I develop a general argument that the quantifier-variable-binding approaches available within the Frege model are unable to analyse a whole class of donkey pronouns. I call this the double-bind problem. I demonstrate this by looking at two major approaches to anaphora–E- and D-type pronoun theory and Discourse Representation theory–but the conclusions apply to other approaches.
Keywords: DRT, Dynamic Semantics, D-type pronoun, double bind, situation quantifier, variable