Discourse Markers and Rhetorical Relations

  • Discourse Markers and Rhetorical Relations

  • Introduction to Rhetorical Structure Theory (RST):

    • Formulated by Mann and Thompson in the early 1980s, RST offers a framework for analyzing the structure of discourse by focusing on the relationships between segments of text. The theory has prompted debates on the exact number and categorization of rhetorical relations utilized in analysis.

    • A significant paper in this field examines discourse markers as essential signals of rhetorical relations, utilizing corpus studies that dissect textual examples from various genres.

    • Findings from these studies reveal that between 60-70% of the rhetorical relations identified within discourse are not overtly indicated by discourse markers, highlighting the complexity of understanding these connections.

  • Coherence in Discourse:

    • Coherence is achieved through coherence relations that create connections and bind various parts of discourse together into a unified whole. It involves understanding the implicit relationships that may not be explicitly signaled.

    • Recognition of these coherence relations is crucial for effective text comprehension, as they guide readers in processing the information and inferring meanings.

    • Discourse markers play a pivotal role in assisting users in identifying and recognizing coherence relations, thereby enhancing the reader's ability to follow the argument or narrative thread effectively.

  • Types of Rhetorical Relations:

    • Paratactic and Hypotactic Relations:

      • Rhetorical relations identified through RST can be broadly classified into paratactic (equal status) and hypotactic (subordinate or dependent) relationships. These relations can be seen in how sentences or phrases function together to convey meaning, either by juxtaposition or through supporting structures.

  • Recursive Nature:

    • Rhetorical relationships are recursive; they can build upon each other, creating layers of meaning throughout the text. This recursive nature allows for complex structures where relationships can influence other relationships across larger segments of discourse.

  • External and Internal Recognition:

    • The recognition of relations involves multiple mechanisms, including morphological (study of word forms), syntactic (sentence structure), semantic (meaning of words and phrases), and pragmatic (contextual usage) processes. Each mechanism plays a significant role in how readers interpret and understand the discourse.

  • Discourse Markers Defined:

    • Discourse markers, although difficult to define succinctly, are generally understood as linguistic signals that link pieces of discourse together. These markers help clarify the relationships among statements and enhance overall understanding by indicating how different segments of text relate to one another.

    • Common examples of discourse markers include conjunctions (such as "and," "but," "or") and transitional phrases (like "for instance," "in contrast," "furthermore") that influence comprehension and guide readers through the argument or narrative.

  • Types of Marking for Coherence Relations:

    • Morphological: Tense is a crucial marker for establishing temporal relations, indicating when events occur relative to one another.

    • Syntactic: The sentence mood (indicative, imperative, subjunctive) can guide pragmatic interpretation by revealing the speaker's attitude towards the proposition.

    • Semantic: Certain verb meanings are indicative of causal relations; for instance, the use of verbs like "cause," "lead to," or "result in" makes explicit the nature of the relationships between actions or events.

    • Pragmatic: Implicature, which refers to implied meanings not explicitly stated, establishes non-explicit relations that may rely on shared context or cultural knowledge between the reader and the text.

  • Corpus Studies Overview:

    • Two corpora were analyzed in depth: a collection of spoken conversations and a set of newspaper articles. These studies aim to explore the relationships marked by discourse markers and their statistical significance across both spoken and written genres.

    • First Corpus: Scheduling Dialogues:

      • This corpus involves 30 task-oriented conversations centered on scheduling topics, specifically designed to examine and annotate rhetorical relations effectively.

      • Findings from this corpus indicated that only 31% of the identified rhetorical relations were explicitly marked by conjunctions, suggesting that the use of discourse markers in natural dialogue is relatively limited.

      • Additionally, the study revealed that top-level rhetorical relations, such as concession and condition, demonstrated varied frequencies of marking, with certain types being more prevalent in dialogue than others.

    • Second Corpus: Newspaper Articles:

      • RST-annotated articles from the Wall Street Journal were similarly analyzed for rhetorical relations, aiming for a comprehensive understanding of written discourse across formal genres.

      • This analysis yielded an overall marking level of approximately 43%, with a more diverse range of signals, including punctuation, being considered.

      • Significant differences were observed between frequently marked relations, such as concession, and infrequently marked relations, like background or elaboration, emphasizing the nuances in how different types of discourse are structured.

  • Discussion and Conclusion:

    • The study raises questions about the ways in which relations are signaled within texts and acknowledges that the recognition of unsignaled relations remains a complex and challenging aspect of discourse analysis.

    • Texts exhibit genre-specific expectations, which guide readers in interpreting relations differently, depending on the context of the discourse.

    • Future research should delve deeper into how themes of comprehension and understanding are constructed, especially concerning unsignaled relations, to enhance the practical applications of RST in various fields such as linguistics, communication, and education.