This paper examines the phonotactic grammar of Korean male and female first names using a maximum-entropy phonotactics learning model (Hayes and Wilson 2008). An analysis of the 7000 most frequent recent Korean names (registered births in the period of 2008-2020, 3500 for each gender) reveals that the Korean names show gender-dependent phonological characteristics, as observed in other languages. Female names are longer, have more sonorants in onset position, and are more likely to end with a vowel than male names. To further compare female and male phonotactic characteristics, maximum-entropy phonotactic grammars were learned from the training data consisting of 3377 names for each gender. The resulting constraints are categorized into word-edge, CV, and VC constraints, and the constraints that are found to differ in male and female names are compared. The results identify the sound sequences that are biased toward either gender. This paper also shows that the predictions made by the maximum-entropy grammar have significant correlations with human ratings.

이 연구는 한국어 학습자들이 의미에 따라 분화되는 종결어미의 억양을 어떻게 판별하고 산출하는지를 밝히는 논문이다. 연구에 참여한 학습자들은 한국어의 억양이 의사소통에 영향을 미친다는 것을 인지하고 있었다. 그러나 한국어 억양을 익힌 방법이 그동안 들었던 수업과는 별 관련성이 없는 것으로 응답하였고 억양 교육의 만족도가 매우 낮았다. 판별 실험에서 오류를 많이 보인 종결어미는 ‘감탄(놀람)’의 ‘–다니’였으며 산출 실험의 경우는 ‘빈정거림’의 ‘-고’였다. 또한 과거의 억양 학습 유무는 억양 판별과는 크게 상관이 없었지만 산출 능력과 유의미한 관계가 있었다. 또 한국어 학습 기간이 길수록 억양 판별을 좀 더 잘했으나 산출 능력과는 관계가 없었다.

The word-initial voiced stops in English are adapted into the two distinctive laryngeal categories of lenis and fortis in Korean (Kang 2008, Oh 2009). The pattern is not easily predicted by phonological contexts, but a significant diachronic change has been reported. The proportion of fortis adaptation has decreased over time (Kang 2008). However, whether adaptation as fortis is still an active process has not been seriously investigated in previous studies. It has been assumed without question that the initial voiced segment in a novel English word can still be adapted new as fortis, and previous studies have treated fortis adaptation as analogous to the native phonological process of word-initial tensification (Oh 2009, Kim 2017). This study challenges this assumption and clarifies whether the fortis adaptation is still active among recent loans. To this end, it diachronically compares the adaptation of each of the 204 most frequent English loans. The results show that the current bifurcation is a mix of older and newer adaptative patterns, where fortis adaptation only exists in the older pattern. In other words, voiced stops in new loanwords are no longer adapted as fortis, and fortis adaptation in the contemporary data is a remnant of the historical pattern. If this analysis is on the right track, the focus should shift towards identifying the factors that facilitate retaining the older fortis forms. This study suggests that the effects of native phonology prohibit them from changing into lenis.

This paper investigates the atypical characteristics of blending found in Korean fandom names. For blends as fandom names, the proper name and the existing word participate as sources, and if clipping occurs, a larger part of the proper name tends to be cut off compared to the existing word. As a result, many blending outputs are phonologically very similar or identical to existing words. In particular, blends with the same phonological form as existing words occupy the largest number in the fandom name dataset. This type of blend was evaluated as absolutely ill-formed, violating the Uniqueness constraint of blending. This paper interprets the creative and playful nature of the blending process as having been reinforced in fandom names to make the newly created name attractive and easy to remember, and evocative of the target proper name. This shows that when analyzing lexical blends, the social and pragmatic purposes of word-formation must be taken into account in addition to phonological and morphological constraints.

