Recognizing and interpreting temporal expressions in open domain texts
D.D. Ahn, Sisay Fissaha Adafre, Maarten de Rijke · 2005
Since the early 1990s, when the third author was a visiting PhD student at Imperial College, he and Dov Gabbay have had many interactions. Usually over the phone, they talked about science, publishing and other initiatives, and invariably their conversations ended on a strategic note. Dov and the third author of this paper share a fundamental vision on the status and future of logic. To have a bright future, the discipline needs to be strongly embedded in external uses and needs. Much of the innovation in logic over the past decades has come from computer science, with new questions, new modeling needs, new reasoning mechanisms, etc. To continue to strive, we believe that logic should be embedded alongside its application areas, with feedback back and forth through measurable evaluations, either theoretical or experimental. It is clear that in many research institutes around the world this is actually how logic and computer science have come to interact. But we believe that logic can and should play a similar role vis-à-vis other scientific areas, both traditional and non-traditional, such as analytic philosophy, law, cognitive science, economics, information science, theology, language, and political theory. In more recent years, the interactions between Dov and the third author have become less frequent, mainly because Dov has been looking at one set of areas for inspiration and research questions (including law and philosophy), while the third author of this paper has turned to another, organized around information access. This paper deals with identifying, and assigning meaning to, temporal information. It touches on at least two of Dov’s longstanding interests, temporal logic and natural languages semantics. There is very little symbolic reasoning in this paper, and the semantics we pursue is shallow at best, but the paper does contribute to our understanding of how to process natural language texts if we are ever going to have systems that deal with the information contained in, for instance, news papers or web pages, in an intelligent way. 2 David Ahn, Sisay Fissaha Adafre, and Maarten de Rijke 1