DOel'MFkT RF 5i'MF ED 023 065 AL 001 364 By -Closs, Elizabeth, And Others Some Copula Patterns in Standard Swahili. Pub Date Mar 67 Note -23p. Available from-Institute of Swahili Research, P 0 . Box 9184, Dar es Salaam, Tanzania. Journal Cit -Swahili: Journal of the Institute of Swahili Research; v37 n1 Mar 1967 EDRS Price MF -$025 HC -S125 Descriptors -Deep Structure, *Grammar, Intonation, Language Patterns, (Languages).Standard Spoken Usage, Structural Analysis, Surface Structure, *Swahili. Syntax, Verbs Identifiers -*Copula Patterns Thisis an outline of the mator characteristicsof sentences like"Ali alikuwa mwalimu" (Ali was a teacher). "Ali ni mwalimu" (Ali is a teacher), and the numerouspattern variations demonstrated by such sentences. Constructionsof this kind are traditionally called copula constructions, and have been classified in terms of (1)predication with a verb, and (2) predication without a verb. The purpose of this study isprimarily to investigate whether the elements mentioned under (Z) are infact sufficient to account for copula constructions without the morpheme "--w-," and whatcombinations of elements are possible. The second part of this outline discusses the structureof and restrictions on expressions in copula sentences.The third part deals with word

order, and the fourth with intonation patterns.(AMM) aridrigalliaiirmorldriamarirmilrirhmamalmesslamiabilltilialowill

March, 1967 ol. 37/1

SWAHILI

JOURNAL OF THE . INSTITUTE OF SWAHILIRESEARCH University College Dar es Salaam

Jarida la Chuo .cha Uchunguzi waLugha ya Kiswahili

Editor: W. H.WHITELEY

U.S. DEPARTMENT Of HEALTH, EDUCATION aWELFARE

OFFICE Of EDUCATION

THIS DOCUMENT HAS BEEN REPRODUCED EXACTLYAS RECEIVED FRON THE

PERSON OR ORGANIZATION ORIGINATING IT.POINTS Of VIEW OR OPINIONS

'STATED DO NOT NECESSARILY REPPISENT OFFICIAL OFFICE OFEDUCATION

POSITION OR POUCY.

Shs. 12/50

L" I;)0 14" ev Published by the Institute ofSwahili Research, P.O. Box 9184,Dar es Salaam co and Printed by Printpak Tanzania Ltd. P.O.Box 9033, Dar es Salaam.

La

ATI001364 1

,

i NOTE ON CONTRIBUTIONS This Journal is publishedtwice a year in March andSeptember. Contributions in general exceed 15,000words, may be written in for any issue, which should not 15th Swahili, English, French orGerman, and should besubmitted not later than November or 15th May in any year,and should be addressed to: The Editor, Institute of Swahili Research, University College, P.O. Box 9184, . Dar es Salaam, Tanzania..

SUBSCRIPTIONS We should be grateful ifsubscribers would do their utmostto insure that their subscriptions are paid duringthe first three months of anycalendar year directly through the office of the Institute. Subscriptions should be sent to: The Secretary, Institute of Swahili Research, P.O. Box 9184, Dar es Salaam, Tanzania.

C. WWI, If 4014.4...4 4D... C.A. ...v..

SOME COPULA PATTERNS IN STANDARD SWAHILI . by Elizabeth Closs, Ali Kondo and Salm Mbaye The following' is an outline of the major characteristics of sentences likeAli anima mwalimu 'Ali was a teacher', Ali ni mwalintu 'Ali is a teacher',and the nume- rous pattern variationsdemonstrated by such sentences. Constructions of this kind are traditionally calledcopula constructions. Mrs. Ashton2 among others has classified them in terms of (a) 'predication with a verb' ; in these sentences -w-'be' is present as in the first quotationabove; and (b) 'predication without a verb' ; no -w- is present, but there is an optional 'invariable copula ni' (Ali ni mwalimu), or a free verbprefix (All yu mwalimu) or a copula ndi- requiring the 'enclitic -0 of reference' (Alindiye mwalimu). The purpose of this study is primarily to investigate whether theelements mentioned under (b) are in fact sufficient to account for copula constructions without -w-, and what combinationsof elements are possible. The second part will go on to discuss the structure of and restrictions on nominal expressions in copula sentences; the third part will be devoted to word order, and the fourth to intonation patterns. It is hoped that the material will be useful to teachers of Swahili as well as tolinguists and that it will stimulate further research into copula constructions. The functional relations between all the sentences under discussion is the same: SubjectCopula PredicatePredicate Nominal. This set of functions is expressed, or 'realized', by several different structuralcategories, and it is these categories and their combinations that must be investigated. They include: (a) nominal expressions (symbolized by N(oun) P(hrase); NP in tura includes the categories of N and various modifiers like Demonstrative; (b) copula expressions (Cop); for the present stage of the discussion, Cop has as its members an optional -w-, andoptional ni, ndl-, Verb (Vpf);3 (c) adverbial expressions (Adv), which include subcategories such as negative adverb (Neg). Other classes will be introduced during the course of discussion. A distinction is made between 'underlying' and 'surface' structures. The difference here is between basic patterns and overtly observable patterns; in some casesunderly- ing and surface structures are different ; e. g. Chiriku huyu ni wangu 'This finch is mine', can be considered as the patternunderlying Wangu ni chiriku huyu 'Mine, that's what this finch is' since the relationship between the items remains the same,although the emphasis is different. On the other hand, Dar es Salaam ni mjimkubwa 'Dar es Salaam is a/the large city' is not considered the base from which Mji mkubwa ni Dar es Salaam 'The big city is Dar es Salaam' is derived. This is because in thefirst sentence mfi can be specific or non-specific; in the second sentencemji can only be specific, and the relationship between the two sentences is therefore not the same. Furthermore the first sentence, as a complete utterance, can occur discourse-initially as well as discourse-medially, while thesecond is limited to discourse-medial position. Both sentences are therefore realizations of different underlying structures,although they happen to share the same lexical items in different order; here underlying and surfice structure coincide. For ease of reference it will sometimes be convenient to use subscript numerals for the two NPs in the patterns under discussion. These numerals will refer to underlying structures unless otherwise specified. In any study .which is concerned with the productivity of language andwith accounting not only for sentences in the data but also for other possible sentences with the same structure, it is necessary to consider impossible as well as possible construc- tions. Only in this way can the limitations of grammatical versus ungrammatical sentences be fully stated. Ungrammatical sentences are preceded by*; those of mar- ginal acceptability by ?.

5

AL 001364 1

simplify description.Parentheses Two other notationaldevices have been used to optional, e.g. N PI (Neg)(Cop) N P2 indicate that the item oritems enclosed in them are negative adverb, followedby optional reads: first phrase,followed by optional followed by second nounphrase. { } indicates thatin any specific copula expression, but not more than one; sentence any oneof the elements inbrackets will be chosen, e.g. Int1 NP reads : either ni orVpf, followed by nounphrase. 1Vpf f PART I: THESTRUCTURE OF 'COPULAEXPRESSIONS' OF - W-: PRESENCE OR ABSENCE realiza- Most sentences of the typeNP1 (Neg) (Cop) NP2permit -w- as one of the absence of -w- is characteristiconly of certain constructions tions of Cop. Obligatory certain lexical where a semanticrelationship of member toclass pertains between realizations of NPI and NP2.4 with any other memberof Cop, although this is rare.It may -W-- may co-occur Swahili ( to the usualrestric- also co-occur with alltense-markers available in of clauses etc.). Tense + -w-contrasts with absenceof -w-; the tions on sequences + -w- and absenceof following illustrate the contrastbetween three 'present tenses' -w-:$ sailor' (he has just started;he 1. Baba awa baharla'Father is becoming a intends to become asailor). sailor' (and started WMtime 2. Baba anakuwa baharla'Father is becoming a back; he is improving). 3.Babahuwabaharla 'Father is at times asailor.' 4.Baba baharla 'Fatheris a sailor.' dictated by Adverbs and clause- Certain restrictions onthe choice of tense are external factors only. Forexample, while sequence. Others areapparently dictated by dada huwa mpishl 'Whenvisitors come, sister is it is possible to say3, or Wager!! wk./a contextually incongruous to say the cook (instead of theusual one)', it is considered times Juma'. Similarly,while Harukihuwa mwana- *Baba huwaJuma 'My father is at the following is not permit- is at times a man' canbe used (abusively), mume liamisi husband'. These restrictions ted: *Hamisl huwamume wangu'Hamisi is at times my are semantic, notgrammatical. before main verb -w-, asbefore any An auxiliary -w- ispermissible, though rare, other main verb; becoming a good child'. 5.Allakawa anakuwamtoto mzuri 'Ali was Tasili had become a translator'. 6.Twillakawaamekuwamkallmanl + main verbconstructions, Neg may combinewith either auxi- As in other auxiliary be identical: liary or main verb. Ineither case the meaning appears to mzurl 'Ali was notbecoming a good boy' (e.g. 7. All hakuwa anakuwa mtoto despite efforts to makehim so). 8.All alikuwahowlmtoto mzurl Ibid.

