한국어 화시표현에 대한 기술주의적 해석

김진웅(Kim, Jinung) 한국어학회 2016 한국어학 Vol.70 No.-

This study includes a few proposals about Korean indexicals. Korean indexicals including personal pronouns, which are linked to the context of utterance, are also interpreted as anaphoric and descriptive uses. Based on Korean data, contra Kaplan(1989), I argue that there is no lexical list of pure indexicals in Korean and that indexicality is a kind of uses, which is interpreted under the context of utterance. It also demonstrates that descriptivism is more resonable theory than referentialism in that the former can provide the proper perspective to Korean indexicality but the latter cannot. In addition, by integrating anaphoric and descriptive uses of Korean indexicals, some categories in Korean can be neutralized. Specifically, first and second person pronoun can be interpreted in a similar way with third person pronoun because all of them can be interpreted as deictically and anaphorically.

자연언어처리에서 윤리적 문제와 해결 방안: 연령 및 지역 편향성 극복의 출발점으로서 방언자료 수집

김진웅(Jinung Kim) 경북대학교 사회과학기초자료연구소 2021 연구방법논총 Vol.6 No.1

이 연구는 자연언어처리 분야에서 윤리적 문제의 발생 원인과 유형을 살피고 기존의 언어자료와 차별화되는 언어 데이터 개발을 통해 데이터 편향성을 완화함으로써 윤리적 문제의 해결 방안을 모색하는 데에 그 목적이 있다. 첨단의 인공지능 시대에 그리스 시대로부터 비롯된 윤리학의 중요성은 더욱 강조되고 있다. 인공지능 채팅봇 테이(Tay)가 사용자들과 대화하면서 학습한 인종차별적 언행 때문에 서비스가 중단된 사건이나 MIT 연구팀에서 개발한 인공지능 노먼(Norman)이 드러낸 싸이코패스 성향은 인공지능의 개발에 윤리적 고려가 얼마나 중요한지를 보여주는 예이다. 인공지능이 윤리적 문제를 일으키는 원인은 편향된 데이터를 바탕으로 학습하기 때문이다. 성, 인종, 지역을 포함한 데이터의 편향성으로부터 인공지능의 결함이 시작된다. 한국어 방언 AI 데이터 구축 사업은 한국정보화진흥원의 감독 하에 진행되고 있는 국가 주도의 데이터 구축 사업이다. 이 사업에서 구축하는 음성 데이터는 기존의 음성 데이터에서 그 절대적인 양이 부족하다는 지적을 받아 온 자유발화를 수집한다는 점에서 그 가치가 매우 크며 지역별 연령별 화자에 대한 분포를 고려하여 수집한다는 측면에서 기존의 데이터에서 나타난 배제에 기인한 편향성을 완화 할 가능성을 제공한다. In the age of AI, ethics is the most popular field. Why is science originating from the Greek era recognized for its importance in the era of advanced artificial intelligence? The case where the service was interrupted due to the racist speech and behavior learned by AI chatbot Tay while talking with users, or the psychopathic tendency revealed by Norman, an artificial intelligence developed by the MIT research team. These are examples that show the importance of ethical consideration in the development of AI. So why does AI cause ethical problems? This is because AI learns based on biased data. The deficiencies of AI begin from the bias of data including gender, race, and region. The Korean dialect AI data construction project is a state-led data construction project under the supervision of the National Information Society Agency. The voice data built in this project is very valuable in that it collects free speech, which has been pointed out that the absolute amount of the existing voice data is insufficient, and it is collected in consideration of the distribution of speakers by age by region. It offers the possibility of overcoming bias caused by exclusions in the data.

메타담화 표지 ‘소위’와 ‘이른바’의 화용적 역할과 기능

김진웅(Kim, Jinung) 담화·인지언어학회 2017 담화와 인지 Vol.24 No.4

In this study, it is proposed that ‘sowi’ and ‘irenpa’ are metadiscourse markers, which cover linguistic expressions used to organize the texts and to communicate authorial attitudes. We utilized a corpus-based methodology to analyze those markers in the newspaper from the Yonsei 20th corpus. A Similar analysis is applied to some quotation marks which co-occur with ‘sowi’ and ‘irenpa’. As pointed out by Gutzmann and Stei(2011), quotation marks are minimal pragmatic indicators which signal that the expression should be interpreted according to its lexicon. They block the ordinary process of interpretation and then open the road to the pragmatic inference. Moreover, the same mechanism can be applied to ‘sowi’ and ‘irenpa’ as well. We compare the uses of ‘attitude quotation’ and those of ‘sowi’ and ‘irenpa’. This means all of them play a similar role as metadiscourse markers and their roles are superfluous relative to different contexts.

