Zambezia (1979), VII (i).SOME PROBLEMS RELATING TO THEINCORPORATION OF LOANWORDS IN THE LEXICONH. CHIMHUNDUDepartment of African Languages, University of RhodesiaTHIS ESSAY IS a summary of observations made by the writer while he wasgoing through Hannan's Standard Shona Dictionary* as part of an attemptto build up a corpus of 'acceptable' loanwords in Shona. The writer foundthat in addition to the common problem of establishing criteria for theacceptability of foreign words,2 a little-studied, little-standardized and little-modernized language like Shona presents other problems for the lexico-grapher: dialectal selection, choice of a suitable format for the dictionary,and considerations regarding its relative usefulness to first and second-language speakers respectively.It must be stressed at the outset that, in the writer's view, the in-consistencies and inadequacies of this Dictionary as outlined below are anindication not of incompetence on the part of the compiling team but ratherof the problems that the complex language situation in the Shona-speakingcommunity creates for the lexicographer. The writer must also point outthat his comments deal mainly with Hannan's handling of loanwords.1.0. The Language SituationBefore these comments are made, the language situation in the Shona-speaking community must be sketched. The language varieties are as re-presented in Figure 1.From Figure 1 it may be observed that a gradual process of dominationby a Zezuru-based variety of Shona is in evidence. One may well claimthat Zezuru is becoming the de facto 'prestige-laden standard language'vis-d-vis, not the other dialects as such, but their sub-dialects or LV as thenumerous 'local patois'.4 For reasons of ethnic identity these LV are pre-served but their speakers adopt certain linguistic features from Zezuru toi M. Hannan, Standard Shona Dictionary (Salisbury, Rhodesia Literature Bureau,2nd edit., 1974).*The problem of acceptability will be discussed later in 4.0.»U. Weinreich, Languages in Contact (The Hague, Mouton, 1953), 1.* T. H. Greenbers, Language, Culture and Communication (Stanford. StanfordUniv. Press, 1971), 185.75176INCORPORATION OF LOANWORDSA1 tLV (other) LV (of Zezuru)(C)ŁLWCKEYLVPL,GVPL2SLOLPVSp, Ps, SPSp(o)Ps(o)PS(o)H,LHLLWC<1). (2), (3)(A) ,(B), (C)(A)(B)(C)Local Variety (Sub-dialect e.g. Duma, Shawasha)Primary Language (Dialect other than Zezuru)General Variety (Zezuru-based)Primary Language ('Standard Shona')Secondary Language (English)Other Language (s) (e.g. Cewa, Afrikaans, Ndebele)Popular Variety (Shona + English and/or Other)Manner of switching or amount of interference (see Note 3)Secondary Language Predominant((o)ŠOther Languages(s)Fnmary Language PredominantBalanced Switching"High', 'Low' in diglossic situationDominant or Upper Language (English)Lower Language (Shona)Language for Wider Communicationintra-group Communication Levels (where 'Group' refersto entire Shona-speaking community)/ u ^~up Communication Levels(sub-) Dialect + (sub-) DialectShona + Other (Rhodesian)InternationalH. CH1MHUNDU 77avoid the use of 'marked' varieties which they think may carry some stigma.*For example, the verb stem -dya (eat) is pronounced in a variety of waysin different parts of the country, but Salisbury at least, most peoplesay [d? ga] as in Zezuru.However Standard Shona (PLJ1 is still struggling to be born, and thiscreates serious problems for a work with such an obviously ambitious titleas Standard Shona Dictionary. The situation is further complicated by afluid social situation in which PL2 is struggling against a faster-growingPV. The PV is itself very unstable, lexically at any rate, being a variety ofadmixtures of Shona (the L) and English (the H) in a diglossic situation,plus interference from other languages (OL) according to the degree towhich individual speakers have been exposed to them. Examples of suchlanguages with which Shona has or has had contact are Ndebele, Cewa,Sena, Afrikaans and Portuguese.The terms 'interference', 'high' (H) or 'low' (L), 'sub-dialect', 'dialect'or 'language', 'native', 'indigenous' or 'foreign', and 'loanwords' or 'borrowedforms' are used in this paper in a technical sense without any evaluative oremotional character. The view taken here is the neutral one that at anygiven point in time the state of a language cannot be better or worse thanany of its previous forms; nor can a language be better or worse than anyother language in the same period of history. Cultures differ, change andinteract, and languages must adopt accordingly 'to suit the occupancy of anew personality'.8 Rather, it is being suggested here that a case can be madefor multiple diglossia in the Shona speech community: (a) along a triglossicpattern from LV through PLa/PV to English (E) (as in Figure 2); and (b)between each variety of Shona and E (as in Figure 3).NOTES TO FIG. 11. PL.2 is not indicated as a LWC because it has no individual existence in naturalspeech flow.2. In the shaded overlap the arrows converge on PV. In the same area are also foundtwo varieties, 'English' and Shona', neither of which is anybody's first language; butwhether a case can be made for a 'Shona-English dialect' is a moot point.3. The term 'interference' in this paper is used in the neutral sense, i.e to refer to in-stances of deviation from the norms of the bilingual's PL as a result of hb familiaritywith SLs.4. The term 'Shona' here is used to refer to any or all of its varieties from LV to PV.»K. G. Mkanganwi of the Deparment of Linguistics, University of Rhodesia,suggests in a personal communication that the result of such attitudes is the emergenceof what may be described as Town Shona (which would be the same in status as PVin this paper).«E. Haugen, The Ecology of Language (Stanford, Stanford Univ. Press, 1972),303. Compare also, M. H. Abduhziz-Mkilifi, 'Triglossia and Swahili-Emrlish bilineualism'Language in Society (1972),!, 197-213.78 INCORPORATION OF LOANWORDSWhat one gets in the end is not a simple case of diglossia between say,LV and GV or PL 7 or between Shona and English, if we accept Fishman'smodification of the notion.8 Nor could a further modification to incorporatethe notion of triglossia adequately describe the manner in which the linguisticversatility of Shona speakers has developed, or is developing, some patternsof usage as social norms.Figure 2TRIGLOSSIAH = ENGLISHL = PL^PV = HLV = LFigure 3MULTIPLE DIGLOSSIAH = E Š E Š EL = LV,x) - PL - PVThe situation then is complex and perhaps the inadequacies of Hannan'sdictionary reflect the practical difficulties that confront any attempt todescribe Shona as a whole. The language is heterogeneous and the socialsituation fluid. The development of a standard dialect is not easy to predictor control in a multiple-diglossic situation because the speakers' socio-inguistic behaviour tends away from purity towards an inter-language inthe overlap area in Figure 1. Hence the real need to give serious attentionio Dorrowed forms and innovations in the language. The whole point inS. °Uttka ^Ptive model of options available to the speaker is tounderline the problems of compiling a dictionary of 'Standard Shona'.ZT7 ' Vefy admirable *ough it is, cannot be regarded as atessand> strictiy speaking' *is not reaiiy a dic-to JlwS ll T tO SUgg6St' however> that no efforts are being made^ Z gUagC Idd Hll T gg' wever> that no efforts are being madegUagC' Indeed Hannan's di<*ionary is one big contribu-d ThTTl°n> Th3t the l3ngUage is beil* «Jat and codified is nottil t u BUKm' educatio"al institutions and the broadcastingŽ to? rUer DeWSpaPers are M contributing in theirways to processes of codification and elaboration, the twin aspects ofHymes' LanPW. Culture and SocietyH. CHIMHUNDU 79language modernization." But this process lacks formal planning (e.g. inthe sense of a national language policy). The result is that although stand-ardization in grammar, orthography and phonology has made some pro-gress, there is some confusion at the lexical level: language elaborationinvolving the addition of technical vocabulary, for instance, is laggingfar behind language codification. Language elaboration is an on-going pro-cess10 but it is being managed mainly at the ideolectal level. This is pro-bably why the general tendency is to switch from Shona to English everynow and then, because English acts both as the language of specialized in-formation and of wider communication." On Fergusson's scale, then, Shonacould be placed at 'W.I' ('written I') where use is only for 'normal writtenpurpose'.1*Hannan's handling of the situation will now be discussed.1.1. The Second Edition. On the whole it can be said that Hannan wasbolder in the second than in the first13 edition of the Shona dictionary asfar as the incorporation of new words is concerned. That some attentionwas paid to the effective attitudes of the mass of the speakers at the PVlevel at the expense of 'purity'14 is indicated by the numerous borrowingsand creations contained in the second edition. As linguistic diffusion is anon-going process each new edition should presumably contain new entriesfrom PV. Some of the entries contained in the 1974 edition would pro-bably have been rejected in 1959-61 on puristic grounds. Others had noteven been created or borrowed. And now, only four years after the publica-tion of the second edition, one feels that a lot more loanwords and innova-tions could be included in another revised edition without risking a warwith the purists. To that extent, then, it can be said that Shona dictionarieswill always be 'out of date' and future compilers must always consider newentries from PV.A good 8 per cent of all the entries in the second edition are of foreignorigin, English is the main donor language, accounting for 64 per cent of»W. Whiteley, 'Introduction' to Language Use and Social Change (London,Oxford Univ. Press for International African Inst., 1971), 1-2.<°Cp. A. R. Diebold, 'Incipient bilingualism', in Hymes, Language, Culture andSociety, 495-508, on lexical borrowing and general acculturation. Here reference isnot to the fully integrated loans which are learnt from childhood, but to the morerecent ones which are learnt through subsequent contact." Cp. E. A. Nida and W. L. Wonderly, 'Communication roles of languages inmultilingual societies', in Whiteley, Language Use and Social Change, 57-74.i* G. A. Ferguson, 'The language factor in national development', in F. A. Bicefed.}. Study of the Role of Second Languages <n Asia, Africa and Latin America(Washington, Centre for Applied Linguistics, 1962), 9.is M. Hannan. Standard Shona Dictionary (London, Macmillan, 1959 ; reprintedwith corrections, 1961).i« J. R. Rayfield, The Languages of a Bilingual Community (The Hague, Mouton,1970), 105. Cp. also, O. Jespersen, Growth and Structure of the English Language(Oxford, Blackwell, 1938), 81.8OINCORPORATION OF LOANWORDSall these loanwords. Among the constituent classes or form classes thenouns are outstanding as the most hospitable to newcomers, accounting for79 per cent of all the loanwords. Most of these loan-nouns have been adoptedin classes 3, 5 and 9 where reference is non-human, and in class 1/la15where reference is human. The most productive noun classes in this respectare 9 and 5, accounting for 53 per cent and 29 per cent of the total loan-nouns respectively.18For reasons of space, the full implications of these figures for Shonain linguistic terms cannot be stated here. These figures are only intendedto indicate the general direction in which the total composition of thelanguage is changing, and only in this connection one or two notes may bemade here.Nouns are most hospitable to newcomers because in a contact situationbetween H and L, the vocabulary of L proves most deficient where materialculture is concerned. It is also very probable that the influences exerted bySL and OL and PV are greater than is indicated by these figures because,in addition to outright transfer of elements, the semantic function of nativeterms may be broadened through the influence of another language. Theuse of imba to refer to any type of (modern) building, of zuva to refer todate, and of chidimbu to refer to 'fraction' or 'percentage' are good examplesof this. For the most part Hannan gives these new meanings although in somecases (e.g. in the case of imba) he fails to do so.Another problem for the compilers appears to have been the incidenceof archaisms and preferred loans. Loan creations as such (e.g. hangurwa formscmt dzimudzangara for 'radio, radiogram' and chitundumutseremutsereior rocket) do not occur as often as their directly imported equivalents(e.g. bhmkitsi redhiyo/redhiyogiramu and roketi). That these direct loansare more easily popularized is probably because they are introduced at thesame time as the items to which they refer, while the loan creations arenew coinages stimulated by a need to match designations in English. Theseare generally used only by those people with some special interest in theT^fit consciousiy avoid using 'foreign words' wheneverof wwL? ? tWu SitUati°n is that there has ^wn a large numberThe Sictiona'rvh " " 'h°n°Wed S^Ž' for nat^ Shona terms,i S ^J^^T^ °f **"- Another result of this du^t, e.g. between mugwagwa and rodhi <l^-te-nS: v::vit\rtd 2+2a+2b - ««*«** <*-«.the loanword counting and the caJr , V^T6 ca'"'^ed to two decimal placesPossible that one or more of them cSu* have not been revised- I* » thereforedegree of possible error is felt tote negKg-bT W°nfC but> h°Wever this may be- theH. CHIMHUNDU 81English, 'road'. Unless the new term is consciously rejected (e.g. teachersgenerally penalize their pupils for using the Fanagalo words chikaju (food)and manje (now) ), the old word may bs discarded as its content is fullycovered by the new: e.g. kobiriŠ»peniŠ»sendi for 'penny, cent', and chi-doonyeraŠ»dhisikodhi for 'discord' (in singing). Yet another result of thisduality may be specialization of content. For example, whereas chiremba isthe Shona word for 'doctor', many people usually use chiremba to refer tothe n'anga (herbalist), while the loanword dhokota is used to refer to theEuropean-type medical practitioner. In Karanga badza. means both 'plough'and 'hoe', but now the first meaning occurs only rarely, while the loanwordgeja < Nguni is the more common word for 'plough'.While a dictionary cannot be expected to describe all these possiblerelationhips between so-called equivalents, it is being suggested here thatsimply indicating that given terms are equivalents is not enough. At leastthe more obvious archaisims or rare forms (such as bhahari < Arabic, 'sea'and bhangeni '< ?, 'very strong, well brewed beer') could be indicated assuch. This is a point which is overlooked in Hannan's dictionary. The wholepoint is that, however desirable it may be, it is not possible to preserveall the words of language in speech. Even without interference from otherlanguages, some words in a language will simply become obsolete as newcoinages are popularized. Low frequency of words, or homonymy, or theloss of expressive force by affective words are all internal factors that con-tribute to the innovating process and the shedding of older forms. Thisis not altogether undesirable and the lexicographer must not ignore thesenatural processes.2.0 Dialectal Selection. The language situation described in 1.0 meansthat dialectal bias is unavoidable even at the lexical level. This bias tendsaway from LV and PL, towards a Zezura-based PV. It is the latter of whichHannan's dictionary is a representation, and from that point of view thetitle Standard Shona Dictionary is misleading. It refers more to an idealto be built on PL2 (which would doubtless be influenced by PV) but whoserealization is neither close nor guaranteed. At best Hannan's work repre-sents a long step at the beginning of a journey towards that goal. In the'Introduction' the compilers indicate that their entries were drawn fromKaranga, Manyika, Korekore (Budya) and Zezuru (p. vii). Perhaps the titleA Dictionary of Central Shona would have been more suitable. A dictionaryof Standard Shona would have to draw items from at least one more dialect,Ndau.The status accorded to Ndau, Rozvi and Kalanga is both unclear andunsatisfactory. Presumably Kalanga was not represented because of low in-telligibility and some structural dissimilarities from Central Shona. Perhaps'Rozvi proper' was excluded for similar reasons. By 'Rozvi proper' here is82 INCORPORATION OF LOANWORDSnot meant the 'marked' variety of the Karanga-speaking Rozvi people (e.g.those of Chief Gumunyu, formerly of Bikita but now living in Gokwe).This 'Rozvi proper' is difficult to place as a sub-dialect17 and very few ofthose who claim to be Rozvi can claim proficiency in it. For that reasonits claim for representation in Standard Shona is weak. But the position isdifferent with Ndau. Ndau is commonly regarded as a major dialect ofShona but it was not even represented in the compiling team.18There are some words that are typically Ndau but occur often enoughin Central Shona: e.g. Ndauwe, interjective (Sir), which is indicated as 'K Ko'(Karanga and Korekore). Another example, twara, v.t. (carry) could perhapshave been indicated 'M Nd' (Manyika and Ndau), and not just 'M\ Bonore,5, 'maize cob' is also 'M Nd' and not just CM'. The compilers gave so muchattention to loans from the Nguni cluster of languages. Since Ndau has moreinfluence from Nguni than any other dialect and at the lexical level manyitems are found both in Ndau and Nguni'9 this connection could be quiterevealing for Central Shona if more attention were paid to Ndau vocabularyin the future.The compilers not only found it convenient to ignore Ndau, Kalangaand Rozvi, but also treated them as foreign languages as the few entriesfrom these dialects show. For example: chipunha, 1, 'Shave spirit' is indicatedas < Ndau'; murisana, 1, 'Boy' is indicated as ' < Rozvi'; and ndebwa, 9,Matter of concern. Case' is indicated as "< siKalanga''. This treatmentplaces the three dialects on a par with Venda, Sotho, Lozi, Swahili, and soon, the entries from which are correctly indicated 'from'( < ).The general argument being put forward here is that, whereas inrespect of structure Ndau, Kalanga and Rozvi may be expected to adjustm the direction of Central Shona, at the lexical level Standard Shona mustadmit more items from these dialects.fLiJnZ??- °f f1?Ž- Even where the origin of a word is c^In rTn' ldentlf"fon °{the so«rce or donor language is not always possibleŁ17Z TT >nTfails t0 detemine the source and he simpJy ^atesfrom EnJl*T 'IfT °f the banW0rds for wh*h he gives the source aretee are he S Af"kaanS <A)' Fana^° <B and Nguni (Ng). Presumablywere Zm* ^T" T* WWch Hannan and/°r members of his teaŽS^wntre WOrdsef 1^ C°UM ^ identify the interii^al co^-where words from these languages were concerned. Relatively few areof Kalanga.>»CDK&Mk °n> <2ndedk) viH!Tith. aP«Mnary NoŽ' NdL?tt* ,°f *'e Morphology of Substantives in Nrtau,thesis, 1973), 48. n wdau ph°nology' (Univ. of Rhodesia, unpubl. M.Phil.H. CHIMHUNDU 83the loanwords indicated as having been borrowed from Arabic, Swahili,Cewa, Portuguese and Sena which must also have had quite an impact onShona. It is quite possible that most of those loanwords simply marked 'For'were borrowed from these languages. Some etymological work along thisdirection could prove rewarding. One would probably find that the 64 percent figure for loanwords from English is too high if all the older loanwordsfrom the other languages could be identified. While borrowings from Englishare easily indentifiable because most Shona speakers know some English,borrowings with more obscure origins tend, if and when accepted in speech,to become fully integrated and lose their interlingual identity. Most speakersare unaware, for example, that the following are borrowings from differentlanguages: kwereta, chituta, shumaira, chikerema, chingwa, tsvigiri, chinotoand bweza.Another line of inquiry that could also prove rewarding would be todetermine the role played by the Nguni (in particular the Zulu and Sotho)in transmitting English and Afrikaans words into Shona during the earlyyears of White settlement (especially since Fanagalo seems to be Nguni-based).Hannan's assumption that many words from English (e.g. foshoro (shovel) )were borrowed via Nguni needs to be tested before the Nguni can bedescribed as agents of borrowing for Shona. It is possible that Hannan'steam's familiarity with some Nguni language(s) may well have given themsome preconceptions respecting source languages. It seems that where thereis similarity of form between Shona and Nguni words, Hannan tends toindicate ' < Ng', e.g. for -kupuka, v.i. (become rich) and -ora, v.t. (gather)and -vonga, v.t. and i. (thank (religious)). Could these verbs simply becases of coincidence Š especially if we consider that the description 'Nguni'is conveniently vague, and that Nguni and Shona are related in CommonBantu anyway?One inconsistency noted relates to Hannan's indication to sourcelanguages for different verb forms built on the same radicals. The verb stem-nata, v.t. (drink (greedily) ) is indicated ' < Ng', but the extended form-natira, v.t. and i. (drink by sucking) is indicated 'KZ\ But surely the sameradical -not- in the former has only had the perfective extension -ira addedto it in the latter form. These two forms are either both foreign or bothnative.3.0. The FormatAs far as the lay-out of entries is concerned two areas require comment:the handling of homonyms and allocation to constituent classes.3.1. Homonyms. Homonyms are linguistic forms that have the samephonemic shape but differ in meaning, e.g. mota, cl. 5 (boil) and mota, 9(motor-car). The problem is in deciding whether a given difference in mean-ing amounts to homonymy or simply represents narrowed or extended mean-ings of the same linguistic form where the guide to exact meaning is context.84INCORPORATION OF LOANWORDSOne meaning may be taken as normal or central (just as one allomorph orallophone may be chosen as basic) and the others as marginal (metaphoricor transferred). Here the practical situation gives the clue to which of thevariant meanings is intended.20But what may be viewed as variants of the same form by some speakersmay be viewed as homonyms by others. That such situations occur is aproblem that is acknowledged in linguistics, the general assumption beingthat: 'In certain communities (speech-communities) some utterances arealike as to form and meaning.'2' This could be construed to mean that eachlinguistic form has a constant and specific meaning Š which, of course, iserroneous but is useful because it is convenient.Hannan, then, could be excused for listing series of unrelated meaningsunder single entries where others (e.g. the present writer) would ratherhave all homonyms entered separately. However, the main contention here isthat this convenient practice is resorted to unnecessarily in some obviouscases. ^The following pairs of homonyms are all entered as single lexicalitems by Hannan:a)b)c)d)e)e.g. kotom, 9,koroni> 9, <A: Koring, 'wheat'koroni, 9, <E: 'Crown' (money)kot[> 9, <E: 'Court of law'koh> 9, <E: 'Quart' (measure)minzi> 9, < E: 'Minced meat'minzi, 9/io, <E: .Means.pomba, v.t., (native) 'Wind round. Bandage'pomba, v.t, <E: 'pump'vhesi, 5 & 9, <E: 'Verse'vhesi, 5 & 9, <E: 'Vest'.UnreIated and in some cases if2? ggeS ^ mCmnt (a) °r the mOddS"" different (b' c> e>- A Unction must be madeS and Slngle f°ms With several but related ««**cotton' or 'Cotton thread' or 'Crochet thread'.*"**"»*Allen and Unwin, 1933), 149-51.H. CHIMHUNDU 85Some cases of allocation to the wrong constituent class were noted. Thefollowing examples are all used in adverbial or adjectival function but havebeen entered as nouns: manje (now); futi (again); hadhi (hard); seti (nicely);mbaimbai (later); mbichana (little, small amount) and nyowani (new).A number of borrowings were entered as members of one constituentclass (e.g. as verbs) but their equivalent forms in another constituent class(e.g. as nouns) were not entered. It must be emphasized that the presentwriter is not suggesting that every borrowed nominal has a verbal equivalentor vice versa.22 Even where such equivalents do occur, frequency of usemay be different. Some forms like weti, 9 (urine) and -hwina, v.t. (win) arevery common while their equivalent forms -weta, v.i. (pass urine) and hwini,9 (win, prize) have not gained general currency. It would be understandableif the latter two were