Writer-reader Interaction in Written Discourse: A Parallel Corpus-based Investigation of English-Persian Translation of Metadiscourse Features in Legal and Political Texts

Authors

  • Mehrdad Vasheghani Farahani

DOI:

https://doi.org/10.14456/nvts.2021.18

Keywords:

Translation Studies, writer-reader interaction in written discourse, metadiscourse features, metadiscourse features in translation, corpus-based translation studies, legal and political texts, Persian and English languages

Abstract

Metadiscourse features are the elements by which interactions between writer and reader and between speaker and audience are constructed, established, and directed. These features as the writer- reader interaction (s) belong to the second level of meaning and have nothing to do with the content of the message as the first level of meaning. Considering this dichotomy, the objective of this parallel corpus-based study was to quantitively and qualitatively compare the distribution of metadiscourse features as well as analyze the writer-reader interaction of written discourse in translating legal and political texts from English into Persian. For this reason, a wide range of different steps were taken. First, for classifying and analyzing metadiscourse features, Hyland’s model (2005), which is divided into two sub-categories of interactive and interactional metadiscourse features was used. Then, for analyzing and extracting metadiscourse features, Sketch engine corpus software was used as the corpus tool. Next, for extracting and analyzing the metadiscourse features in the corpus of the study, the ELRA (European Language Resources Association) parallel corpus as well as a second do it yourself (DIY) parallel corpus were used as the data gathering sources. These two corpora were combined into one corpus in order to compile a unified, large, balanced, representative, and uni-directional parallel corpus of English and Persian language pairs in legal and political texts aligned mostly at sentence and paragraph levels. In line with the parallel corpus, a second Persian monolingual corpus was created which functioned as a reference corpus. This monolingual corpus was created so that the metadiscourse features of the English corpus as well as their translations into Persian could be compared with those of the non- translated original texts in the Persian language.

Once the corpora were compiled and ready for analysis, first the Persian monolingual reference corpus was analyzed manually and scrutinized line by line in order to detect any single token of metadiscourse features. After that, their distribution and frequency were analyzed and calculated by using Sketch engine corpus software. The tokens of metadiscourse features were classified based on Hyland’s model. Then, the same process was performed for the English-Persian parallel corpus. In the final stage, the instances of metadiscourse features were compared in a three-dimensional model. The metadiscourse features of the Persian language (reference corpus) were analyzed to see how they were used in Persian. They were then compared to those found in the parallel English into Persian corpus to see how they were used and distributed in English texts and in their Persian translations. In the final stage, metadiscourse features of the English-Persian corpus were compared with those of the Persian monolingual corpus to see how these features were used in the English language, in Persian translated texts and in Persian non-translated original texts.

The results of this comparison revealed that in terms of type-token ratio, the Persian language ratio was 77:15 which means that the Persian monolingual corpus had a relatively high level of linguistic complexity. In addition, the type-token ratio of the English corpus was 89:19 which was more than that of the Persian monolingual corpus. This ratio showed that the English corpus had a lower number of repetitions and a higher lexical density when compared to the Persian monolingual corpus. However, the type-token ratio of the Persian translations was 91:12, which means that the Persian language like the English corpus and unlike the Persian monolingual corpus had a high level of lexical density and lower number of repetitions.

In terms of the distribution of metadiscourse features, the quantitative results of this study revealed that the English corpus had more tokens of metadiscourse features as compared to that of the Persian translations. In addition, the corpora of both English and Persian had more interactive metadiscourse features than interactional metadiscourse features. This inclination towards interactive metadiscourse features was similar to the Persian monolingual corpus. Moreover, the quantitative analysis showed that in both corpora and in the interactive category, transitions, frame markers and code glasses were the most frequently used metadiscourse features, while in the interactional category, boosters, self-mentions, and hedges were the most used metadiscourse features.

The qualitative analysis of the concordance lines, however, demonstrated, that in translation from English into Persian, there were traces of implicit change, deemphasis change, disinformation change, and invisibility change (four types of changes) as not all of the metadiscourse features were translated. Finally, it can be said that although both English and Persian languages belong to the writer-oriented category of languages, due to the heterogeneous distribution of metadiscourse features from English to Persian, the interaction between writer and reader in the process of translation from English into Persian in legal and political texts changed, meaning that the translations into the Persian language were closer to a more reader-oriented language. The results of this study are hoped to be found useful for researchers and scholars in such fields as translation studies, corpus-based translation studies, corpus linguistics, text analysis as well as contrastive linguistics.

Author Biography

  • Mehrdad Vasheghani Farahani

    Leipzig University, GERMANY

Published

2023-04-20

Issue

Section

Abstracts of PhD Theses

Similar Articles

1-10 of 406

You may also start an advanced similarity search for this article.