[This document can be acquired from a sub-directory coombspapers via anonymous FTP and/or COOMBSQUEST gopher on the node COOMBS.ANU.EDU.AU] The document's ftp filename and the full directory path are given in the coombspapers top level INDEX file] [This version: 4 August 1993] -------------------------------------------------------------------------- File 4. There are 9 parts of this thesis: from 0 to 8 -------------------------------------------------------------------------- A Grammar of Garadjari, Western Australia by Anna Kristina Sands 1989 Bachelor of Arts Thesis Department of Linguistics, Faculty of Arts Australian National University, Canberra ACT 0200 4. VERBAL MORPHOLOGY A verb in Garadjari is made up of a verb root followed by a tense-aspect suffix and subject-object cross-reference suffixes. Optional suffixes may follow these. 4.1 Root Structure Verb roots in Garadjari are either simple or compound. Simple roots may be either monosyllabic or multisyllabic with the most common verbs generally being monosyllabic. At least two monosyllabic roots have a longer, disyllabic, root that is used with certain tense endings. ie. nja- 'to see' has a longer root njangu- ( eg. njanga-njari 'see-CONT' ) and yu- 'to give' has a root yungu- ( eg. yungu-njari ). Note that both of these roots belong to the same conjugation, the -ng-conjugation. Compound roots are formed with auxiliaries which are generally monosyllabic verb roots. These follow a pre-auxiliary element which may be nominal, or have originated from a nominal, though I have limited proof that this is actually so. There is for example a noun djari 'foot' which may be combined with the monosyllabic root ma-, ( ie. djari ma-rnu ) to mean 'to track'. Other examples are limited though there are many examples which may easily be imagined to have nominal pre-auxiliaries. If we look at an example of a compound verb with the assumption that the pre-auxiliary element is actually a nominal we see that the compound verb actually resembles a transitive verb with the object appearing before the verb. We can compare a compound verb with a clear example of a verb preceded by an object. eg. (1) mabu dja-rnu good put(AUX)-past he made it good (2) djuru nja-ngu watersnake see-past he saw a watersnake As the cross-reference marker for the third person singular object is zero ( see section 4.6 ) this does not come into it and we cannot tell if there is actually a hidden cross-reference marker in the compound verb or not. Thus, the only surface difference between the two sentences is in the semantic cohesion of the elements. In sentence (1), with a compound verb, there is a strong link between the pre-auxiliary element 'good' and the verbal auxiliary. These two combine to form a complete idea and the two elements may not very easily be replaced. In sentence (2), however, the elements are obviously only loosely linked and each may easily be replaced by any other element from the same word class. The real grammatical difference between these two sentences may be found when one looks at possible word order. Now in sentence (2) it would be quite possible, and in fact more usual, to reverse the word order and to say nja-ngu djuru . This would represent the unmarked word order in Garadjari, while in (2), as it stands, emphasis is being placed on the object. Sentence (1), however, seems to represent the only possible combination of these elements, at least in the data. We therefore have a real difference between the two sentences and we can say that sentence (1) represents a compound verb due to its semantic cohesion, limited ability to swap with other elements from the same class and its rigid word order. Now mabu dja- is a compound verb where we can see the nominal origin of the pre-auxiliary element and the semantics make sense, ie. 'good' + 'put / make' = 'to make something good'. This, therefore, seems to be a marginal compound verb. However, we have other compound verbs where there is only limited independent existence of the pre-auxiliary elements ( in the data at least ) and the semantics do not necessarily fit the individual components. For example, mira ga- 'to look', is a compound verb made up of ga- an auxiliary meaning 'to carry' and mira . We do not get mira occuring as an independent word. It is thus difficult to see how the individual components could add up to make a verb meaning 'to look'. These verbs can truly be counted as compound verbs as they show : - rigid order of the components - lack of independent nominal existence of the pre-auxiliary element - semantic unity of the elements - grammatical unity of the elements, ie. no element may intervene between the pre-auxiliary and the auxiliary 1 As has been mentioned, the pre-auxiliary element of most compound verbs has limited independent existence. Unlike mabu dja- , from which mabu may possibly be extracted as an independent noun, in most compound verbs the pre-auxiliary element never occurs without its auxiliary. In other compound verbs the pre-auxiliary may appear alone but it always has verbal semantics, ie. we get what looks like a noun, ( it does not take tense inflection ), acting as a verb semantically. This situation is very common for some compound verbs which very rarely appear with the actual auxiliary. For example, gardaga is an element meaning 'sleep'. It may be modified by an adjectival nominal, eg. (3) wubardu gardaga little sleep they had a little sleep but it appears semantically as a verb, ie. it appears in verbless sentences with a subject. eg. (4) gala gardaga ngali then sleep us-two so, let us sleep Now, gardaga forms two compound verbs. ie. (5) gardaga djari-nju sleep become(AUX)-past he went to sleep (6) gardaga dja-rnu sleep put(AUX)-past he put it down To determine whether the independent usage of gardaga is nominal or verbal is very difficult. As seen, gardaga may be modified by an adjective, which might suggest that it is a nominal. But gardaga , and in fact most other pre-auxiliary elements may not appear as subject or object to a verb. When they appear they always appear in verbless sentences and behave semantically as verbs. However, this may simply be an accident of the data or a function of the semantics of the element, ie. the most common pre-auxiliary elements that may stand alone are; mabu = good gardaga = sleep(ing) garibila = stand(ing) widi = play(ing) None of these are nominals that one would expect to stand as the subject or object of a verb under normal circumstances. I have found, however, one example of one of these pre-auxiliary elements taking a nominal inflection. ie. (7) ginjangkudjara widi-ngka they two play-LOC hese two were playing However, I also have an example of the same element taking a verbal inflection, ie. (8) gala widi-njari then play-CONT then there was playing around as well as forming the preverbval element of a compound verb, ie. (9) widi ma-rnu-bula play do-past-3 Du S they 2 played and having independent existence as a semantic verb, ie. (10) gala widi widi then play play then they played Thus, widi appears not only as the pre-auxiliary element of a compound verb, with or without the auxiliary, but appears with nominal inflections and verbal inflections. Is widi , thus, a nominal or a verbal element, or are there two homologues ? I have one more example of a pre-auxiliary element taking a verbal suffix. Compare : (11) djurngu dja-rnu-djanagu show put(AUX)-past-3 Pl D he showed it to them (12) djurngu-djanagu show-3 Pl D he showed it to them The pre-auxiliary element djurngu may appear with the auxiliary and tense or without the auxiliary in which case it still takes the dative cross-reference marker but takes no tense although tense is usually compulsory for all verbs. To complete the picture there are many semi-verbal ( ie. semantically verbal, as are the pre-auxiliary elements ) elements that never appear with an auxiliary in the texts. These appear most often reduplicated. eg. (13) djawid djawid sneak sneak they sneaked (14) gawid gawid silence silence they were silent 4.2 Conjugations Garadjari has four verb conjugations though it appears that a change is in process whereby two of these might be falling together. 2 The two conjugations in quesiton are the -l- and -n- conjugations and these share the past suffix which is the most common of the suffixes to occur in the texts. Thus, falling together could be occuring to level out a system that is already partially similar. The other two conjugations are the -ng- and -0- conjugations. The names for the conjugations are taken from the linker that is used between the verb root and the future suffix -gu . Monosyllabic verb roots that act both as simple verbs and as auxiliaries in compound verbs may belong to a different conjugation in their different usages. eg ga-ngu carry-past = -ng- conjugation yira ga-rnu investigate(AUX)-past = -l- conjugation ma-n-gu-bula take-FUT-3 Du S = -n- conjugation gurda ma-l-gu-bula fill(AUX)-FUT-3 Du S = -l- conjugation We can do a comparison of the different conjugations with regard to transitivity and the number of members. The following results are obtained : Chart 4.1 Conjugation Breakdown -ng- -0- -n- -l- members 7 11 4 10+ transitive 100% 40% 75% 80+% monosyllabic nja- djai- ma- nga- roots (see) (come) (get) (eat) yu- ya- dja- (give) (go) (put) ga- (carry) bu- (hit) From these results I would suggest that the -ng- and -n- conjugations represent closed classes of mainly monosyllabic roots while the -l- conjugation is an open class of mainly tranisitive disyllabic roots while the -0- conjugation is the open class to which most intransitive roots belong. This can be compared to other languages from the Western Desert area with very similar results [ see Dixon, 1980, section 12.2 ] . 4.3 Transitivity Garadjari verbs are either transitive or intransitive. However, as subject and object are generally optional in a sentence and are only included when they represent new information the strictness of the transitivity is very difficult to gauge as verbs very often appear with no overt subject nor object. I do have evidence, however, that transitive verbs may appear without an object in a construction that is different from a missing understood object ( see section 5.2 ). Compare : (15) nja-ngu duru gala bina gari-nju see-PAST bird then hear(AUX-CONT)-PAST he saw a duru bird then he heard it (16) bina gari-nju hear(AUX-CONT)-PAST he heard something In sentence (15) the objectless verb follows a previous clause ( whether this be a part of the same sentence or of a previous sentence ) from which the missing object is understood to be taken from. The previous clause supplies the missing object. However, in sentence (16) we have no previous clauses from which to take the object. Bina gari-nju can stand as a contextless clause meaning simply that 'he heard'. What he heard is of no importance - he heard a sound. 3 Thus, though both missing objects are a discourse device, in sentence (15) it is a device of convenience and reduction of redundancy, in (16) it is a device that has semantic import. Transitive verbs may take up to two core noun phrases, the agent in the ergative case and the object in the absolutive ( with zero marking ). eg. (17) mirin-dju gadi bu-ngu garliguru Mirin-ERG cut(AUX)-PAST bullroarer(ABS) Mirin cut the bullroarer. (18) wiru waran-dju djinda-rnu leaf(ABS) one-ERG break-PAST one of them broke off a leaf Intransitive verbs take a subject in the absolutive. eg. (19) barda gurga dja-rnu-ra sun(ABS) arise put(AUX)-PAST-COMP the sun rose (20) guyi bara ma-rnu animal(ABS) run(AUX)-PAST he animal ran away In addition there are verbs that take an indirect object in the dative. eg. yinda- 'to tell'. However, we seem to get variation as to whether we have the dative or absolutive for the recipient and the ergative or absolutive for the teller, ie. we have the following : Teller = absolutive, Recipient = dative (21 barnangu yinda-rnu-bulagu Barnangu (ABS) tell-PAST-3 Du D (DAT) Barnangu told them 2 Teller = absolutive, Recipient = absolutive (22) yinda-la-la Barnangu Gargidja tell-IRR-3 Sg D (DAT) Barnangu (ABS) Gargidja (ABS) Gargidja told Barnangu Teller = ergative, Recipient = dative (23) gargidja-lu yinda-rnu-la Gargidja-ERG tell-PAST-to him (DAT) Gargidja told him Teller = ergative or absolutive, Recipient = dative (24) mulal yinda-la-nja-la-gu not tell-IRR-1 Sg S-3 Sg D (DAT)-PURP I will not tell him In the majority of cases we have a dative recipient and we cannot tell whether the teller is nominative or ergative because there is no overt noun phrase ( ie. as in sentence (24)). 4.4 Verb Derivation Transitive verbs may be derived from nominals with the addition of the auxiliary ma- 'to take / get' . It generally means 'to make X'. eg. (25) maru ma-rnu song make-PAST he sang a song (26) lalga ma-rnu dry make-PAST he dried it (27) mabu ma-rnu good make-PAST he made it good Intransitive verbs may be derived using djari- 'to become'. eg. (28) marda djari-nju tired become-PAST he got tired (29) lalga djari-nju dry become-PAST it became dry (30) maba djari-nju lost become-PAST it got lost Note that these form compound verbs as have already been discussed in 4.1. I have found only one example of a possible derivational suffix in the data and this forms a transitive verb. ie. (31) gangara-mbi-n high-become-PRES going up high Here the -mbi seems to be a suffix with a meaning very similar to the auxiliary djari-. 4.5 Verbal inflections Garadjari has three tenses; past, present and future, a continuative aspect which may be used instead of a tense and seems to have no inherent tense, and an irrealis inflection. These suffixes are all mutually exclusive and only one may occur on any given verb, though every verb must have an inflectional suffix or a suffix that may replace the inflection. In addition there are several other suffixes that may replace the inflectional ending but which do not differ for conjugation and so will be discussed in section 4.7. Following is a chart of the inflectional endings for each conjugation. Chart 4.2 Conjugational Inflections -l- -n- -ng- -O- past -rnu -rnu -ngu -nju future -l-gu -n-gu -ng-gu -gu present -n -nja -nja 0 irrealis -la -ra -wa,-ra -ya continuative -ri -njari -njari -njari 4.5.1 Past The past inflection is the most common in the texts and indicates an action that was carried out before the time of the speech action. When several verbs in a row use this inflection the implication is that these events are sequential. eg. (32) mirin-dju gadi bu-ngu garliguru , dulga-ngu , mabu dja-rnu , Mirin-ERG cut(AUX)-PAST bullroarer, split-PAST , good(AUX)-PAST bilgabardi-ngu , gala yaga-rnu crack-PAST, then leave-PAST Mirin cut a bullroarer, he split it off and made it good. It cracked so he left it. 4.5.2 Future The use of the future inflection is not restricted to strictly future time but is very often used in the texts where one would expect the past tense. It seems to be used as a discourse device ( perhaps for assigning prominence to a particular segment of text ) because it is very often used in large chunks of text. Where it appears on a single isolated verb in the text it always indicates future time. eg. (33) mira gardi-gu-rna-ra look(AUX-CONT)-FUT-1 Sg S-COMP I will go look The future suffix also has a meaning of a customary or habitual action. It is used at least once to indicate a habit. ie. (34) djarin-dja barda-ngka yilba-gu every-LOC sun-LOC throw-FUT everyday he would throw ( the people ) It is also combined with the suffix -ngala 'might' to indicate a customary or habitual past action, see section 4.7.5. The future is very often combined with the first person plural subject marker and has the meaning of 'let us do X'. eg. (35) ya-ngu-li go-FUT-1 Du S let's go 4.5.3 Present The present tense is used when the time of the speech act and the time of the action described are coincidental. In the texts, therefore, it is used generally only in quoted speech. eg. (36) mulal ngayidju-lu ga-nja birimala not I-ERG carry-PRES 1 Sg S birimala It is not for me to carry the birimala object (37) ya-nja-nja go-PRES-1 Sg S I am going The two verbs nja- 'to see' and yu- 'to give' take an augmented stem before the present ending, ie. njanga- and yungu-. As in English, the present tense is also used for the immediate future. It also appears occasionally in the texts where one would expect the past tense, but unlike the future, it appears on isolated verbs. eg. (38) ngurnu djai-0-bula wulguru-ngka yonder come-PRES-3 Du S shade-LOC yonder, the 2 approached the shade 4.5.4 Irrealis The irrealis suffix is used both in the imperative and the negative constructions though no confusion can occur because a negative verb is