8, which is appended to some words of this class in Chwana, Mozambique, etc., cf. 553, 554. 538. — BA- in Rotse, and probably in Nyengo, perhaps also in a few words in Mpongwe. N. B. In Mpongwe the classifier PA- seems to be regularly replaced by GO-. Besides, in this language the mechanism of localtives has lost much of its regularity. 539, — BUA; or simply BU-, in Angola. 54;0. — PA- commonly in Karanga, Senna, Yao, etc. 'Nl B. In Herero it seems that the regular form of this classifier should also be PA. EXAMPLES (Continued.) between inside time or place) yesterday (last night) to-morrow (in the morningr) Tonga (p)a-kati mu-kati (p)a-mue (p)e-jilo (p)e-junza Bisa Gogo ha-li-gati ... Sagara ha-gati ha-mue ... ,ha-usiku (?) Shambala '.. . ... Boondei . .. ... ha-mue kelo-? Nyanyembe ha-gati m-gati ha-mo h-igolo Sukurha- ... ... . .. Kamba wa-kati kati (ya...) wa-mue 10 Swahlli wa-kati (?) kati (ya...) pa-moja ..- . .. Nika ndsL-fti va-menga dzana ... Senna pa-kati m-kati pa-modzi • .. ... Karanga pa-kati mu-kati pa-mue mpera pe-jiro (?) ... Ganda wa-kati wa-mu e-guro Zulu pa-kati ka-nye pe-zolo Xosa pa-kati ka-nye pe-zolo Herero p-o kati m-o kati pa-mue Bihe p-o kali m-o kati. Mburida ... ba-sindele (?) he-mene (?) Lojazi ' " - ?" he-mene (?) Rdtse ba-kaci. -"'... ' ■ "be-goro be-onda Guha ■ ■• ... Rua ... • •> ... Angola bu-a-xaxi mil-a-xaxi bu-a-moxi ■ •< ... Corigp va-na kati mu-na kati va-moxi e zono (?) Nywenaa • .• Yao pa-kati m-kati pa-ippepe ... Kilinaane v-arre m-arre *.. ... Mozambique v-ari e-ri-ari (ya) va-moka . .. Chwanaproper fa-gare mo-gare (ma-abane) (ka moxo) Suto fa-gare mo-gare Mpongwe go gare ... Fan ... ■ ■;>■ ... ... Dualla ■■•••■ i Ferna-Biiian ua-muela ''':''••■ .m-padi ... 126 South-African Bantu Languages. But in this language the articles «, o, are kept after locative classifiers. Hence the forms PE = PA-E, and PO = (PA-0). 541. — Concerning the mode of connecting this classifier with the stem, it may be remarked that in many words the non-locative classifier does not disappear altogether. Thus in Tonga we find (P)a-nst " down " = (p)a-mVL-se, from muse " the ground " 'i(/>)ey ulu "up" = (p)a-\\-julu, or rather /«-ij-«/« (cf, 256), from ij-ulu " the sky ". Cf 559. In fact the classifier /*/4- is joined immediately to the stem only when the same stem is that of an adjective, as in (p)a-fu(p)i " near ", from -fu(p)i " short " (cf. 601 *), (p)a-mue " together " from -mue " one " (cf. 792). EXAMPLES. (Continued.) bef ore,in front behind where ? whither? whence? In the house Tonga ku-ne-mbo mu-sule (p)a-li? ku-li ? mu Uganda Bisa ku-menso ku-numa ... ... ... Gogo ku-mwandp ku-mgongo ... ... Sagara ku-mwande ku-nyuma ho-ki ? ■ mu numba Shambala ..* ... ... ... ... 3oondei ... nyuma-? ha-i ? ku-i ? nyumba-«i Nyanyembe ku-mbele ku-huma ... ... mu numba SukUma ku-mbele ku-mpirimu ... ... Kamba ku-longuisia (?) ...' wa ? ... Swahlll ... ... wa-pi ? ... nyumba-«« Nika mbele nyuma ... ... Senna pa tsogolo ku mbuyu ku-pi ? m'nyumba Karanga ku-mbiri ... pi"(?) , ku-pi ? mu mumba Ganda mbele nyuma wa?(^wa-pi) ... mu nyumba Zulu pa-mbili e-mva pi? ... e ndl-j«t Xosa vni "in the road ". N. B. I. In this language, as in Chwana, this suffix -wj/or ni is also found after the locative classifiers VA and U. Ex. m-iva-nsi or u-wa-nii or va-nupa-ni " in the house " ; u-bingu-ni " in the sky '', etc. 2. The suffix -ni is replaced by -na in the word mu-hi-na, or m-ki-na, " inside " {== Tonga, mu-nsi, cf. 174). 554. — -E- + suffix -ini in Kafir. Ex. e ndlele-m " in the road " (= e ndlela-ini N. B. In Kafir the suffixing of -ini or -««g causes the various changes of consonants described in n. 122. Ex.: e inlonyeni"''m\\\& mouth" (from u mlomo "the mouth"), e mlanjeni " in the river " (from u mlambo " a river "), e ngutyeni " in the blanket " (from i ngubo " a blanket "), e zinsatsheni zam " among my children " (from insapo " the child- ren of... "), e tnahlwentsheni " among the poor " (from a mahlwempu " the poor "). 555. — Suffix -ni without prefix in Swahili, Nika, and Kamba. Locative Classifiers and Prepositions. 129 § 4. Plural Locative Classifiers, 556. — Strange to say, we find some appearance of a plural locative classifier. Thus in Swahili we find the wOrd for " place " rendered riot only hy fia-hali, but also by ma-kali. I cannot explain this otherwise than by saying that ma-kali was originally a sort of plural oi pa-kali, unless the prefix ma- in ma-kali be considered as being of foreign importation. Again, in Tonga, Senna, and in some other languages, we find the expression " at night " rendered by ma- n-siku (= Kafir e b-suku), from bu-siku " night ". This is either a plural form, or a contraction for mu-a-n-siku, which is not probable, 557. — ^'- -S- The Swahili vior&ma-hali is treated as if it had the classifier PA. Ex. TslLa-hali -p-ole " in every place ". § 5.' Effects of the Locative Classifiers on the other Prefixes of the Substantives. 558. — There is a great variety in the effects produced by locative classifiers on the prefixes of the nouns to which they are prefixed or suffixed, or vice versa. Let us just notice the most important : — 559. — 1° In Tonga and in most other languages the locative classifiers in some cases weaken the classifier MU- of classes MU-BA and MU-MI, as well as the classifiers (L)I-, (I)N- and BU-, often causing them to be reduced to the mere nasal n, but seldom to disappear altogether. Ex. : , — From muse " the ground " : A-n-si, Ku-n-si, Mu-n-st. N. B. In this example the further change of e to i is caused by the accent being displaced (cf. 271). ■ From mn-lilo " fire " : mu-n-dido " in the fire "- N. B. With regard to the change oiliodci. 285. iV directly causes the change of the first / to d, while the second /is also changed to d by attraction. From 'm-ganda " a house " : ku-nganda " towards the house " ; munganda " in the house ". From Uala " a sloping ground " : (pJe-Ma " on the side " ; Hu-fala " above " ; mu-tala " on sloping ground ". From \yulu " the sky " : (pjeyuiu "upon"; koyulu "on high"; moyulu "in the air". From bu-^<7/« " good land " : (^Ja-u-l>oiu " on good land ". Cf. ma-n-stku " at night " from hu-siku. TV. B. In Angola we find even tnu-&-lunga « in the sea ", from iLa-lunga " the sea ". Cf. Chatelain's Gr., p. 87. t36 Sduih- African Bantu Languages. 560. — 2° On the contrary, the locative classifier MU- is often weakened when occurring before the classifier MA-. Ex, u-manzi or m'manzi = in the water (cf. 279). 561. — 3° Something more remarkable is to be noticed in the application of the laws concerning monosyllables to which the use of the locative classifiers gives place. Thus it may be remembered how the law of avoiding monosyllables had given us in SwaihiWn-so " the loins " , n-ia or n-cka " a point ", n-tior n-chi " land ", etc., (cf. 389); and in Chwana n-ilu or en-tlu " a house ", n-ku or en-ku " a sheep ", etc., (cf. 392), instead of the monosyllables so, ta, cka, ti, cki, tlu, ku, etc. Now, when locative classifiers are prefixed or suffixed to these words, the initial n- or' en- is no longer required by the law of avoiding monosyllables. Hence the locative forms of the same words are in Swahili,not n-so-ni "in the loins ", \i\}X.so-ni; not n-chi-ni ox n-ti-ni " on the ground ", but chi-ni or ti-ni ; ^Xz.; and in Chwana, not mo-n-tlu-n^ " in the house ", but mo-ilu-n^ (in the Suto dialect tlu-n^), etc. § 6. On the Use of the Locative Classifiers. 562. — 1° In Tonga, and in the larger number of the Bantu languages, the locative classifiers serve to form those locative sub- stantives which correspond to most of our adverbs of time and place, such as " down, up, below, yesterday ", etc., etc., and to our compound prepositions, such as " be-fore, in-side, a-side, a-midst, with-in ", etc. Only, as has been mentioned above (530), and as will be more fully explained further on (755-764), it should be well kept in mind that from the Bantu point of view they are substan- tives, and that, consequently, when they are equivalent to such compound prepositions as the above, they generally require to be completed by various connective particles. The Tonga say, for instance: u-a-kala kunsi kua manzi " he remained under water ", not u-a-kala kunsi manzi. 563. — 2° The locative classifiers do duty for most of our simple prepositions ; then in most cases there is no objection to separating them from their noun. In Tonga, and in the larger number of these languages, (P)A means " on, flat on, close to, etc. ", thus expressing properly a relation of close proximity, as of things which are face to face. PA Locative Classifiers and PHpositions. 131 is also used when mentioning the determined time of an action. KU implies distance, or " receding from ", or again " coming from some distance to..- " It may be rendered according to the cases by " to, from, among, over, compared tp..., etc. ", MU means properly " in ". Ex. : — Oo4<. — (P)A. U-a-yala a buenga, he went along the edge of the water. Ba-a-mu-bika a mzt-^ilo, they put him over the fire. Ta ku-kpnduadilu-sek, no work is done on the day of the new moon. Ba-lia in-sima ejunzajunza, they eat porridge /« the morning. A mi-lia, on feast days. Ba-a-bika n-zoka mu-nkomo a mu-liango, they put a snake in a bag on the doorway. Ba-a-bika n-zoka a mu-biri, they put a snake round their body. 060. — KXJ.Mu-oyaua Zezauza ku ba-ntu, the spirit of God comes7«; men. Inyue-no mu-a-kaya koj'ulu. .., you who have gone /o heaven. Ba-a-ka tuba kq mu-iue, they turned white at the head i. e. , their hair turned white. Ba-lavu bakVi ba-bua ba-akue, lions are among his dogs. Ba-ana ba-ta toligua ku Burumbu, the children are taken to the land of the Rotse. Ba-kede ku Kafuefue, they live on the Kafuefue. 066. — MU. Tu-njizie ma-anza ta'manzi, let us put the hands into the water. U-la njila mu nganda, he enters into the house. Ba-sangu ta be-zi m'mumi, the sangu (kind of spirit) do not come into the town. U-a-fua mu nganda i-a-kue,^ he died in his house. U-kede mu cisua, he lives in an island. Ba-la kala mu mabue, they live in the rocks (in caves). 00 7 . — N. B. I. In Senna PA seems to be often used where the Tonga use KU Likewise, in Ganda WA (= PA) and in Congo VA (= PA) are often used where the Tonga would prefer KU. 568. — 2. Of course all these principles concerning the proper use of the locative classifiers are not much applied in the languages where the mechanism of the locatives is considerably, or even altogether, disturbed, such as Swahili, Chwana, Mpongwe, etc. § 7, Prepositions which are not Classifiers. 569. — There remain to be noticed a few particles which, having nothing of the nature of classifiers, may be considered as prepositions proper. These are : — t32 NA in Kafanga do. in Kafir do. in Swahili, etc ^£ or A/14 in Ganda NDIya. Senna (Shire, Tette, etc.) tiE (alias NLE) in Mpongwe.... LEm Chwana YA in Lower Congo Karanga i-a-ri ba-xano ba-tlA-ntu mue. ba-kabayana no Ngaru {=na-u Nguam). ba-nu be- J a ne noboro. fa-fupi na mi-riango iria. South-African Bantu Languages. 570. — 1° A connective preposition which means properly " with ". Often it renders our " and " before substantives. Its principal forms are : — A in Tonga (£,,0, by assimilation) Ex. ba-a-ka jana ka-cece a ba-nyena Maria. They found the child ■with his mother Mary! (Mat. 2, ri). baka bona nona na matnae Maria, (do.) babona u m-ntana no nina (= na-« nina). (do.). waka m-mona m-toto na Maryamu mama yake. (da. ) balaba o niwana ne Maryamu nyina. (do), naonakamwana ndi Maria amai ace. (do.) ■ufayexC ont^ivana lie Maria yi ngiye. (do.) bafumana niwana le mae Maria, (do). Yo mwana aku ye lekwa..., your child and his things (= ya nnvana ... ya,e lekua... 571, — Other examples : Tonga Tu-a-li ba-sano o u-mue, we were five with one, i.e. six. Ba-a-ka yasana a Nguaru, they fought with Lobengula. Ba-niu be-eza e in-iobolo, the people came with guns. A-fu' a mi-liango ilia, close by those holes. o72. — N. B. I. In Tonga I find this preposition A sometimes replaced by ANE, as if this were a more emphatic form. 2. In Karanga, Angola, Herero, etc., NA or NI changes to NE, NO, when combined with I, E ; O, U. Likewise, in Lower Congo YA changes to YE, YO, in the same cases. 573. — 2° A preposition which marks properly the instrument and the material cause. It may be rendered variously in English by " with, through, by means of, by, " etc. In Tonga and several other languages this preposition does not differ from the preceding. It differs from it in Kafir, Chwana, Swahili, etc. Hence its principal forms are the following : — /4inTonga(£,0,by assimil.) Ex. be-ense ba-tula i-sumo, bazo-o-Jua^ i-sumo. (Mat. 26, 52). Whosoever takes the sword, shall die of the sword. banu barire batura fumo, boofana.futno. (do.) onse awo omwe atenga mpeni..., adza mwaiika ndi mfeni... (do.) waodu wi btkng' okivara, wibejowo It 'okwara. (do.) boonse a bapete u tnkonto, baya kufa ngo mkonio (=nga a mkonto). bolle bacweren% sabole, batla bolawa ka sabole. (do.). [(do.) Tto wole watwaao upanga, wata.kufa kwa upanga. (do.)(Ct. twa-fa na n-daa, we are dying from hunger). 57'4. — ^. B. This preposition is frequently used before locative expressions in Chwana and Kafir. It seems then to convey the notion of " an interval " between two places, or that of" a certain direction " followed. Ex. in Kafir : Uye nga. pifia f (Chwana : Oik ka kae f) " Which way has he gone ? " 575. — Other examples : Tonga Ba-a-inka e in-zila im-pia (= a in-zila,..), they went by a [new road. U-a-fua e in-zala (= a inzala), he died/raw hunger. Yaka t^bu-longo, build with mortar. NA in Karanga NDI in Senna TIE {NLE) in Mpongwe. iV(?^ in Kafir ICA in Chwana NA and KIVA in Swahili. . Kafir ba-hamba nge ttdlela entsha =nga [i ndlela) wa-fa nge n-dlala (= nga indlala) yaka ngo bulongo (= nga-2<...). ' LocatiDe Classifiers and Prepositions. 133 N. B. In Sdnna the instrumental preposition NDI is sometimes replaced by the locative classifier PA. 576. — 3° An equiparative preposition which means " as, like ". Its principal forms are : — ANGA in Tonga. Ex. Mu-ade u-bede anga in-cefo (or ni-ncefo, cf. 583), the muade is like arsenic. INGA or KALA in Angola. Ex. ...inga be-ulu or kala he-ulu, as in heaven. NGA or NGA-NGA in Kafir. Ex. Lo m-fo u nganga lowo, this man is as big as that. JAKA in Chwana. Ex. Obua jakamogokve, he speaks like his brother. N. B. These particles are also used as conjunctions before verbs with the same meaning as above. 577. — 4° A possessive preposition which is practically equi- valent to our " of ". Its proper form is -A in all the Bantu languages, excepting Mpongwe and other languages north of the Congo. Ex. in Tonga : -a Leza " of God ", -a mu-ntu " of a man ", -a bu-longo " of mud ", etc. , , This preposition changes to -E or -O, according to the general rules of contraction and assimilation, when it happens, to be imme- diately followed by i, e, or \y^ u, 0. Ex. in, Tonga: -e in-^ombe " of -a cow " (= -a tn- gombe, cf 249), -0 liise, " of his father " (= « mse, cf 249). Ex. in Kafir : -e n-komo " of a cow " (= a i nkomo), -O m-ntu " of a person " (= a 2^ -m-ntu). Besides this, the possessive expressions thus formed are treated as if they were a kind ol determinative adjectives. Hence it will be seen further on that they are not immediately joined to the substantive which they determine, but are connected with it by a connective pronoun, such as u in the expression mi-cila yx-a nm-lavu, " the tail of a lion ", or i in the expression mi-cila \-a ba-lavu " tails of lions ", etc., cf 743. 578, — jV. B. I. In some Tonga proper names the possessive particle -a seems to be replaced by -na, as if this were a fuller or more primitive form. Ex. Si-n.a.-meja " Man (or father, or son) of tUsks ", Si-na.-mpondo or Si-a.-mpondp " Man (father, son) of horns ", etc. It may be that/etyrnologically speaking, the possessive particle -a is related to the connective particle a or na (570). 2. With regard to the use of the particles kua, kwa, ka, ga, etc., in possessive expres- sions, cf. 783. § 8. The Particles -LI, -NA, etc., in Locative Expressions. 579. — We often find in locative expressions such particles as -li, -na, etc., which might be thought to be prepositions, or parts of prepositions, but in reality are verbal forms equivalent to our " to 134 South-African Bantu Languages. be ", or " to have ". As they will be shown in their proper place (1040-1046) to have this value, it will suffice here' to sta,te the fact that, when the word which should immediately follow a locative classifier is a pt-onoun, or a substantive which has no classifier proper, such as Leza " God ", iaia " my father ", uso " thy father ", uzse " his father ", etc. (cf. cl. BA, 342), then in Tonga the copula /z (1025) is inserted between this classifier and the following pronoun or substantive. The Karanga use in almost all the same cases the particle na " to have ". In the same cases the Chwana use the locative pronoun g-o, and understand the copula after it. In Senna and Ganda the copula ti is used as in Tonga, but before a greater number of substantives. In Congo the particle na " to have " is used as in Karanga, but before all sorts of substantives ; etc. etc. 580. — Ex. Tonga Karanga Uakafugama kuli Leza, he knelt down to God. Ukede kuli uise (or kuli nguise), he lives with his father. ' Uaka inka kuXx imue nyika, he went to another place. , Muli Leza..., in God... Kafir : Mkulu kunawe, he is taller than you. SwAHiLi : ...kitna.ye, ...relating to him. Senna: Pidaficei paVisttlo..., when he came to the hare,... Congo : F«na kaii, between ; ;««na kati, inside, etc., etc., 1040-1046. § 9. Etymologies. — Varia. 581, — There is every reason to believe that the locative clas- sifiers belong to the most prirpitive elements of the Bantu languages. PA- conveys the notion of " opposition between two things ", or " their facing each other ", or " the application of the one upon the other ", and consequently of " close proximity ". It seems to be related to the verb -pa " to give ". KU- conveys the notion of " receding from, going aside ". It is related to the verbal suffix -uka, which forms neuter expansive verbs (1080), to the adjective -kulu " great ", " ancient ", and to the corresponding verb -kuta " to gxow out ". Cf. 468. MU- conveys the notion of " intimate union ", of " things which are within one another ". It is related to the adjective -mue " one ". Cf. 725. Hence its change to -ni or -tni, which has its parallel in the change of -mue to -n^ye in Kafir (122). Uakafugama kuna, Reja, Ugere pana. fate, Uakeja kun&mwe nyika, Muna, Rejd. Locative Classifiers and prepositions. 135 The etymology of several of the examples which have been given under n. 553 has just been explained in nn. 541, 559. We may complete here the notions there given.' 1 . Pa-nsi " down ", lit. " on the ground ", ku-nsi " below ", etc. From mu-se " the groiind ". It may be remarked that the word (p)a-fisi is generally used after the verb -kala " to sit " (Chwana -nna or -dula), just as we generally say " to sit down ", not simply " to sit " Hence the mistake of several scholars who give us such verbs as ku-kalansi, u-kalathi, u-kalati, etc. " to sit down ",when they should decompose .them into ku-kaVansi, u-kaPatki, etc. In Chwana the word /«-«j, which was originally identical with the Swahili ii-ni or chi-ni (= Tonga mu-nsi), has come to be used not only for the Tonga pa-nsi " on the ground ", as in go-dula ien% " to sit down ", but also, as it seems, as a purely expletive particle, somewhat like our " down " in vulgar English. And in the expressions koa tenq, ka fa iens, etc., it seems to mean " inside ", when we might rather expect it to be equivalent to the Tonga. pa-nsi or rather to the Kafir n£a pa-ntsi " downwards ". Perhaps this anomaly is only apparent, as it may be that in these expressions the word ieng does not answer to the Swahili iini or cAini, but to ndani " inside ", lit. " in the belly ", from i-dda or n-da " belly ". It rriay also be remarked that the Bantu pa-nsi has given to Chwana the word le-fatshe " the earth " (Senna pa-nsi), which at first sight might have been thought to have nothing in common with teng. This again shows what a mixed language Chwana is. Cf. 753. The Kafir word e zantsi " below " means properly " where it comes down ", from -za " to come" and n-isi (= n-si = Tonga mu-se) " the ground ". Its Chwana equivalent ka tla-se is formed in the same manner, as the Chwana verb -tla " to come '' is the equi- valent for the Kafir -za (173, 195). Here therefore the Chwana element which means " ground " is no longer ie as in te-n% nor tshe as in lefatske, but se. 2. \n pa-fu(p)i" near " the element Ju conveys originally the notion, of " death, the end of a thing ''. The meaning of the element pi is not clear. ' Considered in the light of the phoijetic laws it should be related to -pia " to burn ". Cf. 541, 6oi. 3. In (p)a-iize, ku-nze, " outside ", the stem nze means properly " approaching ground ". It is related to -za " to come " and to in-zila " a way, a path ''. 4. In (p)ejulu, kojulu " above ", etc., the word ij-ulu " the sky " means lit. " the open expanse ". Cf. 468(3), 503(8), etc. The verb -jula means " to open ". 5. (P)a-kati " between". From ka-ti" the centre ", 529(4). The Swahili wa-kati, which should be the equivalent for the Tonga, pa-kali; seems to have come to mean exclusively " a timg, the time of... " 6. (P)a-mue " together ". From -mue " one "■ y. (PJejila " yestevda.y ", more properly "last night", lit. "at bed-time", from i-io " bed ". The Kafir pe^zolo means lit. " at the time of stretching oneself out ", from . ku-zola " to stretch oneself out ". 8. (P)ejunza " to morrow ", more properly " to-morrow morning ". From the element ju, notion of" opening " (cf. ku-jula " to open "), and -za or iza " to come '', which implies the notion of " something future ". 9. Ku ne-mbo " in front ". From im-bo " the front side of the body ". 10. Mu-sule " behind ". The word i-sule " the back side " seems to be derived from the elements su, notion of " disappearing ", and le, notion of " length, distance". Cf. 439(5). W. (P)a-li? "Where?"" (whence probably pi?) leaves the thought suspended, and probably contains the classifier LI- with a reference to orientation, ii e. to an indefinite position of the Sun. Cf. 421, and 800, 808. Most of those prepositions which are not classifiers (569-578) seem to have been originally verbal forms related to the auxiliaries ya " to go " (911), enda " to go " (cf 918 and 939), &ala " to sit " (cf. 941 and 944), ^_^« "' to be inclined to... " (cf 995), etc. XII. -- Copulatitje "SttUvcs before Bu&stantiioes. 582. — Among the numerous manners of expressing the copula in the Bantu languages, most of which will be studied together in another chapter, there is one which is to be noted here, because in some languages it is a mere modification of the prefixes of the substantives. Its proper effect seems to be that of verbalizing nouns, i. e. changing them into expressions which have more of the nature of verbs than of that of substantives, as if we should say in English " this 6/eeds ", instead of " this (is) 6lood ". Its proper form in Tonga, and some other languages of the interior, is a mere nasal sound, m or n nasal, prefixed to classifiers. In some cases it is a full nasal syllable, viz. nga, or n^-u, or ni. In Kafir its form varies as the classifiers themselves. In Senna, Chwana, Swahili, etc., it has the same form before all sorts of nouns, etc. 583.- ■ Ex. : — Tonga Kafir Senna Cl. MU'- {m)-mu-ntu ngu m'-ntu ndi mu-niu it is a man. „ — ngu Leza ngu Tixo ndi Mu-lungu it is God. „ BA- ra-ba-ntu nga ba-niu ndi a-ntu those are men. MU?- {ta)-mu-cila ngu m-sila ndi n-cira it is a tail. MI- \tn)-mi-cila yi mi-sila ndi mi-cira those are tails. „ IN- m-n-^ombe yi n-komo ndi mombe it is a cow. „ (Z)IN- nzi-n-^ombe Z\ n-komo ndi (zi)n-gombe those are cows. „ (L)I- tx-di-ianga li tanga ndi tanga it is a pumpkin. MA- {va)-ma-tanga nga ma-tanga ndi ma-tanga those are pumpkins {or)nga:-ma-ianga BU- rorbu-kande bu fyw-ala ndi bu-adua it is beer. KU- {n)-ku-lia ku ku-fya it is food. LU- n-du-anja lu Iw-andle it is the sea. CI- a-ci-bula si si-tulo ndi ci-bura it is a seat. ZI- n-zi-bula zi zi-tulo ndi b^i-bura those are seats. KA- {vi)-kapamba ndi ka-imuana it is a baby. TU- {n)-iu-cece those are babies. „ loc.(P)A- VCi-pa-fui n6xpa-fupi it is near. „ ,, KU- {n)-ku-le ku ku-de ndi ku-tali it is far. „ „ MU- {Ta)-inu nganda ...B&ndli-ni ndi m-nyumba it is in the house. 584;. — N' B. In general, mere nasals which precede hard consonants or m are practically not heard, unless they be .immediately preceded by a vowel which supports them. Hence it is that in the above Tonga examples n or m are in some cases put between brackets, because at the beginning of a sentence, or after a pause, they would not be perceived. 585. — It is impossible to make out to what extent the copula- tive prefixes of Tonga are used in the languages of the interior. Copulative Prefixes before Substantives. '■- ' 137 because nobody that I know of has even adverted to their existence. However it can be traced in Khutu, a language spoken inland from Zanzibar, in Bisa, in Guha, etc. Thus in Bisa (Last's PolygL, p. 135) we find u-lzmi'^2i tongue ",pl. ni-n-dimi, andni-m6ua "a dog ", where it is pretty evident that m is not, a classifier, but the copulative prefix, so that m-m-dua must be rendered literally by " it is a dog ", and ni-n-dimi by " they are tongues ". Likewise, in Guha, Stanley has the word va.-du-ato, which he renders by " boat, canoe ", but the exact rendering must be " it is a canoe ", since the proper word for " canoe " is sivcv^Iy bu-ato, etc. N. B. It will be seen further on that the copulative prefixes of Tonga are used in Senna before pronouns fcf 656* and 1035J. 586. — Copulative prefixes of the same reduplicative sort as those of Kafir are met with in Kaguru, Gogo, Nyamwezi, etc. For Kaguru this is evident from Last's Kaguru Graminar, where we find, pp. 47 and 50, a complete series of reduplicated pronouns such as zi-zo, lu-lo, li-lo, chi-cho, etc., " it is it, it is they ", answering exactly to their Kafir equivalents zi-zo, lu-lo, li-lo, si-so, etc. (= Tonga nzi-zio ndu-luo, ndi-lio, nce-co, etc., 662). Likewise in "Last's, Polyglotta, p. 222, we find the Kaguru expressions di-kumi " it is ten ", di-kunda " it is nine ", where we should have only kumi, kunda, if these meant simply " ten ", "nine ", etc. 587. — Invariable .copulative prefixes similar to the Senna NDI are used in Chwana, Swahili, Karanga, etc. The Chwana form is KE. Ex. Ke mo-tho " it is a man ", Jse ba-lotsana " they ai'e rascals ", )^& ba-thaba-nchu " they are people of Thaba-nchu ", etc. (Crisp's Gr., p. 52). The Swahili form is NL Ex. Ndugu. yangu ni sultani, " my brother is the Sultan ". N. B. We shall see later on that in Swahili NI is apparently replaced by NDI before pronouns. In Karanga the regular form of the copulative prefix seems to be NDI, as in Senna. 588. — There is no evidence of any prefix which can be iden- tified with the above in Herero, Angola, Congo, etc. In M pong we the particle NE is sometimes used with a copulative meaning. Ex. Wao ne mande? " Who are they ?" (= Tonga Boo rci-banif) XIII. — mtz Batticles tobicl) introDuce SutJStantitics aftEC fiaissitie Verbs. / 589. — Bantu languages fall under three classes with regaM to the manner of introducing the name of the agent after passive verbs. Some make use of the msirumenla/ preposition (Tonga A, Karanga NA, etc., § 572). Such are Tonga, Karanga, Swahili, etc. Others make use of the copulative prefixes just described. Such are Kafir and Chwana. Swahili admits also of this construction. Others join such substantives to their verb Without any particle. Such is Ganda. Such is also Zulu, which departs on this point from the Kafir construction. Ex. Tonga : U-a-ka zialigua a Maria, he was born of Mary, lit. he was begotten by Mary. Karanga : U-a-ka-^warwa na Maria, do. Swahili : Isa a-ka-ongozwa na Roho (or ni Roho) . . . Jesus was led by the spirit. . . (Mat., 4, i). Congo : Idi'iu kwa ngandu, it was eaten by a crocodile (Bentley's Diet. p. 29). Chwana : Go-boleiswenQ ke Morena, it was said by the Lord. Kafir : ...kwa-iiwa yi nkosi, do. Zvixs : ...kwa-tiwa i nkosi, do. Ganda : ...Isa na-a-twalibwa o Moyo mu dungu, Jesus was led by the spirit into the desert (Mat., 4, i). etc., etc. XIV. — T()e Su0ire0 of Suastantioes. 590. — In the Bantu languages the suffixes of substantives have very little importance from a grammatical point of view, be- cause, unlike the suffixes of our classical languages, they have no influence on the construction of sentences. The only noticeable exception to this is that of the locative suffix -ni or -ini, which, according to what has been said, has in Swahili and some other languages the same ruling power as other locative classifiers, e. g. nyumba-ni vciw-ako " in thy house " (= mu-nyumba vsxvT-ako). However some stems may be noted which aremore easily appended than others to substantives as suffixes. Such are : — 591. ana or -nyana, which has already been described as forming the regular diminutives of some languages. Ex. in Tonga : mu-kuiu-axiA " an elder brother", lit. " the elder child " (517, 518). -kulu " great, elder ". Ex. in Tonga: «M<;-kulu " his grandfather ". The Suffixes of Substantives. 139 592. kazi (Rolse -kati or -ati, Mozambique -ari, Kafir -azi or -kazi, etc.) = " female ". Hence in Tonga w«-a«a-kazi " wife ", lit. " child female ", or more exactly " female member-of-the-family ". N. B. In Kafir when the substantive to which -kasi'is suffixed has no distinction of sex, this denotes fecundity, beauty, or excellence. Ex. u m-ii-'kazl " a fine tree ". 593. ike or -ke (Yao -che, Herero -tye, etc) = " small ". Ex. mu-an-'ike " a small brother "(519). 594;. — Less important suffixes in Kafir are -ra " something like", and -ndini, a sort of vocative suffix. 595, — ^- ^- ^" Kafir and Chwana the addition to a word of the suffixes which begin with a vowel causes the phonetic changes described in nn. 122 and 202-207. Ex. in Kafir : u m-lanjamL " a small river " (Chwana ma-lacwaxka), from u mlambo " a river " (Chwana mo-lapo),inkonya.na. " a calf" (Chwana kgon%wsina), from i nkomo " one head of cattle " (Chwana kgomo), etc. XV. — Onomatopoetic Substantives. 596. — We meet in these languages with a pecuHar kind of onomatopoetic substantives, which, though having no classifiers, deserve special attention, were it only because they seem to give the key to the formation of a large number of other words. These onomatopoetic substantives are used principally : — a) by themsel- ves, as exclamations ; — b) after the verb -ti " to say, to do ", as in masekua a/ita ka ati l^uakuakua. " when ducks cry, they say •kuakuukua; — c) after a certain number of other verbs, as in mulilo ulasarara piri-biri-biri "fire gives a red blazing fiame " . Some authors prefer to class this kind of word as adverbs. But, considering that they generally do duty as direct objects of verbs, .they are substantives rather than anything else. Examples in other languages : — Kafir : Wati tu, lit. he did tu, i. e. he kept silent. Unibona wati sa, the maize did sa, i. e. was spread about. Senna : CMko charira chonchoncho, a calabash sounds "like chonchoncho, i. e. gives a hollow sound. , etc., etc. N. B. A whole list of such onomatopoetic words may be seen in the Rev. Alexander Hetherwick's Yao Grammar, p. 77-79. Cf. also Rebmann's Kinyassa Dictionary (passim). XVI. — Ifaria. 597. — The classifiers which have been described in this chapter are the very marrow of the Bantu languages, as may be judged from a single glance at n. 42. Adjectives, verbs, determinatives of all sorts, vary exactly as the classifiers of their nouns, thus giving to the sentences a clearness which has perhaps no parallel in any other language. Hende, for any one who wishes to study a Bantu language, the importance of learning first how to analyse substantives, that is, how to distinguish in them the classifier or determining element from the stem or determined element. 598. — We have already stated (245) that many languages of the Niger, the Gliinea Coast, and even Senegambia, are semi-Bantu^, and cannot be explained properly without some knowledge of the purer Bantu languages. This is particukrly true in the matter of substantives. ' It is no rash assertion to say, for instance, that such words in Ibo of Lower Niger as n-ri " food ", n-ti " an ear ", nwa " a child ", on-wu " death ", u-ia " a bow ", w-anyi " a woman ", ma-du " people ", e-kiti " the middle ", e-lu " above ", etc., are closely related to the Tonga ku-lia or ku-ria " food ", " to eat " (52*), ku-tui " an ear " (462*), mu-ana "a child" (322*), ku-fua " death", "to die "(52*), bu-ta " a bow " (453), mu-kasi " a woman '' (322*), ba-ntu " people " (322*), f'/ja-^a^z " in the middle " (5J3*), (p)ejulu " above " (533*), etc. ; and that, consequently, the Ibo prefixes of substantives, a, e, i, 0, u, n, are, like similar prefixes in Mpongwe, mere remnants of the old Bantu classifiers. (Cf. Grammatical Elements of the Ibo language, by the Rev. J. F. Schon, London, 1861). Likewise, or rather a fortiori, when we find in the scanty avail- able collections of the Avatime language of the middle Niger (?) such words as o-no " a person ", o-nyime " a man (vir), plur. be-; 0-dshe " a woman ", plur. ba- ; li-gume " one head of cattle ", plur. e- ; li-tukpo " the head ", plur. e- ; ko-^kpa " an ear ", plur. ba-; ki-nemi " the tongue ", plur. bi- ; li-we " the sun ", plur. e- ; etc., it is not difficult to recognise in them transformations of the Bantu words mu-ntu " a person " (322*), mu-alume " a man " (322*), mu-kazi " a woman " (322*), in-gombe " one head of cattle " (385*), mu-tue (alias li-tue) " the head " (366*), ku-tui " an ear " (462*), lu-limi " the tongue " (469*), i-zuba (Dualla i-we) " the sun " (410*), etc. And it is even easier to see that the prefixes of such Varia. 141 Avatime words are radically identical with the Bantu classifiers. (Cf. Zeitschrift fur afrikanische Spra,chen, 1887-88, pp. 161-188, and 1889-90, pp. 107-132.) What we say of I bo and Avatime can be extended to many other so-called Negro languages. Cf. n. 830. 599. — This thought has also occurred to me sometimes, that, notwithstanding all prejudices to the contrary, ,several ' Semitic prefixes, such as MA- in the biblical names of tribes and men, MA-, MI-, M-, I, etc., mma-bbul " deluge" (Chwana ma-bula), ma-ddd, " knowledge ", ma-t'mon " a treasure ", mason- " food ", ma-kon, m'^-kunah and t'-kunah " a place ", mi-k'loth " perfections ", ta-k'litk " perfection ", mi-kHhabh " a writing ", m^-dan" disputes ", t'-shubah " the return ", f-skurah " a present", etc., A- in a-don " a lord " (Zulu in-dund), E- in e-sheth " a married woman " (Chwana mo- sadi), etc., etc., and, in general, such prefixes as these to substan- tives, participles, and locatives, may be found to be distantly related to the Bantu classifiers. This, however, is a mere suggestion. GDapter III. ADJECTIVES. 600. — The student may have noticed above (nn. 39-43) that in Bantu every determinative of a substantive requires a prefix, which is no other than that of this substantive, or part, or a fuller form, of it. Hence it is, for instance, that in the examples under n. 42 we find the determinative " your " rendered by u-ako in mu- ana u-ako " your child ", by ba-ako in ba-ana ba-ako " your children ", h^ i-ako in mi-samo i-ako " your trees ", by a-ako in ma-sekua a-ako " your ducks ", hy ku-ako in ku-tui ku-ako " your ear", hysi-ako'm zi-ntu zi-uko " your things ", etc. Hence also, the Tonga equivalent for our adjective " bad " is mu-bi in mu-ana mu-bi " a bad child ", ba-bz in ba-ana ba-bi " bad children ", mi-bi in mi-samo mi-bi " bad trees ", ma-bi in ma-sumo ma-bi" bad spears ", ku-bi in ku-tui ku^bi " a bad ear ", zi-bi in zi-niu zi-bi " bad things ", etc., etc. 601. — Another most important principle is that — if however we do not consider all the Bantu languages, but only the larger number of them — these people must be said to be far from agreeing with us in the distribution of the various determinatives of substantives. Basing their own distribution of these on a prin- ciple of logic which we ourselves overlopk, they have one kind of construction for the few determinatives which express nature, dimension, age, or in, general the quantitative, intrinsic , and com- paratively permanent properties of things, such as old, young, big, thin, tall, short, etc., and another kind of construction for all determinatives whatev'er which are expressive of colour, sensible qualities, position, relations, or in general of the external or chan- geable qualities and relations, such as white, red, clean, dirty, near, far, mine, thine, etc. In other words, the Bantu treat differently the determinatives which properly express being (intrinsically), and those which express being with (having or belonging to), or being like... The former alone are adjectives proper. If we consider neither Adjectvues. 143 Swahili nor Angola or Congo, but the generality of the Bantu languages, we may, put nearly all such adjectives under the heading of Quantitative adjectives *- The others may therefore be termed Non-quantitative. 602. — N- B. I. In Swahili and a few other Coast languages, in which foreign influence is particularly felt, some adjectives which do not refer to anything like quantity are treated nevertheless as quantitative. 603. — 2. In Angola and Lower Congo the "notion of quantitative adjectives seems tp have been lost altogether. In these languages most adjectives pass as possessive express- ions, and consequently we shall not treat of them in this chapter, but in the next. (n. 780). *THE MOST USUAL QUANTITATIVE ADJECTIVES. C^c\€\i\ nicely fat, lean, large, ancient,' small \J\J\J\X pleasant.fine poot; bad great great Tonga -botu -nono -bi -pati -kulu -nini Bisa ... ... ... -nini(.?) Gogb -swamu ... i -bi -baha -dodo Kaguru -swamu -nogo (?) -bi -kulu -kulu -dodo Shambala -edi -tana (?) -wi -kulu -kulu (?) -dodo Boondei -edi -tana -baya -kulu ... -dodo Nyam-Birezi -iza -soga ' -wi -kulu -nikulu -do Talta -rani -rifu -lagelage -baa ... -chahe Kamba f -cheo \ -tseo -nene -vii -nene -uu or ku -nini Swahili -ema -nono -baya -kubwa -kuu -dogo PobLomo ... ' -nona ^wi_(?) ... ... -tyutyu Nika j ^' -dzu -nonu -(m)bi -bahe -kulu -didS -tide Senna ... ..• -kulu ... -ngono Karanga -buya -naki -bi -urwana -urwana -cecana Ganda -lungi -mene -bi -kulu -kulu -tone Xosa-Kaflr (-hie ' -tie -bi -kulu -kulu -ncinci Zulu-Kafir ... -hie i-tle -bi -kulu -kulu -ncane Herero -ua -bua -ua -bua -vi -bi -nene \ -kuru -titi Bihe -wa -wa -mi -nene' , -al6 -titu K-wango -bwa -bwa -bi -kamakama r -ndondo Rotse \ -wavva (-bwa -wawa -bwa -i -bi -nene ■ ■ ... -nini Guha ... ... -keC) R,ua -ampi -nune -bi ... -sheshe Yao -bone -koto -chimwa -kulungwa -chekulu -nandi Mozambique ... ... -ulupale -ulupale ... Ch-wana \ '' 2. ... -ntle -be, -shwe -golu -golugolu -nyenyane » -mpe -kgolu -kgolukgolu , ' » Mpongwe \ ' -bia -bia -bia -bia -be -be -volu -polu -lungu -nungu -ango -yango Fan ... ... -be -nene .., ... Dualla -lodi -bi ... -kuon -sadi Fernandian -boke -lile'(.?) -roterote -boloolo -koko (?) N. B. Concerning Angola and Lower Congo, cf. n. 603. \t I. — Quantitatiwe HDjectities. § I. Adaptation of the Quantitative Adjectives to the DIFFERENT CLASSES OF SUBSTANTIVES. 604. — Quantitative adjectives, such as -lanfo " long ", -pia " new ", -kulu " ancient ", -paiz " large ", and the like, incorporate, as a rule, the classifier of their substantive, expressed or understood. Ex. IN Tonga : Cl.MU-ntu : laa-n/utavL-lan/o, a tall man. Cl.BA-ntu: ba.-n/u hA-lan/a, tall men. „ si/an/umiuemu.-lanfo,a.longcameleoB. „ ha-si/antaml'ue'ba.-lanfi>,\oTigcsLmeleons. ,, MU-ciLA! mu«/amu-/a«/», a long tail. ,, MI-ciLA: mi-«73 mi-Za^j/i, long tails. THE MOST USUAL QUANTITATIVE ADJECTIVES. (Continued.) long, short, old young, alive. abundant. tall snaall ne-w ■whole many Tonga j -lanfo ( -danfo -fuefui -nene -pia -umi -ingi or -nji Blsa -tali ... ... ... -ingi Gogo -tali ... -pia(?) ... -ingi Kaguru -lefu -guhi -dala -sia -gima -engi Shambala -tali ... ... -hia -gima -ingi Boondel -le -jihi -dala -hia -gima -ngi Nyam-wezi 'lihu -guhi -lala -dala -pia -panga -ingi Talta -lele -vui -kale -ishi -engi Kamba / -acha \ -adza -guwe -tene -via -ma -ingi Swahili -refu -fupi -kukuu -pia -zima -ingi Pokoino -yeya ... -bfya Nlka 1 I' . -re -(n)de -fuiii > -kare » -via -pia -zima » -ngi » Senna ■tali -fupi ... -pia -inji Karanga -refo -fupi ... -psa -penyo -nji Ganda ( -wanvu ( -panvu -mpi -daa f-gia \-pia -lamu -ngi / -ninzi \ -ninji Xosa-Kaflr -de -futshana -dala , -tsha ... Zulu-Kafir -de -fupi -dala -tsha -ningi Herero -re -de -supi -nene f-pe \-be -ingi Bihe ... -ale ... Kwango ... ... ... ... Rotse -canana -bia ... Guha -la .». ... ... -ingi Rua -lampi -ipi -nunu -umi ... Yao -leu -jipi -chekulu -wisi 1 -jumi ( -yumi ( -jinji ( -ymji Mozambique ■kani ... -kana ... ( -inchi \ or -injeni Chwana ^' -lele ■kutshane -sha ... -ntsi -telele -khutshane ... -ncha ... » Mpongwe 1 1^' -la -pe -lungu -ona -enge -da » -nungu -yona ... -yenge Fan ... -chun ... Dualla ... ... ... ... ... Fernandlan ... ... -boloolo -nkenke Quantitative Adjectives. 145 CL.(I)N-GOMBE:in-^0OTfen-rfa«X388)alongcow. Cl..(ZI)N-GOMBE:in-^«OTMzi)n-aii, the sun is great. JEzi zintu zi-li zi-botu „ ezi zintu x\.z\-botu, these things are good. El nyika i-li m-botu „ ei nyika nim-boiu, this ground is good. £f nyika tinsi m-botu „ ei nyika tinsi rivca-botu, this ground is not good. etc., etc. Karanga. U u-li n-ju;a (= Tonga ue u-li mu-enibezi), thdu art young. Irie nyika tobe m-buyana na 1 (= Tonga Inyika ilia tinsi m-botu na ?) Is not that ground good ? etc. 619. — 2° In Ganda, and in most of the other Eastern languages, the copula seems to be generally expressed by the particle li or its equivalent in affirmative clauses. Concerning negative clauses nothing certain is to be found. Ex. In Ganda : Gwe o-kia-li mu-lamu, (while) thou art still alive... 620. — 3° In Swahili and Mozambique the copula seems to be generally understood before adjectives of quantity when they are used as predicates. Ex. : — Swahili MozAMurQUE We hu kufa, tn-zima. Kana mi mi m- zima . . . Weyo ra-gumi, kukivali, thou art not dead, but alive. Kana minyo gi m-gumi..., if I am alive... (Rankin's Makua Tales, p. 23). 621. — 4° In Kafir generally neither copula nor relative prefix Quantitaiive Adjectives. 149 is expressed, at least in the present tense, and the predicate adjective is usually for clearness' sake placed at the head of the clause. Ex. Mninji u mbona, the maize is abundant; yi-de lo mtikaktllu, this tree is very high; In-da/a le nkomo, this- cow is old; Siba-is^a, we are young; etc. 622. — Likewise, in Chwana the copula is generally understood in the present tense, but its connective pronoun subject is expressed. Ex. Afo^se o mo-ntle, lit. the town it (is) pretty ; le-tseba le \e-golo, lit. the pigeon it (is) great, etc. (Cf. Rev. W. Crisp's Gr., p. 55). 623. — In Herero quantitative adjectives seem to require an article or relative particle before them, even when they are used as predicates. Ex. Owamio mu-nenep'ove, lit. " I am one older than you." II. — I?on=quantitatitie HDjectitieg. 624. — Leaving aside possessive, demonstrative, and numeral adjectives, as well as certain others, all of which will be dealt with in the next chapter, we may mention here a particular kind of adjective which radically are or have been substantives and which are treated in a somewhat peculiar manner. Such are for instance : — In Kafir -. bomvu '' red ", mhlope " white ", mnyama " black ", and other adjec- tives expressive of colour, as well as several others, such as 7izuht " deep ", -hanzi " wide ", etc. In Chwana : molemo " good ", thaia " strong ", etc. 625. — ^- ^- '• I ^^ not certain that such adjectives exist in Tonga and in the generality of the Bantu languages. However it is probable that we should consider as such in Tonga the word lu-lozi " straight " 626. — 2. In Kafir bomvu is properly the ancient substantive bo-mvu, or more probably bu-omvu, which means " red clay " (cf. the word for " red ground " mo-mvu in Nyengo, »z-*« in Chwana, mo-vu in Yeye, mo-pii in Rotse, U-bti in Lojazi, etc.). The substantive ti m-hlope still exists in Zulu, and means properly " the white of the eye " U m-nyama means properly " an enclosure '', or " the rain-bow ". N-sulu [= li-zitlu (cf. 414) means " the sky ", etc. Likewise, in Chwana mo-lemo means properly " straightness, goodness " ; thata, (= n-tata, cf. n. 390) means " strength ", etc. 3. Thus it maybe noticed that in general such adjectives contain already in themselves a classifier. 627. — It is peculiar to this kind of adjective that they are immediately appended to the copula when this is expressed, or to 1^0 South-African Bantu Languages. the pronoun subject of the copula when this is understood, without first incorporating the classifier of their substantive. Ex. : — ,Kaf[r : Si bomvu, we are red (not si ba bomvu, cf. supra, n. 62 1, si-ha.-tsha, we are young). U-ya ku-ba bonivu, he will be' red (not u-ya kuba mu-bomvu). U-mntu 'obomvu, a red man, lit, a man who (is) red (not u mntu o mu- bomvu). Chwana : Ke thata, I am strong (not ke mo-thata). etc., etc. 628. — N. B. I. In Bantu a great many of our adjectives are rendered by verbs. Ex. Tonga : Mwitu tia-ka tuba ku mu-tue, a man who has white hair, lit. who has become white at th6 head, (from ku-iuba, to turn white). Muntu uiede, such a man, a certain man, lit. a man who has done so, who is so, (as pointed out by a motion of the hand). Tede is the perfect oiku-ti, to say so..., to do so... Kafir : U vmtwana o-lungile-yo, a good child, lit. a child who has turned ^ , , out straight, (from ku-lunga, to become straight). — U-lungiU, '': he is good, is the perfect of ku-lunga. 2. In Angola and Congo nearly all adjectives are treated as possessive expressions, cf. 780. III. Compatatitie.s atiD Bupeclatities. 629. — 1° In Bantu comparison causes no changes in the adjectives themselves, as if they were essentially comparative, but it is shown either by the context itself, or by some other means, for instance — 630. — «j By the use of a locative expression which may then be said to be comparative, as in the above Herero example : O wami mu-nene p'ove, lit. " I am old with respect to you ", i. e. " I am oldfer than you ". Ex. : — Tonga ; Ei nzila nindanfo kuli ndilia, this road is longer than that, lit, this road is long with respect to that. Kafir : Ndi mde ku-we, I am taller than you, lit. I am tall with respect i/'Sx'; to you- I ■ ■ M-Jutshane lo mntu kwa bakowabo, this woman is smaller than her relations. M-kulu lo e milanjeni yonke, this (river) is larger than all the others. j631- — b) By the use of the verb ku-pita " to surpass ", or an equivalent for it (in Chwana g-o-fela, in Angola ku-beta, etc.). Comparatives and' Superlatives. 151 Ex. In Chwana : Pitse e elhata go-feta eeo, this horse is stronger than that, lit... is strong to surpass that one. 632. — 2° Superlatives, or intensive adjectives, are generally obtained by repetitions or by laying a particular stress on the prin- cipal syllable of a word. Ex. : — Tonga : Matanga maingi-maingi, or oftener maingiingi, very many pumpkins. Karanga : Mapuji manji-manji, very many pumpkins. Kafir : hnfene e zi-ninji, very many baboons. A particular stress is laid on the first z of -nuiji. N. B. I. The reduplicative adjectives nini "small "jfuefui " short '', etc., are applica- tions of the same principle. 2. We find in Kafir reduplications of the stems of substantives which convey the same notion as our adjective " genuine ". Ex. i-cubacuba " genuine tobacco ", from i cuba " to- bacco ''. m 633. — There are various other manners of expressing inten- sity, e. g. by the use of the adverbial adjective ku-nene " greatly ", or, in Kafir, ka-kulu " gi^eatly ", or by the use of an intensive verb, such as ku-botesia " to be very good ", from -do^u " good ", etc. (cf, 1079). 634. — A particularly interesting manner of expressing super- latives, at least in Kafir, consists in denying that> a thing is what it is with respect to the quality which it possesses in a high degree. Ex. A si mntu u kuba mhle, lit. " he is not a man (with respect) to being beautiful ", i. e. " he is a marvel of beauty ". Qapter IV. PRONOUNS. 635. — Here again we must remember that there are in the generality of the Bantu languages eighteen categories of substan- tives distinguished from one another by classifiers expressed or understood, and that, consequently, there is a proportionate number of pronouns which cannot be used indifferently. Foreigners in general attend very little to this, and the immediate consequence of it is that natives, anxious to speak like the white man, often come by degrees to neglect entirely what constitutes the proper beauty and perfection of their own language. This effect is very noticeable in several coast languages. It goes to its extreme limit in certain Northern semi-Bantu languages. And perhaps in Bantu languages in general the disturbances in the pronominal system are the best criterion of the amount of foreign influence on them in past times. 