Parallel Text Corpora Research Articles

Automatic extraction of named entities in the Chinese-Russian corpus of parallel and comparable texts on political topics

Automatic extraction of named entities in the Chinese-Russian corpus of parallel and comparable texts on political topics

Same yet different

Abstract The article compares the distributional differences in the use of the partitive object cases in Estonian and Finnish via multifactorial modeling in contrastive research using the European Parliament parallel text corpus. Based on previous contrastive research on Finnic, we expected the principles of object case marking to be similar for Estonian and Finnish (confirmed), and the partitive objects to be more numerous in Estonian than in Finnish (not confirmed, as countable objects with scalar verbs proved less likely to be partitive in Estonian). We hypothesized that multifactorial modeling in contrastive research design could help identify the causes for variation and unfold subtle differences between related language systems. Since preferences related to grammatical voice and constituent order revealed subtle differences between the systems, this hypothesis was confirmed.

HYBRID DISTANCE-STATISTICAL-BASED PHRASE ALIGNMENT FOR ANALYZING PARALLEL TEXTS IN STANDARD MALAY AND MALAY DIALECTS

Parallel texts corpora are essential resources in linguistics and natural language processing, especially in translation and multilingual information retrieval. The publicly available parallel text corpora are limited to certain genres, types and domains. Furthermore, the parallel dialect text is scarce, even though they are important in the analysis and study of a dialect. Collecting parallel dialect text is challenging because dialects typically appear in the form of speech and very limited dialectic texts exist. Moreover, there is no standard orthography in most dialects. The contributions of this paper are threefold. First, the paper describes a methodology in acquiring a parallel text corpus of Standard Malay and Malay dialects, particularly Kelantan Malay and Sarawak Malay. Second, we propose a hybrid of distance based and statistical-based alignment algorithm to align words and phrases the parallel text. The results show that the precision and recall values of the proposed alignment algorithm are more than 95% and better than the state-of the-art GIZA++. Third, the alignment obtained were compared to find out the lexical similarities and differences between Standard Malay and the two studied Malay dialects, contributing valuable insights into the linguistic variations within the Malay language family.

A Corpus-Based Study on the Translation Style of Four English Versions of Fu Sheng Liu Ji

ABSTRACT Based on a parallel corpus of Chinese-English texts, this article utilizes corpus tools to examine the translation styles of four versions of Fu Sheng Liu Ji from three levels and nine variables. It is found that different translators have different perspectives and focuses when dealing with the same source text, thus presenting different translation styles. At vocabulary level, Sanders’ translation has the highest lexical density, Black’s translation has the richest lexical variation, and Pratt’s translation has the lowest lexical richness. At the sentence level, Lin’s translation fluctuated more obviously in the use of alternating long and short sentences, while the changes in Sanders’ translation are less obvious, and Pratt’s translation has the lowest average sentence length, which means this version tends to use more short sentences to translate the original text. At the level of discourse, Black’s translation has a more compact plot and a tighter contextual connection, making the translation more readable and narrative. The distinct styles of the four translations are mainly due to the difference in the translators’ aims and their historical backgrounds.

EPISTEMOLOGICAL POTENTIAL OF TARGET TEXTS: A CASE STUDY OF THE RUSSIAN-JAPANESE PARALLEL FICTION CORPUS

This paper discusses the possibility of using a parallel text corpus to study grammatical issues of the target language, and the epistemological value of target texts is argued. Parallel text corpora are more often used in their comparative aspect, even though they have a significant content and epistemological potential for the systematization of linguistic knowledge about the target language as well. This work, carried out as a part of a general study of Japanese agentivity, describes an attempt to research the corpus material to identify recurrent patterns in the target language and to determine the distribution regularities of agentive features in Japanese. This paper was carried out using the material from the author’s parallel Russian-Japanese corpus of fiction texts. The main criterion used to collect this dataset is the markedness of sentences with inanimate subjects, which is imposed by such cognitive and cultural factors as the linguistic and social subject-object hierarchy: human beings, as the highest creatures on the animacy scale, could not be affected by an inanimate subject. The set of grammatically non-identical components of the syntactic structure of the source and target texts identified in the material of the corpus became the starting point for a detailed analysis of the mechanism for distributing syntactic and semantic roles in a Japanese sentence.

