This project, which ran from April 2004 to May 2007, developed broad-coverage computational lexicons and
grammars for two verb-initial languages, Welsh
(Celtic) at the University of Essex, and
Malagasy (Austronesian) at Oxford University.
The project was a part of PARGRAM, a
collaborative project to develop computational grammars within a
shared linguistic framework and based on common linguistic
The syntactic framework of reference for PARGRAM is the highly stable
and mathematically well-understood theory of Lexical Functional
Grammar, a constraint-based theory of linguistic structure.
Participants use the XLE grammar
development platform, an efficient and powerful system developed and
maintained at the Palo Alto
Research Center in California, which faithfully implements LFG
Papers and presentations on the Welsh work can be found
Data and resources for Welsh can be found here.
The Welsh Grammar can be run using the University of Bergen
Interface (access to Welsh Grammar version 13/06/07). For instructions on how to do this and further information,
Other Welsh resources:
Historical Corpus of the Welsh Language
Welsh Numerals Project . An AHRC-funded project developing out of
this ESRC-funded Welsh grammar.
Data and resources for Malagasy can be found here.