Welcome to the homepage of the current release of the Nigel generation grammar for English. Nigel has been under development since the early 1980s, when it was used within the Penman project for English generation. It was mainly developed by Christian Matthiessen on the foundation of work by Michael Halliday. Since then many people have contributed to various parts of its coverage. For references to the underlying linguistic grammatical description, which has been worked out in far more detail than has so far been integrated into the computational version, see here.
Although the new release includes some substantial extensions in coverage, the most obvious difference to preceding versions will be the accessability of the documentation. Documentation is essentially example based. Each area of the grammar should be covered by sets of examples, expressed in terms of semantic input expressions for the grammar (so-called Sentence Plan Language expressions: SPL), that `exercise' the area of grammar in question. Examples include both simple expressions reflecting the area being documented and more complex examples drawn from real texts or applications.
Examples included so far (along with their current generation status: this
is being steadily revised) are the following:
(CLICK
RIGHT to DOWNLOAD!!) |
|
the Penman Nigel exercise set | download file |
the Penman Nigel Reuters examples | download file |
the Penman primer examples (navy sentences) | download file |
the Komet Dictionary of Art style sentences | download file |
HealthDoc example texts | download file |
Example comparatives | download file |
Example concessions (TechDoc, Brigitte Grote) | --- |
Further example sets in preparation include:
The grammar as a whole is organized into functional regions. These can be inspected using the grammar explorer tool. Examples of views that can be found there are given here. In addition, use of the grammar explorer allows any constituent from any of the generated examples to be related back to the precise path through the grammar network that caused that constituent to be generated.
For full grammar development the KPML development
environment is available.
Note:
IMPROVING THAT COVERAGE PROVIDES MANY GOOD OPPORTUNITIES FOR WORK RANGING FROM UNDERGRADUATE DISSERTATIONS TO PhDs AND BEYOND! ANY CONTRIBUTION, EITHER IN THE GRAMMAR OR IN PROVIDING ADDITIONAL SETS OF SEMANTIC INPUTS, IMPROVES THE RE-USABILITY OF THIS RESOURCE FOR THE GENERATION COMMUNITY. IF YOU ARE INTERESTED: GET IN TOUCH! THERE IS PLENTY TO DO; WORKING ON ANY AREA IS ALSO A VERY GOOD WAY OF GETTING TO KNOW THIS PARTICULAR APPROACH TO GENERATION IN DETAIL. |