LinguisticBantiu.pdf

8/11/2019 LinguisticBantiu.pdf

1/11

Issues in African Languages andLinguistics: Essays in Honour of

Kay Williamson

dited y

E. 'Nolue Emenan jo

Ozo-mekuri Ndimele

National Institute for NigrianLanguages, Aba


2/11

Special Issue ofThe Nigrian Language Studies

P.M.B. 7078Aba, Nigeria.

1995 National Institute for Nigrian Languages. Aba.All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means ,lectron ie or mechanica , including photocopying, recording, or any information storagc andretrieval system, without the permission in writing from the copy right ow ner except in the caseof brief quotations embodied in critical articles an d reviews.

ISBN 978-2494-06-2

Published byNational Institute for Nigrian

Languages, Aba.P.M.B. 7078Aba, Nigeria.

rintedby Central Books Limited Plot 1 Ewuru Industrial Es ta te

P.O. Box 1379 Agbor Delta State.


3/11

Kay RM. Williamson in 1969

Professor Kay R.M. Williamson in 1994

111


4/11

JOK J y

J

Bantu

The harder I try the inore difficult I find it to say which nominal phrases are syntacticobjecta in Bantu. Th common tests of objecthood aie not quite^st^sfactory. In this paper Ilist some alternative diagnostic tests. The results are not easy to interpret; die old proMetnsremain arid new ones areadded. ,> -r

Objects have to be distinguished from non-objects. Non-objects come in tw o dTerentkinds: subjects and adjuncts. Furthermore, several^kinds of objects may have to be

i each pther, espectaUy if we concde that a sentence m ay contain more

Since Jinite'verb forms include a subject concord (SCd) regarcDess of th prsence orabsence of a lexical subject in the sentence, it is easy to establish which noun phrase is thesubject of a given sentence: It is the one that agres with the person or dass of the SCd. Iaccept this as a diagnostic feature fpr subjects, at least for the purposes of this paper, eventhpugh jere y prpblems ith thi defuiitio^ Soi Ig sito nwy,c that locativesubjects are'not quite what they seem,u> be. and theie i{are languages where, in certaiorelative verb formst the apparent SCd agres not with the subject but with the antecedent ofthe relative clause.

Having dealt with subjects, I now turn to properties which distinguish objects fromadjuncts. Three tests are widely used to dtermine the objecthood of a noun phrase (cf.Hyman and D uranti 1982):

word order (access to the position immediately after the verb) diticization (as object prefix) subjectivization (through passivization)

In (1), the word kamba 'rope' shows all three properties relative to the verb -kate 'eut'.However, if we replace kamba in the first sentence by, say, Icimya 'silence', il would thenbe gui te impossible to apply the same transformations.

173


5/11

Essaya in Honour ofKayWilliamson

I I ) SwahihG42d

a-me-kata kamba she bas eut a ropea-me-i-kata he bas eut itkamba i-me-katwa th rope bas been eut

rhese tests sometimes help to distinguish objecte from adjuncts, but they do not alwaysnswer ail questions. Diffrent teste may produce conflicting dues, and thus lead to degrees

types of objecthood, a topic that bas recntly inspired a flood of papers on "objectisymmetries".There is another problem conceming th interprtation of such tests. How can we be certainthat th acceptability of any of thse transformations (through changing the word order,pronominalization and passivization) tells us anything about the syntax of the originalsentence? Consider the following case of "symmetrie objects".

(2) Haya E.22 (Hyman and Duranti 1982)akah' omwaait' bitooke. She gave the child bananas.akah' bitook' mwana. She gave bananas to the child.

akamh' bitooke. She gave him bananas. .a ka bib' mwtfana. She gave them to the child.omwtfn' akahbw' bitooke. The child was given bananas.bitooke bikahtfa'bw' mwaana. The bananas were given to the child.

In such a case, where both complments have access to the post-verbal position, to objectcliticisation and to subjectivization, I maintain that it is still possible that the twocomplments in the first pair of sentences may have different syntactic status. As we shallsee, such diffrences are overtly marked in some Bantu languages just as they are markedin English by the use of a prposition.

There is a kind of markers that may teil us something about objecthood, and most of themcan even be observed without recourse to transformations. The problem is that these markersidentify a class of complments that is wider than what we'are used to regard as objecte and

for which we seem to have no use in our syntactic descriptions. There are at

least four such

properties:

object case marking by tone metatony of verb form s shortened verb forms transitive agent nouns

The first of these markers is a particularly powerful diagnostic device because it involves amarking of the noun itself The other three markers appear on the verb. The first three are

directly observable in the appropnate sentences; the fourth one involves a derivationalprocess of nommal isa t ion None of Ihem provides a test that could be applied to just anyBantu l anguage , K it u S of ihem have at least a regional spread within Bantu

174


6/11

Schadeberg: Object Diagnostic in Bantu

OBJECT CASE MARKING BY TONE occurs in an area along th Atlantic Ocan from Gabon Zone B) in th north to Namibia Zone R) in th south; it is not dear how far inland thisrgion extends. The syntactic and semantic dtails differ from language to language. ForUMbundu (R. 11), I have described the situation approximatdy as follows (1986):

Each noun appears in two tonally distinct forms, one starting with a high tone form A)and one starting with a low tone form B). Form A is used for all nouns as the citationform and with a predicative function. Form A is also used for nouns with an augment(which are most nouns except proper names, certain kinship terms and some borrowedwords) when they are used as die first nominal object of an affirmative verb in a mainclause. Form B is used in all other environments. More predsdy we can list the followingenvironments:

(3) UMbundu R. 11

OC Object Case, form A with nouns having an augment) is used for: the first complement of an affirmative verb in a non-subordinated clause this includes locative complments it also includes the locative noun (dass 17) referring to the agent of a passive verb it also includes basic and locative complments of various verbs 'to be' it also includes infinitives used as complments of auxiliary verbs it also applies to the first nominal complement of a verb having an OCd referringto another complement

CC (Common Case, always form B) is used for:

all subjects the second and any following) complement of any verb form this includes a nominal complement af ter a verb with a post-ditic pronominalcomplement the complement of a negative verb form the complement of a verb in a subordinated clause (mostly relative verb forms) subordinated infinitives complments of infinitives the infini live itself being in the O C form when it isthe complement of an affirmative auxiliary verb in a main dause)

the complement of a progressive verb form such forms are probably derived froman auxiliary followed by an infinitive) complments placed into pr-verbal position and (rarely) subjects placed into post-verbal position adjuncts nouns outside the prdication frame)

All these cases have been documented in the article mentioned above; in (4) I repeat just afew examples for further illustration:

(4) UMbundu R . 11 Schadeberg 1986)

175


7/11

Essays inHoaour ofKayWUaaiscai

ottjaK y-4-lh imj* epako9.parentOC SCd9-PST-give IchildreaOC SAuttOCThe parent gave die children a fruitoojali i- va-ha pako9.pajientCC SOi9-OCd2-give 5JnritOCThe parent gives thema fruiti-tnde-nw* v-UviSCdl-leavejOPT-PROl8 lA-3.mc.OCMay heget out of the trapnda w44fle ombis...if SCdl-PST-eatPF 9/ish.OCIf s/hehadeatenthefish ...

Note that UMbundo adrm'ts nkte than on object concon) before the verbstem, but it oeveradmits two nominal complments boh marked for object case (C),METATONY U the tenp nstd for those cases where *e tone of at final (inflecttonal) suffixof certain verb forins is (underlyingly) high before an object and low odierwise. This ishow Meeussen reconstructs tor of |he,nal suffix -a of UK infinitive (1967; 11IX Thegeograpbjcat spread of inelaferiy has not been described in die literaiure, bi&theiact thatMeeussen reconstructs it for Protobantu indicates that it has some significant distiibutioiL Ihave found metatony described for the infinia've and some Ae* veWforafor languages of zones D and L. I shall use Stapften' (1964)grarnnw ofexempficadon. From the description it is cJear that the tgh w fuialan object and. the low tone final *- before pause. Kote that ii is quite dear itatMs is nota phonological conditioning since numerous other words, indnding verb rfemm, areconsistently ending in either high or low tones irrespective of their being in"the*^d


8/11

Sdtudebetg: Object DnceOa in Battu

SHORTENED VERB PORMS we a typical feature of southcrn Bantu languages (Nguni S.40,Soa S30;;Viid iS,20), bot sometjjing snar bas also been described for Interiacustrinelanguages e.g., Rm&i J.62). Again, I am not aware of the exact spread of this

short nVTh linetal bl&orosiiplified statement about their distribution is that the shortfonn ia nsed befoie a oontplement and the long form dsewheie. I shall illusttate dus withson eijtampjes ftm Xhosa (S.41), taken fiom the pedagogical grammar by Louw andJuhat p S; une not marked) and from the study of Xhosa synax by Du Rtessis andVisser (1992; tone maited).Tbcp*4fe two fernes in Xhosa wtch xbibtt a loog and a short form: the Present and the

shown in (6aj. It is obvious at the derence between tfae.twofonns^s not oattioned phonologicaUy. Tts is pattic^ajrly ;>eiepr "in, 4 Prsent

^ does not alf cit the -ttl'pf^ ' f e w o , ; d e t t n y n e s )the^ ^ woni4iitenai teiito ffi The gen distribution of d^e twp fonnsgiven above bas to be fwdir qualified It a|JF>eat tt> b'troe that the short form can neverappear in sentence final position but

(6b) Ie ntnU f-hCn dWphini the giri is returning from town [short fonn], Mn*M.y**lU*wBi tbepeoplcarecryinguidiuidinoagfonn]


9/11

Essays In onour S KayWWiamsan

t-

Xhosa is a language with rather free word order, but changing the place of the noun phrasesbas repercussions on th choice between th long and th short verb forms. Both grammarswhich I consulted ascribe thse diffrences to discoursfunctions and/or emphasis; I won'targue with this but I still suspect that diffrent syntactic structures are involved, too.

TRANSITIVE AGENT NOUNS occur in many East African Bantu languages. The situationis very clear in Nyamwezi F.22) which I shall use for exemplification (cf. Maganga andSchadeberg 1992:173-175).

The stem of th normal, independent agent noun is formed from th verbal base by thaddition of either one of two aUomorphs: -i in gnral, but - after th causative extension-i- and th passive extension -u-. (The high tone of the second allomorph is probably atrace of the original high tone of the two extensions.) JHowever, when the agent noun bas acomplement, another final morphme is suffixed: -. This auffix bas an underlying low tonebut may be realized as high by the regulr opration of High Tone Shift.

(7) Nyamwezi F.22 (Maganga and Schadeberg 1992)

(7a) mtummi worker


10/11

Schadeberg: Object Diagnostics In Banni

syntacc relaonship between the verb and (some of) its complement(s). No real analysisof any one language has been presented, much less a comparative overview of thetypological variation. If any conclusion is indicated at all, it would be that Bantu languagesrecognize a type of syntacc relationship which is wider than our traditional category ofobject, including some but not all of our category of adjunct. I think we should take seriousaccount of such concrete morphological markers in syntactic description (and theory).

Fmally, I would like to add a comparative historical hypothesis. It occurs to me that thefirst two of the four complement marking stratgies presented in this article may well berelated. I have said before that the tonal marking of the noun, as described hre forUMbundu and attested in all language groups along the Atlantic Ocan (except thenorthernmost ones belonging to Zone A), has its origin in the augment (also called "pre-prefix" or "initial vowel"). In the northern part of this area, the segmental, vocalic body ofthe augment has disappeared, leaving behind its high tone which became incorporated intothe noun. Metatony, then, could very well be the incorporation of the same floating hightone into the preceding verb form. Much will depend on the geographical spread ofmetatony, which, ideally, should be adjacent to but not overlapping with the area wherecomplment nouns are marked tonally.

Shortened verb fonns, at first sight, appear to have nothing in common with the augment,object case marking by tone and metatony of verb fonns. Recently, bowever, Creissels (in

press) has shown that in Tswana (S.31), the dichotomy between long and short verb formshas a tonal parallel in other tenses. He also notes the similarity with the facts from Rundifind adopts the syntactically oriented tenus "conjoint" and "disjoint" forms (cf. Meeussen1959). We therefore cannot rule out that further comparative research will establish a linkbetween conjoint verb forms and metatony, too.

Xo reconstruction exists for the final suffix "-a of the transitive agent noun. If it had a hightone it would be tempting to think that it, too, could be a remuant of the augment, beingidentical with the final suffix of the Infinitive when followed by a complement. But this isnot what the data from Nyamwezi suggest.


11/11

Essays la Honaur ofKoyWUliamson

Rfrencesoil Creissels, Denis. In press. La tonalit des finales verbales et la distinction entre forme,n verbales conjointes et formes verbales disjointes en tswana. To appear in Linguistiqu

Africaine,

j Du Ressis, J. A., and M. V isser. 1992. Xhosa syntax. Pretoria: V ia Afrika.v .Hyman, Larry M., and Alessandro Duranti. 1982. On the object relation in Bantu. In Syntak and Semantics, 15: Studies in Transitivity, ed. by Paul J. Hopper and Sandra Au Thompson, pp. 217-239. New York: Academie Press.

^ Louw, J. A ., in coopration with J. B. Jubase. 1978. Handboek van X hosa. JohannesburgEducum.

aiM Maganga, Clement, and Thilo C. Schadeberg. 1992. Kinyamwezi: Grammar, textsul vocabulary. (East African Languages and Dialects, l.) Cologne: Rdiger Koppe.7 Meeussen, A. E. 1959. Essai de gram m aire rundi. (Annales, 24.) Tervuren: Muse Roy

de l'Afrique Centrale.

Meeussen, A. E. 1967. Bantu grammatical reconstructions. In Africana Linguistica lll(Annales, 61), pp. 79-121. Tervuren: Muse R oyal de l'Afrique Centrale.

7 Schadeberg, Thilo C. 1986. Tone cases in UMbundu. A fricana Linguistica X (Annales121), pp. 423-447. Tervuren: Muse R oyal de l'Afrique Centrale.

Stappers, Leo. 1964. Morfologie van het Songye. (Annales, 51.) Tervuren: Muse Royal tkl'Afrique Centrale.

n

i

LinguisticBantiu.pdf

Documents

Transcript of LinguisticBantiu.pdf