Poster A34, Thursday, August 16, 10:15 am – 12:00 pm, Room 2000AB

Differentiating three different types of double subjects/Nominatives in Korean: An ERP-based study

Kiyong Choi1, Daeho Chung2, Wonil Chung3, Say Young Kim2, Bum-Sik Park3, Myung-Kwan Park3;1Kwangwoon University, 2Hanyang University, 3Dongguk University

Building on Lee’s (2014), Kim’s (2015), Kim, Kim, and Yoon’s (2015) offline judgment tests of double Nominative-Case-marked subject (DNS) constructions in Korean, we performed an online ERP (event-related potential)-based examination of processing three different types of DNSs by Korean native speakers. In particular, the target word in these three constructions is the second Nominative Case-marked subject NP. The experimental materials consisted of 180 sets of 3x2 factorial design with 3 types (Property-type, Family Relation-type, and Adjunct-type DNSs) and 2 Case (Nom and Gen/Loc), constructed on the model of KKY (2015). They, however, were composed of double Nominative subjects situated in right dislocation contexts, schematically represented below. Conditions (i)-(ii): Property—Genitive/Nominative: kwake-ey yeypp-ess-e, yengi-uy/-ka maum-i. past-in kind-Pst-Informal, Yengi-Gen/Nom heart-Nom ‘It was kind in the past, Yengi’s heart.’ (iii)-(iv): Family Relation—Genitive/Nominative: kwake-ey yeypp-ess-e, yengi-uy/-i tongsayng-i. past-in pretty-Pst-Informal Yengi-Gen/-Nom sister-Nom ‘She was pretty in the past, Yengi’s sister.’ (v)-(vi): Adjunct—Locative/Nominative: kwake-ey cekes-se, sewul-ey/-i chatul-i. past-in a few-Informal Seoul-Locative/-Nom cars-Nom ‘There were a few in the past, Seoul’s cars.’ Eighteen (male: 12) native Korean undergraduate students participated in the experiment. The descriptive data of the offline 4-scale (1: not at all acceptable, 4: definitely acceptable) acceptability task for the 6 conditions are as follows: (Condition (i)): (P)roperty SSs (single subject) - 3.7; (ii): P-type DNSs - 2.4; (iii): Family Relation (FR) SSs - 3.6; (iv) FR-type DNSs – 1.5; (v): (A)djunct SSs - 3.4; (vi): A-type DNSs – 2.2. The ANOVA results of the offline task showed significant effects of factors such as type (F(2,34)=12.26,p<0.001), Case (F(1,17)=178.27,p<0.001), and type*Case interaction (F(2,34)=12.11,p=0.001). Korean speakers rated the acceptability of P-type DNSs higher than that of A-type DNSs, and the acceptability of P-type DNSs higher than that of FR-type DNSs. ERPs were measured at the second subject NP in DNS conditions and the post-Genitive/Locative subject NP in single subject conditions. In Property-type condition there was an LAN effect in the 400-500 ms interval. In contrast, in Family Relation-type condition there was a P600 effect at the posterior in the 600-700 ms. Furthermore, in Adjunct type condition there was a marginal P600 effect at the posterior in the 600-700 ms. The results indicated that the LAN effect in Property-type condition is an index of difficulty in carrying out theta-identification between the double Nominative subjects based on the same Nominative Case marking (Yoon 1990). The second Nominative subject in this type of DNS enters into ‘agreement’ relation via theta-identification with the first one. The P600 effect in Family Relation-type condition is a signal of syntactic revision/reanalysis that has been reported to arise in sentences involving garden path. The second Nominative NP in this type of DNS is reanalyzed as a subject after the first one has been. Likewise, the P600 effect in Adjunct-type condition also reflects syntactic revision/reanalysis, notwithstanding a marginal P600. The second subject in this type of DNS is not in sharp contrast to its counterpart after the Locative particle –ey ‘in’.

Topic Area: Grammar: Syntax