Speaker Corpus Research Articles

The Semantic Function of “Erqie” in Cross-Language Communication and its Usage Distribution : Take Korean and English Native Speakers as an Example

In long-speaking turn declarative discourse communication, non-native speakers use “Erqie” more frequently than native Chinese speakers. In the corpus of non-native speakers, its procedural meaning plays a leading role, and is mostly used to mark continuation or sequential relation between the information before and after, and play the function of general coordinate connective . In the corpus of native Chinese speakers, its conceptual meaning plays a leading role, and it is mostly used to indicate the logical semantic relationship and horizontal addition relationship of the front and back information. Affected by the progressive meaning of “Erqie”, its core discourse function is not only the continuation of the previous topic, but also the beginning of the follow-up information, emphasizing the relevance of the previous and subsequent information. Therefore, “Erqie” has a strong connection function. In addition, the discourse delay function and the nonsense cohesive discourse construction function of “Erqie” are unique to the corpus of non-native speakers, and occupy a considerable proportion.We believe that the above-mentioned different characteristics are affected by the general coordinate connectives in the English and Korean mother tongue system, and are also related to the speaker

/r/-sandhi in the speech of Queen Elizabeth II

This paper looks at the use of /r/-sandhi in the speech of Queen Elizabeth II. Potential contexts of /r/-sandhi were identified and analysed for the presence or absence of rhoticity and glottalisation in a corpus of Christmas speeches over a period of seven decades. The results show that the Queen avoids intrusive /r/ altogether but that she uses linking /r/ in most potential cases, that glottalisation is common when /r/-sandhi is not used, and that linking /r/ and glottalisation can also co-occur. A comparison with a longitudinal corpus of speakers also shows that the Queen resembles group-level trends in the case of linking /r/ but differs in the case of intrusive /r/. The results also indicate that a number of phonetic and usage-based variables influence linking /r/ usage, including previously unreported factors such as vowel quality collocation frequency. The findings are discussed in the context of usage-based theory.

Usage of Korean Idiomatic Expression in Terms of Language and Culture Literacy – Focusing on the Spoken Corpus of Native Korean Speakers

Usage of Korean Idiomatic Expression in Terms of Language and Culture Literacy – Focusing on the Spoken Corpus of Native Korean Speakers

The Information Structure–prosody interface in text-to-speech technologies. An empirical perspective

Abstract The correspondence between the communicative intention of a speaker in terms of Information Structure and the way this speaker reflects communicative aspects by means of prosody have been a fruitful field of study in Linguistics. However, text-to-speech applications still lack the variability and richness found in human speech in terms of how humans display their communication skills. Some attempts were made in the past to model one aspect of Information Structure, namely thematicity for its application to intonation generation in text-to-speech technologies. Yet, these applications suffer from two limitations: (i) they draw upon a small number of made-up simple question-answer pairs rather than on real (spoken or written) corpus material; and (ii) they do not explore whether any other interpretation would better suit a wider range of textual genres beyond dialogs. In this paper, two different interpretations of thematicity in the field of speech technologies are examined: the state-of-art binary (and flat) theme-rheme, and the hierarchical thematicity defined by Igor Mel’čuk within the Meaning-Text Theory. The outcome of the experiments on a corpus of native speakers of US English suggests that the latter interpretation of thematicity has a versatile implementation potential for text-to-speech applications of the Information Structure–prosody interface.

A Comparative Analysis of the Use of If-conditionals in Korean Learner Corpus and Native Speaker Corpus

A Comparative Analysis of the Use of If-conditionals in Korean Learner Corpus and Native Speaker Corpus

Hedging in spoken conversations by Norwegian learners of English

Hedging is an important aspect of pragmatic competence, but it is also a complex phenomenon that is difficult to master. This has resulted in underuse of hedging strategies by language learners, and many learners limit their hedging repertoire to a few strategies. This study compares the use of 10 hedging expressions commonly used in informal spoken English, a bit, I mean, I think, just, kind of/kinda, like, might, probably, thing(s), and you know, in data from LINDSEI-no, a corpus of Norwegian advanced learners, and LOCNEC, a comparable corpus of native English speakers. Norwegian learners typically show a high level of grammatical competence, but research on their pragmatic competence is limited. This study adds to the empirical research on Norwegian advanced learners by comparing the use of these expressions in native and learner language. The results indicate that as a group, Norwegian learners underuse hedging strategies, but when each expression and individual variation are considered, the picture is more nuanced. In fact, several of the Norwegian learners’ hedging practices partially overlap with several of the native speakers’ practices concerning hedging frequency and types of hedging strategies used.