이 연구는 뉴스 이용자들이 미디어 언어를 제대로 이해할 수 있는 방법을 모색한 논문이다. 보도기사도 언어로 이루어진 텍스트이기 때문에 언어적 기준으로 미디어 언어를 평가하는 방법을 제시하였다. 그 기준은 정확성, 균형성, 명시성, 완결성이다. 정확성은 어휘, 문장구조, 언어 표현 등으로 평가한다. 균형성은 기사에 등장하는 대상에 대한 균형적 서술, 논리적 균형성 등으로 분석한다. 명시성은 취재원을 인용할 경우 그 실체를 확인할 수 있는 언어 표현으로 파악한다. 완결성은 사건에 대한 정보를 육하원칙으로 충분히 제공하고 있는가를 중심으로 판단한다. 이러한 기준으로 보도기사를 평가하면, 균형성이 부족한 편향보도, 명시성이 부족한 허위보도, 완결성이 부족한 왜곡보도로 나누어진다.

The study deals with the pitch patterns of the first two syllables in phonological phrases that appear in spontaneous speech in Korean. In this study, spontaneous speech data from 7 female Seoul speakers were analyzed. The results show that phonological phrases starting with both L(L-PP) and H(H-PP) tones have a positive pitch slope line. The difference in the slope value was not a valid index to distinguish between L-PP and H-PP. The most stable indicator was the pitch of the first two syllables, especially the first syllable. The pitch showed a clear difference between L-PP and H-PP. In most cases, the pitch of the first two syllables in L-PP was lower than that of H-PP. In general, if the first syllable is realized at a point lower than 50% of each speaker’s pitch range, it can be regarded as L-PP. Conversely, if the first syllable is realized at a point exceeding 50% of the speaker’s pitch range, it can be regarded as being realized as H-PP. These findings should be considered in further studies of the pitch patterns of phonological phrases.

본 연구는 자발적 독서 태도와 관계된 변인으로서 가정 및 학교 환경과 스마트폰 사용 변인의 설명력을 밝히고, 각 변인의 상대적 영향력을 탐색하는 데에 목적이 있다. 그간 독서 태도는 가정환경 및 학교 환경 변인에 의해 설명된다는 사실이 보고되어 왔다. 최근 스마트폰 사용과 같은 매체 사용과 독서 태도의 관계에 대한 관심이 고조되고 있는 가운데, 본고는 가정환경, 학교 환경, 가정 내 스마트폰 사용량, 독자 본인의 스마트폰 사용량, 스마트폰 취득 시기와 자발적 독서 태도와의 관계를 살펴보고자 하였다. 이를 위하여 중학교 3학년 학생 902명을 표집하여 설문조사를 실시하고, 단계적 다중회귀분석 방법을 활용하여 분석하였다. 연구 결과 학교 환경 변인, 가정환경 변인, 스마트폰 사용량, 스마트폰 취득 시기 변인이 독서 태도를 설명하는 유의한 변인으로 나타났고, 가정 내에서의 스마트폰 사용 정도는 유의하지 않았다.

이 글은 19세기 전후의 ‘쯤’의 여러 측면을 살피자는 데 목적이 있다. 다시 말해 ‘쯤’의 형태와 의미, 문법 범주 등에 대해서 알아보고, 또 ‘쯤’이 어느 시기부터 문법소로 쓰이기 시작했는지를 알아보자는 것이다. 이 시기에는 ‘즘, ㅅ즘, ᄶᅳᆷ, 쯤’ 등 여러 형태가 등장한다. ‘즘’은 18세기 말에 등장하여 20세기 초까지 생산적인 양상을 띤다. 19세기 말까지는 그 분포와 기능에서 의존명사로서의 지위를 가진다. 20세기 초에는 ‘ㅅ즘’이 등장하는데, ‘ᄶᅳᆷ’보다 뒤에 등장한다는 특징이 있다. ‘ᄶᅳᆷ’은 19세기 말에서 20세기 초까지 생산적인 양상을 보이는데, 분포와 의미에서 문법소로서의 양상을 띤다. ‘쯤’이 문법소로서의 경향을 보이기 시작한 것이 이 ‘ᄶᅳᆷ’의 단계부터라 생각된다. ‘쯤’은 20세기 초에 집중 등장하는데, 주로 신소설 작품에서 볼 수 있다. 그리고 이 시기의 ‘쯤’은 분포와 의미 등을 고려할 때, 그 범주는 접미사로 파악된다.

