Zambezia (2000), XXVII (i).THIS IS THIS, AND HERE ARE SOME EXAMPLES:VERBALISERS AND EXTENSIONS IN SHONAA. A. JEFFER1ESDepartment of Linguistics, University of ZimbabweAbstractAfter a brief discussion of the goals of scientific theory, especially inLinguistics, the article explores the scientifically arrived at morphologicalcategories, 'verbaliser' and 'extension' in Shona. It reviews existing descriptiveaccounts of these categories and suggests that, if they were treated asmembers of a single category rather than as they have been treatedtraditionally, a variety of previously unanswerable questions could begin tobe addressed through new research. The proposed recategorisation wouldtreat 'verbaliser' as the less productive members of the set of tonelessderivational morphemes which, otherwise, consists of 'extensions'.As we engage in the scientific study of language, some of us linguists [inLangacker's (1975) terms, the 'theory people')] are predisposed to focusour efforts on using linguistic data for building theories, and others of us(in Langacker's terms, the 'language people') prefer to direct our effortstowards using theories to understand more deeply the linguistic data weobserve. Still, even though we spend much of our time being immersed intheories of one sort or another, we often do not spend much time thinkingin general about what scientific theories are or what it is that they do forus Š perhaps, more precisely, to us Š in our explorations of languageand linguistic systems. I want to share some of my thoughts about someof these things.1The general characteristics of scientific theories are well known andaccepted. 1) They account for (or describe and explain) precisely andaccurately what they are supposed to account for in coherent, internallyself-consistent and testable ways; 2) they do not yield information thatclashes with knowledge gleaned from related scientific enquiries; and 3)they are marked by simplicity, or 'elegance'. Despite these agreed-uponcharacteristics, an interpretation of 'scientific theory' is not monolithic,I am taking responsibility for this article because its wording and any mistakes in it aremine. Still, it was inspired directly by two Departmental seminars presented in the pastyear by Mr K. G. Mkanganwi, and indirectly by the on-going work being undertaken by theALLEX (African Languages Lexical) Project which has produced a general, monolingualcomputer-aided Shona dictionary (Chimhundu, 1996) as its first step towards documentingthe nation's languages. As the article developed, it was fed by insights offered especiallyby Mr K. G. Mkanganwi, Mr K. Mawomo, Mrs M. Mawema, Mr E. Chabata and Professor N.C. Dembetembe.12 VERBALISERS AND EXTENSIONS IN SHONAand there are a number of different scientific frameworks within which'what it is' is defined. For my part, I look at 'scientific theory' the way thatPopper (1965; 1979) does. According to Popper, in addition to thesegenerally accepted characteristics, an essential characteristic of ascientific theory is that it addresses 'true' problems that cannot beaccounted for by existing scientific theories.He also says that, because it is essential for scientific theories to betestable, it is characteristic of scientific theories to contrast only withbeliefs (like those found in fields such as philosophy of science), and notwith other, 'non-scientific' theories which, in his view, do not exist. Forhim, and for me, if a theory is testable in scientific ways, then it isscientific; if it is not, then it is a belief.But what is at least as important than thinking about theircharacteristics is thinking about what theories are for Š what they aresupposed to do. One of the primary goals of scientific theories, at leastaccording to Popper's view, is to yield new facts about what is going onand how things are. My favourite way of showing this is drawn fromGeology, in the area of tectonic plate theory. Before tectonic plate theorybecame accepted in the late 1950s, the facts we had about things geologicalincluded descriptive facts regarding the coincidental shape lines of theEastern coast of Africa and the Western coast of South America. The factswe had about things biological included descriptive facts about sharedplants and creatures, plus such speculations as 'the creatures and plantsthe two continents had in common must have swum or flown across Š orbeen blown or carried across Š the ocean that separates them'. Aftertectonic plate theory became accepted, we had a variety of new facts,facts which explained more about these things, and others as well. Popperwould predict that, one day, something, an anomaly, will be discoveredthrough tectonic plate theory that will create a problem for it, and thatthinking about the anomaly will inspire a new and better theory that willyield new facts and better explanations of what is going on and abouthow things are with respect to things like continental drift.Another primary goal of theories, again according to Popper, is tobecome refuted Š to be proven, through what he calls 'recalcitrant data',as being mistaken. He says that we can only get closer to truth with ourscientific facts if we keep getting better and better theories that comecloser and closer to being true. In other words, we should aim to havetheories that have the potential for being falsified by data, rather thantheories which can only be confirmed or theories which, under themistaken belief that the theories are commensurable with one another,can even distort the data when our goal is to discover a theory that ismore 'elegant' than another that accounts for the 'same' things. Further,he says, it is our responsibility as scientists to keep stretching in theA. A. JEFFERIESdirection of truth, even though we know in advance that we will never beable to actually capture it through combining our data with our theoriesand treating them in scientific ways, and even though we also know thatour theories may not yield any new facts at all and/or that they might bewildly mistaken and/or that they are actually beliefs because they turnout not to be testable against the data we find in the 'real world'.Nonetheless, we remain with our responsibility as scientists to use ourminds and our training to get closer to truths that, if they are reachableby people at all, are reachable not by our minds, but by our spirits.I would also like to make two other observations about theory becauseof their relevance to the topic of this article, the problematic categorisationof verbalisers and extensions in Shona. One of these is that the theoryyou have about something at least shapes what you see in that something.The other is probably the same observation, but it is applicable to othercontexts: when a theory about something has been accepted, theacceptance itself has an effect on what can be seen in that something bythose who have accepted it. I will be using parts of Fortune's (1984)descriptive analysis of Shona verbalisers and extensions to show what Imean by 'accepting a theory has effects on what can be seen aboutShona', because Fortune's analyses of Shona, in one way or another, havelaid the groundwork for almost all of the work that has been done on thelanguage, and most of the facts we