The massive auditory lexical decision database: Acoustic analyses of a large-scale, single speaker corpus

The massive auditory lexical decision database: Acoustic analyses of a large-scale, single speaker corpus

한국대학생 영어 작문 코퍼스에서의 지칭 표현: 응집성을 중심으로

This study investigates the use of cohesive devices in the corpus of Korean college students’ compositions. In particular, it attempts to provide a comprehensive analysis of the lexico-grammatical contexts in which cohesive referential devices are employed in argumentative essays. For this purpose, a Korean EFL corpus and a native speaker corpus were analyzed based on Gray’s (2010) coding scheme. The coding scheme included variables such as the place and unit of the antecedent, the grammatical role and place of demonstratives, and their preceding or following structures. The results of the analysis showed that Korean EFL learners and native writers were similar in that they tended to use demonstrative pronouns as a subject and demonstrative determiners as objects/complements. Besides this general tendency, non-prototypical uses of the determiners were observed. The learners’ corpus manifested an overuse of an extended antecedent which could reduce the clarity of meaning. The most frequent verb following demonstrative pronouns was the copula (i.e., be) in both corpora. Finally, the learners tended to use a much smaller number of abstract/shell nouns than their native speaker counterparts.

Valency Dictionaries and Chinese Vocabulary Acquisition for Foreign Learners

Valency is a major source of lexical errors in foreign language learning. Accordingly, the research question is how the syntactic and semantic properties of a word can be retrieved from the corpora and represented in a Chinese valency dictionary to facilitate foreign learners' vocabulary acquisition. Within the three aspects of the valency framework — logical-semantic, syntactic and semantic-pragmatic valency — this study examines 60 cases of Chinese lexical misuse extracted from the HSK (Chinese Language Proficiency Test) Dynamic Compositions Corpus. The results suggest that the majority of cases of misuse occur in the dimension of semantic-pragmatic valency and that this semantic-pragmatic misuse can be ascribed to various factors such as semantic collocations, emotive variables, text styles, registers, and other contextual factors. The results are then utilized as syntactic, semantic and pragmatic information to be presented in a Chinese valency dictionary. Specifically, the results obtained from a case study of a misused word by referring to a large-scale native Chinese speaker corpus help retrieve a relatively full list of complementation patterns, based on which the study designs a Chinese valency entry that embodies three basic elements — quantitative valency, qualitative valency and valency patterns.

Contrastive Corpus Analysis: Investigating the Use of Opinion Verbs in Italian by Greek Learners

In this paper, we investigate the usage of opinion verbs in Italian as a foreign language and any errors this usage may cause, taking into account intralanguage and interlanguage factors. First of all, we analyze the frequency of opinion verb types, in a Greek corpus of native speakers and we investigate the frequency of the same type of verbs in an Italian Translator Learner Corpus compiled by the same learners. Secondly, we explore to what extent there are errors regarding the use of opinion verbs. The data will show that learners, in some cases, commit errors because of their mother tongue influence while in others they avoid using opinion verbs because of this particular lack of knowledge.

한국어 학습자를 위한 부사 ‘꼭’ 교육 연구

The purpose of this study is to make clear the functions and meanings of the adverb ‘kkok’, and then to set up the educational contents for teaching it. Through an analysis of four kinds of Korean dictionaries, it turns out that the adverb ‘kkok’ has two functions such as a manner adverb and a stative adverb, and six meanings such as compulsory action, voluntary action, recognition, accuracy, similarity, and force. There are 268 ‘kkoks’ in three kinds of Korean language textbooks, and about 95% among them are manner adverbs. On the other hand, about 67% of 3,500 ‘kkoks’ used in the corpus of Korean native speakers are manner adverbs. Meanwhile, in the test of choosing synonyms of ‘kkok’ carried out by 88 Korean undergraduates and 75 foreign undergraduates, differences of recognition on synonyms of ‘kkok’ between two groups are about 10% bigger in stative adverbs than in manner adverbs. This means that the stative adverb ‘kkok’ should be explicitly and systematically taught in Korean language education. Based on the corpus of Korean native speakers, the teaching level of each meaning of the adverb ‘kkok’ and educational contents for teaching it are presented as follows: teaching voluntary action or accuracy in level 1, compulsory action in level 2, recognition or similarity in level 3, force in level 4, and also providing practice activities replacing ‘kkok’ with its synonyms or practice adding ‘kkok’ in sentences. (Yeungnam University)

