In many languages, arguments such as subjects and objects enjoy substantial freedom in terms of the form in which they are realized: a noun or pronoun with or without a following or preceding functional particle, a bare noun, a null-form argument, etc. This talk is concerned with caseless subjects in Korean, that is, those subjects that occur without functional particles signaling case or discourse function. An interesting and challenging problem for theoretical approaches to case is that caseless and case-marked subjects are in systematic contrast as to their interpretation. This makes Korean a Differential Subject Marking (DSM) language in Aissen’s (2003) terms, wherein some subjects are marked with formal particles while others are not, depending on the semantic and pragmatic features of the subject.

This talk will focus on a hitherto unexplained property of caseless-subject clauses triggering a direct perception interpretation. I will first present evidence from conversation data demonstrating that caseless subjects predominantly occur in clause types that have an agent directly identifiable in the here and now, whereas nominative-marked subjects are most productively used in clause types wherein the identification of an agent cannot be grounded in the here and now. Based on this evidence, I will propose a new account of DSM in terms of cue reliability (Rosch & Mervis 1975; Levshina 2021), arguing that the association of caseless subjects with seemingly unrelated features such as direct perception in the here and now, agentivity, definiteness, tense deficiency and a simple, thetic interpretation (the most preferred information structure associated with caseless-subject clauses) follows from an economical use of formal particles motivated by the reliability of cues for identifying or predicting the grammatical, referential, or discourse status of an argument NP or a clause containing it. When such cues are weak, sentences are likely to convey information that is less predictable and redundant; speakers prefer to mark the subject by overt particles in such sentences, as using case markers would lead to more uniform information density than leaving them unmarked.

These preliminary results support efficiency-based accounts of patterns of grammar, and underscore the importance of communicative efficiency in explaining and motivating grammatical rules and language use (Hawkins 2004, Haspelmath 2008, Jaeger 2010, Lee 2010, 2016, 2021, Lestrade & de Hoop 2016, Levshina 2021).

Hanjung Lee received her Ph.D. in 2001 from Stanford University. She held a research position at the University of North Carolina at Chapel Hill, USA and a faculty position at the University of Minnesota-Twin Cities, USA. She is currently Professor in English linguistics at Sungkyunkwan University, Korea. Lee's work has been concerned with dynamics of probabilistic grammar and competing motivations that shape grammatical rules and patterns of language use such as on-going competitions between the speaker's interests and the addressee's needs, or between constraints imposed by grammar and those imposed by online processing. She has published papers exploring how these competitions impact a wide variety of systems (case marking, word order, politeness forms, lexical and grammatical choices) in leading journals including Natural Language and Linguistic Theory (2003), Cognition (2007), Journal of Pragmatics (2007, 2013) and Journal of Linguistics (2016, 2022). She has also published numerous articles in influential handbooks and edited volumes such as The Handbook of Korean Linguistics (2015, Wiley-Blackwell), Series in Natural Language and Linguistic Theory (2008, Springer), Series in Generative Grammar (2004, Mouton de Gruyter), Palgrave Studies in Pragmatics, Language and Cognition (2004, Palgrave Macmillan), and Studies in Constraint-Based Lexicalism (2001, 2003, CSLI Publications). Currently, she is conducting research projects on i) competing motivations in the grammar of ditransitives and ii) the role of communicative efficiency in motivating and explaining case-marking systems using both data-oriented and experimental methods.