Identifying High-Level Organizational Elements in Argumentative Discourse
23966Authors
nan
Abstract
Argumentative discourse contains not only language expressing claims and evidence, but also language used to organize these claims and pieces of evidence. Differentiating be- tween the two may be useful for many appli- cations, such as those that focus on the content (e.g., relation extraction) of arguments and those that focus on the structure of arguments (e.g., automated essay scoring). We propose an automated approach to detecting high-level organizational elements in argumentative dis- course that combines a rule-based system and a probabilistic sequence model in a principled manner. We present quantitative results on a dataset of human-annotated persuasive essays, and qualitative analyses of performance on es- says and on political debates. 1
Keywords
Full Content
2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 20–28, Montr´eal, Canada, June 3-8, 2012. c⃝2012 Association for Computational Linguistics Identifying High-Level Organizational Elements in Argumentative Discourse Nitin Madnani Michael Heilman Educational Testing ServiceJoel Tetreault Princeton, NJ, USA {nmadnani,mheilman,jtetreault}@ets.org Martin Chodorow Hunter College of CUNY New York, NY, USA martin.chodorow@hunter.cuny.edu Abstract Argumentative discourse contains not only language expressing claims and evidence, but also language used to organize these claims and pieces of evidence. Differentiating be- tween the two may be useful for many appli- cations, such as those that focus on the content (e.g., relation extraction) of arguments and those that focus on the structure of arguments (e.g., automated essay scoring). We propose an automated approach to detecting high-level organizational elements in argumentative dis- course that combines a rule-based system and a probabilistic sequence model in a principled manner. We present quantitative results on a dataset of human-annotated persuasive essays, and qualitative analyses of performance on es- says and on political debates. 1 Introduction When presenting an argument, a writer or speaker usually cannot simply state a list of claims and pieces of evidence. Instead, the arguer must explic- itly structure those claims and pieces of evidence, as well as explain how they relate to an opponent’s ar- gument. Consider example 1 below, adapted from an essay rebutting an opponent’s argument that griz- zly bears lived in a specific region of Canada. The argument states that based on the result of the recent research, there proba- bly were grizzly bears in Labrador. It may seem reasonable at first glance, but ac- tually, there are some logical mistakes in it. . . . There is a possibility that they were a third kind of bear apart from black and grizzly bears. Also, the explorer ac- counts were recorded in the nineteenth century, which was more than 100 years ago. . . . In sum, the conclusion of this argument is not reasonable since the ac- count and the research are not convinc- ing enough. . . . The argument begins by explicitly restating the opponent’s claim, prefacing the claim with the phrase “The argument states that.” Then, the sec- ond sentence explicitly marks the opponent’s argu- ment as flawed. Later on, the phrase “There is a possibility that” indicates the subsequent clause in- troduces evidence contrary to the opponent’s claim. Finally, the sentence “In sum, . . .” sums up the ar- guer’s stance in relation to the opponent’s claim.1 As illustrated in the above example, argumenta- tive discourse can be viewed as consisting of lan- guage used to express claims and evidence, and language used to organize them. We believe that differentiating organizational elements from content would be useful for analyzing persuasive discourse. 1The word Also signals that additional evidence is about to be presented and should also be marked as shell. However, it was not marked in this specific case by our human annotator (§3.2). 20 We refer to such organizational …
Content truncated to 500 words
Labeled Text
Argumentative discourse contains not only language expressing claims and evidence, but also language used to organize these claims and pieces of evidence. Differentiating between the two may be useful for many applications, such as those that focus on the content (e.g., relation extraction) of arguments and those that focus on the structure of arguments (e.g., automated essay scoring). We propose an automated approach to detecting high-level organizational elements in argumentative discourse that combines a rule-based system and a probabilistic sequence model in a principled manner. We present quantitative results on a dataset of human-annotated persuasive essays, and qualitative analyses of performance on essays and on political debates. 1
Publication Details
- Paper ID:
- 23966
- Year:
- 2012
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan