Discourse Parsing — Rhetorical Structure Analysis
Discourse Parsing (Rhetorical Structure Analysis) · Also known as: rhetorical structure analysis, RST parsing, PDTB parsing, Söylem Ayrıştırma (Discourse Parsing)
Discourse parsing is a natural-language-processing task that models the rhetorical relations between sentences and paragraphs of a text — relations such as cause, contrast, and elaboration — and represents them as a tree structure. It works within established frameworks, principally Rhetorical Structure Theory (RST), introduced by Mann and Thompson in 1988, and the Penn Discourse TreeBank (PDTB), released by Prasad and colleagues in 2008.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Discourse parsing fits when you have text documents of sufficient length and want to understand how their parts relate rather than just what words they contain. The text must be in English or another supported language, and each document should be long enough to contain real inter-sentence structure. It is well suited to explanatory and descriptive study of how a text is organised, and is not meaningful on isolated fragments with no discourse to parse.
Strengths & limitations
- Makes the rhetorical architecture of a text explicit, revealing how spans relate rather than only what they say.
- Grounded in established, well-documented frameworks (RST and PDTB) with shared relation inventories.
- Produces a structured tree that downstream tasks such as summarisation, coherence assessment, and argument analysis can build on.
- Requires English or another supported language; coverage outside well-resourced languages is limited.
- Needs documents of sufficient length — short fragments have little discourse structure to parse.
- Identifying relations is intrinsically difficult and remains an open research problem, so parser output can be noisy.
Frequently asked
What is the difference between RST and PDTB?
Rhetorical Structure Theory (Mann & Thompson, 1988) builds a single hierarchical tree from nucleus-satellite relations spanning the whole text. The Penn Discourse TreeBank (Prasad et al., 2008) annotates discourse relations around connectives with a sense hierarchy, without committing to one global tree. Discourse parsing can target either framework; fix the choice before you start.
How long does a text need to be?
Long enough to contain genuine inter-sentence structure. A few isolated sentences offer little for the parser to relate, so discourse parsing is meant for documents, not fragments.
Does it work in languages other than English?
It requires English or another supported language. Resources and parsers are strongest for well-studied languages; applying an English parser to other-language text degrades the relations it recovers.
What does the output look like?
A discourse tree (or set of labelled relations) showing how spans connect — for example which span is the cause, contrast, or elaboration of another. That structured representation is what you read and what downstream analysis consumes.
Sources
- Mann, W. C. & Thompson, S. A. (1988). Rhetorical Structure Theory: Toward a functional theory of text organization. Text, 8(3), 243-281. DOI: 10.1515/text.1.1988.8.3.243 ↗
- Prasad, R., Dinesh, N., Lee, A., Miltsakaki, E., Robaldo, L., Joshi, A. & Webber, B. (2008). The Penn Discourse TreeBank 2.0. Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC 2008). link ↗
How to cite this page
ScholarGate. (2026, June 1). Discourse Parsing (Rhetorical Structure Analysis). ScholarGate. https://scholargate.app/en/text-mining/discourse-parsing
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Argument MiningText mining↔ compare
- Sentiment AnalysisText mining↔ compare
- Text ClassificationText mining↔ compare