enesyila/UD_Ottoman_Turkish-DUDU

Summary

An Ottoman Turkish dependency treebank annotated in UD style. Created by Enes Yılandiloğlu.

Introduction

This treebank comprises 25,010 words in 2,552 sentences that are firstly automaticaly annotated via Machamp (Van der Goot et al., 2021). During the training phase, multiple modern Turkish UD treebanks were used. and then manually corrected in a systematic way. Randomly shuffled sentences were written between 14th to 20th century in various genres such as fiction, news, article, registry record, and religious preach. Unfortunately, for this version, the genres can not be told apart by sentence ids. The order of the sentences is chronology based rather than genre based, the earliest written sentence is at the top. In this treebank, Ottoman Turkish transcription alphabet is used.

Currently, this is the largest Ottoman Turkish treebank in UD framework.

• —Initial release in Universal Dependencies in 2024-05-15 v2.14.
• —The treebank can also be accessed via Universal Dependencies.

Please cite this work: Yılandiloğlu, E., & Siewert, J. (2025). DUDU: A Treebank for Ottoman Turkish in UD Style. In Š. A. Holdt, N. Ilinykh, B. Scalvini, M. Bruton, I. N. Debess, & C. M. Tudor (Eds.), The Third Workshop on Resources and Representations for Under-Resourced Languages and Domains (RESOURCEFUL 2025) (pp. 74–79). University of Tartu Library.

Contact: enes.yilandiloglu@helsinki.fi

Acknowledgments

I am immensely grateful to Fatma Elcan for her tremendous help in providing me with sentences.