Eo Ipso – Automated Internal Reconstruction

DSpace Repository


Dateien:

URI: http://hdl.handle.net/10900/133321
http://nbn-resolving.de/urn:nbn:de:bsz:21-dspace-1333214
http://dx.doi.org/10.15496/publikation-74674
Dokumentart: PhDThesis
Date: 2023-08-15
Language: English
Faculty: 5 Philosophische Fakultät
Department: Allgemeine u. vergleichende Sprachwissenschaft
Advisor: Jäger, Gerhard (Prof. Dr.)
Day of Oral Examination: 2022-05-30
DDC Classifikation: 491.8 - Slavic languages
Other Keywords: Interne Rekonstruktion
Sprachrekonstruktion
Computerlinguistik
Vergleichende Sprachwissenschaft
Historische Sprachwissenschaft
Historisch-Vergleichende Sprachwissenschaft
Digital Humanities
Historical Linguistics
Comparative Linguistics
Computational Linguistics
Ancestral State Reconstruction
Computational Historical Linguistics
Internal Reconstruction
ISBN: 978-3-946552-74-1
License: https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode.de
Show full item record

Abstract:

This volume aims to contribute to the study into the method of internal reconstruction of ancient language stages in two particular ways. Firstly, it seeks to give texture to the theoretical aspects of internal reconstruction, including its problems and basic requirements. The second part deals with the machine implementation of internal‑reconstruction methods. Different methods that have been proposed in the literature are presented and their suitability for an adequate computational implementation is discussed. The emphasis is set on the phonological reconstruction of sound change. In total, six different methods are implemented with German and Proto‑Indo‑European data. The results of the automated internal reconstruction are evaluated quantitatively and qualitatively against a gold-standard.

This item appears in the following Collection(s)

cc_by-nc-nd Except where otherwise noted, this item's license is described as cc_by-nc-nd