Symbol Declarations in Mathematical Writing

We present three corpus-based studies on symbol declaration in mathematical writing. We focus on simple object denoting symbols which may be part of larger expressions. We look into whether the symbols are explicitly introduced into the discourse and whether the information on once interpreted symbols can be used to interpret structurally related symbols. Our goal is to support fine-grained semantic interpretation of simple and complex mathematical expressions. The results of our analysis empirically show the potential benefit of using larger discourse context in automated disambiguation of mathematical expressions.