Language Corpora Research Articles

CODIT. Une nouvelle ressource pour l’étude de l’italien dans une perspective diachronique: applications dans le domaine morphologique

This paper presents a new diachronic corpus of the Italian language, i.e., CODIT (COrpus Diacronico dell’ITaliano ‘Diachronic corpus of Italian’). After a description of the corpus design, the second part of the article is devoted to show possible applications of the corpus in the morphological field, specifically in the domain of evaluative morphology. We highlight the corpus’ strengths and limitations through the diachronic analysis of two affixes, i.e., sopra- and -oide. The study contributes to shed light on the dynamism of evaluative morphology by focusing on the loss of the intensification function suffered by sopra- and the emergence of an approximating value that currently affects -oide.

A nameless relativ BY EXAMPLE: the degree of grammaticalization

Abstract: In the article, in relation to the nameless relativ, one of the syntactic criteria for determining the degree of grammaticalization of a significant word is considered by EXAMPLE. The relevance of the study is determined by the need to determine the place of each derived unit of the service type in the language system. The object of research is a prepositional unit BY EXAMPLE, which is included in the group of nominal relatives. The subject of our research is the degree of grammaticalization of the nameless relativ BY EXAMPLE. The purpose of the study is to determine the grammaticalization stage at which the nameless relativ is located BY EXAMPLE. The paper uses a descriptive method and a corpus method of collecting research material. The scientific novelty consists in the first described models of interaction of a prepositional-case combination, FOR example, with adjectives included in its composition. Based on the analysis of the linguistic material presented in the National Corpus of the Russian language, two main types of models have been established – with a pronominal and a non-nominal adjective. It is shown that a non-nominal adjective in such a model acts in two functions: the actant name function and the modifier function, which is typical for a significant number of similar structures with other prepositional derived units. It is revealed that personal pronominal adjectives also indicate the name of the actor. It is determined that demonstrative pronominal adjectives do not have named functions. Their entry into the model indicates the absence of a service function in the combination FOR EXAMPLE. The article concludes that the nameless relativ, by example, is only at the initial stage of grammaticalization.

An empirical evaluation of phrase-based statistical machine translation for Indonesia slang-word translator

<span>The use of slang (non-standard language), especially in social media, is increasing. It causes reducing the level of understanding when communicating because not everyone understands slang (non-standard language). The purpose of this work is to develop a slang-word translator. The other objective is to find the minimum number of sentences and </span><span>BiLingual Evaluation Understudy</span><span> (BLEU) score used as a benchmark to determine that the translation is understandable. The approach used in this project is a Phrase-based statistical machine translation (PBSMT) approach, suitable for low resource language, with a dataset of 100,000 sentences taken from the comments column of several online political news portals. The comments are then manually translated to produce a parallel corpus of non-standard language-standard language. The sample sentences are taken from the dataset then distributed using questionnaires to obtain the human understanding level regarding the translation result. The result of the implementation is a BLEU score of 64 and the minimum number of sentences to have an understandable machine translation is 500. The conclusion drawn from the distributed questionnaires is that humans can understand the sentences produced by the translation machine.</span>

Identifying the influence of transfer learning method in developing an end-to-end automatic speech recognition system with a low data level

Ensuring the best quality and performance of modern speech technologies, today, is possible based on the widespread use of machine learning methods. The idea of this project is to study and implement an end-to-end system of automatic speech recognition using machine learning methods, as well as to develop new mathematical models and algorithms for solving the problem of automatic speech recognition for agglutinative (Turkic) languages. Many research papers have shown that deep learning methods make it easier to train automatic speech recognition systems that use an end-to-end approach. This method can also train an automatic speech recognition system directly, that is, without manual work with raw signals. Despite the good recognition quality, this model has some drawbacks. These disadvantages are based on the need for a large amount of data for training. This is a serious problem for low-data languages, especially Turkic languages such as Kazakh and Azerbaijani. To solve this problem, various methods are needed to apply. Some methods are used for end-to-end speech recognition of languages belonging to the group of languages of the same family (agglutinative languages). Method for low-resource languages is transfer learning, and for large resources – multi-task learning. To increase efficiency and quickly solve the problem associated with a limited resource, transfer learning was used for the end-to-end model. The transfer learning method helped to fit a model trained on the Kazakh dataset to the Azerbaijani dataset. Thereby, two language corpora were trained simultaneously. Conducted experiments with two corpora show that transfer learning can reduce the symbol error rate, phoneme error rate (PER), by 14.23 % compared to baseline models (DNN+HMM, WaveNet, and CNC+LM). Therefore, the realized model with the transfer method can be used to recognize other low-resource languages.

Natural and elicited: Sign language corpus linguistics and linguistic ethnography as complementary methodologies

Natural and elicited: Sign language corpus linguistics and linguistic ethnography as complementary methodologies

Diverse and styled image captioning using singular value decomposition‐based mixture of recurrent experts

AbstractWith significant advances in vision and natural language processing, the generation of image captions becomes a need. Mathews, Xie, and He extended a new model to generate styled captions by separating semantics and style. In continuation of their work, here, a new captioning model is developed, including an image encoder to extract the features, a mixture of recurrent networks to embed the set of extracted features to a group of words, and a sentence generator that combines the obtained words as a stylized sentence. This Mixture of Recurrent Experts (MoRE) system uses a new training algorithm that derives singular value decomposition from weighting matrices of Recurrent Neural Networks (RNNs) to increase the diversity of captions. Each decomposition step depends on a distinctive factor based on the number of RNNs in MoRE. The used sentence generator gives a stylized language corpus without paired images. Besides, the styled and diverse captions are extracted without training on a densely labeled or styled dataset. MoRE on the COCO dataset generated diverse and stylized image captions without the necessity of extra‐labeling and improved descriptions in terms of content accuracy.

SYMBOLIC SIGNS OF THE HAPPINESS CONCEPT: LINGUOCULTURAL ANALYSIS

The purpose of the article is to determine the complex of symbolic signs of the happiness concept in its structure.
 Objectives of the article: 1. To analyze the language material and identify the symbolic signs of the happiness concept. 2. To present the classification of symbolic signs of the happiness concept.
 The relevance of the work is determined by the understanding of the symbols of happiness related to Russian and other linguocultures. The scientific novelty of the work lies in the fact that this article represents the first experience of addressing the topic of symbolic signs of the happiness concept in the aspect of primordial and borrowed symbols of Russian linguoculture.
 The research material. The National Corpus of the Russian language (www.ruscorpora.ru) is the source of the language material.
 The main research methods include the descriptive, conceptual and interpretive ones.
 The article examines a significant part of the structure of the happiness concept, i.e. its symbolic signs. The symbolic signs of the happiness concept were combined into four blocks: 1. symbols of happiness of the Russian linguoculture, 2. symbols of happiness among other peoples, 3. international symbols and 4. individual symbols of happiness. The language material represents an extensive actualization of the symbolic signs of the happiness concept in the first block.

Variation in the use of constructed action according to discourse type and age in Finnish Sign Language

This paper presents a study of the use of constructed action (CA) in the stories and conversations of adult Finnish Sign Language (FinSL) signers of different ages. CA is defined here as a type of depiction in which a signer enacts the actions, feelings, thoughts and utterances of discourse referents with different parts of their body. Most studies on CA in sign languages have been done on the basis of signed storytelling, and little is known about how the use of CA varies in different discourse types. The use of CA has also been noted to vary between individual signers, but we do not yet know much about the socio-individual phenomena that may be linked to this variation. In the present study, we investigate whether the use of CA is different in the stories and the conversations of adult FinSL signers, and whether younger (18–39) and older (50–79) adult signers use CA differently in stories and conversations. The study is based on the manual annotation of video data recorded for the Corpus of Finnish Sign Language (Corpus FinSL) and the analysis includes frequency descriptions and statistical analysis. The data show that there is a statistically significant difference in the use of CA in stories and conversations, and that in storytelling, older adults use significantly more CA than younger adults. This difference between the two age groups is particularly evident in the overt uses of CA. The work presented here points toward the need for more investigation of CA in non-narrative discourse contexts and in the language use of signers of different ages.

