You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

 
Poster B18, Wednesday, November 8, 3:00 – 4:15 pm, Harborview and Loch Raven Ballrooms

VOS Preference in Truku Sentence Processing: Evidence from Event-Related Potentials

Masataka Yano1,2, Keiyu Niikuni1, Hajime Ono3, Sachiko Kiyama1, Manami Sato4, Apay, Ai-yu Tang5, Daichi Yasunaga6, Masatoshi Koizumi1;1Tohoku University, 2Japan Society for the Promotion of Science, 3Tsuda University, 4Okinawa International University, 5National Dong Hwa University, 6Kanazawa University

Introduction: In many languages with flexible word order, transitive sentences in which the subject precedes the object have been reported to incur a less processing cost during sentence comprehension compared with those in which the object precedes the subject. This observation brings up the question of why this subject-before-object (SO) order should be preferred in sentence comprehension, together with the related empirical question of whether this preference is universal across all human languages. Methods: We conducted an event-related potential (ERP) experiment to address these two issues by examining the word order preference in Truku, a Formosan Austronesian language spoken in Taiwan. The syntactically basic word order of Truku is verb-object-subject (VOS) as shown in (1a) and (2a), and SVO is derived by preposing S to the sentence-initial position as shown in (1b) and (2b). Truku has a symmetrical voice system (also referred to as Focus System). In Actor Voice (AV), S refers to an ACTOR, as in (1), while in Goal Voice (GV), S is a GOAL or a PATIENT, as in (2). (1) a. AV-VOS: qmqah emqliyang niyi ka embanah niyi. kick.AV blue DET NOM red DET ‘The red kicks the blue.’ b. AV-SVO: embanah niyi o qmqah emqliyang niyi. red DET FOC kick.AV blue DET (2) a. GV-VOS: qqahan embanah niyi ka emqliyang niyi. kick.GV red DET NOM blue DET b. GV-SVO: emqliyang niyi o qqahan embanah niyi. blue DET FOC kick.GV red DET In the ERP experiment, 25 native speakers of Truku were auditorily presented these four types of sentences (18 females, M = 61.6, SD = 12.6). The participants were asked to judge whether the event described in a sentence matched a picture presented after the sentence. Results: The result showed that SVO elicited a larger positivity compared to VOS at the third region (i.e., SVO versus VOS), irrespective of VOICE. This result indicates that the syntactically basic word order, VOS has a processing advantage over SVO in Truku. Discussion: This result aligns well with the theory that accounts word order preference in terms of grammatical factors, such as syntactic complexities (e.g., Dependency Locality Theory, Gibson, 1998). Since SVO has a more complex syntactic structure in Truku, this theory expects SVO to be more difficult to process than VOS. The larger positivity for SVO reflects an increased processing cost to integrate a fronted S with its original position. On the other hand, our result is not consistent with the theory that accounts word order preference by appealing to the relative order of ACTOR and PATIENT. If SO preference that has been reported in previous studies reflects a preference for the ACTOR-before-PATIENT order, we expect that SVO should be easier to process than VOS in Actor Voice. This prediction was not borne out by our result. Conclusion: The SO preference in sentence comprehension reported in previous studies may not reflect a universal aspect of human sentence comprehension. Rather, processing preference may be language-specific to some extent, driven by syntactic differences in individual languages.

Topic Area: Grammar: Syntax

Back to Poster Schedule