CC BY
3916. Марфша Ж. Мовна асощацш як поняття психолшгвютики та лшгвостилютики / Ж. Марфiна // Вюник Львiвського унiверситету. Серiя «Фшоло-пчна». - 2004. - Вип. 34. - Ч. 2. - С. 366-372.
17. Недашшвська Т. С. Семантична конструк-цiя асощативного поля (на матерiалi асоцiативного експерименту) / Т. С. Недашк1вська // Психолшгвь стика: Зб. наукових праць ДВНЗ «Переяслав-Хме-льницький педагогiчний ушверситет iменi Григорiя Сковороди». - Переяслав-Хмельницький : ПП «СКД», 2010. - Вип. 7. - С. 126-134.
18. Попова З. Д. Лексическая система языка / З. Д. Попова, И. А. Стернин. - Воронеж : Изд-во ВГУ, 1984. - 148 с.
19. Попова Н. Б. Концептуальное представление семантического пространства многозначного слова / Н. Б. Попова // Вестник Челябинского государственного университета. Филология. Искусствоведение. - Вып. 52. - 2011. - С. 114-117.
20. Соссюр де Ф. Курс общей лингвистики / ред. Ш. Балли, А. Сеше; пер. с франц. А. Сухотина. - Екатеринбург : Изд-во Уральского ун-та, 1999. -432 с.
21. Соссюр Ф. Труды по языкознанию / Сос-сюр Ф. - М. : Прогресс, 1977. - 696 с.
22. Уфимцева А. А. Опыт изучения лексики как системы (на материале английского языка) / А. А. Уфимцева. - М. : Издательство Академии наук СССР, 1962. - 287 с.
23. Щур Г. С. Теория поля в лингвистике / Г. С. Щур. - М. : Наука, 1974. - 256 с.
24. Atkinson, K. (2016). Life After Life. Retrieved from Royallib.com.
25. Boyne, J. (2011). The Absolutist. Retrieved from Royallib.com.
26. Boyne, J. (2015). The Boy at the Top of the Mountain. Retrieved from RuLit.me.
27. Boyne, J. (2006). The Boy in the Striped Pyjamas. Retrieved from Royallib.com.
28. Deese J. On the Structure of Associative Meaning / J. Deese // Psychological Review. - 1962. -Vol. 69 (3). - P. 161-175.
29. Harris, J. (2001). Five Quarters of the Orange. Retrieved from Royallib.com.
30. Rommetveit R. Words, Meaning, and Messages: Theory and Experiments in Psycholinguistics / R. Rommetveit. - New York : Academic Press, 2014. - 336 p.
31. Soloshenko O. D. Modules on English lexicology (for independent work and distant learning) / O. D. Soloshenko, Y. A. Zavhorodniev, M. M. Latyk. - Lviv, 2008. - 290 p.
32. Waters, S. (2009). The Little Stranger. Retrieved from RuLit.net.
33. Yule G. The Study of Language / George Yule. - 4th edition. - New York : Cambridge University Press, 2010. - 319 p.
PARALLEL CORPORA IN TRANSLATION STUDIES
Umerova M. V.
Ph.D. in Linguistics, Associate Professor National Research University Higher School of Economics
ABSTRACT
The present paper discusses the use of parallel corpus in translation, the norm of translation, teaching and learning material, translation model evaluation using corpus-based tools. Translation memory (TM) systems within Computer-Aided Translation (CAT) courses are incorporated in translator's professional training. TM systems are part of everyday working environment, even though professional translators would like to learn more about the potential of corpora for translation. Overall, the practical use of corpora and corpus technology is of paramount importance. It is necessary to assess to which extent corpora are valuable for translation quality in professional and academic environments
Keywords: translation technology, corpus, translator training, universals of translation, translation norms, source text, target text
Corpus and corpus linguistics have been gradually developed. Application of corpus has broadened the horizon of applied linguistics and provided a new model for translation and foreign language teaching. Parallel corpus provides a powerful tool for translation studies and promotes translation teaching. Since the 1990s, Mona Baker [1], Gideon Toury, Miram Shle-singer, Kristen Malmkjaer [7] and other translators have been describing the nature of translation by using corpus. In 1995, the research center for translation of UMIST headed by Mona Baker founded the first comparable corpus - Translational English Corpus (TEC). At present, many large-scale parallel corpora are being established on an international scale. There is an opportunity not only to compare similarities and differences between two different languages, but also to carry out
extensive translation teaching and research. The paper discusses positive effects of parallel corpus on universals of translation, the norm of the translation, the style of a translator.
