Interpreting or describing? Measuring verb abstraction

The paper describes the results of machine learning experiments with verb classification according to the Linguistic Category Model (LCM)[1]. The LCM typology is a well-established tool to measure language abstraction, linked to sentiment and applicable in sentiment-analysis related areas. Our goal is to create automated methods of recognizing LCM verb classes. The method, demonstrated in the Polish language, turns out to be very promising, especially given the upper bounds set by inter-annotator agreement.