Harnessing Artificial Intelligence for Health Message Generation: The Folic Acid Message Engine.

BackgroundCommunication campaigns using social media can raise public awareness; however, they are difficult to sustain. A barrier is the need to generate and constantly post novel but on-topic messages, which creates a resource-intensive bottleneck.ObjectiveIn this study, we aim to harness the latest advances in artificial intelligence (AI) to build a pilot system that can generate many candidate messages, which could be used for a campaign to suggest novel, on-topic candidate messages. The issue of folic acid, a B-vitamin that helps prevent major birth defects, serves as an example; however, the system can work with other issues that could benefit from higher levels of public awareness.MethodsWe used the Generative Pretrained Transformer-2 architecture, a machine learning model trained on a large natural language corpus, and fine-tuned it using a data set of autodownloaded tweets about #folicacid. The fine-tuned model was then used as a message engine, that is, to create new messages about this topic. We conducted a web-based study to gauge how human raters evaluate AI-generated tweet messages compared with original, human-crafted messages.ResultsWe found that the Folic Acid Message Engine can easily create several hundreds of new messages that appear natural to humans. Web-based raters evaluated the clarity and quality of a human-curated sample of AI-generated messages as on par with human-generated ones. Overall, these results showed that it is feasible to use such a message engine to suggest messages for web-based campaigns that focus on promoting awareness.ConclusionsThe message engine can serve as a starting point for more sophisticated AI-guided message creation systems for health communication. Beyond the practical potential of such systems for campaigns in the age of social media, they also hold great scientific potential for the quantitative analysis of message characteristics that promote successful communication. We discuss future developments and obvious ethical challenges that need to be addressed as AI technologies for health persuasion enter the stage.

An exploration of the structural and colligational characteristics of lexical bundles in L1–L2 corpora for English language teaching

Multi-word expressions are formulaic language universals with arbitrary and idiosyncratic collocations. Their usage and mastery are required of learners of a second language in achieving naturalness. However, despite the importance of multi-word expressions to mastering a second language, their syntactic architecture and colligational possibilities have received little attention in English language teaching (ELT). This study examined lexical bundles, a type of multi-word expressions, to understand their structure and co-occurrence possibilities with other syntactic elements. It was aided by an automated frequency-driven approach using two corpora, the British component of the British National Corpus – a first language (L1) corpus – and a purpose-built Nigerian Media Discourse corpus – a second language (L2) corpus. Two items of lexical analysis software were used to extract three-word lexical bundles with a minimum of 50 frequencies per corpus. The syntactic structures of the identified lexical bundles were determined, and their in-corpus usages were analysed for their colligational characteristics. Results showed that both corpora had instances of general and genre-specific lexical bundles (LBs) with varying frequencies. Five categories of lexical bundles with different structural patterns and peculiar colligational characteristics were identified in the study. Since lexical bundles are more frequently found in both L1–L2 texts, mastering how they are used will further enhance the teaching of English as a second language. The teaching of lexical bundles as a multi-word expression is therefore recommended in ELT as a way of enhancing learners’ proficiency and naturalness in English.

"Do You See and Hear More? A Study on Telugu Perception Verbs".

Verbs of perception describe the actual perception of some entity and it is emphasized by earlier researchers that lexicon in languages is conceptually-oriented and is necessary for our daily communicative needs. In this paper, we demonstrate and explain, which among the perception verbs have the higher frequencies of all the five senses (vision, hear, smell, taste, touch) by using a Telugu corpus and self-rating task. This study shows a greater lexical differentiation when compared to studies done using English corpus and other languages. Based on our analysis-vision, followed by hear are the most commonly used verbs in daily communicative needs by the Telugu speakers as compared to touch, taste, and smell; The inconsistency in usage of other senses are not identical to the vision and hear in other studies, it may be due to sampling and methodological variations in the corpus of different language, but in common these two senses play a key role in perception verbs. The study of Telugu perception verbs may give more interesting facts and insights into the cognitive linguistics paradigm.

