Comprehensive Supersense Disambiguation of English Prepositions and Possessives

TL;DR

This paper presents a comprehensive annotation scheme and dataset for disambiguating English prepositions and possessives using broad supersense classes, improving interpretability and consistency in semantic relation analysis.

Contribution

It introduces a unified annotation scheme and task for supersense disambiguation of prepositions and possessives, uniting them under a common class inventory.

Findings

High interannotator agreement demonstrates annotation reliability.

Supervised methods achieve encouraging disambiguation results.

The scheme effectively distinguishes lexical contribution from contextual role.

Abstract

Semantic relations are often signaled with prepositional or possessive marking--but extreme polysemy bedevils their analysis and automatic interpretation. We introduce a new annotation scheme, corpus, and task for the disambiguation of prepositions and possessives in English. Unlike previous approaches, our annotations are comprehensive with respect to types and tokens of these markers; use broadly applicable supersense classes rather than fine-grained dictionary definitions; unite prepositions and possessives under the same class inventory; and distinguish between a marker's lexical contribution and the role it marks in the context of a predicate or scene. Strong interannotator agreement rates, as well as encouraging disambiguation results with established supervised methods, speak to the viability of the scheme and task.