Abstract. The tools of corpus linguistics have become indispensable for research in descriptive translation studies (DTS), which aims to describe the characteristics of the translation process, and translational texts. Machinereadable corpora of translated texts are crucially important since they can yield statistically signifi cant results that underpin the fi ndings of empirical studies. Baker's (1993) seminal paper gave new impetus to translation research as it has re-calibrated the goals of DTS to study and uncover the particular properties of the so-called "third code" (Frawley 1984), i.e. the language of translated texts, with the help of computerized corpora. The present study, after providing a brief overview of international and Hungarian corpus linguistic research, introduces the Pannonia Corpus Project developed by Eötvös Loránd University's Translation Studies Doctoral Programme, which was created to make a Hungarian translation corpus, containing millions of words, available for translation researchers. The Pannonia Corpus (PC) is a multi-modal corpus: it contains translated, interpreted, and audiovisual texts. It represents a diverse array of texts of specialized and literary genres, refl ecting modern language use and the current state of the translation industry. The PC provides researchers with a vital opportunity as its multimodality, diverse textual make-up, and substantial size are unparalleled in the Hungarian context. Until now, there were no large corpora available to researchers that could have facilitated qualitative as well as quantitative research, satisfying the demands of modern translation studies research in Hungary.

This paper discusses a corpus of so-called carmina gratulatoria, or gratulatory poems, composed by 17th-century Hungarian peregrines in Hebrew. The paper argues that this understudied type of text offers a valuable resource for a wide range of fields due to the sociohistorical, cultural, and linguistic information the texts contain.

This paper presents an analysis of the structures the discourse marker vajon forms in translated Hungarian fiction. Although translation data has been deployed in the study of discourse markers (Aijmer & Simon- Vandenbergen, 2004), such studies do not account for translation-specific phenomena which can influence the data of their analysis. In addition, translated discourse markers could offer insights into the idiosyncratic properties of translated texts as well as the culturally defined norms of translation that guide the creation of target texts. The analysis presented in this paper extends the cross-linguistic approach beyond contrastive analysis with a detailed investigation of two corpora of translated texts in order to identify patterns which could be a sign of translation or genre norms impacting the target texts. As a result, a distinct, diverging pattern emerges between the two corpora: patterns of explicit polarity show a marked difference. However, further research is needed to clarify whether these are due to language, genre, or translation norms.