636. — An element essential to every pronoun of the third person is a form derived from the classifier of its substantive. This element is what we shall term the connective pronoun, because its proper function is to connect verbs and determinatives with their substantive. I. — Connective Bronouns. 637. — The connective pronouns are a kind of proclitic par- ticle prefixed to verbs and verbal expressions in order to point out their subject and their object. When we come to relative, possessive, and other determinative expressions, we shall see that most of them, from the Bantu point of view, are considered as verbal expressions, and consequently require also connective pronouns before them. In this article we consider only how these pronouns are formed, and how in their most ordinary use they are prefixed to verbs in absolute clauses. To give a general notion of the essential difference which exists between fhem and substantive pronouns, it may be said that they Connective Pronouns. 153 are equivalent to the French /e, hi, il, lis; me, te, le, les, etc., while substantive pronouns rather answer to the French mot, toi, lui, eux, etc. Ex. (Mu-ntu) n-hde, (the man) he is asleep, (French : // dort). (Bantu) h&lede, (the people) they are asleep, (French : ih dorment). (Lu-sabila) hi-lede, (the baby) it is asleep. (Ndi-ue) u-bonide, (you) you have seen, (French : (toi) tu as vu). (Me)n6.i-ha.-bornde, (I) /have seem them, (French : {mo\)jeles ai viis). 638. — Concerning the use of these connective pronouns the most important thing to be observed is that the fact of expressing the substantive subject of a verb does not dispense from expressing the connective pronoun before the same verb. Ex. Leza VL-kede, God lives, lit. God he lives. Ma-'lozni a.-la sisia, the Rotse are very black, lit. the Rotse they are very black. Bu-izu ta hu-ci-wo, there is no more grass, lit. grass it is no more there. Ba-anike beesti ba a-fiia, our brothers are dead, lit. our brothers they are dead. § I. Forms. 639. — Below may be seen comparative tables of the various connective pronouns in the principal Bantu languages according to the different classes and persons *. Theire are a few columns in * COMPARATIVE TABLE OF CONNECTIVE PRONOUNS. Tonga Kaguru Boondei Nyam-wezi Kamba Swahili Pokomo Senna Karanga Ganda Kafir Herero Rotse Angola Congo Yao ivlozambigue Chwana Mpongwe Dualla 1'=' person. Si ndi. ng- n ni ... ni, n ni, n ni ni, n ni ndi ndi. n nzi, nyi, n | ndi( ngi, z.) ndyi mbi ni, I ngi ngi, 1, n ni, n ki ke, mi n(a) n, n Plur. tu chi tu, ti tu tu tu hu ti ti tu, li si •tu tu tu tu tu ni re azwe di 2'' person. 3" person : CI. MU-BA. Sing. Plur. Suij. ObJ. u, ku mu u, ku m(u) u, ku m(u) u, ku mu u, ku m(u) u, ku m(u; ku, ku mu u, ku mu u, kit mu 0, ku mu u, ku ni u, ku mu u, ku mu u, ku mu, nu u,o, — nu, lu u, ku m(u) u, u m(u) 0, go lo, le 0, ... all we 0, ... Sing. Snbj. Obi. u, a, mu ba yu, a, ka. mu wa yu, a, m wa u, a, mu wa yu, a. m(u) ma, a u, a, m(u) wa (ty)u, ka, viu wa u, a. m(u), n (w)a u, a. m(u), (u)n ba u, a. mu ba u, a, e, m(u) ba, be i> mu ve u, a. .'., a u, a, mu a 0, a, e, m, n be u, a, m(u) wa u, a, m(u) ya, a 0, a. mo ba a. w(i) a, ba Piur. 154 South-African Bantu Languages. which it is important to distinguish objective from subjective forms. For clearness' sake such objective forms are printed in italics. In the other columns no such distinction is to be made, as the objective forms do not differ from the subjective. N. B. The A'a/fr pronouns set in black letters are found only in participial expressions, 640. — As may be readily seen from these tables, most con- nective pronouns have almost the same form as the corresponding classifiers. A great exception to this principle is found in the pro- nouns which correspond to such classifiers as contain m or «,yiz. MU, MI, MA, IN. For in most languages these classifiers commonly drop their m ox n when they are converted into pronouns> keeping it almost exclusively in the objective pronoun MU of cl. MU-BA. Strange to say, Lower Congo, Mpongwe, Dualla, and some other western languages differ on this point from the others by keeping the m or the n in most of those same pronouns. This difference, is all the more remarkable as we have seen in the chapter on sub- stantives that in the Mpongwe classifiers the consonant m is ge- nerally dropped, and in the Congo classifiers it is often weakened to n nasal. 64:1, — N.B. I. Modern Angola agrees in several instances with Lower Congo with .regard to retaining the m in the connective pronouns /««, ma, mi. COMPARATIVE TABLE OF CONNECTIVE PRONOUNS. (Contd.) Cl. MU-MI. Cl. INZIN. Cl. LI-MA. Cl. BU. Cl. KU. Sing. Plur. Sing. Plur. Sing. Plur. Sing. Sing. Tonga u i zi li a bu ku Kaguru u i zi li ga bu ku Boondei ^ u' i zi di ya u ku Nyamwezl gu i zi li ga u ku Kamba u i zi i ga u ku Swahili u i zi li ya u ku Pokotno u i zi dji ya tyu (?) ku Senna u i zi ri a bu ku Karanga u, tin i ji ri a bu u Ganda gu g' hgi zi li ga bu ku Kafir u, Wtl \,yi 1 zi li a, e, wa bu ku Herero u vi zi ri (y)e, we u ku Rotse u ... ... ... li a u ... Angola u, mu i i ji ri ma u ku Congo mu mi i ji di me, ma u ku Yao u ji ji si li ya u(?) ku Mozambique u Chi (?) i chi (?) ni a u u Chwana e e di le a bo ^N Mpongwe w(i) m(i) y{i) s(i) ny(i) m(i) w(i) w(i) Dualla / mu mi ni, e i di,li ma bu,bo ... Connective Pronouns. 155 2. The Herero pronoun w" corresponding to cl. MI is also interesting. 3. Probably in Ganda, Yao, Kafir, Mozambique, etc., the consonants^, j,''w,y, etc., in the pronouns gu, ji, 'wu,yi, etc., are merely euphonic (295). The Rev. F. W. Kolbe thinks that some of them are vestiges of primitive consonants which have been weakened. 642. — The subjoined tables of pronouns exhibit only regular forms independent of phonetic laws. To complete it, it will suffice to apply the general principles of Bantu phonetics which have been laid down in the first chapter of this work. Thus the pronoun ki of Kaguru, Swahili, Ganda, etc., will be changed to c or ck before vowels according to n. 258 ; the pronouns u, mu, ku, tu, bu, lu, will be changed in many languages to w, mw, kw, tw, bw, Iw, etc., before vowels ; likewise, before vowels the pronouns i, li, ri, si, etc., will in some languages be changed to^, ly, ry, zy, etc., and in others to y, I, r, ^r, etc., etc. Cf. principally nn. 247-^98. Q^tS. — ^- B- '" the same tables, it should be observed that in Kafir, Chwana, and Congo, the three locative classifiers are referred to by the pronoun ku (Chwana jfo), instead oipa (va, fa), ku (go) and mil (mo). The same takes place sometimes in Tonga and several other languages. § 2. Connective Pronouns prefixed to verbs as Subjects. 6-44. — As a rule every verb in an absolute clause requires a connective pronoun before it to point out the substantive subject. COMPARATIVE TABLE OF CONNECTIVE PRONOUNS. (Contd.) Cl. LU. Cl. CI-ZI. Cl. KA-TU. Locative Classes. Sing. Sing. Plur. Sing. Plur. PA KU MU Tonga lu ci zi ka tu pa ku mu Kaguru Ii(?) ki bi ka ,,, wa Boondei lu ki vi ka ha kii mu Nyam'wezi lu ki fi(?) ka tu ha ku Kamba u ki ka tu mu Swahili u ki vi ka pa ku, y- mu, y- Pokomo tyu ki vi bfa Senna ... ci pi (.?), bzi ic'a pa mu Karanga' ru ci }wi ka tu pa ku mu Ganda lu lu ki si bi - zi ka wa ' ku KaHr ku Herero ru tyi. vi ke,"(Vu) tu pe ku mu Rotse . ... si ... ku Angola lu lu ki ki i i ka fi tu lu ku mu Congo ku Yao lu chi i ka tu pa ku mu Mozambique u lo se chi(?) di va u m Chwana ... ^.° 1 ... Mpongwe w(i) j(i). y(i) ... Dualla ■ I.e. 1 bi, be ... io ... ... CI. IN-ZIN CI. CI-ZI Cl.KA-TU 156 South-African Bantu Languages. Ex. Tonga Kafir \ Me T\.6\-la yeya nawo, Mnandi-an^anjalo..,, As to me, I think'so... ' j /sue a tw-lie to-onse, Tina, ma s\-tye sonke. As to us, let us eat all together. ( lue M-a-ka ia, fVenavr-eia, Vo\x, you haxe stolen. ' \ Inyue, mu-Zade a It ? JVma, ni-ZikH pina ? Yo\i, y/heie do you live} I Le%a M-kede m'manzi, UQamaia w-klelie mamini, God (he) lives in the water.. [ Ba-buetaha.-zuatingulio, A Baiue a ba. amiatin^io, the BueCtieyJv/emno clothes. ,. j Mu-longa a-zuide, U mlamlw u-zele, the river (li) is full. \ Mi-longai-zuide, I milambo \-zeU, the rivers ('Mey^ are full. In-%omlie \-a-inka ku-H? I nkomo y-emkapina ? Where did the cow go to? In-gombe zi-a-inkaku-H? 1 nkomo z-emka pina ? Where did the cattle go to ? n T T lU 4 i 1-zuba W-a-salala, I langa \i-babele, the sun (it) is scorching. ( Ma-tanga a.-bolide, A ma-tanga a.-bolile, the pumpkins (they) are rotten. CI. BQ. Bu-izu bu/a zua, U tyanihw-ya vela, the grass (il) is coming up. CI. KU. Ku-fua ku-zoo-siia, U ku-fa kn-ya ku-jika, death (it) will come. CI. LU. Lu-iimi \\x-la Ittma, U I'W-ivii la-ya lutita, the tongue (it) bites. Ci-bula ci-a-ua, I si-tulo si-wile, a chair (it) has fallen. Zi-bula zi-aua, I zi-tulo z-a-wa, the chairs (they) fell. Ka-pamba ka.-la lila the baby (it) is crying. Tu-pamba bi-la lila, the babies (they) are crying. Loc. PA. Pa-/ffl pia a-nsi (rare), Ku-j/a tsha pa-nsi, it is warm on the ground. Loc. KU. YL'O.-la pia ku-nsi, Ku-j/a Ma ^«o»to', it is warm below. Loc. MU. Mu-ngandamw-lapia, Ka-ya tsha e ndlini, it is warm in the house. Similar examples might be given for all the other Bantu lan- guages. But they would present no remarkable diflference. 645. — Pronouns are often omitted before certain auxiliary forms of the verbs, as will be seen further on (nn. 873 and sqq.). 646. — Some peculiarities have to be noticed with regard to the pronouns of the first person singular and those of class MU-BA, viz. : — 1° In Chwana, Swahili, etc., the full form of the pronoun of the i''^' person singular is reduced to n before certain auxiliary forms of the verbs. Ex. IN Chwana : H-ka reka I may buy (= ke-^a reka). IN Swahili : "H-ta rudi, I shall come back {= ni-ta rudi). 647. — ^- B- '■ In Tonga the pronoun of the first person singular seems to be omitted in certain negative forms beginning with si. Ex. si-yaniii" I do not like". 648. — 2. In Lower Congo the law seems to be to replace the full form ngi or ngy by i or y before such auxiliary forms of the verbs as begin with a vowel, and by « before such auxiliary forms as begin with a consonant. Ex. : ngi-enda" I may go", y-a-yenda " I went ", n-kw-enda " I go ". Cf. Bentley's Congo Grammar. 649. — Of course wherever the pronoun of the first person is thus reduced to n nasal, the immediate consequence of it is the application of all the phonetic laws relative to that sound. Thus in Nyamwezi we have ?;-di-/? isue sue, ■isu ndisue Kaguru Boondei anye mimi ... ase swiswi se, swi. -itu -itu ... Nyam-svezi nene ne isu tui, -isu Kamba Swahili ninye mimi nye, -kwa mi, -ngu ndimi nisi si si si, _ swi, si, -itu -itu ndisi Pokomo mimi mi swiswi swi ... ... Nika mimi mi, -ngu ndimi suisui sui, -ihu ndisui Senna Karanga ine eme ne, {■"^'^ '"=' \.ngu me, -ngu ndine ndime ife isu fe, su. ■tu ■idu ndife ndisu Gaoda rize iige ... fwe, fe fe Kaflr mna m (^ mi) ndim tina ti, •itu siti Herero oami ami, -ndye owami ete ete, -itu oete Angola eme ami ... etu etu ... Congo mono me ... yeto • •■ -ito Yao une ne, -ngu uwe we, -itu ... Mozambique minyo mi, -ia f dimi \ diminyo hiyano hena. -ihu ... Chwana nna me, {-ka) kenna ( rona ( chona Idho, (---^ho) ke rona Mpongwe Dualla mie mba mie, -mi -mi ... azwe biso zwe, -jio -su ••• i6o South-African Bantu Languages. § I. Forms. I. Enclitic forms. 657. — The enclitic forms of the substantive pronouns are the simplest of all. The principle of the formation of most of them is very plain from the subjoined tables, viz. : in most classes of nouns they consist of a connective pronoun and the suffix o, blended together with the usual contractions. Thus, in cl. MU-MI we find u-o or w-o in the singular, and i-o ox y-o in the plural, where u or w, and i ox y, are the connective pronouns of the same class, while o is the suffix proper to substantive pronouns. 658, — Important apparent exceptions to this principle may be observed in the enclitic pronouns of cl. MU-BA, and in those of the I"* and 2^ person. For the ending o shows itself in a few of them only. But the divergency between the mode of formation of these pronouns and that of the others may not be so great in reality SUBSTANTIVE PERSONAL PRONOUNS. (Continued.) 2'' Person. Singular. 2' Person. Plural. si bb _g . I"" a bib -3 u IS .5 '•3 u k ° "o "3.5 a a ij ifi a ° -a ui c O ■" W ^8 <4H 73 W CO . CO Tonga iue ue, -ko ndiue imue mue, -ino ndinyue Kaguru agwegwe gwe, -ko ... anye nyie, -inu «.. . - Boondei wewe we, -ko ' ... nwinwi nwi. -inu - ... Nyam-wezl wewe we, -ko ... imue mue. -inu ... - Kamba niwe we, -go ... inywi nywi. -inyu .*■ Swahili wewe we, -ko ndiwe nyinyi nyi. -inu ndinyi Pokomo wewe we ... nywinywi nywi ... Nika ... -ko muimui mui. -inu ... Senna iwe we, -ko ndiwe imue mue, -nu ndimue Karanga ewe we, -0 ndiwe ■ino ... Ganda gwe ... mwe mwe ... Kafir wena we, -ko nguwe nina ni. •inu nini Herero ove ■oye ene ene, -inu oene Angola eye e enu enu Congo ngeye nge, -kti ... yeno ... -ino Yao . ugwe gwe, -ko umwemwe mwe. -inu Mozambique weyo we, -0 diwe nyenyu nyenyo -inyu ... Chwana wena 0, (g<^)go ke wena 1 lona ( nyena lo. -eno ke lona Mpongwe awe o, we, -0, anwe nwe. -ni ... Dualla wa . . . on^o binyo -nyu ... Substantive Personal Pronouns. i6i as it seems to be at first sight, as the following considerations may show : — 659. — 1° The fullest and more primitive forms of the pronouns in cl. MU-BA, and in the i'^^' and 2^ person, seem to be the following : — Common form. I'" PERS. SING.: ?««« (perhaps ??ii5;j«) whence me, mi nye(i2z) ne (73, etc.) I'st PERS. PLUR. : sue (or tue,fue, etc.) 2* PERS. SING.; ,, tie (whence we, o, 265) 2'' PERS. PLUR.: ,, tmtelvfheacenyzve, 122) Cl.MU-BAsiNG. ; ,, ««(whence««,>'«,/«,etc.) CI.MU-BAplur.;,, ^a!o(vihencefo,wao,etc.) After the possessive particle. -ngu or K^e.This with the poss.part.gives -a-ngu mine, whence -a-;«ja(273) etc. ■isum -itu. ) J ») -eszf, {=a-isu)j ours. ■ko. •a-Zeo, thine. tno, -mu. ) )» -£nu{ =a-znu), yom'S kue{v/heBce ■ke,-ce) ) )) -ct-Jiue.^ his. bo ) J) •a-boj theirs: 2° Considering that almost all these' forrns end in ue or o, reduced in some cases to u, and comparing them with the substan- tive pronouns of the other classes, most of which take o as their suffix, it may be said that we have here nothing else than an SUBSTANTIVE PERSONAL PRONOUNS. (Continued.) 3" person. Cl. MU-BA. Singular. Plural. bb bj] c . i™ g bb '■5 ti.S -3 ' ii.a % ■3 .5 a,cs 1 u ■3.S tn a ° 1=! ■ Ul ff -s is w- Ug ^ w ug C/1 u 1) en u Tonga uwe ue, -h^e ngue abo bo mbabo Kaguru yuyu yu, -kwe wao Boondei yeye ye, -kwe wao Nyamvrezi uwe ue, -iue awo wq K.amba miya ya, • -kwe acho cho, -iyo Swahlli yeye ye, -ke ndiye wao ndio Pokomo tyetye tye ;■ \' ... • •• wao Nika -kwe ao Senna •iye ■ ye, -cAe . ndiye iwo wo ndiwo Karanga lye ye, -e ndiye iwo wo ndiwo Ganda ye ye be bo Kaflr yena ■ye, -ke nguye bona bo ngabo Herero eye, eye e, ;■ ., ■ -« . • . owo, ovo wo Angola muene g ■•- ' ene a Congo yandi -fuii yau yau ... Yao jue jo (?), -kwe wao wao ... Mozambique ( yoyo yena Ihit, - yayo yayo Chwana ene e, (-ga)gwe ke ene bone bo ke bone Mpong-we aye e, ye, -ye ... wao wao Dualla mo -u •■■' babo babo 1 62 South-African Bantu Languages. application of the general phonetic principle of Bantu that ue and o are convertible in given cases (265). 660. — Hence the general law of the original formation' of simple substantive pronourts in Bantu may be expressed by the following formula : — Connective pronoun + suffix -M^ ox -o. N. B. The presence oik in -ko " thee " ^nA-kue " he " after the possessive particle a is perhaps merely euphonic, or, to be more exact, is intended to prevent contractions which might interfere with clearness of expression. II, Self-Standing forms. 661. ■ — Great dialectic divergencies are noticeable in the forma- tion of the self-standing substantive pronouns. However they all seem tp be applications of the one and same great principle of avoiding monosyllabic self-standing words (44). For, admitting this to be the correct view of the subject, we find that in order to maintain this principle : — SUBSTANTIVE PERSONAL PRONOUNS. (Continued.) Gl. MU-MJ CI. IN-ZIN Singular^. ho Plui -al. Sing ular. bb Plui al. bi bb ^ s , bb g , bij a , '='1 c , bs '•3 J ■i.S ■■3 =4.9 '•5 u "1.5 ■5 ta.S 1 "o "3 S e- 3 ■ Ij .s.S Q,ot - a! "u 3 S c "u ■3.E n.3 . . bie bio ... ke ko Kafir sona so siso zona zo zizo . .. ... Herero ,otyo tyo ... ovio vio ... oko ko otuo tuo Angola ... ... Congo kiau kio . . . yau yo fiau fio twau two Yao che cho ye yo ke ko tue tuo Mozambique cho chicho ... Chwana shone sho ke shone cone CO ke cone ... ... Mpongwe ... ... ... ... DuaUa ... ... ... ... ... ... Substantive Personal Pronouns. 165 2. In Herero I find otuami'-^ it is'I ". Oete is also probably a copulative pronoun of the i'« person plural, and oene one of the 2'' person plural, as if the article o had the same power as the copula. § 2. Use OF THE Different Forms. i. Self-standing forms. 663. — Substantive personal pronouns are used in their self- standing form principally to express contrast or emphasis ( = French moi, toi, lui, eux, etc., before or after verbs). Ex. : — Tonga : lue mulozi, lit. thou, thou art a sorcerer. Mu-zoo-jana baaka sika, inyue ka muli lede, you will find that they came while j'(7«, you were asleep. Bo ba-la tuba, iue u-la sia, they are white, (but) Ae, he is black. Isue iu-lt ba-nini, izio (zi-pembele) n-zipati, (as for) us, we are small, but they (the sea-cows) they are big. 664. — Senna : Ene ndi-na kala, I, I remain ; iue u-nakala, thou, thou remainest ; iye a-nakala. SUBSTANTIVE PERSONAL PRONOUNS , (Continued.) CI. LU. LOCATIVK CLASSES. Singular. 1 CI. PA. CI . KU. Gl. MU . oi- . bb ^ bi fcJ3 ~ .c. , M c bo a bb C bb '3 ^cj rt.S ■-5 c i^.S '■B u rt.B .-I ^3 u rt S as Tj 1 Tj s.S 3 Tj ■3 .3 D-rt c3 'u in C ° *?• w c -t; (n a -s Ul jg -2 "4- ' "a M W W w ^§ m 00 en c/: U Tonga lo ndulo wo mpowo oko ko nkuko mo Kaguru ... k) lulo (?) ho haho(?) ko Boondel ... ho Kyam-wezi , lo ho ko mo Kamba ... Swahill wo ndio papa po .ndipo ko ndiko mo ndimo Pokomo ,'.. bfo Nika ... vo ko Senna ... ... po ndipo ko mo Karanga iro ro ndiro po ndipo ... Ganda rwe rwo we wo gie gio mwe mu Kafir lona lo luio ... kona ko kuko Herero oruo ruo opo po oko ko omona mo Angola ... . .. Congo luau lo vau vo kwau ko mwau mo Yao . lu luo ... pe po kwe- ko mwe mo Mozambique ... vavo vo Ch-esrana lone lo ke lone gone I go ke gone Mpongwe ... ..^ ... Dualla ... 1 66 South-African Bantu Languages. he, he remains ; ife ti-na kala, we, we remain ; imue niu-na kala, you, you remain ; iwo a-na kala, they, they remain ; kala-ni imue, do ye lemain, you, etc. 665. — Karanga : , Isu ti-riba-cecana, iju n\u-kuruana; we, we are smallj they (e. g. sea-cows) they are big. Ibo batipe,thcy (e. g. the men), they said no. W-no-penga iue, thou art mad, thou. Imue mu-a-fana Reja ; you, you are like God. lye, ua-ru-ba xe n-kuruana ; he, he was a great king. 666. — Old Angola (from F. deCoucto's Angola Catechism, Rome, i66i) : Nga-ku-sawile iye ngana yami, I have offended you, you my Lord (page 6). Bene, okitatu Mao ; they, the three of them (p. ii). Ehue, ne atu ossololo, you and all men (p. 17). Mu-ftg-t/ie ime ngana yenu, do ye know me, me^jovx Lord (p. 17). 667. — HfiRERO (from Dr. Biittner's Mdrchen der Ova-Herero in the Zeitschrip fitr afrikanische Sprachen, 1887-88) : Ku-tura ete Ko uvi, to deliver us from evil (p. 294). Ka pve, it is not thee (p. 190). 668. — SwAHiLi (from Dr. Steere's Swahili Tales, London 1889) : Wewe ingia ndani, go inside, thou. Wewe nani'i or weye nanil Who art thou ? (p. 338). Wa-toka wapi, wee? Mimi natoka mjini kwetu. Where dost thou come from, thou ? l, I come frorn our town (p. 338). Ku-nywa wewe, drink thou (p. 358). Wa-ka-enda vivio hivio, thus they went, lit. they went it, that (manner), (p. 342). Papa (papo (i*)) /%«/« = here, (lit. (at) it, this place). 669. — Kafir: Ku-ya harnba mna, lit. There will go myself, i. e. I will go myself. Nda ku-ku-bulala wena, ukuba uisho, lit. Thee I shall kill, if thou saystso. Kwaba 7ijalo ukufa kwa lo mfo: bati ke bona, bapuma emanzini, such was the death of that man ; as to thejn, they came out of the water. 670. — Ganda (from the Grammaire Ruganda) : Tu na sika gue o-kia-li mulatnu? Lit. Shall we come into power when thou, thou art still alive? Nze bue ndia imnere, sikkuta, I, when I eat porrid^, I cannot be satiated. 671. — Chwana : ' ' Nna, ka-rejalo, lit. r, I said so (Crisp's Gr., p. 13), etc., etc. 672. — : ^- ^- I- In some languages, viz. in Karanga, Herero, Chwana, Mozam- bique, etc., substantive pronouns are also used regularly in their self-standing form after the preposition which means " with, and, also ", viz. na or ne in Karanga and Herero lU in Mozambique, le in Chwana, etc. (570). 'Substantive Personal' Pronouns. i6j Ex. : Karanga : Ne-ebo, ba-ka-ba banji, they also became numerous. Herero : N'eye a kotoka, and she came bacjc ; n'owo va-tre, and they went. Mozambique: Ni-minyo gi-nahagoloa,^l too, I shall come back. Ni-yena a-kala na mwaraui, he too had a. wife. IN Chwana : Le-ene, he too ; le-bone, they too, etc... 673. — 2. In Chwana and Mozambique, substantive pronouns are used regularly in their self-standing form after several other prepositions or particles (cf. Crisp's Chwana Gr., p. 13). '''': II. Enclitic forms. 674. — The reader may remember first that all the other forms of substantive pronouns, contain at least originally the enclitic form. This is found also doing duty regularly either as a noun or as a determinative in many other expressions which vary according to the different languages. Thus, in Tonga, we find it : — a) After the prep, a " and, with, also ". Ex, : Baainka a-ue kunganga, they went to the doctor with him, (lit. they went he-also to the doctor). N. B. Concerning the forms of the pronouns after the other prepositions in Tonga, cf. 688 and 1040- 1041. / b) Before or after verbal expressions without emphasis. Ex. I Ue u-ii... He (the man already mentioned), he says ... Ba-lapelela sue, they pray for us. c) Before locative expressions. ' Ex. : Ta mu-zooinki 'k.o ku-lta, do not go there, lit... to it, that (place). d) In such expressions as u-ci-liwo, he is still there ; u-a-li ko lu-bela, he was there from the beginning, etc. 675. — In Karanga, Swahili, Kafir, Senna, Angola, etc., we find these enclitic forms of pronouns in the same cases as in Tonga, though not so often before verbal expressions, and in several others, more particularly after prepositions in general, and j, often before numbers, as also before the words which render our " all ". Ex. : — Karanga : NJoonda na-yo (ijira) lit. I shall go by it (the road), Bariubeja ku-na-su, men came to us. ; 676. — Angola : 'O ngana yekdla na-Q, the Lord is with thee {Catechism, p. 2).' Ku-tundana.\o, to ^IxeXdeiit (lu-kuako, the arm) (p. 23). i68 South- African Bant ft, Languages. O mvenye uae uaikWo o kukatula mo o nayenyo... lit. his soul went thither to draw fromtherdn the souls:.. [ibid., p. .27). O inussa ueiu tube o, lit. our food give it {to) us (ibid., p. i). Jtuxi ngiriela yo, the sins I have committed (them) (p. 54). QTin , SwAHILI : Looo! simba u-mo ndani, Oho ! lion, thou a.x\. there inside. Yu-tno ndani, he is there inside. Na-mi, and (or) with me; na--\nt, and (or) with thee, etc... A'-mi, it is not /,■ «-ye, it is not he, etc. . . Ki-suni-li-cho «a-cho, the knife I have..., lit. the knife I am //with it (cf. 733). Zo zote (njia), all the roads, lit. they all (the roads). 678. — Senna: Si-nt, it is not /. Mba-pita-Y& nkati, and he entered inside. Mba-pita na-yo ( mbuzi) n-fiyuinba, and he entered the house with // (the goat). 679. — Kafir: Yiza-ai, come ye. Wk-O ku-ni, he is there nt&xyou. Yiza na-m, coirie with tne.' A-si-ye, it is not he. ^ A-si-\o hashe, it is not a horse, lit. it (is) not it, horse. 680. — Ganda : iVij-nge, and (or) with me. N'a-wangula wo e mpagi, lit. and he drew out there a pole. 681. — .3° In Herero after prepositions and locative classifiers we do not as a rule find enclitic, but self-standing pronouns. Enclitic pronouns are found however in locative expressions of a. different kind. Ex. N'u-i-ko, and he gotsoS (there). - A-rire ty'a-tua mo m'o ndydtu, lit. and she put it in therein in the sack {ZeUschrifi, 1887-88, p. 190). 682. — 4° In Chwana enclitic substantive pronouns are found almost exclusively after the preposition na " with ". Ex. Na-ho " with them ", na-o " with thee ", etc. The locative pronoun eo is often used after a negative copula. Ex. Gaa-eo, he is not there (= Tonga taa-'ko, Kafir a teko). 683. — 5° In Mozambique enclitic substantive pronouns are found principally after a negative copula. Ex. ICa-vo, he is not there, ( = Swahili ha-]s.o). Minyo o-^'-hio Amrani, I am not he, Amran, (= Swahili mimi si-ye Amrani). N. B. Self-standing pronouns are used regularly in most other cases. Ex. Ni-minyo • gi-na hegoloa, I too, I shall come back (Rankin's Makua Tales, p. 2) etc. Substanti've Personal Pronouns. 169 684. — From all this are excluded possessive expressions. For in these almost all the Bantu languages -agree in regularly using enclitic pronouns. Ex.: IN Tonga : Ingombe zw-ngu, 2/etu, ziemx, s/a-bo, etc. my, -our, your, their cattle, etc., (cf. 659). III. Copula-containing forms, 685. — i(? These copula-containing forms are used generally before substantives, or independently, to assert identity with a par- ticular and determined person or thing. Ex.: Tonga, TCaranga, Senna : Iwe ndi-ue Marani, You are Maran, lit. You, // is you, Maran. SwAHiLi ; Wewe ndi-we Marani, do. Mozambique : Weyo di-we Marani, do. Kafir : Wena ngfu-we Marani, do. Chwana : Wena ke-wena Marani, do. 686. — N. B., I. A similar construction in Herero is the following : Owami Kaare,l am Kaare, lit. It is I, Kaare. Cf. 662. 687. — 2. We may observe in Tonga the difference between such expressions as lue mu-lozi and Ndiwe mu-lozi. The first means only,: " You are a sorcerer " ; the second means : " You are the sorcerer (I am looking for) "- Tonga, idiom : Nceco ci nda-ta fiiambana ktt-za, " that is why I have not hurried to come ". Ci-ntu " a thing " is here understood. 688. — 2° In Tonga these copula-containing pronouns are also used regularly after all prepositions and locatiye classifiers, though not always after the pardcle a when it means "and " (cf. 674), nei- ther after the possessive particle a (684). Ex.: Tu-la kondua a ngue, a-mbabo, we shalLrejoice with him, with them. U-a-inka ku-li ndilio (i-saku), he is gone to him (to the devil). N. B. With regard to the insertion of It between ku and ndilio, cf. n. 1040. U-a-lapela a-nzio (in-gubo), he wears them (clothes) when praying, lit. : he U-bed' anga ndi-me, he is like me. [prays with them. N. B. I do not know that these peculiar constructions have been noticed as yet in other Bantu languages. § 3. Varia. 689. — 1° In Tonga, the suffix -nya " self ", equivalent to the Kafir -na^ Chwana -na or -ne, Ga.nda -^nna, Mozambique -nyo, etc. (cf. 824), is often added to substantive pronouns for the sake of greater emphasis. Ex.: Tu-ia kondua a-nguenya., ( Leza), we shall rejoice with him himself (God). Ncedo-nys. cio, (ci-ntu), that is the very thing. ' •«r^;ji^|»j«^ 170 South-African Bantu Languages. 690. -; — N. B. I. In the last two expressions ncecia and nzinio are copula-con- taining pronouns, while cio and zio are enclitic pronouns. 2. Tonga idiom ; Mpawo-nya na aka amba, immediately after he had spoken.,. 691. — 2° In Tonga the suffix -bo is generally appended to substantive ' pronouns of the i"^**^ and 2^ person when they are preceded by the particle « " and, also ". Hence a-sue-ho, we also; a-nytie-\iQ, you also ; a-e-ho ( = a-ue-ho), thou also ; a-mde-ho { — a-me-ho), I also. This suffix -do is radically identical with -jnue " one, another " (n. 792). A', B. I. Likewise in Kara.nga na-su-bo tonda, " we shall go, we also ", and in Senna ifie-hve " I also ", i/e-bve " we also ", etc. 2. Kafirs use in similar cases thp prefix ^ua- " also ". Ex. kua-mna " I too " kwa-iiiena " thou also '', etc. 692. — 3° In Ganda we find a sort of dual formed in the same manner with the SM^y;. -mbi " two (cf. 792) ". "Ex. fe-mbi " both of us " ; bo^mbi " both of them " etc. (cf. 794). III. — X)cmon»tratit)e ficonoung. 693. — The various forms of demonstrative pronouns are distri- '|., butable into fundamental, emphatic, and copula- containing forms*. * FUNDAMENTAL DEMONSTRATIVE PRONOUNS. Class MU-BA. Singular : MU-ntti Plural BA-n/u 1"' Position 2'^ Pos. S"* Pos. I"' Position • 2" Pos. 3" Pos. Tonga (o)yu (o)uno (o)yo (o)ulia aba (a)bano abo (a)balia Kaguru ayu yuyo yudia wawa ... wawo wadia NyamWezl uyu uyo awa ... awo Boondei uyu uyo yuda awa ... awo wada Kamba .1.. uya tiyu uuya aya awu aaya SwahlU huyu . .. huyo yule hawa hawo wale Pokomo huyit huyo huyude hawa ... hao hawade Senna uyu • .. uyo ule awa ... awo ale Karanga (i)oyu (e)ondia ... ... Ganda uyu ono oyo oli bano abo ball Xosa- Kafir lo lowo, lo Iowa, la aba ... abo abaya Zulu-Kaflr lo lona lowo loya laba ... labo labaya Herero (i)ngui ... j nguini ( nguina (i)mba ... ( mbeni Imbena Angola iu (i)o (i)una awa 00 {i)ana Lower Congo oyu oyo ona aya owo ana Yao (a)ju (a)jino (a)jojo (a)jula (a)wa (a)wano (a)wo (a)wala Mozambique ula, ola uyo. ole ala ayo ale Chwana eo ( eono ( eona eoo . eole ba ( bano "l bana bao bale Mpongwe ... wino WOI10 ... witto wono ... Demonstrative Pronouns. 171 The student's attention is particularly called to the fact that our adverbs " here ", " there ", " yonder ", are rendered in Bantu by the demonstrative pronouns which correspond to the locative Glasses PA, KU, and MU. § I. Fundamental Forms. 694. — In Bantu grammars the' fundamental forms of demon- strative pronouns are generally distributed into pronouns expressive of proximity, pronouns expressive of things already mentioned, or of limited distance, and pronouns expressive of greater distance. This certainly is not a correct view of t+ie subject, at least in those languages oh which the greatest amount of reliable materials is available. My informants of various tribes all agreed in distri- ; buting these pronouns as follows :■ — 1° Pronouns expressive of proximity /o thepei^son speaking, or, as we may call them, demonstrative pronouns of the i^^t position. Ex. in Tonga : eli sekua, this duck (near me). In some languages these pronouns have two forms, the one without any suffix, ?is aba in aba,bantu " these people ", the second FUNDAMENTAL DEMONSTRATIVE, PRONOUNS. (Continued.) Class MU-MI. Singular : MU-cila Plural : xMI-cila I"' Position 2^ Pos. 3'' Pos. 1'^' Position 2" Pos. 3° Pos. Tonga (o)yu (o)uno ■(o)yo (o)ulia ei (e)ino 1 — (e)yo (e)ilia JKlaguru au uo udia ai ... lyo idia Nyamwezi ugu ... ugo ii io ... Boondei ... unu uwo uda . inu iyo ida Kamba ... ' uya ,uyu uuya iya iyu iiya Swahili huu ... huo ule hii hiyo ile Pokqmo huu huo huude hii hiyo hiide Senna uu ule ii iyo > ile Karanga oyu oyo (e)ondia iyo ilia Ganda guno ogo guli gino egio gili Xosa-Kafir lo lowo,Ip Iowa, la ie leyo, lo leya, la Zulu-Kaflr lo lona lowo' Iowa le lena leyo leya Hererb ' (i)mbui ( mbuini \ mbuina (i)imbi ... ... f mbini \ mbina Angola ia ... (i)0 (i)una eyi oyo in a Lower Congo owu owo owuna ami emio emina Yao (a)u (a)uno (a)oo (a)ula (a)ji (a)jino (a)jo (a)jila Mozambique (ula (una uyo ole chila chile Ch-wana J ono i ona 00 ole e (eno \ ena eeo ele Mpongwe ... wino wono ... yinn yono 172 South-African Bantu Languages. with the suffix no (na, la), as bano in bantu bano, which means also " these people ". 2° Pronouns expressive of proximity to the person spoken to, whatever be the distance from the person speaking, or demonstra- tive pronouns of the 2^ position. Ex. "EXlo .sekua, that duck (near you). Almost all these pronouns end in -0. 3° Pronouns expressive oi distance from both the person speaking, and the person spoken to, or demonstrative pronouns of the ■^^ position. Ex. Hlilia sekua, that duck (far both from me and from you). 695, — N. B. This then is the correct division of demonstrative pronouns, at least in Tonga, Karanga, Kafir, CHwana, and Senna. That the same may be said of Swahili and Angola can be safely established by considering that in the safest specimens of native literature in these languages the demonstrative pronouns ending with the suffix ■0 are used almost exclusively with reference to position near the person spoken to. There is no difficulty with regard to the pronouns of the first or the third position. 696. — Ex. In Swahili (from Steere's Swahili Tales, London, 1889) : Page 20. Ume kwenda kwa harrjako hapo. You have gone in a hurry thither (where j(?a are). do. Nangojea hiyo tttmbako, I am waiting for that tobacco (which I say is near j'»«). do. Kilwa kicho kita kuuma, that head (of yours) will ache. 'P3L.%ti(>.Ah ! mume wangu,... maneno yayo kwa yayo sikuzote! Ah! my husband, every day those woid.s {0/ yours), those same words. FUNDAMENTAL DEMONSTRATIVE PRONOUNS. (Continued.; Class IN-ZIN. Singular : IN-^ombe. Plural : {Z)IN-^ombe. I"' Position. 2'' pos. 3" pos. I"' Position. 2" pos. 3" pos. Tonga ei (e)ino eyo (e)ilia ezi (e)zino ezio (e)zilia Kaguru ai ... iyo idia azi ... zizo (?) zidia Nyain-wezi ii ... io ... izi izo azia Boondei inu iyo ida izi ... izo zia Kamba iya iyu iiya ziya ziyu ziiya Swahili hii iyo ile hizi . .. hizo zile Pokomo hii ... hiyo hiide , hizi hizo hizide Senna ii iyo ile izi ... izo zile Karanga ei iyo (e)ilia (i)oji ijo ((e)jilia Ganda eno eyo eli **• zino ezo zili Xosa-Kaflr ie ... |!r {ir ezi ezo (eziya leza Zulu-Kaflr le lena leyo leya lezi lezo leziya Herero (i)ndyi f ndyini (ndyina (i)nira fOtateni Angola eyi oyo ina eji 0)0 jina Lower Congo eyi eyo (ey)ina e|i ... ezo Tt Yao (a)ji (a)jino (a)jo (a)jila (a)si (a)sino (a)sio (a)sila Mozambique ila iyo(.?) ile ciiila ... chile Ghwana e f eno (,ena eeo ele tse ( tseno ) tsena tseo tsele Mpongwe ... yino yoiio ... xino xono Demonstrative Pronouns. 173 Cf. ibidem, page 26, line 1 2 ; p. 36, 1. 7, 8 and 36 ; p. 40, I. 3, 9, lo, etc. N. B. At page 40, line 8, of the same work, the pronoun hilo in nywmba hilo " this house ", might be thought to create a diSiculty ; but in reality it is a misprint for hii. Nyumba hilo is in, no sense correct, because nyumba is a word of cl. IN, while hilo is of cl. LI. 697. — Ex. In Angola : ' O messo ae 00, those eyes {of yours). From Father de Coucto'S C«/,, p. 3. 1. Formation of these pronouns. 698. — As may be easily seen from the subjoined tables, the most general formula for' the formation' of these demonstrative pronouns, is as follows : — r none, or -no (-na, -la), for the i"' position. A kind of article -|- connect, pr. -|-suffi.x-| -ofor the 2'' position. ( -lia ('or la, na, ya, le) for the 3'' position. The article seems not to be used at all in Chwana, Mozambique, or Mpongwe. In the other languages it is left out only in given cases, which vary according to the different languages. The forms of the same article are also various, viz. a in Yao; a, '/ 174 South- African Bantu Languages. live pronouns of Zulu always contains /, whereas in Xosa the /is only heard as a rule in such pronouns as -have no other proper consonant. , 0QQ_ — jV. B. I. In Angola the pronouns baba, bobo (of class PA), kttku, koko (of class KU), and mitmu,. momo (of class MU), are properly reduplicative pronouns (cf. 705). If the simple pronouns corresponding, to these existed in Angola, they would be apparently aba, obo; oku, qko; omu, omo. 2. The demonstrative pronouns in Karanga seem to have two articles, the one ordinary, viz. a, e, or 0, the other emphatic, viz, t. More information is wanted as to this language, one of the most interesting of the Bantu family. 3. The presence of h in the articles of the Swahili pronouns is probably due to Arabic influence. Possibly the presence of / in the corresponding Zulu articles is due to some ', ancient influence of the same sort. 4. I consider it as probable that the suffix -o for pronouns of the 2^ position was originally identical with the pronoun «« or ,4i? "you "of the 2'' person singular. Perhaps the suffix -»(? for pronouns of the i'" position was also identical with the pronoun -ngu, the possessive form of the i '" person singular. The suffix -le foi- pronouns of the 3^ position means " far ". The ftillform lia is probably a compound of le "iax" -\-a, demonstrative in the distance,' 2. Use and place of these pronouns. 700. — First, demonsti-ative pronouns can be used substantively as self-standing words. Ex. : — Tonga : Oyu iitu-lozi, this (man) is a sorcerer. FUNDAMENTAL DEMONSTRATIVE PRONOUNS (Continued.) Class BU. Class LU. BU-siku. LU-luni. , 1"' Position. 2" pos. 3" Pos.- i^ Position. 2" Pos. 3<'Pos. Tonga obu (o)buno obo (o)bulia olu (o)luno oluo (o)lulia Kaguru au uo , udia alu lulo(.?) ludia Nyamwezi uwu ... uwo ... ulu ulo Boondei ... unu uno (?) uda lunu luno (?) luda Kariiba , uu • .. uyu uuya uu uyu uuya Swahili huu huo hule huu ... huo hule Pokomo hutyu hutyo hutyude ... ... Senna uu ... uo ule ulu ulo Karanga (i)obu ubo (i)oru ... Ganda buno obwo b'uii runo orwo ruli Xosa-Kaflf obu ... ■ obo f obuya tob4 olu olo (oliwa (olwa Zulu-Kafir lobu . .. lobo lobuya lolu lolo loluya Herero (i)mbui ... f mbuini t mbuina (i)ndui ... {[En£ Angola iu (i)una olu olo luna Lower Congo owu . >. ' owo (o)wuna olu ... olo (o)luna Yao (a)ju (a)juno ao (a)ula (a)lu (a)Iuno (a)lo (a)lula Mozambique uu ula (?) ... ule uu ... ... ule Chwana JO Jjona Ijono JOO jole lo ( lono 1 lona loo lole Mpongwe wonn wino ... ... wiiiB wono ... Demonstrative Pronouns. J 75 Aba mba-kazoaxa bako, these (men) are thy priests. Ba-yuni ba-a wano, the birds of this" place, lit. of here. Nda-ka inka okulia, I went there. N. B. Lino or ellno,'demonstrative pronoun of cl. LI, and ino or eino, demonstrative propoun of cl. IN, are often used independently to render our adverbs " then, now, immediately ''. Ex. Ndi-li-wo\vaa, I shall be there Wz>e'. SWAHILI : Una fanya nhii hapo? What are you doing there? (Rankin's Makua and Swahili Tales, ^.S). Huyu sikondoo, this is not a sheep {ibid.^^ p. 5). Kule kdondeni, there among the sheep {tbid.\ p. 7). ' IVakakaa liule, they 'remained there (ibid., p. 9). Mozambique : Una vara sheni-va."? What are you doing here ? {ibid., p. 4). Hoyo kahiyo ibwitibwiti, that is not a sheep, (ibid., p. 4). 0-madani, there among the, sheep {ibid., p. 6). etc., etc. 701. — Secondly, in the generality of the Bantu languages, when demonstrative pronouns are used adjectively, they seem to be placed somewhat indifferently before or after their substantive. In Chwanaand Ganda they seem to be always placed aftef. Ex.: — FUNDAMENTAL DEMONSTRATIVE PRONOUNS. (Continued.) Class CI-ZI. Singular ; CT-niu Plural : Zl-ntu i"^' Position 2" Pos. 3" Pos. 1'^ Position 2'" Pos 3" Pos. Tonga eci (e)cino ecio (e)cilia ezi (e)zin0 ezi, (e)zilia, Kaguru achi ■ .. chicho(?) cliidia avi vivo (?) vidia Nyamwezl iki icho iti ... ifo ... Boondei iki iko kia ivi ivio via Kamba kiya kiyu kiiya iya iyu iiya Swahili hichi hicho kile hivi hivio vile Pokomo hityi hityo ' hityide hiwi hivvyo hivide Senna ici icio cire fibji ■ j ibso lipo ( bsire Ipire Karanga (i)oci (i)ocino ... (l)0{Wl ijwo Ganda kino ekio kiii ... bino ebio biii Xosa-Kaflr esi - eso f esiya • (.esa ezi ' ezo J eziya leza Zulu-Kafir lesi ... leso lesiya lezi lezo . V leziya Herero (i)hi .iKina ^ (i)mbi ... ( mbini \ mbina Angola eki okio kina eyi oyo ina Lower Congo eki ekio (e)kina eyi eyo (ey)ina Yao (a)chi (a)chino (a)cho (a)cila (a)i (a)ino (a)yo (a)ila Mozambifljue ... ila ile ... chila chile Chwana se j seno (sena seo •sale tse J tseno { tsena tseo tsele Mpongwe jino jono ... ... yitio yona 176 South-African Bantu Languages. Tonga ' Karanga Baakafua e inzala oyu tnuaha, Bakafa mjara muaga oyu, they died from hunger this year. Ilia nyika iiinsi mboiu na 1 \n^nyika tobembuyanana na ? Is not that ground good ? Mbuzie mukuarana angU oulia, Mtt-buje nkuru anga eondia, Ask that brother of mine. Ei nkani iamana Mawui aa apera, This story is finished, 702. ^ — Examples tdken from Rankin's Arab Tales translated from Swahili : — SwAHiLi : Mozambique : Kila mmoja katika wale wezi (p. 2) Moz' a z^'ale... weyi , Each one of those thieves... ... akatia zile dinari. (p. 4), ... kuliela ole vizurugu, ...and he put those ^'i^J pieces of money inside. Wakija,\\3iVidLivezi...(^.ii), Ala weyi yarod..., when these thieves shall come... Napaa huyu amekuja. (p. 6), Na-nazoro ola ahoroa, yet this gazelle has come. Tulize sisi paa huyu, na-ufito huu, Ntuiniheri nazoro ola, ni-mtali ola, ni-mwalu na kisu hiki. (p, 8). ola. Sell us this gazelle, and this stick, and this knife. 703. — Other examples : — Ganda : E kifananyi kino kia anil Whose is this likeness ? (New Testa- ment). Swahili : Ya-nani sanamu hii ? do. Chwana (Suto) : Secwanco sena ke sa-mants, ? do. Mpongwe : Edidi zina za-iiiande ? do. FUNDAMENTAL DEMONSTRATIVE PRONOUNS (Continued.) Class KA-TU. Singular KA-samo. Plural : TU-samo ', ' V' Position. 2'» Pes. 3" Pos. i"^' Position. a"* Pos. f Pos. Tonga aka (a)kano ako (a)kalia otu (o)tuno otuo otulia Nyamwezi aka ako ... utu uto Boondei aka ako kada ... Kamba "(?; kayu kaaya twii (■>) tuyu tuuya Karanga ako ... (i)otu Ganda kano ako kali (Cf. Class BU) Herero (i)nga J ngeni 1 ngena f^O f suini ( suina Angola akp. oko kana otu oto tuna Lower Congo efi(52i) efio (e)fino otu (475) ... oto (o)tuna Yao (a)ka (a)kano (a)ko (a)kala (a;tu (a;tuno (a)tuo (a)tula Demonstrative Pronouns. 177 Angola ; Senna : Kafir : Mo kiluiji eki kia masoxi, in this vale of tears {Angola Cat., p. 2). Ndoko kadzuke luku ii, go and Wash this spoon. Yopula i nyama le, or Yopula le nyama, take this meat out of the pot. § 2. Emphatic forms. 704. — 'In the generality of the Bantu languages great stress is laid sometimes on the last vowel of the demonstrative pronouns of the 3d position in order to express great distance. Ex. In Tonga : okulia, there (far); muntu oulia, that man (far). In Kafir : paya, there (far) ; u mntu Iowa or lowaya, that man (far). In Kamba : mtu uuya, that man (far). (Last's Kamba Gr., p. 28). In Swahili : mii u\t, that tree (yonder, far away). (Rev. P. Delaunay's Swah. Gr., p. 31). 705. — In Swahili, Kamba, etc., another kind of emphatic demonstrative pronOun is formed by reduplicating their full forms. Such pronouns lay stress on the strict identity of a thing. Ex. In Swahili : Akalala palepale, and he slept at that very place. Mto uleule, that very river yonder. In Kamba : Umama paapae, you may stand just here. 706. — In some other languages, as also in Swahili, emphatic FUNDAMENTAL DEMONSTRATIVE PRONOUNS. (Continued.) The two classes KO ( non-locative and locative). Non-locative class : KU-tui Locative class : KV ■nsi I"' Position 2" Pos. 3" Pos. I"' Position ,2'* Pos. 3" Pos. Tonga oku (o)kuno oko (o)kulia oku (o)kuno oko (o)kulia Kagurii aku ... kuko(?) kudia Boondei ... kunu kuno(?) kuda ... kunu kunoC.?) kuda Nyamwezi uku uko uku ukunu uko ikudia(?) Kamba ... kwaa kuyu kuuya . .. kwaa kuyu , kuuya Swahili huku ... huko kule huku huko kule Senna uku ... uko kure uku uko kure Katanga (i)oku ... ... (i)okuya (ijoku okuno oko (i)okuya Ganda ... kuno okwo kuli ... eno eyo (l)oko Xosa-Kaflr oku ... oko okuya (l)oku okuya Zulu-Kaflr loku ... loko lokuya loku ... loko lokuya Herero Angola (i)ngui oku ... oko j ngumi 1 nguina kuna (i)ngui kuku nguno koko ( (ilnguini \ (ijnguina kuna Lower Congo oku ... oko (ojkuna oku oko (ojkuna (a)kula Yao (a)ku (a)kuno (a)ko (a)kula (a)ku (a)kuno (a)ko Mozambique uu ... ... ule ngwe Ghw^ana ... ... ... •,'• koa koano koo koale Mpongwo ... ...■ guno gogo ,178 South-African Bantu Languages. forms are often obtained by adding to the simple demonstrative pronoun a substantive pronoun of some kind or other. Ex, In Kafir : Yiyo-le i ndlela, this is the very road (you are looking for). § 3. Copula-containing Forms. 707. — We find in Bantu two distinct kinds of demonstrative expressions which contain the copula. Those of the first kind render our " it is this, this is it, it is that, " etc. Those of the second kind render our" there he is, there she is, there it is, "etc. First kind. 708. — Those of the first kind, which we find in Tonga, Kafir, Senna, Chwana, etc., are mostly formed according to the same principles as the copula-containing personal pronouns-. Ex. In Tonga : Ng'-oj'/^, ng-oyo, ng-ouh'a mu-ntu, it is this, that person. Nz-«z/, nz-ezi'o, az-ezilia n-gombe, it is these, those cows. Mp-awa " it is here " ; mp-awo " it is there ", etc. In Kafir : Ngu-Io, ngvi-loit'o, ng\x-lowa m-niu, it is this, that person. 1^-eli, \-elo, l-eladada, it is this, that duck. etc., etc. FUNDAMENTAL DEMONSTRATIVE PRONOUNS (Continued.) The locative classes (P)A and MU. PA •nsi. MU-nsi. I"' Position. a"" Pos. 3" Pos. i'^' Position. a-^Pos. 3" Pos. Tonga awa ( (p)ano 1 awano awo (a)walia oniu (o)muno onio (o)mulia Kaguru baha . baho hadia .,. ... Nyamwezl haha haho ... Boondei hanu aho hada umu umo mda Kamba waa wayu waaya .... ... ... Swahili hapa hapo pale humu humo mle Pokomo habfa habfo habfade ... Senna apa pano apo pai'e muno J^aranga la 3r Congo imbique ana CI. KA-samo ka ke..'.'(-ko) ka Ike ku ka, ku kuna ka CI. TU-samo tu tu tu tu tuna tu CI. LU-llmi lu, ndu lulo o... (-we) u ... -lo ro(mue) e rwe (o... (-yo) ( olona ndu lu luna lu lo... (-ng) CI. CI. CI. (P)A-nsi KU-nsi MU-nsi (p)a, mpa ku mu haho he... (-ho) kwe,..(-ko) mwe... (-mo) ...-po ... -ko ... -mo po(tnue) ko(mue) mo(mue) we gie e... (-yo okona mue pu ku mu bu ku mu vana kuna muna . pa ku mu va fa e(?) 1 84 South-African Baiitu Languages. MU; -ini, -ng, -m, (548-555), are closely related to the adjective -mue " one ", and must therefore be said to signify primarily " one, together with ". 726. — 7° In Swahili also the relative particles are identical with the substantive pronouns. But they have this peculiarity, that, instead of introducing the relative clauses,as in most other languages, they are suffixed to their first verbal form, even when this is a mere auxiliary. (See examples n. 733)- 727. — 8° The relative particles in Chwana do not differ from the simplest forms of the demonstrative pronouns. But in this' language the suffix -;?g ( = Kafir -jj/^, 723, 725) is generally appended to the principal verb of a relative clause. In Congo the relative particles look like demonstrative pronouns of the third position without their initial article (693*). N. B. More information is wanted on the proper forms of the relative particles in the other languages. § 2. Use of Relative Particles and Construction of Relative Clauses in General. 728. — The proper use of relative particles, and in general the construction of relative clauses, is the main difficulty in all Bantu languages. That of treating it here is considerably increased by the variety of the languages with which we are dealing, and by their divergencies on this very point. For clearness' sake we may consider separately : 1° The relative clauses in which the antecedent is represented by the subject of the verb, as in Mu-ntu u-a-fua... or u mu-ntu u-a-fua,.., " the man who is dead... " — 2° The relative clauses in which the antecedent is represented by an object of the verb, as in Mu-ntu ngu nd-a-bona, " the man whom I have seen. " Hence : — 1. Relative clauses in which the antecedent is represented by the subject of the verb. 729. — First Construction (without a relative particle). In most Bantu languages, when the antecedent is represented in the relative clause by the pronoun subject of the verb, this pronoun alone generally does duty as relative pronoun, and no relative par- ticle is used. This is the usual construction in Tonga, Karanga, Relative Pronouns. 185 Angola, Mozambique, Kaguru, Kamba, Nyamwezi, Mpongwe, and the Suto dialect of Chwana. In Kafir these relative clauses without a relative particle are found only after antecedents which have themselves no article. 730. — Ex. : — Tonga : Monze u-la busia bantu hSL-a-ka fua, Monze can raise to life people who are dead. Ndi-ue u-a-ka- ndi-loela mu-ana yx-a-ka fua ejilo, it is you who had bewitched my child wM died yesterday. Ni-n-gatiga nm-ntu u-sonda, lit. it is a Nganga, a man who smells, i. e. a Nganga is a man who smells people. Ue, tata uesu u-a-ka tu-bumba, Thou, our Father, -who didst form us. Ba-la sondela ba-rlu ba/a fuide, they come near, the persons who are riot dead. KARANGA-.y/ i-pone ixindi yi-tio-psanga nda? Where are the muircats which like to go ? Old Angola : Esue tn-ekala ko uze ou..., We who live in this world... (Father de ;Coucto's Cat., p. 34). Mozambique : M-tu a,-ruele, the man who went. (Chauncy Maples' Gr., p. 56). Kaguru : Mu-ntu any-enda, the man who loves me. (Last's Gr., p. 47). Kamea : Mu-ndu a.-ny-enda, the man who loves me. (Last's Gr., p. 28). Nyamwezi : N-zwik z-a-za, lit. hair which is red, i, e, red hair. Mpongwe : Nyare yire veno, i-nyare %\re veno, the ox 7iihich is here, the oxen which 3.re here. (Ms'" Le Berre's Gr., p. it). Suto : Leseli le-ienQ go uena... the light that is in thee. (Mat., 6, 23). Kafir : A-si m-ntu w-tanda a ma-hashe (not ...o-tanda), he is not (a) man who likes horses. Ndi-teta la m-ntu vt-hamba paya, (not .."o-hamba), I mean that man who is walking yonder. Kangela ela dada li-« m-lanjeni (not ...e Use...), look at that duck which is in the river. Wena, \x-hamba-ze (not o-hamba-ze), you who walk naked. 