ВІДТВОРЕННЯ ЗАЙМЕННИКІВ ЯК МАРКЕРІВ НЕОЗНАЧЕНОСТІ В ДВОМОВНОМУ ПАРАЛЕЛЬНОМУ КОРПУСІ: ЗІСТАВНИЙ АСПЕКТ

The article continues a series of publications devoted to the application of a corpus method to the study of translation transformations from the contrastive perspective. The author analyzes indefinite pronouns in the parallel corpus of texts (English-Ukrainian language pair), and finds out the specifics of strategies for their reproduction in Ukrainian. The parallel corpora contain a palette of strategies used by translators, as well as repertoires of translation equivalents with examples of their contextual use, thus providing a valuable resource for contrastive cross-linguistic studies. Pronouns are classified according to a systemic structural grammatical features, corresponding morphological and functional, specified person, object, feature, or their absence. The study proved that the variability of translation of the linguistic units under analysis is usually determined by the taxonomy of indefinite pronouns (semantic and structural aspects), as well as their functionality traits within a particular text fragment and context in general. Among a scope of interlingual transformations in the parallel corpus, the study revealed that two translation transformations prevailed: logical/semantic development and differentiation of meaning. Interstingly, the compensation of losses method was not used at all, which can be explained by indefinite pronouns’ isomorphic features – presence of similar equivalents in the Ukrainian language to express the category of uncertainty. Motivated exclusively by linguistic factors, indefinite pronouns demonstrate a high level of intertranslatability, do not require additional contextual clarification and can be compensated by other translation transformations.

SPEEDSER: A Possibilistic System for Query Disambiguation, Expansion and Translation

We design, implement and assess in this paper a new architecture of a possibilistic mono- and cross-language information retrieval (IR/CLIR) system. The latter is useful to experiment query disambiguation, expansion and translation processes in both IR and CLIR frameworks. We take advantage of possibility theory to overcome the problems of query disambiguation and expansion in an uncertain and imprecise IR/CLIR context. We investigate the impact of combining possibilistic query disambiguation with expansion on IR/CLIR efficiency. A co-occurrence graph representation is exploited to quantify the similarity between query terms and their semantically close words (expansion task) or between query terms and their possible meanings (disambiguation task). We extend the possibilistic mono-language query disambiguation approach to a cross-language framework. We conduct a set of experiments using ROMANSEVAL data collection, the French–English parallel text corpus Europarl and the CLEF-2003 French–English IR/CLIR test collection. Results highlight some statistically significant improvements of our possibilistic approaches when compared to some state-of-the-art IR/CLIR works.

THE TERM COMBINATION AND THE METAPHOR IN THE OFFICIAL BUSINESS DOCUMENT: COGNITIVE ASPECT

THE TERM COMBINATION AND THE METAPHOR IN THE OFFICIAL BUSINESS DOCUMENT: COGNITIVE ASPECT

Morphology & word sense disambiguation embedded multimodal neural machine translation system between Sanskrit and Malayalam

Machine translation is the process of conversion from one language to another using computers. In spite of the existence of several traditional machine translation systems like rule-based, statistical etc., the development of neural machine translation has significantly improved translation quality. However, the morphological richness and agglutinative nature of Indian languages hampers the usage of neural machine translation for translation among Indian languages despite its potential to overcome the difficulties posed by conventional translation systems. Hence the combination of neural machine translation with morphology, part of speech tagger and word sense disambiguation is proposed to address these challenges and adapt it for bidirectional translation between Sanskrit and Malayalam. Conventional neural machine translations are unimodal systems that utilize textual data for translation. Translation quality of conventional neural machine translation is enhanced with a novel multimodal bidirectional neural machine translation system with text & speech as the two modalities based on parallel text corpus and speech database. The best performance score is obtained for two level fusion, the feature vectors extracted from speech signals using different transforms (Wavelet transform, 1-D sequency mapped real transform & 1-D GCD based mapped real transform) are fused and which is further fused with a context vector from the text modality. Automatic and manual review methods are employed to assess quality of translation. Findings from the experiments reveal that addition of speech modality improved the overall quality of translation. BLEU scores of 43.89 and 42.72 are obtained for Sanskrit to Malayalam and Malayalam to Sanskrit multimodal translation respectively.