This study aimed to investigate the realization of Korean lateralization in /l-n/ sequences. In previous studies, progressive lateralization has been considered a compulsory phonological phenomenon in Korean, and has not attracted much attention. However, in actual speakers’ speech, non-assimilation patterns are observed in /l-n/ sequences. A reading task was conducted with 20 Korean speakers in their 20s in order to acquire the realization of /l-n/ sequences. The realization of progressive lateralization was shown to differ depending on individual speakers and experimental words/phrases. The greater the morphological boundary between /l/ and /n/, the less likely was lateralization. In addition, in principle, lateralization is realized even if a phonological phrase is located internal to an /l-n/ sequence, but more non-assimilation patterns have been observed in tokens where phonological boundaries occurred. Gender and speech rates also influence the realization of lateralization. Male speakers lateralize /l-n/ sequences more stably than female speakers. Moreover, the rate of lateralization increased when the speech was faster.

이 글은 조사 ‘도’가 결합한 양보 관계 접속어미의 특성을 밝혀 한국어교육에서의 문법 설명력을 기술하고자 한 것이다. 한국어교육에서 양보 관계를 드러내는 문법형태로 기술되어 있는 ‘-아도’, ‘-고도’, ‘-는데도’, ‘-으면서도’는 ‘접속어미+도’의 구성을 보인다는 것이 특징이다. 그리고 이들 양보 관계 접속어미는 대개 중급단계의 교수요목인데 조사 ‘도’가 결합하지 않은 ‘-아’, ‘-고’, ‘-는데’, ‘-으면서’는 초급단계의 교수요목이라는 것도 특징이다. 이 연구에서는 이러한 특성에 주목하여 한국어교육에서 조사 ‘도’ 결합형 양보 관계 접속어미를 효과적으로 교육할 수 있는 이론적 특성에 대해 고찰하고자 한다. 이 연구는 한국어교육에서 양보 관계 접속어미 교육 방안 마련의 이론적 토대가 될 수 있다는 데 그 의의가 있다.

This article aims to overcome the limitations of classical analyses of compensatory lengthening in Middle English and offer a viable alternative solution, where deletion of the coda consonant and subsequent vowel lengthening constitute separate steps in derivation. The basis of our analysis is that Hayes (1989)’ analysis fails to capture the universal nature of the phenomenon. Also, it is noted that the principle of unmarked one-to-one mapping between tiers is violated in his theoretical framework. I make a critical review of Torres-Tamarit’s (2016) account for compensatory lengthening in Harmonic Serialism. That mechanism turns out not to be satisfying in that its first step of the derivation starts with a structure that should not be allowed to undergo vowel lengthening at all, and this point had already been called into attention by Hayes. Besides, Torres-Tamarit’s analysis can be further reduced by the elimination of debuccalisation. Staying under the framework of Harmonic Serialism, I provide a new universal constraint responsible for assigning different indexes to segments when they belong to different syllables. This markedness constraint of universal nature will be instrumental in our analysis of compensatory lengthening throughout this article. We bring to light more examples from Latin illuminating the workings of this new constraint on compensatory lengthening.

The aim of the current paper is to investigate how Case is assigned to non-canonical nominal expressions which are distinct from Case-marking on canonical subject and object arguments. More specifically, we focus on Case marking to nominalized verbs in Korean Long-Form negation constructions. As other nominal expressions get Case in Korean, the nominalized verbs with the -ci marker in Long-Form negation constructions can also get either NOM or ACC. However, the distributions of Case marking in this paper show that Case markers attached to the nominalized verbs are not randomly assigned, but it is systematically given depending on syntactic and semantic properties of the nominalized verbs. This paper proposes two distinct conditions as follows: the negated auxiliary verb anh (i) assigns only ACC Case to its nominalized verb complement or (ii) assigns either NOM or ACC Case allowing free Case alternation. To solve the puzzle of distinct Case assignment, this paper argues that agentivity is the important factor in deciding Case marking on the nominalized verb. Evidence from the data further proves that if nominalized verbs have +Agent feature, then the feature can percolate up to the Case assigner anh, and it eventually forces to assign only ACC Case back onto the nominalized verb. If not, the auxiliary verb anh cannot have +Agent feature and assigns either NOM or ACC.

This paper takes a corpus-based approach and examines the linguistic properties of two Korean nominalizers -(u)m and -ki. From the Sejong Treebank corpus, all the sentences with -(u)m and -ki are extracted. Twenty linguistic factors are manually encoded into the extracted sentences. Then, all the encoded data are statistically analyzed with (binary) logistic regression. Although we take a monofactorial analysis, we obtain a good statistical model whose C value is 0.956. Through the analysis, the followings are observed: (i) -(u)m and -ki are used with the ratio of 1:9 in Korean, (ii) among twenty linguistic factors, only ten factors are statistically significant, and (iii) not only the verbs which take -(u)m and -ki as a complement but also the verbs which merge with these two nominalizers also play important roles in the determination of nominalizers.

In addition to the widely known Free Choice Items (FCIs), there exists another type of anti-specific item, known as existential indeterminate or Referentially Vague Items (RVIs) such as some-X-or-other in English. Such polarity items are characterized to be rather semantically non-emphatic, and their non-emphaticness is modeled as referential vagueness as a speaker-based felicity condition of minimal, non-exhaustive variation. Both FCIs and RVIs are anti-specificity phenomena, relying on a speaker’s epistemic judgment, but distinct in that, whereas FCIs require exhaustive variation, RVIs require partial, non-exhaustive variation. However, the landscape of specificity in signed languages has been less well-defined. In this background, the main goal of current study is twofold: first, by identifying an RVI sign in ASL, equivalent to RVIs in English, I support the necessity of the notion of referential vagueness to correctly capture the meaning and distribution of non-emphatic, non-exhaustive NPIs occurring in nonveridical contexts; and second, I show how both English and ASL exhibit a remarkable case of semantic-o-pragmatic extension from anti-specificity (as RVIs) to anti-veridicality (as metalinguistic negation).

This paper proposes a weighted constraint grammar analysis that best models English syllabification as reported in Eddington et al. (2013). After we examine the factors governing the syllabification of medial consonants in two-syllable words, a set of constraints are formulated to address these factors. Based on these constraints, an initial grammar is constructed, to which we submit distribution information about medial syllabifications obtained from Eddington et al.’s (2013) syllabification survey data. We then perform learning simulations using Exponential Noisy Harmonic Grammar, a weighted constraint grammar designed especially to deal with constraints with negative weight (Pater 2009). As a result of the learning, an output grammar is produced in which each constraint is assigned a numerical weight. From the output grammar, we generate predicted syllabifications, which are then compared with observed syllabifications to evaluate the success of the analysis. We measure the predictive performance of the proposed analysis in terms of Root Mean Square Error and R², both of which provide a positive assessment.

