Translation of “and” in a parallel TedTalk corpus of English, Czech, Hungarian, Lithuanian and French: functions and omissions

Abuczki, Agnes;Burkšaitienė, Nijole;Crible, Ludivine;et.al.
(2018) Cross-linguistic Discourse Annotation: Applications and Perspectives (TextLink2018) — Location: Toulouse, France (19.March.2018)

Files

No attached file found for this publication.

Details

Authors
  • Abuczki, Agnes
    Author
  • Burkšaitienė, Nijole
    Author
  • Crible, LudivineUCLouvain
    Author
  • et. al.
Abstract
In this paper we report on the methods and findings of a multilingual corpus study focusing on the functions of and in English and its translations into Czech, French, Hungarian and Lithuanian, in a selection of TedTalks. Firstly, we outline the functions of the discourse marker and, as it has been previously described in the literature. Secondly, we describe our annotation scheme (Crible & Degand’s in press) and our results. Discourse markers have several functions: they connect single text segments into a compound unit, they express a semantic type of relation (contrast, reason, instantiation, etc.) and express various pragmatic functions. Highly frequent among them, the additive discourse marker and encodes very little information in its core meaning. Yet, it is used in a variety of contexts where additional meanings can be identified, such as contrast or consequence. According to Schiffrin (1987), and has two basic discourse uses: coordination and continuation. It is a structural device for building text which coordinates ideas and units, however, it has little semantic meaning. Besides, and also has contrastive uses, e.g. We tried to win. And we lost. It can also preface the outcome of a reason, e.g. That’s one game I remember because we had a driveway and, like we would hide, and they would walk around the driveway? Y’know? And I- I remember it so distinctly. (Schiffrin 1987:129.) Moreover, and can connect events: [POSITION EVENT] and [SUPPORT EVENT] and [EVENT], and it can also connect reasons or two pieces of support at a higher level of idea structure: [POSITION SUPPORT 1] and [SUPPORT 2]. And can connect a general conclusion drawn from a list of specific events which are asyndetically connected, e.g. I uh I go on trips with ‘em, I bring ‘em here, we have supper, or dinner here, and I don’t see any problem because I’m workin’ with college graduates. (Schiffrin 1987:130.) And often links structurally similar clauses as well and it doesn’t favor tense switching (unlike temporal connectives). And also has pragmatic effect: as a marker of speaker-continuation in interaction (which is a consequence of the speaker’s situated context-bound use). It can be used to (try to) reopen an interactional unit whose completion has earlier been interrupted (turn-taking) or to continue/return to a previous question as a request for elaboration (turn-giving): [QUESTION 1] [ANSWER] and [QUESTION 1].On the other hand, and can be used to link pre-arranged questions in a question agenda: [QUESTION 1] [ANSWER] and [QUESTION 2] (Schiffrin 1987:146). In this research study, all occurrences of English and as a discourse marker along with their translations into Czech, French, Hungarian and Lithuanian have been manually identified in a selection of TedTalks. The originals and their translations were annotated in each language by two experts, following Crible & Degand’s (in press) functional classification. Crible & Degand’s taxonomy of domains and functions of discourse markers, which is specifically designed for annotating discourse markers used in spoken discourse, consists of four main domains. First, the ideational domain is linked to “states of affairs in the world, semantic relations between real events”. Second, the rhetorical domain is linked to “the speaker’s meta-discursive work on the ongoing speech”. Third, the sequential domain is linked to “the structuring of discourse segments, both at macro- and micro-level.” Finally, the interpersonal domain is linked to “the interactive management of the exchange, in other words, to the speaker-hearer relationship” (Crible 2014:18). In addition to the domains, eleven functions can be assigned to the discourse markers, including addition, contrast or specification. Domains and functions are independent: any domain can apply to any function and any function can apply to any domain. According to Crible & Degand (in press), annotators “can choose to start at domain-level or function-level, to annotate both levels simultaneously or independently, and could even decide to stop at one level if a particular domain DM token is under-specified for the other level” (2017:20) which enhances the annotation process. Also, the authors believe that this system vouches for reliable annotation (high inter-annotator agreement), because of the reduced number of labels and the independence of the two levels (i.e. domains and functions). We address the following research questions: (1) What is the functional spectrum of and in English TedTalks? (2) How is and translated in Czech, French, Hungarian and Lithuanian? Are specific functions of and associated with specific