PATTERNS WITH NI: but permissible in allcopula cons- The formative ni(negative si) is optional, tructions.' Its functionis specification of thecopula relationship, sailor). 9.Baba ni baharla 'Fatheris a sailor' (precisely a 6

( talkative person'. 10.&mist anakuwa ni msemaji'Hamisi is becoming (exactly) a Ali alikuwa ni mwalimu'Ali was (exactly) a teacher'. the necklace was just 12. Kutaja kidani kullkuwa nibahatl nasibu 'Mentioning a gamble'. all co-occur,Neg is If -w- andnico-occur,nialways follows; whereNeg, -w-,andni supported by either -w- ornl,apparently without inmeaning: 13. Baba allkuwa :1 baharla'Father was not a sailor'. 14.Baba hakuwa ni bahariaIbid. andni,may not alwaysbe in close proximity to each The two formatives, -w- position za- other; for example, anAdvmay be shiftedfrom its usual position after NP2 to .ons between -w- andnifor stylistic effect: tied there was nothing 15a.Ulikuwa ni usingizi siku hiyo'It was just sleep that day' (i.e. but sleep that day). may 'It was that day just sleep.' ric b.Ulikuwa siku hiyo ni usingizi the of PATTERNS WITH VERBPREFIX: This element is not used byall speakers of StandardSwahili except in fixed aphorisms or expressions likeYutayart'He/she is ready'. In a randomsample of educated Coastal Swahilispeakers, mostly university students,two out of twelve Vpf in productive constructions ;$ someothers felt it was .rejected the use of any kind of archaic or `substandard', butused it in their own speech. Itis very rarely found in the written language. In Vpf is mutually exclusivewithni,and likealspecifies copula relationship. Addition, it implies the relationshiphas persisted over some periodof time and tMt it will continue to do so. It isoften used to express 'X is Y byprofession'. for some time be) a sailor' ISe 16.Baba yubaharia'Father is (has been, and will (by profession). a is 'Cultivation is (always has been and will be) hardwork'. say 17. Kuchbnba ku kazi is (always) a 18.Mtambua ndwele yu mganga'One who diagnoses an illness doctor'. Contrast withnican clearly be seenin: husband' (can be used 19. Yeye yumume wangu'He is (has been and will be) my only by someone stillmarried). used by someone married, or 20. Yeye atmume wangu`He is my husband' (can be by a widow, e.g. in pointing to apicture). andWewe Often, however, the Vpf/ntdistinction is neutralized.Wewe at mtoto ;nbaya 'You are a naughty child', forexample, alternate freely exceptwhen u mtoto mbaya if so, only at is the speaker wishes to imply.thechild is bad only in this one instance; e. possible, but even then the sentenceWewe ni moto mbayais ambiguous. Occasionally Vpf may co-occurwith -w-, but this constructionis primarily through limited to sentences with adverbbado'still'. Then Vpf implies continuation the period referred to: `Last year Fatuma was still 21.Mwaka Jana Fatuma alikuwa yumwanafunzi bado a student'. 22.John alikuwa yu mume wangubado'John was still my husband'. dr. 7 C

gilikaaliftiirdsildwililiiiiiiiiiiiii

When -w- and Vpfco-occur, Vpfmustfollow .(note the parrallel withni). Neg may also be supported by either -w- orVpf; Neg -I- -w- + Vpf sequences are,however, extremely rare ; K in fact does notaccept them : yet' (by profession). 23. ?Baba hakuwa yu bahariabado`My father was not a sailor 24. ?Sisi tulikuwa hatu bahariabado 'We were not sailors yet'. The negative form of Vpf isactually very uncommon in anysentence except in constructions like: 25.Hatu wapishi 'We are not cooks'. 26.Ham wanafunzi 'You are notstudents'. The third person singularnegative form : 27.Hayu mpishi 'He is not acook'. is considered archaic, and thethird person hawa is notaccepted at all, probably because it could be ambiguouswith Demonstrative Pronoun +NP2 (e.g. Hawa walimu 'These are teachers', *Hawawalimu 'They are not teachers'). Vpf may not co-occur withndio to be discussed below.Further restrictions on depend on certain NPI . . NP2 co-occurrences(q.v. Part II, 87), Vpf if it is used at all, is not or are imposed byidiolect, style, and the semanticdomain of 'continuity'. Vpf in identification copulas(those with restricted to first and second persons allow Tu wapishl Nominal) as Mrs. Ashton suggests, exceptin the negative. She would 'We are cooks', U mpishi`You are acook', but not Yu mpishl 'He is acook' ;9 the latter constructions are, however,acceptable. Vpf is also not and several other similar shape as the prefix restricted to those verb prefixeswhich have the same morphological of the nominal with which they arein concord, as Loogman suggests;m if dialectally acceptable at all, all forms ofYpf are permitted, other thanthe negatives discussed above, and any Vpfin concordwith the plural of the mlwaandil/ma classes (since they are ambiguous withthe genitive formants in concordwith these classes).

1 PATTERNS WITH NDI-INDIYOINDIO: . Textbooks on Swahili usuallydiscuss a copulative ndi- (negativeJO which takes however, mention two otherforinatives which also thi '--0 of reference' ; they do not, specifier, and occur in NP1. .NP2 position: ndlo (negativesio), which is another ndiyoi (negative siyo). The latteris not to be confused withthe well-known homony- ndlyo, (negative slyo)ltis(not) so'. The three formants appear mous sentence Adverb characteristics of occurrence in to be morphologicallyrelated; they share the syntactic mutually exclusive. They must,however, be treated as copula sentences, and they are different separate elements sincethey have contrastingfunctions and have niarkedly co-occurrence restrictions: ambiguous with 28.Mtoto ndlye khunbe 'Thechild (referred to) is a creature' ; 'The child is the creature(referred to)'.1I creature' (used almost ex- 29.Mtoto ndlyo kiumbe 'Achild is (as you say) a clusively with high stress andhigh intonation on ndlyo, q.v.Part IV, IP II :x; limited to statements). (usually without stress 30.Mtoto ndio kiumbe 'Achild is (defined as) a creature' on ndio; used inquestions and statements). (A) NDI-: Mrs. Ashton says of thisformative: by the appropriate concord,forms a pronominal The -0 of reference, preceded Class. or adverbialcomplement of the copula exceptin the Singular Personal Note the replacement of ni byndl-, which takes the word stress.(ni as copula is never stressed).

8 :

,,,,,,,,...... ,...... 6...a...... ".. ..4.4-40.4"11Ma.... oft.,,...... ,2=0...... ,L, A10.1.114181114"11...._...... 1ftebyallei...... 44,.....,,...

