Comparing Probabilistic, Distributional and Transformer-Based Models on Logical Metonymy Interpretation

In linguistics and cognitive science,Logicalmetonymiesare defined as type clashes be-tween an event-selecting verb and an entity-denoting noun (e.g.The editor finished the ar-ticle), which are typically interpreted by infer-ring a hidden event (e.g. reading) on the basisof contextual cues.This paper tackles the problem of logicalmetonymyinterpretation, that is, the retrievalof the covert event via computational methods.We compare different types of models, includ-ing the probabilistic and the distributional onespreviously introduced in the literature on thetopic. For the first time, we also tested onthis task some of the recent Transformer-basedmodels, such as BERT, RoBERTa, XLNet, andGPT-2.Our results show a complex scenario, in whichthe best Transformer-based models and sometraditional distributional models perform verysimilarly.However, the low performanceon some of the testing datasets suggests thatlogical metonymy is still a challenging phe-nomenon for computational modeling.