Research On Acquisition Of English Causative Alternation Under The Combined Usage-based And Rule-based Approach

The causative alternation refers to a language fact in which a change verb has both causative and inchoative construction variants,and the patient of the causative construction shares the same semantic role as the agent of the inchoative construction.The past research have generalized the relationship between alternating variants and the alternating rules of change verbs,but still failed to reasonably explain the syntactic differences and asymmetric alternation of some verb members.Moreover,although native speakers can grasp complex causative alternation,it still needs to be further investigated whether second language learners with limited input can successfully acquire the causative alternation.Few of the current research have taken the causative alternation as an independent phenomenon,rather most focus on the dyadic unaccusative verbs without including alternating monadic unaccusative and unergative verbs into the research category,or have systematically studied the four overgeneralization of verbs,namely taken overpassivization,undercausativisation,underpassivization,and overcausativisation as inter-related problems of alternating verb acquisition and the explanation on influencing factors for these problems still lacks the supporting theory of acquisition approach.To sum up,this paper tends to study the following three questions:(1)What are the more reasonable alternation mechanisms and rules for the change verbs in the causative alternation?(2)When Chinese college students acquire English causative alternation,is it difficult for them to distinguish the inchoative from passive constructions and the causative from the analytical causative constructions when each pair expresses similar meaning,thus having overpassivization and undercausativisation problems?Is the emergence of them conformed to the usage-based approach(affected by item abstraction,entrenchment,statistical preemption)or rule-based approach(affected by verb semantics and L1 transfer)?(3)When Chinese college students acquire English causative alternation,is it difficult for them to distinguish the alternating verbs from non-alternating ones,thus having underpassivization and overcausativisation problems?Is the emergence of them conformed to the usage-based approach or rule-based approach?To explain change verbs in causative alternation,based on verbs' syntactic distribution in real corpus,this study uses Lexical Construction Grammar and Causal-aspect Model to make a new interpretation of the event framework,alternation mechanism and semantic rules of change verbs.To explore overpassivization,undercausativisation,underpassivization and overcausativisation,the experiments design grammatical judgment tasks based on the context of alternating constructions and verb alternation rules respectively,using quantitative and qualitative methods to investigate Chinese college students' overgeneralization errors in the tasks.For the quantitative method,the study applies multi-factor mixed design analysis of variance to investigate the interactive effects of item abstraction,entrenchment,statistical preemption,verb semantics and L1 transfer.And later through the post hoc test,the simple effect of each factor is analyzed.For the qualitative method,this study adopts an interview questionnaire to effectively separate the L1 transfer from the preemption effect sharing the same error representation.Finally this study comes to the following results through the above research on explanation of causative alternation and acquisition experiments:(1)Change verbs consist of external caused(dyadic unaccusative)verbs and internal caused(monadic unaccusative and unergative)verbs,respectively evoke encyclopedia semantic frames of external and internal events,and follow opposite alternation mechanisms when interacting with inchoative and causative constructions.When a verb signifies an external event,the frame of the verb contains cause and patient mapped to subject and object position and the verb has causative representation,the subclass of which with descriptors showing binary semantic features to describe[+independent change]and[+definite change]can enter inchoative construction by cutting external argument.When a verb signifies an internal event,the frame of the verb only contains patient directly mapped to subject position and the verb has inchoative representation,the subclass of which with degraded internal argument can enter causative construction by adding external argument.In addition,in order to integrate with causative construction,a motion manner verb needs to combine with a prepositional phrase of location to obtain[+definite change]feature,hence merge with abstract causative construction's meaning.(2)College students have overpassivization and undercausativisation problems,and the reason for overpassivization is that they are influenced by prototypical scenario association during item abstraction,so tend to choose passive construction reflecting more typical transitive scenario when the causative behavior is mentioned in the near context.The low level are also affected by the entrenchment and preemption effect so that they do better with high frequency verbs than low ones,better with monadic unaccusative and unergative verbs in lower contingency with causative constructions than dyatic unaccusative verbs.The high level are not influenced by frequency and make significantly fewer mistakes than the low level,which demonstrates although the high level meet fossilization of high-frequency verbs to some extent,they can still have abstract mental representation of different frequency verbs due to close attention to low frequency ones,whereas the low level could only obtain verb schematization in contexts with strong cognitive hints,and although master high-frequency verbs,still have to judge low frequency ones in light of verb lexical meaning and the connection preference between verbs and constructions.The reason for undercausativisation is that the high level's judgments on internal caused verbs and the low levels'on all the verbs are affected by entrenchment.But different from overpassivization,the-high level are greatly affected because this kind of verbs have high agentivity,and their association with causative constructions representing external cause event is atypical and marked.As for the low level with overall limited input,their usages of all the verb classes have to lean on the statistical inference of sample accumulation.The high level are not affected with dyadic unaccusative because of their abundant input and competence to associate verbs with transitive scenarios.Moreover,the two levels are both affected by a waxing-waning effects of preemption and L1 transfer,i.e.with the vocabulary input(diminishing from high to low English level,from high to low frequency)decreasing,the degree of the learners affected by the preemption effect will go down while the transfer effect will go up.(3)College students also have underpassivization and overcausativisation problems in acquisition of causative alternation.The root for underpassivization is that the high level'judgments of the non-alternating asymmetric verbs,and the low levels' judgments of all the non-alternating verbs are all affected by the entrenchment.The high and low level still lag behind in acquiring a small portion of low-frequency verbs,since it is hard for them to pay attention to the complex syntax performance of the asymmetric verbs when the input is insufficient,and the low level are also affected by L1 transfer,because of the their lack of the sentence abstraction ability and too much reliance on first language.The root for overcausativisation is that the high and low level are influenced by the entrenchment effect in judging non-alternating monadic unaccusative verbs because there are great differences among members of the subclass light emission verbs,therefore whether right differentiation can be made is determined by the quantity of input.The high and low level are also affected in the same way by the semantic rules,which means they are all at their best in acquiring non-alternating unergative verbs,second in non-alternating monadic unaccusative verbs,while still fall short in the two alternating internal caused verbs and unergative verbs without adding prepositional phrase of location.This is due to the stable connection between the internal caused verbs and independent change events and compared to monadic unaccusative verbs,students are more sensitive to the meaning of unergative verbs with higher agentivity.The poor performance in acquisition of alternating internal caused verbs is owing to the weak association between the verbs and causative constructions,resulting in a very low frequency of causative construction and severe shortage of positive evidence in the input.The backward situation with unergative verbs without prepositional phrase is because students barely pay attention to the asymmetric syntax performance of these verbs in implicit learning or get access to it in explicit learning.In the non-alternating nonce-verb experiment,except for the high levels' judgments of internal caused verbs,and the low level students'judgments of 'fihh' verb,the two level's judgments of other verbs are all affected by the pre-emption effect,while restricted by verb semantics and cognitive concept in the meantime.Based on the above experiment,this research proves that compared with native speakers,the high and low level Chinese college students face different degrees of overgeneralization problems in acquisition of the causative alternation and the acquisition process is in conformity with joint path of usage-based and rule-based approach,being affected by the interactive influence of item abstraction,entrenchment,statistical preemption,verb semantics and L1 transfer.Therefore,the study further proposed that the usage-based approach should be followed so as to assist learners in taking advantage of frequency effect from multiple perspectives,including improving the frequency of input and output,using Zipf's law for reference,and using structural priming appropriately,and the rule-based approach should also be followed so as to encourage learners to acquire construction knowledge of form-meaning/function pairing in an all-round way from various aspects,including optimizing the input of explicit construction knowledge,generalizing overall abstract construction rules of verb items,and acquiring verb mini-constructions in a hierarchical manner.