have about Shona have arisen out ofFortune's analyses. For this to have happened, Fortune's work must havebeen, and must continue to be, both acceptable and accepted by scientistswho base their descriptive and theoretical work on the categories heestablished, and I continue to believe that this groundwork deserves tobe so because Fortune's work covers so much, so carefully and in somuch detail.Still, I claim that accepting Fortune's theories about how Shona isstructured has consequences. The most important of these consequences,given the influence of the work, is that Fortune's analyses facilitate ourseeing certain facts about Shona, and they prevent our seeing a variety ofother things that are at least potentially facts.I suspect that they do so, not only because 'accepting his theoryabout something constrains what you can see in that something', but alsobecause of his style of presentation, which seems to me to be patternedin such a way that it is easy to believe that whatever he says is receivedwisdom; that is, it is easy to see his analyses as being correct, as beingtrue, as being what there is to say about the something in question. Whathe shows us through his work tells us something like 'Here is the name ofthe category, here is the list of things that fit it perfectly, and here are lotsof perfect examples. As for the imperfect examples I have which requiremore research and deeper understanding, here is my note that says theyare imperfect examples.'4 VERBALISERS AND EXTENSIONS IN SHONAThe pattern could be abstracted out and abbreviated as: 'This is this,and here are some examples', and what it does is to demonstrate theexistence of the category through the listed examples. What we accept,and what shapes our own work and our own thinking about Shona, is theset of facts the analysis yields; in other words, those facts about Shonawhich comprise the theoretical categories Fortune posited for Shonacombined with the data he amassed so responsibly. Having been exposedto his analysis, we see that there is a category 'verbaliser',2 which consistsof certain things, and that there is another, quite different, category,'extension', which consists of certain other things. Having accepted hisanalysis, we come to believe in the existence of verbalisers and extensionsin Shona, and our beliefs in them shape all sorts of other things webelieve about the structure of the language.Even so, the question could arise: 'Does Fortune's use of the "this isthis" approach draw us into digging more deeply into the structure of thelanguage so that we can become even more aware of it, or does it,because we have memorised "his facts" about the language, actually havethe effect of preventing us from digging more deeply into it?' This is aquestion that Mkanganwi has raised indirectly (see his article in thisissue). It is a good question, and one which I would like to ask again in thecontext of Shona verbalisers and extensions.Through Fortune's analyses, we have come to accept the idea thatverbalisers and verbal extensions in today's Shona are two differentcreatures. One creature, the verbaliser, brings words from a variety ofword classes into the word class 'verb'. For instance, see the exampleslisted in 1.1. VERBALISERS3Other Word Classes Verbsparu 'tearing' (IDEO) -paruka 'get torn'-parura 'tear' (something)svi-i 'be dark'(IDEO) -sviba 'be dark'makoko 'pot scrapings' (N) -kokota 'clean pot with finger,consuming remnants'shamwari 'friend' (N) -shamwaridza 'be friendly towards;befriend'-kobvu 'thick (ADJ) -kobvumara 'be thick'2 According to Dembetembe (1987, 9), the use of the term 'verbaliser' in linguistic studiesof Shona comes from Gowlett's (1967) work on Lozi.3 Abbreviations used in this work are as follows: IDEO Mdeophone"; N 'noun'; ADJ 'adjective;s/o 'someone'; s/th 'something'; FV 'final vowel'; ADV adverb; V 'vowel'; C 'consonant'.A. A. JEFFERIES 5On the other hand, the verbal extension does something else: itcreates different types of verbs out of words that are already verbs, as,for instance, those listed in 2:2. VERBAL EXTENSIONSVerbs Other Verbs-ita 'do' -itika 'happen'-sunga 'tie, bind, arrest' -sungwa 'be tied, bound, arrested'-netsa 'annoy' -netsana 'annoy one another'-chengeta 'keep, guard, look -chengetedza 'look after'after, preserve'-bvuma 'agree, acknowledge', -bvumira 'allow, permit'admit, assent'-bata 'touch, hold, seize, take -batisa (intensive) 'hold firmly',captive, acquire, practise' (causative) 'cause to hold'Fortune's analysis here is generally congruent with Guthrie's (1962)tripartite distinction between verb stem types in Bantu languages. InGuthrie's view, there are simplex verb stems which consist only of roots,or radicals, complex verb stems which consist of a radical plus a fossilisedand unanalysable suffix, and extended radicals which are those verbswhich are derived with an analysable suffix. These latter, generallyproductive, affixes comprise most of what we would call extensions inShona and the former, the fossilised ones, comprise mostly what wewould call verbalisers.Although they function in ways that accomplish different things, theextension and verbaliser creatures are similar in one major respect: theyboth provide Š or have provided Š means through which new words,have been Š and/or can be Š created. This point does not come up veryoften when we focus our attention on the syntactic relations associatedwith several of the verbal extensions, for example, as part of a claim thatit is the syntax which brings them into the verb, because, when we dothat, it is easy to see extensions as inflections which get attached to anold word, 'the base', so that 'the base' can fit into new syntactic patterns,rather than as derivational kinds of things which create new lexical items,or words, that name different kinds of events. The syntactic view of thesemorphemes is something that Fortune would agree with,4 but I am not yetSee, for example, his (1984, 22) description of extensions in general: 'Extensions aresuffixed to the R [radical] they extend. Since it is properly VPs [verb phrases] and notVERBALISERS AND EXTENSIONS IN SHONAconvinced of it. The addition of some verbal extensions predictably addsto, subtracts from or otherwise modifies the argument structure ofsentences, yes, and, by doing so, could be seen to be functioning in somesort of inflectional and grammatical way. But, even so Š at best Š thisapproach yields only part of the picture. Not all extensions bring aboutdifferent argument structures in sentences, and that is not all that any ofthem do, as Khamisi (1985), for example, shows with respect to Swahiliextensions, and as you can see in the Shona examples given in 3:3. EXTENSIONS AS DERIVATIONAL MORPHEMESOld WordsNew Words-ziva 'know'-sunga 'tie, arrest'-cheka 'cut'-rara 'sleep*-daidza 'call'-shuma 'hand over s/th to s/o'-tya 'fear'-ziv-isis-a 'know very well'Root+Intensive+FV (no structuralchange; meaning shift)-sung-unur-a 'untie, unarrest'Root+Reversive+FV (no structuralchange; meaning shift)-chek-erer-a 'cut into tiny pieces'Root+Perfective+FV (no structuralchange; meaning shift)-ra-dz-a 'put s/o in a grave'Root+Causative+FV (structuralchange; meaning shift)-daidz-ir-a 'call out loudly'Root+Applied+FV (given this senseof the verb, there is no structuralchange; meaning shift)-shum-ir-a 'worship and praise' (e.g.,God) Root+Applied+FV (unexpectedstructural change; meaning shift)-ty-is-idz-ir-a 'frighten'Root+Causative+Causative+Applied+FV (not as much structural changeas would be expected; meaning shift)merely Rs which are transformed by extensions, the same question occurs here asoccurred in the treatment of Rs derived from ideophones. Are extended Rs morphologicalconstructions which can be properly treated at the level of the R? The answer here, asthere, appears to be that extended Rs, which are morphological constructions with ICs[immediate constituents] consisting of (1) the extension and (2), in whole or in part, theR of the transformed VP, do result from the transformation as the result of a reconstructingof relationships with the extended VP.'A. A. JEFFER1ES 7The recent work developed by Mkanganwi (1995) and Chabata (1997)on Shona derivational morphology shows that we come closer tounderstanding 'what is going on and how things are' by explicitly treatingextensions as derivational morphemes.If here too we focus on the idea that the set of verbalisers and the setof extensions comprise sets of derivational morphemes, we need areminder of what derivational morphemes are said to do if we are goingto explore them more deeply. Katamba (1993, 47), for example, says thatsuch morphemes form new words either 'by changing the meaning of thebase to which they are attached' (as Fortune would describe the functionof verbal extensions in Shona); or 'by changing the word class that a basebelongs to' (as Fortune would describe the function of verbalisers inShona). This is a pattern we find quite easily in a wide variety of languages.In English, for example, a derivational morpheme like -ly (as in happily,ADJ to ADV) or -ness (as in happiness, ADJ to N) serves productively tobring 'outsiders' into a particular word-class, but a derivational morphemelike un- (e.g., unhappy, ADJ to ADJ) or -ish (e.g., happyish, ADJ to ADJ)affects 'insiders', words that are already in the word class in question. Ina language like English, the two categories of derivational morphemesproductively operate in mutually exclusive environments: either only tobring outsiders 'in', or only to modify insiders that are already 'in' there.The verbalisers given by Fortune (1984) Š which he says serve tobring words from other word classes into the word class verb Š aredescribed as consisting of single consonant, or C, morphemes, vowel-consonant, or VC, morphemes, or combinations of morphemes, as appearsin 4, which shows his listing of them.4. FORTUNE'S (1984, 13-21) VERBALISING MORPHEMESOccur with Ideophones-b--dz--k--m--mb--n--nd--nh--ng---ny--p--r--t--V--z--ik--ir--k-ir--m-ar--m-adz--m-ar-adz-t-er--dz-ir--dz-er-Occur with Adjectives-k- -v- -mar--p- -s-VERBALISERS AND EXTENSIONS IN SHONAOccur with Nouns-k- -t- -mar-r- -dz- -dz-an-List 5 presents my annotated version of Fortune's (1984) list of verbalextension morphemes Š or those derivational morphemes which hesays make new verbs out of 'old' verbs:5. FORTUNE'S (1984, 23-30) VERBAL EXTENSIONS-at-, -et- (contactive [indicating contact]; no longer productive)e.g.,-sungata 'tether'; cp -sunga 'tie'-chengeta 'preserve, take care of; cp -chenga 'keep'-ik-, -ek- (extensive [the action named by the unextended root isprojected in space]; no longer productive) e.g.,-rumika 'bleed'; cp -ruma 'bite'-tondeka 'point to'; cp -tonda 'face'-is-, -es- (intensive [heightened, more vigorous or effectiveaction]; can be reduplicated) e.g.,-batisa 'hold firmly'; -batisisa 'hold very firmly'; cp-bata 'hold'-irir-, -erer- (perfective [action/process named by the unextendedroot is carried through to completion, sometimes, evento excess]) e.g.,-sekerera laugh on and on'; cp -seka 'laugh';-dyiririra 'eat a lot'; cp -dya 'eat'-urur-, -oror- (repetitive [repeated action or action completedthoroughly]) e.g.,-tukirira 'scold/curse over and over'; cp -tuka 'scold';-ononora 'see s/th clearly, scrutinise; cp -ona 'see'-ar-, -arar- (stative [state as a result of some action/process]; nolonger productive) e.g.,-omarara 'become dry, hard'; cp -oma 'be dry, hard'-an- (associative [action/process/state named by theunextended verb is 'common to a group']; no longerproductive) e.g.,-ungana 'assemble together', cp -unga 'collect*-ik-, -ek- (potential [action is possible/easy; state resulting froma process or action]) e.g.,-goneka 'be feasible'; cp -gona 'be able to';-itika 'happen'; cp -ita 'do'A. A. JEFFERIES-w-, -iw-, -ew- (passive [action/process named by the unextendedverb is undergone; impersonal and locative subjectscan occur]) e.g.-piwa 'be given'; cp -pa 'give';-bikwa 'be cooked'; cp -bika 'cook'-an- (reciprocal [action named by the unextended verb isperformed mutually by agents upon one another]) e.g.,-dana/-danana 'like/love/need one another'; cp -da 'likelove/need'-ir-, -er- (applied [action/process named by the unextended verbtakes place 'with special reference to some person, thing,event or place, which is the complement of the extendedverb phrase']) e.g.,-wira 'fall into, onto'; cp -wa 'fall from';-fanira 'be fitting'; cp -fana 'resemble';-chengetera 'look after for s/o'; cp ' -chengeta 'look after'*-y-, -idz-, -edz- (causative type (a): ['participation by an agent in someactivity; a personal participation involving action by anagent on someone or something']; the *-y- allomorphnever appears; a variety of allomorphs [-dz-; -ts-; -sv-;-zv-; -nz-; -idz-] depending on the preceding consonant;occurs with a limited number of roots, notably with thoseending in -k, -r, -p, -b, -v, -mb, -nd, and -ng) e.g.,-rambidza 'forbid'; cp -ramba 'refuse';-fadza 'make s/o happy'; cp -fara 'be happy';-mutsa 'wake s/o up deliberately'; cp -muka 'wake up'-is-, -es- (causative type (b): [someone is made to act, or someoneis made to be acted upon; causation is relatively indirect;can also add an instrumental sense; has the potential toadd an argument]) e.g.,-farisa 'greet'; cp -fara 'be happy';-garisa 'make s/o sit'; cp -gara' be seated, stay at, live at';-mukisa 'make s/o wake up without intending to'; cp-muka 'wake up'An examination of the verbaliser morphemes in 4 and the extensionallomorphs in 5, shows that the forms of at least some of the verbalisersresemble the forms of at least some of the extensions, particularly withrespect to the consonants they contain.This is part of the observation that got me thinking about developingthis article. The rest of the observation is demonstrated in a list ofrelated sets of words which Fortune (1984) presents and which I havereproduced as 6:10VERBALISERS AND EXTENSIONS IN SHONA6. FORTUNE'S (1984, 16) LISTING OF VERB PAIRS DERIVED FROMIDEOPHONESIdeophoneparu'tearing'simu'rising'pfudugu'uncovering'kwachanu'coming away fromwall'bwodo'dislocating'undu'moulting'tepfenu'slackening'tasanu'being straight'mwau'breaking'nyandu'disordering'Verbs Derived withVerbaliser -k--paruka'get torn'-simuka'rise'-pfuduguka'get uncovered'-kwachanuka'fall from wall'-bwodoka'get dislocated'-unduka'lose feathers'-tepfenuka'be slack'-tasanuka'be straight'mwauka'get broken, torn'-nyanduka'be dishevelled'Verbs Derived withVerbaliser -r--parura'tear (s/th)'-simura'lift (s/th)'-pfudugura'uncover (s/th)'-kwachanura'take (s/th) fromwall'-bwodora'dislocate s/th'-undura'pluck (s/th)'-tepfenura'loosen (s/th)'-tasanura'straighten (s/th)'-mwaura'break, tear (s/th)'-nyandura'disorder (s/th)'Fortune's observation about these related forms is that -k- and -r-,respectively, bring intransitive and transitive meanings to the ideophonicbase, but my observation is: isn't it interesting that -k- brings an intransitivesense to its set of derived verbs, and that -r- brings a transitive sense toits set of derived verbs, since that is not very different from what the -ik-/-ek- and -ir-/-er- extensions do with extended verbs?To test this as an idea that could have possibilities, let us look at theeffects of the potential -ik-/-ek- and the applied -ir-/-er- have on someunextended verbs. In 7, I have selected a few random samples that Iexpect will not serve as 'perfect examples'.A. A. JEFFERIES117. SAMPLE VERBS EXTENDED BY THE POTENTIAL AND THE APPLIEDEXTENSIONSVerb StemVerb Stem plus Potential-ziva 'know' -zivika 'be knowable, get known'(Fortune, 1984, 25)-gura 'cut off; cut to pieces' -gurika 'be broken, as in dried things'(Chabata, 1997, 53)-netsa 'annoy' -netseka 'be troubled'(Kangira, 1997, 28)-rova 'beat' -roveka 'be beaten' (Kangira, 1997, 28)-tora 'take' -toreka '(able to) be taken'(Kangira, 1997, 28)-ona 'see' -oneka 'be seen; visible'(Kangira, 1997, 28).Verb Stem(Hannan, 1981)Verb Stem plus Applied(Hannan, 1981)-rova 'beat'-gura 'cut off; cut across'-oma 'become hard, dried,parched, difficult'-gara 'be seated, live at,stay at, last, stay away from'-mira 'stand, wait'-tora 'take'-rovera 'fasten by striking; nail'-gurira 'take a short-cut to/from'-omera 'adhere (by drying onto s/th)'-garira 'lie in wait for X, ambush X,watch over X, watch for X, sit on'-mirira 'wait for s/o, s/th'-torera 'take for or from s/o'Even for those examples that do not demonstrate perfectly what I amseeing here, what is most striking to me is that the verbalisers -k- and -r-in 6 signal differences between the concepts named by the ideophonesand those named by their derived forms that are at least parallel with thedifferences signalled between the concepts named by the unextendedverbs and the concepts named by the verbs which have been extendedby -ik-/-ek- and -ir-/-er-, as we see in 7. Specifically, the -k- verbaliser in 6signals the addition of the potential's sense of 'result of action or process',and the resulting intransitive verb is usually unaccusative (see Kangira,1997, for discussion), with a patient or experiencer or theme as subject,just as the -ik/-ek- extension does in 7.12 VERBALISERS AND EXTENSIONS IN SHONAThe verbaliser -r- in 6 signals a transitive sense, with an agent orinstrument as subject, and it signals that an object must be includedamong the arguments associated with the derived form, in parallel withwhat the applied extension signals in at least some sentences.5There is a problem with this similarity. Morphemes are not supposedto overlap that much. They are supposed to signal different units offunction and meaning with different phonological shapes. If the shapes ofeach pair of these verbalisers arid extensions resemble one another somuch, and if they function to derive new words and they mean suchsimilar things, then why is it necessary to place them in two completelydifferent categories, as if they could not possibly be uses of the samemorpheme in somewhat different contexts?The most obvious reason could be that the derivations in 6 areexceptional derivations and that what 1 have shown with Fortune's set ofexamples departs from the norm. But closer examination suggests thatthis may not be the case. In the appendix, I have listed all of Fortune's(1984) examples of verbalised ideophones which take -k-, -dz- and -r- (i.e.,those verbalisers which resemble most closely what appear to be currentlyproductive verbal extensions), so that you can examine them on yourown. Briefly, the words derived with -k- and -dz- seem to me to be fairlystraightforward: -k- (intransitive; result of process or action; potentiality);and -dz- (causing/making a sound or a sight which is describable by anideophone). With the -r- derived forms, each instance of-r- seems to markone of three things: bringing transitivity to the stem, requiring thespecification of direction-towards/away-from in the argument structure(both like the applied extension) or expressing positionality Qike Fortune'sold, now non-productive, -ar- extension). What it does not seem to markwith these ideophone bases is the benefactive/malefactive relation thatis associated with many -ir-/-er- extended verbs, something that shouldbe explored more deeply later on.If these three Šk-, -dz-, -r- Š turn out to be productive 'verbalisers'these days (and we will need to find out if they are), then at least for thesemorphemes, there is reason to see Fortune's split categories as actuallybeing a single category which derives verbs. Sometimes, they help tobring 'new' words into 'verb' with specifiable characteristics, and,sometimes, they make 'new' verbs out of 'old' verbs with at least some ofthe same specifiable characteristics. In this respect, they are unlike those5 The forms derived by the -iy-/-er- extension may not be quite so obviously parallel withthe forms containing verbalisers as they are with -k- and -lk-/-ek-. A possible reason forthis is highlighted by Myers (1984, 192) when he observes in another context thatideophones, unlike base verbs,"... have no argument structure", a claim that will need tobe examined carefully in future research that takes into account their collocations withintroducer phrases.A. A. JEFFERIES 13'extensions' like -an-, -w- etc. which serve only to make 'new' verbs out of'old' ones, and they are also unlike those 'verbalisers' which are nolonger used productively to bring words from other word classes into theword class 'verb'.This last observation leads me to another reason that could explainwhy Fortune's verbalisers and extensions have been treated as if theywere different creatures. It could have come out of that underlyingassumption about what derivations do which is captured in the 'either'class-changing 'or' class-maintaining idea Katamba (1993) expresses inhis discussion of derivational morphemes. Maybe the relation is morelike 'and/or', as Crystal (1991) and Robins (1989) suggest. It could be thatat least -k-, -dz- and -r- serve in both class-changing and class-maintainingfunctions, and Shona could be a language that presents a clearer case ofthis dual function than English does.6Regarding some of Fortune's other verbalisers, I wish someone wouldstart some historical and comparative work so we can find out moreabout which of today's verbalised ideophones and nouns represent frozenforms: the results of having been derived by yesterday's productiveextensions. That -at- (contactive), -an- (extensive) and -ar- (stative) arelisted by Fortune as synchronically non-productive extensions issuggestive, especially when they are clearly at least related to, on bothsemantic and 'shape' grounds, today's -t- verbalised forms, and at leastsome of the -n- and -r-, verbalised forms, respectively. The stative, -mar-,which is possibly a combination of