Analysis of semantic relations and predicate-argument structure is one of the core pieces of any system for natural language understanding. This field has a long history, but has also undergone significant change in recent years. The advent of systematically defined resources for semantic structures has enabled the creation of large corpora annotated according to these standards, and automatic systems trained using the resulting data. Given the large amount of work in this area from groups around the world, it can be difficult for the uninitiated to get a sense of the state of the art; we hope that the overview we provide in this volume will be helpful to newcomers to the field as well as to practitioners in other areas of natural language processing. While some background in computational linguistics will be helpful to the reader, we have assumed very little specific technical knowledge.We have envisioned this volume as the initial basis of a graduate seminar, possibly including a variety of students from computer science, linguistics, and related fields.We hope that it will provide an appropriate springboard for progress on the exciting new challenges we outline in our conclusion.
Much of following chapters arises from our own work in this area, and would not have been possible without our collaborators, in particular:Dan Jurafsky,Paul Kingsbury,Karin Kipper Schuler, Hoa Trang Dang, Szu-ting Yi, Edward Loper, and Julia Hockenmaier.We are also grateful to the organizers of CoNLL who have showcased this task in several consecutive meetings. Finally, we wish to thank Katrin Erk, Susan Brown and two anonymous reviewers for their comments on our draft, and Graeme Hirst for his guidance in the development and editing of the book.