731. — Second Construction (a relative particle before the relative clause). This is the usual construction in Kafir, Chwana proper, Senna, Ganda, Yao, and Lower Congo. I find also examples of it in Tonga, but with this peculiarity, that the relative particle is placed before the antecedent itself, not after it as in these other languages. With regard to Kafir and Chwana we must remember that a suffix, viz. yo in Kafir, nQ in Chwana, is in given cases appended to the verb (723. 727)- 1 86 South-African Bantu Languages. 732. —Ex. Kafir : Ngu m-ntu o-tanda a ma-hashe ( = o u-tanda...) he is a man who likes horses. Ngu m-ntu o-ndi-tanda-yo, he is a man who likes me. N. B. Here the antecedent m-ntu being preceded by an article, the relative clause likewise requires a relative particle. CiiWANA PROPER : Lc-scdi ^& k-Un^ mo go wena. (Cf. Suto : le-ieli le-lenq, go uena, supra n. 730.) the light that is in thee, lit. the light that which is in thee. - Dinku tse di-timetsens, the sheep that have strayed. Monna eo o-na aka re-bolelela, the man who could have told us. (Rev. William Crisp's Gr., p. 52.) Ganda : A ba-niu a ba-genze, the people who have gone. C mu-ntu y-a-ja ( = e-y-a-ja) the man wAi? is coming. (Cf. French Ganda Gr., p. 30). Senna : Ku-um'ka kb-mue ku-li mw-a iwe, the light that'\s, in thee. (Mat., 6. 23)' * Mu-ana o-jnue wa-sua n-diro u-a-tawa, the child which was crying has^one off. (Rev. Father Courtois' Tete Grammar,^./^'].) Yao : Nyumba yij-a-gwile Use j-a-piU inoto, the house which fell yester-% day has been burnt. (Rev. A. Hetherwick's Gr., p. 34). Lower Congo : E n-taudi in 'o-kuizanza, the child which is coming. (Rev. Father Visseq's Gr., p. 25). • Tonga : U mu-ntu u-a-keza ejilo, the man who came yesterc^ay, lit. he the man who... 733- — Third Construction (a relative particle appended to the first verb of the relative clause). This is the regular construption in Swahili and Boondei. It is also met with in some Senna dialects. Swahili : Ki-su hi-kata-cho, the knife which cuts. J^i-su h'-na-cho anguka, the knife which is falling. Ki-su hi-li-cho anguka, the knife which has fallen. Ki-su hi-taka-cho anguka, the knife which will fall. Ki-su ki-si-ch.0 kata, the knife which does not cut. BooNDEi : Mu-ntu e-za-ys, the man who is coming. Mu-ntu enda-y& eze, the man who will come. 734;. — N. B. I. In Boondei this construction is generally coupled with the second. Ex. Muti we kugva-o, the tree which fell. 2. The Kafir construction with the suffix -yo, and the Chwana construction with the suffix -«g, may well be compared with this. 2, Relative clauses in which the antecedent is represented by an object of the verb. 735. — Here again we may distinguish two kinds of construction. In the first kind the antecedent is represented in the relative clause Relative Pronouns. 187 le relative particle alone. In the second kind the antecedent is lied either before the verbal stem by an objective pronoun nective), or after the verb by a substantive pronoun. 36. — First construction (the antecedent represented in the ive clause by the relative particle alone placed at the very nning of the clause). his is the usual construction, at least for affirmative clauses, in ga, Karanga, Angola, Yao, Senna, Ganda, etc., when the ive particle" represents the direct object of the verb of the ive clause. Ex. Tonga : Ka mu-cita zi-niuzi ndi-yanda, Do ye the things which I, like. Ka u-ndi'pe ci-niu ci nd-a-amba. Give me the thing which I have said { = Ka u-ndi-ps,--. cf. 274). I-sekua li nd-a-ka bona ejilo..., the duck which I saw yesterday „. Ndiue na u nda-ka bona ejilo I Is it you whom I saw yesterday ? B. In such clauses, whether for the sake of clearness or that of euphony, we often hose nasalized forms of the relative particles which contain the copula, viz. ngu, idr,n}i, mba, etc. (721), instead of u, a, li, i, ba, etc. Ex. /-/z7z' isekiia nA-u-amba ? li-u-amba) Which is the duck which you mean ? Herero : E purura ndi u-a-tora, the purura ivhich thou hast carried off. Kt) Angola : O y-uma -y-a-tu-tuma, the things which they order us. (Father, de Coucto's- Cat.) iRN Angola: O mbua i ng-a-jiba..., the dog which I have killed. (Cf. Hdli Chatelain's Gr., p. 95). Karanga : 1 nyika i nd-a-ka lebereka..., the ground which I have said. Yao ; Nyumha]\ tw-a-weni liso..., the house which we saw yesterday. Ganda : A ba-ntu be tu-laba, the men whom we see. Senna : Ma-u o-mue na-nena, the words which he says. B. In clauses of this kind in Senna the connective pronoun subject of the verb is ally understood, as in the preceding example, in which na-nena is for u-na nena. 37. — Second construction (the antecedent recalled by a second loun besides the relative particle). 'his is the usual construction in Tonga, and the other languages mentioned, when the relative particle represents an indirect :ct of the verb. I find it also in Tonga in negative clauses when antecedent represents a a^zV^tif object. 1 Kafir, Chwana, Swahili, and Kamba, it is the usual construction all kinds of relative clauses in which the antecedent is repre- ed by an objept of the verb. In Yao it seems to be as usual as first construction (Rev. A. Hetherwick's Gr., p. 34). Ex. : — 1 88 South-African Bantu Languages. Tonga : Ba-la loa mu-ntu u ba-ta vaw-yandi, they bewitch the man whom they do not like, lit. the man him they do not him like. In-gubo zi a-lapela anzio-., the clothes in which he prays, lit. the clothes ihem he prays with /hef>i. N. B. In such clauses the connective pronoun u of class MU-ntu is changed to a, 650. Old Angola : ... ne pango y-a-iu-fuila na-yo, ■ ■ ■ and the manner in which he died for us, lit. ... and the manner tha( he died for us with //. 738.- Kafir : Zi-ye pina i nkomo e ndizi-tengile-yo7 Where are the cows which I have bought? lit. They have gone whither, the cows that I them bought ? In-gubo a ba-iandaza na-zo, the clothes in which they pray, lit. the clothes that they pray with them. N. B. r. Kafir idioms : Ezi nqanawe rX-hamba a belungu or ...z\-hamba u beltingu nga-zo, lit. these ships (with) which go white people, i. e. which white people go by. — Hamba u-yo ku-ba eza nkabiiX-lima aba battfu. Go to steal those oxen vyhich those men are ploughing with, lit. ...(with) which are ploughing those people. — In such constructions, where that which should be the object of the verb is apparently made the subject, there is a great deal of analogy with the Tonga construction, only the real subject is understood. 2. Kafirs say, for instance : I nkomo a.-zi-tengile-yo (== / nkomo a.-a-zi-tengile-yo = ...o-a-zi-iengile-yo), " the cows which he has bought " ; and likewise : i nkomo a.-wa-zi- teiiga-yo = " the cows which he bought ", thus replacing by a the relative particle of class MU-nin. 739.— Chwana : Mo-lelo o re-o-iukisifse-n:^, the fire which we have kindled. Tlhobolo e kefudile-nq /ia-eone, the gun with which I have shot, lit. the gun that I have shot with it. (Cf. Rev. W. Crisp's Gr., p. 18.) Swahili : Neno gani a-li-\o li-sema? What is it that he says? Lit. Which (is) the word he is if saying it ? Kamba : Ka-indo ka ni-na )s.Si-onie iyo, the insect which I saw yesterday, lit the insect that I saw // yesterday. (Cf. Last's Kamba Gr., p. 29.) 740, — To complete this matter, we must add a word on the possessive relative " whose ", and the like, viz. " of which " and " of whom ". As a general principle it may be said that in Bantu the clauses which contain such a particle have a construction similar to that just described. Ex. : — Tonga : Ba-li ku-li bantu ba n-zim-pongo zi-a-\>o ezi? Where are the people whose goats these are? lit. ...the people they it is the goats oi them \htst. Chwana : Kgosi & le-fatshe e-lenSja-eone, the chief whose land this is, lit. the chief that the land is that of him (Rev. W. Crisp's G/:, p. 18). 741. — The usual Kafir construction equivalent to this is some- what idiomatical. Ex. : — Yi-nto e zandla zi mnyama, he is a man whose hands are black, lit. he is a thing 7vhich (has) hands that are black. Relative Pronouns. 189 42. — ■'V. B. Though these are the main principles which preside over the uction of relative clauses in the Bantu languages, it remains true that this point es further study. I have at hand several grammars in which these delicate questions to have been carefully avoided. I have others which in this matter are by no means le. V. — Bconouns in fiogseissiKJe ej:pre0sions. § I. General Principle. 5t3. -^ In most Bantu languages possessive expressions are ed by placing the particle -a before substantives and pronouns. ; from mu-ame " a king " we obtain -a mu-ame " the king's " of the king ", and from 60 " they " we obtain -ado " their ". g thus formed, these expressions are treated as if they were a of relative clause, or, in other words, as if the possessive par- -a were properly a verb meaning " to belong to..., to appertain ". Hence they require connective pronouns as well as relative ies. . IN ToNG.i : I nakaz' u-a mu-ame, the king 's wife, lit. the wife which (is) of the king. takazi ba-a mu-ame, the king's wives. la u-a mu-lavu, a lion's tail. 'a i-fl ba-lavu, tails of lions. 5/ x-aho, (z)im-bizi zl-ako, thy horse, thy horses. ., etc. 44. — In those languages however ' which require relative teles of various kinds in certain relative clauses (731), these icles are not generally required before possessive expressions, xceptions to these principles will be seen further on (761 and 778). Thus, in Kafir we have u-mfazi w-ako, thy wife (not fazi o w-ako, 732 and 775), and in Chwana we halve mo-sadi ■{ago, thy wife, (not mo-sadi eo o-a gago, 7Z^). 45. — ^- B. I. As may be seen from the examples just giyen, the principles re. to possessive expressions in general are applicable as well to the possessive ives -angu " my ", -ako " thy '', -akue " his ", -esu " our ", -enu " your ", -ado r ", -awa, -ayo, -alio, -alo, etc. " its ", etc. (656*, 658, 659). 46. — 2. In Mpongwe the possessive particle -a is not heard in ordinary possess- tpressions. Ex. Mboni yi ngowe (not mboni ya. ngowe), the chief's goat. But it is ed in possessive adjectives, as in ATboni ya-mi, my goat. 190 South-African Bantu Languages. § 2. Connective Pronouns Suppressed. 747. — Before possessive expressions such connective pronouns as consist of a mere vowel, viz. u, i, or a, are sometimes sup- pressed. Thus we may hear in Tonga mu-ana a-ngu " my child " for mu-ana M-a-ngu, tat'esu " our father " for iaia n-esu, etc., in Kafir i bokw* a-m " my goat " for i-bokwe y-a-m, etc. 748. — This, combined with various other principles, has pro- duced in several languages a remarkable series of nouns of rela- tionship, as may be seen from the following table : — Tonga Shambala Guha Kafir Herero Ganda Ch-wana Swahlll Mpongwe etc., etc. ray father tata tate tata f tatk \bawo tate ( sebo ( kitange rre babangu rere thy father his father uso uise IXO ixe so se u yihlo u yise iho ihe kito (?) kite (.') rrago babako ? rragwe babaye my mother ba-ma in-lala maju uma o mama ( mange ( nyabo mme mamangu ngi yami thy mother ba-nyoko nyokwe nyoko u nyoko o nyoko nyoko mmago mamako ngi yo his mother ba-nyena nine nina u nina ina nyina mmagwe mamaye ngue 749. — ^- S- I- Most of these words are easily analysed.- Thus in Tonga uso = u-si-a-o ==u-isi a-koj u-tse = u-isi a-e ^u-isi a-kwe ; ba-nyoko = ba-mau-a-ko (cf. 122), etc. In tata, ba-ma, the, possessive pronoun is understood. The word for " his mother " in Tonga, Shambala, etc., seems to be derived from the element anya, notion of " giving the breast ", and -ana " child ". 750. — 2. As has been said in n. 143, in Tonga the words for " mother '' are generally used in the plural number instead of the singular as a mark of respect. In some other languages on the contrary the words for " mother " may be used in the singular number, but not so the words for " father "- Thus in Mozambique the word athithi " my father " is a plural of class MU-BA, and in Kafir tata is generally used as a plural of class IN-ZIN. Hence we may hear sometimes tata z-am " my father ", tata z-ako " thy father " (== u yihlo), etc. In Senna bbth the word a-tatu " father" and a-tttai " mother " seem to be used always as plurals of cl. MU-BA. \ 751. "-~ 3' ^^ some languages the words for " father '' are oftener brought under cl. IN-ZIN than under cl. MU-BA. This is the case particularly in Angola, Nika, Swa- hili, etc. 752. — 4- Saba or Bawo is apparently borrowed from Ar.^bic or from another Semitic language, and in some languages it is not used properly with the meaning of " father ", but with that of " sir, master ", or as an honorific title. The true Bantu word for " father " is taia or rara {tate, n-tate, etc.) 753. — 5- The Rev. W. Crisp {Secoanu Gr., p- 21) notices some contractions in various nouns of relationship in Chwana which show distinctly that this language is Pronouns in Possessive Expressions. 191 sgnated with words .borrowed from several others. Tims the possessive expression -eno ir, yours ", is borrowed from Tonga, or Karanga, or Kafir, to form the words rraeno ir father ", mmaeno " your mother ", etc. A-ke'{= a-nge) "mine ", is borrowed Karanga to form the words moro-ake " ray son ", nnake " my younger brother " laranga nonange), mo-gatsake " my spouse " (= mo-gadi-ake, cf. 205), etc. 'S^, — 6. In Ganda, among other e.xpressions similar to those above mentioned, lay notice baze " my husband ", baro " thy husband ", etc. (French Ganda imar, p. 26). § 3. Possessive Expressions after Locatives. 55. — Locative expressions give rise to a great variety of >truction for the possessive expressions whicli depend on them. IS : — 56. — 1° In Senna, Nyamwezi, Karanga, Mozambique, etc., lessive expressions which depend on locatives regularly admit connective pronoun corresponding to the locative classifier of r noun. Nyamwezi : Ha-numba h-a wawa " at the house of my father "• Ku-numba ku-a wawa " towards the house of my father ". Mu-numba mu-a wawa " in the house of my father " Karanga : Ku-mberi ko \winu ^wirire " before all things '', pe-juru pan-sece ( =pa-un-sece) " on the earth " ; pakaii ^tnjizi m-biri " between two rivers ", mu-kati mu-e-muniba " inside of the house ". Senna : Pa-kati pa akazi " amidst women " ; pa-maso pa-ace " before his eyes " ; apano pa-kati pa pili pa ini-sozi " here in this vale of tears '' ; m-mimba mw-anu " in your bosom "; ku-viusa ku-a nzou " at the abode of the elephant ". 3ZAMBIQUE : Va-zulu va-ia " over it '' ; mu-Mna m-a-ia " inside of it ", etc, 57. — 2° In Tonga, Herero, Angola, Kongo, etc., the possess- expressions which depend on locatives admit only in a few ;s the connective pronoun corresponding to the locative prefix : e commonly they require the connective pronoun corresponding be, proper classifier of their noun : in some cases they require no lective pronoun at all, principally when the locative expression rmed with an adjective. 58. — Ex. Tonga: With 2. locative pronoun v Ba-lala l^u-nsi ^\x-a-manzi ;y (the Mbunda) can sleep at the bottom of the water ". Enda ku-nembo ku- " walk before me ". ith a non-locative pronoun : A ]s.di-ti ka-a ma-cedo " in the middle of night ". Ku zii-a-bo " in their villages ". Mu n-ganda i-a-ko (or through assimilation mu- ida &-a-ko) " in thy house ". 192 South-African Bantu Languages. Without any connective pronoun : Ba-la njila mu-kati a-manzi " they go into the water '. U-kede ku-iala a Si-ongo, ku-nsi a Mukuni " he lives above Siongo (Victoria ' Falls), below Mu-kuni ". Mu-nsi a muse (or mu-ns'a muse " under the ground "). 759. — Other examples : Herero : Tua p'-e kunia rV mu-vero (Kolbe's Did.) " put it down behind the door " ; Ko me/io yoye " before your eyes " ;.,^V taw-rungu u-e " before hira " ; m'o ka-// k.'o meva " in the midst of the water ". • Angola (from F. de Coucto's Catechism, 1661). Mo kumbi rl-akufua ku-etu " in the hour of our death ". Mo ki-luiji ekiki-a ma-soxi " in this vale of tears ". U-ekala ko rorbando y-a lukuakolu-a kuria " he sits at the right hand... " Congo. Mwia di-ambu di-a... " on account of... "; kuna iQse In a... " before the face of... "etc. (Cf. Bentley's Diet. p. 612). JV. B. It is worthy of notice that, the diminutive classifier ira having been lost in Congo, the ancient expressions formed with ka-ti " middle, centre " are now connected with fol- lowing substantives by the pronoun ku, as in mu-na 'k.a.-ti ku-a... " in the centre of ". This connects Congo with Kafir, Chwana, etc. 760. — 3° In Swahili the possessive expressions which depend on locatives require different connective pronouns according to the meaning of the locative expression. Ex. Xaii y-a njia . " in the middle of the road '' ; juu y-a-ke " over it '', etc. (as if katiandj'uu were nouns of the class IJVoi MA). Kati ka He jiwe " in the middle of that stone " (as if kati were here a noun of class KA, a class nearly obliterated in Swahili). Nyuvtba-xa. Ww-a-ke " at his house " ; kanwa-ni kw-a Muungu " from the mouth of God " (as if the locative suffix -ni were here equivalent to ku). Mi-kono-ni mw-a-o " in their arms " ; ...uvuli-ni vaw-a ma-uti " in the region of darkness " (as if the locative suffix -ni were here equivalent to mu). Ma-hali pa raha " the place of rest " (as if ma-hali were a locative of class PA). 761. — 4° In Kafir and Chwana, where the mechanism of locatives is perhaps still more disturbed than in Swahili, the possess- ive expressions which depend on old locative expressions are in most cases connected with them by means of the pronoun ku (Chwana go). Other locative expressions require the connective pronoun corresponding to the proper classifier of the noun which they contain. Ex. KafiS. 'Pezi/\i^-a-ko " over thee ". 'E.-calem kvr-a-m " at my side ". 'Ezantsi k.vr-en-taba " below the mountain ". — E-a-dliniy-a-ko " in thy house "- Chwana. Kwantle ga motse ( = go-a motse) " outside the town ". Fa-gare ga ba-sadi " amidst women " { = ...go-a ba-sadi). Mo-ienz ga lesaka ( = go-a le-saka) " inside of the kraal ". — Mabogom aona " in their arms ". 762. — 5° In Mpongwe the locative particle ^lO generally does duty for all the locative classifiers of the other Bantu languages, and Pronouns in Possessive Expressions. 193 1 — ' ~ as a mere preposition. Hence in this language, when possess- :pressions follow locative expressions, the connective pronoun . is required is, as a rule, that which corresponds to the Rer of the noun which is preceded by the locative particle. Go nyumiy-ani " behind you ". 3. — N. B. In the same language some ready-made expressions remind one of liar constructions of the other Bantu languages. Ex. AH ga impumi Uani " within irs " (cf. Swahili kati)/.s...., Tonga a ka-tik.a..., etc., n. 758) ; m-pangini ga ndego : quality of friend " (cf. Chwana supra), etc. 4. — Concerning the locative expressions which mean " near " far ", it must be noticed that in nearly all the Bantu languages ire generally followed by the preposition which means " with " ja «, Karanga, Kafir, etc. na, Chwana le, etc.). iRANGA : pa-fupe n^-mwnba ( = t\3i-imumba) " near the house " Kafir : ku-fupi aa-m " near me " ; ku-de no »/// ( = na-u mti) " far from the tree ". Ihwana : kgakala le rona " far from us " ; ga-uchwane le motse " very near the town " ; etc., etc. i5. ■ — Finally, in this matter we again find applications of the pie of avoiding monosyllables. For, when possessive express- should be reduced to mere monosyllables, this is avoided • by appending them as suffixes to the preceding word, or by :ing a relative particle to the possessive expression. The first ese forms may be remarked principally in Ganda, the second aranga. ANDA : O Moyo-gwe " his heart " ; mu-kcino-gwo " thy hand " ; e kanzu-yo, " thy cloth ". ANGA : Ba-nona \iZ.-b-e " their brothers " ; zina li-ri'-o " thy name ". d6« — ^- B. I. In Karanga the possessive is expressed by a suffix when it con- ic consonant proper. Ex. Nona-\xo " thy son " (= nana u-a-o). 37. — 2. Following a somewhat similar principle, Kafirs generally say u-m-ntan, my child ", i n-gubw' am " my blanket ", etc., instead of u m-ntana w-a-m, i ngubo etc. m VI. — BLclatitie anD fioggessitie ejcpresgions uscD ButJgtantitielp. 768. — Relative and possessive expressions may be used sub- stantively, viz. as subjects or predicates. Supposing, for instance, that a Tonga had spoken of " feasts " mi- Ha, he may say : I-e im- pewo nji-a-ku-sanguna, lit. " (those) of winter are (those) of the beginning, i. e., are the first ". In order to understand the formation of such expressions in Bantu, we must distinguish between those languages which have articles and those which have none. Hence : — 769. — \'^ In those languages which have no ar/icie, such as Tonga proper, Senna, Chwana, Swahili, etc., when such relative and possessive expressions are used substantively as subjects or objects, they appear to have commonly the same form as when used adjectively. When used as predicates they require before them a copula-containing relative particle. These laws however suffer ex- ceptions, and rnay require tb be modified when reliable materials are more abundant. Ex. In Tonga : A. Relative clauses. "Without a copulative prefix. "With a copulative prefix. U-a-ka sanguna, ngu Monze, he who began Monze x\%\l-a-ka sanguna, lit., Monze it was Monze. . is he who began. K-luma batitu, nga-masaku (or m'masaku), Ma-saku, nga.-Iuma ba-ntu, the devils are lit. they who bite men are the devils. they who bite men. 1-la inka a bantu babi(in-zila), nji li-a mu-lilo ; i-la inka a baboiu, nji li-ei kukondua, lit. that (road) which goes with bad people, that is it which has fire; that which goes with good people, that is it which has happiness. 770. — B. Possessive clauses. U-a ku-sanguna mu-ntu, ngu-Adamoi the Adamo ngu-a ku-sanguna, Adam was the man of the beginning (i. e. the first first man. man), was Adam. ZA-enu (zintu) nzezi, yours are these Nzi-2r«««, ^z/z/w///, they are yours, these (things). ' things. Li angu (isekua),ndeli,mme is this(duck). E^lisekua ndi-ti-angu, this duck is mine. Lu-a Leza (lu-zubo), m-baame, lit, that Ba-ame, ndu-Iu-a-Leza, the chiefs are (the race) of God, they are the chiefs. God's race. Ci-ako (cinii:) need, thine (thing) is this. Eci ciniu aci-ako, this thing is thine. Relative and Possess. Expressions used Substantively. 1 95 771. — N. B. I. In Tonga, for clearness' sake relative and possessive clauses very often adntiit that form which begins with the copula-containing relative particle. We have seen above that this is done particularly when the relative pronoun is the object of the verb (736). It is also done regularly when the relative or possessive clause is of some length containing several distinct words. This may be considered as a sort of bracket construction. 'E.Ti. Baa jaya mberere n^a. ku-pa-ila "they have killed the sheep for the sacrifice " (lit... "it is that of the sacrifice ") -jBa-lalia nyama nja ku-sunsiainsima " they eat the meat which has been cooked with the porridge " ,(lit. " ... it is that of flavouring the porridge ".) 2. This kind of bracket construction seems to be particularly frequent in Herero.. 772. — Examples in other languages : — ^. Without copulative prefix. Relative clauses. Senna : Muka, u-guliise ciomwe uli na-cio, go, sell what thou hast (Mat., 19, 21.) QviynMiK: Ea,o-bapatse tseo-nanfi na-co, do. SwAHiLi : Twaai-li-yo yako, take what is thine (Mat., 20, 14.) etc. Cf. Mat., 5, 3-10, in the various translations of the New Testaipent. Possessive clauses. Chwana : Cula ecugago or cula eeo eagago, take what is thine (Mat. zo, 14.) Mpongwe : Wong' i-ya, do. SwAHiLi : Y-a nani sanamu kii...? ... YaKaisari. Whose is this image...? Cesar's. (Mat., 22, 20) eta B. With a copulative prefix. Chwana : Pitsa e e-ihubegilenn ke-« mosetsana oiia a-e-reka neogala, the pot which is smashed is that which the girl bought last year. (Rev. W. Crisp's Gr., p. 18). Karanga : Ndi-r«-fl Wange (ru-jubo), it isWange's (family). ,; Swahili : Ufalme wa mbingu m wao, theirs is the kingdom of heaven. (Mat., 5, 10). SuTO : Lemololena ke la maml ... Ke la Kesare. Whose inscription is this? Caesar's. (Mat,, 22, 20). 773. — ^- B. It is remarkable that in some Senna dialects, though the copula before substantives, is generally «^z (587), nevertheless before possessive expressions it has forms similar to those of Tonga (721). Ex. N-kaziuyu ngn-anif Ngu-a/tga. This woman, wkose is she? Mine. Mi-adia ii aji-ani? Nji-aHga. These canoes, wAos^ are they ? Mine. Ci-kazi ici rxax-ani? Nci-anga. This bottle, wAose is it ? Mine. 774. — 2° Those languages which in given cases ,/4az/^ an, article before substantives require a relative particle, or a simple article, in similar cases before possessive and relative expressions when these are used substantively. Thus : — 196 South- African Bantu Languages. 775. — A) In Kafir such expressions require the relative par- ticle 0, e, or a (723), where substantives would require the articles, z, or a, Ex. O warn u mfazi mtsha, o wako mkulu, my wife is young, yours is old. O sebenza kakulu, ndim lo{ = o usebenza. . .), lit. he who works much, it is myself. Reciprocally, no relative particle is used where substantives require no article (317). Ex. Wetu,yiz'apa (not o wetu...). Our (friend), come here. Yinina, belu (not ...a betu) ? What is it, our (friends) ? Lo warn u mfazi (not lo o warn), this wife of mine. A si warn u mfazi (not a si warn), it is not my wife. - : Where substantives require to be preceded by a copulative prefix (58)2) the possessive and relative expressions are likewise preceded by a copula-containing relative particle, viz. n^o in cl. Ml/, nga in classes BA and MA, ye in classes .^7 and IN, le in cl. LI, lo in cl. LU, bo in cl. BU, ko in cl. KU, ze in classes ZI and ZIN, se in cl. SI. Ex. Lo mfazi ngo ka bani?. This woman, whose is she? Eli Cuba le li ka bani? Le lam. This tobacco, whose is it ? It is mine. La ma-hashe nga ka bani? Nga weiu. These horses, whose are they .' They ^' are ours. 776. — B) In Herero, Angola, and Congo, the same sort of expressions require an article where substantives require one. Ex. Herero : ruvio o ruandye, the knife is mine. Cf. ruvio ruandye, my knife. (Kolbe's Did.) Congo : E yame mbele ivididi, my knife is lost, lit. mine knife is Jost, or it is my knife that is lost. (Bentley's Gr., p. 523). Angola: O yatuma Santa Ngeleja..., that which Holy Church commands... (De Coucto's Ca/., p. 6). N. B. This last example exhibits a relative clause. Possessive clauses used substantively have no article in Father de Coucto's Catechism. 777. — In Ganda expressions of the same sort, when not used as predicates, generally require a simple article ; in a few cases they prefer a peculiar kind of relative particle which much resembles the demonstrative pronouns of the first position in several languages, viz. orUy ebi, eyi, etc, Ex. Mudu wange Miiganda, o wo Musoga, ray slave is a Ganda, thine is a Soga (French Ganda Gr., p. 25), Genda otunde ebi bio (not simply e bid), go and sell what is thine (Mat,, 19, 21), Twala eyiyo (not simply eyo), take what is thine (Mat,, 20, 14). lative and. Possess. Expressions ttsed Substantively. 197 V^hen used as predicates, they remind us of the Kafir construction ve noticed by requiring as a kind of copula that kind of substan- pronoun or relative particle which ends in e (724). c. O bwakabaka o bwo mu gulu bwe bwabwe (Kafir U bukiimkani ba se zulwini abo), the kingdom of heaven is theirs. Cf. 830. [. — B.elatit)£ anD Boissessitie erpression0 equivalent to our KDjectiues, 78. — It has been mentioned above that in Bantu adjectives Der are comparatively few in number, and that their apparent t is supplied principally by relative clauses and possessiye ressions. I now add a few remarks for a better understanding lis principle. Thus : — ' In Tonga the words which correspond to our adjectives ressive of colour, sensible qualities, exterior form, etc., are itly verbs, such as ku-tuba " tjo be white ", or more exactly " to )me white ", ku-salala " to be red ", ku-sia " to be black ", ku- ma " to be straight ", ku-pia " to be hot ", etc. Hence' the :ctives " white, black, hot, " etc., of our languages pass aly as verbs in Tonga. Absolute clause : Ma-nzi a-a-pia, the water is hot , lit... has become hot. Relative clause: Ndi-yanda ma-nzi a-a-pia, I want hot water, lit... water that has become hot. f Absolute clause : Ei n-zila i-luleme (-lukme = -pei{ect oi-lulama), this road is straight. Relative clause : Inka e in-zila i-luleme, go by the direct road, lit. . . by the road ^ that is straight. B. Expressions of the same kind are found in all the Bantu languages. 79. — 2° In Angola and Congo most of the expressions which espond to the quantitative adjectives of the other Bantu lan- jes (601, 603) have the form of possessive expressions. Such for instance, -a ?n-bote ( — a mu-bote or perhaps -a bu-bote) lod. ", lit. " of goodness ", -o-nene [—a u-nene= a bu-nene) eat ", lit. " of greatness ", -o-be {= a u-be) " new ", lit. " of mess ", o-^utu { = a u-kulu) " old " lit. " of growth ", etc. Zme ngi mu-tu u-a mbote, I am a good man ( = Tonga : ndime mu-ntu mu-bolu). lye u-o-nene, thou art great ( = Tonga : iwe mu-nene). 198 South-African Bantu Languages. 780. — N- B. Expressions similar to these, but for different kinds of adjectives, are found in nearly all the Bantu languages. Ex. Kafir : lyem le-nene " a true, genuine medicine " ( = iyeza la i-nene''^& medicine of a truthful man ", from i-nene " a gentleman, a man who does not ' cheat "). Hence in Kafir ngo kwenene " in true language^ truly " ( = ngo ku-teta kw-e-nene, lit. " in the language of a gentleman "), SwAHiLi : M-iu w-a choyo " a greedy person ", (lit. " a man of small heart", from ch-oyo = ki-oyo " small heart " ; cf. m-oyo " heart "). etc., etc. VIII. — jJtonouns referring to Substantities unDerstooD anti fironoun0 useo as Conjunctiong. 781. — Connective pronouns and others are often used with reference to substantives which are entirely understood, being not even expressed in preceding sentences. Their meaning must then be made out from the context. The number of the substantives w^hich may be thus understood is however limited. In Tonga they are principally the following : — In class LI : i-zu6a " the sun, a day. " Ex. lA-a ku-sanguna li-a in-sipi, the first S; , (day) of the week. » MA : »?a-«2rz '* water. " Ex. U-a-yala a-bu-enga, pa a.-ka selelela, he went along the bank, where it (the water) rushes down. » BU : ^«-«/4« " night. " ^\. A-lahu-d..., when it (night) has not yet cleared up... » KU : notion of action, time or tnanner. Ex. Ta-^in-kondua a lu-sek, no work is done on the day of the new moon, (lit. // is not worked ) » CI : ci-ntu " a thing. " Ex. Nci-nyamanzi CO o-yeya ? {= d u-eya, 251). What is »y that you are thinking about? N. n. I. In Kafir the word z n-te " a thing, " being of class IN,Xht connective pronoun used with reference to this word, when it is understood, is t. Ex, : X-mnandi u ku-ncokola na-we, " it is pleasant to chat with you. " 2. In Tonga the plural zi-nlu " things " is seldom understood. But in Angola its equi- valent j/-ma" things " appears to be as often understood as the singular ki-ima " a thing ". Ex. I ua-ngi-dele, nga-i-ria kia " (the things) which you gave me, I have eaten thetn already ". (Hdli Chatelain's Ki-mdundu Gr., p. 143). Pronouns referring to Substantives understood. 199 In class LU : lu-zuho " family, race. " Ex. Ba-leya bala tulua \\x-a-baana. The Lea are deprived (by the Rotse) of their children, lit. of that (part of their fa- mily which is) children. — Ndulo- ndulo ndu-lu-a Davide {ci. 770), it is David's own (race). N. B. In Kafir the connective pronoun lu'is often used with reference to u-suku "day" understood, exactly as in Tonga li is used with, reference to i-zuba. Ex. O lu-« si-tatu, \\x-e sinne, etc. " the 3" day (of the week), the 4* day ", etc. 782. — Hence some ready-made locative expressions which have the form of possessive expressions', such as voM-a-kale " to the bottom "; lit. " unto the (inside part) of the end " ; ku-a-kate " for ever", lit. "to the (time) of the end "; k\x-a-Mpande "at Mpande's (place) ", etc. Of course in such locative expressions the connective pronoun cannot be understood as it often is when its antecedent , is expressed (757). Ex. 'Ba-lapela Mpande ka be-enda a bu-botu k.U-a-kaIe, they pray to Mpande that they may go in happiness for ever. Mu-nari u-a-njila tau-a-kale, Livingstone went in right to the bottom (of the water). 783. — Locative expressions of the same description as the preceding are commonly found in nearly all the Bantu languages. Ex. Kafir : kw-etu, at our place ; kw-ake, at his place ; kw-a Gcaleka, at Gcaleka's place ; etc. SwAHiLi -.kw-etu, at our place; mw-etu, in our house; kw-a mamae, at his mother's place ; etc. Ganda -.eib-ange, at my place; ew-a, at thy place ; ew-e, at his place; etc. Cf 546. TV. B. In Kafir and several other languages, when those, substantives of cl. MU-BA which have no classifier in the singular, as u-yise " his father ", u Sa-rili " King Kreli ", etc. (342), have to come into possessive expressions, they are first made into possessive locative expressions of the kind just described. Ex. i nkomo za 'k.-wa.Rili " Kreli's cattle ", lit. " the cattle of Kreli's place ''. This particle kwa is in Kafir sometimes contracted to ka. Ex. U mnf a kwcuTixo or Umnt' a ka Tixo, " the child of God ", lit. " the child of God's place ". Ga {== go-d) is used in Chwana where Kafir has kwa or ka, and in a few other cases. The regular use of this particle before certain possessive pronouns is particularly remark- able, as in batho ba ga.-gwe " his people " (Tonga 6an/u baakue) ; pitse ea sa.-go, " thy horse " (Tonga im-bizi iako). Pronouns used as Conjunctions. 784. — As a result of the principles just laid down, some relative particles have come to be regularly used ,as conjunctions. 200 South- African Bantu Languages. They may then be considered as referring to certain notions of time, place, or manner of thought, understood. Ex. : — iLI (referring to i-zuba, the sun) = " when ", with re- ference to a determined moment of the day. Ex. Baamuzika li bu-cia, they bury him when night is clearing up. 785. — (P)A = " when ", with reference to successive actions. Ex. Pa-« kafua muntu, bala muzika, when a man is dead, they bury him. -« N. R. The Swahili relative clauses which correspond exactly to the preceding contain the relative particle -po, in accordance with the genius of this language (726). Ex. Tu-li--^Q ku-ja..., " when we had come "... (Tonga Pa tu-eza...) (P)A = also " where ". The other locative particles, viz. ku and mu, may likewise be used as conjunctions. Ex. "ParCi-ka tuba, (in the part of the body) where he is white. Mu nd-a-ka njira, where I went in. A'. B. I. Cf. in Mozambique : Ya no-kelaka, wherein I am entering ; u (= ku) no-kuma nchua, where the Sun comes out. 2. Cf. in Herero : Ku me-kara, where I stay ; o n-dyuo mu tu-rara, the house wherein we sleep. 3. Cf. in Ganda : To-tnanyi -we (or gie) n-sttla. You do not know where I live ; U-a-laba nyumba in-we n-sula f Have you seen the house wherein I live.'- — We, gie, jnwe are relative particles corresponding respectively to the locative classifiers wa, e, mu {= pa, ku, mu). Cf 719 and 540, 546, 552. etc., etc. 786. — BU = " supposing that..., if... ". Ex. Bu tu-bona u-bereka nawo.... Suppose we see you working thus... N. B. I. Cf in Chwana BO == " as if". Ex. A o-lirajalo bo fiqoanya?ia ? Do you act thus, as if you were a child ? (Rev. W. Crisp's Gr., n. 74). 2. Cf in Ganda the relative particle Bwe " if". Ex. Bwe o-no-genda ruegulo, o-no-iuka kiro, if you go this evening, you will arrive during night. Ne tt\iLaba-lia,ti ba-kkuta, even if they eat, they will not be satisfied. (French Ganda Gr., p, 40J. 787. — CI = " while, if ". Ex. Ci tu-bereka..., while we are working. N. B. I. Cf in Herero : tyi 7na mu-tyiwa, if ye know ; tyi iu-a-kara koyenu, when we were with you. (Rev. F. W. Kolbe's Diet.) 2. In Herero the relative particle (i)ndu (referring to o ru-veze " time ") is used in the same manner for " when, while". Ex. ...ndu ma tii-ya, when we come. 788. — KA == " if, when, while, and ". Ex. Ba-lia ka ba-ti..., they eat saying at the same time... Siabulvngo u-a-toligua a Leza ka a-ci lu-sabila, Siabulongo was taken up by God while he was still a baby. Pronotms referring to Substantives understood. 20 r £a-aka sika, inyue ka mu-7i-lede, they arrived while you were asleep. Ka alt a-fuefui, uti,.., ka a-li ku-le, nti..., if he be near, he says..., if he be far, he says... Cf. 970. APPENDIX ON THE LUNDA LANGUAGE. 788'^'^. — While reading over the last proofsheets of this article I received Henrique Augusto Dias de Carvalho's " Methodo pratico para fallar a lingua da Lunda (') ", a most valuable addition to Bantu literature. As I had till then only a few pages of this work, my conclusions on this important language were limited to the few remarks laid down above in nn. 143 and 144. Complete as the same work now is, it furnishes good materials for comparison with these pages. I. Phonetics. — Lunda has a great deal, in common with Angola, Lower Congo, and Mbamba, more particularly with the last. Its most remarkable feature seems to be the uncertain sound of certain voWels, and the transition of some others to a. Ex. ku-mana " to see " (Tonga ku-bond), mwlambo " a lip " (Tonga mu-lomo, Angola mu-zumbu, Dualla mo-lumbu, cf. 360*), n-zavo " an elephant " (Tonga inzovu), etc. The following verbs may' be compared with the table of examples under n. 52 : ku-hia " to steal ", ku-mana " to see ", k'ovua " to hear ", ku-fua " to die ", ku-da " to dawn ", ku-fika " to arrive ", kw-eza " to come ", ku-jala " to dress ", ku-nvala " to beget ", ku-nua " to drink ", ku-dia " to eat ", ku-lald " to lie down ", ku-dila " to cry ", ku-dima " to hoe ", ku-su?na " to bite ", ku-neta or ku-leta " to bring ", ku-enda " to walk ", Tiu-tuma " to send ". //. Substantives. — ■ Lunda has the r 2 classes of substantives described in a pre- vious chapter. Ex. : — CI. MU-A (= MU-BA) : Muntu " a person ", a- ; mw-ana or mona "a child", a- ; viH-kaje " a woman ", a- ; mu-ata '< a chief ", a- ; mu-roro " a servant '', a- ; Nzambi " God " ; Mu-kuarunda " a Lunda ", a-; tatuko" father " ; maku " mother ". CI. MU-MI : Mu-jimba " the body ", mi-; mu'-Mla " the tail ", mi-; mu-tue " the head ", mi- ; mu-lambo " a lip ", mi-; mu-xima " the heart '', mi- ; mu-iondo " a tree ", mi- ; mu-pueji " a stream ", mi- ; mu-vo " a year ", mi-. CI. N-JIN : Mbiji o\ nama " meat "; n-gombe " one head of cattle ",jin- : v'l-pembe " a goat ", jim-; n-zolo " a fowl ", jin-; n-naka " a snake ", jik- : n-zavo " an elephant " ; jin- ; n-vula " rain " ; n-jila " a path ", jin-. CI. (D)I-MA : Z'/-«:i<7 " a day ", ma-; di-su " an eye ", mesu; di-zeu " a tooth", ma-; di-fupa " a bone ", ma-; di-yala " a stone ", ma- ; di-jina " a name ", ma-. CI. (B)U-MA : Ma-rufo " wine " ; u nga " flour "; u-ato " a canoe " ma-u- ; u-cuko " night " ; mema " water " CI, KU: Ku-hua " to fall ". Only infinitives of verbs seem to belong to this class. CI. LU-JI(N) : Lu-dimi oxRu-dimi " the tongue ''. According to Carvalho the plural of this word is ji-dimi, and, in general, the plural classifier of this class is not jin, but//. It seems scarcely possible' that this should be correct. The plural classifier of lu-dimi must be jindimi, and, in general, if the n of the classifier jin is not heard in some words, it must be only before hard consonants, according to nn. 151 and 283. ' : i. Lisboa, Imprensa nacional, 1890. 202 South-African Bantu Languages. CI. CI-I (= CI-ZI) : Ci-ounia " a thing ", /-/ ci-kanda " a hide ", /-; ci-lalo " a bridge ", /- ,• i-kumbo " a hut ", i-. "■ CI. KA-TU : Ka'-kungi " a youth ", tu- ; ka-sme " fire ", ka-bwa " a dog ", fu-. liocative classes PA, KU, and MU : Fa-xi or pa-nd " down "; mu-ixim "within"; paisuipa " near " -jpa-lepa " far " -jpola " outside " ; pa-kaxi " between "; pe-uro " upon " ; ku-nyima " behind " ; mu-cikumho " in the hut ". The author gives also the locative classifier BU. But is not this again a mistake? BU is the Angola classifier corresponding to the Lutida PA. Hence, for instance, when he says, p. 159, that " down " is rendered into Lunda by paxi ox boxi, must not his words be understood in this sense that boxi is the Angola equivalent for the hunda. paxi? I find in Lunda the two interesting locatives po/o and kolo ( = pa-ulo and ko-ulo), both of which mean " a place ". I have as yet noticed their exact equivalents only in Chwana in the words/e/o and golo. Cf. 537. /// Adjectives. — The laws for the adjectives which I term quantitative, such as wape " good ", -ipe or impe " bad ", -jima " great ", kiepe " small ", -iepa " long ", -ki " new ", etc., are the same as in Tonga (596)- Ex. mu-tondo mu-jima " a large tree ". Pronouns. — The connective personal pronouns seem to be ni " I ", u " thou ", u " he ", iu " we ", nu ox mu " you ", a " they '', u, i, lu, di, etc. The substantive personal pronouns seem to be ami " I ", eye " thoii ", endi " he ", ecu " we ", enu " you ", ene " they ", etc. But in Carvalho's work I remark a certain inconsistency in the forms of those pronouns which correspond to classifiers containing m (cf n. 649). Thus I find ma-zui ama ai-mi " these words of mine " (p. 205) next to ma-i ma nzolo " eggs of fowls" (p. 51), ma-ciko ma.-oso " all the days " (p. 227) next to ma-dho ama s^-oso " all these days '' (p. 231), mu-tue vaxxa mona " the head of the child " (p. 209) next to mu-t'ue u-ei " thy head " (p. 223), mu-jikita ou " this work " (p. 136) next to mu-lambo omu " this present " (p. 135), etc. Other conclusions on Lunda will be introduced into the following pages. Strange to say, many words in Lunda remind one of the languages which are heard near Delagoa Bay. IX. — J^umeralg. § I. Bantu Numeration. 789. — As far as I have been able to verify, counting among the Bantu is done principally with the aid of fingers. Old Kafirs, for instance, seldom express a number by the proper word, but they show it by a motion of the hand which they accompany with the expression zi-nje " they are so many... " or ba-nje, mi-nje, ma-nje, etc., according to the class of the things in question. N. B. The following is in general the meaning of the principal signs : — Raising one of the small fingers alone : counts i T> » with the next » 2 » » with the next two > 3 » » with the next three » 4 > the five fingers of a single hand, » 5 » the thumb alone or the thumb of one hand with the five fingers of the other » 6 T> the thumb with the index » 7 » » with the next two fingers...... » 8 » > with the next three fingers 1> 9 Both hands laid flat against one another » 10 Ten is a. kumij two tens (26) are 2 ma-kumi or opening both hands twice ; three tens (30) are 3 ma-kumi or opening both, hands three times, etc. One hundred in Kafir is a kuhi, i. e. " a large number ". In many languages it is 3.gana. 790. — There are however also for the different numbers the proper words or expressions, which may be used when required. These are partly adjectives of one kind or another, partly sub- stantives. Thus among the Tonga and other tribes of the interior, there exist numeral adjectives up to five, but 6 is five-and-one, 7 is five-and-two, etc. Ten is expressed by the substantive ikumi, a hundred by ma-kume-kumi, which is a superlative, of " ten ". Beyond that there are in Tonga only " tens withput number ", makumi a-ta balui. In most of the other languages there are numeral adjectives up to 6, and substantives or foreign words for the other numbers. In a few languages " a whole man " is " twenty ". In general South-African natives will see at a glance that one of their goats or head of cattle is missing even in a very large flock or herd. Yet they are very slow at counting properly, until they have been taught pur own methods, which, it may be remarked, they adopt readily. 791. — On the point of numeral adjectives the Bantu languages 204 South- African Bantu-Languages. go two different ways. Most of them, like Tonga, usually treat them as pronouns, so that they incorporate connective pronouns, not classifiers. Others, like Kafir, treat them as quantitative adjectives, so that they incorporate classifiers, not connective pronouns. 792. — I subjoin comparative tables *, which exhibit in their bare form the numbers i, 2, 3, 4, 5, 6, 10, and 100, in the principal Bantu languages. Where two forms are given for one number in the * COMPARATIVE TABLE OF NUMBERS. 1 2 3 4 Tonga -mui(i4) -bill -tatu -nne Bisa -mo -wili -tatu -ne Subia -moina -bere -tatue -ne Ungu -mwe -wili -tatu -ne Hehe -monga -wile -datu -tai Bunga -weka -sona (?) -lila(?) -dai Gogo -monga -bili (-yete) - -datu -ne (-ena) Kaguru -mue -ill, -bili -datu -nne Kondoa -mosi -ili, -pili -tatu -ne Sharabala -mwe -ili -tatu -nne Bobndei -mwenga -idi -tatu -nne Zeguha -mwenga -idi -tatu . -nne Kami -mosi -ili, pili -tatu -nne Taita -mojoeri -bili -datu -nne Nyaniwezi -mo -wiri -datu, -yatu -nne Sukiima -mo -wiri, -bili -datu, -yatu -nne Nyambu, -mwe -wili -datu -ne Ganda -mo -bili -satu -nya (-na) Kamba -mondi -eri -datu -na SwahUi -moja -wili, pili -tatu -nne Pokottio -manda -wi, pili -hahu, -tahu -nne Nika -motsi -iri, -biri -hahu, tahu -ne Senna -bozi -wiri, piri -tatu -nai Karanga -muempera -biri -tatu -nna Xosa- Kafir -nye -bini -tatu -nne Zulu-kaflr -nye -bili -tatu -nne Herero -mue -vari -tatu -ne Blhe -mosi -vali -tatu -kwana K-wengo -morika -bari -tatu -nana Rotse -mue -yeri -atu -nne Nyengo -moya -bitri -ato -nne Guha -mo -wiri -sato -nna Rua -mo -biji -satu -nne Angola -yari -tatu -wana Lo-wer Congo -moxi -ole -tatu -ya Nywema -mo -fi -satu -neng Yao -mo -wili -tatu m-cheche Komoro -monsi ' -bili ..., taru -nne Mozambique -moka -ili (-nli), pili -raru, taru -chexe Chwana -ngwe -bedi, pedi -raro, tharu -nne Mpongwe ■mori -wani, -baili -raro, tvaro -nai Dualla -wo, po -ba -lalu -nei Fernan^lan -bul^ -iba -ita (?) -ela (-ele) Lunda -mue -adi -sato -nyi Numerals. 205 same language, they cannot be used indifferently, but the first-placed generally is the regular one, while the second is that used under a nasal influence (55-59). Where however the second is between brackets, it is merely a dialectical modification of the first. N. B. The elements -ka or -nga, •jz'or -zi, etc. in the words of the column for the number " one " originally meant " only, exclusively, by itself" (814-818 and 824). The Bantu word for " one " is properly -mue (variously changed to -mo, -ng,we, -nye, -bo), etc. COMPARATIVE TABLE OF N,UMBERS. (Continued.) 5 6 10 100 Tonga. -sano -sano a-mue i-kumi ma-kumekumi Bisa -tano -tano na-mo i-kumi i-gana Subia -Janue -omoiana (?) kume Ungu -tanu -kaga kumi ka-gana, Hehe -hano -tahdatu chumi chi-gana Bunga -fundo ■mfu li-hundu ... Gogo -hano -tandatu i-kumi i-gana Kaguru -sano -tandatu i-kumi i-gana Kondoa -sano -tanda kumi gana Shambala -xano -tandatu kumi gana Boondei -xano -tandatu kumi gana Zeguha -xano -tandatu kumi gana Karoi -thano -tanda i-kumi i-gana Taita -sa!nu -tandatu i-kumi i-gaona Nyamwezi -tanu, -hanu mkaga i-kumi i-gana Sujiuma -tano, -hano -tandatu i-kumi i-gana Nyambu mxa mkaga i-kumi i-xana Ganda -jano (-tano) rakaga kumi ki-kumi Kamba -Irano -thandatu i-kumi i-yana Swahili -tano sita (Arabic) kumi gana Pokomo -zano, -dsano -handahu, -tandahu kumi' gana Nika . -dzano -handahu, tandahu . kumi gana Senna -sano (-canu) -tandatu kumi dzana Karanga -xano -xano na-mue kumi makume-makumi Xosa-Kaflr -hlanu, -tlanu -tandatu i shumi i kulu Zulu-Kafir -hlanu -tandatu i shumi i kulu Herero -tanu -hamboumue mu-rongo e-sere Bihe -t,anu e pandu ekwi cita Kwengo -tanu -tanu na-mo li-kumi Rotse -tanu aombomue li-kume Nyengo -tanu -temoya (?) ni-kume (?) Guha -tano -tanda kumi gana Rua -tanu -samba ki-kwi Tia-twa Angola -tanu -samanu (ri)-kuinyi kama Lower Congo -tanu -sambanu ■ ■ e kumi ' n-kama Nyivema -tano -samalo vum lu-kama Yao m-sanu m-sanu na-mo li-kumi ma-kume li-kumi Komoro -sano ... tandaru kumi i-jana Mozambique -thanu -thanu na-moka ni-kumi ma-kumi (?) Ch-wana -tlanu -rataro, thataro le-shume le-kgolo Mpongwe -tanij -tyani o-rowa i-gomi n-kama Dualla -tanu ■d-um Fernandian -ito ito la buli biu Lunda -tano -sambano di-kumi ci-tota 2o6 I „ South-African Bantu Languages. § 2. Formation' and Use of the Numbers from " One " TO " Six ". 