Subjecthood and Grammatical Relations in Korean: An Experimental Study with Honorific Agreement and Plural Copying

31st Pacific Asia Conference on Language, Information and Computation (PACLIC 31), pages 206–213 Cebu City, Philippines, November 16-18, 2017 Copyright c⃝2017 Ji-Hye Kim, Yong-Hun Lee and James Hye-Suk Yoon Subjecthood and Grammatical Relations in Korean: An Experimental Study with Honorific Agreement and Plural Copying Ji-Hye Kim Yong-hun Lee James Hye-Suk Yoon Department of English Education Department of English Literature & Language Department of Linguistics Korea National University of Education Cheongju, Korea Chungnam National University Daejeon, Korea University of Illinois, Urbana-Champaign U. S. A. jkim@knue.ac.kr yleeuiuc@gmail.com jyoon@illinois.edu Abstract The present study investigated the following: i) how NPs bearing differing GRs behave with respect to two proposed subject diagnostics – Honorific Agreement (HA) and Plural Copying on adverbs (PC) and ii) whether scrambling allows non- Subject GRs to control these properties. An experimental investigation using Magnitude Estimation (ME) was conducted. The result revealed that the sentences with Subject NP controller got higher acceptability scores compared to non- Subject NP controllers for both diagnostics and that scrambling did not have an effect on acceptability. While both HA and PC showed a similar pattern of preference for Subject controllers, the contrast between Subject and non-Subject controllers was more pronounced with HA. 1 Introduction: Subjecthood diagnostics in Korean The question of whether Grammatical Relations (GRs) such as Subject or Object are universal has been a matter of debate. While there are theories that posit GRs are theoretically central notions (Relational Grammar, Lexical Functional Grammar), there are others that do not countenance them at all but instead try to derive properties traditionally attributed to GRs from other aspects of the organization of a sentence, such as c- command among arguments (Government and Binding theory, Minimalist Program). Another debate has centered on whether GRs, as primitives or as derived notions, play a role in the grammar of all languages. Li and Thompson (1976) famously argued that there are languages where the syntactic articulation of a clause does not reference GRs but discourse relations like Topic instead. Topic- prominent languages like Chinese are argued not to utilize GRs at all, whereas Subject-prominent languages like English employ GRs centrally in the syntactic articulation of a clause. They argued that Korean may be both Topic and Subject prominent, given that it possesses signature properties of both types of languages. Against this backdrop, Sohn (1980) has argued that Korean is only Topic- prominent, with the notion of Subject playing no role. Sohn’s (1980) position has been an outlier, however. The vast majority of generative works on Korean assumes that GRs, whether as primitive or as derived notions, are central in the grammar of Korean, and many properties of Korean have been analyzed using the vocabulary of GRs and related ideas. The most extensive defense of the role of GRs/Subjects in the grammar of Korean comes from works in the RG tradition, where a representative list of properties identifying Subjecthood (Subjecthood diagnostics) was proposed (Youn 1990, Gerdts 1991, Gerdts & Youn 2001, etc.). (2) Subject Diagnostics in Korean a. Controller of optional …

Content truncated to 500 words