A semiautomatic lemmatisation procedure for treebanks. Old English strong and weak verbs.

摘要

The aim of this paper is to present a semiautomatic lemmatisation procedure implemented on database software that is compatible with the morphological tagging of treebanks. The language of analysis is Old English, for which parsed corpora are available but they are not lemmatised. The scope of the paper includes both strong and weak verbs. After describing the lemmatisation procedure, the paper discusses the results of automatic searches and compares them with the lists of inflectional forms by lemma provided by other lexicographical sources. The conclusions insist on the limits of automatic lemmatisation and the compatibility with treebank parsing.