Some aspects of the syntactic encoding of nouns in a computational lexicon – the STO project
Abstract
In this paper I first give a brief introduction to the syntactic representation of the STO lexicon. Hereafter I focus on some ofthe problematic phenomena that we encounter during the syntactic encoding ofnouns. In principle each noun with one subcategorisation frame is encoded as one syntactic unit even ifthe noun has two or more semantic readings. Two different syntactic manifestations of the same semantic reading of a noun should however be encoded in the same syntactic unit. When the subcategorisation frame is established, the limits between arguments and modifiers can be difficult to distinguish. Normally modifiers are not considered part of the subcategorisation frame but if corpus search reveals some frequent modifiers, these can be encoded in line with the arguments. Noun compounding is a productive word formation process in Danish. The first element of a compound can fill in the argument slot ofthe second element, changing the subcategorisation frame, but this is not always the case. Genitive is only encoded when it is part of the subcategorisation frame. For nondeverbal nouns it can be difficult to decide when to encode a genitive. The choice is highly dependent on the number of occurrences in corpora.
Session
Computational Lexicography and Lexicology
Keywords
BibTex
@InProceedings{ELX02-015, author = {Sussi Olsen}, title = {Some aspects of the syntactic encoding of nouns in a computational lexicon - the STO project}, pages = {159-168}, booktitle = {Proceedings of the 10th EURALEX International Congress}, year = {2002}, month = {aug}, date = {13-17}, address = {København, Denmark}, editor = {Anna Braasch and Claus Povlsen}, publisher = {Center for Sprogteknologi}, isbn = {87-90708-09-1}, }