793. — 1° According to what has been previously noticed, the numbers from " one " to " six " in Tonga, Bisa, Herero, Kamba^ Nyamwezi, Ganda, Nyambu, Guha, Rotse, etc., incorporate the connective pronoun corresponding to the classifier of their noun, and consequently their construction is essentially identical with that of possessive expressions. This however is remarkable, at least in Tonga, that such numbers often admit before themselves, me- rely, as it seems, for the sake of emphasis, a copula-containing relative particle, and then the connective pronoun which should follow them immediately is generally understood, so that we hear, for instance, uli a ci-to nci-mue " he possesses one ford " (not... nci-ci-mue). Ex. : — Tonga: A) Without a copula-containing relative particle : Baatiku muntu \i-miie..., they said to one man... Bali e ingorna zi-iatu, they have three drums. Uaka cita (miezi) i-taiu..., (miaka) i-nne, he remained there three Hlonths..., four years. B) 'With a copula-containing relative particle ; Miezi iejinza nji-sano a i-mue, the months of the. rainy season sxtfive and one. jBakede kule, imtezi ngu mite kuli Zuanga, they live far, at one month's distance from Wange. Bisa: Mabuzi gn-wili, two fowls. (Last's Polygl., p. 138). Herero : O muhona yu-mue, one Lord ; o rutu ru-mue, one body ; n-gamburiro y{-7nue, one faith, etc. (Rev. F. W. Kolbe's Herero Diet., p. 349). N. B. Instead of yu-mue, yi-mue, we should expect regularly u-mue, i-mue (cf. 639*) ; the presence of the initial j' is probably due to the fact of -mue being monosyllabic, and the consequent necessity of not exposing the whole adjective umue or imue to be sounded as a monosyllable through contractions or elisions (44). KamCa: Mundu yu-mwe, one man; inti u-mwe (wu-mwe }), one tree ; mi-longo i-tatu, three tens (i, e. 30), etc. N. B. 'tiotict yu-mue for u-mue, as in Herero. Nyamwezi : Ma-kumi a.-wiii (not ma-htmi ma.-wiii,SteeTe's Coii.,p. 4g),tv/o tens (20). GaNda : Nagamba mudu-we o-inu na bakazi-be ha,-satu..., and he said to one of his slaves and to three of his wives..., etc. NvAMBt; : Ma-kumi a.-wili, two tens ( = 20). Last's FoL, p. 160, etc., etc. Numerals. 207 794. — 2° In Kafir, Chwana, Mozambique, Swahili, etc., the numbers from " one " to " six " are treated as quantitative adjectives, and consequently incorporate the classifiers of their nouns (cf. 604). Ex. IN Kafir. A) Nurabers used as epithets (616). Wa-tenga a mayeza a ma.-^afu, he bought three medicines. N. B. The numeral nye or nye-ga " one only " causes its noun to be used without an article, and consequently does not admit itself any relative particle (6i6). Ex. Una mfazi xa-nye-ga,he has a single wife. B) Numbers used as predicates (618). Mangapina a mahashe apo? M.a.-iaiu. How many horses are there near you?Three. Other examples : Chwana: Ba-sadiba ha-raro batla sila, the three women will grind ; cf. Basadi hBrraro, the women are three. (Cf. Rev. W. Crisp's Secoana Gr., page 27). Swahili: Meno yake raZrwili, his two teeth. Dinari tatu { = ...ntatu, cf. 282), f , three coins. Mozambique : j?/«w awe ap men/t ( = ma,-ifilt), his two teeth; aiu a.-ianu nenlt ( = ...na-a-inli) five and two men ( = 7 men), etc. N. B. After substantive prdnouns numerals are treated somewhat as suffixes in Kafir, Chwana, Swahili, etc. Ex. in Kafir : bo-ba-bini, both of them, lit. they both ; zo-p-tatu (in-kojno), the three of them (cattle), etc. (675). Hence in Ganda that kind of dual formed with the suffix -mbi (== -bili) " two ", (692). § 3. Formation and Use of the Numbers above " Six ". ^ 795. — The numbers above " six ", when they are not complex (796), are generally substantives, and, as such, require Various par- ticles before them, according as they are self-standing, or predicates, or in apposition to other nouns. Ex. : — Tonga : 1-kumi lie imbefe're { = li-a-itnberere) or Jmberere kumi, ten sheep; Imberere ndi-Aumi or zi-h'-ikumi, the sheep are ten. Kafir : In-komo e zi li-shumi, lit. cows they which (are) a ten, i. e. ten cows. Jn-komo zi li-shumi, the cows are ten. etc., etc. § 4. Complex Numbers. 796. — In complex numbers, such as " five and two (seven)", "ten and one (eleven) ", " a hundred and three", etc., care is always had to give to every number \\s, proper prefix. Ex. : — Tonga : Ndabona iagombe zialimsL-kuminga-nnea zi-lalu, I have seen cows which were 4 tens (40) + 3 (cows), where it may be noticed that 2o8 South^African Bantu Languages. nga-nne agrees with the tens (ma-kumi), while zi-iatu has to agree with the cows (zi)n-%ombe. Old Angola : Mi-rw va.a.-kuim 2,-tatu ne-itatu, 33 years, lit. Years tens (which are) three + three (years) ; A-tatu agreeing with ma.-kutm and i-tatu ' with mivo, etc., etc. § 5. Ordinal Numbers, and Numerical Adverbs. ' 797. — In Bantu ordinal numbers are possessive expressions proper./" First " = " that of the beginning ", 2^ = " that of the second place (or order) " ; 3'^ = " that of the 3"^ place {or order) ", etc. Ex. I-zuba li-.a ku-sanguna, the I'^'day, lit. the sun of the beginning. I-zuba lia bu-biri, the 2^ day, lit. the sun of the second change, etc, etc. In several languages numbers are changed into quasi-ad verbs by prefixing to them one of the classifiers KA- or KU-. Ex. in Tonga : ka-mue " once ", ka-bili " a second time ", ka-tatu " a third time " (526) ; ku-bili " in two parts ", ku-tatu " in three parts ", etc. The negative particle before the number " one ". 798. — In Kafir and Bihe I find here and there before the number " one " a pe- culiar negative particle which does not seem to be used in any other position. Its form is na in Kafir, la in Bihe. Apparently it means " not even " Cf 570. Ex. : — Bihe : /ia u-mue " no one "j la-kumwe " nowhere ", la ci-mwe " nothing ", etc. Stover's Umbwidu Gr., pp. 40-41 , Kafir : Ngu bani na ongatshoyo... f Na m-nye. Who is the man who can say....'' No one. X. — Intettoffatitie Bronoung, anl) Various »eterminatities. 799, — Interrogative pronouns, and most of those determinatives which we usually term " indefinite pronouns " in Bantu generally incorporate the connective pronoun of their class, exactly as de- monstrative pronouns and possessive expressions. Only in a few languages some of them incorporate classifiers, not connectiv^e pronouns, being thus treated as quantitative adjectives. They also present in their use several peculiarities, varying according to the different languages. I shall touch on the most striking only. Interrogative Pronouns. loc^ \ I. The pronoun " How many?". 800. — The Bantu equivalent ior our " How many ? " is -nga-pi? lit. "going where .? going how far? " from -pi? " where?" and -«i« or -nga " go ". This is pronounced -nga-pi ia Kafir, Herero, Karanga, Senna, etc. nga-iornga-hi'm Tonga, Bisa, Subia, Kaguru, Shambala, etc., -nga {nga-i ( ? )) in Gogo, Nyamwezi, Guha, etc.; -nga-vi {ka,-vi(^')) in Mozambique, ka-e\n Chwana. Its equi- valent is -lingwa in Yao ; -meka in Ganda, -mia in Mpongwe, -kwa in Congo, -anaia in Yi^.mh^i, ku-xi m Angola, etc. This pronoun is treated exactly as the numbers from " one " to " six " : hence it incorporates a connective pronoun in certain languages, a classifier in the others (791). Ex.: — Tonga : Mi-samo i-ngai? how many trees ? Kafir : Mi-ii mi-ngapi? » N. B. In Angola Ki-kuxi? is used instead oi A-kuxi? in class A-ntu. Ex. A-ntu ki-kuxi? How many persons ? (Hdli Chate'lain, Zeitschrift, i88g-go, p. 304). >§ 2. The Pronoun and Adjective " What ? What sort of... ? " 801. A). — Originally the simple fortn corresponding to our " What?" was essentially, in the generality of Bantu languages, the bare classifier of the word which means " a thing ", or " things ", though a little modified in some cases, according to certain phone- tic laws. Hence we still have the following : — Pronoun " What ? " Cf. " Thing " or " things ". Tonga : nzi ? „ zi-n/u, things. Herero : /yt? or vi? „ tyi-na, a thing ; vi-tia, things. SwAHiLi : ki? „ ^\-tu, a thing. Gai«3A : ki ? „ yd-ntu, a thing. Congo : nki 1 „ ki-wa, a thing. Kafir :«/? ,, in-/o, a thing. LuNDA : ecil „ ci-ouma a thin^. Instead of the simple nsif the Tonga prefer to use generally nyama-nzi? \\\.. " what sort of meat ? what sort of stuff ? " In like manner, instead of the simple nif the Xosa-Kafirs prefer in most cases nto-ni? " what thing ? " Ex. : — Tonga : Udta-tiz\ ? or ucita njyamanzi ? What are you doing ? Herero : Maungura tyi ? What is he working at ? Motya vi ? What are you saying ? (Kolbe's Diet) Kafir : Wati-m ? What did he say ? Uteta ntoni ? What do you mean ? 14 2IO South- African Bantu Languages. An interrogative suffix, for instance -na in Kafir, -ke in Herero, is often added to such pronouns, as in general to other interrogative expressions. Ex. in Kafir : Uti nina ? What does he say ? Utetg, ntonina. ? ^yhat do you mean ? In some other languages the original pronoun for "what.?" is either more transformed than in the preceding or borrowed from a neighbouring language. Thus we have : — ' Chwana : -«g (= Kafir -ni). Ex. Obatla-n^} What do you seek ? Angola : -nyi {= do.). Senna : -//>'/ (probably for ci). Ex, Unafuna-dyil What do you want? 802. — Several of these particles have also a selif-standing, originally a copula-containing form. Such are in Kafir jj'm.'' "What is it? " (sometimes tyinif), in Chwana enQ (Ex. ke enQ? " What is it?"), in Swah'iYi liinif, in Ganda kiki?, in Tonga. m-nya7nanzzf, etc. 803. B) — The pronouns which are used for " What? What sort of... ? " either as adjectives, or with reference to a determined class, generally incorporate the connective pronoun of their class. In Kafir they incorporate its classifier. In a few languages they are invariable. These pronouns are the following : — Tonga : nyamanzt? with a connective pronoun. Ex. Uyanda musamo u-nyamanzi ? AVhat sort of tree do you want ? Angola : -anyi? (lit. of what?), with a connective pronoun. Ex. jEne ngimutu u-anyil What sort of man am I ? (Hdli Chatelain, ZeiiscAr., p. 204). N. B. In Angola the pronouns of the locative classes are bu-nyif ku-nyif mu-nyif not bte-anyif, etc. {ibid.) ' j Chwana : -anii ? (lit. of what?) with a connective pronoun. Ex. Selo se ke s-ans? What sort of thing is this ? (Crisp's Gr., p. 19). '^ Senna : -am'? with a connective pronoun. Ex. Mu-adia (ng)u«/«'/ What sort of canoe ? N.B. In Herero -ani? means " whose? ". Ex. muatye ingui o u-a«/.? This bag, whose is it? (Rev. F. W. Kolbe's Herero Did., p. 547). In Senna -a«« may also beused with the same meaning (Cf. 773.) Yao : -achil (lit, of what?). Ex. M-kalo vr-achi? What kind of knife? (Rev. A. Hetherwick's Yao Gr., p. 35). Kafir : -tiil with a classifier. Ex. / zinti za mti m-ni ? Sticks of what sort of tree ? N. B. In Kafir this adjective -ni causes its substantive to be used without an article. Herero : -ke? or -ngel with a classifier. Cf. Rev. F. W. Kolbe's Diet., p. 543. SwAHiLi : ganil invariable. Ex. Mtu gani? Kiiu ganil What sort of man ? What sort of thing? PoKOMO : ga ? invariable. Congo : nkial invariable, followed by its noun; etc., etc. Interrogative Pronouns. 211 § 3. The Pronoun " Who ? ". 804. — The pronouns for " Who? " are the following : — Tonga : u-an\ ?, pi. ba-ani ?, or with the copula ngu-.am ?, pi. mba-ani ? Ex. Ngu-a«« izina liakol What is thy name ? lit. Who art thou (with regard to) th^ name ? Cf. IN Nguru : Zina diako mbwanif What is thy name? (Last's Polygloiia, p. 47). Karanga and Senna : Sing, anil, or with the copula ndi-ani? Ex. in Karanga : Jina lirio ndi-anil What is thy name? Ganda : Anil pi. bani'i Ex. Aniadze? Who has come? Herero : Sing, ani^ at yani. N. B. Vrohahly yani f --= the Tonga copula-containing form ngu-ani? Mozambique : U-pani? pi. a-panil Kafir : U-bani? pi. o-banil, or with the copula ngu-bani? pi. ngo-banil Ex. Igama lake ngu-bani-na 1 What is his name ? Chwana : Mangi pi. bo-mang ? Ex. Ke bo-manii ? Who are they? Mpongwe : Mande ? pi. wa-mande ? Ex. Wa-mande mongi xino ? Who are these people ? Angola and Lunda : Nanyi'i pi. a-kua-nyi? Congo, Swahili, Boondei, etc. : Nani? Ex. in Congo : Nkumbu andi nanil What is his name ? 305. — ^- ^- '• " Whose? " is generally rendered by a possessive expression regularly formed from the above. Ex. in Kafir : I gania li-ka banif " Whose name?" Here it may be remembered that in Kafir the prefix of possessive expressions before individual names is ka, not -a, as it is ga in Chwana (783, TV. B.) 2. It may be noticed that here again in the forms of these pronouns Mpongwe differs more from Congo than from Chwana and Mozambique. (Cf. 213). § 4. The Discriminative Pronoun " Which } " 806. — As far as we can judge, in the generality of the Bantu languages the discriminative pronoun " Which ? " is rendered by an expression which means literally " the one which is where ? ", viz. ti in Tonga, ne in Herero, pi in Kafir, etc., with a connective pronoun. Ex. : — Tonga:-//? with a connective pronoun. Ex. U-yanda ci-bula cili? Which chair do you want ? lit... the chair which is where ? N. B. This particle -It? being a mere monosyllable generally requires its connective pronoun to be strengthened by a sort of relative particle when such a connective pronoun should be otherwise a mere vowel. Ex. Mu-samo o-«-//.? " Which tree?" instead of mu-samo u-Hf Probably for the same reason, when it is preceded by one of the copulative relative particles ngu, niba, ndu, etc., this does not cause the connective pronoun to be dropped. Ex. Ngu-u-li (mu-satno) ? " Which is it (the tree)? " not simply Ngu-lif Cf. Ngu- a-kafiia, it is the one which is dead {769). 212 South- African Bantu Languages. 807. — Other languages : — Hkrero : -ne? with a connective pronoun, and a relative particle in some cases. The same in Karanga. Kafir : -pi? with a connective pronoun. Ex. U-funa si-hlalo si-pi ? Which chair do you want ? N. B. When the connective pronoun is a mere vowel, it is strengthened by a semi-vowel placed before it. Ex. U-funa in-ii -wru- pi? "Which tree do you want? " Hence, with the copula : A^«-wu-pi u-mii o-ivu-funa-yo ? " Which is the tree you want ? " Cf. My Outline of a Xosa-Kafir Gr., p. 39. SwAHiLi : -pi? with a connective pronoun, yu-pi ? in class M-tu (cf. 806). ' Chwana : -fef do. Ex. Motho 0- fe ? Which man ? Angola : -ebil do. Ex. Ki-fua ki-ebi ? Which manner ? Congo : -eyi 1 do. Mpongwe : -e2 do. Ex. Nagu y-e ? AVhich house ? A-dombe m-e? Which sheep ? Kaguru : -ahokil^iX. . of where }), with a connective pronoun. Ex. IVa-ntu wahoki ? < Which men? 808. — In Tonga " When? " is rendered by izuba li-li? " Which sun ? " or simply li-li?, the word izuba being then understood, (cf. 782), and in Herero by ru-ne (0 ru-vese), lit. Which time ? This is probably the origin of the word for " when ? " in several other Bantu languages (Kafir nini?, Ganda di? Swahili linif, Chwana len^ ?) etc. When our "When?" means "Which year? Which season? etc., " it is rendered in Bantu languages by a full expression, as in Tonga ; Mu-aka ouli ? Which year ? N. B. The Tonga self-standing form a-li? "where?" is also properly the discriminative interrogative pronoun of class PA (536). § 5. Interrogative Pronouns used Indefinitely. 809. — In Kafir, and probably in several other languages, in- terrogative pronouns are often used with an indefinite meaning, such as " no matter who, no matter where, etc. " Then they are generally reduplicated. Ex. IN Kafir : Wena /5a-bani-bani..., thou, (child) no matter of whom,.. Way a pipipi..., he went, no matter where. § 6. The Pronoun and Adjective " All, Whole ". 810. — In Tonga " all " is rendered by -onse (sometimes -^«.fi? after a or e, 250) preceded by a connective pronoun. Contrary to Various Pyonouns. 213 most other pronouns, it has forms proper to the I'^sf and 2^ person, at least in the plural number. But in the 2^ person plural no-onse is used instead of mtt-onse. Examples : Iswe to-onse, all of us; Inywe noonse, all of you ; Batitu bo-onse or he-ense, all men ; niu-samo u o^se (or through assimilation o-onse), the whole tree, etc. A'; B. I. Ko-o»se (= ku-onse), when self-standing, means " always, permanently ". Ex. Nebombua u-la-kunka ko-onse, the river Nebombua flows permanently. 2. In many instances the word -onse is not so well rendered by " all " as by " Wii toge- ther (I and you, you and they, etc.). " 811. — The construction of the word for " all " in most other languages is essentially the same as in Tonga. But its stem varies as follows : Kafir -onke, Senna -onse, or -ense, or -onsene, or -ensene, Chwana -oilhe, Congo -onso or -nso, Swahili -ote, Mozambique -oteni, Angola -eselele or osololo or -ese, Mpongwe -odu, Ganda -onna or -enna, etc. In several of these languages a substantive pronoun is often used as a sort of determinative before this adjective when already completed by its connective pronoun, principally when it means " whoever, whatever ". Ex. In Swahili : Y& yvte atakaye pita, mpige. Whoever will pass by, strike him. A'. E. The stem of this adjective being a dissyllable beginning with a vowel, its form in class Mu-ntu has become somewhat iriregular in some languages, for instance, in Swahili jc-o/^, not w-ote, in Gut^Ah y-enna, not w-onna or w-enna, etc. 812. — In Herero and Karanga a particle is always required as a kind of determinative before this adjective, viz. a- in Herero, where the stem itself is -he, and bn- in Karanga, where the stem itself is -rire. Ex. : — Herero : O vandu z.ve-he, all men ; o-vina s,-vi-he, all things. Karanga : Ixindihw-jirire, all the muircats ; inyika hu-i'-ri're, the whole earth. 813. — Some Bantu languages have a sort of superlative form of this adjective which means " whoever, whatever, any one ". Ex. in Angola : Mutu u-ese u-ese, any man (who...) § 7. The Pronoun A-like " Alone, By himself ". 814. — 1° In Tonga the following is the formula of the express- ion which renders " alone " : — a + connective pronoun + like. Ex. Lesa a.-B.-like, God alone; Aba bantu a-ha-like, these men alone. N. B. In such expressions a is the kind of preposition described in n. 573. It is equivalent to our " by " in " by himself " 214 South-African Bantu Languages. We hear in the i"' person singular ndime e-ndike (=...a-(i) like) "It is I alone ", and in the 2^ iue alike ( — a-u- like) " thou alone "; the other persons are regular. 815. 2° In Ganda the stem of the word rendering " alone " is -okka or -ekka, in Karanga -d^a or -e\^a, in Kafir -edwa or -odwa. These follow the same laws as the Tonga -onke " all " {-onna in Ganda, etc.), and have their proper forms even in the i^'^t and 2^ person of the singular number. Ex. IN Kajfir : mna nd-edwa, I alone, = Ganda nze nz ekka = Karanga : ewe nd oha -jj/ena y-edwa, he alone, = Ganda ye y-ekka. N. B. As the Kahr stem -edwa " alone " seems to be foreign to Bantu, may it not be thought to be related to the word edua or dtia " one " in Fiji.? This reference to a Papuan language might seem out of place here, if it stood isolated. But it is warranted by several other signs of distant relationship between Bantu and several languages Oceania. (See Introduction, 3=" section). ' 816- — 3° In Qhwana the same word is -osi or -est. In some cases it admits before itself the particle ka ( = Tonga a, 573). In others it follows the same laws as the Kafir -edwa or -odwa. Ex. Ke nna ka-n-w/j It is I myself ; Ke-bone \i-osi it is they alone. 8X7, — N. B. In all these langtages the same pronouns are sometimes used to render " himself, itself ", etc. 818. — 4° In Herero " alone " is rendered in the first person singular by -erike, in the other cases hy peke (invariable). Ex. Mba- ende erike " I went alone"; ma-kara peke "he stays alone". It may also be rendered by poru- followed by a possessive expression varying according to the class. Ex. Poru-andye, "by myself"; poru-oye " by yourself " ; poru-e " by himself ", etc. 819. — S'' In Swahili " alone " is rendered by ;>i?y^^ (invariable) or hy peke y- followed by a possessive expression. Ex. peke y-angu " by myself " \peke y-ako " by yourself ", etc. Cf. in Mozambique yek-aga " by tnyseU", yek-eku " by ourselves ", etc. § 8. The Pronouns A-a-la-kue "He also ", A-ba-la-bo " They also ", etc. 820. — In Tonga a series of expressions rendering " he also, they also", etc. is formed according to a formula somewhat similar to that of the expressions for " alone ", viz. : — a + connective pronoun + /« + substantive pronoun. Various Pronouns. 215 Ex. Leza a-a-la-kue, God also ; aba bantu a-ba-la-bo, these people also. 321. — N. B. I. Notice -kue instead of -«* in class MU-ntu. In the other classes we have a-u-la-o (mtt-cila), a-i-la-io (mi-cila), a-li-la-lio (i-ziibci), etc. In the i'"' and 2" person the expressions corresponding to these are a-mbe-bo (= a-me-bo) " I also ", a-e-bo (a-ue-bo) " thou also ", asne-bo " we also ", a-nytie-bo " you also ". (Cf. 691). I fail to see distinctly the exact value of the particles a and /a in these expressions. 822, — 2° Possibly expressions of the same sort exist in Yao, as I read the fol- lowing in Rev. A. Hetherwick's Yao Grammar, p. 37 : " -alakwe, with the characteristics (connective pronouns) of the first cluss (Mu-ndtt), is frequently used in the sense of" this person ", " he ", " those persons '', " they ", and may be used as representing the third personal pronoun. K^. Ana-wani ajiu'i/ejangati ja.a.laK'we, "Who stole? Was is not he?" 823. — 3° ^" most other Bantu languages the expressions " he also, they also ', etc. are generally rendered by the preposition which means " with, and ", followed by a substantive pronoun. In Kafir the particle /lua " also " is often used along with such expressions. Ex. K"wa u inlambo wa-hamba na-wo, or U mlambo wa-hamba'krvirsi-na.- ■wo, " the river also went along ". § 9. The Pronouns rendering " Self ". I 824. — It has already been noticed (655) that in Bantu " him- self, itself, themselves " after- verbs are regularly rendered by a connective objective pronoun, such as si in Tonga, i in Chwana, ri in A-ngola, etc. Again, it has been noticed that in certain languages 'the same expressions are rendered after nouns by the same pronouns which render " alone" (817). But there is also in Bantu a special particle for " self ", viz. -nya in Tonga, -ene in Angola, -ini in He- rero, -enyewe in Swahili, etc. In Tonga -nya is suiifixed to substantive pronouns. Ex. a-ngue- nya, with him himself. In more emphatic expressions the substantive pronoun is again repeated after -nya. Ex. Ngue-nya-yxe, it is he himself; nce-co- nya-co, it is the very thing, etc. 825. In Angola -ene, in Herero -tnz, in Swahili -enyewe, etc., are preceded by connective pronouns. But in Angola mu-ene is used in class MU-tu instead of u-ene, and mu-nne in the locative class M U instead of mu-ene. In Herero all such pronouns require the article before them, e. g. veni (cl. VA-ndu), zeni (cl. ZO-ngombe), etc., and, in class MU-ndii, mu-ini is used instead of c u-im. In Swahili mw-enyewe is used in the three classes M-tu, M-ti, and Ursiku. 2i6 South- African Bantu Languages. N. B. Ill Swahili similar expressions are formed with -etiyi " one who has... ", followed by a determinative. Ex. Mw-enyi ku-penda " the same who loves " ; Ki-tu ch-enyi m-fOirongo '■ a round thing ". § lo. The Pronoun -mbi " Other, Different, Foreign ". 826. — In Tonga and Kafir the word " different " is rendered by -mbi, preceded by the projaer connective pronoun. But, because this stem is monosyllabic, the connective pronoun generally requires to be strengthened when it should be otherwise a mere vowel ; is it nor dropped after the copula. In Chwana the form of this pronoun is -pe (185). Ex. : — . .Tonga : Tinsi ngue pe, ngu-u-mhi. It is not he, no, it is another (man). . Kafir : Asiye, hayi, ngu-'wu-mbi or ngo ivu-mhi, do. Chwana : Ga ke na sepe, I have nothing else. 827. — The equivalent of this pronoun is in Herero -ai'ue with a connective pronoun. Ex. va-ndu v-arue, other people (foreigners, strangers). In Yao it is -ine, also with a connective pronoun. Ex. mu-ndii ju-ine, another man, etc. § II. The Pronouns " One... another ", " Some ... others ". 828. — In Tonga the expressions " one... another " " some... others " are rendered by the numeral adjective -mue " one " repeated. Ex. U-mue uati..., u-mue uati, the one said,.., the other said... Ba-?nue baasiala, ba-mue baainka, some remained behind, others went. N. B. Hence the repetitive expression -mueimd, " few, scarce ''. Ex. Bantu bamuemtii, few men. (Cf. Superlatives, n. 632). 829. — Likewise in Kafir they are rendered by -nye " one " with the proper classifier and an article. Ex. O m-nye wasala, mnye wemka, one remained behind, the other went. In Ganda they are rendered by the quantitative adjective -lala, repeated. Ex. M-.lala a-lia, m-lala talia, the one eats, the other does not eat, etc., etc. Hetrogpect on tbe Hrticle. 830. — We have seen in a previous chapter (321,4), that the nearest approach in pure Tonga to the article of Kafir, Angola, Congo, Herero, Ganda, etc., is a kind of relative particle occa- sionally placed before substantives as if to determine them. From this fact I there might have dra;wn the conclusion that Bleek had rightly considered the Kafir article as having originally been " a pronoun derived from the derivative prefix (classifier) which it precedes ". {Compar. Gr., p. 153). But this conclusion I have reserved for this place, that I may the better show in what relation the various kinds of the Bantu particles now reviewed stand to one another. 'Y\\& classifiers, which are essentially a kind of adjective or generic nouns, are the, basis of the whole mechanism of Bantu with respect to nouns and pronouns. The most elementary of all the pronouns is \ki^ connective pronoun, which in the various classes of the 3d person is itself nothing else than the classifier, weakened in some instances, strengthened in others, of the substantive which it represents (639). The connective pronoun, when emphasi- sed and made into a word, no longer a mere particle, becomes a substantive pronoun (656). This substantive pronoun maybe used in most Bantu languages as a relative particle (718) and then it becomes again a kind of enclitic or proclitic particle. It is properly from such relative particles that the article is derived in most of the languages in which it exists. And this is only natural, as articles are to substantives what relative particles are to relative clauses (774). Hence, for instance, the Kafir u nt-ntu " a person " might originally have been rendered by " he person ", exactly as o-tetayo is still exactly rendered by " he who speaks ". It is from the same relative particles, or directly from the connective pronouns, that demonstrative pronouns ?ir& derived (698). Thus we find that the derivation of the various kinds of pronouns in Bantu agrees perfectly with what might be suggested by rea- son itself, and by their natural relation to each other. In connection with this conclusion I notice that the Bantu demonstrative pronouns have become a kind of article in certain semi-Bantu languages. To borrow an instance from Wolof, a lan- guage of Senegambia, in this language an article consisting of a consonant and a vowel is generally appended to substantives. In the plural the consonant is always y, but in the singular it is in 2i8 South- African Bantu Languages. most cases the initial consonant of the substantive, exactly as the consonant of demonstrative pronouns in Bantu is regularly that of the corresponding classifier. The vowel is i for things which are near (ii^st position), u for things which are at some distance (2^ position), a for remote things. (Dard's " Didionnaire Frangais- Wolof, 1825 ", p. XIX). Itcan liardly be doubted that such articles were 43rig^aa% identical with the Bantu demonstrative pronouns. Ex. Singular. Plural. marre-mi^ the river (here) marre-yi, the rivers (here) marre-mu, the river (there) marreyu, the rivers (there) tnarrema., the river (yonder) fnarre-ya,, the rivers (yonder) daaba-d.{, the lion (here) daaba-y\, the lions (here) daalia-&a, the lion (there) daaba-yvL, the lions (there) daaba-As., the lion (yonder) daabaya., the lions (yonder) m/'gue-si, the leopard (here) satgue-yi, the leopards (here) saigtie-sa, the leopard (there) satgue-yu, the leopards (there) saigue-ss., the leopard (yonder) saigue-yai, the leopards (yonder) etc., etc. In general, African natives, endowed as they are with keen senses, and little accustomed to consider abstract notions, are fond of vivid descriptions, in which motions of the hand coupled with demonstrative pronouns necessarily play a prominent part. For instance, a native will seldom be heard using a vague expression like this : " He lost one eye " ; but, as he noticed which eye was lost, he will say : " This eye of his died ", pointing at the same time to one or the other of his own. Likewise, instead of telling you that there is a three hours' distance between two places, he will say : "If you start when it (the sun) is there, you will arrive when it is there ", and he will show you at the same time different points of the sun's course. The same remark accounts for the general use of motions of the hand and demonstrative pronouns to express numbers (789). When my native informants had to enumerate objects of the same kind, I never heard such expressions as " the first, the second, the third ", etc., but " the first " was expressed by " this " with the little finger pointed out, " the second " was also " this ", with the second finger pointed out, " the third " was again " this ", and so forth. The same remark again may account for the variety of descript- ive auxiliaries which will be dealt with in the next chapter. chapter V. ON VERBS. 831. — It is no easy task to coordinate my materials on verbs. On the one hand, the peculiarly descriptive Bantu turn of mind has introduced into the conjugation so great a variety of particles ; on the other, it is so hard to obtain directly from natives proper inform- ation as to their exact value ; besides, the correspondence of these particles in the various languages is so far from being plain, that in matter like this one does not see how to avoid either con- fusion or misleading connections. The plan which I have finally adopted is to consider : — 1° The fundamental forms of the simple verb. 2° The various auxiliary forms. 3° The copula. 4° The derivative verbs, including the passive voice, causative forms, etc. I. — FunDamental Forms of tbc Simple Verb. § I. Principal Parts of the Verbs in Bantu. 832. — We have here to attend principally to five sources of modification in the verb, viz. : — 1° The form of the verbal stem itself, according as it is mono- syllabic or polysyllabic, beginning with a vowel or with a consonant (837, 840, 843, etc.). 2° The class and person, as also in some cases the object, of the verb. This point has already been elucidated in the preceding chapter (63.7-655). For the convenience of the student in the follow- ing pages the connective pronouns are generally either set in different type from the rest, or separated from the verbal stem by a dash. 220 South-African Bantu Languages. 3° The difference of mood. Here we may distinguish four moods, viz. : — a) The indicative, naturally expressive of an actual fact, as tu- bona..., " we see... " , b) The> subjunctive, expressive of a fact still in the mind, as... tu-bone.... "that we may see ". c) The imperative, which might also be referred to the one or the other of the previous two, according to its various forms, as the quasiindicative bona, and the subjunctive imperative u-bone, both -of which mean " see thou ". d) The infinitive, or substantive mood, as ku-bona ''to see ". 4° Duration in the indicative mood. Here we distinguish two stages, viz. : — a) The transient or non-permanent stage. Ex. Tubona..., " we see..." b) The permanent or perfect stage. Ex. Tu-bonide "we have seen ". 5° The difference of actuality, according as the clause is afifirmative or negative. Ex. Afifirmative form : tu-bona..., " we see ". Negative form : ta-tu-boni " we do not see ". 833. — Thus, considering the variations which affect the verb in its endings, we are led to distinguish in most Bantu languages four principal parts, or different forms, of the verbs, viz. : — 1" An indicative, imperative, and participial form, ending in -a, as bona " see ", tti-bona... " we see ", ku-bona " to see ". N. B. There may be coupled with this form in Angola and in most other Western lan- guages a form ending with a sort of mute, or indifferent vowel, which varies as the vowelsound of the penultimate, as in Angola -jitni, from -ji7na, in tu-a-jimi (Tonga iu-a- zima) " we have extinguished ", and -nu for -nu\x, from -nua, in iu-a-nn. (Tonga tu-a-nyud) " we have drunk ". 2° A negative form ending in \ {e or i, 270), as -bone or -boni in ta tu-bone... or ta tu-boni " we do not see ". 3° A subjunctive form, which is also imperative, ending in e, as -bone in a tu-bone " let us see ". 4° A perfect form ending, in the larger number of the verbs, in -ide or 'i/e, as -bonide in tu-bonide " we have seen ". In Kafir and several other languages there may be added to these a fifth, ending in -anga, as in Kafir -bonanga in a si-bonanga " we have not seen ". Fundamental Forms of the simple Verb. 221 Hence, for instance, if we were to recite the principal parts of the verb ku-bona " to see ", we should say: bona, boni, bone, bonide, and in Kafir we should add bonanga. 834;. — N. B. I. In Swahili there are many exceptional verbs borrowed from Arabic, or from other foreign languages, which have a form ending in / where it should end in a, as ku-hinni " to refuse to give ". In the other languages there are very few such verbs. We may however notice in nearly all of them the verb ku-ti " to say, to do " (perfect -iede in Tonga, -te and -tile in Kafir, etc.). The form of this verb in Chwana is go-re (172 and 200). We may notice also in Tonga and several other laiiguages the verb '-kiizi or ^K-a2z"to know" (without a perfect, as far as I know). Another remarkable verb in Kafir is ku-tsho " to say so " (with reference to something already said or done). Its perfect form is -Iskilo. In Ganda this verb hais the form -tyo. 2. In Swahili there is no such perfect form as -bonide. § 2. Fundamental Forms derived from BONA. 1. Imperative form BONA " see ". 835. — In nearly all the Bantu languages there exists for the second person singular an imperative form which regularly is the bare form ending; in a, as bona " see ". In most languages the addition to this form of a pronoun which means " you " {-m, -ini, -enu, etc.) produces an imperative form for the 2^ person plural. Ex. See thou See ye See thou See ye Tonga bona Kafir bona bona«/ Kaguru langa langefii Herero muna BOONDEI ona onani ROTSE mona Kamba ona onaz Angola mona nion^«« Swahili ona onsLfii Congo moria ««mona Nik A ona Yao wona vvona«« Senna ona onani Mozambique ona onani Karanga wona vronani Chwana bona bona«jj Ganda labba Mpongwe yena yena«/ 836. — -^^ ^- I" Nyamwezi the forms corresponding to these have generally the suffix -gain the singular, -ge in the plural, as wonaga. " see ", pi. wonage. In some verbs, principally in those which end in -ia in the infinitive, these suffixes are replaced respect- ively by -ja and -j'e, and then various phonetic changes often take place, as in sumaja. " consent " (cf. ku-sumia, to consent), oiaja. " bake " (cf. ku-ocha, to bake). Cf. Steere's Collections for a Handbook of the Nyamwezi Language, pp. 67 and 64. 837. — The effect on imperatives of the phonetic laws relative to monosyllabic stems, and to such as begin with a vowel (44 and 46, n. 2), is remarkable in most Bantu languages, principally in the from the stem -ja, (( come " ■iza ff come thou, come ye " -J'« jj ,. -ja J) It, „ -eza )» -tla »j -tlha n „ -ya ti „ -iza i» „ za, - iza ]) 222 South-African Bantu Langtiages. verb " to come ". We may notice particularly the following forms : — Ganda : jangu, NvAMWEZi : nzagu,i^\. nzoji Heuero : indyo, pi. indyoni SwAHiLi ; njoo, pi. njooni BoONDEi : soo{ = n-soo), pi. nsooni Chwana proper : eintlo, pi. tlanfi SuTO : tlho, pi. tlhottQi MpoNGWE : yogo, pi. yogoni Lower Congo : wiza, pi. tiniza Kafir : yiza, pi. yizatii etc., etc. 838. — In Swahili the other monosyllabic verbs, and a few of those which begin with a vowel, take in the singular the prefix ku-, which is probably the pronoun which means " thou " (639*). Ex. ku/« (from -la) "eat thou ". There is no plural form properly corres- ponding to this ; for such imperatives as kuleni " eat ye " must be referred to the subjunctive imperative form (855). 839. — In Lower Congo there are probably no monosyllabic verbs. Verbs which begin with a vowel take in the imperative singular {2^ person) the prefix w- " thou ", as.w-e«a^a " go thou ". The plural is regular. 840. — In Kafir the verbs which begin with a vowel take the prefix jj/-, and monosyllabic verbs the prefix jz. Ex. y\ya " go thou " (from -ya), yenza " do " (from -enza). 841. — In Senna u is prefixed to monosyllabic verbs. Ex. ndya " eat ", ndyani " eat ye ", umua " drink ", nmuani " drink ye ". In the same language the verb ku-enda " to go " becomes in the imperative ndoio " go thou ". In Angola also we find this form ndoko next to nde, which has the same meaning (cf. 938). 2, Indicative form NDI-BONA " I see ". 842. — This form is obtained as a rule by prefixing the various connective pronouns (639) immediately to the form ending in -a. Ex.: — Fundamental Forms of the Simple Verb. 223 I see thou seest he sees ■we see you see Mfl'ibona «/langa «dabba ? tiiuhone mulange mulabbe mwone mhone ««mona /ohone mwone 6ahone a/alange ^alabbe a/aone ^abone ^abona ^rtbone /awone 858. — Thus it may be seen that in this form the connective pronoun of cl. MU-ntu is generally a. Here again Congo differs from the generality of the Bantu languages in having o instead of a. And, singularly enough, in the same language the vowel-ending of this form is a instead of «z si ona«z (ine) si «rfzone etc. ^zzzThis is one more of those instances which show the close relation between classifiers and auxiliaries. Cf. 970, 971. 989. — In Ganda the form of the same auxiliary is kia. It may also be rendered variously. Ex. : — Still : Tu-kia It lalamii, We are still alive, (= Tonga tu-ci It baumi). While : Tula ivano, tu-kia genda Mbuga, Remain here while we go to Mbuga. Not at all (with negation): Si-kia lia, T do not eat at all (I no longer eat?) Cf. Grammaire Rvganda, p. 46. 9Q0. — In Rotse its form is si. It equals " still ". Ex, : — 0-si tenda, he still does (= Chwana o-sa dira). 991. — In Kafir it has one iorm sa, which properly means " still, yet ", and another form sd, which properly means " already ". The latter form is used principally to introduce participles; Ex. : — Still : Ndi-sa ba-bona, I see them yet. , Ba-sa pilile, They are still in good health. No longer (with negation) : A ndi-sa ba-boni, I no longer see them. Already : Se be-qala u ku-lima, Already they begin to plough. Ms^ u kti-gibilisa le nja nga matye, uya ku-yi-bona se yi kufupi, se yi-kwela kiiwe. Never throw stones at this dog, you would soon see it close to you, already coming up to you. Kw-a-yiwa kw-a-fikwa se be-m-iulele, Tlje people went and came up to them, when they had already killed him. ' 992. — Itl Chwana the same auxiliary has the double form sa, which means " still ", and kile (Kafir and Suto se), which means " already ". To these may be added the form ese, which with a negative particle means " not yet .". Ex. : — Still: Osa-bona or o-sa ntse a-bona, he still, sees. (With regard to ntse, see n. 945). Already (in Chwana proper) : N-kile k-a-bona, o-kile o-a-bona. I have already seen, thoti hast already seen (= Kafir se ndi-bonile). • Already (in Suto) : Sehpe se-se se-beiloe. . . The axe has already been piit . . . Not yet : Ga ke-ese ke-bona, I have not yet seen. 993, — In Swahili the same auxiliary is pronounced ki. It seems to be, like the auxiliary /&« (971), a prefix inseparable from the prin- cipal verb, and to form exclusively a kind of participial expression. Ex. : — 17 258 South-African Bantu Languages. Paka a-ka-hama kjia tiyingine, a-ki katnata, The cat removed to another road, continuing to prey in the same way. (St. Swah. Tales, p. 248). Ni-ine amj)iwe ni-ku-pe khati, u-kiishe soma tu-fanye safari, I have beea. told to ' give you the letter, and that, ivhen you have finished reading it, we should start on. our journey. (Ibid., p. 152.) Jua li-ki ckiva wa-ka-fanya khema zao, wa-ka-lala. When the sun set they got' their tents ready, and slept. (Ibid.\ p. 158.) A-M-pita mtu hu-m-la, Whenever a person passes, he eats him. 9Q4:. — I cannot make out with certainty to which verbs these auxiliaries origin- ally belonged. The Tonga form ci, as well as the Ganda -kia, and the, Kafir and Chwana -sa, seem to belong to the verb -cia " to dawn •' (Ganda -&ia, Kafir -sa^td, cf. 52*), but this, while explaining the meaning " already ", would give no reason for the meaning " still ". On the other hand the Tonga ci may well be also a contracted form of -ietfe, which is the perfect of -ia/a " to sit, to remain ", and I suspect that, if a Tonga, for instance, were asked to .develop the notions implied in the sentence u-ci bona " he is still seeing ", he would render it by u-kede u-boiia " he remair^s seeing. '' Again, the Kafir j£ " already " is sometimes replaced by sele, which is the perfect of -sala " to remain behind " (=t= Tonga -siala), and, consequently, may well be derived from this verb, or from the Tonga -sia " to leave behind ", a simpler verb from which -siala itself is derived. Hence, finally, I am of opinion' that the auxiliaries ci, ki, kia, sa, se, etc., are in fact related to the various verbs -cia, -siala, -kede, etc., andJhat in some cases they have more of the meaning of one, in other cases of another, viz. where they mean " already ", or " to begin ", they originally were no other than the verb -cia " to dawn ", and where they mean " still ", or " to continue ", they are more directly connected with -sia " to leave be- hind ", or with -kede, the perfect of -kala. Cf. 502. < § 13. The Auxiliary NGA. 995. — I do not yet know whether this auxiliary is used in Tonga. It is found in Kdfir, Swahili, Pokomo, Herero, Ganda, etc., where it generally forms a kind oi hypothetical or conditional i&nsG.. It is derived from -nga " to wish " [originally " to bend the body, as when entering a Kafir hut (?)"]. Its perfect and subjunctive form is -7t£e. Ex. : — 996. — Kafir : U-ngaya na? Ndi-ngaya. Do you feel inclined to go ? I do. Nge ndi-si-za, ndixakekile, I wished to come, but I could not do so. Ezo nlaka zi-ku-fumene e matoleni azo, a kit-nge tandi u kupinda iiye apo zizalele kona, If these birds found. you near their young, you would not like to go again where they have their nest. U-ya %-azi na i nkomai Ewe, ndi-nga z-azi. Can you manage cattle ? I think I can. A maxalanga u-nga-H, u kiiba ute waya e-si-tya, aza a-ku-bona, a-nga ku-tya, lit. Vultures, you would say, if you went (near them) while eating and they saw you, that they have a mind to eat you. N. B. In cl. MU-ntu Kafirs say in affirmative clauses A-nga-bona " he may see, he would see ", not -a-nga-bona, but in relative clauses they say o-nga-bona " who may see " Ex. Nanku u mntu o-nga-ya, Here is a man who may go. Auxiliaries. 259 997. — SwAHiLi (from Steere's handbook, p. 139) : A-nga-nia, " though he be " or " he would be ''. Ni-nga-li penda, " I should have loved "or " had I loved... " Cf. 1002. Si-nga-U penda, " I should not have loved " or " had I not loved... " Kama u-nga-li ku-iva po hapa, ndugu yc^ngu a-nga-K pona, If you had been here, my brother would have got well. Ni-nge penda, " I should love " or " if I did love." Si-nge penda, ha tu-nge penda, " I should not, we should not, love " or " if I did, if we did not, love ". Kama u-nge ku-wa na akili, niali yakti u-nge dumu nayo, lit. If you were with wits, your property you would continue with it, i. e. if you were a man of sense, your property would still be yours. N. B. Nga is used only before monosyllabic verbs, the others require nge. 998. — PoKOMO ( Zeitschrift, 1888-89, P- 180) : Ni-nge dsakka, I should have loved. Ta ku-nge dsakka, We should not have loved. Yeo ni-nge ku-dsa mudsina, luka muntu tywangu ka-na-iveza, To day I should have gone to town, but my man could not (go). 999. — Herero (Zeitschrift, 1887-88) : Nga tu-tyite vi? \^ U-nga tu-tyite ?77 (?)], What do you wish us to do ? P. 191. Ngatu-zepere mumue,'Ltt us kill together. P. 200. Nu nga tu-zepe mama, And let us kill our mothers. P. 200. 1000. — Chwana: N-ka reka, o-ka reha, o-ka reka, I am, thou art, he is, inclined to buy. Ke-ne n-ka reka, I would have bought, etc. 1001. — An auxiliary like to the above, though perhaps more expressive of wish, is singa in Ganda, sinka in Ratse. Ex. : — Ganda ( Grammaire Ruganda) : Singa n-dia (^ n-singa n-lia), I should like to eat. P. 38. Singa tu-a-genze. We should have gone. P. 38. ■.Singa nina emmere, singa n-dia kakano, if I had food, I would eat now. P. 39. Singa w-a-genda edda, iv-a-ndi-tuse kakano. If you had gone before, you would already have arrived. P. 39. RoTSE (Livingstone's Mss.) : U-sinka ko-i-ba (= Chwana o-batla go-i-polaea or o-ka i-polaea), He is nearly killing himself 1002. — •^- B. I. The fact that in Swahili the auxiliary nga, and in Ganda the auxiliary sinka, is used not only in the apodosis, but also in the protasis of conditional sentences, must probably be explained by considering that in fact both may include the notion of some sort of wish ; so that, for instance, the Ganda sentence jz'^^a nina e mmere, singa ndia kakano might be rendered literally by " I wish I had food, (because) I should like to eat now ". 26o South-African Bantu-Languages. 1003. — 2. In Karanga the verb d(X " to wish " may in some instances be con- sidered as an auxiliary nearly equivalent to the Kafir nga. Ex. u-no-da gara (= u-no-da {li)ugara\ he wishes to remain. § 14. The Auxiliary TI "to Say ". 1004. — 1° In most Bantu languages the verbs which mean " to speak ", as -ambola in Tonga, are seldom used without being followejJ, and, as it were, completed by the verb -ti " tp say " (Angola -ixi, Chwana re, 172), as if such verbs did not mean properly " to utter sounds ", but only " to open the mouth ". This principle is generally extended to other verbs which express an act of the mind or the will. It may also be noted that generally there is no pause in Bantg after the verb -ti, but it is joined immediately to the sentence which is to it what a direct object is to a transitive verb. 1005. — Examples : — Tonga : Muame u-a-ka ambola u-a-ti " A mii-inke '', The king spoke and said : " Go away ". (Lit. the king opened the mouth (?) and said...) Ba-a-ka amba ku-H " Tu-la inka ''. They spoke to say : " We go ". V Ba-Ia nvuika kti-ti/a..., they are heard saying... {N'. B. Ku-tua " to be said " is the passive form of ,4?^-/'/, 1047). Ndi-ydnda ku-ti " iurli ba-ingi ", I am glad we are numerous. : Senna : Amakabuzi a-longa okaoka ku-ti'' Mba i-ende ". The shepherds said one to another: " Let us go ". Angola (X^hatelain's J^mibutidu Gr., p. 147) ; U-ainbela n'mhatu u-ixi..., He spoke to the woman, saying... Chwana : Ke-a-itse go-re tnotho, I know that you are a man. etc., etc. 1006. — N. B.\. The translator of S' Matthew's Gospel into Ganda has at every page the expression na agamba-nti... "and he spoke saying..." I do not see how this can be correct, because nti is a form of the i''"' person singular, and means " I say ", not " he said ". It seems that the connective pronoun should vary according to class and person, as nti " I say ", o-ti " thou sayest '', a-/i " he says ", /u-/i " we say ", etc. Cf. Grammaire Rtiganda, p. 2 1 . 1007. — 2. In Kafir some verbs prefer to be followed by u ku-ba or n-ba (Zulu ti-ma, 978), rather than by u ku-ti. Ex. Babum u ku-ba " u na mahashe na f ", lit. They asked to say (more literally " to stand ") : " Have you got horses? " Babuza u kuti... would also be correct. In Chwana go-re is likewise replaced after some verbs by /a (= Zulu utiia = Kafir ti-ba). Auxiliaries. 261 1008. — In Tonga, and still more in Kafir and Chwana, the verb ti (Chwana re, perfect rile) is much used as an auxiliary to introduce conjunctive clauses. Then in most cases it is praciicglly . equivalent to our conjunction " when ", or it completes some par- tible or expression which has this meaning.' Literally it means " to do so (as follows...) ". Cf. 834. Ex. : — 1009. —Tonga: Utmte miiezi a-ti u-ze, oyu u-fue, ba-cite milia {=... u-a-ti u-ze...), lit. When the next moon comes, and this is dead, they will make feasts. Na a-ka ti a-fue, 7nulUo iila pia, ba-a-mu-ten/a, When he is quite dead, the fire be- gins to flame, and they burn him. 1010. — Kafir: Yi nymnafiazi u mvundla e-ii, y-a ku-vuka, i-papateke, The hare is an animal which, when it awakes, is all nervous from fear. (Lit. which does so, when it awakes...) / mbovane zi-hlala e sidulini, a-ti u mntu u kuba «-te iv-a-hlala, zi-me zonke. Ants live in an ant-hill, so that if a man happens to sit upon it, they all come up. Lit... so that he does so, a man, if he has done so he sat, they all will stand up, N. B. In this sentence it would be equally correct to say... -zi-ti, u mntu u kuba u-ie wa-hldla..., lit. ;.. so that they do so, if a man... However, through some sort of attract- ion, Kafirs generally prefer to give to the auxiliary ti the connective pronoun of the verb of the incident clause which follows it closely rather than that of the principal verb which is more distant. The same may be noticed in the following examples ; — U mfazi w-a-landela i ndoda, zv-a-ti a ku-fika, y-a-iii ndoda u kuba " U-funa nto 7iina1 ", The woman followed the man, and when she came up to him, he said : " What do you want ? " Ba-ti, ba ku-gqiba, y-a-j!