Session Detail

Back

Talk Session 1-1: Cognitive Linguistics

Sep. 1, 2017 13:30 PM - 15:00 PM

Room: Archimedes Room
Session chair: Tai-Li Chou
Entrenchment and Creativity in Chinese Quadrasyllabic Idiomatic Expressions

Presentation Number:111.02Time:13:45 - 14:00Abstract Number:0071
Shu-Kai Hsieh1, Tai-Li Chou1,2, Chia-Lin Lee1, I-Wen Su1, Chia-Rung Lu1, Te-hsin Liu3, I-Ni Tsai3, Benjamin T'sou4
1Graduate Institute of Linguistics, National Taiwan University, Taipei, Taiwan
2Department of Psychology,, National Taiwan University, Taipei, Taiwan
3Graduate Program of Teaching Chinese as a Second Language, National Taiwan University, Taipei, Taiwan
4Department of Linguistics and Translation, City University of Hongkong, Hong Kong, China


As an idiosyncractic and indispensable part of language, idioms/idiomatic expressions have gained increasing attention. Traditionally, idioms are defined as multi-word units for which the semantic interpretation is not a compositional function of their composing units. The idea of schematicity in idiomatically combining expressions have a different status across linguistic theories. Usage-based accounts (incl. construction grammar, pattern grammar, emergentist and alike) have reasonably shown that idiomatic expressions as pairings of form and meaning are entrenched in a speaker’s mind (Croft and Cruse, 2004; Croft, 2012), and the defining properties of them singling them out from other MWEs are their conventionality and metaphoricity. In this paper, we focus on a special type of idiomatic expressions of even length called Quadrasyllabic Idiomatic Expressions (QIEs) in Chinese, and explain their variations with reference to the interaction of construction semantics and exploited ontological knowledge. We select QIEs with symmetric numbers within the construction, which is estimated that they are the most frequent and productive ones, and divide them into idiom-QIE (’chengyu’) and prefabs-QIE. 96 Idiom-QIEs and 96 prefabs-QIEs are selected based on the required language resources tailored for Chinese QIEs with suggested lowest level of frequency. We create corresponding pseudo-idiom/prefabs-QIEs with character replacements of different semantic distances to measure the effect on comprehensibility. Corpus evidence and human ratings suggest that the comprehension and processing of nonce QIEs emerges from the interaction of construction and lexical semantics. The result of behavioral experiment shows that semantic distance affects the speed of comprehension with the construction entrenchment. However, for those QIEs with idiomaticity, semantic distance leads to no major effect. We show that Chinese QIEs provide an ideal testing ground for the empirical investigation of the functional linguistic notion of entrenchment in processing multi-morphemic strings. A Pilot study on fMRI experiment also shows some interesting findings.


 
The Acquisition of the Implicit Syntax of Mass/count Nominals by L2 Learners of Mandarin Chinese

Presentation Number:111.03Time:14:00 - 14:15Abstract Number:0039
Panpan Yao
Linguistics, Queen Mary University of London, London, UK,

Language acquisition is a challenging issue for late second language learners (L2ers), and it is debatable whether the L2ers could have native-like representations and grammars (Jiang, 2011 vs. Foote, 2011). Very rare research investigates L2ers’ real time processing of implicitly acquired knowledge using natural languages. Based on Cheng & Sybesma (1998), the classifier-adjective (Cl-Adj) order is an implicit syntactic cue for the mass/count classifiers in Mandarin. Previous off-line studies (Barner, et al., 2008, 2010, 2012) found that different classifiers can be used to generate mass/count interpretations by native Mandarin speakers. However, it remains unclear whether the Cl-Adj order also contributes to the mass/count distinction, and how native speakers and L2ers make use of this implicit syntactic cue to process NPs in real time. The current study used the Visual World Paradigm (VWP) to look into how native Mandarin speakers and late Dutch-Mandarin learners interpret mass/count NPs with neutral classifiers online, manipulating the Cl-Adj order, and comparing nouns which were judged by native Mandarin speakers as typical count nouns (e.g., spoon) and typical mass nouns (e.g., pebble). The results show that native speakers make rapid use of the Cl-Adj order to generate a ‘massified’ interpretation and that late L2ers have also acquired and make use of this implicit syntactic cue. Even though late Dutch-Mandarin learners could use the Cl-Adj order as a syntactic cue for ‘massified’ interpretations, it took them slightly longer than native speakers. Also, L2ers were not sensitive to the semantic difference between different classifiers, or the connections between classifiers and their associated nouns. In general, the current study found that late L2ers have acquired the syntactic cues for mass/count interpretations through implicit learning. However, L2ers require more time and/or information to make use of these cues and are more sensitive to the lexical meanings.


 
Spatial Language and Cognition Across Adult Lifespan: Taiwan as a Test Case

Presentation Number:111.04Time:14:15 - 14:30Abstract Number:0118
Yen-Ting Lin1, Hui-Chen Hsiao2
1Department of Linguistics, University at Buffalo-SUNY, Buffalo, New York, USA
2Department of Chinese as a Second Language, National Taiwan Normal University, Taipei, Taiwan


This paper presents an investigation of the nexus between ageing and spatial cognition by examining the use of spatial frames of reference in Taiwanese populations. Existing cognitive neuroscience research typically did not test for possible age factors in preferential reference frame use (Klencklen et al., 2012). It has been suggested that allocentric frames may be more difficult to retrieve from spatial memory subject to normal ageing than egocentric frames based on the results found in European languages (Antonova et al., 2009; Lithfous et al., 2013; Ruggiero et al., 2016). The test languages exhibit distinct features in spatial reference in small-scale space: TSM speakers shows an allocentric preference, using cardinal directions or external landmarks, while MC speakers show an egocentric preference, projecting the viewer’s perspective onto the object. The current study extended the research design of Bohnemeyer et al (2015) to explore the age effect in the use of spatial reference across monolingual and bilingual populations. The research methods comprised a discourse study in a referential communication task and a recall memory experiment (Levinson & Schmitt, 1993). The sample was divided into young (18-30 yrs), middle-aged (31-60 yrs) and senior (61yrs and above) groups. Regression analyses indicated that all three age groups were significantly different from one another in the discourse study and that the senior group and monolingual group had an interaction effect in the recall memory study. This finding suggests that the preferential reference frame is population-specific and advocates the need to examine the language experience of a given population. These studies make valuable contributions to the documentation of variation in spatial cognition in a multilingual context and to understanding the factors driving such variation. By investigating changes in preferred strategies across the lifespan, they also lay crucial groundwork for the study of neural networks involved.


 
A Comparison of the Role of Top-down Factors on Local Reading Processes in Braille and Print

Presentation Number:111.05Time:14:30 - 14:45Abstract Number:0040
Ronan G. Reilly1, Inthraporn Aranyanak2, Ralph Radach3, Christian Vorstius3
1Computer Science, Maynooth University, Maynooth, Ireland
2Computer Science, King Mongkut Institute of Technology Ladkrabang, Bangkok, Thailand
3Psychology, Wuppertal University, Wuppertal, Germany


The purpose of this study was to compare the effects of top-down task demands on low-level information processing in sighted and braille reading. The reading task either required a shallow or deep reading of the text and the aim of the study was to quantify the impact of differential task demands on information pick-up in two radically different reading modalities: sight and touch. Previous studies of sighted reading had shown a dynamic interplay between eye movement control and top-down demands (Radach, Huestegge, & Reilly, 2008). It was an open question whether a similar degree of interactivity held for braille. To study braille readers’ behaviour we designed a finger-tracking system utilising affordable components, yet providing high temporal and spatial resolution finger position data (Aranyanak & Reilly, 2012). An SR Research EyeLink 1000 eye tracking system was used to track eye movement in sight readers. The results reveal similar top-down effects on different modalities. Word-viewing times as measured by a range of metrics were significantly shorter for simple verification questions, as opposed to responding to the complex comprehension questions. The more demanding task of answering comprehension questions caused a consistent elevation in the tendency to reread in both tactile and visual reading. Because of the apparent “noisy” nature of hand movements, there is some debate as to the detectabilty of lexical or supra-lexical effects in the finger movements of braille readers (Hughes, Van Gemmert, & Stelmach, 2011). The results described here support the view that braille readers respond dynamically to the demands of the reading task and that these responses are readily detectable by a suitable tracking device. Another of our studies has also shown a similar responsiveness to lexical factors such as word frequency (see Aranyanak & Reilly, 2013).