currently non-productive morphemes,-(a)m and -ar-, as Fortune (1957, 224-5; 1984, 15) notes, could also havebeen productive in Shona at some point. When some research has beendone, it may turn out that many verbs that we now see as beingproductively derived with 'verbalisers' are actually left-overs from anearlier time.At least some of the others could be as productive as -k-, -dz- and-r-. The -v- verbaliser, for example, could be a particular manifestation ofthe morpheme -va/-ve 'become', as in -nyorova 'become soft' (cp -nyoro(ADJ) 'soft') and -fovova 'become shrivelled up' (cp fovo (IDEO)'shrivelled'. We might even find that the consonants found in some of the6 Even English, however, provides a non-productive example. Robins (1989, 242) notesthat, although English -Iy usually serves with a class-changing function, it also has a class-maintaining function with a few words. He gives kindly as an example where ADJ+lyyields either another ADJ (e.g., a kindly act, with a sense that is closer to its source word,like, which is also found in N+ly ADJs like manly) or an ADV (e.g., she acted kindly, withthe [more generalisable and less relevant sense], 'manner of). Bybee (1985, 82-4), likeKatamba (1993), distinguishes between the two types of derivation, using how the relevancecriterion applies to each to separate them. Her approach to morphology in general,however, stresses the idea of gradience (e.g., 'inflectional' and 'derivational' are gradient,rather than bipolar, characteristics). As a result, it is difficult to say whether or not shewould agree that class-changing and class-maintaining are bipolar types.14 VERBALISERS AND EXTENSIONS IN SHONAother auxiliary verbs (i.e., -na, 'be with' -ri, 'be' -nga, 'be able' and -ti 'say')participate in 'verbalising' as well . . . something that is reminiscent ofGivon's (1971) 'far-fetched hypothesis' which claims that Bantu extensionsall originally came from verbs.My final observation about the consonants serving a 'verbalising'function is that the *-y- extension shown in 5 above seems to modify a listof root-final consonants (i.e., -k, -r, -p, -b, -v, -mb, -nd, -ng) that, except for-nh, includes the rest of the verbalisers listed in 4 that I have not evenspeculated about here. Only a new theory tested against relevant datawill help us know more about what is going on here.If the idea is correct that some morphemes function in ways thatbring words from other classes into the word class 'verb' Š i.e., 'verbalise'Š and create new words from verbs that are already verbs Š i.e. 'extend'Š then there are implications. The first has to do with the morphologicalshapes of several of the verbal extensions. In most descriptions, thepassive (-w-/-iw-/-ew-), causative (-idz-/-edz- plus *-y- modified stem-finalconsonants; -s-/-is-/-es-), applied (-ir-/-ei*-) and potential (-ik-/-ek-) extensionsare described as VC, with V representing -i- or -e-, since extensions aresubject to word-internal vowel harmony processes.7 Of these, the passiveand the causatives are the only extensions which have a C allomorph,especially in Zezuru where the 'contracted' C forms seem to be mostwidely used (see, e.g., Fortune, 1984, 25). The passive and the causativesappear as -w- or Ł*- (or some other consonant which has been modified bywhat Fortune refers to as -*y- respectively) if and only if, the immediatelypreceding sound can combine with it (or replace it distinctively) to forma syllable onset, and so long as the derived stem does not yield amonosyllabic word. If verbalisers are manifestations of the samemorphemes that give us Š or at one time gave us Š extensions, then Ithink it could be worthwhile to think again about the status of the vowelin the VC allomorphs of these derivational morphemes.What I am tempted to say about this is a claim that Myers (1994) hasalso made with reference to what he specifies as the causative morpheme,although he made it for different reasons than I do, and there isconsiderably more certainty expressed in his work than I am willing toexpress here without more research having been done. Given what I7 For discussion, see, for example, Myers (1987). It also occurs to me that the operation ofvowel harmony in affixation may suggest the presence of 'not-very-strong' vowels incertain affixes, which, in addition to depending on the nucleus of a neighbouring syllablefor vowel shape also depend on the tone of a neighbouring syllable for their surfacetones. This is an observation I make in Jefferies (1990) in connection with certainallomorphs of the copulative, 'possessive' and some of the adverbialising prefixes inShona, where, again, these inflectors depend on 'their' stems for vowel shape and howtheir underlying high tones are realised. I hope someone will explore these phenomena,especially in the context of Optimality Theory, where structure preservation is takenseriously, in the sense that what is generated by the linguistic system is seen to emergewithout modification.A. A. JEFFERIES 15know so far, I agree with Myers that the vowel in the -s- causative isneeded more for phonotactic reasons than it is for lexical identification.For the lexical identification of this morpheme, as well as for otherswhich have iC/eC allomorphs, all you need is the consonant, as thecommonly used passive, causative and Fortune's verbalised forms showin 8:88. C FORMS OF SOME DERIVATIONAL MORPHEMES IN CONTEXT-w- -sungiwa/-sungwa 'be tied, arrested'; cp -sunga 'tie, arrest';-batiwa/-batwa 'be held' cp -bata 'hold, touch' (Mkanganwi, 1995,71)-s- -batisa/-batsa 'cause to hold'; cp -bata 'hold, touch' (Mkanganwi,1995, 71)-potesa/-potsa 'cause to go around'; cp -pota 'go around' (e.g., ahill) (Mkanganwi, 1995, 71)-pfupisa 'shorten'Š cp -pfupi (ADJ) 'short' (Fortune, 1984, 21)-k- -svetuka 'jump'; cp svetu 'jumping';-bhururuka 'fly'; cp bhururu 'flying';-pflmbika 'ripen fruit' cp pfimbi (N) 'hole dug for ripening fruit'(Fortune, 1984, 18-21)-dz- -pururudza 'ululate'Š cp pururu 'ululating sound';-chochodza 'run, of a cock' -cp cho cho cho 'running, of a cock';-shamwaridzana 'be mutually friendly' cp shamwari (N) 'friend'(Fortune, 1984, 18-21)-r- -gomera 'groan'Š cp gome 'groaning sound';-marangura 'scratch surface'-cp marangu 'scratching surface';-kokora 'scrape pot'; cp makoko (N) 'pot scrapings' (Fortune,1984, 18-21).The data in 8 suggest that vowels are needed primarily for pronouncingthe derived forms. The C form of passives and causatives, for example,can occur only when the immediately preceding segment of the8 In Fortune's (1984) listing, the only productive extension containing a vowel that contrastswith another containing at least one -i- is -urur-, which is also subject to vowel harmony.What he calls the morpheme -irir- is not among those extensions which have contractedforms. Fortune (1957,220-1) also lists reversives -Vnur-a and -Vnuk-a, which mark valencein addition to 'reversivity' and where V is the same as the last root vowel. There, he says,they have contracted -ur-a and uk-a forms, but none of these appear in his 1984 listing ofextensions, since, in the interim, he had apparently re-analysed -(V)-nu- as 'belonging' tothe ideophone, rather than to the verb (see, for example, Fortune, 1984,153; for additionaldiscussion, see