Annotation of clausal functional information for semantic retrieval

The study of language functions is closely associated with the semantic and pragmatic aspects of language. While data driven approaches have been successfully applied on retrieval of functional-semantic information at the discourse level, the work at the clause level is still largely absent. In this paper, we annotate an initial corpus with Systemic Functional Linguistics, a prominent framework for the analysis of language functions at the sentence/clause level. The annotated corpus makes it possible to train a classifier to automatically classify functional processes at the clausal level. With an initial computational resource, the linking and interoperation between the two levels of functional information is now possible, giving rise to a range of potential applications in functional/semantic retrieval.