A multilingual corpus study of the competition between<scp>past</scp>and<scp>perfect</scp>in narrative discourse

The western European present perfect is subject to substantial crosslinguistic variation. literature, however, focuses on individual languages or comparisons of a restricted number languages. We piece together the puzzle and do so in data-driven way by comparing use through parallel corpus based French novel L’Étranger its translations Italian, German, Dutch, Spanish, British English, Modern Greek. introduce showcase Translation Mining, software suite combining database with annotation analysis tools. Mining allows us generate descriptive statistics tense across but also visualize variation multidimensional scaling component link we find underlying data integrated setup. confirm that competes past reveal fine-grained scalar nature To complete puzzle, ascertain dimensions variation, ranging from lexical compositional semantics dynamic pragmatics. 1