Computational and Linguistic Issues in Designing a Syntactically Annotated Parallel Corpus of Indo-European Languages
Dag Trygve Truslew Haug, Marius L. Jøhndal, Hanne Martine Eckhoff, Eirik Welo, Mari Johanne Hertzenberg, Angelika Müth · Trait. Autom. des Langues · 2009
This paper reports on the development of the PROIEL parallelcorpus of New Tes- tament texts, which contains the Greek original of the New Te stament and its earliest Indo- European translations, into Latin, Gothic, Old Church Slav ic and Classical Armenian. A web application has been constructed specifically for the purpo se of annotating the texts at multiple levels: morphology, syntax, alignment at sentence, depend ency graph and token level, informa- tion structure and semantics. We describe this web applicat ion and our annotation schemes. Although designed for investigating pragmatic resources,the corpus with its rich annotation is an important resource in contrastive and historical Indo -European syntax and pragmatics, easily expandable to include other old Indo-European langu ages. RESUME. L'article decrit le developpement du corpus aligne PROIEL,qui couvre le texte ori- ginal grec du Nouveau Testament et les traductions latine, g vieux-slave et armenienne. Pour faciliter la creation du corpus, nous avons developpe u ne application web qui permet l'annotation des textes sur plusieurs niveaux: morphologi e, syntax, alignement de phrases, syn- tagmes et mots, structure informationelle et semantique. D ans l'article nous decrivons cette application web ainsi que nos schemas d'annotations. Bien q concu pour l'etude des ph eno- menes pragmaticaux, l'annotation tres riche des textes a re sulte a une ressource importante pour l'etude comparee and historique du syntax et pragmatiq ue indo-europeen, et le corpus pourra facilement etre etendu a d'autres langues indo-euro peennes.