>uza i nkosi... When they had finished, the king asked... The following, on the contrary, is an exainple in which the connective pronoun of // is necessarily that of the principal verb. ' Zeni-tind-a ku-biza, ni-pume... lit. X)o ye so, when I call, do come out. 1011. — Chwana (Rev. W. Crisp's Secoana Grammar, n. 68) : Etla re (or ke-tla re) ke-tsamaea ke-go-bitse, When I go I will call you. E-a-re (or e-rile, or ke-a-re, or ke-rile) ke-fitlha kiva molacwanen^, k-a-timela, When I got to the brook, I lost my way. § 15. The Auxiliary BUYA " to Come Back ". 1012. — In Tonga I can find no evident example of the verb -bola " to come back " used as an auxiliary proper. But its Kafir equivalent -(5«j/« is often used as such under various forms, such as buya, buye, ba, be, bi. In like manner, its Chwana equivalent boeaor boa often appears under the forms ba, bo, and bile. We find the same 262 South-African Bantu Languages. auxiliary in Mpongwe under the forms y^ and ?/^. This auxiliary may be rendered practically into English in some cases by " back ", in others by " again ". Ex. : — 1013. — Kafir: W-aputna e manxitti, w-a-be w-a-ngena (or ...w-abuye w-a-ngena, or w-a-buya - w-a-ngena, 874), He went out of the water, and went in again. W-o-puma a-be a-ngene (or a-h' a-ngene), He will come out and go in again. U-V u-ye pina ? (= U-be u-ye pitta ?), Where do you come back from 7 Lit. You come back, having gone whither ? N. B. In the last example, and in the like, there is no evident sign that this auxiliary be (= buye) is essentially different from the other auxiliary be, which has been described in n. 978. Hence, though it be more probable that the latter is related to the verb -via " to stand ", it may also be correct to derive it from btiya. 101-4i, — Chw"\na (Rev. W. Crisp's Secoana Grammar, p. 38 and sqq})-. Ke-bile ke-a-reka, I buy again, lit. I have come back (and). I buy. Ke-tla ba ke-reka, I shall buy again. Ke-a-bo ke-reka, I am buying again. (Crisp renders it : "I am bu5ring as usual. ") N-tla bo ke-reka, I shall buy again. (Crisp : " I shall be buying. ") N-ka bo ke-reka, I would buy again (Crisp : " I would (or should) buy. ") 1015. — Mpongwe (Mgr Le Berre's Mpong. Gr., p. 134 and sqq.) Mi fa dyena, I see again. Mi a-fo dyenaga, I was seeing again. N-a-vo dyeni mie, and I saw again, etc., etc. § 16. The Auxiliary MANA or MALA "to Come to an End". 1016. — The verb -inana or -mala (52* and 280) is derived from ma " to stand " and the suffix -ata = ka/a " to sit, to remain ". Hence etymologically it properly means " to stop, to stand at the end ". From this are derived some idiomatic uses and meanings of it in various languages. Thus we find : — In Tonga: mane "until". Ex. l/a-iiku-tua inseke, mane zi-a-mana, he was grind- ing corn, until it came to an end. In Ganda : maze (perfect of ?nala) " already ", " finally ". Ex. Y-a-maze ku-genda, he was already gone ; oba o-maze o-n-dagira, since you have finally betrayed me. In Senna : mala " afterwards, then ". Ex. A-mala a-famba, a-famba. Then he went and went. In Kafir: mana " to continue to ". E.'c. Man 'u kii-ndi-nceda, continue to help me; u-man 'u kn-gqita..., he is passing continually... etc., etc. Auxiliaries. 263 1017. — ^- ^- I" Swahili the same idiomatic use is noticeable in the reib -isha " to finish ". Ex. Tende zi-me kw-isha liiva na ndege (= zi-me kiv-isha ku-liiva...\ The (^ates are already eaten by birds (Steere's Swah. Tales, p. 220). § ij. Various Auxiliaries. 1018. — A good number of other verbs might be mentioned' as being ofti^n used idiomatically in various languages. But this is not the pjace to dilate on thenr, because in their idiomatic use nothing i$ common to any large number of languages. It will suffice to say that many of our adverbs are rendered into Bantu by such verbs. Thus, in certain cases " soon " will be rendered into Kafir by -hlalela " to sit upon ", e. g, uhlalel'u kujika " he will soon arrive ", lit. " he sits upon arriving " ; and into Congo by -vita " to pass by " (= Swahili -pita), e. g. oyatidi wa-vita kw-iza " he will soon come " (Bentley's Gr., p. 693). In like manner some Kafirs continually use the verb -sttka " to get off" with the meaning of our " then, straightway ", or simply to express g, change of idea or determination, etc., etc. III. — JF&e Verbs " To Be " anD " To Bane ". 1019. — it was necessary in the chapter on substantives (582) to mentibn some peculiar forms of the Bantu copula. We now go on to state what remains to be said on this matter. I think that originally there was no verb in Bantu which expressed simply the act of being, and which consequently could be termed properly a copula, or substantive verb. Hence it is that in the present stage of development of these languages : — 1° In many cases in which we make use of the verb " to be " nothing of the kind is expressed in Bantu, and the predicate is joined immediately to the subject. — 2° In other cases we find in Bantu verbs or particles which corres- pond to our " to be ", but these vary according to the facts expressed, and they always include some peculiar mode of being, such as posi- tion, or situation, in addition to the fact of being. These verbs and particles are in fact no other than those which we have seen used as auxiliaries in the preceding article. Hence : — 264^ Soutji-African Bantu Languages. § I. Copula Understood. 1020. — In Kafir, Chiwana, and Karanga, the copula is gene- rally understood in absolute clauses of the present tense, except in those cases in which it is rendered by one of the particles mentioned in nn. 582-588. In nearly all the Bantu languages the most notice- able case in which the copula is understood is when it would be followed by one of the locative pronouns ko, po, mo, or their equi- valents. Ex. : — Kafir : Mninji u mbona, the maize is abundant. 0.1. 621. U ko e ndlini or u se ndlini, he is in the house. Chwana :, O moiho, thou art a man ; montle, he is good-looking. O mo tlun% he is in the house. Karanga : E-t-obe nyika i-li kiile, i pafupi, the ground is not far, it is near. (Lit. it is not a ground which is far, ii is neat). Irienyikct i mbiiya, that ground is good. SwAHiLi : Bwana yu-ko wapi^ Yu-ko koondaiL Where is the master of the house ? He is with the sheep. BoonDei : Yuko kUt'angu, he is at my house, etc., etc. 1021. — In Herero an article often acts as the copula. Ex. Owami o muhona, I am a king ; Ka ove ? Is it not thou ? § 2. The verbal Forms LI, LE, ELE, IRI, etc., used as THE Copula. 1022. ■^- The most usual form of the copula is li or ri in most Bantu languages, flf? in Lunda,/? in Rua, iri in Angola, le or cle in Chwana, la in Nyambu, etc. This is in fact no other than the form which we have found used as an auxiliary in n. 929. Considered etymologically, this form is to ila what in Tonga jivafe " seated " is to -kala " to sit down ", viz. a sort of perfect form. Hence it is that in some languages it admits in certain cases the perfect suffix -k, or an equivalent for it, as rire in Herero, zrile in Angola, /zji in Yao, etc. Nevertheless, there are difficulties as to its original meaning, because there is no such verb as -z/a in the generality of the Bantu languages. We find this verb regularly used in Angola only, and even in this language its exact fneaning is not quite plain. H6\i Chatelain in his Kimbundu (Angola) Grammar renders, it by " to do, to say ", which sheds very little light on the The Verbs " To Be ''and " To Have ". 265 matter, as it may be used only in a few given cases to render " to do " and " to say ". My opinion js that the original Bantu ^u-z/a was the applicative' form (1065) of ku-ya or M-a, "to go, to act ", and meant properly " to act towards obtaining a certain effect ". We have explained above (919) its probable connection with ku-enda " to go ". What seems to confirm thisview is that the perfect of kw-enda " to go " is given as being -e/igin Lower Congo (Bentley's Gr., p. 642). . 1023- — ^- ^- '■ In Old Angola okuila is often used to render " that is to say " .{= ku ku-/i'm Tonga, oku kii-itm Kafir, cf. 1004). 2. In Kafir the nearest equivalent to the Tonga copula It is ye, which is a perfect form of the verh'kuya "togo"(cf. 913-917). Likewise, in Herero-r«>i? and -r/ are often replaced by -e, which seems to be originally identical- with the Kafir /f. 102-4. — Whatever be the etymology and the original value of the copula It or ri, the fact is that it is treated as if it were a sort of perfect form, and consequently it is never used in future nor in imperative clauses. In Tonga, Karanga, Senna, Yao, etc., it may generally be used to render the copula in affirmative present and past clauses. In Chwana, Angola, Herero, and Swahili, its use is more limited. The use oi ye as the copula in Kafir is also limited. Ex.:— . 1025. — Tonga : . Ndi-li muumi, I am in good health ; iu-li basano, we are five ; u-li mu nganda, he is in the house. Nd-a4i munini, I was small ; ba-a-li wo, they were there ; tu-a-ka li hasano, once we were five ; ndi-ci H muumi, I am still in good health, etc. Yao : ndi-li... I am; n-a-Iiji..., I was... Senna : Muztitigu a-li ku tnusa, the master is at home. Karanga : U u-li vjuja, thou, thou art young ; t-a-ri baxano, we were five ; e-i-o-be nyika ili kule, the ground is not far. (Lit. it is not a grpund which is far.) 1026. — Herero : U-n pi? Where art thou ? U-a-ri pi? Where have you been ? Ve-ri pi? Where are they? A.-rire tyi mb-a-i or a-e tyi mb-a-i, and then I went, lit (it was that I went). TV. B. I suspect that the same e acting as the copula mqst be seen in such examples as the following : O zondti ze pi (= Kafir i gtisha zi-ye pi na ?) " Where are the sheep ? ", md\eye pif (= Kafir a mafuia «-ye pi na ?) " Where is the fat .? ", etc. Modern Angola : '' , Eye u-eri {= u-a-iri) maju kala matemu, lit. thou art (as to) the teetli like hoes, i. e. thy teeth are like hoes, or O maju ma-ku-iri kala matemu, lit. the teeth are (to) thee like hoes. Cf. H^li Chatelain's Kimbundu Gr., p. loS. 266 South- African Bantu Languages. Mpongwe : Mi-are-v'eno. I am here, etc., etc. 1027. — Languages in which the use of this copula is more limited: SwAHiLi : Relative clauses, as Kanzu zi-li-zo ndefu, shirts which are long. Chwana: Past clauses, as Ba-ne ba-U mo tlun%, they were in the house. Kafir : Pefore locative expressions, as / n-komo zi-ye pi na ? Where are the cat- - tie ? — Past clauses, as z-a-ye zinkulu, they were large. N. B. In such Kafir clauses the copula may as well be understood. § 3. The Verb KU-BA " to become, to come to be ", used as THE Copula. 1028. — In the article on auxiliaries, considering that the auxiliary da interchanges with ma, we treated them as having been originally one and the same verb. Here it matters little whether this view is correct or not. The fact is that the verb -da (-wa in^sr^tlTIi, and some other languages) is one of those most fj^eflfBently used to render our " to be ". Particular attention jjewever has to be paid -tolhis, that properly speaking the •form ba is expressive of an act which is stilJ in progress, iiot oTan act already accomplished. Hence, generally, ba will be more exactly rendered by " to becomej to come to be " than by " to be ". Hence also, as a mere consequence of this, the present " I am, thou art", etc. is not rendered by ndi-ba, 7<-(J«, etc., but by past or perfect forms, such as nd-a-ba, u-a-ba, ete., ndi-bede, u-bede, etc. The principal parts of this verb are in Tonga : ba, bi, be, bede. ,, ,, in Ganda : ba, be, badde. etc., etc. Cf. 833. Examples : — 1029. — TONGA: Nd-a-ba. Tua-ba basano, now we are five, lit. we have come to be five. Nd-a-ka ba. Bo ba-a-ka fua, ba-a-ka ba kuli Leza, ba-a-ka ba a haana baakue. Those who are dead are now with God, .they are among his children, lit... they have gone to be with God, they have gone to be with his children. Ndi-zoo-ba. Tu-zoo-ba bakazoasa baako, we shall be your priests. Ndi-be. 6''-^« ;««^£;/«, be good, lit. become good. 1 Ndi-bede. il/7, w/»fl^tf is (a poison) like arsenic. : 1030. — N. B. In most of these languages the construction of this verb presents nothing essentially different from Tonga. In Swahili, and a few other languages, the fact of its being .monosyllabic causes it in certain cases to take the prefix ku^ according to The Verbs " To Be "and " To Have ". 267 n. 853. In Xosa-Kafir the substantives which follow the verb -da require that kind of copulative prefix which has been described in n. 583. Ex. ; — SwAHiLi : Manem-ye jj/a-me ku-wa uwongo, his words are false, ' Kafir : U-ya ku-ba yi nkosi, you shall be king. § 4. The Verbs -KALA and -NNA or -IN A " to sit ", USED AS THE CoPULA. 1031. — The verb -ka/a (Old Angola -e^ala^), which means properly " to sit " (52*), hence " to remain ", is used as the copula in several languages, principally in Angola, Lower Congo, and Mozambique. But, besides the copulative notion, it always implies a determined local meaning. Cf. Heli Chatelain, Zeitschrift, 1889-90, p. 164. There' is nothing very peculiar in its forms except in Modern Angola, where ng-a'la, u-ala, tu-ala, etc. mean " I am, thou art, we are ", etc., w\\'\&~ngi-kala, ti-kala,tu-kaia, etc. mean " I shall be, thou wilt be, we shall be ", etc. (908, 975). The perfect form is kedi in Congo, kexioK kexile'xn Angola. Ex. : — OuD AnGQiA.: N^amU y-rMa mo atu. atatu (Cat., p. 8) lit. God is in three persons. Modern Angola : Khe kit tu-ala o kifuxi kie, lit. Let thy kingdom come where we are. Chat. Kiinb. Gr., p. XX. Lower Congo : E nsusu kuni ji-kalanga mo 1 How many fowls are here ? Rev. H. Bentley?s Gr., p. 691. \ Mozambique : A-kala mtu, there was a man ( = Swahili Pa-li ku-wa na mtu). Rankin's Arab Talis, p. 4. Ku-kalaviqlimu nmlubale (= Swahili A-ka-wa shekh mkuu). Ibid. • 1032. -r- In Chwana there are two verbs which mean " to sit ", viz. -du/a and -nna. The one used for the copula is -nna. Its perfect is -ntse. Ex. : — Go-tla nna senile, it will be nice. Ga kf-a nnajalo, I am not so. Ke-ntse jalo, I am so (== Tonga ndi-kede nawo, Kafir ndi-hlelinje'). IS. 1033. — - We find in Angola, Lower Congo, and Kaguru, the verb ku-ina, which probably is et'ymologically one with the Chwana go-nna. It meians " to be habitually ". Cf. 945-947. In Angola it seems to be used exclusively in its perfect form -ine or -ene. In Kaguru the form kw-ina means ' '.' because ", exactly as ku-ba in Kafir (982). Ex. : — 268 South-African Bantu Languages. Lower Co'ngo : K-ina vava or ki-na vava, it is here. Beiitley's Gr., p. 690-691. Modern Angola -.Erne ng-ene..., I am... ChaiAain's Kimbuiidu Gr., p. 107. Old Angola : Ku-iiie ringimulongat Is there anything else? Cat., p. 10. 1^. B. Hence probably the suffix -ene " self, same ", n. 825. § 5. The Verb -ENDA used a.s the Copula, 1034. — The verb -enda " to go " in Tonga, and its equivalents in other languages, are used for the copula in some instances. Ex. : — Tonga : U-enda maya, he is naked, lit. he goes naked. Kafir : U-hamba ze, do. do. § 6. Various Copulative Particles. 1035. — Looking back to the various copulative particles which have been mentioned in previous chapters, we- may now consider most of them as being more Contractions or modifications of the various forms which have just been described. Thus : — I" The Swahili copula ni, e. g. in ni Sultani, it is the Sultan (590), . is probably a modification either of the copula li (1022) or of the copula -ine {1033). The same may be said of the Tonga copula ni or n (583), e. g. in ni-ngombe, it is a cow. Ndi before pronouns in Swahili and Tonga, as in ndi-ue " it is thou ", stands probably for n-di = n-tz — m-tt, in which n or ni is the copula proper, while /z is a kind of article or classifier. Cf. 661. 2" The Senna and Karanga copulative particle ndi, as in ndi moto " it is fire ", is probably directly derived from the perfegt of -endd, and thus stands also in close relation to the copula li (1022). 3° In Kafir, ngu mntu, nga bantu, and 7tga matanga (583), pro- bably stand for ni u mntu, ni-ba bantu, ni-a matanga = li-u mntu, li-ba bantu, li-a matanga. On the contrary, in such expressions as si si-tulo " it is a chair ", etc., the copula is dropped, but its sub- jective, pronoun is retained. Hence si si-tulo — si-(li) si-tulo. 4° The Chwana copulative particle ke [ = nge, 190) might be thought to have been originally identical with the Senna ndi, Mo- zambique tAi. But this would be the only example of the phonetic change of tA or jtd to k. More probably it stands for ntse, the perfect of -nna " to sit " (1032), as we find in Chwana tse interchanging in some instances with ^e, as in -/ceua or -tsena (Tonga -njila, Kafir -kena, 52*) " to go in ". The Verbs " To Be " and " To Have ". 269 § 7.. The Copula in Negative Clauses. 1036. — In negative clauses the copula is rendered in some cases, principally when the clause is not in the present indicative tense, by the regular verbs -ba, -^kala, -enda, etc., and then it pre- sents no special difficulty. In other cases it is rendered by the negative auxiliaries which have been mentioned above (875-891), with or without other par- ticles, and then we have to notice some peculiar constrydtions. In Tonga particularly we have to notice the regular use of the auxiliary -insi together with the negative particle ta or si (976). In Mpongwe we may remark, among other constructions, the use of the form -jele (= Tonga -kede, perfect of -kala), before which the negative particle is understood. In Chwana, Swahili, Angola, Herero, etc., the negative particle by itself does duty as the negative copula. In Ganda the copula li is retained together with the negative particle. In Kafir the auxiliary si (875) is sometimes used together with the other negative auxiliary a or nge ; etc., etc. Ex, : — •Tonga : Ti insi pdimulozi, I am not a sorcerer, lit. it is not (that) I am a sor- cerer. Ei nyika ti insi mbotu, this ground is' not good. Sinsi nyika i-li kule, this ground is not far, lit. it is not a ground which is far. Ganda : Si-li-ko Me n-jogera, there is nothing for me to say, lit. it is not there Vifhat I may say. Swahili : .Si-ye,\i is not he ; si-mi, it is not I ; si-mo, it is not therein, etc. Herero: O «j?a«^a'^a-jv, this is not the village. Kafir : A si mntu or a si ye mntu, he is not a man ; a ka ko, he is not therein ; ... e-tige mntu, ... not being a man (cf. 851) ; ..." e-n'ge ko, not being there. (N. B. Notice that nga is thus changed to nge before the words which are not verbs.) Chwana ; Ga ke motho, I am not a man, lit. not I (am) a^ man. Motse ga montle, the town is not pretty, Kene ke-se molemo, I was not good. Angola : tat 'enu ki sob'e, your father is not a chief. Erne ngi 7nulu ami {ntgaiwe. particle understood, = erne \d ngi mutu ami), I am not a man. Cf. Chatelain's Kimbundu Gr., pp. 51-56. Mpongwe : Ga mie, it is not I ; ga we, it is not thou. Mi a-jele..., 1 a.m noi...; o-jele..., than axt not..., tic. , Cf.' Msr Le Berrc's' Grammaire Foiigouee, -p]). 108- 121. \ 27Q , South-African Baniu Languages. § 8. The Verb " To Have. " 1037. — It may be laid down as a general principle that in Bantu the verb " to have " is rendered by the copula followed by a preposition which means " with ", viz. a in Tonga, jv« in Congo, ni in Angola, na in Swahili, Kafir, Karanga, etc., etc. Cf. 570. The copula is sometimes understood, according to n. 1020. The prepo- sition itself is generally not understood in any language, except in Lower Congo. Ex. : -^, Tonga : Ndi-li a baana, I have children, lit. I am with children. Ganda : Miti tu-U na gio, we have trees, lit. trees we are with them. Chwana ; Xe-na le piise ( — ke-nna le piise), I have a horse. Angola : Eiu tu-ala ni tunzo, we have little houses' (at present). Etu tuene ni tunzo, we have little houses (habitually). Cf. H^li Chatelain's Kimbundu Grammar, "i^. 107. Mozambique : A-kala na mithaku minjeni, he had much property. Swahili : A na maneno makubwa, he has great words. Karanga : Ndi na tunyutii, I have little birds. Herero : U no vanatye {— u na o...), he has children. Kafjr : Ndi no mfazi ( = ndi na u...), I have a wife. Congo : Bakediyo madia mengi, {= ya madia), they had much food. Mbele zam ng-ina sau, I have my knives, lit. my knivzes I am (with) them. Cf. Bentley's Congo Gr.,-^. 691. N. B: In Kafir affirmative clauses the preposition na is generally understood when the substantive which follows it is followed itself by a determinative. Then also this substan- tive takes no article. Ex.: Yinto e zandla zimnyama, he^ is a rnan whose hands are black, lit. he is a thing which (has) hands which (are) black. This is the usual construction in Kafir for " whose, of which ". Cf. 740, 741. 1038. — As "to have " is generally rendered into Bantu by " to be with ", so " not to have " is generally rendered by " not to be with ". Tonga seems to prove an exception to this principle, since in the clauses which contain " not to have " we generally find that peculiar verb ku\ which we have already seen coupled with negative particles to render our " never " (964). Ex. : — Tonga : Ta-bakue ngubo, they have no clothes. Ei nzila ti i-kue bantu pe mu-li ei, lit. this road it has no men at all in it, i. e. there is nobody on this road. Swahili : Si 7ta chuma, I have no iron, lit. not (I am with iron). Hu na, thou hast not ; ha tu na, we have not, etc. Angola : Ki tueny-etu ( = ki tu-ene etu) ni kitari, we have no money (habitually). .^«/«-a/'«/« «/^/V««, we have no money (at present). Cf. 1037. The Verbs " To Be " and " To Have ". 271 Herero : Hi no ruveze, I have no time, lit. not I (am) with time. Ka pe no mundu, there is no rnan, lit. not there has a man. Kafir : A ka na hashe, he has no horse, lit. not he (is) with horse. Chwana : Ga ke na pitsi (not ga ke na k pilse, 1037), I have no horse. Ke-ne ke-se na sepe, I had nothing. ' Ga go na sepe, there is nothing, lit. the place has nothing. Congo ; Ke bena ( = ba-ina) ya madia ko, they have no food. Mpongwe : Mia-jele ni..., I have hot..., Cf. 1036. etc., etc. 1039. — Sometimes, in Tonga and Ganda, the verb lia " to eat " is used With the meaning of " to have, to possess ". Ex. : — Tonga : Miaka koci kede, koci lia buame, lit. (all) the years you live, eat the kingdom so long, h. e. possess the kingdom till the end of your life. N. B. In this sentence, if the verb " possess '' were rendered by " be with ", in Tonga one would say koci ba a buame (= ka u-ci ba a muami) instead of koci lia buame (= ka u-'ci lia buame). Ganda : Biiganda buno mif-bu-lie, lit. this Buganda eat it (ye), h. e. jjossess (ye) this Buganda. {Grammaire Ruganda, p. 83). '■ f ,'' • § 9. The Verbs " To Be " and " To Have " IN Locative Expressions. 1040. — We find in Bantu some quite idiomatic constructions for locative expressions when their locative particle {pa, ku, mu, etc.) is followed by a pronoun, or by a substantive without classifier, such as Leza " God ". For such expressions as " to me, from me, near God, to God, " etc. are rendered in several languages by "where I am, where God is ", etc., and in a few others by " the place A^hich has myself, the place which has God ", etc. This prin- ciple explains a large number of very puzzling expressions. In Lower Congo and some other languages it is extended to all sorts of .substantives. Ex. : — 1041. — Tonga: Uakafugama ku It Leza, he knelt down to God, lit. where is God. Ukede ku-li. uise, he lives near his father, lit. where is his father, Mii-li ei nzila, on this road, lit. wherein-is this road. N. B. In Tonga one may also hear : Mti-zoo-ba mbu-li Leza " you shall be like unto God ", lit. " like as is God ", and other similar expressions. 1042. — Modern Angola (H^li Chatelain's Kimbundu Gr., p. 113) : Ngondo kuiza ku-al'enu, I shall come to you, lit. where are you. ' .. .ku-al--eme, by, "from, to me, lit. where am I. , 272 .■ South-African Banhl Languages.- N. B. Hence, even after passive and quasi-passive verbs : Riosomke ku-aVeme, it ha been written by me. 1043. — Senna, etc. : Senna : Pida ficei pa-li sulo, when he cam© near the hare, lit. where is the har< Ganda : ^agenda e-ri lubare, they went to a doctor, lit. where is a doctor. lO^^;, — Karanga, etc.: ICarangA: Ugere pana tate (= Tonga ukede kuli uise), he lives near his father, li (at the place) which has his father. Takeja ku-no Eva, it came to Eve, lit. to (the place) which has Eve. SwAHii,! : Palikuwa na mitt..., there was a man..., lit. a place had a man... Kafir : Mkulu ku na we, he is taller than you, lit. he is tall at (the place) whic has you. 1045. — Congo: Mtt-na nso, in a house, lit. within (the place) which is the house. N. B. The original meaning of the particle na in such expressions seems to be entire! obliterated in Lower Congo. 1046. — Chwana: Ea kwa go 'inago, go to your mother, lit. go where you mother (is). Tla mo go nna, come to me, lit. come inside where I (am). .Cf. Crisp's Chwana Gr., pp. 70-71. The view which this author has taken of certai locative expressions does not seem to be altogether, correct. Thus, among other thing he has not sufficiently attended to the fact that mo implies the notion " inside ", whic kwa does not. Cf. 563. IV. — Deritiatioe If ertjg. § i. Passive Verbs. 1047. — Leaving aside Angola and Mozambique, the'genera law in Biantu for the formation of the principal parts of the verb: in the passive voice is to insert -u- or -w- before the final vowel o the active voice. Ex. in Tonga: ku-luma " to bite ", pass, ku-lumm " to be bitten ". The principal exceptions to this law are the following : — 1048. — 1° In Tonga the full element inserted is generalh -igu: Ex. ku-jatigua " to be seized ", from ku-jata " to seize. " Th< insertion of the simple -u- seems to be admitted nearly exclusivel; for the verbs which end in /a, da, or ma. 1049. — 2° In Ganda the element insei'ted is generally -ibw or -ebw-, according as the preceding syllable contains a short vowe Derivative Verbs, 273 (a, i, or u) or a long one (e, or 0). Ex. ku-sulibwa " to be cast ", from ku-sula " to cast " ; ku^emebwa" to be felled ", from ku-tema " to fell ". The insertion of the simple -w- seems to be admitted exclusively for certain verbs ending in /a or ra, and this only in certain tenses. 1050. — 3° In Boondei the element inserted is -tgw- for verbs ending in two vowels and a few others. (Woodward's Gr., p. 41.) 1051. — 4° In Yao the element inserted is -zlw- for certain verbs. (Which ?). Hetherwick's Gr., p. 40. 1052. — 5° In Kafir the element inserted is generally -zze;- for monosyllabic Verbs, and for such dissyllabic verbs as begin with a vowel. In the same language the passive form corresponding to the active bonanga is bonwanga, that corresponding to the SLCtive. bonile is boniwe, and that corresponding to the active boni is bonwa. Cf 833. 1053. — 6° In some languages, principally in Chwana and Kafir, the addition to the verbal stem of the sufiExes wa, we, or iwa, iwe, causes in certain verbs considerable phonetic changes, according to nn. 122 and 202-207. Ex. : — Kafir : u ku-tunywa u ku-hlatywa u ku-bunjwa u ku-kutshwa u ku-lityalwa (u ku-betwa) To send : u ku-tuma. To stab : u ku-hlaba. To shape : u ku-bumba, To turn out : u ku-kupa, To forget : u ku-libala, To beat : « ku-beia, etc., etc. pass. ' Chwana : go-roma, pass, go-ron^wa go-ilhajwa go-tlhaba, gp-bopa,- go-bocwa go-lebala go-beisa go-bediwa 1054. - - Exampl es : — Active Forms : bona bonanga bonaga boni bone bonide Tonga bonigua bonigui bonigue bonidue Gandajp^,. zala^ to beget zalanga. zala zale zadde -'.-{rwa zaiibwanga (?) --Irbwa zalibwe zaliddwa ^' Boondei onwa onwa onwe ^. Kaguru langwa langagwa langwe langwe langigwa Yao wonwa wonagwa wonwa (?) wonwe > Nyam-wezi wonwa wonagwa wonwe wonwe wonilwe Swahili onwa ... onwi onwe EL. Kafir bonwa bonwanga bonwa bonwe boniwe L. Congo monwa monangwa (?) monwa (?) monwe mwenwe(?) Chwana bonwa ... bonwe boncwe(?j etc., etc. iS 274 South-African Bantu Languages. 1055. — A somewhat different kind of passive verbs is obtainec by suffixing to the verbal stem -ika, or simply -ka, a suffix whicl changes regularly to -ike, ik\, ikide, etc., according to tense anc mood. This suffix is pronounced ia or ea in Mozambique, according to n. 175. In Kafir and some other languages the same suffix hai generally the form -eka, and in a few verbs the form -kala (Chwan; -fala or -hala, Angola -ala or -ana). 1056. — Properly speaking, the difference between passiv« verbs ending in ua and those ending in ka is that the former sup' pose a personal or external agent, while the latter suppose either z natural or internal agent, or that the act expressed by the verb i; done naturally.. For instance, in Tonga bonigua would be usee properly when speaking of a person who brings himself into view while bonika would better be said of a mountain or something els€ which from its very position naturally comes into view. The same distinction exists in Kafir between bonwa and bonakala. This dis- tinction does not seeni to be so well observed in some other lan- guage's. 1057. — ^- B. I. When active verbs end in •»/« or -una this sort of passive form is generally obtained by changing the final la or na into ka. Ex. in Tonga : ku-a ndtila " to break open ", pass, ku-anduka. 2. In Mozambique and Angola there seems to be no other regular way of forming pas sive verbs than the one here described. However, we may notice in Angola another pas sive ending, viz. -ama, principally for verbs which in the active voice have the ending ■eka. 1058. — Tonga : Ku-nvua " to hear ". Pass, ku-nvuika " to be heard ". Ku^amla " to speak ". Pass. Citonga cila ambtka, Tonga is easy to be spoken. Kafir : U hu-tanda "to love". Pass. U mntwana o-tandwayo ngu nina, a child which ii loved by his mother. — Umntwana o-tandekileyo, a lovely child, (a child tha is natiirally loved). U kurbona " to see ". Pass, u ku-bonwa... " to be seen by.., ", w ku-honakala " tc appear, to come into view "- , 1059. — Angola : Ku-jikula " to open ". Pass, ku-jikuka " to be opened ". Ku-mona " to see ". Pass. ,ku-»wneka " to appear ". Ku-bengeleka " to render crooked ". Pass, ku-bengalala " to get crooked ". Notio that this ending -ala causes the vowel of the preceding syllable to be change( from e to a. The ending -«««, which is only a phonetic modification of -ala (280) has tlie same effect. Ex. ku-temeneka " to provoke ", pass, ku-temanana " to ge angry ". Cf. Hdli Chatelain's Kimbundu Gr,, p. 98. Derivative Verbs. 2^^ 1060. MoZAMBlQUfi : U-abela " to cook for... ", Pass, u-abelia or u-abelea " to be cooked for... " U-ona " to see "- Pass, u-onia or u-onea " to appear ". Ganda : , Ku-labba " to see ". Pass, ku-labbika (perfect -labbise " to appear ", etc.. etc.) 1061. — Etymologies. The passive suffixes -ika, -eka, -ea, -kala, -ala, -ana, are nothing else tha,n' the verb -ekala, or -kala, " to sit " (52*)., It may thus be seen what considerable changes one and the same theme may undergo according as one or other of its conso- nants is dropped or weakened. A little short retrospect also will show what important parts the theme ekala plays in Bantu languages. We have just seen it used as a passive suffix. We had seen it a little before acting as the copula (1031), and as an auxiliary, in the various forms -ala, kana, and probably ei, ki, etc. (Cf. 941, 994, etc.). We haye also found it giving us the classifier ^'» (527), and perhaps the classifier ci (502). Finally the word -eka " self " probably belongs to the same theme. 1062. — With regard to the passive endings -_^««, -<5?^. In mosi other cases it sounds more like i, and then in Tonga the sound of the following / approaches that of r (17). Examples : — 1067. —Tonga: .- Ku-tila, to pour water. Kutilila, to pour water into... Ku-leta, to bring. Ku-letela, to bring for (some one or some purpose). Kuua, to fall. Ku-uila, to fall upon... Ku-fugama, to kneel down, Ku-fugamena, to kneel down/<7r... Uletela nziinyamal What are you bringing meat/^r? Ka mululapelela. Pray ye/o/- us, (from -lapela, pray). Ndiwe uaka ndiloela niuana. It is you who bewitched my child, lit it is you who ' bewitched to me the child. (From -loa, bewitch.) Matezi utilila paa Ceezia, The River Matezi flows into the Zambezi near Ceezia's place, lit... pours (its waters) into (the Zambezi).., (From -tila, to pour water). jV., B. The applicative form of -za "to come " is -zida " to come for ". Ex. Muazida nzi? What have you come for ? 1068. — Karanga : U-ja, to come. U-jira, \ja coxa& for . . . U-ta, to do, to make. Ultra, to make^r... U-tanga, to begin. U-tangira, to begin for... Uleba, to speak. U-lebera, to speak ^;'... U-xoba, to call. U-xobera, to caW for... 1069. — Kafir : Ku-lala, to lie down. Ku-lakla, lo lie in waity5?r... Ku-peka, to cook. Kupekela, to cook for... Kufa, to die. Kufda, to &\s.for.... Hence the passive ku-felwa, to be deadi/^/"..., i. e. to lose by death... Ex. Wafelwa ngu nina, He lost his mother, lit. he was A^dAforhy his mother. 1070. — Angola : Kusumha, to buy. Ku-sumbila, to buy^;-... Kutuina, to send. Ku-tumina, to %e.nifor... Derivative Verbs. 2^'^ Ku-banga, to do. Ku-bangela, to do for. . . Ku-soneka, to write. Ku-soiiekenaj to write ^r... 1071. — Lower Congo: ■ (Ku-)sumba, to buy. (K^u-)sumbila, to buyy^;-... (Ku-)boka, to call. (Ku)-bokela, to caller... (Ku-)noka, to rain. (Ku-)nokena, to rain (?«... 1072. — Other languages : Senna : Ku-lima, to till. Ku-limira, to till/i?^... Yao : Xu-tola, to carry. Ku-iolela, to carry y^jy. . . BooiiDii : Ku-leta, to bring. Ku-letela, to hi'mg for... Nyamwezi : Ku-enha, to bring. Ku-enhela, to bring^^. .. etc., etc. Hi — Causative verbs. 1073. — Causative verbs are properly expressive of the efficient cause that determines an act. The most common causative suffix is -isia, -isa, or -ixa, accordiiig to the different languages. In Mozam- bique it is -ika, according to n. 174. In Yao, Boondei, Congo, and Angokj it is -isa after short vowels (i, u, a), esa after long vowels (e, o,a). Ex. : — Tonga : Ku(g)ua, to fall. Ktt-guisia, to cause to fall, to bring down, Ku-nyua, to drink. Ku-nymsia,Xo force to drink. Ex. Babue bala guisia meno imbooma, The Bue knock out the teeth of boas, lit. cause to fall the teeth (to) boas. Balozui bala nyuisia viuade balozt, The Rotse force sorcerers to drink Muade (a kind of poison). Yao : Ku-ienda, to do. Ku-tendesia, to cause to do. Ku-kamula, to seize. Ku-kamulisia, to cause to seize. Congo and Angola : Ku-sumba, to buy. Ku-sumbisa, to cause to buj . Ku-zola, to love. Kri-zoksa, to cause to love. Lunda : Ku-sota, to look for... Ku-sotexa,Xo tell to look for... Kukuata, to hold. Ku-kuatexa, to help. Ku-xika, to arrive. Ku-xikixa, to cause to arrive. Boondei : Ku-hela, to cease. Ku-helesa, to cause to cease. ' Ku-gua, to fall. Ku-guisa, to cause to fall. Kaguru : Ku-gala, to bring. Ku-galisa, to cause to bring. Kafir : Ku-buya, to come back. Ku-buyisa, to bring back. Ku-anya, to suck (milk). Kuanyisa, to suckle. Chwana : Go-loma, to bite. Go-lomisa, to cause to bite. SwAHiLi : Ku-panda, to climb up. Ku-pandisha, to take up. Mozambique : U-thepa, to increase. U-thepiha, to cause to increase, etc., etc. N. B. The Nyamwez! equivalent of this suffix -i^a seems to be -ta. Ex. ku-zima " to go out ", ku-zimia " to extinguish " ; ku-oha " to suck ", ku-ohia " to suckle ". (Steere's Collections, p. 73). 278 South-African Bantu Languages. 1074. — The endings -ka (Chwana -go), and -ta (Chwana -rq), in most languages become -sia or -sa in the causative form. Ex. : — Tonga : Ku^kunka, to flow. Kukunsia, to cause to flow. ^ Ku-oluka, to fly. Ku-olusia, to take up in a flight. Ku-kuata, to marry. Ku-kuasia, to cause to marry. BooNDEi : iKu-eguta, to be satiated. Ku-egusa, to satisfy. Yao : Ku-sauka, to suffer. Ku-sausia, to punish. Kafir : I^u-goduka, to return home. Ku-godusa, to send home back.. Ku-ambata, to put on a dress. Xu-ambesa, to clothe (some one). Chwana : Go-coga, to awake. Go-cosa, to awaken. Go-apara, to put on a dress. Go-apesa, to clothe (some one). Swahili : Ku-anguka, to fall. Ku-angusa, to cause to fall. Xu-fuata, to follow. Xu-fuasa, to cause to follow, etc., etc. 1075. — The ending -la in several languages becomes -sia or za (Chwana tsd) in the causative form, as if the influence of the / softened the harder sounds -sia or -sa. Ex. : — Tonga : Xu-njila, to go in. Xu-njizia, to bring in. Xu-lila, to weep, to cry. Xu-lizia, to play (an instrument), lit. to cause ' to cry. Ganda : Xu-agala, to love. Xu-agaza, to cause to love. Nyamwezi : Xu-manila, to be accustomed. Xu-maniza, to atcustom. Swahili : Xu-tembea (= ku-iembela,' ?>?>), to walk. Xu-tembeza, to bring out for a walk. Kafir : Xu-sondela, to come near. Xu-sondeza, to bring near. Senna : Xu-lila, to cry, to spund. Xu-lidza, to cause to sound. Chwana : Go-gakaia, to be provoked. Go-gakaisa, to provoke, etc., etc. 1076. — Likewise,, in some languages the causative suffix cor- responding to -na is regularly -nya. Ex. : — Yao : Xu-songana, to come together. Xu-songanya, to gather together. ' Nyamwezi ; Xu-lina, to rise. Xu-Unya, to raise. Ganda : Xu-wona, to recover. Xu-wonya, to cure. Chwana ; Go-tlkakana, to meet. Go-tlhakanya, to bring together, etc., etc. 1077. — The suffix- ika " to set " also appears in some words as a causative suffix. It then causes various phonetic changes. Examples in Tonga : — Xu-kala, to sit. Xu-kazika, to put some one in a sitting posture. Xu-ma, to stand. Xv-bika, to set a thing standing, i. e. to place. Xu-pia, to boil, to burn. Xu-jika (= ku-pika, 52*), to cook, to boil (trans.). N. B. Ku-sika " to bury " seems to be a causative form of the non-reduplicative form of -lala " to lie down ", just as -kazika is the causative of -kala. Derivati've Verbs. ' 2^]^ 1078. — Etymologies. The suffix -ika, though active in mean- ing, probably is related to the verb -kala " to sit ", no less than the passive suffix -?'/§« (1061). ' The suffix -isia seems to be the same as the v£rb -sia " to leave, to part with, (52*) ". From this meaning is naturally derived the causative one of" imparting ". It may be noticed by the way that the causative word u-zse "his father" (748), lit. "the one who leaves him behind "„also contains the element st'a. III. — Intensive verbs. 1079. — In Tonga and a few other languages we find intensive, or quasi-superlative, verbs,, which imply that a thing is done with great attention, or well, or with persistency. In form they much resemble causative verbs ; in many instances the context alone will tell whether a verb is causative or intensive. Their regular ending is -zsia in Tonga and Yao, -z'dza in Senna, -tsa in Chwana, etc. More expressive endings are -z'szsta in Tonga, -ichisia (?) in Yao, -isidza in Senna, -isisa in Chwana, etc. Ex. : — TbNGA : Xu-amba, to say. Ku-ambisia, to say well. Ex. Uaambisia, m.uame, You have said well, sir. — Ku-ambisisia, to speak with perfection. Xht-langa, to look. Ku-tangisia, to look attentively, to compare. Ku- langisisia, to consider very carefully. Ex. Uazilangisisia inkaba, He looks at the dice, studying them very attentively. Yao : Ku-gumilisia, to cry aloud exceedingly. (From ku-gumila 1) Senna : Ku-lira, to cry. Ku-Uridza, to cry perseveringly. J^u-liristdza, to be most obstruslvci importunate. LuNDA : Ku-tala, to look. Ku-talexa, to compare. Xu-londa, to speak. Kit-lon-dexa, to explain. Chwana : Go-feta, to surpass. Go-feiisa or go-fetisisa, to be much above, etc., etc. We may couple with intensive verbs such reduplicative forms as kii-endenda, to walk about, to journey. (From ku-enda, to go, to walk.) IV, — Reversive and expansive verbs. 1080. — Reversive verbs express the undoing of what is ex- pressed by the simple, as " to tie — to untie " in English. Expan- sive verbs imply expansion, or dilatation, or ejection. Reversive and expansive verbs agree in taking identical suffixes. Their aetive ending is -ula, or, in a rediiplicated form, -ulula 28o South-African Bantu Languages. (Chwana -ola, -olold). These according to certain phonetic laws become respectively in some instances -ola or -plola, and in othei instances -una or -ununa, -ona or -onona. Their regular passive ending is -uka (Chwana -oka), according to n. 1057, or -uluka, (Chwana -olokd). "Examples : — 1081. —Tonga: Ku-lima, to dig. Ku-limula or ku-limuna, to dig a crop out. Ku-ziia {'=ku:vua), to come out. Ku-vula, to breed, to multiply. Ku-zuata, to dress, to tie the dress. Ku-zula, to undress, Ku-jala, to shut. Ku-jula, to open ; ku-juka, to be opened. Ku-fuanda, (?). Ku-fuandula, to open a spout. Kw-fuanduluka, to spout out. 1082. — Angola (Hdli Chatelain's Kimhundu Gr., pp. 101-162) : Ku-beteka, to incline. Ku-betula, to raise. Ku-bandeka, to unite. Ku-bandulula, to separate. Ku-jitika, to tie. Ku-jituna, to untie. Ku-kuta, to bind. Ku-kutununa, to unbind. Xu-spkeka, to join. Ku-sokola, to disjoin. Ku-fomeka, to sheathe". Ku-fomona, to unsheathe 1083. — Other languages : LuNDA : Ku-sala, to do. Ku-salununa, to undo. Kafir : ■ Xu-hlamba, to wash. Ku hlambulula, to wash out all diit. Ku-hlamluluki to be cleansed. Chwana: Go-bofa, to bind. Go-bofolola, to unbind. Go-huna, to tie. Go-Hunolola, to untie. J etc., etc. V. — Reciprocal verbs. 1084. — In nearly all the Bantu languages reciprocal verb are derived from the others by appending to them the suffix -ana Ex. : — Tonga : Ku-nvua, to hear. Ku-nvuana, to hear one another, to agree. Ghwana : Go-aina, to touch. Go-amana, to touch one another, Yao : Ku-suma, to trade. Ku-sumana, to trade with one another, Kafir : Kuianda, to love. Ku-iandana, to love one another. Ganda : Kiv-agala, to love. Kw-agalana, to love one another. BooNDEi ; Ku-kunda, to like. Ku-kundana, to like one another. Lower Congo ; (Ku-)tonda, to love. ( Ku- )iondana, to love one another. SwAHiLi : Ku-penda, to love. Ku-pindana, to love one another. 1085. — Conclusion, There is unmistakably an essential dil ference between the general notion implied by verbal suffixes an( Retrospect. 281 that implied by auxiliaries. But, until we have somewhat more abundant data to go by, it will be no easy task to define this differ- ence exactly. If however I am not mistaken, auxiliaries generally imply a notion of time. Respectively they imply that an action Is taking place now or took place before, lasts a long or a short time, was never done or was done once, still lasts or is already accom- plished, etc., all of them notions which come under that of difference of time. Verbal suffixes, on the contrary, are rather either relq.tional or include relation, and cannot be said to contain the notion of either time or duration. Passive verbs, for instance, suppose an agent and a patient ; applicative verbs suppose a subject and an object ; cau- sative verbs suppose an efficient cause acting upon a subordinate agent ; intensive verbs, bieing superlative, imply comparison with what . is usual and common ; expansive and reversive verbs bring back the mind to a contrary action ; reciprocal verbs suppose at least two agents acting one upon the other, all of them notions which come under the head of relation. BLetrospect on Htit)et60, Prepositions, anD Conmnctions. 1086. — The student might have expected to find here a chapter on adverbs, prepositions, and conjunctions. But the analytical method which we have followed throughout has already brought under his notice most of the particles which might have found their place in a chapter thus headed. Those which have not been men- tioned are for the most part found only in a few languages, and I do not know of any which may not be readily explained by the principles laid down in the preceding pages. To sum up all that refers to those which we have come across, the notions which we render in English by prepositions are express- ed in Bantu partly by particles, which may also be termed prepo- sitions (569-578), partly by locative classifiers (563-567), partly by verbal suffixes (i 065-1 071). Our adverbs for the most part are not rendered in Bantu by invariable particles, but partly by locative expressions (533*), partly by locative pronouns (693), partly by 282 South-African Bantu Languages, auxiliaries subject to the, same changes as other verbs (873-1018), partly by variable verbal suffixes (1079). A few conjunctions exist in Bantu, but most of them have retained something of the nature of auxiliaries (939, 940, 944, 958, 984, 985). Of the other particles which correspond to our conjunctions, part are still auxiliaries pro- per (943, 955-958, 972, 978, 982, 1008, etc.), part are relative par- ticles (784-788). Hence the student who wishes to take a correct view of any Bantu language, must, as it were, first forget all that he knew concerning the division of the parts of the speech in classical languages. Other minds and other shapes of thought entail other grammatical sys- , terns. Hirst Hppcntit^-. ETHNOGRAPHICAL NOTES IN TONGA DICTATED BY NATIVES. The following pages cannot claim to be considered as good specimens of the Tonga style in general, because my informants were not the best I qould have wished for, and still more because my slow writing under their dictation naturally made them shorten both narratives and sentences. I am, however, encouraged to give them hexe by the fact that they contain a large number of sentences in which the thought is Shaped otherwise than it would be in English, and thus weH deserve the student's attention. The italics between brackets (a, b, etc.) refer to notes at the end of this appendix. ! I. ON THE ROTSE. Malozui nga akede mu Luizi, ku- tala a Basubia. Bayanda mulilo. Baame baao m-Balumbu. Mbabb ba- nyuisia balozi muade. The sorcerers. Aba balozi mbantu baloa, ball a masaku, bazua masaku. Mbuli ci tubeleka, ci tuzuete ezi ngubo, umue muntu uakubona, uati " Nguazuata ngubo zinono oulia mu- ntu. " Ualangisia, uati " Uerede kufua ", ko kuti '' Afue oulia muntu. " Oy.u ta amunvuide uaambola nabo, uainka, liafua mu nganda iakue. Bantu ba-amuzika li bucia, baamulila. Oyu mulozi mansiku mbuli lino ua kutola mo inzule iakue. The accusation. Beenzinyina baati " Ualumua a nzi muntu uHa a afua? Caa mpoo uabona isaku caafua. " Umue muntu uati'ku umue muntu The Rotse (") are the people who live on the Zambezi, above the Su- bia ('). They are fire-worshippers. Their chiefs are Lumbu ("). It is these whp give sorcerers to drink the poi- son called mudde. These sorcerers are people who kill by charmsi They have devils, they let out devils. It is as if, (for instance), while we are working, wearing these clothes, some one had seen you and said : " That is a man who has fine clothes on. " He looks fixedly and says : "Be thou bewitched for death. " That is : " May he die yonder man ! " This (other man) did not hear him speaking thus ; he goes off, and dies in his house. People bury him in the morning and weep over him. He, the sorcerer, at night, just as now, goes to dig out his clo- thes. The parents say : " What was that man bitten by C) the day he died ? It is because he saw a devil that he died, " One man says to another : 284 First Appendix. " Ndiue uaka ndiloela muana uaka fua. " Ue uati " Pe, tinsi ndime. " Ue uati " Tuia ku baame, ku Balumbu. " Bala inka a ue ku Balumbu. T/te ordeal. Baasika, Balumbu bala mubika mu julu, a busanza. Baabika tusamo, tumue tuasimbua, tumue tuayalua etala. Kunsi a busanza baa- bika mulilo. Ue uli kede a busanza. Balumbu baati " Ue 'mulozi. " Ue uakasia uati " Pe, tinsi ndi mulozi. " Baati " Unyue musamo oyu, muade. " Muade ula tuba, ubed 'anga ni ncefo. Uabueza (?) muntu, uenyua. Ka all mubotu, ta aci fui pe, ula luka ; ka ali mulozi, muade uamukola, uaandula mutue. Ula cisa, uci zezela, uci ua. Mulilo ula pia, bala mutenda. Muntu ta anvuide mulilo, uaka fua. Ordeals with thieves. Baati ku mu- ntu umue " Uaka ba. " Uati " Pe, tinsi ndime pe, nguumbi. '' Baati " Tunjizie maanza mu manzi. " Baa- jika manzi aapia. Beense baanjiia maanza. Uasupuka lukanda mubj, bo pe, tinsi lutete luboko. The kings of the Rotse. Muame ua Balumbu ngu Liuanika. Sebitunyana nguaka sanguna. Uali kufua, kuanjila muana uakue Segeletu. Uali kufua Segeletu, ueza Sipopo a Malozui uati " Ndime Sipopo ", uanjila mu buame. Uali mubotu, uaka cita miaka njisano e inne, ua kujayigua. Muciu uakue " It is you who bewitched my child who is dead. " The other says : " No, it is not I. " The other says : " We go to the Lords, the Lumbu. " They go with him to the Lumbu. When they arrive, the Lumbu put the man up in the' air on a scaffold. They put poles, some fixed in the ground, others laid above. Under the scaffold they place fire. He (the ac- cused) is sitting on the scaffold. The Lumbu say : " You are a sorcerer. " The (man) denies em.- phatically, and says : " No, I am no sorcerer. " They say : " Drink this poison, (this) muade. " The muade is white, it looks like arsenic. The man takes the cup (?), and drinks. If he is good, he will not die of it, he will vomit ; if he is a I sorcerer, the muade contracts his face, and breaks his head : he burns with pain, totters, and falls. The fire then blazes, and they burn him. The man did not feel the fire, he was dead. They say to some one : " You have stolen (such a thing). " He says : " No, it is not I, it is some one else. " They say : " Let us put our hands into water. " They heat water until it boils. They all (the accusers and the accused) put their hands into it. The thief's skin blisters, the others (feel) nothing, their skin is not even softened. The king of the Lumbu is Liwa- nika. Sebituane (') was the first. When he died, his son Sekeletu came on .When Sekeletu died, Sipopo came with Rotse warriors, saying : " I am Sipopo ", and he came into power. He was a good man, he reigned nine years (>), and then was killed. It was Ethnographical Notes in Tonga. 