Argumentation Synthesis following Rhetorical Strategies

Proceedings of the 27th International Conference on Computational Linguistics, pages 3753–3765 Santa Fe, New Mexico, USA, August 20-26, 2018. 3753 Argumentation Synthesis following Rhetorical Strategies Henning Wachsmuth ∗ Manfred Stede ∗∗ Roxanne El Baff † Khalid Al-Khatib † Maria Skeppstedt ‡ ∗∗ Benno Stein † ∗ Paderborn University, Paderborn, Germany, henningw@upb.de ∗∗ University of Potsdam, Potsdam, Germany, stede@uni-potsdam.de † Bauhaus-Universität Weimar, Weimar, Germany, <first>{.<last>}+@uni-weimar.de ‡ Linnaeus University, Växjö, Sweden, maria.skeppstedt@lnu.se Abstract Persuasion is rarely achieved through a loose set of arguments alone. Rather, an effective delivery of arguments follows a rhetorical strategy, combining logical reasoning with appeals to ethics and emotion. We argue that such a strategy means to select, arrange, and phrase a set of argumentative discourse units. In this paper, we model rhetorical strategies for the computational synthesis of effective argumentation. In a study, we let 26 experts synthesize argumentative texts with different strategies for 10 topics. We find that the experts agree in the selection significantly more when following the same strategy. While the texts notably vary for different strategies, especially their arrangement remains stable. The results suggest that our model enables a strategical synthesis. 1 Introduction The primary use of arguments in natural language is to persuade others of a stance towards a controversial topic (Mercier and Sperber, 2011). According to Johnson and Blair (2006), an argument is logically cogent if its premises are relevant as support for its conclusion, individually acceptable, and together sufficient to draw the conclusion. In real life, however, argumentation is by far not only about logic (Allwood, 2016). Without a rhetorical strategy, arguments will hardly ever unfold their persuasive effectiveness. Following Aristotle (2007), we see a rhetorical strategy as the purposeful encoding of three means of persuasion in a well-arranged and well-phrased speech or text: logos (providing logically reasoned arguments), ethos (demonstrating good character and credibility), and pathos (evoking the right emotions). Listeners or readers then decode the encoding, forming their view of the author’s logos, ethos, and pathos. In the realm of the area of computational argumentation, rhetorical strategies are particularly relevant for technologies that synthesize argumentative text and that aim to deliver arguments effectively. Existing argument mining research largely focuses on the logical structure of arguments, identifying their units (premises vs. conclusions) and relations (support vs. attack). Recently, a few studies have tackled strategy-related aspects, such as explicit expressions of ethos (Duthie et al., 2016) and the effects of logical and emotional arguments across audiences (Lukin et al., 2017). So far, however, strategies have not been considered in argumentation synthesis, which altogether has not received much attention (see Section 2). In this paper, we study the role of rhetorical strategies when synthesizing argumentation. In particular, we consider monological argumentative texts where an author seeks to persuade target readers of his or her stance towards a given topic, such as news editorials and persuasive essays. Conceptually, we argue that an author synthesizes a text of such genres in three subsequent steps: 1. Selecting content in terms of argumentative discourse units …

Content truncated to 500 words