A distributional study of negated adjectives and antonyms

Abstract

Comunicació presentada al Fifth Italian Conference on Computational Linguistics (CLiC-it 2018); celebrada del 10 al 12 de desembre de 2018 a Torí, Itàlia.

In this paper, we investigate the relation between negated adjectives and antonyms in English using Distributional Semantics methods. Results show that, on the basis of contexts of use, a negated adjective (e.g., not cold) is typically more similar to the adjective itself (cold) than to its antonym (hot); such effect is less strong for antonyms derived by affixation (e.g., happy - unhappy). In questo lavoro, analizziamo la relazione fra aggettivi negati e antonimi in inglese utilizzando metodi di Semantica Distribuzionale. I risultati mostrano che, sulla base dei contesti di uso, la negazione di un aggettivo (ad es. “not cold”; it.: “non freddo”) e tipicamente ` piu simile all’aggettivo stesso (“cold”; it.: ` “freddo”) che al suo antonimo (“hot”; it.: “caldo”). Tale effetto e meno accentuato ` per antonimi derivati tramite affissi (ad es. “happy”-“unhappy”; it.: “felice”- “infelice”).

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 715154), and by the Catalan government (SGR 2017 1575).