Dembetembe, 1987,11). You will see in 6 that -k- and -r-, even when affixedto -u- final stems, are interpreted, not as reversives/conversives of the meanings of thebase form, but as valence markers. As Mkanganwi in his seminar earlier in the year,pointed out, we need new ways of looking at these phenomena if we are going tounderstand them more deeply.16 VERBALISERS AND EXTENSIONS IN SHONAmultisyllabic stem, which is usually a consonant, is something -w-, *-y- or-s- can combine with to create an onset. Unless -k- and -r- are preceded bya vowel, they must occur in the VC allomorph form since they combinewith nothing to form syllable onsets. With Fortune's 'verbalisers', thesegment of the noun, ideophone and adjective which immediately precedesthese morphemes is usually a vowel. The 'verbaliser' itself, on its ownthen, would serve as a syllable onset in these derived forms.In going through Fortune's examples of verbalised forms which containmonomorphemic verbalisers (1984, 13-21), I noticed three patterns:1) the most typical is that the base appears in its full form, and theverbaliser combines with the final vowel -a to serve as the final syllableof the derived verb (as we saw in 8 above);2) in a few derived forms, the final vowel of the base form is dropped andreplaced with iC/eC; and3) in a few other derived forms, the final vowel of the base form isretained and the iC/eC verbaliser occurs as well.In 9,1 list most of the few iC/eC forms that Fortune (1984) offers asexamples.9. FORTUNE'S (1984, 113-21) iC AND eC VERBALISER MORPHEMES9Ideophone Verbs with -ik- or -ir-rukutu 'being weak' -rukutika 'be weak'vhukutu 'dashing to pieces' -vhukutika 'dash to pieces'parapata 'waking up suddenly' -parapatika 'wake up suddenly'tsakata 'vanishing' -tsakatika 'get lost, vanish'tande 'stretching out' -tandira 'stretch out'ndure 'stinging' -ndurira 'sting'dzimu 'extinguishing' -dzimira 'burn grass to prevent thespread of a fire'teka teka 'swaying from side' -tekaira 'sway from side to side' (no FVelision)rita rita 'walking aimlessly' -ritaira 'walk aimlessly' (no FV elision)Adjective Verbs with -es- or -ek--chena 'white' -chenesa 'whiten'-shoma 'few' -shomeka 'be scarce, insufficient'Note that, with some of these, elision occurs; with others, it doesn't. The examples alsoshow that coalescence does not occur with these forms (see -tekaira or -ritaira), thoughit could be argued that it does with -chenesa (from ADJ -chena). Regarding this lastpossibility, Fortune (1984, 21) notes that verbalised ADJs whose root vowel is mid takethe-eC- 'verbaliser morpheme', on false analogy with what would happen if the 'verbalisers'were 'extensions'.A. A. JEFFERIES 17With most of these, the final vowel of the ideophone is dropped, andthe ideophone's final syllable onset (-t, -nd, r, m or k) is a consonantwhich is listed as one of the verbalisers in 4 above. There is, then, at leasta possibility that the base forms are fossilised deverbatives which havebeen further derived as ideophones. Fortune does list a few others, butall of these are clearly preceded by other verbalisers, as you can see in10, wh'ch lists the rest of his iC/eC examples.10. VERBALISER + iC/eC DERIVATIONS IN FORTUNE (1984, 13-21)ko ko ko 'tying to; securing' -ko-m-er-a 'tie to; secure'go go go 'groaning' -go-m-er-a 'groan'fe fe fe 'blowing of wind' -fefe-t-er-a 'blow'tsvu tsvu tsvu 'sound for urging -tsvutsvu-dz-ir-a 'urge on a dog'on a dog'tiki tiki 'perspiring' -tiki-t-ir-a 'perspire'Especially in 10, the iC/eC 'verbalisers' are indistinguishable from'extensions', yet they are listed as examples of particular verbalisermorphemes. I am curious about this, since it is not characteristic ofFortune to list, without comment, examples of a particular categorywhich at the same time are potentially Hstable in another category, as isthe case with these forms.If the vowel in what I see here as iC and eC allomorphs is neededmore for pronunciation than for signalling part of a unit of derivationalmeaning, then we have an opportunity to test whether or not epenthesismight be in operation. The alternative, which may be equally plausible, isto see the -i- or -e- being deleted whenever it can be deleted: after othervowels (usually, but not always, as we saw in 9) and between consonantswhich can be adapted to serve as syllable onsets. Information whichcould help us choose the better of these two alternatives would need toarise out of something independent of them, and it would be good if thechoice of one or the other could solve a problem we see elsewhere in thegrammar.The tonal structure of verbs in their citation form might be an areawhere there is information that could help lead us to a choice. Unlikewords in other word classes, regular Shona verbs conform with tonalconfigurations, rather than, as is the 'usual' case, having tones assignedto each syllable, perhaps in accordance with principles that are not yetunderstood at all. In their citation form, regular verbs10 are classified as10 There are a variety of irregular verbs that do not have what I am referring to as the'regular' tonal configuration for verbs. Mkanganwi has pointed out that, in addition to theso-called 'defective verbs' like -dai (HL) and -daro (HL), there are verbs like -dzungira18 VERBALISERS AND EXTENSIONS IN SHONAbeing either H, with the first (up to a maximum of three) syllables of thestem carrying high tones and the rest being low, or L, with all syllablescarrying low tones. The syllables which carry the derivational morphemes,then, depend on the tone assigned to the root for their own tones. As aresult, they appear as underlyingly toneless morphemes, as they haveoften been described in the literature (see, for example, Fortune, 1985,21; Odden, 1981; 1984; 1986; Myers, 1987; Jefferies, 1990; Kenstowicz,1994).In the context of the content word lexicon of a level tone language,where it seems to be the case that tone is associated with morpheme-and word-sized meaning units (see, for example, Leben, 1978; Hombert,1986), this is curious: lexical tones are assigned to the syllables whichrealise the morphemes of content words in order to help signal differentwords. So why would the syllables carrying certain derivationalmorphemes in this level tone language be inherently and relativelysystematically toneless? So far, we do not have an explanation for that.We just know from observation that they seem to be toneless, and weknow that it is easier to handle them within generative frameworks if wetreat them as being toneless (see, for example, Kenstowicz, 1994, 332). Ifthere is not necessarily a nucleus in those syllables which carryderivational morphemes when lexical tone is associated with the word,then there is a reason for H-tone and L-tone patterns to spread on H andL verbs respectively. In other words, lexical tone assignment could bemade independently of the specification of syllabification.11 Conceivingof derivational morphemes that apply to verbs as C units which takeepenthetically inserted vowels whenever they are needed or wanted,may help us to explain, rather than merely describe, regular verb tonalpatterns.