These forms are often spoken of as the"Emphatic" forms of the copula ni- may the nominal, pronominal or adverbial :ver, (sic). They are used to define more sharply antecedent to which they refer.12 on). The following points may be noted: (a) Ndi - does not replace ni, it occurs withit (even if only rarely). (b) Ni is frequently stressed (cf. its usewith IP II :x, Part IV). (c) While -0 may function as apromninal or adverbial , this is by no means the onlyfunction available to it. It may form a subject pronominal as in Ndiye ni msichana'That one (referred to) is a girl'; or it may be used in a limited appositional construction, as in 28. Furthermore,it is not restricted to copula constructions, althoughit is most frequently found there: MN ndiye allkwenda sokoni 'That man (referred to) went tothe market'. Ibly When ndi- ± -0 is used as an optionallimiter of a nominal it may occur only Iwo once in any onemain clause pattern. We therefore find either: NP1 (Neg) (ndi- + -0) (...-w-...) ifniIANP2 on k 1 Vpf f I where -0 is preceded by concord withNP1 or: not :ate NP1 (Neg) (...--w-...)({4if})(ndi- + -0) NP2 .ishi :ter where -0 is preceded by concordwith NP2.13 In either case the nominal expression 3ot with which nth- + -0 is in concord may notbe expanded by the modifier Demonstra- :fix tive: Illy ied 31. *Mtu huyu ndiye ni mganga 'That manreferred to is a doctor'. ley but: 32. Mtu huyu ni ndlye mganga 'That manis the doctor referred to'.

wherendi- + -0 limits NP2. , 4s Sentences with ...-w-...ndi-+ -0 -I- NP2are quite comMon.On the other hand .'.so the sequence ...ni + ndi-- -I- -0 + NP2is only marginally acceptable in simple .nd sentences.; but it is fullyacceptable if a relative clause modifies NP2: ly. is a matth'. :ar 33. Kipawachake nacho !Wald (a) 'His talent referred to 1n (b) 'His talent is the match referredto'. as referred to is/(was) :nt 34. lapowachoke ndicho (kilikuwa) ni kifani 'His talent exactly a match'. the match referred Co'. h 35. lapowachake killkuwa ndicho kifani 'His talent was tukiklzungumza 'A/the child is exactly the 36. Mtoto ni ndicho kiumbe tulichokuwa : . creature we were talkingabout'. Intonational grs.wpingis the same whetherndi- + -0 is in concord with NP1 or NP2; i.e. ndi-.. .occurs in theintonational group of the Cop. In this andits failure to co-occur with itself in asimple sentence, and in its ability to supportNeg, it contrasts with Demonstrative and other expansionsof NP (these participate in the NP intonotional group): 37. Mvulana huyu mtoto wangu 'This boyis my child'.

1 38. Mvulana ndlye mtoto wangu The boyreferred to is my child'. On the other hand, in its association with thepronominal -0 of reference and in its : ability to be directly dominated by NP inSubject, Predicate Nominal and Adverb functions, it is unlike the specifying formants of Cop.Some speakers associate certain _ 9 l -0 with ni + appositivePersonal Pronoun. E.g.Chiriku ndiye occurrences of ndi- + Chlriku ni yeye ndege The finch referred tois a bird', is thoughtof as connected with it referred to, is abird'; Alt hakuwa ndiyempishl wake ? ndege 'A finch, specifically mpishi wake ? of his ?' is connectedwith All hakuwa ni yeye Was Ali not that cook not his cook ?' Was Ali, specifically hewhom we have referred to, of Swahili have not At this stage, when thepronominal and adverbial systems analysis can be Oven.Pending further been adequatelyinvestigated, only a tentative has study, ndl- + -0 can beconsidered as an appositionalreferential pronoun which strict restrictions on uooccurrencewith nominal modifiers.

(B) NDI Y02: Another formative which is partof the referential systemof Swahili discourse in ndlyo2. It occurs optionallyin strings of the type: NP1 (Nes) (ndiy02) (. - -w-- .)({Vpf})NP2

(q.v. Part IV). WhenNeg is where it is mostly accompaniedby IP II:x or IP II :ii:a (cf. 41). present it is alwaysattached to ndlyo2, not to -w-,nt, or Vpf. negate the (Cop)NP2 Ndlyo2 is used in answers totruth questions to affirm or Very rarely it may beused to affirm or negateVP, also proposition of the question. repeated in the answer in answer to atruth question (cf. 44).NPI is frequently not (cf. 42), or the word orderof the answer may beshifted (cf. 43). 39 Q.Salmon nt mwanafunzi? 'IsSaimon a student?' stuthnt'.(i.e. Salmon, A.(a) Salmon ndiyo ntmwanqfunzi 'Salmon is a certainly = a student). student' (i.e. Saimon, (b) Salmon styo ntmwanafunzt 'Saimon is pot a certainly not mc a student).

40 Q.Salmon st mwanafunzt? 'IsSaimon not a student?' student' (i.e. Salmon, A.(a) Salmon ndlyo n1mwanafunzt 'Salmon is a certainly vist student). (b) Salmon siyo nimwanafunzt 'Salmon is not astudent' (i.e. Salmon, certainly not a student). Baba hakuwa nt baharla? 41 Q.Baba allkuwa si baharta?freely alternating with

Was father not a sailor?' , sailor'. A.(a) Baba ndlyo altkuwa ntbaharta 'Father was a (b) Baba styo altkuwani baharta 'Father wasnot a sailor'. (N.B. no alternation with*Baba ndlyohakuwa nlbaharla or *Babanellyo alIkuwa .slbaharta). match?' 42 Q.Ktpawa chake kIlIkuwakifant? Was his talent a A.Ndlyo kiltkuwa ktfant 'It was amatch'. 43 Q.Klpawa chake Want? 'Ishis talent a match?' talent'. A.Ndiyo Want klpawachake 'It was a match, his the market? (i.e. Is it 44 Q.Salmon ndlyo alikwendasokont? 'Did Saimon go to true: weat to themarket?).

10

williipswes.11111111.11,/}1111,41.1 i I 1 A.Ndiyo alikwenda sokoni 'He certainly went to the market'. The yo of this formant may be consideredpronominal. The question arises whether ndiyo2 should be considered aspecialized form of the appositive mu + 0 discussed above, or whether it should betreated separately. In the interests of simplicity of description and explanation of the partial similarityin function and behaviour between ndiyo2 and the Sentence Adverb ndiyo, to bediscussed below, ndiyo2 is here classified together with Neg as a Preverbal Adverb. Assuch, its domain is (Cop) NP2 and occasionally the verb phrase. Decision as to the statusof yo is left for further investi- gation of pronominals. As a formative of the referential system, ndiyo2 has adouble function. In the answer, ndlyo2affirms Predicate relationship, or (supported by Neg) negatesit. But whether the formative affirms or negates the propositionof the previous question depends on whether the question did or did notcontain Neg. Ndlyo2 in A. affirms (Cop) NP2 in Q. if Q. was affirmative; it negates itif Q. was negative. Conversely, Neg ndlyo2 in A. negates (Cop) NP2 in Q. only if Q. was affirmative;it affirms it if Q. was negative.Table I is provided as an abstraction of 39-41 to 'demonstratethe gross features ofndiyo2 structures discussed in this section :

39 Q. X ni Y A.a. X ndiyo ni Y (affirms Q.) b. XNeg ndiyo ni Y (negates Q.) 40 Q. XNeg ni Y A.a. X ndiyo ni Y (negates Q.) b. XNes ndiyo ni Y (affirms Q.) 41 Q. XNeg w ni A.a. X ndiyo w ni Y (negates Q.) b. XNeg ndiyo w ni Y (affirms Q.) TABLE I: ND1102 CONSTRUCTIONS Ndiyo2 contrasts with the Sentence Adverb (Neg) ndlyo, which occurs sentence- initially in statements and sentence-finally in questions. This. Sentence Adverbhas its own mid-level intonation group, separated by comma pause from the restof the. utterance (contrast ndlyo2 in 39-44 which occurs in the sameintonation pattern as the following NP). This Sentence Adverb is also part of the referential system,but has as its domain the whole previous sentence; hence (Neg) ndlyo, may co-occurwith e.g. ... (Neg) nl.. . Ndlyo, is not used primarily in copula constructions; it imposes no context- restrictions on the sentence with which it is in construction. 45 Q.Kodl nl muhlmu ? 'Is tax a "must"?' 4 A.(a) Ndlyo, kodl ni =him 'That is the case, tax is a "must" '. (b) Slyo, kodl nl muhlmu 'That is not the case, tax is not a "must"'. 46 Q.Kodl sl muhimu ? 'Is tax not essential ?' A.(a) Ndlyo, sl muhlmu 'That is the case, it is not essential'. (b) Slyo, nlmuhlmu 'That is not the case, it is essential'. 47 Q.Salmon si mwanafund tullyemzungumza ? 'Is Saimon not the student we were talking aLout ?'

