Native and non-native processing of structural ambiguities
Marina Y. Sokolova · ePrints Soton (University of Southampton) · 2019
The main goal of this thesis is to approach adult second language L2 acquisition through the analysis of the processing behavior of non-native speakers of English and Russian at the intermediate level of L2 proficiency. Specifically, the thesis investigates how sentence processing in L2 is performed and whether the attested processing behavior can be explained by the existing theories of L2 acquisition. The thesis claims that in both native and non-native languages sentence processing is incremental with a structural parse being its initial stage. Structural parse combines top-down and bottom-up algorithms of mental structure-building which complement each other throughout the sentence. The dissertation believes that L2 acquisition is UG-governed and the human parser is sensitive to the properties of the L2. It is within the capacity of the human parser to spot the minimal differences between the L1 and the L2, assign new meaning to the L2-specific properties and start their acquisition. The experimental part consisted of two self-paced reading experiments investigating (a) monolingual speakers of English and Russian, and (b) L2 speakers of English and Russian at the intermediate level of L2 proficiency. The L2 participants were divided into two sub-groups to be tested in either their respective L1s or L2s. The results of the experiments show that in both native and non-native languages, sentence parsing begins with a top-down structural prediction. This prediction undergoes a bottom-up check for grammatical fitness of the incoming constituents. During the grammatical check, the structure can be amended if needed and a new projection is generated. The cycle repeats. The study established that L2 speakers are sensitive to highly salient L2-specific linguistic phenomenon. However, the participants mainly show L1-like behavior in the L2. The findings add to the existing literature in L2 acquisition. The thesis goes in line with the theories that argue for L1 transfer in L2 acquisition and shows that at the intermediate level of L2 proficiency, the parser is implementing the L1 parsing hypothesis to the new linguistic material. The thesis captures the stage of L2 acquisition, when the parser starts showing sensitivity to L2-specific linguistic cues. It means the effects of L1 transfer are being overcome and target-language-like behavior in the L2 has started developing. Besides theoretical implications, the dissertation contributes to the methodological aspects of studies in L2 processing. It shows that using bilingual speakers tested in the L1 as a control group for the bilinguals tested in the L2 provides a clearer picture of the non-native behavior and is more informative for psycholinguistic research.