285 nguaka mujaya. Uaka cija uanjila mu buato, ua kufuida mu kasua afui a munzi uakue. Uayasigua e intobolo. Pa akafua Sipopo, baainka ku Ci- lumbu, baamubuzia kabati " U^e . muame. " Baati " Ucite itatu. " Kabe baati " Miaka k'oci kede, k'bci lia . buame. " Cilumbu uati " Pe, ta ndiya- ndi buame buenu, " Mpawo kuanjila Muanaena. Uati " Ndime Muanaena. " Uaka cita muaka ngumue, baamujaya. Mua- naena nguaka yasana a Bambala, pa aka fua Sipopo. Bambala bakede kunsi a Babue, pa lutilila a Kafuefui, ku Buzungu. Bapalua meno. Mu- zungu uabo ngu Manuele. Boonse baiyasana a bukali, boonse baakafua ua kumana musili uabo. Mpawo baaka kala. Masotane nguali muame ua Beciseke. Uci li muunii. , '. Mpa aka fua Muanaena, baanjizia Liuanika. Ngoci li wo. The Mambunda. Makuango ali bantu ba Liuanika. Bakede mu talel' elino ku Mababe, ka bajaya mansui a li mu manzi. Bali a tuato tunini, bala njira mukati a manzi, ka bajaya insui a mazui (?), ka bazitola kuli Liuanika. Kuategua balala kunsi kua manzi. Ngaongao nga Mambunda. Depredations of the Rotse. Balumbu bamue bakede ku Ciseke, bamue ba- kede kutala a Basubia. Balatola ku Kangombe baana baa Balea a baana his nephew who sought his death. He (Sepopo) fled, got into a boat, a,nd went to die on an island near his city. He was shot with a gun. When Sipopo died, they went to Cilumbu (^), and asked him, saying : " Be king. " They added : " Try three years. " Again they said : " (All) the years you shall live keep the power. " Cilumbu said : " No, I do not want afriy kingship over you. " Then it was that Muanaena came in. He said : " I am Muanaena. " He reigned one year, (then) they killed him. Muanaena was (the king) who had a quarrel with the Mbala after Sipopo's death. The Mbala live below the Bue, where the Zambezi receives the Kafuefue, in the Portuguese ter- ritory. They file their teeth. Their lord is Manuel (*). They fought fu- riously on both sides, and died in great numbers, until their powder was exhausted. Then they sat dbwn. Masolane was at the head of the people of Sesheke. He is still alive. When Muanaena died.they elected Liwanika. It is he who is still there (as king). The Kuan go are subjects of Liwa- nika. They live on, this side (of the Zambezi), on the Mababe (river and flats), killing the large fish that is in the water. They have small canoes, (with which) they go into the water and kill fish with a special kind of assegai (?), taking them (then) to Liwanika. It is said that they can sleep at the bottom of the water. It is they that are called Mbunda ('). Of the Lumbu some live at Seshe- ke, others above the Subia. They take children of the Lea (^) and the Ngete (*) to the white people of the 286 First AppendiXi baa Marlgete ku bantu batuba, ka baula ntobolo, ka beza a maato ka baza ku jaya bantu. Bakalanga bala- koniba, Masukulumbue ala yasana, Batonga tabakombi ta bayasani, bala zubuka a maato, ka baza kukala mu talel'elino, ka ba,ti, a bata ci yowi Balumbu, baye kubola ku minzi iabo. Bihe, and sell them for guns ; then they come in canoes to kill people. The Karanga submit (to their ex- actions), the Shukulumbue Q fight, the Tonga neither submit nor fight, but they cross (the Zambezi) in ca- noes, and come to live on this side (the southern bank of the Middle Zambezi), returning (afterwards) to their homes, when they no longer fear the Lumbu. II. ON THE KARANGA. The Karanga 'chiefs. Bakalanga bamue bakede ku Bulumbu, ball a baame Taalimui a Nyamezi, baanza kunvua cigululu. Bamue bakede ku Bupunu. Mbavumbe aba, mbabiia baa Nguaru. Bamue bakede ku Bu- tonga.ngu Zuanga muame uabo.Oyu muame ta akue uaka komba kuli ngumue kusanguna. Monze, muame ua Batonga, uati " Ukombe kuli ndime. " Oyu Zuanga uaka.kaka, uati " Kana nkufua, ndila fua. Sikue ndila komba. " Monzje uatuma ba- lavu kuli Zuanga.. Balavu baaluma bantu baa Zuanga, uakomba. Oyu Zuanga nguise uali muame mupati ua Bunyayi boonse. Uise ua Nguaru nguamubeja ua mujaya. Wange's priests. Zuanga uli a ba- kajoaxa ('). Leza nguaka ti " Aba mbakajoaxa baako, Banerukoba, Ne- tombq, Bampire. " Bo mbapati, bali baame. Beeza kuli Zuanga, baati " Sue tuaba bakazi baako, ta tuzi kuyasana, ta tukue sumo. " Mba- Part of the Karanga nation live in the Rotse territory, they have as chiefs Taalimui and Nyamezi ; they are beginning to understand the Ko- lolo language. Others live in the Tebele territory. These are the Vum- be, they are the dogs (") of Loben- gula. Others live in the Tonga territory, they have Wange (") as chief. This chief did not submit to any one at first. Monze, a king of the Tonga, said : " Pay homage to me. " This Wange refused (to do) saying : " If it is death, I can die. Never will I submit. " Monze sent lions against Wange, the lions bit Wange's men, he submitted. This Wange's father was the para- mount chief of the whole Nyayi territory ( = Monomotapa). The fa- ther of Lobengula deceived and killed him. Wange has cacices ("). It is God who said : " These are thy cacices, the families of Nerukoba, Netombo, and Mpire. " They are old men and chiefs. They came to Wange, and said : " We have become your cacices, we will not fight, we have no spears. " i. Bakajoaxa is a Karanga word. If it were adapted to the Tonga pronunciation it would be sounded Bakatoasa. Ethnographical Notes in Tonga. 287 bonya bo bacita milia imvula iue ; bala icita kabili muaka ( = mu mua- ka ?) ngumue, imue mu mpewo, imue ejinza ; ie impewo njia kusanguna. The seasons. Umue muezi uati uze, oyu ufue, libe jinza, bacite milia, ipe kulia bantu, balime : ie jinza miezi njisano a umue. Liamana jinza, iaba mpewo ; njinne : oyu upola bantu, oyu ngua milia, ei nimpeo luzutu. Liamana mpewo, ciaba cilimo ; nji- bili. Eciamana cilimo, liaba jinza, ia ua mvula, liadilima mvula. The feasts. Mpa a milia boonse baame baa Makalanga beza kuli Zu- anga bazoolapela mvula. Baana baa- kue bala lizia ngoma, ka baziana. Zilila ziti kdi'ndili-kdindili-kdindili lingandanda-lihgandanda-linganda- nda kdi-kdi-kdi kdfndili Zuanga uasandula uazuata zimue ngubo zia muzimo zi alapela a nzio. Ula njila mu nganda ili a muzimo ia Ciloba. Oyu 'muntu mubotu uaka fua ciindi : uaka ziala banyena baaZuanga. Ngue unjira muakale Zuanga, ngue aalike a bakazi baakue. Ta tuzi ci nyamanzi ci o'cita mukati a nganda. Uazua uafugama ansi, uati guada ('), ula lapela Leza, ka ati " To kubomberera, tate bedu, su bana babo " ; ko kuti "Tulakukombelela,tuli baana baako, kootupa (2) mvula. " It is these same (people) who offer ' the feasts (sacrifices ?) to bring down rain ; they offer them twice a year, the first in winter, the second in summer; the winter ones are the first. When another moon conies and this one is dead, it will, be the rainy season C) (summer and autumn), when feasts will take place to give food to the people, and they will till the ground : the moons of the rainy season are five and one (in number). When the rainy season is over, winter comes, it lasts four (moons) : this (the first) refreshes the people, this (the second) is that of the feasts, these (the third and fourth) are only wind. Winter over, spring comes ; it lasts two (moons). Spring over, the rainy season comes (again), rain comes, the (sky) showers (copious) falls of rain. It is on feast-days that all the chiefs of the Karanga come toWange to pray for rain. His children (') ( = people) play musical instruments, and dance. The (instruments) sound like kdi'ndili-kdi'ndili-kdfndili lin- gandanda-lingandanda-lingandanda kdi-kdi-kdi kdi'ndili Wange then puts on other clothes, those of a spi- rit, in which he offers his prayers. He goes into the house which con- tains the spirit of Ciloba. This was a good man who died long ago, he begot the mother (ancestors C) ?) of Wange. Wange alone goes inside, he and his cacices. We do not know what he does inside the house. He comes out, kneels down, prostrates him- self, and prays God, saying (in Ka- ranga) : " To kubomberera, Tate bedu, 1. Guada, from -gua oneself ". a. = ka u-iu-pa. • fall ", ' on " and ida " belly ". Hence " to fall on one's belly, to prostrate 288 First Appendix. Bakajoaxa baakue mba bayasa, mbelele e isumo, imbelele ia kupaila (kupa ila?), ka baisinza, ka babika mu ndido, ka baitenda, ka bapaila, ka babanda Leza. Oku nkupaila kuabo : bala tila manzi a bukande, ka bati " Inyweno muaka fua ciindi, muaka ya kuli Leza, ka mutufuga- mena kuli ngue, ka mutukombelela, ka mutulapelela bubotu. " Mpawo balia ka bati " Tulia mubili ua Le- za. " Ta ulii koozuete (2) oyu hosi, uauzoola, uaubika ansi. Bakajoaxa luzutu baaka kala ku kupaila, abalike. Zuanga aalike ula langa. Baana baakue bala lizia ngo- ma. Bakajoaxa bala lia ei nyama, Zuanga ta ailii pe. su bana babo " (lit. " We adore Thee, our Father(i),we are Thy children); " that is to say : " We adore Thee, we are Thy children, give us rain. " They are his cacices who slaughter a sheep with an assegai for the sacri- fice (remission of sins ?). Then they skin it, put it on the fire, roast it, and offer up the sacrifice to propi- tiate God. This is their manner of offering the sacrifice : they pour water and beer (upon the roasted sheep i*) ('), saying : " You who died long ago, and who went to God, kneel down for us before him, pay homage for us, and ask happiness for us. " Then they eat saying : " We eat the body of God " ('). You do not eat with your hat on, you take it off and put it down. (All this time) the cacices have been there alone to offer the sacrifice. Wange alone is present (lit. looking). His children are playing music. The cacices eat this meat, Wange does not eat of it. 3. ON THE TONGA. How the Tonga obtain rain. Ba- tonga ta bakue milia, bala pundula. Bala inka ku Monze,ka batola mbe- lele e impongo, ka bati " Moonze ! Tuaka komba kuli ndiue, tu baana baako. Siabulongo ! Sikazimena ! Mpandayo! Muana uaLezal Miiana ua Mpande ! " Monze ualapelela baana baakue kuli Leza, imvula iaua. Monze, a favourite of the Son of God, Oyu Mpande ngue Muana ua The Tonga have no sacrifices, they are heathens. They go to their chiefs and bow down for rain. Many chiefs go to Monze ("), taking to him sheep and goats, and saying : " Moonze ! We have paid homage to you, we are your children ! Siabulongo ! Sikazi- mena ! Mpandayo ! Child of God ! Child of Mpande 1 " Monze prays to God for his children, and rain falls. This Mpande is the Son of God He lives in the air, in the rain-bow. i. Lit. "our fathers ", plural of dignity. 2, = ka uiuete. Ethnographical Notes in Tonga. 289 Leza. Ukede mu julu, mu mpini-ci- ongue. Uaka tola Monze ka aci lusa^ - bila, uaimuolusia, uamukazika mujii- lu. Kabe uamuselezia ansi ; uaua kuti po, wati " Ndila leta mvula, ndaambolana a Leza, uati ' Ka mun- dilapela kutede ; ta mucite citede, caamuima kulia, caaka cila mvula ', ko kuti ' caaikasia imvula '. Mucite nabo, zlntu zi ayanda Leza, ula mupa mvula. " Mpawo baacita, imvula iaua. God's abode, Batonga bati Leza ukede 'u manzi, mu Siongo. Munari, Miinkua, Munjilisimane, uaka ya"" kuli ngue.uanjila muakale, uaka zua. tlati " Ndime muana a' Leza, rldila njila awa ". Bo baati " Pe, t'insi ula njila, ula fua. " Ue uati " Pe, t'insi , ndila fua. " Mpawo uanjila, uayala a ,. buenga, pa akaselelela, uanjila 'u ':fi manzi, uazua. God's justice. Leza uli muziiiio, ta tumuboni. Ula nvua zintu zionse : uaamba zintu zibotu, uanyua : uaa- mba zintu zibi, uanvua. Bo baamba zintu zibotu, uya kubabika bubotu kojulu. Inzila nzibili:ei njitola bantu bacita zibi, njili a mulilo ; ei njitola bantu bacita zibotu, bayanda, njili a bubotu, njili a kukondua. Prayers to the dead. Bantu baaka fua ciindi baaka ba kuli Leza, baaka ba a baana baakue. Baame bala lapela kuli mbabo mu minzi iabb, bala la- pela ka tuehda a bubotu kuakale, ka bati " Ka mutulapelela kuli Leza, ka mutufugamena kuli ngue, asuebo He once took up Monze when still a baby, he made him fly up and re- main in the air. Afterwards he let him down. He fell with a sound like po, and said : " I bring rain ; I have spoken with God who said : ' Pray to me in such a manner ; do not do such a thing ; this has stood in the way of your food, this has made rain scarce ', that is to say : this has pre- vented rain. Do thus, (do) the things God wants, he will give you' rain. " Then they did so, and rain fell. The Tonga say that God lives in the water at Siongo (*) (= Victoria] Falls). Livingstone, a. white man, an Englishman, once went to him, he went in to the h]Ottom, and came out. Hehad sdid : " I am a child of God, I can enter therein. " The people said : " No, you cannot enter therein, you will die. " He said : " No, I shall not die. " Then he went in, he went along the bank up to where the" water rushes down, he went into the . water, and came out. God is a spirit, we do not see Him. He hears all things : if you say good things, he hears (theni) ; if you say bad things, he hears (them). To those who say good- things he will give happiness in heaven. There are two roads : this is the one which takes people who do evil, it has fire ; this is that which takes people who do good, who love; it has happiness, it has rejoicings. The people who are dead long ago have gone to God, they have been received among his children. The chiefs pray to them in their villages, they pray that we may go with hap- piness to the end, saying : " Pray ye for us before God, kneel down 19 290 First Appendix.] tuzooende nzila mbotu ili a kukon- dua. " Monze raising the dead. Monze ula busia bantu baaka fua, ingombe, im- belele. . , Uati" A muze.a muzoolange bantu beenda bee cifndi, ndizooba- busia ba ndaamba. " Uama nkolia (?) ansi, inyika iaanduka. Ino bo baakeza bantu baalanga ansi ka basondela. Bala bona bantu baaka fua bali mu kuendenda, imbelele, beense ba- nyanxa, balavu, ihyati, ingombe... Monze uati " Ka mugona mansi- ku ", uati " A ta buci, muazua anze, ka mucibuka, muazoojana baakasika inyue ka mulilede, ka bali baciabuka anze. Mujike kuliia, muzoolie a mba- bo. " Bo baazicita ezi zintu, baajika ku- iia, baabika mu ndido, boonse baati " A tuHe toonse tusonone maala ", ko kuti " Tuanjilile a amue. " Baa- buzia bqumi ka bati " Muta no zui muoyo ", ko kuti "Muta no yowi kua kufua. " The Tonga doctors. Muntu usonda ninganga. Pa aka fua muntu been- zinyina baakue baati " Tuende ku kusonda. " Baainka, beeza ku nga- nga.baabuzia baati "Tuyandal 'ube(?) anze. " Ni nganga iazua afize, iya kusondela a mbabo, iasonda, iasonda. Ka all afuefui muntu uaka loa, inganga iati " Oyu'mulozi. " latis " lue mulozi, uaka loa utede, uakede for us before him, that we also we may go (by) the good road which has happiness (^). " Monze can raise dead people,' cat- tle, sheep, (etc.). He says : " Come and look at men of former times walking, I will raise up" those I men- tion (i. e. So-and-So). " He then stri- kes on the ground' with a stick (?), and the ground opens. Then the people who have come look down, coming near the edge. They see people who were dead walking, (as well as) sheep, all sorts of animals, lions, buffaloes, cattle, (etc.). Monze says : " Sleep during the night ", and he adds : " Before day- break, if you come out when just getting up, you will find that they have come (up here) while you were sleeping, and that some are still rising up (?) outside. Do you cook some food, that you may eat with them. " The men do these things, they cook food, they put it on the fire, they all (the living and the i;isen) ■ say : " Let us eat together, and mix our nails " ; that is to say : " Let us throw them (our nails) one with ano- ther. " They (the risen) encourage the living, saying : " Do no let out your hearts " ; that is to say : " Do not fear to die. " A man who smells is (called) a nganga (') ( = doctor). When a man is dead, his parents say : " Let us go to smell. " Then they go, they come to a nganga, and ask him (out), say- ing : " We wish you to come out- side ". Then the nganga comes out and, approaching close to them, he smells and smells. If the man who has bewitched (the dead person) is Ethnographical Notes in Tonga. "ZC^X kutede. Ka ali kule, iati " Awa ta akue mulozi, muaka musia ko 'u munzi uenu. Muinke kuabede. " Ila baambila izina, iati " Ngu ndaba, u- tede. " Mpawo baainka kuabede, baya kumuita, baamunanga.baati" Ndiue mulozi, ndiwe uaka loa ndaba. " lue uakasia, uati " T'insi ndime mulozi. " Bo baati " A tuende. " lue ta akaki kuinka. Mpawo baainka a ue' ku nganga. Beeza ka lici zua izuba, ta bezi e isikati. Mpe eza i nganga iabualila nkaba nzisano a imue. Jio, jio, kua, ziaua, ziya ziti ka. Iati " A muzijate, a muzibuabile. " Boonse bala zijata, bala zibualila, inganga ia kuzifunda inkaba. lakanyua misamo iazio, ia zooba ngariga. Bamue ta baizi pe. Iati inganga " Ndiue mulozi. " Ue uti " Ndime t'insi ndi mulozi. " Iati " Uzibualile aebo. " Nguenya mulozi ula zijata, uazibualila katatu. Uazi- langisi.sia munganga, uli mu kubua- lila muntu. Ni baaka mana kubua- lila, inganga iabalemba mpemba ba t'insi balozi, mulozi iamulemba ma- sizi. Mpawo bo bala tuba nkuniu, iue ula sia ntaamu. near, the nganga says : " This is the sorcerer. " And (to him) he says : " You are the sorcerer, you have bewitched (that man) in such a manner, you were sitting in such a place. " If he be far, the nganga says : " There is no sorcerer here, you have left him there in your village. Go back to such a place. " He tells them his name, saying : " It is So-and-So, such a person. " Theri the people go to the said place, going to call him, they get hold (?) of him, and say : " You are the sorcerer, it is you who have be- witched So-and-So."The man denies strongly, saying : " It is not I (who am) the sorcerer. " The mep say : " Let us go. " He does' not refuse to go. Then they go with him to the nganga. They come when the sun is just rising, they do not come in the mid- dle of the day. When he (the sorce- rer) comes, the nganga shakes dice five and one (in number). Ji9 ! Jio ! Kua ! They fall, they disperse, they . stop. He says : " Take them your- selves and shake them. " They all take them and shake them, while the nganga studies them. He has former- ly drunk their science (lit. their trees or medicines, 378) in order to become a nganga. The other people under- stand nothing of it. The nganga says : " You are the sorcerer. " The man says : " I am no sorcerer. " The nganga says : " Do you also shake them. " Then the sorcerer also takes them, and shakes them three times. The nganga looks fixedly while the man is shaking them. And when they have finished shaking, the nganga paints in white those who are no 292 First Appendix. Tame snakes, pythons, and croco- diles. Babue mBatonga bakede ku- tala a Bambala. Ta bazuati ngubo, beenda maya. Bati, iajati'gua imboo- ma, baaipumbaila a rrfubili, baaiza- mbaila zambi zambi, mutue uazopso- ndela nabo. Bala ialila bantu, baati " Inka uka lume ndaba muntu. " Bamue baabika inzoka mu nkomo, baaituma ko kuluma bantu. Bamue, baajata intale a musamo, ta baijayi, baaibuzia, baati " Ka ijate muntu u bata muyandi. " Muntu ute- ka manzi, intale imujata. Bamue bali a nzoka anga(?) babua. Baabika nzoka mu nkomo a mulia- ngo. Uaisia uainka ku mpompo, mu- ' ntu bu eza uanjila mukati uazooba, inzoka iamusingila azoomujane mui- ni ue inganda. sorcerers ; as to the sorcerer, he paints him (with) charcoal. Then they have their forehead all white, and he, he is quite (?) black. The Bue are those Tonga who live above the Mbala. They wear no clo- thes, they go naked. When they have caught a boa, they coil it up round their body, they coil it round and round, so that its head should be near by so (as shown by a gesture). They throw it on people, saying : " Go and bite So-and-So. " Some put a snake in their wallet (°"), and send it to bite people. Others, when they have caught a crocodile by means of a charm, do not kill it, but ask him to catch a man whom they do not like. This man draws water, the crocodile catches him. Others use snakes as dogs ("). They put a snake in a .bag at the door (of their hut). They leave it and go so- mewhere : (then), if a man comes in- side to steal, the snake keeps him in until the master of the house may find him. NOTKS. [a) The Rotse. — The Rotse, or Ma-rotse, or Ba-rotse, are well known from the des- criptions of Livingstone, Holub, and Father Depelchin. According to Livingstone they call themselves Baloi, or Ba-loiana. Ba-rotse is the Chwana pronunciation of the same word. The Tonga call them Ma-lozui. It is not without interest to find them described by the Tonga as fiie-worshippars. We Icnow from ancient Arab geographers that the fire-worshippers of Siraf oa the Persian Gulf used to trade with South-Africa at least as early as the 9'" century of the Christian Era, and we still find the Parsees all over the east coast, principally at Mozambique. Putting these facts together, I am inclined to think that Parsee traders or slave dealers, starting at an unknown time from the East Coast, have pushed their way as far as the Upper Zambezi, and grouped together those blacks who now form the Rotse nation. I should not even be astonished if the word Ba-rotse were merely a phonetic adaptation of the word Parsee to Chwana pronunciation. {b) The Subia. — The Subia are a Tonga tribe that used to be found between the Victoria and the Gonye Falls. Incorporated into the Kololo Empire about the year 1840, they have naturally become the subjects of the Rotse ever since these destroyed the Kololo. But ill-treated, and continually robbed of their childrea by their new masters in Ethnographical Notes in Tonga. 293 their old homes, they began to seek new ones. They are now found in great numbers, mixed up with other tribes, between take Ngami and the Zambezi, principally on the Mababe River. id) Their chiefs are Lumbtt. — Whenever I meet in Tonga that Bantu sound which is intermediary between / and r, I adopt the /. Otherwise the word Lumbu might as well be spelt Rumbu. The word Balumbu, ox Ba-rumb«, seems to mean " white people ",' or iriore exactly " yellow people ". Hence, if it be correct to say that the Rolse nation has been formed by Parsees from the East, the modern Lumbu mentioned in these notes are probably no other than their descendants. The Ba-lvmbu of my Tonga informants are probably the same as the white A Ba-lamba repeatedly alluded to by the traveller Anderson in his " Twenty-five Years in a Waggon " (Vol. I, p. 247 ; vol. II, p. 200, etc.). (d) What was the tnan bitten 'by? — On the Zambezi whoever dies young, unless killed in battle, is by the natives supposed to have been bewitched or poisoned, as they cannot imagine that a man may die a natural death before he has reached a goo4 old age. This execrable notion dooms to death every year hundreds of imaginary sorcerers. A sorcerer is called mu-iozi in Tonga, un-doi in Karanga, mo-roi m. Chwana, u m-tagati in Tebele, 11 m-takaii in Xosa, un-Jiri in Senna, etc. Xe) SebitUane. — As is well known from Livingstone's Travels, this truly great man was the founder of the Kololo Empire, ^e died in 1851. My informants knew no distinction between the Kololo and the Rotse Empire. (/) He reigned nine years. — Sipopo, alias Sipopa, was not a Kololo, but a Rotse. A short time after the. death of Sekeletu, which occurred in 1864, he came down upon the Kololo, destroyed them all, and reigned paramount on the Upper Zambezi. ( g) Cilumbu. — I do not know who this Cilumbu is who has so much influence among the Rotse, but I suspect that he is a black from the Bihe. {K) Manuel. — This must be Manuel Antonio de Souza, capitao mor, formerly of Zumbo, now of Gorongoza. In the Portuguese East-African possessions, the chiefs are called Ba-zungu, which, whatever its etymology may be, is a synonym of Baptized Christians, baptism being considered as the marlc of a chief, or child of God. The name oi Ba-mbala, Or Ba-mbara, which is given by the Tonga to the subjects of the Bii-zufigu, must probably be identified with Amhara,vih\ch in Abyssinia is a synonym of Christian. {i) The Mbunda. — As.has been mentioned in a previous note (p. 30), the word Mbtinda is applied to many different tribes. This word properly means " people of the back ", i. e. " the West " (See Introduction, I). The word Kwango has been misspelt Kwengo at pp. 30, 3r, and 10-14, of this work, as I now find that th.e Ma-kwengo of my informants are different from their Ma-kwango, and probably are not even a Bantu tribe. (/) The Lea. — The Lea are a Tonga tribe dwelling round the Victoria Falls. They have submitted to the Rotse. One of my informants was a Lea. {k) The Ngete. — The Ngete, also known as Nkete, Nketa, Kheta, Khete, Nqeii, whence, with the classifiers MA- and BA-, Ma-nketa, Ma-ngete, Ba-n^eti, etc., are a very indus- trious tribe inhabiting the Rotse Valley from the Gonye Falls to near the confluence of the Nyengo River with the Zambezi. They are particularly remarkable for their works in iron and wood. If I may believe my native informants, their language differs less from , Rotse than from Tonga. In all probability they are related as a tribe to the no less industrious Ba-kete of the Lu-lua Valley, whose beautiful plantations have been described by Bateman in the " First ascent of the Kasai ". (l) The Shukulumbue. — This tribe is located on the Upper Kafuefue River. They were described by my Zambezi informants as being very fierce. They will allow no white man to visit theircountry. Dr. Holub, the only European who ever reached it, was robbed by them of all his effects, and forced to retrace his steps southwards. («) They are the dogs of Lobengula. -- Wherevef Mohammedan customs have pene- trated in South Africa, the native chiefs divide their subjects into " children " and " dogs "- 294 First Appendix. As a consequence of their being meie " dogs ".those Karanga who have accepted Loben- gula's rule, are not allowed to possess cattle. Fine herds of these may well be seen under their care, but they all belong to the king. («) Wange. — This chief, also called Wankie, was repeatedly said by my informants to be the legitimate representative of the house that ruled for centuries over the whole Bu-nyai, or the Empire of the Monomotapa. I cannot conciliate, this with the claims to the same honour of the chief Catoloza, or C^taloze, who in Livingstone's time had his residence at some distance to the west of Tette, unless these opposite claims be the result of an ancient scission of the Karanga nation, which has not been recorded by history. Wange's chief town is situated at the southernmost point of the Upper Zambezi. He is said to be a very good man. But, pressed on one side by the Rotse, on another by the Tebele, and on another by the Tonga, whose territory he has invaded, he has none of the power of his forefathers. {o) Wangahas cacices. — When, on the first day of January 1561, the venerable Father Gongalo da Sylveira, S. J., reached the court Of the Monomotapa, Onde Gongalo morte e vituperio Padecera pela Fd sancta sua, {Lusiads, X, 93), he found the place already occupied by Mohammedan emissaries, called caches, the very men who by dint of calumnies soon caused him to be put to death by the so-called Em- peror. This readily explains why the customs of the Karanga, who in those times were the ruling tribe in those parts, are mostly borrowed from the Mohammedans. For, though the emperor, repenting of having sacrificed Father Sylveira to the hatred of the Moham- medans, is said to have driven them out of his Empire in the year 1569, and to have then sincerely desired to live as a Christian, nevertheless, from want of Christian teachers he retained most of his Mohammedan practices. (^) When thif moon is dead, it will be the rainy season. — This was written on Septem- ber 3, 1884, the 13* day of the moon. Therefore, as the Karanga year begins with winter, it must be said to commence in March or the beginning of April. {q) His children — Wange, being a good chief, calls all his subjects " his children ". {/) He begot the mother {ancestors ?) of Wange. — I do not know whether ba-nyena, lit. " mothers '', is here a plural of respect (cf. n. 343), or a real plural. If it be a plural of respect, Ciloba must be said to have been the grandfather of Wange. (j) Upon the roasted sheep (f) — It may be that they pour it simply on the ground. Old Kafirs used to make such libations round the enclosure in which the sacrifices took place. (t) We eat the body of God. — This remnant of Father Gongalo da Sylveira's short stay at the court of the Monomotapa is a good specimen of the religious eclectism of the Ka- ranga. I also find that ever since the days of this glorious Martyr, the kings of those parts were never recognized as such until they had received something like baptism. (DerNeiic Welt-Bot, 1748, n. 555, p. 106). {ft) Mome. — This chief went to meet Livingstone on his first journey from Sesheke to the East Coast. After having saluted the great traveller according to the Tonga fashion by throwing himself on his back and roUing from side to side, he made him several pre- sents, and passed a whole day in his company. Livingstone thought him to be as good- natured a man as could be. {Missionary Travels, pp. 552-555). His sacred animal is the buffalo, as that of the old Karanga kings was the hippopotamus [n. 461 (10)]. {v) Child of God ! — Lest more importance should be attached to this expression than it has in reality, it may be remarked that it is here a mere compliment, or " name ", as Kafirs say, just as the other expressions Sikazimena, Mpandayo, etc., the meaning of which is not clear to me. Chiefs are very generally termed Children of God, as are Christians in general, and whoever is considered lobe of white, or the divine, race. It happened to me once, after having given a loaf of bread to a poor old Kafir woman, to Ethnographical Notes in Tonga. 295 hear her burst into the following expressions of thanks : Nkosi! Dade ! Mia ka Tixo ! 'Mia kct Rultimente ! Solotomana ! that is : " Lord ! Father ! Child of God ! Child of the Government ! Solotomana ! " The last expression was considered by Kafirs as my proper name. {x) God lives at Stonga. — " At three spots near these falls ", says Livingstone, " three " Ba-toka (= Ba-tonga) chiefs offered up prayers and sacrifices to the Ba-rimo (= Tonga " Mi-zimo). They chose their places of prayer within the soun^ of the roar of the cata- " ract, and in sight of the bright bows in the clouds... The p!ay of colours of the double " iris on the cloud, seen by them elsewhere only as the rainbow, may have led them to " the idea that this was the abode of the Deity. " {Missionary Travels^ London, ] 857, p. 533.)' (y) The road wliich has happiness. — These to all appearances are prayers to ask for material, not eternal, happiness. {z) A man who smells is a nganga. — The Bantu practice of smelling described in this passage (Tonga ku-sonda, Kafir ku-ni(ka) exists in the larger number of the Bantu tribes. In the hands of the chiefs it is the most powerful arm for getting rid of the men who are in their way. (aa) In their wallet.. -^ No Kafir ever goes about without his httle bag or wallet made out of the skin of some little animal. He puts together in it tobacco, pipe, knife, small tools, and in general whatever he can pick up for his use. One of the worst kinds of un- , politeness is considered to he that of asking a man what he has in his bag. (M) Oth'ers use makes as dogs. — This singular custom of using snakes as dogs has its counterpart in the use of snakes as cats among the Kafirs of Gazaland. We read in Father Depelchin's " Trois ans dans VAfrique Ausiraje ', p. 71, that in the hut in which Father Law died, " there lived two snakes, the one a cobra three feet long, thick as " an arm, the other smaller, which used to fulfil the duties of our eats in Europe by keep- " ing at a distance the mice and rats which would make their appeara!nce at every " corner ". : i:**';,- , Heconti HppenDijCi SPECIMENS OF KAFIR FOLK-LORE. Kafirs are in possession of a large number of traditional tales in which the heroes ?ire not animals, but human beings. No such tales seem to be known by the other Bantu tribes. Neither do I find anything like them in any version of Pilpay's Fables. One of the most remarkable features of most of them is that they contain parts that are siing. It might even be thought that in several of them the story is merely' the frame of the song. N. B. I. \The division of the short melodies that occur in these tales into intermixed bars of 3, 2, or 4, beats each; is not intended to express a rigorous rhythm as in European music, but mel-ely to set otif those notes which bear the musical accent. Hence, though the relative value of the notes must be kept at least approximately in rendering these tunes, what is more important is that the first beat of each bar be accented. 2. The italics between brackets {a, b, c, etc.)' refer to notes at the end of each tale. First Tale. INTAKA: ENYA A MASI. Wati u mfo, ngo'mnye u mhla, wati e mfazini, ma kaye e masimini, i alime. Waya ke, wafika, walima, wa- goduka. Yafika i ntaka ku la ndawo ayilimileyo, yati : — • Allegt'Hto. THE BIRD THAT MADE MILK ("). Once upon a time a man told his wife to go to hoe in the gardens ('). So she went, she arrived, she hoed, and came home back. Then a bird went to the place which held been hoed, and sang : i e£ Ui^ Tya - ni ba le ntsi Tya ni ba le nta " Tyani ba le ntsimi, cididi ! Tyani ba le ntaka, cididi ! " Bapuma u tyani, kwa ngati be kungalinyvyanga. Yafika i ndoda yati : " Ulime pi ? " Wati u mfazi : " Ndilime apa. " Yati i ndoda :" Uya- xoka, a kulimanga. " Yatsho, ya se imbeta ngo mpini. Walila. Yambiza i ndoda yati : " Yiza mi, ka, ci ci di di di! di! " Grass of this garden, shoot up. Grass of this bird, shoot up. " And the grass came up : it was as if no spot had been hoed. The hus- band came and said : " Where did you hoe? " The woman said : " I hoed here ". The husband said : " You He, you did not hoe ". So he said, and then he struck her with the handle. And she cried. Her specimens of Kafir Folk-Lore. 297 silime. " Waya ke, balima, balima, bagoduka. Yafika i ntaka, yati : — " Tyani ba le nlsimi, cididi ! Tyani ba le titaka, cididi ! " Betu, kwa ngati be kungalinywa- nga. Bati ke baya kusasa, a bayibona i ndima. Wati u mfazi : " I pina ke i ndima? " Yati i ndoda:" O ndibdnile, mfazi, ub' unyanisile ; uz' undimbele ke uvelise i sandla sodwa. " Wayenza ke lo nto u mfazi, wagoduka. Yafika i ntaka yati citi citi, yanya- tela e_ sandleni se ndoda, yayibamba. Yati i ntaka : " Ndiyeke, ndi yi ntakana enya a masi." Yati i ndoda: " Ka wenze ke, ntak'am, ndibone. " Yati pudlu i ngqaka e sandleni. Yagoduka nayo, yafika, yati ku m- fazi ma kahlapibe u mpanda ayifa-, ke kuwo. Wayifaka ke u mfazi. Wati akugqiba u kuwuhlamba yazalisa u mpanda nga masi. Bavuya kakulu, kuba ba belamba, bafumana u ku- hluta. Baya kulima, bashiya a bantwana e kaya. Aba bantwana a magama abo o mkulu waye ngu Ngencu, o mncinane waye ngu Notuncu. Wati u Ngencu : " Ma siye kwa bantwana, sibaxelele le ntaka." Wati u Notuncu ; " UbawD ub 'ete ze singa baxeleli, uya kusibulala. " Wati u Ngencu : " Hlal' uti tu, ntwanandini inolwini. " Wayeka u Notuncu, kuba uyoyiswa. husband then called her and said : " Come, let us hoe. " So she went ; they hoed and hoed, and then went back home. The bird came then, and, sang : " Grass of this garden, shoot up. Grass of this bird, shoot up. " Dear me ! it was as if no spot had been hoed. So, ^hen they came in the morn- ing, they saw ; no place hoed. The woman said : " Where is the work done (yesterday) .' " The husband said : " Oh ! I see how it is, my wife : bufy me then in the ground, so as to leave the hand alone out. " The wo- man did so, and went back home. The bird cartie, and picked here and there, till it trod upon the man's hand, and he got hold of it., The bird said : " Leave me, I am a bird that makes milk. " Thd man said : " Make some then, tny bird, that I may see. " So it made thick milk on his hand. He went home with it, and when he arrived he told his wife to wash a milkpail and to put it into it. So the woman put it there, and when she had finished washing the milk- pail, the bird filled it with milk. And they rejoiced greatly, because they were hungry and they had found plenty. They went to work in the field, and left the children at home. The names of these children were Nge- ncu for the elder, and Notuncu for the younger. Ngencu said : " Let us go to other' children, to tell them of this bird. " Notuncu said : " Our father told us not to mention it to them, otherwise he would kill us. " ' Ngencu said : " Hold your tongue, 298 ■ Secona Appendix. Waya kubaxelela. Wati ke, akubaxelela, bati : " Ma siye. " Baya kufika, bayirola e mpan- deni. Wavakala u Ngencu esiti : " Ka wukangele i ntaka ya ko kvvetu. " Yati i ntaka : " U kuba ndi yi ntaka ya ko kwenu, hamba liyo kundibeka e buhlanti. " Wayitata waya kuyibe- ka e buhlanti., Yafika yati e buhlanti, ma kayibeke e lusaseni, wayibeka. Yesuka yapapazela yemka. Wavakala u Notuncu esiti : " Na- ntso i nto e nda ndiyixelela, ndisiti siya kubetwa. Uya yibona na ke imka nje ? " Basuka babaleka aba bantwana be bezo kuyiboniswa, be- mka. Yavakala i ntaka ihamba esiti : " Ndiyekwengu Ngencu no Notun- cu. " Yatsho yada ya malunga ngo^ yise lowo. Wavakala u mfazi : " Na- ntso i ntaka yako isiti " iyekwe ngu Ngencu no Notuncu. " Yati i ndo- da : "iMs'u kuyinyebelela i ntak'am. A bantwana bam bangati, ni u kuba ndibayala kangaka, kanti ba kwenza i nto embi kangaka ? " Bagoduke bafike ekaya. U mfazi akangele e mpandeni, afike ingeko o kunene. I ndoda i sel' ibiza a bantw- ana : " Ngencu no Notuncu ! ", ba- sabele. Iti ; " Yizani apa. " Baye. Iti bakufika, ibuze i ntaka. Ati u No- tuncu : " lb' ikutshwe ngu Ngencu. " Ati ke u yise, akutsho u Notuncu, arole i ntambo, ati " uya kubabula- la. " Bakale a bantwana. Avakale u mfazi esiti : " Yinina, Songencu, u- ngade ubulale a bantwana nga masi?" you lying little creature. " So Notun- cu yielded, as she was frightened. And he went to tell them. . So when he had told them, they said : " Let us go. " When they came, they took it out of the milkpail. Ngencu shouted out, saying : " Look at the bird of our place. " The bird said :" If I am a bird of your place, gb and put me in the kraal ".^He took it, and went to put it in the kraal. When in the kraal, it said he should put it on the fence, and he put it there. Straightway the bird took to flight, and went off. ^ Notuncu then cried out, saying : " There is just what I told you, when I said we should be beaten. Do you see it now going offthus? " Straight- way the children who had come to see it began to run, and went off". The bird was heard saying while going : " I have been let off by Ngencu and Notuncu. " It kept say^ ing so till it passed near that father of theirs. The woman cried out : " There is your bird saying it has been let off'by Ngencu andNotuncu." The husband said : " Don't you speak ill of my bird. How could my children have received from me so strict in- structions and yet do so bad a thing.'" Then they go back and arrive. The wife looks in the milkpail, and finds no bird in it certainly. The hus- band then calls out for the children : " Ngencu and Notuncu ! " ; they answer. He says : " Come here you. " They go, and when they come he inquires for the bird. Notuncu says : " It has been let off" by Ngencu. " The father, when Notuncu has' said this, draws a rope, and says he is going to kill them. The children cry. specimens of Kafir Folk-Lore. 299 Ivakale isiti i ndoda : " Nda kukubu- lala wena ke, u kuba utsho. " Ayeke u mfazi, alile. Ifake i ntambo, iyo kubaxoma e mlanjeni e mtini o pezu kwe siziba. Emke, ibaxome. Iti i ntambo iqauke. Bawe e sizibeni apo batshone kona, be nga bantu bo mla- mbo. Bakwazi u kuzalisa. Kwati, nge linye i xesha, kwafika i lizwe, baya kuwela a bafazi. Bawu- zalisa. Bavakala a bafazi besiti : " Vulela, Ngencu no-Notuncu. " Ba- bavulele, a bafazi bawele. Bati ba kuwela bawuzalisa. Afika a madoda, bawuzalisa. Ava- kala esiti : " Vulela, Ngencu no No- tuncu. " Apela ke ' a manzi, angena ke a madoda. Ati, akubona ukuba a pakati, wafika uyise Iowa way'eba- bulele. Bawuzalisa. Avakala a manye a madoda : " Puma, mfondini, wa ubulela ntonina wena a bantwana ? " Wapuma wauta ke u mlambo. Awela ke lo madoda ; wasala yedwa lo nlntu way'ebabulele a bantwana bake. Yada yabonakala i vela i mpi. Wavakala esiti : " Vulela, Ngencu no Notuncu. " Bati : " Oko wa usjbula- la ! " Wavakala ekala,. yafika i mpi, yambulala, wafa ke kwapela. The woman cries out, saying : " What is that, father of Ngencu.? Would you go so far as to kill children for milk ? " The man bursts forth, say- ing : " Then I shall kill you yourself, if you speak thus. " The woman in- sists no more, and sheds tears. The man ties (the children) with the rope, intending to go and hang them up near the river on a tree that is over a pool. He goes and hangs them up. But the rope breaks, and the children fall into the pool. There they disap- pear, they are turned into river-men, with power to produce floods. ' Then, at one time, there happened to be an invasion of the enemy ; the women went to cross the river, but the. rivermen filled it up. The women then cried out, saying : " Let us pass, Ngencu and Notuncu. " And they opened a way through, and the wo- men crossed over the river. When these had crossed, they filled up the river again. The men came also, then the riv- ermen filled the river. The men cried out, saying : " Let us pass, Ngencu and Notuncu. " So the water disap- peared, afid the men went in. But, when they wiere half-way, the father who had killed them arrived. They filled the river again. Then the other men shouted out : " Get out, you man, why did you kill your children ? " He went out, and the river dried up. Those men then crossed the river, and he remained alone, the man who had killed his children. At last the invading army was seen to appear. The man raised his voice, saying : " Open for me, Ngencu and Notuncu. " They said : " Why ! You who killed us ! " He burst out shout- 300 Second Appendix. Kwaba njalo u ' kufa kwa lo mfo wabulala a bantwana bake nge. nxa ya masi. Bati ke bona, bapuma e manzini, bafuna u nina. Eamfumana, bahlala naye, ba se bcsiya ngo ku- hamfca e mlanjenj. Ndiya pela apo. ing. The enemy came, slew him, and he died ; that was the end of him. Such was the death of that man who had killed his children for the sake of fflilk. As to them, they came out of the river, and went to look for their mother. They found her, and remained with her, but kept the power of going into the river, I stop there ('). NOTES. (a) Twoqther versions of this tale have been published by Geo. M'= Call Theal in his delicious little work, entitled " Kafir Folk-Lore ". Both of them want the interesting con- clusion of the one here given, but they complete it in some other parts. {b) A man told his wife to go to hoe in ike gardens. — Among the XosaTKafirs the work was formerly so divided that men had the care of the cattle, and women that of the gar- dens. The introduction of the plough has naturally thrown upon the men part of the gar- den-work. (c) I wonder whether this tale has not its parallel in Stanley's Legend of the Tanganyi- ka {Dark Continent, ch. XIX). In both we first see gardens cultivated by a man and a woman ; then a marvellous supply of food, heaven-sent fish on the Tanganyika, heaven-sent milk among the Kafirs ; then the precious secret betrayed to a visitor, in the one case by the woman, in the other by the children of the house ; then punishments by the loss of the treasure and further calamities, a flood on the Tanganyika, a flood and war together among the Kafirs. SeconD Tale, U MLONJALONJANI NO DADE WABO NE MBULU; Kwati ke kaloku kwako u Mlon- jalonjani e ne singqi. Wati ke u dade wabo : " Uhlell nje, u ne singqi na ?" Wati : " Yiza, ndokuqaqe lonto. " Wati yena : " Hayi, nda kufa. " Wati : " Hayi, mnta ka mama, uya kuti nina, uza kwaluka nje .' " Wati ke : " Ewe, ndiqaqe. " Wati ke qaqa qaqa nge zembe, Wati yena : " Shushushu ! ndafa. MLONJALONJANI, HIS SISTER, AND A MBULU ("). Once upon a time there was (a boy called) Mlonjalonjani, who was hunch-backed. His sister said to him : " Such as you are, are you really hunch-backed ? " She added : " Come that I cut that hump off you. " He said : " No, I should die. " She said : " No, child of my mother. What will you doj as you are going to be cir- cumcised ? " He said : " Well, cut it off". So she cut, and cut, with an axe. He said : " Oh dear 1 Oh dear ! I am specimens of Kafir Folk-Lore. 301 mnta ka bawo. " Wati ke : " Yima, se yiza kumka. "Wati ke qaqa qaqa. Wati : " Shushushu, ndafa. " Wati ke : " Se yiza kugqitywa, se yiza kumka. " Wati qaqaqa. Yawa ke. Wati ke, ya kuwa, wasuka wafa." Wabaleka ke u dade wabo, waya kuxela ku yise no nina u kuba u Mlonjalonjani ufile. Beza ke u yise no nina, beza belila. B'afika batshisa ke i ndlu, bazifaka e ndlini, bazitshisa nayo Zati ke i ntombazana zemka zilila, zaquba i nkomo za ko wazo, zahamba ke zaya ku lo nina. Wasuka u mhlaba wahlangana, kwasuka kwa mnyama. Batike: — Andantino. dying, cliild of my father. " Siie said: " Patience i lit is nearly, off. " So she . cut again. He said : " Oh dear ! Oh dear ! I am dying. " So she said : " It is nearly finished, it is nearly off. " She cut again and the hump fell d6wn. But when it fell down, he died. Then his sister ran, and went to tell her father and mother that Mlon- jalonjani was dead. So the father and the mother came shedding tears. When they reached their hut, they set fire to it, shut themselves in it, and burnt themselves with it. So the girls went away crying. They drove before them the cattle of the place, and went in search of their mother. Suddenly the earth was covered with a thick fog, and it got dark. So they sang : S^ESE li =^- Qa - bu - ka m-ga- da, mba-nga-mba-nga! Si - fe - le ma we-tu, i -A- i=jiE=5=i^r£S 3= (tir). W^ — ?- ihba-nga-mba-nga ! -^ — 5>- U - zi -tshi - Si-.bu-le- le Si - m-qa se Mlo qa " Qabuka, mgada ('), mbangambanga ! , Sifele If) ma (3) wetu, mbangambanga ! Uzitshise ne ndlu yake, mbangambanga ! Sibulele Mlonjalonjani, mbangambanga ! Simqaqa singqi sake, mbangambanga ! " Wasuka u mhlaba waqabuka. Bahamba ke, bahamba, bahamba, bahamba, bava kusiti roqo roqo roqo ne ndlu ya-ke, mba-nga-mba-nga ! •nja - lo-nja-ni, mba-nga-mba-nga! si-ngqi sa-ke, mba-nga-mba-nga ! ", Open out, earth, alas ) alas ! We have lost our mother, alas ! alas ! She has burnt herself with her hut,alas!alas! We had killed Mlonjalonjani, alas ! alas ! By cutting off his hump, alas 1 alas ! " Then the earth opened ouf. So they went and went ; they went and went, until they heard a sound 1. Mgada is a word used only by women for m-hlaia. 2. Regularly we should have felwe, not -fele; but; as I never could perceive the w, I have thought it better not to insert it. Possibly also si-fele is for u-sifele, lit. " she is dead for us. " 3. Ma, poetical f6r ngu ma, if si-fele stands for si-felwei for « Ma, if si-fele stands for u-sirfele. 302 'Second Appendix. roqo pantsr kwe litye e sidulini. Ya puma ke le nto yati.: " Nifuna nto nina ? " Bati bona : — " Sifele ma wetu, mbangambanga ! etc. {as above). ■" Yi mbulu lo nto. Yati : " H^mbani ndinikape, ndinise ku lo nydko. " Bahamba ke. Yati yakufika e zibu- kweni e likulu, yati : " Na kuhlamba, u kuba nowavile (') (a manii). " Ba cancata ke e matyeni, bacancata. Yasuka i mbulu yati ngcu ngo msila, yati ke tshizi. Yati ; " Hlambani ke, niwayile nje. " Bahlamba ke, watata i mpahla zabo,- wazingxiba zona. Bati ke : " Zis'i mpahla zetu. " Wa'ti : " O! ka nihambe, nina mbuka wa nina ? " Ba hamba ke, bafika ke nga ku lo mzi. Bati ke : " Yis'i mpahla zetu. " Wati ke : " Ni na mbuka wa nina ? " Basika ke baziqab' u daka. Baha- mba ke. Bafika ke ku lo mzi. Yati ke le nto, le mbulu i no msila, yati : " Yipani o mgodwanja (2) u kutya. " Bapiwa ke. Kwatiwa : " Ma bayo ku linda a masimi atyiwa zi ntaka. " Bahamba ke kusasa, baya ku linda. Lati i xego : " Tsayitsayibom ! Nanzo, mgodwanga. " Zati i ntomba- like roqo, roqo, roqo, coming from under a stone in a hill. So that thing came out, and said : '• What are you looking for ? " They sang : " We have lost our mother, alas ! alas ! etc. (as above) ". That thing was a mbulu. It said : " Go on, I will lead you the (right) way, and bring you to that mother of ' yours. '' So they went on. When the mbulu came to a great ford, it said : " If you are touched b.y water, you must go in and bathe. " So they walk- ed on tottering and tottering from stone to stone. Suddenly the mbulii struck the water with its tail, and splashed it. Then it said : " Go in^^and bathe, since you have been touched by water. " So they went in. Then the mbulu took their clothes and put them on himself. They said : ." Let us have our clothes. " It answered : "Just go on. What can you complain of? " So .they went on. When they came near that village, they said : " Let us have our clothes. " It said : " What can you complain of .' " Then they smeared their body with clay, and they went on. They reached that village. Then th^t thing, that mbulu with a tail, said : " Give food to these offsprings of dogs. " They received food. Then they were told to go and watch the gardens that were being eaten by birds. So they went to watch in the , morning. An old man said : " Tsayitsayi- bom (') ! There they (the birds) are I. JVi>wavile=ai-wavile. The change of » to is the result of a partial assimilation with the following «i. .!. U tHgodwaHja, pi. mgodwAnja, is a compound word derived from « m-gado ' ' breed " and i nja ' ' dog. " specimens of Kafir Folk-Lore. 303 zana : " Tsayitsayibom ! Nanzo, Ma- belengambonge (») : — " Sifele ma wetu, ijibangambanga ! etc. ". (the same as before). Lati i xego : " He ! " Bagoduka ke baya e kaya ngo kuhrwa. Alaxela ela xego. Yona ke i mbulu yahlala e kaya. Kiwabuzwa i ndaba, yati " Kjjsapi- liwe," benga boni ingesiyo ntombaza- na ke, iyi mbulu. Yapuma ne nkosi ke, yaya kulala e ndlini yayo. Yati i ne sisu, yati : " Ncincinu, ndifun'i qwili (°). " Yafika ke i mpuku. La lise ko i xego ke, lati : " Yi mbulu le, u msila lo ufun' i mpuku wpna." Alaxe- la noko. Kwasa ke, zapinda ke i ntomba- zana, zaya'kulinda kanjako. Lati i xego : " Tsayitsayibom ! nanzo, mgo- dwanga. Zayidla i ntsimi kakade, iayitshitshela. " Bati bona : "Tsayi- tsayibom ! NanzcT, M^belengambo- nge:— . near you, breed of dogs. " The girls Said : " Tsayitsayibom ! There they are near you, Mabelengambonge : " We have lost our mother, alas ! alas ! etc. " (the same as before). The old man said : " What is that?" So they went home in the evening. The old man said nothing. As to the mbulu, it had stayed at home. They asked it the news. It said : " Our health is good yet. " They did not see it was not a girl, but a m,bulu. So it came out with the chief, and went to sleep with Jiim in his hut. It said it had a belly-ache. Then it said : " Ncincinu (3), I want a medicine. " Then a mouse came. The old man was still there. He said : " That is a mbulu, that tail wants niice {'). " But he did not tell anybody. Morning came; the girls went again to watch. The old man said: " Tsayi- tsayibom ! there they are, breed of dogs. It is a long time, already that they are eating off the garden. They are going to finish it altogether."They said : " Tsayitsayibom ! therethey are near you, Mabelengambonge : i 3!=iE J: Si fe tu, mba nga mba nga! i _1|S_. m E^==:^ ^^_z 3: 4^ U - zi Si -bu Si - m Sa 3a Wa - si tshi se ne ndlii ya le le Mlo - nja lo nja qa - qa si - ngqi sa ha - mba si - fu na hla - nga - na ne mbu hlu - ta mpa - hla ze ke, ni, ke, ma, lu, tu. mba - nga - mba - nga mba - nga - mba - nga niba - nga - mba - nga mba - nga - mba - nga mba - nga - mba - nga mba - nga - mba - nga Si - hie - li zi - tye - ni ze zi - nja, mba - nga - mba - nga 1 .1. Mabelengamtonge \s the proper name of the old man. 2. , I qwili, a word seldom used, is a synonym of iyeza. 3. Ncincinu seems to be the proper name of the chief. 304 Second Appendix. Sifele ma wetii, mbangambanga ! Uzitshise ne ndlu yake, mbangambanga ! ' Sibulele Mlonjalonjani, mbangambanga ! Simqaqa singqi sake, mbangambanga ! Sahamba sifuna ma, mbangambanga ! Sahlangana ne mbulu, mbanga.mbanga ! Wasihluta mpahia zetu, mbangambanga ! Sihleli zityeni (') ze zinja, mbangambanga ! " Bagoduka. Wati u Mabelengam- bonge e nkosini : " Ungandinika nto nina, ndokuxelela i nto ? " Yati i nkosi : " Ndinga kunika i nkomo. " Wati : " Ndi na mazinywana apina o kutya i nkomo ? " Yati : " Ndoku- nika i bokwe. " Wati : " Ndi na ma- zinywana apina o kutya i bokwe ? " •Yati : " Ndokunika i nqweniesha? " Lati ke i xego : " Ndi na singqana sipina so kungxiba i nqwemesha ? " Yati ke- : " Ndokunika u kobo. " Wati ke : " Kauti sibon§; " Baluga- lela ke, walutya ke. Wati ke: " Eza ntombazana ziti zifelwe ngu ma wazo, zahlangana ne mbulu, yazihluta i ngubo zazo. " Kwatwa ke ku la mbulu : . " Ma u dimbaze. " Yangena ke e si seleni. Agalelwa ke a manzi ashushu kuyo. Yasuka yati pundlu e siseleni, yati : " Ndiwadle kade a we nkonazana. " Kukupela kwayo ke. We have lost our mother, alas ! alas ! She has burnt herself with Tier hut, alas ! alas ! We had killed Mlonjalonjani, alas ! alas ! By cutting off his hump, alas ! alas ! We went in search of our mother, alas ! alas ! We met with a mbitlu, alas ! alas ! He robbed us of our clothes, alas ! alasj We now sit in the mangers of dogs,alas ! alas !" They went home, Mabelengim- bonge said to the king : " What will you give me, and I will tell you a thing ? " The king said : " I shall give you a cow. " The man said : " What remnants of teeth are left to me for eating a cow ? " The king said : " I shall give you a goat. " The man said : " What remnants o^ teeth are left to me for eating a goat ? " The king said : " I shall give you a loin- cloth. " The man said : " What loins are left to me to gird them with a loin-cloth ? " The king said : " I shall give you millet. " The man said : " Let us see. " So they poured out the millet, and he ate it. Then he said : " Those girls say that, having lost -their mother, they went in search of her, and met with a mbulu which robbed them of their clothes. " So they said to that mbulu : " Go and take Kafir corn out of the pit. " Then it went into the pit. Hot water was poured over it. But it jumped out of the pit, saying : " I have more than once played tricks of young girls. " That is the end of it. NOTKS. Another version of this tale has been given by Mr. G. M= Call Theal in his " Kafir Folk-Lore. " It contains no song. (a) Mbulu. — The mbulu is a fabulous being, supposed to live near the rivers and to 1. Zityeni, poetical for e zityeni. Likewise, in tlie preceding lines, seveval articles are poetically omit- ted. Thus, Mlonjalonjani stands for u Mlonjalonjani, singci for i singci, ma for u ma, and mpakla for »' i Specijneiis of Kafir Folk-Lore. 305 be fond of playing tricks on young girls. Its essential feature is a (ail. In all other respects it has the appearance of a human being. Some Kafirs identify it with the Gqongqo, described in the following tale. {b) Tsayiisayibotn. — In Kafirland the principal occupation of women in summer time is to watch over the gardehs, so as to prevent the birds, principally a small kind of finch, from eating the Kafir corn which is then ripening. Their usual stratagem for driving the birds away is merely to make a noise by clapping the hands. The exclamation " Tsayit- sayibom ! " is what they are often heard to shout out when they wish to warn one another of the presence of birds in various quarters of the field. (c) That tail wants mice. — In Kafir lore the tail of the mbulu is supposed to be parti- culary fond of mice. In Mr. Theal's version, the episode of the mouse comes, perhaps more naturally than here, only at the end of the tale. The people of the place, having then been told already by the old man that the supposed girl is a mbulu, wish to ascertain the truth of the assertion, and, to obtain their purpose set snares, in which the mbulu's tail gets fast while pursuing mice. , n^birn Tale. A MAGQONGQO NO QAJANA. THE GQONGQOs(«) AND QAJANA (3).' Kwati ke kaloku i nkomo ze nkosi zamita ("). Za Ij shumi. Zazal'e zinye, a yazala e nye. Yasika, lo mhla ya- zala, yazala i nkwenkwe. Ngu Qaja- na i gama la le nkwenkwe. Kwatiwa ma kaaluse i nkomo. Zati ke i nkomo kusasa zapuma e buhlanti. Yati le nkwenkwe : — Allegretto {quasi Allegro). Once upon a time ten cows of the king conceived. All of them calved except one. But the day she calved, she bore a boy, who received 'the name of Qajana. He was told to herd the cattle. So in the morning the cattle went out of the kraal, and the boy sang : Ro qo za - ni, ro qo " Roqozani, roqozani u kuhamba {bis) ". Zahamba ke i nkomo, zaya e hla- tini. Kwati, nxa zityayo, kwafika a magqongqo beza kuziba. Wati o mnye : " Kodwa uyazazi na ? " Wati o mnye : " A ndizazi, siqelile u ku- dla a banye a bantu tina. " Wati o mnye : " Mna ndiya za^i. " Afika ke la magqongqo, aziquba, ha - mba (bis) " Range yourselves to go, range your- selves (bis) ". So the cattle left the place, and went to the kloof '''. While they were grazing, there came gqo.ngqos, who wanted to steal them. One of them said : " But do you know how to manage cattle ? " Another .said : "I don't know, our own custom is to eat other people. " Another said : " I do know. " So they came, those gqongqos ; I, With some Kafir tribes a more usual form of tliis word is semita (Gr. u , 274). 3o6 Second Appendix. azahamba. Azibeta, azibeta, azibeta, azibeta, ada asika ancama agoduka. Yiyo le nkwenkwe yazigodusa • nkomo, isiti : — is fs I ,_ ^ ^S_ ii ^- they tried to drive off the cattle ; they beat and beat them, they beat and beat, until at last they gave up resisting, and went homewards. It is that boy who made them go home by singing : -A H ^ ^>— . — f --? Ro qo za ni, ro qo ku ha - mba {bis). ii ^± 3= =3- =q:: Ni ya bon' u ku " Roqozani, roqozani u kuhamba (bis). Niyabon' u kuba nifile (bis). " Utsho e zinkomeni za ko wabo. Zahamba ke zaya e kaya zafika. Kwasengwa ngo kuhlwa ke, kwa- sengwa i ntlazana. A zapuma i nko- mo. Yati ke : — " Roqozani, roqozani u kuhamba (bis) " (Sung as before). Zahamba ke zaya e hlatini, zafika ke, zatya ke e hlatini. Afika a magqongqo kanjako, azi- beta, azibeta, azibeta, azibeta. A za- hamba. Yati i nkosi ya magqongqo : ".Kahifune e zi nkomeni, zingabi zi no mntu ozitetelayo. " Bafuna ke, basuke ke babona le nkwenkwe i ku nina. Bati : " Bonga. " Yati yona : " A ndikwazi. " Wati o mnye : " Bon- ga, ndokuhlaba ngo mkonto lo. " Wati ke : — ba ni fi le [bis). " Range yourselves to go, range your- selves (bis). You see that you are killed (bis). " Thus he spoke to the- cattle be- longing to his village. So they went homewards, and arrived (safely). The evening milk was drawn, and the morning milk was drawn '■". They did not go out. So the boy sang (as before) : " Range yourselves to go, range your- selves (bis) ". Then they started, and went to the kloof, where they began to graze. Again came the gqongqos, they beat and beat them, they beat and beat. They refused to go. Then the chief of the gqongqos said : " Just look well among these cows, may be there is somebody who directs them." So they looked and found that boy near his mother. They said : " Spell. " He said : " I do not know how to spell ". One of them said : " Spell, or I shall stab you with this spear. " Then he sang : ba ndi le (Ms). specimens of Kafir Folk-Lore. 507 " Roqozani, roqozani u kuhamba (bis). Niya bona u kuba ndifile (bis). " Zahamba ke i nkomo zitinjwa nga magqongqo. Yasuka e nye i nkabi e nkulu a yahamba.Bati ke: " Kwedini, bonga le nkabi. " Yati le nkwenkwe:' " A ndikwazi. " Bati ke bona : " U ya kwazi. " Yati ke le nkwenkwe : — i Et E^ " Range yourselves to go, range your- selves (bis). You see that I am dead (bis). " Then the cattle went, being driven ''away by the gqongqos. But one old ox refused to go. So they said : " Boy, spell this ox. " The boy said : " I don't know how to do so. " Th«y said: "You do know. " Then the boy sang, _-| 1 ^^ — Jy 1= E3E Wa qe qe qe qe ku ha - mba (bis). i E^ 115 ^ =|- U ya bon' u ku " Waqeqeza ('), waqeqeza u kuhamba (bis). Uya bon' u kuba ndifile (bis). " Yahamba ke le nkabi, yema kwe nye i ndawo, bati : " Bonga, kwedini." Yati : — " Waqeqeza, etc. (ike same as before). " Yahamba ke, yafika e mlanjeni, yafika yema. Bati ke : " Bonga, kwe- dini. " Yati ke : — " Waqeqeza, waqeqeza u kuwela (bis). Uya bon' u kuba ndifile (bis). " Yawela ke, yahamba, bayiquba. Yati ya kufika nga se buhlanti, a ya- ngena. Bati ke : " Bonga, kwedini. " Yati ke le nkwenkwe : — " Waqeqeza, waqeqeza u kungena (bis). Uya bon' u kuba ndifile (bis). " (Sung as the previous spells.) Yangena ke. Batata i ntambo,beza kuyixela. Bayirintyela. Yasuka, a ya- rintyeleka. Bati : " Bonga, kwedini. " Yati ke : — ba ndi fi - le (bis). " Take the trouble to go, take that trouble (bis). Thou seest that I am killed (bis). " So the ox went, but it stopped at another place. They said : " Spell, boy. " He sang : " Take the trouble, etc. (the same as before)". So the ox went on ; but, when it came to the river, it stopped. They said : " Spell, boy. " So he sang: — " Take the trouble to cross, take that trou- ble (bis). Thou seest that I am killed (bis). " So the ox crossed the river and went on. They drove it before them. But when it came near the kraal, it refused to go in. They said : " Spell, boy. " So he sang ; " Take the trouble to go in, take that trouble (bis). ' Thou seest that I am killed (bis). " So it went in. They took a riem <'', in order to go and slaughter it. They pulled. But it could not be drawn. They said : " Spell, boy. " So he sang ivv , 1. In another version of this tale I heard u-ya-geqena, which is more regular, but not so well adapted to the rhythm. , 3o8 Second Appendix. t -^ \~\ - Wa - qe - qe -■ za, wa - qe qe za u ku ri - ntye le ka {bis). /7\ :f,%—4 J ^- — lN 1 ^ =ij 1 _ £ \ ■■■ ^ ,1 4 * m ■ — : U ya bon' u ku " Waqeqeza, waqeqeza u kurintyeleka {iis). Uya bon' u kuba ndifile (di's). " Yarintyeleka ke. Bayihlaba apa e siswini ngo mkonto, a wangena u mkonto. Bati : " Bonga, kwedini. " Yati ke : — " Waqeqeza, waqeqeza u kuhlatywa (iis). Uy abon' u kuba ndifile (Ms). " XSu^g as the previous spells!) Wangena ke u mkonto e siswini. BayihUnza ke bayigqiba. A kwatyi- wa ne ntwana e ngcingci, baya kuyibeka e ndlini. Bona baza kum- ka. Bati baya kuhlamba i sisu e Iwandle, bobuya ngo kuhlwa. Bemka ke, bashiya i xekwazana(') e kaya, liza kugcina i nyama na la 'nkweokwe. Yasuke ke le nkwenkwe, ba kum- ka.yatata if) a mafuta, yawapeka e ziko, anyibilika, Yasuke yatata u, mcepe, yaka e mafuteni, yawanika eli xekwazana e shushu. Lati lona : " Ndakutsha. " Yati yona : " Sela. " Lasela, lati : " Ashushu. " Yati le nkwenkwe : " Sela, " ngo msindo. La- sela. Yati yona : " Kwaza. " Lati i xekwazana : " Hu ! i nkomo ziyem- ka. " Yapinda yaka kanjako, ingxa- mele u kuba ze linga kwazi u kuteta. Yalita a mafuta, yati : " Kwazi. " ba ndi fi le {bis). " Take the trouble to be drawn, take that trouble {pis). Thou seest that I am killed {bis). ■' So the ox was drawn. They tried to stab it here in the belly with a spear. But the spear could not go in. They said : " Spell, boy. " He sang : " Take the trouble to be stabbed, take that trouble {bis). Thou seest that I am killed {bis). " So the spear went in into the bel- ly. They skinned the ox and prepa- red it. But not the least bit of it was eaten then, they only went to put it down in a hut. Then they left the place, saying that they were going to wash the tripe ^^ in the sea '«■', and that they would be back at sunset. So they started, leaving a little old woman at home to watch over the meat and over that boy. As soon as they had left, the boy took fat, and cooked it at the fire- place until it melted. Then taking a large spoon, he took out some of it, and presented it quite hot to the old woman. The woman said : " I shall be burnt. " The boy said : " Drink. " She then began to drink, but she stopped, saying : " It is too hot. " The boy said with an angry tone : " Drink. " She drank. The boy said : " Scream (now). " The old woman said : " Whew ! the cattle are going I. It seems that the right spelling of this word should be i xegw<,.ana, not i xekwazana. but I have thought better to spell U as I heard it pronounced. It is derived from i .r,go " an old man ". with the remmme suffix -azt ani the dimmutive suffix -ana (591 and 592). ». Ta/a is the usual pronunciation of the word which is commonly written iaiata. specimens of Kafir Folk-Lore. 309 Lati: " Awu ! " Yapinda kanjako, yati : " Kwaza. " La linga kwazi u kukwaza, litshile nga mafuta. Yasuke ke le nkwenkwe, yati : — i off. " He dipped again into the fat, wishing to make her unable to utter a sound. He poured it into her (throat), then said : " Scream. " She said : " Au ! " He did the same once more, then said : " Scream. " She could not scream, she had been burnt by the fat. Then that boy sang : __J , i^_ j\_ '-T- i Wa ,V qe qe qe qe ku vu ka {bis). --X '^mk U - ya bon' u ku " Waqeqeza, waqeqeza u kuvuka (bis). ' Uya bon'^u kuba bemkile (bis). " Yavuka ke le nkabi ixeliweyo. Yaziquba ke le nkwenkwe i nkomo zonke, igoduka nazo. Yati, ya kufika nazo e kaya, kwa- tiwa :" Be ziye pina lo nyaka wonke?" Yati ke yona : " Za zibiwe. " Kwati- wa ke : " Ulibele (') yi nto nina we- na .' " Yati ke : " Nam be ndimkile nam. " Kwatiwa ke : " Kulungile. " Wona a magqongqo afika e kaya, inkomo zingeka. Ati : " Madlebedlu- mbi (2), i nkomo ziye pina? " A kakw- azi u kuteta. Asuke ke la magqongqo enz' i zi- bata. Yaya ke le nkwenkwe, yaya e zi bateni, yafika kubanjisiwe i ntaka. Yati ke yakulula ke e nye i ntaka, yabanjiswa ke ngo mnwe. Yati : " I ! ub' i sandla sam siye pina ? " Yatiwa go ke nga so, oko kukuti, ziti i zibata ba be mki le {iis). " Take the trouble to rise again, take that trouble (bis). Thou seest that they are gone (bis). " So that ox which had been slaugh- tered rose again. Then the boy drove all the cattle before him, and went home with them. When he got home, the people said : " Where have the cattle been all this long time ? " He said : " They had been stolen. " The people said : " Where were you then ? " He said : " I too, I had gone with them. " So they said : " All right. " As to the gqongqos, when they came home, they did not find the cattle there. They said (to the old woman) : " Madlebedlumbi, where are the cattle ? " But she could not speak. So they went and laid snares. That boy then went where the snares had been laid, and found birds caught in them, but, while he loosened one of them, he was caught himself by one finger. He said:" Hee ! Where do you want to take my hand to (3)? " i. This is a participle. It means lit. ," You having delayed... 2. This is the proper name of the old woman. It means lit. ' Iiong-ears, ". 3. Lit. " You steal my hand that it may go whither.' " " Ears that eat another person ' 3IO Second Appendix. zimbambe. Yati : " I ! ub 'e sinye i sandla sam siye pi? " Yatiwa go ngesi sandla. Yati : " I ! ub' u mlenze warn uye pi ! " Yatiwa go ngo mnye u mlenze. Yati : " I ! ub' o mnye u mlenze wam uye pi ? " Yatiwa go nga lo mlenze. Yati : " Hub' i ntlokw'am iye pi } " Yatiwa go nga yo. Yati : " I! ub' u mlomo wam uye pi ? Yatiwa go nga wo. Afika ke a magqongqo a mabini, ati : " E ! siya mfumana namhlanje u Qajana. " Wati : " Ndikululeni ize ndife. " Ati : " Hayi, uya kubaleka. " Wati yena : " Hayi, a ndisa kubaleka." Bamkulula ke. Wati : " Basani i nya- nda ze nkuni zibe mbini, ize ndife. " Bati : " Hayi, uya kubaleka. " Wati : " Hayi, a ndukubaleka ("). " Bavuma ke bazibasa ke. Wati : " Vutelani no babini, ize ndife. " Bati : " Yi nto nina lo nto? Ungxamele u kuze ubaleke. " Wati : " Hayi, a ndukubaleka. " Bati ke, bavutela ke, wabafaka bo babini e mlilweni i ntloko. Wati ke e zinkomeni za ko wa- bo: — But that hand did " g-o ", that is to say, it was caught in the snares. He said : " Hee 1 Where do you want to take my other hand to ? " He was caught by that hand. He said : " Hee! Where do you want to take my leg to ? " He was caught by that leg. He said : " Hee ! where do you want to take ray other leg to? " He was caught by that leg. He said : " Hee ! where do you want to take my head to? " He was caught by the head. He said: " Hee ! where do you want to take my lips to? " He was caught by the lips. Thereupon came two gqongqos, who said : " Aha ! we have caught him to-day, this Qajana. " He said : " Loosen me, that I may die. " They said : " No, you would run away. " He said : " No, I shall no more run away. " So they loosened him. He said : " Set fire to two bundles of wood that I may die. " They said : " No, you would run away. " He said : "No, I shall not run away./' So they consented and lit the fire. He said : " Blow, both of you, that I may die. " They said : " What is that ? You only want to run away. " So they blew the fire ; then he sent them both into it head-forward. Then he said to the cattle of his own village : i ^- ± ^- Ro qo __> -_^ =< iz qo za ni u ku - go du - ka (dis). /7\ 3^3^ E^3^ Ni ya bon' u ku " Roqozani, roqozani u kugoduka (iis). Niya bon' u kuba batshile (dis). " ba ba - tshi - le (dis). " Range yourselves to go home, range your- [selves (3ts). You see that they are in the fire (dis). " • 1. This is for a ndiyi kuialeka. I specimens of Kafir Folk-Lore. 311 Zagoduka ke. Zona i nkomo za zibiwe nga magqongqo, waziquba, waya nazo e mzini we nkosi ya ma- gqongqo. Wafika wati : " Ndafumana ezi nkomo zibaleka. Ndazinqanda ke, ngabi zezenu. " Kwatiwa : " Ewe. '' Yati ke i nkosi ; " Ma ke uye kuza- lusa. " Wemka nazo ke u kuya ku- zalusa. Wati ke, a kumka nazo ke, wazityoba, wazityoba e mitini, wazi- tyoba,wazityoba,wazityoba. Wabuya ke wati : " Nkosi, le nkomo yandi- hlaba, ma yixelwe. " Yaxelwa ke. Yati ke i nkosi : " Hamba uhlambe eli tumbu e mlanjeni. " Wahamba ke, wati ke yena kruntsu, kratya, wati : " Qweqwede! Bonela, sale, i tumbu le nkosi a ndilityi. " Wemka ke nalo walisa e nkosini, wahamba esiti : " Eyi ! Eyi ! isele yandipanga. " Walinika ke i nkosi, wemka ke, wati uya e zinkomeni. Wahamba ke, wabona u msi uqu- ma nga se hlatini. Waya ke, wafika kungeko madoda, i li xekwazana" lodwa lipeka e nye i nyamakazi.Wati: "Molo, makulu!" Wati: " Maku- lu, yopula i nyama le. " Layopula ke, bayitya. Wayitya u Qajana le nyama ; wati, a kuyitya, wati : " Ma- kulu, kunjanina? Ma senze i ntlonde yo kupekapekana. " Lati i xekwaza- na : " Ewe. " Lamfaka ke lafaka u Qajana e mbizweni. Wavakala u Qajana esiti : " Makulu, ndopule. " Wamopula ke uninakulu. Walitata ke u Qajana eli xekwazana, waliti fungu, walifaka ke ng'e ntloko e ma- '" So they went home. As to the cows which had been stolen by the gqon- gqos, he drove them before him, and went with them to the kraal of the king of the gqongqos. When he came to the place, he said : " I have found these cows running away, I have brought them back, thinking they might be yours. " The people said : " Yes. " Then the king said : " Go and herd them. " So he went to herd them. When he had gone, he drove them deep into the bush, he drove them deeper and deeper. Then he came back (with one cow), and said : " King, this cow is vicious, it should be slaughtered. " It was slaughtered. The king said : " Go and wash this tripe in the river. " He went and bit off a piece ; he found it raw ; then he said : " Qweqwede! See here, frog, the tripe belonging to the king, I cannot eat it. " He took back to the king what was left, saying on the way : " Oh dear ! dear ! a frog has robbed me. " So he handed it back to the king, and went away, saying that he was going to see the cattle. On his way, he saw smoke coming up from the direction of the forest. So he went in that direction. When he came, he fouiid no men there, but only a little old woman who was cooking venison. He said : " Good morning, grandmother. " He added : " Grandmother, take the meat out of the pot. " So she took it out, and they sat down to eat it. Qajana ate most of it. When he had done, he said : " Grandmother, what do you think of this ? Let us play at cooking one another. " The old woman said : " Yes. " So she put Qajana into the pot. Soon he cried out, saying: 312 Second Appendix:: nzini ashushu, Lati : " Shu ! ndatsha, ndopule, Qajana. " Wati : " Yitsha. " Lati : " Hu ! ndatsha, mntan'am Qa- jana." Wati yena : " Vutwa. " Kade lisitsha, wafuna i siciko, wacika. La- sike lavutwa ke. Walopula ke, wali- gcuba, wafaka le nyartia yalo e mbi- zeni kanjako. Le mpahla yalo wali- tata, wambata ke, wahlala ke. Afika ke a madodana, ati : " Yo- pula, ma." Wati: '' Hayi, yopulani, banlwana bam. " Bayopula ke, bayi- gqiba,bayitya. Wati ke wapuma wati : " Ndzebe, badla nina. " Basuka, ba kuva lo nto, bamfunza nge zinja. Wafika u mjambo uzele,'wasuka wa- zenza i sikuni. Afika ke a magqo- ngqo, lati ke e linye : "Ma ke sigibise- le i zikuni. " Bazitata ke, bazigibit>ela pesheya.U Qajana wagibiselwanaye. Wafika nga pesheya, wazenza i nkwenkwe kei kanjako, wati ke : " Ndzebe, nandiweza. " " Grandmother, take me out. " The grandmother took him out. Then Qajana took her and thrust her head,- forward into the boiling water. She said : " Oh dear ! I am burning, take me out of the pot, Qajana. " He said: " Burn on. " She said: " Oh dear! My child Qajana, I am burning. " He said : " Get done. " When she had been burning a long time, he looked for the lid, and covered the pot. So the woman got done. Then he took her out, peeled off her skin, and put the meat back into the pot. He also took her clothes, put them on, and sat down. When the young men came, they said : " Mother, take the meat out of the pot. " He said : " Take it out yourselves, my children. " So they took the meat out of the pot and ate it. Then he went out, saying : " The fools ! they have eaten* their mother. " As soon as they heard this, they chased him, setting dogs after him. He came to a river which was full, he then transformed himself into a log of wood. When the Gqongqos came, one of them said : " Let us throw logs of wood across. " So they took the logs, and threw them over to the other side. Qajana was thrown also, and thus came to the other side. Then he turned himself into a boy again, and said : " Fools 1 you have helped me across " (*). NOTES. (a) — The Gqongqos. — In Kafir lore the Gqongqo (or Kongo, or Qongqongqo) is a sort of wild man of the woods with ears as long as a man's hand, always described as a man- eater. He is distinguished from the ordinary cannibal, who in Kafir is called izim. It strikes me that probably the notion of the Gqongqo is not purely fabulous. My Tonga informants used to designate certain Bushmen tribes which are still in existence under the name of Ma- ngvko, and as there are no clicks in the Tonga language, there is every appearance specimens of Kafir Folk- Lore. 3 1 3 that this word was originally identical with the Kafir noun a Ma-gqongqo. This again taay have some connection with the double f?ct that in ancient Aiabic geographies sever- al South-African tribes are described under the name of Wa-kwakwa ('), and that in these same geographies the Wa-kwakwa are considered as being related to the Chi- nese (°), who go themselves by the name of Gog and Ma-gog. Further in the same line of analogies, the Cape colonists used to call certain semi- Hottentot and semi-Bushmen tribes " Hottentot Chinese " ; and the most rerharkable feature of the language of the Bushmen is that the words generally change their meanings by admitting different accents somewhat as Chhiese does. I wish to draw no conclusion from these coincidences ; I on- ly notice them as being not devoid of interest, and as giving some weight to the thought that the notion of the Gqongqo may be derived from history. Pursuing the same range of ideas, I wonder whether these various words Gqongqo, Ngoko, Kwakwa, Gog, etc., are not related themselves to the name of the Gogo tribe (Ma-^ogo), which is found inland •from Zanzibar, all the more as the Chinese once occupied an island near the Zanzibar coast, and it would be astonishing if their name of Gog and Magog had not been pre- served by one or other of the tribes that had more intimate connections with them at that time. — The long ears of the Gqongqos remind one of the custom which some South- African tribes have of stretching the lobes of their ears by means of copper weights. (Jj) Qajana is one of the most popular heroes of Kafir lore. The facts related here are only a few of his exploits. The very same stories which are told of Qajana are sometimes attributed to Hlakatiyana. Possibly these two heroes are in reality one and the same. They are as it were the Samsons of Kafir lore. The\ characteristic feature of Qajana is cunning and love of revenge. He is not precisely a specimen of courage, and his revenge , falls mostly on poor old women. This represents unfortunately one of the worst traits in the Kafir character. For, as a rule, these people cannot be said to be very respectful or kind to old women. They told me themselves that in former times it was not uncommon, when women wSre getting incapable of doing any more work, to send them to draw water from the river, and then to make them jump into it. {c) The cattle... went to the kloof. — In South-Africa we term kloof 2i sheltered valley bedecked with trees. This is properly a Dutch word. {d) The evening milk... and the morning milk. — These are common expressions to mark the two most important times of the day (cf. Homer's vux-xo? a.)i.ok-\uf, Iliad, XXII, 317 ; Od., IV, 841). In Kafirland cows are generally milked first at sunset when they come from grazing : they are not milked at dawn, but they are then generally let out of the kraal for one or two hours to enj oy in its neighbourhood the short grass that has been refreshed by the dew of the night, and it is only after this that they are milked again to be led afterwards to more remote and richer grazing grounds. Hence the word i ntlazane, or i ntlazana, which means properly " small grass, " has come to be applied to the tirtie of milking cows in the morning. The word kraal is of Dutch origin. The Kafir kraal, u buhlanti, is an open round inclosure, sometimes built with stones without mortar between them, but more commonly made only with thorn-bushes. {e) A riem. — This is another Dutch word. Kafirs have nothing like European ropes. The only strong thing of some length they know of for tying or dragging anything is a kind of thong or leather strap prepared in a special manner, and in South-Africa called a riem. i (/) Kafirs are particularly fond of the tripe of clean animals. But even those among them who make light of the old custom of not eating unclean foodj such as pork, monkeys, eels, etc., would never for anything in the world touch pig's tripe. {g) In the sea. — I have three different versions of this tale in my hands, and in all I. Kwakwa is now the name of that arm of the Zambezi on which Kilimane is built. There is also in Gazaland south of Sofala a tribe still Icnown under the name of Wa-kwakwa . ' 2. Cf. fntroduction. 3H Second Appendix. th|(«e of them, it is not in a river, but in the sea that the Gqongqos a.re supposed to wash the tripe of the ox. This, I think, can be explained only by saying that in Kafir lore the proper dwelling-places of the Gqongqos are supposed to be somewhere near the sea. iji) The conclusion of this tale is common to several others. It is somewhat abrupt. But this is one of the characteristic features of most Kafir tales that they are brought to an end precisely when the hearers would be glad to hear something more about their heroes. Fourth Tale. TaNGA-LO-MLIBO ('). Kwaka kwako u mfazi enga hambi e mini, aze ahambe e busuku. La li ngu Tanga-lo-mlibo i gama lake la ko wabo. Wati eya kuzekwa yi ndoda, wati " a kahambi e mini, " Yati i ndoda : " Mziseni noko, anga hamba e mini, ahambe ngo kuhlwa. " Yam- zeka ke. Waze ke wazala u mntana. Yati i ndoda yake nge nye i mini, yaya kuzingela. Washiyeka ke lo mfazi e kaya, ne xego, ne ntomba- zana. Lati i xego : " Hamba undike- lele a manzi e mlanjeni. " Wati yena : " A ndihambi nge mini, ndihamba, e busuku u kuya e mlanjeni. " Wati u yise : " Ndincede, mntan'am, nda- qauka 11 ngxano, ndiyafa. " Wati ke yena : " Nanga a masi. " Lati i xego : " A ndiwafuni, ashushu. " Wati : " Nabu u tywala. " Lati : " A ndi- bufuni, bushushu. " Wati ke watuma i ntombazaria u kuya kuka e mlan- jeni, wati|: " A ndihambi e mini mna. " Yeza nawo ke a manzi i ntombazana. Lati i xego : " Ashushu. " Lati : " Hamba undikelele vvena, mntan' am. " Wati : " A ndihambi e mini. " TANGA-LO-MLIPO ("). There was once a woman who used never to go out by day, but to go out afterwards at night. Her name at home (*) was Tanga-lo-mlibo. As she was about to be married to a man, she said she could not go out by day. The man said : " Bring her all the same to me : she will not go out in the daytime, but only after sunset. " So he married her, and in time she bore him a child. One day her husband went to hunt. She was left at home with the old man (i. e. her father-in-law) and a young girl. The old man said : " Go to draw water for me from the river. " She said : " I never go out by day, I go to the river at night only. " Her father (-in-law) said : " Have pity on me, my child, I am panting with thirst, I am dying. " She said : " Here is sour milk ('). " The old man said : " I do not want it, it is too hot. " She said : " Here is Kafir beer (■*) " He said : " I do not want it, it is too hot. " So she sent the little girl to go and draw from the river, saying : " I can- not go myself by day. " So the girl came back bringing water. But the old man said : " It is too hot. Go and draw for me yourself, my child. " She said : " I cannot go out by day. " i. Proper name. It means properly " pumpkin of the tender shoot ", i. c. " first pumpkin ', as if the name had been given at the time the first pumpkhis of the year were to be seen in the fields. Specimefis of Kafir Folk-Lore. 315 Wade wahamba u Si-hamba-nge- nyanga, washiya usana e kaya. Waya ke ecatazela, wafika e mlanjeni. Waka ngo mcepe, wasuke watshona. Waka ,ngc mbiza, yasuke yatshona. Waka nge sitya, sasuke satshona., Waka ngo mpanda, wasuke watshona. Waka nge qiya, yasuke yatshona. Waka nge sikaka, sasuke satshona. Wasuke wenjenje waka nge sandl' esi, watsho- na naye wonke. Ushiye u ^sana Iwake e ndlini e htombazaneni. Waze ke wavakala !o mntana elila. Yasuke i ntombazana yamsa e mlanjeni. Yafika, a yabona u nina. Yema nga pezu ko mlambo, yati : — At last Si-hamba-nge-nyanga (the walker by moon -lig"ht^ went, leaviag her babe at home. She went tottering all the way. When she reached the river, she tried to draw with a large spoon ; it sank. She tried to draw with a pot ; it sank. She tried to draw with a basket («) ; it sank. She tried to draw with a jug; Jt sank. She tried to draw with her kerchief ; it sank. She tried to draw with her apron ; it sank. Fi- nally she did thus, dipping this hand of hers ; she sank herself with her whole body. She had left her babe at home in the hands of the little girl. After a time the child was heard crying. Then the girl took it to the river, but she did not see the mother. She then stood on the bank of the river, and sang thus : Andantino. m -^-^H^ litt: :&= ^ 2p3. U-ya li - la, u-ya li - la, um-nta-n'-a-ko, S'-ha-mba-nga-nya-riga. U-ya-li - nga. i 5^S^3E (ter). Pu - ma e mla - nje Umnta - n'a - kw - e - li - la Vel!u - ze ku ma - nyi ni, nje, Si ha-mba - nga - nya - nga. Si - ha-mba - nga - nya - nga. , Si ha-mba - nga - nya - nga. " He is crying, he is crying, thy child, Si- hamba-nga-nyanga (bis). Come out of thie river, Si-hamba-nga.npn- ga, "'' As thy child- is crying thus; Si-hamba-nga- nyanga. Show thyself, and give him thy breast, Si- hamba-nga-nyanga. " Thus the girl spoke. The woman then showed herself in the pool, and Wati, nx' aza kupuma e mlanjeni, before coming out of the river, she wati : — sang thus : 1. Here nyanga is used without its article i. Hence Si-hamia-nga-nyanga, whereas we had above Si- Immba-nge-nyanga (= Si-hamba nga i nyanga). 2. Contraction for u mntana wako elila. ' " Uya lila, uya lila u mntan'.ako, Sihamba- nga-nyanga (■) {6is). Puma e mlanjeni, Sihamba-nga-nyanga, U mntan' akw' elila (") nje, Sihamba-nga- nyanga. Vel' uze kumanyisa, Sihamha-nga-nyanga. " Yatsho ke i ntombazana. Wati ke yena u mfazi, wavela ke e sizibeni. 3i6 Second Appendix. i ~»: ifs; 1*=: U - ti ma ndi - ti Ndi - pu - me e mla - nje ni, ni, No No ta ta nda nda - la ndlo la - ndlo -3- vu? vu? ^^}^^^^^^l?^l (" ^^^- Nde Pe zu Nde - nzi ■ Pe zu lu we lu we nga pe - zu ngu-ba pe - zu U-nxii-tu-me a ma-nzi e - mi A-ndi S'ha - mba nga-nya U-ndi PjE - ZU Be - ta Pe zu A - ku ■ tu lu me lu ■m- si Uti ma nditi ni, nga-ngo pe - zu ngo ca pe - zu ku yi bom, lu, wo, lu, ' ni, nga, ca, lu, na, lu, se. Notanda-la-ndlovu' (")? Ndipume e mlanjeni,' Notanda-la-ndlovu ? Ndenziwe ngabom, Notanda-la-ndlovu, Pezulu pezulu, Notanda-la-ndlovu. Ndenziwe ngu bawo, Notanda-la-ndlovu. Pezulu pezulu, Notanda-la-ndlovu. Unditume a manzi e mini, Notanda-la- [ndlovu. A ndi Sihamba-nga-nyanga, Notanda-la- [ndlovu ? Unditume nga ngpca, Notanda-la-ndlovu, Pezulu pezulu, Notanda-la-ndlovu. Beta ngo cana, Notanda-la-ndlovu, Pezulu pezulu, Notanda-la-ndlovu. A kumsi ku yise, Notanda-la-ndlovu ? " Watsho u Si-hamba nge-nyanga. Wapuma ke e mlanjeni, wamanyisa lo mntana, wabe wangena e manzini. Wati : " Uz' unga baxeleli a bantu b'e kaya u kuba lo mntana ke ndama- nyisa. " Yagoduka ke le ntombazana. No - ta-nda - la-ndlo vu. No - ta-nda - la-ndlo - vu. No - ta-nda - la-ndlo vu. No - ta-nda- la-ndlo vu. No - ta-nda - la-ndlo vu. No - ta-nda - la-ndlo vu ? No ta-nda - la-ndlo vu. No - ta-nda - la-ndlo vu. No - ta-nda - la-ndlo - vu. No - ta-nda - la-ndlo vu. No - ta-nda - la-ndlo vu ? " What dost thou want me to do, Notanda- [la-ndlpvu ? That I should come out of the river, No- [tanda-la-ndlffvu ? My fate has been brought about intention- [ally, No-tanda-la-ndlovu, Above and above, No-tanda-la-ndlovu. It has been brought about by my fether, [No-tanda-la-ndlovu. Above and above, No-tanda-la-ndlovu. He sent me for water in the daytime, No- [tanda-la-ndlovu. Am I not the Walker-by-moonlight, No- [tanda-la-ndlovu ? He sent me as if with a stick, No-tanda-la- [ndlovu, i\bove and above, No-tanda-la-ndlovu. Beat the child with rushes, No-tandla-la [ndlovu, Above and above, No-tanda-la-ndlovu. Why dost thou not take him to his father, [No-tanda-la-ndlovu.'' " ThusSi-hamba-nge-nyanga spoke. Then she came out of the river, gave her breast to the child, and went back into the water. She said : " Do not tell the people at home that I did give my breast to the child. " The girl went home back. Night I. This is the proper name of the girl. It means " Mother of Elephant-track ' specimens of Kafir Folk-Lore. 317 Kwahlwa ke, kwasa i mini, kwaba nge ntlazane, walila u mntana. Yam- sa ke i ntombazana ku nina kanjakq. Yema nga pezu ko mlambd, yati: — " Uyalila, uyalila, etc. (the same as before). " Wavela ke u nina e sizibeni, wati : " Uti ma nditi ni, etc. (iAe same as before). " Wapuma ke, wamanyisa lo mnta- na, wabe wangena e manzini, wati : " Uz' ungatsho u kuti ndamanyisa e kaya. " Yagoduka ke i ntombazana, yaya e kaya no mntana. Kwabuzwa : " Lo mntana umnika nto nina .'' " Yati : " Ndimnika u kutya. " Kwatiwa : " Hayi, xela. " Yati ke yona i nto- mbazana : " Wanyisiwe ngu nina. " Wati ke u yise : " Ub' epumile e mlanjeni ? " Yati ke yona : " Ewe. " Yatsho ke yalila no yise. Wati u yise: " Ze siliambe ne ntambo ngo mso, siye kumrola, simrolele apa. " Kwati ke, kwa kusa, yahamba i ntombazana na madoda, yaya e mla- njeni. Yema pezu ko mlambo kanja- ko, yati : — came, then dawn, tlien full daylight, and then the child began to cry. So the girl took him back to his mother. Again she stood on the bank, arid sang (as before) : " He is crying, he is crying, etc. " So the mother showed herself in the pool, and sang (as before) : " What dost thou want me to do, etc. " Then she came out, gave her breast to the child, and went back into the water, saying: "Do not tell anybody at home that I have given him the breast. " So the girl went home back carry- ing the child. This question was ask- ed : " What do. you give to that child ? " She said : " I give him food to eat." The people said : " Impos- sible, tell the truth. " Then the girl said : " He has been suckled by his mother. " So the father said : " Then she catiie out of the river ? " The girl said " Yes ", and she shed tears to- gether with the father. The father said : " Let us go with riems to-mor- row, to drag her hither. " So on the following morning the girl went with the men in the direct- ion of the river. Once more she stood on the bank and sang thus : J^J^ rjiiirt d^= -I- I ll=^E^E^^^^i=£Ei zdi+i; ■J-J^L-Jy :*zM=t S m la, u-ya li - la, um-nta-n'-a-ko, S'-ha-mba-nga-nya-nga. U-ya-li nga. eSeS^^^e^eI (6 times). Pu - ma e mla-nje ni. Si We - nzi we nga - bom Si Pe - zu - lu pe - zu - lu, Si Wa-turnywa a ma-nzi e mi ni, Si Ka-nt'u-ngu S'ha-mba nge nya - nga, Si U mntan' ak'u - ya - li - la, Si - ha-mba-nga - nya - nga. - ha-mba-nga - nya - nga. - ha-mba-nga - nya - nga. - ha-mba-nga - nya - nga. Si - ha-mba-nga - nya - nga. Si - ha-mba-nga - nya - nga. 3i8 Second Appendix. " Uya lila, uya lila, u mntan' ako, Sihamba- nga-nyanga (bis). Puma e mlanjeni, Sihamba-nga-nyanga. Wenziwe ngabom, Sihamba-nga-nyanga, Pezulu pezulu, Sihamba-nga-nyanga. Watunywa a manzi e mini, Sihamba-nga- nyanga. Kanti u ngu Sihamba-nga-nyanga, Siham- ba-nga-nyanga. U mtan' ako uya lila, Sihamba-nga-nyanga.'' Akapuma. Emka ke a rtiadoda, Yasala i ntombazana, yati : — i .-^- " He is crying, he is crying, thy child, Si- hamba-nga-nyanga {bis). Come out of the river, Si-hamba-nga-nyan- ga. Thy fate has been brought about intention- ally, Si-hamba-nga-nyanga, Above and above, Si-hamba-nga-nyanga. Thou wast sent for virater in the daytime, Si-hamba-nga-nyanga. Yet thou art the Walker-by-moonlight, Si- hamba-nga-nyanga. Thy child is crying, Si-hamba-nga-nyanga.'' The mother did not come out. So the men went away. The girl remain- ed behind, and sang again : ^ N -i^^^^E^ ^ U mntan' ak' u ya li Vel' u - ze ku ma - nyi " U mntan' ako uyalila, Sihamba nga [nyanga. Vel' uze kumanyisa, Sihambarnga-nyanga. " Wapuma wamanyisa u nina, wabe wangena e manzini. Yagoduka ke le ntombazana. Yafika yati : " Uke wapuma e mva kwenu. " Kwasa ke,yaya i ntombazana, yaya na madoda kanjako. I ntombazana yahamba pambili, a madoda ahamba nge mva kwe ntombazana. Afika ke a madoda, azimela. Yema i ntomba- zana nga pezu ko mlambo kanjako, yati : — " Uyalila, uyalila, etc. (the same as the day [be/ore). Wavela u Si-hamba-nge-nyanga, wati : " Ndi ma manwele. Ndiyoyika ngati uze na bantu. " Yati le nto- mbazana : " Hayi, andizanga na ba- ntu. " Wapuma ke wamanyisa. Wa- bonwa esamanyisa nga madoda. Yasuke i ndoda yake yati ruquruku la, Si ha - mba-nga - nya nga. sa. Si ha - mba-nga - nya nga. " Thy child is crying, Si-hamba-nga-nyan- fea, Show thyself, and come to g^ve him the [breast, Si-hamba-nga-nyanga. " The mother then came out, gave her breast to the child, and went back into the water. The girl went back home. When the girl came, she said : " She (the mother of the child) came out after you had gone. " Morning came. The girl went back with the men as before. She walked in front, and the men walked behind her. When these came near the river, they hid themselves. The girl stood again on the bank of the river, and sang: " He is crying, he is crying, etc. (as on the [preceding day). Si - hamba - nge - nyanga showed herself and said : " I feel my hair standing on end upon my head. I fear you have come with other people. " The girl said : " I have not come with anybody. " Then the mother came out and gave her breast to the specimens of Kafir Folk-Love. 319 nge ntambo e mqaleni. Bamrola ke bambekisa e kaya e ndlini. Kwa u mlambo wahamba nawo, ulandela lo mntu ubanjweyo. Seza i siziba sahlala e zantsi ko mzi. Kwaya kutengwa i zinto e zintsha, ne ziko- tile, ne qiya, ne lokwe, ne kumtye. Zabekwa ke e mlanjeni. Sahlala, asemka. Yati yeza i nkomo e bomvu ibaleka, yaya e sizibeni, yabuya le nkomo. Sahlala ke i siziba. Wati u Si-hamba-nge-nyanga : " Tumani u mntu u kumxelela u ma u kuba ndatshona e mlanjeni. " Kwatunywa i nkabi. Yati ya kufika, yati i ndoda ka Si-hamba-nge-nya- nga : " Nkabi, ndikutume na? " Yati " Mmo. " Kwatiwa nku, yabetwa yapuma ke. Yatunywa i bokwe. Kwatiwa, ya kufika: " Bokwe, ndikutume na?" Yati : " Me. " Kwatiwa nku, yabe- twa, yapuma ke. Yatunywa i nkuku. Kwatiwa : " Nkuku, ndikutume na ? " Yati : " Ewe. " Kwatiwa : " Uye kuti nina t " Yati : " Ndiya kuti : — Allegretto. child. She was then seen by the men. Her husband rushed up, and threw a rope round her neck. So they dragged her, and brought her home into the hut. But the river also went along, fol- lowing the person who had been sei- zed. The pool went to fix itself at the foot of the kraal. The people went to buy new things (-^), tinvessels, an apron, women's clothes, and crocke- ry. They were put into the river. But it remained there, and would not go away. Then a red cow came run- ning, and went into the pool ; but it came back, and the river did not move. Si-hamba-nge-nyanga said:" Send somebody to tell my mother that I sank down into the river. " The people wanted to send an ox. When it came, the husband of Si- hamba-nge-nyanga said : " Bullock, shall I send thee ? " The ox only bellowed. So they struck it, and it went out. Then they wanted to send a goat. As it came, somebody said : " Goat, shall I send thee ? " It only said : " Bay 1 " They struck it, and it went out. Then they wanted to send a cock. Somebody said : " Cock, shall I send thee ? " The cock said " Yes. " The people said : " And what wilt thou say? " The cock said : " I shall say : i ; rapido. --A- (ter). Ndi za Tanga - lo - mli bo U tsho - ni le " Ndiza kubika Tonga-lo-mlibo utshonile, Utshonile e mlanjeni. " ku - bi - ka u - tsho ni le, e mla - . nje ni. " I come to report That Tanga-Io-mlibo has sunk down, Sunk down into the river. " 320 Second Appendix. Kwatiwa : " Kulungile. " Yahamba ke. Yati ya kufika ku lo mzi ka Si- hamba-nge-nyanga, kwatiwa : " Uti nina ? " Yati : — They said : " All right. " So it went away .When it reached the birth-place of Si-hamba-nge-nyanga, the people said : " What hast thou to say ? " It sang : i Recitativo. ^^ zk --J ^^sTTh T- -i*=^' d;i= X $ Ku - lu - ku Rapido. ku ku ! (■) Ndi-nku - ku - nje. (quater). A - ndi nku ■ - ku Ndi - zo Tanga - lo - mlibo U - tshonile e ya ku ku tsho mla be bi ni nje twa. ka. le. ni. i ^= E^- U - nga ndi bii " Kulukukuku ! Ndi nkuku (°) nje, A ndi nkuku ya kubetwa. Ndizo kubika Tanga-lo-mlibo utshonile, Utshonile e mlanjeni. Unga ndibulali nje {bis). " Yatsho ke i nkuku, walila u yise ka Si-hamba-nge-nyanga. Wati u nina, a kuva, wati : " Hamba siye e mlanjeni ku lo mntana wetu, sika- ngeie u kuba simtenge nga nto nina. '' Bahamba ke, bafika ku lo ndoda ka Si-hatnba-nge-nyanga, bati : " Kuxe- Iwe i nkomo e md,aka, ifakwe e mlanjeni. " nje {6is). Kwaxelwa ke i nkomo e mdaka, yafakwa ke e mlanjeni. Semka ke i siziba, saya kuhlala e ndaweni yaso. Ipelile ke. la - li " Kulukukuku ! I am a cock as you see. I am not a cock to be beaten. I have come to report That Tanga-lo-mlibo has sunk down, Sunk down into the river. Do not chase me in this way {it's). " Thus the cock spoke. The father of Si-hamba-nge-nyanga shed tears. Her mother, when she heard this, said : " Let us go towards the river to that child of ours, that we may see how we may buy her back. " So they went. When they came to the place of that husband of Si-hamba- nge-nyanga, they said : " A black cow must be slaughtered, and thrown into the river. " Then a black cow was slaughtered^ and thrown into the river. Then the pool went oflf back to its proper place. That is the end of the story. J. Imitation of the cock's crow. X. Poetical for «rfj yi nkuku. specimens of Kafir Folk-Lore. 321 NOTKS. [a) Tanga-lo-mlibo. — This is one of the most popular tales in Kafirland. I have my- self collected six different versions of it., Here I give the fullest of the six. A version differ- ent from every one of mine has been published by Mr. Theal in his " Kafir Folk-Lore, " pp. 56-66. The most peculiar feature of Mr. Theal's version is an introduction explaining how it happened that Tanga-lo-mlibo could not go out in the daytime. (d) Her name at home. — This name is opposed to that of Si-hamba-nge-nyanga, which this woman is going to receive at her new home. It is very common with Kafirs to have different names in different places. They are particulary careful to assume a ne\fr name when they go to work for white people, in order that their master may know as little as possible concerning their antecedents. Hence, among other causes, the great difficulty which is sometimes experienced in identifying thieves. if) Sour milk. — This is the principal food of every Kafir who has a sufficient ritimber of cattle. It is nothing else than coagulated milk from which the whey has not been remov- ed. It is kept in skin-bags, which men alone are allowed to touch, and which are well shaken in order to break the little lumps, whenever milk is poured out of them. Every time the cows are [milked, the fresh milk is poured into these bags, where, mixed with the old milk, it ferments rapidly without any further trouble. Sour milk is the most refreshing drink Europeans can have in South-Afrika. id) Kafir beer. — Kafir beer is the same as the Abyssinian doorah. The ordinary kind is made out of Kafir corn, which is a kind of sorgho. The corn is first soaked in water, then left to sprout until the sprouts are nearly half an inch long. Then it is spread out in the sun to dry. When quite dry, it is mixed with an equal quantity of corn that has not sprouted. The women then kneeling before a flat stone a little hollowed out pound this corn on it with a small oval stone. The malt thus obtained is cooked in water till it boils, and left to stand in barrels for a day or two. Over night a little malt that has been kept is thrown over the liquid, to set it into fermentation. The following, day the beer is strained through a small bag of wicker-work, which allows most of the substantial elements of the flour to pass with the liquid. The beer thus prepared, though a little sour, is a beve- rage not to be despised. Of course it cannot be kept more than two or three days. Kafir women are often valued as wives according to the quality of the beer they make. Some Kafirs have the bad taste to pour some bottles of brandy into their barrels of beer. This certainly does not improve it. («) She dipped a basket. — Kafirs know how to make wicker-work baskets, which, not leaking in the least, may be used to hold not only milk, but even water. (f) The people went to buy new things. — Undoubtedly this is a sentence that does not belong to the original version of this tale, as it mentions severalobjects which the natives have learned to know of only through Europeans. It shows how Kafir lore is being transformed under new conditions of life. Hlpl)abettcal Intie;j:. N. B. The numbers which are preceded by p. refer to the pages in the Appendices. Those in italics tekt to the paragraphs in the Introduction. The others refer to the paragraphs in the Grammar. Those which are accompanied' by an asterisc (*) refer to the comparative tables sub- joined to various paragraphs. A, how pronounced, 6. — = U, 273. ( — how changed before a vowel, 249. — when accented, 309. a, article, 317. a, classifier, 337, 535. a, pronoun,639*,65o,6si,737, 858, 995-1003. a, relative particle, 718, 723, 724, 781. a, auxiliary, 875, 892-910. a, preposition, 570, 573, 743, 746. a, conjunction, 785. a, prefixed to various pronouns, 812, 814, 820, 821. -a, verbal ending, 833, 842-854. . • above, 133, 530, 533*, 541, 581(4). , , ' Abulfeda, no. abundant, 601*- Abyssinian tribes, origin of, 72. ACCENT, 301-312. — its influence on the forms of the words, 444. — its effect when transposed, . - 468(1), 559. accustomed, 1075. actions, 454. ^ ADJECTIVES, 600-634. — rendered by relative, or pos- sessive expressions, 778-780. ^ ADVERBS, 873, 1086. affinity, how close between the various Bantu languages, 246. affirmative clauses, 832. •'; afterwards, 1016. j * -aga, verbal suffix, 870.. again, 1012-1015. aged, 67. agent, name of, after passive ; - - verbs, 589, 1042 ('A''. \5.;. agglutinative languages, I08. Agisumba, Agisyraba, (p. ., A agree, to, 1084. ', ■'' agriculture, 454. '•;. ' air, 377- . . a-ka, auxiliary, 966. •aka, verbal suffix, 870. -ala, verbal suffix, 1031, 1037. alive, 601*. all, 136, 194,250, 810-812. alone, 250, 814-818. along, 564. ALPHABET, 3-38. already, 987, 991, 992, 1016, 1017. also, 691, 819-823. always, 810. amidst, 758-763. among, 554, 563, 565, 758-763. -ana, nominal suffix, 518, 590. ■ana, verbal suffix, 1057, 1084, ancient, 601*, 236. and, 570, 971, 972, 939.' Anderson, on certain ruins, 18. -ans ?, 803. anga, preposition, 576. ■anga, verbal suffix, 833, 870. Angazidja language, 212. Angola cluster of languages, 14.. — sources for its study, fj. Angola language, — — its phonetic features, 146-158. — its article, 317, 321,;, — how its purity has been pre- served, ^S, ani?, 803, S04. animal, 525. animals, names of, 358,401, 483. ankle, 402. another, 827-829. answer, to, 128. ant-hill, 491. 503(9)- any one, 813. ■anjii?, 803. appear, to, 1058-1061. APPLICATIVE VERBS, I065-IO72. Arabic words in Swahili, 85. Arabs in Africa, 83-gj. -ari, nominal suffix, 592. arm, arms, 81, 179, 230, 232, 414, 462*, 468, (2), 484. armlet, 482. arrive, to, 52*, 105. arrow, arrows, 99, 186,372, 525. ARTICLE — — its forms and use, 317-321 — before relative clauses, 774, 776, 777- — in Herero, 616, 623. — in Ganda, III. — in Nyambu, 119. — not found in Swahili, 84. -^ how combined with other particles, 572. — its origin, 830. artificial objects, 372 404. as, 576, 944- as if, 786. ashes, 420. ask, to, 126, 206. aspiration, replacing a suppress- ed consonant, 298. assimilation of vowels, 249-255, 263, 276, 277, 290, 612. assimilation, predominant in Bantu, 299. at, 565. at last, 929: at night, 556. ■ati, nominal suffix, 592. augmentative nouns, 425-428, 430. authority, 458. AUXILIARIES, 873-1018. — their general meaning, 1085. — when they must be separated from the principal verb, S92 (N. B.) Avatime language, 598. awake, to, 173, 179, 1074. axe, 504*, 525. Azauia, V, euphonic, 294, DZ= Z after N, 288. dza, auxiliary, 948; -dza, verbal suffix, 1075, 1079. d%i, classifier, 496. * dzi, reflexive pronoun, 655. E, how pronounced, 10, 271, 272, 1066. — how changed before vowels, 252. — = A, 274. — = A-I, 249. — = /, 200. e, article, 317. e, classifier, 331, 367, 369, 390, 411.492,493- e, locative classifier and prepo- sition, 543. e, reflexive pronoun, 655. e, connective pronoun, 639*. e, relative particle, 718*, 723, 724. e, indefinite pronoun, 955. -e, verbal ending, 833, 862, 872, 874. ■ea, verbal suffix, 1055, 1060. ear, ears, 72, 143, r6i, 211, 232, 243, 462*, 465, 468(1). ear-ring, 525. earth, 89, 90, 1 47. eat, to, 52*, 178, 838, 841. eci>, 801. -ede, verbal suffix, 860. Edrisi, gi, no. egg, 67, 161, 419. eight, 789, 796. ej, classifier, 492. -«/ia, verbal suffix, 1055, io6l. ■ela, verbal suffix, 1065-1072. eland, 358. elder, 153. ele, copula, 1022. -eUla, verbal suffix, 1065-1072. elements, 374. elephant, 126, 133, 208, 214, 233. 38s* elision of vowels, 249, 255, 256, 6l2. embers, 504*. emphasis, 302, 663, 704-706. empire, 458. •ena, verbal suffix, 1065-1072. enclitics, 656. end, 471. enda, auxiliary, 918-940. enda, copula, 1034. endings of the verbal forms, 833- -«»c, 825, 1033. enja, enje = enza, enee, 289, enjoyment, 487. enter, to, 191, 221. enyo, auxiliary, 947. eo, locative pronoun, 682. epithets, adjectives as, 613-617. ■era, verbal suffix, 1065-1072. •erera, verbal suffix, 1065- 1072, eri, copula, 1026. ese, auxiliary, 992. euphonic letters, 641 (3). evil spirit, 410*, p. 283. evening, 421, , ever, for, 782. EXPANSIVE VERBS, I080-IO83. extinguish, to, 147, 1073. eye, eyes, 76, 89, 126, 133, 136, 143. 174. 194. 208, 228, 230, 236, 410*, 414, 419, 439(4)- eyelids, 99. F, how pronounced, 11. -= G, 77, 77. — = H, 177- — = ■ff', 177, 233. — = P, 180. — = iV^, 99. — = SH, 177. — = r, 128, 133. — suppressed, 66, 225, 233. fa, classifier, 537- face, 228, 230, 4 14, "440*, 453, 457.461(1). fall, to, 462*, 1067, 1073, 1074. family, 482. Fan language, 6g. — its phonetic featnres,23 1-237. far, 96, 533*, 764. Farini's description of certain ruins, 18. fat, 66, 75, 80, 93, 133, 177, 220, 225, 440*, 461 (11). fat, adj., 601*. father, 126, 228, 322*, 342, 365 (7), 748-753. 1078. feasts, 379, p. 287. feather, 373. female, 136; 164, 211, see wo- man. fermented drink, 440*, 461 (3). Fernandian group of languages, — sources for its study, yo. Fernando Po languages, 14. — their phonetic features, 238- 241. fi, classifier, 496, 520, 521. field, 464. fifth, 797. Fiji islands, gs, 815. file, to, 236. finally, 929, 1016. fine, 601*, finger, fingers, 133, 164, 23c, 373- finish, to, 1016, 1017. Fiote language, see Congo (Lower), fire, 162, 211, 232, 239, 366*, 374, 384 (8), 504*. firewood, 73, 80, 188. fire-worshippers, 9^, p. 283. fish, 136, 161. first, 797. five, 789, 792*. flame, 232, 529 (5). flaton, 563. flea, 471, 52S. flesh, 385*, 402. flour, 440*, 455, 461 (6). flow, to, 1074. Alphabetical Index. 327 fluids, 454. fly. 523. fly, to, 1074. fo, auxiliary, ipiz, lOiS- FOLK-LORE of the Kafirs, pp, 296-321. follow, to, 1074. food, 176. Gog, p. 313. Gogo language, passtm. — its phonetic features, 77. Gogo tribes, p.. 313. gold-trade, 8o-g6. Golden Meadows, 362 (2), 72. Goncalo daSylveira,Father,95, P-' 294- foot, 81, 88,233,243,462,468(3). good, 220, 601*, 624, 628, 77Q. for, 1065 force, to, 1073-1078. ford, 499. forehead, 402, foreign, 8z6, 827. foreign, words, 406. form, to, 185. four, 789, 792* fourth, 797. fowl, 333, 385*, 409 (4). from, 563, 575. fruit, fruits, 419, 439 gore, 1004-1012. gothic letters, 5. GQ, how pronounced, 37. Gqongqo, pp. 305-313- grass, 105, 440*, 456, 461 (2). gravel, 485. great, 601*, 779. greatly, 633. greedy, 780. - green, see grass, 440*. grinding-.stone, 517- ground, '502, 581 (l), 626. full, to become, 52*, 67, 94, 99, grow, to, 468. 173. ^, pronoun, 639*. future tenses, 907, 908, 912-916, Guha cluster of languages, 14., 920-928, 930, 937, 946, 967. — sources for its study, 56. G, ®, how pronounced, 12. G= F, 75, 77- — =/, 221. - = -s-, 175. Guinea languages, related to Bantu, 588. guUey, 207. gun, 404. Gunda language, 97. — euphonic, 113, 294, 295,297. gut ta percha, 83 ga, pronoun, 639* Gwamba language, 210. ^a,before possessive expressions, Gwamba tribe, 2() (foot-note) 783' ga, negative auxiliary, 875, sqq. -ga, verbal suffix, 836, 920. Gabiin River, 246. Ganda cluster of languages, 14.'' ... sources for its study, 37. Ganda language, passim. ■ its phonetic features, 1 1 1-1 18. — = TV, 284, Gweno language, 2 12. GX, how pronounced, 38. vy, how pronounced, 13. - = A 73- - = F, 177. - = K, 73, 123, 177, 211. — its article, 317-319- Gangi cluster of languages, 14. — sources for its study, 32. garden, 374. gardens, Kafir, p. 296, p. 300. gather, to, 77, 1076. GC, how pronoiinced, 36. genders in Bantu, 313. genuine, 632(2), 780. gi, pronoun, 639*. Gindo language, 103. ; give, to, 147. go, classifier, 465, 543. go, pronoun, 639*. — in locative expressions, 579, 1046. go, to, 193, 250, S39*-S4i,843, 911-939. go in, to, 52*, 250. Goa, 246. Goanese in South-Africa, 96, Z46 (foot-note), goat, 139, 142, 162, 164, 185, 205, 220, 385. gobane, 982. — = NS, 194. -= P, 73> 177, 148. — = S, 123. — = 7', 73. 93. "4- ha, clas.sifier, 535, 537. ha, pronoun, 639*. habitual tenses, 947. hailstorm, 498. hair, 139, 174. ■hala, verbal suffix, 1061. half, 500. hammer, 404. hand, 77, 415, 419. handle, 180, 372. happiness, 457. hatchet, 522. hatta, conjunction, 958. have, to, Q04, 1037- 1039. he, 637, 639*, 656*. he who, 718- head, 131, 144, 161, 164, 211, 229, 232, 238, 366*, 3S4 (3), 384 (3), 468 (I), 471- heal, to, 77. healthy, 67. God, 86, 105, 322»,339,365(6). hear, to, 133, 137, 152. '86, .— how described, p. 289. 238, 250, 1058, 1084. -^ his abode, p. 289. , heart, 139, 209, 223, 366 , '— eating his body, p. 294 (t). 384 (6). heifer, 418. hen, 126, 139, 188, 214, 358. hpre, 211, 693,, 693*. Heiero cluster of languages, 14. — sources for its study, 48. Herero language, passim. — its article, 317-319. — its phonetic features, 125-130, 133- — its peculiar accentuation, 304. Herodotus, on South-Africa, 81. hi, negative auxiliary, 875, sqq. hide, 491*, 500, 503 (7). high, 425. hill, 126, 491*, 500, 503 (8). him, 639*, 653, 656*. Hinzua language, 211. hippopotamus, 99, 161, 233. — a sacred animal, 461 (10). his, 211, 745, sqq., 7(>^-Tn. HL, how pronounced, 33. — its use in Kafir, 121. — = TLH= .y, 208. Hlakaiiyana, p. 313. hoe, 410*, 422. hoe, to, 52*. Homer, on South-Africa, 8r. hold, to, 172. honey, 118,455. honey-bee, 471. hoof, 194, 471. horn, 419, 484. horse, 401, 454. horses in South-Africa, 72. hot, 778. Hottentot-Bushman languages, 2-5. Hottentots, 3. house, 243, 385*, 390, 409(4). 533*- how many?, 211, 800. hundred, 789, 792*. hunger, 66, 88, 96, 97, 115, 126, 195, 208. hijsband, 96, 214, 223, 235, '241, 322*, 365(2)' hut, 133, 500. hut, burning one's, p. 301, hyaena, 128, 139. ,/, how pronounced, 14. * — = .£, 200. — = N, 198, 285, 414. — = £/, 275, 276. — combined with a consonant, 257-258. — how changed before vowels, 255-258. — change to Y before vowels, 255. — elided before vowels, 255-256. — transposed, 285. — initial, dropped, 250. 31, how pronounced, 14. — when used; 271. — how changed before vowels,' 251. i, article, 317. ;, classifier, 369, 390, 411, 492. 495. 496, 522. 328 :%?i'- South-African Bantu Languages. i, connective pronoun, 639*. i, reflexive pronoun, 655. i, relative particle, 718. i, indefinite pronoun, 955. -i, negative ending, 872. I. 637, 639*, 656'. -la, verbal suffix, 1065-1072, 1073 (N. B.) Ibo cluster of languages, 14. — sources for its study, 44. -ibwa, passive suffix, 1049, 1062. ■ichisia, verbal suffix, 1079. -ide, verbal suffix, 8,60. if, 786, 787, 788, 943, 963, 970, 997, 1001, 1002. ■igua, -igwa, passive suffix, 1048, 1050, 1062. ■iha, verbal suffix, 1073-1078. ■ika, verbal- suffix, 1055, 1061, 1077. ■ike, nominal suffix, 593. -ila, copula, 1022, I023. -ila, verbal suffix, 1065-1072. -He, verbal suifix, 860. im, classifier, 388. immediately, 690. imperative mood and tenses, 832, 835-841, 855-859, 859, 873, 880, 906, 938, 968. IN-MA sub-class of substan- tives, 385. IN-ZIN class of substantives, 385-409. in, classifier, — — • its transformations, 386- 390- — its etymology, 407. — its use, 399-406. in. 533. 564. 565. 544- in front, 533*, 542, 581(9). in the air, 533*, 541, 581(4). m the house, 533*, 553, 561. in the mouth, 554. in the road, 552. in the river, 554. in the sky, 553. ina, copula, 1031. incline, to, 1082. increasfe, to, 1073. indefinite pronouns, 809. Indifins in South-Africa, g6. indicative clauses, 842, sqq., 832, 876, sqq. infinitive forms, 466, 832, 853- 854.' inga, preposition, 576. ■ini, locative suffix, 532, 548, 554. — compared with -yo, 723, inscriptions in South- Africa, 18. insi, auxiliary, 976. — copula, 1036. inside, 533*, 549, 581 (i, 5). instrument, 573. intensive adjectives, 632. intensive notion, 502. INTENSIVE VERBS, IO79. INTERROGATIVE PRONOUNS, I 799-809. INTERJECTIONS, 596. into, 566, 1065. inea, auxiliary, 976. ■ira, verbal suffix, 1065- 1072. iron, 403. iron ore, 455. ■isa, -isha, -isia, verbal suffixes, 1073-1079. isha, auxiliary, 1017. -isidza, -isisa, -isisia, verbal suffixes, 1079. island, 525. Isubu language, 227, 229. it. 637, 639*, 653. 656*. it is, 582-588, 656*, 662, 685, 707- , , its, 745, sqq.,-^6Z-fTj. -iwa, passive suffix, 1052, 1062. -iza, auxiliary, 948, 949, 952. y, how pronounced, 1 5. — = BE, Bl, BO, BU, 122, 202, 178, 445. — = LE, LI, 178, 205. — = Z, 52*, 63, 89, 106. — euphonic, 294, 295. — suppressed, 81. — a favourite in Yao, 68. —ja, auxiliary, 948, 949, 951, 963, 964. jackal, 358. jaka, preposition, 576. Javanese in South-Africa, jy. ji, classifier, 394. ji, connective pronoun, 639*. //, reflexive pronoun, 655. ji, relative particle, 718*. jin, classifier, 394. Jinga tribes, y6. join, to, 1082. joint of the arm, 524. Jorge (Father), 20. journey, 454. just, -974, 987. JW= BW, 202, 1053. K, how pronounced, 16. — = C, 150. — = F,■\^^, 233. — = G, 175, 214. — = H, 123, 211. — = NG, 190, 479. — = NJ, 191. — = SH, 175. -=T, 244. — = TS, 1035. — = TY, 214. — = V, 233. — = W, 175. — suppressed, 175, 211, 225, 233, 290, 292, 559. ^i, how pronounced, 16. — its use, io5, 133. KA- TU class of substantives, 504 -sag- ha, classifier, 504-529. — its use, 524-527. — its original meaning, 527. ka, non-classifying prefix, 345, 347. 35°. 527. ka, connective pronoun, 639*. ka, relative particle, 7 1 8*. ka, before possessive expression, • 783-, ka, auxiliary, 875, sqq. , 965-975, 1000. ka, conjunction, 788, 970. ka, preposition, 573, 574. — ka, suffix in numbers, 792. — /5o,' verbal suffix, 1055, 1061. Kafir beer, p. 321. Kafir cluster of languages, 14. — sources for its study, sy. Kafir folk-lore, pp. 396-321. Kafir language, passim. — its article, 317, 319. — its peculiar sounds, 33-38. — its phonetic features, 120-I24. kala, auxiliary, 941-947, kala, copula, 1031. kala, conjunction, 943, 944. kala, preposition, 576. — kala, verbal suffix, 1055, 1061. Kamba cluster of languages, 14. — sources for its study, 41. Kamba language, passitn.' — its phonetic features, 81-83. kana, conjunction, 943-944. ka-nga, before nouns, 515. Kangombe, p. 285. Karanga cluster of languages, 14. — sources for its study, s8, Karanga language, passim. — agglutination in it, 254. — compared with Kamba, 303. — its phonetic features, 104-108, '33- . Karanga nation, pp. 286-288. kali ka, 758-760, 763. — kazi, nominal suffix, 592. As, pronoun, 639. ke, copulative particle, 587, 1035. ke, auxiliary, 875, sqq., 965-975. — ke, nominal suffix, 593. Keish, the prince of, gi. Kele language, 227, 230. key, 404. KG = NK, 188. — = NG, 189. KH = NF, 187. KI = CI= TYI =SI= SE, etc., 260, 492. A'T, changed to C or CH, 258- 259. ki, for ka, 250. ki, classifier, 462. ki, connective pronoun, 639*. ki, relative particle, 718*. ki, auxiliary, 986, 993, 994. ki ?, 801. kia, auxiliary, 989, 994. kid, 418. kile, auxiliary, 992. Kilima Njaro, 246. kill, to, 81. kindness, 457. king, fs (foot-note), see chief. Ki-rimba, Ki-zimba, 9. kloof, p. 313. Alphabetical Index. 329 knee, 500. kneel, to, 1067. knife, 404, 512. know, to, 834. ko, pronoun, 674 (f), 675-681, 1020. ICoelle, 83. Kololo laflguage, 131 (3), 169. Kololo tribes, p. 293. Konde language, 14. 'r^ sources for its study, .#6. Kongo, 13, see Congo. koo = ke-ku, 967. kraal, p. 313 (d). KU-MA class of substantives, 462-468. ku, classifier and preposition, 462-468, 533-581. — its transformations, 463-465, 542-546. — its use, 466-467, 562-568. — its original meaning, 468, S8i. — dropped, 154. ^«, connective pronoun, 639*. ku, relative particle, 718*. ku, referring to a substantive understood, 781. ku, copulative prefix, 583. kua, locative particle, 783. Kua group of languages, J2. — its divisions, and subdivisions, If- — its proper features, 169, and - passim. — its origin, 97, 246. Kua tribes, their origin, 97. kuba, 982. kta, auxiliary, 9B4. — , negative copula, 964, 1038. .kiiila, kuina, 1023. ktili, 580, 1004-1012. — kulu, nominal sufHx, 591. ,Kumbi tribes, 50. kumi, 792*. kuna, 580, 1OC4-1012. Kush, 72, -js-n- . kwa, locative particle, 543, 546, • 691 (2), 783. kwa, preposition, 573. Kwakwa River and tribes, P- 313- Kwana tribe, 246. Kwango language, 14. — sources for its study,5o. ■— its phonetic features, 132-133. Kwango tribes, 108, p. 293. Kwengo, see Kwango. Z, how ponounced, 17. — = N, 280. — = JI3, 214. — = T, 220. — = Z, 165. - — suppressed, 81, 88, 96, 97, 123, 224, 233. la, classifier, 471, 545. - la, auxiliary, 918-922. la, demonstrative suffix, 698. lake, 499. land, 389. LANGUAGES OF SOtlTli-AFRICA - — their general division, i. — their elegance, 75. — their names, 484, 491*, 497) 503 (2), 14. large, 215,425, 427, 601*. last night, 533, 581 (7). ' ' Last, J. T., 145, z.xA passim. laugh, to, 82, 133, 136, 174. Zj? =7,205. le, classifier, 411. le, pronoun, 639*. le, auxiliary, 918, 930-936. le, copula, 1022. /«, preposition, 570,940. — le, element in Words, 409(2). le, suffix of demonst. pron. 698. Le Roy, p'ather, g4. Lea tribe, p. 285, p. 293. lean, 601*. leave, to, 52*, 136. leave behind, to, 174. lest, 974. Z/ = /=//, 143, 178. li, for la, 250. ! LI-MA class of substantives, 410439. /«', classifier, 395, 410-415, 426. — its transformations, 410-415. — its use, 417-418. — its original meaning, 429- 435- — omitted, 411. — omitted before adjectives, 614. li, connective pronoun, 639*. — referring to a substantive understood, 781. li, relative particle, 718*. //, auxiliary, 918, 930-936. li, copula, 619, 1022, 1024, 1025. li, particle in locative express- ions, 579, 1040-1043. //, conjunction, 784. HI, 806. ha « to have », 1039. — Ha, suffix of demonst. pron. 698. — lia, 401. lie down, to, 52*, 1069. life, 440*, 461 (9). light-hole, 491*, 503(10). liji, copula, 1022. like, 576, 944. like, to, 1084. Limpopo River, 246. lion, 358. lip, 366*, 384 (4). lip-rings, phonetic effects of, 50, 2ip. LITFRATURE, BANTU, 16-70. liver, 500. livin'g-place, 374. Livingstone, p. 289, p. 294, p. 295. — ■ his spelling, 141. Liwanika, p. 284, p. 285. lo, classifier, 476, 490, 511. lo, pronoun, 639*. -Lobengula, p. 286. LOCATIVE EXPRESSIONS, 53O- 581. — containing adjectives, 605. ■ — how enjphasized, 704-705. — requiring certain particles inserted, 1040-1046. — their peculiar value, 43. — used as comparatives, 630. ^- what concord they require, 643, 674, 67s, 755-767. loins, 389, 390. Lojazi language, 132-133. Ion, classifier, 398. long, 214, 601*. lord, 502, see chief, look, to, 1079. loud, 423. Lower Congo, see Congo, love, to, 1058, ro73, 1075. 1084. LU= CW, 205. LU-ZIN cXass of substantives, 469-490. lu, common classifier, 469-490. — its transformations, 470-473. — its use, 475-476, 481- 488. — its original meaning, 4S9. — dropped, 472. lu, connective pronoun, 639*. lu, relative particle, 718* lu, referring to a substantive understood, 781. lu, copulative prefix, 583, Luba language, 14. — sources for its study, 59. — its phonetic features, ,143. Luiana language, 141. Lumbu people, pp. 283-286, P- 293. Lunda language, 14, 788°'=, — ■ sources for its study, JJ. — its phonetic features, 143- 144, 788'='=. M, how pronounced, 18. — = B, 240, 804. — = MO = MU, 279, 240, 560. — = N, 281, 607. — = dropped, 292, 640. 'M = MOB, 279. m, classifier, 328, 367, 551-553- for », 389. f)i, connective pronoun, 639*. m, copulative prefix, 583. MA sub-class of substantives, 442-444. md, classifier, 322, 338, 355- 356, 416-438, 442-4S0. — its transformations, 416, 442- 444. ■ — its use, 355-356, 417-428, 454, 480. — its original meaning, 436- «^- -= * ma, pronoun, 939 . ma, auxiliary, 977, 978, 980, 981. , 330 South-African Bantu Languages. ma, verbal suiRx, 1063. Macrobians, 81, 82. madsd = madzi, 272. Magogi p. 313. Ma-gqongqo, Ma-kalanga, Ma- kua, etc., see Gqongqo, Ka- langa, Kua, etc., maize, 500. mala, auxiliary, 1016. Malays in South-Africa, jy. man, 322*, 340, 365. mana, auxiliary, 1016. mang 1, 804. Mianuel Antonio de Souza. p. 285, p. 293. many, 454, 601*. mark, national, 490. marry, to, 1074. Marvels of India, Book of the, 90. Masai languages, o. Masai tribes, 79. ' Mashcjiialand, 7j, see Shona. Mas'oudi — — on the name of the king of the Zindj, 365 (2). — on the origins of the Bantu, 73, 8t), <)3. — his trustworthiness, /J. Matapa-metsi, 461 (10). match, 504*. may, 966, sqq. MB.= P,iSs. mba, relative particle, 718*, 721, 771. mba, auxiliary, 984. Mbala language, 102. Mbala tribes, 50. Mbamba language, 146-151. Mbangala language, 157. Mbara, see Mbala. mbu, relative particle, 718*, 721, 771. mbulu, pp. 300-305. Mbunda, Mbundu, 13, 132. Mbunda languages, 14.. — sources for their study, 50-53. Mbunda tribes, 77, p. 285, p. 293. me, classifier, 369. me, pronoun, 639*. me, auxiliary, 977, 979, 980. me, conjunction, 910, 985. me, 639*, 653, 656*. meal, 487. meat, 73, 196, 385. medicine, 378. meet, to, 1076. Melanesian languages, 95. melekh, Semitic word, 339. metal, 403. metempsychosis, p. 114 (foot- note). m-falme, 75, 365 (2). Mr = y.f=E, 164. mi, classifier, 366-384. — its transformations, 369-370. — its etymology, 384. mi, pronoun, 639*. 'mi, conjunction, 985. mice, how the mbulu is fond of, P- 305- ^ midday, 136. middle, 139, 504*. milkj sour, 454, p. 321. milking among Kafirs, p. 313, p. 321- mimosa-tree, 371. mine',. 768-777. mines, gold-, 80-82, 84. Mi-zimo, S6. Mlonjalonjani, the tale of, pp. 300-305. MO = NGW, 204, 207. mo, classifier, 328, 367. mo, pronoun, 639*. Molokh, 86, 339, 365 (6). monkey, 358. Monoraotapa, 461 (10), p. 294. MONOSYT.LABIC STEMS — — how accented, 45, 310. — their peculiar laws, 44, 45, 283, 284, 310, 325, 368, 413, 444. 464, 478, 561, 608, 652, 661,832,837, 841,843,851, 853, 866, 867. Monze, the chief, p, 286, pp. 288-290, p. 294. MOODS, 832. moon, 81, 126, 136, 164, 173, 214, 232, 366*, 374, 384 (10), 481. Moon, mountains of the, 82, ■ 266. morning, 533*, 581 (8). mother, 126, 322*, 342, 365 (7), 748-753- motions of the hand, 362, 789, 830. mountain, 425, 491*, 503 (8). mouth, 164, 223, 366*, 384 (4, 10), 504*. Mozambique cluster oflangua-' ges, 14. — sources for its study, 63. Mozambique language, passim. — ■ its affinity with Chwana and Mpongwe, 12, 169. — its phonetic features, 169- 208. Mozambique tribes, 50 (foot- note). nipi, relative particle, 718*, 721, 771. Mpongwe cluster of languages, 14. — sources for its study, 67. Mpongwe language, passim. — compared with Nywcma, 164-167. — its affinity with Chwana and Mozambique, 12, 169, 213. — its phonetic features, 59, 213-218. — its traces of an article, 321 (3)- MU= m= Ny, 122, 581, 792. — = O, 164. MU-BA class of substantives, 322-365. MO-Ml class of subslantivesy 366-384. MU, locative class, 533-581. viu, classifier-' — ■ — its transformations, 323-332, , 367-368, S47-SSS- — its use, 357-360, 371-380, 562-568. — its original meanings, 361, 381, 581. — weakened, 559, 560. mu, connective pronoun, 639*. mu, relative pkrticle, 718. MUA = NYA, 268. muade, poison, 378, p. 284. Muanaena, p. 285. muave, see muade. 'Mue, 792*, 820. ■mue, relative suffix, 725. — = yo = Kg, 725. muene, 825. mwenyewe, mwenyi, 825. multiplied, to be, 137. multiply, to, 108 1. Muhmgu, Muungu, 339. muiia, 580. murder, 423. musical instruments, p. 2S7. mw-, classifier) 326, 367. my, 745. -f??- N, JB, how pronounced, 19. N=I, 152, 198,285. — = L, 280. — = M, 551, 552. — = MU, loi, 107, 153, 559. — = NY, 196. — = Z, 232. — suppressed, or half-suppress- 6d, 282, 551, 584, 640. \/V Class of substantives, see IN- Z/JV class, n = un = mu, classifier, 327, 367, 551- « (= in) classifier, 388-389,391, 477-480. n pronoun, 639*. «, copulative prefix, 583, 1035. na, prefix in substantives, 347, 35°- . . na, auxiliary, 91S-929, 946. na, preposition, 570, 573, 578, 682, 704, 940, 1037. na, in locative expressions, 579, 1040-1046. — in Congo, 537, 542, 549. na, conjunction in Ganda, iii, 939- na " to have , 1037. — na, locative suffix, 553- — na, pronominal suffix, 689, 698. nail, 525. naked, 1034. name, 99, 115, 147, 173,220, 410*, 423. Nano language, 132, 133. — its article, 321 (2). NASALS — — suppressed, or half-suppressed. Alphabetical Index. 331 77, 78, 151, 166, 183-196, 225, 283, 607. — their peculiar influence on following consonants, 51, 52, ■JS-Sg- 67, 69, 73, 77, 78, 80, 83. 93. 94. 95. "4, 127.129. 140, 170-198, ai4-2i6, 283, 412, 479y 608, 609, 649. nations, names of, 322*,-365 (5) «". 773- ND = T, TH, 193. ndi., connective pronoun, 639*. ndi, relative particle, 718*, 721 771. ndi, copulative pre6x, 583,587, 1035- ndi, preposition, 570, 573, 940. -ndini, vocative suffix, 595- NDL = TL = NZ, 208. Ndonga language, 132-133. iidu, relative parHcle, 7i8*,72i, 771. ndyi, pronoun, 639*. ne, auxiliary, 918, 930-931. ne, copulative particle, 588, 1035. ««, H«, preposition, 570, 573, 940. tt«, conjunction, 939. near, 533*, S41, 581 (2), 764. neck, 136, 142. negative auxiliaries, 875-891. negative clauses and tenses, 832, ■ 833, 872, 875-891, 964, 967, 976. — containing the copula, 1036. — containirgtheverb "tohave," 1038. negative particle before num- bers, 798. negative notion, 502. Negro languages, 830. nest, 500. nestling, 418. never, 046, 960-962, 964. : new, 94, IDS, 127, 129, 137, . 147, 601*, 779. NF = KH, 187. NG = K, 190. — dropped, 210. N<& = MU, 1053. — = NI, 200. -n%, relative and locative suffix, 532. 552. 718*, 727, 734, . ■ 802. nga, relative particle, 718*, 721, 771. nga, auxiliary, 875, sqq., 995- , 1003.' nga, copulative prefix, 583, 1035. nga, preposition, 573, 574, 576. •nga, suffix in the number " one ", 792. -tiga, verbal suffix, 920. ngai?, 800. '^' nganga, 385*, 409 (i), p. 290, p. 295. nganga, preposition and con- junction, 576. ■ngapi?, 800. nge, auxiliary, 995, sq^. Ngete tribe, p. 286, p. 923. -»g8, locative suffix, 553.. »gu, relative particle, 718*, 773- ngu, copulative particle, 583, 1035- K<&1^ = MO, 204, 207. NI= MU = N&, 581, 200. ■ ni, classifier, 387, 411. ni, pronoun, 639*. ni, ■ copulative particle, 583, 587. 1035. ni, preposition, 1037, ni, conjunction, 939. ni, 801, 803. ■ni, locative suffixy 532, 548, 5S3i 555, 590. -ni, pronoun suffixed, 855. nice, 601. Niger languages, related to Bantu, 598. night, 128, 174, 440*, 443, 453, 456, 461 (5)- night, at, 556. Nika cluster of languages, 14. — sources for its study, yt?. Nika Hngas^ge, passim. — its phonetic features, 92-96. Nika tribes, no. Nile, sources of the, Sa-Ss. nine, 789, 796. NJ=K= TS, 191. Njenii language, 141 (3). nji, relative particle, 7I8*, 721, 771, 773- NJW = MBW, 1053. fika, nominal prefix, 527. nki?, 801. nle, conjunction, 939. nna, auxiliary, 929, 941-947. nna, copula, 103 1. -nna, pronominal suffix, 689. -nne, 792*. no, nominal prefix, 347. no, not, 872, 875-891, 1036. no longer, 987, 991. no more, 987. no one, 798. -no, demonstrative suffix, 698. . noo = 7ie-ku, 946. non-quantitative adjectives, 601, 624-628. North-African languages, $. nose, 133,402. not at all, 964, 989. not yet, 929, 960, 963, 974, 987, 992. nothing, 798. now, 421. nowhere, 798. Northern traders in South- Africa, fg-Si. NS = H, 194. - = T= TH= 7Zjy=X,i94. Nsundu people, 10. NT= T= TH, 192. iitsi, auxiliary, 945. NUMERALS, 789r798. numeration in Bantu, 789, 830. NV = C, 186. NY= MU, 122, 581, 1053. — = 7V, 196. ny, classifier, 478. NYA = MUA, 268. nya, nominal prefix, 465. -nya; pronominal suffix, 689. Nyambane cluster of languages, 14. — sources for its study, 63. Nyambane language, 210. Nyambu language, 1 19. Nyamwezi cluster of languages, 14. — sources for its study, 3^. Nyamwezi language, /ajj;/«. — its phonetic features, 73-76 Nyamwezi tribes, 50 (foot-note). -nyana, nominal ^liffix, 518,590. Nyanja, or Nyassa, language; see Senna. nye, 829. Nyengo language, 142, nyi, pronoun, 639*. Nyika, see Nika. -nyo, pronominal suffix, 689. Ny wema cluster of languages, 14, — sources for its study, 57. Nywema language, passim. — its phonetic features, 163- 168. NZ= D = TH= T = TL, 210, 195. — = DZ, 21 r. — = N = NY, ig6, 197. — = x, 195. mi, relative particle, 718*, 721, 771. nzi7, 801. 0, 3D, how pronounced, 20. — • = A-L, or A-0, 249. — = A-KU, 907. — = MU, 164. — = U, 200, 262.. — = UA, WA, UE, WE, 265, 270, 659. — = UO, 263. 0, article, 317. 352, 353. 0, classifier, 330, 331, 367, 449, 465, 480. 0, connective pronoun, 639*. 0, relative particle, 718*, 723. 724- -0, pronominal suffix, 659, 698. oba, 982. object of a verb, how expressed, 653. . , — in relative clauses, 735-742. OBJECTIVE VERBS, See APPLI- CATIVE. of, 577, 589- Ogowe River, 246. oil, 66, 75, 177, 220, see fat. ojo, auxiliary, 949. ■oka, -ola, ■olola, ■oloka, verbal suffixes, 1080-1083. old, 601*. old man, 139, 500. old women ainong Kafirs, p. 3!3- olo, auxiliary, 920, 942. 332 South-African Bantu Languages. on, 544, 563, 564, 565, 1065. once, 525, 797, 974.' ondo, auxiliary, 918, 937, '949. one, i22, 235,789, 792*. one... another, 828-829. one who, the, 718. O'Neil, on South-African in- scriptions, iS. • ONOMATOPOETIC WORDS, 596.' open, to, 1059, 1081. Ophir, 8$: ordeals for sorcerers and thieves, pp. 283-285. ordinal numbers, 797. ore, 455. ORIGIN OF THE BANTU, yi-IOO. ORTHOGRAPHY, regarding the separation of the words, 892 (N. B.J. ostrich, 388, 390, 401. other, 826-829. our, 745, sqq. ours, 768-777. outside, 136, 195, 208, 533*, 581 (3).' over, 531, 563, 564. ox, 202, see cow. oxen, pack-, 72. P, how pronounced, 21. — = B = V=H^ 137, 138, 166, 180, 211, 215. — = BF, 94. — = F, 180. -=/.89. — = ]\dB, 185, 479. — = MP = PH, 184. — = NV, 186. — weakened, or suppressed, 64. 69, 73> 74, 77, ii7> 148, 292. pa, locative classifier, S33-581. — its transformations, 534-541. — its use, 562-568. — its meaning, 581. pa, connective pronoun, 639*. pa, relative particle, 718*. pa, negative auxiliary, 875, sqq. pa, conjunction, 785. paddle, 142, 500. palm-leaf, 500. pana, 580. Papuan languages, ^y. Parsees in South-Africa, pj. p. 292. participles, 851, 910, 931-935, 979, 980, 993- particles changing their forms, 291-298. Paruain, 8s. PASSIVE VERBS, IO47-I063, — the name of the agent after them, 589, I042. past, 498. past tc-nses, 892-907, 930-936, 909, 910, 917, 966. path, 97, 142, 147, 385*. 409 (4), see road. Paz, Ss. pe, pronoun, 639. />e,;negative particle, 872. pebble, 504*. Pedro Dias, Father, see Bias, people, 210, 232, 322*, 340. perfect forms and tenses, 70, 860-871, 892, 904, 905, 981. permanently, 810. Persians in South-Africa, pj, p. 292. person, 133, 139, 142,192, 223, 241, 243, 244, 322*, 340, 365 (I). PERSONAL PRONOUNS, 635- 692. persons, names of, 357. personified, things, 360. PH = MB = P, 184. PHE = CHW, 203. PHONETIC CHANGES — — their general laws, 47-59, 247-300. — their main cause, 50. — specimens, 52*, and passim. ■ — in the various languages, 60-246. — in the perfect form, 861. — caused by suffixes, S96jl053- P/ = TSH, 122. pi? Soo. pierce, to, 202. pig, 233. pipe, 404. P>t, 193, 377- place, 403, 537, 5S6, 782, 783. place, to, 1077. plain, 374. plant, to, 122. pleasant, 601*. plural, for the singular, 343, 344, 750- PO = CW, 203. point, 389. Pokomo cluster of languages, 14. — sources for its study, 40. Pokomo language, passim. — its phonetic features, 92-96. Polynesian languages, jy. pool, 374. poor, the, 554, 601*. porridge, 404. POSSESSIVE ADJECTIVES AND EXPRESSIONS, 57.7, 659, 684, 743-780. pot, 500, 525. potato, 500. pot-clay, 455, 525. pour water, to, 1066. powder, 378. Prasos, Cape, 96. pray, to, 1067. prayer, 487. prayers of the Bantu, 365 (6), pp. 287-290. Predicates, adjectives as, 618- 623. Prefixes of the substantives, see CLASSIFIERS. PREPOSITIONS, 530-581, 674- 688, 940, 1086, present tenses, 930-936, sqq. priest, see cacice. privative notion, 502. productive notion, 502. PRONOUNS, 635-830. proper names of persons, 346. provoke, to, 1059. PU = TSH, 122. pumpkin, 415. Pun, land of, 84. , put out a light, to, 147. Q, how pronounced, 37. qa, 792, 794. Qajana, the tale of, p. 305, p. 313- quantitative adjectives, 601. question, 206. P, how pronounced, 23. — = T, 172, 2IO, 211, 214. — = 2 = Z», «73-. ra, nominal prefix in Timneh, 471. ■ra, suffix, 594. race, 482. raft, 482. rain, 66, 73, 79, 107, 114, 116, 123, 143, 186, 224, 243, 385*, 403, 409(4). rain, how obtained, pp. 287-290. rain, to, 1071. rain-bow, 375. raise, to, 1076, 1082. raising the dead, Mcnze, p. 290. razor, 478. RE, = TSH, 206. re, pronoun, 639*. re, auxiliary, 875, sqq., 1004- 1012. Rebmann's mode of spelling, 92, lOI. — his enthusiasm for the Nyassa language, 98. RECIPROCAL VERBS, I084. recover, to, 52*, 1076. red, 624. red clay, red ground, 626. reduplicative forms, 632(2), 1079. REFLEXIVE PRONOUN, 655, 198. Regga language, 14. — sources for its study, 36. RELATIVE CLAUSES AND EX- PRESSIONS. — — used substantively, 768-777. ^ negative, 878. — their construction, 717-742. — forms of the verb in them, 844. „ ^ RELATIVE PARTICLES, 718-746. — in what they differ from rela- tive pronouns, 718. — their forms, 719. — their etymology and nature, 830. — their use before relative clauses, 728, 731-742, 769- 777- Alphabetical Index. 333 — their use before possessive ' expressions, 769-777. — their use before adjectives, ' 605, 616-617. ^RKLATIVE PRONOUNS, 717-746. teinain, to, $2*, 126, 941-947, 1031-1033. repetitive tenses, 946. return, to, 1074. REVERSfVE VERES, I080-I083. rhinoceros, 500, 502. W, classifier, 411. •«', connective pronoun, 639*. ri, reflexive pronoun, 655. «', copula, 1022, 1024, 1026. rich, 233. riem, p. 313. rile, auxiliary, 1008. rire, copula, 1022, 1026. ring, 404. rise, to, 1076. river, 203, 366*, 374, 384(9), 470*, 490(2), 525. rivers, names of, 466, 470*, 486. Roa, or Rwa, people, 3. road, 72, 88, 97, 126, 142, 147, 187, 195, 224, 239. rod, 404. roof, 482, rope, 470*, 485, 490(3)- ropes, how made, 490(3). Rotse cluster of languages, 14. — sources for its study, s^- Rotse language, f.assim. — its phonetic features, 135- 141. Rotse nation, <)3, pp. 283-286, p. 292. round, 564, 1065. rouse, to, 174. row, 482. ru, classifier, 490, 509. ru, pronoun, 639. Rua language, 14. — - sources for its study, $8. — its phonetic features, 145. ruins in South- Africa, 18. Runda, see Lunda. S, S), how p.onounced, 24. — = C, I76i — = A =7,136. — = 13,82. — = F, 410*, 656*. — = H, 123, 174. — = %iW, 105. — =T, 90, 128, 174, 239- -^=TLH, 174. ' — = TSU, 174. — = TY, 214. ^ = X, 106, 133, 136. — — Z, 214. — suppressed, 174. sa, nominal prefix, 347. sa, auxiliary, 875, J??., 9^6, 991. 992, 994- Saba, the queen of, SS- Sabseans, 83-86, 365(6). sacrifices, pp. 286-288, p. 294, saddle, 500. Sagara cluster of languages, 14. — sources for its study, 34. Sagara language, passim. — its phonetic features, 77. Sagaraland, 82. same, 825. sand, 485. sandy ground, 420. -sano, "j^i*. Sasos, 80, 85. satisfied, to be, 1074. saw, 422. say, to, 834, 1004- loi I, 1079. se, common classifier, 492. se, locative prefix, 5S3. se, nominal prefix, 348. se, pronoun, 639*. se,. auxiliary, 875, sqq. ggt, 992. sea, 470*, 478, .486, 488. sea-cow, 99, see hippopotamus. seasons, p. 287, p. 294. seat, 491* SCO, 503(4). Sebituane, p. 284, p. 293, Se-chwana,- Se-kololo, Se-suto, etc., see Chwana, Kololo, Suto, etc. second, 797. see, to, 52*, 126, 179, 835. seeds, 185. seize, to, 107, 108, 1048, 1073. Sekeletu, 1095. self, 689, 816, 824-825. Semi-Bantu, 598, S30. Semitic languages, distantly related to Bantu, 599. send, to, 52*, 172, 216, 232, 220, 1070. Senegambia languages, related to Bantu, 598. Senna, ^3, no (foot-note). Senna cluster of languages, 14. — sources for its study, 30. Senna language, passim. — its importance, 103. — its partial loss of the classi- fiers £/and LU, 380. — its suppressed nasals and double consonants, 94, 412. sensations, 405. separation of the words, 892 (N. B.). servant, 211, 322*, 363(2). Sesheke, p. 285. seven, 789, 796. shadow, 375. shaft, 478. Shambala cluster of languages, 14. — sources for its study, 43. Shambala language, passim. — its phonetic features, 78. shame, 79, 82, 126. 151, 194, 208, 214, 385*, 409 (4). shape, to, 52*. sheathe, to, 1082. sheep, 390, 401. shield, 192, 202, 404, 500. shoe, 194, .208, 404. Shona country, identified with Siyuna, no. Shona language, 1 10. Slwna tribes, 50 (foot-nnle). short, 187, 500, 502, 601*, 632 (I)- should, 395, sqq. shoulder, 468, (2). shout, to, 175. Shukulumbue nation, p. 286, p. 293. shut, to, io8i. SI = X, 174. SI, classifier, 396, 492. 523. si, nominal prefix, 347. si, pronoun, 639*. si, auxiliary, 875, sq.q., 986, 990, 994. SI copulative prefix, 583. side, 83. side of a river, 421. sin, classifier, 396. singa, sinka, auxiliary, looi. Siongo, ^4, 499, p. 289, p. 295. - Sipopo, p. 284, p. 285, p. 293. Siraf, 5y. sit, to, 52*, 70, 94i-947> 1031- 1033. six, 789, 792*, 796. Siyuna, 7^, no. skeleton, 161. skin, 419, 484. sky, 115, 126, 410*, 420, 439 (7). slave, 143, 418. small, 228, 525, 595, 601*, 632 (I), smali-pox, 455. smoke, 147, 455. snake, 68, 162, 175, 195, 197, 233. 38s*. 427. , snakes, charmed, p. 292. snakes, used as dogs, p. 292, p. 295. snow, 454. iiofala, 73, 8s. Sofala language, 102. soil, 374. Solomon, Ss- some, 828-829. son, 322*, 364, 365 (3), 747, 753, see child, soon, 1018. sorcerers, p. 283, pp. 290-295. soul, 232, 365 (6), 375. sound, 423. sound, to, 1075. soup, 378. SOURCES FOR THE STUDY OF THE BANTU LANGUAGES, 76- 70. Souza, Manuel Antonio de, p. 285, p. 293. soyi, to, 179. speak, to, 113, 1004-1011, 1058, 1079. spear, 80, 410*, 422, 439 (5). spirit, 232, 365 (6). spirit, pernicious, 410*- spit, to, 80. springbock, 203. stand, to, 977, 1016. 334 South-African Bantu Languages. Stanley, i68. star, 40.-5, 525. start,,, to, 527, 965. steal, to, 52*, 139. stick, 1)04*. still, 987, 989, 990, 991. sting, 389. stone, 72, .122, 126, 133, 139, 141 (2), 243, 380, 4io», 414, 415, 420,438,439(7), 485. stool, 491*, 503 (4). stop, to, 977, ioi6. story, 69, 8d. straight, 220, 625, 626, 778. straightway, 971, 1018. stranger, 826. straw, 443. strength, 173. string, 470*, 478, 490 (3). strong, 624, 626. stamp, 491* 503 (6). f«,. classifier, 478. su, nominal prefix, 348. sub-classes ofsubitantives, 315. Subia language, 65. Subia tribe, p. 285, p. 292. subject, how expressed, 638, 644. subjunctive mood and tenses, 832, 8SS-859, 955-958, 969- SUBSTANTIVES, 313-599. SUBSTANTIVE PERSONAL PRO- ' NOUNS, 656-692, 830. SUBSTANTIVE VERB, I0l8, Sqq. such, 498, 628. such, to, 1073. suckle, to, 1073. suffer, to, 1074. suffixes of substantives j S90-595. summer, 499. sun, 136, 141(2), 162,211,239, 243. 410*, 421,430. 438. 439 (2). Sun, Table of the, 81, 82. Sundi tribe, 10. SUPERLATIVES, 632-634, IO79. superlative, quasi-, pronouns. 813. suppose that, supposing that, 786, 988. surpass, to, 631, 1079. Sato language, 169. Suto nation, 365 (5), /o. Swahili cluster of languages, 14. — sources for its study, 42. Swahili language, passim. — its phonetic features, 84-91. swallow, 151. sweep, to, 122. Sylveira, Father Gonjalo da, p. 294. T, how pronounced, 25. — = C, or CJI, 90, 139. — = ff, 93. — = ir, 244. — — L, 220, 232. — = JVD, or NL, 193, 479. — = NJI, 479. — = NT, 192. — = j\rz, 210. — = R, 172, 198, 210, 211, 214. — = S, go, 174, 239. — = TLH, 949. — = Z, 136, 239. ta, auxiliary, 875, sqq., 948, 949. 953- Tabele, see Tebele. table, 422, 500. Table of the Sun, 81,82. tail, 96, 150, 153, 176, 366*, 384 (2). Taita cluster of languages, 14, — sources for its study, 38. Taita language, passim. — its phonetic featiiiias, 80. take, to, 528. tall, 601*. tankhara, 8o. ■tano, 792*. Taprobana, g6. ■tatu, 792. te, auxiliary, loio. tear, tears, 126, 136. Tebele language, see Kafir. Tebele nation, p. 286. tear, 105. teeth, see tooth. teeth, filed, or knocked out, ^2, 50, 209, 210. Teke language, 159-162. Tekeza language, 200. ten, 131, 175, 225, 789, 792*. Tette, see Senna. TU = NT, 192. — i NZ, 195. that, 693-716. Theal, .«= Call, p. 300. thee, 639*, 653, 656*. their, 745, sqq. theirs, 1b%-T]T. them, 639*, 653, 656*. then, 1016, 1018. there, 693, 693*. these, 693-716. they, 637", 639*, 656*. thick, 500, 502. thieves, ordeals for, p. 284. thine, 175, 211, 768-777. thing, 176, 491*, 497, 502, 503 (I), 781 (I), 782 (2). third, 797. this, 693-716. thorn, 525. those, 693-716. those who, 7I8. thou, 637, 639*, 656*. three, 172, 198, 210, 211, 214, 220, 232, 238, 244, 789, 792*. throat, 484. through, 573. thunderclap, 42J. thy, 745. sqq- //, classifier, 512. ti, pronoun, 639*. /I, peculiar verb in Bantu, 596, 834, 1004-1011. ti, negative auxiliary, 875, sqq. tie, to, 1082. tiger, 517. till, 955-959, 97i- till, to, 52* {lima), 1072. time, 504, 787. times, 3, 4, etc., 797. Timneh language, 471. Titles of dignity, 356. TL, how pronounced, 32, 33, — = NDL = NZ, 195, 208. — = TLH, 20%. — its use in Kafir,- 121. TLH, how pronounced, 32. -=S=NS = T, 174. 949- tla, llha, auxiliary, 948^ 955, 9S8. Tlhaping language, 169. to, 563, 1065. tobacco, 632. together, 533*, 541. S8i (f^), 810. to-morrow, 142, 421, S33*, S8i (8). Tonga cluster of languages, 14. — sources for its study, sg. Tonga language, passim. — ■ its phonetic features, 62-64. — taken as standard, 2. Tonga tribes, 2g (foot-note), 50, pp. 288-295. tongue, 133,143, 145, 243,470*, 478, 484, 488, 490(1), 583. tools, 372. tooth, teeth, 126, 143, 145, 205, 233, 235, 243, 410*, 419. tortoise, 205, 418, 500. touch, to, 1084. town, 136, 142, 147,229. tr.ide, to, 1084. Trade in South-Africa, 79-99. transient actions, 832. transitory tenses, 971, 972, 974. tree, 153, 164, 206,210, 211, 366*, 371.425, 526. tribal names, 322*, 376. tripe, p. 313. Troglodytes, 3. Troglodytica, So. true, 780. TS = C, 2o5. — = DJ, 205. — = K, 1037. — = N/, 191. — = NZ, 195. — = 3- after iV, 288. TSH= pu, or PI, 222. — = RE, 206. — = S, 174. Tshagga cluster of languages, ^.#. — sources for its study, 6^. Tshagga language, 211. tsho, peculiar verb, 834. TU= TSH, 144. tu, connective pronoun, 639*. hi, relative particle, 718*- Tua people, 3. tusk, 419, 484. /a-K^a,- nominal prefix, 515. twice, 504* (a second time), 797. two, 131, 14s, 217, 233, 7S9, 792*. Alphabetical Index. 335 TY = BU, oxBI, 122, 445. •tye, nominal suffix, 5x9, 593. tyi, classifier, 492. tyi, pronoun, 639*. tyi, conjunction, 787. tyi^, 801. iyu, pronoun, 639. TYW= BU, 122. U, how pronounced, 26. ~ = A, 273. — = /, 276. — = MU, 279, 560. — = 0, 200. — = UA, 332. — = UE, or WE, 265. — = UI, 267. — = W,2(>i. — elided, 264. — how changed before vowels, 261-268. ^- Suppressed, 278. «, article, 317. », common classifier, 329, 367, 448; 465, 490, 509. u, locative classifier, 543. «, connective pronoun, 639*. «, relative particle, 718*. UA = OA, 262. ■ua, passive suffix, 1047, 1053, 1062. uia, 1007. Z/E = 0, 265, 659. ■tte, pronominal suffix, 660. -tiAtiy verbal suffix, 1050, 1083. 22 /^»jiz, 982, 1007. u kuii, 1007. « liuze,- 959. -ula,-Ulula,-ulaka, una, -ununa, verbal suffixes, 1057, 1080- 1083. un, pronoun, 639*. unbind, to, 1082, 1083. undo, to, 1083. undress, to, 108 1. Ungu cluster of languages, 14, — sources for its study, 33. unite, to, 1082. unproductive aninials, 418. unsheathe, to, 1082. untie, to, 1082.- until, 9SS-9S9, 1016. up, 562. , Upa?, 8s. upon, S33*.S4i, 581 (4)- lis, 653, 639*, 656*. V, ^, how pronounced, 27. K = jSK,94,99. — = A", 233. — = P, 137, 148, 180, 211, 215- — = 2,91,63. — = suppressed, 66. va, common classifier, 338. va, locative classifier, 537- va, pronpun, 639*. Vasco de Gama, gs. ve, pronoun, 639*. vein, 402. verbal suffixes, 1047-1084. Verbs, 831-1085. very, 633, 6S9, 824. very far, 533*. vi, classifier, 369, 494, 496, 510, 522. vi, pronoun, 639*. OT.?, 801. Victoria Falls, 74, 499, p. 289, P- 295- village, 136, 142, 147, see town. Viti cluster of languages, 14. — sources for its study, 31. Viti. tribes, gs- vo, auliiliary, 1012, 1015. VOWELS, changed, 559. — elided, 49, 76. — indifferent, 833. — remarkably ■ firm in Bantu, 48. — weakened, 235, 892, 904, 905. VOWEL-STEMS, or stems which begin w,ith a vowel, 46 (3). — their peculiar laws, 46 (2), 67, 87, 113, 415, 478, 611, 648, 832, .837-841, 843, 851. VU, meaning of the element, 409(2). vulture, 358. Vumbe tribe, p. 286. W, how pronounced, 28. — = K, 175. , — = P, 272. — = U, 28. ■ — euphonic, 295. •w, classifier, 448, 465, ,490. •w, pronoun, 642. wa, common classifier, 336. wa, locative classifier, 536. wa, pronoun, 639* -wa, passive suffix, 1047, 1053, ;io62. Wakwak, fz, 84, 93, p. 312. walk, to, 52*, 1075. wallets, p. 295. Wange, the chief, pp. 286-288, p. 294. "Wangwana, 246. Wankie, see Wange. wa?h, to, 174, 208, 1082. watch over, to, 193. watching over gardens, p. 305. water, 89, 115, 133, 136, 143, 195, 210, 211, 440*, 454, 461(10). watercourse, 470*, 490(2). we, pronoun, 639*, 656*. we, 128, 214, 637, 639*, 656*. well, 499. what ?, 801-803. whatever, 813. when, 784, 785, 787, 788, 902, 970, 993, 1008-1012. when?, 808. whence?, 533*, ■581(11). whenever, 993. where?, 533*, S8i(ii)- where, wherein, 785. which, 717-746. which ?, 8o6-iio7. while, 787, 788, 970, 988, 989. white, 624, 628, 778. white man, 355, 365(0,432. whither?, 533*, 581(11). who, 717-746. who ?, 804. whoever, 813. whole, 601*, 810-812^, whom, 717-746. whose?, 740-741, 805. wi, pronoun, 639*- wicker-work among Kafirs, p. 321. wide, 624. wife, 173) 322*, see woman. wild beast, 502. wind, 69, 73, 77, 79, 137, 142, . I49> '84, 405. window, 503(6). wine, 440*, 455, 461(3). wing, 419. winter, 137, 142, 405, see wind. wish, to, 126, 995, 1003. with, 570, 573, 575. within, 194, 533.533*. 58i(i)- wizard, 143, see sorcerer. Wolof language, 830. woman, 142, 143, 211, 239, 243, 322*, 365(4). wool, 139, 440*. work, 379. would, 995, sqq. wound, 423. write, to, 1070. ' writing, 487. writing ip South-Africa, 16, ly. written language, when differ- ing from the spoken, 253. wu, classifier, 447. wu, pronoun, 639*. X, how pronounced, 29, 38. — = NS, 194. — = NZ, 195. — = SI, 174. Xosa language, passim, see Kafir. , — its peculiar sounds, 33-38. — its phonetic features, 120- 124. Y, how pronounced, 30. — = Z, 96, 97. — = M, or N, 295. y, classifier, 496. y, pronoun, 639, 642. ya, pronoun, 639*. ya, auxiliary, 911-917, 920. ya, preposition, 57°- -ya, demonstrative suffix, 698. Yansi cluster of languages, 14. Yansi language, 159-162. Yao language, 14. — sources for its study, ^7. — its phonetic fea.tures, 66-72. — its peculiar plurals, 354. — its relation to Chwana, 72. 336 South-African Bantu Languages. ye, auxiliary, ^i6, 917. year, 366*, 379, 384(11). yesterday, 421, 533*, '581(7). .yet, 991. Yeye language, 109. yi, pronoun, 639*. ju,, suffixed to verbs in relative clauses, 723, 734. yonder, 693, 693*. you, 637, 6,30*, 656*. young, 601 . young otiinimal, 418. your, 745, sqq. yours, 768-777. -c. youth, 504*. yo = ye, 916. ^«, pronoun, 639*, 652. Z, how pronounced, 31. — = BZ = BV, 99, — = /), or R, 173, 220. — = Z), or T, 136. — = DY, 123, 214. — = DZ, 94, 99. — =/, 89, 106, us, 143.147- — = Z, 165, 173 (cf. 9), 211, 214, 232. V = ^, 232. ■i-= T, 239. — = V, 91, 115. — = F, 126, 225. — suppressed, 66, 81. 5, how pronounced, 31. — its use, 105, 123, 133. jfl, classifier, 490. ia, auxiliary, 948, 950. -za, verbal suffix, 1079. Zambezi, 486. Zaramo language, 14. — sources for its study, 45. ze, auxiliary, 955. zebra, 133, 136,401. Zend], see Zindj; zi, classifier, 491-503. . — its transformations, 496. — its use, 497^501. — its original meaning, 502. zi, connective pronoun, 639*. zi, reflexive pronoun, 655. zi, relative particle, 718*. zi, copulative prefix, 583. ■si, suffix in the number ' ' one ", 792. -zia, verbal sulfix, 1079. Zimba, Zimbawe, Zimbabye, Zimbaze, p. Zimbabye, or Zimbaze, ruins. — — inscriptions found on them, 18. — their description, 7j. — their origin, 73-74. zin, classifier, 385-409. — its transformations, 391-398. — its use, 399-406, 477-480. — its original meaning, 408. Zindj, who they are, 8. — their ancient trade, 80. — their origin, 72. zon, classifier, 397. zo-o = za-ku, 950. zu, meaning of the element, 409(2). Zulu language, see Kafir. — its difference from Xosa, 124. — its peculiar sounds, 33-38. Zulu tribes, origin of the, oj", 365(5)- Zumbo language, 102. \wi, classifier, 496. JOTJ, pronoun, 639*. H. ffi. ». G. DescMe, De Brouwer, and C°., Bruges, Belgium. ,^ J' ^ J* V^ti^' '>> -^ i^ "^ 4^! ^ tfl* -. ^ ■>> , A 5 ^ 4?! ~*^" 3V ^-^