Presentation 2017-07-22
A model of semantic linguistic composition under conditions of morphosyntactic underspecification: the case of eventive iteration construal
Maria Mercedes Pinango, Yao-Ying Lai,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) It is observed that in English the composition of a semelfactive predicate, jump, and a for-adverbial, such as 1_"Frances jumped for an hour," engenders an iterative reading (Frances jumped repeatedly for an hour). Previous studies report that this iteration generation engenders processing cost (vs. 2_"Frances ran for an hour"). The cost is captured by the Partition Measure (PM) hypothesis, which argues that for-adverbials contain a universal quantifier that quantifies over the subintervals of the interval that they denote (e.g., hour). A non-infinitesimal partition measure interpretation of the subintervals yields an iterative reading, construed out of the local lexical meaning of the verbal denotation (e.g. jump). An infinitesimal partition measure yielding a continuous reading (one jumping event), implausible in (1), is nonetheless always available and cost-less. Crucially, this account predicts that an iterative reading emerging from the composition of a for-adverbial with a "durative" predicate, as in 3_"Frances ran for 10 years", also engenders cost from the necessary PM construal, this time from the entire sentential context. For-adverbial composition thus demands a partition measure obtained from structured individuals--entities with structured subparts. The relation between the sub-parts, i.e., non-infinitesimal vs. infinitesimal, yields the iterative vs. continuous readings observed. A Self-Paced Reading_Study.1 and an fMRI_Study.2 from our lab show that, as predicted, both iterative conditions--punctual_1 & durative_3--are costly vis-a-vis no-iteration_2. fMRI results also show brain activation consistent with the source of partition measure: lexical vs. sentential, and convergent deactivation, consistent with the partition-measure search. The observed PM retrieval process evidences a model of real-time sentential semantic composition that demands (a) rich lexical meanings whose full realization depends on key contextual information; (b) a search process that uses lexical vs. sentential locality as a processing complexity metric, and (c) a conceptual structure system that ultimately resolves the underspecified lexical variables.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Semantic composition / Underspecification / for-adverbials / event construal / partition measure / self-paced reading / fMRI / lexical meaning and context
Paper # TL2017-21
Date of Issue 2017-07-15 (TL)

Conference Information
Committee TL
Conference Date 2017/7/22(2days)
Place (in Japanese) (See Japanese page)
Place (in English) NINJAL
Topics (in Japanese) (See Japanese page)
Topics (in English) Human Language Processing and Learning
Chair Masami Suzuki(KDDI Research)
Vice Chair Tadahisa Kondo(Kogakuin Univ.) / Chiaki Kubomura(Yamano College of Aesthetics)
Secretary Tadahisa Kondo(Kobe Gakuin Univ.) / Chiaki Kubomura(Kyoto Univ.)
Assistant Nobuyuki Jincho(Waseda Univ.) / Noriaki Takada(Ferris Univ.)

Paper Information
Registration To Technical Committee on Thought and Language
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A model of semantic linguistic composition under conditions of morphosyntactic underspecification: the case of eventive iteration construal
Sub Title (in English)
Keyword(1) Semantic composition
Keyword(2) Underspecification
Keyword(3) for-adverbials
Keyword(4) event construal
Keyword(5) partition measure
Keyword(6) self-paced reading
Keyword(7) fMRI
Keyword(8) lexical meaning and context
1st Author's Name Maria Mercedes Pinango
1st Author's Affiliation Yale University(Yale Univ.)
2nd Author's Name Yao-Ying Lai
2nd Author's Affiliation Yale University(Yale Univ.)
Date 2017-07-22
Paper # TL2017-21
Volume (vol) vol.117
Number (no) TL-149
Page pp.pp.43-43(TL),
#Pages 1
Date of Issue 2017-07-15 (TL)