The use of TM systems has become fully integrated in translator training programmes. Translation didactics acknowledged that corpus use forms part of wider translation, the use of corpus processing tools (concordancers) to search through corpora as part of translation resource. The use of TM systems dominates in professional translator's environment, the role of corpora and concordancers as a translation resource needs to be disseminated in translation profession [9, 131].
Translation corpus has a guiding significance for exploration of universal laws and rules of translation,
exploration and verification of universality of a target text. Mona Baker and others propose using parallel corpus to study features of translation and confirm the following hypothesis of the universality of translation: simplification, which means that translators simplify the language and information unconsciously; explication, which means that translators tend to make the content clearer in translation; normalization, which means that translation tends to be intermediate rather than to the periphery [1, 227].
Translation norms can be reliably analysed by using corpus and computer technologies. In specific social, historical and cultural environment, translation has some regular features, which is Mona Baker's translation norms or Gideon Toury's operational norms. Accumulation of general characteristics of translation language is the foundation of translation concept construction and translation norms: it is clearly seen that norms of translation behavior can be different due to various social cultures. Through corpus, translation strategies and language differences between female and male translators can be studied, certain language features and translation strategies can be referred to some specific translation categories (novels, news, speech).
Parallel Corpus produces a paramount influence on the research of the style of translation and translator. To avoid subjective assumption, research of the style must be supported by quantified actual corpus: certain expression form and language performance are preference of translators and appear repeatedly, relying on experience or observation. Through automatic labeling and processing of computer software, multifaceted comparison and research of lexical density, sentence length, word frequency, collocation mode and specific word frequency can be made. It is essential to reveal translator's unique language habits, language industry preference, usage of special syntactic structure and punctuation, etc. For example, Mona Baker made a comparative study of translation styles of famous English translators Bush and Clark by corpus. The source text which was translated by Bush comes from three writers whose languages belongs to different variants of Spanish (novel and biography genres). The source text which was translated by Clark comes from two Arab writers with different styles. Analysis of word frequency, average sentence length, stylistic variation and use of various word forms of say, Mona Baker convincingly proved the existence of the so-called "translator's brand" [1, 229].
Corpora can serve as tools to evaluate translation quality, for instance to identify the extent to which student translations differ from professional translations their usefulness may also be considered with regard to their added value on the translation process itself [3, 27]. Lynne Bowker conducted a noteworthy study of usefulness and impact of corpora on the quality of students' translations. Her pilot study is aimed to reveal how a specialized monolingual native language corpus could be compared with conventional resources such as dictionaries. Fourteen English native-speaker students, enrolled in an Applied Computational Linguistics programme, translated two French texts into English using conventional resources on the one hand and a corpus
together with a concordancer on the other. The results showed that corpora help enhance students' translations by providing information missing from dictionaries, especially regarding terminology choice and idiomatic expressions [2, 31].
A growing number of studies have tackled the usefulness of corpora at an experimental level by collecting data produced by students specializing in translation. The efficiency of corpus use for students during terminology processing in the field of earth sciences was investigated and it was found out that "besides an obvious and expected positive outcome on the validation and translation of terms (...) there is an interesting positive influence of corpora in the translation process on elements other than those related to terminology, such as collocations and various genre- and discourse-based features [5, 7]. In particular, the added value of parallel corpora through specific use of translation memories versus bilingual concordance search within CAT tools and external sources was measured. All studies proved the added value of corpora on terminology, for retrieving idiomatic expressions and identifying discourse-based features pinpointed the major benefit of concordance search and external sources compared with translation memories.
Actually, conducting such experimental studies involves considering issues that particularly relate to: 1) defining among the different stages of the translation process which stage(s) to incorporate in the study i.e. understanding of the subject field, lexical, terminological, phraseological and syntactic accuracy - in relation with conceptual and linguistic information retrieval -as well as stylistic fluency and appropriacy; 2) selecting various corpora to be assessed and to be compared (monolingual, comparable, parallel corpora), online corpora, online translation databases (e.g. Reverso Context) as well as Google search; this issue is of major importance as it has an impact on the corpora and corpus processing tools to integrate into a translation classroom based on their efficiency on student translations; 3) determining the criteria to consider in order to measure usefulness of the corpora under study (for instance, which linguistic criteria).
Conducting such experimental studies also involves devising an adequate methodology at the experimental level that includes determining which instructions students are given - especially in terms of providing feedback on their search and on their use of electronic resources; the linguistic items or portion of text to evaluate.