시제, 양태 그리고 한국어의 양태 종속

김진웅 ( Jinung Kim ) 연세대학교 언어정보연구원(구 연세대학교 언어정보개발원) 2016 언어사실과 관점 Vol.38 No.-

이 연구에서는 양태 종속 현상을 중심으로 양태 표지인 ‘-을 수’, ‘-을 것’과증거성 표지인 ‘-지’가 다양한 맥락에서 상호간에 영향을 주고받으며 그 해석이 제약을 받거나 허용되는 양상을 살피고 원인을 분석하였다. 1980년대 후반부터 양태 종속 현상은 맥락 의존 의미론에 대한 관심과 더불어 영어자료에 대해 연구가 시도되었고 2000 년대에는 일본어 자료에서도 동일한 현상이 존재한다는 점이 밝혀진 바가 있다. 하지만 한국어 자료를 바탕으로 한 연구는 아직 본격적으로 진행된 적이 없다. 이 연구의 일차적인 목표는 한국어에 양태 종속 현상이 존재하는가의 여부인데 다양한 예문을 통해서 한국어에도 양태 종속 현상이 존재한다는 점을 확인할 수 있었다. 한국어에서 양태 종속 현상이 일어나는 일차적 원인은 의존명사를 수식하는 관형절의 시제에서 비롯된 가능 세계로의 해석이다. 양태 종속 현상은 선행 발화와 후행 발화 사이의 시제의 일치만으로 설명할 수 없는 예외들을 발견할 수 있었다. 이는 한국어의 담화 표지들에 의해 나타나는 수사적 관계를 통해서 이유를 설명할 수 있었다. 이들 수사적 관계의 영향을 설명하기 위해서는 분절 담화 표상 이론의 논의를 수용했다.

구어말뭉치에 나타난 대등연결어미 운율의 양상

김진웅(Kim, Jinung),박상훈(Park, Sang-Hoon) 담화·인지언어학회 2013 담화와 인지 Vol.20 No.3

The purpose of our study is to demonstrate Continuation Rises(CR) in Korean by analyzing the prosody of coordinating connective endings and their distributions. The motivation of our study is that CR is affected by physiological conditions which decide the mechanism of human utterance and also is cross-linguistically found in many other languages such as French, Dutch, and Spanish. Specifically, we annotated additional information to ‘New Yonsei Corpus 2’ to extract the proper tokens of coordinating connective endings and examined their prosodies. Our result showed that ‘H’ occurs more frequently than ‘L’ with coordinating connective endings. Moreover, we found that ‘M’ usually co-occurs with ‘imperfect Intonation Phrase’ with coordinating connective endings and that ‘H’ is correlated with dependency of antecedent clauses on following clauses.

화용표지로서 인용부호의 사용 양상과 유형 분류

김진웅(Kim Jinung) 한국어학회 2017 한국어학 Vol.77 No.-

The purpose of the study is to elucidate the roles and characteristics of quotation marks based on the pragmatic approach. Traditionally, the quotation in Korean is subdivided as direct and indirect quotation. However, Those categories cannot reflect the whole reality of various linguistic phenomena related to quotation marks. In fact, linguistic expressions like syllables, morphemes, words in addition to utterance can appear between quotation marks and nonlinguistic expressions like sounds and symbols also do. Accordingly, anything can appear between quotation marks. Moreover, I suggest types of quotation depending on syntactic and pragmatic criteria. Specifically, I show five kinds of them such as pure quotation, direct quotation, name quotation, mixed quotation, attitude quotation. I demonstrate that the pragmatic approach is more convincing than semantic one to explain the uses of quotation marks in Korean.

