Argument Mining

Mathilde Janier has co-authored over 15 publications, principally dealing with the annotation and modeling of argumentative dynamics in debate and dispute mediation. Her work mainly focuses on argumentation in dialogical contexts. Patrick Saint-Dizier is Senior Researcher at CNRS – IRIT Toulouse, France. His work is based on logic, language, argumentation, natural language processing and logic programming. He is the author and co-author of 11 books on these topics.

• Preface xiChapter 1. Introduction and Challenges 11.1. What is argumentation? 11.2. Argumentation and argument mining 41.3. The origins of argumentation 71.4. The argumentative discourse 81.5. Contemporary trends 10Chapter 2. The Structure of Argumentation 132.1. The argument–conclusion pair 132.2. The elementary argumentative schema 142.2.1. Toulmin’s argumentative model 142.2.2. Some elaborations and refinements of Toulmin’s model 172.2.3. The geometry of arguments 182.3. Modeling agreement and disagreement 202.3.1. Agreeing versus disagreeing 202.3.2. The art of resolving divergences 232.4. The structure of an argumentation: argumentation graphs 252.5. The role of argument schemes in argumentation 272.5.1. Argument schemes: main concepts 272.5.2. A few simple illustrations 282.5.3. Argument schemes based on analogy 292.5.4. Argument schemes based on causality 302.6. Relations between Toulmin’s model and argumentation schemes 312.6.1. Warrants as a popular opinion 322.6.2. Argument schemes based on rules, explanations or hypothesis 342.6.3. Argument schemes based on multiple supports or attacks 352.6.4. Causality and warrants 37Chapter 3. The Linguistics of Argumentation 393.1. The structure of claims 403.2. The linguistics of justifications 453.3. Evaluating the strength of claims, justifications and arguments 473.3.1. Strength factors within a proposition 493.3.2. Structuring expressions of strength by semantic category 513.3.3. A simple representation of strength when combining several factors 523.3.4. Pragmatic factors of strength expression 533.4. Rhetoric and argumentation 593.4.1. Rhetoric and communication 603.4.2. Logos: the art of reasoning and of constructing demonstrations 613.4.3. Ethos: the orator profile 623.4.4. Pathos: how to persuade an audience 63Chapter 4. Advanced Features of Argumentation for Argument Mining 654.1. Managing incoherent claims and justifications 654.1.1. The case of justifications supporting opposite claims 664.1.2. The case of opposite justifications justifying the same claim 674.2. Relating claims and justifications: the need for knowledge and reasoning 674.2.1. Investigating relatedness via corpus analysis 684.2.2. A corpus analysis of the knowledge involved 694.2.3. Observation synthesis 724.3. Argument synthesis in natural language 744.3.1. Features of a synthesis 754.3.2. Structure of an argumentation synthesis 76Chapter 5. From Argumentation to Argument Mining 795.1. Some facets of argument mining 795.2. Designing annotation guidelines: some methodological elements 815.3. What results can be expected from an argument mining system? 825.4. Architecture of an argument mining system 835.5. The next chapters 84Chapter 6. Annotation Frameworks and Principles of Argument Analysis 856.1. Principles of argument analysis 866.1.1. Argumentative discourse units 866.1.2. Conclusions and premises 886.1.3. Warrants and backings 896.1.4. Qualifiers 896.1.5. Argument schemes 906.1.6. Attack relations: rebuttals, refutations, undercutters 906.1.7. Illocutionary forces, speech acts 926.1.8. Argument relations 936.1.9. Implicit argument components and tailored annotation frameworks 956.2. Examples of argument analysis frameworks 976.2.1. Rhetorical Structure Theory 976.2.2. Toulmin’s model 986.2.3. Inference Anchoring Theory 996.2.4. Summary 1026.3. Guidelines for argument analysis 1036.3.1. Principles of annotation guidelines 1036.3.2. Inter-annotator agreements 1046.3.3. Interpretation of IAA measures 1056.3.4. Some examples of IAAs 1066.3.5. Summary 1076.4. Annotation tools 1086.4.1. Brat 1086.4.2. RST tool 1096.4.3. AGORA-net 1106.4.4. Araucaria 1106.4.5. Rationale 1116.4.6. OVA+ 1126.4.7. Summary 1136.5. Argument corpora 1146.5.1. COMARG 1156.5.2. A news editorial corpus 1156.5.3. THF Airport ArgMining corpus 1156.5.4. A Wikipedia articles corpus 1156.5.5. AraucariaDB 1156.5.6. An annotated essays corpus 1166.5.7. A written dialogs corpus 1166.5.8. A web discourse corpus 1166.5.9. Argument Interchange Format Database 1166.5.10. Summary 1176.6. Conclusion 118Chapter 7. Argument Mining Applications and Systems 1197.1. Application domains for argument mining 1197.1.1. Opinion analysis augmented by argument mining 1207.1.2. Summarization 1207.1.3. Essays 1207.1.4. Dialogues 1207.1.5. Scientific and news articles 1207.1.6. The web 1217.1.7. Legal field 1217.1.8. Medical field 1217.1.9. Education 1217.2. Principles of argument mining systems 1227.2.1. Argumentative discourse units detection 1237.2.2. Units labeling 1237.2.3. Argument structure detection 1247.2.4. Argument completion 1257.2.5. Argument structure representation 1257.3. Some existing systems for argument mining 1267.3.1. Automatic detection of rhetorical relations 1267.3.2. Argument zoning 1267.3.3. Stance detection 1277.3.4. Argument mining for persuasive essays 1277.3.5. Argument mining for web discourse 1277.3.6. Argument mining for social media 1287.3.7. Argument scheme classification and enthymemes reconstruction 1287.3.8. Argument classes and argument strength classification 1287.3.9. Textcoop 1297.3.10. IBM debating technologies 1297.3.11. Argument mining for legal texts 1297.4. Efficiency and limitations of existing argument mining systems 1307.5. Conclusion 131Chapter 8. A Computational Model and a Simple Grammar-Based Implementation 1338.1. Identification of argumentative units 1348.1.1. Challenges raised by the identification of argumentative units 1348.1.2. Some linguistic techniques to identify ADUs 1358.2. Mining for claims 1398.2.1. The grammar formalisms 1408.2.2. Lexical issues 1428.2.3. Grammatical issues 1458.2.4. Templates for claim analysis 1488.3. Mining for supports and attacks 1508.3.1. Structures introduced by connectors 1508.3.2. Structures introduced by propositional attitudes 1518.3.3. Other linguistic forms to express supports or attacks 1528.4. Evaluating strength 1538.5. Epilogue 154Chapter 9. Non-Verbal Dimensions of Argumentation: a Challenge for Argument Mining 1559.1. The text and its additions 1569.1.1. Text, pictures and icons 1569.1.2. Transcriptions of oral debates 1569.2. Argumentation and visual aspects 1579.3. Argumentation and sound aspects 1589.3.1. Music and rationality 1599.3.2. Main features of musical structure: musical knowledge representation 1609.4. Impact of non-verbal aspects on argument strength and on argument schemes 1619.5. Ethical aspects 162Bibliography 163Index 175