Language Typology and Syntactic Description, Volume I: Clause Structure, Second edition
Download 1.59 Mb. Pdf ko'rish
|
Lgg Typology, Synt Description v. I - Clause structure
ira-wampaki- ‘throw/send something
toward someone’ is a ditransitive verb, derived from the transitive verb root wampaki- ‘throw/send something’, as in (176a), plus the allative applica- tive prefix ira-, which adds the meaning component ‘toward someone’. Note that, while the verb in (176a) is formally transitive, with two pronomi- nal affixes, wa- (ix.sg[ −a]) and n-(3sg[+a]), that of (176b) is ditransitive, having three pronominal affixes, realizing two [ −a] participants, one with the absolutive prefix wa-(ix.sg[ −a]) and the other with the dative suffix - mpun (3pl[ −a]), the latter corresponding to the [+oblique] participant in (176a). A ditransitive verb derived by applicative formation is indistinguish- able in grammatical behaviour from a basic underived ditransitive verb like ŋ a- ‘give’. This process of derivation illustrated in (176) can be represented as (177): 440 William A. Foley (177) [ + A ] [ – A ] [ ] wampaki- VTR ‘throw/send’ z [+oblique] n- 3 SG [ + A ] wa- IX . SG [ – A ] mpu-nampan 3 PL -toward ira- ALL [ + A ] [ – A ] [ – A ] ira-wampaki- VDTR ‘throw/send toward’ z> [–oblique] n- 3 SG [ + A ] wa- IX.SG [ – A ] -mpun 3 PL [ – A ] [–oblique] [–oblique] [– oblique] [– oblique] Applicative constructions in asymmetrical ergative–absolutive languages have additional complications. This is due to the one pivot per clause constraint. Because the unmarked pivot choice for asymmetrical ergative–absolutive lan- guages is the [ −a] argument, the one pivot per clause constraint prohibits multiple [ −a] arguments in these languages. We have already seen in section 2.3 that these languages proscribe underived basic ditransitive verbs, so that verbs like ‘give’ in these languages are formally transitive. The same holds for verbs derived by applicative formation. They cannot be formally ditransitive. So when an applicative derivation does operate on a transitive verb root in these languages, it also has the effect of a backgrounding passive, so that the [ −a] of the underived transitive verb appears as [+oblique], and the derived applicative verb is only formally transitive. Consider these examples from Dyirbal: (178) (a) balan d ugumbil ba ŋ gul ya a- ŋ gu det.abs woman.abs det.erg man-erg ba ŋ gul yugu- ŋ gu balga-n det .instr stick-instr hit-tns ‘The man is hitting the woman with the stick’ A typology of information packaging 441 (b) bala yugu ba ŋ gul ya a- ŋ gu balgal-ma-n det.abs stick.abs det .erg man-erg hit-applic-tns bagun d ugumbil-gu det .dat woman-dat ‘The man is hitting the stick against the woman’ Dixon (1972) Example (178a) is a normal transitive construction with the transitive verb balgal- ‘hit’ with two [ −oblique] arguments, the [+a] ya a- ŋ gu (man-erg) ‘man’ and the [ −a] dugumbil ‘woman’. As Dyirbal is an asymmetrical ergative– absolutive pivot language, the pivot is the [ −a], which has absolutive case, the case of [ +pivot]. The clause also contains a [+oblique] instrument partici- pant np, yugu- ŋ gu (stick-instr) ‘with the stick’ (the ergative and instrument case forms are homophonous, but Dixon (1972) presents convincing arguments that they are distinct; ergative is the case of [ −oblique] nps, while instrumen- tal occurs with [ +oblique]). Example (178b) is the corresponding applicative version of (178a), in which an applicative verb form balgal-ma- ‘hit with’ is derived. This derived form presents the former [ +oblique] instrument yugu ‘stick’ now as [ −oblique] and [−a]. As [−a] it necessarily assumes pivot status and appears in absolutive case. Because of the one pivot per clause constraint, the [ −a] of (178a) dugumbil ‘woman’ can no longer be pivot and so, to prevent this, the applicativization also has the effect of a backgrounding antipassive, blocking the linking of the [ −a] participant dugumbil ‘woman’ to [−oblique], so that it necessarily appears as [ +oblique], in (178b) in the dative case, the prototypical case of [ +oblique] [−a] nps in Dyirbal antipassive constructions. 4.4 Summary of clause-internal packaging constructions We can now summarize our discussion of passive, antipassive and applicative constructions as in Table 7.4. Passives and antipassives are defined universally as blocks on the normal linking patterns in lexical entries. They derive new lexi- cal forms through overt derivational affixation or other morphological processes in the verb or verbal complex which prohibit the unmarked linking patterns of the arguments [ +a] and [−a], so that one of these fails to link to a [−oblique] function, [ +a] for a passive derivation and [−a] for an antipassive. This, de facto, forces these arguments to be realized as [ +oblique], and if this is all that occurs or is intended, then a backgrounding construction results. If, however, the point of blocking the linking of either the [ +a] or [−a] to [−oblique] is to meet certain clausal syntactic constraints, particularly those germane to pivots, then a foregrounding construction results, presenting the remaining [ −oblique] argument as the sole one of a derived intransitive verb and thereby conferring the syntactic properties of pivots upon it. It should be pointed out that, while 442 William A. Foley Table 7.4 Summary of voice constructions Passive Antipassive Applicative Core definition [ +a] = [−oblique] [ −a] = [−oblique] [ +oblique] = [−oblique] Backgrounding [ +a] = [+oblique] [ −a] = [+oblique] Foregrounding [ −a] = [+pivot] [ +a] = [+pivot] foregrounding passives and antipassives are diagnostic and prototypical of pivot languages, they may not be restricted to them. It is conceivable that pivotless languages could possess foregrounding passive constructions, if the purpose of the derivation is to get the [ −oblique] [−a] argument to assume some of the grammatical properties of the [ +a] argument, the most prominent argu- ment in the argument structure (some Bantu languages may indeed be pivotless languages with exactly this property). When this occurs with a marker in the verbal expression we can say it is a passive construction. Such a derivation, of course, would not be necessary via foregrounding antipassives, because [ +a] arguments, by a universal algorithm, are already the most prominent argument in the argument structure and the controller and target of many constructions. Applicatives differ from passives and antipassives in that they are not blocks on the linking of normally subcategorized [ +a] or [−a] arguments to [ −oblique], but the introduction of normally [+oblique] non-subcategorized arguments into the argument structure of a derived verb and linked to [ −oblique] status. The arguments introduced by the applicative affixes are realized as [ −a]. This is normally straightforward in most types of languages and has little sys- tematic effect on the lexical entries of the derived verbs, but in asymmetrical ergative–absolutive pivot languages, the one pivot per clause constraint requires the effect of a subsidiary antipassive to force the [ −a] argument of the under- ived verb into [ +oblique] status, so there is only one [−oblique] [−a] argument for the derived applicative verb, and thus one pivot. Download 1.59 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling