Skip to main content

Computational Morphology

Introduction to the main approaches, tools, and methods related to computational modeling and learning of morphology and morphophonology.

This course will examine these major questions:

  • How do we model our linguistic knowledge in morphology and morphophonology computationally?
  • Are such models something we could learn automatically from either raw data or annotated data?
  • What do our computational models tell us about the nature of morphology?

The course has both a practical and a theoretical component. On the practical side, we want to explore current methods to model morphology computationally and what kinds of applications can be derived from such models. Many natural language processing systems rely on a morphophonological component at some stage, reflecting the importance of adequate and accurate treatment of this part. The goal is to achieve fluency in translating linguistic generalizations into a working computational model.

From a more theoretical perspective, computational morphology and morphophonology studies the computational properties of sound and word-formation patterns. Investigating these properties formally helps distinguish actually occurring patterns from logically possible patterns. Computational models are both accurate and explicit at the same time and thus allow for deeper insight into the nature of phonological and morphological processing from a cognitive perspective. Computational analyses also contribute to the ongoing discussion on the requisite formal power of theoretical models of word-internal patterns in linguistics.

Computational morphology and phonology are also very active current research topics. An aim of the course is to familiarize ourselves with contemporary questions and research results. The main topics to be discussed are:

  • Modeling phonology and morphology
    • Formal analysis
    • Finite-state implementations of generative models
    • Applications of morphophonological models
  • Learning
    • Formal analyses of learnability
    • Supervised and semi-supervised learning
    • Unsupervised learning