always preceded by the negative particle mulal . eg. (39) mulal ya-ra not go-IRR he did not go (40) mulal nja-wa-ya ngaba not see-IRR-3 Pl S water they did not see water The irrealis inflection replaces any other tense inflection and therefore the negative has no tense associated with it. The irrealis suffix when used without the negative particle can be taken to be the imperative construction. eg. (41) wanda-ya stay-IRR stay! I have found only three examples of the imperative in the -ng-conjugation and it appears that there may be some confusion as to the form of the suffix. ie. (42) yu-wa give-IRR nja-wa see-IRR mira ga-ra watch-IRR Here we have -wa in two verbs and -ra in the other. This data is too limited to draw any conclusions from it and the -ra is possibly even a mishearing or false transcription. The irrealis suffix may be combined with the third person non-singular subject markers to form the non-singular imperative. As the third person singular pronoun suffix is zero we might like to say that the imperative is actually formed by the irrealis suffix plus the subject suffix. It is therefore no different to all the other inflectional endings which must also take the subject suffix. eg. (43) ya-ra-0 go-IRR-3 Sg S go! (44) ya-ra-bula go-IRR-3 Du S you 2 go! (45 ya-ra-ya go-IRR-3 Pl S you plural go! The imperative may be freely combined with the object or dative suffixes eg. (46) yaga-la-nja leave-IRR-1 Sg O leave me! (47) dja-ya-dja come-IRR-1 Sg D come to me ! 4.5.5 Negative Imperative The negative imperative combines the irrealis suffix with the purposive suffix -gu . -Gu follows any subject or object suffixes. Note that unlike the imperative that takes a third person subject marker the negative imperative takes the second person marker. This implies that this is actually an example of a negative construction and not an imperative. eg. (48) mulal djinga-la-nja-n-gu not come-IRR-1 Sg O-2 Sg S-PURP do not come after me! (49) mulal nga-la-mbula-djinja-gu not eat-IRR-2 Du S-3 Pl O-PURP don't you 2 eat them 4.5.6 Continuative The form of the continuative is -njari for all conjugations except for the -l- conjugation for which it is -ri , though very often verbs from the -l- conjugation use -ri and verbs from the -n- conjugation use -njari . This is just part of the general crossing over that occurs between these two conjugations. The final -ri of the suffix changes to -ra where it precedes the -ya third person plural cross-reference marker. The monosyllabic verb roots nja- 'to see' and yu- 'to give' take an augmented root before the continuative, ie. njanga- and yungu- respectively. The continuative aspect is used to denote an action whether it be in the past, present, or future, ie. no distinction is made. The continuative aspect does not necesarily denote that an action is ongoing or durative but may be used where an action's consequences are durative, eg. (50) yaga-ri gara leave-CONT west he abandonded him in the west (51) gardaga djari-njari-bula sleep become(AUX)-CONT-3 Du S they went to sleep or where an action is repeated many times in the story, eg. (52) njanga-njari warayinj gardu see(AUG)-CONT another young-man he saw another young man though it may be used for a durative action. eg. (53) warara minma-ri-ra gardu-gu old man wait-CONT-COMP young man-DAT the old man waited for the young man Very often, though, the continuative aspect is used where one would expect the past eg. (54) ngayidju gagara warara yinda-ri I east old-man say-CONT 'I'm going east' the old man said Thus, the continuative is very different to the English equivalent. The continuative suffix may also be used in compound verbs to derive a new verb, ie. we may have two contrasting compound verbs differing only in the presence or absence of the continuative suffix. eg. (55) mira ga-ngu-bula-djinja marungu look(AUX)-PAST-3 Du S-3 Pl O man they 2 looked at the men (56) mira gari-nju-djinja look(AUX-CONT)-PAST-3 PL O he watched them Here we have a contrast between 'to look at' without the continuative and 'to watch' ( a durative action ) with the continuative. 4.6 Cross-Referencing Suffixes Immediatly following the tense inflection you get the cross-referencing suffixes that agreee with the subject, object, or indirect object of the sentence. Though the suffixes are obligatory the noun phrases themselves that they agree with are optional and are generally only included if they represent new information or are added for emphasis. Following is a chart of the cross-referencing suffixes. Chart 4.3 Cross-Referencing Suffixes first second third subject sg -rna,-nja -n 0 du -li -mbula -bula pl -nja -njura -ya object sg -nja 0 du -ngalinja -bilinja pl -nganinja (-njura) -djinja dative sg -dja -ngalgu -la du -ngaligu -bulagu pl -nganjagu -njuragu -djanagu It is useful at this stage to do a comparison of these forms with the free form pronouns, see section 3.3. We see that in the singular forms, ie. the most common forms, the cross- reference markers bear no resemblance to the free form pronouns. We have first person singular ngayidju as a free form with cross-referencing markers -rna , -nja and -dja . In the second person singular we have njundu compared to -n and -ngalgu and in the third person singular ginjangka as compared to zero ( 0 ) and -la . However, if we exclude the third person from the analysis and look at the dual and plural forms we can see a definite pattern. ie. Chart 4.4 Comparison of Pronouns with Cross-Referencing Suffixes free subject object dative 1 du ngali -li -ngalinja -ngaligu 1 pl (ngawa) -nja -nganinja -nganjagu 2 du njumbala -mbula 2 pl njura -njura (-njura) -njuragu It seems if we exclude the first person plural, for which I am not at all sure of the free form, that the subject - object cross-referencing markers are very definitely based on the free form pronouns. In the subject cross-referencing marker we have dropping of the initial syllable of the free form in two out of three forms, and we get -nja added to the free form for the object and -gu for the dative. These results suggest that the cross-referencing system of suffixes has developed from the pronouns. In the third person there is no resemblance between the free form pronouns and cross-reference markers, but we note that third person pronouns are not reconstructed for proto Australian. Therefore, we could say that the cross-referencing system developed at at time when the third person pronouns were not yet developed and they then developed independently. Note that the third person dual subject cross-referencing suffix -bula is a common dual suffix in Australian languages. 4.6.1 Subject The subject cross-referencing suffix always occurs before the object or dative cross-referencing suffix except where these are in the first person singular, in which case the dative or object suffix precedes the subject suffix. eg. (57) yinda-ri-dja-n say-CONT-1 Sg D-1 Sg S you said to me (58) mulal yinda-la-nja-n-gu not say-IRR-1 Sg O-2 Sg S-PURP do not say to me The subject cross-referencing suffix is used for both the transitive and intransitive subject and thus differs from the ergative - absolutive system employed in nominals. Some examples of how it is used are: Transitive subject : (59) ngaba yarnda-rnu-bula water search-PAST-3 Du S they 2 searched for water (60 nja-ngu-ya bidi see-PAST-3 Pl S hole they saw a hole Intransitive subject : (61) ya-rnu-bula go-PAST-3 Du S they 2 went (62) bara ma-rnu-ya run (AUX)-PAST-3 Pl S they ran away The first person singular cross-referencing subject suffix has different forms depending upon which inflectional ending it appears with. The use of the first person singular subject also causes changes in the inflections. I have the following examples of this : In the future : First person singular subject marker is -rna with the future suffix -gu . eg. (63) mira gari-gu-rna-ra look(AUX)-FUT-1 Sg S-COMP I will go look (64) njuma yilba-gu-rna-ngala egg throw-FUT-1 Sg S-might I think I will throw the egg In the present : First person singular marker is -nja and tense marker may be -nja or zero. eg. (65) ya-nja-nja go-PRES-1 Sg S I am going (66) gala ngayidju-lu gulja bidja ma-0-nja-ra then I-ERG south search(AUX)-PRES-1 Sg S-COMP so I am going to search in the south These are the only examples of the present tense that I have found ( though ya-nja-nja occurs three different times ) so I cannot determine what is actually happening. Note that bidja ma- is an -l- conjugation verb and therfore we would not expect it to take a zero present tense. In the past : First person singular and past combine to form -njan. Note that as first person singualar is -nja elsewhere we might like to say that -nja was the first person singular marker here. However, this would require saying that -n was a marker of the past and that the past inflection came after the subject marker. This does not occur elsewhere. Some examples are : (67) njanga-njan djibi see(AUG)-PAST 1 Sg S djibi I saw a djibi bird (68 guyi ma-njan animal get-PAST 1 Sg S I was getting animals In the continuative : I have only one example of this but the suffix is the same as in the future ( ie. -rna ) for the first person singular subject. eg. (69) binaguba yanka-ri-rna djara grub get-CONT-1 Sg S self I am going to get grubs for myself In the negative : The negative and the first person singular subject seem to be combined in -nja . eg. (70) mulal njanga-nja ngayidju-lu not see(AUG)-IRR 1 Sg S I-ERG I cannot see (71) mulal ngayidju-lu ga-nja birimala not I-ERG carry-IRR birimala I do not carry the birimala object 4.6.2 Object The object marker immediately follows the subject marker, if there is one, unless the object is first person singular in which case it precedes the subject marker. The object-marker cross-references the transitive object. eg. (72) bu-ngu-bula-bilinja-la cut-PAST-3 Du S-3 Du O-la they 2 cut the 2 ( bullroarers ) (73) garibila djai-nju -djinja-gudji stand come(AUX)-PAST-3 Pl O-then thereupon he stood them up (74) nja-ngu-n-nja see-PAST-2 Sg S-1 Sg O you saw me The form -nja seems to be closely associated with the object as it appears on all object cross-refernce markers except for the third person singular, which is zero, and possibly not on the second person plural. The form -nja does not appear on free form pronouns in Garadjari as these follow a basically ergative - absolutive case system with no case marking on the object. The subject-object markers, however, with a nominative - accusative system, do show case marking on the object with the form -nja.. This may reflect a time in the history of the language when free form pronouns followed a nominative - accusative case system as we see that the majority of Australian languages not only show a nominative - accusative system on bound pronouns, where they occur, but a nominative-accusative system in free form pronouns with a case marker -Nha ( ie. -nja in single laminal languages ) on the object [ Dixon, 1980 ] . 1 I have only one possible example of the second person plural object and it is marginal whether this has been translated correctly. ie. (75) miljura-ngka dja-rnu-njura hole-LOC put-PAST-2 Pl O (?) I put you into the hole This has been translated as''I put you into the hole', however, there is no free nor bound pronoun to indicate the first person subject. One would expect that the second person plural object would more likely be -njuranja . 4.6.3 Dative The dative marker agrees with an indirect object, a noun phrase in the dative. The dative marker follows the subject, except where the dative is the first person singular, in which case the dative marker precedes the subject marker. I have found no examples of the dative co-occurring with an object so I do not know whether such a situation is possible and what would be the relative orders of the dative and object markers. As -nja appears to be associated with the object so -gu is connected with the dative except in the first and third person singular. This is consistent with the use of -gu on nominals as a marker of the dative. In the texts the dative cross-referencing suffix is used in the following situations : (76) yinda-ri-dja-n tell-CONT-1 Sg D-2 Sg S you tell to me (77) mulal dja-ya-dja-n-gu lani not come-IRR-1 Sg D-1 Sg S-PURP don't come close to me (78) djurngu dja-rnu-la show(AUX)-PAST-3 Sg D he showed it to him (79) djurara njidi-ra-ya-la (?) crouch-CONT-3 Pl S-3 Sg D they crouched over him (80) marungu-gu maru-gu ya-ra-ya-djanagu man-DAT song-DAT go-CONT-3 Pl S-3 Pl D they went in quest of men for their song (81) magur wira ga-ya-ngaligu road show(AUX)-3 Pl S-1 Du D they showed us the way 4.6.4 Benefactive The benefactive is formed by adding -ra to the dative, with the exceptions of the first person singular which has a possible benefactive of -njara and third person singular which inserts the -ra before the dative marker -la , ie. the benefactive is -ra-la . I have only one example of the possible first person singular benefactive, ie. (82) guyi mayi yu-wa-njara animal herbage give-IRR-1 Sg B (?) give to me animals and vegetables and three examples of the third person singular, ie. (83) ga-ngu-ra-la garbina djarndu-gu carry-PAST-3 Sg B shield woman-DAT he carried a shield to his wife (84) djanba-gu ya-rnu-ra-la djanba-DAT go-PAST-3 Sg B he went along tracks of the Danba man (85) ya-rnu-ra-la djina-gu go-PAST-3 Sg B foot-DAT he went along tracks In all of these examples the benefacitve meaning is not so clear ( ie. assuming the traditional benefactive being to do something for the benefit of someone else ), however, McKelson notes that in the Njangu dialect the third person singular benefactive is -ra-la and that the benefactive and dative suffixes are freely interchangeable. Therefore, it is not unreasonable to assume that such a situation might take place also in the Nadja dialect. Some examples of the certain benefactive are : (86) ga-ngku-n-djanagura ngaba binga-nga carry-FUT-3 Pl B water conch-LOC he carried water for them in a conch shell (87) rumbanj bunga-nju-bulagura nest fall-PAST-3 Du B nests fell down for them 4.6.5 Reflexive / Reciprocal Garadjari has two reflexive particles which may follow the subject suffix. It has an object reflexive -warninja and a dative reflexive -warnungku and these replace the object and dative markers respectively in a sentence. eg. (88) bira-rnu-bula-la cover-PAST-3 Du S-3 Sg D they covered it ( in the fire) --> bira-ri-warnungku cover-CONT-REFL D he covered himself (89) yinda-rnu-la tell-PAST-3 Sg D he told him --> yinda-l-gu-nja-warnungku tell-FUT-2 Du S-REFL D let's talk to one another (90) djinka-rnu-djinja strike-PAST-3 Pl O he hit them --> djinka-l-gu-li-warninja strike-FUT-2 Du S-REFL O let's strike one another (91) yilba-rnu-0 warngu throw-PAST-3 Sg O stone he threw a stone --> mulal yilba-ya-warninja-gu not throw-IRR-REFL O-PURP men are not to throw each other 4.7 Other Suffixes 4.7.1 -nba / -nma Capell [ 1977 ] describes a suffix -ba which denotes a recent past. I have found limited examples of a similar suffix -nba in the texts which could denote the recent past though with limted data it is not possible to be sure about the meaning. eg. (92) yaga-nba-li maba djari-nga bidi-nga leave-REC PAST-2 Du S lost become(AUX)-LOC 2 hole-LOC we must have lost it in the hole (93) yilba-nju-nba-ya gan-gardi throw-PAST-REC PAST-3 Pl S up-towards they threw it up high This suffix presents a problem because in sentence (92) it replaces the tense inflection, whereas in sentence (93) it co-occurs with the inflection. I have no real answers as to what is going on. -Nma seems to be a suffix with the same meaning which is used when the subject is first person singular. This replaces the subject marker and tense inflection. eg. (94) yaguru yaga-nma walgulu dog leave-REC PAST 1 Sg S thirsty I left a dog thirsty (95) waradja gurlu dja-nma one miss(AUX)-REC PAST 1 Sg S I missed one However, there are some examples which do not fit with a meaning of the recent past. ie. (96) djai-nba-gu-li come-REC PAST (?) -FUT-2 Du S let's come together (97) ngana guli-gu yinda-nba-n what fight-DAT tell-REC PAST (?)-2 Sg S why do you call out fighting ? (98) djarndu gudjara yaga-nma-bilinja-ngala woman two leave-REC PAST(?) 