to be excluded from the list of entries (although theyhave in fact been entered). But in the case of such commonly used wordsas -ticha, v.t. (teach), wari, 9 (worry) and -noka, v.t. (knock), the decisionto exclude them seems either arbitrary or an oversight, since their equallycommon equivalent forms ticha, la (teacher), -wara, v.i. (worry), and noki, 9(knock) were entered. In fact the verb -wara is probably less common thanthe noun wari.Within the constituent class of nouns Hannan must have had someproblems in deciding on allocation to noun classes especially since manyloan-nouns are ambivalent in classification. These five nouns, all borrowedfrom English, all occur as class 5 or 9: karenda (calendar); kenduru (candle);goridhe (gold); girama (grammar) and vheserina (vaseline). Because ambiva-lence in classification is made complex by the complex language situationdescribed in 1.0., the treatment of these nouns needs extra care.Hannan is not consistent in his classification of such nouns. Where theambivalence is due to dialectal differences (e.g. vhiki and bhuku are 9 inKaranga and 5 in Zezuru, PLa and PV) it is understandable if one optionshould be abandoned in the name of standardization. But in other caseswhere such ambivalence occurs at all levels from LV to PV, Hannan"s choiceof class can only be described as arbitrary. Examples are:rivhi (leave) given as 5 but also occurs as 9rejimendi (regiment) given as 5 but also occurs as 9jerigadhi (prison warder) given as 5 but also occurs as lagadhi (guard) given as 5 but also occurs as lajuzi (stooge) given as 5 but also occurs as larobhoti (robot) given as 5 but also occurs as 9raisi (rice) given as 9 but also occurs as 5patapata (sandal) given as 9 but also occurs as 5." Such a suggestion would be ridiculous in view of the observations that havealready been made about relative receptivity to newcomers in 1.1.86 INCORPORATION OF LOANWORDSAmbivalence between 5 and 9 (with no secondary meaning involved)is particularly common. The explanation for this is simple. The bases forclassification of loan-nouns are semantic (e.g. la if 'human' and 3 if 'tree')and phonological where the initial sound corresponds with an independentprefix (e.g. chinwa is 7 and uroja is 14).Z3 But since borrowing tends towardssimplification of forms and relaxation of grammatical rules24 not only dothe majority of loan-nouns enter the zero-prefix classes 5 and 9 (whichrenders addition of an independent prefix unnecessary), but also, whereneither semantic nor phonological considerations are overriding, speakersfeel free to use 5 or 9 concords: e.g. when referring to kitibhegi, napukeni,tangi, dhuku. Hence the ambivalence. This is a common practice whichmust represent effective attitudes of the speakers which the lexicographermust take into consideration. Loan-nouns, then, need more careful treat-ment in this respect.28The tendency towards relaxation of grammatical rules during theborrowing process creates yet another problem: the occurence of irregularsingular-plural pairings such as the following:(i) Singular 1/1 a _y Plural 6Kesi" -> manesimupurisa ^ mapurisahweta -> mahweta(ii) Singular 9 ->. Plural 6bhebhi + mabhebhichukasi + machukasiHere the problem for the lexicographer is how to indicate these irregulari-ties. Hannan resolves this problem by indicating the independent prefix oftne plural form or giving the complete plural form although he does notdo this consistently (e.g. he only indicates la after neni <E: 'nanny').1 his seems satisfactory enough but perhaps it would be more economicalsimply to indicate the irregular plurals by the numbers of their appropriateclasses, especially since these borrowed forms seem to be quite versatile.**Tht -as 9 and la; matiridha (blue soanWTQ !!' ,Hannan classifies katakwara (cultivator). » A different type of^verStt 'i?^ 1*5 and «dnitH (minute) as 5 and 9.ton. Single **te^%ri^*j£^M^*n«d* h in phonological assimila-and spelt variously, e.g. bhambo/M^h /Bu!e\may each Vield a rePlira pronouncedBhatbheri/Bhaibhere/BhaibhiT /bh°mbu: ^burera/sumburera/sambureni;H. CHIMHUNDU 87Chukasi and bhebhi, for example, may also be class 10. The plural formsof these could then be indicated 'pi: 10 and 6'. But perhaps this is not avery important point.4.0. AcceptabilityIn any language there will always be the older, fully integrated loans andthe more recent, controversial ones. These groups represent the two extremesof a continuum, as it were. For that reason it is not possible at any givenpoint in the development of a natural language to draw up a list of accept-able 'new' words.27 The general tendency seems to be to reject the latestimportations and innovations on puristic grounds. But given time, many ofthese 'rejects' will gain general currency and become fully integrated in thelanguageŠeven so much so that their foreign source will be forgotten or justcease to be relevant (except to