UBC Graduate Research

Verbal agreement of datives in Nyangumarta Weber, Natalie 2012-12

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


42591-Weber_Natalie_Verbal_agreement_Nyangumarta_2012.pdf [ 148.38kB ]
JSON: 42591-1.0075739.json
JSON-LD: 42591-1.0075739-ld.json
RDF/XML (Pretty): 42591-1.0075739-rdf.xml
RDF/JSON: 42591-1.0075739-rdf.json
Turtle: 42591-1.0075739-turtle.txt
N-Triples: 42591-1.0075739-rdf-ntriples.txt
Original Record: 42591-1.0075739-source.json
Full Text

Full Text

Verbal agreement of datives in Nyangumarta Natalie Weber 24 December 2012 1 Language introduction Nyangumarta is a Pama-Nyungan language (Marrngu subgroup) spoken in Western Australia. It has a rich system of nominal case morphology and verbal agreement. There are three series of verbal pronominal suffixes (Sharp 2004:244) which agree with subjects, objects, and indirect objects, respectively. They inflect for person, number, and an inclusive/exclusive distinction for first person dual and plural. A chart is given in the appendix. There is additionally a verb-final suffix -a, which Sharp analyzes as a purposive marker: ‘The general function of this morpheme is to indicate that an action was performed for a particular purpose although when it attaches to the indirect object verbal pronouns it indicates benefactive or possessive’ (Sharp 2004:247). However, there are many counter-examples where purposive goals (1b), benefactives (2b), and possessed objects (3b) do not necessarily trigger the -a suffix on the verb. (1) a. ‘Come’ Palama that milpa–nyi come–nfut karlaya. . . emu ‘The emu was coming back. . . ’ (Sharp 2004:175, 5.26) b. ‘Come’ + Purposive goal (no -a) Jinta other kuyi–rrangu meat–pl milpa–nyi–yi come–nfut–3plS [parrja–na–ku look–nmlz–dat mirti–ja–ku. run–abl–dat ] ‘The other animals came to watch the race.’ (Sharp 2004:134, 4.65) (2) a. ‘Hit’ Wirla–rna. hit–nfut ‘S/he hit it.’ (Sharp 2004:65, 2.35d) 1 b. ‘Hit’ + Benefactive (no -a) Wirla–rna–ngu. hit–nfut–2sgD ‘S/he hit it for you.’ (Sharp 2004:68, 2.41) (3) a. Possession (with -a) Pungka–nya–ngu–a. fall–nfut–2sgD–phave ‘Yours fell.’ (Sharp 2004:248, 7.15d) b. Possession (no -a) Ka–nganya–rni take–prs–1sgS janinyi 3plO nyungu–rrangu this–pl ngaju–mili–rrangu 1sg–gen–pl partany–karrangu child–pl kuyi–karti. meat–all ‘I’m taking my children for meat.’ (Sharp 2004:137, 4.81) Clearly the semantic meaning of -a is more complicated than previously charac- terized. I will show that there are two syntactically distinct homophonous -a suffixes in Nyangumarta, and this homophony has until now obscured their function. The first -a is a low applicative (Pylkkänen 2002) which denotes a transfer of possession relation between two DPs and which I equate with phave (Harley 2002). The second -a is an indirect object agreement marker which agrees specifically with inanimate or non-specific animate arguments, and which I illustrate specifically with benefactives. In section (2) I summarize the recent literature on applicatives and lay out di- agnostic tests to distinguish high and low applicatives. In section (3) I analyze Nyangumarta benefactives as high applicatives. In section (4) I turn to a double object construction in Nyangumarta and show that -a in this construction is a low applicative and can be equated with Harley’s (2002) phave. Section (5) discusses instances of verb-final -a which are not low applicatives, and section (6) concludes. 2 Applicatives Pylkkänen (2002) observes that applicative constructions have very different syn- tactic properties cross-linguistically, even when their surface appearance is similar. For example, both English and Chaga have double object constructions involving an applied benefactive argument, but only in Chaga can such a benefactive be added to an unergative verb. 2 (4) English a. I baked a cake. b. I baked him a cake. c. I ran. d. *I ran him. (5) Chaga a. N–a̋–̋ı–lyì–í–à FOC–1s–PRS–eat–APPL–FV m–kà 1–wife k–élyá 7–food ‘He is eating food for his wife.’ b. N–a̋–i–zrìc–í–à FOC–1s–PRS–run–APPL–FV m–bùyà 9–friend ‘He is running for a friend.’ (Bresnan and Moshi 1993:49–50) The syntactic differences reflect an underlying semantic distinction. The English double object construction contains a low applicative which introduces an indirect object low in the syntactic structure. It specifically denotes a ‘transfer of possession’ relation between the direct and applied objects; the reason it cannot occur with unergatives is because it necessarily involves a direct object. The Chaga benefactive contains a high applicative, which relates the benefactive argument to the event. It is thus similar to vP (Chomsky 1995; Kratzer 1996), which relates an external argument to an event. The difference between the two is schematized below. (6) High appl (Chaga) vP He v wife APPLben eat food (7) Low appl (English) vP I v bake him APPL cake These two types of structures make different syntactic predictions about the available lexical verbs, argument structure, passivization, and secondary predicates. These can be used as diagnostics to determine whether a given applicative is high or low. 3 2.1 Verb semantics Since low applicatives create a thematic relation between two DPs, they cannot occur with verbs like ‘eat’ which involve destruction of the theme, nor static verbs like ‘hold’ which do not allow a transfer of possession. Predictably, double objects in English are ungrammatical with these types of verbs. (8) English (low applicative) a. *He ate the wife the food. b. *John held Mary the bag. Because the applied object in high applicatives stands in a relation to the entire event itself, the semantics of the verb should not make a difference. This is why the benefactives in Chaga (5) and Luganda, shown in (9) are allowed for a events involving destruction of the theme or no transfer of possession. (9) Luganda a. Mukasa Mukasa ya–som–e–dde PAST–read–APPL–PAST Katonga Katonga ekitabo. book ‘Mukasa read Katonga a book.’ (Pylkkänen 2002:25) b. Katonga Katonga ya–kwaant–i–dde PAST–hold–APPL–PAST Mukasa Mukasa ensawo. bag ‘Katonga held the pot [sic] for Katonga.’ (Pylkkänen 2002:25) 2.2 Transitivity diagnostic Because low applicative heads denote a relation between the direct and indirect object, they should be unable to appear in structures that lack direct objects. The prediction is that low applicatives should not be able to combine with unergatives while high applicatives can, and indeed, the double object construction in English cannot be used with unergatives, while benefactives in Chaga (5) and Luganda (11) can. (10) English *I ran him. (11) Luganda Mukasa Mukasa ya–tumbu–le–dde PAST–walk–APPL–PAST Katonga. Katonga ‘Mukasa walked for Katonga.’ (Pylkkänen 2002:25) 4 In addition, low applicatives cannot be used with unergative verbs with implicit object, such as ‘bake (a cake)’, even when the verbal semantics of transfer of posses- sion are otherwise correct. Albanian benefactives are an example of a high applicative and show that they can occur with unergative verbs which have an implicit object. (12) English *I baked him. (Intended: I baked his cake / I baked him a cake) (13) Armenian Drita D.nom i cl pjek bakes Agamit A.dat (rrepat). turnips.acc ‘Drita bakes (turnips) for Agim.’ (McGinnis 2001:4) 2.3 Passivization Another asymmetry concerning applicatives is that low applicatives only allow pas- sivization of the indirect object, as shown by the contrasts in (14) for English and (15) for the Kinyarwanda locative applicative. (14) English a. Bill was baked a cake by John. [ApplPlow, IO passivized] b. *A cake was baked Bill by John. [ApplPlow, *DO passivized] (15) Kinyarwanda (Low locative applicative) a. Ishuûri school ry–oohere–j–w–é–ho sc–send–asp–pass–asp–loc t igitabo book n’úúmwáalímu. by.teacher ‘The school was sent the book by the teacher.’ b. *Igitabo book cy–oohere–j–w–é–ho sc–send–asp–pass–asp–loc ishuûri school t n’úúmwáalímu. by.teacher ‘The book was sent to school by the teacher.’ (McGinnis 2001:3) High applicatives, such as the benefactive applicative in Kinyarwanda, allow pas- sivization of both the theme and the indirect object. (16) Kinyarwanda (High benefactive applicative) a. Umukoôbwak girl a–ra–andik–ir–w–a SP–Pres–write–Appl–Pass–Asp tk íbárúwa letter n’ûmuhuûngu. by.boy ‘The girl is having the letter written for her by the boy.’ 5 b. Ibárúwak letter i–ra–andik–ir–w–a SP–Pres–write–Appl–Pass–Asp umukoôbwa girl tk n’ûmuhuûngu. by.boy ‘The letter is written for the girl by the boy.’ (McGinnis 2001:2) There are no known passive constructions in Nyangumarta, so this criteria cannot be used as a diagnostic. I do not discuss it in the following sections, but it would be interesting to explore further. 2.4 Secondary predication Depictive secondary predicates can be predicated of subjects and objects, but not when the external argument is explicit (as in passives), nor of the indirect object in double object constructions (see Williams 1980 and Baker 1997). (17) a. John ate the meat raw. b. John wrote this letter drunk. c. *This letter was written drunk. d. *I talked to Sue drunk. e. *John told Mary the news drunk. Pylkkänen (2002) rejects the small clause analysis of secondary predicates with control in Williams (1980), because under this analysis indirect objects should be possible controllers, even though they cannot have a secondary predicate (Koizuma 1994). Instead, she adopts a complex predicate analysis, where the depictive phase combines with the verb directly (Cormack and Smith, 1999; Yatsushiro, 1999; Geuder 2000). The semantics are such that transitive verbs and vP can combine by Predicate Modification with secondary predicates, but low applicatives cannot. High applied arguments should then always be available for depictive secondary predication. Nyangumarta allows discontinuous modifiers, but it is unclear if they have the same restrictions as secondary predicates. Modifiers usually have the same case mor- phology as the head that they modify, and there are examples of ergative, absolutive, and dative secondary predications. I don’t have enough data to see whether or not these can predicate objects with implied external arguments or the indirect object in the double object construction discussed in section (4) below. 3 Nyangumarta benefactives Benefactives in Nyangumarta are marked with dative case morphology and trigger indirect object agreement on the verb. 6 (18) a. Mirtawa–lu woman–erg kampa–rna cook–nfut mayi. food ‘The woman cooked the food.’ (Sharp 2004:360) b. Partany–ku child–dat kampa–rna–lu cook–nfut–3sgD mayi vegetable.food mirtawa–lu. woman–erg ‘The woman cooked the food for the child.’ (Sharp 2004:360) The benefactive construction in Nyangumarta patterns as a high applicative. Benefactives can be used with verbs which imply no transfer of possession, such as kalku-L ‘have, keep, take care of’1. Other good verbs to test this with would be warli -L ‘hold’ and verbs which destroy the object such as nga-NY ‘eat’. (19) Kalku–lu keep–imp janaku! 3plD ‘Take care of it for them!’ (Sharp 2004:256, 7.40) (20) Wirla–rna–ngu. hit–nfut–2sgD ‘S/he hit it for you.’ (Sharp 2004:68, 2.41) The benefactive can occur with unergative verbs, such as yarnta-L ‘write (lit. spear)’ in (21) and kurnta-L ‘sing’ in (22). Since low applicatives cannot occur with unergatives, this is evidence that the benefactive is a high applicative. (21) Mili–mili–ngi red–paper–loc yarnta–rna spear–nfut janaku 3plD muwarr word paliny–ju. 3sg–erg ‘He wrote it in a book for them.’ (Geytenbeek and Geytenbeek 1991) (22) Julurr corroboree kurnta–la–ji–npi–li sing–imp–1sgD–2sgS–ant ‘You might sing a corroboree song for me.’ (Geytenbeek and Geytenbeek 1991) 1Verbs are cited followed by a hyphen and the verb class they belong to. 7 4 Nyangumarta double object constructions Nyangumarta also exhibits a double object construction. It differs from the benefac- tive construction in that the indirect object DP is unmarked (instead of dative) and triggers accusative verbal agreement (instead of indirect object). This construction occurs with the verbs mira-L ‘relieve’, jami -L ‘withhold’, and yu-/yi -NG ‘give’. The verb mira-L ‘relieve of’ has the schema ERG relieves ABS of DAT when the absolutive argument is not third person singular2. (23) a. Wayilpila–lu white.fellow–erg mira–rni–yi relieve–nfut–3plS nganinyi–a 1pl.excO–phave warrarn–ku. country–dat ‘The white man has taken the land from us.’ (Sharp 2004:343, 10.112) b. Kunyma–na–kata–lu tie.up–nmlz–char–erg mira–rna–yi relieve–nfut–3plS nganinyi–a 1pl.excO–phave yukurru–rrangu–ku dog–pl–dat nganarna. 1pl.exc ‘The policemen took our dogs from us.’ (Sharp 2004:343, 10.109) Note that the dative theme only occurs with verb-final -a. In an analogous construction using the locative case, there is no verb-final -a. (24) Kunyma–rna–kata–lu tie.up–nfut–char–erg mira–rna relieve–nfut nganinyi 1pl.excO yukurru–rrangu—ngu. dog–pl–loc ‘The policeman took away our dogs from us.’ (Geytenbeek and Geytenbeek 1991) The verb jami -L ‘withhold’ has the schema ERG deprives ABS of DAT when the absolutive argument is not third person singular. (25) Jami–lama–rna withhold–fut–1sgS janinyi–a 3plO–phave marrngu–rrangu person–pl ngajarri–rrangu stranger–pl kuyi–ku. meat–dat ‘I deprived the strangers of meat / I didn’t give any meat to the strangers.’ (Sharp 2004:344, 10.114) 2I ignore the problem of third person singular, which occurs as a dative or locative argument for all three verbs. There are not enough examples for me to make any generalizations. 8 The verb yu-/yi -NG ‘give’ is unusual in that both the indirect object and the di- rect object have absolutive case morphology. The schema is ERG gives ABS(recipient) ABS(theme) for non-third person singular recipients. (26) Kuyi meat yu–ngkuluma–rna give–fut–1sgS pulinyi–a 3duO–phave ngaju–lu 1sg–erg pulany. 3du ‘I will give those two some meat.’ (Sharp 2004:342, 10.105) It can also occur with other transitive verbs, such as kampa-L ‘cook’. It is unfor- tunate that in this example the theme is not overt. I suspect it would have dative morphology like those discussed above. (27) Warrukarti–lu night–erg kampa–lami–nyi–a cook–fut–1sgO–phave wika–nga. fire–loc ‘S/he will cook (the grasshoppers) for me on the fire tonight.’ (Sharp 2004:123, 4.14) Pylkkänen (2002) distinguishes two types of low applicatives. Both involve a transfer of possession relation between two individuals, but for low recipients the direction of transfer is to the possession of the indirect object, while in low source recipients the direction of transfer is from the indirect object. This analysis is dif- ferent from that of Landau (1999), who argues that so-called ‘possessor datives’ are a different construction entirely from double object constructions. The English dou- ble object construction is a low recipient applicative. Other languages, like Korean, allow source indirect objects. (28) English a. Bill sent the man the letter. b. John gave the girl the flower. (29) Korean Totuk–i thief–NOM Mary–hanthey Mary–DAT panci–lul ring–ACC humchi–ess–ta. steal–PAST–PLAIN ‘The thief stole a ring from Mary.’ (lit: the thief stole Mary a ring) Interestingly, double object constructions in Nyangumarta are used for both di- rections of transfer. For the verb mira-L an object is removed from the possession of the indirect object, and for the verb jami -L an object is withheld from the rightful 9 possession by the indirect object. For the verb yu-/yi -NG ‘give’, the indirect object comes into the possession of the direct object by means of the transfer3. If the double object construction in Nyangumarta can only be used in situations with a transfer of possession, then it should be impossible with static verbs and verbs of destruction. Unfortunately the sources do not have negative data on this point, but I predict that the following should be ungrammatical. (30) Predicted ungrammatical *Pulany–ju 3du–erg warli–rna hold–nfut pulu 3duS nganinyi–a 1pl.excO–phave nganarna 1pl.exc kurtan–ku. bag–dat ‘They (two) held us the bag.’ (31) Predicted ungrammatical *Mayi–ku vegetable.food–dat nga–na–rna eat–pst–1sg pulinyi–a 3duO–phave ‘I ate them the food..’ Syntactically, low applicatives should only be able to apply when there is a direct object. Unfortunately I do not have negative evidence, but I predict that the double object construction should be ungrammatical with unergatives, including unergatives with an implied object. (32) Predicted ungrammatical *Julurr corroboree kurnta–rna sing–nfut nganinyi–a 1pl.excO–phave ‘He sang us a corroboree song.’ (33) Predicted ungrammatical *Kampa–rna cook–nfut nganinyi 1pl.excO–phave #‘He cooked us (damper).’ (Shoul only mean ‘He cooked us (alive)’.) All instances of the Nyangumarta double object construction involving -a are ones where the indirect object c-commands the theme. Harley (2002) argues that the applicative which introduces the indirect object in these cases is phave, the same element which is used in possession constructions. She claims that only languages which allow possessors to c-command possessees, either with phave or a verbal ele- ment like HAVE which bundles phave into a verb, will allow double object construc- tions where the goal c-commands the theme. Languages which lack Phave do not 3‘Give’ is very unusual, as most verbs with recipients, paths, or locations have locative arguments. 10 allow possessors to c-command possessees, and only have double object constructions where themes c-command goals. The double object construction in Nyangumarta discussed above is one in which the goal c-commands the theme. By hypothesis then, Nyangumarta should allow either a verb or a prepositional possessive construction in which goals c-command themes. This is indeed the case, as one way of denoting possession is with the verb kalku-L ‘have’ (which also can be used to mean ‘keep, take care of’). (34) Marlu many yukurru dog kalku–rnu–n keep–nfut–2sgS janiny? 3plO ‘Do you have many dogs?’ (O’Grady 1964:18) (35) Jurni–lu kestrel–erg kalku–nul–pulinyi keep–rempst–3duO ngalyun–jirri. woman–du ‘A Kestrel had two wives.’ (Dodd and McKelson 2007:80) If the low applicative -a is phave, then we might expect to see it in other types of possessive structures, such as possessor datives. Since possessors must be animate, it also predicts that the indirect object of Nyangumarta double object constructions cannot be inanimate. I leave these and other predictions for next semester, however. 5 Inanimate dative agreement There is a homophonous verb-final -a which is not used in double object construc- tions. Specific animate dative arguments show indirect object agreement on the verb. When the dative argument is non-specific or inanimate, then -a is used. For example, in sentences where the benefactive is inanimate, as for ngurra-ku ‘camp-dat’ below, then the verbal agreement is -a (compare to (18) above). (36) Kulpa–nya–marninyju arrive–nmlz–as wika fire ma–na–yirni–a get–nfut–1pl.excS–inanD ngurra–ku. camp–dat ‘Before getting back we got firewood for camp.’ (Sharp 2004:390, 11.68) There are a range of other uses of the dative, including purposive goals of motion verbs, and the objects of transitive verbs like ‘search for’ and ‘wait for’, as well as intransitive verbs like ‘like’. In all cases, inanimate and non-specific animate objects trigger -a on the verb, while specific animate objects trigger the corresponding indirect object suffix. 11 (37) Purposive goal a. Pala–nga that–loc pirirri–jirri man–du ya–na go–nfut pula–lu 3duS–3sgD maja–murniny–ku. boss–own–dat ‘And then the two men went for their boss.’ (Sharp 2004:148, 4.137) b. ya–nal–pulu–a go–rempst–3duS–obv pala–ku that–dat–meatdat kuyi–ku ‘The two went [for meat].’ (Dodd and McKelson 2007:76) (38) ‘Wait for’ + dat a. Mima–rna–ya–lu wait.for–nfut–3plS–3sgD pirirri–rrangu–lu man–pl–erg maruntu–ku. goanna–dat ‘The men waited for the goanna.’ (Sharp 2004:125, 4.25) b. Kulpa–nya return–nfut pulu 3duS mima–rna wait.for–nfut pulu–a 3duS–inanD wirlarra–ku. moon–dat ‘Those two returned and waited for the moon (to come up).’ (Sharp 2004:349, 10.130) (39) ‘Like’ + dat a. Pirirri man karri–nyi–lu like–nfut–3sgD mirtawa–ku. woman–dat ‘The man likes his woman.’ (Sharp 2004:354, 10.158) b. Pupuka frog karri–nyi–a like–nfut–inanD kuyi–ku. meat–dat ‘The frog likes meat.’ (Sharp 2004:354, 10.157) None of these are low applicatives. The purposive goals and emotive verbs occur with intransitive verbs, whereas we know that low applicatives may only occur with verbs with an internal object. None of the examples involve a transfer of possession relation, nor do they relate two individuals to one another. Furthermore, there is evidence from (23a) and (23b) that the applicative does not vary based on animacy like these datives. I conclude that the -a in these cases is not the same as phave -a used in double object constructions. 6 Conclusion This paper is a first investigation of verb-final -a in Nyangumarta. I have argued that it is not a purposive marker, but is in fact two different homophonous suffixes. One of the suffixes is a low applicative which is used in double object constructions and 12 in certain possessive contexts. It patterns like the phave posited by Harley (2002). The second suffix is part of the verbal agreement paradigm an agrees with inanimate and non-specific animate dative arguments. Recognizing that instances of -a are not homogenous allows us to make testable predictions about the syntax and semantics of Nyangumarta datives and applicatives. 13 Appendix: Verbal agreement suffixes The symbol V represents a vowel whose vowel quality is predictable and determined by rules of vowel harmony. Person Subject Object Indirect 1sg -rnV -nyV -jV 1du.inc -li ngalinyi ngaliku 1du.exc -layV ngalayVnyi ngalayVku 1pl.inc -nyV nganyjurrVnyi nganyjurrVku 1pl.exc -yirnV nganVnyi nganVku 2sg -n(pa) -ntV -ngu 2du nyumpulV nyumpulVnyi nyumpulVku 2pl nyurrV nyurrVnyi nyurrVku 3sg -Ø/-rrV4 -Ø -lu/-lV (anim), -a (inan) 3du pulV pulVnyi pulVku 3pl -yV janVnyi janVku Table 1: Nyangumarta verbal pronouns Bibliography Bresnan, J. and Moshi, L. (1993). Object Asymmetries in Comparative Bantu Syntax, pages 50–93. Number 38. Center for the Study of Language and Information. Chomsky, N. (1995). The Minimalist Program, volume 28 of Current Studies in Linguistics. Dodd, T. and McKelson, K. (2007). Nganarna Nyangumarta Karajarrimili Ngur- ranga [We Nyangumarta in the country of the Karajarri]. Wangka Maya Pilbara Aboriginal Language Centre, South Hedland. Geytenbeek, B. and Geytenbeek, H. (1991). (interim) Nyangumarta-English dic- tionary, with English-Nyangumarta finderlist, may 91. Ts, Pilbara Aboriginal Language Centre. Harley, H. (2002). Possession and the double object construction. Linguistic variation yearbook, 2(1):31–70. Kratzer, A. (1996). Severing the external argument from its verb. Phrase structure and the lexicon, 33:109–137. 4This suffix is used in the Northern dialect only. 14 Landau, I. (1999). Possessor raising and the structure of vp. Lingua, 107(1):1–37. McGinnis, M. (2001). Phases and the syntax of applicatives. In PROCEEDINGS- NELS, volume 31, pages 333–350. O’Grady, G. N. (1964). Nyangumata grammar. Number 9. University of Sydney. Pylkkänen, L. (2002). Introducing arguments. PhD thesis, Massachusetts Institute of Technology. Sharp, J. (2004). Nyangumarta: a language of the Pilbara region of Western Aus- tralia. Pacific Linguistics, Canberra. 15


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items