SWELL-2020 (22nd-23rd December, 2020)

SWELL (SNU Workshop on Empirical and Laboratory Linguistics) held at SNU


The linguistics department of SNU successfully held the first SWELL (SNU Workshop on Empirical and Laboratory Linguistics) at SNU on 22nd and 23rd of December, 2020.

The following is the schedule for the workshop:

The presenters of the following talks have gladly consented to share the content of their talks in the form of slides or abstracts.


Day 1 (December 22nd):


  1. Psychopy Tutorial by Gyuhwan Lee (SNU)

Link to the material


  1. Eye-tracking Workshop 1 by Jeong-Ah Shin (Dongguk University)

The first workshop on Eye-tracksng methodology presented by Jeong-Ah Shin from Dongguk university was entitled “Eye-tracking in Language Research”

Link to the material


  1. Past Tense Morphology and the Choice of Connectives in Korean Counterfactual Conditionals: An Experimental Approach by Jiyeong Kim (SNU)

Abstract:

Counterfactual conditionals convey the speaker’s belief that the proposition in the antecedent is contrary to fact. As Iatridou (2000) claimed, past tense morphology is the hallmark of subjunctive conditionals cross-linguistically, which is also the case for Korean (Han,2006). Besides, Korean has two types of conditional connectives, -myen and –tamyen. The connective –tamyen has been related with the speaker’s hypothetical or irrealis attitude while –myen is a conditional connective in any type of attitude. However, some debates have arisen regarding the hypothetical property of -tamyen since it is freely interchangeable with –myen and it can also be used in a situation where the antecedent clause is quoting another’s utterance. This paper examined how past tense morphology and two different types of connectives in Korean conditionals contribute to counterfactual interpretation using an experimental method.

Forty Korean native speakers participated in a truth judgement rating (7-point Likert scale). A conditional sentence, differing in two levels (Past vs. Non-past, -myen vs. –tamyen in antecedent) was first presented to the participants as a context. The participants then judged the following sentence indicating realization or non-realization of consequence in the prior conditional. If the antecedent of a conditional sentence is interpreted as having a strong counterfactual meaning, realization of a consequent will be more likely to be judged as false l while the unrealized condition will be true.

The results show that counterfactual meaning in Korean conditionals is affected not only by the use of past tense morphology but also by the choice of connectives. Significantly lower truth rating was observed in the realized consequent condition when past tense was marked in the antecedent. Furthermore, counterfactual interpretation was reinforced with –tamyen when past tense appeared.


  1. Modal Semantics of Korean Bouletic Verbs by Mingyeong Choi (SNU)

Link to the material


  1. Island effects and scrambling: an on-line self-paced reading study by Heejeong Ko (SNU) & Kitaek Kim (SNU)

Link to the material


  1. The Processing of Korean Relative Clauses by Soyeon Kang (SNU), Gyuhwan Lee (SNU) & Sunwoo Jeong (SNU)

Link to the material


Day 2 (December 23rd):


  1. Exploring the relation between alternation and phonotactics of Sundanese liquid assimilation and dissimilation by Yoona Yee

Link to the material


  1. Glottal stop and high tone in Korean by Hayeun Jang (SNU)

Link to the material


  1. Eye-tracking Workshop 2 by Jeong-Ah Shin (Dongguk University)

The second workshop on Eye-tracking methodology presented by Jeong-Ah Shin from Dongguk university was entitled “Shortcuts to Eye-tracking Experiments”

Link to the material


  1. The prosodic effects on the fronting of alveolar consonants in Korean by Mira Oh (Chonnam University) & Soohyun Kwon (SNU)

Link to the material


  1. Voice quality research using the Electroglottography (EGG) by Seunghun J. Lee (ICU / University of Venda)

Link to the material


  1. Combining acoustic and articulatory findings in phonological analysis by James Whang (SNU)

Abstract:

This talk proposes an ultrasound experiment that is designed to collect high-resolution articulatory and acoustic data simultaneously, starting with Japanese high vowel devoicing as a test case. There is an ongoing debate on the phonetic/phonological status of devoiced vowels in Japanese, and a survey of previous studies suggests that the disagreement stems in large part from methodological differences; studies that argue that Japanese devoiced vowels are deleted rely primarily on acoustic data while studies that argue against deletion rely on articulatory data. Whang, Shaw, and Kawahara (2020) analyzed the acoustic recordings accompanying the EMA study by Shaw and Kawahara (2018b) and found that although Japanese speakers sometimes fully retain the lingual gesture for devoiced vowels, no trace of the vowel is detectable in the acoustic signal after coronal consonants. In other words, Japanese high vowels are articulatory devoiced but acoustically deleted in certain contexts. Such results reveal a disconnect between articulatory and acoustic findings and a need for experiments that analyze both data types simultaneously. This talk proposes an ultrasound experiment due (i) to its low level of invasiveness that allows for more naturalistic speech and (ii) to the relative silence of the machinery that allows for clean, time-aligned acoustic recordings. By combining articulatory and acoustic methods, the experiment will give a more holistic view of Japanese high vowels, but more importantly, extending this approach to other phonological phenomena is expected to provide deeper insight into the link between production (articulatory) and perception (acoustic) grammars.


  1. How tight is the link between alternations and phonotactics? by Jongho Jun (SNU), Hanyoung Byun (SNU), Seon Park (SNU) & Yoona Yee (SNU)

Abstract:

It has been argued that some phonological alternations apply to abide by phonotactic restrictions, whereas alternation processes with morphologically derived environment effects (MDEE) exhibit a mismatch with phonotactics. However, as phonotactic well-formedness can be gradient rather than categorical, there might be cases in which lexically attested sequences which undergo an alternation across morphemes are not entirely phonotactically well-formed.

To explore the link between alternations and phonotactics in detail, we focus on three processes in Korean which may apply only in derived environments: palatalization, vowel harmony and compound tensing. In order to find out the extent to which these processes are supported by their corresponding phonotactics, we conduct a human behavioral experiment by collecting Korean speakers’ non-word acceptability judgements. Furthermore, we conduct a computational modelling of Korean speakers’ phonotactic grammar to examine whether the non-word acceptability judgements can be projected from the Korean lexicon.

Our primary results show that not all alternation processes match well with the relevant phonotactics. However, acceptability ratings by Korean speakers show a high correlation with the predicted ratings from the learned grammar, if the lexicon employed consists of both native and Sino-Korean words. This suggests that Korean speakers’ phonotactic knowledge is indeed projected from the lexicon.