 
Acquisition of Chinese Classifiers by Preschoolers with Congenital Hearing Impairment: Influences of Linguistic and Cognitive Factors

Presentation Number:111.06Time:14:45 - 15:00Abstract Number:0042
Ming Lo1, Yi-Xiu Lin1
Speech and Hearing Science Research Institute, Children’s Hearing Foundation, Taipei, Taiwan,

This study aimed to examine (1) the acquisition of count-classifiers and mass-classifiers by learners who have limited hearing and verbal comprehension abilities, and (2) whether the acquisition of the two types of classifiers is influenced by the learners’ vocabulary size and working memory capacity. A count-classifier (e.g., 朵 /duo3/ in 一朵花 /yi1 duo3 hua1/, a flower) must be used with a noun, which makes vocabulary of nouns a prerequisite for learning the semantic and syntactic characteristics of a count-classifier. A mass-classifier (e.g., 盒 /he2/ in 一盒花 /yi1 he2 hua1/, a box of flowers) is typically a name of an object and can be used without other nouns. However, it takes extra cognitive effort to recognize that the object name can also be used to denote the quantity of the entity named by a noun. Therefore, it was hypothesized that acquisition of count-classifiers is mainly predicted by a learner’s noun vocabulary while acquisition of mass-classifiers is affected by not only a learner’s vocabulary size but also the learner’s working memory capacity. The participants consisted of 40 hearing-impaired preschoolers. They have joined a rehabilitation program adopting auditory-verbal approach, and the number of count-classifiers, mass-classifiers and ordinary nouns that each participant has acquired was calculated. Moreover, the participants were equally divided into two groups according to their working memory spans (high-span vs. low-span). Regression analyses showed that the size of noun vocabulary accounted for acquisition of count-classifiers in both groups of participants (high-span: p < .05; low-span: p < .05). However, the size of noun vocabulary accounted for acquisition of mass-classifiers in the high-span group (p < .01) but not in the low-span group (p > .10). The results support the hypothesis and its implications for language acquisition of children with hearing impairment will be discussed at the conference.


 
Speakers' Trade-off Based on Communicative Efficiency

Presentation Number:111.07Time:Abstract Number:0181
Dongsu Lee1, Hongoak Yun2, Daun Kim1, Upyong Hong1
1Konkuk University, Seoul, Korea
2Gachon University, Incheon, Korea


Kurumada & Jaeger (2015) claimed that speakers hold comprehension consideration in the expense of their production cost. Using Japanese transitive sentences in a spoken-recall paradigm, they demonstrated that Japanese speakers were more likely to attach case markers to animate objects than to inanimate objects (Exp’t 1), in order to make it clear that the second nouns are patients rather than agents. However, we failed to replicate the animacy effect in a written-recall paradigm with Korean. Instead, we found that speakers’ production preferences were remarkably influenced by the distributional patterns associated with case markers and word orders. First, we conducted the written-recall production experiment in which word order, the animacy of direct objects, and the ellipsis of object case markers. Second, using the Korean spoken corpus, we computed the degree of surprisal to encounter sentences corresponding to our manipulation. Our key findings were as follows. 1) Overall, Korean speakers strongly preferred to produce case markers and to rearrange non-canonical OSV to canonical SOV word orders. 2) Korean speakers did not show particular preference due to animacy. 3) We observed that surprisal was significantly proportional to speakers’ behaviors for the use of object case markers and the use of SOV order. Of interest, the degree of surprisal based on OSV sentences strongly predicted speakers’ bias to attach case markers increasingly, whereas it predicted speakers’ willingness not to change word orders. On the other hand, the degree of surprisal based on sentences with null case markers strongly predicted speakers’ bias to change word orders from OSVs to SOVs, instead of attaching case markers. Our findings showed that speakers diligently exploited the distributional information, attempting to achieve a trade-off between production ease and comprehension goal.