Abstract
We present a novel semantic framework for modeling linguistic expressions ofgeneralization— generic, habitual, and episodic statements—as combinations of simple,real-valued referential properties of predicates and their arguments. We usethis framework to construct a dataset covering the entirety of the Universal
Dependencies English Web Treebank. We use this dataset to probe the efficacy oftype-level and token-level information—including hand-engineered featuresand static (GloVe) and contextual (ELMo) word embeddings—for predictingexpressions of generalization.