Norwegian in-service teachers’ perspectives on language corpora in teaching English

This study aims to explore potential reasons why the use of the tools and methods of corpus linguistics are not prevalent in English teaching in Norway, using the research question What do in-service English teachers in Norway find useful about corpora and what do they find challenging? The study provides interview data from in-service teachers, contributing to our understanding of the in-service perspective on corpora. The research design consists of teaching corpus use in seminars for in-service English teachers (featuring LancsLex, the concordancer AntConc and the OANC), integrated into a language course that is part of a further education programme, and semi-structured interviews with four of the students who took the course, during which they also interacted with Netspeak, SKELL and COCA. As with previous research, the in-service teachers found corpora particularly useful for teaching and learning vocabulary, and found challenges to use which are categorized here as usability (criticism of AntConc), IT challenges (a lack of IT skills among teachers), learner-corpus interaction challenges (the complexity of software and concordance lines for pupils; pupil uninterest in language), and lack of teacher need (mistakes being “obvious” to teachers in the lower years). The article discusses some implications of these findings.
 Keywords: English language teaching, pedagogical corpus application, corpora 
 
 

Analysis of the Full-Size Russian Corpus of Internet Drug Reviews with Complex NER Labeling Using Deep Learning Neural Networks and Language Models

The paper presents the full-size Russian corpus of Internet users’ reviews on medicines with complex named entity recognition (NER) labeling of pharmaceutically relevant entities. We evaluate the accuracy levels reached on this corpus by a set of advanced deep learning neural networks for extracting mentions of these entities. The corpus markup includes mentions of the following entities: medication (33,005 mentions), adverse drug reaction (1778), disease (17,403), and note (4490). Two of them—medication and disease—include a set of attributes. A part of the corpus has a coreference annotation with 1560 coreference chains in 300 documents. A multi-label model based on a language model and a set of features has been developed for recognizing entities of the presented corpus. We analyze how the choice of different model components affects the entity recognition accuracy. Those components include methods for vector representation of words, types of language models pre-trained for the Russian language, ways of text normalization, and other pre-processing methods. The sufficient size of our corpus allows us to study the effects of particularities of annotation and entity balancing. We compare our corpus to existing ones by the occurrences of entities of different types and show that balancing the corpus by the number of texts with and without adverse drug event (ADR) mentions improves the ADR recognition accuracy with no notable decline in the accuracy of detecting entities of other types. As a result, the state of the art for the pharmacological entity extraction task for the Russian language is established on a full-size labeled corpus. For the ADR entity type, the accuracy achieved is 61.1% by the F1-exact metric, which is on par with the accuracy level for other language corpora with similar characteristics and ADR representativeness. The accuracy of the coreference relation extraction evaluated on our corpus is 71%, which is higher than the results achieved on the other Russian-language corpora.

From Badger to Bubble: Oxford Children's Dictionaries and the Oxford Children's Language Corpus

From Badger to Bubble: Oxford Children's Dictionaries and the Oxford Children's Language Corpus

Investigating user group behavior patterns in digital examinations

Computer Mediated Communication (CMC), as well as Electronically Mediated Communication (EMC), is becoming a staple feature of many working environments. More recently, Computer Mediated Assessment (CMA) in the form of Testing Assessment Systems (TMS) has joined the ranks of arrangements where learning, information exchange and assessment take place in digital landscapes, where on top of deploying assessment, they constitute a meeting ground for teachers, assessors, and students/learners. Any such system has the potential to offer numerous other opportunities including amongst others a training platform, a system for creating learner portfolios, a vetting system for rating scale construction or a multimedia annotated corpus of learner language. Additionally, information stored in a TMS may be used for research purposes, collaborative arrangements for professional development, vetting and training solutions. This paper presents what has emerged from developing procedures and functionalities in digital exams, and what may appear as attractive opportunities offering insights into the functioning of digital assessment and digital testing environments. While, admittedly, functionalities presented here have been gradually, and while in most cases, added ad hoc to the system, over the course of eleven years since the environment was first put into operation, have come, not only to complement one another, but also to function as a coherent system. In their current state, they offer practical solutions in quality control, training, research, and systems forensics.

