13. – 18. Dezember 2015, Dagstuhl-Seminar 15512

Debating Technologies


Iryna Gurevych (TU Darmstadt, DE)
Eduard H. Hovy (Carnegie Mellon University – Pittsburgh, US)
Noam Slonim (IBM – Haifa, IL)
Benno Stein (Bauhaus-Universität Weimar, DE)

Auskunft zu diesem Dagstuhl-Seminar erteilt

Dagstuhl Service Team


Dagstuhl Report, Volume 5, Issue 12 Dagstuhl Report
Gemeinsame Dokumente
Dagstuhl's Impact: Dokumente verfügbar
Dagstuhl-Seminar Wiki
Programm des Dagstuhl-Seminars [pdf]

(Zum Einloggen bitte Seminarnummer und Zugangscode verwenden)


Why do people in all societies argue, discuss, and debate? Apparently, we do so not only to convince others of our own opinions, but because we want to explore the differences between our own understanding and the conceptualizations of others, and learn from them. Being one of the primary intellectual activities of the human mind, debating naturally involves a wide range of conceptual capabilities and activities, ones that have only in part been studied from a computational perspective in fields like computational linguistics and natural language processing. As a result, computational technologies supporting human debating are scarce, and typically still in their infancy. Recent decades, however, have seen the emergence and flourishing of many related and requisite computational tasks, including sentiment analysis, opinion and argumentation mining, natural language generation, text summarization, dialogue systems, recommendation systems, question answering, emotion recognition/generation, automated reasoning, and expressive text to speech.

This Dagstuhl seminar was the first of its kind. It laid the groundwork for a new interdisciplinary research community centered around debating technologies - computational technologies developed directly to enhance, support, and engage with human debating. The seminar brought together leading researchers from relevant communities to discuss the future of debating technologies in a holistic manner.

The seminar was held between 13 and 18 December 2015, with 31 participants from 22 different institutions. The event’s sixteen sessions included 34 talks, thirteen themed discussions, three system demonstrations, and a hands-on "unshared" task. Besides the plenary presentations and discussions, the program included several break-out sessions and mock debates with smaller working groups. The presentations addressed a variety of topics, from high-level overviews of rhetoric, argument structure, and argument mining to low-level treatments of specific issues in textual entailment, argumentation analysis, and debating-oriented information retrieval. Collective discussions were arranged for most of these topics, as well as on more forward-thinking themes, such as the potential and limitations of debating technologies, identification of further relevant research communities, and plans for a future interdisciplinary research agenda.

A significant result of the seminar was the decision to use the term computational argumentation to put the community’s various perspectives (argument mining, argument generation, debating technologies, etc.) under the same umbrella. By analogy with "computational linguistics", "computational argumentation" denotes the application of computational methods for analyzing and synthesizing argumentation and human debate. We identified a number of key research questions in computational argumentation, namely:

  • How important are semantics and reasoning for real-world argumentation?
  • To what extent should computational argumentation concern itself with the three classical rhetorical appeals of ethos (appeal to authority), pathos (appeal to emotion), and logos (appeal to reason)? Is it sufficient to deal with logos, or is there some benefit in studying or modelling ethos and pathos as well?
  • What are the best ways of dealing with implicit knowledge?

A number of discussion questions at the seminar followed from these points, particularly in relation to the data and knowledge sources required for implementing and evaluating computational argumentation systems. For example, are currently available datasets sufficient for large-scale processing or for cross-language and cross-domain adaptation? Can we reliably annotate logos, ethos, and pathos? In any case, what sort of data would be considered "good" for a shared task in computational argumentation? Is it possible for computational argumentation to repeat the recent successes of "deep" natural language processing by employing shallow methods on large masses of data? How does cultural background impact human argumentation, and is this something that computational models need to account for? Finding the answers to these and other questions is now on the agenda for our burgeoning research community.

Summary text license
  Creative Commons BY 3.0 Unported license
  Iryna Gurevych, Eduard H. Hovy, Noam Slonim, and Benno Stein


  • Artificial Intelligence / Robotics
  • Data Bases / Information Retrieval
  • Society / Human-computer Interaction


  • Debating
  • Human-machine interaction
  • Interactive systems
  • Discourse and dialogue


In der Reihe Dagstuhl Reports werden alle Dagstuhl-Seminare und Dagstuhl-Perspektiven-Workshops dokumentiert. Die Organisatoren stellen zusammen mit dem Collector des Seminars einen Bericht zusammen, der die Beiträge der Autoren zusammenfasst und um eine Zusammenfassung ergänzt.


Download Übersichtsflyer (PDF).

Dagstuhl's Impact

Bitte informieren Sie uns, wenn eine Veröffentlichung ausgehend von Ihrem Seminar entsteht. Derartige Veröffentlichungen werden von uns in der Rubrik Dagstuhl's Impact separat aufgelistet  und im Erdgeschoss der Bibliothek präsentiert.


Es besteht weiterhin die Möglichkeit, eine umfassende Kollektion begutachteter Arbeiten in der Reihe Dagstuhl Follow-Ups zu publizieren.