Extracting accent information from Urdu speech for forensic speaker recognition

This paper presents a new method for extraction of accent information from Urdu speech signals. Accent is used in speaker recognition system especially in forensic cases and plays a vital role in discriminating people of different groups, communities and origins due to their different speaking styles. The proposed method is based on Gaussian mixture model-universal background model (GMM-UBM), mel-frequency cepstral coefficients (MFCC), and a data augmentation (DA) process. The DA process appends features to base MFCC features and improves the accent extraction and forensic speaker recognition performances of GMM-UBM. Experiments are performed on an Urdu forensic speaker corpus. The experimental results show that the proposed method improves the equal error rate and the accuracy of GMM-UBM by 2.5 % and 3.7 %, respectively.

The Use of Make in a Korean Learner Spoken Corpus

The present study explores the various uses of the high-frequency verb ‘make’ by Korean learners of English. Specifically, we focus on the usage of this verb revealed in a Korean learner spoken corpus and compare its result with the native English speaker corpus (SBCSAE) and a Korean learner written corpus (Kim, 2002). A combination of manual and computer-aided approaches using software called AntConc 3.5.8 is used for analysis. The results demonstrate that there are significant differences in the usage of ‘make’ in the two spoken corpora under study. Specifically, while Korean learners tend to overgeneralize the causative and productive meaning of ‘make’, the preferred use of ‘make’ among English speakers is the delexical one. In addition, the frequency of productive meaning of ‘make’ by the Korean learners of English is noticeably higher as compared to the results of the written corpora. Taken together, the results of the present study allow us to draw the following two conclusions. First, in speaking, Korean learners use ‘make’ differently from native English speakers; second, the Korean learners’ uses of ‘make’ in speaking are significantly different from those in writing. Moreover, the current study implies pedagogical implications. That is, students should be given a corpus-driven teaching to learn how to use high-frequency delexical verbs correctly occurring in natural contexts, comparing native English speaker spoken corpus. (Incheon National University)

Modal Usage of Advanced Korean EFL Learners, Young Korean EFL Learners, and Native English Speakers: A Corpora Comparison Study

Building on Seog and Choi’s (2018) documentation of the early to intermediate interlanguage development of Korean learners’ English modal use, the current paper focuses on the highly advanced Korean learners and sheds light on the spectrum of interlanguage of Korean EFL learners in acquisition of the English modal verbs. For investigation and analyses, two learner corpora, the KNU Student English Learner Corpus-Written (KSELC-W) and the KNU English Learner Corpus (KELC), and a native speaker corpus, the Corpus of Contemporary American English (COCA), are utilized. Corpora comparisons and log-likelihood tests are conducted to reveal differences in the modal verb usage. The results show that in contrast with the early or intermediate young learners of KELC, the advanced learners of KSELC-W overuse the preterite forms of modal verbs such as could, would, and should. In addition, the learners productively and accurately used would to deliver the hypothetical sense while rarely using the modal verbs to express the epistemic sense. This study can represent the overall process of the Korean learners’ acquisition of English modal verbs.

Informal Correspondence by Greek Learners of the Italian Language: A Study Based on Learner Corpora, Native Corpora and Textbooks

The aim of this study is to compare various lexical structures between a learner corpus of students with Italian as a foreign language and a reference monolingual Italian corpus. More specifically, the first is a learner corpus (part of a wider learner corpus) comprised of Greek students studying Italian as a foreign language while the second is the CWIC reference corpus of native Italian speakers. The research findings help us explain the role of didactic material in comprehending linguistic structures that are found in informal letters/emails and, moreover, they provide us valuable information regarding the use of the same lexical structures by native speakers.

Universals and transfer in the acquisition of the progressive aspect: Evidence from L1 Chinese, German, and Spanish learners’ use of the progressive-ingin spoken English

AbstractThis study investigated the effects of learners’ first language (L1), lexical aspect of verbs, and proficiency levels on their use of the English progressive aspect. It analyzed spoken data from learners of three different L1s (Chinese, German, and Spanish) in an international learner corpus (LINSEI), in comparison with native speech in a comparable native speaker corpus (LOCNEC). The analysis reveals that regardless of learners’ L1 and proficiency levels, their use of progressive markings is predominantly associated with activity verbs (prototypes), supporting the association prediction of the Aspect Hypothesis (Andersen, Roger W. & Yasuhiro Shirai. 1994. Discourse motivations for some cognitive acquisition principles.Studies in Second Language Acquisition16(2). 133–156). Contrary to the fourth prediction of the Aspect Hypothesis, both intermediate and advanced learners use stative progressives, and especially intermediate L1 Spanish learners overuse stative progressives, indicating a complex interaction between L1 and proficiency on non-prototypical form-meaning associations. The results suggest that L1 effect, lexical aspect of verbs, and proficiency levels jointly drive tense-aspect acquisition.