Лингвокогнитивные аспекты исследования метафоры глагола 煎 (жарить) в китайском языке / Linguo-cognitive aspects of studying the metaphor of the verb 煎 (fry) in chinese language

The article presents the metaphorical mappings of the verb 煎 (fry) in Chinese language, and attempts to analyze the mechanismsof emergenceof the corresponding metaphors. Applying the formula of interaction between the cognitive structure of the source domain and the cognitive structure of the target domain (A is B), developed by J. Lakoff and M. Johnson in the framework of the theory of cognitive metaphor, and using materials from authoritative Chinese dictionaries and texts from the online corpus of modern Chinese language, the author reveals the main metaphors associated with the studied verb in Chinese.

Compilation, Analysis and Application of a Comprehensive Bangla Corpus KUMono

Research in Natural Language Processing (NLP) and computational linguistics highly depends on a good quality representative corpus of any specific language. Bangla is one of the most spoken languages in the world but Bangla NLP research is in its early stage of development due to the lack of quality public corpus. This article describes the detailed compilation methodology of a comprehensive monolingual Bangla corpus, KUMono. The newly developed corpus consists of more than 350 million word tokens and more than one million unique tokens from 18 major text categories of online Bangla websites. We have conducted several word-level and character-level linguistic phenomenon analyses based on empirical studies of the developed corpus. The corpus follows Zipf’s curve and hapax legomena rule. The quality of the corpus is also assessed by analyzing and comparing the inherent sparseness of the corpus with existing Bangla corpora, by analyzing the distribution of function words of the corpus and vocabulary growth rate. We have developed a Bangla article categorization application based on the KUMono corpus and received compelling results by comparing to the state-of-the-art models.

Референциальное кодирование культовых символов в паремиях разноструктурных языковых систем восточных лингвокультур

The study focuses on linguocultural and psycho-linguistic analysis of mechanisms employed to encode archetypal images within the process of verbal representation. The most representative in this case are considered paroemiological units containing cult components or mythologems, which are the result of the top level of categorization and associative abstraction. For a case of typical languages featuring different structures employed to present various bases of reflexive reality categorization, we have opted for Turkic and Sino-Tibetan (Kazakh and Chinese) languages. The relevance of the article relies on an attempt taken in order to fill the gap in the comparative study of verbalization processes for archetypal cult semantemes in linguocultures that are based on nature-, anthropo-, as well as sociocentric positions of mythologization. The basis of a consistent interlanguage linguocultural comparison involving Chinese and Kazakh paroemias, helps reveal the interdependence of referential encoding based on the figurative and emotive aspects of metaphorical transference, as well as the initial historical-cultural, areal and existential components of the worldview. The aim of this study is to determine the typological similarity and ethno-cultural specifics of the actualization mechanisms for cult and mythological components in language systems of different structures.The authors conclude that the specific morphotype of Sino-Tibetan languages determines socio and anthropocentric models of encoding cult and mythological components, whereas the dissected type of categorizing The paroemiological units were selected subject to a stratified sampling in the corpus of national languages, in view of the frequency and accentuation criterion of the structure mythonymic elements. The comprehensive methodology of the study, including comparative structural, typological, historical and etymological methods, is enhanced through separate operations of hermeneutic and interpretive analysis, which contributes to the disclosure of consituative associates. The authors conclude that the specific morphotype of Sino-Tibetan languages determines socio- and anthropocentric models of encoding cult and mythological components, whereas the dissected type of categorizing agglutinative Turkic languages points at the primacy of nature-centric models of metaphorization, despite the historical, cultural and existential similarity.

