The <Yonsei Dictionary> from a Korean Informatics Perspective

서상규 연세대학교 언어정보연구원 2018 언어사실과 관점 Vol.43 No.-

이 논문에서는, 1986 년 이래 <연세한국어사전>의 편찬 과정의 전개와 국어정보학의 태동과 발전의 흐름을 소개하고, 올림말 구성에 초점을 맞추어 국어정보학적 분석을 시도하였다. 특히 <연세한국어사전>의 편찬 과정에서 특히 올림말의 선정에 잦기(빈도) 정보를 활용한 것에 착안하여, 실제로 이 사전의 올림말 구성이, 오늘날 더욱 진전된 정교한 말뭉치 주석(분석)의 결과에 비추어 볼 때, 과연 얼마만큼의 타당성을 갖는가 하는 문제를 다루고자 하였다. 이를 위해서, 약 5 만 개의 올림말로 이루어진 <연세한국어사전>의뜻풀이는 어떠한 낱말로 이루어졌으며, 과연 이 사전의 올림말로 뜻풀이를 어느 정도나 이해할 수 있는가를 밝히기 위해서, 이 사전의 전체뜻풀이문을 추출하여 말뭉치로 구성하여 분석하였다. 이를 통해서 다음과 같은 사실을 알 수 있었다. 뜻풀이말뭉치에는 모두 21,490 개의 단어와 형태(조사와 어미 등)가 쓰였으며, 그 빈도수를 다 합하면766,322 개가 된다. 모든 뜻풀이의 95.1%의 사용 어휘가 <연세한국어사전> 즉 YKD(1998)의 표제어로 수록되어 있으며, 이들이 사용된 전체 빈도수의 합계는 뜻풀이 말뭉치 전체의 98.7%에 이른다는 놀라운결과를 얻게 된다. 서상규(1998)에서 밝힌 바와 같이, 이 사전이 실제말뭉치에서 사용 빈도가 높은 중요 어휘를 추출하여 표제어로 수록하였다는 사실이 입증되는 것이다. YKD 사전의 표제어 중 실질어휘는 총 47,877 개이다. 이 실질어휘중에서 실제 사전의 뜻풀이에 사용된 것은 18,105 개로 전체의 약37.8%에 불과하다. 이 점을 미루어 본다면, YKD(1998)은, 의도했든 의도하지 않았든, 뜻풀이 사용 어휘를 매우 적극적으로 통제하였다고 평가할 만하다. 문법형태 표제어의 출현 비율은 평균적으로 48.4%로 나타나서, 앞서살펴본 실질어휘 표제어에 비해서, 10% 이상 높다. 즉 문법형태는 뜻풀이에 훨씬 더 자주 사용되었다는 것이다. 한편, YKD 의 표제어가 실제 말뭉치의 사용 양상을 얼마나 충실히반영하고 있는지를 알아보기 위하여, 문어와 구어 균형 말뭉치(각 100 만 마디)의 주석 결과를 비교하였다. 성격이 다른 두 개의 균형 말뭉치(구어와 문어)의 단어 목록과 일치하는, YKD 의 표제어들의 빈도수의 합은, 전체 말뭉치의 무려 90.7%를차지하는 것으로 나타났다. 이 사실은, <연세한국어사전>이 실제로 말뭉치의 분석에 기반을 둔 국어정보학적 편찬 방법에 충실했다는 사실을뒷받침해 준다. 이와는 반대로, 올림말의 약 34%에 해당하는 16,637 개의 단어(형태)는, 2 개의 말뭉치에 나타나지 않았고, 특히 구어 말뭉치의 사용 단어가 소극적으로 반영되어 있음을 발견할 수 있었다.

힐버트 변환을 이용한 디지털 위상천이기의 성능 분석

서상규,정봉식,Seo, Sang Gyu,Jeong, Bong-Sik 한국융합신호처리학회 2013 융합신호처리학회 논문지 (JISPS) Vol.14 No.1

