Discourse Classification into Rhetorical Functions for AWE Feedback

Abstract

NLP-based technology constitutes the backbone of automated analysis of writing and feedback on grammar, usage, mechanics, style, organization, coherence, and content. However, analysis of rhetorical intent is a challenging NLP problem that remains to be addressed in order to facilitate the learning of academic genres. This paper reports on the development of the analysis engine for the [TOOL], an AWE program designed to provide genre and discipline-specific feedback on the functional units of research article discourse. Unlike traditional NLP-based applications that categorize complete documents, [TOOL]’s analyzer categorizes every sentence in the text as both a rhetorical move and step using a 17-step schema. The paper first reviews the approaches to and implementations of automated discourse categorization in order to provide a background for the automated genre analysis methodology. Then, it describes the construction of a cascade of two support vector machine classifiers trained on a multi-disciplinary corpus of annotated Introduction texts. Lastly, it demonstrates how this categorization approach was applied to the generation of feedback on the rhetorical functions of the research article genre. This work not only demonstrates the usefulness of NLP for automated genre analysis, but also paves the road for future AWE endeavors and forms of automated feedback that could facilitate construction of functional meaning in writing.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 35 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.