A-Scrambling and Subjecthood in Korean: An Experimental Approach

A-Scrambling and Subjecthood in Korean: An Experimental Approach James Hye-Suk Yoon Ji-Hye Kim* Department of Linguistics Department of English Education University of Illinois, Urbana-Champaign U. S. A. Korea National University of Education Cheongju, Korea jyoon@illinois.edu jkim@knue.ac.kr Abstract The present study investigated whether the local/A-scrambling of non-subject NPs to clause-initial (a potential subject) position can turn them into controllers of subject properties such as Honorific Agreement and Plural Copying. The study was motivated by findings which showed that A-scrambled non-subjects can control some subject properties, such as reflexive binding, from the scrambled position. An experimental study was conducted with Korean native speakers using Magnitude Estimation to test the acceptability of the sentences with subjects and scrambled non- subjects functioning as the controller of these properties. Overall results revealed that scrambled non-subjects cannot control these subject properties, despite the fact that scrambling could have placed them in a subject position (SpTP). Implications of the findings for the research on subject properties in Korean and more generally are discussed. 1 Introduction: Scrambling in Korean Since Saito (1985), it has been known that local scrambling in Japanese/Korean has both A and A’ properties. However, under Saito’s analysis of scrambling as adjunction to TP, A-scrambling behavior was difficult to account for, since adjoined positions are by definition A’-positions. Utilizing later developments in the analysis of clause structure, Miyagawa (2001) offered an elegant solution to this conundrum by analyzing A- scrambling as movement/substitution to SpTP, an A-position, with the thematic subject staying in SpvP, under the assumption that subjects may, but need not, move to SpTP in languages like Japanese/Korean. It is notable that in Miyagawa’s analysis, the A- scrambled object occupies a position typical of subjects and is predicted to behave like a subject, given that SpTP is the canonical derived/high subject position across languages (McCloskey 1997). A-scrambled objects do indeed control certain properties typical of subjects, such as the binding of anaphors and having wide scope over the base subject. The object controls these properties only when A-scrambled but not when it remains in-situ, as well-documented in both languages. These facts seemingly lend support to Miyagawa’s analysis that takes A-scrambling to be an instance of movement to a SpTP, a subject position. However, work on A-scrambling has shied away from making an explicit claim that A-scrambled non-subjects end up in a subject position. This may be because (anaphor) binding and wide scope are not the only properties controlled by subjects. There are other subject properties. For Korean, control of honorific marking –si on the predicate (Honorific Agreement, HA) and of the copied plural marker –tul on constituents within the VP (Plural Copying, PC), are also deemed to be the prerogatives of a subject (S. Song 1995, J. Song 1997, Cho 2000, Choe 2004, Choi 2010, etc.). Nevertheless, an object fronted by A-scrambling (as diagnosed by anaphor binding into the base subject) fails to act as the controller of these properties, as seen below. (1) a. *Kim-kyoswu-nim[1]-ul [caki[1] Kim-professor-HON-ACC self citohaksayngtul]-i ___ chaca ka-si-ess-ta advisees-NOM …

Content truncated to 500 words