11

vo. :

t

indeed the student'. A.Siyo, ndiyo mwancifunzi'That is not so, he is which it may be added as a tag, It is interesting to notethat in truth questions, to and a preceding affirmativesentence is often taken tobe a combinationofndiyoi statement in English: sarcastic, as is an affirmative tagquestion after an affirmative is that the case?'(dubious, 48.Mume stara, ndiyo?'A husband is a protection, sarcastic, cf. EnglishA husband is a protection,is he?). is that not the case?' 49. Mume stara, siyo ?'A husband is a protection

(C)NDIO: Like them it is a specifier ndiois mutually exclusive withniandVpf. The formative unlike them it may occurbefore of copula relationship(specifically definition); but or after -w-. Most commonlyndiois used in answers to contentquestions: SO Q.Mtoto ni nini?'What is a child?' A.(a)Mtoto ndio kiumbe'A child is defined as acreature'. (b)Mtoto ni kiumbe'A child is a creature'. 51 Q.Chiriku ni nini?'What is a finch?' A.Chiriku ndlo ndege'A finch is defined as abird'. kiumbe,or is a SOa expresses definition,while 50b impliesmtotois an example of member of the classkiumbe. enquire about definition. The construction may alsobe used in truth questions to "What is "fish"?' 52.Samakindio nini?'What does "fish" mean ?', or 53.Kuiba ndio nini?'What is the meaning of"stealing"?' In both instancesNegis very rarely found. truth questions and in A related usage is incorrections, negative responses to negative is frequently used.In a situation where someonestates echo questions. Here else asks whether it is one,the thatJiwe ni kiumbe'A stone is a living creature' or correction or answer mightbe: living creature'. 54.Jiwe do kiumbe'A stone is not (defined as) a An incredulous echomight be: creature?!' SS.Jiwe ndio kiumbe ?'A stone is defined as a before or after Whenndioco-occurs with -w-it may, as mentioned above, occur it, apparently with nodifference of meaning: Chiriku ndio hakuwa jiwe 56.Chiriku hakuwa ndio jiwein free variation with 'The bird was not defined as astone'.

ABSENCE OF COPULA: ni, Ypf, ndio,is optional for all Cop, withits members --w- andthe specifiers sentences havingcopula function, with threerestrictions: Preverbal Adverb, then atleast one member (a) IfNeg isthe only realization of ofCopmust be present to supportit. ndiyo2is also not, the construction maynot (b) WhenCopis not chosen, and This is undergo a word order shiftsuch thatNP1 + Pred> Pred +NP1. NP1would be ambigubus withNP1 + NP2. presumably because NP2 + be shifted to Therefore,while Baba ni baharia'My father is a sailor' may

12 ; NI baharla baba ' A sailor, that's what myfather is', Baba haharla may not be shifted to Baharla baba. The latter utterance occurs,but only meaning 'The tag, sailor is my father', never 'A sailor, that's what myfather is' (q.v. Part III). to be :fish: (c) When Cop is not selected, then the onlyintonation patterns available are those with a break between NP1 and NP2 (q.v. PartIV, Types II:i and II:ii). ions, Because of this restriction there is noambiguity in the spoken language between sentences and phrases. In writing, however,such an ambiguity does arise, cf. 57.411 mwallmu (a) 'Ali is a teacher'. (b) 'Ali, the teacher'. sifier vfore It is particularly common when Predicate Adjectiverather than Predicate Nominal is present : 58.Mtoto mehafu (a) 'The child is dirty'. (b) 'The dirty child'. In writing, therefore, copula sentences without Cop areavoided.

CONCORD: The generitl rule for concord is that main verbs, auxiliaries,verb , modi- fiers and limiters of NP all agree (if present) in number andclass membership with the noun realizing NP1 (the Subject). Limiters andmodifiers of NP2 are the only items which show concord with NP2. is a Sentences which show concord between Cop and NPI arestatistically actually rather few, since ni alone is by far the most commonrealization of Cop. Among those sentences investigated in which concord was markedin Cop, a few were found which countered the general rule outlined above; in these concord wasdictated by NP2. For example, beside the expected: $9 a. KIlimo hOd kilikuwa kazi 'This digging was work'. there is.also : id in tes b. nitno hikl Ilikuwa kazi 'This digging was w6rk'. the with concord between Verb Prefix and NP2. Concord withNP1 in (a) emphasizes NPi, concord with NP2 in (b) emphasizes NP2. Furthermore, Mfeels that concord with NP1 implies the speaker assumes the predicate relationshipis known, but that if concord is with NP2 then the speaker assumes the predicaterelationship is unknown and that he is providing the listener with new information.Therefore (b) would be used perhaps in response to someone who did not know whatwork was, or who thought cultivation was a soft option, but (a) would be aredundant remark, used after perhaps when coming home tired from the field. It is notpossible to derive (b) as an of some sentencelike Kazi iikuwa kilimo hiki 'The work was thisdigging' jiwe since kazi is semantically spechic in this sentence, whilekazi in 59b can have specific or non-specific reference. Other examples are: all 60 a.Kusaidlana ku uungwana `To help one another is humanity'. b. Kusaidiana u uungwana `To help oneanother is humanity'. ither 61 a.Mtoto yu klumbe 'A child is a living being'.

flot b. Mtoto ki klumbe 'A child is a living being'. is c.Yu klumbe mtoto 'A living being, that's what achild is'. Vp2. 31 to d. Ki klumbe mtoto 'A living being, that's what achild

13 14-I.A. j la

Where both NP1 and NP2 belong tothe same morphologicalclass, the sentence is amSiguous. The tendency toassociate first position in the sentencewith emphasis, however, causes most suchambiguous sentences to be interpreted asof type (a). with NP2 There is a wide disagreement among usersof Vpf as to where concord is permissible, except in thefollowing set of sentences: 62 a.Biriani i chakulaTiriini is afood'. b.Biriani ki chakula 'Biriani is afood'. accepted Vpfat all re)ected 62aand claimed only 62b Six out of the ten speakers who fact that was possible.Perhaps this special concordrelationship is attributable to the *tont is a loan word (andalso a loan dish). Native foodsdo not appear to allow concord with chakula so readily. Forexample, some of the speakerswho insist on 62b also insist on the (a) form,if Vpf is to be used at all, in: rice/porridge/cassava is a food'. 63 a.Walilugalilmuhogo u chakula 'Cooked rice/porridge/cassava is a food'. b.Walilugalilmuhogo ki chakula 'Cooked 64 a.Samaki yu chakula 'Fish is afood'. b. Samaki ki chakula 'Fishis a food'. M does, however, allow all the(b) forms as well as the (a). SUMMARY OF PATTERNS: This survey suggests that atleast the underlying'structural patterns charted in Table II must be established for thesimple copula sentence in Swahili.

Se tence

Predication (Sen.Adv) NP

(Prey. Adv) (Copi) NP

(Cop verb) (Specifier)

/\ ni N (Mod) (Mod)(Neg) (ndiyo2) (Aux) (-w-) ndio Vpf

TABLE II: UNDERLYINGSTRUCTURES

The table shows thestructural groupings of categoriesin the copula pattern 'Sentence' consists ofoptional Sentence Adverb +Noun Phrase + Predication: the optional Preverbal Adverb +optional Copula + Noun Predication consists of choices not discussed in this Phrase, etc. Not includedin the table are other possible Adjective or Adverb insteadof NP2, or the choice ofVerb Phrase paper, such as include the concord relations, instead of (Cop) NP2.Furthermore, the table does not restrictions on Negative Adverb, etc.,the optional position ofndio before nor various be regarded as -w-, nor theoptional appositional ndi- +-0, since these may %II obligatory or optional operations onthe basic structure outlinedhere.

14 e

i

PART II: THE STRUCTUREOF 'NOMINAL EXPRESSIONS' So far it has been assumedthat NP can be realized as anyN(oun) in copula constructions. Certain grossrestrictions must, however, bemade on the realizations of NP and also on the co-occurrenceof such realizations acrossCop. In broadest terms, NP maybe realized as any N with itsclass-number prefix, and optionally any one or more of a seriesof modifiers includingDemonstrative, Adjective, Possessive, etc.14 Otherwise it maybe realized as a pronominalform such as Dem(onstrative)- Pro(noun),Pers(onal) -Pro(noun),Inter (rogative) -Pro(noun). verbal expression may function as anNP, i.e. NP may 62b A third possibility is that a be realized by a nominalization.This nominalization takes theform of kwamba or : that infinitive construction kuM(ain)Kerb). lbw kuwa + Sentence, or of an 4 on More specifically, however,various restrictions pertainbetween realizations of NP as follows : 'of. BOTH NPs ARE REALIZED AS .. .N... : toe. When both NPs are realized as .. . N.. . ,both are usually membersof the same sub-category of N, e.g: both will beanimate or inanimate or abstract .How far further generalization is possiblewithin terms of the syntactic structureof the language is, however, unclear. For example,that Mchlcha n1 mboga 'Spinachis a vegetable' is acceptable while Mchicha ni mt1'Spinach is a tree' is consideredmarginal, appears to be semantic, not syntactic. Two exceptions to the generalization maybe noted. Any .. .N.. . in NPposition :d in may co-occur withkitu 'thing' in NP2 position (thiskind of sentence is usually inter- rogative): 65. Mwallmu ni kitu gani ? 'Whatkind of a thing is a teacher?' 66. All nikltu gani ? 'What kind ofthing is Ali?' When across (Cop) from Nanimitekitu is usually pejorative, and so maybe considered a special deviant usage.The other exception is also deviant.There is a set of sentences in which any sub-category of N may occurwith any other: 67. Mama n1 kero 'Mother is a nuisance'. 68. Ufinga n1 nyumba ya glza 'Ignoranceis a house of darkness'. . These constructions are almost allproverbial or metaphorical and so partof a special style and linguistic environment. In most cases number remains constant across(Cop); some singular mass nouns I such as chakula 'food' may, however, co-occurwith : \tod) i 69 a.Vlaz1 n1 chakula 'Potatoes arefood'. b. * Viazi n1 vyakula 'Potatoes arefoods'.

ONE OR MORE NPs ARE REALIZEDBY PRONOUNS: 1 Pronouns are unlike Ns in that mostof them are restricted in positionand co-occur 'ern with only a few modifiers. If only oneNP is realized as PersPro orDem-Pro, then the it must be NP1:. oim this 70. Yeye ni mlevi 'He is a drunkard'. 'rase 71. Huyu n1 ndege 'This is a bird'. ons, .ore./ On the other hand, if only one NPis realized as Poss-Pro, then it mustbe NP2: 1 as 72. Ndege ni wangu 'The bird ismine'. a

15

,

- Forms like Mlevi nt yeye andWangu nt ndege do occur,but they are special emphatic versions of 70 and 71, or elsethey occur as embeddingsin complex sentences such as Nyumba yako nzurt, yangu mud sana'Your house is good,mine especially good', in either case this order cannotbe considered part of theunderlying structure (q.v. Part III). Nan! 'who' and /gni 'what' arethe only Inter-Pro formsinvestigated here. Nan! occurs as aninterrogative pronominal for any N'human': 73 Q. Nan! ni rnwalimu 1 'Who isthe teacher ?' A. e.g. Bw. Mwakatoga'Mr. Mwakatoga', or Huyu'This man'. 74 Q. Mwalimu ni nani? 'The teacheris who?' A. e.g. Ibid. 75 Q. Nan! nt mume ?Who (whichof you) is a husband?' A. e.g. Mimi 'Iam'. 'non-animate' or of a nominali- Mai is commonly used asthe Inter-Pro form of N zation (e.g. infinitive orkwarnbalkuwa nominalizations). Itis restricted to Nhposition when only one NP is realized by aninterrogative: 76 Q. Nostra ni nint? 'Anger is what?' A. Hasara 'Loss'. 77 Q. Karl ni nini? 'Work is what?' cultivate A. NI kulea mwana, kuchfrnba na .. .'It is to bring up children, to and.. .' Cop, Me form of Inter-Pro is chosen notaccording to concord relationship across The syntactic but amording to the expectationof syntactic subcategory in the answer. sub-categorization of prime relevanceis the distinction human vs.non-human; animate vs. non-animate is of norelevance here, although elsewherein the system (e.g. concord with Adjectives,Demonstratives, Vpfs) an Nwhich is animate and non-human takes on the humanconcords (m-lwa- class). Forexample: 78 Q. Samakt ni ntni? 'Fish is what?' A. Chakula (non-human, non-animateN) 'Food'. 79 Q. Samakt nt nint? `Fish is what?' A. Mnyama (non-human, animateN) 'An animal'. of Fish ?). 80 Q. Sarnakt nt nani? 'Fish is who ?'(i.e. 'Who is playing the part A. Salma (human, animate N)`Salma'.

NOMINALIZATIONS: Infinitive nominalizations of the typekuMV are common. Either orboth NPs be may be realizedby such a nominalization;if only one is, then the other must realiz -W by an N belonging to aclass including: kart 'work' desturi 'habit' hakt 'justice' kawaida 'habit' 'aim, intention' wajibu 'necessity' shabaha 'aim, intention' jernbo 'matter' madhumunt

16 'haw shlda `difficulty' tnaana 'reason' ,411 as 'reason' (1'; kosa 'error' sababu ballad 'luck Jlbu 'answer' Kulea mwana ni kazi `To look after a child is work'. Nani 81. 82. III kwnpa uwezo Rats kumteua mtu kuweia kuendelea kazi ni Jambola maana sana 'So that to give thePresident the power to choose a man to bring about progress in work is a matter of great importance'.13 83. Madluanuni ni kuwapokea na kuwatazama mpaka waondoke 'The intention is to welcome them and look after them until they leave.' 84. Kazi ni kulea mwana `The job is to bring up (the) children'. 85. Kumsaidia kuondoa maradhl .ni kumweka mahall ambapo pakukaa `To help him eliminate disease is to put him in a place where he can live'.16 86. Kucheka ni kiffurahl `To laugh is to rejoice'. A special subset of kuMV nominalizations is that in which both NP, and NP2 dominate a ku-nominalization such that MV is in both cases realized as w-- no NP 'be with NP'.17 In such sentences both kuwa no may be deleted, or else justthe kuwa no dominated by NP1: 87 a. Kuwa na mume ni kuwa na shoga 'Having a husband is having a friend/ A husband is a friend'. b. Mwne ni kuwa na shoga Ibid. c. Mume ni shoga Ibid. Since the NPs are derived from nominalized sentences, there is no restriction on the subclass of N selected across (Cop) in 87 b and c. Of particular interest is the fact that when either or both kuwa na are present, then Vpf may not occur; when neither is present, however, Vpf is permissible: 87 d. Mume yu shoga 'Having a husband is having a friend/A husband is a friend'. e. *Kuwa na mume ku shoga Nominalizations with kwambalkuwa + Sentence occur as realizations ofNP2 only. They appear across (Cop) from any member of the class of Ns outlined at thebeginning of this section. Usually this N is modified by an AdJ such as muhimu'important' or by a genitive phrase (a NP). Copulas with kwambalkuwa nominalizations require special study.Kwamba is intuitively related to and necessity ; a sequence of co-ordinate verb phrases is expected within the nominalization, or at least Verb Phrase + Adverbof Manner. Kuwa is intuitively connected with indefiniteness and lack ofnecessity ; a simile verb phrase is expected. Bothsubordinators can introduce an indicative or a subjunctive sentence. Since the subjunctiveoften implies necessity, there is a problem in the analysis of kwamba sentences with subjunctives in the dependant clause.Note, however, that 88 b and c both have subjunctive dependant clauses, but only the one introduced by kwamba signals obligation. A set of contrasting sentences, including contrast with a kuMV nominalization is provided by : 88 a. Jambo tnuhimu kabisa ni kutengeneza madirisha 'A very important matter is to fix the windows' (generalization).

