Language Typology and Syntactic Description, Volume I: Clause Structure, Second edition
participants who are acted upon by the instigating agent of a causative and then
Download 1.59 Mb. Pdf ko'rish
|
Lgg Typology, Synt Description v. I - Clause structure
participants who are acted upon by the instigating agent of a causative and then produce the effect described: (130) a. ə y-n ə ma-bu sel pi he-nom I-acc money gave ‘He gave money to me’ b. ə y-n ə ma-bu teb ə l ilh ə lli he-nom I-acc table caused.to.push ‘He made me push the table’ These uses are subject to the restriction that none of the other arguments of the verbs be able to take bu; if this condition isn’t met, the locative d ə is used instead, and can be used in any event (so both instances of bu in (130) could be replaced with d ə ). 214 Avery D. Andrews There are several complexities in this system which we can’t look at here: a. The predictable uses of n ə and bu are optional. The full circumstances are not entirely clear to me from Bhat’s discussion, but it seems that, for example, any of the case-markers in (128) could be omitted. b. The markers have additional uses to indicate strictly pragmatic functions, in which case they are placed after instances of the markers that are signalling semantic roles (Bhat (1991:126–30)). Now, considering the issue of a-subjects, a proponent of the universality of grammatical relations could suggest that the marker n ə , in its function as a semantic role marker, applies only to a/s, that is, a-subjects (since this language has no passive rules), therefore providing some evidence for the relevance of an a-subject concept. The counter-argument is that the distribution of the n ə marker can be characterized in purely semantic terms, along the lines of ‘instigating and intending agent’ (someone who does something because they want to do it). This would be expected to prevent n ə from appearing on a causee agent, because such an agent is being described not as doing something because they want to, but because the instigatory agent makes them do it. So it is certainly plausible that the distribution of n ə , insofar as this is related to semantic roles, might be determined directly by its semantic role of volitional agent, rather than involving a-subject or other grammatical relations as an abstract intermediary. Similarly, the distribution of bu might well be determined by a semantic role, although it is not so clear from the evidence given exactly what that role would be. But an indication that a semantic role rather than a grammatical relation might be the crucial factor is provided by certain negative sentences, which can have bu rather than n ə on their agents: (131) a. ma-bu laktre he-acc came.not ‘He didn’t come’ b. layriksi ə y-bu padri book.this he-acc read.not ‘He didn’t read this book’ Bhat suggests that the accusative is motivated by an implication that some outside influence affected the agent, preventing them from performing the action (Bhat (1991:122–3)). This is evidence that the distribution of bu is determined by a semantic role along the lines of ‘something that is influenced’, rather than by a grammatical relation such as ‘object’. Bhat considers various other phenomena beyond case marking which might involve grammatical relations in Manipuri, and finds no evidence The major functions of the noun phrase 215 that they do. For example, there is a participial construction which doesn’t allow both clauses to contain non-coreferential actors/causers, but does allow both clauses to contain coreferential actors/causers, only one of which is expressed: (132) a. Ra:ju akki tandu be:yisida Raju rice brought(past.ptcpl) cooked ‘Raju brought the rice and cooked it’ b. *Ra:ju akki tandu hari be:yisida Raju rice brought(past.ptcpl) Hari cooked ‘Raju brought the rice and Hari cooked it’ Bhat (1991:75) However, if only one clause contains an actor/causer, or neither clause does, then no coreferential argument is required: (133) a. avanu be:gane bandu namage tondard a:yitu he(nom) early came(past.ptcpl) us(dat) trouble became ‘We were troubled by his coming early’ b. mara biddu ma:du muiyitu tree fell(past.ptcpl) roof broke ‘The tree fell and the roof broke’ Obligatory ellipsis and understood coreference of an argument frequently pro- vides evidence for a grammatical relation, but not in this case, because the constraint against this construction of having non-coreferential actors/causers appears to be statable in entirely semantic terms. The conclusion is that no grammatical relations at all, including a-subject, are required to describe the grammatical structure of this language. Although this is a very interesting result, it is important to keep in mind that it is inherently difficult to prove a negative, and a few dozen pages of a single investigator’s work can’t provide conclusive proof that grammatical relations truly play no role at all in the language. An example of a potential issue might be whether you could say something such as: (134) ma-bu l ə ppi he-acc cried ‘He cried [because of something somebody did to him]’ If this is acceptable, then the account of the accusative cases in (131) would be corroborated. If not, then it might be a problem to devise a meaning for -bu that allowed (131) while excluding (134), and, consequently, there might be a role |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling