parser is being developed to examine lexical items (L) and phrases
in respect of one another and in respect of the sentence (S)
as a whole. In short, I am trying to parse a natural language
(NL) rather than to parse tags which are appended by speakers
of a language to a corpus.
date, I have created lists of closed-class words on the assumption
that such lists are part of the competence of a NL speaker. The
next stage applies optimality theory to the issue of L having
ambiguous semantic and phrasal associations within an S.
recommend the work of O.D. Macrae-Gibson and J.R. Lishman, whose
scanned ASPR is one of the more ambitious attempts at this kind
of thing. Their work is available from the Oxford Text Archives.
If you have any comments or questions or suggestions, please
feel free to write at firstname.lastname@example.org.