(HLH), -jirichanya (HHLH) and ^ogodoka (LHLH) which have citation form patternswhich, for at least some speakers, do not conform with the 'regular' patterns that aretypically described in the literature. Exploring how the tones of adopted and coinedverbs are treated in their citation forms, however, seems to expose the 'regular, productive'patterns, as shown, for example, in Chimhundu (1983).1 That this could be happening is supported in fast speech pronunciations of words likemasikati '(good) afternoon'. Even when the -i- of the syllable -si- weakens to the point ofdeletion, thereby contracting the word to three syllables (i.e. to ma-ska-ti), the pitch levelof the next syllable (-ska-) is not lowered. Rather, it maintains the same pitch level that itwould have had before 4- was deleted. Another, less direct, kind of evidence comes fromHombert (1986) who reports on word-game experiments conducted with speakers oflevel and contour tone languages. When asked to reverse the syllables of disyllabicwords, the level tone language speakers, unlike the contour tone language speakers,reversed the segmental parts of the syllables, but not the tonal contour of the word (inparallel with what happens with English word stress patterns with syllable reversals: e.g.,FOOTball to BALLfoot. In contrast, the contour language speakers constructed reversalstaking the pattern equivalent to *ballPOOT.) The conclusion drawn was that, while toneis associated with meaning units like words in level tone languages such as the onesunder study, it is associated with syllables in the contour tone languages of the subjectsof the study.A. A. JEFFER1ES 19We know from C. Harford's paper on disyllabification in Shona12 thatif there is need to epenthesise in Shona, the vowel of choice for doing sois the high front vowel /if. Is it a coincidence that the allomorphs of theverbal derivational morphemes I am referring to here Š potential, applied,causatives and passive Š use /i/ to derive verbs? Or could somethingelse be happening? Again, this is something that would need someresearch. If my observation about this turns out to be correct, then it maybring out those kinds of recalcitrant data for existing theories Š such asone that would claim that, as a matter of principle, either vowel harmonyor epenthesis applies to vowels, for example Š that move ourunderstanding of what is going on forward.The other implication of the 'extension equals verbaliser' proposalcomes out of my suspicion that positing 'verbalisers' as being differentfrom 'extensions' is probably one of the many things that make it difficultfor us to explore the role of ideophones in the grammar. Fortune's (1962,1971; 1985) approach to ideophones is that they are used 'expressively',that is, that they exist at the phonaesthemic edges of the language. Insupport of this, he argues that there are limited class-maintainingderivational processes which apply to ideophones, that they do not takeinflections, that their internal phonology does not conform with 'regular'phonology, and that ideophonic constructions tend to interrupt the'regular' downdrift patterns of 'regular' sentences and are used to signala dramatic (as this is opposed to a 'normal') style of speech. Although Iam not aware of a statement that makes the claim succinctly,13 I sensethat Fortune sees 'verbalisers' as those morphemes which 'bring'ideophones under the control of 'the regular grammar' of Shona. So far,the only work that begins to look at this highly productive class of wordsas comprising 'proper lexical items' that operate within the grammar ofShona is Tafangombe (1997). Perhaps the 'verbaliser equals extension'view will also help us take ideophones into account, especially when weare exploring the grammatical properties and/or lexical projections ofShona words. In addition, when we study things that have to do with'extensions', such as morphological causatives or transitivity or object12 C. Harford (1998) "Disyllabification in Shona", a seminar presented to the Department ofLinguistics (Harare, University of Zimbabwe).13 He sometimes comes close to making such a statement, but has not, to my knowledge,actually done it. For example, 'There are many correspondences in form betweenideophones on the one hand and verbs and nouns on the other where ideophones formpart of a verb or noun in various ways. These seem to be clear cases of derivation andthere are a series of governing suffixes for deriving verbs (Fortune, 1962, 21)'. He goes onto note that, 'Ideophonic verbs and nouns are, of course, inflected normally, and are partof formal speech. Whereas the ideophone may have exaggerated features of pitch andlength as compared with those of normal speech, these are not found in the correspondingderived verb or noun forms' (Fortune, 1962, 24).20 VERBAUSERS AND EXTENSIONS IN SHONArelations, we may need to redefine what the relevant data is and examinedeideophonic verbs in addition to those extended verbs we currentlylook at as a matter of course.I hope that scholars will undertake some research in this area. At amore general level, I hope the questioning I have done here encouragesscholars to question the facts we have about Shona and to take newquestions about them into account. The questions I have asked hereaddress real problems, given my goal of trying to understand what isgoing on and how things are, and if the observations I have made yieldsomething that helps us move our understanding forward, I would bepleased. If they end up yielding nothing useful, or yielding only mistakenimpressions, I would also be satisfied, as long as the research that wouldtest them is actually carried out. In either case, we will have learntsomething.I have demonstrated something by using Fortune's selection ofexamples to argue for a 'this is this and here are some examples . -. .'position that is quite unlike Fortune's own. This is that we cannot trustinglyaccept the facts that result from the 'this is this...' approach to linguisticanalysis. I do not believe we can do this with any approach discovered sofar. Whether the analytical tools we use to arrive at the categories thatgive us our facts are relatively concrete, as Fortune's are, or highlyabstract and/or heavily formalised, as they are in the majority of generativeapproaches, they can easily suggest to us that they yield immutabletruths about language in general, and/or about Shona in particular, ratherthan what they actually yield: facts that result from combining our theorieswith our data in ways that are acceptable within the particular scientificframework that suits us best. Since what we accept as facts shapes whatwe see in our linguistic research, my belief about the nature of factscould help prevent us from leaping too quickly to conclusions about howthings are and what is going on in the real world of human languagestructure.ReferencesBYBEE, J. L. (1985) Morphology: A Study of the Relation Between Meaningand Form (Philadelphia, John Benjamins).CHABATA, E. (1997) 'Applying the Predictability Criterion to ExtendedVerbs: A Study of a Headword and Sense Selection Problem in ShonaLexicography' (MA Dissertation, Harare, University of Zimbabwe,Department of African Languages and Literature).CHIMHUNDU, H. (1983) 'Adoption and Adaptation in Shona' (D. Phil Thesis,Harare, University of Zimbabwe, Department of African Languagesand Literature).