translations? (3) Which uses of and tend to be omitted in the translations? A pilot study has been carried out and its first results point at regular tendencies regarding implicitation, multiple translation equivalents and functional shifts of and across languages. The methodological decisions made in the course of research are related to the choice of the corpus and the annotation method. These choices have been determined by the aim of the research, that is, to annotate the English spoken discourse marker and, to compare its functions with its counterparts in Czech, French, Hungarian and Lithuanian as well as to analyze the translations of and into Czech, French, Hungarian and Lithuanian. For these reasons, the multilingual translations of TED Talks were chosen. This choice was made on the grounds that parallel texts are considered to be ideal for optimal comparability between languages as they provide more flexible and accurate ways to compare discourse markers (Zufferey & Degand 2013). The choice of the functional approach to be used for this investigation was predetermined by the specific nature of discourse markers, which covers some specific features, e.g. even though most languages possess discourse markers, they have a high degree of contextual variation (Crible & Degand in press). The empirical research consists of two stages. Initially, the discourse marker and is compared to its Czech, French, Hungarian and Lithuanian counterparts by applying Crible & Degand’s taxonomy of domains and functions of discourse markers. Then, the translations of and into Czech, French, Hungarian and Lithuanian, found in the annotated samples, are analyzed. The research reveals that in the annotated sample, the discourse marker and with its counterparts in Czech, French, Hungarian and Lithuanian is mainly used in the ideational domain, marking/connecting factual information, and in the sequential domain, representing the structuring of local and global units of discourse and less often in the rhetorical domain which is related to the speaker‘s subjectivity. Cross-linguistically, the uses of and in the sequential domain are the most frequently omitted, whereas and’s operating in the ideational domain are usually preserved. The reason could be the attempt to observe the continuation function to avoid juxtaposition and connect the utterances in a basic way or it might be due to constraints of the translation by subtitles. When and in sequential or rhetorical addition is translated in French, et is annotated by many different labels. In Lithuanian the rhetorical domain also uncovers a number of differences. Here we find rhetorical contrast rendered into the Lithuanian o which could represent both addition and contrast. Rhetorical consequence is also observed which is mostly related to the whole argument. The reason could be the nature of the domain since it is related to the speaker‘s subjectivity. A striking regularity observed cross-linguistically is the use of and for topicalization of the previous focus (in the relation of addition): En There'd be a huge spread in her scores. [And] [actually] it's this spread that counts. Cz V jejím hodnocení by byl velký rozptyl. [A] [právě] na tomto rozptylu záleží. Hu Nagy lesz a szórása a pontoknak. [És] ez az a szórás, ami számít. Li Jos balai būtų visiškai pasiskirstę. [Ir] [išties], svarbus būtent tas pasiskirstymas. Fr omission of the discourse connective Similarly, and is regularly used in many languages as means of expressing temporal succession of two events: En We give ourselves a little bit of time to play the field, get a feel for the marketplace or whatever when we're young. [And] then we only start looking seriously at potential marriage candidates once we hit our mid-to-late 20s. Cz Dopřejeme si čas na hraní a průzkum toho, co je k mání, dokud jsme ještě mladí. [A] vážný konkurz na kandidáty pod čepec zahájíme až po pětadvacítce. Li Leidžiam sau šiek tiek išsilakstyti, kol esam jauni, leidžiam suprasti, kas yra rinkoje, ar panašiai. [Ir] tuomet vėlesniame dvidešimtmetyje pradedame į vedybų kandidatus žiūrėti rimtai. Hu, Fr omission of the discourse connective In sum, our study of the functions and translations of and in a parallel corpus across five languages has shown the most frequent functions for and in the particular register of Ted Talks, as well as systematic tendencies regarding its translation. In particular, and is most frequently omitted in the translations when it expresses sequential addition, that is, a basic continuity function. This study is the first step of a larger research project on the uses of underspecified connectives such as and. In future work, we will replicate the analysis on a larger sample, and investigate additional connectives such as now, so or but.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Abuczki, A., Burkšaitienė, N., Crible, L., & et al. (2018). Translation of “and” in a parallel TedTalk corpus of English, Czech, Hungarian, Lithuanian and French: functions and omissions. Cross-linguistic Discourse Annotation: Applications and Perspectives (TextLink2018), Toulouse, France. https://hdl.handle.net/2078.5/175018