17 .....1.111=11111..ailarum...... :

\ ukuta kabisa ni kwamba watengenezemadirisha na watomee b. Jambo muhimu fix the windows andplaster 'A very important matteris that they should the walls' (expressesobligation). muhimu kabisa ni kuwawatengeneze madirishaIbid. (except used c. Jambo to fix the windows). as suggestiononly; 'they' may choose not tajanye jambo la maana'Our need is to dosomething 89. Haja yetu ni kwamba useful' (and we areobligated to do it). 'His reply is that he(probably) won't come'. 90. Jibu lake ni kuwa hail Here Kwamba and kuwa maythemselves introduce akuMV nominalization. to be introducersof appositive sentences,perhaps trans- kwamba and kuwa appear noticeable that, like sentenceswith kuMV latable as ' NP is asfollows: Sentence'. It is kwambalkuwa, they aregeneralizations related topeople in general, (cf. 88a) but no spwifies particular not to people inparticular (contrast 88b, cwhere watengeneze introduces an appositiveregarded as a necessity or known persons).While kwamba kwamba is asso- introduces an appositiveregarded as a likely act; oblig,ation, kuwa unrepeated ones: ciated with repeatedactions, kuwa with single, ni kwamba kutengenezamadirisha no kutomea 91 a. Jambo muhitnu kabisa should fix ukuta 'A very importantmatter is thefollowing: that people the windows andplaster the walls' (ingeneral). madirisha kesho 'A very b. Jambo muhimu kabisa nikuwa kutengeneza following: that someoneshould fix the windows important matter is the recognizes that tomorrow' (a single action,important, but the speaker it may not be carriedout). kujua kanuni za Tanu,kudumisha uhuru, no kuongeza 92. Wajibu wao ni kwamba know uchtani wa nchi 'Theirduty is the following:that everyone should freedom, and further the economyof the rules of Tinu,should perpetuate the nation'.

PART HI: WORDORDER

of NA (Prey. Adv) (Cop)NP2 constructions areof three The word order patterns NP2(Prev. Adv) (Cop) NA types: (i) basicorder: NP + Predication;(ii) inverted order: where NP is realized asPen-Pro, Foss-Pro,Dem-Pro or a restricted to sentences order: Predication +NP kwambalkuwa + Sentencenominalization; (iii) shifted (shift of (ii)]. [shift of (i) or (Prey.Adv.) (Cop) NPI + NP2

BASIC ORDER: apply only Part II, strict grammaticalrestrictions on word order As indicated in kwambalkuwa + sentencenominahzations. in special cases ofPronominal forms and Normal, i.e. unemphatic,order is as follows: N (Modifier) (Prey. Adv) (Cop) N (Modifier) kuMV kuMV I nani nani nini Pers-Pro I J 1 Poss-Pro Dem-Pro r kwambal Sentence 1 kuwa f 2 subject to the context- of I may co-occurwith any member of 2, where any member in Parts I and II. restrictions onsubclassification outlined

18

; tukuta INVERSION: where there is restriction onthe realiza- Inversion occurs onlyin those sentences kwambal ;.quster Pro-forms (excluding,however, nano and tion of NP,i.e. in sentences with apparent. Noparticular emphasis is, kusva + Sentence.Elsewhere inversion is only laugh is to be 1 ___, by either of the sentcncesKucheka ni kufurahl 'To lisou fcir example, signalled laugh' and neither canbe taken as happy' and Kufurahl nikucheka 'To be happy is to the other. On the otherhand, we have thefollowing sets: 4iing the base of 93 a. Wewe ni mwalimu 'You are ateacher'. that's what you are'. b. Mwalimu ni wewe 'Ateacher,, Hayo si mwisho 'Thisis not the end'. .Here 94 a. is'. Tans- b. Mwisho sl hayo `The end,that's not what this 44BI book is mine'. I :sera!, 95 a. Kitabu ni changu 'The is'. b. Changu ni kitabu 'Mine,that's what the book i Ityculour hakuna giza 'A matterof wonder is that there ,ISSO- 96 a. Jambo la ajabu ni kwamba is no darkness'. is no darkness, thatis b. Kwamba hakuna giza nljarnbo la ajabu 'That there :twusa a matter ofwonder'. Id fix constructions in Sentences like 96b are rare asthere is atendency to avoid complex barely acceptable notbecause of subclass orsemantic sentence-initial position. 97b is especially 1I very imcompatibility, but becauseof the length of theinitial NP construction, .Jows when compared to thelength of the NP followingCop: 3 that ambalo ninaliomba 'Thisis the very great 97 a. Hill ni ombo langu kubwa sana pray'. leza prayer that I hill. mow b. ?Ombo langu kubwaSanaambalo ninallomba ni 3y of SHIFT: inverted sentence mayundergo shift. Where a Any acceptable basic sentence or whether on clearly particular NP +Predication constructionis not permissible, grammatical or on semanticgrounds, its shift is also notpermissible. ree of 93a). 93 c. Ni mwalimu wewe 'Ateacher, you are' (shift teacher' (shift of 93b). or a d. NI wewe mwallmu'It's you who are a NP 94 c. Si rnwlsho hayo 'Notthe end, is this'. d. Si hayo tnwisho 'Notthis, is the end'. there is no darkness that Nay 96 c. Ni kwamba hakuna gizajambo Ia ajabu 'It's that AIL is a matter of wonder'. 98 a. Kulba ni kuasi 'To stealis to commit a crime'. steal'. c. Ni kuasl kulba 'Tocommit a crime, that is to (unacceptable to M and K 99 a. *Kuasl nl kulba 'Tocommit a crime is to steal' since kuasl to themrequires enumeration, askuasi refers semantically to an activity moregeneral than that referred toin kuiba). c. N1 kulba kuasi. pronominal forms, are notentirely The translations,introducing as they do many position of the emphasis incombination with wordorder that At. satisfactory. It is the counts.

19 The tendency of Swahili speakers toavoid complex constructions at the beginning of the sentence accounts for the fact thatwhile 96a, d are common, 96h, c are rare. Relative frequency or infrequency of acertain base form may have no aearing onthe relative frequency or infrequency of aninverted or shifted form ; the criterion is whether the base is acceptable at all,and whether or not any complex realizationof NP is at the beginning in the surface structure. Shifts must always be accompanied by anintonation break after Pred (or after (Preverbal Adv) (Cop) NP1 in the case ofshifted inversions), and also by ndlyo2 or some realization ofCop (presumably to avoid ambiguity withNP1 NP3 construc- tions). Only question-word copulas with Nalpermit absence of any realization of Cop:

100 a. Meza (Cop) ?dal? `What is a table?' b. *Mid (Cop) meza ? c. (Cop) ?dal meza 'ft's what, a table?' (shiftof 100a). d.*(Cop) meza nbil? (shift of 100b). Since 100b is impossible as thequestion-word is invariably a realization ofNP3 if it is the only ?dal in the sentence, NMI'ma? is unambiguously derived byshift from !dna nini ? and cannot be the inversion*led (Cop) meza ?.