C-E Translation of Scientific Paper Abstracts’ Academic Vocabulary Based on a Contrastive Analysis of Parallel Texts —Take Journals in Aviation Area as Example

Abstract is an important part of the academic paper, which can attract readers and provide information. However, there are many problems in the abstract writing of scientific papers. This study will analyze 32 abstracts, to study the characteristic of English abstracts written by Chinese researchers, and compare it with those written by native speakers through establishing parallel text corpus. Then, the study provided effective translation strategies and writing reference for researchers.

A MODEL OF THE ENGLISH-UKRAINIAN SUB-CORPUS OF TEXTS OF NATO, UN AND WTO OFFICIAL AND BUSINESS DOCUMENTS

A MODEL OF THE ENGLISH-UKRAINIAN SUB-CORPUS OF TEXTS OF NATO, UN AND WTO OFFICIAL AND BUSINESS DOCUMENTS

БЪЛГАРСКИТЕ ДЕЗИДЕРАТИВНИ ГЛАГОЛИ И ТЕХНИТЕ УКРАИНСКИ СЪОТВЕТСТВИЯ В ПАРАЛЕЛЕН КОРПУС / BULGARIAN DESIDERATIVE VERBS AND THEIR UKRAINIAN COUNTERPARTS IN A PARALLEL CORPUS

All uses of Bulgarian desiderative verbs of the type «glagoli» ‹ми› се ‘feel like ‹verb›ing’ (e. g., спи ми се ‘be sleepy’, работи ми се ‘feel like working, be in a mood for work’) and their prefixed phase derivatives were excerpted from a parallel corpus of Bulgarian and Ukrainian fictional texts. In the corresponding Ukrainian sentences, the specific semantics of the construction is most often expressed either by an independent lexical item – the verb хотітися ‘want, feel like’ or хотіти ‘want’ – or by a combination of a reflexive verb and an optional dative experiencer, as in (мені) працюється ‘I want to work, feel like working; I work eagerly; my work is going well’. The latter counterpart is of particular interest due to its structural similarity to the Bulgarian desiderative verbs. The comparison shows that the constructions in the two closely related languages are considerably different in meaning (which is the reason why this correspondence is only the third most frequent) and that although they have a lot in common (as attested by their lexical diversity), translators tend to overestimate the degree of similarity between them. Keywords: Bulgarian language, desideratives, dispositional constructions, reflexive–dative constructions, Ukrainian language

Prefabricated orality in theatre translations: An overview based on an English–Spanish parallel corpus

Abstract Prefabricated orality is an inherent trait of dramatic texts, given their written-to-be-spoken nature. From the perspective of translation, it has been observed in the literature that rendering this specific mode of discourse into another language poses a major challenge (Baños & Chaume, 2009). However, scant attention has thus far been paid to how prefabricated orality is realised linguistically in translations of theatre plays. This article sets out to offer an overview of syntactic and lexical-semantic features that mirror spoken discourse in a parallel corpus of theatre texts. More specifically, by drawing on an analytical framework proposed for audiovisual texts (Baños, 2014: 414), it aims (1) to verify the presence and incidence of orality markers in the original plays and their translations, and (2) to identify tendencies in translation techniques. Our findings reveal that playwrights and translators resort to a wide range of linguistic features typical of spoken discourse, especially vocatives, repetitions, discourse markers, intensifiers and deixis. In addition, results bring to light the use of different translation options and a possible compensation strategy.

Selectives (“topic markers”) on subordinate clauses

AbstractThis typological study based on data from a parallel text corpus is a two-step investigation of selectives (“topic markers”). First, a set of selectives in 81 languages from all continents is compiled on the basis of their occurrence with emphatic personal pronouns in contrast constructions. In a second step, it is explored how this set of markers is used across 19 subordinate clause domains. The results indicate that, despite much crosslinguistic diversity, the distribution of selectives across subordinate clauses is strongly constrained. Selectives in subordinate clauses are distributed following a tendency scale (no strict hierarchy, but no blatant exceptions): general relative clauses rank highest followed by conditional and temporal clauses with concessive and purpose clauses ranking lowest. No postposed subordinate clauses attract selectives. It is further found that selectives tend to occur at the end of the constituent which they have scope over where there is minimal risk of scope ambiguity. Despite the frequent occurrence of selectives on conditional clauses, selectives tend not to be conditional connectives unless this happens to be their grammaticalization source.

Реалії у «Вулику» Каміло Хосе Села та його перекладах

The article applies modern digital methods of text analysis to analyse realia’s functioning in Camilo Jose Sela’s novel “Beehive” and their rendering into the Ukrainian language in two translations: Petro Sokolovskyi and Anatolii Sobutskyi in 1979 and S. Borshchevskyi in 2011. The authors review previous linguistic research on the definition of the national-specific vocabulary, its grouping by thematic principle, and analyse the works on realia in the Spanish-Ukrainian language pair. The procedure of studying the functioning of realia in the novel and finding their equivalents in the analysed translations involves the use of corpus linguistics methodology, namely, creating a corpus of parallel texts using the cloud translation management system Memsource and using the AntConc corpus manager to generate word lists and search for the necessary contexts. Considerable attention is paid to the procedure of identifying in the source (Spanish) text mono- and polylexemes with a national-specific meaning, their classification, and search for equivalents in translations of different years in Ukrainian. Based on the analysis of the source text’s corpus employing the AntConc programme, a list of realia was formed. Later, the way they are rendered into Ukrainian was analysed. Most often, realia are presented through transliteration, substitution, omission, and various types of translation.

Застосування корпусних технологій у навчанні галузевого перекладу

Corpus technologies (corpora of English and Ukrainian texts and tools for their processing) represent modern specialized discourse and facilitate searching for and comparing different units of translation, which makes them a useful tool for both practicing and trainee translators. The purpose of this article is to determine the role and place of corpus technologies in teaching specialized translation on the example of the oil and gas industry. Comparative and parallel text corpora are characterized. The paper reveals methods of applying mono- and bilingual comparative and parallel corpora and corpus managers for acquiring knowledge about genre-stylistic features of texts; developing skills to distinguish a term and determine its collocation profile and semantic preference; analyze translation techniques; translate collocations, complex noun constructions, verbal phrases, and abbreviations. Examples of relevant exercises and tasks that should be performed at the translation training stage are given. Further research should be aimed at integrating corpus-based tasks into the translation practice stage involving the implementation of a translation project.

On word order variation and informationstructure in peninsular spanish and italian ‘why’-interrogatives

In this paper, we investigate the effect of information structure on word order in Italian and Peninsular Spanish ‘why’-interrogatives, and whether these two languages differ from each other. To this end, we conducted two empirical studies. In a parallel text corpus study, we compared the frequency of the word order patterns ‘why’SV and ‘why’VS, as well as the distribution of focal and non-focal subjects in the two languages. In order to get a deeper understanding of the impact of the information structural categories focus and givenness on word order in ‘why’-interrogatives, we conducted a forced-choice experiment. The results indicate that word order is affected by focus in Italian, while it is not determined by any information structural category in Peninsular Spanish. We show that Italian and Peninsular Spanish ‘why’-interrogatives differ from each other in two ways. First, non-focal subjects occur preverbally in Italian, while they occupy the postverbal position in Peninsular Spanish. Second, Italian reveals a lower level of optionality with respect to word order patterns. Even though we find a high preference for the postverbal position in Peninsular Spanish, we argue that this limitation is related to a higher flexibility regarding word order in Peninsular Spanish than in Italian which does not allows for ‘why’VSO in contrast to Peninsular Spanish.

