Scope Disambiguation Is a Ranking Problem
Christoph Rzymski (Jena)

Order of evaluation undoubtedly plays a crucial role in natural language
systems. The order in which we process, and perceive, information,
fundamentally influences our understanding of any composite natural
language expression. This presentation argues that a better
understanding and a more thorough application of order of evaluation
concepts helps with tackling the unwieldy and error-prone task of
(automated) scope disambiguation. Specifically, I argue that
scope-bearing elements can be ranked, with each ranking permutation
representing (potentially) different semantic and pragmatic properties
and interpretations. Said rankings, in turn, can be used to build and
train systems for scope disambiguation.