Corpus-based Extraction of Collocations for Near-Synonym Discrimination

By November 23, 2016,
Page 367-374
Author Mariusz Kamiński
Title Corpus-based Extraction of Collocations for Near-Synonym Discrimination
Abstract This paper shows a method of extracting collocations from a corpus for the purpose of discrimination of near-synonyms. Information on synonym discrimination, which is often part of usage notes appended to entries in general dictionaries, specifies points of differences between words of similar meaning. Using computer technology, one can automate the process of collocation retrieval and reduce the amount of time and effort needed to prepare such information. The paper emphasises the importance of examining collocational behaviour when establishing meaning differences, but it also underlines the necessity of a critical analysis of existing records of word meaning, in particular dictionaries. An evidence from a parsed corpus is used in order to find collocations for a set of synonymous adjectives (false, counterfeit, artificial, bogus, forged) standing either in attributive or predicative positions. The corpus findings are compared against definitions in MacMillan Dictionary online.
Session Lexicography and Corpus Linguistics
Keywords collocation; synonymy; discrimination; extraction; corpus; parser
author={Mariusz Kamiński},
title={Corpus-based Extraction of Collocations for Near-Synonym Discrimination},
booktitle={Proceedings of the 17th EURALEX International Congress},
address={Tbilisi, Georgia},
editor={Tinatin Margalitadze, George Meladze},
publisher={Ivane Javakhishvili Tbilisi University Press},