Representation models for aggregating linguistic information: Issues and analysis