Use of English Futures, Will vs. Be Going to: A Corpus-Based Comparison Study

The present study investigates usage of English futures, will and be going to, by Korean EFL learners and American native English speakers. The examined usage data are extracted from the native speaker corpus, the Corpus of Contemporary American English (COCA), and two learner corpora, the Kyungpook National University (KNU) Student English Learner Corpus-Written (KSELC-W) and the KNU English Learner Corpus (KELC), and compared to determine interlanguage developmental patterns and usage similarities or differences between native speakers and nonnative speakers. In support of existing literature, the findings from the current study reveal that the American native English speakers of COCA significantly overuse both will and be going to in the spoken register compared to the combined written registers. Furthermore, the findings also indicate that although the writing samples of the learner corpora were written during formal EFL education settings, will and be going to usage by the Korean EFL learners closely resemble the usage data of the combined written-all registers of COCA. Finally, semantic analyses show that the advanced learners of KSELC-W use be going to quite correctly with the semantic senses in the present form. In contrast to the learners in Coates (1983), they use be going to to deliver the sense of epistemicity rather than the root intention senses.

Anomaly detection based pronunciation verification approach using speech attribute features

Computer aided pronunciation training tools require accurate automatic pronunciation error detection algorithms to identify errors made by their users. However, the performance of these algorithms is highly dependent on the amount of mispronounced speech data used to train them and the reliability of its manual annotation. To overcome this problem, we turned the mispronunciation detection into an anomaly detection problem, which utilize algorithms trained with only correctly pronounced speech data. In this work we adopted the One-Class SVM as our anomaly detection model, with a specific model built for each phoneme. Each model was fed with a set of speech attribute features, namely the manners and places of articulation, extracted from a bank of binary DNN speech attribute detectors. We also applied multi-task learning and dropout approaches to alleviate the overfitting problem in the DNN speech attribute detectors. We trained the system using the WSJ0 and TIMIT standard data sets which contain only native English speech data and then evaluated it using three different data sets, a native English speaker corpus with artificial errors, a foreign-accented speech corpus and a children's disordered speech corpus. Finally, we compared our system with the conventional Goodness-of-Pronunciation (GOP) algorithm to demonstrate the effectiveness of our method. The results show that our method reduced the false-acceptance and false-rejection rates by 26% and 39% respectively compared to the GOP method.

A Case Study of Prosody of Chinese Synonyms

Semantic prosody, an important factor in the study of modern collocation, plays an important role in the Teaching Chinese as a Foreign Language (TCFL). By using retrieving function of online corpora and software, AntConc3.2.4w, this research compares the collocation and prosodies of two synonyms, "tiao" (choose) and "xuan" (select), in a large scale Chinese corpus BLCU Corpus Center (BCC) and an interlanguage corpus HUAYU. It is found that in the corpus of native speakers, "xuan" sometimes collocates abstract and negative nouns while "tiao" appears with concrete and neutral ones. However, foreign learners are confused by these synonyms, always using them with concrete and neutral nouns. This research also proves that the application of semantic prosody can enrich the content of lexical research and innovate vocabulary teaching methods in TCFL.

A Corpus-Based Study on “Delexical Verb + Noun” Collocations Made by Korean Learners of English

Collocation is one of the essential elements that must be acquired in the second language classroom to reach an advanced level of English proficiency. However, L2 learners of English tend to have difficulty in using native-like collocations. It has been reported that “delexical verb + noun” collocations are particularly difficult for L2 learners from various L1 backgrounds, owing to the fact that delexical verbs are characterized by little or depleted meaning. The present study uses corpus-based analyses to investigate how Korean learners of English use delexical collocations. It compares a learner corpus comprising essays written by Korean university students with a native speaker corpus as a control. The findings of a quantitative analysis indicate that Korean learners significantly underuse delexical collocations and that it would be hard for them to reach a native-like command with regard to this language element. The findings of a qualitative analysis show that Korean learners make quite a few collocational errors with delexical verbs, and that these errors seem to be caused by both interlingual and intralingual influences. These results offer meaningful pedagogic implications for language teachers and researchers in the field of English language teaching.