Studying the Distribution of Fragments of English Using Deep Semantic Annotation

We present a preliminary study on how to use deep semantic annotation, namely the Boxer statistical semantic parser, that is capable of producing FO semantic representations for English sentences, to understand the distribution of families of socalled fragments of English. In particular, we try to answer the questions, relevant for the field of natural logic, of whether and how the semantic complexity of those fragments (viz., the computational complexity of the satisfiability problem of their FO semantic representations) correlates with their frequency.