Inference through Alternative-Set Semantics
Gann Bierner, Bonnie Webber · Edinburgh Research Explorer (University of Edinburgh) · 2000
We show how alternative set semantics can be used to analyze a class of lexical items frequently found in natural language requests to search engines and databases (including other (than), such (as), and besides as well as better , faster , and other comparatives ). We show how the analyses allow us to infer properties of entities, as in the bolded phrase above where other licenses the inference that better and faster are comparatives.