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
|
participant will be marked [ +pivot]; all other core arguments are [−pivot]. Thus, we have the following distribution of features: A typology of information packaging 391 (80) [ – oblique] [ + oblique] [ + pivot] [ – pivot] This set of features defines the array of grammatical relations for English and many other languages. Any participant which is assigned an [a] feature from its governing verb is also assigned the feature [ −oblique]; otherwise it gets [ +oblique]. This defines the first tier of grammatical relations and appears to be universal. Then, among the participants which have been assigned [ −oblique], one is assigned [ +pivot], the rest [−pivot]. The basis for this assignment in English and other languages is straightforward: assign [ +pivot] to the most prominent argument in the argument structure. When there is more than one argument in the argument structure, as with transitive or ditransitive verbs, the most prominent argument is the one linked to the [ +a] participant via the precedence algorithm [ +a] > [−a] > [ ], so that the [+a] will be assigned [ +pivot] and the sole or double [−a] argument(s) [−pivot]. Using this feature analysis, traditional English grammatical relations can be described as: (81) SUBJECT OBJECT OBLIQUE – oblique + pivot – oblique – pivot [ + oblique ] A sentence like (82) Egbert presented Mildred with an award would have the following representation: (83) lexical semantic structure: present : someone causes someone to have something [ + A ] [ – A ] [ ] argument structure: present grammatical function structure: – oblique + pivot – oblique – pivot [ + oblique ] 392 William A. Foley All languages have grammatical functions because all languages have a con- trast of core vs oblique nps ([ −oblique] vs [+oblique]). This is the minimal system of grammatical functions permissible and it does appear that there are languages in which this minimal system is all that there is. In other words, there are languages – indeed very many, perhaps the bulk of them – in which the concept of pivot as defined above, and therefore the features of [ ±pivot] for core arguments of their verbs, is entirely absent. Yimas is a very strong candi- date for such a language. In example (74) above, we noted constructions like nonfinite complements in which the most prominent argument of an argument structure is the target for control, e.g. the [ +a] argument of a transitive verb and the sole argument of an intransitive verb, be it unergative ([ +a] argument) or unaccusative ([ −a] argument). There are also constructions in Yimas, how- ever, in which the [ −a] argument of a transitive verb or the sole argument of an intransitive or unergative unaccusative verb is the target of the construction – for example, the scope of elevational/directional affixes: (84) (a) kay i-na-l-ampu-n canoe.viii.sg viii.sg.s-prog-down-float-pres [ −a] ‘The canoe is floating down there’ (b) ka-mpu-tra-ya-n likely -3pl.s-about-come-pres [ −a] ‘They can come about’ Elevational/directional affixes indicate the spatial coordinates of the conceptual event described by the clause, notions like ‘up’, ‘down’, ‘in’, ‘out’, ‘about’, etc. With intransitive verbs it is the location/direction of the single argument which is predicated, but with transitive verbs it is always that of the [ −a] participant, as the transitive equivalents of (84) demonstrate: (85) (a) kay nan-l-arm-na- ŋ kan-i canoe.viii.sg imper.pl-down-board-imper-pc-viii.sg [ +a] [ −a] ‘You few board the canoe down below’ *‘You few down below board the canoe’ (b) nmpi ay-cra-wampak- a- ŋ kt leaf.vii.pl hort.pl-about-throw-imper-pc [ −a] [ +a] ‘Let us few send messages about’ *‘Let us few about send messages’ A typology of information packaging 393 In both cases it is the location of the [ −a] participant which is specified by the elevational/directional affix; the canoe is down in (85a) and the messages are to be sent about in (85b). The single argument of (84) corresponds to the [ −a] argument of the transitive verbs in (85) and both are under the scope of the eleva- tional/directional prefixes. There seems to be no coherent notion of pivot in these Yimas data; consider for illustration, just transitive verbs: control in nonfinite complements targets [ +a] arguments, while scope of elevational/directional affixes targets [ −a] arguments. But in fact the picture is even more complicated. Control in nonfinite com- plements and elevational/directional scope in Yimas are unusual in showing a preference among core arguments, albeit different core arguments. The large majority of grammatical constructions in Yimas show no such preferences: any core argument, regardless of its function, is a potential target. A good example is clause chaining. Unlike Iatmul and most other Papuan languages (see (73)), Yimas does not restrict clause chaining to situations of coreference among the most prominent arguments in the argument structure. Dependent verb clause chains can be formed as long as there is a coreferential core argument: (86) (a) tmal kray-mpi ya-kay-am-wat amtra sun.v.sg dry-dep v .pl.p-1pl.a-eat-habit food.v.pl [ −a] [+a] ‘The sun having dried it, we always eat the food’ (b) panpan-tat-mpi mnta narma ŋ pound sago-start-dep then woman.ii.sg ŋ ka-pu-k-mp-n pia-n-i-k-nakn go by land-toward-irr-vii.sg-obliq talk p-3sg.a-tell-irr-3sg.du [ −a] [+a] [ −a] ‘(He) starting pounding sago and then (his) wife comes toward him, she tells him’ In (86a) the argument shared by the two clauses is amtra ‘food’, functioning as [ −a] in both clauses. For (86b) the shared argument between the dependent clause and the final independent clause is the understood panmal ‘male, hus- band’. This functions as the sole [ +a] argument in the first dependent clause with the intransitive unergative verb panpan-tal- ‘start pounding sago’, but as a [ −a] participant in the final independent clause. Clearly, these clause chaining constructions are much more freely formed in Yimas than in Iatmul – indeed, on any shared core argument. Investigation of many other Yimas constructions would force similar con- clusions. The fundamental principle organizing Yimas grammar is the con- trast between the grammatical functions [ −oblique] versus [+oblique]. Oblique |