Š (1996). Duramazwi ReChiShona (Harare, College Press).CRYSTAL, D. (1991) A Dictionary of Linguistics and Phonetics, 3rd ed.A. A. JEFFER1ES 21(Oxford, Basil Blackwell).DEMBETEMBE, N. C. (1987) A Linguistic Study of the Verb in Korekore(Harare, University of Zimbabwe Publications).FORTUNE, G. (1957) An Analytical Grammar of Shona (Cape Town,Longmans, Green and Co.)-Š (1962) Ideophones in Shona (London, Oxford University Press).Š (1971) 'Some notes on ideophonic constructions in Shona', AfricanStudies, XXX, 237-57.Š (1984) Shona Grammatical Constructions, Vol. 2,3rd ed. (Harare, MercuryPress).Š (1985) Shona Grammatical Constructions, Vol. 1, 3rd ed. (Harare, MercuryPress).GIVON, T. (1971) 'On the verbal origin of the Bantu verb suffixes', Studiesin African Linguistics, 2 (ii), 145-63.GOWLETT, D. F. (1967) 'The Morphology of the Verb in Lozi' (UnpublishedMA thesis, University of the Witwatersrand, Johannesburg).GUTHRIE, M. (1962) 'The status of radical extensions in Bantu languages',Journal of African Linguistics, 202-220.HANNAN, M. (1981) Standard Shona Dictionary, 2nd ed. (Harare, TheLiterature Bureau).HOMBERT, J. M. (1986) 'Word games: Some implications for the analysisof tone and other phonological constructs', in J. J. Ohala and J. J.Jaeger (eds.) Experimental Phonology (Orlando, Academic Press),175-86.JEFFERIES, A. A. (1990) 'Beyond Tone: Functions of Pitch in Shona' (PhDDissertation, Gainesville, University of Florida).KANGIRA, J. (1997) 'The Application of the Unaccusative Hypothesis toShona Intransitive Verbs' (BA Honours Dissertation, Harare, Universityof Zimbabwe, Department of Linguistics).KATAMBA, F. (1993) Morphology (London, Macmillan).KENSTOWICZ, J. (1994) Phonology in Generative Grammar (Cambridge,Blackwell, MA).KHAMISI, A. M. (1985) 'Swahili Verb Derivation' (PhD Dissertation,University of Hawaii).LANGACKER, R. W. (1975) 'Modern Syntactic Theory: Overview andPreview', Paper presented at the First International Conference onLinguistics in Central and Southern Africa (Salisbury, University ofRhodesia).LEBEN, W. R. (1978) 'The representation of tone', in V. Fromkin (ed.).Tone: A Linguistic Survey (New York, Academic Press), 177-219.MKANGANWI, K. G. (1995) 'Shona: A Grammatical Sketch (Department ofLinguistics, Unpub. Manuscript, University of Zimbabwe).MYERS, S. (1987) 'Tone and the Structure of Words in Shona' (PhD22 VERBALISERS AND EXTENSIONS IN SHONAŠ (1994) 'Epenthesis, mutation, and structure preservation in the Shonacausative', Studies in African Linguistics, 23 (ii), 185-216.Odden, D. A. (1981) 'Problems in tone Assignment in Shona' (PhDDissertation, Urbana, University of Illinois).Š (1984) 'Stem tone assignment in Shona', in G. N. Clements and J.Goldsmith (eds.) Autosegmental Studies in Bantu Tone (Dordrecht,Foris), 255-80.Š (1986) 'On the role of the obligatory contour principle in phonologicaltheory', Language, LI1, 353-83.POPPER, K. R. (1965) Conjectures and Refutations: The Growth of ScientificKnowledge (New York, Harper Torchbooks).Š (1979) Objective Knowledge: An Evolutionary Approach (Oxford,Clarendon Press).ROBINS, R. H. (1989) General Linguistics, 4th ed. (London, Longmans).TAFANGOMBE, C. (1997) 'The Functions of the Ideophone in Shona' (BAHonours Dissertation, Harare, University of Zimbabwe, Departmentof Linguistics).FORTUNE'S (1984,Verb-bhidhirika-bhururuka-bwodoka-cheneruka-chenuka-dimika-doka-dzedzereka-foshoka-kocheka-kotoka-kwachanuka-ndoka-parapatika-paruka-payika-pepereka-petenuka-pfuduguka-pfumburuka-piripitika-ribiduka-rotomoka-rukutika-saruka-simuka-sungunuka-svetuka-tasanuka-tendeuka-tepfenuka-tepuka-tereka-togoka-tsakatika-tsvukurukaA. A. JEFFERIESAPPENDIX2313-21) -k- AND -ik-/-ek- VERBALISER EXAMPLESGloss'roll over''fly''get dislocated''be whitish''be covered with dust''hint, allude, be metaphorical''set, go down''totter''pour out''hang up''disperse''fall from wall''go out, as fire''wake up suddenly''get torn''hang up, suspend s/th''wave in wind''get unfolded''get uncovered''be greyish''stampede, run together''gallop''talk loudly in sleep, orthoughtlessly''be weak''choose''rise''get untied''jump''be straight''turn around''be slack''sway''place (pot) on fire''emit smoke1'get lost, vanish''be reddish'Ideophonebhidhiri bhidhiribhururubwodocheneruchenudimidodzedzerefoshokochekotokwachanundoparapataparupayipeperepetenupfudugupfumburupiripitiribidu ribidurotomorukutusarusimusungunusvetutasanutendeutepfenutepu teputeretogotsakatatsvukuru24VERBALISERS AND EXTENSIONS IN SHONA-tuturuka-unduka-vhozhoka-vhukutika'swell''lose feathers''gush out''dash to pieces'tuturuunduvhozhovhukutuVerbGlossADJ-kobvuka-pfupika-shomeka-tsveneka'be thick''be short''be insufficient''be clean'-kobvu-pfupi-shoma-tsveneVerbGlossNoun-pfimbika'ripen fruit'pfimbiFORTUNE'S (1984, 13-21) -dz- VERBALISER EXAMPLESVerbGlossIdeophone-bhabhadza-bhararadza-chechedza-checheredza-chochodza-dadadza-dededza-dhiriridza-dododza-gogodza-kekedza-kukuridza-kerekedza-ngururudza-ngwerengwedza-njereredza-pfipfidza-pururudza-sesedza'flap wings''shout, bleat''trim''run with a hoop, like abicycle rim''run, as a cock''run, as a dassie''toddle''thunder, roar''hammer''knock''cluck''crow''cluck''growl, as a leopard''tinkle''sing, as a cicada''squeak''ululate''trot, as a dog'bha bha bhabhararachecheterechechererecho cho choda da dade de dedhirirrrdo do dogo go goke-ke-ke-ekukurigokere kerengurrrngwere ngwerenjererepfi pfi pfipur-r-rse se seA. A. JEFFERIES25VerbGlossNoun-shamwaridzana 'be friends with one another'shamwariFORTUNE'S (1984, 13-21) -r- AND -ir-/-er- VERBALISER EXAMPLESVerb-chachura-chakura-davira-dzamura-dzedzemura-fefetera-gachira-gakaira-gomera-hukura-jajura-kakaira-kambaira-kambura-kokorora-kwegura-marangura-mhara-momotera-nanaira-ndonyera-ndurira-n'aira-n'un'unura-nyura-pazaura-ritaira-saira-sesemura-tandavara-tandira-tanhauraGloss'take from pot''munch, like a pig''answer''take a big handful''cut out a large section''blow, as wind''receive''gulp down noisily''groan''bark, as a dog''take from pot greedily''drink noisily''crawl''scoop food" with fingers''croak, as a frog''grow old''scratch surface''perch''swarm around''move painfully, slowly''plant a shoot''sting''glitter''gnaw''sink''pull down, here and there(as a building)''walk aimlessly''push''cut out a section''stretch out''stretch''remove nuts from a plant,legs/wings from a locust'Ideophonechachuchaku chakudavidzamudzedzemufe fe fegachigakaigomehukujajukakaikambaikambukokoro/kokororokwegumarangumhamomotenanaindonyenduren'ain'un'unu n'un'ununyupazaurita ritasaisesemutandavatandetanhau26VERBALISERS AND EXTENSIONS IN SHONA-tekaira-tikitira-tonhofa-tonongora-tsamura-tsvairatsvutsvudzira-vhaira-yangarara-zvambarara'sway from side to side''perspire''be cold''shell nuts''take a pinch''sweep''urge on a dog''show off'float''lie prone'teka tekatiki tikitonhotonongotsamutsvaitsvu tsvu tsvuvhaiyangarazvambaraNotes: Fortune (1984, 24) observes that the -ar-/-arar- stative extension' ... is akin to the verbaliser -r- which combines with ideophones toindicate posture'. Adding -r- results in transitivising some ideophones inhis examples. Movement to/from and position seem to be the mostcommon features of the other derived forms he presents.VerbGlossNoun-kokora'scrape pot'makoko