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
while
+ gerund construction is interpreted as if it had a subject coreferential to the subject of the matrix clause (note that if while is omitted, we immediately understand the subject of the gerund to be coreferential with the object rather than the subject of the matrix clause). On the basis of (52, 53) alone, one might venture an alternative account, in which it is the agent rather than the subject of the while + gerund construction that is suppressed, and the subject of the while construction is understood as being the same as the agent rather than the subject of the main clause. In this kind of account, we would have a direct connection between the overt form and the meaning, without an intervening level of grammatical relations. This possibility may be discounted on the basis of sentences such as John felt apprehensive while being wheeled into the operating room, in which the overt and ‘understood’ subjects are not agents, and even more strongly by examples in which the while + gerund construction is combined with the passive construction: 170 Avery D. Andrews (54) a. The student watched the guard while killing the prisoner b. The student watched the guard while being killed by the prisoner c. The student was watched by the guard while killing the prisoner d. The student was watched by the guard while being killed by the prisoner It is the subject of the matrix that is understood as the subject of the gerund, regardless of the semantic roles involved, and of how unusual the situation described is. It also seems that no well-defined pragmatic notion such as topicality is the conditioning factor, although this is hard to show conclusively, since pragmatic functions are generally more elusive and less well understood. For example, in a sentence such as A guard tortured the prisoner while watching television, it seems pretty clear that the prisoner can be the topic. Nonetheless, the principle for the interpretation of the while + gerund construction continues to operate as before. Phenomena such as these illustrate the need for a level of syntactic structure at which abstract grammatical relations such as subject are defined, which are distinct from semiotic concepts, and which are significant for the functioning of grammatical rules. From a theoretical point of view, there are three major possibilities for the analysis of while + gerund constructions. The first is that the gerund has no subject in syntactic structure, but that the principles of semantic interpretation treat it as if it had a subject coreferential with that of the matrix (main) clause. Second, the gerund might have a subject in the syntactic structure which is coreferential with the matrix subject, but which does not appear in the overt form of the sentence. The third possibility is that the theory of sentence structure characterizes the np in matrix subject position as the subject of both the main clause and the gerund. The choice between these possibilities is a complicated question, which does not concern us here. What matters here is that, whatever approach is taken, it is clear that the notion of subject plays a central and obvious role in the description of the constructions: it is the subject of the subordinate clause that is obligatorily omitted, and the subject of the matrix that obligatorily serves as its ‘controller’, that is, as the np that is understood as the subject of the subordinate clause. Subject ellipsis can often be used to provide more evidence about gram- matical relations when the coding features are equivocal. In Warlpiri, there are counterparts to the while + gerund construction that show that this language has a subject grammatical relation (one expressing s and a functions) in spite of the inconsistent testimony of the coding features. These are ‘infinitival’ subordinate clauses (adverbial or relative in sense), in which no auxiliary appears, but an ‘infinitival complementizer’ is attached to the verb, which then appears finally The major functions of the noun phrase 171 in the infinitival phrase, and can’t be reordered within it (there is, however, a possibility of nominals within the infinitive phrase ‘leaking’ out of it into the matrix (Laughren (1989))). Many of the infinitival complementizers require suppression of the comple- ment subject, imposing various conditions on what it may be understood to be coreferential with. One of these is the complementizer kurra, which expresses action simultaneous with that of the main verb, and imposes the condition that the complement subject be coreferential with a non-subject (preferably object) argument of the matrix: (55) a. Ngajulu-rlu-rna yankirri pantu-rnu, ngapa nga-rninja-kurra I-erg-1sg(subj) emu(abs) spear-past water(abs) drink-inf-while ‘I speared the emu while it (not I) was drinking water’ b. Ngarrka-rna nya-ngu wawirri panti-rninja-kurra man(abs)-1sg(subj) see-past kangaroo(abs) spear-inf-while ‘I saw a man spear a kangaroo’ c. Ngaju ka-rna-ngku marri-jarri-mi nyuntu-ku I(abs) pres -1sg(subj)-2sg(obj) grief-being-nonpast you-dat murrumurru nguna-nja-kurra(-ku) sick lie-inf-while(-dat) ‘I feel sorry for you while you are lying sick’ d. Karli-rna nya-ngu pirli-ngirli wanti-nja-kurra boomerang(abs)-1sg(subj) see-past stone-elative fall-inf-while ‘I saw the boomerang fall from the stone’ The infinitival verbs of (55a–b) would take ergative subjects if finite, those of (55c–d) absolutive. The examples also illustrate a variety of semantic roles for the omitted subject and its controller. It is crucial to the argument that kurra requires (rather than merely permits) omission of the subject: since nps can be rather freely omitted in Warlpiri, if a complementizer merely permits an omitted argument in its clause to be understood as coreferential with one in the matrix, without actually requiring omission and understood coreference, we could simply say that the omitted argument was an ellipsed anaphoric pronoun which happened to be corefer- ential with an np in the matrix clause (this would often be permitted by the usual principles governing null anaphora). There would then be no syntactic phenomenon specifically associated with the subject of a -kurra complement. The more general point is that what needs to be shown is some difference in omissibility from ordinary clauses. In English, for example, nps aren’t freely omissible, so the possibility of omission in the while + gerund construction is enough to make an argument for a grammatical relation, whereas in Warlpiri |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling