The Issues of Giving Meronyms in the Corpus

Abstract

This thesis about meronyms, which represent part-whole relationships and its several challenges when are given in a linguistic corpus. One major issue is ambiguity in part-whole relations, where the same part may belong to multiple wholes depending on the context (e.g., “wheel” as a part of both a car and a bicycle). Additionally, the paradox of inconsistencies arise when meronyms exist at different hierarchical levels, making it difficult to determine the appropriate level of annotation (e.g., “leaf” as a part of “branch” vs. “tree”). Another challenge is context-dependent meronyms, where a word functions as a meronym only within specific domains (e.g., “petal” is a part of “flower” only in botanical contexts). Furthermore, lexical variation and synonymy complicate the notion, as different terms may refer to the same part in different linguistic or dialectical differences (e.g., “tire” vs. “wheel” for a car component). Moreover, multiword and compound expressions present difficulties in corpora, as some meronyms appear in phrases rather than single words (e.g., “front door” as a part of a house). Lastly, cross-linguistic variation makes lexemes cognitively complicated, as languages differ in how they express part-whole relationships (e.g., some languages use a single term for both “hand” and “arm) such as in Uzbek language. Identifying these challenges requires clear explanations, hierarchical structuring of meronyms, and clarify concrete strategies to ensure consistency in corpus analysis.