КОРЕФЕРЕНТНІСТЬ У ПОЕТИЧНИХ ТВОРАХ ЛЕСІ УКРАЇНКИ

The category of coreference plays an important role in the creation of texts, as it is based on fundamental principles of text organization. The linearity of a text and the non-linearity of a situation described in the text are closely related to the mention of elements of a situation with another nomination of this object. Therefore, chains of coreferential names appear in the text and, having a binary relation, partly cause the emergence of a secondary nomination, which is an anaphoric relation with the primary nomination. The purpose of this article is to study the category of coreference in the poetic speech of Lesia Ukrainka based on the corpus of the Ukrainian language represented on mova.info. The object of the research is the idiostyle of Lesia Ukrainka’s poetry, and the subject is the functioning of the category of coreference. The categorical essence of coreference is characterized, the types of representatives of real objects, their functional and stylistic varieties and typical objects of representation in a poetic text are determined, and the most frequent structures are considered. Software for working with coreference has been developed and implemented. It has a user-friendly interface, which allows searching, sort, and doing the quantitative processing of the collected information according to the needs of researchers. On the material of 153 texts, 1520 referentially identical pairs are established. The coreferentiality index and the index of coreference coverage of the text proposed in this article helped to quantitatively assess the saturation of the text with the category of coreference. In addition, the classification of coreferential relations is composed: 1) identity; 2) collection gap; 3) part-whole; 4) predicative identity; 5) metonymy; 6) association. A dependency tree of a sentence made it possible to determine the location of coreferential pairs in the sentence, the position relative to each other, the belonging of the segment with the coreferent or the referent to the complicators of the sentence, i.e., a subordinate clause or phrases. This representation also helped to establish the type of syntactic relationship between the coreferent and the referent, the type of syntactic-semantic relationship between them, the presence of referent-dependent words, the intersection of the coreferent and the referent in gender and number. Automatic extraction of coreferential pairs from dependency trees allowed building intersentential coreference.

НАЙНОВІТНІШІ ЕВОЛЮЦІЙНІ ПРОЦЕСИ У ФРАЗЕОЛОГІЧНОМУ ЛАДІ УКРАЇНСЬКОЇ МОВИ

The article aims at the systemic analysis of the newest changes (in 2021) that have happened in the phraseological corpus of the Ukrainian language. In particular, the focus is set on the appearance of new set expressions on the basis of traditional and occasional patterns; on the formation of mottos that undergo phraseologization; on the changes in easily identified (with minimal transformations) and allusive (with maximum transformations) sayings, proverbs, famous people's, masters' of the artistic word apt expressions, slogans cultivated in different historical periods and their transformations into anti-proverbs, anti-sayings, anti-expressions, anti-slogans that acquire new senses which do not coincide with the background knowledge in different parameters and often contradict it. The intra- and extra-lingual analysis allowed thorough characterization of the phraseological model "numeral тисяча + noun – a synchronic precedent name", which is a formal-grammatical plagiarism. The study logically reveals the proprietary function of this model (Юлина тисяча, Вітина тисяча, Петрова тисяча, Вовина тисяча), its meaningful hypero-hyponymic specificity, the forming of which is determined by the influence of extra-lingual (political-financial) reasons. The research showed iso-functional relations constituted by onyms Тимошенко, Янукович, Порошенко, Зеленський (sound combination-substituter Зе graphically modified in various ways), substantivized adjective Зелений, adjective "зелений" with secondary occasional possessive semantics, as well as possible periphrastic representations, variational modifications that appear as a result of updating, reducing, expanding the lexis structure and various methods of intraparemic and interparemic contamination. The article reveals the cognitive mechanism of Coronavirus phraseology; shows its connotative potential, firstly, pejorative; identifies its means of explication: the use of an adjective derived from an onym in diminutive and augmentative-pejorative forms, decapitalization of a component of an antroponym, the use of inverted commas, secondary nomination, and the use of jargonisms. This background makes the language law of expressiveness and creativity of the whole nation and its representatives brighter and allows explicating the relation between possible transformational processes in the corpus of paremias and language game and its typical comic effects – mockery, irony, sarcasm.