화용적 추론에 기반한 환유의 해석

김진웅(Kim Jinung) 한글학회 2017 한글 Vol.- No.315

이 연구에서는 기존의 한국어의 환유 연구에서 주도적인 역할을 한인지언어학적 관점의 한계를 지적하고 화용적 추론을 통한 환유 해석을 모색하였다. 기존의 환유 모델에 대한 대안으로서 적합성 이론의 핵심적인 기제인 임시 개념을 도입하여 이를 한국어 환유 해석에 적용하였다. 결과적으로 환유를 화용적 함수나 개념적 사상으로 이해하는 대신에 화자와 청자가 소통 과정에서 최대한의 적합성을 발견해 나가는 추론 과정으로 이해할 수 있었다. 또한 한국어에서 나타나는 지시적 환유와 관습적 환유의 해석 과정을 보이고 환유라는 언어 현상이 일어나는 동기를 제시했다. In this study, I indicate the limitation of the cognitive linguistics, which overwhelmingly affects the account of metonymy in Korea and try to explain metonymy based on pragmatic inference effects. I accept the proposal by relevance theorists that metonymy can be properly explained by means of ad hoc concepts and account for metonymy in Korean. As a result, metonymy can be understood as the communicative process in which a speaker and a hearer arrive at the goal of the maxim of relevance, not as pragmatic function or conceptual mapping. Lastly, I introduced the emergence and interpretation of referential and conventional metonymy and showed what motivated metonymy.

집단 차별어의 의미론과 화용론

김진웅(Kim, Jinung) 담화·인지언어학회 2024 담화와 인지 Vol.31 No.1

This study adopts research methods from the philosophy of language to present the categorical properties of slurs and to compare and evaluate the various arguments raised by competing theories. Slurs belong to the subtypes of unethical or hate speech that have been studied in the literature, but they have distinctive characteristics that make them an independent category. First, slurs are characterized by the fact that it targets a homogeneous group of people. While profanity, which falls under the category of unethical expression, shows discrimination or hatred regardless of the individual or group, slurs are limited to a single group. Because the group they refer to is clear, slurs are paired with neutral correspondents that refer to the same audience. In addition, the negative attitudes contained in slurs are directed at a specific group, rather than at an individual. Other important characteristics of slurs include scopelessness, appropriation, and autonomy of attack. In this study, I examine the features, advantages, and disadvantages of various semantic, hybrid, and pragmatic approaches. The significance of this study is that it broadens the horizon of research on theoretical trends in the study of slurs based on the perspectives of linguistic philosophy and semantics.

말뭉치언어학 관점에서 신문기사 범죄사건 명명 방식 분석

유희재(Yu, Huijae),유성희(Yoo, Seonghee),김진웅(Kim, Jinung) 담화·인지언어학회 2016 담화와 인지 Vol.23 No.2

This study undertakes a corpus-based methodology to analyze the naming process of crime cases in the headlines of Korean newspapers. The most used piece of information in the name of crime cases was information about the crime itself, often realized as ‘murder’ or ‘rape’ cases. This information was often used together with information about the location of the crime, which was found to be the second most used piece of information. This tendency can be explained by the optimization of conveying information - more general information about the crime type together with more specific information about its location can refer to a specific instance of crime most effectively. It was found in the data that in the naming of crime cases, information about the perpetrators of the crime was used much less frequently than information about the victims of those crimes, with later analysis revealing that only in unusual cases were details of criminals given. Information about the victims on the other hand was used widely, especially when the victims were female and of a young age. Following these findings, we underline that the ‘ideal’ victims of crime for Korean newspapers to report are young women, and that the media tend to emphasize the victims of crime rather than the criminals.

형태 분석을 위한 메신저 텍스트 처리 방안

안의정 ( Ahn Euijeong ),송현주 ( Song Hyunju ),김진웅 ( Kim Jinung ) 한국텍스트언어학회 2020 텍스트언어학 Vol.49 No.-

This study presents a methodology that facilitates UGC morphological analysis by normalizing non-standard texts, and shows that the performance of existing NLP tools can be improved through it. The KakaoTalk messenger texts were morphologically analyzed using Utagger and their error rates were checked over a total of three steps. In order to increase the accuracy of morphological analysis, it is necessary to grasp the characteristics of the messenger data. In the case of messenger data, the spacing norms are not observed, and punctuation marks are often omitted. Variants that are written as they sound or whose letters are partially modified are used, and typos are also frequent. In addition, various abbreviations are used, new words and dialects appear, and new sentence final endings are frequently used. These linguistic features can cause an error of POS and should be considered in the process of normalization, The error rate was changed from 35.1% to 17.7% after due to wrong interword space detection, and the error rate was reduced by 12.2% due to spelling checking. The remaining errors were the case of misanalyzing the sentence final endings as a connecting endings due to the absence of punctuation marks, and errors related to new proper nouns, new words, and interjections.