본 논문에서는 멀티암(multi-arm) 스파이럴 안테나용 디지털 위상천이기(digital phase-shifter)를 힐버트 변환(Hillbert transform)을 이용하여 설계하였다. 힐버트 변환은 입력신호에 포함된 모든 주파수 성분을 $90^{\circ}$ 위상천이 시키며, 퓨리에 변환(Fourier transform)과 역퓨리에 변환(Inverse FIT)을 통해 구현된다. 디지털 위상천이기는 ADC(Analog-digital converter)로 샘플링된 입력신호에 힐버트 변환을 적용하여 위상차가 $90^{\circ}$인 두 신호를 생성하고, 이 두 신호를 이용하여 입력신호의 위상을 천이위상만큼 천이시키게 한다. 힐버트 변환 기반의 디지털 위상천이기는 Xilinx사의 System generator로 설계되었고, 입력 잡음, FFT 포인트 수, 샘플링 주기, 입력신호의 초기위상 및 천이 위상각 등에 따른 위상천이 성능을 시뮬레이션 하였으며, Matlab 결과와 비교하여 일치함을 확인하였다. In this paper digital phase-shifter for multi-arm spiral antennas was designed by using Hilbert transform. All frequency components in input signal are phase-shifted for 90 degree by Hilbert transform, and the transform is implemented by FIT and IFIT. Digital phase-shifter generates two signals with phase difference of 90 degree by using Hilbert transform from input signals sampled by analog-digital converter(ADC), and then the input signal is phase-shifted for a given phase by using two signals. Hilbert transform based on digital phase-shifter is designed by Xilinx System generator, and the effects of input noise, FIT point, sampling period, initial phase of input signal, and shifted phase are simulated and its results are compared with Matlab results.

Overexpression of ADP-glucose pyrophosphorylase (IbAGPaseS) affects expression of carbohydrate regulated genes in sweet potato [Ipomoea batatas (L.) Lam. cv. Yulmi]

서상규,Su-Hyeon Bea,전병기,김선태,권석윤,김선형 한국유전학회 2015 Genes & Genomics Vol.37 No.7

ADP-glucose pyrophosphorylase is crucial in starch metabolism. The open reading frame of IbAGPaseS gene is comprised of 1539-base pairs and encodes a polypeptide of 512 amino acid residues. The deduced IbAGPaseS protein sequence was homologous with that of other AGPase proteins of various plant species. Also, the IbAGPaseS gene appears to be a member of multiple gene families and further, the differences in the copy numbers among the ten cultivars were considered to lead to differences in starch contents. IbAGPaseS was expressed in whole tissues and revealed high activity in thick-pigmented roots and developing tuberous roots. RT-PCR analysis indicated that the main role of IbAGPaseS is starch biosynthesis accumulating sink tissues during sweet potato root development. The IbAGPaseS cDNA was introduced to sweet potato under the control of the CaMV 35S promoter using Agrobacterium-mediated gene transfer techniques, thereby affecting the expression levels of carbohydrate regulated genes and sporamin gene. These results suggest that IbAGPaseS affects carbohydrate gene regulation.

석인 선생의 옛말 연구

서상규 국립국어연구원 1998 새국어생활 Vol.8 No.3

최 현배의 <우리말본> 첫재매 유인본 연구

서상규 한글학회 2010 한글 Vol.- No.289

The purpose of this study is to first construct a detailed bibliography of the mimeograph editions of Urimalbon Cheotjaemae (the first chapter) held at Yonsei University and clarify the characteristics of each, and then to identify the value of each edition of Urimalbon Cheotjaemae through an analysis of how they were produced. In particular, through a comparison of the bibliographical characteristics, handwritten writing styles, order of contents, and degree of similarity of the main terminology and descriptions used in each mimeograph edition, this study aims to clarify the stages and forms of the production process of Urimalbon Cheotjaemae(1929) and Urimalbon(1937). Through this analysis, it is proven that a newly discovered mimeograph edition of Urimalbon Cheotjaemae is a different book compared to previously known editions. In addition, the bibliographical and content characteristics of this book are described, and the similarities and differences, production periods, and production order of the two mimeograph editions of Cheotjaemae are discussed. 이 논문의 목적은 연세대학교에 소장된 <우리말본> 첫재매의 유인본 의 구체적인 서지와 특징을 밝히고, <우리말본> 첫재매의 성립 과정에서 이들이 어떠한 가치를 갖는가를 밝히는 데 있다. 특히 유인본의 서지적 특성, 손글씨의 글씨체 비교, 차례 구성, 주요 갈말(용어) 및 기술 내용과 표현의 일치도 등을 비교함으로써, <우리말본 첫재매>(1929)와 <우리말본>(1937)의 성립 과정 단계와 모습을 알아보고자 한다. 이를 통해서 새로이 발견된 <우리말본> 첫재매의 유인본이 기존에 알려진 유인본과는 다른 책이라는 사실을 밝힌다. 아울러 이 책의 서지적, 내용적 특성을 기술하고, 첫재매의 두 가지 유인본의 공통점과 차이점, 그리고 성립 시기와 유인본 간의 선후 관계 등을 고찰한다.

바이오매스로서의 억새에 대한 연구 동향

서상규,이정은,전서범,이병현,구본철,서세정,김선형 한국식물생명공학회 2009 식물생명공학회지 Vol.36 No.4

The carbon dioxide concentration of the atmosphere is projected to increase by almost 50% over the first 50 years of this century. The major cause of this increase is continued combustion of fossil fuels. As a result, the significant changes in climate that have already occurred will be amplified, in particular a global temperature increase. Renewable energy production has a central role to play in abating net CO2 emissions to a level that will arrest the development of global warming. Especially, biomass crops are becoming increasingly important as concerns grow about climate change and the need to replace carbon dioxide- producing fossil fuels with carbon-neutral renewable sources of energy. To succeed in this role, biomass crop has to grow rapidly and yield a reliable, regular harvest. A prime candidate is Miscanthus, or Asian elephant grass, a perennial species that produces over 3 metres of bamboo-like stems in a year. Miscanthus species are typically diploid or tetraploid. Hybrids between species with different ploidy levels result in the highly productive triploid hybrids, M. × giganteus. Here we will detail the Miscanthus characteristics desired of a biomass fuel crop.

백 춘범 지은 [단어결합과 어울림 연구] 평설

서상규 한글학회 1997 한글 Vol.- No.237

Development of an efficient protocol for high-frequency regeneration system in Hibiscus syriacus L.

서상규,류선희,Yang Zhou,김선형 한국식물생명공학회 2017 JOURNAL OF PLANT BIOTECHNOLOGY Vol.44 No.2

A rapid and efficient in vitro regeneration system was established for Hibiscus syriacus L. The successful regeneration protocol employs induction of shoot organogenesis on leaf, petiole, and root explants. Among the various plant growth regulators evaluated, thidiazuron (TDZ) was the most effective for inducing rapid shoot formation. Most efficient shoot regeneration frequency was obtained from Murashige and Skoog (MS) media containing 0.01 mg/L TDZ. Regeneration efficiency was highest in the roots, and lowest in the leaves. A combination of 0.01 mg/L TDZ with benzyladenine (BAP) markedly improved the frequency of shoot differentiation from the root (up to 98%) and petiole (up to 88%) explants. Furthermore, leaf and petiole explants showed the highest frequency of shoot induction in half-strength MS media containing 0.01 mg/L TDZ and 1.0 mg/L BAP, while root explants formed the greatest number of shoots when 0.01 mg/L TDZ and 0.1 mg/L BAP were added to half-strength MS media. Although the frequency of shoot differentiation from leaf explants was only 50%, the leaf is considered the most efficient plant organ for use in tissue culture because leaves are easier to obtain than roots and petioles. Our findings show that various organs of H. syriacus can be used for plant regeneration, and the protocol developed in this study may be applicable in the horticulture industry.

창작 탈놀이극의 연기방법론 연구 - ‘왼손이’와 ‘날아라 장수매’의 공연사례를 중심으로 -

서상규 한국연기예술학회 2017 연기예술연구 Vol.10 No.2

“연기란 배우가 등장인물이 되어 관객에게 이야기를 들려주고 보여주어 관객에게 감흥을 일으키는 예술 적 행위이다”라고 일반적인 정의를 내릴 수 있겠다. 그러나 언어와 인물 내면이 연기의 중심인 스타니슬 랍스키의 사실주의 연기, 신체의 유연성을 강조하는 메이어홀드의 생체역학(biomechinics)3) 연기, 배우와 등장인물간의 거리두기를 통해 소외효과(alienation effect)를 강조하는 브레히트4) 연기론 등등, ‘연기란 무 엇인가’에 대한 기본 정의는 맥을 같이하여도 연기 방법론은 시대에 따라, 연극사조에 따라 혹은 연출가 의 사상에 따라 그 연기방법론을 달리하여 연구되고 발전되어왔다. 창작 탈놀이극과 탈놀이 연기방법은 우리나라 민속예술인, 탈춤, 풍물굿, 무속굿, 전통무용, 민요, 판소 리 등과 지난 70-80년대에 성행한 마당극에 그 뿌리를 두고 있다. 창작 탈놀이 연기는 현실을 재현하는 사실적 연기와는 달리 몸의 신체적 표현을 중시하는, 탈놀이 고유의 특징적 연기 방식이 자리 잡고 있으 며 그 제 요소들의 방법론적 중요성은 매우 크다 할 것이다. 본 논고에서는 우리나라에서 70년대부터 자생적으로 나타난 창작 탈놀이극의 연기 방법과 창작 탈놀 이극 두 편의 공연에 적용된 사례를 통해 탈을 쓰고 놀 때, 탈을 가지고 놀 때에 나타나는 탈놀이 연기의 특징과 한계 그리고 앞으로의 발전가능성을 살펴보고자 한다. Acting is generally defined as “an artistic activity in which a story is told by means of its enactment by an actor or actress who adopts a character to inspire audience”. However, various acting methodologies have been studied and developed according to the times, according to the theater theory, or according to the idea of the director even though the basic definition of ‘what acting is’ is the same. Examples are such as; Stanislavsky's realism focused on language and inner sides of characters, Mayerhold's Biomechanics emphasizing the flexibility of the body and Brecht’s theory which emphasizes the alienation effect through the distance between the actor and the characters, Creative Mask play and its acting method are rooted in Korean traditional fork arts such as mask dance, poongmulgut(percussion music), moosokgut(shamanism), korean dance, folk song, pansori(solo epic chant) and madangkuk(open air play) which was flourished in 70s and 80s. In contrast to the realistic performance that reproduces reality, the act of creative play has a unique characteristic of acting, which emphasizes the physical expression of the body and its methodological importance is enormous. In this paper, I will discuss the acting methodology of Creative Mask play originated from 1970s in Korea. In addition, I will examine the characteristics, limitations and possibility of future developments of this Creative Mask play.

<우리말 말수 사용의 잦기 조사>의계량적 연구

서상규 연세대학교 언어정보연구원 2020 언어사실과 관점 Vol.51 No.-

This study is intended to reveal that Hyunbae Choi’s research of 『Frequency Research of Korean Vocabulary Use』 (1956) is a valuable achievement which was conducted long before the start of modern field of Korean language informatics based on computer and corpus. First, we analyzed and described the structure and characteristics of the vocabulary list, and the vocabulary classification system and the description methods to distinguish meanings in 『Frequency Research of Korean Vocabulary Use』 (1956). Based on this result, we quantitatively analyzed the characteristics of the vocabulary of the 1950s. In order to compare the quantitative features of the vocabulary of the 1950s represented by this data with the features of the Korean written and spoken language at the end of the 20th century, we used three corpora in ‘New Yonsei Corpus (nYsc)’ (1 type of written language corpus and 2 types of spoken language corpus: each corpus contains one million words). 『Frequency Research of Korean Vocabulary Use』 (1956) and New Yonsei Corpus 1 (nYsc1) are both written language corpus, and New Yonsei Corpus 2 (nYsc2) and New Yonsei Corpus 3 (nYsc3) are both spoken language corpus. Based on the data to be compared, we analyzed the quantitative characteristics of Korean language in two aspects: the distribution and composition of words by part of speech, and the distribution by etymology (type of words). By comparing the distribution of vocabulary by part of speech in terms of the number of words and its ratio, we found a remarkable difference in nouns, pronouns, adverbs, and interjections. In the written language corpora, 『Frequency Research of Korean Vocabulary Use』 (1956) and nYsc1, the proportion of nouns was higher than that in the spoken language corpora (nYsc2, nYsc3). However, the ratios of pronouns, adverbs, and interjections were higher in the spoken language corpora at the end of the 20th century (nYsc2, nYsc3) than the written language corpora (1956, nYsc1). In terms of the sum of frequency and its ratio of the distribution of vocabulary by part of speech, a large difference can be found in nouns, pronouns, adverbs, and interjections. This aspect is like numbers of words, but the difference is much larger. However, a new feature that has emerged in terms of the sum of frequency is that the ratio of particle’s sum of frequency is higher in written language (1956, nYsc1) than in spoken language. On the other hand, as a result of analysis from the etymological distribution, in terms of the number of words and its ratio, it was characterized that the ratio of Chinese words was significantly higher than that of Korean or foreign words. The ratio of Chinese words in written language is higher than that of in spoken language, and the ratio of Korean words in spoken language was higher than that of in written language. However, it shows the opposite features in terms of the sum of frequency and its ratio. In other words, it can be seen that the relationship between Korean and Chinese words viewed by etymology is inversely proportional to each other. In other words, in terms of the number of words alone, Chinese words appear almost twice as many as Korean words, but when viewed as the sum of the number of frequency actually used, the relationship is the opposite, Korean words appear at a rate that is three times higher than that of Chinese words. Another remarkable result is that the difference between written and spoken language is very clear. Therefore, in the spoken language corpus, whether it is the number of words or the sum of the number of frequency, it appears at a much higher rate than the written language. Through this analysis, we once again confirm that it is a very important achievement to analyze the characteristics of Korean language in the 1950s reflected in this data, along with the Korean language informational value of 『Frequency Research of Korean Vocabulary Use』 (1956).