An issue of major relevance relates to determining the criteria to consider in order to measure the usefulness of corpora. Studies rely either on different levels of translation acceptability (three levels termed as accepted, to be revised and unacceptable or yes/ no/ partly - criteria for a correct translation) or on an error-annotated corpus. Analyzing translations consists in comparing translations taking into account the criteria mentioned. However, the so-called "blind faith" in the corpus needs to be tackled in the corpus-based translation classroom where the emphasis should be put on raising student linguistic awareness since using a corpus as a translation resource which requires interpreting
the data. In this regard, it is necessary to suggest an introductory course on corpus linguistics outside translation classroom in order to help students realize the potential of corpora.
Bilingual parallel corpus is a reference tool and a working platform for translation teaching. It is an example of abundant and various bilingual inter-translation of a certain word and phrase, common bilingual inter-translation with a special structure, multi-translation material which can be extracted randomly. Taking the common structure of bilingual inter-translation retrieval as an example, there are some special patterns and structures in source and target languages, textbooks fail to demonstrate how to deal with them, however, students can get more perceptual knowledge from abundant example sentences provided by corpus. In translation, compared with literary texts, non-literary (informative) texts tend to structural correspondence. In translation teaching, students can make inter-translation retrieval for bilingual corpus and feel inter-translation context. After perceptual knowledge of students to the structure of translation is enhanced, they can get a good teaching effect with teachers' proper induction and explanation. Parallel corpus can also provide a text with multiple translations. With parallel corpus platform and tools, translation teachers can store different versions of translation that can be used at any time for comparative study. When students read a translation, they cannot find its characteristics, but after comparing many versions they may discover new things so as to inspire and help students. Moreover, corpus retrieval tool can obtain statistics quickly and accurately and provide a reliable statistical description for teaching practice. For example, the token ratio to observe the richness of words of different translations based on the corpus of similar size: the larger the numerical value of token ratio and average word length, the richer the vocabulary, and the greater the complexity of the vocabulary.
A self-built person is able to build a parallel corpus which is applied to translation teaching and constantly expand and develop it in a certain period as long as there is some basic computer knowledge, such as text editing, format conversion, internal code conversion and master the positioning retrieval software, which manages the retrieval corpus, such as Concordance, for instance, and the software that is suitable for teaching display. Only when a teacher sets up positive thought and tries to approve the use of corpus in translation teaching, translation teaching can be improved. Some people even regard Advances in Computer Science Research, volume 73 360 bilingual parallel corpus as an interpreter training workshop to enhance the students' ability to understand the original text and translate them in a more fluent target language.
The technology at translator's disposal in his/her working environment includes TM systems and other CAT tools, machine translation tools as well as web-based resources such as dictionaries and termbanks, portals, directories of dictionaries and glossaries, specialized search and meta search engines as well as corpora. While translation technologies have proven to be indispensable for professional translators, not all of
them have been used to the same extent and surveys on the use of translation technology by professionals have shown that not all the existing technology varieties were successfully adopted by professionals [2, 43].
Most importantly, not all surveys have integrated the use of corpora by translators - the common ground of all surveys being TM systems - and while "in today's market, the use of technology by translators is no longer a luxury but a necessity if they are to meet rising market demands for the quick delivery of high-quality texts in many languages" [2, 45], there seems to be little or no demand from the translation market requiring translators to use corpora and concordancers in particular. There is no particular pressure from the market to train translators to use corpora which does not encourage the integration of corpora in translator training and may be counterproductive for teachers involved in the training of future translators.
However, a number of surveys has set out to uncover the use of corpora and corpus technology by professional translators. Bowker carried out a literature survey of publications produced by Canadian translators' associations and analyzed a database of job advertisements to "determine how many Canadian employers are seeking candidates who are familiar with corpus-based resources" [2, 41]. Her major results show that Canadian professional translators are more interested in TM systems than in corpora. As for the job advertisements for translation-related positions, 62.3% stated some general degree of computer literacy as required and a mere 15.1% were aimed at candidates with knowledge of specialized translation software. However, the advertisements never mentioned the word "corpora" as part of their requirements.
A few years later, a survey on the use of corpora by professional translators was conducted within the Mellange project [8, 7] with a total of 740 respondents - the vast majority were from the UK (567) and France (125) and a limited number from Germany (25), Italy (19) and Spain (4). Overall, the data collected brought to light that corpora and concordance use were not common practice among professional translators and translators (53.4%) reviewed electronic corpora and used search facilities in word processors instead of searching them with software. The survey showed that professional translators found it time-consuming to build corpora, however, 83.7% had major interest in the potential of corpus use.
Surveys of the Spanish market provided data on the use of corpora as translation resources from 526 Spanish translators of various ages and varying amounts of experience. The overall results suggest that integrating skills related to corpus use when designing translation courses combined with literature on building and exploiting corpora "are beginning to bear fruit, as almost 50% of respondents stated they use corpora sometimes/ often/ very often in their work" [8, 9]. Another noteworthy result is that concepts related to corpora and how to exploit them may remain poorly defined among professional translators, "or at least have rather diffuse boundaries" [8, 11]. It was found that professional translators had little explicit knowledge of
corpora. In this respect, the relationship between corpora and professional translators claim that professional translators actually resort to corpora as they look for parallel texts using search engines or search previous translations. A survey was conducted in order to assess the translators' use of corpora in Switzerland based on the gap between the scholars and the translators' point of view, it found that respondents to the survey hardly use corpora as such on a systematic basis and use Google as a mega-corpus; the survey also found that translators are rather confused when asked to define what a concordancer is. It looks like defining what a corpus is as well as defining the corpus-related concepts should be part of a core programme of any translator training institution. Finally, respondents to the survey among corpus users said they favoured specialized and multilingual corpora over general and multilingual corpora, as well as stable corpora, corpus users acknowledge the benefits of corpora and require specific training, which calls for a greater collaboration between academic and professional circles. The interest non-users have in the potential corpora offer should greatly motivate translator trainers "who are the real transmitters of this type of methodology" and foster the dissemination of corpus benefits. For instance, there is information on professional translators using corpora and concordancers but it proved fruitless as it did not seem to have any review or report on the use of corpora by translators themselves (in contrast with a number of reviews on CAT tools), nor professional association literature (online forums, for instance).
Therefore, providing both future translators and professional translators with empirical evidence on the value of corpora is very likely to serve both groups and raise their awareness of corpora on translation quality since to start using corpora is important for translators, they realize that corpora have the potential to help them find answers to questions for which there are often no clear answers in dictionaries, glossaries, Google searches and other tools and resources they are accustomed to using. Dissemination of a corpus-based approach in translator working environment will hopefully bear fruit and a widespread uptake in professional practice. Overall, the future of the usefulness of corpora in translation will continue to be shaped by empirical studies addressing the problem of translation quality which will undoubtedly contribute to enrich the field and provide answers to issues related to corpora and teaching in particular.
Translation corpus provides a new tool for translation studies, expands the research scope of translation and provides the data of quantitative analysis and the support of quantification for translation studies based on qualitative research. This point is confirmed in the research of universality of translation based on corpus. In the field of translation teaching corpus can provide abundant and complete translation examples and coherent contexts, which provides an objective and convincing basis for translation teaching. Evidently, there are some problems in corpus translation research and its application. The use of corpus makes translation theorists pay more attention to the universality of translation, complicated problems in translation should not be taken as secondary.
References
1. Baker, M. Corpora in translation studies: An overview and some suggestions for future research. Target (1995): 223-243.
2. Bowker, L., McBride, C., Marshman, E. Getting more than you paid for? Considerations in integrating free and low-cost technologies into translator training programs. Redit 1 (2008): 26-47.
3. Cappelle, B., de Clercq O., de Sutter, G., Loock, R., Plevoets, K. A corpusbased, statistical approach to translation quality. Journée d'études Traduction & Qualité 2016: corpus et qualité. Université Lille 3 (2016).
4. EMT Expert Group. Competences for professional translators, experts in multilingual and multimedia communication (2009).
5. Heylen, K., Verplaetse, H. Parallel Corpora for Medical Translation Training: An Analysis of Impact On Student Performance. CULT Conference, 2629 May, Alicante (2015).
6. Krüger, R. Working with corpora in the translation classroom. Studies in second language learning and teaching (2012): 505-525.
7. Malmkjaer, K. Linguistics and the Language of Translation (2005).
8. Mellange Corpora & E-learning Survey Results. ITI Bulletin (2006).
9. Rodríguez-Ines, P. Evaluating the process and not just the product when using corpora in translator education. Corpus Use and Translating. Ed. A. Beeby, P. Rodríguez-Inés and P. Sánchez-Gijón. Amsterdam: John Benjamins (2009). 129-149.
