Name of the teacher: Francis Bond
Contact details of the teacher
- Affiliation: Palacky University, Olomouc
- Email address: Francis Bond
- Website: https://fcbond.github.io/
Level: Intermediate
The target group is PhD candidates and second-year RM students who do have general linguistic knowledge but who have not (yet) written papers in the specific subject of your course.
The course assumes basic knowledge of lexical semantics and the python programming language. Students should bring their own laptops to class.
In this course, students will explore how lexical graphs can be used to represent and analyze word meaning across languages. We will focus on WordNet, and use it as a case study in how semantic relations—such as synonymy, antonymy, hyponymy, and meronymy—can be encoded in a structured network. Through hands-on exercises, students will learn how meanings are grouped into synsets, how these synsets are linked by semantic relations, and how such networks can model lexical organization in a way that supports linguistic inquiry.
A central theme of the course is cross-linguistic comparison. Using the Open Multilingual WordNet (OMW), students will investigate how different languages carve up conceptual space, identifying cases of one-to-many sense correspondences, lexical gaps, and differences in morphology. This exploration highlights both the strengths of graph-based approaches—explicit structure, interpretability, and linguistic grounding—and their limitations, including issues of coverage, granularity, and cross-lingual alignment.
In the final part of the course, we will contrast these symbolic representations with distributional and embedding-based approaches to meaning. Students will see how large-scale statistical models capture lexical similarity in high-dimensional space, and will discuss how graph-based and vector-based representations can complement one another in modern research.
The course is hands-on and skills-focused. By the end, students will be able to navigate WordNet and OMW data programmatically, use them to test linguistic hypotheses, and critically assess how different computational models represent meaning. The methods introduced will provide a foundation for empirical research in lexical semantics, computational linguistics, and cross-linguistic analysis.
The reading material for this course has not yet been finalized and will follow later.
The material to be used in your course should be accessible to students. This means that you should either use published work that is accessible online (please include doi’s) or that can be bought easily (e.g. a handbook). If you need to include older or unpublished work, please make sure that students can download this from your webpage, or prepare a zipfile to send to your students a month before the course begins.
The norm for the readings during the school is not much more than 10 pages a day, which in practice amounts to one short article or chapter per day. Please be aware that students are taking two courses per week, so that they do not have too much time to do readings during the week.