1 Sg S-3 Du O-might he was going to leave the two women 4.7.2 -ni -Ni denotes that the action takes place 'here' or at the place of the speech action. eg. (99) ya-ra-ni go-IRR-here come here! (100) wanarda-ya-ni djarla stay-3 Pl S-here now they are staying here now 4.7.3 -gudji -Gudji is used to link two verbal actions where the second occurs immediatly after the first. It therefore occurs most often on the second clause of a sentence and is translatable as 'thereupon'. eg. (101) wigaman njirani djinga-rnu-gudji chickenhawk behind pursue-PAST-then the chicken hawk came up behind and thereupon pursued him (102) ginjangka ma-rnu njuma djur djur bardi-nju-gudji he take-PAST egg slide slide descend-PAST-then he picked up the egg and thereupon slid down (103) nguru nguru ma-rnu djungku bara ma-rnu-gudji steal steal take(AUX)-PAST firestick run-PAST-then he stole the firestick and thereupon ran away It may also occur, however, upon the initial verb of a sentence in which case it does not suggest that the action immediately follows the previous action, but can be translated more as 'and then'. eg. (104) bara ma-rnu-bula-gudji djarla wanardi-bula gurwalja run(AUX)-PAST-3 Du S-then now staying-3 Du S sky and then they ran away and are now staying in the sky (105) wubardu-gu-njanu gala lani djari-nju-gudji little-DAT-one then close become-PAST-then and then they came close to the little one 4.7.4 -ra -Ra is the most common of the verbal suffixes and occurs on a large proportion of verbs in the texts. Because of it's common occurence in a great variety of situations and tenses it's meaning cannot be easily pinpointed and this would suggest that it may be aspectual in nature. Capell [1977] suggests that it may be a completive aspect and this is certainly consistent with the information in the Laves texts, though Laves translates verbs with or without -ra exactly the same. The completive aspect does not mean that the action is over and done with but rather seems to mean that the whole of the action is being talked about, whether this action may be finished or not. We get -ra in all tenses. Past : (106) djula bargara ya-rnu-ra blind turkey go-PAST-COMP a blind turkey came up (107) ngaba-gu garli-nju-ra bidi water-DAT dig-PAST-COMP hole he dug a hole for water (108) garndi-nju-bula-ra baga-ngka climb-PAST-3 Du S-COMP tree-LOC they climbed up the tree Present : (109) dambal ma-nja-ra djinka-rnu anthill take-PRES-COMP break-PAST he seized the anthill and broke it up (110) gala ngayidju-lu gulja ya-nja-ra then I-ERG south go-PRES 1 Sg S-COMP so I am going south Future : (111) gala ngayidju-lu mira gardi-gu-rna-ra then I-ERG look(AUX-CONT)-FUT-1 Sg S-COMP so I will go look for it (112) ma-ngu-li-ra-ngala mayi-gardi take-FUT-2 Du S-COMP-might herbage-ALL let's go get some food Continuative : (113) warayinj-gardi djari-njari-ra bari another-towards become-CONT-COMP eyes he turned his eyes away (114) marungu marlu njanga-njari-djinja-ra djina man many see-CONT-3 Pl S-COMP foot he saw the tracks of many men 4.7.5 -ngala -Ngala is a suffix that may be used in many situations. It co-occurs with the future to mean customary or habitual action. eg. (115) wubardu gawuru nja-ngku-n-ngala lara-gadja little twig see-FUT-2 Sg S-habit sideways-like wandu-gu-n-ngala stay-FUT-2 Sg S-habit you always see a little twig ,wave your head sideways and remain there (116) nja-ngku-n-ngala see-FUT-2 Sg S-habit you can always see (us ) It is also used in questions, eg. (117) ngana nja-n-gu-ngala what see-FUT-might what can we see ? (118) ngana nja-ngu-ngala what see-PAST-might what did we see ? However, in it's most common usage, it can be translated as 'might'. It indicates that some doubt exists as to the carrying out of the action. eg. (119) marala djinka-rnu-ya-ra-la-ngala Marala kill-PAST-3 Pl S-COMP-la-might they must have killed Marala (120) njuma yilba-gu-rna-ngala egg throw-FUT-1 Sg S-might I think that I will throw the egg (121) ngayidju yara gurga dja-ngala-gu I dark arise(AUX)-might-PURP I might arise at night 4.7.6 -nnga Like -ngala , -nnga may be used in questions. eg. (122) wandja-gardi ya-nnga where-towards go-FUT Q where might I go ? (123) wai njundu gardaga djari-nnga QUEST you sleep become(AUX)-2 Sg S-FUT Q do you want to sleep ? However, unlike -ngala , which co-occurs with a tense inflection, -nnga replaces the tense. As I only have two examples of each suffix in questions and the -ngala questions are in the past and present tenses and the -nnga questions are in the future, it could be that -nnga actually is a form of the future used in questions. 4.7.7 -gu -Gu is the purposive suffix. It may be added to verbs in any tense. Where it occurs on the second clause of a sentence it means that the first action was carried out in order to carry out the second. eg. (124) marndi gadjari mulal njanga-nja-gu birimala walking far not see-PRES-PURP birimala they went far away so that they would not see the birimala object (125) bidja ma-nja-ra nuwa ma-ngala-gu search(AUX)-1 Sg S-COMP same get-might-PURP I will search to perhaps get the same The purposive may also occur on the only verb of a sentence and here it adds some sense of purpose or deliberation on the part of the subject in carrying out the action. eg. (126) ngayidju yara gurga dja-ngala-gu I dark arise(AUX)-habit-PURP I will always arise at night (127) mirling ganing bara ma-rnu-gu spear down run(AUX)-PAST-PURP it ran away with the spear in it (128) gurga dja-rnu-gu ngura gani-ngu arise(AUX)-PAST-PURP ground under-ABL the earth rose up from underneath (129) yarnda-rnu-gu mirlin-dju drive-PAST-PURP spear-INST he drove him with a spear (130) mulal ma-nja-gu wangai not get-PRES-PURP alive he couldn't catch him alive I have one example of the purposive occuring on the verb after an auxiliary ( see section 5.5.1 ). Here the auxiliary takes on the meaning of 'should'. eg. (131) ngayidju ma-nja-n wanda-rna-gu I AUX-PAST 1 Sg S stay-1 Sg S-PURP I should have stayed The purposive suffix is also used in the negative imperative construction. This construction is basically a negative with the purposive suffix added to the end of the verb. eg. (132) mulal yinda-la-dja-n-gu not say-IRR-1 Sg D-2 Sg S-PURP do not say to me 4.7.8 -la -La is the most problematic of the Garadjari verbal clitics and it occurs in a wide variety of situations. There are several verbs that seem to always take -la . ie. bu-ngu-la = he cut it gama-rnu-la = he tasted it djanima-rnu-la = he covered it bira-rnu-la = he covered it over Some examples of the usage of -la in sentences are : (133) djai-njari-la warlibi-ngka come-CONT-la camp-LOC he came into camp (134) djanima-ri-la djura-ngka cover-CONT-la fire-LOC he covered them in the fire (135 yira ma-rnu-djarnagura-la name(AUX)-PAST-3 Pl B-la he named it for them (136) gulba-ya-warnungku-ra-la return-3 Pl S-REFL D-B-la they went back from each other (137) barda ngalba-nju-ra-la sun set-PAST-COMP-la the sun set (138) marala djinga-rnu-ya-ra-la-ngala Marala kill-PAST-3 Pl S-COMP-la-might they must have killed Marala The problem of the suffix -la and what it may represent will be discussed in chapter 6. 4.8 Order of the Suffixes Though there is no example in the texts of a verb that has all of the possible verbal suffixes, or even most of them, the relative position of most of them to one another can be determined. Firstly, the verb stem may have one of three forms : a) the verb root by itself, eg. yangka- 'to hunt' b) a pre-auxiliary element plus an auxiliary, eg. mira ga- to look' c) a nominal root plus a derivational suffix or auxiliary, eg. gangara-mbi- 'to become high' The verb stem is always and obligatorily immediately followed by the inflectional endings that differ according to conjugation, ie. those endings presented in chart 4.2. This ending may, however, be replaced by one of the suffixes -nba , -nma or -nnga which represent the near past, near past first person singular and question future respectively. eg. (139) yaga-nba-li maba djari-nga bidi-nga leave-REC PAST-2 Du S lost become(AUX)-LOC hole-LOC we must have lost it in the hole Immediately following the inflectional suffix we get the cross-referencing markers ( with the exception of the third person singular dative marker ) with an internal ordering as follows : -dja (1 Sg D) + Subject marker + Object marker + -ra ( Benefactive ) -nja (1 Sg O) Dative marker Reflexive marker For examples of these see section 4.6. The suffix -ra , the completive aspect, follows the inflectional suffixes. eg. (140) bara ma-rnu-bula-ra run(AUX)-PAST-COMP they two ran away The next slot in the verb is taken up by the suffix -la covering both the third person dative and other roles as discussed in chapter 6. eg. (141) gurga dja-lgu-warnungku-ra-la arise(AUX)-FUT-REFL D-COMP-la he got himself up Following this we have the suffixes -ngala 'might' and -gudji 'then'. eg. (142) gaunda-ra-ya-ra-la-gudji call-CONT-3 Pl S-COMP-la-then thereupon they called out (143) marala djinga-rnu-ya-ra-la-ngala Marala kill-PAST-3 Pl S-COMP-la-might they must have killed Marala I have no examples with both of these suffixes together in the data. The final suffix appearing in the data is the purposive suffix -gu . This always appears as the final element on the verb. eg. (144) ngayidju yara gurga dja-ngala-gu I dark arise(AUX)-habit-PURP I will always arise at night There is one final suffix and this is the suffix -ni 'here'. This comes after the cross-reference markers, however, its relative position to the other suffixes that follow is unknown because it dos not co-occur with them. eg. (145) ya-ra-ni go-IRR-here come here! In summary we can diagram the order of the verbal suffixes as follows : VERB ROOT + -bi INFL CROSS-REF. -ra -la -gudji -gu -ba -ngala -nnga -ma -ni 1 Note that this conclusion is reached through the data and this data may represent trends only and not grammatical fact as examples are limited. To discover that word order was not so rigid, that the elements could be seperated or that cross referencing of the preverbal element with the auxiliary occured ( ie. if we had a situation where we had plural preverbal and we got the plural object marker turning up on the auxiliary ) the analysis could be changed. 2 The alternative to the possibility that the two conjugations are falling together is that the speakers sometimes get confused between the two conjugations and do not know to which conjugation a verb belongs. As the mix-up is always between two specific conjugations, the -l- and -n- conjugations, it seems likely that if confusion occured commonly it might be beetween all four conjugations and not simply the two of them. 3 This is parallel to the situation in English where a missing object of a transitive verb is taken to be the most socialy normal thing to do eg. he ate = he ate a meal he drinks = he drinks alcohol Unlike English, however, where a very specific object is understood, I think that in Garadjari the understood object is very much more general and has no limitations to it. 1I would therefore propose that at the time that Garadjari developed a system of obund pronominal forms ie. the cross-referencing system, the pronouns inflected on an nominative-accusative systemwith a -nja object case. Upon the establishment of the subject-object cross referencing system, free pronouns become less frequent they began to be remodelled on the nominal system and now follow the ergative-absolutive case system. 2 The reason for the locative case suffix here on the verb is unknown and is discussed further in section 7.1.4. -------------------------------------------------------------------------- end of file