Languages in Contrast - Volume 20, Issue 2, 2020

This article outlines the beginnings of corpus-based contrastive studies with special reference to the development of parallel corpora that took place in Scandinavia in the early 1990s under the direction of Stig Johansson. It then discusses multilingual corpus types and methodological issues of their exploration, including the tertium comparationis for contrastive studies based on different types of corpora. Some glimpses are offered of recent developments and current trends in the field, including the widening scope of corpus-based contrastive analysis, concerning language pairs as well as the kinds of topics studied and the methods used. The paper ends by identifying and discussing some challenges for the field and indicating prospects and directions for its future.

The aims of this paper are to analyse differences in the degree of lexical variation (type/token ratio and hapax/token ratio) of reporting verbs in reporting clauses placed medially or in postposition in English, French and Czech fiction and to evaluate their consequences in translation, especially in regard to explicitation/implicitation. We expect that, in translations from a language with a low degree of lexical variation of reporting verbs into a language with a high degree of lexical variation, the frequency and the degree of explicitation will be higher than in translations involving languages less different with respect to lexical variation. The analysis, relying on data extracted from the InterCorp multilingual corpus, proposes a classification of reporting verbs based on the type and amount of information conveyed, which allows evaluating the degree of explicitation operated in translations. The results show that most shifts involve only the neutral reporting verb say/dire, replaced by a stylistically more specific synonym or by a verb explicitating information obvious from the context. This suggests that modifications of reporting verbs in translation are motivated primarily by respect for the stylistic norm of the target language and the degree of acceptability of the repetition of the neutral reporting verb.

The aim of this article is to shed new light on the use and translation of English and Swedish questions in fiction by using a combination of parallel and comparable corpus data extracted from the bidirectional English-Swedish Parallel Corpus. In particular, the study examines questions containing a question mark (QMquestions) categorised into wh-interrogatives, polar interrogatives, alternative questions, tag questions (including those with invariant tags), declarative questions, wh-fragments and non-wh-fragments. The parallel analysis shows that most QMquestion types are more often translated congruently into English than into Swedish. The focus is on types with low mutual correspondence scores: fragments, tag questions and declarative questions. The comparable analyses concern both bilingual contrasts between the original texts and monolingual contrasts between the translation and original subcorpora in both languages. The bilingual analysis aligns with several preliminary findings in the parallel analysis, e.g. the favouring of tag questions and some types of wh-fragments in English. The monolingual analysis reveals both over- and underuse in translations and points to a strong effect of source-language influence.

In many languages, conjunctive adjuncts (e.g. however, therefore) are syntactically mobile. Several corpus-based contrastive studies have shown that languages differ in the positions that they tend to prefer for conjunctive adjuncts. However, the studies available have formulated general cross-linguistic differences in placement for languages as wholes, without considering the possibility that such contrasts may be influenced by register. The objective of this paper is to investigate and compare the placement patterns of English and French conjunctive adjuncts of contrast in two written registers (viz. editorials and research articles) in order to measure the impact of register variation on the differences between these two languages. The results suggest that, although register variation plays a significant role on placement within each language system, language is a better predictor of placement than register, since cross-linguistic differences in placement between English and French are stable across communicative situations. In a second stage, the results obtained in the comparable corpus study are complemented with the analysis of translation data, with a view to assessing the translators’ degree of awareness of the inherent word order preferences of the target language. The study is grounded in the framework of Systemic Functional Linguistics and relies on the notions of Theme and Rheme to describe conjunctive adjunct placement.

This paper explores both comparable and translation data from the fiction part of the English-Norwegian Parallel Corpus (ENPC) in a new way. Rather than studying fiction as a unified register, we investigate to what extent fiction can be seen to contain (at least) two distinct registers – dialogue and narrative – and to what extent this may have implications for contrastive studies based on a corpus such as the ENPC. Token counts show that, although the texts are predominantly narrative in nature, the Norwegian texts are even more so than the English ones. On the basis of word lists, two items proportionally more frequent in dialogue and that had previously been studied on the basis of the fiction texts in the ENPC were identified and chosen for further scrutiny: there and see. Results from these two case studies uncover some differences in the use of there and see in dialogue vs. narrative, most conspicuously for see where its preferred use in dialogue is the cognition sense and in narrative the perception sense. For there, a noticeable difference is the choice of verb in the Norwegian translations of existential there-clauses in dialogue and narrative. In narrative, verbs other than verbs of existence are sometimes chosen, while this is never the case in dialogue.