Ditransitive constructions, expressing an agent, a recipient and a theme, are syntactically realized in various structures within and across languages (Malchukov et al. 2010). Korean is known to have two ditransitive constructions that vary in case marking on the recipient argument, i.e., the canonical ditransitive construction Jisu-kaNOM Mina-eykeyDAC chayk-ulACC cwu-essPAST-taDECL and the double accusative construction Jisu-kaNOM Mina-lulACC chayk-ulACC cwu-essPAST-taDECL ‘Jisu gave Mina a book.’ They are often used in the study of crosslinguistic comparison as counterparts of the English dative alternation. Previous research, however, reported huge imbalance between the two Korean variants in their frequencies of occurrence and speakers’ acceptability of the constructions. In this context, we conducted two experimental studies that investigate the perception and production of the constructions, respectively. Study 1 examined the factors that may influence the acceptability of the constructions, i.e., verb type and -cwu benefactive morpheme. Study 2 investigated whether the double accusative ditransitive construction is a valid construction ever produced by Korean speakers, using sentence completion experiment. The results showed that Korean speakers judge the double accusative ditransitives as highly unacceptable but they do not completely avoid them in sentence completions. We discussed the results in the context of theoretical and psycholinguistic perspectives.

This paper discusses the analysis of the Romance anticausative proposed in Alexiadou et al. (2015) and expands their analysis to inherent reflexive verbs in Romance languages so that a uniform account of both anticausatives and inherent reflexives in Romance languages are provided, further shedding lights on antipassive constructions in non-ergative-absolutive languages. Based on observations on inherent reflexives in Alboiu et al. (2004), we propose a neo-Davidsonian structure for inherent reflexives adapting Alexiadou et al. s (2015) analysis of anticausatives. That is, while in normal transitives the object introducer has the feature set [λx, D], in inherent reflexives the object introducer merely has the feature [D]. In brief, then, we propose that the SE anaphor is introduced in object position and the theme is merged as an adjunct PP. The SE anaphor merely satisfies the need for an element with a D feature in object position but does not receive a theta-role. Under our proposal transitive reflexives, anticausatives, and inherent reflexives in Romance receive a monolithic analysis, bringing us toward a uniform analysis of Romance SE.

This paper argues, from the perspective of labeling, that the epistemic predicates kolyeha- and kancwuha- taking small clause complements headed by -(u)lo have different verbal domain structures, in that the verbal domain with kolyeha- can have either a proleptic object or a raised object, while that with kancwuha- can only have a raised object. Contra Chomsky (2015), we argue that this results from the difference in the strength of the root based on the (un)availability of genitival nominalization containing those verbal noun Rs kancwu is a weak R, while kolye is a strong R. Using this modified labeling theory, we account for disparities between the verbal noun kolye and the verbal noun kancwu in scrambling and ellipsis, despite the fact that these two Rs convey the same semantic meaning as epistemic verbs.

To account for the positional (final vs. non-final) asymmetry observed in elliptical contexts (Park 2005, 2013, Park and Shin 2014, Ku and Cho 2014, Chung 2015a, b, among others), several proposals have been made in the literature including An’s (2016, 2018) extra deletion approach, Ahn and Cho’s (2017a, b) repetitive gapless right dislocation analysis, and Chung’s (2015a, b) oblique merge approach. This article addresses a novel type of fragment called an elliptical contrastive topic construction (ECTC, Chung 2020) and examines which of the theories proposed thus far best accounts for the characteristics of the construction. In particular, it will be shown that neither An’s nor Ahn and Cho’s theory properly captures certain important syntactic and semantic properties of the ECTC, while there is some room for Chung’s approach to accommodate them.

One basic restriction for multiple subject constructions in Korean is that it is preferable for the first, non-argument subject to be linked to a gap in the second, argument subject position. In this paper, I first show that the effect of this condition can be explained in terms of the Specified Subject Condition (Chomsky 1973) if we assume that multiple subject constructions involve A-movement. Positing A-movement, however, will face problems with various locality constraints in that the posited A-movement can violate the constraints on both A- and A’-movements, and I explore this issue in the latter part of the paper. An answer for this problem, I suggest, can be found if we acknowledge that non-clause-bounded A-movement is syntactically possible in Korean and that the Tensed S Condition and the Specified Subject Condition have a processing nature, similar to Island Constraints.