(one overhead per page)
Overheads from my plenary talk arguing for the necessity and utility of systemic-functional treebanks given at the 30th International Systemic Functional Linguistics Congress in Lucknow in December 2003.
While functional linguists in general, and systemic functional linguists in particular, have concerned themselves with both their traditional concerns, for example, language teaching, as well as opening up new areas for exploration, such as multimodal meaning and appraisal, there has been a radical revision of the abilities, requirements and methods adopted in computational approaches to language study. What can only be described as vast sums of money have been invested by research-funding bodies in the broad area of human language technology or 'linguistic engineering'. Here we see international research and development projects active in all areas from corpus design, through lexical and grammatical resource development, to semantics and contextual modelling. So far, however, systemic-functional linguistics has had very little contact with these developments---and has, accordingly, benefited from these investments similarly little.
For better or worse, funding programs have moved on from research and experimentation and increasingly call for standardisation and quantifiable results. The question of whether systemic-functional linguistics can avoid being 'standardized' out of the linguistic picture is therefore of some significance.
In this talk we survey the current state of human language technology and linguistic engineering and examine the requirements and tools that are now being established. We will explicitly ask the questions of whether systemic-functional linguistics has anything to offer to this development and whether it has anything to learn from it.
|
Acrobat pdf | (one overhead per page) |
download (zipped, 4Mb) |
|
Acrobat pdf | (6 overheads per page) | download (2.45 Mb) |
Thanks are due to the Deutsche Forschungsgemeinschaft (DFG) for funding this trip to Lucknow.