Poster C18, Friday, August 17, 10:30 am – 12:15 pm, Room 2000AB

Subject and Object Asymmetry in Korean Scrambling: An ERP Study

Oh Eunjeong1, Wonil Chung2, Myung-Kwan Park2, Sanghoun Song3, Euhee Kim4;1Sangmyung University, 2Dongguk University, 3Incheon National University, 4Shinhan University

This study examines the effects of scrambling either a subject or object associated with (‘floating’) numeral classifiers ((F)NCs) in Korean by using the event-related potentials (ERP) paradigm. The experimental materials consisted of 360 sets of 12 items, which vary in terms of three factors such as (i) the grammatical role S(ubject) vs. O(bject) that (F)NCs associate with, (ii) the type of Case/F(ocus) particle marker on FNCs (Case-less vs. N(om)/A(cc) Case-marked vs. F-particle-marked), and (iii) the (non-)application of subject or object scrambling, with FNCs schematically represented in English below. i) [park-in dog-Nom bread-Acc 2-FNC ate] I heard. ii) [park-in dog-Nom bread-Acc 2-FNC-Nom ate] I heard. iii) [park-in dog-Nom bread-Acc 2-FNC-Foc ate] I heard. iv) [park-in bread-Acc dog-Nom 3-FNC ate] I heard. v) [park-in bread-Acc dog-Nom 3-FNC-Acc ate] I heard. vi) [park-in bread-Acc dog-Nom 3-FNC-Foc ate] I heard. Using the materials, we investigated the following two questions. First, is there a difference between 'scrambling-less adjacent' and 'post-scrambling stranded' versions of NCs? Second, is there a difference between Case-less, Case-marked, and F-particle-marked FNCs in light of the effects of 'post-scrambling stranded', compared to 'scrambling-less adjacent', versions of NCs? Twenty Korean speakers (male: 12, mean age: 22; range:19-28) participated in the ERP experiment. The ANOVA results on the descriptive data of the offline acceptability task showed that in the Subject/Object condition there are significant main effects of such factors as displacement (F(1,25)=14.13,p<0.001); (F(2,50)=11.93,p<0.001) and Case/F-particle (F(2,50)=11.67,p<0.001); (F(1,25)=11.93,p<0.001). ERPs were measured at the NC in each condition. In overall ANOVA, in the subject condition there are significant effects of displacement at the 50-200 ms and 500-630 ms interval, while in the object condition there are marginally significant effects of Case/F-particle at the 50-200 ms and 300-450 ms interval. We found that, first, the Caseless FNCs in S/O scrambling sentences relative to those in scrambling-less ones elicited P600, besides N100 and LAN only in the subject condition. The Case-marked FNCs registered P600 in the subject, but not in the object condition. By contrast, the F-particle-marked FNCs in scrambling sentences elicited N100 and P600 in the subject condition, and marginal P600 for the object condition. First, our ERP results point to a robust subject vs. object asymmetry (Saito 1985). As this asymmetry gave rise to P600 effects, it is taken to be a reflection of a syntactic anomaly. Second, our ERP results also point to a Nom Case-less FNC vs. Nom Case-marked FNC asymmetry, only in the subject condition. Nonetheless, FNCs in all the three levels of subject condition induced P600 effects, relative to 'scrambling-less adjacent' NCs. By contrast, FNCs in all the three levels of object condition evoked no significant P600 effects, relative to 'scrambling-less adjacent' NCs. Third, there were N100-like effects in the Nom Case-less or F-particle-marked subject condition. These effects are also known to overlap with the mismatch negativity (MMN; Nataanen, Gaillard, & Mantysalo, 1978) effects. These effects apparently arose when the subject-related FNC is not matched with the subject NP in terms of Case.

Topic Area: Grammar: Syntax