1. Experiment & Results (± honorific features vs. main/embedded subject with Emb.Verb-honorific)  Experimental conditions  Self-paced reading time Participants: 37 native Korean speakers Materials: 40 sets of experimental sentences Procedures: non-cumulative moving-window  Eye-tracking Participants: 44 native Korean speakers Materials: 40 sets of experimental sentences Procedures: calibrated for every stimuli Eyelink 1000 Plus  ERPs Participants: 26 native Korean speakers Materials: 120 sets of experimental sentences Procedures: word-by-word, non-cumulative presentations (SOA: 600 ms; ISI: 200 ms) Neuroscan Synamps self-paced RTs At the critical verb (W5) At W6 (spill-over) : go-past durations : first pass RTs Interaction t = 2.13 emb subject t = -2.36 main subj t = -2.53 At the critical verb (W5) At W6 (spill-over): go-past RTs Interaction p <.05 Interaction t = 2.08 Facilitatory intrusion effects in subject-verb honorific agreement in Korean Nayoung Kwon 1 & Patrick Sturt 2 1 Konkuk University, 2 University of Edinburgh nayoung.kw@gmail.com Background & Research questions Main subjEmb subjW1W2W3W4W5W6W7W8 HHChair-nompresident-nomon.time-atmeeting-accstart-si-compmeeting.roomdoor-accclosed NHHJinswu-nompresident-nomon.time-atmeeting-accstart-si-compmeeting.roomdoor-accclosed HNHChair-nomInho-nomon.time-atmeeting-accstart-si-compmeeting.roomdoor-accclosed NH Jinswu-nomInho-nomon.time-atmeeting-accstart-si-compmeeting.roomdoor-accclosed ‘The chair/Jinswu closed the front door of the conference room so that the president/Inho could start the meeting on time.’ This research was supported by National Research Foundation of Korea (NRF-2014S1A2A2028232).  Memory retrieval is content addressable (Lewis & Vasishth, 2005; Lewis et al., 2006; McElree et al., 2003; Van Dyke & McElree, 2006)  Potential targets in memory are activated in parallel in response to retrieval cues.  Facilitatory intrusion : Reading time penalty for a mismatching dependency could be reduced due to the presence of a partially matching distractor (Wagers et al. 2009; Vasishth et al. 2008; Xiang et al. 2009) (a) The musician who the reviewer praise won the prize. (b) The musicians who the reviewer praise won the prize. reading times at praise: (b) < (shorter) than (a)  Similarity based interference Processing difficulty that occurs when the intended dependency target completely matches the retrieval cues, but where there is also a partial match with the distractor (Badecker & Straub, 2002; cf. Chow et al., 2014; Dillon et al., 2013) (a) John thought that Bill owed him another chance. (b) John thought that Beth owed him another chance. reading times at him: (a) > (longer) than (b)  Korean  SOV word order with case marking & impoverished verbal agreement except for subject honorific agreement  Subject honorific suffix –si– is optional and can be omitted (a) but  when used, should agree with the subject in honorific feature (b)  cannot be used with a subject of low social status (c) a)Grandpa-nomTV-accwatch-decl(optional) b)Grandpa-nomTV-accwatch-si-decl c)*Kid-nomTV-accwatch-si-decl  Subject honorific violation in Korean elicits a P600 (Kwon & Sturt, 2015).  Goal of study: To investigate retrieval processes in Korean using subject-verbal honorific agreement Discussion & Conclusions  These results suggest i) the subject-verb honorific agreement in Korean is susceptible to facilitatory intrusion effects by structurally illicit but feature- matching attractors ii) the intrusion effect is strong such that it can be found even when the main subject did not intervene in the embedded subject-verb dependency either linearly or structurally iii) the effect is not limited to ungrammatical sentences but facilitates the processing of already grammatical sentences.  Overall, these results suggest that, despite apparently different verbal morphology, similar retrieval mechanisms (e.g., the content- addressable-retrieval: McElree, 2000; Lewis & Vasishth, 2005) underlie the processing of subject-verb agreement across languages.