Sliding Window and Parallel LSTM with Attention and CNN for Sentence Alignment on Low-Resource Languages

A parallel text corpus is an important resource for building a machine translation (MT) system. Existing resources such as translated documents, bilingual dictionaries, and translated subtitles are excellent resources for constructing parallel text corpus. A sentence alignment algorithm automatically aligns source sentences and target sentences because manual sentence alignment is resource-intensive. Over the years, sentence alignment approaches have improved from sentence length heuristics to statistical lexical models to deep neural networks. Solving the alignment problem as a classification problem is interesting as classification is the core of machine learning. This paper proposes a parallel long-short-term memory with attention and convolutional neural network (parallel LSTM+Attention+CNN) for classifying two sentences as parallel or non-parallel sentences. A sliding window approach is also proposed with the classifier to align sentences in the source and target languages. The proposed approach was compared with three classifiers, namely the feedforward neural network, CNN, and bi-directional LSTM. It is also compared with the BleuAlign sentence alignment system. The classification accuracy of these models was evaluated using Malay-English parallel text corpus and UN French-English parallel text corpus. The Malay-English sentence alignment performance was then evaluated using research documents and the very challenging Classical Malay-English document. The proposed classifier obtained more than 80% accuracy in categorizing parallel/non-parallel sentences with a model built using only five thousand training parallel sentences. It has a higher sentence alignment accuracy than other baseline systems.

Identifying constitutive articles of cumulative dissertation theses by bilingual text similarity. Evaluation of similarity methods on a new short text task

Abstract Cumulative dissertations are doctoral theses comprised of multiple published articles. For studies of publication activity and citation impact of early career researchers, it is important to identify these articles and link them to their associated theses. Using a new benchmark data set, this paper reports on experiments of measuring the bilingual textual similarity between, on the one hand, titles and keywords of doctoral theses, and, on the other hand, articles’ titles and abstracts. The tested methods are cosine similarity and L1 distance in the Vector Space Model (VSM) as baselines, the language-indifferent methods Latent Semantic Analysis (LSA) and trigram similarity, and the language-aware methods fastText and Random Indexing (RI). LSA and RI, two supervised methods, were trained on a purposively collected bilingual scientific parallel text corpus. The results show that the VSM baselines and the RI method perform best but that the VSM method is unsuitable for cross-language similarity due to its inherent monolingual bias.

STAGES OF CREATING PARALLEL CORPUS OF ENGLISH-UZBEK SIMILES

In recent years, corpus linguistics has been mentioned in the scientific literature as the main tool for the elaboration of dictionaries and grammar manuals, the creation of corpus analysis, the practical use of corpus, as well as the statistical study of linguistic analysis using corpus. Articles, textbooks, manuals on the creation of general and special electronic corpus have been published, but as the Uzbek language is a relatively new field of corpus linguistics, there is a need for research on some issues. The development of corpus linguistics and the increasing focus on statistical methods of processing linguistic materials have led to the development of a number of techniques related to the usage of parallel or similar texts in different languages. This article is devoted to the stages of creating bilingual parallel corpus and explains the peculiarities of creating parallel corpus of texts. The creation of parallel corpus also has its own research stages. Statistical research in linguistics has a long history, especially since the advent of computers in the fifties of the twentieth century, research in this area has grown rapidly. The purpose of the study is to analyze the scientific views on the creation of the linguistic supply of the program of translation of similes on the basis of parallel texts (English-Uzbek, Uzbek-English), to study the linguistic basis and to study the lexical-semantic relationship of similes in bilingual vocabulary. This article deals with the creation of a parallel corpus of literary texts with English and Uzbek similes, analysis of lexical-semantic relations of similes, formation of a database of text segments based on translation alternatives of similes, scientific proof of linguistic models of English similes in translation and the effect of mentality is scientifically substantiated in the example of the translation of similes. A bilingual simile dictionary also plays an important role in creating a parallel corpus of similes. There is a great need for electronic dictionaries in the field of literary translation. The creation of bilingual and multilingual dictionaries, which embody the subtle nuances of the word, serves as a very important resource for translators, writers, poets, and language users.