A Bayesian model of Turkish derivational morphology

Date

2023-7-21

Author

Kunter, Utku Can

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

204
views

155
downloads

Cite This

Building on an extensive review of the psycholinguistics literature and Turkish Derivational Morphology (DM), we propose a novel structure for representing DM in three hierarchical layers: segmentation, lexical selection and derivation. This proposal involves laying a belief structure over the traditional morphological structure of DM. We call this novel structure the Conventionalized Structure (CdS). We develop a computational model of morphology processing based on CdS using Bayesian Belief Networks (BBN). We present an algorithmic implementation for this model that learns and accurately represents new lexical items, recognizes affixes and tracks the salience of each item probabilistically. We carry out trials on this model with realistic observation lists and observe that model predictions are in line with the findings in studies in psycholinguistics. To support our claims and methodology, we carry out an extensive study of Turkish DM, looking into both Modern Turkish and Orkhon Turkic. We also look into the distributional semantics of derivational affixes and observe a high degree of regularity. In order to represent the complex semantics arising from interactions between morphemes, we use the categorial grammar framework. We build a baseline grammar, based on which we construct observation lists for exploration trials. While we focus on Turkish DM, we do not make any language-specific assumptions, our methods and results should be generalizable to other languages with segmental morphology.

Subject Keywords

Derivational Morphology, Categorial Grammar, Bayesian Belief Networks

URI

https://hdl.handle.net/11511/104888

Collections

Graduate School of Informatics, Thesis