A qualitative model for time granularity

This paper addresses the problem of granularity in temporal representation in the context of text analysis. In contrast with other fields where granul arity levels are essentially quantitative, in natural language the differen t levels are not always precisely defined and granularity is of a more subt le and qualitative nature. A model is proposed for representing such phenom ena, based on time units and time units intervals. Time units represent chu nks of time which are considered indivisible at a given granularity level, but which may be refined and contain other time units or intervals at a hig her granularity level. The structure of the set of time units and intervals is studied and a relation algebra is defined which extends the traditional Alien's Point or Interval Algebra. Weak and strong representations of this algebra are proposed. We claim that such representations, called grained t emporal structures, are adequate for coping with dynamic qualitative change s of granularity. A logic with restricted quantifiers is proposed for forma lizing temporal knowledge and examples are discussed which show the relevan ce of the model for natural language analysis.