PART IV: INTONATION PATTERNS" /(ntonation)P(atterns) are the same for underlying sentencesand their inversions. They are also the same for shifts andshifted inversions:The rules in this section are therefore specified for the overt Noun Phrase -I-Predication structures, not for the underlying structures. The main /Ps associated with copulaconstructions have the following features: ii) they are continuous or else brokenby brief pause; (ii) they end in afall or a rise; ih) these rises and falls are themselves of two types :staccato or norestaccato rise,and au from a mid or from a high point ; (iv)non-continuous /Ps begin level or non-level.

P > oonnt. tuno s uous {fall}

fall > fhiFhfsll

rise > f staccato rise 1 lnon-staccato risef

continuous > f level 1 1 non-level f

(Square Brackets indicate simultaneouscategories; the arrow indicates 'realized as'.) Continuous /Ps (/P Type I) are thosespread without interruption over surface structure Noun Phrase +Predication. They are limited to unshifted sentencesand to those in whkh at least one memberof (Preverbal Adv) (Cop) is present.Non-conti- nuous /Ps (Type II)consist of two parts with predictabledomains (a special pattern with three parts will be discussedseparately). The domains are the surface structure: ....momimatirwirAfthi4orroeraiiMwki64.4mm.wk.viet.v.Ia..,,r,...ir

Predication + NounPhrase. Noun Phrase +Predication and Prey. Adv and Cop areabsent must benon-continuous. rare. All sentencesin which both a the is mid level (IP Type i) are common.The intonation niis IPs with alevel beginning in the case ofshifts, :aof through NP1, or beginning of the utteranceat least mid level anti rise onthe from the patterns (Typeii) begin at Predication. Non-level with a stressablesyllable through the first NP.19If the first NP starts after stresseusyllable of the N of or intonation naturallystarts high. the questions; with statementsand content-word Mid falls (Typea) are associated time. All falling Cop: `truth-questions' whenasked for the first high falls(Type b) with high point, havefreevariants in .which whether they startfrom a mid or down patterns, penultimate (stressed)syllable, or else as a step the fall occurseither during the stressed to final(unstresses:I) syllable. from rises and slight stress on (Type c) arecharacterized by smooth Non-staccato rises they are associatedwith rhetoricalquettions (usuallyunstressed) syllable; markedly staccato, the final Staccato rises(Type d) arecharacterized by and echo-questions. often accompaniedby phonetic glot- !it is final syllables,the staccato slightly stressed, questions asked asecond time orin anxious, 'rckm associated with talization;. they are ,final fWIy stressedsyllable of the ut-. debate. All risesbegin on the the final animated etc. slight drop inpitch before thefinal syllable, but mance;there may be a pitch as the highpoint of theprevious sYllable starts atapproximately the same fades upwards. syllable and Different sets of combinatic ns ofthe typesmentioned above occur. 311S. All possible attitudes. : ale combinations areused tosivtardifferent the IP TYPEI: converiation ; in certain contexts Continuous IPs areusually used incasual, 'light' rudeness. JO: worthless topic,unconcern,petulance, even Om may implydismissal of a unbroken IP usedfor careful,deliberate constructed withndlo is the Ind Only in answers :vet speech. kuchimla 'Our yetu-iatrirs"ffitimbd, in freevariationwith Karl yetu ni La.Kai (might be used as adisgruntled reply). .job is to dig' (answer to e.g. decoration is whata/the tabje is' Pomba ndiomad 'A Mem iii nint?'What is a/thetable?). bored). husband?' (implicationthat speaker is Mums nininIPWhat is a ntkuchimbh 'Is Icuchimbh, in freevariation withKarl yetu I:b.Karl you nt question asked forthe first time). our jobto dig?'(casual question, or with Mume ni slam'A husbandis a Mume nt slamin free variation protection r (echo). Mume ni staro'Is a husband a .....f""Mume ni sutra,in freevariation with anxious, lively 1,;(1. asked a secondtime, or used in protection?'(rsiulant, or disCussion). they are used IP TYPE II:i: beginnings are themost common; Non-continnous 1Pswith level disregard, annoyance, usually withoutovertones of in deliberate,thoughtful speech, etc.: 21 II:i:a. Karl yetu ni kuchlmbh'Our job is to dig'.

Nani ni mwalimu'Who is the teacher?'

H :i:b. Mume ni stara'Is a husband a protection?' (serious question asked for the first time)

II thc. Fatuma ni mwalirnu'Fatuma is a teacher?' (echo).

II:d. Mume ni stara'Is a husband a protection?' (second time).

IP TYPE II :H. These patterns are rare; they are used to emphasize the quality ofNP1in respect ofNP2. NP1 (Prey. Adv) (Cop) NP2constructions with this IP combination may be roughly translated as:NP1 is the greatest of NP2where 'greatest' implies quality, not quantity.

OEM 1I:ii :a.All ni mwallmi'Ali is the greatest of teachers'.

NI haslra dhambi'Anger, that's the greatest of sins'.

II:ii:b.BIrlanI ni chakulir'15 biriani the greatest of foods?' (question asked first time).. ow/"/ II:ii:c.Ali ni mwallmu/Mnif'Ali is the greatest of teachers?' (echo). II:ii:d.Joshua alIkuwa mplshlWas Joshua the greatest of cooks?' (insistent).

OTHER INTONATION PATTERNS There are two other IPs, one continuous, the other non-continuous, but otherwise outside of the patterns described above. The continuous IP starts high and falls throughout the pattern. It is used exclu- sively in content-word questions, sometimes pejoratively:

I:x. Mume ni Oil'What is a husband of all things?'

Nord ni mane 'Whoamongst you is a husband?

The non-continuous pattern has the distinctive of high stress onCop,and division into two or three predictable parts. It starts level and the fall is from mid. If either the formativesndlyo2orndl- -0occurs, or -w-, then there are two parts, high stress occurring onmil.. .or else the stressable syllable of -w-; when onlyVpfor niare present then there are three parts, the specifiers receiving high stress followed by pause. All sentences withndiyo2have either this pattern (l.x) or else IP II:x is usually associated with emphatic speech and signals emphasis on NP2; it is not usually used withndio.

II:x.KuchIntha ndlyo kcal yetu'Digging, that is our job'.

Ndiyo kullkuwa nikad yetr kuchlmbi'It was exactly digging that was our job'.

Kuchlmba ni kad yetu'Digging is our job'. 22 .

' /OW ,.00111111111111116116110"

ADDENDA

Two limited constructions maybe noted which do not appear tobelong to the system sketchedabove. They are of considerableinterest for study of the syntactic AM for scope of Neg. (A) There is a set of sentences withthe surface structure of :1 NP Dem- Pro and a broken risingintonatiOn similar to IP II:d., but nbt the same since there are two stepsfrom a low beginning, rather than one from amid beginning. Despite the negative si, these sentences areInterpreted as affirmative, and despite the staccato rise at the end they areinterpreted as exclamations or statements, not asquestions: St chakula hicho.`This is food' (to someonewho does not know what 'this' is, e.g. to someone who sees an objectand asks whether it is a decoration, : respect or admonishing achild). I °lay be Ay, not --r (not e.g. Saimon). SI HamIsi yule 'This is Hamisi'

St shamba hilo 'This is a field'. chakula with the same meaning; these sentences can 1 There is no associated Ilk& 81. therefore not be regarded as shifts.

asked (19 Another set of constructions is associatedwith the same categories and formatives: st+NP+ Dem-Pro, butused exclusively with a variant of IP II:ii:a. In this variant the rise on NP isapproximately the rise of two steps, not one. When heard in discourse, the pattern ismid, very high, mid falling, rather than low,high, low falling, with expressivelengthening of the stressed high syllable. Sentenceswith this intonation are exclamatory,expressing amazement at the size or quantity ofNP: St chakula hick; `That's an enormousquantityof food' (implying disbelief at the quantity). SI1"* shamba hili 'That's a huge field'(disbelief again). :exclu. Such sentences should be examinedin. connection with other exclamations such as Nzuri vlbaya 'Very good' (lit. 'Goodbadly') where lexical items usually semantically negative or pejorative are used to signal strongpositive approval.

FOOTNOTES: I. This is a modified version of a moretheoretical paper: Elizabeth Closs, 'Somecopila cons- 2. and tructions in Swahili', Journalof African Languages(forthcoming).

I Parts, We are much indebted to W. H. Whiteley for hisassistance and valuable criticism. Many by his article, 'Some problems of the syntax of ;Wor of the questions discussed were stimulated sentences in a Bantu language of EastAfrica,'Lingua9:148-74 (1960). Thanks are due also to 4.Idby H. E. Lambert, Joan Maw and E. Polomefor their helpful suggestions. ibt -ugly Most bf the material was provided by SelmaMbaye (M) a nativeof Zanzibar and Ali Kondo (K) a native of the Baymoyo region.Further data wai contributed by G. Mhina, A. M. Khamisi, J. Rasnadhard; the newspapersUhurnand Nix's/me, the National Assembly records for 1964(Parliamentary Debates, National Assembly Reports, TheUnited Repgblie of Tanga- nyika and Zanzibar, Dar es Salaam,1964), and various short stories printed in Swahili were also cowshed.

2. E. 0. Aditon,Swahili gPrlintMOI (incledkg intonation),London, 19472; for sections relevant to the discufsios in this article seespecially pp.92-5, 17940, 205-13, 281-3. See also A. Loogman, and syniax,Pittsburgh, Pa., 1965, pp.40S-1S.

23 This is often called the Subject Prefix; that term has been abandoned here since concord is occasionally shown with the Predicate Nominal, not the Subject, cf. the discussion of sentences 59-64.

4. At this stage it is impossible to establish clear grammatical criteria for `class-member' sente- nces such as Mafua ni ugonjwa `A cold is a disease', Kilimo ni kart `Cultivation is work'. These sentences are all generalizations which do not permit the use of -w- or the expansion of NP by specifying modifiers (e.g. demonstrative, possessive); they are all ambiguous with sentences expressing equivalence or deftnition, where -w- is possible and also expansion of the NP by specifying modifiers: Kilimo (hild) ni kazi `This cultivation is work', Kilimo kililcuwa kaziThe cultivation was work'. The ambiguity is hard to account for on grammatical grounds without a clear framework of contextual restrictions (e.g. position within cliscourse; familiarity of the spatkers and hearers with the topic of discussion). It may, however, be noted that many (though not all) the class-member type sentences allow expansion by the frame: Noun Se + ni + moja .1- katika .1- Noun PI, e.g. Mtoto ni klumbe `A child is a creature' may undergo paraphrase expansion to Mtoto ni mmoja katika viumbe `A child is a creature/ A childione of the (class of) creatures'.

S. The apparent contrast between `be' in 3 and `become' in 1, 2 is oneof translationand semantics only. No structural distinction is made that would justify setting up two grammatically discrete verbs.

6. Ashton, Swahili grammar, p.92, claims that `when the complement identifies the subject' Vprmust be used for 1st and 2nd persons, ni for the 3rd. No sentences have been found in the dialect here under discussion where NP1 ni NP2 was considered unacceptable. In fact Wewe ni mtoto `You are a child', Sisi ni wapishi 'We are cooks', Sisi ni watu dhaIfu 'We are weak mortals' are preferred to the corresponding Vpftypes.

7.It also occurs in a few non-copula constructions; in these it also emphasizes or specifies the relationship to the Predication of the particular NP with which it occurs: Momela Farm ni imefanywa National Park `Momela Farm (specifically) was made into a National Park', Momela Farm imefanywa ni National Park `Momela Farm was made (specifically) into a National Park', Mama amefanywa (kuwa) ni muuguzi 'My mother has bwome (exactly) a nurse.'

IL The two were from Zanzibar. This might seem to agree with the entry under yu in F. Johnson, ed., A standard Swahili-English dictionary, O.U.P., 1939, that Vpf is not usual in Zanzibar and adjacent coastal regions; on the other hand 1 wo of those who accepted Vpf are from Zanzibar, including M who accepts Vpfperhaps more readily than any of .the other speakers consulted.

9. Ashton, Swahili grammar, p.92. See also ft.6.

10. Longman, Swahili grammar.. . , p. 409.

11. Ambiguity is due to the fact that both NPs are realized by members of the noun class taking mlwa class prefixes; ndi .1- 0 could therefore be in concord relation to either NP. 12. Ashton, Swahili grammar, p.179.

13. In colloquial speech an invariable ndiyo seems to be gaining ground, cf. Hiki ndicho kitabu? freely varying with Hiki ndiyo kitabu? Is this a boolc?' This ndlyo does not have the same function as either ndiyol or ndiyo2.

14. For a detailed study of the structure of NP cf. Judith Olinick, University of Wisconsin Diss. (in progress). She postulates a maximal or near-maximal string containing the following adjectival types with watoto `children' as the nominal head : hawa watoto wangu wazuri sena wo wote wengine gani watano `which five other of these very good childrenof mine', i.e. DemonstrativeNounPossessiveAdjective (including relatives)IntensifierQuanti- fiertQuantifier2Interrogative AdjectiveNumeral. In the phrase cited the Demonstra- tive hawa has been moved to prenominal position; in short nominal groups, however, it is often post-nominal.

IS. This sentence, which is an excellent example of embedded kuMV nominalizations, occurs in the Parliamentary Debates, 1964, p.19. It is a counter-example to the tendency for avoiding complex NPs in initial position; see discussion of sentences 96, 97.

24 1

4 v11 i

, 16.Parliamentary Debates,1964, p.46. _cord la awes 17. Most constructions of this 'copulaof association' type can be optionally expandedby kama 'like' immediately preceding N/12: (Kuwa na) mama(n1)(kamo) (kuwa na)shoga'Having a husband is like having a friend'. In practice, however,kamaisrarein the absence ofkuwana is, therefore, un- .,. essem- except in proverbs and other sayingsof a different stylistic mode.Kama 'Their success is our : in productive derived sentences like: Ufarisi wao niforgja yetu i expected in the kuwana ';anion happiness',thud si hoja'Beautydoes not count', but Wreadily acceptable t;$ With forms of these sentences. On theothrr hand, kamois quite often heard in sayings like:Mke- 'Like mother, s: .-sion of mwensi ni (keno) adui'A co-wife is an enemy',Moto wo nyoka ni (kama) nyoka like being with a snake), Mapenzi ni (kama) %- Ikuwa likedaughter' (lit. Being with the child of a ;flake is ',lands futa'Love is fat'. In proverbs NPi may also berealized as any kuMV; this is, however, not t. :dinky characteristic of productive usage: Kujibu ni(kama)(kuwa no) dhahabu 'Getting a reply is 1 a chat (like) getting gold'. ' cloud 3may related by paraphrase to causatives. The causative -Wore/ Several of these constructions are further verbs most often associated with these constructions are-kr- 'bring' and-sababish-'cause' ; i et Dada alikuwa kero'My sister was nuisance' and theparaphrases'Kuwa nadada kulilcuwa I There is a further possibleparaphrase Dada ana 1 (kuwa no) kero, Kuwa no dada kulileta kero. 1.tantics kero(lit. My sister is with a nuisance). All theseconstructions pose problems which need ..itically further study.

> I18. No attempt is made at detailedcharacterization or phonemicization; only theessential ilbiece outlines are provided. The IPnumbers do not correspondto Mrs. Ashton's.For a phonemi- t,i in the cization of Swahili intonation, see M. H.Watkins, 'Swahili phonology, with special tti rig emphasis on intonation patterns',Monograph series on linguage and linguistics,Georgetown, , 11 :25-42 (1958). I few loans) is characterized by stress on thepenulti- 1 19. The Swahili word (except in the case of a 1is the mate syllable, none on the final syllable. ni Park', tato a -11y) a

i'm-1421.1tar : from zkers

iv? 4me

NS& iifts WIC , anti. ,tra- 4 is

'Ifp :mg i i 25

i _