the academic as a matter of historical in-ference). This creates serious problems for the lexicographer and especiallyone who has to work in a complex language as that described in 1.0.In the opening paragraph of his 'Introduction' (p. vii) Hannan statesthat one of his team's aims was 'to provide as complete a list as could bemade ... of the words used in speech and in the writing' of the four dialectsthey handled.One can only try to update language in the lexicon but a variety of pres-sures require that one be selective when handling words used in speech,especially at the Popular Variety level. It may be necessary to become arbit-rary. Here the writer must emphasize that his comments respectng selectionor omission of specific words by Hannan are based on his impressions as afirst-language speaker (who could be wrong) rather than on any scientificstudy to test people's attitudes about their use.Criteria for acceptability can be determined, but faithfully to adhereto them would be impracticable. Greenberg suggests three 'indices of assimi-tion': (1) phonological and grammatical adaptation; (2) frequency of usein various socio-economic strata and occupational groups; and (3) the judge-ments of speakers regarding the degree of 'nativeness' of forms of foreignorigin.28 The first can be worked out by the linguist easily enough, but (2)and (3) would require a mammoth survey. Such a survey would have tobecome a permanent institution, continually reviewing and updating thevocabulary. Further, it must be maintained at the same pace as socio-linguistic change. Although perhaps desirable, such a project is practicallyimpossible, if only because of the possible costs involved."Haugen, The Ecology of Language, 335.as Greenberg, Language, Culture and Communication, 187-8.88INCORPORATION OF LOANWORDSSince attitudes are so much involved with the whole issue almost any-body will always point out words that 'shouldn't be there', words that 'mustbe there', or words that 'must be acceptable if this one is acceptable'. ListsA, B and C below should illustrate this point. In 'A' are words entered byHannan but which the writer feels are controversial. In list 'B' are wordswhich the writer feels should have been entered but were not. List 'Ccontains more words which were not entered but which the writer feelshave the same claim for entry as those in list 'A'chukasibhurekaenaemachecha!dhaizavhayabhepawenazemenakimbarisoritichatimbadhigiridhenjamushekanjanichunimurandadhuwetan'anzaThehas toowords offilling orselection iin relationv.i.&t.v.i.inter.v.t.v.i.inter.pron.v.t.5&9inter.v.t.959inter.inter.91v.i.v.t.LIST A:(<E:dice)(<?)(<Ng)LIST B:LIST C(<F/Ng)(<F/Ng)(<E:tune)'girl (-friend)''take a break''earn''aim''Hurry up!''trick, deceive''go''Sorry''you''examine (medical)''type of brick*'sorry''teach''timber''degree''colour red, e.g. on robot''well''How is it?''girl-friend''chap, boy-friend''die''eat, scrounge'?Z * *" "*orto a, "* ** *" Stmdard Sh°na Diction^^ords, but rather that the treatment of loan-^ VerlaP °f PV area' Whether these be »^"' is inconsiste"t. While the problem ofse f *** ^ *" 1OanW°rdS Can be ^^izedset of statuses. Each category could then be givenH. CHIMHUNDU 89the same treatment and designations such as 'slang' or 'Chimanjemanje' or'ChiHarare' could then be used in the dictionary. At least this would bebetter than arbitrary selection and would provide better guides for usage,as many users of the dictionary (e.g. in schools) may take the view thatwhat is in the dictionary is acceptable and what is not is unacceptable.5.0. UsefulnessIn conclusion, a few remarks may be made regarding the dictionary's use-fulness to first and second-language speakers. Again the views expressedin this section are impressionistic.The first point to note is that meanings are given in English. A commoncriticism is that the dictionary is only useful to English-speakers, or thatat any rate a good command of English is a prerequisite for profitable use ofthe dictionary. It is claimed that the average first-language speaker oftenfinds himself looking up 'the meaning of the meaning' in an English diction-ary. A further suggestion is that a Shona dictionary must also give the mean-ings in Shona.In respect of these comments the writer has come to two conclusions:(a) that a closer examination of the dictionary and of the Shona speechcommunity as a whole shows that the criticisms are not very valid;and(b) that a dictionary of Shona 'in Shona' is not a practical propositionat the moment, unless only a small, simple volume is envisaged.The following entries, taken at random and reproduced as in thedictionary, should suffice to illustrate both points:(i) munzungu [LLH] KoZ n 3sp small tree: Cassia singeana. cp kaudziudzi Ko (B); mudyamhunguK; mutandanyoka M.(ii) -shanangura [L]K vt Choose. Select, cp -sarudza K Ko MZ(iii) tsapata [LLL] KZ n 9 Worn-out sleeping-mat, cp rutsekete M. Tsapatarukukwe hazvienzani nokuvata pasi: An old sleeping mat is not asbad as sleeping on the ground (i.e. better an inferior article thannothing at all) prov.The descriptions that follow the entries are clearly intended to be use-ful to everybody: the tone patterns, form classes and English meanings aregiven mainly for the benefit of the second-language speaker; cross-referencesto equivalents and near equivalents in different dialects (excepting Ndau,Kalanga and Rozvi; see 2.0) are given mainly for the benefit of the first-language speaker; and in many cases, examples of use in sentences aregiven for the benefit of both. The main body of this volume (pp. 1-757) isof great value to all categories of speakers. This 'Shona-English' section,then, is intended for the general user.90 INCORPORATION OF LOANWORDSIt is the supplementary sections which are intended for specific groups.The 'English-Shona index' and the 'English-Shona index to ideophones' areof use only to the second-language student of Shona. One cannot thinkof situations in which a native Shona speaker may look up items in thesesections unless he is translating from English.29 Then there is the 'Index toBotanical Names' which the specialist will find particularly useful. There isalso a lot of grammatical information for the student of Shona in the descrip-tions given after the Shona entries in the main section and in the tables of'Class Prefixes, Predicatives and Qualificatives' and of 'Verb Forms' both ofwhich are given in the preface. Of particular use for teachers and studentsalike at the secondary school level (and at an even higher level for those learn-ing Shona as a second language) are the 'Guiding Principles for WordDivision in Standard Shona' also given in the preface.30As for the suggestion that a unilinguial Shona dictionary is a must, thefollowing observations should be noted. Firstly, the native Shona speakerwho finds himself having to consult a dictionary is presumably literate inboth Shona and English. English is a compulsory subject at school and intheir everyday lives Shona people generally find themselves compelled by avariety of pressures to learn and use some English. Therefore, to the extentthat all its users are likely to be literate in English, the dictionary's usefulnessto the first language speaker (even without considering the cross-referencesin Shona) is greater than might appear to be the case.Secondly, it must be pointed out that the statement of meanings is theweakest point in the study of any language. Meanings are very unstable.Theoretically, one hundred per cent precision in the statement of meaningswould require advancement of the natural languages to a stage where ascientific classification of the entire 'speaker's world' Š including emotionsand concepts ŠŁ would be possible. This is an ideal which is impossible torealize. Therefore, in practice, meanings of linguistic forms are given bydemonstration (e.g. 'this is ...'), by translation (e.g. from English to Shona)and by circumlocution (i.e. in the manner of a unilingual dictionary), ex-cept where definition is possible in terms of some science.31On a unilingual Shona dictionary, the options would be circumlocution,where indigenous forms are concerned, and translation, mainly from English.But owing to the lack of equivalents, such a dictionary must allow a consider-able amount of partially assimilated English forms to explain new concepts2» Or, unless he belongs to a family which has gravitated towards what W ELambert describes as 'the other linguistic-cultural group', 'A social psychology of bili-ngualism' in Whiteley, Language Use and Social Change, 104.so Given as originally set out by G. Fortune in A Guide to Shona Spelling (Salis-bury, Longman, 1972).»< Cp. Bloomfield, Language, 140.H. CH1MHUNDU 91and technical terms,32 Even then it is doubtful if such a form of Shona couldcope with all the codes of natural sciences. Such an attempt would producea huge volume which would still fail to satisfy the purists among the pro-tagonists of a unilingual Shona dictionary.This being the position, the writer feels that a bilingual dictionary likeHannan's (with improvements, of course) is the most practical solution forthe moment. That a unilingual dictionary is desirable is not contestedhere Š in fact it is being taken for granted. There is no reason why firstattempts cannot be made now. But any such attempts should be realistic.Perhaps a start could be made with a small, simple dictionary without tech-nical terms. Improvements could be made in the future. As the languagedevelops it should become capable of describing more and more conceptsand subjects. However, this is a process that must be awaited rather thanprescribed, and present indications are that this process will be slow.s* Gp. Nida and Wonderiy, 'Communication roles of language*'. 57-74, onlanguages of specialized information.