http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : Michael Frazier (검색결과 3 건)
- 무료
- 기관 내 무료
- 유료
-
김나연,Michael Frazier 담화·인지언어학회 2023 담화와 인지 Vol.30 No.4
The understanding of elliptical constructions, in which material not present in the phonological or written form of a sentence is nonetheless present in its interpretation, presents a number of challenges for the sentence processing system. Specifically, upon identifying a Noun Phrase Ellipsis construction, readers must identify the antecedent in the memory and retrieve its contents (Kim, Brehm, and Yoshida 2019, Kim and Frazier 2023). In this study, based on Kim and Frazier (2023), we investigated the time-course of elliptical dependency resolution with strict/sloppy identity ambiguity using an eye-tracking while reading experiment, to distinguish what is retrieved in antecedent retrieval at earlier from later stages in processing and how retrieval proceeds. Our results revealed that the strict identity interpretation is preferred at initial stages of processing, indicating that coreference information is retrieved from memory (Kim and Frazier 2023). However, the sloppy identity interpretation shows clear signs of being entertained in later eye-tracking measures as well as in later regions, suggesting that coreference information is temporally prioritized over morphological cues in cue-based retrieval system of content-addressable memory access (Lewis, Vasishth, and Van Dyke 2006; Martin and McElree 2008, 2011). Our result supports Kim and Frazier’s (2023) overall conclusions regarding the parser’s prioritization and the representational content it can access in antecedent retrieval, and extends this finding to later stages of antecedent-ellipsis integration, when information accessible to the parser less rapidly can be integrated into the ongoing parse.
-
The Effect of Antecedent Complexity in the Processing of Noun Phrase Ellipsis
김나연,Michael Frazier 담화·인지언어학회 2022 담화와 인지 Vol.29 No.4
The current study investigated the effect of antecedent complexity on the time-course of antecedent retrieval for Noun Phrase Ellipsis (NPE) with the aim to understand whether the structural content associated with the antecedent have an impact on the online processing of NPE. We found that while NPE structures are more taxing to process in real-time than similar constructions with pronominal anaphora, the complexity of the antecedent yielded no observable reading-time differences nor differences in accuracy. This pattern of results is broadly compatible with both pointer-based and cost-free copying accounts of ellipsis processing (Frazier and Clifton 2001; Martin and McElree 2008, 2009, 2011) but inconsistent with approaches to ellipsis resolution in which the antecedent must be re-built in the ellipsis-site. However, crucially, the absence of an observable antecedent-competition effect (Van Dyke and McElree 2006; Jäger, Engelmann, and Vasishth 2017) is suggestive evidence in favor of the cost-free copying accounts (Frazier and Clifton 2001, 2005). Future work should attempt to expand the range of complexity-based manipulations (in both the degree and the nature of complexity) to ensure that results of this kind are reliable across a range of antecedent structures.
-
Processing of Referential Non-Isomorphism in Noun Phrase Ellipsis
김나연,Michael Frazier 한국언어학회 2023 언어 Vol.48 No.1
Anaphoric dependencies can span the boundary of an ellipsis-site, as in sentences like “Bob’s photo of himself from the birthday party was beautiful like Pablo’s [Noun Phrase Ellipsis] from the pool party”. These dependencies are known to exhibit a systematic ambiguity between an interpretation where anaphors interpreted in the deleted portion of the sentence can be interpreted as bound by either their nearby candidate antecedent for the reflexive in its local clause proximal to the NPE-site or by the structurally analogous noun phrase in the antecedent clause. In the above example, the ambiguity is whether the reflexive in the ellipsis is co-referent with Pablo or with Bob. This study presents the results of an experiment on the real-time processing of this so-called strict/sloppy-identity ambiguity to examine which reading the parser preferentially adopts. The study finds that the strict reading is preferred on-line, indicating that coreference information is available to the parser, and suggests implications for this finding to better understand the grammatical representation of coreference.
KCI
DBpia연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
